Lighting Control-Technology and Applications
This Page Intentionally Left Blank
Lighting
Control
–
Technology
Applications
Robert S. Simpson
and
CONTENTS
Focal Press An imprint of Elsevier Linacre House, Jordan Hill, Oxford OX2 8DP 200 Wheeler Road, Burlington MA 01803 First published 2003 Copyright© 2003, Robert S. Simpson. All rights reserved The right of Robert S. Simpson to be identified as the author of this work has been asserted in accordance with the Copyright, Designs and Patents Act 1988 No part of this publication may be reproduced in any material form (including photocopying or storing in any medium by electronic means and whether or not transiently or incidentally to some other use of this publication) without the written permission of the copyright holder except in accordance with the provisions of the Copyright, Designs and Patents Act 1988 or under the terms of a licence issued by the Copyright Licensing Agency Ltd, 90 Tottenham Court Road, London, England W1T 4LP. Applications for the copyright holder’s written permission to reproduce any part of this publication should be addressed to the publisher
Permissions may be sought directly from Elsevier’s Science and Technology Rights Department in Oxford, UK: phone: (+44) (0) 1865 843830; fax: (+44) (0) 1865 853333; e-mail: [email protected] You may also complete your request on-line via the Elsevier Science homepage (www.elsevier.com), by selecting ‘Customer Support’ and then ‘Obtaining Permissions’ British Library Cataloguing in Publication Data Simpson, Robert S. Lighting control : technology and applications 1.Electric lighting – Control I.Title 621.3’2 Library of Congress Cataloguing in Publication Data A catalogue record for this book is available from the Library of Congress ISBN 0 240 51566 8
For information on all Focal Press publications visit our website at: www.focalpress.com
Printed and bound in Italy
iv
CONTENTS
Contents
Trade marks and disclaimer Acknowledgements Preface
vii viii ix
Part 1 Foundation chapters 1
Electricity and light
1
1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 1.10 1.11
Electricity Electrical units and components Electrical distribution Power factor Control of electric power Electromagnetic compatibility (EMC) Light The eye, how we see light Measurement of light Color Measurement of color
1 4 15 24 29 36 39 44 45 50 53
2
Lighting electronics
58
2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8
Electronic principles The diode The transistor The thyristor, triac and GTO Analog and digital The integrated circuit and ASIC The microprocessor Programmable devices
58 63 68 81 85 93 102 110
Part 2 Lamps 3
Everyday lamps
112
3.1 3.2 3.3 3.4 3.5 3.6
Non-electric lighting The incandescent lamp Tungsten halogen lamps The fluorescent lamp Compact fluorescent lamps Special purpose fluorescent lamps
112 113 115 119 123 124
4
Arc lamps
130
4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8
High intensity discharge lamps Mercury vapor lamps Sodium and high pressure sodium lamps Metal halide lamps Compact source metal halide lamps High pressure mercury vapor lamps Xenon arc lamps Arc lamp classification
130 130 133 136 138 139 140 141
5
Special purpose lamps
144
5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8 5.9 5.10 5.11
Induction lamps Flat lamps Neon lamps Electroluminescent lamps Light emitting diodes (LEDs) Lasers Ultra-violet lamps Infra-red lamps Flash tubes Fiber optics and lightguides Video displays as lightsources
144 148 149 151 153 159 162 163 164 166 171
Part 3 Lighting components 6
Electromagnetic components
172
6.1 6.2 6.3 6.4 6.5 6.6
172 186 192 204 207
6.7
Principles of transformers and inductors Transformers for lighting Ballasts for fluorescent lamps Ballasts for HID and arc lamps Ignitors and starters Lighting control by transformers and ballasts Power factor correction
7
Electronic components
215
7.1 7.2
Circuit elements Electronic ballasts for fluorescent lamps
215 219
210 211
v
CONTENTS
7.3 7.4 7.5
Electronic ballasts for HID and arc lamps Electronic transformers Lighting control by electronic transformers and ballasts
242 245 248
Part 4 Dimmers and control systems
12
Architectural lighting control systems 374
12.1 12.2 12.3 12.4 12.5 12.6 12.7 12.8
“Setting the scene” Manual versus automatic control Single channel control Small multi-channel control systems Large multi-channel control systems Switching systems Centralized versus distributed systems Emergency and safety
374 375 376 377 381 383 390 391
13
The merging of “architectural” and “entertainment” lighting control
396
8
Dimmers
250
8.1 8.2 8.3 8.4 8.5 8.6
Introduction to dimmers Non-electronic dimming Thyristor and triac dimmers Transistor dimmers Electromagnetic compatibility (EMC) New developments in electronic dimming
250 252 257 269 272 276
9
Control signals and protocols
280
13.2 13.3
9.1 9.2 9.3 9.4 9.5 9.6 9.7
Introduction Analog control Digital control Standard protocols for lighting control Networks and buses Computers in lighting control Cordless control
280 280 283 288 303 332 334
10
Why lighting control?
343
10.1 10.2 10.3 10.4 10.5
The practical role The esthetic role The energy management role Influence of legislation Lighting design
343 344 345 348 351
11
Stage and entertainment lighting control systems 353
11.1 11.2
Basis of stage lighting control Simple multichannel controls for entertainment Memory consoles Live versus automatic Control of moving lights Control of color Large scale entertainment lighting control
11.3 11.4 11.5 11.6 11.7
vi
353 354 356 362 363 370 372
13.1
User demands and the influence of designers Automatic lighting control in public shows and public areas Control of exterior lighting
396 398 400
14
Energy management and building control systems 402
14.1 14.2 14.3
Principles Sensors and timers Switching versus dimming, control algorithms Local versus central control Impact of lighting on HVAC Power quality Integrated versus separate lighting control Monitoring systems
14.4 14.5 14.6 14.7 14.8
402 403 412 417 417 417 418 418
Part 5 Applications 15
Architectural applications
420
15.1 15.2 15.3 15.4
The home Integrated home control systems The workplace Meeting rooms, conference centers, and auditoria Places of worship Museums, art galleries and libraries Visitor centers and exhibitions
420 428 431
15.5 15.6 15.7
436 447 449 454
CONTENTS
15.8 Hotels, hospitals and institutions 15.9 Restaurants, bars and pubs 15.10 Illuminated signs
459 463 464
16
Functional applications
469
16.1 16.2 16.3 16.4 16.5 16.6 16.7 16.8 16.9 16.10
Retail spaces Agriculture and horticulture Manufacturing processes Healthcare Simulation On water In the air On the road On railways Control rooms
469 472 473 474 475 476 481 490 505 508
17
Entertainment applications
17.1 17.2 17.3 17.4 17.5
510
Small stages Large stages Television Touring shows Outdoor shows, Son et Lumière, pyrotechnics 17.6 Stadia, arenas, sporting facilities 17.7 Theme parks 17.8 Entertainment within retail 17.9 Discotheques, dancefloors and clubs 17.10 Conclusion
510 514 520 524
Some suggestions for further reading Table of acronyms Index
547 550 552
526 534 539 542 544 546
Trade marks and disclaimer In order to ensure that this book can be of practical use, it has been necessary to mention many commercial products by name. The inclusion or omission of any particular company’s products does not imply any endorsement or comment by the publisher. Where summaries are given of manufacturers of different kinds of equipment, these are intended as examples only, no claim is made that such summaries represent any kind of comprehensive directory. This book is intended as a source of information only. Care has been taken to verify the accuracy of all information contained herein, but neither the author nor the publisher can take responsibility for the consequences of using the information, or for any errors or omissions. Example circuits, devices and techniques may be the subject of patent protection or patent application. Their publication in this book does not imply any license for their use. Any references to standards and protocols (whether public or proprietary) are intended to give readers an introduction to their nature and operation. There is no implication of any license to use them, and current standards specifications and details of any licensing associated with them must be obtained from the sponsoring body concerned. All trade marks are acknowledged. Where known they are identified in the text by TM or ®.
vii
CONTENTS
Acknowledgements
The author acknowledges the help given to him in preparing this book by many industry colleagues from around the world. As far as possible application illustrations and case histories are acknowledged within the text. Unacknowledged diagrams and photographs are mostly from Helvar and Electrosonic. Among the many individuals who have contributed information and help, special thanks are due to to the following, arranged in company alphabetical order: Individual independent consultants
Brian Legge Thomas Baenziger
ADB (aviation lighting) Ira Jackson AIM Aviation (Jecco) Ltd Rolf Startin Anytronics Ltd Bob Hall Artistic License Wayne Howell Art2Architecture Peter Fink Arup Acoustics Sam Wise Arup (Manchester office) John Waite Avolites Richard Salzedo British Library Michael Wildsmith Building Research Establishment Michael Perry Carr & Angier Paul Covell CCT Lighting David Manners Claude Lyons Jim McIlfatrick Color Kinetics Melissa Connor The Deep (Hull, UK) Dr David Gibson Delmatic Stephen Woodnutt Derungs Licht Claudio Roth DHA Lighting Design Adam Grater Dynalite Dannielle Furness ECS Philips Lighting Controls Chris Holder Electrosonic Ltd Yvonne Hegarty, Daniela Simonides ETC Fred Foster Firework Shop John Stapleton Fisher Marantz Stone Scott Hershman Rob Schoenbohm Focus Lighting Inc Paul Gregory Genlyte Controls Jason Moreno Helvar (Finland) Teijo Viljanen, Markku Nohiu Eeva Harjula Helvar (Germany) Ingo Sommer Helvar (UK) Alan Jackson, Trevor Forrest Austen Conway, Dr Scott Wade, Mel Collins Howard Industries Mike Dodds
viii
IBL Leviton Lighting Architects Group
Peter Saunders Breda Potter Jonathan Speirs Mark Major, Iain Ruxton LSI Projects Ltd Nick Mobsby Lumisphere Products Bob Myson Lutron Brian McKiernan MEM and MEM250 Richard Hunt, H. Milligan Osram Hans Jörg Schenkat, Verena Roemer P. Ducker Systems Ltd Richard Thomas Philips Marc Segers, Holger Moench Peter Woodward, John Rothery Pinniger & Partners Miles Pinniger Project Interational Richard Dixon Pulsar Light Cambridge Andy Graves, Paul Marden Pyrodigital Consultants Ken Nixon Quo Vadis Ltd Michael Stott Relco Daria Fossati repas AEG Dirk Buchholz Royal National Theatre Great Britain MikeAtkinson Schott Fibre Optics UK John Meadows Starfield Controls Wayne Morrow Strand Lighting Vic Gibbs, Ivan Myles Sutton Vane Associates Mark Sutton Vane Technical Marketing Ltd Andy Collier Teknoware Jari Tabell Thorlux Lighting Terry Fletcher Tunewell Transformers Derek Price, Glen McGovern Tridonic Stewart Langdown Kate Wilkins Lighting Design Kate Wilkins Vantage Controls Andy North Vari-Lite Europe Ltd Samantha Dean,Colin Brooker Waltzing Waters Douglas Tews, Michael Przystawik The Watt Stopper Inc Joy Cohen WRTL Exterior Lighting Tom Thurrell Wynne Willson Gottelier Tony Gottelier Wybron Inc Brandon James Young Electric Sign Co Graham Beland, Blake Gover In addition, grateful thanks to Paul Ashford for driving “Pagemaker”, Noel Packer (of Helvar UK) and Paul Ashford for drawing most of the diagrams and Maggie Thomas (of Electrosonic) for help thoughout the project. Finally, thanks to the staff at Focal Press for their support over three years.
CONTENTS
Preface
Artificial lighting is part of our daily lives; in the modern world there are few activities which take place without it. While there are many books on the subjects of light and lighting, there are few that cover the subject of lighting control. Those that do look at the subject from a limited perspective, for example that of stage lighting. Within the practical limits set by its size, this book is intended as a review of all methods of lighting control. It covers all the current technologies, and gives application examples from many aspects of our daily lives. It is intended for all those who already work in the lighting industry, for designers and consultants, and for the sophisticated end user. It is also intended as a training resource for those new to the industry. No significant prior technical knowledge is assumed. The book is written for the “intelligent layman”, and mathematics are kept to a minimum. In order to make it a complete resource Part 1 is in the nature of a “foundation course” to give the necessary background to those with limited (or no) knowledge of the basics of light, electricity and electronics. It can be skipped by engineers and technicians who already have this knowledge. Part 2 is a review of light sources. Unless you understand how a particular light source works, you cannot appreciate how to control it. Again much of the information in Part 2 may be well known to some readers, although there could be a few surprises arising from recent developments in traditional sources, and from the arrival of completely new sources.
Most light sources need some kind of “load interface”, for example a ballast or transformer, between them and the electricity supply. Part 3 reviews these “lighting components”, most of which are available in both electromagnetic and electronic form. In recent years there have been major advances in electronic lighting components, and they now form an essential part of many lighting control systems. Part 4 is the technical heart of the book, covering dimmers and control systems. The dimmer or dimming interface is now only part of lighting control. With modern systems often embracing thousands of lighting “channels” spread over a large building complex, a basic knowledge of network and computer technology becomes necessary. Lighting control is used for practical, esthetic and energy management reasons. Part 5 reviews how the technologies described in the first four parts of the book are applied in practice. The aim here has been to cite examples, without attempting to be exhaustive, representing best practice from different parts of the world. Lighting control has been part of my life for over 50 years, from operating slider dimmers for school plays, to being a member of a specifying team for the latest generation of digital dimmers. I hope that my enthusiasm for the subject is reflected in this book, and that even if you are an experienced lighting practioner, you will find something useful or unexpected within it. R.S.S.
ix
This Page Intentionally Left Blank
Chapter 1
ELECTRICITY AND LIGHT
Electricity and light Part 1 – Foundation chapters The first two chapters are intended for those who have little or no knowledge of light, electricity and electronics. They include definitions and background technical information which is assumed in the succeeding chapters. Restrictions on space, and the use of only the simplest mathematics, mean that the introduction is cursory, but nonetheless it should help those new to the subject (and those of us who have forgotten our school physics and simple electronics, and just want a reminder!). Readers with a technical qualification in electronics or electrical engineering can safely omit the sections on electricity and electronics, but may still find the sections on light and color useful.
1.1 Electricity Both electricity and light are mysterious. We are well aware of their effect, but, if asked the question “what is electricity?” or “what is light?”, we have difficulty giving a convincing answer. We find it easiest to think in terms of a model, or models, which make it possible for us to understand how an electrical or optical device works. Provided the model, or “mind picture”, is consistent, this is a quite acceptable way to proceed – indeed it is the only way for the layman to get an insight into physical processes. In this chapter we describe the simplest models of electricity and light to the point where the methods of controlling them can be understood. The first model we need is that of the structure of matter. The atomic theory holds that the smallest unit of matter is the atom. In a compound there are several different kinds of atom, but in an element all the atoms are the same. All atoms are themselves made up of fundamental particles. Modern physics has produced an alarmingly long list of these, but for our purposes it is sufficient to know of only three: • the proton, a massive particle which carries a positive electric charge • the neutron which has the same mass as the proton, but has no charge • the electron which carries a negative electric charge equal and opposite to that of the proton.
The model of atomic structure which is sufficient for our purposes imagines that atoms have a planetary structure: a heavy nucleus consisting of a mixture of protons and neutrons, surrounded by orbiting electrons. The resulting atom is electrically neutral since the number of electrons in orbit exactly matches the number of protons. Figure 1.1 gives some examples of elements which are relevant to our subject. An important point to notice is that the electron orbits are not haphazard. The model shows that the electrons occupy “shells”, each one of which can only contain a defined number of electrons. As we move through the elements, from hydrogen the lightest, through to the heavy elements such as uranium, one unit of charge is added for each element. The shells get filled up in sequence, and this has the result that only some elements have a “complete” outer shell. It is the nature of the outer shell of electrons which determines many of the electrical and chemical properties of different elements. When an atom, for example a copper atom, has only a single electron in its outer shell, this electron can be easily dislodged – a so-called “free” electron. Such a free electron can be influenced by an electric charge. Electric charges have as their principal characteristic the fact that like charges repel each other, and unlike charges attract each other. An atom which has “lost” its free electron will itself aquire a
1
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 1.1 Simplified models of atomic structure. (a) Hydrogen, with one proton and one electron. (b) Carbon, showing two groups of electrons in elliptical orbits. (c) Simplified version of Carbon, showing positive charge of the nucleus balancing the negative charges of the electrons, and the K and L “shells”. (d) Neon, an insulator, with two full K and L “shells” of 2 and 8 electrons. (e) Silicon, a semiconductor, with a partially filled outer M shell having only four out of a possible maximum 18 electrons. (f) Copper, a conductor, with the three inner shells filled, but with the outer N shell having only one electron, out of a possible 32 maximum for this shell.
positive charge and is said to be positively ionized; it will then tend to attract a free electron to balance the charge again. However, free electrons are quite mobile in metals, and if an electric charge is applied across a length of metal, it results in a movement of free electrons; which we now identify as an electric current. Figure 1.2 shows the idea of an electron flow within a metal wire or conductor. The measure of the rate of flow of electric current is the Ampere, often abbreviated to Amp. For our mind picture of electricity it can be thought of as a flow of 6.26 × 10 18 (that is 6.26 multiplied by 1,000,000,000,000,000,000) electrons per second, although for reasons explained later this is not the actual definition of an Ampere.
2
The measure of electrical charge is the Coulomb. It is defined as the charge transferred when a current of one ampere flows for one second. Since the free electrons creating an electric current are themselves charged particles, the coulomb is equal to the total charge carried by 6.26 × 1018 electrons. Next for our model of electricity flow we need to have an idea of what can “push” the free electrons along the conductor. Figure 1.3 shows a simple electrical circuit where a battery is the source of electro-motive force or e.m.f. measured in Volts. When the first electrical discoveries were made, it was a convention that electric current flowed from the positive terminal of a battery to the negative terminal. This convention remains today, even though our “electron flow” model of electricity
ELECTRICITY AND LIGHT
Wire Conductor
POSITIVE CHARGE THIS END attracts electrons
-
+ Electron flow Conventional current flow
= Nucleus with full inner shells and free electron in outer shell
Figure 1.2 The idea of free electron flow creating an electric current. At the positive end of the conductor there is a deficiency of electrons, so the free electrons move to fill the space. Conversely at the negative end, there is a surplus of electrons “pushing”.
shows the free electrons flowing the other way. The more volts applied, the stronger the current. It is also
found that the conductor heats up while the current flows, and that this heating effect is proportional to the square of the current. The tendency for an electrical conductor to restrict current flow is called its resistance and is measured in Ohms. Table 1.4 shows how the Ohm is defined by the amount of heat produced by a current of one Ampere, and how in turn the Volt is defined as the Potential Difference across a resistance of one Ohm when it is carrying a current of one Ampere. The most useful and most easily remembered relationships which link the main electrical units together are: V = IR Ohm’s Law (Volts = Current in Amps × Resistance in Ohms) which can also be expressed: I=
V R
and
R=
V I
and the power dissipated in a resistance: W = I 2R
Figure 1.3 The concept of potential difference down the length of a uniform conductor. At any point in the wire, the potential difference is proportional to the length of wire. On the right a hydraulic analogy shows that water pressure drops in a pipe in a similar way. If the tap is closed, then the level in all the columns would rise to match the level in the tank.
3
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
(Power in Watts = Current in Amps squared Resistance in Ohms)
×
which Ohm’s Law allows us to more conveniently remember as: W = VA
(Power in Watts = Potential Difference in Volts × Current in Amps) Figure 1.3 shows the battery as a source of e.m.f. (measured in Volts) providing a potential difference (also measured in Volts) across the electrical conductor which we can think of as being some kind of heating element. Down the length of the conductor the potential drops in proportion to the length – if you take a Voltmeter to measure the voltage it gets lower along the length of the conductor. A simple hydraulic analogy is shown of water from a header tank losing pressure down the length of a pipe. In the real world the e.m.f of the battery measured on open circuit (nothing connected) only equals the p.d. across the load resistance when connected if the
battery itself has no internal resistance. If it does have internal resistance, then the actual voltage across the load will be lower. An automobile battery is an example of a battery with very low internal resistance; it can deliver a current of several hundred Amperes at 12 Volts to run a starter motor. A battery set for a portable compact disc player may also deliver 12V as an open circuit e.m.f., but its internal resistance limits the current available (and the actual p.d. at the load). See Figure 1.4.
1.2 Electrical units and components 1.2.1 Units Having introduced the concept of the flow of electricity, it is now necessary to be more rigorous about how it is measured. The way in which we measure things depends finally on having agreed reference standards, so that
Figure 1.4 A car battery has negligible internal resistance, so can deliver a full 12V to a lamp load. The same lamp load connected to a small 12V battery might show only half the current flowing through it because of the battery’s internal resistance. (The value shown here has been chosen for simplicity) Thus the p.d. across the load is much less then the e.m.f. of the open circuit battery. The idea that a source of power can itself limit the available current is important.
4
ELECTRICITY AND LIGHT
Quantity Length
Unit Meter
Symbol m
Mass Time
Kilogram Second
kg s
Plane angle
Radian
rad
Solid angle
Steradian
sr
Temperature
Degree Kelvin
K
Defined as The length of the path travelled by light in a vacuum during a time interval of 1/299,792,458 of a second. The mass of the prototype kilogram kept at Sèvres in France 9,192,631,770 periods of vibration of the Caesium 133 atom (formerly 1/86,400 of the mean solar day). The angle subtended at the center of a circle by an arc of equal length to the radius (r). The ratio of a circle's circumference to its diameter (which equals 2r) is S. Thus 360° is equivalent to 2S radians, and one radian = approximately 57.3°. The solid angle subtended at the center of a sphere by an area numerically equal to the square of its radius. Since the surface area of a sphere is 4Sr2, a sphere subtends 4S steradians. The temperature scale where each degree is numerically the same size as that on the Celsius scale (i.e. there are 100 degrees between the freezing and boiling points of water) but whose zero point is Absolute Zero (the lowest temperature possible where all molecular movement ceases) 0K = 0010273°C approximately.
Table 1.1 Fundamental units
there is consistency between different methods of arriving at the same result. For example the heat generated by a gas heater or an electric heater must be measured in the same units; or the power developed by an electric motor must be comparable to the power developed by a gasoline engine. It would be very confusing if we had different units for both, although in practice this can happen! For example most people like to talk in terms of horsepower (derived from “imperial” units of measurement) for the capability of a motor, whereas it is in fact easier to work in kilowatts (derived from metric units) when working as an engineer. Most electrical engineering is now based on SI (Système International d’Unités) units. These are based on the MKS or meter, kilogram, second system, whereby the primary units of length, mass and time are used as the basis of all other units. Thus, while we could define the flow of electric current in terms of the number of electrons passing a particular point per second, it is actually agreed to define it in terms of the mechanical force created by a current carrying conductor, because this way we can relate it back to the MKS primary units. Originally the meter was related to the dimensions of the earth, and represented by a
physical standard of platinum–iridium kept in Paris. Similarly the second was related to the mean solar day. However, the demand for greater accuracy in measurement has resulted in the meter being related to the velocity of light, and the second to the emission of radiation from a particular atom. Table 1.1 lists the fundamental SI or MKS units, and Table 1.3 shows how other commonly required Prefix
Symbol
Multiplier
Example
PicoNanoMicroMilliCentiDeciUnitDekaHektoKiloMegaGigaTera-
p n P m c d da h k M G T
10 -12 10 -9 10 -6 0.001 0.01 0.1 1 10 100 1000 10 6 10 9 1012
Picosecond Nanometer Microfarad Millisecond Centiliter Decibel Hectoliter Kilogram Megawatt Gigahertz Terabyte
Table 1.2 The common unit modifiers.
5
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Quantity Frequency
Unit Hertz
Symbol Hz
Force
Newton
N
Energy (or Work)
Joule
J
Power (or the rate of doing work) Pressure (force per unit area)
Watt
W
Pascal
Pa
Defined as The number of repetitions of a regular occurrence within one second. Formerly cycles per second. The force which, when applied to a mass of one kilogram, gives it an acceleration of one meter per second per second. Force in N = Mass in kg u Acceleration in m/s 2 The work done when a force of one Newton acts for a distance of one meter in the direction of the force. Energy and heat are directly equivalent, so heat is also measured in Joules. (The SI avoids the use of the calorie; but a Kilogram-calorie, beloved of dieticians and the heat required to raise a kilogram of water by one degree Celsius, is found by experiment to equal 4187 Joules.) Work being done at the rate of one Joule per second. (The imperial based horsepower is 550 ft lb/sec, which converts to 746 watts.) When a force of one Newton is applied across an area of one square meter, the force is one Pascal. (One Bar is 10 5 Pa. Comparing with imperial measurement, one Pa is 1.45 u 10 -4 lb/sq. in.)
Table 1.3 Derived units (mechanical).
mechanical units are derived from them. Table 1.4 then shows the main electrical units, and here it can be seen that by relating the unit of electric current to physical force, and the unit of resistance to measurable heat energy, all the electrical units are “tied back” to the SI fundamental units. Many of the standard units are inconveniently big or small for certain types of measurement. They are modified by prefixes denoting powers of ten. Some of particular interest to lighting are frequencies such as kHz and MHz; and short wavelengths such as nanometers. Table 1.2 summarizes these prefixes.
of heat or pressure so that it becomes conducting. This will be discussed further in Chapter 3. Another category is the semi-conductor, the best known example of which is silicon. Pure silicon at room temperature is an insulator. But under certain conditions of heat or local impurities it is possible to detach a free electron from the incomplete M shell, at which point silicon becomes a conductor. Semiconductors are the basis of modern electronics, and are discussed further in Chapter 2.
1.2.2 Conductors and insulators
While we usually want a conductor to conduct electricity as well as possible, and an insulator not to conduct it at all, there are times that we need to limit the current flowing in a circuit. This is done with a resistor, which is simply a device which has a known resistance in ohms. Resistors take many forms (Figure 1.5.) At one extreme a heating element for a cooker or electric fire is a resistor, at the other modern electronic circuits use tiny resistors often with values in the kilohm or megohm range. The size of a resistor will depend on the way in which it is to be used. Because a resistor dissipates heat in direct proportion to its resistance in ohms, and in proportion to the square of the current going through it; it must be designed to get rid of this heat. A resistor in an
In Section 1.1 we identified copper and other metals as being good conductors of electricity, because of the availability of free electrons. Different metals have different conductivity, for example copper, silver and aluminum are good conductors; iron is a conductor, but is poor compared with copper. The opposite of a conductor is an insulator which in theory should not carry any electricity at all. In Figure 1.1 the gas Neon is shown as an insulator, because its electron shells are full and tightly bound to the nucleus. However, the perfect insulator does not exist. A very strong electric field can “break down” an insulator, or alternatively an insulating gas or compound can be ionized under certain conditions
6
1.2.3 Resistors
ELECTRICITY AND LIGHT
Quantity and symbol Electric Current I
Unit Ampere
Unit Symbol A
Electric Charge Q Electric Resistance R
Coulomb
C
Ohm
:
Electric Potential V
Volt
V
Electric Energy E
Joule
J
Electric Power P
Watt
W
Defined as One Ampere is the constant current which, if flowing in two infinitely long parallel conductors, of negligible cross-sectional area, in a vacuum and placed one meter apart, creates between them a force of 2 u 10-7 Newton per meter length. The quantity of electricity transported when a current of one Ampere flows for one second. The resistance of a conductor is one Ohm if a current of one Ampere flowing for one second generates one Joule of heat energy. When a resistance of one Ohm carries a current of one Ampere, the potential difference across the resistance is one Volt. Ohm's Law can be stated as V=IR The physicist James Joule determined that the heating effect in a wire conductor was proportional to the square of the current flowing I, to the resistance of the wire R, and to the time the current flows t, i.e. E vI2Rt (symbol v means “proportional to”) As shown above the definition of the Ohm is based on E = I2Rt when E is in Joules, I is in Amperes and t in seconds One Joule per second, so, from the basis of E above Power in Watts = I2R Applying Ohm's Law, the alternative Power in Watts = Current in Amperes u Potential Difference in Volts
Electric Capacitance C
Farad
F
Electrical Inductance L
Henry
H
Magnetic Flux )
Weber
Wb
Magnetic Flux Density B
Webers per square meter (or Tesla)
Wb/m2 (or T)
Or Watts = Amps u Volts is the more convenient and easiest form to remember. If the potential across the plates of a capacitor rises to one volt as a result of being charged with one coulomb, then its capacitance is one Farad. In practice this unit is too large for most purposes. The microfarad (PF), and nanofarad (nF) are common measures in electronics. A circuit (or coil) has an inductance of one Henry if an e.m.f. of one volt is induced across it when the current changes at the rate of one Ampere per second. Again this is a large unit and mH and PH are widely used. The magnetic flux which, linking a circuit of one turn, produces in it an e.m.f. of one volt as it is reduced to zero in one second at a uniform rate. But more strictly derived from the definition of flux density. A magnetic field has a flux density of one Weber per square meter if a conductor placed at right angles to the field and carrying a current of one Ampere, has a force of one Newton per meter acting on it.
Table 1.4 Electrical and magnetic units, showing how they are derived from the mechanical units in Tables 1.1 and 1.3
7
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 1.5 Examples of resistors. The circuit symbols for resistance, and some simple laws about using resistances in series and parallel. A variable resistor is a two terminal device; a potentiometer has a similar construction but is a three terminal device; a voltage is applied across the ends of the resistance element, and the moving contact has a varying potential difference (voltage) according to its position along the resistance.
Figure 1.6 Resistance varies with temperature. On the left a 100W tungsten lamp has the voltage presented to it changed by a variable resistor. Current through, and voltage across, the lamp are measured by A and V. The resistance can be calculated, and plotted against the voltage, and this shows that at proper operating temperature the filament has more than ten times the resistance it does when cold.
8
ELECTRICITY AND LIGHT
electronic circuit may only need to dissipate a few microwatts, or at most a watt or so. A resistor for use as a current limiter in a lighting or motor circuit may have to dissipate a kilowatt. There is a complication about resistance which is highly significant in lighting. In most materials it varies with temperature. Figure 1.6 shows a simple experiment in which a 100W light bulb is connected to a 230V mains supply via a suitable variable resistance to change the potential across the lamp. A voltmeter measures the volts across the lamp, and an ammeter measures the current through it. Using Ohm’s law it is possible to calculate the resistance for different voltages. The graph shows that when no current is flowing the lamp has a resistance of 50 ohms – but when the lamp filament is running at its rated operating temperature, its resistance has risen to over 500 ohms. The variation of resistance with temperature is defined by the temperature coefficient of resistance. The resistance of pure metals rises with temperature; whereas the resistance of carbon, silicon and insulating materials drops with temperature. Special alloys (such as Eureka, a 60/40 mixture of copper and nickel) have practically no change in resistance over a wide range of temperature and are used in applications where the resistance must remain constant.
I=
V R
1.2.4 Capacitors A capacitor (also formerly known as a condenser) consists of two conducting surfaces separated by an insulating layer called the dielectric. The plates (or electrodes) can be flat metal plates, and the insulator between them can be air; but as a practical component capacitors usually consist of thin plastic films on which layers of aluminum have been deposited, or are thin metal sheets with an electrochemically derived insulating layer – see Figure 1.7. An “empty” capacitor has no potential difference between its electrodes, but if an electric field is applied across them the capacitor “charges up”. The process is shown in Figure 1.8(a).
Figure 1.7 The principle of a capacitor (a), and (b) a possible construction using metallized film or paper.
9
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
At first the switch S is in a center-off position. There is no voltage across the capacitor C. If the switch is closed to terminal b, a battery is connected to the capacitor through a resistance R. Measurement of the current through the capacitor by ammeter A, and the voltage across it by Voltmeter V, gives the curves shown in Figure 1.8(b) with respect to time. The initial current is high, but as the voltage across the capacitor rises, the current drops to zero. The “charging time” is found to be limited by the resistor R. Double the resistor value, and the initial charging current halves, and the time to full charge is doubled. If the switch is returned to its center-off position, the voltmeter V will continue to show the full battery voltage across the capacitor – and in a perfect capacitor it will hold the charge indefinitely. Now if the switch is turned to the d position, the current and voltage will behave as shown in Figure 1.8(c) as the capacitor discharges. Figure 1.8(d) shows a hydraulic analogy of the operation of a capacitor. In the top cylindrical chamber there is a rubber diaphragm stretched across the cylinder (= the dielectric). In the bottom cylinder the piston can increase the pressure on one side of the diaphragm and move more fluid into the top chamber (= the battery applying an e.m.f.) If the force is removed from the piston, the diaphragm springs back and forces the fluid back again (= capacitor discharging). No fluid can get from one side of the diaphragm to the other under normal circumstances. However, if it is subjected to excessive pressure, it bursts. Exactly what happens when a capacitor is subjected to too high a voltage. Table 1.3 defines the ability to store charge in a way which can be expressed as a simple equation: Capacitance Ch arg e in Coulombs = in Farads Potential Difference in Volts
or
Q =C V
Practical capacitors used in lighting and electronics are usually measured in microfarads or mF. Their capacity can be related to their physical dimensions by the equation: C = ε0 ε r
10
A d
Figure 1.8 Charging and discharging a capacitor. Figure described in the main text.
ELECTRICITY AND LIGHT
where C is the capacitance in Farads, A is the area of the capacitor plates in square meters, and d is the distance between the plates in meters. ε0 is the permittivity of free space (or vacuum) a quantity which can be determined experimentally as having a value of about 8.85 × 10-12 F/m. It relates the electric force, or field strength across the capacitor plates measured in V/m to the electric flux density measured in Coulombs per square meter. (It also has a deeper significance, referred to in Section 1.6.) εr is the relative permittivity of the dielectric material. For a vacuum this is unity (one). For other materials it is higher, for example for paper it is 22.5. Plastic films have a relative permittivity of 4-6. Sometimes ε0εr is shortened to ε and is referred to as the dielectric constant of the material concerned. From an electrical power point of view a most important characteristic of a capacitor is that in any practical circuit containing even the smallest amount of resistance, it is not possible to change the voltage across a capacitor instantaneously. Figure 1.9 gives some more summary information about capacitors.
same happens with magnetic fields produced electrically – a force is produced which causes the affected items to repel or attract each other. The definition of the ampere given in Table 1.4 is based on the idea that any current carrying conductor creates a magnetic field, but also a conductor carrying a current in a magnetic field has a force exerted on it. An inductor consists of a coil of wire, usually, but not necessarily, wound on a core with
1.2.5 Inductors Electricity and magnetism are inextricably mixed. The fundamental discoveries of electromagnetism were: • when a conductor carries an electric current a magnetic field surrounds the length of the conductor. • if a conductor is moved in a stationary magnetic field, then an electric current is induced in it. Conversely if a magnetic field moves with respect to a stationary conductor, an electric current is induced in the conductor. This is the basis of electricity generation. • if a conductor carrying an electric current is placed across a magnetic field, a force is exerted on the conductor. This is the basis of electric motors. Any magnet has a “north” and “south” pole (named after the way small magnets align themselves with the earth’s magnetic field, as in a compass) and we are all familiar, from magnetic toys, with the way that like poles repel, and unlike poles attract. The
Figure 1.9 More about capacitors.
11
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 1.10 A simple inductor, showing the magnetic field. The field is more intense when there is an iron core, and the iron magnetizes.
ferromagnetic (or iron-like) magnetic properties. The details of different kinds of inductors relevant to lighting control are given in later chapters, for now
it is sufficient to see what happens when an inductor is placed in a simple circuit – in the same way we considered capacitors in 1.2.4 above. Figure 1.10 shows a simple inductor, and the way a magnetic field is created by it. In Figure 1.11(a) a circuit is shown containing an iron cored inductor, arranged in a way that confines the magnetic field largely within the core. The circuit also contains a resistor R and two ammeters, one of which is “center-zero”. Closing the switch S connects the battery to both the resistor and the inductor, and results in the currents shown in Figure 1.11(b). The current through R jumps immediately to the value determined by Ohm’s law. But that through the inductor takes time to reach a steady level (which does indeed reach a value limited by the resistance of the coil – but only when the steady state is reached does Ohm’s law apply as usual). What slows the current rise? The current produces a magnetic flux in the core; but because this flux is changing, it itself introduces an e.m.f. back in the coil. Lenz’s Law states that when this happens, the e.m.f. is always in the opposite direction to that which is creating the flux. This effect limits how fast the current can rise in the coil – in an inductor it is not possible to change the current
Figure 1.11 Experimental circuit showing the effect of inductance in a DC circuit (a). The switch on and switch off currents are shown in (b). IL is the switch-on current through the inductor, and IR the switch-on current through the resistor. Ib is the switch-off current, which goes through the resistor in the opposite direction to IR and arises from the back e.m.f. of the inductor. The magnetic flux in the inductor core builds up during switch-on, and collapses at switchoff.
12
ELECTRICITY AND LIGHT
instantaneously (compare the capacitor – where it is not possible to change the voltage instantaneously). If switch S is now opened, an interesting thing happens. The original source of e.m.f. has gone, so the current drops in the coil – but this creates a change in flux again and, therefore, an e.m.f. The only way for its resulting current to go is through the resistance R – the center-zero ammeter now shows a decaying current in the opposite direction to that originally flowing through R. What happens if there is no resistance R in the circuit? Opening the switch S produces a rapid current change, therefore a rapid flux change. Therefore, in turn, a big “back e.m.f.” which in practical circuits can result in arcing (sparks!) at the switch. Even in the simplest circuits using inductive components precautions have to be taken to limit the effects of back e.m.f. 1.2.6 Magnetic units In the discussion on capacitors, the concept of permittivity was introduced to describe how different materials are affected by an electric field. A similar idea applies to inductors and magnetism, where permeability is the description of how materials are affected by magnetic fields. Since magnetic properties are fundamental to ballasts and transformers used in lighting control, some understanding of magnetic units is important.
We have described an electric current in a circuit as being due to the presence of electromotive force. In electromagnetism we can postulate that in a magnetic circuit a magnetic flux is created by the presence of magnetomotive force (or m.m.f.) caused by a current flowing through one or more turns in a coil. Since the m.m.f. is proportional both to the current and the number of turns, the unit of m.m.f. in the MKS system is the ampere-turn. Figure 1.12 shows a toroidal coil with T turns carrying a current I. The mean length of the magnetic circuit is l meters. The magnetizing force H is defined as the m.m.f. per unit length, so: H = IT/l
The strength of a magnetic field is termed its magnetic flux density B. This is measured in Webers per square meter, or Tesla, and is defined by relation to the ampere, see Table 1.4. To see how B and H are related we need to do a small thought experiment. Figure 1.13 shows a long thin conductor in a vacuum A. It is carrying current of one ampere in the direction of the paper. Let us suppose that the return path of this current is a long distance away, so any magnetic field the return path generates does not affect things. The magnetic flux created by the current is in the form of concentric circles. What is sometimes referred to as Maxwell’s corkscrew rule says that the direction of the flux is clockwise if the current is flowing away from you as in Figure1.13. In this figure just one line of flux is shown, at a distance of one meter from the conductor. Since the conductor (and its return partner) form one turn, the m.m.f. acting on the flux path is one ampere-turn. The length of the flux path (2πr where r is one meter) is 2π meters. Thus the magnetizing force acting at one meter radius is: H = 1 Amp
× 1 Turn /2π meters
H=1/2π
ampere-turn/meter
or:
Figure 1.12 Defining the magnetizing force H. Here a toroidal inductor has magnetic circuit length L meters, has a coil of T turns and is carrying a current I Amperes.
ampere turns/meter
Now let us imagine a second conductor A´ sited one meter away from A, and also carrying one
13
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
represented by the symbol μ0 and applies not only to vacuum but also to non-magnetic materials. The magnetic flux inside a coil such as that in Figure 1.12 or 1.10 is greatly increased if an iron or other magnetic core is introduced. Magnetic materials are defined by their relative permeability, μr such that: B= μrμ0H
Figure 1.13 Defining the magnetic flux density B. Imagine the conductor A as perpendicular to the paper, with current flowing into the paper. Maxwell’s corkscrew rule says the magnetic flux is in a clockwise direction.
ampere. The definition of magnetic flux density gives us the equation: Force on conductor in Newtons = Flux density in Wb/m2 X length in meters X current in amperes. So for A´ we have: Force per meter = B (Wb/m2) × 1 m
× 1A=B
Newtons
But we know the force acting under these circumstances, because our thought experiment has replicated the conditions which define the ampere (see Table 1.3). The force is 2 X 10-7 N. So now we can say that: the flux density B at one meter radius from a conductor carrying a current of one ampere is 2 × 10-7 Wb/m2. And further, if we compare the flux density at A´ to the magnetizing force at A´, we now have: B 2 × 10 −7 wb / m 2 = H 1 / 2 π ampere − turn / m
or
B = 4 π × 10 −7 H
The ratio B/H for the conditions we have defined is referred to as the permeability of free space. It is
14
While the relative permeability of air and other non magnetic materials is 1, that for special nickel– iron alloys can be as high as 100,000. If measurements are made on different materials of the effect of increasing magnetizing force H on the flux density B, a graph of the kind shown in Figure 1.14 is found. Here it is assumed that the experiments started with unmagnetized material. The increase in flux density follows the curve OAC, but beyond a certain value of H the value of B reaches a limit. The material is said then to be magnetically saturated. If H is now reduced to zero, the value of B does not go down to zero, but retains a remanent flux density OD (usually between 60 and 75% of the maximum). A reverse magnetizing force OE, known as the coercive force is required to return B to zero. If the reverse force is increased, point F can be reached where the material is now saturated in the other direction. If the cycle is continued, by reversing H again, the curve follows FGC. The complete loop resulting from a double reversal of magnetizing force is called the hysteresis loop. The “fatter” the curve, the more effort is needed to take the magnetic material through a flux reversal. 1.2.7 Inductance The unit of measurement of the ability of an inductor to slow down a rise in current, its inductance, is the Henry, defined in Table 1.4. For an air cored coil of the kind shown in Figure 1.12 the inductance is: • proportional to the square of the number of turns of wire T. • proportional to the cross sectional area of the core a.
ELECTRICITY AND LIGHT
same as if the flux varied linearly along FOC, and, therefore, that the inductance of the coil becomes: Induc tan ce L =
μ r × 4 π × 10 −7 × aT 2 Henrys l
1.3 Electrical distribution 1.3.1 Direct current sources
Figure 1.14 The relationship between flux density B and magnetizing force H is not linear, and displays hysteresis.
• inversely proportional to the length of the magnetic circuit l. In fact: Induc tan ce L =
4 π × 10 −7 × aT 2 Henrys l
When the air core is replaced by an iron or other magnetic core, the situation becomes very complicated. The magnetization curve Figure 1.14 shows that the variation of flux is not linearly proportional to the current (which causes the magnetizing force) – so a coil can have a lot of different inductance values depending on the range of current variation. However, our interest in inductance is primarily concerned with alternating current (described in Section 1.3.2) where the current changes direction repeatedly. If μr is the relative permeability of the core material corresponding to the maximum value of the flux (point C on Figure 1.14) and it is understood that the relevant use of the inductance involves repeated flux reversals (to point F) then it is assumed that the value of the inductance is the
In the concepts presented so far, only circuits using Direct Current or DC have been mentioned. With the exception of the changes in current or voltage introduced by capacitors or inductors, the circuits have been “steady state” with current flow conventionally seen as flowing from positive to negative. DC is essential for the operation of most electronic components, and is most familiar to us as being supplied by batteries. A battery is in fact a group of electrochemical cells whose action is to convert chemical energy to electrical energy. Primary cells or batteries are “use once” devices, exemplified by the common zinc/carbon dry battery. Secondary cells are rechargeable devices, where the chemical reactions leading to the generation of electricity can be reversed, and are exemplified by the standard automobile battery.
1.2v 300mAh battery used in cordless control
3.6V 4Ah rechargeable battery used in emergency light fitting
12V 11Ah sealed lead acid rechargeable battery
Figure 1.15 Examples of batteries. Battery capacity is given in ampere-hours, thus a 2Ah battery can sustain a current of 2A for one hour or 500mA for 4 hours.
15
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 1.15 shows some examples of batteries. The demands of portable electronic equipment, emergency power supplies and electric traction have resulted in considerable developments in battery technology.
DC is also derived by power conversion from AC, and this is described in Chapter 2. It can also be generated by DC generators or dynamos, but this is not now generally relevant to lighting.
Trigonometry In our treatment of electrical units and electricity, we are keeping mathematics to a minimum; using where possible simple algebraic formulae with each symbol clearly defined. In describing alternating current it is necessary to use simple trigonometry. For those who have forgotten the basic relationships, here they are: In the right angled triangle in Figure 1.16 the angle θ is identified, as are the three sides of the triangle: • the hypotenuse which is the longest side opposite the right angle, the length of which is identified here as H. • the adjacent side which is the side next to the angle θ. Identified here as having length A. • the opposite side, which is the side opposite to the angle θ, identified here as having length O.
Figure 1.17 A force F acts on an object at angle θ to the horizontal. If the line OF represents the magnitude and direction of the force; it can be resolved into horizontal and vertical components of forces H and V, represented by the lengths of OH and OV. Conversely, if there are two forces acting on an object, there is a resultant force whose magnitude and direction lies between them.
The following relationships apply: O2 + A2 = H2
(Pythagoras’s Theorem)
The Sine of the angle θ = O/H The Cosine of the angle θ = A/H The Tangent of the angle θ = O/A Sine, Cosine and Tangent are abbreviated to Sin, Cos and Tan in mathematical expressions. A mnemonic for remembering the above is:
Figure 1.16 Simple trigonometry. Defining the sine, cosine and tangent of angle θ.
16
Old Always Other Hands Help Aviators
ELECTRICITY AND LIGHT
An important use of trigonometry is in “resolving” forces, and in defining the contribution of a vector quantity (a vector quantity is one which has both magnitude and direction) in a particular direction. Figure 1.17 shows the principle of resolving a force into two different components. A force of F Newtons is acting on an object at an angle θ. If we want to know the horizontal and vertical components of the force, we draw the rectangle shown, with the length of the diagonal corresponding to the magnitude of the force. The lengths of the sides of the rectangle now represent the horizontal and vertical components of the force, and it can be seen that:
F cos θ is the horizontal component, and F sin θ is the vertical component But there is an important concept here, which will be appreciated as we look further into how alternating current, described in the next section, behaves. That is that, if you have two vector components acting in two different directions, you have a resultant single vector – and while this is quite easy to appreciate when dealing with a force, it may not be so obvious when you are dealing with an electric current.
1.3.2 Alternating current The electricity which arrives in our home is Alternating Current or AC. The reason for this is that it is both easier to generate, and easier to distribute. To understand why this is so, it is necessary to understand how an alternator (or AC Generator) works. Figure 1.18 shows a rectangular loop conductor rotating in a magnetic field. A conductor cutting magnetic flux generates an e.m.f. but here the conductor is not always cutting the flux at the perpendicular. When the loop is vertical the conductors are parallel with the flux, and no e.m.f. is generated, when the loop is rotated 90° the loop is cutting across the flux and maximum e.m.f. is generated. The instantaneous e.m.f. is proportional to the sine of the angle at which the conductor cuts the magnetic flux. e.m.f. generated in one side of loop = Blv sin θ volts
Where B is the flux in Wb/m2, l is the length of one side of the loop, v is the velocity of the conductor through the flux and sin θ is the resolved component of that velocity perpendicular to the magnetic field. Without needing to worry too much about the details of this equation, it is clear that the more and longer conductors there are, and the faster the loop rotates, the more volts will be produced.
Figure 1.19 shows how a sine wave is produced by this action. It also shows another “mind picture” of how a rotating vector produces the sine wave. A vector quantity is one which has both magnitude and direction (for example speed is a scalar quantity, having magnitude only, whereas velocity is a vector quantity which must always be specified in both magnitude and direction). The magnitude in both cases can be km/hour or m/s or m.p.h. as appropriate. Notice here how the use of angular velocity ω, measured in radians per second simplifies the mathematics, since there are 2π radians in each revolution. A version of the device shown in Figure 1.18 could be used for generating AC by connecting pickup brushes to both ends of the conducting loop as shown in the diagram. (Indeed a very rough DC could be derived by having a “commutator” arrangement which switched the direction of current to the outgoing circuit every half revolution.) However “real” alternators of any size work the other way round. They have static conductor coils, which are then easily connected to the outside world without the need for any pick up brushes; and the moving magnetic field is produced by rotating electromagnets. These are powered by DC, traditionally generated by a dynamo on the same shaft as the main alternator, but now also derived by converting from AC.
17
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 1.18 Creating an alternating current. On the left a loop conductor being rotated in a magnetic field (shown in the position of rotation θ=90°). On the right it is shown that maximum e.m.f. is generated when the conductor cuts the flux at right angles (rotation angle 90°) and no e.m.f. is generated when the conductor is moving in the same direction as the flux lines (rotation angle 0°).
Figure 1.19 If the value of r sin θ is plotted at equal increments along an axis, a sine wave is the result, as shown on the left. On the right we see the idea of a rotating vector producing a sine wave. The horizontal axis can represent both the angle of rotation and time (if the angular velocity is known). If the vector length represents a current of maximum value IM the waveform is that of an alternating current.
18
ELECTRICITY AND LIGHT
Figure 1.20 The single phase alternator. The left hand diagram shows the principle with a single 2-pole magnet. The right hand diagram shows an 8-pole magnet system. In a real alternator the magnets are electromagnets fed by a DC supply.
Figure 1.20 shows the principle of an alternator. AC does not have to be sinusoidal (sine wave shaped) and many other waveforms exist, especially in electronics. But for power distribution it is the most efficient waveform, so alternators are designed to give a sinusoidal output. (To be technically correct it should be noted that the alternator construction implied by the two arrangements shown in Figure 1.20 would not give sinusoidal output without some refinement.) To make better use of the alternator frame, and to assist with power distribution, AC is often produced in three phase form. Here the alternator simultaneously produces three different sine waves, but each of these is timed to peak at different times. The idea is shown in Figure 1.21, and a real alternator is shown in Figure 1.22. In the USA, electricity is generated at 60Hz (cycles per second) in Europe it is generated at 50Hz.
1.3.3 AC circuits In Section 1.2 many of the basic rules about electricity were given in relation to DC circuits. Unfortunately they need some “tweaking” when applied to AC. First we need to know if Ohm’s law and the simple power calculation rules still apply. Clearly it is no use taking “average” current and voltage, since these are both zero if the waveform is symmetrical. The way forward is to say that the effective value of an alternating current is that which produces the same heating effect in a resistance as does a direct current of the same numerical value. We can do a comparison by “slicing up” an AC waveform. In Figure 1.23 we show n instants at which current is measured in a half cycle. The instantaneous heating effect at each instant is in2R. So we can say
19
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 1.21 The principle of the 3-phase alternator. The top diagram shows how the three sets of stator coils are set at 120° (or 2π/3 radians) apart, the bottom diagram shows the relationship of the three waveforms.
In words, the current equals the square root of the mean of the squares of the current. The r.m.s. or root mean square values of current and voltage are used whenever AC supplies are specified. Ohm’s Law and the relationship Watts = Amps × Volts both work when applied to resistive loads. Some other aspects of the sine wave are shown in Figure 1.24. An interesting point arises with 3-phase AC. Usually each phase can be considered as an independent supply, requiring two wires. In most electrical systems one side of the supply is connected to earth or “ground”. It is referred to as the Neutral connection, and the other side is called the Live (or “Hot”) connection. You get an electric shock off the live connection, but not off the neutral connection (provided it is well bonded to earth) because the neutral is at the same potential as you are if you are standing on the ground. In a 3-phase system the three neutrals are obviously joined together if they are all to be at ground potential. If you look carefully at a 3 phase waveform as in Figure 1.21 you can see that at any instant the three currents add up to zero – for example if at any one moment there are two positive currents and one negative, the sum of the two positives
that the average heating effect over the whole half cycle is: 2
2
2
2
i 1 R + i 2 R + i 3 R + ..... .... + i n R n
Now if we say that I amperes is the Direct Current through the same resistance R which gives the same heating effect as the average heating effect of the AC, we have: 2
I 2R =
2
2
2
i1 R + i2 R + i3 R + ..... .... + in R n
Therefore: 2
I =
20
2
2
i 1 + i 2 + i 3 + ..... .... + i n n
2
Figure 1.22 A power station generator under construction. This one is from Siemens AG, and is rated at 1,000 MVA, with output at 27 kV. It uses hydrogen cooling to achieve a comparatively small frame size. The rotor carrying the rotating electromagnets is being inserted into the stator assembly, the coils of which collect the induced alternating current.
ELECTRICITY AND LIGHT
Figure 1.23 Finding the heating effect of AC. Imagine the sine wave to be divided into many small sections, of current value i1 i2 i3 and so on to in. Each little chunk of current will make its own i2R heating contribution.
exactly equals the negative. The vector representation of currents helps make this clear. A 3-phase system with common neutral is referred to as a “star” or “Y” system, and its feature is that if the load on all phases is equal there is no neutral current. Power system design usually calls for balancing the load on the three phases in order to minimize or preferably eliminate neutral current. This is of great significance in lighting systems. While the r.m.s value of the voltage on a single phase in a 3-phase system is the same for each phase
Figure 1.24 Peak, r.m.s. and half wave average values of a sine wave. For all symmetrical waveforms the form factor is the ratio of r.m.s value to average value (1.11 for a sine wave) and the crest or peak factor is the ratio of the peak value to r.m.s. value (1.414 for a sine wave).
Figure 1.25 3-phase electricity. (a) shows the concept of three simultaneous rotating vectors, 120° apart. (b) shows a 3-phase alternator with Y output and Y connected loads; and (c) shows a 3-phase alternator with delta output (ǻ) and delta connected loads.
21
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
(for example 230V in Europe and 115V in USA in the home) the voltage across phases is much higher (and more dangerous). The r.m.s voltage across two phases is √3 × the single phase voltage (about 398V when the single phase voltage is 230V). There is another way of distributing 3-phase AC, arising from the fact that the vector sum of the e.m.f.’s is zero. This is to do away with the “neutral” connection altogether, and connect the alternator windings in a “mesh” or “delta” format. Figure 1.25 shows the idea. Vector considerations of the current show that the current in each load line is √3 times the current in the phase winding of the alternator. The arrangement is particularly suitable for powering big electrical motors. It is also encountered in some electical distribution systems (for example in Norway and on board many ships) and this has some consequences for lighting control systems. 1.3.4 Transformers In any simple electrical system, the power delivered to the load is the product of the voltage and the current. If we want to transfer 10kW of power, we can do it in several ways:
The voltage conversion process is done by a transformer. A transformer extends the principle of the inductor described earlier. A typical construction is shown in Figure 1.26. A coil is wound on an iron core and is fed from an AC supply. A second coil is mounted on the same core. When the current changes in the first coil or primary winding, it produces a change in magnetic flux, and this flux change must cut the conductors of the second coil. A current is, therefore, induced in the secondary winding, through mutual inductance – the process where two coils share the same magnetic flux. In the description of the inductor in Section 1.2.5 its behavior was discussed in relation to the application or removal of a direct current. This showed that the current reached a steady value and the coil circuit obeyed Ohm’s law once the steady state was reached. If we were to apply DC to a transformer, the same thing would apply. The current induced in the secondary would be momentary, and all that would happen is that the transformer would cook due to the high current and low coil resistance of the primary.
• at 10 Volts, we would need a current of 1,000A • at 200 Volts, the current is a reasonable 50A • at 10,000 Volts, the current is only 1A High currents mean very big conductors if there is not to be significant heat loss in the conductor due to the conductor’s resistance. The heat effect is proportional to the square of the current, so a cable resistance of one ohm would result in 2,500W being dissipated by the 50A current, but only 1W by the 1A current. Thus for electrical distribution the power generated at the power station’s alternator at around 20,000 Volts (20kV) is “stepped up” for primary distribution to as much as 400kV. Local distribution may be done at 11kV or 33kV, requiring a “step down”. Final distribution to the home and office requires a further step down to 230V or 115V depending on where you live. Figure 1.26 Construction of a transformer.
22
ELECTRICITY AND LIGHT
Figure 1.27 Example of a transformer used for lighting – a step-down transformer for low voltage tungsten halogen lamps. (Photo from Relco.)
But with AC the current, and, therefore, the magnetic flux, is constantly changing. There is a continuous back e.m.f. limiting the current in the primary, and the current in the secondary follows that of the primary. In a perfect transformer (and in fact most power transformers are quite efficient) all power put into the primary can be extracted from the secondary. For reasons explained in Section 1.4, transformer ratings are always given in VA (or VoltAmps). Figure 1.27 shows a 300VA transformer intended for use for lighting operating at 12V. This means that if it is fed with AC at 230V, the primary current is 300/230 = 1.3A. At the secondary, rated for 12V, there is a total of 300/12 = 25A available. The Ampere/Turns must be the same for the primary and secondary (clearly the transformer cannot actually create energy). Therefore:
In the example transformer, the 230V primary winding might have 400 turns, thus to give 12V at the secondary, the secondary winding needs only 21 turns. The turns ratio determines the voltage ratio. Good transformer design confines most of the magnetic flux close to the transformer core, so all of it is used to induce e.m.f.’s into the conductors. Power transformers working at 50 or 60Hz use special transformer iron alloy to provide the core. Being metal, this is a conductor – therefore the moving magnetic flux must induce electric current into it. The current so induced is called an eddy current because it circulates within the metal. Eddy currents cause heating, like any other current, and could be responsible for serious losses in transformers (and, of course, other electromagnetic devices, including alternators and motors). The problem is solved, if not entirely eliminated, by using a laminated construction for the transformer core. See Figure 1.28. The laminations are insulated from one another, and in this way any eddy current is confined to a single lamination. An example illustrates how effective the technique is. Suppose we substitute a single core piece with ten laminations: • the e.m.f. per lamination is only one tenth of that generated in the solid core.
Primary × Primary Secondary × Secondary turns amps = turns amps
But also: Primary × Primary Secondary × Secondary volts amps = volts amps
So: Secondary volts =
Secondary turns × Primary volts Primary turns
Figure 1.28 Electromagnetic devices working at 50/60Hz use iron alloy laminations to minimize the effect of eddy currents.
23
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 1.29 The autotransformer has only a single tapped winding.
• the cross sectional area is also reduced to one tenth, so the resistance goes up ten times; this reduces the current by a further factor of ten. The current per eddy path is thus reduced to 1/100th of what it would be in the solid core. In fact the eddy current loss is proportional to the square of the thickness of the laminations, so within reason the aim is to use the thinnest laminations practicable. An important feature of most transformers is that primary and secondary windings are completely separate; the power transfer is achieved entirely by induction. For many applications this is essential for safety or signal isolation reasons. In an autotransformer only a single tapped winding is used (Figure 1.29). Transformers used at high frequencies work on the same principle but use a different construction, described in Chapters 6 and 7. The iron alloy used in low frequency transformers has pronounced hysteresis, making it unsuitable for high frequency operation.
In an AC circuit with a resistive load, the current in the load is exactly in phase or “in step” with the applied voltage. But this is not the case with capacitance and inductance. Figure 1.30 shows a sine wave of current flowing through a pure inductance. The instantaneous value of current is Imax sin ωt. Now the voltage across an inductance is given by the inductance multiplied by the rate at which the current is changing. Mathematically this can be stated as: e =− L
di dt
Where e is the voltage, L the inductance in Henrys, and di the change of current in small time dt. The negative sign arises because the induced e.m.f. is in the opposite direction to the applied current. The rate at which the current is changing at any instant corresponds to the instantaneous slope on our sine wave current curve. If we plot the value of this
1.4 Power factor 1.4.1 Reactance The description of the transformer has introduced the idea that some loads may behave differently to simple resistances when connected to an AC supply. If we consider what happens when a pure inductance (i.e. an inductor with no resistance and no capacitance) or a pure capacitance (a capacitor with no inductance and no resistance) is connected to an AC supply we find the voltage and current get “out of phase”.
24
Figure 1.30 AC flowing through an inductance. The top sine wave is the current. The bottom curve is a plot of the gradient (or slope, or rate-of-change, whichever you find easiest to imagine) of the current sine wave. It turns out to be a cosine wave, which is simply another sine wave 90° (or π/2 radians) out of phase with the first one.
ELECTRICITY AND LIGHT
slope against time, we find we get another curve looking the same, but out of phase with the current curve. It is a cosine curve, which is exactly the same as a sine curve, but 90° (or π/2 radians) out of phase. It is at its maximum when the current is zero, and at zero when the current is at its maximum. This sounds counter-intuitive, but what is happening is that the external AC supply puts energy into the inductance, but the induced e.m.f. puts the energy back into the supply. A mechanical analogy is that of the pendulum – when the pendulum bob is stationary, as it is changing direction at its highest point, it has maximum potential energy (voltage) but no movement (current); and as it passes through its lowest point, it is moving fastest (highest current) but has no potential energy (voltage). What limits the current through a perfect inductor? Is there an equivalent of resistance? There is, it is called the inductive reactance with symbol XL. It is the ratio of r.m.s. voltage to r.m.s. current, and is, therefore, measured in ohms. We can derive the formula for reactance in several ways, but the easiest is using simple calculus. Start with the equation: i = I max sinωt
Which is the description of our current sine wave. Differentiate this with respect to time to obtain the slope of the curve at any time (remember this is the rate of change of current with time). di = ȦI max cos Ȧt dt
(This is a mathematical shorthand for describing the lower curve in Figure 1.30. It says that any point on the curve representing a particular value of di/dt, or rate of change of current, can be calculated by the instantaneous value of ωImax cos ωt.) Going back to the equation e = - L di/dt, we can substitute the general expression for rate of change with the curve value (ignoring the - sign): e = −L
di = ω LI max cos ω t dt
For any angle θ, Cosine θ = Sine (θ + 90°) so we can write: e = ωLImax sin( ωt + 90° )
But in magnitude terms we know that Imax sin (ωt + 90°) is the same as Imax sin ωt (but advanced by 90°) and this equals the instantaneous current i. So we have in magnitude: e = ωLi
We have defined the inductive reactance XL as e/i so we have: XL =
e ωLi = = ωL i i
Now ω is the angular velocity of the current vector in radians per second, but this is the same as 2πf, where f is the frequency of the supply. Even if you have not followed the mathematics completely, the end result X L = 2πfL
is interesting. It tells us that the reactance is not simply proportional to the inductance, but also to the frequency of the AC supply. The higher the frequency, the higher the reactance. The same kind of analysis can be applied to a pure capacitance. This time we start with the equation: i=C
dv dt
That is to say, the instantaneous current equals the capacitance multiplied by the rate of change of voltage dv/dt. This time it is easiest to start with the voltage sine wave, and create the the current waveform from it (by the same procedure of plotting the new curve of instantaneous dv/dt). We find that, whereas with inductance the current lags the voltage by 90°, with a capacitance the current leads the voltage by 90°, and that the capacative reactance XC is defined by:
25
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Xc =
1 1 ( ohms ) = ωC 2πfC
Again the reactance depends on frequency. This time, however, it is inversely proportional – the higher the frequency the lower the reactance. Our “mind picture” of what is happening with reactive loads such as capacitors and inductors must not get carried away. It is easy to think of a capacitor “carrying” current through itself, but actually charge is building up cyclically on either side and then falling away. With an inductor, current flows in, to be returned by the back e.m.f. Both devices are temporary stores of energy, and in their pure form do not take heat energy from the supply. The fact that in both cases the current is 90° out of phase with the voltage means that average VA product is zero. Pure inductive and capacitive loads are sometimes referred to as “wattless”. 1.4.2 Impedance In any real circuit there is no such thing as a pure inductor or a pure capacitor. Any such device, including any real-world connections to it, includes resistance. Depending on the frequency of operation, an inductor may include significant, in the sense of reactive, capacitance and a capacitor may exhibit some inductance. The combination of the reactance and resistance in a circuit is referred to as impedance, and is frequency dependent. It is represented by symbol Z. The rules for calculating impedance, and for calculating currents in an AC circuit are more complicated than for simple DC circuits. We could use several different mathematical treatments to explain them, but the simplest way to visualize what is happening is the vector method. Let us start with the problem of adding two AC voltages, which are not in phase, but separated by a phase difference of angle φ. What is the resultant voltage? In the vector model we imagine two rotating vectors, each generating a sinusoidal AC voltage. We draw a line OE1 representing the first voltage. The length of the line represents the peak voltage E1, and the instantaneous voltage we designate e1. Figure 1.31(a) shows the idea. (At this stage we are
26
Figure 1.31 Adding two vector quantities (AC voltages) by the parallelogram method. (a) shows the principle for two rotating voltage vectors with phase difference φ. (b) shows an r.m.s. vector diagram, where V = 0.707E, and the vectors are stationary because time is not involved with r.m.s. values.
not bothered about the instantaneous phase situation of e1 but at any time e1 = E1 sin θ, where θ is the instantaneous phase angle.) Now we do the same for the other voltage, which we envisage as having a peak value of E2 and an instantaneous value of e2. The instantaneous sum of the two voltages can be expressed as: e = e1 + e 2 If we create a parallelogram from the two original vectors, we see that the diagonal of the parallelogram has an instantaneous value corresponding to e. Therefore we can say that the resultant voltage has a peak value E, corresponding to the length OE of the
ELECTRICITY AND LIGHT
We can generalize this approach for r.m.s. voltages and currents. Figure 1.31(b) shows the corresponding r.m.s. vector diagram (where r.m.s. values are 0.707 of peak) and the vectors are considered stationary. In practice an inductor is a device that has both inductive reactance and resistance, and can be considered as an inductance and resistance in series as shown in Figure 1.32(a). Figure 1.32(b) shows the vector relationship between the resistive voltage (which must be in phase with the current) and the pure inductive voltage, which lags 90° behind it. The line voltage V is the diagonal of the rectangle, and leads the current by φ°. The figure may be re-drawn as a “voltage triangle” as Figure 1.32(c). But we know in each case that the relationship between voltage and current is V = IR for a resistance, V = IX for a reactance and V = IZ for the circuit impedance. Pythagoras’ theorem gives us: (IZ)2 = (IR)2 + (IX)2 which can be restated as: Z2 = R2 + X2 or Z = √ (R2 + X2) From the above, the “voltage triangle” idea applies equally to impedance, as Figure 1.32(d). 1.4.3 Power factor
Figure 1.32 (a) shows a resistance and inductance in series. (b) shows the applied voltage V being made up of V1 across the resistance, in phase with the current; and V2 across the inductance, 90° ahead of the current. (c) shows the corresponding voltage triangle, from which we can derive the impedance triangle (d).
diagonal. Its phase relationship to the constituent voltages clearly depends on both their respective magnitudes and the phase relationship between them, and can be measured directly from the parallelogram model.
In a typical AC circuit with reactance present, a vector calculation of the resultant current and voltages done on the basis of Figure 1.32 will usually yield a VA product which is higher than the power in watts being delivered to the resistive load. The power factor in an AC circuit is defined as: watts r .m .s .Volts × r .m .s . Amps
or power factor =
watts VA
But watts = I2R and VA = I2Z.
27
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Table 1.5 Impedance formulae and power factors.
So power factor = R/Z, and from the impedance triangle we can see that: Power factor = cos φ where φ is the phase angle difference between current and voltage. When the current follows, or lags, the voltage, there is said to be a lagging power factor – as is the case with inductive loads. A leading power factor applies for capacitive loads. Table 1.5 shows the impedance and power factor for the common circuit elements. Any circuit involving inductance, especially electric motors and electromagnetic fluorescent lamp ballasts, will have a poor power factor of, typically, around 0.5. Why “poor”? Electric utilities companies don’t like low power factors because it means that their generation and distribution plant must be rated for the maximum VA product taken, but their customers only pay for the watts actually used. In practice inductive loads, with lagging power factor,
28
are the problem – and the problem is solved by applying power factor correction using capacitors to restore the power factor to near unity. This is an important technique for most kinds of discharge lighting. Today the term “power factor” is used to cover all “wattless” activity. As will be discussed later, many modern power control systems introduce harmonics onto the power line, or, looked at another way, they switch the power on and off very rapidly. The power company only gets paid for the “on” time, and cannot divert the resource during the “off” time which is measured in milliseconds. So they don’t like harmonics, any more than they like their fundamental sinusoidal current and voltage being out of phase. These factors are highly significant in lighting. When cos φ is quoted on a piece of electrical equipment (typically a fluorescent lighting ballast) it is defining power factor in the traditional way described above, related to a standard power frequency of 50 or 60Hz. The alternative power
ELECTRICITY AND LIGHT
factor symbol is λ which is the ratio of watts to all wattless components however they arise.
mechanically power assisted, and have arrangements for arc quenching.
1.5 Control of electrical power
Where there is to be a repeated switching action at comparatively high frequency (for example in animating an electric sign) the ordinary switch is no longer appropriate. Before the advent of electronic switches the best that could be done was the use of motor diven switches with “brushgear” contacts or of mercury switches where connection was made through a “liquid contact” of mercury. Today all high speed switching is done by power electronic components such as the thyristor or power transistor, described in Chapter 2.
The control of electric power is divided into two: • at its simplest to be able to connect power to a device (whether it be a giant electric motor, an electric kettle or a bedside lamp) and, just as important, the ability to disconnect it. • at a more sophisticated level, the ability to regulate a power flow. In this case we may want to adjust the speed of the motor, adjust the temperature of the water in the kettle, or the brightness of the lamp. In this section we review some of the control elements which are relevant to lighting. 1.5.1 Switches The humble light switch is the most familiar electrical control device. It is no more than a device which makes a physical break in the electrical circuit. However: • The switch contacts must not only be able to carry the running current, they may have to absorb an “inrush” current on switching on (and, indeed, must be able to withstand a short circuit current matching the rating of the fuse or circuit breaker protecting it) and, more importantly, suffer some arcing when switching off. This is particularly the case for DC, and also for inductive AC loads like fluorescent lamps and motors. • The switch contact must have negligible resistance. If it gets corroded by arcing, its resistance goes up, I2R heating sets in, and its resistance goes up still further. Eventually it fails, but it may have caused a fire in the meanwhile. For this reason switch contacts are usually made with precious metal alloys, and are constructed with a “wiping” action to assist maintaining a clean connection. A spring “snap action” is also included to make a fast circuit break. • High power switches used in electrical distribution have their contacts immersed in an insulating oil; or are mounted in an inert or otherwise controlled atmosphere. Any “open air” switches are
1.5.2 Fuses and circuit breakers All electrical circuits require protection against faults which might either damage equipment or injure people. Fault protection is divided into different types: Short circuit protection. This protects against short circuits. For example a stage lighting dimmer needs good short circuit protection if a badly wired temporary cable creates a “short”. Short circuits can result in very high currents, so must be cleared quickly.
Figure 1.33 Power factor correction capacitors, as used in fluorescent and high intensity discharge lamp circuits.
29
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Overload protection. This is to protect against conditions where the actual load is greater than that for which the circuit or control device is designed. For example where someone puts a 3kW lamp load onto a 2kW dimmer, or where a fault like a “shorted turn” develops in a transformer, which could lead to a fire if not detected. Life safety protection. This is a form of protection against lethal electric shock. In this case the equipment may be functioning correctly, but someone may have inadvertently touched a “live” part of it. The best known protection device is the fuse, and this is based on the simple principle of having a deliberately weak link in a circuit which is destroyed under fault conditions. A simple fuse is a single strand of thin tinned copper wire. When a particular current is exceeded, the I2R heating effect melts the wire, and the circuit is broken. When the wire breaks an arc develops between the two pieces of wire, so the circuit is not actually broken until the arc is extinguished through further destruction of the wire. The simple fuse is, therefore, an imprecise device, especially as it changes its characteristics with age due to oxidation of the wire. Typically a wire fuse blows at about 200% rated carrying capacity. It is therefore acceptable for short circuit protection, but not for overload protection. Fuses on their own do not provide protection against electric shock. The high rupturing capacity or HRC fuse is a more satisfactory device. For power circuits it is housed in a strong ceramic body to withstand the thermal and mechanical stresses arising from both normal running and fault conditions. The body is filled with granulated quartz, and the fuse element is silver. When the fuse blows the arc is immediately quenched because the silver and quartz together yield a high resistance compound. HRC fuses are precision devices. They can run at 100% rated current capacity continuously, but in some versions can be relied upon to blow at only 120% rated current. When HRC fuses are subjected to short circuit, they have a “clearing time” consisting of the time taken to melt (“pre-arcing” time) and the time taken for the arc to extinguish (“arcing time”). Their
30
5x20mm glass fuse rating 100mA - 10A Breaking capacity [email protected]
5x20mm sand filled ceramic body fuse rating 100mA - 10A Breaking capacity [email protected] Panel mounted fuse holder for 5x20mm fuses; used in electronic and electrical equipment Consumer HRC fuse ceramic body rating 5A - 45A Breaking capacity [email protected] General purpose European style fuse ratings 1A - 63A in different sizes Breaking capacity up to [email protected] Industrial HRC fuse rating 2A - 200A in different sizes Breaking capacity up to [email protected]
Industrial HRC fuse carrier
Figure 1.34 Examples of fuses and their carriers.
characteristic is shown in Figure 1.35. It is obviously important that the “cut-off” current which is let through is lower than that which can be withstood by the equipment being protected. Usually the assymetric cut-off current is appreciably less than the prospective r.m.s. short circuit current. For
ELECTRICITY AND LIGHT
example a 60A fuse encountering a prospective short circuit of 50,000A should in practice cut off at 7,000A. All distribution systems need protection devices which have discrimination. Thus when protection is cascaded so, for example, one big fuse is used to protect a distribution panel with several lower rated fuses, it is important that the correct fuse blows when there is a fault. Thus the bigger fuse must have a higher cut-off current than the smaller one, otherwise a failure in a sub-circuit will blow the master fuse instead of the circuit fuse. (The same argument applies to circuit breakers.) When overcurrents involve fusing times of less than 10ms on a 50Hz supply, a different rating is used. This is based on the let-through energy measured as I2t, where I is the current in Amps and the time in seconds. In order to have correct discrimination, a family of fuses should have separated I2t, as shown in Figure 1.36. Small fuses are widely used for the protection of electronic equipment, both in the high voltage supply side and on the low voltage side. They represent the lowest initial-cost form of protection, but, of course, they are destroyed when they blow. In most cases a fuse failure in such equipment is a signal that something is seriously amiss.
Figure 1.35 The clearing time of a fuse.
Figure 1.36 Fuse discrimination measured by I2t in a family of fuses.
In power circuits fuses are used in primary protection, for example at the power intake to a house, and in some circuit protection (for example in the UK where the 13A portable appliance plug has a ceramic bodied fuse in it). Fuses are also widely used for the short circuit protection of electric motors. However, the majority of electrical distribution for lighting is based on circuit breakers. A circuit breaker is an automatic switch which automatically switches itself off when there is an overload. Circuit breakers use two different principles of operation. A thermal circuit breaker uses the heat of the current passing through it to heat a bi-metallic strip. Bi-metallic strips bend when heated due to the different rates of thermal expansion of each metal (the same action as in a thermostat). The strip triggers a spring release mechanism when it has heated up sufficiently. Thermal circuit breakers are useful for long term overload protection, but are imprecise because their tripping characteristic can depend on the load being carried and on ambient temperature. Magnetic circuit breakers rely on electromagnetic action. A coil carrying the fault current exerts a force on a plunger to trigger the release. This action can be very fast, resulting in nuisance tripping on loads which have an inrush current. This can be avoided by damping the plunger action hydraulically.
31
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 1.37 The miniature Circuit Breaker or MCB. Diagram from MEM Circuit Protection and Control (Delta Electrical Ltd.)
In lighting systems local cicuit protection is usually by miniature circuit breakers or MCBs. These have a combined thermal and magnetic action to deal with both short circuits and long term overload. An electrical distribution panel will usually have a group of MCBs connected to a bus bar, a solid copper or aluminum conductor which in turn derives its supply from a larger circuit breaker. The larger breaker will usually be what is known as a moulded case circuit breaker or MCCB. It is of a much heavier construction than the MCB and is intended for interrupting higher fault currents. Some MCCBs have adjustments permitting fine adjustment of the overload and short circuit trip currents, and trip times. Distribution panels may be single phase or three phase. Most work on 3-phase + neutral supplies, so the neutral conductors for all the loads are connected to a common bus bar, and protection is only single pole in the “live” conductor. There is always a requirement to “balance” 3-phase loads, i.e. to ensure that each phase takes the same load and neutral current is minimized. However in some countries like the UK there is also a safety requirement that limits the proximity of circuits on different phases (in the home, for example).
32
Sometimes “isolators” are fitted to electrical equipment. These are physically similar to MCBs or MCCBs, but have no trip element. They do, however, isolate both the live and neutral connection. The Residual Current Device or RCD is the device used for interrupting circuits for reasons of safety. It results in the RCB and the RCBO, or Residual Current Breaker, and Residual Current Breaker with Overload Protection. The principle of the RCD is that in any circuit the current flowing in the two wires (i.e. “live” and “neutral” in an AC system) should be equal. If it is not, then it is assumed that some current has “gone missing” – maybe through someone’s body. The RCD has current sensing coils in both legs of the supply, if it detects an imbalance, it trips. The principle is shown in Figure 1.39. The current which the human body can withstand without severe physical effect or even death is small, so individual RCDs are usually set to a current of around 30mA. There is a problem about RCDs in lighting control. Most systems have some unavoidable current imbalance, often a “leakage to ground”. (RCDs used to be called earth leakage trips by some.) This is quite often to do with the suppression of RFI (radio interference), or it may have to do with signalling circuits, or even the nature of the lamp load. While it is not a problem for single circuits,
Figure 1.38 Examples of Moulded Case Circuit Breakers or MCCBs. Some of these devices allow the user to set the tripping parameters. Photo from MEM.
ELECTRICITY AND LIGHT
As an aside it has to be said that designers of power control equipment based in the 115V areas of the world sometimes under-estimate quite how “stiff” a European 230V supply can be. Conversely European designers can be mean in their provision of the high current connections needed for 115V, and sometimes don’t allow for the much wider percentage voltage swings met in practice. 1.5.3 Relays and contactors
Figure 1.39 Principle of the Residual Current Device or RCD.
since it is easy to set the RCD to a safe value while still permitting the required out-of-balance current; it does become a problem in multi-channel systems. These may require RCDs on each channel or small group of channels. If RCDs are to be used in cascade the back-up RCD must clearly have a much higher trip current than that of the subcircuit RCDs. The RCBO combines a thermal overload trip with an RCD, so it can be used for complete circuit protection. Power supplies can differ greatly, and this can have an influence on the effectiveness of protection systems especially where electronic power control devices are being used. Supplies are popularly characterized by their stiffness. A stiff supply is one with a low source impedance – typically one near the step-down transformer, and most likely to be found in large buildings with their own sub-station. With stiff supplies the prospective short circuit currents are much higher, whereas a “soggy” supply, usually involving long supply cables, cannot deliver high fault currents. Stiff supplies are to be preferred because the supply voltage does not change with load and the performance of protection equipment is more predictable. Soggy supplies can result in significant volt drop as more load is added – the “brown out” phenomenon.
In the home it is quite satisfactory to have a switch which is directly in the power line to control a lamp. But wherever there is a requirement to control high power, to provide multiple control points, or have remote control, there is a need for something different. A relay is the simplest remote switching device. Figure 1.41 shows that it consists of a switch contact which is actuated by an electromagnet. A small current through the coil causes the core to magnetize and attract the switch actuator. Removing the current causes demagnetization and the opening of the switch. The important principle is that the control signal can be completely isolated from the power it is controlling and usually, but not necessarily, is of much lower power than the power being controlled. Figure 1.42 shows some examples of relays. Some points worth noting are:
Figure 1.40 The “Memshield” MCB shown here can be converted to a Residual Current Breaker with Overload Protection or RCBO by the addition of auxiliary module.
33
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 1.41 The principle of the relay.
• relays can have AC or DC coils. Obviously the DC coil relays must have a high resistance because once the coil is magnetized the control circuit is only resistive. In an AC relay the coil is an inductive load, so the aim is the lowest practicable impedance to limit heating in the coil. • relays are not limited to single contacts. They can have multiple contacts in “Normally Open”, “Normally Closed” and “Changeover” configurations. • special relays are available for switching very low currents and high frequency signals (e.g. data signals, audio and video signals). Their construction is quite different to those used for power switching. • most DC relays use a magnetic core with low remanence to ensure quick demagnetization when current is removed. Some relays have high remanance, or are polarized with a permanent magnet, to provide bistable operation – i.e. even if the control power is removed, the relay stays operated. This usually involves the use of two coils, one to operate the relay, the other to release it. • in lighting control relays are usually being controlled by some electronic circuit. This means that they are typically operated with 24V DC coils. When the power is removed from the coil, the back e.m.f. may be sufficient to destroy the control electronics, so relay coils are usually fitted with suppression diodes. The diode (see Section 2.2) can
34
carry current only one way, so in normal operation carries no current. The current from the back e.m.f. being in the opposite direction to the main control current is then fed back into the coil where it dissipates as heat. See Figure 1.43. • AC relay coils are not so bad, but again can produce a voltage “spike” which can both create interference, and damage the control circuit. With AC the direction of the current is indeterminate, so a “snubber” RC (resistor-capacitor) network is used. An alternative is a device called a voltage dependent resistor which normally has a high resistance, but goes low resistance when it is subjected to a voltage above a certain threshold. • it is not only the relay coil which can cause voltage spikes. Clearly the switching contact will usually be producing even bigger ones. Assuming that the load being controlled is AC, then an RC Reed relay, used for switching low level control signals. 5V DC coil, 500mA 50V single pole contact.
24V DC coil, single pole 20A 230V AC contact, used for circuit isolation in a professional automatic dimmer.
24V DC coil, 1A 100V dual changeover contacts. Used for control signal isolation in show control equipment.
Plug in general purpose power relay. Twin changeover contacts 10A 230V AC. Choice of AC (24230V) or DC (12-24V) coils.
Figure 1.42 Examples of relays.
ELECTRICITY AND LIGHT
network across the contact can be used for suppression. This is good practice in any large system, and wherever sensitive electronic equipment is installed; but it has to be admitted that in many installations no such snubbing network is installed. Big relays, switching high currents are, for some reason, referred to as contactors. These usually, but not necessarily, have AC coils working at line voltage. They are often used as part of lighting control systems for primary power switching. They are also the basis of conventional electric motor control (outside the scope of this book, but often requiring three phase operation with a switch from star to delta connection of the motor windings as the motor runs up to full speed). 1.5.4 Power control Power control for lighting is reviewed in detail in Parts 3 and 4 of this book, so here a simple summary of the possibilities will suffice. Clearly any power control system must vary the current passing through the “load”. At this stage it is easiest to think only in terms of a resistive load like a tungsten filament lamp.
Figure 1.43 The suppression diode reduces the high voltage back e.m.f. from relay coils. An RC network is often used to suppress interference from relay switching contacts.
Figure 1.44 Examples of contractors suitable for lighting control from MEM Circuit Protection and Control.
Figure 1.45 summarizes the possibilities for AC. A variable impedance can be introduced into the circuit to vary the current. At its simplest this could be a variable resistance, but this is wasteful and another possibility is a variable inductance. A means of providing a variable voltage could be provided. For example an autotransformer with a sliding contact tapping. A high speed switching device could be imagined which worked so fast that it could vary the electric power simply by switching it on and off. We would not see the effect of this because the thermal inertia of the lamp filament would keep it glowing even when the power is disconnected. This method, unlikely as it may seem, is the basis of most dimming systems today. The AC waveform is “chopped” to deliver as much or as little power as needed. The load receives continuous power at full on, no power at off and 100 or 120 (depending on
35
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 1.45 Control of AC power.
where you live) bursts of power per second of a variable duration for any setting in between. An electronic power converter could be used. There might be an advantage in having a device which could not only regulate the current, but could in fact change the power regime. For example a 230V 50Hz supply on the input might be changed to a variable 0-24V 50Hz supply on the output; or a 115V 60Hz supply on the input converted to a controlled output at 100V 30kHz. Electronic power conversion is now an important part of lighting control.
or even when an old automobile passes near by. The fact that such interference is actually much less noticeable today than it was some years ago is largely due to concerted international efforts to reduce or eliminate its effects. The huge proliferation in electrical and electronic devices has led to a realization that all such equipment must be “compatible”, in that the safe and effective operation of one device must not itself lead to the unsafe or unreliable operation of another. The term electromagnetic compatibility, often abbreviated to EMC is used to describe this hopedfor harmonious relationship. The compatibility must be of various kinds: • any electrical device must not itself emit interference (in practice above a certain minimum level) • notwithstanding this, any electrical device must itself be immune to interference (again, in practice, below a certain maximum level) • the emission and immunity must apply both to conducted and radiated interference. Conducted interference is fairly easy to understand. This is any interference which comes up the power line, or comes into to equipment by virtue of a wired control connection. The description of inductance given in earlier sections shows that it can create back e.m.f.s which can result in high voltages leading to switch arcing and other undesirable effects. An “unsuppressed” inductive load can, on switch-off, create a voltage “spike” with
1.6 Electromagnetic compatibility (EMC) Most of us are familiar with the effects of electrical interference. Maybe your computer has “crashed” after a lightning strike, or when you unplugged another electrical device on the same circuit. Or you have noticed your radio or TV has a buzz or crackle when certain other electrical devices are operating,
36
Figure 1.46 The concept of the electromagnetic wave. It consists of simultaneous electric and magnetic fields exhibiting a transverse wave motion. The electric field E is orthogonal (at right angles) to the magnetic field B. Here the graph x axis is time, the y axis is the momentary value of E and the z axis the momentary value of B.
ELECTRICITY AND LIGHT
In words, the velocity of the wave motion is related to the relative permeability and the relative permittivity of the medium it is passing through (see Sections 1.2.4 and 1.2.6). Here the power -½ is the same as the reciprocal of the square root -1/√x. In the special case of free space or vacuum, the velocity is denoted by c, and: c = (μ0ε0)-½
Figure 1.47 Transverse wave motion, and the definition
a peak which might be more than double the normal line voltage peak. No wonder your computer objects. Conducted interference is not necessarily due to disturbance on the power line. Interference can be induced into control cables, even when there is no direct electrical connection. Radiated interference arises from the indiscriminate generation of radio waves. It might be wondered what radio waves have to do with switches and dimmers, but, in fact any electrical device is involved in some way or another with electromagnetic radiation – a posh name for radio waves (and other kinds of radiation). When the fundamental discoveries about electricity and magnetism were made in the nineteenth century the experimentally observed results were codified into a number of mathematical equations. Around 1865 James Clerk Maxwell compared Gauss’ theorems as applied to electrostatics (electric fields, like those found in capacitors) and magnetic fields, with Faraday’s and Lenz’s laws of electromagnetic induction and Ampère’s law for magnetomotive force. He realized that the presence of both an electric field and a magnetic field would result in a wave motion. The electromagnetic wave which he predicted was demonstrated experimentally by Heinrich Hertz about 20 years later. Maxwell showed that the velocity v of the wave motion in would be: v = (μμ0εε0)-½
Substituting the actual values of μ0 and ε0 gives value for c of around 299,790 km/second (about 186,000 miles per second) which had already been determined experimentally as the velocity of light. Maxwell realized, therefore, that his as yet undemonstrated electromagnetic waves were probably related to light. Any wave motion has the relationship: v = fλ where v is the velocity, f the frequency (number of oscillations per second) and λ is the wavelength, or the distance between successive “in phase” points on the wave – see Figure 1.47. Radio waves are generally defined as that part of the electromagnetic spectrum where the frequency runs from 1–10 12 Hz, and wavelengths from kilometers to fractions of a millimeter. AM Radio works around 1MHz, FM radio around 90MHz, UHF Television around 600MHz, and digital mobile phones in the GHz region. If lighting control is normally concerned with controlling power at 50 or 60Hz, why does this have relevance to equipment operating at much higher frequencies? Most electrical waveforms, especially those occurring in electronic control equipment, are not sinusoidal. However, Fourier showed that any complex waveform could be synthesized by adding together a number of different sine waves. A harmonic of a fundamental sine wave is one which has twice, three times or any multiple of the fundamental frequency. Figure 1.48 shows what happens when harmonics are added to a fundamental.
37
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 1.48 The harmonics of a fundamental wave motion. Only the second and third are shown, but the principle applies to any multiple of the fundamental. Complex waveforms can be analyzed into the sum of a large number of sine waveforms. Square and sawtooth waveforms are widely used as current and voltage waveforms in electronics and are rich in harmonics.
Fourier’s bad news was that when there is a steeply rising waveform, as might happen when a sine wave voltage is switched at its peak, there can be a huge number of harmonics generated. Digital computers and microprocessors depend for their operation on square waves where current is being switched on and off as fast as possible. Fourier says that a real square wave would have an infinite number of harmonics. Any electrical or electronic equipment can produce electromagnetic radiation by virtue of the fact that the components and conductors within it produce electric and magnetic fields. If the waveform is complex, the resulting radiation could be high frequency even if the fundamental frequency of operation is low (as in AC power systems). The problem is to design the equipment so that the resulting electromagnetic radiation is minimized. Many components can be designed so that stray fields are minimized. In transformer design the aim is always to keep the magnetic field within the core material. Electric fields can be contained by electrostatic shielding, which in simple terms is no more than a conducting surface, for example a metal can or a metallic conducting paint on a plastic surface. The way in which cables are constructed is
38
critical; twisted pair signal cables can be constructed so that fields from the two conductors cancel out and radiation is minimized. Most power control equipment requires RFI suppression components to be included in their design. Radio Frequency Interference suppression usually involves the use of small capacitors and inductors which have a low or high impedance at the high frequencies concerned. These can be used either to “short out” the high frequencies, for example by capacitor coupling to ground (earth) or reduce high frequency currents by inductive impedance, for example by using ferrite rings round the cable. The lumps you see in cables connecting a personal computer to peripheral equipment are an instance of this technique. Conducted interference signals can be measured in volts, so standards for limiting them, or for defining immunity to them can be laid down in terms of the voltage at different frequencies. Radiated signals can be defined both by their electric field strength and by their magnetic field strength. It is more usual to use the electric field strength E, measured in volts per meter (V/m, or its multiples, like μV/m) since the voltage developed in an antenna (aerial) is directly related to E, and can be measured
ELECTRICITY AND LIGHT
directly. However, magnetic field strength H, measured in amperes per meter (A/m) is one of the signals measured in EMC tests. The full requirements for EMC cover not only electromagnetic phenomena, but also magnetic fields, inrush currents, electrostatic discharge and other quantities. EMC is of great importance in lighting control equipment, and is referred to in the appropriate sections of this book. In particular a review of the main legal requirements is given in Section 8.5.
1.7 Light 1.7.1 The nature of light In some ways light is more mysterious than electricity. While a simple introduction to electricity allows us to use a single model or mind picture, light needs different models depending on the aspect of interest. Isaac Newton proposed that light existed as “corpuscles”. These would be radiated in straight lines by luminous bodies, and would act on the eye’s retina. Straight line optics work well for some aspects of optical design, especially when the components are large, like mirrors and lenses. So the idea was not without merit. Huygens, backed by experimental results from Young and others, demonstrated that light must have a wave nature. Observed effects like diffraction could not be explained any other way. The wavelengths were found to be very short. At the time there was a problem in describing what the waves travelled in – we can understand waves as ripples on water, or pressure waves in air, but how did the waves travel through a vacuum? It was proposed that an all pervading substance called the “ether” was the carrying medium. The wave theory received a big boost, and the problem of the ether disappeared, when James Clerk Maxwell proposed that light was a form of electromagnetic radiation as described in Section 1.6. Electromagnetic radiation is seen as having a very wide spectrum, and visible light occupies only a small part of it as shown in Figure 1.49.
Figure 1.49 The electromagnetic spectrum.
However Einstein, Planck and others showed that the photo-electric effect, where light falling on certain substances could cause an electric current to flow, could only be explained by a 20th century variation of the corpuscular theory. This new version was proposed as the quantum theory of radiation, because it requires that energy only travels in defined packets or quanta. The quantum of light was, and is, called a photon. Modern physics brings together both the quantum and wave theories by stating that every moving
39
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
particle of mass has associated with it a wave whose length is given by: wavelength = h/mu where h is Planck’s constant (6.626 × 10-34 Joule seconds) m is the mass of the particle, and u is the velocity of the particle. The equation comes with the not altogether helpful rider that you cannot actually determine all the properties of a particle simultaneously. Heisenberg’s uncertainty principle tells us that if you know where something is, then you cannot know its precise mass – or if you know its mass, you don’t know where it is. If you try and apply the equation above to everyday objects, you obtain meaningless results; but when applied on the atomic scale things begin to make sense. A dual “mind picture” or model of light is needed: • for most practical applications, the wave nature of light, expressed as part of the electromagnetic spectrum, is the simplest model. • for some aspects of the behavior of light, the quantum or photon model is needed, on the understanding that the photon itself has a defined wavelength. The photon model becomes more comprehensible when the model of the atom described in Section 1.1 is recalled. What happens if an electron jumps from one shell, or orbit, to another? The idea that such a move can only happen between one defined energy level and another fits well with the quantum concept. The idea emerges that: • when an electron drops from a high energy level to a lower one, a photon is emitted. • similarly a quantum of energy can be absorbed, and when this happens an electron moves from a lower orbit to a higher one. Sometimes referred to as moving from a normal orbit to an excited orbit. The energy of a photon is defined as: E2 - E1 = hν where: E2 is the energy associated with the excited orbit,
40
E1 is the energy associated with the normal orbit, h is Planck’s constant, ν (Greek letter nu, NOT our letter V) is the frequency of the radiation emitted as the electron moves from Level 2 to Level 1. E2 - E1 is conveniently measured in electron-volts. An electron-volt (eV) is the work required to move an electron through a potential difference of one volt. It is equal to 1.603 × 10-19 Joules. In lighting we are concerned with the wavelength of the emitted light, since this will determine whether or not the radiation is visible. Since we know that: c = νλ where c is the velocity of light (approximately 3 × 108 m/sec) λ is the wavelength, and ν the frequency; we can, by putting in the value of Planck’s constant, re-write the photon energy formula in the form: Wavelength λ in nm = 1239.76/energy in eV An example relevant to lighting is the fact that electrons surrounding mercury atoms (when in vapor form) can be made to emit photons with energy 4.89 eV. The formula shows us that this results in radiation of wavelength 254nm, well in the ultra-violet part of the spectrum. 1.7.2 Wavelength conversion Concepts which are needed to understand how the lightsources described in Part 2 work include the ideas of wavelength conversion and blackbody radiation. The energy needed to raise the energy level of electrons can be imparted in several ways. One significant way is from heat. Heat within an object is rapid motion of its atoms; the energy from this motion can change the energy levels of the outer electrons. Another way is by the application of an electric field; and yet another is by bombardment with electromagnetic radiation. An example of this last case in action is the fluorescent lamp (described in more detail in Chapter 3.) In this lamp the invisible 254 nm radiation from
ELECTRICITY AND LIGHT
example is sodium. Low pressure sodium vapor emits a characteristic monochromatic yellow light (actually two “lines” very close together at 589.0 and 589.6 nm) and this can be seen if salt is dropped on a flame, or in the low pressure sodium street lamp. Individual atoms have characteristic spectra. In order to get useful light for illumination purposes, it is necessary to have a source giving multiple lines, or, ideally, a continous spectrum. 1.7.3 Blackbody radiation Figure 1.50 Concept of transitions between different energy states. On the left, transitions between normal and different excited states. On the right, transitions between different excited states only.
mercury is directed at a phosphor, which is excited by the radiation. The excited electrons in the phosphor then drop back to a lower level; but the drop is not as great as the 4.89 eV of the original photon. It is more like 2.3 eV, resulting in visible radiation around 545nm. (Note that this is just an example – phosphors emit a band of radiation over a range of wavelengths.) This is an example of “wavelength conversion” whereby an efficient method of producing an invisible electromagnetic radiation (in this case UV) is used to stimulate the production of visible radiation, via, in this case, a phosphor. The outer electrons of atoms can have several excited states. Figure 1.50 shows in simplifed form what can happen. Electrons can be raised from their normal state to one of several excited states. But in addition there can be transitions between excited states. The result of this is that any atom has a characteristic absorption spectrum and emission spectrum. Each transition, whether between different excited levels, or between an excited level and normal level, is associated with a given frequency and corresponding wavelength (approximately 254 nm in the case of the mercury example already given). In lighting our interest is in those transitions which result in visible radiation. Reference is often made to spectral lines. This is because, when light from a source is analyzed using a spectroscope, it is seen as a series of colored lines. A well known
A metal heated to a high temperature glows; it emits visible radiation. A blackbody radiator is one whose intensity and spectral properties are dependent solely on its temperature. In fact the blackbody radiator is also a blackbody absorber – in that it absorbs all incident radiation of all wavelengths. It does not reflect any incident radiation. The blackbody is an ideal concept; it radiates more total power and more power at any given wavelength than any other lightsource operating at the same temperature. Figure 1.51 shows blackbody radiation curves for a variety of temperatures. From the curves it can be seen that a blackbody needs to be at a temperature of above 2,000K to give out useful quantities of visible radiation. Its important attribute is that the output spectrum is continuous. Note that for any given temperature there is a radiation peak at a particular wavelength, which gets shorter as the temperature rises. Also, as the temperature rises, the total amount of radiation increases. Real sources of radiation can never achieve the performance of a blackbody; but the blackbody is used as a performance reference. The spectral emissivity of a radiator is a wavelength dependent quantity which is the ratio of the output of the radiator at a particular wavelength, to the output of a blackbody at the same wavelength. Figure 1.52 shows the comparative performance of the perfect blackbody and tungsten (used in filament lamps) at 3,000K. The color temperature of a radiator (for example a tungsten filament) is the temperature at which a blackbody would have to be for its output to match
41
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
control, and for which it is useful to have a basic understanding. While the blackbody absorbs all radiation falling on it, all real surfaces reflect some or most incident radiation. Some definitions are: Specular reflection is where light is reflected in the manner of a mirror. Specular reflectance is the proportion of light reflected expressed as a percentage of the incident light. Specular reflection obeys the Laws of Reflection illustrated in Figure 1.53. Specular reflectance can range from around 7% for a normal sheet of glass (which transmits most light) to 95% for a surface aluminized mirror. Polished stainless steel and chromium plate have reflectances of around 60%. Diffuse reflection arises when the surface is rough. In this case the light is scattered in many directions. A special case is the perfect diffuse reflector which reflects any incident light equally in all directions. A matt white painted surface is an example of a near perfect diffuse reflector. Most
Figure 1.51 Blackbody radiation curves at different temperatures.
as closely as possible that of the chosen radiator. The concept works well for any continuous spectrum source, for example an incandescent filament or sunlight, but breaks down when applied to sources with discontinuous spectra. Examples of color temperatures are: Typical tungsten filament lamp: Noon sunlight: Cloudy sky; average daylight:
2,850K 4,900K 6,700K
1.7.4 The behavior of light The subject of optics is beyond the scope of this book, but nonetheless there are aspects of the behavior of light which are of importance in lighting
42
Figure 1.52 Comparison of the radiation from a blackbody at 3,000K, and the corresponding performance of tungsten.
ELECTRICITY AND LIGHT
surfaces exhibit a combination of diffuse and specular reflection. Refraction describes what happens when light passes from one transparent medium to another; for example when going from air into or out of glass, or from either into or out of water. Figure 1.54 shows light passing through a transparent glass prism. Light which does not arrive normal (at right angles) to the surface is bent or refracted. The following apply: • The angle of incidence (i in the figure) is the angle which the incoming ray of light makes with normal; similarly the angle of refraction (r) is the refracted ray in the new medium. Assuming the ray is going from a less dense medium to a denser one, r is less than i. • Refractive index μ is defined by
μ = Sin i / Sin r • The refractive index applies to different pairs of media, but the refractive index vacuum/glass is almost exactly the same as that of air/glass; so in practice it is common to refer to the refractive index of a material referred to air. • The wave theory of light shows that in fact the refractive index is the ratio of the velocity of light in Figure 1.54 Light being refracted through a prism (a). The dispersion of light into its different colors by a prism (b). The critical angle at the glass/air surface within a prism (c).
Figure 1.53 The Laws of Specular Reflection (Snell’s Laws). The incident ray, reflected ray and the line at right angles (normal) to the surface are all in the same plane; and the angle of incidence equals the angle of reflection.
a material to the velocity of light in a vacuum (or air, where the slowing down of light is negligible). • Within most materials there is a slight difference in the velocity of light for different wavelengths. This is why a prism can be used to split light into its different wavelengths or colors, because the bending effect is different for each color. • Transparent materials have a critical angle. From Figure 1.54 it can be seen that the emerging ray diverges from normal (i.e. gets closer to the prism surface). The internal ray can finally be at an angle (the critical angle) where theoretically the emerging ray would correspond to the plane of the glass surface. Beyond this angle the light is internally
43
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
reflected, with the exit glass/air surface acting as a mirror. In binoculars and camera viewfinders prisms are used to reflect light. Glass and certain plastic materials such as acrylic and polycarbonate sheets are made to be transparent; but may also be made translucent. Here either surface treatment or the nature of the material itself causes multiple scattering of the light. In luminaire manufacture such material is used as a diffuser to increase the apparent size of a light source (and reduce its point brightness). Typical luminaires make use of reflection, refraction and diffusion in order to make the best use of the light from a particular source. Figure 1.55 shows examples. The wave theory of light predicts the possibility of interference. If two wave motions of identical wavelength meet, then their individual amplitudes can be added together. This means that if the wave motions are in phase they reinforce each other, but if they are out of phase, they cancel each other out.
ceiling tile luminaire with prismatic panel (refraction)
ceiling luminaire using both reflection (center) and diffusion (either side round tubes)
luminaire using aluminum reflectors
An anti-reflection coating is a practical manifestation of interference. Such coatings are used to minimize specular reflections in showcase glass, or to increase light transmission through lenses. The principle is shown in Figure 1.56. This shows the specular reflection off the first surface being cancelled out by a second wave exactly half a wavelength out of phase. The second wave has been refracted into the thin film, reflected off its other internal surface, and refracted out again – the film thickness being such that by the time it emerges, it is out of phase with the initial reflection. A thin film of different thickness could augment the reflection instead of destroying it. Multi-layer coatings are used to improve transmittance or reflectance. For example a 4-layer coating on a surface coated mirror (designed to increase reflectance) can raise reflectance above 99%; conversely triple coated lenses for cameras have a reflectance of less than 1%. The same technique can be used to create interference filters. By using multiple thin films (each layer of a different refractive index, and different thickness) it is possible to create a filter. A high pass optical filter passes high frequency (short wavelength) light, and reflects long wavelength light. A low pass optical filter does the opposite. In the lighting field such filters are important. They make available hot and cold mirrors, as found in some reflector spotlights and projection devices. A cold mirror reflects the full visible spectrum, but transmits infra-red radiation, which in the example spotlight application, would go out the back of the reflector instead of on to the illuminated object. Dichroic filters are efficient filters with a very sharp transition between transmission and reflection • for example having a transmittance band of 400– 500nm blue, and reflectance band 500–700nm yellow.
1.8 The eye, how we see light Figure 1.55 A luminaire using a fluorescent lamp may well make use of a combination of reflection, refraction and diffusion to make the most effective use of the light from the tube.
44
The human eye is an incredible device – but the eye on its own is only a light collecting mechanism, it requires the brain to make sense of what is received.
ELECTRICITY AND LIGHT
Figure 1.56 Principle of the anti-reflection coating.
Figure 1.57 is a much simplified diagram of the eye, and identifies: • the cornea, the transparent outer layer which, with the tear layer, protects the eye, but also provides the first refractive element of the eye. • the iris, which regulates the amount of light entering the eye, by changing diameter. It does this in the range of about 2-8mm, representing a range of 16:1. This alone is not sufficient to cover the huge range of illumination encountered. Our eyes also have to “adapt” to the lighting condition, but this takes time, whereas the iris acts instantaneously. • the lens, which focuses light onto the retina. By use of the ciliary muscle the lens can change focal length, depending on whether it is looking at a near or distant object. • the retina, which receives the imaged light. It contains the receptors for light and produces the electrical signals for the brain to process. • the fovea, more or less at the center of the retina. This is the part of the eye by which we see most detail, but represents a very small part of our overall field of view. • the optic nerve. This communicates the information received by the retina to the brain. Because it occupies a small part of the retina, we have a “blind spot” at this point.
The retina is found to have two different kinds of light receptors, identified as rods and cones. Rods provide us with scotopic or night vision, and are very sensitive to light. There are no rods in the fovea. All rods contain the same photopigment, so they see in “monochrome”. Photopic or day vision is done by the cones. Of six million of these, more than one million are located at the center of the fovea, representing only a 1° field of view. All these have an individual connection to the brain. The remainder are spread across the retina, and, like the rods, report to the brain in groups. The cones are in three classes, each with a different photopigment. Figure 1.58(a) shows the spectral response of the eye as a whole, and here it is clear that the scotopic and photopic responses peak at different wavelengths. Figure 1.58(b) shows that for the three different kinds of cone there is a significantly different response. It seems that, between them, the cones deliver to the brain achromatic (non-color) information and separate hue information (not unlike a color TV).
1.9 Measurement of light In the same way that we needed objective methods of measuring electrical quantities, there is a need
Figure 1.57 The eye.
45
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
(a)
(b)
Figure 1.58 Spectral response of the eye.
for the objective measurement of light. In fact the measurements can still be based on the MKS system, but it is necessary to introduce one more fundamental unit, and the SI system uses the Candela as that unit. A major problem about measuring light is the practical necessity of relating the measurement to
46
human vision. As shown in Figure 1.58 the human eye has a different response to different wavelengths of light. By experiment it was found that the wavelength of maximum luminous efficiency, that is the wavelength which produces the maximum response in the eye for the minimum power input, is approximately 555nm. At all other wavelengths our eyes have a lower photopic spectral luminous efficiency V(λ), where λ is the wavelength. If V(λ) is unity (1) at 555nm, then it is found to be about 0.5 at 510 and 610nm, and as little as 0.00006 at 385 and 760nm. The Candela, the SI unit of luminous intensity is today linked to the human eye’s peak response, and is tied back to the power radiated in Watts; its official definition is given in Table 1.6. (Note that Table 1.6 also gives the definitions of total power radiated, independent of the eye’s response.) Our appreciation of the brightness of a light source is defined by its luminance, expressed in Cd/ m2. The definition is given in Table 1.6 – the bit about orthogonal projection is telling us that you get a different brightness if you view the source at an angle. Luminance figures are very important in lighting design and in the specification of electronic displays. Table 1.7 gives some examples of luminance. The lumen is the unit of luminous flux. This is a unit which many lay people have heard of (especially from salesmen who probably have no idea what it means) but which in fact is one of the more difficult units to understand fully. At 555nm, the lumen is fairly easy – since it is tied directly to the candela (see the definition in Table 1.6) – a one candela point source emits 4π lumens. At 555nm it is easy to calculate the number of lumens which can be derived from a given power (or Watt) input. But what happens at other wavelengths? The efficacy of a light source is defined in two ways. The luminous efficacy of radiant flux is measured in Lumens/Watt, and is the ratio of the luminous flux in lumens to the total radiant flux in Watts. More useful in the lighting business is the luminous efficacy of a source of light, also measured in Lumens/Watt, and the ratio of the luminous flux in lumens to the total lamp power input.
ELECTRICITY AND LIGHT
Quantity and symbol Luminous Intensity I
Unit
Luminance (or 'Photometric brightness') L Luminous Flux F
Candela per square meter (also known as the 'nit') Lumen
Cd/m2
Illuminance E (also Illumination) Radiant Intensity Ie Radiance
Lux
lx
Watts per steradian Watts per steradian per square meter Watts
W/sr
Radiant Flux )e Irradiance E
Candela
Watts per square meter
Unit Symbol Cd
lm
W/(sr.m2) W W/m2
Defined as The luminous intensity of a 555.016nm (or 540 u 1012 Hz) source which has a radiant intensity in a given direction of 1/683 Watts per steradian, when measured in that direction. (Formerly defined as 1/60 of the intensity of a square centimeter of a blackbody at the temperature of solidification of platinum.) The intensity of a source in a given direction, divided by its orthogonally projected area in that direction. An isotropic (one which emits radiation equally in all directions) point source of intensity one Candela produces a total luminous flux of 4S lumens. But see also main text which describes how, when related to input power, lumens are wavelength dependent. The concentration of luminous flux falling on a surface. One Lux is is one Lumen per square meter. Radiant power emitted by a point source in a given direction. The radiant intensity of a source in a given direction, divided by its orthogonally projected area in that direction. Radiant power of a source at all wavelengths Radiant power incident on a surface. (Compare Illuminance)
Table 1.6 The measurement of light. Radiant intensity, radiant flux, radiance and irradiance are physical units based on power and apply to all wavelengths. Luminous intensity, luminous flux, luminance and illuminance are luminous units, which are wavelength dependent and based on the human eye’s response. They are tied back to physical units through the Candela.
Everyday examples of the latter are around 14 lm/W for an ordinary tungsten lamp and 85 lm/W for a fluorescent lamp. To derive the luminous efficacy of a source which is emitting many wavelengths, it is necessary to add the contribution at each wavelength. For photopic vision the formula is:
770
Lumens = 683 Σ PλVλΔλ 380
where the Σ sign is telling us to “sum for all wavelengths between 380 and 770nm”; and:
47
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Source
Ft. lambert
Nit
482,000,000 Sun at meridian 1,750,000 Sun at horizon 600 – 2,300 Sky, typical 2,600 Fluorescent lamp 150 TV screen (CRT) 14-18 Motion picture screen
1,700,000,000 6,000,000 2,000 – 7,800 8,800 500 48-62
Table 1.7 Examples of luminance.
Pλ is the spectral power in watts of the source at wavelength λ. Vλ is the photopic luminous efficiency at wavelength λ. Δλ is the interval over which the spectral power is measured. To be really thorough these “photopic” lumens should be added to a separate calculation for “scotopic” lumens. In this case the initial multiplier is 1,700, the wavelength range is 300 to 770nm, and there is a quite different set of values for scotopic luminous efficiency. In practice scotopic vision is only of importance at very low light levels, so is not usually considered when discussing electric lighting. There is a maximum theoretical efficacy for an “ideal” white light source. This is a source which radiates light at constant power over the complete visible part of the spectrum, but does not radiate power outside it. The maximum efficacy is approximately 220 lm/W. An important measurement in lighting design is illuminance – this is a measure of how much light is falling on a surface per unit area. Today the accepted measure is the Lux or Lumen per square meter. However, some industries still use measurements based on imperial units where the illuminance is measured in Footcandles (or lumens per square foot) and luminance is measured in FootLamberts (where a surface emitting one lumen per square foot is said to have a luminance of one FtL). Table 1.8 shows the relationship between the imperial and metric units. Figure 1.59 may help visualize the units. It shows a sphere of radius r, where we know that the surface area of the sphere is
48
4πr2. At the center of the sphere is a point source of one candela, which we know is emitting 4π lumens. Let us assume the inside of the sphere is totally absorbing, so the inside surface is ONLY illuminated by the point source, and not by any cross-reflection. If the radius of the sphere is one meter, then the flux falling on one square meter of its surface (= r2) must be one lumen, and the illuminance on that square meter one lumen/square meter or one lux. Another way of looking at it is that the flux density is 1 lm per steradian, which applies whatever the radius of the sphere. Thus if the sphere radius is only one foot, the illuminance on the inside surface will be one lumen per square foot or one footcandle. It would be very tedious if, every time we wanted to measure luminous flux or illuminance, we had to do it at many different wavelengths. Fortunately this is not necessary. Most light measurements are now made using solid state detectors based on silicon. The principle is that the absorption of a photon creates a free electron to contribute to an electric current. By having a detector of a known area, it is possible to measure the incident illuminance. Detectors respond to radiation in the visible spectrum (and, if required, either side of it). The detector is usually “corrected” so that its response closely matches the Photopic Luminous Efficiency characteristic of the eye. This means that the complete instrument or Photometer can directly read out in Lux, see example in Figure 1.60.
Quantity Luminance Illuminance To convert Nit to Ft Lambert Ft Lambert to Nit Lux to Ft Candle Ft Candle to Lux
Metric
Imperial
Nit or Cd/m2 Lux or lm/m2
Ft Lambert Ft Candle or lm/ft2 Multiply by 0.2919 3.426 0.0929 10.76
Table 1.8 Relation between metric and imperial units.
ELECTRICITY AND LIGHT
Figure 1.59 Derivation of units of illuminance.
A Luminance Photometer is similar, but it includes an optical system, like a telescope, which focuses a source onto the detector. The Commission Internationale de l’Éclairage or CIE is the body which lays down how lighting measurements are made. Their publications not only define how instruments are calibrated with reference to the SI standards, they also include the agreed luminous efficiency curves. One problem characteristic of illumination is glare. This does not lend itself to easy measurement because its effects are subjective. In practice it is a significant factor in lighting design, so various efforts have been made to codify its effect. We see things not only by their brightness, but also by their contrast relative to their surroundings. We can easily see black lettering on a white background, but making out white lettering on a white background is difficult if not impossible. Generally the higher the illumination, the better we see things; but if luminance levels vary greatly, for example when driving in the dark and having to contend with oncoming headlights, a high luminance relative to the surroundings causes our perception to break down. Discomfort glare is when the lighting situation is uncomfortable, disability glare is when the lighting situation means that it has become impossible to do
the intended task. Usually both apply at once. For example low winter sun streaming into an office can be both uncomfortable and make carrying out a normal office task impossible. Glare can arise in all lighting schemes. A person sitting in a low ceilinged office may have a well illuminated desk from an overhead luminaire, but also be distracted by the glare from an overhead luminaire some distance in front of him. Specular reflections in VDU screens arising from sources of relatively high brightness can make the screens impossible to use. Various methods are used to quantify the effects of glare, to assist lighting designers eliminate its effects. They are all related to defining the highest levels of luminance for a given source size, which can be tolerated within a given level of illumination for a particular activity. The USA use the Visual Comfort Probability (VCP) system, the UK the Glare Index system and the CIE has introduced the Unified Glare Rating (UGR). Glare is relevant in lighting control, because its effects can sometimes negate the intended results of a control scheme.
Figure 1.60 Example of a photometer used for making illuminance measurements.
49
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
1.10 Color 1.10.1 Introduction Sections 1.8 and 1.9 above have warned us that the eye does not work in an obvious way. Different wavelengths of light are seen with a different sensitivity; and the eye sees the visible spectrum with three different receptors. Worse still, our perception of color varies according to the field of view (i.e. the angle subtended at our eye by the object.) This means that any measurement of color must take into account how the eye works, and cannot be based on simple power spectrum measurements alone. In practice the situation is made much worse because the apparent color of an object will depend not only on its own properties, but on the spectral distribution of the source. As an extreme example an object which is seen as blue in daylight looks black or gray when illuminated by a low pressure sodium lamp (with its monochromatic yellow light). This section describes some color phenomena, and gives definitions of color attributes. Section 1.11 describes how color is measured. 1.10.2 Primary and secondary colors The way the eye/brain combination sees color has the interesting result that the simultaneous presentation of two colors makes the brain think it is seeing a third. For example if separate red and green spotlights are both directed at the same surface, the eye sees the surface as yellow. Figure 1.61 shows how three primary colors, red, green and blue can be mixed: red + green = yellow green + blue = cyan blue + red = magenta red + green + blue = white The two-color combinations so derived are called the secondary colors, they each represent “white minus a primary color”:
50
yellow = white - blue cyan = white - red magenta = white - green This phenomenon is the basis of all aspects of color manipulation. Additive color mixing is where primary colors are mixed to achieve a desired color. There are many everyday examples related to lighting and display: • color television depends wholly on additive color mixing. The TV screen consists of a large number of small dots which, at a normal viewing distance, the eye integrates into a single image. The dots are in red, green and blue triads. Each small zone on the screen has its own color value determined by the mix of red, green and blue. • a stage backcloth can be illuminated by cyclorama floodlights in sets of three, each set red, green and blue. By setting different proportions of each, any color (within the gamut of the primaries) can be achieved. • lamps can be “tuned” to have an appropriate output. Thus, for example, a fluorescent lamp might be deemed to be too “cold”. This can be corrected by adding a phosphor which emits in the red part of the spectrum. The red will not be seen separately, but only as a “warming” of the white. Color mixing can also be subtractive. In additive mixing, light of different colors is added (to “black”) with the result that the perceived light is brighter than its constituents. Subtractive color mixing works the other way, it takes light away from white, so the result is darker than its constituents. The mixing of color paints, and the color printing process are examples of subtractive color mixing. The use of color filters is another subtractive process. A theatre spotlight with a tungsten filament lamp gives “white” light; a color filter removes all colors other than the wanted color. A problem can arise when filters are applied to light sources with uneven spectral output. Putting a red filter round a fluorescent tube may only go to show how little red there is in the light it is emitting.
ELECTRICITY AND LIGHT
Standards, the Optical Society of America, British Standards and DIN have developed alternative systems to Munsell, but use similar principles. This type of system is practical for many real world tasks; but does not relate color directly to physically measurable quantities. However, it is possible to relate Munsell (and similar) color descriptions to their equivalent CIE Chromaticity value, described in Section 1.11. Figure 1.61 Primary and secondary colors.
1.10.4 Color temperature
1.10.3 Aspects of color, the Munsell system
Color temperature was defined in Section 1.7.3. Strictly the term applies only to light sources with continuous spectra, and which lie on the blackbody radiation curve or planckian locus described in Section 1.11 below. However it is convenient to use it for other sources, and in this case it is referred to as the correlated color temperature or CCT. The CCT is the temperature of a blackbody whose chromaticity (defined in Section 1.11 below) most closely matches that of the source concerned. Computer programs and graphs are available which give “isotemperature” lines on the chromaticity chart, so if the chromaticity of the source is known, it is possible to derive its CCT. However, the further away from the blackbody locus the source lies, the less meaningful the derived CCT. Lamp manufacturers refer to lamps being “warm” or “cold” according to their CCT. For example:
An early attempt to quantify color was made in 1915 by the American teacher, Albert Munsell. It is still in use because in practice it provides an easy method for architects and designers to compare colors. In the Munsell system, color is defined by three attributes: Hue: this is simply defining the perceived color as being near to a principal color like red, green, blue, yellow or to a mixture of any two. The Munsell system defines five principal and five intermediate hues, but allows 100 hue steps in all. Value (also brightness): This expresses whether the surface is reflecting (emitting) more or less light. It can also be considered as the “whiteness” of the color, and is measured on a scale of 0 (black) to 10 (white). Chroma (or Saturation): is a measure of the “colorfulness” or intensity of the color. Munsell allows 20 (or more) steps from neutral through to highly saturated. The Munsell system is represented both by a three-dimensional “color atlas”, and by a collection of carefully standardized color chips. The ideas behind the Munsell system are illustrated by Figure 1.62, but note that this is for illustrative purposes only and cannot show the actual colors. Color matching using Munsell color chips can only be done with a standardized source (preferably daylight) but, of course, such chips can also be used to judge the performance of different light sources. Other institutions, including The Inter-Society Color Council working with the National Bureau of
CCT less than 3,300K is “warm” (example, tungsten lamp) CCT 3,300–5,300K is “intermediate” (example, cool white fluorescent lamp) CCT greater than 5,300K is “cold” (example “daylight” fluorescent and some metal halide lamps) The effect of color temperature is quite dramatic. If you are in a room lit by tungsten lighting, a small fluorescent fixture fitted with a “daylight” lamp looks very blue; conversely, if you are in a daylit space, an ordinary lamp looks somewhat orange. The effect is particularly notable in photography and television. A film intended for outdoor use is usually “balanced”
51
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
52
MIRED filter corresponds to a change of 1,800K. But it is easy to work out that a +100 MIRED filter applied to a 10,000K source, makes a 5,000K change in color temperature.
1/
2/
4/
Chroma 6/ 8/
10/ 12/
14/
9/ 8/ 7/ 6/ Value
for 5,500K; that is to say that it will give faithful reproduction of colors if the scene is, in fact, illuminated by daylight, or its artificial equivalent with the same color temperature. If the film is used on a studio set illuminated with tungsten lighting at 2,800–3,200K, everything comes out very orange. While films balanced for studio tungsten lighting (at around 3,200K) are available, it is often the case that film and TV lighting cameramen must make use of filters to modify the color temperature of studio luminaires. Figure 1.51 showed the characteristics of blackbody radiation. Mathematically the graphs illustrate: • the Stefan–Boltzmann Law which states that the amount of radiation emitted by a blackbody is proportional to the fourth power of the temperature (so the amount of radiation increases enormously with temperature). • the Wien Displacement Law which states that the wavelength of peak emission is inversely proportional to the temperature. The latter leads to the observation that perceived equal changes in color are more accurately expressed by a measure of the reciprocal of the color temperature (i.e. 1/T). In order to have manageable units, the reciprocal megakelvin is used. This is simply 1,000,000 divided by the color temperature. The same unit is also known as the micro reciprocal degree or MIRED. Thus, for example, a source of 4,000K has a MIRED value of 250. If you just specified a color temperature correction filter as, for example, “blue” to correct a low temperature source, you would not be able to predict easily what its effect would be, since it would be quite different on each source. But if you specify a filter by its MIRED value, it is easy to calculate exactly what its effect will be – regardless of the color temperature of the source to be modified. As an example; suppose the available source is 5,000K (=200 MIRED) and it is required to apply a filter to convert to 3,200K (=312 MIRED), the filter has to be +112 MIRED. A filter to go from 3,200K to 5,000K would correspondingly, have to be a -112 MIRED filter. For this particular color temperature change, a 112
5/ 4/ 3/ 2/ 1/
Figure 1.62 The Munsell system of color classification. The diagram shows the main hue classification and the way in which value and chroma apply to a single hue.
ELECTRICITY AND LIGHT
CIE Color Rendering Index Ra >90 80001090 60001080 40001060
20001040
Typical Application
intervals, in practice 5nm. The most commonly used sources are:
Where accurate color matching is required; e.g. print inspection Where good color judgement is required, or where appearance is important; e.g. retail display Where moderate color rendering is required; e.g. commercial premises Where color rendering is not important but where a marked distortion of color is unacceptable; e.g. warehouses Where color rendering is of no importance, and a marked distortion of color is acceptable; e.g. some road lighting
Standard Illuminant A: Representing tungsten at a color temperature of 2,856K. Standard Illuminant B: Representing noon sunlight at a color temperature of around 4,870K (now superseded by Standard Illuminant D50) Standard Illuminant C: Representing average daylight (cloudy sky) with an average color temperature of 6,700K (now superseded by Standard Illuminant D65)
Table 1.9 The CIE Color Rendering Index.
1.10.5 Color rendering There is a need for different kinds of lamp to be classified according to their color rendering abilities. Clearly the monochromatic low pressure sodium yellow lamp has non-existent color rendering ability; and a tungsten filament lamp has quite good color rendering. The most common method of classifying color rendering is that of the CIE, the Color Rendering Index or CRI, designated Ra. The method uses sample surfaces of eight or fourteen Munsell test colors (described in 1.10.3 above.) Each surface is illuminated in turn by a standard reference source, and then by the source under test. At each stage the spectral reflectance (i.e. the reflectance at different wavelengths) of the test color is measured. The reference source is deemed to have a CRI of 100, and the test source is rated against it for each part of the spectrum. The problem here is, clearly, “what is the reference source?” The CIE have defined several in terms of their spectral power distribution. To be of any use in practice this has to be done to quite fine
Spectral power distribution tables are also available for other “daylight” illuminants, such as D55 and D75. Table 1.9 gives examples of CRI. Clearly the fact that different reference sources can be used to determine CRI means that a user must know which reference has been used.
1.11 Measurement of color 1.11.1 CIE Chromaticity Diagram The achievement of a desired color by mixing primary or secondary colors leads to the idea that any color can be described numerically as three tristimulus values. However, simply relating color to, for example, three spectral lines corresponding to optimized red, green and blue is not sufficient. The reasons can be summarized as: • such an arrangement cannot easily describe spectral colors (other than the three primary colors chosen). This is intuitive, since additive color mixing is always “lightening” the illuminated object. A simple analysis shows that any system based on this idea would need “negative” color values; which might be an acceptable mathematical trick (and can, in fact, be used in electronic analogs of color) but is not satisfactory for real world work. • the gamut of colors which can be described by three “real” primaries is limited. This can be seen by drawing an imaginary color triangle as in Figure 1.63. Points R G B represent the primaries, and any
53
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
point within the triangle represents colors which can be produced by mixing them. Point W in the middle represents the white light (equal energy of each of Red, Green and Blue). Colors which might exist outside the triangle (including pure spectral colors) cannot either be created by the chosen primaries, nor be described purely with reference to them. The CIE got round this problem with their Chromaticity Diagram, originally produced in 1931. The idea here is to postulate three imaginary primary colors denoted X, Y and Z. Any color is then defined by the proportions it contains of each, the values for any color being denoted by the chromaticity coordinates, designated x, y and z. Because it is always the case that: x+y+z=1 any color is, in fact, defined by its x and y coordinates on the diagram. Figure 1.64 shows the basis of the diagram. It starts with the spectrum locus which is the horseshoe shaped curve. On this curve appear the pure spectral colors, and the principal wavelengths are identified in the figure. The line across the bottom of the horseshoe represents the magentas and purples (which are not spectral colors, but mixtures of blue and red). The outer lines of the diagram represent saturated colors.
Figure 1.63 Simple color triangle
54
Figure 1.64 The CIE chromaticity diagram showing the spectrum locus, and how an individual color is defined.
To help understand the diagram, it is shown in various forms. Figure 1.65 shows it in color. Here it can be seen that hue is dependent on the position relative to the spectrum locus, and saturation is dependent on how near the edge of the diagram the color is. Superimposed on it are two color triangles. Points R, G and B correspond approximately to the phosphors used in cathode ray tubes (CRTs) to produce a color TV picture. The triangle RGB defines the limit of the colors which can be reproduced. As a matter of interest earlier CRTs used more widely spaced primaries, theoretically giving a wider color gamut. However, today’s phosphors have been chosen as the best compromise in respect of overall image brightness and color range. Points R’, G’ and B’ correspond approximately to the triangle available in a theatrical luminaire used for color wash effects. In this case the color filters used are dichroic. Figure 1.66 shows the diagram with the planckian, or blackbody, locus superimposed. This line represents the appearance of a blackbody at increasing color temperatures. It clearly shows how
ELECTRICITY AND LIGHT
Figure 1.65 The CIE chromaticity diagram in color. Obviously this is an approximation, limited by the printing process. See main text for description of the triangles.
Figure 1.67 Some examples of isotemperature lines shown on the CIE chromaticity diagram. The number against them is in MIRED or recripocal MK.
“red” tungsten filament lighting is when compared with other “hotter” sources. Figure 1.67 shows examples of isotemperature lines, used to derive correlated color temperatures (see Section 1.10.4). 1.11.2 Luminance, metamers, color spaces It would be correctly argued that the introduction of the CIE chromaticity diagram in 1.11.1 has oversimplified matters. The diagram as it stands is twodimensional. While it correctly deals with hue and saturation, it does not properly deal with brightness or luminance. A more rigorous understanding of tristimulus values is needed for this. The aim is to define the light from a source, or from a reflective surface, with tristimulus values which uniquely describe its strength and color. The chromaticity co-ordinates x, y, and z are defined by:
Figure 1.66 The CIE chromaticity diagram showing the planckian, or blackbody, locus.
x = X/(X + Y + Z) y = Y/(X + Y + Z) z = Z/(X + Y + Z)
55
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 1.68 The CIE primary color matching functions. The figure is an approximation. CIE issue two different sets or figures depending on field of view. Each set is given for a range of standard illuminants. Note here that the colors used to show x , y and z are simply to distinguish them in the figure.
X, Y and Z are the amounts of the imaginary primary colors needed to match a specified color. They are defined by equations which in words can be summed up, as: X is the sum for all wavelengths between 380nm and 780nm of the measured spectral energy, multiplied by the spectral tristimulus value x. Similarly for Y and Z. The values of, x, y and z for different wavelengths are shown graphically in Figure 1.68 assuming an equal power spectrum (i.e. the maximum power at all wavelengths is equal). The CIE issues a table
56
giving precise values at 5nm intervals. In practical applications (for example in the film studio) X, Y and Z can be measured by portable instruments like the one shown in Figure 1.69. This does the necessary calculations to show the results in various ways in addition to tristimulus values – for example the x and y co-ordinates on the Chromaticity Diagram, with a separate reading for brightness. Color matching is bedevilled by the fact that it depends not only on the source, but also on the characteristics of the object being illuminated. Two visually indistinguishable lightsources with the same spectral compositions are said to have a spectral match. But two observed lights may appear to be visually indistinguishable despite having quite different spectral compositions. Such a color match is called metameric, and is conditional. For example two items illuminated by a tungsten lamp may appear to have an identical color; however, when illuminated by daylight they appear to have different colors by virtue of their different spectral reflectances. This phenomenon is of great importance in lighting design. For example, a stage or interior lighting designer can only do his or her job with the actual dress or furnishing fabrics to be used. “Rehearsal” dresses might look quite different in color terms under the final intended lighting, even
Figure 1.69 A portable tri-stimulus color meter from Minolta. This one shows luminance, the x and y color coordinates, and also temperature.
ELECTRICITY AND LIGHT
Figure 1.70 Adding a third axis to the chromaticity diagram allows luminance to be included.
if the “rehearsal” and “real” fabrics looked the same under daylight. The y function in Figure 1.68 corresponds to the photopic response of the eye. Thus the Y value is a measure of the brightness (luminance) of the source or illuminated object. A third axis can be added to the chromaticity diagram as shown in Figure 1.70. Here lighter colors appear above the point representing their chromaticity at a height representing their lightness. The Y axis “touches down” at the point representing the chromaticity coordinates of the illuminant. Figure 1.70 illustrates the concept of color space, a three dimensional representation of both color and luminance. Its problem is that it does not correlate all that well with perceived color differences, in that regions of apparent small color difference use up a lot of space, and some areas of major perceived difference are all “bunched up”. To get over this, in 1976 the CIE introduced other color spaces known as CIELAB and CIELUV. In both systems L* (lightness index) and either the a*,b* or u*,v* chromaticness indices (the way in which the perceived color appears more or less chromatic) are related to X, Y and Z by a set of equations. In industry the two systems co-exist because CIELAB has been found most suitable for
describing the color of objects, while CIELUV is used for emissive devices like TV screens. Figure 1.71 shows in diagramatic form the CIELAB color space. Notice that black now has a defined place on the map, at the origin. The L* , a* and b* axes can be considered as representing “white”, “blue-yellow” and “green-red” respectively. The alternative CIELUV is similar (L* is the same for both) but uses a different set of transform equations. In many practical systems (especially those concerned with electronic display and printing) somewhat simpler color models are used. These are normalized so that, for example, measurement of red, green and blue primaries (for display) or cyan, magenta, yellow secondaries (for printing) is weighted. This is so that each can be scaled from 0–1 or 0–100%, yet still sensibly yield a full color gamut on a simple incremental basis. The weighting is tied back to the CIE color spaces. Color, and the many different attempts to describe it, is a complex subject. This brief review can do no more than introduce it, and give a first insight as to why lighting in practice often gives surprising results. The reading list includes books that give more comprehensive coverage.
Figure 1.71 The CIELAB color space. The color of the point P is described by three quantities. Lightness L* measured on the vertical axis; Chroma C*ab measured as a*, b* co-ordinates, and Hue, measured as the angle h.
57
Chapter 2
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Lighting 2.1 Electronic principles 2.1.1 Introduction In this chapter the intention is to give sufficient background information to understand the basics of electronics as applied to lighting control. We build on the ideas of Chapter 1. Electronics differ from electrical engineering mainly in the power levels and frequencies involved. Traditional electrical engineering and electromechanical systems involve items such as motors, generators, switches, relays, transformers etc. all operating at something similar to mains frequency alternating current. Electronic engineering, on the other hand, is normally characterized by having no moving parts (other than items such as isolating switches) although sometimes the electronic device is itself connected to a “traditional” electromechanical unit such as a relay or motor. Most electronics operate at low voltages, and with comparatively low currents. They may also operate at high frequencies. For example a microprocessor may work from a 3.3V supply, consume only a few milliamps, and operate at a frequency of many MHz. In lighting control we encounter power electronics which is where electronic techniques are used to handle significant electrical power. The principle is usually that all the “clever stuff”, for example logic and control, is done at low power, i.e. in the normal electronic environment, but that this is connected to various electronic power devices which translate the intelligence into brute power. The range of power electronics is very wide. At the low end a small inverter to operate a fluorescent lamp from a battery may only be controlling a few watts. At the high end the motor controllers used on modern trains are electronic inverters handling more than a Megawatt. In lighting control we do not usually get into the Megawatt range for a single control channel; light-
58
electronics ing electronics is mainly concerned with power control in the range 5W–10kW (for a single channel). Complete systems may, indeed, handle Megawatts, but always as multi-channel systems. 2.1.2 Some basic electronic operations Electronics are highly complex, but can be thought of as being based on a “building block” principle. However, in the same way that all electric motors work on similar principles, yet a 1MW train motor is significantly different in construction to a sewing machine motor; so the electronic building blocks, such as electronic switches, may start from the same principles, yet require very different designs to realize for different applications. Without, yet, considering how we are going to either create or use the devices concerned, let us postulate some items which could be useful. For example: The electronic switch. Section 1.5.3 introduced the idea of the relay, where a small signal current is used to control a bigger current. In the relay the action is electromechanical – there is still a physical means of breaking the circuit. In the electronic switch the principle is similar to that of the relay, except that the switching element is solid state, it has no moving parts, and no contact gap. Comparing the electronic switch to its relay equivalent might well raise the following questions: How is the control signal isolated from the power it is controlling? Typically transformers are used to provide signal isolation. Opto-electronic components are also used. Is it possible to have an electronic switch with no resistance? Obviously if there is a voltage drop across the switching element, heat will develop within it. There always is a drop, but in the best devices it is small. Most power devices need mounting on a heat sink to take away the excess heat. Is the “off” position of the switch really off? where safety is concerned any leakage could be dan-
LIGHTING ELECTRONICS
(a)
High current Load
Low current signal
Electronic switch
Large AC signal through loud speaker
(b)
Amplifier
Small A.C. Signal
Figure 2.1 (a) shows the concept of the electronic switch, where a low voltage, low current signal is used to instruct the electronic switch to apply full power to the load. Although shown for DC switching, the principle also applies to AC. (b) shows an amplifier, where a small varying voltage is amplified to a larger voltage.
gerous. Most systems have a separate isolating switch which is used when complete isolation is required. Does the switch work faster than its electromechanical equivalent? Yes, and this is its main advantage. It can switch millions of times without wearing out, and it can switch very fast indeed. The amplifier. The concept of an electronic am-
plifier is similar to the electronic switch. In the switch a small signal of fixed magnitude “lets through” a current which is only limited by the circuit impedance. In Figure 2.1(a) the switched current through the load will be determined by the voltage V, and by the total resistance represented by the load, the switch itself, and the internal resistance of the battery supply. (If it was an AC circuit we would be talking about impedances instead of resistances.) In Figure 2.1(b), the electronic switch is replaced by an amplifier. In this device the current let through is not constant, but varies in proportion to the control signal. No input signal, no output. 20% input signal, 20% output current, 100% input signal, 100% planned output current. Amplifiers are normally used with varying or alternating currents, the most obvious manifestation being audio equipment where very small signals derived from radio detectors, phonograph pick-ups or CD decoders are amplified so we can hear them when converted back to acoustic energy by a transducer, for example a headphone or a loudspeaker. The rectifier. There is often a need to turn AC into DC. Most electronic equipment (or at least its control and logic side) works from DC, so it is either battery powered, or is powered from a step-down transformer whose AC output is converted to DC. The start of the conversion process is by the rectifier. This can be thought of as a special kind of switch, which does not require any signal to operate, but only lets current through in one direction. Figure 2.2 shows the action. The output of a simple rectifier is DC, but is the result of half wave rectification. In practice it needs cleaning up. A full wave rectifier can be constructed in several ways, but the most common method is the bridge rectifier. Figure 2.3 shows how four rectifying elements are used. The positive going part of the AC waveform goes through one rectifier, then the load, and out through a second rectifier. The other two rectifiers provide a blocking action to ensure that the current only goes through the load in one direction. When the negative going part of the AC waveform arrives, the rectifier pairs exchange roles. Rectified AC has a high ripple content. In order to make it suitable for circuits needing a DC supply,
59
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Anode
Anode
Cathode
Cathode
Rectifier
Rectifier
Load
Load
Load
AC Supply
AC supply
Waveform at Current flow positive half cycle Anode
Waveform at anode
Waveform at cathode
Negative half cycle Waveform at Cathode Waveform at load
Figure 2.3 The most common full wave rectifier, the bridge rectifier. Figure 2.2 The half wave rectifier. Pulses of DC flow through the load.
it must be smoothed, and the simplest way of doing this is to add a capacitor. The capacitor value is chosen to minimize the ripple, but precautions have to be taken to ensure that the initial charging current does not exceed the rating of the rectifier. Figure 2.4 shows the idea. The oscillator. Electricity as described so far has come in two flavors, DC and AC. We have also admitted that there may be electrical signals used for carrying information or control instructions in the form of continually varying voltages. In some circuits there may be a need for a high frequency AC signal. This might be needed either because it was intended to operate an electrical device at a much high frequency than normal mains frequency, or because a precise means of timing was required by virtue of the accuracy of the signal. For this purpose we need some kind of oscillator. Electronic oscillators played an important part in the development of radio, producing sinusoidal waveforms. Today they still play an important role in electronics, but techniques have changed considerably.
60
2.1.3 Resonance, tuned circuits and filters In Chapter 1 we encountered capacitors and inductors in isolation. Now we consider what happens when the two meet. Figure 2.5 shows an AC circuit with Addition of smoothing capacitor
Load
Waveform at load
Ripple
Original Smoothed
Figure 2.4 Using a capacitor to “smooth” the output of a sine wave rectifier.
LIGHTING ELECTRONICS
R
L
C
I
The nett result of the two reactances, since they are in direct opposition, is OB-OC, and is shown as OD. The supply voltage V must be the vector sum of OA and OD, shown as OE on the diagram. A little bit of Pythagoras shows: OE2 = OA2 + OD2 = OA2 + (OB-OC)2
V
If we convert the vectors into the corresponding electrical quantities: V2 = (RI)2 + (ωLI - I / ωC)2
B
Factor out the current I to give: Voltage Vector
V = I √ [R2 + (ωL - 1 / ωC)2]
D
E
Restate as: I = V / √ [R2 + (ωL - 1 / ωC)2]
O
I
I
A
But I = V / Z where Z is the circuit impedance. Therefore in the series circuit shown, the combined
Current Vector
1
C
Q=
resistance, capacitance and inductance. If V and I are the r.m.s. values of the voltage and current respectively, we can say that the voltage across the resistance R is RI volts, shown as OA in the vector diagram, and is in phase with the circuit current I. We know that the voltage across the inductance is ωLI (i.e. the inductive reactance multiplied by the current) and that it is 90° in front of the current, and shown as OB in the diagram. Similarly the voltage across the capacitor is I/ωC but lagging the current by 90°, and shown as OC in the diagram. For the purposes of the diagram OB is considered greater than OC.
100
R Low Q 10
Current I/I R
Figure 2.5 At the top an AC circuit containing resistance, inductance and capacitance in series. Below is the corresponding vector diagram. OD is the voltage across L and C combined. OA is the voltage across R and is in phase with current I.
High Q
Z2 L
0.5
0 0.5
1.0
Frequency f/f0
1.5
Figure 2.6 Current in a series resonant circuit. Maximum current possible is that due to the resistance only IR. The Q of a resonant circuit can be shown to equal the ratio of inductive reactance at resonance to the resistance. In the diagram f0 is the resonant frequency
61
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
R C L I
fo
Figure 2.7 In a parallel resonant circuit resonance is achieved when φ equals zero; impedance is then at a maximum, and current at a minimum.
reactance arising from the inductance and capacitance is ωL - 1/ ωC, or inductive reactance minus the capacitive reactance An interesting thing happens when we juggle the values of L, C, and ω (i.e. the frequency) such that the two reactances cancel out. The current in the circuit is now limited only by the resistance, and it is possible for the voltages in the circuit to be many times the nominal supply voltage. When this condition arises the circuit is in resonance. The resonant frequency of the circuit is given by the condition that the two reactances cancel out, i.e.:
ωL - 1/ωC = 0 or ω2 = 1/LC but ω , the angular frequency, = 2πf where f is the frequency, so the resonant frequency is: f = 1/2π√(LC) The current carried by a circuit such as that of Figure 2.5 varies with frequency in the manner shown by Figure 2.6. The narrowness of the current peak and the ratio of the maximum to minimum current represent the Q of the circuit.
62
If a similar analysis is carried out on a circuit where the inductance is in parallel with the capacitance, an interesting result is found. Again there is a resonant condition, and the formula for the resonant frequency turns out to be exactly the same. However, whereas in the series resonant circuit, the total impedance is at a minimum at resonant frequency; in the parallel resonant circuit the impedance is at a maximum at the resonant frequency. See Figure 2.7. The resonance phenomenon is widely used in electronics. Tuned circuits are used in radio – for example a circuit containing an inductance and a variable capacitor is used to “tune in” to a particular frequency – if the circuit has a “high Q”, it will accept the signal at the required frequency, and ignore signals either side of it. LC combinations can also be used as the basis of oscillators. While an LC combination on its own cannot maintain a continuous electrical output because of the inevitable resistance in the circuit; it can be the determinant of an oscillator frequency. A transistor is used to pulse energy into the circuit, but the transistor itself is triggered by the resonance. However, it is difficult to make frequency-accurate oscillators this way, and most oscillators used in lighting control, for example as the master timing element for a microprocessor, are based on quartz crystal oscillators. These use the natural mechanical resonance of the crystal as their basis. Finally LC combinations can form the basis of filters. In fact some filtering can be done using only RC combinations, for example simple low pass and high pass filters – but LC filters are more specific and effective for particular frequencies. They can use both series and parallel arrangements – for example a parallel arrangement can block a particular frequency, or a series arrangement can be used to divert or short-circuit unwanted signals. Such filters are of great importance in lighting control and power electronics. 2.1.4 Active and passive components In electronic circuits components are broadly categorized as being active or passive. Passive devices are those which require no energy to operate, other than
LIGHTING ELECTRONICS
the applied voltage or “signal”. They generally present an impedance, and are represented by inductors, resistors and capacitors. Active devices are those which change the nature of an electrical signal or waveform, and allow the realization of the building blocks described above. They require an external energy source; examples are the transistor and the thyristor. The diode is a special case; although it is usually a passive component, certain types of diode or diode circuit make it an active component.
2.2 The diode 2.2.1 Background The diode is a two terminal device which provides the function of the rectifier described in 2.1.2 above. In the early days of electronics low current, high frequency diodes, as used in radio sets, were constructed as vacuum tubes (valves to the Brits) or as point contact crystal diodes (used in “crystal sets”). Medium power, low frequency diodes, as used in battery chargers, were made using selenium or copper oxide rectifiers; and high power diodes, suitable for providing DC for traction purposes (e.g. the London Underground) used mercury arc rectifiers. However, for all but highly specialist or replacement purposes, these have all been superseded by the semiconductor junction diode. It is this component, first realized in germanium, but now mainly in silicon, which is of importance to lighting electronics. 2.2.2 Semiconductors Most active electronic components start with extremely pure crystalline silicon, produced by growing a big crystal of silicon from a seed crystal, and then zone refining it. This is a local heating process where the heating is progressively moved down the length of the crystal; it has the effect of making all the impurities move down to one end. The impure end is cut off, and the remainder sliced up into wafers for further treatment. Figure 1.1 showed that the silicon atom has four electrons in its outer orbit. On its own, this is an un-
Si +4
Si +4
Si +4 Spare Electron
Si +4
+5 P
Si +4
Si +4
Si +4
Impurity Atom
Si +4
Valency Bond
Figure 2.8 An impurity within a silicon crystal lattice can provide an extra electron for conduction.
stable state and stable crystalline silicon arises from a sharing of the outer valency electrons. Covalent bonds between neighboring atoms result in each atom having eight electrons in its outer orbit, which is a stable state. Electrical charge neutrality is maintained because the additional electrons are only “borrowed”, and there is no nett change in electrical charge. Crystalline silicon at room temperature is an insulator. If extra energy is given to a silicon atom, for example in the form of heat, it is possible for an electron to break away from its valency bond, and become a free electron able to conduct electricity. However, the atom left behind now has a positive charge. This is referred to as a hole, and the movement of holes, arising from free electrons jumping from one atom to another in succession, is another form of conduction – in the opposite direction to the electron flow. Silicon (a Group IV element in the chemical periodic table; Group IV also includes carbon and germanium) is referred to as tetravalent because of its four valency electrons. Group III elements, such as aluminum, gallium and indium, are trivalent with three valency electrons, and Goup V elements, such as phosphorus and arsenic, are pentavalent with five valency electrons. Silicon’s ability to conduct electricity is changed dramatically if it has a few atoms of another element
63
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
dispersed within it. By deliberately adding impurities, it is possible to create two different kinds of conducting silicon. n type silicon is silicon doped with a very small amount a Group V element. n stands for “negative” where conduction is predominantly by negative electrons. Figure 2.8 shows that the effect of a pentavalent impurity is to provide a free electron which can wander round the crystal lattice. p type silicon is achieved by doping with a trivalent impurity, creating additional holes which provide “positive” conduction. In both types of conducting silicon electrical neutrality is maintained overall. In the one case a lot of fixed positively charged nuclei are balanced by the same number of wandering negatively charged electrons, and in the other a lot of fixed negatively charged nuclei are balanced by the corresponding number of holes. In any real silicon conduction is by both electrons and holes. However, in n type silicon the majority carrier is the electron, with the hole being the minority carrier. In p type silicon the situation is reversed.
P-type
n-type
+++ +++ +++
- - - - - - -
++ - + - ++ - + - ++ - + - ++ ++ ++ -
+ - + - + - -
(a)
(b)
(c)
Depletion Layer
p
n
Holes
Electrons
2.2.3 The semiconductor junction diode An interesting thing happens when a connection is made between n silicon and p silicon – a so-called pn junction. At the junction, free electrons on the n side are enticed into occupying holes on the p side, and vice versa. This has the effect of creating a depletion layer where there are no mobile current carriers. A barrier potential difference is set up, this is about 560mV for silicon. What happens if we try and pass a current through such a junction? Figure 2.9 shows that if the positive terminal of a battery is applied to the p side, the holes are pushed back across the junction; in fact provided the external voltage exceeds 560mV the depletion layer is eliminated and current can flow. On the other hand, if the battery connection is reversed, the migration effect at the junction is enhanced, and in effect the depletion layer gets wider, preventing current flow. Figure 2.10 shows the diode characteristic, where forward and reverse currents are plotted against ap-
64
+
- >0.6V
(d)
Current Flow
(e)
-
+
No Current Flow
Figure 2.9 The p-n junction. (a) shows the separate p and n type silicon, each with its mobile carriers. (b) shows that when they are placed together the carriers at the junction migrate across it. This has the effect of creating a depletion layer which has no carriers (c). If a battery of greater than 560mV is connected to forward bias the junction, the depletion layer is broken down and current flows (d). But if the battery is connected the other way round, the depeletion layer is made wider (e) and no current flows.
LIGHTING ELECTRONICS
Current
Reverse Voltage -40
0.5
1
Forward + Voltage
Reverse Breakdown
Figure 2.10 The silicon diode characteristic curve. The negative voltage scale is condensed; the actual reverse breakdown voltage depends on the diode construction. It might be 40V as shown, or several hundred volts.
plied voltages. From this we can derive a number of important parameters. Most of these parameters are just as important for transistors and thyristors, and are highly relevant to the practical realization of lighting control electronics. Forward voltage drop. This arises because of a combination of the depletion layer voltage and the additional resistance of the junction material and connection to it. In any power component the aim is to keep this voltage to an absolute minimum. In operation any voltage drop at the device will result in a V2/R heating effect. Reverse leakage current. A p-n junction is not a perfect rectifying device. There is always a small reverse leakage current arising from several causes. As the temperature rises minority carriers can diffuse into the depletion region. In a real diode there are practical problems at the physical edges of the junction where surface contaminants can cause unwanted conduction. Finally the main cause is intrinsic conduction within the depletion layer; this is the conduction by silicon’s own free electrons. This is highly temperature dependent, since it is the thermal energy which creates the free electrons – the current can double for every 10°C rise in temperature. In
power electronics the aim is always to minimise the leakage current. Reverse breakdown voltage. Figure 2.10 shows that when the applied reverse voltage exceeds a certain value, the current rises very fast – a so-called avalanche effect. The effect of the electric field is to increase the velocity of the mobile carriers; when this exceeds the thermal drift velocity they have enough energy to knock otherwise stable valency bond electrons into conduction, creating hole-electron pairs. These pairs can then themselves create further pairs, leading to the avalanche. Any real semiconductor device is rated to ensure that it does not break down. Real world ratings are derived from the above characteristics of the device; they are usually given in graphical form, because most are temperature dependent and/or inter-dependent. Examples are: Maximum reverse voltage. This must be considerably less than the breakdown voltage. It is shown as three different values all related to temperature; one is a continuous value, which the device can withstand indefinitely. The second is peak repetitive voltage, higher than the continuous voltage, but limited as to the percentage of time it can be applied. For example it might correspond to the peak value of an AC waveform. The third is an absolute maximum voltage which can only be allowed occasionally in the life of the device, for example a transient voltage surge. Maximum current. Again this is specified three ways. Maximum continuous current is that which can be carried continuously at a specified temperature. Peak repetitive current applies for a limited time. A combination of these two ratings is used to derive the achievable performance in an AC circuit (for example a bridge rectifier) and the result is then expressed as an “average” or “r.m.s.” rating. The current waveform has a significant bearing on the achievable result. The absolute maximum rating is for occasional events, and, in the case of power diodes (and power transistors and thyristors) is used for determining the correct method of protection against short circuit and overload.
65
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Power dissipation. A diode or other power electronic device dissipates heat according to the current being carried and the voltage drop across the device. For example in a bridge rectifier, the current has to go through two diodes in each half of the sine wave cycle. The total drop might be around 1.1V, and if the bridge is carrying 5A r.m.s, the total dissipation will be 5.5W. The heat must go somewhere, and power devices dissipating more than a very few watts are normally mounted on a heatsink to dissipate the heat. Thermal resistance. In respect of a diode rectifier, the junction to case thermal resistance is a measure of the difference between the junction temperature and the external temperature of the casing. It is measured in °C/W. Thus, for example, if the junction temperature is 120°C, the case temperature is 100°C, and the device is dissipating 5W, it has a thermal resistance of 4 °C/W. This information is needed when designing heatsinks, or determining the actual ambient temperature at which it is safe to run a device. The junction diode is available in a huge range of ratings and packages – from small high frequency, low voltage diodes carrying milliamperes to giant power devices carrying thousands of amps at high voltage. Some examples are shown in Figure 2.11. 2.2.4 The zener diode In addition to the standard diode mainly used as a rectifier, there are a number of special purpose diodes. The first of these is the zener diode. This name has now been given to all kinds of voltage reference diodes, although, strictly, it should apply only to diodes showing the zener effect. Voltage regulation diodes work as normal diodes in the forward direction, but have a precisely defined breakdown voltage in the reverse direction. i.e. they do not carry current in the reverse direction until a particular voltage is reached, whereupon they conduct. Varying the concentration of n and p carriers in the junction varies the width of the depletion layer. Heavy doping results in a narrow depletion layer, and this results in a very high electric field across the
66
a) small signal diode
c) four diodes in same package to make a bridge rectifier
b) individual low power diode used in electronic ballast or transformer
d) stud mounted power diode
Figure 2.11 Examples of silicon diodes and their packaging.
junction. The zener effect is when this field is so great that it gives electrons sufficiently high energy to break away from their valency bonds and become conducting electrons. It happens below about 5V. If the p and n layers are lightly doped, the depletion layer is wide, and the critical field for the zener effect cannot be reached. However, the diode still suffers from the avalanche effect. Figure 2.12 shows the difference between the avalanche and zener effects in reverse biased junctions. Voltage reference diodes are diodes made with precisely known avalanche breakdown or zener voltages. If they are placed across a supply, they will “clamp” the voltage. Such diodes are made in the range 1–300V. The low voltage diodes are based on zener effect only. Those between 5 and 7 volts on a combination of zener and avalanche effect, and those above 7V on avalanche effect. The zener voltage decreases with temperature, while the avalanche voltage increases with temperature. Voltage regulation diodes around 5.6V combining both effects have negligible variation with temperature since the effects cancel out, and are, therefore, particularly useful in providing a stable voltage reference.
LIGHTING ELECTRONICS
Current Symbol for Zener diode
Ca5V Voltage Avalanche
Zener
Zener slope is more gentle
Figure 2.12 The zener characteristic is more gentle than the avalanche. The symbol for a voltage reference diode is the same, whichever mechanism is used.
Zener diodes are used in power supplies to ensure correct output voltages. For example in an automobile the output of the alternator is rectified to DC, but then regulated by a zener diode to ensure that the battery does not receive an excessive charging voltage. Clearly if a zener diode “clamps” a supply, the excess energy has to go somewhere. Zener diodes are, therefore, often power semiconductors, designed to dissipate heat. A special application of the voltage regulation principle is in voltage surge suppression. If two zener diodes are placed back to back across an AC supply, no current passes through the diodes if the peak voltage of the supply does not exceed the diode rating. If a voltage “spike” comes along, then the diode pair conducts – one half conventionally, the other by breakdown, and the load is protected. Clearly the pair must be designed to absorb the energy of the spike. 2.2.5 The light emitting diode Chapter 1 introduced the idea that if an electron changes energy level, then a quantum of energy is absorbed if the new level is higher, or is emitted if the new level is lower. At a semiconductor junction, electron/hole recombination results in a quantum of
energy being released as an electron drops from its conduction band energy level to the valence band. In silicon this mechanism does not produce any visible light; however in the compound semiconductor Gallium Arsenide (GaAs) it does. The wavelength of light produced varies according to the dopant used. Nearly pure GaAs emits in the near infra-red region, but doping with phosphorous, zinc oxide or nitrogen can produce visible red, green and yellow radiation. Light emitting diodes or LEDs are important in lighting control for several reasons. • The infra-red LED is used as the light source in opto-couplers. It is also used as the radiator in some cordless remote control systems. • The conventional colored LED is widely used as an indicator lamp in control panels and mimic diagrams and within indicating switches and pushbuttons. • LEDs of more complex construction than the simple junction diode are now available to give a wide range of colors with respectable conversion efficiency. They are becoming valid sources of light to take their place alongside conventional lightsources. 2.2.6 The photo-diode In the same way that a semiconductor junction can emit radiation (i.e. an electric current is converted into visible light) so the reverse is possible. The absorption of a photon can raise an electron to the conduction band, so light can be converted to an electric current. Silicon p-n diodes respond to visible and infrared radiation, and can be used as light detectors and as the basis of light measuring instruments. When modulated light is to be used (for example in infrared remote control systems) the speed of response of the simple diode is not fast enough due to capacitance effects, so in this case a p-i-n or PIN construction is used. Here the simple p-n junction is separated by a thin layer of intrinsic conduction (pure or very lightly doped) silicon. Photodiodes normally work in the reversed biased mode, the reverse current in excess of leakage current being directly proportional to the incident radiation. However, if they are operated in forward
67
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
individual LED in clear housing for general purpose indication
7 bar LED indicator used for channel number indication in dimmers and moving lights
2.3 The transistor 2.3.1 The Bipolar Transistor The Bipolar Junction Transistor has two different constructions, p-n-p and n-p-n. The principle is that of placing two semiconductor junctions together as The p-n-p Transistor
high power cluster of LEDs used in signs and video displays
photo transistor fitted in light level sensor unit
p-type
Emitter E
photo thyristor used to provide isolation in professional dimmers
conducting mode an e.m.f. of about 0.5V is developed, creating a solar cell. This alerts us to the fact that photo-sensitivity is an inherent property of silicon semiconductor devices, so unless it is a required function, all silicon electronic devices must be protected from light (or other electromagnetic radiation) An opto coupler is created when an LED is packaged next to a photodiode (or, more often, a phototransistor or photothyristor; however, the principle is similar) within a lightproof housing. In this case an electrical signal fed to the LED will be replicated as another electrical signal coming from the photo-detector. This method gives total electrical isolation between input and output. It is widely used in lighting control to separate low voltage control signals from high power/high voltage circuits. Figure 2.13 shows some opto-electronic components.
68
p-type
Base B
E
Figure 2.13 Examples of opto-electronic components.
n-type
C
Collector C
E
C
Symbols
B
B
The n-p-n Transistor
n-type
p-type
Emitter E
n-type
Base B
E
C
B
Collector C
E
Symbols
C
B
Figure 2.14 The bipolar junction transistor. The circuit symbols with the circle are used for discrete components. Those without the circle are used when the transistor is built in to an integrated circuit, usually with many other transistors.
LIGHTING ELECTRONICS
Narrow emitter-base depletion layer due to forward bias
Wide base-collector depletion layer due to reverse bias
Emitter
Base
Collector
Hole Flow n-type
p-type
IE
>0.6v For Conduction
p-type
IB
IC
VBE +
-
VCB +
-
Figure 2.15 The operation of the p-n-p transistor showing the biasing arrangement. In an n-p-n transistor all voltages and current flows are reversed, and main current is carried by electrons instead of holes.
shown in Figure 2.14. Both types of bipolar transistor work in a similar manner. The reason they are called bipolar transistors is that they depend on both positive and negative majority carriers for conduction – holes in the p type silicon and electrons in the n type silicon. The bipolar transistor is usually just referred to as a transistor, whereas other types are always referred to with some additional description. It has three electrodes, the emitter, the base, and the collector. Its importance lies in the mechanism of current gain, whereby a small current in the base results in a big current in the collector. Figure 2.15 shows how a p-n-p transistor is biased. (The description which now follows also applies to n-p-n, but then the biasing arrangements would be reversed in polarity, and all references to holes would become references to electrons, and viceversa.) In the diagram batteries are shown as the power supplies for simplicity, and the direction of current is shown conventionally as flowing from positive to negative. Back in the mid – 19th Century, Kirchhoff set out laws about what happens when different electric currents meet at a point; and his first
law states that if several conductors meet at a point, then the total current flowing towards that point is the same as that flowing away from it. It is intuitive that this is so, since if it was not, the meeting point would accumulate an electric charge – and apart from no such charge being measured in practice, it is not easy to visualize how such a charge could be held. Applying the law to the transistor circuit, we have: IE = IB + IC i.e. the current flowing through the emitter connection equals the sum of the currents through the base and collector. The transistor is special because IB is very small compared with IC, and in use the effect is that small changes in IB make big changes in IC. The effect of forward biasing the emitter-base junction is to narrow its depletion layer; whereas the reverse biasing of the collector-base junction is to widen its depletion layer. The description of the p-n diode showed that if VBE exceeds 0.6V, the depletion layer between emitter and base reduces to zero width. The positive potential on the emitter repels the surplus positive holes
69
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
PNP Transistor
IC
IC
Current carrying limit
Imax
IB
50
VCE VBE
Power dissipation limit
IE
25 %Duty Cycle
75 100
Breakdown voltage limit
Secondary breakdown limit IC mA
Vmax
gm at P is slope of graph at P
Figure 2.17 The safe operating area for a transistor. Note that it changes with duty cycle.
IC
P
5
VBE 0 550
600
650
mV
VBE
hFE = IC /IB is the large signal or DC current gain; important for switching applications. hfe = 'IC/'IB is the small signal or AC current gain; where ' signifies a small change; important for audio amplifiers etc. gm = 'IC/'VBE is the transistor mutual conductance (measured in Siemens or mho, reciprocal ohms) relating small changes in base-emitter voltage to collector current. VCEO(SUS) is the maximum collector-emitter voltage which can be sustained without breakdown, measured with open circuit base. (Breakdown voltages are also specified with other base biasing conditions.) ICEO is the collector-emitter leakage current when there is no base current. It is highly temperature dependent. VBE(SAT) the base-emitter saturation voltage, and VCE(SAT) the collector-emitter saturation voltage are the respective voltage drops at the point when increasing the base current gives no further increase in collector current. The conditions under which they are measured must be specified. PD is the maximum power dissipation of the device (at a specified temperature) in principle the product of VCE and IC. But in fact transistors must operate in their safe operating area 0010 see Figure 2.17.
Figure 2.16 Some parameters relating to the bipolar transistor. The diagrams relate to p-n-p transistors.
70
VCE
which are attracted by the negative base region. The base-collector junction is reverse biased by the larger potential VCB – so the collector is negative with respect to the base. The widened depletion layer effectively prevents holes in the collector crossing to the base, but does not stop the flow of holes in the other direction, since they are attracted by the heavily negatively charged collector. Once conduction has started, the great majority of the holes starting from the emitter make it through to the collector, but a small proportion combine with free electrons in the base region, to form a small base current. The fundamental property of the transistor is the static current gain, or common emitter forward current transfer ratio designated hFE where: hFE = IC /IB Typical transistor gains are in the range 100–300. From a “mind picture” point of view, you can think of a small base current being “amplified” to a large collector current; or a small base current controlling the larger collector current flow. Actually it is the base emitter bias voltage which determines the collector current, and hFE which in turn determines the base current. Figure 2.16 lists some of the common parameters relating to a transistor. It is important to understand that many parameters are inter-dependent and temperature dependent. Figure 2.17 shows the idea of a the safe operating area for a transistor. For any particular tempera-
LIGHTING ELECTRONICS
Lamp Load
+12V I C=120mA
I B=1mA ON 0.75V
OFF
hfe =120
0V
Figure 2.18 The transistor as a switch. The 1mA control signal switches 120mA through the n-p-n transistor.
ture the device will break down due to excessive voltage Vmax or excessive current Imax. There would normally be a straight line maximum power dissipation curve, but in a transistor there can be an uneven distribution of current, arising from temperature variations across the device and irregularities in construction. This can lead to secondary breakdown arising from local thermal runaway; so the safe operating area is less than expected. Many applications of transistors in power circuits require the transistor to operate without carrying current continuously; its duty cycle is less than 100%. In this case the safe operating area is increased as shown in the figure. From the description of how a transistor works, we can see that it could be used to make some of the building blocks referred to in 2.1.2. Figure 2.18 shows the transistor being used as a switch, where a very small current is used to switch a large current on and off. In practice the low power switch shown in the circuit would most likely itself be another, smaller, transistor being driven by some kind of logic circuit – possibly being driven very fast to provide fast on/ off switching. In practice if we want to use a transistor as a switch, we want the voltage across it to be the full line voltage when it is “off”, and to be as little as possible when “on” – since any remaining voltage will result in heat being dissipated within the transis-
tor. To achieve the minimum voltage requires the transistor to be driven to saturation which is the point at which an increase in base current produces no more increase in collector current. Figure 2.19 shows a transistor being used as a current regulator. Here the base bias voltage is simply being varied between the point at which the transistor is just not conducting, to the point at which it is passing maximum current. The problem here is that the transistor is now operating as a variable resistance – it is not in saturation. Figure 2.20 shows the transistor being used as a voltage amplifier. Here the idea is to produce an output which is a magnified but perfect copy of the the original input signal. The input is applied to the base of the transistor, and here shown as a 0.05V swing, superimposed on the base bias voltage (which must be there, otherwise there would be no conduction at all.) The output is shown as a varying voltage at the collector, arising from the change in collector current. The collector is biased at well below the maximum collector voltage to allow a voltage swing to develop. For a typical “discrete” transistor the 0.05V input voltage swing might produce a 7.5V swing in the collector voltage – corresponding to an amplifier voltage gain of 150. Notice that the output voltage is out of phase with the input, the higher the collector current, the lower the collector voltage.
Lamp Load
Rotary 0.8V Potentiometer
+12V
IC Varies with change in VBE
0.65 0.6V
VBE 0V
Figure 2.19 Replacing the input switch with a rotary potentiometer gives a current regulator. A resistor network ensures that the potentiometer only operates in the active range 0.6–0.8V.
71
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
VCC
IC
VRC=ICRC
RC
VC VOUT=VCC-ICRC VIN=VBE
IE
Input 0.07
VBE=
VIN
VBE 0.65
mon to both input and output. RB1 and RB2 set the bias voltage at the base, and are chosen to carry enough current such that the base signal current is insignificant compared to it, and therefore does not itself affect the bias voltage. CIN and COUT are capacitors which block the DC bias, but have a low AC impedance at the frequencies of the signal being amplified, they simply isolate the wanted signals from the amplifier. RE biases the emitter (which in turn implies an increase in the needed base bias). This is done so that if there is an unwanted rise in collector current due to high temperature, the voltage drop across RE increases, and has the effect of reducing the base-emitter voltage and, therefore, the collector current. CE is needed as a decoupling capacitor to pass the amplified AC current, which would otherwise be limited by RE.
Base Bias
IC
+VCC
RC C OUT
RB1
0
Time
C IN V OUT
Output 12
V IN
CE
RB2 RE
V
CC
0V
VC 6
0
Collector Bias
VC=
VOUT
IC Collector current with no input
Figure 2.20 The transistor as a voltage amplifier. Note that all values are examples only. Vb
2.3.2 Amplifiers 2.3.2.1 Class A and Class B amplifiers Figure 2.21 shows the components needed to surround a transistor to make it work as a Class A audio amplifier stage in a common emitter configuration. This simply means that the emitter terminal is com-
72
V IN Figure 2.21 The Class A amplifier. The transfer characteristic is linear, because the transistor is biased to ensure this is the case.
LIGHTING ELECTRONICS
The Class A amplifier has low distortion; that is to say the output voltage waveform closely matches the input waveform; but it is inefficient because current must flow at all times. The transistor is biased so that the relationship between the base input signal and collector output is linear. The inefficiency does not matter in small signal work, or in low power audio amplifiers; but it matters a lot in high power amplifiers and power conversion equipment. A Class B amplifier works with minimum quiescent current (by biasing transistors to cut-off) and, as a result, can have high efficiency. Such amplifiers work in push-pull mode, where one transistor operates in the positive half cycles, and another one in the negative. Figure 2.22 shows the principle of the Class B amplifier. The output transistors are a complementary pair, one being p-n-p, and the other n-p-n. The variable resistor VR sets the bias condition for both
Nett input
(V IN- B VOUT) V IN
A
B
V OUT = A (V IN
-
BV OUT)
-BV OUT
Figure 2.23 The concept of negative feedback applied to an amplifier.
output transistors so that for zero input signal, no collector current flows. The diodes D1 and D2 provide compensation for the variation of VBE with temperature. Note how the output waveform is distorted, arising from the crossover distortion arising from using both transistors on the non linear part of their characteristic curve. 2.3.2.2 Negative feedback; Classes C and D
+ VCC TR1
VR D1 D2
C OUT
V IN
TR2 0V
IC
I OUT
In real audio amplifiers a hybrid Class AB approach is used, not as efficient as Class B, but much more efficient than Class A. Real power amplifiers also make use of negative feedback to stabilize the gain and reduce distortion. Figure 2.23 shows an amplifier with open loop gain (the gain with no feedback) A. A proportion B of the amplifier output is subtracted from the input, therefore the actual input voltage becomes VIN - BVOUT; so now the actual output voltage becomes A(VIN - BVOUT ). System gain is the actual output voltage divided by the original input voltage, and can be stated as:
VB2 VB1
V IN
Figure 2.22 The Class B amplifier. The transfer characteristic is non linear, because both transistors operate both in their linear regions and in their non-linear cut-off regions.
Overall Gain = A/(1 + AB) This expression is known as the closed loop gain. It has an interesting result. If A is 100, and B is 0.1, the overall gain is 9.1; if A is 200 and B remains 0.1, the gain only changes to 9.5. Practical power amplifers are built with A very large, in which case the approximation; Closed loop gain = 1/B
73
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Signal IN
High speed switch
Output power transistors
Filter V OUT
Output waveform before filter
Minimum output
Maximum signal output
Intermediate output
Figure 2.24 The concept of the Class D amplifier.
applies; i.e the gain is independent of the actual gain of the underlying amplifier, and depends only on the feedback ratio. Feedback also reduces distortion. If an amplifier has distortion N% in its open loop state, negative feedback reduces this to: Distortion = N/(1 +AB) % However, this only applies if the distortion does not, itself, affect amplifier gain. Unfortunately crossover distortion falls into this category. Efficient Class C amplifiers are used in radio frequency applications. In this case the transistor is biased so far “off”, that only the peaks of the input waveform are amplified. Tuned loads restore the input waveform. Of more relevance to power control is the Class D amplifier. In Class D the transistor(s) act only as a switch, i.e they are on (saturated condition with minimum volt drop) or off. To work as an amplifier this requires that the switching frequency is far higher than the frequency being amplified. One possibility is to use a square wave of constant frequency with variable mark to space ratio. For example the full
+VCC
RB
width wave represents 100%, and the narrowest “spike” represents zero. Figure 2.24 shows the idea. The high speed switch generates a train of pulses of varying width, depending on the strength of the input signal. These are amplified by the output power transistors, and the result is filtered to eliminate the high frequency components and restore the original waveform. 2.3.2.3 The emitter follower There are a number of other amplifier terms which can be encountered in lighting control electronics, and a few of them are summarized here. The emitter follower is used as a buffer amplifier in signal distribution. Instead of being in the common emitter configuration used in the circuits described so far, the transistor is used in the common collector configuration shown in Figure 2.25. In such a circuit the output follows the input in phase and amplitude. That means that, as a voltage amplifier, the emitter follower (so called because the emitter voltage follows the base input voltage) has unity gain, so it might well be asked what use it is. Its virtue lies in the fact that it has a very low output impedance. Conversely it has a high input impedance. In practice this means that it does not load, and thus distort, its source; but itself can tolerate a wide range of loading. 2.3.2.4 The Darlington pair In power electronics conventional transistors suffer from the fact that at high currents they have low gain. This means they need large base currents, and, as a C
B
TR1
1
~
D1
Input high Z
D2
OUTPUT RB2
Low Z
Figure 2.25 The emitter follower.
TR2
~
RE 0V
74
R1
R2 E
Figure 2.26 The Darlington pair.
LIGHTING ELECTRONICS
ts tf IC
90%
10% td
tr OFF
ON Turn on time
Turn off time
Figure 2.27 Turn-on and turn-off times in a transistor. td is a delay in start of collector current due to collector and emitter depletion capacitance; tr is the rise time, determined by junction capacitances and the carrier transit time in the base region. ts is the storage time, the time needed for excess charges in base and collector regions to recombine as the transistor comes out of saturation. tf is the fall time, determined by junction capacitance and negative base current.
result, complex base drive circuits. The Darlington pair shown in Figure 2.26 goes some way to solving the problem. TR1 is used as the base driver for TR2 . The practical circuit includes resistors R1 and R2 to prevent the transistors amplifying their own leakage current. Diode D2 protects the output transistor from reverse voltages. Diode D1 removes the charge carriers stored in the base of the output transistor, and thus speeds up its turn-off. Turn-off and turn-on times in switching transistors are critical, since they result in slow operation and heat dissipation. They are caused by charge storage and capacitance effects, as shown in Figure 2.27. While “power Darlingtons” can be made up from discrete components, they are usually supplied as a single semiconductor structure.
identical in charactersistic and in variations with temperature (achieved in practice by them both being in the same integrated circuit and encapsulation). The fact that all changes due to temperature etc are equal, and that emitter current is shared in R3 (the “tail”) means that any changes affect both transistors equally. The long tail pair amplifies the difference between the two inputs V1 and V2 . If V1 = V2 then it is clear from the figure that Vout is zero. As an example suppose V1 = 0.12mV and V2 = 0mV, and that the measured output Vout = 60mV. Then the amplifier is said to have a differential gain of: Ad = 60/0.12 = 500 Now suppose there is an unwanted noise signal of 1mV which affects both inputs simultaneously. Such a signal is referred to as being common mode. Now it is the case that V1 = 1.12mV and V2 = 1mV. It might be that now the output, instead of being 60mV has risen to 62mV. The common mode gain, which refers only to the extra 1mV signal is: Ac = (62 – 60)/1 = 2 The measurement of an amplifier’s ability to amplify differential signals while rejecting common mode signals is referred to as its common mode rejection ratio (or CMRR) and equals Ad/Ac. In the example its value is 250. In integrated circuits much higher values are achieved, so great that it is inconvenient to represent them as simple numbers. A typical “op-amp” (to be described in Section 2.6) has a CMRR of over 30,000 – a figure so big that it is easier to use a logarithmic measure, the dB; for example +VCC R1
VOUT TR1
TR2
V1
2.3.2.5. The long tail pair; differential amplifiers The long tail pair or differential amplifier is at the heart of DC amplifiers and many AC control signal amplifiers. The bipolar transistor version is shown in Figure 2.28. The two transistors TR1 and TR2 are
R2
IN
V2 R3 0
Figure 2.28 The long tail pair.
75
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
VGS
+ Gate G p
VIN n
Source S
p
Figure 2.29 A center-tapped transformer producing a balanced signal.
90dB. (The dB or deciBel is described in Section 2.5.) A differential amplifier is fed in one of two ways. A “single ended” input could be applied as V1, whereas V2 could be tied to 0V. Alternatively a balanced input could be provided to both inputs simultaneously. Similarly at the output there is a choice of an in-phase (from TR2), or out-of-phase (TR1) output signal, or a balanced signal. A balanced signal is one where the same signal is fed on two wires, but each half is exactly out of phase. A “traditional” way of achieving a balanced signal is to use a center-tapped transformer as in Figure 2.29. The term common mode referring to an unwanted signal affecting two input wires simultaneously is of considerable significance in lighting control, especially with reference to EMC and digital control signals. 2.3.3 The Unipolar Transistor 2.3.3.1 The JFET Unipolar transistors depend for conduction on only one type of carrier. They are also called field effect transistors or FETs. This is because the current flow is determined by electric field – in practice just a voltage. In a bipolar transistor the gate current can be significant, whereas in a unipolar transistor it is negligible. Unipolar transistors come in several varieties. Figure 2.30 shows the standard biasing arrangement for an n-channel junction field effect transis-
76
Drain D
+
VDS
Depletion regions
S
G
D
p n channel p Substrate
JFET
+ D
G
G
D
p Channel
+
S
S n Channel
+
C
B
C B
n-p-n
E
p-n-p
E
+
BIPOLAR
Figure 2.30 The n-channel JFET. The concept (top), the principle of construction (center), and the comparison of JFET circuit symbols with those of the bipolar transistor (bottom).
tor, or n-channel JFET. A p-channel JFET, would, like its bipolar counterpart, simply have n and p regions reversed, and the biasing reversed. Figure 2.30
LIGHTING ELECTRONICS
also shows the symbols for JFETs and compares them with those for bipolar transistors. The channel is the conduction channel between the source and drain electrodes. The gate electrode is shown as being either side of the channel. Being of the opposite polarity to the channel, it sets up a depletion region at its junctions with the channel. As the source–drain voltage is increased, the depletion region gets bigger, until a point is reached when both halves meet – the pinch-off voltage. At this point the current through the channel is limited to those electrons which can sweep through the depletion layer – so there is no further increase in current with increase in source-drain voltage. (Other than when a gatechannel breakdown occurs.) Figure 2.31 shows the characteristic of the JFET for different levels of bias. 2.3.3.2 The MOSFET The fact that the FET works by the influence of an electric field raises the possibility that there need be no actual conduction path from the gate electrode. The Insulated Gate Field Effect Transistor (IGFET) uses this idea; in practice its manifestation being the Metal-Oxide Silicon Field Effect Transistor (MOSFET). In the manufacture of semiconductor devices, there are a number of processes used. The process starts with the wafer of pure silicon (Section 2.2.2) ID
VDS PINCH IDSS
VGS=0 -1V -2V -4V
Gate to channel breakdown
VDS
Figure 2.31 The JFET characteristic showing variation of drain current with drain to source voltage for different levels of gate bias. The dotted line shows the pinch-off voltage.
but this can then be treated in several different ways: Oxide growth. If the silicon is exposed to oxygen or steam it oxidizes, producing a layer of Silicon Dioxide (SiO2). This is an insulating and protecting layer. It can be grown and removed (by acid etching) at various stages of production. Epitaxial growth. This is growing more silicon crystal on the existing base. It is done by exposing the silicon to Silicon Tetrachloride vapor (SiCl4) which disocciates when it meets the hot silicon. If the vapor also includes an impurity (for example Phosphorus Trichloride) then an n or p type silicon is grown. Diffusion. The introduction of impurities as part of epitaxial growth is really only applicable to large areas. If small areas are to be treated, then the impurities enter the silicon by diffusion. The silicon is held near its melting point so the impurity atoms from the vapor can easily enter the crystal lattice. Areas which are not to be diffused are protected by an SiO2 layer, which itself is put down using a photolithographic process. Metal formation. In this part of the process metal is evaporated or sputtered on to the silicon to provide interconnection between parts of the semi conductor device, or to provide connection to the outside world. Again photolithography is used to create the protective layer needed to ensure the metal only goes where it is needed. From this brief description, the construction of the JFET shown in Figure 2.30 can be appreciated. So can the fact that SiO2 , being an insulator, allows the idea of a gate electrode which acts through an insulating layer to be realized. The MOSFET is, in reality, made in two ways, shown in Figure 2.32. This identifies two kinds of MOSFET. The depletion mode version, with a narrow channel, works in the same way as the JFET, but has a much higher input impedance. In the alternative (and, in practice, standard) enhancement mode version there is no continuous n-channel; if there is no bias on the gate, no current flows. Forward bias must be increased to a threshold voltage before conduction starts. At the threshold voltage an inversion layer is created under the gate electrode and (in the figure example) an nchannel is created.
77
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
S
G
D
Si02
G
p Substrate
Channel S
Sub S
Sub G
DRAIN CURRENT
D
n
D
Si02 n
D
n
G
p
Sub S
Sub
Figure 2.32 The top figure shows the construction of the depletion mode MOSFET and its symbol. The bottom diagram shows the enhancement mode MOSFET. Note how the symbols clearly show the absence of direct connections.
MOSFETs have a number of advantages over bipolar transistors: • Because they do not have minority carriers (i.e. all the current is carried by electrons or holes, but not by a mixture) there is no minority carrier storage effect, which delays switching times (see Figure 2.27). MOSFETs are therefore faster than bipolar transistors. • MOSFETs are voltage controlled, with negligible gate current. Their gain is much higher than that of bipolar transistors. • MOSFETs do not suffer from secondary breakdown effects; their safe operating area can be extended to be solely power limited. • This is because as the temperature increases, the bulk resistivity increases, so there is no thermal runaway. This has the subsidiary benefit that devices can be connected in parallel to increase current capacity,
78
GATE TO SOURCE VOLTAGE
Figure 2.33 Example transfer characteristics of an n-channel enhancement mode power MOSFET. The linear relationship between gate voltage and drain current make them ideal amplifiers. T is the junction temperature.
without the need for any kind of current sharing components. All these considerations mean that the MOSFET is the preferred device for power electronics. Figure 2.33 shows their transfer characteristic. The linear characteristic and high mutual conductance make MOSFETs excellent amplifiers, and their fast switching speed makes them good switches. The high gate impedance of MOSFET devices makes them susceptible to damage from electrostatic discharge. Some devices are fitted with internal protection, using back-to-back zener diodes. But in most cases it is necessary to ensure that they cannot be subjected to static electricity. 2.3.3.3 Power MOSFETs The construction of the standard MOSFET shown in Figure 2.32 results in a thin “horizontal” conductive channel with limited current carrying ability. It also has a low breakdown voltage, largely dependent on the thickness of the insulating SiO2 layer. While the construction is suitable for small signal work, power MOSFETs use a different construction shown in Figure 2.34. The inversion layer, and current flow, is
LIGHTING ELECTRONICS
“vertical”. Because the conducting channel is formed within both p and n diffusions, devices of this kind are also referred to as being double diffused or DMOS devices. The figure also shows that this construction of MOSFET provides a drain-source p-n junction diode between the n epitaxial layer and the p layer. If this diode is forward biased (i.e the transistor is reverse biased) it conducts. Some circuit designs take advantage of this parasitic diode, especially for the diversion of reverse voltage surges. One of the most important MOSFET parameters is RDS(ON) this is the drain-source resistance when the device is switched fully on. In any high current application a high value of RDS(ON) results in inefficiency and a lot of heat. A number of different strategies are used to limit this internal resistance. Each construction can have some additional benefit, for example higher frequency operation, or the use of less silicon in its manufacture. In practice MOSFETs intended for power control are made up of thousands of MOSFET cells, all operating in parallel. The concept works because the positive temperature coefficient of resistance of the conducting channels ensures equal current sharing.
G
n channel
S
p+
n+
n epitaxy
n+ substrate
D Conduction by electrons
Inversion layer
D
G S
Figure 2.34 In power MOSFETs the conduction channel is vertical. The symbols + or - applied to n and p mean heavy or light doping with impurities. The figure shows an n channel MOSFET with conduction by electrons. Note that conventional current flow is in opposite direction to electron flow.
Source metallization
Silicon gate
Si02
P
P n epitaxy + n
y ax pit e n +
n
D
Silicon gate
Drain metallization
Parasitic diode p-n junction D
Figure 2.35 Power MOSFETs are made with a cellular structure. Examples are TMOS® (originally from Motorola, who no longer participate in this market) that uses rectangular cells (left), and the HEXFET® construction from International Rectifier with hexagonal cells (right.)
79
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
2.3.4 The Insulated Gate Bipolar Transistor
G S
p+
An objective, which has only been realized practically in the last few years, has been to create a device which combines the virtues of the MOSFET, such as negligible input current drive and a wide safe operating area, with those of the bipolar transistor – in particular that of low saturation voltage. Such a device is known as an insulated gate bipolar transis-
n+
-
n epitaxy
p
p
Construction
D
Inversion layer
Conduction by electrons
Figure 2.36 CoolMOS® from Infineon (formerly Siemens) uses a different diffusion pattern to reduce the sourcedrain resistance. RDS(ON) rises linearly with blocking voltage, not exponentially as is the case with conventional MOSFET.
The gate connection is made of poly-crystalline silicon instead of metal; this is possible because the gate operates at high impedance anyway. Manufacturers use different cell geometries, and different cell sizes and spacing to achieve their claimed best result, but the principle is the same for all of them. Figure 2.35 illustrates typical cell construction. A problem with MOSFETs is that at high forward voltages the epitaxial layer must be thick. RDS(ON) rises exponentially with blocking voltage, which has limited the application of MOSFETs. A recent development, exemplifed by Infineon’s CoolMOS® shown in Figure 2.36, is the introduction of vertical p stripes into the epitaxial drift region. This has the effect of providing blocking not only in the vertical direction, but also horizontally, which in turn allows the layer to be made thinner. The construction is claimed to make the relationship between RDS(ON) and blocking voltage linear instead of exponential, resulting much smaller devices for a given power handling. MOSFETs made this way are claimed to offer competition to the IGBT described in the next section.
80
G
E
n+ substrate
p+
n+
n epitaxy
p+ substrate
D Conduction by electrons Conduction by holes Symbol
C G E C
Equivalent Circuit
G
E
Figure 2.37 The IGBT. The figure emphasises the point that conduction is by both majority and minority carriers. The equivalent circuit shows that the device is in principle a p-n-p transistor driven by an n-channel MOSFET, but that the construction introduces other parasitic devices.
LIGHTING ELECTRONICS
tor or IGBT. The idea is similar to that of the Darlington pair, where one device is used to control another – in this case a MOSFET input transistor driving a bipolar transistor. The construction, symbol and equivalent circuit is shown in Figure 2.37, drawn to emphasize that conduction is bipolar. In practice the manufacture of IGBTs is a similar process to that of making MOSFETs. One manufacturer points out that simply by changing the starting materials and varying some process steps, it is possible to make IGBTs with the same photolithographic mask set as is used for making power MOSFETs. The IGBT is fast to switch on, but slower to switch off, with a current “tail”, due to minority carrier storage in the epitaxial layer.
At the individual component level there are exotic high frequency transistors used in RF applications, and special low noise transistors which are well outside the scope of this book. A useful variant encountered in lighting control is the photo-transistor, a close relative of the photodiode described in Section 2.2.6. By having a transistor as part of the diode circuit it becomes easier to set minimum reponse levels – for example the phototransistor can be biased so that it does not conduct at low light levels.
2.4 The thyristor, triac and GTO 2.4.1 The thyristor 2.4.4.1 Basic construction
2.3.5 Other transistors It is clear that transistors exist in many forms. Each of the major types can be made in several different ways to match a particular application. Many small and medium power devices include additional circuitry for protection or more sophisticated control. Such circuitry can be added by the manufacturing process – for example by diffusing on more conducting and insulating layers to make additional active components.
small signal transistor used in control circuits
MOSFET power transistor used in electronic ballast
Figure 2.38 Examples of transistors used in lighting control.
The MOSFET and IGBT have made big inroads into the power control business, especially where high frequency is concerned. However, the workhorse of electronic power control is the thyristor. The full name of the main variant is the reverse blocking triode thyristor. It is particularly significant in lighting control as the basis of electronic dimmers. The thyristor was invented at General Electric (USA) in 1957, some ten years after the invention of the transistor at Bell Laboratories. It was originally called the Silicon Controlled Rectifier, and the acronym SCR is still widely used. Figure 2.39 shows that the thyristor is a four layer device, with three electrodes anode, cathode and gate. Its symbol gives a clue as to how it behaves. In the reverse current connection (positive voltage applied to the cathode) no current flows, exactly as would be the case for a conventional diode. If the reverse voltage is greatly increased, then avalanche breakdown can occur, again exactly as for a diode. If the thyristor is forward biased, but no gate signal is present, no current passes under normal operation. If the forward blocking voltage is exceeded, the device will, however, go into conduction. The thyristor will also conduct if there is a positive gate current. Once the thyristor goes in to forward conduction, it continues to conduct, even without a gate signal, unless one of the following happens:
81
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Anode
Anode
A
p n
G
Gate
Gate
p n
K
Cathode Cathode
Two transistor Model
A p
n
p
n
p
K n
current flow between anode and cathode is limited to the transistor leakage current. If a current IG is introduced at the gate, and if the gain in the basecollector loop of T1 and T2 exceeds unity, the loop current is maintained regeneratively, driving both transistors into saturation. At this point the thyristor is said to be latched. The gains of the transistors are current dependent, so once the gate current has started the regenerative action, the anode current increases sufficiently that the gate current IG can be removed without the transistors coming out of saturation. Thyristors are bipolar devices. They have the advantage of very low on-state voltage. However, minority charge carriers must be removed before the thyristor can block an applied voltage, so switching times are long compared with, for example, MOSFET transistors. This means that in practice thyristors are used at mains supply frequencies.
G
A
2.4.4.2 Ratings
T2
Figure 2.41 shows the characteristics of a thyristor. All thyristor parameters are temperature dependent, so, as with transistors and diodes, device data is given as a set of graphs showing how the device behaves at different temperatures. Section 2.2.3 introduced a number of parameters relevant to power control devices, a few more are introduced here. The rate of change of voltage or current, denoted by the differential calculus symbols dV/dt and dI/dt
G
T1
IG K
Figure 2.39 The thyristor. Construction, symbol and twotransistor model.
• either the current through the device drops below a minimum holding current. • or the voltage across the device is changed in polarity, when it will block the reverse voltage. Notice that, unlike a transistor, the gate has no effective proportional control of the current. If the gate signal is above a minimum, the thyristor goes to full conduction, with current limited only by the circuit impedance. Figure 2.39 also shows that the thyristor can be thought of as a pair of transistors, one n-p-n and the other p-n-p. They are connected such that the base current from one transistor is derived from the collector current of the other. If there is no gate current,
82
40A thyristor used in professional dimmer encapsulated module containing two thyristors triac used in consumer dimmer
heavy duty triac used in professional equipment Figure 2.40 Examples of thyristor and triac construction.
LIGHTING ELECTRONICS
Iforward
ON
IL
IG>0
IH
Vreverse
IG=0 Vforward
Avalanche breakdown
Vbreakover
Ireverse
Figure 2.41 Thyristor characteristics.
respectively can be important. dV/dt, measured in Volts/Second, or, more usefully, V/μs, is relevant, since a thyristor can be forced into conduction by high dV/dt. This arises because p-n junctions have a capacitance, and with a change of voltage a capacative charging current flows: iC = C × dV/dt If dV/dt is high, the charging current may be high enough to trigger the thyristor on. When a thyristor is turned on, the whole active area cannot be turned on simultaneously. The area nearest the gate connection turns on first and conduction “spreads” across the device. If the anode to cathode current rises too quickly, only part of the thyristor may be conducting, and this causes localized heating, possibly followed by device failure. So an excessive rate of forward current change dIf /dt can damage a thyristor. On the other hand, provided gate ratings are not exceeded, a high rate of change of gate current dIG /dt helps a thyristor to turn on quickly. The protection of a thyristor by fuse or circuit breaker is determined by its I2t capability, a measure of its ability to let through energy (see also Section 1.5.2). For reasons which will become clear, the figure is usually quoted with reference to the duration of one half cycle of alternating current (i.e. 10ms for 50Hz or 8.3ms for 60Hz). For example, if a thyristor is quoted as having an I2t of 26A2 seconds on the
8.3ms basis (a typical figure for an 8A thyristor), then it could let through an average √3132 = 56A for a half cycle of 60Hz mains. This leads to another important characteristic which is the peak forward surge current. This is the maximum peak current that the device can stand as a single non repetitive event, and again is usually quoted with reference to a half cycle of mains current. For example the 8A device referred to in the previous paragraph might be able to stand a peak of 80A within a half cycle of 60Hz (provided that, over the whole half cycle, the I2t rating was not exceeded.) This is an important parameter for lighting control, because tungsten lamps have a low cold resistance, resulting in inrush currents which can be 10–14 times running current. While the major ratings of the thyristor are related to the anode-cathode current path, the gate has its own set of ratings in respect of maximum and minimum gate voltage and current; maximum and average gate power etc. 2.4.4.3 Control of AC power The most commonly used arrangement for thyristor AC power control is the back-to-back pair shown in Figure 2.42. If neither thyristor has any gate signal, then the pair block any current from going through the load. If both thyristors have a permanent gate “on” signal, then the load will receive maximum current – the current being limited by circuit impedance and the nature of the load itself, and only very slightly by
AC Supply
Firing control Circuit
Load e.g. Lamp
Figure 2.42 AC power control using a pair of thyristors.
83
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
V
Voltage across thyristor
Supply voltage
Current through thyristor
I
Trigger angle G
Conduction angle
Normalized voltage power fraction
Figure 2.43 Conduction angle in AC power control by thyristor.
1.4 1.2
Peak
1.0 RMS
0.8 0.6
2.4.4.4 The triac Average
0.4 0.2
0
20
40
60
80 100 120 140 160 180 Conduction angle
Figure 2.44 Variation of peak, r.m.s. and average voltage with firing angle for AC power control by a thyristor pair. The voltage scale is normalized to r.m.s. sine wave voltage, so to get an actual voltage, multiply by 230 for European supplies and 115 for USA supplies. The red line shows power output for a resistive load.
84
the thyristors (which have a forward voltage drop of 1 – 2V). One thyristor carries the positive going half cycle, and the other carries the negative going. If the gate signal is discontinuous, and is applied at some point in the half cycle, the appropriate thyristor starts by blocking current, but then switches on. The thyristor continues to conduct for the remainder of the half cycle, but then, as the polarity of the AC waveform changes, self commutates, i.e. it ceases to conduct because of the polarity reversal. The second thyristor similarly controls the other half cycle. Assuming that both thyristors are symmetrically controlled, it becomes possible to continuously vary the power to the load by changing the point in the AC cycle at which the gate “on” signal is given. This firing point is defined by the delay or trigger angle δ and the conduction angle α as shown in Figure 2.43. The “angles” relate to the rotating vector described in Figure 1.19. Figure 2.44 shows how peak, r.m.s. and average voltage varies with firing angle. It also shows how power output varies for a fixed resistive load. Note, however, that, a tungsten lamp does not have a fixed resistance, so a tungsten lighting power curve would be a little different. Nonetheless the fact remains that the majority of the power change happens between conduction angles 30° and 150°. Details of thyristor power control and firing circuits are given in Section 8.3.
The triac, also known as the bidirectional triode thyristor or the TRIode AC semiconductor switch is a five layer device as shown in Figure 2.45. It can be considered equivalent to a back-to-back thyristor pair in a single package. Because it can conduct current in both directions, reference to an anode or cathode would be meaningless, so the main terminals are designated MT1 and MT2. The triac can be triggered by current flowing into (+) or out of (-) the gate terminal. This results in there being four possible operating quadrants for the triac as shown in Figure 2.46. In practice quadrants 1 and 3 are the most sensitive. Quadrant 4 requires a higher
LIGHTING ELECTRONICS
2.4.4.5 The GTO and other special thyristors MT1 n
Gate
MT1 n
p
Gate
n p n
MT2 MT2
Figure 2.45 The triac.
gate current than the others, so is not used. Many practical circuits operate with quadrants 2 and 3, since then a negative going trigger pulse is used for both halves of the AC sine wave. While triacs are convenient low power control devices, they are not as robust as the equivalent thyristor pair. There can be problems in recovering blocking capability, especially at higher temperatures or with high dV/dt.
+
2
+
MT2+
1
The conventional thyristor and the triac are widely used for lighting control and their application is discussed in more detail in Chapter 8. There are a number of other “special” thyristors or thyristor-like devices, which have application in other areas. A brief summary of some of them is as follows: • The gate turn-off thyristor or GTO has a gate arrangement whereby positive current into the gate switches the thyristor on, and negative current taken from the gate switches it off. In the two transistor model, the arrangement can be thought of as being one where taking current from the gate breaks the regenerative cycle. GTOs have good forward blocking, but relatively poor reverse blocking capability. • The photothyristor uses light to trigger the thyristor – in effect a p-n junction is used as a photodiode to provide the gate current. Such thyristors are ideal for high power work in electrically noisy environments. At the other extreme small photothyristors have taken over from transformers as control isolation components in medium power applications (including dimmers.) • The insulated gate controlled thyristor like its transistor counterpart, has reduced gate drive requirements while maintaining low forward volt drop at high voltages.
2.5 Analog and digital 2.5.1 Analog
IG
-
-
IG
G-
G+
IG 3
+
MT2-
+
IG
Figure 2.46 The operating quadrants of the triac.
4
It is already clear that, for lighting control, electronics can do two things for us. Power electronics can be used for the direct control of electric power, generally by using high speed switching techniques. In turn the power electronic devices can be controlled by very low power electronic control devices or systems. These help make practical systems by, for example, allowing long distance remote control, or by bringing some “intelligence” to the control. A dictionary definition of “analogy” is the “agreement or correspondence between two things otherwise different”. In analog electronic control the
85
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
correspondence is between an electrical quantity, usually current or voltage, and some other physical quantity. An example already cited is that of an audio amplifier; where the chain is: • a transducer (in this case a microphone) which turns sound pressure level into a varying electrical voltage. The voltage is thus the analog of the sound. • an amplifier which increases the electrical signal. In this case it is clear that the amplification process must be one which does not distort the signal. • another transducer (in this case a loudspeaker) which turns the amplified electrical signal back into an audible sound. The power output of the amplifier can be controlled by a simple variable resistance at the input, which directly limits the strength of the signal reaching the amplifier. But a more sophisticated method would be to have a voltage controlled amplifier or VCA at the input. Such a device would then allow long distance remote control of the sound level, without the need to take the sound signal itself to the level regulating control. The small DC signal controlling the amplifier is an example of an analog control signal (representing the sound volume) controlling another analog signal (representing the sound signal itself.) In lighting control we can use the same idea; where a small remote control potentiometer carrying a tiny current, might produce a voltage variation of 0–10V representing the desired lighting level. This is then applied to an electronic dimmer, sited at a place most convenient for electrical power distribution, which does the actual power control using components like thyristors or IGBTs. 2.5.2 Digital The analog signal is continuously variable between limits set only by the nature of the apparatus concerned. Our 0–10V signal can, if the equipment is of sufficient precision, be set to any fractional voltage required, for example 5.467V. The nature of a digital signal, on the other hand, is that it can only adopt discrete values. For example, the 0–10V signal might only be available in integral voltages, 0, 1, 2, 3..V etc.
86
Despite the marketing man’s hype of all things digital, we actually live in an analog world. Our senses of sight, sound, smell and touch work on a continuous basis. In any practical digital system the divisions between the discrete levels must be so small that we cannot detect them. For example, on today’s CD (compact disc) 65,536 discrete sound levels are recorded between zero and maximum output in order to fool our ears that the sound pressure variation is continuous. In practice all digital electronic equipment works by using only TWO discrete levels. This has a number of advantages. In any form of signal processing, the analog signal can suffer in many ways: • electrical noise can be added to the signal, thus degrading its quality (resulting in noisy sound or flickering lights). • the signal can suffer from distortion. The description of transistors showed that their transfer characteristics are not always linear, so this is easy to understand. • the signal can suffer lack of precision due to the processing device. For example the transistor control circuit for a dimmer might give a different output at different temperatures (since all transistor characteristics vary a lot with temperature.) However, if only two signal levels are used, let us say 0V and 5V (OFF and ON; or 0 and 1), there is little likelihood they will be confused. If the detector of the ON signal is set to look for anything over, say 4V as ON, and anything else as OFF, it means that the cable carrying the signal can have a lot of noise on it without impairing the information carried. Distortion and precision (in level) are not an issue, because we are only looking for two quite discrete states. Many of the principles of digital electronics were understood in the 1920s and 30s; and early computers based on vacuum tubes were based on them. However it was the advent of the high frequency transistor and the integrated circuit which made digital electronics practical. The transistor’s ability to block a voltage on the one hand, and then to go to a saturated switched-on condition on the other, is the key to achieving the 0 and 1 conditions required by digital signals.
LIGHTING ELECTRONICS
2.5.3 Binary “Digital” is derived from digit, our fingers or toes. It implies a method of expressing numbers or counting to base 10. Our normal numbering system is referred to as being of decimal notation, where each “digit” in a number can assume only one of ten values 0–9. Although not wholly impossible, it would be difficult to make a precision electronic device able to adopt ten different states. With a transistor this might involve working the transistor in inefficient intermediate states between off and saturation, resulting in the dissipation of a lot of heat. To eliminate this kind of complication, digital electronics uses the binary notation for counting, where each digit can assume only one of two values 0 or 1. Binary digits are referred to as bits. Table 2.1 shows the relationship between binary and digital notation. It also introduces the idea of resolution. For some tasks low resolution is quite sufficient; for example many lighting control systems are based on 8-bit resolution (256 separate levels.) However, if moving lights are digitally controlled, and are working over a 180° arc, 8-bit resolution results in noticeable “jumps” as the device moves; after all, each step is nearly a degree. Therefore a higher resolution is needed. In the CD referred to earlier 16-bit resolution is used; a new generation of audio disc (based on DVD technology) is using 24-bit. Digital signals can be presented in two ways. Parallel data presents all the bits simultaneously on the requisite number of separate signal cables. Computers usually have a parallel bus architecture, for example 8, 16, 32 or 64-bit, requiring the corresponding number of connecting wires between each part of the computer. Data is transferred between the different parts of the computer in small chunks referred to as bytes. In small computer and microprocessor work, 8-bit bytes are used and references to memory capacity in bytes assumes 8-bit bytes. In a computer, however, the architecture could use a 32-bit parallel bus for either single 32-bit bytes, or for four 8-bit bytes. Parallel operation is fine for use within the confines of a computer. It requires very precise timing, so the bus structure includes a clock line. This car-
ries a clock signal which defines the moment at which the main bus should be examined for data. It is not practical to extend parallel signal buses for any distance, because the timing between different bus lines can vary so much. For this reason any longer distance communication is done using serial data signals. Here the bits are transmitted in sequence, with some identifying signal to show the start of each byte. Groups of bytes are referred to as words. It is now clear that all digital electronics is based on numbers. If we need to process any information, or measure a physical quantity, or control any device, we must do it by numbers. Examples of code jargon that is frequently encountered in electronics are as follows. Binary Coded Decimal or BCD is a coding technique sometimes used for input and output devices. Here any number is stored as decimal, with one 8-bit byte allocated to each decimal digit. Only the first ten binary numbers, 0000 to 1001 are used. This is obviously wasteful for computing, but is convenient for some applications. For example illuminated number displays, like those used on petrol pumps or train indicators, might use BCD for each decimal
Decimal and binary equivalents 0 0000 1 0001 2 0010 3 0011 4 0100 5 0101 6 0110 7 0111 8 1000 9 1001 10 A 1010 11 B 1011 12 C 1100 13 D 1101 14 E 1110 15 F 1111
Each power of 2 adds one bit 1 bit = 21 2 2 bits = 22 4 3 bits = 23 8 4 bits = 24 16 5 bits = 25 32 6 bits = 26 64 7 bits = 27 128 8 bits = 28 256 9 bits = 29 512 10 bits = 210 1024 For example “10 bit resolution” can represent 1024 different levels
Table 2.1 The relationship between binary and decimal notation. The letters A-F against decimal 10-15 are the hexadecimal equivalent (see text).
87
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
digit section of the display. Hexadecimal Code. Here all binary numbers 00000000 through 11111111 (0–256 Decimal) can be represented by a two digit number to base 16 (0– FF hexadecimal). Decimal 0–9 are designated as normal, decimal 10–16 are designated by letters A–F. Hexadecimal or “hex” is used by computer programmers as a shorthand for describing 8-bit bytes. It is much easier to recognize, and less prone to transcription error, than the equivalent binary string of 1s and 0s. ASCII Code (American Standard Code for Information Interchange). In its original form this represents: • the digits 0–9 • the letters A–Z upper case • the letters a–z lower case • punctuation marks • some symbols (e.g. +, $, %, @ etc). • 32 commands and identifying symbols. The symbols include items like STX or “start of text”. The commands include printer commands like CR or “carriage return”. All fit into only 7 bits (128 different characters.) So, as an example, the letter “Z” (upper case) has the code 1011010. (This binary number is also equivalent to decimal 90 and hexadecimal 5A.) ASCII code has been extended in both standardized and non-standard ways to cope with different alphabets and frequently used symbols. 2.5.4 Analog to Digital Conversion. ADCs and DACs Because the real world is analog many digital systems need a method whereby an analog signal is turned into its digital equivalent and vice-versa. For example a digital audio system must, at some point, create a signal suitable for connecting to a loudspeaker; or a light sensor detecting light levels may be required to convert its analog output signal into a digital signal suitable for computer processing. The devices required are called analog to digital converters (ADCs) and digital to analog converters (DACs). The ADC’s job is to turn a continuously varying
88
analog waveform into a series of numbers. The process is illustrated in Figure 2.47. The idea is to end up with a set of numbers which, if plotted as a graph against time, would reconstruct the original analog waveform. The ADC process is to sample the waveform at frequent intervals, and convert the measured signal level (usually voltage) to the nearest digital value which can be stored; a process called quantization. It is clear from the figure that if only a few samples are taken per second, and if only a few values are allowed (e.g. only 4-bit quantization), and then a graph is plotted from these figures, the result is a “staircase” version of the original. This may be satisfactory for some applications, but is usually unacceptable. The performance of the quantization process is determined by two factors. The sampling frequency and the quantizing level. The quantizing level is the same as bit resolution described in 2.5.3. The requirement will depend on the application – many bits for critical applications like audio, less for lighting and video. The sampling frequency determines the highest frequency that can be sampled. It is intuitive that if one wants to sample an audio frequency of, say 9kHz, it will be necessary to use a sampling frequency which is higher than 9kHz – but how much higher? Successive digital bytes OUT
Analog In
1 0 1 1 0 0 1 0
ADC
1 1 0 1 0 1 1 0
1 1 0 1 1 1 0 0
t
t 1 0 1 1 0 0 1 0
1 0 0 1 1 0 1 1
1 1 0 1 0 1 1 0
1 0 0 1 1 0 1 1
1 1 0 1 1 1 0 0
DAC
Successive digital bytes IN
Figure 2.47 Principle of the ADC and DAC.
Analog Out
LIGHTING ELECTRONICS
Signal IN analog
Read clock
Reference voltage
R 1 R 255 to 8 encoder
Potential divider
2
R 254 R 255
Digital out 8 bit
Differential comparators
Figure 2.48 One of the methods of creating an ADC. This is the “flash” converter.
Nyquist’s theorem shows that the the Nyquist frequency, the highest frequency which can be accurately sampled, is half the sampling frequency. Therefore a sampling frequency of at least 18kHz would be needed to convert the 9kHz audio. In fact, in the CD system, a sampling frequency of 44.1 kHz is used to achieve an audio bandwidth of 20kHz. In a lighting control system, where one might be using an analog device such as a slider or sensor, as input, the sampling frequency could be comparatively low. In a video system requiring digital video, the sampling frequencies are much higher (e.g. 13.5MHz). An ADC is a complex device which can only be realized practically as an integrated circuit. An example which is easy to understand (but uses the most circuitry) is the flash converter shown in Figure 2.48. The principle is that the incoming voltage is compared to a known voltage. An 8-bit (256 level) converter has a reference voltage source which is fed to
a chain of 256 resistors. Therefore the voltage at the end of each resistor is known. The device then contains no less than 255 differential comparators. These are not unlike the differential amplifier described in Section 2.3.2.5. They have two inputs, and in this case are designed only to give an output if the sampled voltage exceeds the reference voltage. As can be seen from the figure, the sampled voltage is fed to one input of all the comparators. The other input of each comparator is connected to one of the 255 reference voltages. The result is that, if a continuously varying waveform is fed to the device, there is a continuously varying pattern of “on” signals appearing at the outputs of the comparators. In fact each comparator is fitted with a switch controlled by a clock signal, so that the outputs are only “read” at the fixed sampling frequency. It is then a comparatively simple matter to convert the pattern of 255 comparator outputs to an 8-bit digital signal. The digital output is normally stored in some kind of output register, so it can be extracted when required by the device which is processing the data. The flash converter has been described because the concept is simple. “Slower” ADCs use “sample and hold” techniques (and much less circuitry) where the incoming voltage is momentarily sampled and “held” in a capacitor. The “held” voltage is then measured by using a single comparator and measuring the time taken for the capacitor to discharge to a reference voltage. The counter that does the timing produces a digital signal as output. DACs work on similar principles, but in reverse. For example by having available a number of binary weighted currents which are switched into a common load according to the digital signal. The selected currents are summed, and together represent the analog of the digital signal, Figure 2.49. This abbreviated description of DACs and ADCs is not intended to be definitive, but rather to illustrate how circuit building blocks can be used to meet a particular function. Both DACs and ADCs are widely used in electronics, but use significantly different constructions according to the bandwidth and accuracy required. Their use in lighting control is one of the less demanding applications.
89
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
R R
R
R
R
R
R
R
OUT ANALOG
2R
2R 2R
0
1
2R 2R
0
0
2R 2R 2R 2R
1
1 0
Transistor switches operate as 8 bit signal
importance in lighting. Some light sources cannot be continuously controlled; for example attempting to vary the light output of an LED by simply varying the applied voltage does not work very well. However, LEDs can be switched on and off very fast. Human eyesight has the attribute of persistence of vision whereby, above Modulating waveform
AM
1
Resistor network gives binary weighting to current flow
Reference voltage
Figure 2.49 One of the methods of creating a DAC.
Constant frequency carrier
2.5.5 Modulation The word modulation in an electronic context refers to the means by which source information is modified or encoded in order to match the means of transmission or storage. An easy example to understand is amplitude modulation as used by AM radio. An AM radio signal may be transmitted by an electromagnetic wave of, for example, 600kHz. Clearly a continuous transmission imparts no information, but by modulating the 600kHz carrier, information can be conveyed. In AM radio this is done by the simple expedient of varying the strength or amplitude of the carrier by the analog of the audio signal to be transmitted. The problem with AM is that the signal strength must vary in order to send the information. More efficient use of transmission power, and a better signal, can result from methods whereby the signal strength is constant. One way of achieving this is by frequency modulation or FM. It is intuitive that, for such a system to work, the carrier mean frequency must be much higher than the modulating frequencies. Another possibility is to use some form of pulse modulation. One frequently used example is pulse width modulation or PWM. This is used in Class D amplifiers (Section 2.3.2.2). Again the pulse frequency must be much higher than the frequency of the signal to be transmitted. PWM is of considerable
90
Frequency varies as moulating signal
FM
Constant amplitude carrier
Original waveform PAM
t Pulse amplitude corresponds to instantaneous value of original waveform, sampled at interval t
1
0
- Measure amplitude at intervals - Convert to binary code - Send as serial digital signal
PCM
0
0
1
1
0
1
0
t 1st sample
Figure 2.50 Methods of modulation.
1
0
1 1
2nd sample
LIGHTING ELECTRONICS
a certain frequency, our eye/brain combination integrates successive light stimuli. This is the basis on which film and television works, whereby, for example, 30 frames are presented each second and we see a continuous image variation. The same principle can be used to, apparently, vary the intensity of light. Clearly if the pulse frequency is too low, we just see successive flashes of light. At intermediate frequencies we see flicker which is extremely uncomfortable. The subjective effect of flicker is not only dependent on frequency, but also on other factors such as ambient light level, contrast etc, but once the switching frequency is high enough, we simply see an apparently steady source of illumination. For lighting control purposes the effective use of PWM depends on the source being used, but would usually need a minimum frequency of 120Hz. An intermediate system of modulation is pulse amplitude modulation or PAM where the source analog waveform is sampled at discrete intervals. Instead of transmitting the complete analog waveform, sample pulses only are transmitted. In fact, such a method is rarely used, but it does form the basis of the commonly used pulse code modulation or PCM. In PCM the magnitude of each pulse is converted into a binary number – exactly the process already described for the ADC. Once we are in the digital domain, all kinds of modulation systems become possible; for example delta modulation uses only the differences between successive amplitude values in order to reduce the required bandwidth. Highly sophisticated methods are used to eliminate errors, and ingenious methods used to increase the data throughput of any transmission channel. But most digital modulation systems start with PCM, whatever method is then used to attach the resulting signal to the carrier medium.
Volt). Even more confusing, it seems to turn up in unlikely places. Most people are aware of it as a measure of sound, but, less obviously, it turns up as a measure of radiated signal strength and as a measure of performance of data cables, both of which are relevant to lighting control. Sound levels may also be relevant, especially if control equipment is fitted with noisy fans, or emits a buzz. The decibel or dB is a measurement of ratios. It was originally applied to sound where, because our hearing is logarithmic, it was easier to express changes in sound levels using the logarithm of the ratio of two powers. It is best to start with the acoustic definition. The Bel was the original unit, where: B = log10 (P2/P1) and P1 is the reference power P2 is the new power. log10 is the common (base 10) logarithm, where numbers are expressed as their equivalent as the power of 10. In practice the Bel was found to be too big for sound intensity (defined as the energy passing through one square meter normal to the direction of propagation) so the unit used is the decibel where: Sound Intensity in dB = 10 log10 (P2/P1) Here the power is in Watts/m2, and the reference power P1 is taken as the threshold of hearing, the lowest sound intensity which a healthy young human being can detect. It is usually taken as 10-12 W. Sound intensity is a difficult quantity to measure, so it is more usual to measure sound pressure level or SPL. But this is proportional to the square root of the intensity, so the equation becomes: SPL in dB = 20 log (ρ/ρr)
2.5.6 The decibel A unit frequently encountered in electronics and audio is the decibel. It is often wrongly treated as though it was a physical quantity, which can be measured in physical terms related to fundamental units (like a
Where ρ is the pressure and ρr is the reference pressure corresponding to the threshold of hearing – taken to be 20μPa. Table 2.2 gives some examples of sound pressure levels. The log10 of one is zero (i.e. 100 equals 1) so when ρ = ρr = 20μPa, we have a reference SPL of
91
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
SPL in dB 0 20 40 50 60 70 80 90 100 110 120 140
Actual Example SP 20PPa Threshold of hearing 200PPa Background in recording studio. Quiet countryside. 2000PPa Whisper at 2m Average suburban area 6310PPa Conversation at 1m General office 0.02Pa Restaurant, store 0.063Pa Radio/TV in the home 0.2Pa Busy street 0.63Pa Heavy truck at 6m Symphony orchestra fff 2Pa Disco 6.31Pa Pneumatic drill at 1m 20Pa Modern aircraft at takeoff 200Pa Military jet at take off at 30m Threshold of pain
Table 2.2 Sound pressure levels.
0dB. It is now easy to see why the logarithmic method of measurement is convenient. The 140dB range of sounds covers our normal experience, with quite easy reference points over a reasonable scale – especially when you consider that we can only just detect a 2dB change in SPL. However 140dB represents a range of sound pressures with the ratio 10,000,000:1. Note that without a reference level, a dB figure cannot give us an absolute value for a quantity. As with the human eye, the human ear does not hear all frequencies equally. The phon is the unit of loudness. At 1kHz one phon is equivalent to 1dB SPL; so if SPLs are only measured at 1kHz, the loudness in phons is the same as SPL in dB; but at all other frequencies they are different. The phon is, therefore, analagous to the lumen, in that its value depends on a human attribute (sensitivity to wavelength of sound, instead of wavelength of light). The fact that our ear response to sound varies with frequency means that using linear measurements of sound is misleading. Noise and sound pressure level measurements are weighted to be meaningful. The
92
most commonly used weighting curve is the A weighting curve shown in Figure 2.51. Measurements made using it are referred to as dBA. Where noise is an issue for lighting control equipment (e.g. in ballasts) it is usual to see the noise figure quoted in dBA. Early designers of audio amplifiers realized that when analog electric currents and voltages were used to represent sound, the dB notation was useful. In this way the effect of, for example, a 6dB change in signal level, gives a good idea of the aural effect, in a way which a statement of increase in voltage or power would not. When dB are applied to electrical quantities the relationship only holds good if the circuit impedance is the same for both values. Also, as with the acoustic dB, the relationship is Power ratio in dB = 10 log10 (P2/P1) where P1 is the reference power; but, because both current and voltage are proportional to the square root of the power, voltage or current ratios are defined by: Voltage ratio in dB = 20 log (V2/V1) Current ratio in dB = 20 log (I2/I1) A Weighting dB +10 0 -10 -20
-30 -40 -50 -60 10
20 30
50
100
500 1000
5000 10000 Hz
Figure 2.51 The A weighting curve for acoustic measurements.
LIGHTING ELECTRONICS
The dB has been found to be useful for electrical and electromagnetic quantities other than those directly related to sound. So it is common to see figures for the attenuation of all kinds of signals down a cable being referred to in dB. Likewise measurements of electric field strength are conveniently referred to dB. “Electrical dB” must, like their acoustic counterparts, have a reference level. Some of these are shown in Table 2.3.
2.6 The integrated circuit and ASIC 2.6.1 Introduction Many electronic circuits are made up of discrete components. Often, in order to get a workable circuit, the circuits become quite complex. This is because a realworld circuit needs additional components; for example to compensate for variations in temperature, to eliminate the effects of voltage spikes, or to stabilize a power supply. When high powers are involved it is generally necessary to use discrete components, but when a circuit is only required to operate at low power, the idea of an integrated circuit becomes realizable.
Notation dBV dBmV dBPV/m dBW
dBm
Reference Level 1V (for voltage or current a 6dB change is approximately u2, and a 10dB change u3) 1mV 1PV/m (electric field strength) 1 W (for power a 3dB change is approximately u2, and a 10dB change u10) 1 milliwatt (in audio circuits dBm is usually referred to 600 ohms. At this impedance the equivalent Voltage 0dB is 0.775V)
Table 2.3 Some “Electrical dB” with comments.
The description of the MOSFET in Section 2.3.3.3 indicated how this might be possible. The MOSFET power transistor consists of hundreds or thousands of individual transistors all operating in parallel. But a similar technique can be used to create individual transistors, diodes, zener diodes, resistors and capacitors together with a circuit interconnection pattern. Within integrated circuits, capacitors and resistors of defined value are difficult to make. For this reason many applications of ICs require external passive components for setting precise operating parameters. Also there are many cases where it is easier to make an apparently more complex circuit, using many transistors and few or no passive components, than it is to make a simpler circuit using fewer transistors with passive components. Early integrated circuits were comparatively simple, providing a simple “building block” with the equivalent of only a few transistors. Such circuits are still an essential part of everyday electronics but, as production techniques have improved, the possible component density within integrated circuits has reached a previously unimaginable level. The largest integrated circuits consist of millions of transistors. A broad classification of integrated circuits, or ICs, is as linear and digital ICs. Linear ICs are mainly amplifiers, and can be considered analog devices. Digital ICs consist mainly of gates of different complexity, where the signal paths of the circuit assume only one of two binary values, high or low (also expressed as 1 or 0). Most references to ICs are to monolithic ICs, where the IC is made on a single chip of silicon. The chip is then packaged in a hermetically sealed plastic or ceramic package which includes the lead out connections. For some applications much larger hybrid ICs are made. These consist of an assembly of several chips and other components mounted on a ceramic substrate which is then sealed as a whole. Hybrid ICs are widely used in applications requiring both low level control signals and power signals; for example medium power audio amplifiers and engine management electronics. They also have a place in lighting control.
93
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
+ Supply typical +15 V
Inverting input Output
+ Non-inverting input High resistance typical >2M :
Low resistance typical 75 : Supply typical -15 V
Open loop gain typical 200,000
Figure 2.52 The operational amplifier or op-amp.
2.6.2 Linear integrated circuits 2.6.2.1 Operational amplifiers Depending on the application, linear ICs are based on either bipolar or unipolar transistors. The best known example of the linear IC is the Operational Amplifier or, more familiarly, op-amp. The name comes from its original discrete component ancestor, used to describe amplifiers carrying out mathematical operations in analog computers. The op-amp is a high gain differential amplifier. Figure 2.52 shows the symbol, and it is easily identified as the basis of the comparator in Figure 2.48 and the output amplifier in Figure 2.49. An op-amp should have an output of 0V when both its inputs are at equal voltages, but in real world op-amps there is usually an input offset voltage to achieve 0V. Some op-amps have an input null facility to ensure that the 0V output condition is given when the inputs are equal. Most op-amp circuits use negative feedback. As explained in Section 2.3.2.2 this makes the gain of the resulting circuit independent of the open loop gain of the amplifier, which, in the case of op-amps, is very high. Figure 2.53 shows how an op-amp can be used as an inverting amplifier (where the output is of opposite polarity to the input) and as a non-inverting amplifier. In the case of the inverting amplifier having its
94
(+) input set at 0V, feedback causes the circuit to stabilise with the (-) input being at 0V too. The circuit action is as if there is a direct path between the (-) input to the 0V rail. The (-) input of the op-amp is said to act as a virtual earth (ground). Apart from straightforward signal amplification, op-amps form the basis of many useful circuit elements. A few examples are: • as a voltage adder. In Figure 2.49 an op-amp produces an output equal to the negative of the sum of the input voltages in a virtual earth circuit. • as a subtractor, or straightforward differential amplifier, where the output is the difference between the two inputs. • as a voltage follower (similar to the emitter follower) where the voltage gain is unity. This is used as a signal buffer, with very high input resistance, and low output resistance. Its advantage over a transistor follower is that the output voltage is exactly the same as the input, whereas there is always an offset with a simple transistor. • as the basis of a constant current circuit. Here the op-amp controls the base current to a transistor. RF RA
VIN
VOUT
RB
+
voltage gain
0V
RF RA
Inverting amplifier
RF
VOUT VIN
RB
+ voltage gain RA
R A 000e RF RA
0V Non-inverting amplifier
Figure 2.53 Operational amplifier circuits. In this case the power connections have been omitted to simplify the diagram.
LIGHTING ELECTRONICS
15V RE
R
Constant current circuit
+ R
RL
0V
IOUT constant even if RL changes
15V R
Inverting Schmitt trigger
VIN
VOUT
VT
+ RF
R
When VIN=VT
Positive feedback
0V
2.6.2.2 Phased locked loop
First order active filter using non inverting op-amp RF VIN
VOUT R
+ C
0V
RA
R
threshold voltage. Such circuits are essential in digital applications where an input signal may be degraded or noisy, and where it is essential to achieve a definite “on” signal when a threshold is exceeded (and a similar “off” signal as the input reduces). Schmitt triggers can, of course, be made from discrete transistors, but the op-amp version can be designed with great precision in respect of threshold voltages and hysteresis – in this case the difference between the on and off triggering voltages. Schmitt triggers use positive feedback in the op-amp circuit to achieve the snap action. • as the basis of active filters. Section 2.1.3 described how reactive components could be made the basis of filters. In small signal work, high pass, low pass and band pass filters are often needed, and opamps can be used as the basis of active filters. These can be based on simple resistance-capacitance networks to select the frequency pass bands. They use selective negative feedback to eliminate the unwanted frequencies.
Another common linear IC is the phase locked loop, or PLL. This device is sometimes called the electronic flywheel, because it maintains an accurately timed output even if the input has some jitter or is occasionally discontinuous. It is particularly useful for maintaining accurate clock signals in digital systems. In serial data systems it is not practical to send
-3dB frequency fc then -6dB per octave
1 2 SfcC
INPUT Nominal f
Phase Detector
Amplifier + filter (damping)
Divide by N
VCO
Figure 2.54 Some applications of the op-amp.
Negative feedback ensures that the emitter current is constant, which in turn ensures that the current flowing in the collector load is constant – even if the load resistance changes. • as the basis of Schmitt trigger circuits. The Schmitt trigger is a device which produces a fast “switch” output when an input signal exceeds a set
OUTPUT F locked
OUTPUT NxF
Figure 2.55 The phase locked loop or PLL.
95
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
a separate clock signal, so it may be necessary to derive a clock signal from the transmitted data. The PLL can do this. At its heart is a voltage controlled oscillator or VCO; this device is an electronic oscillator whose frequency is determined by a voltage control signal. The control voltage in this case is derived from a phase detector. This compares the phase of an incoming signal with that of the oscillator. The complete PLL provides a device which ensures that the output is at the same frequency and same phase as the input should be. The output is at a constant amplitude regardless of the input amplitude. By using an arrangement whereby the VCO operates at a multiple of the input frequency the flywheel effect is increased, so the oscillator continues to provide an output even if there are discontinuities at the input. 2.6.2.3 Signal switches Many control applications require devices which can route control signals. In audio-visual applications these could be audio or video signals (either digital or analog) and in lighting control systems it could be the routing of serial control signals or analog control signals. The idea that a transistor can be used to do this, as opposed to a relay, has already been introduced. The most common arrangement is the CMOS transG Negative gate signal p-channel MOSFET VIN
VOUT
mission gate. CMOS construction is described in the next section. Each transmission gate consists of a complementary pair of MOSFET transistors which together constitute a bi-directional switch. Control is by complementary gate signals applied to the two transistor gates. Note here the need to distinguish between the “gate” electrode of a transistor, and the function of a circuit block as a logic “gate”. It should be clear from the context which is being referred to. Such switches are usually supplied as integrated circuits containing many such transmission gates. These may be independent but, more usually, the circuit is a multiplexer where either many input signals can be routed to a single output, or a single input can be routed to multiple outputs. Such devices form the basis of electronic patch panels. The name derives from old telephone practice where exchange operators would manually “patch” one circuit to another by inserting cords with jack plugs into a routing panel. With integrated circuits of this kind important attributes are the fan out and the propagation delay. Fan out describes the number of connections which can be made at the output. Ideally it should be possible to connect as many as required, but this obviously depends on the loading. For this reason circuit elements take advantage of the MOSFET properties of high input impedance (meaning that they present a small load to the signal source) and low output impedance (meaning that they can feed many high impedance loads). In practice the input impedance is limited by capacitance, meaning that the loading is heavier than would otherwise be expected. Clearly the problem gets worse as the signal frequency increases. Capacitance is also primarily responsible for a slight delay in any signal going through the gate, so propagation delay can become an issue when a signal has to pass through several of them.
Symbol n-channel MOSFET G
Positive gate signal
Figure 2.56 The CMOS transmission gate makes a bidirectional switch. Here simplified symbols are used for the MOSFETs. See also Figure 2.66.
96
2.6.3 Digital integrated circuits 2.6.3.1 Logic gates An extended description of binary logic operations is beyond the scope of this book. Only a simple example is described here, but from the example it
LIGHTING ELECTRONICS
Name and Logic Symbol
NOT _ A
AND AxB
NAND (= NOT AND)
___ AxB
OR A+B
NOR (= NOT OR)
___ A+B
EX-OR (= Exclusive OR)
A† B
EX-NOR (= Exclusive NOR)
____ A†B
Circuit Symbol
Truth table A 0 1 A B 0 0 0 1 1 0 1 1 A B 0 0 0 1 1 0 1 1 A B 0 0 0 1 1 0 1 1 A B 0 0 0 1 1 0 1 1 A B 0 0 0 1 1 0 1 1 A B 0 0 0 1 1 0 1 1
Z 1 0 Z 0 0 0 1 Z 1 1 1 0 Z 0 1 1 1 Z 1 0 0 0 Z 0 1 1 0 Z 1 0 0 1
A S B
C0
Figure 2.58 The half-adder.
for the most common gates are shown in Figure 2.57. This shows that, for example, an AND Gate with two inputs has a 0 output if any of the inputs are 0; and ONLY has a 1 ouput if both inputs are 1. Using a combination of gates it is possible to carry out both simple and complex mathematical operations. The idea can be understood with reference to the commonest arithmetical operation, that of adding two numbers. Figure 2.58 shows a half adder which can add two inputs A and B. The “Sum” output S is determined by the Exclusive OR gate, which means that it will be 1 if A or B = 1; but will be zero if both A and B are 1 (or both 0). The parallel AND gate will give a “Carry” ouput CO only if both A and B are 1.
A B
Ci
S
Figure 2.57 The common logic gates shown for two inputs A and B. Z is the output.
should be possible to understand how much more complex arrangements could be built up. Logic gates obey very simple rules. All inputs and outputs can only adopt one of two values 0 or 1. The different type of gate, for example AND Gate, OR Gate etc. have different rules which can be expressed in a truth table. The truth tables and symbols
C0
Figure 2.59 The full adder.
97
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
+V
S 1
0
1
J Clock In
Q
J
0
1
Q
J
Q
J
C
C
C
C
K
K
K
K
y2
1
R
Q
y4
y8
y16
Outputs
Q
Figure 2.62 Linking JK flip-flops together to make a
Figure 2.60 The SR flip-flop or bistable.
This arrangement is called a half adder because it cannot deal with a carried input from a previous stage of addition. The logic arrangement gets more complicated for a full adder. Figure 2.59 shows that one way to achieve it is by using two Exclusive-OR gates and four Not-AND or NAND gates. The logical expressions which describe the operation of gates are called Boolean Algebra. Each expression is quite simple, but highly complex operations can be carried out by using multiple gates. Integrated circuits are available that provide the common gate functions. Often there are several gates on one chip, or one chip provides a complete arithmetic function. Thus one can obtain a quad 2-input AND IC, which means that it contains four 2-input AND gates; or a 4-bit adder with carry. The manufacturing process used for such chips tends to favor the production of one kind of gate over another. It is easier to construct an IC with a large number of gates of a single kind than it is to construct one with gates of different J Q
4-bit counter/divider. By setting J = K = 1, each flip flop “toggles” on successive clock pulses. The arrangement can be used to divide down, or count the clock pulses. kinds. By using Boolean Algebra it can be shown that the half adder of Figure 2.58 can be made using five NAND gates. While this is more gates than the two-gate arrangement, it is actually easier to make as an IC. Note that gates are not confined to two inputs. A three-input AND gate will only give a 1 output if all three inputs are 1. A four input OR Gate will give a 1 output if any of its four inputs are 1. A special case arises when there is an element of feedback in the logic. This leads to monostable, bistable and astable elements. Of particular significance is the bistable element, often referred to as a flip-flop. Flip-flops can be made using a number of different gate combinations; an example that is easy to understand is the SR flip-flop shown in Figure 2.60 The Set-Reset flip flop can be considered as a four terminal device, with inputs S and R, and outputs Q and Not-Q. In the example using NAND gates, S and
Change only on falling clock pulse
Parallel Outputs
Serial data IN
Clock Q
D
D
D
D
C
C
C
C
K J=K=0
Q
1
J=0 K=1
Q
0
J=K=1
Serial Data Out
No Change
J=1 K=0
Clock
Outputs change to opposite sense
Figure 2.61 The JK flip-flop.
98
Q
Figure 2.63 A shift register made up from D flip-flops.
LIGHTING ELECTRONICS
R are normally high (1) and initially Q and 0003Q are respectively 1 and 0. If R goes low (0) the device momentarily goes unstable, but settles when the gates’ logic truth tables are again satisfied resulting in Q now being 0, and 0003Q being 1. The device stays in this state, even when R reverts to being 1. It can be seen that successive low (0) inputs on S or R cause the device to successively adopt one or other of the stable states. In practice flip-flops are used as the basis of registers and counters. Their bistable nature allows them to store numbers. However, when a number of them are used together, there can be timing difficulties, so practical flip-flops are more complex. An example is the JK flip-flop shown in Figure 2.61. Internally this consists of two sets of flip-flops, but externally can be considered as a clock controlled bistable. The idea is that the flip-flop can only change state internally on the rising clock-pulse, and whether it does so is determined by the state of the J and K inputs at the time. Once the change has been made, it does not matter that the J and K states then change again. However no output change (if there is to be one) is given at Q until the falling edge of the clock pulse. Whenever Q changes 00030003 Q changes to be opposite to it. While the JK flip-flop introduces a delay, the delay is precisely determined by the clock. Thus when numbers are transferred between one set of flip-flops and another, it is always done with precision, with no danger that the temporary instability of an individual flip-flop as it changes over could affect the process. Figure 2.62 shows how JK flip-flops can be joined together to make a counter. Within integrated circuits it is possible to use simpler flip-flops called D or Data flip-flops. Here the output changes on the rising edge of the clock pulse (i.e. as the clock goes high.) It only changes if there has been a change in the data input since the last clock pulse. A common requirement is the shift register shown made up from D flip-flops in Figure 2.63. Here data shifts one step along the chain each time there is a clock pulse. It is clear that such a register could be used in several ways, for example:
• to convert a serial data stream into a parallel stream. In this case an 8 bit word could be clocked in to the register, and the parallel ouputs read once the whole word was in place. • to delay a serial data stream by a set time. In this case the output would be taken from the serial output shown, but the number of flip flops used would correspond to the number of clock cycles of delay required. 2.6.3.2 Logic gate construction In the above section the function of gates was described in logic terms, without reference to how the gate is actually made. It is quite possible to work the other way and, for example, to describe the electronic action of a circuit using two transistors, and to show that such a circuit is bistable. As already intimated, the actual logic gates created in integrated circuits are partly determined by the manufacturing process involved. Figure 2.64 shows that there is potentially a wide choice - and in fact the diagram has been simplified to show only the main families. The three main technologies are: • TTL or Transistor Transistor Logic. • CMOS or Complementary Metal Oxide Silicon. • ECL or Emitter Coupled Logic. TTL was the first logic family. It is fast, but is quite critical in its power requirements in that the supply must be kept within very close limits. For many applications it has given way to CMOS, as the
Unipolar FET
n - MOS
Bipolar
CMOS
Bi-CMOS Hybrid
Saturated
UnSaturated
TTL
ECL
Figure 2.64 The technologies available to make digital integrated circuits.
99
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
INPUT A
Q1
+ VCC
Q3
Q2
INPUT
VDD
A B
Output Z = A B
B Output Z = A B Q4 0
0V
Figure 2.67 The NAND gate executed in CMOS.
Figure 2.65 The circuit of a TTL NAND gate.
latter’s performance has improved. Many ICs originally offered in TTL are now offered as “TTL compatible” CMOS. An example of a TTL gate is shown in Figure 2.65. This is a NAND gate, such that output is 1 unless both inputs are 1. Q1 is a transistor with two emitters. If both are set high (1) Q2 is switched on, resulting in Q3 being forced off, and Q4 on. The output is then low (0) fulfilling the NAND condition. If either input is low, Q1 conducts, the base of Q2 goes low, switching it off. The states of Q3 and Q4 are reversed, making the output high (1). The double transistor output arrangement is used to ensure that rise and fall times of the output are minimized – clearly the NAND condition itself is already satisfied earlier in the circuit. The complete circuit helps make the point that practical circuits are more comG
Polysilicon Gate D
S
n
G S
n
SiO2
p
p n well
p
n channel MOST
P Channel MOST
Figure 2.66 The structure of CMOS transistors. Although a common substrate is used, the use of “wells” allows both types of MOSFET to be made together. In this figure the drains are shown linked, as they would be in the CMOS inverter shown in Figure 2.68.
100
plex than might be expected. CMOS is by far the most widely used technology for general purpose application. It can work at a wide range of voltages; from very low, making it suitable for battery powered equipment, to relatively high, making it suitable for signals in electrically noisy environments. CMOS gets its name from the fact that both nchannel and p-channel FETs are made on the same substrate. There are alternative ways in which this can be done, but the most common is the use of wells. For example n-well CMOS is made by starting with a p substrate and creating on it a number of n areas or wells. Within each well area a p channel FET is made; whereas on the ordinary substrate n channel FETs are made in the normal way. See Figure 2.66. A NAND gate based on CMOS has the outline circuit as Figure 2.67. This shows two n channel transistors in series, tending to pull the output down to 0, and two p channel transistors in parallel tending to pull the output up to 1. If both inputs are 1, the two series transistors switch on and connect the output to ground (low or 0). The other transistors will be off because their gate-source voltage is zero. In all other cases the output is high. At least one of the series transistors will be switched off, and at least one of the parallel transistors will be switching the output to VDD (high or 1). A simple CMOS element is the NOT gate or inverter. This can be achieved with a single pair of complementary transistors. In Figure 2.68 if the input is low (0) the effect is the p channel transistor switches on raising the output high (1). When the input goes high, the switching arrangement is reversed. Now the
LIGHTING ELECTRONICS
+VDD
S p channel VIN A
VOUT
D D n channel
A
S
0V
Figure 2.68 The CMOS inverter or NOT gate.
p channel transistor switches off because its gate and source are both at VDD, and the n channel transistor switches on bringing the output low. This simple element forms the basis of a wide range of computer ICs because joining two inverters together creates a bistable. The principle is shown in Figure 2.69. In the complete D type flip-flop (or latch) additional transmission gates are shown. One provides the feedback connection to give the bistable action. The other lets in the new data. The transmission gates are clocked in antiphase, so the feedback loop is only disconnected when new data is clocked in. ECL is the fastest logic family. It operates bipolar transistors in an unsaturated regime to achieve very fast switching times, but requires high power. It is used in fast computers, but is not likely to be encountered in lighting control applications. The principal circuit element used is the long tail pair described in Section 2.3.2.5. 2.6.4 Application Specific Integrated Circuits Integrated circuits have evolved from circuits carrying a few gates and simple amplifiers to LSI and VLSI (Large Scale and Very Large Scale Integration) construction. A huge range of standard ICs is available; from gate assemblies to memories and complete computer devices (which are both reviewed in the next section). Many electronic designs can be realized using
“off-the-shelf” ICs in combination. However the result may not be cost-effective, or may represent a compromise. When large quantities, or special performance, are required, it may be better to use an ASIC or Application Specific Integrated Circuit. ASICs can be made in many forms. They can be digital, analog, mixed signal or hybrid. They can use different forms of construction depending on whether they are logic devices, signal processing devices or devices for power control. Section 2.3.3.2 gave a brief overview of the processes used in the manufacture of transistors. To the processes described there must be added ion implantation as a means of introducing the p and n impurities. This method is more complex than diffusion, but can give better distribution of the dopant which becomes very important at the the miniscule dimensions of transistors within an IC. Ions are simply atoms of an element with either an electron missing (positively charged ion – for example B+, a positively charged Boron atom for producing p silicon) or an additional electron, a negative ion. Ion implantation consists of placing the silicon wafer in a vacuum chamber and directing a beam of ions at it. Since the ions are charged, they can be attracted by an opposite charge. The beam can be controlled, and the dosage measured by measuring the ion beam current. Feedback
0 or 1
0 1 1 0 Principle of bi-stable Gate switches feedback loop
C Clock C
Bistable D Data IN
Gate switches data
Q OUT Q
Figure 2.69 The D flip-flop constructed from CMOS inverters.
101
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
The production of ICs is, therefore, highly complex. Typically between 10 and 14 major process steps are involved in their production, each with many supporting steps, and each requiring very high repeat accuracy. Commodity ICs are cheap only because they are produced in huge quantities, and because one silicon wafer can carry maybe hundreds or thousands of identical chips. ASICs are produced as custom ICs and semicustom ICs. Custom ICs are designed from scratch and require a full knowledge of the semiconductor manufacturing process. Custom ICs of any complexity are extremely expensive in terms of their NRE or Non-Recurring Engineering charges, and in terms of development time. For this reason most ASICs are semi-custom. In this case the user does not have to know in detail how the device works. The manufacturer offers the user a set of cell libraries and a CAD (Computer Aided Design) program to enable him to design the ASIC round the standard cells. The “cells” are commonly used circuit elements. these could be as simple as a single transistor – but actually are more complex items, for example complete gates, bistables, multiplexers and input/output circuits. IC production involves the use of many different masks in the lithographic process, and the idea of semi-custom ASIC production is that several stages of manufacture are common to many different designs. The custom element of the design can often be restricted to the masks needed to create metalization layers providing the interconnection between circuit elements. Cell based designs have the cells in rows with channels between them. It is in the channels that the interconnections are made. For logic applications the customizing is made simpler in two ways. Most logic applications simply require standard gates; so in this case a gate array is produced. Large numbers of standard gates are produced in the lower level of the silicon, with the channel interconnections requiring only a few process steps. The idea is taken further by including through connections or vias in the gate construction. The idea is a miniature version of how multi-layer printed circuit boards are made. When this is done, there is no
102
need to leave space for the interconnection channels, the resulting channel-less construction sometimes being referred to as a sea of gates. The interconnections are now on the top layers of the IC and are the only customized part of it. Where complex logic is required ASICs have, to some extent, lost out to the programmable devices described in Section 2.8. However, they are still very effective in high volume products of medium complexity, and are likely to be found in some lighting control products.
2.7 The microprocessor 2.7.1 Computers The idea of computers is not new. The first electrical or electronic computers or calculators were designed to carry out a fixed task, and were in some case successors to earlier purely mechanical devices – for example the simple cash register. The significant development, which occurred mainly in the 1940s, was the idea that electronics could not only make calculation much faster, but would allow the introduction of the programmable computer. An early office calculator is an example of a fixed program device. In it the data, the sums of money being entered, could be variable, but the processes applied to the data were fixed; for example simply adding up money totals. In a programmable calculator or computer the operations to be carried out on the data are themselves changeable or programmable; so the device not only works with stored data, it works under stored program control. So, for example, the data might be treated differently according to different input parameters. In a period of only a few years computers have become commonplace to the point where it is unlikely that anyone reading this book does not have access to one. For this reason it is easiest to start with describing what, until recently, would have been considered a very powerful machine. Today’s laptop computer has considerably more computing power than an office main frame computer of 20 years ago.
LIGHTING ELECTRONICS
Monitor
Graphics controller I/O controller
Hard disk drive
Floppy disk drive
CD-ROM drive
DVD drive
HDD controller
FDD controller
CD drive controller
DVD drive controller
CONTROL BUS
DATA BUS
ADDRESS BUS
Microprocessor CPU
Address bus register
Internal bus Serial Parallel I/O ports
Program Stack counter pointer
Keyboard controller
Instruction and status registers
Data bus register
A A A A A A A A A A A A A A A A A A
Keyboard
A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A
Mouse
A A A A A A A A A A A A A A A A A A A A A A A A A A
Accumulator
General register
Control
Clock
mouse
Figure 2.70 Block diagram showing the principal components of a typical personal computer system.
Figure 2.70 is a simplified block diagram of a typical personal computer system. Within a computer most data is transferred by a parallel bus, since this gives a higher speed, and is practical over the short distances involved. Usually the data is carried over more than one bus, typically: • a control bus used by the central control to give control commands to the peripheral devices. • a data bus which carries data relevant to the application. This could be data being read off a hard disc drive, for loading in to the Random Access Memory or RAM. • an address bus. This carries the addresses of memory locations, much like a postal address. If the central control places an address (for example a location in the RAM) on the address bus, then the data currently on the data bus gets read in to it. The keyboard and mouse are the main methods by which the user communicates with the computer. These create alphanumeric, symbolic and positional data. Some applications benefit from additional or alternative input devices; for example a joystick or a
graphics tablet. For many applications, especially in the fields of automation and even lighting control, it is necessary for the computer to be able to communicate with other external devices, and it does this through the Input-Output or I/O Port. In a personal computer the I/O port typically provides two serial ports and a parallel port. The parallel port is usually used by a printer. The serial ports are usually EIA32 (see Section 9.3) although for special purposes other ports, for example the IEE1394 Firewire® high speed digital video port, might be fitted. Many computers are also fitted with USB (Universal Serial Bus) ports that provide much higher data transfer rates than the “traditional” serial ports. All activities within the computer are closely regulated by the central clock. Every time data is transferred, or an address read, it is always done synchronized to a clock cycle. This way it is known that a byte of data on the data bus has been transferred only to the address on the address bus, and has not been cut in half! But the outside world is not synchronized to the
103
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
clock. So the I/O port is fitted with a Universal Asynchronous Receiver Transmitter or UART. The UART is a special type of data register. It receives data from the outside world and stores it until it receives the instruction to put it on to the data bus. Similarly for outgoing data it receives data from the data bus, and sends it out at the correct speed for the external serial data link. The computer is fitted with several memory devices of varying capacity and speed of access. The ones identified in the figure are: ROM or Read Only Memory. This is an IC with permanent data which cannot be changed. It is used to store the program which starts up the computer when it is switched on, and various other housekeeping routines. The start-up routine is sometimes referred to as the bootstrap or booting routine (derived from the expression to “pull oneself up by one’s own bootstraps”). RAM, already referred to, is the computer’s working memory. It is a collection of memory chips that allow data to be transferred very quickly. The RAM will normally hold both the current application program and its associated data – but if it is not big enough, data must also be exchanged with the hard disk. In the 1960s a main frame office computer was considered powerful if it had 8 kilobytes of RAM, resulting in an enormous amount of data shuffling between the RAM and external tape stores. Today even a modest personal computer may have 128 Megabytes of RAM or more. Hard disk drive. This device stores large quantities of data; for example all the different applications programs currently needed by the user, and all their related data files. Data is stored magnetically on one or more magnetically coated discs, using the same principles as magnetic recording tape but with much higher storage densities. 20 years ago a hard disc drive storing a few MB was very expensive; today a drive storing several GB is standard. Access to data on the hard disk drive is not as fast as it is to RAM, but is still respectably quick, whereas access to data on any kind of external drive is comparatively slow.The hard disk drive is a permament part of the computer and, except in special applications such as video servers, is not usually exchangeable since the
104
storage medium is an integral part of its construction. The Floppy Disk Drive on the other hand is intended for data exchange, allowing computer files made on one machine to be used by another, and allowing data to be backed up so it can be recovered in the event of computer failure. Typical floppy disks have a capacity of 1.44MB. While in the early days of personal computers a floppy disk holding a few hundred kB was considered adequate, it is now the case that floppy disks are suitable for simple alphanumeric data, but inadequate for applications involving graphics and mixed media. Many types of high capacity drive exist, based on either magnetic tape or on some form of optically read disc. The CD-ROM drive is currently the most common large capacity device for data exchange. The CD-ROM itself works on the same principles as the audio Compact Disc (CD) but with additonal error correction algorithms to ensure the integrity of the data. Originally only available in ROM form, it is now available as a writeable medium, either write once for data which must be protected or re-writeable. The CD drive has a capacity of 650MB per disc and will remain a standard for some time, but is already being challenged by the DVD or Digital Versatile Disc. This takes the basic idea of the CD further by storing the data at a density unachievable when the CD was invented. In its first form DVD has a typical storage capacity of around 5GB, but multi-layer and double sided versions are forecast to increase this greatly. Most DVD drives fitted to personal computers also read CDs. In order that one can see what is going on in the computer there needs to be some kind of display. In personal computers this is provided by a monitor. Traditionally these have been based on a variant of the Cathode Ray Tube or CRT as used in TV sets. Portable computers, and, increasingly high end desk top computers, use flat screen displays usually based on Liquid Crystal Displays or LCDs. The display is served by a special memory device or graphics controller. This is a high speed RAM devoted to supporting the display. The display itself is capable of displaying a defined number of pixels or picture elements, for example an XGA display
LIGHTING ELECTRONICS
shows 1024 × 768 = 786,432. In a high specification computer each such pixel needs 24-bits of data to specify its color and brightness (8-bits for each of red, green and blue). The computer itself, referred to as the Central Processing Unit or CPU, is centered round the arithmetic/logic unit or ALU that performs the calculations. It is supported by a number of registers and temporary memories. The number and size of these is determined by the computing power needed, but always includes items like the program counter which keeps track of where the computer is in the program routine it is supposed to be carrying out, the instruction register, which holds the current instruction for the ALU, and the accumulator which carries the data being currently operated on by the ALU. Today the entire CPU of a typical personal computer is carried in a single microprocessor that embodies the ALU and all its associated control and register devices. Early personal computers needed a chip set at their heart since it was not possible to get all the central functions on to a single chip. 2.7.2 Computer programs The complexity of a computer arises from the incredible speed it operates at, and the huge amount of data it manipulates. However, every single operation is very simple, consisting as it does of moving binary data from one store to another, and carrying out a simple mathematical operation while doing so. A computer program is a series of instructions executed in a set order. Often the instruction set changes as a result of either the result of a calculation being to tell the CPU to jump to a different part of the program, or of an interrupt, an external signal which takes priority. Each instruction is itself no more than a set of binary numbers which itself specifies a simple action, for example: “Load the accumulator with contents of memory address BB29” (here the address is given in hexadecimal, since this is much easier to write, and less prone to transcription error, than the actual binary address which would be expressed as two 8-bit bytes 10111011 00101001). Or another example: “Subtract 34 from the contents of the accumulator”.
When a computer program has been written as a series of binary or hexadecimal instructions, it is said to have been written in machine code. In low cost microcontrollers with limited memory capacity, or where a high speed of program execution is needed, it may be necessary to write the program this way. However, for most applications this would be tedious. The next step up from machine code is an assembler language. This simplifies the programming process by using mnemonics and labels. Thus the mnemonic DEC might mean “decrement the contents of the accumulator by 1”, or the label KEYC, identify the memory address of the C character on a keyboard. Both machine code instructions and assembler languages are specific to particular microprocessors. Again, assembler language can be appropriate for programming microcontrollers, but is still tedious for sophisticated applications. There is a need for a high level language. This takes the mnemonic idea much further by allowing programming commands to be entered in text form. For example a program line might consist of: 44 LET Lumens = Effic*Watts Here the number 44 is simply the line number in the program. A new variable called “Lumens” is being defined by the instruction LET, and itself is the product of two variables which have already been defined. If they have not been defined the program would stop. In this case the existing variables are “Effic”, the efficacy of a class of lamp, and “Watts” the power rating of a particular lamp. When a high level program is written, the result is source code. This includes not only the high level language instructions, but also plain text remarks which explain what each instruction is intended to do. This makes de-bugging, the removal of program errors or “bugs” easier, and is essential when a team of programmers is at work, or when programs have to be changed some time after their first issue. If a program has been created using a high level language, there is a need for an interpreter. When the program runs, the interpreter looks at each
105
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
program line and converts it to the equivalent machine code instruction. For some applications running programs with an interpreter is satisfactory, but clearly it is an inefficient way of doing things. This is particularly the case when some simple command is repeated many times. So when it comes to running complex programs the procedure is to use a compiler. A compiler is a computer program that takes the high level source code and converts it directly to machine code. It does this as a separate operation. Once a program has been compiled, the final program installed by the user is a machine code program, and thus highly efficient. However, the development of compiled programs is a complex task, since the programmer doesn’t know whether his program works or not until it is compiled, and if it then fails to work, he can have some trouble working out what the problem is. Actual compilers, for example a compiler to convert a program written in the high level language C++ into a form suitable for running on an Intel Pentium™ microprocessor, have sophisticated tools to help the programmer analyze what is happening. The microcontrollers referred to in Section 2.7.4 might only run a single program. However, personal computers run several programs concurrently. These are in three categories: • The ROM based programs which get the computer under way, usually including the Basic Input Output System or BIOS and the booting routines. Programs which are carried in ROM (and its programmable variants) are referred to as firmware, distinguished from hardware, the physical construction of the computer device, and software, the changeable applications and operating programs not permanently resident. • The Operating System which determines how the computer carries out general tasks, particularly how it operates its memories, displays and Disc Operating System or DOS. Currently the best known operating system is Microsoft’s Windows®. But there are many others, some of which are more suitable for networked applications or real time tasks. • The Applications Program. For most users this is the only one that matters, because this is the one car-
106
rying out the specific task, whether it be word processing or doing a lighting design. 2.7.3 Memories One of the keys to the success of microprocessors has been the incredible pace of development in semiconductor memory technology. There is now a wide range of memory types, each with its own acronym. The main ones are summarized in Table 2.4. Semiconductor memories are defined as volatile or non-volatile. A volatile memory is one which loses all its data when the power to it is disconnected. A non-volatile memory retains its data even when power is disconnected. Memory chips are available in a wide range of sizes, from 256 bits up to 256 Megabits (and beyond.) They come in different configurations to match different bus arrangements, so, for example, a 16,384 bit memory might be organized to store 2,048 8-bit bytes. The concept is shown in Figure 2.71. This greatly simplified diagram shows a 4 bit address bus able to address 16 different memory locations. When a location is addressed, it is connected to the 8-bit data bus. In a ROM all that can happen is
Acronym RAM DRAM SRAM S-DRAM VRAM ROM PROM EPROM EEPROM Flash E2 FIFO
Full Name Random Access Memory Dynamic Random Access Memory Static Random Access Memory Synchronous Dynamic Random Access Memory Video Random Access Memory Read Only Memory Programmable Read Only Memory Erasable Programmable Read Only Memory Electrically Erasable Programmable Read Only Memory see text abbreviation for EE as in EEPROM = E2PROM First In First Out
Table 2.4 The acronyms used to describe the most common types of semiconductor memory.
Memory locations
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
ADDRESS DECODER
Address bus
LIGHTING ELECTRONICS
Data bus
Figure 2.71 Principle of the semiconductor memory.
that the data in the memory location can be read on to the bus. In a RAM, data can be read from the location, but can also be written into it, over-writing any previous data. RAMs operate in a read-write cycle, being either read-enabled or write enabled. In the former case its data is put on the data bus, in the latter it takes data from the data bus. One way to achieve a ROM is to have a single transistor at each cross point in the diagram, as shown in Figure 2.72. When the address decoder makes a memory location line go high, the transistor switches on and puts a signal on the data output line. Clearly if all the transistors in one location switch on the only output that can be given is 11111111. In a mask programmed ROM the manufacturing process includes a step whereby designated transistors are disabled, so each memory location can give any required pattern of 0 and 1. Mask programmed ROMs are expensive to set up, but cheap to manufacture in large quantities. In a PROM the same principle is used, but this time the user can decide which transistors are to be disabled. This is done by selectively applying a high voltage to the collector of transistors which are to be disabled. This then blows a fusible link in the emitter connection. The standard PROM is referred to as a one time programmable or OTP device. Both the masked ROM and the PROM are clearly non-volatile devices, but once data has been written in, it cannot be changed. During the development of
a product the ability to change data is essential; for example while debugging firmware. Even when products are fully developed, the ability to change the firmware to introduce new features can be important. The EPROM allows this. In this case the transistor is a MOSFET with a difference. It has a “floating gate” which can be charged up by applying a reverse voltage to the drain. The reverse voltage pulse is sufficiently big to give some of the electrons carrying the current enough energy to pass through the SiO2 layer and to charge up the gate. The negative charge is then “trapped” and will stay there for years. In Figure 2.73, G1 is the normal gate connected to the memory location line, G2 is the floating gate. If G2 is negatively charged it has the effect of increasing the gate G1 threshold voltage from around 1V to 5V. This effectively means that the transistor is OFF, and stores a 0. However, the charge on the floating gate can be erased by exposing the transistor to intense ultra-violet radiation; allowing the EPROM to be re-programmed. PROM and EPROM programmers are standard devices designed to transfer data into non volatile memory as firmware. For short and medium production runs it is quite practical to program devices as part of product production, and both PROMs and particularly EPROMs are widely used in lighting control equipment. Data bus Address +V
1
0
1
1
0
0
1
1
Figure 2.72 The operation of a ROM or EPROM, showing how a transistor is used to indicate a 0 or 1 at a particular memory location.
107
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
S
D
G1 G2
n
n
SiO2
P
S
D
G1 G2
n
n
SiO2
p
Figure 2.73 Construction of the eraseable transistor memory cell. The top diagram shows the EPROM transistor, and the bottom diagram shows the EEPROM transistor.
Both devices require programming to be done as an off-line process, which may not be convenient. The EEPROM allows the programming to be done in situ. In this case the layer of SiO2 between the second gate and the drain is made very thin, less than 10nm compared with around 40nm for normal gate. The charging of the gate is now achieved without the need for such a high energy. When the programming voltage is applied to the drain, quantum mechanical tunneling (taking advantage of the fact that, like photons, electrons have a wave nature) gets the electrons through the insulator. Electrical erasure is achieved by applying a pulse of opposite polarity to the drain, a procedure which is obviously much quicker than taking the chip out of the equipment to flood it with UV. “Traditional” EPROM and EEPROM need quite long writing cycles, and are not intended for frequent
108
re-writing. In the last few years great advances have been made in programmable memories, and this has led to the introduction of Flash memories. In principle these are similar to EEPROM, but are intended for frequent re-writing, and have a comparatively fast write cycle. Within computers standard RAM is volatile, and only carries data when the power is on. It is much cheaper to make and is much faster than any kind of EPROM, because more memory cells can be packed on to a smaller area of silicon. There are many varieties of RAM, the main ones being: DRAM or dynamic RAM. Here the 0s and 1s are stored as capacitive charges, which affect the gating of a transistor. However, these charges can leak away, so DRAM has to be continually refreshed. A typical 16 Mb DRAM may need 1024 refresh cycles every 16 milliseconds. There are a number of different architectures for DRAM, but a typical DRAM memory cell uses two transistors. SRAM or static RAM. This has the advantage that it does not need the refresh activity. It is based on flip-flops, described in Section 2.6.3.1, so a typical memory cell needs four transistors. The circuitry is more complex within the SRAM, but the external circuitry is made much simpler. In many applications requiring non-volatile memory, battery or capacitor backed static RAM is used where EPROM is not appropriate – where, for example, the memory contents changes as part of the equipment operation and the need is to guard against power failure and protect the current contents of the memory. If capacitors are used, they are very high capacitance, of the order of 1F, but only rated at 2.5 or 5V. However, this technique is giving way to the use flash memory. VRAM is an example of a special kind of RAM, referred to as a dual port RAM. VRAM itself is video RAM, and is the kind of memory used in graphics controllers to support the display screen. The need for dual port operation arises because the data coming in to the memory is regulated by the computer clock, so comes in at the convenience of the CPU. However, the data going out, representing what is to be seen on the display screen, needs to go out at timings determined by the display itself. Thus the memory must be of a kind where writing and
LIGHTING ELECTRONICS
reading can take place simultaneously, and with different timing. 2.7.4 Micro-controllers Microprocessors have now developed so far that there is a great difference in capability between the extremes of the market. A lot of modern electronic products, including all sophisticated lighting control devices, are based on microprocessors. However, the microprocessors used are not often the same kind used in computers. They are more usually microcontrollers. Microcontrollers do not need all the paraphernalia associated with a full personal computer, but do need to be optimized for the task in hand. A typical microcontroller includes, all on one chip: • a CPU (of limited capacity compared with a personal computer’s CPU). • clock circuitry (but usually an external crystal is used to provide the clock master timing). Typical microcontrollers operate at a clock frequency of 1– 20MHz (compared with 500MHz to 3GHz for a personal computer). • provision for program memory. This is loaded by the user, using OTP, EPROM or flash programming. Typical program capacity is 4–16KB. Many smaller microcontrollers are made with mask programmed ROM once product development is complete. • RAM, typically in the range 128 bytes to 2kB. • ability to address further external memory, up to 64KB. • parallel I/O port, typically 32 bit wide. • serial I/O port(s) providing UART like facilities • provision for, typically, 4–32 interrupt sources. In a lighting control device an interrupt could come from the zero crossing of the AC sinewave, or as a signal from a presence sensor. • (sometimes) 4–8 analog I/O ports. • one or more timers. One of these is usually a watchdog timer which issues a signal if processor activity ceases for any reason. Programming of microcontrollers can be done using high level languages with suitable compilers, and is assisted by various forms of emulator. This
can be a software model of the microcontroller running on a larger computer, or can be an in-circuit emulator, which allows an actual microcontroller to be used while at the same time allowing the programmer to see what is happening within it. Such programming often requires the examination of the contents of individual registers and data streams. All microprocessors have an instruction set which cover their basic operation. For each operation there is an operation code or opcode. Microprocessors used in personal computers have quite a large instruction set which simplifies programming, but uses up processing power and slows down operation. The microprocessors used in microcontrollers are often examples of RISC Processors, or Reduced Instruction Set Computers.
2.7.5 Programmable Logic Controllers Computers, microprocessors and microcontrollers live in a world of their own, where everything happens very fast at low voltage and low current. But in the real world there is a need to interface with other devices. Within a product a microcontroller might be required to control a power device – for example an electronic dimmer might use a small microcontroller, in turn interfaced by suitable isolating components, to a power device such as a thyristor or IGBT. Within a system, as opposed to an individual
Figure 2.74 An example of a programmable logic controller.
109
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
product, there is often a requirement to carry out logical sequences, but in a difficult electrical environment. For this purpose the Programmable Logic Controller, or PLC, has been developed. These devices vary in complexity, but generally embody the following features: • a central processor module, usually based on a microcontroller; but in the bigger models based on a PC core. • industrial modular construction; suitable for an industrial environment. • all solid-state with non-volatile program memory. • input and output modules giving a high level of isolation, and a range of power handling capabilities. PLCs are used in industry for the control of all kinds of industrial processes. They are also used for lighting control, usually when combined with the control of other items.
2.8 Programmable devices 2.8.1 FPGAs The different concepts and technologies which have individually given us the microprocessor, the standard logic chip and the programmable memory can be used in other ways to produce various kinds of programmable ICs. For many applications such devices can be used in place of ASICs. Because they are “off the shelf” and are well supported by high level programming tools, they give product designers a quicker way to market, and lower overall costs, even if the cost of the IC itself is higher than its ASIC equivalent. The devices work by providing the user with a number of logic gates, input and output lines and, in the case of the larger devices, a multiple bus structure. The biggest devices have sections that are optimized for use as local memory, or even as complete microprocessors. The way in which the various elements can be connected together represents the “programmable” aspect of the device. In the simpler de-
110
vices standard EPROM one time programming, the blowing a fuse to make a disconnection, is used; but this may be accompanied by the use of programmable anti-fuses to make a connection. The technique here is to create vias between two conducting layers separated by a very thin insulating layer. The insulation can be locally destroyed by the application of a strong enough electric field. In the more advanced devices, transistors are used to determine the signal routing. The transistors are arranged to operate collectively as a static RAM which is loaded on power-up from a non volatile memory (usually a set of flash memory on the same chip). Different manufacturers use different processes and architectures to achieve what they consider to be the optimum combination of capacity, flexibility and speed. Many different acronyms have been used to describe programmable devices. Most powerful ones are now collectively referred to as FPGAs or Field Programmable Gate Arrays. While the basis of the FPGA is quite simple and easily understood, the huge number of combinations possible means that the only practical way to program a big one is using computer aided design. The manufacturers provide the necessary programming tools, that also ensure that design rules are followed, and the timing of signals taking different routes through the array can be accurately predicted. Otherwise there could be the problem that data which should be “in sync” actually emerges from the FPGA out of synchronization. Small programmable devices offer a limited number of useable gates, and are similar in cost to EPROMs. Large programmable devices offer several hundred thousand gates, but may cost hundreds of dollars. FPGAs are used in logic intensive operations requiring high speed real time operation. An example of practical use is in digital video and graphics image processing. So, for example, big LED video screens make extensive use of FPGAs to convert the incoming digital video signal into the huge number of individual control signals needed to provide PWM control to vary the brightness of each individual LED.
LIGHTING ELECTRONICS
2.8.2 DSPs The Digital Signal Processor is a special kind of microprocessor optimized to the real time processing of signals which until recently were analog. Many analog circuit elements, especially filters, have a digital counterpart, whereby the digitized signal is operated on mathematically to achieve the same effect. For this reason one of the most widespread uses of DSPs is in professional (and high end consumer) audio equipment. Obviously, if this idea is to work, the processes must take place at frequencies very much higher than those being treated. DSPs are also used in such diverse applications as motor control and image processing. DSPs are designed to carry out repetitive tasks with maximum efficiency. Many applications require the same calculation (with varying input data) to be carried out at very high speed, so some of the features likely to be included are: • the ability to deal with both fixed point and floating point calculations, in order to ensure that maximum possible accuracy is maintained, while at the same time not getting hung up on out of range results. • additional arithmetic hardware. This is likely to include one or more hardware multipliers to ensure
that multiplications can be carried out in a single clock cycle. The conventional ALU has to carry out many cycles to achieve multiplication. Much if not most DSP work involves successive multiplication (but this is not exclusive to DSP, FPGAs can also be programmed to carry out hardware multiplication). • additional bit manipulation hardware, for example the barrel shifter used to instantly shift bit arrays to left or right. This might be used to ensure that the results of floating point calculations are conformed for the next process (e.g. an audio signal might have been processed using 40-bit floating point calculations to minimize added distortion, but need conforming to a 16-bit output to match CDs). • additional buses, for example a “results data bus”, to allow simultaneous complex operations to be executed in a single clock cycle. This chapter has presented a very brief review of the basis of modern electronics relevant to lighting control. It is amazing that the behavior of different kinds of slighly impure silicon joined together can give us everything from an electronic switch, capable of switching a Megawatt, to a single memory chip able to hold the contents of a book – but that is now the case. Modern lighting control is taking selective advantage of electronic developments to create new products and new control concepts.
111
Chapter 3
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Everyday
lamps
Part 2 – Lamps The next three chapters describe the different kinds of lamps or “light sources” likely to be encountered. It is increasingly the case that the lightsource determines the method of control, so a good understanding of how lightsources work is essential for understanding lighting control design. Here sources are divided into “everyday lamps” of the kind that most of us are already familiar with; arc lamps which until recently were only used in industrial applications but which are now encroaching on many commercial applications, and special purpose lamps.
3.1 Non electric lighting In his book Lighting Historic Buildings (see Reading List) Derek Phillips reminds us that our forbears managed without electric light, yet were often very skilled at producing surprising and functional lighting effects. Architects understood how daylight worked, and designed their buildings to use daylight effectively. Today the effective use of daylight is once again considered important, on the grounds of both energy efficiency and creating pleasant places to live and work. Lighting based on flame, technically pyroluminescence, has been the only source of “artificial” light available to the human race until recently. Light from flames arises from the high temperature chemical reaction between oxygen and the “fuel”; energy is released from the excited atoms and molecules, and some of it is in the form of light. The amount depends both on the temperature and the fuel being burnt. An acetylene flame does not give much light compared with the input energy – as little as 0.2 lm/W. By comparison, the old style flashbulb (which works on the basis of combustion) using zirconium as a “fuel”, gives 56 lm/W. Fires were found to be inefficient sources of light, and the idea of using a concentrated fuel for lighting purposes resulted in the candle and the oil lamp. The arrival of gas lighting in the 19th Century
112
brought entirely new concepts to lighting. The most obvious is the idea that the fuel for lighting is supplied from a distance; in practice both much more efficient and safer than carrying a fuel reservoir next to the lamp. Less obvious are two other points: • the invention of the gas mantle. This device, placed at the point of gas ignition, creates an intense white light; much brighter than the simple gas flame. The mantle, coated with rare earth salts, exhibits fluorescence, and converts some of the thermal radiation from the gas flame into visible radiation. The mechanism is similar to, but not the same as, that of the fluorescent lamp. • gas lighting permitted remote control of lighting. Admittedly pilot lights had to be used to ensure that any gas always ignited; but in principle many of the ideas we have about lighting control today were understood in the age of gas. Light from electricity was demonstrated in the early 19th Century, just as gaslight was becoming established. However its early manifestation was the carbon arc lamp. This source of illumination went on to be widely used where high brightness sources are required – for example searchlights, film projectors and film set lighting; but was impractical for “everyday” lighting. The carbon arc is created by placing two carbon electrodes together, passing a high current between them, and drawing them slightly apart. Light is radiated by various mechanisms. Principal is incan-
EVERYDAY LAMPS
descence, arising from the high temperature of the electrodes, but in addition there is luminescence from the vaporized electrode material and its combination with air.
3.2 The incandescent lamp. Practical artificial electric lighting had to wait until the late 1870s, when Joseph Swan in England and Thomas Edison in the USA invented the enclosed incandescent lamp. This solved the problem of high temperature material catching fire in the atmosphere. The incandescent lamp, or, more correctly, the incandescent filament lamp was, until comparatively recently, the main electric light source. For many people, especially in the home, it still is. Its great flexibility, variety of possible designs, pleasing light and ease of control mean that it will continue to play an important part in the lighting world. Incandescent lamps consist of a filament mounted either in a vacuum, or in an inert gas (one which does not chemically react with the filament). Early electric lamps used a carbon filament, but this was both fragile and inefficient. Tungsten is now the main filament material because it is easily worked, has a low vapor pressure and has a high melting point. The melting point of other metals is generally too low to produce useful incandescence while solid. Light is produced by passing an electric current through the tungsten filament and using the i2R heating effect to raise the filament temperature sufficiently to produce incandescence. While tungsten’s behavior when heated is similar to that of a blackbody (in that it produces a continuous, but not uniform spectrum of light) its total radiation is considerably less than that of a blackbody at the same temperature (see Figure 1.52); and anyway only a fraction of the total radiation of a blackbody can be in the visible spectrum. At its melting point of 3,655K tungsten has a luminous efficacy of around 53 lm/W. Practical tungsten lamps have efficacies in the range 8–30 lm/W depending on their construction and the filament operating temperature. A 100W general service lamp operates with a filament temperature of 2,700K, giving 12.5 lm/W. Photoflood
lamps may operate at 3,400K, giving around 30 lm/ W, but with a very short life. The tungsten filament fails mainly due to evaporation. The same energy which is exciting the tungsten to emit light can also impart sufficient velocity to some molecules that they escape – and generally appear as a blackening on the bulb wall. Evaporation can be reduced by having other molecules, in Optional silica diffuser coating
Glass envelope
Filament Molybdenum support
Nickel lead wires
Dumet wire
Glass pinch seal Cement Fuse in glass sleeve Screw base contact Exhaust tube Glass insulator
Coiled filament
Solder contact
Coiled coil filament
Figure 3.1 Construction of the general service tungsten filament incandescent lamp. Most use the coiled coil filament and have a gas filling. Low wattage lamps use a simple coiled filament in plain vacuum. Dumet wire is a compound iron/nickel/copper wire suitable for making a gas tight connection through glass.
113
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
the form of a gas, present to exert a pressure to prevent the escape. The choice of gas is a compromise. If its thermal conductivity is too high, it simply conducts heat away from the filament to the lamp wall, reducing the light output. It needs to have as high a molecular weight as possible to do its job. It must not ionize under normal conditions within the lamp. In practice the economic gas filler choice for most lamps is argon, with some nitrogen to reduce the tendency to ionization. Such ionization is particularly likely to happen when the filament breaks, resulting in a high arc current and quite possibly a shattered lamp. For this reason general service lamps are fitted with a fuse in the glass filament support. In order to reduce its overall length, and to minimise the need for supports (which conduct heat away from the filament) the tungsten filament is coiled. A filament which might need to be 90cm (3 ft) long can be tightly coiled to have a length of 5cm (2 in). This coiling also helps reduce the conduction of heat away from the filament by the filler gas. Still greater efficacy is achieved for medium power lamps by using a coiled coil construction. Figure 3.3 shows the relationship between nominal voltage, lamp life and light output for incandescent lamps. From this it is clear that life can be greatly extended by “under running” the lamp, but the light output drops by a much greater percentage than the voltage drop. For purely decorative applications this may not matter – but for any application where the cost per lumen is important, in terms both of power used and lamp replacement costs, it is better to run lamps at rated voltage. General service tungsten filament lamps are designed for a nominal life of 1,000 hours. Rough service, and “long life” lamps are rated for 2,000 hours or more, but have significantly lower efficacies. The resistance of tungsten varies greatly with temperature (see Figure 1.6). The cold resistance of a lamp can be around 14 times lower than the resistance at rated voltage, resulting in a high inrush current for the first few cycles after switching on. It takes between one tenth and half a second for the current to drop to the running value. In practice circuit impedance, and the impedance of the supply, reduces
114
Figure 3.2 The higher the molecular weight of the gas filling, the longer the life. The lamps illustrated above, which are incorporated in the Lumisphere X-24 Low voltage low voltage festoon system use a Xenon filling. Whilst the filling is too expensive for “everyday” domestic type lighting, it is justifiable for use in inaccessible places where extremely long life is essential. The life of these 24V 8.5W lamps is 20,000 hours and the glass lamp is mounted within a protective polycarbonate globe; the whole unit having an IP 68 protective rating.
the actual peak current; however control equipment, especially dimmers, should be designed to withstand the full inrush current. Most general purpose tungsten filament lamps are designed to run directly on the available supply – whether it be 12V in an automobile, 115V in the USA, 100V in Japan or 230V in Europe. The higher voltage lamps must use a thinner filament for a given power rating which has two disadvantages; the obvious one of fragility, and the less obvious one that the filament structure is less compact, and, therefore, less suitable for optical focusing. Tungsten filament lamps running on line voltage are easily and effectively dimmed. When used for decorative schemes, dimming can considerably prolong lamp life, not only by virtue of the “underrunning” of the lamp, but also because the “soft start” that can be imparted by automatic dimming reduces thermal shock. The quasi blackbody radiation characteristic of tungsten filaments results in a color temperature change as lamps are dimmed. In most practical applications this is not a problem, since the resulting “warmer” light seems subjectively more appropriate at lower light levels.
EVERYDAY LAMPS
180
300
160 200
140 Percent 120 Current 100 Power Lumens 80
100 Percent Life 80
Current
60
er Pow
40
60 ens Lum
20 40
50
60
70
40
Lif e
80
90
100
110
20 120
130
0
Percent nominal voltage
Figure 3.3 The relationship between light output, current, input power, lamp life and operating voltage for a typical tungsten lamp. The discontinuity in the life curve is due to a change of scale in the graph below 100%.
One problem can arise in systems using thyristor dimmers. These “chop” the mains power supply, and the resulting steeply rising waveform can cause a corresponding mechanical vibration in the filament. An audible “lamp sing” can be the result, particularly if the lamp and luminaire combination has a corresponding resonant frequency. The problem is worse on high current lamps, and as a result is more noticeable in 115V systems than it is in 230V systems. The only practicable cure is substantial filtering (discussed in Section 8.3). Infra-red lamps are a special variant of the tungsten lamp. All incandescent lamps emit a lot of infra-red anyway; so in appropriate ratings they represent an efficient source of heat. The infra-red lamp, made in ratings from around 150W to 5,000W, is especially convenient for some industrial, commercial and consumer applications, especially where heat must be directed or is required rapidly (e.g. in paint curing plants etc). Infra-red lamps have a long life, because at the temperature they are running, the tungsten vapor pressure is low.
3.3 Tungsten halogen lamps A heated tungsten filament radiates energy over a continuous spectrum in the range 300nm to 2,000nm, but the proportion radiated in the visible spectrum (say 400nm to 700nm) is dependent on the filament temperature. The nearer the filament is run to the melting point of tungsten, the greater the efficacy but the higher the rate of evaporation. The question arises, is there any way to minimize or even reverse the evaporation process? The answer lies in the idea of a regenerative cycle, where any tungsten that does evaporate from the filament is somehow returned to it. Tungsten halogen lamps have a proportion of halogen vapor in their gas filling. Halogens make up the highly reactive series of elements which include chlorine, bromine, iodine and fluorine. In the lamp the halogen reacts with the tungsten vapor to form a tungsten halide. This is gaseous, provided the temperature remains above 250°C. When the halide circulates to the filament area, the heat of the filament
115
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Incandescent lamp
Relative energy
Daylight for comparison
IR UV
VISIBLE 400
500
600
700
800
900
1000
1100
I.R. Lamp
Figure 3.4 The spectral power distribution curve of an incandescent lamp shows a preponderance of infra-red. Infra-red lamps, such as the example from Osram shown, are made to produce minimum visible radiation, and to have a long life – in this case 5,000 hours.
breaks the halide down, and the tungsten is deposited back on the filament. This process allows the filament to be run a higher temperature, typically 3,000K, and gives a longer life, typically 2,000 hours at rated voltage. Indefinite life is not achievable, mainly because it cannot be assumed that the re-deposited tungsten will obligingly always land on the weakest parts of the filament. Tungsten halogen lamps are very compact because, in order to maintain the 250°C vapor temperature, the filament must be near the outside wall. This construction has the added advantage that filament cooling by gas convection is reduced or eliminated, and the small overall volume allows the use of the more expensive, but more efficient (because heavier) Krypton as a filling gas. The high temperature, and the corrosive effects of the halogen, mean that ordinary glass cannot be used. The envelope is, therefore, made of fused quartz. This presents some difficulty in making the external electrical connection since any disparity in thermal characteristics between the quartz and the lead-in wires causes the envelope to break.
116
Molybdenum meets the requirement, but has the problem that at above 350°C it oxidizes easily, resulting in a mechanical stress at the external connection. Luminaire design must take this into account, and usually involves some kind of heatsink at the lampholder. The tungsten halogen re-generation cycle only operates at filament temperatures above 2,000K. This clearly has some implications when lamps are dimmed; it could be inferred that lamp life might actually be reduced when the lamp is dimmed. In practice this appears not to be the case, presumably because the rate of filament evaporation is greatly reduced anyway at the lower temperatures. Some experts recommend that, in installations where lamps are maintained at a low light level for long periods, they are periodically run at full output for a short time to activate the regeneration cycle. Certainly this has the effect of removing any tungsten blackening on the bulb wall. In an installation like a hotel public area, where the lighting might be bright at breakfast time and dim in late evening, any possible problem is eliminated automatically by virtue of the daily programmed cycle. The higher filament temperature, and the use of quartz as an envelope, results in the tungsten halogen lamp radiating a significant proportion of ultraQuartz envelope
>250K
Wall zone Formation zone
3,000K
Filament zone
Tungsten filament (W)
W 000e 2I 000e A A 000e WI2 œ w 000e 2 I 000e A WI2 000e A 000e I2
Figure 3.5 Simplified diagram of the tungsten halogen cycle. Here Iodine (I) is shown as the halogen, but this is only an example. The inert gas Argon (A) acts as catalyst, but it is known that impurities in the tungsten and in the filling vapor also play a part in the cycle.
EVERYDAY LAMPS
double ended for flood light low voltage, pinbase
low voltage reflector spotlight
mains voltage decorative
mains voltage reflector spotlight
Figure 3.6 Examples of tungsten halogen lamps. Photos from Osram.
violet. For most applications this is undesirable. It could be a safety hazard, and it definitely hastens the bleaching of pigments. For this reason most tungsten halogen luminaires or lamp assemblies are fitted with a clear UV filter. A comparatively recent development is the “doping” of quartz to render it UV absorbing so some tungsten halogen lamps are available which do not need filters. Mains voltage (110–230V) tungsten halogen lamps have traditionally used linear filaments and have had a double-ended construction. This makes them particularly suitable for floodlight applications, in a wide range of ratings typically 300W–1,000W, but available down to 60W, and up to 2,000W and more. This construction is not suitable for spotlights, so compact-filament single-ended halogen spotlight lamps operating at 12V or 24V are used for this purpose. Ratings are in the range 20–75W. The low voltage operation requires a transformer, which must be sited near the lamp in order to minimize the length of the high current cable run. Common practice for multi-lamp installations is to use one transformer for several lamps. Recent developments in filament manufacture have resulted in a new range of mains voltage compact spotlight lamps. These are not of the same dimensions as their low voltage counterparts, but, nonetheless, allow the creation of compact luminaires without the need for transformers.
For many architectural applications the lamps are supplied with an integral reflector. There can then be some confusion as to the merits of different kinds of reflector. “Cool” reflectors have a dichroic coating to reduce the amount of infra-red radiated by the lamp; but this means that the excess heat goes out the back of the lamp. This may be undesirable since it could over-heat the luminaire or the built-in transformer. Alternative “hot” (actually aluminum) reflectors reflect all wavelengths so the infra red accompanies the visible radiation. This identifies a problem area with lamps. Superficially similar lamps can have different beam angles, wattages and reflectors. A good lighting design can be spoilt by poor maintenance when the wrong lamp from a lamp family is fitted as a replacement. The most recent development in tungsten halogen lamps is the introduction of the halogen infrared lamp. This is a little more “bulbous” than the usual lamp, and has an axial filament with the return lead running down the center of the filament to produce a single ended lamp. The inside of the bulb wall has a thin film coating designed to transmit all visible light, but to reflect all infra-red back onto the filament. This technique can result in an increase in efficacy of as much as 37%. For example a 60W lamp in
Aluminum reflector reflects IR resulting in lower luminaire temperature
Dichrioc reflector transmits IR resulting in lower beam temperature IR
Visible IR
Visible
Figure 3.7 Tungsten halogen reflector lamps are available with two different kinds of reflector. It is important that the correct replacement is used.
117
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Construction Conventional incandescent Standard tungsten halogen Halogen-IR
16000
14000
Lumens 930
75
1030
60
1110
Life 2000 hours 2500 hours
Efficacy 10.9 lm/W 13.8 lm/W
3000 hours
18.5 lm/W
Table 3.1 Comparison of ca 1,000 lumen PAR38 lamps using different filament technologies. Acknowledgement to Osram Sylvania (reporting to the Lightfair® 2000 Annual Lamp Review).
12000
a conventional tungsten lamp in a number of different physical sizes, of which the most widely available is the PAR38. This is supplied in a range of wattages and beam angles; with either conventional filaments or tungsten halogen inserts. It is interesting to compare the performance of different versions, as in Table 3.1. In the case of the halogen-IR version the lamp designers have a choice of pushing for more lumens/ W at the expense of life, but the table shows a dramatic increase on both fronts.
10000
Spot
8000 Candelas
Watts 85
7000 6000 Wide spot
5000 4000
Floods
3000
IR reflected back to filament
2000 Wide floods
1000
90
60
30
0 Degrees
30
60
90
Figure 3.8 The luminous intensity distribution from variants of the same power reflector lamp. The graphs here are representative of the performance of different 12V 50W lamps. The result of fitting incorrect replacements can be to ruin the lighting design.
“standard” tungsten halogen might give 16.5 lm/W, whereas its halogen-IR equivalent can give as much as 24.2 lm/W. The new technology has shifted the balance of lumen output, wattage and lamp life. An example here is the popular PAR (Parabolic Aluminized Reflector) lamp. This lamp was originally produced as
118
Optical coating
Visible light transmitted
Figure 3.9 The principle of the halogen-IR lamp (above) and an example MR16 lamp from Philips using the technology (below). MR stands for Multifaceted Reflector, and its most popular size is MR16.
EVERYDAY LAMPS
3.4 The fluorescent lamp The fluorescent lamp is the most familiar of the large class of lamps referred to as discharge lamps. In these lamps light is created by an electrical discharge within a gas or vapor. The fact that an electric or electromagnetic field within a gas can create light is manifest in the natural world both in lightning and in the “Northern Lights” (Aurora Borealis). In all cases the light creating mechanism relies on the electrical or electromagnetic power input being used to raise the energy levels of electrons. As they fall back to their normal level, they emit quanta of radiation. Most discharge lamps have a significantly higher efficacy than tungsten lamps, and many, but by no means all, have significantly longer lives. The construction of the standard fluorescent lamp is shown in Figure 3.10. Standard straight tubular lamps are designated by their nominal length, which includes the thickness of standard lamp holders, and by their diameter. The diameter is specified as a “T number” which is the diameter measured in eighths of an inch (never mind the metric system! but an eighth of an inch is just over 3mm). So a T12 tube is 1.5in (38mm) diameter, a T8 (currently the most common) is 1in (26mm) diameter and a T5 is five eighths of an inch (16mm) diameter. The smallest diameter tubes generally available are T2 (7mm), but these are classed “special purpose” in this book. At each end of the tube are electrodes, referred to as cathodes. In fact in an alternating current circuit each electrode alternates as cathode and anode, but the emphasis on the name cathode identifies the prime function of emitting electrons. In the normal hot cathode fluorescent lamp (cold cathode lamps are discussed in Section 3.6) each electrode is a tungsten filament of special construction coated with alkaline earth oxides. When heated to around 1,100°C the cathodes emit copious electrons. An anode can attract these electrons, creating a current between anode and cathode. In the tube itself there is a gas filling. The active constituent is mercury vapor at low pressure (around 1.07 Pa or 0.00016 lb/in2) with the addition of Argon (or a mixture of inert gases) at around 200 Pa to help the initial discharge. When a fluorescent lamp
is in its running condition, with an AC supply connected to either end, there is an arc discharge along the length of the tube, and the mercury vapor ionizes. The combination of the ionization itself, and the excitation of the mercury atoms produces electromagnetic radiation at a number of wavelengths – the principal of which are 254, 313, 365, 405, 546 and 578nm. The majority of these wavelengths are outside the visible spectrum and are ultra-violet. So the fluorescent tube uses a two-stage process; the inside wall of the tube is coated with a phosphor. Phosphors are inorganic crystalline compounds with a small proportion of metal, known as the activator. An example is calcium halophosphate with antimony as the activator. Phosphor science is an advanced form of cookery, and much research has been done to identify the phosphors giving the highest efficacy and best color rendering. Modern fluorescent lamps use a mixture of phosphors. The ultra-violet radiation causes the phosphor to fluoresce. The phosphor’s outer electrons are Tube
Length
Nominal Power
Tube Current
T12 T12 T12 T12 T12 T8 T8 T8 T8 T8 T5 T5 T5 T5 T5 T5 T5 T5
2,400mm 1,800mm 1,500mm 1,200mm 590mm 1,800mm 1,500mm 1,200mm 590mm 438mm 1,449mm 1,449mm 1,149mm 1,149mm 849mm 849mm 549mm 549mm
100W 85W 65W 40W 20W 70W 58W 36W 18W 15W 80W 35W 54W 28W 39W 21W 24W 14W
950mA 800mA 670mA 430mA 370mA 700mA 670mA 430mA 370mA 310mA 540mA 170mA 450mA 170mA 325mA 165mA 295mA 165mA
Table 3.2 Examples of standard linear fluorescent tubes in the range 550mm–2,400mm (1.8ft–8ft).
119
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Lamp cap
Double or triple coiled filament with oxide coating
Glass tube
Electrode shield
Bi-pin connector
Mercury vapor + inert gas
Coating on inside wall
Glass/metal seal
Figure 3.10 The standard fluorescent lamp.
stimulated to a higher energy level by the ultraviolet, but due to loss of heat energy, the quantum emitted when the electron drops back to a lower level has a visible, longer wavelength than that of the ultraviolet radiation. Fluorescence is the light emitted while the radiation is present, phosphorescence is light which continues to be emitted after the radiation is removed. The phosphors used in fluorescent lamps exhibit little phosphorescence, but it is sufficient to maintain a subjectively even light output as the current through the tube varies due to the AC supply. The cheaper halophosphate phosphors have a CRI of around 56. Halophosphate performance degrades comparatively quickly with time. Tubes with superior color rendering (CRI 80–85) and much better lumen maintenance use triphosphor technology. Intermediate tubes use a mixture of triphosphor and halophosphate, Figure 3.11. Triphosphors work on the basis of 3-color mixing (as opposed to a continuous spectrum). Three different rare earth phosphors are used; each one having an emission band which has a narrow peak. The peaks are around 610nm (red) 545nm (green) and 450nm (blue) close to the CIE tristimulus values. For CRI better than 85 multi-band phosphors are used, but the better color rendering (CRI > 90) comes at the expense of efficacy. Besides being classified by their color rendering abilities, fluorescent tubes are classified by their correlated color temperature. The standard whites are as follows: warm white CCT 3,000K white CCT 3,500K cool white CCT 4,000K daylight CCT 6,500K
120
The spectrum achieved by some typical fluorescent lamps is shown in Figure 3.12. Fluorescent lamps can convert about 20% of their input energy to visible light, and have an efficacy of between 25 and 80 lm/W, depending on arc length, phosphors and power supply frequency. The way in which energy is used in a typical lamp is shown in Figure 3.14. The fluorescent lamp is a constant voltage device. For a given tube length, the voltage across the arc remains the same, regardless of the current. After the initial arc has been struck, more and more atoms ionize and, without something to prevent it, the current would build up until the circuit fuse failed or the tube exploded. Discharge lamps need, therefore, a device to limit the current through them, and this is referred to as a Sn 02 Conducting layer AL2 03 Barrier layer Halo phosphate Triphosphor blend
Figure 3.11 Fluorescent tubes may have up to four internal coatings. A transparent conducting layer to aid starting, a transparent barrier layer to prevent contamination of the phosphors by the glass, a halophosphate layer and a rare earth triphosphor layer.
EVERYDAY LAMPS
Warm white halophosphate
Daylight multi-band Relative energy
Relative energy
Relative energy
Warm white tri-phosphor
Wavelength
Wavelength
Wavelength
Figure 3.12 Examples of the spectra achieved by typical fluorescent lamps. From Osram. Spectra are diagrammatic only.
ballast. Ballasts are of two principal kinds, electromagnetic (described in Section 6.3) and electronic (Section 7.2). In this section we only discuss the simplest type of electromagnetic ballast to help describe how a fluorescent tube works in practice. Clearly the use of resistance to limit current would be wasteful, so inductive reactance in the form of a choke is the most common way of limiting current. The simplest and most common form of fluorescent lamp circuit, using switched start is shown in Figure 3.15. When power is applied, current flows through the choke, one cathode, the starter, and then through the other cathode. In this configuration the cathodes incandesce, and are hot enough to emit sufficient electrons for the discharge. Phosphor drops back in stages one of which emits visible quantum
UV stimulates phosphor to higher levels
100% electrical energy in 3%
Emits UV quantum
Excited mercury atoms electron drops in level
After a short time the starter switch opens automatically. This causes a high voltage to appear across the tube – a combination of the line voltage and the inductive spike from the choke – and the arc strikes. If, for any reason, it fails to be maintained, the starter switch closes again and has another go. Assuming, however the arc is successfully initiated, an AC arc is maintained in the fluorescent tube. Notice that: • once the starter has gone open circuit, there is no heating current through the cathode filaments. The cathodes continue to be maintained at the correct temperature for electron emission by a combination of the ion bombardment and the arc current. • the simple inductive ballast circuit has a poor power factor. The simplest way of correcting this is by a parallel capacitor of the appropriate reactance.
Hg
UV 53%
Heat 44%
Fluorescence
Heat 79%
Phosphor coating
Figure 3.13 The mechanism of fluorescence within a fluorescent lamp.
Visible 21%
Infra red 37%
Dissipated heat 42%
Figure 3.14 The conversion of energy within a fluorescent lamp.
121
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Choke (electromagnetic ballast) AC Supply
Starter
Figure 3.15 The simple switch start circuit for a fluorescent lamp.
Alternative power factor correction circuits are given in Chapter 6. • the starter switch has a small capacitor across its terminals. This reduces radio frequency interference and helps prolong the starting pulse. The most common starter switch is the glow starter shown in Figure 3.16. It consists of a small gas filled bulb in which there is a bi-metallic electrode and a fixed electrode (or two bimetallic electrodes). Initially these are apart. When power is applied there is no voltage drop across the fluorescent tube, so the full line voltage appears at the starter, causing a discharge between the electrodes. The heat from the discharge bends the bimetallic element(s) so the electrodes touch, and cathode heating begins. But when they touch, the discharge stops, so the heating stops. The electrodes move apart creating the starting pulse. Once the tube is running, there is not sufficient voltage at the starter electrodes to recreate the glow discharge. There are other types of starters. The thermal starter uses a heating element to operate a bimetallic strip. It draws a continuous current, and is slow to operate – but does give reliable starting under cold conditions. Recently electronic starters have been introduced. There are also circuits which work on different principles. Control of the light output of the tube is achieved by controlling the current through it, but below a certain current level the discharge becomes unstable. Dimming of fluorescent tubes is possible with most types, but only by using circuitry which maintains the correct cathode temperature for electron emission. In anticipation of later chapters:
122
• the “old” T12 tubes are satisfactorily dimmed using electromagnetic ballasts and a separate cathode heating supply. • T8 and T5 tubes can only be satisfactorily dimmed using electronic ballasts. • fluorescent dimming is fine for architectural “scene setting” and energy management. It is not recommended for auditorium dimming down to blackout. Below a certain level the performance of different tubes varies, and anyway the lowest achievable level is far above blackout, resulting in the need for a sudden drop in light as blackout is achieved. • the best specified performance of fluorescent dimming is to 1% light level. But in practice this is only achieved with specified tubes at specified operating temperatures. It is much better to plan on 5% or more (but only after checking the tubes are suitable). • the long 6ft (1800mm) and 8ft (2400mm) tubes are not good candidates for dimming. Fluorescent lamps have long lives, typically 6,000 to 15,000 hours. However, life is affected by operating conditions. Frequent switch starting and unsuitable cathode heating circuitry can reduce life. The performance of fluorescent lamps is greatly enhanced by the use of electronic ballasts in place of the traditional electromagnetic ballast. Electronics are also making possible new lamp concepts. The T5 lamps listed in Table 3.2 can only be operated with electronic ballasts. T12 and T8 lamps of the same length are generally interchangeable, and, over a period of many years the T8 replaced the T12 for most applications. The narrower T8 tubes use a mixture of Krypton and
Fixed contact
Bi-metallic contact
Glass envelope low pressure inert gas with ionizer
Protective can
Twist socket contacts
Figure 3.16 The glow starter is the most common fluorescent lamp starter.
EVERYDAY LAMPS
Figure 3.17 Circular and U-shaped lamps are derived from their linear counterparts. While for some luminaire designs they are being superseded by compact fluorescent lamps, the new T5 circular lamps are proving popular. Photos from Osram.
Argon as the inert gas filling. The use of Krypton lowers the losses at the electrodes, and results in a lamp needing less power for a given light output. The T5 lamps listed in Table 3.2 take fluorescent lamp development further and represent a completely new generation. Their narrowness opens up new possibilities in luminaire design. Some points to notice about the T5 tubes are as follows: • their lengths are slightly shorter than their T12 and T8 “equivalents”. • for most lengths two different regimes are offered. For example the 1,149mm length tube is available at 54W (ca 5,500 lumens) or 28W (ca 2,700 lumens). This allows luminaire designers to achieve the right balance of light output and efficiency for any given design. • efficacy is of the order of 100 lm/W. • the conductive coating and the barrier layer applied to the inside of the tube prior to the application of multi-band phosphors play a significant part in the improved performance. This last point needs some explanation. The conductive layer assists the initial ionization in the tube. It is also designed to reflect UV, so any UV which has not been absorbed by the phosphors, gets refelected back into the tube as opposed to being absorbed by the glass wall. The barrier layer protects the phosphors from sodium in the glass which would otherwise shorten phosphor life. Mercury vapor is a poison, so anything which reduces its presence in the atmosphere is to be welcomed. The T5 tube uses only 10% of the mercury than its equivalent T12 of 25 years ago. Just as significant though, is the fact that if a given number of
lumens is produced by either incandescent lamps or modern fluorescent lamps, the incandescent lamps are actually responsible for more mercury vapor in the atmosphere than the fluorescent (even if the fluorescent tube’s mercury is allowed to escape into the air rather than being re-cycled). This is because much more power is needed and most power stations discharge significant amounts of mercury into the atmosphere. The high efficacy of fluorescent lamps resulted in many attempts to create different shaped lamps to allow greater flexibility in luminaire design. The most common shapes are the 20/40/65W T12 “U” shaped lamp, and circular lamps (approx 200mm, 300mm and 400mm diameter for 20/30/40W) based on 30mm diameter (T 9.5!) tubes.
3.5 Compact fluorescent lamps The long life and high efficacy of fluorescent lamps make them attractive for many architectural applications. However, their practical execution as long thin sources is not ideal for all applications. The compact fluorescent lamp uses a tube of T4 or T5 size, but is folded into two, four or even six “fingers” to make a high brightness lamp of short overall length. Such lamps were originally introduced with the idea of replacing GLS (General Lighting Service) tungsten filament lamps, but the range has now been greatly extended so that the largest of them are used in modular luminaires - replacing the straight 600mm (2ft) tube. Compact fluorescent lamps designed for the direct replacement of GLS lamps are two terminal devices, and have an internal electronic ballast. Table 3.3 gives an indication of the input power needed to achieve a given number of lumens using either incandescent or compact fluorescent lamps. Lamps designed for new luminaires are four terminal devices; and these in turn are divided into those which have an internal starter, and those which do not. Internal starter compact fluorescent lamps are used with suitable electromagnetic ballasts; those without a starter may be suitable for electromagnetic ballast and external starter, or may only be suitable for electronic ballasts.
123
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
It is accepted that in luminaires using low power lamps (say below 20–30W total) power factor correction is not usually applied. This is acceptable where only a few lamps are involved. However in schemes using large numbers of lamps, provision for correction must be made. If electromagnetic ballasts are used, the correction can be by central capacitors. Modern electronic ballasts include power factor correction. Compact fluorescent lamps work well when they are installed in luminaires which are designed round their characteristics, and when the designer or user has chosen a lamp of the correct color. Unfortunately there are many cases where they have been used indiscriminately – usually in a misguided attempt at “economy”. Truly grotesque examples of their use, for example by replacing tungsten lamps in candelabra in churches and similar sensitive applications with “daylight” compact fluorescent lamps, are all too often to be seen. Many compact fluorescent lamps are unsuitable for dimming, including most, but not all, of those intended as simple replacements for incandescnt lamps and those with built-in starters. Compact fluorescent lamps use a mixture of old and new filling and phosphor technologies, so if dimming is required it is essential to check that the lamps are suitable for it.
Approximate lamp output in lumens 100 240 400 600 900 1,200 1,500 1,800 3,000
Incandescent lamp load 15W 25W 40W 60W 75W 100W 120W 150W 200W
Compact fluorescent lamp load 3W 5W 7W 11W 15W 20W 23W 28W 42W
Table 3.3 A guide to the equivalence of incandescent lamps and their compact fluorescent equivalents. Note that the appearance of the compact lamp may be significantly different unless chosen to match incandescent. This means choosing lamps with the correct phosphor blend and a CCT of around 2,700K.
124
Conventional base
Electronic ballast
Folded tube
Figure 3.18 Construction of a compact fluorescent lamp intended as a replacement for incandescent lamps. Figure from Osram.
The best way to dim those that are suitable is to use luminaires fitted with controllable electronic ballasts. The same restrictions concerning the minimum achievable light levels apply as to the standard tubular fluorescent lamp.
3.6 Special purpose fluorescent lamps 3.6.1 Variations on a theme. There are a lot of “special” types of fluorescent lamps designed to meet the needs of particular applications and industrial processes. A few of them are listed here. When the fluorescent lamp was first invented, and phosphor science was in its infancy, it was expected that its main application would be as a source of colored light. In practice fluorescent lamps using single color phosphors, red, green and blue are available, but are not widely stocked. Another way of achieving a single color lamp is to sleeve a white lamp with a color filter. However, this can only be done by comparing the filter characteristic with the spectral distribution of the lamp. Tri-phosphor lamps emit at specific wavelengths, and the result of using filters can be disappointing if the filter does not match the peak wavelength. Manufacturers use different trade names for lamps with enhanced or diminished radiation at particular wavelengths. Tubes with enhanced red and blue are
EVERYDAY LAMPS
4 pin lamps for external control gear suitable for dimming
18-80w
225mm-570mm long
10-26w
87mm-150mm long Built in electronic ballast Not suitable for dimming
8-21w
5-15w
46-70mm dia
16-21w
100-120mm dia
Figure 3.19 Examples from Osram of the many variants of compact fluorescent lamps.
used for lighting plants and aquariums. The enhanced wavelengths promote photo-biological action. Lamps with specially chosen red phosphors are used for the illumination of meat and delicatessen products in shops and supermarkets. Lamps with a somewhat yellow cast, which do not emit any UV radiation at all, are used in factories where UV lithography is the main process (e.g. chip manufacturing and other “clean room” processes).
Ultra violet radiation is divided into three bands: UV-A 400nm – 315nm UV-B 315nm – 280nm UV-C 280nm – 100nm Lamps for sun-tan beds and solaria are available in various grades. The “weakest” radiate only at wavelengths 350nm and longer; they are considered safe for long exposure and for maintaining a tan. The middle range lamps radiate from 315nm upwards, creating new pigment, but again, considered safe for long exposure. The “strongest” sun-tan lamps have a more “sun like” UV spectrum containing some UVB, and must only be used for limited periods. Lamps for sun-tan beds and solaria are T12. While short (590mm 40W) lamps are available, most of the market is served by long lamps, for example 80W 1,500mm, 100W 1,800mm, 140W 1,500mm and 160W 1,800mm. The lamps operate at comparatively high current (1A for the 100W and 1.5A for the 140W.) Black glass fluorescent lamps are fitted with a black filter to eliminate almost entirely the visible radiation. They radiate in the range 350–400nm with a strong peak at around 365nm. They are used in processes based on fluorescence, ranging from the detection of cracks in metal to the detection of forged bank notes. They are also used for special effect lighting in the theater and other entertainment applications requiring the use of “black light”. In such applications there may be a call to provide dimming control to allow the black light to be treated like any other light source.
3.6.2 Ultra violet lamps Obviously fluorescent lamps without the phosphor could be used as a source of UV radiation. However, short wavelength UV is dangerous, and anyway would be severely attenuated by the glass. A range of UV tubes is available for different applications.
Figure 3.20 Encapsulite manufacture a range of color sleeves for fluorescent lamps, some of which also provide protection in the case of tube breakage. Here they are used in the Anytronics “Anycolour” fluorescent luminaire.
125
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 3.21 Examples of UV “fluorescent” lamps, from Osram. “Blacklight” lamp used in entertainment applications (above). Compact lamp for insect attraction (below).
Insects are attracted by near UV. Both compact and linear lamps are available which emit light in the 350–400nm range with some visible violet radiation. When fitted in an insect killer, such lamps attract insects onto an electric “zapper” which kills them. The same lamps are used for other applications like the polymerization (“curing”) of plastics and adhesives, and in diazo copiers. Both compact and linear lamps are available which produce UV-C, with a peak at 253.7nm. Wavelengths of this order are germicidal and of comparatively high photon energy, so these lamps are used for: • the disinfection of water, for example drinking water and swimming pool water. • the disinfection and de-odorizing of air, in air conditioning plants. • the disinfection of surfaces in the food and pharaceutical industries. • the erasure of EPROMs. Linear UV lamps of all kinds tend to use electromagnetic ballasts, whereas the compact lamps are more likely to use electronic ballasts.
light distribution were achieved using the reflector fluorescent lamp. In these lamps a reflective layer is deposited over 200°–330° of the tube prior to the deposition of the phosphor. Such a construction is useful for situations like cabinet lighting and the edge lighting of signs where it may not be possible to fit a proper optical reflector. It is also useful in situations where the top of a horizontally mounted fluorescent tube may gather dust and be inaccessible for cleaning. The reflection is not total, and some light gets through the reflective layer; but the great majority is directed out of the 30°–160° gap remaining. There is a slight loss of overall lumens. The tubes are otherwise conventional and are suitable for dimming with suitable control gear. 3.6.4 Instant start lamps It can come as surprise that some hot cathode fluorescent lamps come in versions with only a single connection pin at each end. From the description of the simple starter circuit in Section 3.4 it is clear that, once an arc has been struck and sufficient current is flowing, the cathodes maintain thermionic emission and a high temperature from the ion bombardment
Glass Reflecting layer Phosphor
Normal
3.6.3 Reflector fluorescent lamps The arrival of the T5 family of lamps has simplified the optical arrangement for some applications, allowing slim, compact, luminaires with good light distribution to be produced. Prior to the adoption of long T5 lamps, and still a valid construction, narrow fittings with a unilateral
126
Reflector Approx 30
Typical light distribution pattern
Figure 3.22 Reflector fluorescent tubes have an internal reflective coating over part of the tube. The Philips TL-D has an especially narrow aperture of only 30°.
EVERYDAY LAMPS
Figure 3.23 An example of an instant start hot cathode fluorescent lamp, used in non-pressurized luminaires in semi-hazardous atmospheres. Notice the single pin connection and the conducting strip used to aid starting. The lamps are T12 and are available in nominal lengths 440, 600, 1,200 and 1,500mm. Photo from Philips.
alone, and do not require any continuous heating current. Is it possible to do without the starter? With appropriately designed tubes the answer is “yes”. The starting circuit described in Section 3.4 is known as pre-heat starting. Other starting regimes for hot cathode fluorescent lamps will be described in the chapters dealing with ballasts, but here we can anticipate one type of starting known as instant start or cold start. In this case there is no pre-heating, and starting relies on field emission from the cathodes, requiring a higher starting voltage across the tube. Lamps of this kind are sometimes specified for use in increased safety damp-proof and explosion proof luminaires. The possibility of local heating and arcing from the individual cathode circuits due to poor connection or faulty ballast is eliminated. Instant start lamps need correct conditions for starting. The initial glow discharge can be between an electrode and an outside groundplane before the main arc strikes. It has been found that tubes should have a silicone, non wetting, finish to give a high surface resistance, and should be within 10mm of a grounded metal surface – usually the luminaire itself. An alternative arrangement is to fit the tube with a thin metal strip, connecting both of the metal end caps, which should be grounded. Cold start lamps are not suitable for dimming.
For example in a T8 hot cathode fluorescent tube the optimum electric field for UV generation is around 1 V/cm, so a typical volt drop down the arc is around 100V. But in addition to the arc volt drop, there is a volt drop at the cathode, the voltage required to create ion and electron flow. This drop is referred to as the cathode fall. In a hot cathode lamp thermionic emission of electrons arising from the heating of the cathode to 1,100°C, and the use of an oxide coating providing a low work function (meaning that not so much effort is needed to peel the electrons away) results in a cathode fall of around 10–12V. However, a discharge can still take place in a tube fitted with electrodes which have no heating arrangement. Such lamps are called cold cathode lamps, and in this case the cathode fall can be as high as 50V. This means that the tube must operate at a higher voltage. Cold cathode fluorescent tubes are not as efficient as their hot cathode equivalent for “standard” lengths. For long lengths the electrode loss contribution gets proportionately less. Cold cathode lamps have a long life, because they do not suffer from cathode emission degradation, and are not affected by repeated starting. For these reasons cold cathode fluorescent has a number of specialist applications:
Voltage 500 Cold cathode 400 300 200
Within a discharge tube there is a substantially linear voltage drop down the length of the arc discharge.
mA 50
Hot cathode
A 0m 50
100 0
3.6.5 Cold cathode fluorescent lamps
A 0m 10
500 1000 1500 2000 2500 Arc length mm T8 Tube
Figure 3.24 The operating voltage of hot and cold cathode fluorescent lamps as a function of arc length.
127
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
3.6.6 T2 and miniature fluorescent lamps 25mm
20mm
Fluorescent lamps are normally associated with relatively large physical dimensions and wide diameter tubes. However, there is a significant industry devoted to producing miniature fluorescent lamps. The largest of these are the hot cathode 7mm diameter (T2) linear lamps in 6–13W ratings and lengths of 220–520mm (8.6–20 in). These lamps are, typically, available with the same colorimetry as mainstream T5 lamps, and can, therefore, complement them in lighting schemes requiring unusual luminaire designs. In common with the T5 lamp, these miniature hot cathode lamps require electronic ballasts. The smaller sizes of miniature fluorescent lamps are made in a wide range of tube diameters, including 6.5, 4.8, 4.1, 3.8, 3.0, 2.6 and 2.0mm. Linear tubes are in the range 38–420mm, and nominal ratings from less than 1W to around 6W. These lamps are cold cathode, and are used mainly for instrumentation and backlighting, for example as backlights for LCD screens in portable computers.
Figure 3.25 Construction and typical operating circuit of a cold cathode lamp.
• where a long run of light is needed, without an obvious break. For example cove lighting and other specialist architectural applications. • where custom shapes are needed. • where repeated flashing of the lamp is required. A lot of sign lighting is done with cold cathode; most “neon” sign lamps are not neon at all, but cold cathode fluorescent. One of the reasons that cold cathode lamps are useful for this purpose is that they operate over a wide temperature range, and always start. The light output of hot cathode lamps drops dramatically below 0°C, and at low temperatures they can be difficult or impossible to start. Long cold cathode lamps need relatively high voltages to operate (see Figure 3.24) so are operated from transformer type ballasts or their electronic equivalent. They are suitable for dimming. The cathodes in cold cathode lamps are in the form of a hollow cylinder, as shown in Figure 3.25.
128
7mm hot cathode (Osram)
3.8mm cold cathode (LCD Lighting Inc.)
Figure 3.26 Examples of miniature fluorescent lamps.
EVERYDAY LAMPS
Often these miniature lamps are made in customized forms with multiple bends to give even backlighting. They all work with electronic ballasts; such ballasts are quite specialized, since a typical specification can be as demanding as: • input voltage 12V DC • lamp starting voltage 2,000V • lamp running voltage 480V rms • lamp current 5mA Users of laptop computers might be surprised to learn that such high voltages are present in their machines! For instrumentation use the ballasts may be supplied as dimming ballasts, in which case PWM control is normally used, although some simple ballasts give limited control by varying the input current to the ballast. While cold cathode lamps work down to quite low temperatures, their visible light output then goes down significantly. For some specialist applications miniature cold cathode lamps are provided with heater coils loosely wound round the exterior of the tube. But for most applications it is sufficient to wait for the infra-red, which is still generated at low temperatures, to warm up the lamp. 3.6.7 Long life fluorescent lamps The life of fluorescent lamps is determined by a number of factors. The most significant are: • loss of lumen output due to phosphor degradation. • lack of electron emission arising from cathode failure in hot cathode lamps. In the case of hot cathode lamps useful life is greatly extended by the use of the correct electronic ballast. Cold cathode lamps generally have a longer life than their hot cathode counterparts (e.g. 20,000 hours rated instead of 10,000). Some specialist manufacturers produce long life fluorescent lamps by combining a number of techniques:
Figure 3.27 The “Thermo-LL” lamp from Auralight AB combines a 36,000 hour life with enhanced low temperature performance (to -20°C) by virtue of its glass enclosed construction.
• first and foremost by extra attention to eliminating phosphor contamination during lamp manufacture. • secondly by using a more complex (and expensive) cathode construction designed to maintain electron emission, and to minimise the effect of lamp starting (which otherwise tends to damage the cathode by the stress of field emission). • choice of phosphor. Such lamps may have a rated life of around 36,000 hours instead of the more usual 12,000 hours or so. In this case “life” might be defined as the time taken for lumen output to drop to 70% of initial lumens, and based on a 3-hour switching cycle. Tri-phosphor lamps have a shorter useful life, so long life versions of these might give, for example, 25,000 hours. 3.6.8 Electrodeless fluorescent lamps The extra cost of long life lamps can be recouped on the first lamp change if luminaires are difficult to access – the cost of replacement can easily outweigh the additional cost of long life lamps. Besides cold cathode and long life hot cathode lamps there is a range of electrodeless fluorescent lamps with very long lifetimes. These are described in Chapter 5 under the heading of “Induction Lamps”.
129
Chapter 4
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Arc
lamps
4.1 High intensity discharge lamps
4.2 Mercury vapor lamps
The carbon arc was the first commercially useful high intensity source of light powered by electricity. It consists of two carbon rods which are brought together and then drawn slightly apart while carrying an electric current. An arc is formed between the two and visible radiation results both from the incandescense of the electrodes and the flame luminescence resulting from the vaporizing of the electrodes. Carbon arc lamps operate in the atmosphere and, as a result, have a short burning time and require continuous attention. Nonetheless they were widely used until comparatively recently in film projection, film production and searchlights. Now they have been replaced by xenon arc and metal halide lamps according to application. All other practical arc lamps operate in a controlled atmosphere within some kind of arc tube. The fluorescent lamp already described in Chapter 3 is one kind of arc or discharge lamp, in its case characterized by a long arc length, low pressure of operation within the arc tube, and a low tube wall temperature. The arc lamps described in this chapter have comparatively short arcs and operate under more stringent conditions than the fluorescent lamp. They are generally characterized by high luminous efficacy and/or high point brightness. The main family of lamps such as mercury HID (High Intensity Discharge) and high pressure sodium lamps are widely used for industrial, public area and street lighting. Metal halide lamps can have a pleasing color performance and are made in many different formats, making them suitable for architectural lighting and even for projection. Xenon lamps have a very short arc length and a near continuous spectrum, making them ideal for projection.
The Mercury Vapor HID lamp was the first HID lamp to be widely used commercially. Electromagnetic radiation is created from a discharge within mercury vapor, but the regime is different than that found in the normal fluorescent lamp: • when running, the pressure within the lamp is in the range 200–400kPa (compared with only 1Pa in the fluorescent lamp, and equivalent to 2–4 times atmospheric pressure). • it is not possible to achieve the mercury vapor discharge in a cold lamp. For this reason the lamp also includes argon, and the initial arc is struck as an argon arc. The energy from this discharge vapor-
130
Figure 4.1 The carbon arc lamp, in this case built in to a stage follow spot. Photo from Strand Lighting/Brian Legge.
ARC LAMPS
reflector
elliptical
globe
ply is not great enough to initiate the argon ionization. Provision has to be made to start the discharge with an auxiliary electrode. • once running, the arc produces a much greater proportion of visible light than the discharge given in a fluorescent lamp. Main wavelengths are 405, 436, 546, 577 and 579nm. Together these give a bluish-white light, but it is almost totally deficient in red. • for this reason most mercury vapor HID lamps also include a phosphor layer. The mercury vapor still produces a lot of ultra-violet, and the phosphor can be used to improve the color rendering and give higher overall efficacy.
Standard coating
Relative power
Outer glass soda-lime <125W Borosilicate >125W
450
550
650
Main electrodes
650
Starting electrode
Wavelength in nm
Support
Nitrogen filling
Quartz arc tube
Relative power
Warm coating
450
550
Wavelength in nm
Mercury Argon filling
Figure 4.2 Examples of Mercury Vapor High Intensity Discharge Lamps above; and the spectrum of the lamps, below. Illustrations from Osram.
izes the mercury to get the main discharge going. • metal oxide electrodes are used to provide electron emission. They are not separately heated, and rely on the arc bombardment energy to maintain the correct temperature. • the field provided by the standard electricity sup-
Resistor
Figure 4.3 The construction of the mercury vapor HID lamp.
131
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Lamp L
Ballast
N Power Factor Correction (PFC) Figure 4.4 Electrical circuit for the mercury vapor HID lamp.
A typical construction for a mercury vapor HID lamp is shown in Figure 4.3. The high pressure and temperature needed for the arc requires the use of a compact arc tube, usually made of quartz. This in turn requires molybdenum ribbon connections (for the same reason as the tungsten halogen lamp.) The arc tube itself is housed in an outer envelope for a mixture of practical and safety reasons. The outer envelope is usually made of borosilicate glass. It ensures a more practical outside wall temperature, provides an interior surface on which the phosphors can be deposited, and filters out undesirable ultra-violet wavelengths. The outer envelope itself is usually filled with nitrogen. This prevents oxidation of internal parts and ensures a stable temperature regime for the arc tube. An auxiliary electrode is fitted to aid starting of the lamp. It is fed through a high resistance, and creates a local glow discharge to start the ionizing of the argon. The starting period is characterized by a low voltage across the lamp and a current somewhat higher than the eventual running current. As the lamp warms up an increasing amount of the current is taken by the mercury vapor. It takes several minutes for the vapor pressure to stabilize. If power is disconnected, the lamp will not relight if power is restored quickly. It must cool down until the vapor pressure has dropped to the point where an arc can be established with the available voltage. The cooling down period is similar to the warm up time – a few minutes.
132
Mercury vapor HID lamps have an efficacy in the range 30–60 lm/W and a life of 12,000 to 24,000 hours. Ratings are from 50W to 2,000W, but the most popular sizes are in the 80W to 400W range. They are favored for outdoor and some industrial use because of their comparative simplicity, but the quality of light is not so good as other sources because the light is based on comparatively few spectral lines. The circuit of a typical lamp is shown in Figure 4.4. It is similar to that of the fluorescent lamp, but has no starter. There are limited possibilities for controlling HID lamps – if the current is reduced too much the arc becomes unstable. Some limited control of light output is provided by the use of tapped chokes. Note that the simple circuit of Figure 4.4 applies when the line voltage is sufficient. Chapter 6 describes different types of HID ballast used for various lamp types and line voltages. A variant of the mercury vapor HID lamp is the self ballasted lamp. This was introduced to allow luminaires fitted with high wattage tungsten lamps to operate at higher efficiency without the need to fit ballasts. In these lamps the arc current is limited by a resistance – but the resistance used is a tungsten filament within the outer envelope, and itself acts as an incandescent lamp. The idea has the merit of simplicity, and the incandescent filament helps with color rendering at the red end of the spectrum. However, the resulting lamp does not have a particularly high efficacy – it is barely double that of a tungsten lamp. This is because the arc only strikes at a threshold voltage, so current through the filament is not continuous – it therefore takes a disproportionate amount of power. High pressure mercury vapor lamps are also supplied as sources of UV-A and UV-B. In this case, as in that of the low pressure “fluorescent” tube, the outer envelope is black Woods Glass, which blocks the visible radiation, and lets through the long wave ultra violet. These lamps are used in similar applications to the linear lamps, but where a point source is more appropriate.
ARC LAMPS
4.3 Sodium and high pressure sodium lamps 4.3.1 Low pressure sodium lamp
2
mw per m per 5 nm per 1000 lux
The low pressure sodium lamp generates light from a discharge in sodium vapor at the low pressure of around 0.7–1Pa. “Sodium lighting” is familiar because of its almost monochromatic nature, actually two yellow wavelengths very close together at 589.0 and 589.6nm not far from the eye’s peak sensitivity. Terrible for color rendering, but great on efficacy at between 150 and 200 lm/W. The sodium “D lines” can also be generated if you drop some salt into a flame, when sodium’s characteristic yellow radiation can again be seen. Low pressure sodium lamps are made in ratings from 18–180W, and have a life of between 16,000 and 23,000 hours. The 180W lamp gives around 33,000 lumens. Low pressure sodium lamps need a long arc for low current density and are constructed either double ended, or, more commonly, single ended. The single ended construction requires a U shaped tube. In both cases the tube is made of sodium resistant glass and is mounted inside an evacuated outer
2400 2000 1600 1200 800 400
SOX 55 W, SOX-E 36
300
Outer jacket
Arc tube support Neon and sodium filling
ITO IR reflective layer
Some designs have dimples for sodium condensation
High vacuum Inner arc tube
Coated tungsten cathode
Sodium resistant layer
Lamp cap
Figure 4.6 The construction of the low pressure sodium lamp.
envelope. This is necessary both for safety and to ensure that the main tube wall maintains a temperature of 260°C. Sodium resistant glass is damaged by common impurities, so the actual construction of the arc tube is “two ply”, where a thin coating of sodium resistant glass is applied to the inside of the discharge tube. The outer envelope has a thin internal coating of indium-tin oxide which reflects the IR radiation Lamp
L PFC 400
500
600
700
Leakage reactance auto-transformer
800
Wavelength in nm
Figure 4.5 A low pressure sodium discharge lamp, SOX, from Philips (above) and the spectrum of the lamp (below). This series of lamps is between 300 and 1,100mm long, with ratings between 35W and 180W.
N
Figure 4.7 Electrical circuit for the low pressure sodium lamp.
133
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Coated elliptical 1000W 400mm Mercury free 80W 150mm
Double ended 400W 200mm Clear 250W 250mm Relative power
Standard
450
Relative power
back to the lamp, helping to maintain the optimum tube wall temperature. As with mercury vapor HID, the sodium lamp needs a starting gas. In this case it is 99% neon and 1% Argon at a pressure of around 1,000Pa. Sodium street lights are noticeable for their dull red neon discharge as they warm up. Warming up time is quite long, at around 10 minutes; however, the hot reignition performance is good. Low pressure sodium lamps need around 400– 500V to start. Because of the high starting voltage the ballast needs to be a combined transformer and current limiting reactance; the ballasts used are described in Chapter 6. The lamp’s main application is in street lighting and, like mercury vapor lamps, the need and opportunities for variable output control are limited.
550
650
550
650
Wavelength in nm
Mercury free
450
Wavelength in nm
Figure 4.8 Examples of High Pressure Sodium Discharge Lamps (above) and the spectrum of the lamps (below). Illustrations from Osram.
134
4.3.2 High pressure sodium lamp The high pressure sodium lamp (HPS) operates at a pressure of 5–10kPa. Although uneven, with some pronounced peaks, the spectrum is nearly continuous, resulting in a pleasantly warm light. Curiously the strong sodium D lines at 589nm, which we associate with “sodium” are not present, because at this pressure the D line light is self-absorbed by the gas. The main spectra either side of the absorption dip can be regarded as a broadening of the original lines. HPS lamps have negligible radiation in the ultraviolet, but around 25% of input power radiation in the infra-red. The high efficiency versions can radiate as much as 25% visible radiation. The balance is lost in electrode losses and non-radiated losses in the arc column. The lamp construction is shown in Figure 4.9. At high pressure sodium is highly corrosive, and neither glass nor silica can be used. The arc tube is made of a ceramic, polycrystalline alumina, which is a translucent material transmitting about 90% of the light. The alumina is difficult to work, and cracks easily, making connection to the tungsten electrodes difficult. A ceramic plug is fitted into each end of the arc tube, through which a niobium connection is made. The plug is able to bond with both the alumina and the niobium.
ARC LAMPS
Once again the arc tube is itself mounted in an evacuated outer envelope. Sometimes the outer envelope has a diffusing internal surface to create a larger, lower point brightness, source. Lamps are available in the range 35W to 1,000W, and efficacy is between 45 and 150 lm/W depending on size; efficacy goes up with power rating, levelling out above 400W. A higher sodium pressure improves the color rendering, but decreases efficacy. A higher xenon pressure improves efficacy, but makes starting more difficult in the larger lamps. Life is between 10,000 and 24,000 hours; unfortunately the shorter life applies to the lamps with the best color rendering. The starting of a high pressure sodium lamp is complex. First the starting gas is a mixture of xenon and a sodium-mercury amalgam. The xenon at 20kPa gets the discharge going, and the mercury vapor (60kPa when warmed up) helps raise the pressure. Warm-up time is about 10 minutes, and re-strike time about one minute. Arc tube support and current lead-in
Hard glass envelope
Coated tungsten electrode
Vacuum
Ceramic arc tube
Sodium amalgam and xenon filling
Glass seal Alumina plug
Niobium wire Lamp cap
Figure 4.9 The construction of the high pressure sodium lamp.
L
Ballast Lamp
PFC
Ignitor
N
Figure 4.10 Electrical circuit for the high pressure sodium lamp.
Second, the starting voltage required is much higher than can practically be provided by transformer action. Starting is achieved by using a component called a superimposed pulse ignitor. This device produces a series of high voltage pulses, in the range 2.5kV to 4kV depending on lamp size. Color rendering ranges from R a = 25 for the “standard” HPS, through 60 for intermediate lamps up to 85 for the highest pressure “improved color” versions. The comparatively good color rendering of the improved color HPS lamp means that it is used for architectural applications, and there can be a demand to dim the light, both for energy management and practical/esthetic reasons. It is possible to achieve limited dimming but: • lamps should always be run up to full first. • the range is limited (typically 50% is the very best that can be achieved, and 60% is a practical minimum). • even at the 60% level, color rendering is affected, with significant changes to lamp color. • dimming can shorten lamp life. A recent development is the mercury-free sodium-xenon HPS lamp. In this lamp the xenon pressure is around 10 times higher than in the normal HPS lamp, and no mercury vapor is used. The arc tubes are significantly longer and thinner than the standard type. Starting is more difficult and in practice this type of lamp is only offered in low ratings (35–80W) using an electronic ballast with builtin igniter. This arrangement does allow the current through the lamp to be precisely controlled, both in
135
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
30 0m m
72W compact with ceramic discharge tube
20 0m m
2000W tubular
11 0m
m
1200W PAR 64 reflector
250W single ended for theatre luminaires
250W compact source
50 0m m
180mm
80 m m
400W elliptical suitable for open fitting
400W tubular double ended
40 0m m
10 0m m
respect of absolute current and current waveform. As a result of this such lamps can be offered with a selectable color temperature.
18000W 44mm arc length
Figure 4.11 Examples of Metal Halide Discharge Lamps (photos from Osram). The mm dimension is the approximate overall length.
136
4.4 Metal halide lamps Metal halide lamps can be considered as a variant of the high pressure mercury vapor lamp. In addition to mercury vapor and argon, these lamps contain metal halides. The halides can be a mixture of rare earth halides, usually iodides, or a mixture of sodium and scandium iodide. The mercury vapor radiation is now augmented by that of the metals, which use a regenerative cycle not unlike the one used in tungsten halogen lamps. Once the lamp has achieved operating temperature, the metal halides vaporize. However if the halide vapor reaches the high temperature core of the discharge, it dissociates into the halogen and respective metal, and the metal radiates its associated spectrum. The separated halogen and metal molecules move by diffusion and convection to the cooler parts of the tube, especially the walls, where they recombine, for the cycle to start again. Metal halide lamps have efficacies in the range 75–125 lm/W. Color rendering is significantly better than mercury vapor, and can be tailored by the choice of halides. Life is in the range 6,000 to 20,000 hours, and ratings range from 35W to 20,000W. There are various different approaches to starting metal halide lamps: • all small lamps use high voltage pulse start. • in Europe it has been the practice to use superimposed pulse start ignitors on the larger lamps for many years (3–7kV pulses, depending on lamp type). Recently this approach has gained ground in the USA as well. • however, auxiliary electrode start is also possible on the higher rated lamps when suitable ballasts are used. • but in this case the reactive nature of the filling can lead to electrolysis in the quartz envelope between the main and auxiliary electrode. For this reason some lamps include an arrangement (operated by a bi-metallic switch) to disconnect the starting electrode.
ARC LAMPS
The standard circuit for commercial metal halide lamps in the 35W–1,000W range, when using superimposed pulse ignition, is the same as that for HPS lamps, as shown in Figure 4.10. Metal halide lamps operate at a significantly higher pressure than the normal mercury lamp, in the range 1,000–1,500kPa. They also emit considerable quantities of UV. For both reasons they may only be operated in specially designed luminaires, or be offered in special versions, which filter out the UV, and which can contain the lamp fragments in the case of catastrophic lamp failure. Some lamps using the conventional double envelope construction use UV-stopping quartz, obviating the need for external filtration (while still requiring a shatter proof cover). There are also lamps designed for use in open luminaires, but in this case they have a “third jacket” within the outer envelope to provide the necessary protection. Lamps designed for open luminaires have special bases to prevent the “standard” lamp being fitted by mistake. The range of possibilities afforded by different lamp constructions and halide dosing has resulted in the metal halide lamp being made available in a huge range of formats. Figure 4.11 gives an idea of this range. Some points to note are: • CCT is in the range 3,600–4,200K for the general use lamps. Rare earth dosed lamps achieve a higher CCT in the range 3,800–5,600K
Relative intensity
3
2
1
0
400
600
800
1000
1200
1400
1600
1800
2000
2200
Wavelength in nm
Figure 4.12 The spectral radiant intensity distribution of an Osram HMI 4,000W lamp closely matches that of the sun (gray line). It is an example of a metal halide lamp dosed with rare earths and results in a high Ra of 95. “Standard” metal halide lamps have Ra = ~ 70.
Hard glass envelope
Transparent protective shroud
Metal halide filling
Quartz arc tube
Figure 4.13 Construction of the general service metal halide lamp suitable for open luminaires.
• CRI (Ra) is around 70 for the general use lamps, but can rise as high as 95 with rare earth dosed lamps, at the expense of lamp life. • some lamps use an internally phosphor-coated outer envelope. This converts UV to visible radiation. The technique provides a more diffuse (lower point brightness) source and gives a “warmer” light, but does not significantly improve efficacy. • “single element” dosing of metal halide lamps can be used to create moderately high efficacy color lamps, considerably more efficient than using color filters if the color is acceptable. Indium yields a blue light, and thallium a green light. • metal halide lamps are also available in UV versions for industrial applications. • most metal halide lamps use a quartz (fused silica) arc tube; however some lower rated lamps now use ceramic arc tubes (as used by HPS.) This allows a higher operating temperature and results in improved color rendering. Lamps using ceramic tubes are referred to as CDM lamps, Ceramic Discharge Metal halide. They are now widely used in architectural applications because they are available in low power ratings (e.g. 20W, 35W and 70W).
137
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Metal halide (and other) arc lamps may have some restriction in respect of burning angle. Lamps designed for vertical burning fail if operated horizontally because the arc bows upwards due to convection currents. This both overheats the top side of the arc tube and disturbs the halide cycle. Lamps designed for horizontal burning have shaped or assymetrical arc tubes. Some of the shaped arc tube designs for general use lighting achieve the highest efficacies. The remarks concerning dimming are basically the same as for HPS, with the additional problem that as current is reduced, the color temperature actually goes up, resulting in a more “blueish” light. This is because the metals providing the improved red in the lamp are the first to condense out. However the need for some degree of level control is now appreciated, and both improvements in the lamps and the availability of controllable electronic ballasts mean that dimming of these lamps (over a limited range) is becoming a practical proposition. Metal halide lamps have a warm-up time of between one and four minutes. Most general use lamps are not suitable for hot re-strike and must be allowed to cool for a period of two to ten minutes before restrike is attempted. Some lamps (including all the “compact” types referred to in the next section) are designed for hot restrike. However, such restrike requires a much higher superimposed pulse than normal, in the range 20–60kV depending on the lamp type.
4.5 Compact source metal halide lamps Metal halide lamps are made with a range of arc lengths. At one extreme a 1,500W double ended lamp designed for floodlighting has an arc length of 190mm, at the other extreme the latest generation of projection lamps has an arc length of 1.3mm. Lamps intended for general use have arc lengths of around 40mm for a 250W lamp to 140mm for a 2,000W lamp. Lamps with medium or short arc lengths (compared with the general use type) are referred to as compact source metal halide lamps. The
138
“medium arc” lamps used for applications such as film set lighting have arc lengths from 6mm for 400W lamp up to 44mm for an 18,000W lamp. Typical compact source lamps have arc lengths around 4mm for a 400W lamp up to 15mm for a 2,500W lamp. The short arc allows the lamp to be used in focusing optical systems. Lamps are often supplied with integral ellipsoidal reflectors, and applications are as diverse as: • theatrical luminaires (principally moving light fixtures, with mechanical shutter dimming) • video, data and graphics electronic image projectors (but see also Section 4.6) • fiber optic illumination • film and television lighting • automobile lighting • architectural and display feature lighting Most of these applications demand good color rendering, and much experimentation has gone into the choice of halide mixture to achieve it. Both rare earth iodides and bromides are used. Significant problems with the compact source metal halide lamps are achieving long lamp life, and maintaining color rendering through life. There can be a tendency for different life performance from the different halides, resulting in a color shift during life. This problem is being addressed, and some of the newer projection lamps combine reasonable life with maintained color performance. The higher color temperature needed for these applications is achieved by the use of different halides, particular that of tin. Lamps based on tin halide are good at maintaining color temperature when the lamp is dimmed. Medium arc lamps intended for film set lighting have a CCT of 6,000K. Most other compact source lamps are available with CCTs in the range 4,800–6,500K; but those intended for video and data projection are in the range 7,000– 8,500K. Lamp life depends on the lamp power and arc length; it ranges from around 750 hours to 10,000 hours or more. Some projector lamps (e.g. 400W) are designed for “boosted” operation, whereby they can be operated at up to 50% above nominal running current, but with a sacrifice in lamp life of 50%.
ARC LAMPS
In general compact source lamps use electronic ballasts with integral ignitors, some of which allow limited control. They are able to re-strike instantly when hot. When used in theatrical applications, for example in moving light luminaires, dimming of the metal halide light is achieved using an iris shutter or similar mechanical arrangement. This varies the light output from the lamp, which itself remains full on at all times the luminaire is in use.
4.6 High pressure mercury vapor lamps for projection Compact source metal halide lamps are commonly used as the light source for video and data projectors, but currently the preference is to use xenon arc for the larger projectors, and high pressure mercury vapor arc lamps for small and medium sized projectors. This is because the arc length required is small, typically around 1mm, and it is difficult to get a reasonable lamp life with metal halide. In electronic projectors the object being illuminated is very small, and efficient light collection requires a source of the order of 1 Gcd/m2. The technology used is popularly called UHP (a Philips trademark) Ultra High Performance or Pressure. UHP and its equivalents from other manufacturers is actually a pure mercury vapor lamp; but the lamp is operated at very high pressure (around 20 MPa, about 200 times atmospheric pressure.) The
high pressure yields a continuous spectrum with reasonable red performance, and also gives a high arc voltage (≈80V) despite the short arc length. Color performance remains constant throughout the life of the lamp. UHP lamps are currently made in the range 100– 200W where they have the desirable characteristic of long service life (around 8–10,000 hours for the 100W version.) They achieve this partly due to the use of a small dose of bromine that maintains the tungsten arc electrodes. UHP lamps require dedicated electronic ballasts to provide a special waveform to the lamp that ensures a stable arc that does not “jump”. Such jumping would otherwise shorten lamp life and be visible as a flicker. The ballast also provides the 20kV needed for starting the lamp. Recent developments have reduced the starting voltage to 5kV. Normally 20kV is required to extract electrons from tungsten, but they can also be produced by photo-emission if low wavelength (<270nm) UV is present. The UHP lamp can be fitted with an auxiliary discharge cavity that produces UV which in turn is directed to the main lamp by total internal reflection. This reduces the required ignition voltage to 5kV. Hot restrike would then be a problem, but this is overcome by fitting an external field electrode that modifies the field within the arc tube to allow re-ignition after only a short waiting time.
l
t
Figure 4.14 The special lamp current waveform that ensures stable arc attachment within a UHP lamp. Information from Philips.
Figure 4.15 A UHP lamp in a PAR19 reflector with its electronic ballast. This one uses the low voltage ignition system. Illustration from Philips.
139
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
4.7 Xenon arc lamps The arc discharge in xenon is valued for its continuous spectrum. Although a little peaky around 475nm, the spectrum is continuous between 380 and 760nm with only around ± 20% variation. In fact the peak of radiation from xenon arc is in the near infra red, 800–1,000nm. The light is similar to daylight, with a color temperature of around 6,000K. For this reason it is favored for film projection and for some professional video projection. Big xenon lamps are used in theatrical follow-spots and searchlights. Xenon lamps for projection are constructed with a short arc. Exactly as is the case for metal halide lamps used for projection, the shorter the better. In film and video projection the object being illuminated, for example the film frame or a small LCD (Liquid Crystal Display) panel or DMD (Digital Micromirror Device) may only be 25mm (1 in) wide.
250mm
1600w
+ 410mm
4000w
+ Figure 4.16 Xenon arc lamps used for film projection. The 1600W and 4000W lamps (photos from Osram) are air cooled.
140
100
Rel. Spectral ratiation O
UHP lamps with low voltage ignition are distinguished by a coiled wire round the lamp. This acts as both an external electrode to create the secondary UV discharge, and as the electrode to modify the internal field to assist hot re-strike.
75 50 25 0
300
500
700
900 Wavelength O
1100
1300 nm
Relative spectral distribution of radiation of XBO lamps
Figure 4.17 Spectrum of xenon arc lamp.
This requires optical systems with as near a point source of light as possible. Xenon arc lamps operate at a comparatively high pressure. It can be as much as 5MPa (almost 50 times atmospheric pressure) for small lamps, and is still as much as 1MPa for large ones. This means that great care must be exercised in their handling. For example lamphouses for motion picture projection have interlocks to prevent their opening while the lamp is on, and projectionists must wear protective clothing and eyewear when changing lamps. Unlike the other sources reviewed in this chapter, which all operate as AC arcs, the xenon arc lamp uses a DC arc because this gives better arc stability and longer life. For lamps 50–10,000W the arc voltage is in the range 12–60V, and current in the range 4–160A, so high power lamps need large regulated DC supplies. A typical cinema projection lamp, the nominal 2,000W lamp, runs at 27V 70A. Starting, which is almost instantaneous, is by superimposed pulse, and this may be as high as 50kV. Xenon arc lamps are constructed in a wide range of ratings, from below 100W to above 20kW. Motion picture projection lamps are usually 1kW, 1.6kW, 2.5kW, 4kW, or 7kW, depending on screen size. Lamps up to 7kW can be force air cooled, but, strange as it may seem, lamps from 10kW and upwards actually use a combination of forced air cooling and water cooling of the electrodes. Water cooling involves circulating cold water through each electrode. This can only be done if the water is completely pure (distilled) and is unionized. Clearly there would be a problem if the
ARC LAMPS
water itself conducted electricity, which would be the case if tap water was used. The ceramic reflector xenon lamp is a variation where the short arc xenon lamp is housed in an integral reflector housing with a sapphire window. This produces a lamp which itself is a mini-optical system, greatly simplifying its application to video projectors, medical illumination systems etc. The arrangement also allows safe handling without special precautions. The larger lamps of this kind include considerable heat sinking, and are sometimes offered on a factory exchange basis, whereby defunct lamps are returned to the manufacturer to have the arc tube replaced, while re-using the outer assembly. The efficacy of xenon lamps ranges from around 30 lm/W for the smaller sizes, up to 50 lm/W for the big lamps. The efficacy is, therefore, considerably lower than that of the metal halide lamp. Lamp life is typically in the range 500 hours to 2,000 hours. In common with some metal halide lamps, xenon arc lamps can produce ozone. Ultra-violet radiation in the band 180–220nm converts a proportion of the normal oxygen molecule O2 to the alternative molecule ozone, O3. This immediately combines with nitrogen N2 molecules in the atmosphere to produce various nitrous oxides (NOx). While ozone itself is odorless, the NOx impurities are not and have a characteristic smell, often referred to as the “ozone” smell. Besides being unpleasant, NOx and ozone are considered health hazards. Today most projection arc lamps are offered
Rectifier
~
+ Smooth + regulate Ripple <5%
HF filter
Superimposing transformer
~
Pulse generator (timed)
Figure 4.18 Electrical circuit for the xenon arc lamp.
Figure 4.19 Exchangeable xenon arc lamp module used in professional video projectors. Photo from Digital Projection Ltd.
in “ozone free” versions, which means that the quartz used as the envelope filters out the undesirable wavelengths. However, where this is not the case, it is necessary to provide special exhaust air extract arrangements.
4.8 Arc lamp classification Arc lamps are often referred to by a letter code; so it is easy to come across specifications referring to lamps coded as SOX, HIT, SON, HQI, HPI, HSE and so on. The letter codes can be quite confusing and it is not always easy to determine which kind of lamp is being referred to. The underlying problem is that individual manufacturers have their own code for their version of a particular lamp class; but at the same time there are various international classifications which should be recognized by everyone. The sort of confusion that arises is that the same letter may mean something quite different according to the code being used. For example the letter H means “High pressure” in the ZVEI classification, but means “mercury” (from its chemical symbol Hg) in the proprietary Osram classification.
141
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
For those who need to specify lamps a useful publication is “Lamp Designation System – LBS” published by the German Electrical Industries Association – ZVEI, address details are given in the Reading List. This explains the international coding method in detail. What follows here is a summary of the main points. The full description of an arc lamp is given in the form: A B CDE – FG In general the idea is to use as few letters as possible, so often obvious or redundant items are omitted. In the case of Arc lamps A is a single letter denoting the method of light production. In fact only two are of relevance here: H = High pressure L = Low pressure The B character describes the principal or distinguishing filling of the lamp: M = Mercury vapor I = Iodide (as in metal halide) S = Sodium The CDE codes indicate the lamp construction as indicated in Table 4.1. The – FG characters, which come after a hyphen, give further information about the lamp construction and this runs to a very long list.
142
As an example an HST-DE lamp is High pressure Sodium, Tubular – Double Ended lamp. Table 4.2 lists some common fluorescent and HID lamps showing both their LBS classification and some representative manufacturers’ house equivalents. There are two points to notice in this table. In the case of the fluorescent lamps the first two letters (which would be L and M) are omitted as being redundant. The post-hyphen information is more important here. For example in this position L means Long, E indicates the need for an External starter, and I means that the lamp has an internal Igniter. The HSE-X lamp is a high pressure sodium lamp specially designed as an eXchange lamp to take the place of high pressure mercury lamps.
Letter E ET ER G L M PAR R T TC TCA TCG TCR
Meaning Elliptical lamp Elliptical/Tubular lamp Elliptical/Reflector lamp Globe lamp Linear lamp Mushroom lamp Parabolic Aluminized Reflector lamp Reflector lamp Tubular lamp Tubular Compact lamp Tubular Compact All-purpose lamp Tubular Compact Globe lamp Tubular Compact Reflector lamp
Table 4.1 The designation of some standard lamp constructions.
ARC LAMPS
N am e F luo rescent lam p s
S ym b o l
M arking
S LI
O sram
P hilip s
T ung sram
GE
PL-S
T
C o m p act fluo rescent lam p s 2 fingerstarter
TC
LY N X -S
D ulux S
FD
B iax S
2 fingerw ithoutstarter
TC -E
LY N X -SE
D ulux S/E PL-S
FD /E
B iax S/E
4 fingerw ith starter
TC -D
LY N X -D
D ulux D
FD -D
4 fingerw ithoutstarter
TC -D E
LY N X -D E
D ulux D /E PL-C
FD -C /E
6 fingerw ith starter
TC -T
-
D ulux T
PL-T
-
B iax T
6 fingerw ithoutstarter
TC -TE
-
D ulux T/E PL-T
-
B iax T/E
2 finger17,5
TC -L
LY N X -L
D ulux L
PL-L
FD -L
B iax L
2D -lam p
TC -D D
-
-
-
-
2D
Fluorescentlam ps -circular
T-R
LU X LIN E-ES LU M ILU X TLE
-
C ircline
Flat4 finger
TC -F
-
D ulux F
-
-
-
LST
SLP
SO X
SO X
-
SO X
Lam p
HST
SHP-T
N A V -T
SO N -T
TC F
SO N -T
Elliptical
HSE
SHP-E
N A V -E
SO N
TC L
SO N -E
Internalignitor
HSE-I
SHP/I
N A V -E
SO N
-
Luxalox I
D ouble-ended
HSE-D E
-
N A V -TS
-
-
SO N -TD
S ub stitutive hig h p ressure so d ium lam p s
HSE-X
SHX
N A V -E
SO N -H
HIT
HIS-T
HQ I-T
M H N -T (P G 12)
PL-C
D ouble B iax D ouble B iax
Lo w p ressure so d ium lam p Lam p H ig h p ressure so d ium lam p s
SO N C lassique
Elliptical M etal halid e lam p Single-ended
M B I-T -
Single-ended
HIT
HIS-T
HQ I-T
HPI-T
HqM IF
M B IF
D ouble-ended
HIT-D E
HIS-TD
HQ I-TS
M HN -TD ,HPI-TD
HqM IS
M QI
Interalignitor
HIT-I
-
HQ I-T
HPI-B U S
-
M B I-T
Elliptical
HIE
M (E27)
HQ I-E
HPI
H1M IL
M BI
W ith reflector
HIR
-
HQ I-R
-
HqM IR
-
H ig h p ressure m ercury lam p s
HM E
HSL
HQ L
HPL
HqLI
HqLI
Spherical
HM G
-
HQ L-B
HPL-B
-
R eflectorlam p
HM R
HSR -B W
HQ L-R
HPL-R
-
Elliptical
M B FR
Table 4.2 Examples of lamp codings, showing both the recognized marking system and some individual manufacturers’ codes. The table is representative of general lighting lamps – it is not intended as a complete list. Table from the Helvar catalog.
143
Chapter 5
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Special
purpose
lamps
5.1 Induction lamps
™ lamp 5.1.2 GE Lighting Genura™
5.1.1 Introduction
This lamp was introduced in 1994. It is designed as a replacement for the 100W tungsten reflector lamp and requires no external equipment. It gives the same light output as the 100W tungsten lamp, but consumes only 23W, and lasts at least 15,000 hours, compared to 1,000. It is available in two color temperatures, 2,700K and 3,000K. Ra is more than 80. Figure 5.2 shows the construction and shows how closely it follows the principles of Figure 5.1. The main glass bulb has a re-entrant tube in which the ferrite coil sits (with the exhaust tube for evacuating the envelope going back down the middle of the coil). The inside of the glass has a transparent conductive coating to limit electromagnetic interference. The back part of the glass, and the tube surrounding the
The useful life of conventional fluorescent lamps is primarily limited by the life of the cathodes. A secondary issue is phosphor life, which is partly dependent on the presence of impurities – some of which come from the cathodes. It can therefore be surmised that if it was possible to make an electrodeless lamp, the result would be a lamp with exceptionally long life. In a normal fluorescent lamp a longtitudinal electric field between the two electrodes drives the discharge. The induction lamp is a device whereby an electric field is induced into an ionized vapor, without the need for electrodes. A way by which this can be done is illustrated in Figure 5.1. Here a ferrite coil is carrying a high frequency alternating current. Its magnetic field is continually varying; but we know that if a conductor is placed in a varying magnetic field, a voltage is induced in it. The conductor itself then produces an electric field. If the conductor is itself a plasma, then the field is set up within it. Figure 5.1 shows that the electric field is at right angles to the magnetic field. If the field is strong enough and is within the mercury vapor, then the result is the production of UV radiation, just as it is in a fluorescent lamp. Inductive discharges were known about in the 19th century, and an induction lamp was patented as early as 1907. However, the realization of a practical induction lamp had to wait until the arrival of compact power electronics. The design of induction lamps is a juggling act between getting the physics of the discharge right, achieving a good power conversion efficiency, and meeting EMC requirements. Lamp manufacturers have approached the problem in different ways. Three well established designs are now described.
144
Figure 5.1 The principle behind the induction lamp, showing how a changing magnetic field can induce a corresponding electric field.
SPECIAL PURPOSE LAMPS
coil have a reflective titanium dioxide coating, and the entire inside glass has a phosphor coating. The outside of the glass at the back of the lamp has a discontinuous conducting film (to prevent induced currents) which is grounded. It forms a capacitor with the internal conducting film and this helps to short out unwanted RF. One possible problem with this kind of lamp is that its magnetic field can affect the luminaire; if the luminaire then develops its own field, it will be in opposition to the lamp’s field. In the Genura lamp, a single “shorted turn” is placed at the equator of the lamp, which has the effect of screening the luminaire from the lamp’s magnetic field. The effect of the fixed loading afforded by the shorted turn is taken into account in the lamp drive design. The Genura lamp operates at 2.6MHz. It is fortuitous that this frequency is optimal for the lamp design, and is one for which emission standards are relatively relaxed because broadcast reception is not good at this frequency. While it is comparatively easy to understand how the induction lamp works once the plasma is ionConductive transparent coating Phosphor
Ti0 2 Reflector
Kr + Hg Amalgam fill
Shorted turn Electronic assembly
Outer conducting film
Coil mounted in re-entrant part of bulb
Figure 5.2 Construction of the GE Genura™ lamp.
Figure 5.3 The GE Genura™ lamp.
ized; it is not so easy to understand how the lamp “starts”. In Figure 5.1 the coil will itself produce a weak longtitudinal electric field which gets into the discharge space. Conduction can then be by the small parasitic capacitance of the coil. It must be made sufficient to provide initial ionization. Once the discharge has been started, the azimuthal field takes over. In the case of the Genura lamp starting (and restarting) is subjectively instantaneous, but maximum light output, which depends on the lamp reaching the correct operating temperature, is achieved after about 80 seconds. ™ lamp 5.1.3 Philips QL™ Philips also introduced an induction lamp in the early 1990s; but their approach was different. They reasoned that if the electronic drive circuit was kept separate from the lamp, it would last much longer. So while the lamp envelope and drive coil mounting are similar to the Genura; the overall philosophy is different. The QL lamp is supplied as a “system” to luminaire manufacturers. The system consists of the lamp and power coupler as one sub-assembly, and the drive electronics as another. Philips give the luminaire manufacturers guidance on how to make the completed luminaire meet EMC requirements.
145
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
HF Generator
~
Cable <400mm
Reentrant bulb
Power coupler
Figure 5.4 Principle of the Philips QL™ lamp.
QL lamps are available in 55, 85 and 165W ratings, delivering 3,500, 6,000 and 12,000 lumens respectively. They are thus considerably bigger and more efficient than the Genura, but they are only suitable for new installations. The QL lamp is available in three color temperatures, 2,600, 2,900 and 3,850K. Its most impressive attribute is that it comes with a five year guarantee, and with a rated life of 60,000 hours (nearly seven years 24 hours-per-day running.) Induction lamps are good “starters” and can satisfactorily operate at low temperatures. The QL lamp, for example, starts without difficulty at temperatures as low as -20°C. This makes it suitable for inaccessible outdoor applications. ™ lamp 5.1.4 Osram Endura™
discharge vessel is in the form of a closed ring, like a single “shorted turn” transformer winding. Ferrite cores, which themselves have a winding on them, are clamped round the tube to couple the high frequency energy. The principle is shown in Figure 5.5. In the research leading to the Endura lamp, Osram scientists found that the diameter of the arc tube, the discharge pressure and the arc current were all critical factors. A lamp using similar principles had been patented as far back as 1970; but it was comparatively inefficient, suffering serious core losses. Compared to this earlier lamp, the new Osram lamp uses a much lower vapor pressure, and much higher discharge current. The lamp is in the form of a “rectangular” loop, and is at present available in 100W and 150W versions. The characteristics of the larger of two 150W lamps are listed in Table 5.1. Osram state that in principle the shape of the overall lamp is not critical, although the diameter of the tube making up the lamp is optimum at 54mm. A lesser diameter would result in a less efficient lamp. The rectangular shape has been chosen because it is comparatively easy to manufacture, and results in a lamp suitable for “tile like” luminaires. In principle circular or ovoid lamps would work just as well, and smaller lamps could use just one coupling coil instead of two. The low operating frequency of the Endura lamp simplifies the achievement of EMC. At present the induction lamps on the market are T17 Closed loop tube
Yet another approach has been taken by Osram (Osram Sylvania in the USA). They have created a lamp which is most obviously like a transformer. The
Ferrite core clamped round tube
Figure 5.5 Examples of the Philips QL™ lamp.
146
Kr + Hg amalgam filling
Coil around core fed by separate power supply
Figure 5.6 Principle of the Osram Endura™ lamp.
SPECIAL PURPOSE LAMPS
Lamp parameter Tube diameter Overall lamp length Overall lamp breadth Overall lamp height Supply voltage Rating Discharge power (approx) Ferrite core loss (approx) Luminous flux CRI (Ra) CCT options Discharge current Mean discharge length Buffer gas Gas pressure Filling Min ignition temperature Operating frequency Lamp life
(T17) 54mm 414mm 139mm 72mm 110 or 220V 150W 138W 2% 12,000 lm >80 3,000 or 3,500K >7A 720mm Krypton 33Pa Hg amalgam -25°C 250kHz 60,000 hours
Table 5.1 Characteristics of the 150W 400mm Endura™ lamp from Osram.
designed for fixed power output. In principle there is no reason why they could not have variable output where a separate ballast is used, as with both the QL and the Endura, but efficacy would be considerably lower at low light outputs. Both QL and Endura are stated to have useful lives of around 60,000 hours. However, it seems likely that, as experience is gained in their use, and as valid life tests get completed, the practical life of this kind of lamp may be re-rated to around 100,000 hours. 5.1.5 Microwave lamps Another kind of electrodeless lamp uses microwave radiation to stimulate a plasma, resulting in a lamp with very high output from a small volume. Such lamps have been made as UV generators for chemical curing, and also as high brightness visible sources. One kind of microwave lamp uses sulfur vapor as
Figure 5.7 The Osram Endura™ lamp.
the basis of the plasma. Sulfur can only be used in electrodeless lamps, since it would react with any electrode material. Under high pressure (1MPa) all its UV emission is suppressed, and it emits a white light of CCT 6,500K, with very little IR and UV. The only problem is that for the discharge to be stabilized the containing vessel has to be rotated. A practical lamp consists of a microwave generator and a waveguide to lead the radiation into a reflecting cavity. The cavity is tuned so that standing waves are set up within it, and high electromagnetic fields are concentrated at the point where the bulb vessel is sited. The cavity has a front face to let out the visible radiation, but in order to contain the microwave radiation, it is made of a conducting mesh. Magnetron
Conducting transparent window
Waveguide Cavity Rotating bulb containing plasma
Figure 5.8 Principle of the microwave lamp.
147
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
cacy of 120 lm/ per microwave Watt (about 70 lm/ W as a system) and produces 400,000 lumens. Other models with lower outputs have also been developed, although at the time of writing (2002) the commercial future of this technology is uncertain. Japanese manufacturers have announced that they intend to commercialize production of microwave lamps in the medium future.
5.2 Flat lamps
Figure 5.9 The backlighting of LCD screens is based on several different light sources. Most widely used is compact cold cathode fluorescent. But both electroluminescent and LED sources are also used. Now the achievement of very long life and high brightness (> 1,000 cd/m2) displays may be simplified by the introduction of new flat lamp technology.
The microwave lamp has a long life; in practice the life is limited by the life of the magnetron which generates the microwave. For obvious economic reasons, the makers have chosen to use magnetrons operating at 2.45GHz, the same as those used in microwave ovens. The sulfur microwave lamp as exemplified in a lamp by Fusion Systems has a 28cm diameter spherical bulb rotating at 400 rpm. It operates at an effiDischarge Phosphor coating
Top glass
Spacer
The illumination of Liquid Crystal Displays (LCDs) backlit transparencies and similar items demands a system which keeps the overall combination of lightsource and light modulator as thin as possible. Some applications can allow the use of multiple fluorescent tubes (with hot or cold cathode depending on the size of the item) but this usually results in unevenness in the illumination. Many applications of this kind use edge lighting with lightguides (see Section 5.10) to achieve a thin display. Clearly if some kind of “thin flat” lamp could be developed, which had a uniform light output across its surface, with a Lambertian light distribution, results could be improved. Such flat lamps are now being developed. They are based on UV fluorescence, like normal fluorescent lamps, but use only a Xenon discharge without any mercury vapor. Two examples are described here. The Planon™ lamp, developed by Osram, is only 9mm thick, but delivers a luminance of 6,000 Nit. Current sizes are 10 inch and 18 inch diagonal, but the manufacturers indicate that sizes up to 30 inch diagonal (illuminated area 24 × 18 inches, 610mm×457mm) are practical. Dielectric barrier
Xe fill
Anode
Frame Bottom glass
Cathode Anode
Dielectric coating
Reflecting layer
Figure 5.10 Principle of the Osram Planon™ flat lamp.
148
Cathode Shaped discharge Anode
Figure 5.11 Electrode pattern in the Planon™ lamp.
SPECIAL PURPOSE LAMPS
Color temperature is 6,500–7,000K and life is expected to be 100,000 hours. Depending on the particular lamp and control unit, dimming of the lamp is possible in the ranges 20%–100% and 50%–100%. Hitachi Lighting Equipment in Japan have developed flat lamps for the backlighting of small LCD displays, such as those used in camera viewfinders, vehicle instrumentation etc. The lamps are in sizes of 0.5 inch to 5.2 inch diagonal. They use a xenon discharge in an Argon-Neon-Xenon fill mixture, with pulsed excitation. Because the lamps are small, there are only two electrodes, one either side of the lamp. Driving is from high voltage pulses (800–1,200V) with a typical cycle of 1μs on, 60μs off. Figure 5.12 The Osram Planon™ lamp built in to a prototype luminaire.
Pure xenon discharges have great advantages over combined rare gas/mercury vapor. Mercury vapor is toxic, results in a temperature dependent light output and significantly shortens the phosphor life. Xenon on its own eliminates these problems, but the physics of the discharge are complex, especially if reasonable efficiency is required. The lamp consists of two glass substrates, separated by a glass frame and by spacers. The bottom substrate is thick-film printed with a reflecting layer and with parallel lines of anode and cathode electrodes. The electrodes are encased in a dielectric barrier which limits the discharge current. The top substrate is coated with a tri-band phosphor. The regime under which the dielectric barrier discharges operate is covered by Osram’s patent, and achieves a UV conversion efficiency of around 60%. Instead of applying continuous high frequency energy to the electrodes, the Planon system applies discontinuous excitation pulses, the timing of which is directly related to the kinetic properties of the gas. Gas pressure within the Planon lamp is around 13kPa. Figures 5.10 and 5.11 show the principle of the Planon lamp. Because the multiple discharges are shaped, the resulting light is not completely uniform, so a front diffuser is used. The resulting lamp is slightly thicker than its edgelight counterpart, but significantly thinner than a multi lamp arrangement.
5.3 Neon lamps 5.3.1. The neon glow lamp The neon glow lamp is a low efficacy (0.3 lm/W) low power source used mainly for indication. Some larger types have been made for decorative and warning purposes, but most are in the range of less than one tenth of a Watt up to 2W. Their most widespread use is as mains power indicators. They are convenient for this, because their
Figure 5.13 The neon glow lamp used as a mains indicator. Basic lamp (above); built in to panel mounted indicators (below). Photos from MEM 250.
149
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
nature allows them to be connected directly to mains voltages. Thus a glowing neon indicator is a sure sign that dangerous voltages are about! The standard indicator consists of a small glass envelope, similar to a small torch bulb in size and with or without a conventional lamp cap. The envelope has two electrodes, and has a neon gas fill. Below a starting voltage the lamp acts as an insulator, but once at the critical starting voltage, the glow starts. In common with other gas discharge lamps current would rise uncontrollably unless limited, and with glow lamps a resistor of the appropriate value is either fitted externally, or fitted within the lamp base. If a neon indicator was operated on DC the glow would be seen on the negative electrode; with AC both electrodes glow. 5.3.2 Neon tube lamps When people refer to “neon” lighting they do not have the neon indicator in mind. They are much more likely to think of Las Vegas. The neon lamp is associated with signage, advertising and color. While, strictly speaking, the term “neon lamp” should apply only to lamps containing neon gas, the word has Pinch seal electrode
Alternative ring seal electrode
Electrode insulating boot
Coated tube
Plasma Glass joint
Electrode
Figure 5.14 Neon lamp construction. Figure derived from information supplied by Masonlite.
150
become generic for colored and shaped tubular lamps. As pointed out in Section 3.6.5 most lamps referred to as “neon” are in fact cold cathode fluorescent. A more useful demarcation could be by the method of manufacture. Hot cathode fluorescent lamps, and mainstream cold cathode lamps, are only made on fully automated machinery working a continuous production process. However cold cathode lamps intended for signage and similar applications are made in a batch process. Typically: • one process makes the glass tubing; which is cut into standard lengths. • a second process (carried out by another company) cleans the glass and applies any required phosphor coating. A side process prepares the short lengths of tubing that will be fitted to the ends of the tube and which carry the electrodes. • and finally the tube may be shaped, or spliced to another tube, have the electrodes fitted and be evacuated and dosed with the required filling. This will be done at the sign manufacturers’ or specialist lamp fabricators’ premises. In principle only a few companies supply the basic components, but many companies do the final customized assembly. There is now a bewildering choice in the “neon” market, and it requires some knowledge of the processes involved to get the required lamp appearance. There are four main variables: The type of glass used. Most tubes use a clear glass; but for more saturated colors it can be better to use a colored glass tube. The gas fill. Most use argon/mercury vapor; but neon is also used as the basis of red and orange tubes. The gas pressure used can also affect the color. Pressure is in the range 600–1,600Pa. The phosphor coating. Some tubes (e.g. neon in a clear or red glass tube) do not use a phosphor. Most do. Another technique is “double coating”, where instead of using colored glass, the inside of the clear tube is coated with a color pigment, and then by a phosphor. The tube diameter. “Neon” tubes are supplied in a variety of diameters; between different manufacturers it is possible to obtain tube diameters in millimeter increments from 6–25mm. The majority
SPECIAL PURPOSE LAMPS
Clear glass
Colored glass
Clear or colored glass phosphor coat
Clear glass double coat pigment + phosphor
Neon clear, red Argon clear, blue Argon clear coated, blue Argon green glass coated, green Argon yellow glass coated, yellow
depending on the filling and the tube diameter. Narrower tubes have a high volt drop, and neon filled tubes have a higher drop than argon/mercury vapor tubes. Thus voltages significantly higher than “mains” voltage are needed to operate these lamps. Most neon tubes are fed by conventional transformers with outputs from just below 1kV (optimistically referred to as “low voltage!”) for short tubes, up to 10kV or even 15kV for long tubes; but electronic transformers are also used. The operating voltage is limited by regulations in many countries. More information on neon control gear is given in Chapters 6 and 7.
Argon clear coated, green
5.4 Electroluminescent lamps
Argon clear double coated, gold
5.4.1 Introduction
Neon clear coated, red
Figure 5.15 Examples of the colors available for neon lamps. Derived from information supplied by Masonlite.
of work is done using tubes in the 10–20mm range. Narrow tubes are bright but have a short life. In all a neon supply company may well offer a choice of over 50 colors in a wide choice of tube diameters. In the case of white tubes, results similar to those achieved by normal fluorescent lamps can be obtained, since the same tri-phosphor technique can be used. While the prime manufacturer of tube parts may offer the tubes in standard lengths, the final assembler can make up longer tubes if required. Most tubes use a “fold back” arrangement for the electrodes to facilitate uninterrupted runs of light (see Figure 3.25). The light output and to some extent the color of the tube depends on the running current. Each tube diameter has an optimum running current, and typical currents are 25, 30, 50 and 100mA. 100mA is only used on 20 and 25mm tubes. The operating circuit is as Figure 3.25, but tubes are often operated in series. The required voltage will then depend on the overall tube length. The cathode fall for neon tubes is between 100 and 150 volts at each electrode; and the volt drop per meter length of tube varies between 450V and 1,300V
In the discharge lamps already described an electric field causes an intermediate medium (usually a gas) to emit mainly ultra-violet radiation. The ultra violet then stimulates a phosphor to emit visible light. A question that can be asked is “is it possible for an electric field to create light directly?” The answer is “yes”, and the phenomenon is known as electroluminescence. It is manifested in two principal ways. One is in the Light Emitting Diode, to be described in Section 5.5, and the other is where an electric field is applied directly across a phosphor. In this case the field needs to be very strong, which in practical terms means that the phosphor layer must be thin. Most practical electroluminescent lamps are based on the use of AC, and one way of looking at the electroluminescent lamp is to think of it as a capacitor where a phosphor is mixed in to the dielectric, and where one electrode is transparent. 5.4.2 Practical construction The construction principle of the electroluminescent lamp is shown in Figure 5.16. A conducting substrate is covered with a reflective layer. On top of this is the phosphor layer (actually phosphor dispersed in some carrier). The top electrode is a transparent conducting layer, and then there is an outer protective layer.
151
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Transparent protective layer Transparent conducting layer
Light
The principle is realised in two main ways. A rigid lamp is based on: • a steel substrate. • enamel with additive to improve reflection and dielectric constant as the reflective layer. • phosphor in transparent enamel as the light emitting layer. • tin oxide as the transparent conducting layer. • transparent vitreous enamel as the protective layer. This results in a thin rugged lamp suitable for back lighting and similar applications, but somewhat lacking in flexibility in all senses of the word. Today the more common electroluminescent lamp is based on plastic films, and the layers now become: • an aluminum foil or metallized plastic film substrate. • barium titanate in a plastic binder as the reflector layer. • phosphor as the light emitting layer (in the same binder as used in the reflector layer). • transparent plastic foil coated in indium oxide as the transparent conducting layer. Then the whole assembly is sandwiched between layers of transparent plastic film for protection. The plastic electroluminescent lamp is made in sheet or continuous roll form. In sheets it can, within reason, be cut to special shapes. In the roll form it is possible to get continuous lengths of 500ft (152m) or more. Such continuous light sources are used both
Phosphor layer Reflective layer AC Supply Conducting substrate
Figure 5.16 Principle of electroluminescent lamps.
152
Figure 5.17 Electroluminescent sheet, here incorporated into an aircraft cockpit floodlight. Photo courtesy Ultra Electronics.
for special effect, and for the safety marking of exit routes, stair treads etc. Electroluminescent lamps operate in the voltage range 100–200V AC, and at frequencies 50–3,000Hz. In principle the higher the voltage, and the higher the frequency, the greater the light output. Unfortunately, the higher the frequency the shorter the life. The lamp does not usually suffer catastrophic failure, so useful life is defined as the point at which light output declines to a specified level. Apart from the electrical conditions, life is determined by phosphor deterioration in the presence of impurities (especially water vapor) so the manufacturing process is crucial. The color range of practical electroluminescent lamps is limited. The most suitable phosphors are based on zinc sulfides, with different activating elements. The most efficient lamps give a blue-green light; but blue, yellow-green, orange, white and red are also available. Electroluminescent lamps are not high brightness devices, their value is in their thinness and suitability for applications that cannot be met easily by more conventional sources. Electroluminescent sheets, for example, can be less than 1mm thick. Typical initial luminance is 70Cd/m2 for blue-green and 50Cd/m2 for other colors when operated at 115V 400Hz (this is a representative example, not related to any particular manufacturer). Electroluminescent lamps are “cold” lightsources, operating at around 1–2μA per mm2. A typical dissipation is 6mW/cm2 meaning that the lamp temperature never rises more than a few degress above ambient temperature. The operating ambient temperature range can be as wide as -40°C to in excess of 100°C. Some vendors offer equipment that varies the operating frequency in order to achieve a more uniform
SPECIAL PURPOSE LAMPS
Figure 5.19 Electroluminescent wire. Photo from Anytronics Ltd.
Figure 5.18 Extruded electroluminescent lamp available in long runs; suitable for exit route and stair tread marking. Photos from Electroluminex Lighting Corporation.
light output through life. This is done by detecting capacitance changes as the lamp ages, and changing the frequency accordingly. The useful life of electroluminescent lamps varies widely according to design and application, from as little as 1,000 hours to 10,000 hours and more. Electroluminescent lamps can be dimmed, simply by reducing the RMS AC voltage. However, the relationship between voltage and light output is not linear, and zero light output occurs at a relatively high voltage. A practical problem in any system relying on pre-set voltages to achieve pre-set light levels is that as the lamps age, light output declines for a given voltage. 5.4.3 Other constructions The principle of the electroluminescent lamp does allow of other constructions. One such is the electroluminescent wire. This is based on a copper wire with a phosphor coat. Spiral electrode wires are wrapped round the phosphor, and the whole is encased in a tough plastic jacket. The resulting “wire of light” lends itself to decorative effects. Unfortunately the life of the “lamp” is comparatively short,
with a time to half brightness of less than 1,000 hours. The principle of electroluminescence is used in the AC Thin Film ElectroLuminescent display (ACTFEL). In this case the phosphor layer is deposited as a thin film instead of being within a powder. Phosphors and electrodes are arranged in columns and rows to allow pixel addressing, and the resultant device is a rugged high resolution display that competes with LCD (liquid crystal display) for some applications.
5.5 Light emitting diodes (LEDs) 5.5.1 Introduction Section 2.2.5 introduced the LED as a special variant of the p-n junction. Electron-hole recombination can result in the emission of photons, the wavelength Hole and electron combine to give light quantum
Conventional current flow
Electron flow
Figure 5.20 Principle of the LED.
153
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Top contact GaP window AlIn GaP Active layer
+
Substrate Bottom contact
5.5.2 New generation LEDs +
P-GaN Active layers AlGaN and InGaN n-GaN GaN Buffer layer Sapphire substrate
Figure 5.21 High efficiency LED structures. Red AlInGaP made by epitaxial crystal growth (above). Blue/Green InGaN made by chemical vapor deposition (below). The colors in the diagram are used for clarification.
of which depends on the energy gap between the conduction band and the valence band of the electron. λ = hc/Eg where λ is the wavelength, h is Planck’s constant, c is the velocity of light, and Eg is the energy gap. As an example the case of gallium arsenide (GaAs) gives the following. Eg is 1.49eV, where one eV (electron volt) is equal to 1.602 × 10 -19 Joules. h = 6.6262 × 10 -34 Js, and c = 2.99793 × 10 8 m/sec. Putting these values into the equation gives λ as 8.322 × 10 -7 metres, or 832nm, a wavelength in the infra-red. LEDs used in electronics for power indication, simple numerical displays etc were for many years based only on modified GaAs. GaAsP (Gallium ArsenidePhosphide) produces visible wavelengths in the range 590–660nm, depending on the proportion of phosphide to arsenide. When GaPN (GalliumPhosphideNitride) is introduced, a green of 570nm is produced. While the first generation of LEDs were useful
154
products, they were comparatively inefficient, and could not be considered as proper light sources. Their spectral characteristics were also limited, being confined to red and orange, with an inefficient and yellowy green.
The situation changed dramatically when alternative combinations were introduced. In the late 1980s a new generation of high output LEDs began to appear based on AluminumIndiumGalliumPhosphide. They produce light in the range 590–630nm (amber – red). At the same time blue LEDs based on SiC (silicon carbide) appeared. While the red end of the spectrum was now served by a high efficiency device, the green and blue were still deficient. The available green was inefficient and Gold wire bond LED chip Reflector cup
Cathode lead
Plastic lens
Epoxy dome lens
Anode lead
Silicone encapsulent
Cathode lead
InGaN semiconductor flip chip
Gold wire Heatsink slug
Solder connection
Silicon sub-mount chip with ESD protection
Figure 5.22 LED packaging. Traditional “indicator” LEDs simply mount the chip in an epoxy package (above). High light output LEDs have more efficient optics, and include a generous heatsink slug. This in turn transfers heat to the printed circuit on which it is mounted, which itself may have an aluminum core to assist with heat dissipation. Diagram courtesy Lumileds Lighting.
SPECIAL PURPOSE LAMPS
Figure 5.23 Examples of LED lighting “engines”. The “coin” from Osram (left) and a white light unit from Insta (right).
too yellow. The blue was hopelessly inefficient as a lightsource, although suitable for indication. In 1993 Shuji Nakumura at Nichia Chemical announced the development of a blue LED based on GalliumNitride, closely followed by an IndiumGalliumNitride “family” which could cover the wavelength range 450–525nm (blue – green). Now the stage was set for LEDs to take their place as lightsources for many applications. The construction of high efficiency LEDs is complex because they require multiple layers. The AlInGaP family is made using epitaxial crystal growth. An epitaxially grown AlGaInP/GaP layer is sandwiched between p type and n type GaP layers. The InGaN family is made by chemical vapor deposition of the successive layers onto a sapphire substrate. Cree make hybrid devices where GaN or InGaN are built on to an SiC substrate. Each wafer can be the basis of many thousands
Figure 5.24 LEDs mounted in strips and then placed in a plastic tube can emulate neon lighting. Here an example marketed by Lumenyte International Corporation is shown applied to an Applebee’s outlet.
of LEDs, since the individual LED is as small as 0.25mm × 0.25mm. The LED chip is packaged in various different ways according to power and application, Figure 5.22 shows the principle of the commonest forms of packaging. Individual LEDs are low power devices, operating in the voltage range 1.5 to 4V. Earlier generation LEDs operated at around 20mA with a power dissipation of not more than 0.1W; however the newer generation operate at higher currents, for example 100mA, and dissipate between 0.65 and 1.2W. While it is confidently expected that LEDs will, in due course, have a role in general lighting, their present roles are in applications that benefit from their properties of high brightness and long life. LEDs are “candela” as opposed to “lumen” sources. In its basic form the LED is a highly directional source with a peak intensity in the range 100–1,200mcd. An array of LEDs, as used, for example, in outdoor TV screens, can thus achieve a very high brightness in the range 1,000–7,000 Nit, albeit by using hundreds of thousands of LEDs in a single display. Practical LEDs incorporate lensing to spread the light, and are specified in terms of: Peak brightness. Viewing cone angle, defined as twice the angle at which the intensity has dropped to one half the peak value. Dominant wavelength; for example 626nm red, 615nm red-orange, 605nm orange, 590nm amber, 525nm green, 505nm blue-green and 470nm blue. Spectral half bandwidth, defined as the wavelength range over which the output is not less than half the dominant wavelength output. Typically the “red” range of LEDs has a half bandwidth of 15nm, and the “blue” range 30nm. Manufacturers of LED based luminaires are currently a little coy about luminous flux, because there is so little of it! An individual LED has an output around one lumen; or up to 10 lumens or so in high brightness red devices. So the cost per lumen is very high compared with other sources. Efficacy has been low, as little as 0.01 lm/W for early blue devices and 0.6lm/W for early red devices. Current devices intended as light sources achieve efficacies 13–25 lm/ W for red and green, and 3–5 lm/W for blue.
155
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
The manufacturing process of LEDs results in a considerable “spread” of characteristics between individual LEDs. In order to make practical products it is necessary to sort the production in respect of the major characteristics to achieve a uniform result in multi-LED devices. In critical applications like giant video screens the basic sorting alone is not sufficient, and individual LEDs require calibrating in order to ensure uniform change in intensity under varying drive conditions. Some wild claims are made about the life of LEDs. From some product literature you would think it is infinite, but it is not. Some of the LED products are so new that reliable life data is not yet available. Unless there is a packaging failure due to overtemperature operation, the LED does not fail catastrophically like a filament lamp, but degrades over life. Useful life is generally accepted to be the “half life”, when intensity has dropped to half the initial value. While operating temperature does not greatly affect light output, it does affect life. Any practical luminaire must be designed to ensure that LEDs are not operated above recommended ambient temperatures. Present indications are that single color LEDs have useful lives in the range 50,000–100,000 hours. 5.5.3 White LEDs and other developments LEDs are potentially excellent sources of near monochromatic light. For decorative and signage purposes they can be mounted in strips within colored plastic tube to emulate neon lighting, as shown in Figure 5.24. Applications taking advantage of the LEDs long life, such as emergency exit signs, are not usually color critical, so can use monochromatic LEDs or a mixture of two colors. When white or any intermediate color is needed, it is possible to achieve it by 3-color mixing using 470, 525 and 626nm primaries. Figure 5.25 shows examples of 3-color luminaires that can be programmed to give any required color output lying within the color triangle determined by the dominant wavelengths of the actual LEDs used. However, the use of three different kinds of LED
156
Figure 5.25 Color lighting achieved by 3-color mixing in LED luminaires. The Pulsar ChromaBank™ (above) and the Color Kinetics ColorBlast™ (below)
to achieve white is not really an attractive proposition to luminaire manufacturers. This has led to the introduction of the “white” LED, a hybrid device using photoluminescence. The principle is that a blue (around 470nm) LED is coupled to a yellow phosphor to produce a blueish white. Devices using this structure have shown efficacies of 10 lm/W, Ra around 85 and a CCT of between 6,500 and 8,500K. Newer devices may give as much as 18 lm/W or more.
Figure 5.26 Color Kinetics Inc offer the “Board Family” of color mixing LED engines for OEM use. The different shaped engines permit the production of a wide range of luminaires.
SPECIAL PURPOSE LAMPS
Clearly once this kind of performance can be achieved over a range of wavelengths (and always assuming a sensible price) LEDs will take their place alongside the more conventional sources in general lighting applications. LEDs can be dimmed. Varying the drive current directly affects intensity, but it also affects the dominant wavelength which decreases as drive current is increased. In non critical applications drive current variation can be used to achieve dimming down to around 10%, below which differences in individual devices show up. The preferred method of controlling the output of LEDs is by using pulse width modulation. This ensures that the color output remains constant. PWM can dim LEDs down to below 0.05% of maximum output; essential for applications like big video screens, and important for the emerging architectural and entertainment applications of LED. Figure 5.27 The relative intensity vs wavelength of different single color LEDs (above) and that of the white LED (below). From information provided by Lumileds Lighting.
The present problem with the white LED is life. In common with other phosphor based devices, useful life is limited by the phosphor life. In 1999 the Lighting Research Center at Rensselaer Polytechnic University carried out tests on the then available white LEDs, and found that useful life was around 10,000 hours, but this seems to have improved subsequently. Ultimately an alternative white LED architecture, using AlInGaP semiconductor wavelength conversion layers instead of phosphor may yield a longer lived device. LED research has revealed that the potential efficacy of LEDs is much higher than that achieved in practice. Theoretical internal efficacies range from 75 to 500 lm/W, but the problem is getting the radiation out of the device; much of it gets internally reflected and ultimately absorbed. Agilent, a partner with Philips in Lumileds, have raised the extraction efficiency by abandoning the usual parallel sided LED construction and using instead chamfered sides, such that the device is a truncated pyramid. Individual devices using this construction have achieved over 100 lm/W at 610nm, and produced outputs of 60 lm.
Figure 5.28 LED sources are now being widely used for introducing color into architectural lighting. This is the Fruitmarket Gallery in Edinburgh, where the color lighting is provided by Color Kinetics LED luminaires. These are DMX compatible, so a DMX sequencer is used for control. Photo from the lighting designer, Peter Fink, of Art2Architecture.
157
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
5.5.4 LEP and OLED Cathode (calcium)
The semiconductors and phosphors described in this book have, so far, been exclusively inorganic. This means that they are based on elements we instinctively think of as being mineral and unlifelike, like silicon, zinc and sulfur. There are now organic semiconductors, meaning semiconductors based on carbon chemistry, which covers both life forms and materials like plastics. The idea that one could make a light source directly out of plastic is obviously attractive since it holds out the idea of simplicity of manufacture and great flexibility in application. The work going on in the field of organic light emitters is primarily aimed at the display market. Emissive displays of exceptional thinness and reasonable efficiency are promised. Manufacturing costs should ultimately be much lower than the corresponding backlit LCD equivalent. However, there is no reason why the same technology can not be used in bulk as a light source; possibly providing a more versatile and lower cost alternative to inorganic electroluminescent sources. One of the attractive aspects of organic electroluminescence is that most of the production processes can be carried out in solution – much easier than the complex high temperature processes involved with inorganic semiconductors.
Cathode Electron transport layer Light emission layer Hole transport layer Hole injection layer Anode Glass substrate
Figure 5.29 Structure of a molecular OLED. Diagram adapted from information provided by Cambridge Display Technology.
158
Light emitting layer PPV semiconducting polymer Conducting polymer layer PEDOT + PSS Anode IT0
Glass substrate
Figure 5.30 Structure of a LEP device. The calcium cathode injects electrons into the PPV semiconductor film. The anode, which uses ITO as the electrical connection and a conducting polymer to “match” the energy characteristics of the PPV, injects holes. Electron and hole capture in the PPV results in the formation of neutral excitons, bound excited states that decay by emitting a photon. PPV = poly(phenylene vinyline) ITO = Indium Tin Oxide PEDOT = poly(ethylenedioxy)thiophene PSS = poly (styrene sulphonic acid) Explanation and diagram based on information provided by Cambridge Display Technology.
Although the basic idea behind both is the same, organic light emitters are currently classified in two main types, the OLED (Organic Light Emitting Diode) and the LEP (Light Emitting Polymer). The chemical difference is that OLEDs are based on comparatively “small” molecules, whereas LEPs are based on polymer molecules with very large molecular weights (300,000 and more). While research and development of these devices is now worldwide, early and leading exponents of OLED were and are the Eastman Kodak Company; and of LEP the University of Cambridge, UK, and the spin-off company Cambridge Display Technology. Figures 5.29 and 5.30 show the construction of OLED and LEP devices. At present these are made onto glass substrates, because this material keeps out oxygen and water vapor, either of which destroy the device. It would clearly be much better if the devices could be made on a “plastic” substrate, but unfortunately presently available polymers are unsuitable.
25
10000
20
1000
15
100
10
10
5
1
0
1
2 3 Voltage
4
5
Brightness Nit log scale
Effifacy Lm/w
SPECIAL PURPOSE LAMPS
0.1
Figure 5.31 Perfomance of a green LEP device. This shows that an efficacy of 20 lm/W can be achieved for an output of 100 nit. Higher brightness (1,000 nit) is achieved at the expense of efficacy (14.5 lm/W). Diagram from Cambridge Display Technology.
In order to show the different layers, the diagrams may give the impression that the devices are quite thick. They are, in fact, extremely thin. In a polymer device the light emitting layer is only 100nm thick. Figure 5.31 shows the performance of a green LEP device. The achieved 20 lm/W with promise of more to come certainly makes the device worthy of consideration as a light source. The problems that remain to be overcome are: • life. Lifetimes of more than 10,000 hours are already being achieved, but not on all colors. • color. For display applications satisfactory red, green and blue devices have been demonstrated or are in production. It is not yet clear whether an efficient white light device could be developed.
5.6 Lasers The laser (acronym for Light Amplification by Stimulated Emission of Radiation) is a device which produces an intense monochromatic beam of coherent light. Most light sources produce light of multiple wavelengths, in multiple bursts of radiation that are not in phase with each other and that are multidirectional if not omnidirectional. The laser in its purest form produces monochromatic (single wavelength, single color) light that is coherent, that is to say that all the light produced is in phase, or “in step”.
This results in an intense narrow beam of light which does not diverge. In high power lasers the energy carried by the laser beam is so great that it can be used for the precision cutting of sheet metal. Laser action arises when a large number of atoms or molecules are put into an “excited” state. When an “excited” atom drops from one energy state to another it emits a photon or burst of light energy. In a normal discharge tube like an arc lamp the process is random and the light incoherent. In a laser tube the energy is so great that a very large proportion of the molecules are in the excited state, and when one decays it can stimulate neighboring ones to do the same. This simultaneous lemming-like drop in energy results in a burst of coherent radiation. The laser differs from the normal discharge lamp in that the transition between energy levels is stimulated by radiation of the same wavelength as the emitted light. Some lasers are pulsed lasers, where the energy needed to get the high proportion of molecules into the high energy state is provided by, for example, a flash tube. Display lasers (those used for lighting effects in entertainment and presentation applications) need a continuous form of energy to ensure that they can give a continuous output. Lasers are constructed in many ways. For display work inert gas lasers are used with optical power outputs of typically 5–15W. They consist of a long ceramic tube which forms a resonant cavity for light. At either end there is a highly polished mirror, one of which is only partially coated, and it is from this end that the laser beam emerges. A DC arc is set up in the tube using a regulated power supply (typically taking 30–60Amps per phase from a 3-phase 400Volt supply). The plasma discharge is of sufficient energy to maintain continuous laser action. A continuous water flow of around 13 l/m is required to cool the laser tube. In terms of turning incoming electrical power into visible light the whole contraption is only about .005% efficient. A practical gas laser for display includes magnetic stabilization of the plasma discharge and a means of maintaining the correct gas fill pressure. It also includes a method for ensuring correct mirror alignment. The gas mixtures in the laser are chosen to
159
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Brass outer jacket contains magnet
Helical water flow diverter Ceramic plasma tube Brewster window
Mirror plate mounted on invar rods fixed to the outer jacket Tungsten discs contain plasma
Tube support disc transfers heat to tube wall
Figure 5.32 Spectra-Physics Inc’s Chroma® ion laser used for laser lightshows. Available in 5, 10, 12 and 15W versions.
ensure that there is a means of getting the lasing ions to the right energy in the first place. For example in the Helium-Neon laser (in practice realized as a low power device using RF as the exciter) helium is the energizing gas. One of its energy levels is one where it can only lose energy by collision (loss of kinetic energy.) However, this same level is one at which neon emits infra-red light. So in the He-Ne laser the helium is energized, transfers its energy by collision with neon; whereupon the excited neon emits photons. The photons bounce back and forth between the end mirrors stimulating any newly energized neon to create more photons. An equilibrium is reached when the radiation loss through internal absorption, and from the exit beam from the partially reflecting mirror, matches the gain from the lasing process. Laser technology has now advanced to the point that a single laser can produce multiple wavelengths, and it is this type of laser that is used for display. The beam emerging from the laser is less than 2mm diameter, and may have a divergence of less than .00065 radians. It emerges as polarized light from a Brewster window , constructed from crystalline quartz. A laser display system consists of one or more ion lasers (argon, krypton or mixed) mounted on an optical bench followed by an array of optical ele-
160
ments to manipulate the beam. These include: Dichroic color separators, to separate the colors from a multiple wavelength laser. Shutters to “switch” the beam on and off. The most sophisticated of these are acoustic modulators
Figure 5.33 Coherent Inc’s Star II entertainment laser includes a motorized mirror system for automatically maintaining laser alignment. The mixed gas multi-line PL model produces a total of 3.5W of different wavelengths: 647nm Red 1.1W 568nm Yellow 0.05W 521–515 nm Green 0.8W 488–477nm Blue 0.8W 458nm Deep Blue 0.045W
SPECIAL PURPOSE LAMPS
Figure 5.34 Optical bench for the direction and control of laser beams for lightshows. Photo from Lobo Laser and Multimedia Systems, Aalen, Germany.
that can work at very high (video image) frequencies. An acoustic wave in a crystal can diffract light, and in a modulating system, the straight through light is either absorbed or used for another purpose, and the diffracted light (as much as 80% of the available light is diverted to the first order diffraction angle) is the light that is used in the display. Static beam diverter mirrors. Dynamic beam diverter mirrors. These are usually based on the principle of the mirror galvanometer. The mirror needed to divert a laser beam can be very small, so it is practical to support such a mirror on a thin torsion wire in a magnetic field. Varying the current though the wire varies the angle of the mirror. Light scattering devices. In rare cases a laser is used to produce a fixed light pattern. More usually it produces dynamic light patterns by the programmed operation of the electrooptical components. The programming is done by a computer with suitable external interfaces to the controlled devices. Because of its size, power requirements and need of water cooling, the big gas display laser can be difficult to site within a lighting scheme. For this reason the laser is sometimes sited remotely, with the light being taken to the optical elements by optical fiber. Gas lasers are big and fragile, and solid state lasers would obviously be more convenient. The idea of a solid laser is not new; the first lasers were based
on ruby. A ruby rod with highly polished ends, one with a full reflecting surface and the other with a partial reflecting surface is the basis of a pulsed laser. The rod is surrounded by a helical xenon flash tube which provides the pumping energy. Of more practical use is the semiconductor laser. This can be achieved by extending the principle of the LED. If a p-n junction is arranged such that the semiconductor is polished with two parallel faces at right angles to the junction plane, light created at the junction can reflect back and forth. Further, if the current through the junction is sufficient, the requisite population inversion can occur – this is the condition needed for laser action whereby there are more electrons at conduction band energy than there are at valence band. Small solid state lasers, laser diodes, have been available for many years, most operating in the infra-red. They form the basis of many industrial and consumer products (including laser printers and CD players.) Visible light laser diodes with sufficient output to be used as a light source are a more recent development. They are already being considered as the light source for some types of video and graphics image projection, either as a simple source of lumens, or the basis of a scanning beam system. In such cases
Figure 5.35 The Viper® Green (532nm) laser from Coherent Inc uses remotely mounted infa-red laser diodes. Their output is fed by optical fiber to pump the green solid state laser head. Outputs of 5–10W are available, and the system only needs around 600W of single phase power. Air cooling is sufficient for all but the 10W model which uses closed circuit water cooling.
161
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
used. These have powers in the range 400W–5kW. • for some medical and solar simulation applications hybrid lamps are available. The high power metal halide lamps are tailored to specific applications. Just as in metal halide lamps intended for the visible spectrum, where particular metals give a required color, other metals enhance particular UV wavelengths. For example UV polymerization of plastics, and photo-resist applications like the manufacture of printed circuits need around 350nm, achieved with elements iron and cobalt. Figure 5.36 Laser diode. This one is a 5mW device from Nichia Chemical, it operates at 405nm (violet).
efficacy around 10 lm/W is expected. In due course they may also be suitable as sources for special luminaires, especially those based on fiber optics. Solid state lasers can be coupled to “frequency changing” devices. These are non linear crystals that have the property of doubling the input frequency. This means that efficient high power infra red lasers can have the output converted to a visible wavelength, and it is likely that this type of laser will displace the “traditional” ion laser in the entertainment lighting field.
5.7 Ultra-violet lamps Ultra-violet lamps are special versions of the fluorescent and high intensity discharge lamps already described. As already pointed out in Section 3.6.2, UV is classified in wavelength bands with the shorter wavelengths being more dangerous. As a summary: • low power UV lamps, used for sun-tanning, fluorescence inspection, stage effects, insect attraction etc are based on normal fluorescent lamp construction with ratings in the range 7W–125W. Versions are also available for UV-C, but these should only be used in their proper enclosure for a specific application (EPROM erasure, sterilisation etc.). • when something more like a point source is needed, black glass mercury vapor lamps are used (Section 4.2). Typical ratings are 125W and 250W. • when higher powers and specific wavelengths are required, special versions of metal halide lamps are
162
Figure 5.37 A UV luminaire from Wildfire Inc of Los Angeles, suitable for entertainment and architectural applications. It uses a 400W metal halide lamp. The same company offers a range of fluorescent paints, the ones in the photo are visible under normal light and fluoresce under UV. There are other paints that are clear or “invisible” under normal light, but that fluoresce under UV.
SPECIAL PURPOSE LAMPS
But much better effects (and the almost complete concealment of any kind of visible source) can be achieved by using luminaires which can direct the UV in much the same way as a theater spotlight or floodlight. This avoids the unintentional illumination of other objects or people which could give rise to unwanted or inappropriate fluorescence. This type of luminaire uses either the black-glass mercury vapor lamp, or a 400W metal halide lamp and filter optimized for UV-A. The reflector and lens must also be suitable for UV. Control of the UV light is by shutter. Specialist companies supply a wide range of “invisible” fluorescent paints, which are neutral or white under visible radiation, but are brightly colored under UV.
5.8 Infra-red lamps
Figure 5.38 Example of the use of UV light in scene painting. “Daylight” scene conventionally lit (top). “Night time” scene revealed by UV lighting (below). Illustration from UV Light Technology Ltd. (UK) courtesy of the artist Jim Harper.
Metal halide lamps are also used in medical applications of UV, for curing skin conditions. The hybrid lamps, which combine incandescent and discharge sources, radiate in the range 300– 2,000nm. They are used both as “health” lamps, where the combination of UV and IR is considered beneficial, and as radiation sources for testing materials and equipment in respect of sunlight. As an example Osram suggest that an array of 16 of their 300W hybrid lamps arranged to cover an area of one square meter, and sited 50cm from it, simulates sunlight from a midsummer noon sun with an irradiance of 1kW/m2. It is clearly important that in any weathering test the radiation used has the same proportion of UV in it as experienced from real sunlight. Ultra-violet radiation can be refracted and reflected just like visible radiation. In theater and entertainment applications UV is often used as a “wash” light using simple fluourescent batten type fittings.
Infra-red radiation is the band of radiation between (and overlapping with) visible radiation and the very highest radio frequencies. Wavelengths range from 600nm to 1mm. The filament of a tungsten or tungsten halogen lamp behaves in a similar manner to a black body, and if heated to above 2,000K has peak radiation in the area of 1,000nm (see Figures 1.51 and 1.52). We can feel heat in various ways. If we are in still warm air, then we feel warmth by conduction as the air molecules conduct their movement to our skin. Conduction also applies if we touch a warm pipe or panel. But we also feel heat by radiation when we stand in front of a fire, or receive any source of IR. In this case our bodies absorb the radiation and the process creates heat – i.e greater molecular movement. Short wavelength IR is not absorbed by air over short distances. It also behaves like light in that it can be reflected and focused like light when non IR absorbent reflectors are used. It is, therefore, a useful source of directed heat. IR lamps are simply special purpose tungsten and tungsten halogen lamps. Some of them have considerable visible radiation, since their application benefits from its presence – for example “health” lamps and lamps used in animal husbandry. Others have
163
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
filters to eliminate visible radiation, for example those used for night security surveillance by closed circuit TV by IR sensitive cameras. The attraction of IR as a source of heat is that it can provide localized heat instantaneously and locally. The applications already referred to use reflector type lamps (see also Figure 3.4). Those used for high power heating are special tungsten halogen lamps in quartz envelopes. Examples of use are: • as radiant heaters in entrances to shops, restaurants etc. in winter. Whereas warm air heaters simply lose the warm air to the cold outside, radiant IR heaters give heat that can be immediately felt. (But best to get them on before any ice forms, since the IR can be reflected by white ice!) • as heaters for many industrial processes, for example paint finishing and plastic forming.
Straight tube 20W for warning lights
2400 Joules photo flash U-Tube
Helical strobe tube 20W
Airport runway flash lamp 60 Joule Helical tube in PAR 56 envelope
Figure 5.39 Examples of flash tube construction. Based on information from Amglo Kemlite Laboratories Inc.
164
• in food processing and cooking. Modern hotplates use special IR lamps as the source. Since most IR lamps are simply special versions of tungsten filament lamps, thay can easily be controlled using the same methods. Such control will normally be for functional reasons, but it can also be for special effect.
5.9 Flash tubes Flash tubes are used for a wide range of applications, for example: • photography • industrial stroboscopes for examining rotating machinery etc • aircraft warning lamps • obstruction warning lamps • laser pumping • entertainment stroboscopes They produce high intensity pulses of short duration, and are made in a wide range of sizes. They are based on long arc discharges in xenon. A flash tube consists of a borosilicate glass or quartz tube which, depending on the application, may be straight, U-shaped or helical. The tube is filled with xenon gas at a moderate pressure of around 60kPa. The main electrodes are at the extreme ends of the tube. Flashtubes are designed to operate at various voltages depending on length, from 250V to 3,500V or more. However, xenon gas is normally an insulator; to become a conductor it must be ionized. While some specialist tubes use a superimposed high voltage trigger pulse across the main electrodes, most flash tubes use a separate trigger electrode. This is in the form of a wire wrapped round the exterior of the tube. Small tubes use a conductive coating. When a high voltage (several kV) is applied to this electrode it ionizes the xenon which then conducts. Continuous conduction, if it was possible, would blow up the tube, and anyway would not produce the desired short pulse of light. The principle of the flashtube is shown in Figure 5.40. The voltage across the tube is derived from a capacitor or bank of capacitors. The capacitor must be suitable for discharging a high current in to a low impedance load. The
SPECIAL PURPOSE LAMPS
Flash tube
Trigger circuit
a.c in D.C rectifier
Storage capacitor
Trigger transformer
Figure 5.40 Principle of the electronic flash tube. The trigger coil ionizes the xenon gas in the tube. It then becomes conductive and discharges the energy stored in the capacitor. When the current drops below a threshold, conduction ceases, the gas reverts to being an insulator, and the capacitor re-charges.
trigger circuit is an oscillator producing an AC voltage impulse of 150–400V. This is fed through a stepup transformer to produce the high voltage required to ionize the xenon (4–20kV). The trigger circuit is arranged to operate either on a “one shot” basis, to give a single flash, or on a continuous basis, to give a series of flashes at a precisely defined interval. In photographic applications the emphasis is on producing infrequent individual flashes of maximum energy. In stroboscopic applications the requirement is for rapid repetitive flashes at frequencies from 1Hz to 500Hz. Specially designed tubes can operate at frequencies above 1kHz. A flash tube is rated according to average power input, that is to say the product of the energy per flash and the number of flashes per second. Ultimately the limit is set by the power that can be absorbed by the envelope, which is around 5W/cm2 in free air, with considerably higher ratings for forced air or liq-
Tube type Color temperature Life Flash duration Flash duration Energy per flash Tube rating average
Xenon flash 6,500K 100 million flashes 20001050ms slow speed 100010500Ps high speed typical 180mJ @ 20ms 15W
Table 5.2 Outline specification of a xenon flash tube used in portable stroboscopes.
Figure 5.41 A portable stroboscope for industrial applications. This one has a range of 30–14,000 flashes per minute (0.5–233 Hz). Photo of the Monarch Nova Strobe from Monarch Instrument.
uid cooled tubes. The life of a flash tube is determined by the number of flashes it gives in relation to the maximum flash energy (defined as the explosion energy which bursts the tube within a few flashes). A photographic flash tube might be designed to carry out a few thousand or tens of thousands of flashes, in which case it could be run at around 30% of maximum. A stroboscope tube, intended for millions of flashes, would only be run at 5% maximum. Electronic flash units for photography usually have their flash energy specified in watt-seconds; for example: • flash attached to a camera 100Ws. • self contained studio flash 250Ws. • big studio flash with separate power supply 6,000Ws.
Figure 5.42 Examples of entertainment strobes from Pulsar Light of Cambridge Ltd. The 20W “Jumbo” strobe uses a comparatively small U tube. The 1,500W “Demon” uses a linear tube.
165
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Flash tube nominal W
20 60 180 1,500
Flash Power at 16 flash/sec J
1.25 3.75 11.25 30
Flash Power at 4 flash/sec J
5 15 45 375
Table 5.3 Examples of flash tubes used in entertainment strobes. Derived from the Pulsar Light of Cambridge catalog.
Photographic flash tubes are not intended for continuous operation, and have a minimum time between flashes of 1–15 seconds. Of course the Watt-second is the same as the Joule, and industrial flash tubes are more usually quoted in J or mJ. Stroboscopes are devices for giving flashes of light at precisely known flash frequencies; they are used for observing moving and rotating machinery. When the machine or rotating device appears stationery, the flash frequency is the same as (or a sub-multiple of) the reciprocating or rotation frequency of the device under observation. Examples of flash tube specifications for portable industrial stroboscopes are given in Table 5.2. The flash duration is limited by the impedance of the flash tube, the capacitor value and the presence of any inductance in the circuit. Flash tubes used for aircraft and airport runway applications are rated for low frequency continuous operation. The frequency range is typically from one flash every 10 seconds to a maximum of two flashes per second. For example a 120W warning flash mounted in a PAR56 envelope might be rated for two 60J flashes/second with a life of more than seven million flashes. A different kind of stroboscope is the entertainment stroboscope. This is a low cost device used in discotheques and stage productions. Here the aim is to get the brightest possible flash for the minimum outlay. Flash frequency is limited to the range 0.5– 16Hz, or the stroboscope can be externally triggered. This allows the flashes to synchronize to the beat of music, for example. The flash power can be quite significant when a large number of high power entertainment strobes are used together. Table 5.3 gives some examples.
166
5.10 Fiber optics and lightguides 5.10.1 Fiber optics Strictly speaking fiber optics and lightguides are not lightsources at all, but for many applications they have become the means of delivering light to where it is wanted; so an understanding of how they work is useful if one is to understand how to control their light output. The principle of the optical fiber is shown in Figure 5.43. It depends on the total internal reflection of light arising from using two materials of differing refractive index. A single optical fiber consists of a core surrounded by a cladding of a lower refractive index. The concentric fiber may then itself be placed in a protective outer sheath, or many fibers may share a common sheath. For light to get in to the fiber in the first place, it must be directed at one end. The fiber has an acceptance angle which is determined by the refractive index of the core and the cladding. Light incident outside the acceptance angle may either be reflected off the entry face, or fail to be totally internally reflected inside the core. In Figure 5.43 it can be seen from the internal reflection mechanism that the acceptance angle also determines the angle at which the light comes out the other end of the fiber. In principle optical fiber can be made from any transparent material, for example glass, plastic or even liquid. In practice optical fibers used for lightLight acceptance cone
Sheath (optional)
Cladding
Core
Figure 5.43 The principle of fiber optics.
SPECIAL PURPOSE LAMPS
ing are made either from glass or from plastic polymers such as PMMA (poly-methyl methacrylate). For lighting purposes, fiber is configured in one of four principal ways: • as a single large core plastic fiber. • as a bundle of large core plastic fibers. • as a bundle of small core plastic fibers. • as a bundle of glass fibers. The fibers within a glass bundle are thin, from as little as 50μm to 1mm. Plastic fibers are thicker; the large core fibers can be as much as 18mm diameter. Glass fiber optic components are normally fully factory made in the form of “harnesses”. This is necessary because the ends of the fiber need finishing and polishing. The result is a precision, long life but somewhat expensive product. Plastic fiber can be finished “in the field”, but for some applications its optical performance is not as good. Plastic fiber can discolor, and lasts about 10–15 years. It is not suitable for outdoor use. The performance of fiber optics as a light delivery method is measured by a number of factors, for example: Attenuation over a given length. This can either be expressed as a simple percentage, or as a relative power figure, such as dB/m. It is significant, even the best fiber optic illumination system will lose around 20% of the light over 5m.
Figure 5.44 Examples of the performance of fiber optics. Top, light transmission of Schott Spectraflex® glass fiber harnesses. Middle, the variation in optical attenuation with wavelength of single glass fibers (also from Schott). Bottom, variation with wavelength in a large core plastic fiber (derived from papers given by the LRC at Rensselaer Polytechnic Institute).
Figure 5.45 Fiber optic “harness” from Schott.
167
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 5.46 Examples of illuminators. The SpectraStar uses a 150W metal halide lamp (Left) while the 100SH uses a halogen lamp (50–100W). Both illuminators from Schott.
Spectral response. The transmission of different wavelengths is not uniform (see Figure 5.43). If “white” light is put through a fiber, it may appear “colored”, for example a greenish tinge, at the output, especially if a long length is involved. Bending capability. Thin fibers are best for a short bending radius. Uniformity. Light attenuation is caused by both absorption and by light scattering (instead of the desired reflection.) Scattering is worse on bends. In a multiple fiber system not all the entry light gets into the fibers since, if they are circular, it is clear there must be gaps between them. This loss is referred to as packing fraction loss. In a glass fiber bundle there are various ways of organizing the fibers. For optical instruments they can be coherent, such that they can be used for imaging. For lighting they can be randomized or unrandomized. This is significant when one illuminator illuminates a big bundle of fibers that is then split into several smaller bundles, as happens with pre-prepared harnesses. If the fibers are unrandomized, and if the end illumination is uneven, one bundle may receive more light than another. A randomized arrangement means that each small bundle has its fibers distributed randomly in the big bundle – ensuring that all the small bundles receive the same amount of light. A fiber optic illumination system thus consists of three principal components. The fiber, fiber bundle or fiber harness itself, suitably finished at both entry and exit, an illuminator (that may be designed to
168
illuminate more than one bundle) and, in many cases, exit optics. In addition to the basic components, there may be accessory components, for example a color wheel. The aim of a fiber optic illuminator is to collect as much light as possible from a conventional lamp, and direct it on to the entry end of the fiber or fiber bundle, in such a way that the light is all within the acceptance angle. Studies by the Lighting Research Center at the Rensselaer Polytechnic Institute, Troy, NY have found enormous differences in the effectiveness of different illuminators. One illuminator can get twice as much light into a fiber harness than another illuminator using the same lamp. Their work has shown that effective illuminators must be designed with low numerical aperture or NA. NA = n sin α where n is the refractive index of the material (in this case the core) and α is the most oblique angle of incidence used. For plastic fiber, in particular, low NA is essential for illuminating long fibers. On fiber lengths over 10m there can not only be an improvement of around 55% in overall light transmission, there is also an improvement in color performance. The transmission of different wavelengths in plastic is markedly different for different values of α. The most commonly used sources are tungsten halogen lamps and metal halide lamps. Tungsten
Figure 5.47 Examples of end optics. Adjustable and fixed lens fittings from Schott (above). Waterproof and bollard fittings from Philips (below).
SPECIAL PURPOSE LAMPS
Figure 5.48 Leaky, or side emitting, fiber optics make a good substitute for neon. These examples are from Lumenyte International Corporation. Top photo is of the Franklin County Courthouse, Columbus OH. Bottom photo of a pedestrian bridge in Santa Rosa, CA. Lighting design by Michael Hayden of Thinking Lighting.
halogen has the merit of simplicity, and is suitable for conventional dimming. Metal halide gives higher efficacy, longer life and higher color temperature (usually around 4,200K for this application). Suitable ballasts allow only limited dimming. Most illuminators have fan cooling, so noise can be an issue. Exit optics exist in many forms. In a plastic fiber based “star cloth”, the individual fiber ends are viewed directly. In sophisticated systems for display cases, the exit end of the fiber bundle is coupled to a miniature “luminaire” which itself can provide the normal functions of spotlighting or floodlighting, often by the use of additional lensing. A special form of fiber optics is the “leaky fiber”. Such fibers allow deliberate leakage of light by scat-
tering down their length. They are used as comparatively low brightness substitutes for neon and cold cathode tubing. The system efficacy of fiber optic illumination is low. When all losses are taken into account a tungsten halogen system might give 3 lm/W and a metal halide system 12 lm/W. However, for some applications fiber optics have considerable advantages. • ideal for wet and hazardous locations. The source can be safely mounted in its own suitable environment. The light has no electrical connections whatever. • the exit light is free of IR and UV (but for some fibers filters are needed at the entry to avoid damage to the fiber itself). • easy maintenance for multiple light points. Instead of having to change 50 or 1,000 lamps in a showcase or a starfield, only one conveniently placed lamp needs replacing. • esthetic advantages for some applications, arising from the absence of bulky luminaires. Coupled with this is the ability to get a source of illumination into a position that would be quite impractical with a conventional luminaire. Applications range from entertainment and display, through signage, swimming pools and fountains, to lighting for museums, galleries and showcases. Industrial and professional applications include those in medicine, dentistry, clean rooms, and manufacturing processes. 5.10.2 Light guides There are many lighting systems and luminaires that use a combination of reflection and refraction to achieve a particular distribution of light, and the general subject is outside the scope of this book. However, it is appropriate to include one other device that is a near relative of the optical fiber – but this time with significantly greater dimensions. The tubular light guide or light pipe (promoted in particular by 3M and by TIR Systems Ltd) is another device that works by internal reflection with controlled “leaking” of light. The principle is shown in Figure 5.49. The guide is an acrylic or polycarbonate tube of around 150mm diameter. Light
169
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Cross section
Endfeed light guide
Acrylic housing
Extractor
Input endcap
Extractor Reflected light ray
Mirror endcap
Emitted light ray 0
X Clear prismatic film
Emiting sector
Figure 5.49 Examples of 4in (100mm) and 6in (150mm) diameter tubular lightguides from TIR Systems Ltd. The bigger guide can be up to 40ft (13m) long with single end illumination, and with a 400W HID lamp gives 310 lumens/ft at this length.
Figure 5.50 TIR Ltd use LEDs as a source of illumination in their color changing light pipes.
170
enters one end of the guide, and is normally internally reflected down its length. The far end is mirrored, so light is reflected back again (although for long lengths it is possible to have a lamp at both ends). The inside of the tube is lined with a clear optical film that has a microprismatic surface. This ensures a uniform and efficient spread of light down the tube. Part of the interior of the tube is fitted with an “extractor” film. This provides a diffuse reflecting surface. Part of the reflected light from the extractor is incident on the tube wall at an angle that allows it to escape. The combination of prismatic film and extractor film can be tailored to ensure that most light is emitted from a defined angle. A light pipe of this kind, illuminated from one end only, can give even lighting over a length of over 13m. Cylindrical light guides or pipes generally use metal halide lamps as the original source of lumens. However, they also use less conventional sources. 3M have used the Sulfur Lamp (see Section 5.1.5) as a source, because it makes good optical fit with the pipe, and has excellent spectral characteristics. For narrower diameter pipes used for interior decorative installations, LEDs can form the basis of color changing pipes. The use of special optical films for light control is not limited to cylindrical light guides. Rectangular guides are forming the basis of exceptionally thin
SPECIAL PURPOSE LAMPS
illuminated advertising signs. Instead of using multiple fluorescent tubes across the back of a color transparency, tubes are sited at the edge only, and light is distributed to the back of the transparency by optical film.
5.11 Video displays as lightsources It is easy to overlook the fact that there are sources of electrically generated light other than “lamps” as generally understood. For example, the cathode ray tube (CRT) has been a significant source of light in
the home for 60 years; although that is not its intended role. The surface brightness of CRTs, and now other emissive display devices like plasma display panels, is significant. In the appropriate surroundings they can form part of a lighting scheme, whether they are showing normal video images or are showing color washes for effect. Clearly this idea has only been used in the entertainment environment, but there is no reason why adventurous lighting designers should not create electronically shaped color. Work of this kind is likely to be based on massed LED, but the older technologies should not be forgotten.
Figure 5.51 CRT Video screens as a main source of light. Photo of the “Point After” dance club on the Carnival Destiny cruise liner from Wynne Wilson Gottelier.
171
Chapter 6
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Electromagnetic
components
Part 3 – Lighting components Today most lamps require some kind of “load interface” between themselves and the electricity supply – for example discharge lamps need a ballast, and low voltage halogen lamps need a transformer. The next two chapters look at the common interfaces, Chapter 6 deals with electromagnetic components, and Chapter 7 deals with electronic components. While electronic components are taking an increasing proportion of the market, electromagnetic components still have an important role to play.
6.1 Principles of transformers and inductors. 6.1.1 Introduction In this Chapter we are concerned with electromagnetic components that themselves are in the lamp circuit, and that operate at line voltage and frequency. In practice this means transformers and ballasts. The word ballast is the generic word for any device that limits the current through a discharge tube. The simplest ballasts are inductive reactors or chokes. More complex electromagnetic ballasts combine transformer and choke action. A complete ballast circuit usually includes a capacitor for power factor correction, but sometimes the capacitor is also used as a current limiting reactance. The main design aim with this kind of component is to minimize losses. In lighting components losses are not only wasteful, they can actually be dangerous by producing local hotspots which can result in fire. Legislation in both Europe and the USA is now requiring higher efficiencies in all kinds of lighting equipment, and this is helping to ensure a higher standard of component. (See also Sections 1.2.6, 1.2.7, 1.3.4 and 1.4.1 for definitions etc. that are assumed in this chapter).
172
6.1.2 Transformer action A simplifed explanation of what happens in a transformer starts by looking at Figure 6.1 This shows two coils carrying current, coupled together. Coil 1 we can refer to as a primary coil, and Coil 2 as a secondary coil. In the description that follows, the symbols have the meanings ascribed to them in the figure. The voltage induced in a coil is proportional to the number of turns in the coil and to the rate of change of flux through it. In the SI system of units: e = −N
dφ dt
This tells us two things about transformers. First, that the voltages are proportional to the turns ratio; since, for the mutual flux that is shared between two coils, dφ/dt is the same for primary and secondary. Thus it is the case that: e1 N = 1 e2 N2
Second, that a high voltage can be achieved either by lots of turns, or by a faster rate of change of flux. This can be achieved by working at a higher frequency. Transformers are, in practice, remarkably efficient, so it is valid to start by thinking of a perfect transformer. This is one that has no leakage flux, so
ELECTROMAGNETIC COMPONENTS
Figure 6.1 Basic transformer action.
all flux is mutual, and has no other losses. In such a case, for the primary, the applied and induced voltages are equal, so: e1 = v1 = V1m sin ωt (equation 6.1) but also: e1 = N1 d(φm1 - φm2)/dt So: V1m sin ωt = N1 d(φm1 - φm2)/dt Which can be re-written (by integration) as: φm1 - φm2 = - V1m cosωt/N1ω The expression (φm1 - φm2) represents the mutual flux Φ, so its peak value can be written as: Φm= V1m /N1ω
Therefore the peak applied voltage V1m = ωΦm N1 Substituting this in Equation 6.1 gives us: e1 = ωΦm N1 sin ωt From this the r.m.s. voltages induced in primary and secondary can be written as: E1 = ωΦm N1 /√2 = 2πf Φm N1 / √2 = 4.44 f Φm N1 Similarly E2 = 4.44 f Φm N2 But the mutual flux is the product of flux density and core area, so it is also possible to write: E1 = 4.44 f BmAN1 and E2 = 4.44 f BmAN2
173
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
This confirms to us that the induced voltage is proportional to core area, the frequency and to the maximum flux density. This shows that in principle high frequency transformers can be smaller for a given power rating than their low frequency counterparts. Notice that there is nothing in the transformer equations above that says a transformer must have an iron core. However at low frequencies it is simply not possible to construct an effective transformer without one. Only by using a core material of high permeability is it possible to achieve a sufficiently high flux density, and a high mutual flux with minimum leakage. 6.1.3 Core materials The core material used in a transformer or choke must exhibit the right magnetic properties at the operating frequencies to be used. The B/H curve shown in Figure 1.14 represents the magnetizing effect as DC is applied to an inductor. Corresponding curves for AC get progressively fatter as the frequency is increased, and in practice the simpler materials cannot be used at high frequencies because of hysteresis loss. The use of laminations minimizes eddy current losses, but, again, at high frequencies they become appreciable and a different approach must be used. The principal ferromagnetic materials used for transformer cores are summarized in Table 6.1. The most common material, sometimes referred to as transformer steel, is a silicon iron alloy. A special version of this is grain oriented steel. In such material the magnetic domains within it are aligned, and this has the effect of reducing the magnetizing force needed to achieve a particular flux. The alignment is achieved by cold rolling the steel strip. Standard transformer steel is fine for transformers operating at normal mains frequencies, but is not suitable for audio frequencies or for circuits that depend on saturation. Alloys of nickel and cobalt are used and in all cases the final magnetic properties of the material depend very much on how it is treated. This relates both to the rolling process to achieve lamination thickness, and to heat treatment and cooling cycles. Table 6.1 refers to “square loop” and “round loop” alloys. Round loop alloys have a
174
traditional hysteresis loop like Figure 1.14. Square loop materials have an almost rectangular hysteresis curve like Figure 6.2. Such materials are characterized by a low magnetizing force to achieve saturation, high remanent flux density and low coercive force. For frequencies above the kilohertz range, eddy currents become a serious problem and it is necessary to use a core material that has a high resistivity. This can be achieved by using a powdered core, where, instead of separating the core into laminations, it is separated into particles. Both powdered iron and powdered alloys are used. Ferrites, shown at the bottom of Table 6.1, are non metallic ferromagnetic oxides of the form XO.Fe2O3 where X is a divalent metal, for example Iron (Fe), Zinc(Zn), Manganese(Mn), Nickel(Ni) or Magnesium(Mg). Actual ferrite core materials are usually a mixture of two or more types of ferrite crystals. Ferrites have a resistivity around 1013 greater than iron, and, as a result, eddy current losses are negligible. This allows operation at very high frequencies. The practical problems of ferrites are: • they are brittle, ceramic-like materials, difficult to work. • they have a relatively low saturation flux density. • they have a low Curie point. The behavior of the core materials described here is called ferromagnetic. At high temperatures they cease to be ferromagnetic, and the Curie point of a magnetic material is the temperature at which the change takes place. Ferrite Curie points can be so low that in a practical circuit design it is necessary to take account of operating temperature. 6.1.4 Magnetizing current When a transformer primary is fed an alternating voltage, and when there is no load on the secondary, it takes a very small current. Indeed, if it behaved as a perfect transformer, the primary of an unloaded transformer would behave as a perfect inductance, with no net current taken. What happens when a load is applied to secondary can be imagined as modifying this perfect inductive behavior. The mutual flux cuts through the secondary
ELECTROMAGNETIC COMPONENTS
and induces a voltage across it – if a load is connected, a current flows. As a load is taken from the secondary, the back e.m.f. of the primary is reduced, resulting in a net current flow. In the perfect transformer the mutual flux remains constant. For this to be the case the magnetizing force, and hence the m.m.f. expressed in Ampere Turns must also remain constant. So if the secondary ampere turns rises, the primary ampere turns reduce, correspondingly reducing the back e.m.f. No transformer is perfect, and it is interesting to examine the way the current varies. Figure 6.3 relates the induced voltage to the sinusoidal variation in flux. But the current needed to achieve the variation is NOT sinusoidal. This is because the core must be dragged round the hysteresis loop. The actual magnetizing current waveform will be dependent on the B/H loop for the material being used. Material 4% silicon iron alloy 300104% silicon alloy, grain oriented 40% nickel iron alloy “square loop” 50% nickel iron alloy “round loop” 70-80% nickel, 3-5% molybdenum iron alloy 3-50% cobalt iron alloy Powdered iron Powdered Molybdenum nickel iron alloy Manganese zinc ferrite Nickel zinc ferrite
Figure 6.2 The rectangular or “square” hysteresis loop. Hc is the coercive force necessary to reduce B to zero.
Trade name (all •)
Initial Permeability 400
Maximum Permeability 8,500
Maximum Flux Density 1.6T
Remanent Flux Density 1.1T
Typical Frequency <100Hz
Hypersil
1,500
30,000
1.8T
1.3T
<1,000Hz
Deltamax
1,000
50,000
1.5T
1.4T
<50kHz
Hypernik
7,500
50,000
1.4T
0.8T
<25kHz
Square Permalloy
20,000
100,000
0.8T
0.6T
<100kHz
Supermendur
800
50,000
2.2T
2.1T
<3.5kHz
Polyiron
5
85
0.9T
0.6T
MolyPermalloy (MPP)
30
550
0.6T
2kHz0010 300MHz 1kHz0010 1MHz
Ferroxcube
500
7,000
0.3T
0.2T
<1MHz
Ferroxcube
300
3,000
0.15T
0.1T
10kHz0010 20MHz
Table 6.1 Examples of transformer core materials. The values for permeability and flux density are indicative only, since within each category manufacturers offer a wide range. The trade names are examples – again for each category there are several manufacturers.
175
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
6.1.5 Transformer losses We can now identify a number of factors that make a real transformer differ from the perfect transformer. Even if a transformer has no load on the secondary, it still takes a measurable current (and hence uses some power.) One major component of this is the power needed to magnetize the core first in one direction and then in the other, as shown by Figure 6.3. This is the hysteresis loss that can only be minimised by the choice of the core material. As already described eddy current loss is bound to increase as frequency increases, and is again dependent on the core material and the way that it is presented. The apparent permeability of core materials also varies with frequency. Both hysteresis loss and eddy current loss are measured in W/m3 (Watts per cubic meter) for the operating frequency and material concerned. In practice core material manufacturers also use other units, for example Watts per pound or kg. In the perfect transformer, all flux is mutual. Figure 6.1 shows, however, that there is invariably some leakage flux. Power is still needed to drive this flux, so this represents flux leakage loss. Here the transformer designer tries to ensure that the coils are tightly coupled, and that the magnetic circuit is complete. However, as will be seen later, sometimes leakage flux has its uses. A particular problem with unintended leakage flux is that it may couple to other nearby conductors which means that not only is power wasted, but that there could be undesirable side effects. Both primary and secondary have some resistance, so both have i2R losses. These are referred to as copper losses – distinguishing them from the above mentioned core losses. All the losses are finally expressed as heat. Another source of copper loss is skin effect. The AC resistance of a conductor is found to be higher than the DC resistance. This is because in an AC conductor conduction tends to be only at the surface of the conductor. The reason for this is that the center of the conductor is encircled by more magnetic flux lines, and therefore shows a higher inductance than the outer part.
176
Figure 6.3 Core hysteresis results in a non sinusoidal exciting current, even though both the transformer voltage and flux are sinusoidal. The relationship of flux Φ and current i follow B and H.
Skin effect is frequency dependent. At high frequencies negligible current is carried other than at the surface of the conductor. Conversely for all but the very largest transformers, skin effect can be ignored for transformers operating at 50/60Hz. High frequency transformers of the type used in inverters use various strategems to lessen the impact of skin effect. One is to use Litz wire for the windings. Litz wire is wire made up of multiple individually insulated strands, thus effectively increasing the surface area of the conductor. Another possibility is the use of flat copper strip. In high frequency transformers capacitance becomes an issue giving another source of loss. The arrival of ferrite core materials and high frequency high power switching transistors has meant that significant voltages can be developed, and significant power transferred, from transformers of
ELECTROMAGNETIC COMPONENTS
very small dimensions. This type of transformer is of relevance as a component in electronic ballasts and transformers discussed in the next chapter. In this chapter we are more concerned with “traditional” transformers working at mains frequencies. The typical application of interest is a transformer with a 120 or 230V primary and 12V secondary for feeding a resistive load in the form of tungsten halogen lamps. Figure 6.4 shows a simplified equivalent circuit of a real transformer. This shows a perfect transformer, surrounded by circuit elements representing the various loss mechanisms. The circuit analysis becomes more complicated when the waveforms are not sinusoidal, and when the secondary load is complex (i.e includes inductance and capacitance). The reading list includes books that develop this subject in detail. There are several ways by which the performance of a transformer is measured. In the case of transformers for lighting, where the load is resistive, the measurements are simple. The regulation of a transformer is a measure of how constant the output voltage is under varying load conditions. It is defined as the difference between the secondary terminal voltages when at no-load and full-load expressed as a percentage of the full load voltage.
% Re gulation =
100 ( Open circuit voltage− Full load voltage ) Full load voltage
For typical power transformers with low leakage reactance it can be shown that the regulation is dependent on the resistive losses. As an example it is important that a transformer feeding multiple low voltage halogen lamps must have good regulation, since otherwise the failure of an individual lamp would result in a significant voltage rise. Such transformers must, therefore, be designed with minimum resistive losses. The efficiency of a transformer is represented by the symbol η. It is simply the ratio of the output power developed in the load to the input power put in to the primary. It can be expressed another way: η=
output power output power = input power output power plus losses
It can be shown that for a power transformer operating at constant input voltage, constant power factor but varying load current, the efficiency is highest when the I2R losses are equal to the constant core losses. Like the regulation advantage of low resistance, this point is intuitive if you consider that it is saying that if you want to run high currents with high efficiency, you must have superior core
Figure 6.4 Simplified equivalent circuit of a power transformer showing the principal loss mechanisms. The symbols are mainly defined in Figure 6.1. The core losses are shown as an equivalent resistance RC and the magnetizing current losses are shown as an equivalent inductance Lm1 both in shunt (parallel) with the primary. The current i 2 is the current flowing in the primary as a result of the secondary current.
177
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
materials. But it is a significant point in transformer design when trade-offs are being made between such items as wire size, core material costs and required regulation. For distribution transformers, where the load is variable, the transformer designer may well optimize the design to ensure maximum efficiency at one third or one half full load; since this might well be the typical load. If full load is only rarely required it makes sense to optimize for the load representing majority use. The power factor of a transformer with its load can be shown as: Power factor =
output power plus losses in W input Volt − Amperes
shapes are used to facilitate manufacture and minimize wastage of material, but in all cases the two lamination stacks must be tightly clamped together to ensure completion of the magnetic circuit. Low cost transformers can be made by using transformer steel strip (with at least one side insulated, as it would be with laminations) wound as if it was tape into a rectangular shape. The resulting core is then cut in half to allow the insertion of the bobbin(s) before being clamped together again. Such a construction was widely used in old radio sets etc. to achieve low cost, but is not used for lighting transformers. However, the principle of tapewound cores does form the basis of the toroidal transformers described in Section 6.1.6.3. 6.1.6.2 Wire and impregnation
If the secondary load is complex, then the power factor will be heavily dependent on its nature. However, most lighting loads are resistive, so the nearer to perfection the transformer is, the nearer to unity the power factor. In practice lighting transformers have a slightly lagging power factor, just less than unity, arising from the inductive losses. 6.1.6 Transformer construction 6.1.6.1 Core and shell Practical transformers have to take into account methods of manufacture. For simple power transformers it is easiest to wind the coils on bobbins as a separate operation, and then to slip the ready wound coils over a pre-prepared stack of laminations. The most common forms of construction are referred to as core and shell constructions. In core type transformers the windings surround the magnetic core, in the shell construction the magnetic core surrounds the windings. The principle is shown in Figure 6.5. Although in theory transformer construction allows the primary and secondary windings to be on separate parts of the core, better coupling is achieved by winding the primary and secondary over each other as shown in the diagram. Clearly the laminations have to be in two parts to allow the insertion of the coils. Different lamination
178
In order to occupy the minimum volume the wire used to wind transformers needs to have the thinnest practicable insulation layer. Within any winding there is a voltage gradient, such that neighboring turns have a relatively small potential difference, but the extreme ends of windings may have a high potential difference. Most transformers have an insulating layer between primary and secondary, and some require insulating layers between winding layers within primary or secondary when high voltages are involved. The majority of lighting transformers are step down transformers using comparatively thin primary windings with a large number of turns, and thick,
Figure 6.5 Transformer construction. The common core and shell constructions used in small power transformers.
ELECTROMAGNETIC COMPONENTS
high current, secondary windings of few turns. The wire used is enamel insulated. But behind these simple statements lie a number of practical issues. The insulation grade is chosen both for its insulation properties, and for its maximum operating temperature. Wire insulation can be single enamel, double or even triple enamel, glass/enamel, polyester/ glass and cotton/enamel. Any other insulators used in the transformer, for example the inter-winding layer, must have at least the same operating temperature capability. Not surprisingly insulators like paper and cotton have the lowest temperature rating; glass and high temperature resins the highest. Inevitably there is a temperature rise in transformer windings. Such a rise makes matters worse because increasing temperature results in increasing resistance. When rating the operating temperature of a transformer consideration is given to the ambient temperature (for example 35°C) the temperature rise of the windings (for example as much as 55°C) and provision for “hot spots” (for example in the middle of the coil, where the temperature might be yet another 10°C higher). In this example the total is 100°C. Insulation temperature performance is described by its class. For example, Class A has a maximum of
Figure 6.6 Typical lamination shapes used in transformers.
Figure 6.7 Insulation life as a function of temperature. This shows that transformer and choke life is dramatically reduced if the wire insulation is subjected to high temperature.
105°C, Class B 130°C and Class F 155°C. A typical lighting transformer uses Class B insulation for a maximum winding temperature of 130°C, allowing a maximum ambient temperature of 40°C, and resulting in a typical maximum case temperature of 85°C. The bad news is that insulation life is seriously dependent on temperature. The Class rating is for an expected 10 year life; but if Class B insulation is operated at 160°C it may fail within one year. Completed transformer coils are usually impregnated with an insulating material that is liquid when it is applied but solidifies as part of the impregnation process. The reasons for doing this are fourfold: • it helps conduct heat from inside hot spots to the exterior of the transformer. • it improves the dielectric strength of any interwinding insulation. • it prevents the ingress of moisture that could cause wire corrosion. • it mechanically stabilizes the winding; so the wire is prevented from moving. It also gives some protection against mechanical damage to the winding. To be effective it is essential that there are no air voids in the impregnation, therefore the process is carried out under vacuum.
179
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Most transformer impregnation is now done with polyester or epoxy resins. These cure to the solid as a result of baking at specified temperatures, as a result of a chemical reaction, or as a result of irradiation by UV. Where resin based impregnation is based on the use of solvents, there are environmental considerations. The transformer manufacturer must either have sophisticated traps to prevent the solvent getting into the atmosphere, or must use water-based solvents. Recent developments in resin chemistry have made this possible. 6.1.6.3 Toroidal transformers A transformer construction particularly relevant to lighting is the toroid. Toroidal construction is used for a wide variety of core types, including standard transformer steel, powdered iron and ferrite. The latter
two are mainly used in filter and high frequency components. The toroidal transformer using standard transformer steel has several advantages over its more conventional shell type competitor. The core is simple to make, since it only requires a wound tape construction without the need to stamp laminations. The resulting transformer is flat, making it easier to “design in” to lighting products. Best of all a toroidal transformer has very low leakage flux, so it is an efficient design. The low leakage allows toroidal transformers to be sited closely together, and minimises interaction with other components and metal housings. The windings are on the outside of the core, so heat is easier to dissipate. Generally a toroidal transformer occupies less space and has a lower weight than its shell construction counterpart. Mechanical noise is also reduced.
Figure 6.8 Construction of a toroidal transformer with steel tape core. Illustration courtesy of Toroid International Ltd.
180
ELECTROMAGNETIC COMPONENTS
It is, however, difficult to wind. Standard shell type transformers can have their coils wound on fully automatic machinery that can wind several bobbins at once. A toroid can generally only be wound on a single part machine, and involves pre-winding the wire onto a shuttle which is itself then passed through the core. However, the machinery for toroid winding has progressed to the point where the additional assembly costs are minimal and are usually outweighed by the benefits. Another problem with toroids is the lead-out. Clearly it is easy to insulate primary and secondary over most of their length by simple lapped insulation. The problem only comes at the point where the primary connection must make its way through the secondary at the lead-out point. Double insulation is used on the primary leads. Toroidal transformers are easy to mount. They usually have a simple circular clamping plate, with single bolt fixing, used in conjunction with protective pads to avoid damage to the windings. An alternative is to fill the center of the toroid with resin and provide a threaded insert for bolt fixing. One small point; if the fixing bolt is bolted onto a metal plate, then on no account must the head of the bolt be allowed to touch any other part of the metal casing. This would result in a “shorted turn” secondary for the transformer. The consequences could include total destruction of the transformer, possibly with a fire as an added attraction.
Figure 6.9 Examples of ferrite cores used in making the inductors and transformers used in electronic ballasts.
6.1.7 Special transformers 6.1.7.1 Three phase transformers Three phase tansformers are an essential part of electricity distribution. One way of transforming three phase electricity is to simply use three single phase transformers, but this is rather inefficient. However, in local electricity distribution it is often the case that for a small area a single phase transformer is used;
6.1.6.4 Ferrite cored transformers Whilst ferrite cores are fragile, they have the great advantage that, because they start life as a powder, they can be molded into many different shapes. Thus ferrite cores come in thousands of different varieties, in terms of size, shape and core material. Shapes include E, I, U, straight rod, toroid and pot core. The pot core is a two part construction where the ferrite both passes through the coil bobbin and surrounds it. The choice of shape is made depending on the application and on the required method of mounting. Where small two part ferrite cores are used, they are usually held together by spring clips. Figure 6.10 The “theoretical” three-phase transformer.
181
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 6.11 Examples of three phase distribution transformers. A group of oil cooled “Tunorma” local distribution transformers. Photo from Siemens PTD.
and two other neighboring areas use another two single phase transformers to balance the load. Most large buildings and neighborhoods receive their electricity at three phase, and require local stepdown to mains voltage. For the same reason that three-phase transmission is more efficient than single phase, a three phase transformer is very much more compact and efficient than three single phase transformers. A theoretical three phase transformer is shown in Figure 6.10. Here flux change is shared between the magnetic circuits. For the same reason that, in a balanced three phase circuit, there is no neutral current (see Sections 1.3.2 and 1.3.3) there is no magnetic flux in the center limb of the theoretical transformer. In theory the center limb could be omitted. The theoretical three-phase transformer would be difficult to make. In practice it is possible to make three-phase core and shell transformers as shown in Figure 6.11. The construction does not provide a perfect balance for the three magnetizing currents, because the magnetic circuits are not exactly equal in length, but in practice the balance is very close and achieves the required size reduction. Three phase transformers are configured in different ways according to application. The most common arrangements are: Delta-delta (Δ-Δ) transformers, used as part of the primary distribution network. For example to step up from alternator voltage to distribution grid voltage,
182
or to step down from grid voltage to area distribution (e.g. 400kV down to 33kV.) Delta-star (also known as Delta-Wye or Δ-Y) transformers, used to step down to mains voltage. The outputs have a neutral and are, therefore, suitable for local power distribution. The delta star transformers are reasonably tolerant of out of balance loads, in the sense that regulation does not suffer unduly. Delta zig-zag transformers are used when it is expected that loads may be seriously out of balance. Here the secondaries are wound in two parts, one half being wound on one limb, and the other half wound in opposition on the next limb. Note that in this case, because the flux changes within a limb are not in synchronization, the output phases are not exactly in phase with the inputs – the actual phase shift depends on the allocation of the secondary turns to the two limbs of the transformer. Star-star (Y-Y) transformers are rarely used, since they are the equivalent of three single phase transformers, and suffer from out-of-balance problems. Likewise star-delta (Y-Δ) transformers are not widely used as they represent a rarely required configuration. 6.1.7.2 Harmonic rejection transformers Section 1.3.3 explained that in a three-phase electrical system the neutral current is zero when the loads are balanced. Also the neutral voltage is intended to be at ground (earth) potential. Unfortunately this statement is only true if the loads are simple (resistive equivalent) and do not themselves introduce waveform distortion. Very often this is not the case, and the load introduces harmonics. While even harmonics are not of great concern, odd harmonics, especially third harmonic (150Hz or 180Hz for most electricity supplies) cause a real problem. This is because, far from cancelling out as neutral currents, the odd harmonics are additive. In the worst case this results in a neutral current that is considerably higher than the single phase neutral current. This in turn can lead to big potential differences between ground and neutral (which can
ELECTROMAGNETIC COMPONENTS
introduce an exactly opposite current, and this can be done by introducing a three phase autotransformer of special construction near the load. It is important that such a transformer represent a comparatively low impedance to the unwanted neutral currents, so they are, effectively, fed back into the loads rather than to the supply transformer. The principle is shown in Figure 6.13. Harmonic rejection transformers can result in higher than expected fault currents to ground (earth) and their use does require a careful analysis of the actual circuit conditions; however, they can do a
Figure 6.12 The third harmonic problem in a distribution network. It results in a big neutral current, a PD between ground and neutral and a “trapped” harmonic current in the primary of the distribution transformer, resulting in overheating.
seriously affect electronic circuits) and to overheating conditions in transformers. The electricity supply companies are also not happy with harmonics. One of the principal offenders in harmonic production has been computer power supplies. For reasons explained in the next chapter, distortion from electronic power supplies is now much more closely controlled, so it is less of a problem than it was – but it still remains the case that within a building using large numbers of electronic power supplies (including lighting electronics) third, and higher, harmonics on the mains can cause a real problem. Delta-star (Δ-Y) transformers feeding a large building can end up with a large circulating third harmonic current in the primary (Figure 6.12) as a result. The only way to eliminate such a current is to
Figure 6.13 Introduction of the harmonic rejection tranformer, also known as the “neutral trap”. By using zig-zag winding, it is possible to return the third harmonic currents to the loads, but phase shifted. This results in, typically, an 80% reduction in third harmonic neutral current.
183
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 6.14 A 400A harmonic rejection transformer from Claude Lyons Ltd.
significant “clean up” job, especially if the incoming transformer (Δ-Y) is also wound with harmonic cancellation characteristics. 6.1.7.3 Leakage reactance transformers Most transformers are designed so that as much flux due to the coils is confined to the magnetic circuit as possible, and so that primary and secondary are closely coupled, usually by winding one on top of the other, or even interleaving them. The aim is to minimize flux leakage to achieve maximum efficiency. However, some transformers are deliberately designed to have a controlled amount of flux leakage because it can provide current limiting in the secondary. The effect of not having close coupling, indeed having a “loose” coupling, is for the secondary to become progressively decoupled from the primary as the current increases. The effect is magnified if a flux leakage path is introduced. For applications requiring current limiting, or near constant current operation, this effect can be advantageous. Figure 6.15 shows a transformer with magnetic shunts intended to increase leakage flux i.e. to divert flux that would otherwise be coupled to both primary and secondary. The shunts consist of lamination extensions that do not pass through the coils, and complete a magnetic circuit via an air-gap. The practical effect of this is that at small loads the
184
Figure 6.15 One possible configuration of the current limiting or leakage reactance transformer. The graph shows output voltage plotted against output current in the perfect current limiting transformer
transformer behaves normally, and the shunts carry little flux. As the load increases, the secondary ampere turns force more of the flux into the shunts, until at short circuit current the output voltage is zero. In a perfect transformer of this construction, the secondary voltage/current graph is a quarter circle – equivalent to the secondary current at short circuit being all reactive. Real leakage reactance transformers show some losses, but have the useful current limiting characteristic. It can be seen from the curve that the output current is substantially constant over a quite wide range, and anyway has a limiting value. This characteristic is ideal for some discharge lighting applications. 6.1.7.4 Auto transformers For many applications the autotransformer represents an efficient way of transferring power, since it can be much smaller than its equivalent conventional transformer with isolated windings. This is because in an autotransformer much of the power transfer is by direct conduction as opposed to induction.
ELECTROMAGNETIC COMPONENTS
Figure 6.16 illustrates the principle of the stepup and step-down autotransformer. In the step-up version the voltage induced in the additional winding is added to the supply voltage. In the step down version the induced voltage reduces the supply voltage. Autotransformers can only be used in applications where isolation between primary and secondary is not required. The short circuit current is much higher than in transformers with isolated windings. They are ideal for bucking or boosting an AC supply, when it is required to change the supply voltage by a comparatively small amount. In this case the economy afforded by the autotransformer principle is considerable (compared with the conventional transformer alternative). Most modern electronic equipment now uses electronic power supplies that can operate over a wide input voltage range (e.g. from 100–250V 50/60Hz). However older equipment and appliances still need transformers if it is required to operate European 230V equipment on USA 115V supplies and vice versa. Auto transformers are often used for this purpose because of their compactness, but it must be remembered that transformers being operated from
Figure 6.16 The auto transformer, shown in both its stepup and step-down versions.
230V to give 115V can apply the full 230V to the load if there is a secondary winding failure. For this reason transformers used on construction sites to provide 110V MUST be full isolating transformers (indeed they must now have center tapped secondaries to ensure there is no “live” above 55V rms, which is non-lethal). A special variant of the auto transformer is the variable auto-transformer. These are widely used in laboratories and other applications where a smoothly variable AC supply is required. In this case a single layer winding is carefully done so that adjacent turns of the coil do not touch, and at one point on each turn the insulation is removed. A movable sliding contact brush makes contact with the “bare” part of the winding and provides the variable output connection. The fact that where the brush touches the windings it may actually cover more than one turn is not a problem in practice, since the volts per turn figure is low, and the heating effect in the short circuited turn(s) is minimal. The smoothness of the variable output depends on the number of turns, since this determines the resolution of the output. A typical 230V variable autotransformer has between 300 and 600 turns depending on the rating, so the output changes in around 0.5–1 volt steps. Variable auto transformers can be constructed in a straight-line form, but are most commonly constructed in toroidal form. This is both more efficient and more convenient since it allows a simple rotary control. They are usually supplied either as line voltage transformers, which means that they can deliver any voltage from zero up to the line voltage; or as over voltage transformers. These typically have a 10% additional winding so, for example, a nominal 230V transformer can give an output in the range 0– 250V, which is useful for test purposes or where the supply is sub-standard. Variable auto transformers are sold under such registered trade marks as “Variac” and “Regavolt”. Typical ratings of individual variable autotransformers are from 250VA up to 7.2kVA. They can be ganged together to provide a variable 3phase supply. Higher current and higher voltage operation can be achieved by operating such
185
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Variable autotransformers are also available in motor driven versions, and these can form the basis of automatic voltage regulation systems.
6.2 Transformers for lighting 6.2.1 Introduction
Single phase 1KVA
Bench/portable 650 VA
Three phase 21.6 KVA
Figure 6.17 Examples of variable auto transformers. These are Regavolt™ products from Claude Lyons Ltd.
Transformers used for lighting and intended for operation at normal mains voltage and frequency are conventional in construction, and in principle all the points made in Section 6.1 above apply to lighting transformers. The special type of transformers needed as part of fluorescent lamp ballasts are dealt with in Section 6.3. There can be a conflict in the specification of lighting transformers between the need to have the lowest cost component and the need for long-term performance. The main problem is heat, since heating within a transformer is both potentially dangerous and can greatly reduce transformer life. Lighting transformers are often sited within luminaires or enclosures that restrict ventilation, resulting in the need to operate in a high ambient temperature. Safety is a major issue with lighting transformers. 6.2.2 Transformers for low voltage tungsten halogen lamps 6.2.2.1 Requirement
transformers in series or parallel, although in the latter case small reactors must be placed in the output lines of each transformer to limit circulating currents between transformers. An important point to remember is that a variable autotransformer has a maximum current rating that can only be drawn when the transformer is set near maximum output (or near zero output). If the load is constant impedance, this limitation is no problem since the current taken at any setting will be proportional to the output voltage. But care must be taken with variable impedance loads (including incandescent lamps with their low cold resistance) to ensure that excessive current is not taken at intermediate settings. This could result in hotspots in the winding and its eventual destruction.
186
The great majority of tungsten halogen lamps used for architectural, commercial and decorative applications operate at 12V. There are also some specialist lamps that operate at 6V or 24V. Similarly the great majority have power ratings in the range 20–50W, although lamps with ratings as low as 5W are available, and the higher ratings include 65W, 75W, 100W and 150W. Many applications require that each lamp has its own individual transformer. However low voltage track systems, multi-lamp luminaires and downlighter systems can use one transformer to supply several lamps. Because of the high secondary current involved, and the need to avoid voltage drop by siting the transformers as close to the lamps as possible,
ELECTROMAGNETIC COMPONENTS
Figure 6.18 In the IBL “Wiresystem” the support for multiple low voltage spotlights doubles as the low voltage distribution. In fact the wire is transparent insulated, but pierced at the point of support/connection. Such systems (and low voltage track systems) are limited by the current that it is practical to carry – usually 25A, representing a 300VA circuit at 12V.
the rating of transformers for multi-lamp loads does not often exceed 600VA. At 12V even a short length of wire can introduce a significant voltage drop at the comparatively high currents involved. The distribution of the 12V to an individual lamp or group of lamps must be carefully considered to ensure that the correct, and, in the case of multiple lamps fed from one transformer, the same, voltage appears at all the lamp terminals. Many transformers offered as nominally 12V are actually 11.5V (at full load) in order to ensure that the lamp does not receive an excess voltage under any operating condition, and to prolong lamp life; however, in some cases this can result in lamps not providing the expected light output. Low voltage lighting transformers are run at or close to their full power rating most of the time. They are, therefore, designed for maximum efficiency at full load (I2R losses equal to core losses). Regulation is not an issue for a single lamp transformer intended for a specific lamp rating, since the load conditions are defined – the fact that the open circuit voltage might be 13V for a running voltage of 11.5 is not a problem. But it clearly is an issue for multi-lamp transformers, or for transformers intended to be used for different lamp ratings.
over-heat condition destroy the transformer, it can easily start a fire. The likely location of the transformer (within a ceiling void for example) makes the situation even more dangerous. Thus all low voltage lighting transformers must be equipped with comprehensive protection. This usually consists of separate methods of protection against short circuit and over-heating. The simplest protection against over-heating is the installation of a thermal fuse in the primary. This device is solely temperature dependent, and interrupts the current at a particular temperature e.g. 125°C. Such devices are available in both one-time and automatically resetting versions. “Budget” transformers for single lamps may use the one-time version, but, if it operates, the transformer will need replacing. Multi-lamp transformers and higher quality single lamp transformers use the automatic re-setting type. Some of these require that the power is disconnected to initiate reset, in addition to the temperature dropping. Short circuit protection may be provided by an anti-surge fuse fitted in the transformer casing. This is the safest method, but some users rely on the feed circuit’s own protection. Secondary circuit protection is not normally provided for single lamp transformers; however it is advisable for multi-lamp transformers. The best of these have multi circuit ouputs with a separate fuse for each output leg.
6.2.2.2 Transformer protection The consequences of failure in a lighting transformer are extremely serious. Not only can an over-load or
Figure 6.19 Example of thermal fuses used as protective devices in transformers. These break the circuit at a defined temperature, not at a defined current. Photo from Limitor.
187
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
conventional shell type transformers being used for lighting, but, in order to meet market demand and current safety requirements, the transformer is somewhat disguised in its housing. 6.2.2.4 Toroidal transformers Figure 6.20 Example of lamination based lighting transformer rated at between 50VA and 105VA. The construction is similar to that used for electromagnetic fluorescent lamp ballasts, and lends itself to automatic mass production. The “Slim” range from Relco.
6.2.2.3 Conventional transformers Present day demand is for ever more efficient lighting systems, with minimum energy losses and maximum safety and performance. This has had a side effect of marginalizing the use of conventional lamination based transformers, especially since the most efficient lamination based designs are not as convenient to mount as their toroid equivalent. However, for single lamp and small load applications there is still a place for lamination based transformers in lighting. These are now of designs that lend themselves to automatic mass production, and are usually supplied in special housings to match a particular application or luminaire design. Figure 6.20 shows an example of such transformers, and it is easy to see how the design has to some extent been determined by the production method. Nonetheless the resulting shape is convenient for many applications. Figure 6.21 shows
Figure 6.21 Lighting transformers from Relco of a more conventional construction, but disguised by the plastic housing that provides double insulation. Ratings 30–250VA.
188
The improved efficiency, silence, convenient shape, and low leakage of the toroidal transformer means that it is the preferred construction for electromagnetic lighting transformers. The manufacturers supply the transformer in many ways, depending on whether it is to be built in to a luminaire or housing, or is intended to stand alone. The accompanying figures show toroidal transformers in their unenclosed form, and in various types of enclosure. When selecting a toroidal transformer for lighting the temperature rating is the thing to watch for (as is the case with all other ballasts, a subject that is developed further in Section 6.3.) A rated ambient temperature of 40°C may sound generous, but is really the minimum specification for lighting transformers. An obvious point, but one that
Figure 6.22 Winding toroidal transformers. The split ring passing through the toroid must be loaded with the correct length of wire before winding onto the core can start. Photo courtesy of TCI.
ELECTROMAGNETIC COMPONENTS
Figure 6.23 Typical toroidal lighting transformers from TCI. Notice the availability of a “flat” construction.
is often overlooked, is that the transformer must not be sited in a way that allows the lamp load to heat it. The surface temperature of the transformer at full load is a potential hazard. Many lighting transformers are rated to have a case temperature as high as 90°C (or more). This may not be acceptable in furniture etc., and anyway is a burn hazard. Most toroidal transformers for lighting are rated Class B insulation, with a maximum winding temperature of 130°C. For some applications it may be necessary to specify transformer wire with insulation rated for a higher temperature. 6.2.2.5 Dimming low voltage lighting transformers Low voltage tungsten and tungsten halogen lighting fed by transformers can be dimmed, but there are important caveats. If the dimming is done by simply reducing the amplitude of the AC sine wave, for example by using a variable autotransformer, there is no problem. The lighting transformers continue to function normally on the lower line voltages, and the lamps dim accordingly. However, today most dimming is done by electronic dimmers, and this can have some unexpected side effects.
Figure 6.24 A fully encapsulated toroidal transformer from IBL Lighting. Intended for supplying low voltage track systems, it has a built in anti surge fuse, automatic thermal cut out and generous stud terminals for the load connection.
The majority of electronic dimmers work by switching the AC so that only the required proportion of energy gets through. Thyristor and triac dimmers have leading edge switching, transistor dimmers have trailing edge switching, as shown in Figure 6.26. Two problems arise. First the fast rise or fall of voltage effectively means that the transformer is being asked to work at higher frequencies; thus core losses can increase with resultant overheating. Second, and more serious, if there is any assymetry in the two halves of the controlled waveform, there will be a DC component in the output. This can lead to saturation of the core, and, in the extreme, the transformer primary behaving like a simple resistive load. In the case of poor transformer steel, the DC can even lead to permanent magnetization of the core. In practice the problems can be avoided. The rules are: • use only high quality, grain oriented, transformer steel in the transformer. • ensure that the manufacturer of the transformer confirms that the transformer is suitable for electronic dimming.
189
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
• when using thyristor dimmers, use hard firing techniques (explained in Chapter 8) to ensure symmetry. • ensure that the dimmer is intended for transformer loads. • if using “flashing” for effect, ensure that the flash control results in wholly symmetrical power to the transformer. Provided that the transformer is intended for dimming, then electromagnetic lighting transformers should all work on thyristor, leading edge switching dimmers. Some toroidal transformers are also specified as being able to work on trailing edge transistor dimmers – but this may depend on the dimmer. A story from 20 years ago is a cautionary tale of what can go wrong. A London nightspot installed a large quantity of transformer fed low voltage lamps. These were fed from electronic flashing devices. Quite correctly these were of the zero crossover type – i.e. the switching on of the triac was done at the zero crossing of the sinewave to ensure that
Figure 6.26 Electronic dimming by thyristors and transistors.
harmonics were minimized, and that the transformers received only sinewave power. However, the highly accurate, crystal controlled, flashing routine resulted in each “flash” having an unequal number of positive going and negative going half sinewaves. Thus each flash had a small, but quite measurable DC component. This resulted in the quick destruction of the transformers. 6.2.3 Transformers for cold cathode lighting
Figure 6.25 For some installations it is more convenient to have the toroidal transformer mounted in a metal box. This 600VA example from IBL has full secondary load protection with fused multiple outputs.
190
The lighting transformers described so far have resistive loads on their secondary, and behave substantially as resistive loads as seen by the mains supply. A discharge lamp represents a quite different type of load. If the voltage is not high enough there is no current; but once conduction starts the lamp behaves as a negative resistance. It is necessary to have some means of limiting the current. Cold cathode and “neon” lamps (described in Sections 3.6.5 and 5.3.2) require a high AC voltage to operate. Because of their nature the supply to them must be of a kind that both provides the high starting voltage required, and limits the current. The current limiting transformer, based on the leakage reactance principle described in Section 6.1.7.3 meets the
ELECTROMAGNETIC COMPONENTS
Tube Diameter mm
Volts per meter for Neon filling
Volts per meter for Argon 0010 Mercury filling
9 11 15 18 20
1,330 1,050 730 600 550
930 750 530 400 350
Volts per electrode
150
112
Table 6.2 The open circuit voltage required by neon/cold cathode lamps. Table derived from information in “Neon – a practical handbook” By Peter J. Mason.
requirement. In simplistic terms it can be thought of as a high voltage transformer with a built in current limiting reactance in the secondary. Transformers for this application are specified by: • open circuit voltage • operating current • short circuit current
Figure 6.27 The “Eurosafe” range of transformers for cold cathode and neon lighting from Tunewell Transformers Ltd embodies both open circuit and earth leakage protection. They feature center tapped secondaries. Current range is from 25mA to 200mA and output voltages from 1.5–0–1.5 kV up to 5–0–5kV.
For safe operation of cold cathode lamps it is essential that the lamp is closely matched to the transformer. In many countries there are regulations specifying the performance of this kind of transformer. A transformer may feed a single length of tubing or several lengths in series. The open circuit voltage required to strike the tube(s) is determined by the total length of tube, the gas filling and by the number of electrodes. Table 6.2 gives some examples. The ratio between the operating current and the short circuit current varies according to local practice. Typically the short circuit current is between 110 and 120% of the operating current. Such an arrangement has the advantage of excellent current limiting and, therefore, safety – since the short circuit primary load is only a few percent above running load. However it means that the transformer has to be closely matched to the load in respect of open circuit voltage to ensure reliable lamp operation. Some countries permit a higher ratio – even up to 170%, but while this simplifies the operation of multiple tubes and reduces the number of different transformers needed, it is less safe practice. This is because, in the event of fault, the primary current rises accordingly with possible overheating. Transformers of this kind are usually fitted with additional protective devices. Residual current devices protect against leakage of the high tension voltage to earth (ground), and open circuit detectors shut down the transformer if the secondary is in an open circuit condition for more than a few seconds. The running voltage across the tube when conducting is considerably lower than the open circuit voltage needed to strike it (typically 60%) so detection of the open circuit condition is quite easy.
Figure 6.28 Cold cathode lighting ballasts from Tunewell embody a 1.2kV 100mA transformer and a power factor correction capacitor in a simple housing. Intended for 20mm diameter tubes, 2–3m long.
191
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
The load presented by a transformer fed cold cathode lamp is reactive, with a low power factor. Therefore complete lighting units use power factor correction capacitors, exactly like standard fluorescent lamps. Transformer fed cold cathode lighting can be dimmed; with the same caveats applying to both dimmers and transformers as given in Section 6.2.2.5. Any power factor correction must be prior to the dimmer device, and NOT connected as part of the load. The wiring of luminaires using cold cathode lamps requiring voltages below 1kV can be conventional, subject to cables of the required insulation rating being used. As the voltage gets higher much greater care is needed, since the high electric field created by kilovolt wiring can lead to unexpected current leakage and hazard. Transformers must be sited as close as possible to the lamps, and the special high tension leads kept as short as possible.
voltage and lamp current. Main light output is in the region of negative resistance. For the simple fluorescent lamp circuit introduced in Figure 3.15, the dynamic characteristic results in the current and voltage waveforms shown in Figure 6.30. From this it is clear that any practical ballast circuit must have the following attributes: • the circuit must be able to generate a starting voltage significantly higher than the running voltage. • the supply voltage must be significantly higher than the arc voltage in order to ensure that the
6.2.4 Transformers in series There are some applications where lighting transformers are operated in series. Examples are series regulator ballasts for HID lamps, and series operated incandescent lamps used for airport runway lighting. The former technique is not widely used, and the latter is discussed in Section 16.7.
6.3 Ballasts for fluorescent lamps 6.3.1 Requirements The aim of any lamp-ballast combination is ensure that the lamp operates safely, at maximum efficacy, and with a long life. The lamp should also start quickly, and the combined circuit must not represent an undesirable load on the supply. Ideally the combination should provide some regulation in the sense that changes in supply voltage should not be reflected in correspondingly large changes in light output. The fluorescent tube represents a complex load. Figure 6.29 shows the relationship between lamp
192
Figure 6.29 Relationship between fluorescent lamp voltage and lamp current.
ELECTROMAGNETIC COMPONENTS
• lamps can be operated across phases in a 3-phase supply in order to achieve the higher line voltage. Or a special “lighting” supply can be provided. In the USA industrial and office lighting is operated from 277V supplies. In Canada 347V is used. • ballast circuits can include a step-up transformer to produce the required higher voltage. 6.3.2 Ballast circuits
Figure 6.30 Current and voltage waveforms for the simple reactor ballast circuit.
discharge is reliably maintained as the current changes direction. • there must be provision for power factor correction. • the design should aim to reduce the harmonics arising from the non sinusoidal lamp current. It can be seen from Figure 6.30 that the simple reactor ballast provides a lamp voltage that is out of phase with the current. This is helpful as it ensures sufficient voltage for restrike at each half cycle. A resistive ballast would not only be wasteful, but would result in a significant discontinuity as the supply voltage changed direction resulting in much worse harmonic generation. Figure 3.24 shows that fluorescent lamps operate at around 100V. A mains supply of 200V or more provides the necessary margin for reliable lamp operation using the simple reactor ballast discussed so far. For all parts of the world that have 220–240V as their single phase supply the simple reactor ballast is the most efficient electromagnetic component for the job. In the USA, Japan and other 100–120V countries, the simple reactor is not adequate for tubes of significant length. Two different techniques are used to get round the problem:
In 220V countries most fluorescent lamps are operated from reactor electromagnetic ballasts using the simple starter circuit of Figure 3.15. Such ballasts can be made reasonably efficient, and are inexpensive. While there are moves to the more widespread use of electronic ballasts, the simplicity, low initial cost and reliability of the electromagnetic ballast means that it will be around for a long time. The situation is somewhat different in 110V countries. The ballast is significantly more complex to manufacture, and is less efficient than its 220V counterpart. The difference in manufactured cost between electronic and electromagnetic ballasts is much less, so that in the USA electronic ballasts now have more than 50% of the standard ballast market. By 2010 it is expected that electromagnetic ballasts will no longer be permitted in the USA because of their comparative inefficiency at 110V. One side effect of the more complex ballast construction for lower voltage supplies, is that it allows for different starting methods, some of which do not require a conventional starter. Figure 6.31 summarizes the main circuits used. Three different starting methods are shown. Pre-heat starting, using the conventional starter. Here a current is passed through each cathode for a short time prior to the striking of the arc. Once the tube is running the cathode temperature is maintained by bombardment, and no separate heating current is required. Rapid starting. Here the cathodes are fed a heating current from a separate circuit. This may be from a separate transformer, or, more usually, from additional windings on the main ballast. Such circuits usually have a capacitor in the circuit to help develop the starting voltage. In “energy efficient” rapid start
193
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 6.31 Some fluorescent lamp ballast circuits.
circuits there may be an arrangement to disconnect the cathode heating once the tube has started. Rapid start lamps require a grounded conducting plate within about 12mm of the tube (usually the luminaire structure) for reliable operation, and are silicone coated to reduce the effect of humidity. Instant starting. (See also Section 3.6.4.) Here no heating current is applied at all, and starting relies on a high enough electric field being created in the lamp. In principle the lamps must be designed for this method of operation, although, as will be described in the next chapter, some electronic ballasts
194
“instant” start tubes normally intended for pre-heat or rapid start. Figures 6.31 (a) (b) and (c) show the common pre-heat circuits. (a) is a repeat of Figure 3.15. (b) is a variant suitable for short (<600mm) tubes where two tubes are run in series. (c) shows a circuit using a capacitor in series with a reactor. This arrangement is useful for a number of applications. It results in a leading power factor; thus by mixing lamps with this circuit with lamps with the standard lagging power factor a good overall power factor can be obtained.
ELECTROMAGNETIC COMPONENTS
In such a mixed installation, the slight stroboscopic effect that arises with fluorescent lamps operating at 50 or 60Hz is virtually eliminated because the two sets of lamps have peak light output at slightly different times. The nature of the circuit allows high voltages to be developed, so it is particularly suitable for long tubes. In this connection the capacitor must typically have a voltage rating double that of the mains supply. It must also be of close tolerance (usually specified as ±4% or even ±2%) to ensure correct tube operation. Figure 6.31 (d) shows a resonant rapid start circuit of the kind used, for example, for 2,400mm lamps in the USA. The capacitor provides power factor correction and the high starting voltage. The first choke has two transformer windings to provide heating current to the cathodes (approximately 3.6V). The complexity of 120V ballasts has resulted in them often being designed for more than one lamp. A typical circuit is Figure 6.31 (e). Here two rapid start tubes are run in series from a leakage reactance autotransformer and capacitor. A small capacitor shunts one tube to assist starting, on power-up it momentarily puts the full ballast secondary voltage across the second tube. Sequence starting of two tubes also applies in the instant start circuit shown in Figure 6.31 (f). An auxiliary winding helps start the first lamp. In instant start circuits very high open circuit voltages are developed, and for UL approval luminaires are required to have circuit interrupting lamp holders. These are constructed such that if a lamp is removed, power to the ballast is disconnected. 6.3.3 Reactor fluorescent lamp ballast construction 6.3.3.1 Principles The idea of using a simple reactor to limit the current through a fluorescent lamp is that the reactor should behave like a pure inductance. If it really did so, there would be no energy losses in the current limiting process, and all consumed energy would be within the lamp.
As an example, Table 3.2 shows that the common T8 36W 1200mm tube has a running current of 430mA, indicating a voltage across the tube of approximately 84V. Disregarding the complication that the lamp voltage and current waveforms are not sinusoidal, and treating the lamp as if it was a resistive load, the ideal version of the circuit of Figure 6.30 would have the voltage across the lamp VL as being 90° out of phase with the voltage VB across the ideal choke ballast. The simple vector diagram of Figure 1.32 tells us that, if the supply voltage VS is 230V, the voltage across the choke is given by: VS2 = VL2 + VB2 For the example, this gives a value of the voltage across the choke as 214V. The impedance of the choke, is, VB /I . The current I in this case is the lamp current, so the needed impedance is 498 ohms. The impedance equals ωL, so if it is further assumed that the circuit is running at 50Hz, the required inductance L is given as: L = VB / I × 2πf = 214/0.43 × 2π × 50 = 1.58H So far, so simple – the need for a 1.58 Henry choke seems to be established. But how to make one? and anyway, is this what is really needed? The obvious idea is to make a choke using a transformer core and a single coil. But in practice this does not work. The two principal reasons are: • that the core can easily saturate. If it does so, the choke ceases to act as an inductance, and only its resistance limits the lamp current. • the permeability of the core material varies enormously with both the magnetizing force and with the magnetic flux, see Figure 6.32(a). So the inductance changes with the instantaneous value of the current. Figure 6.32(b) shows that inductance increases with increasing current at first, but having reached a maximum, actually decreases with further increasing current. The ideal core material would, therefore, be a core that could not saturate, and whose permeability is constant under all conditions of magnetizing force
195
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 6.33 A magnetic circuit with airgap.
The inductor with an airgap can be described by analogy to an electrical circuit. Figure 6.33 shows a two part magnetic circuit, one part with an iron core, one part with an air core. In it a flux Φ Weber is developed by an m.m.f. IT Ampere-turns. The relationship between these two quantities is the reluctance of the core – and if m.m.f. is analogous to e.m.f, and flux to current, then reluctance is analogous to resistance. In Figure 6.33 the reluctance of the airgap is in series with the reluctance of the iron core; if we disregard any leakage, the flux (equivalent to current) is the same in both, so:
Figure 6.32 Variation of relative permeability with magnetizing force and flux density for a typical transformer steel (a). The resulting variance in the inductance of an iron cored coil with current (b).
and flux. The material for which both statements are true is air. Reactor chokes for fluorescent lighting use a laminated core, and the same kind of steel as used in small power transformers. The difference in construction lies in the introduction of an airgap into the magnetic circuit. The airgap has the practical result of maintaining an inductance value almost independent of variation in main core permeability, and of preventing saturation. Its dimensions and stability are critical to the correct operation of the choke.
196
Figure 6.34 An example of a standard or reference ballast used for comparative measurements of ballast performance. Its construction is specified by the IEC.
ELECTROMAGNETIC COMPONENTS
LP-4
0A
Type Turns in coil Wire diameter mm Stack length mm Stack height mm Stack width mm Flux path length mm Airgap mm Voltage (test) V Current (test) A EEI
L36A
L2 L36 T
-S
LP-40A 710 0.27 65 26 40 56.8 0.38 178 0.4 C
L36A-S 578 0.30 82 26 40 56.8 0.30 178 0.4 B2
L36 TL2 585 0.385 95 26 40 0.32 178 0.4 B1
Figure 6.35 Three ballasts of differing efficiency suitable for 36–40W fluorescent lamps. Data from Helvar. (Type numbers are for identification only, and have no other significance.) Flux =
m.m.f (reluctance of iron + reluctance of gap)
The reluctance of a material is l/μrμ0a where l is the length of the magnetic path and a the cross sectional area of the core. For the case of Figure 6.33 a is assumed to be the same for both the iron core and the airgap. So the flux is given by: Φ=
IT l1 l + 2 μ r μ0 a μ0 a
(equation 6.2)
where μ0 is the permeability of free space (and air) or 4π × 10-7 , and μr is the relative permeability of the iron core. Clearly the airgap has a much higher reluctance per unit length than the iron core. By getting the relationship between the two path lengths of airgap and iron core l2 and l1 right the choke behaves as a linear inductor, with variances in relative permeability making little difference. Most important, the inductor becomes an efficient energy store; ballast designers talk in terms of the airgap being the key to ballast operation, and the rest of the magnetic circuit
being there to service the gap! The airgap solves the problem of consistent energy storage, and in theory would allow a choke of a specified inductance to be designed from classic equations like those of Section 1.2.7. Unfortunately, however, the idea of a simple inductance value simply does not work because of the non-sinusoidal waveform and non-linear load characteristic. Indeed the problem is so complex that it does not lend itself to simple mathematical treatment. As a result of the complexity of the voltage/ current relationship in a real world ballast, the real world is pragmatic. In order to have a reproducible means of ballast comparison, the IEC define a standard ballast (also known as a reference ballast) against which the performance of real ballasts is compared. Performance measurements of real ballasts must be made in a way whereby the complex waveform does not itself affect the measurements. Today the practical method is to use modern electronic instrumentation. Such instrumentation can make 50,000 instantaneous voltage measurements per
197
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 6.36 Magnetic flux distribution within a ballast of the kind shown in Figure 6.35. The figure shows only half the ballast section (the other half is symmetrical to it). The flux density across the airgap can be seen to be around 0.9T. Figure from Helvar.
second. The same instrument can measure current by measuring the voltage across a known resistance. The resulting measurements can re-create the complete current and voltage waveforms at 50/60Hz, that can be analyzed to give true power performance information. Simple reactor ballasts are designed so that they have a known voltage/current relationship, and conformity testing is done by measuring the voltage applied to a ballast to achieve a specified current. The test is done by using a low impedance sine wave source (essentially a large power amplifier). Going back to the typical ballast mentioned at the beginning of this section, it is interesting to look at the details of some real ballasts. Figure 6.35 shows three different ballasts suitable for 36–40W
198
fluorescent lamps, together with their principal characteristics. The EEI is explained in Section 6.3.5. In principal all the ballasts have the same Volt/ Ampere characteristic, but their efficiency is different. It is an interesting exercise to feed the values applying to a real ballast into equation 6.2. Here theory and practice show quite a good correlation, indicating that the flux density is around 1T (Wb/ m2). The actual flux distribution within a ballast is shown in Figure 6.36. 6.3.3.2 Practice The reactor ballast has been developed over more than 50 years. While its principles have remained the
ELECTROMAGNETIC COMPONENTS
Fluorescent lamp ballasts are made with a comparatively long coil (Figure 6.38) that is dropped into one half of the lamination stack, and is then enclosed by the other half of the stack. In practice a number of different lamination arrangements are used, and some of these are shown in Figure 6.39. It is clear that not all these are “zero waste” designs, and practical maximum waste figures are less than 10%. The reasons for the different shapes include the following: • the power losses in the transformer steel increase markedly with flux density. Typically the loss is 2.5 times as great at 1.5T than it is at 1.0T. Thus flux distribution is important and can, to some extent, be influenced by the lamination shape and dimensions. • The grain orientation of transformer steel that helps give it its magnetic properties is achieved partly Figure 6.37 The stamping of E and I laminations from strip to achieve minimum waste.
same, manufacturing techniques have changed considerably. The majority are now made on automated production lines to close tolerances. The shape of the ballast is determined partly by the requirements of the luminaire manufacturer, and partly by the demands of automated manufacture. Industry standardization has resulted in ballast crosssectional areas and fixing details being agreed. As already indicated in the previous section, basic theory can only give an indication of how a ballast will perform; so in practice many of the rules for ballast design have been derived from experiment. By constructing ballasts with varying dimensions and magnetic characteristics, and measuring their performance, ballast manufacturers have developed graphs that show how power losses or costs can be minimised. For example, ballasts can be constructed using classic E-I laminations. The manufacturer wants to ensure minimum waste, so stamps the “I” out of the “E” – two ways of doing this are shown in Figure 6.37. In both cases the theoretical waste is only the cut for the airgap. Experimental data shows that there is an optimum value for the width of the limb to give minimum loss.
Figure 6.38 An “exploded view” of ballast construction showing how the coil of a ballast is wound in a separate operation before being inserted into the lamination stack. Diagram from Helvar.
199
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
by the rolling process. This means that the steel has a preferred direction of magnetization, generally in the direction of the original rolling. Clearly some strategems for extracting the maximum material out of the strip result in the optimum magnetization direction being transverse to the field. Thus the manufacturer must either have a lamination punching strategy that results in laminations having the right orientation, or must specify a particular performance from the steel supplier. • In any final ballast construction there must be a method of holding all the laminations together and of maintaining the airgap. This can result in the need
Figure 6.40 Poorly wound coils with thick insulation result in a heavier, less efficient ballast with shorter life (above) compared with the the high copper fill factor ballast (below). Diagram from Helvar.
Figure 6.39 The completed construction (above) of the ballast shown in Figure 6.38 is representative of how ballasts are made. The lamination shapes used by four other manufacturers are shown below. Many aspects of ballast construction are protected by patents.
200
for special cutting to provide a grip, or for matching the housing. • The size of the “window” that accepts the coil is significant. In low loss ballast designs it can be optimal to increase the window size. Low loss designs use longer lamination stacks; and the larger window reduces the loss in an individual lamination. Figure 6.39 also shows other features of electromagnetic reactor ballasts. Some designs require only a baseplate to accommodate the laminations, others require a metal enclosure. In either case the lamination shape and the baseplate or enclosure are designed so that the overall result is one where leakage (or stray) magnetic field is minimized. Ensuring that the airgap dimension is accurately maintained throughout ballast life requires a rigid
ELECTROMAGNETIC COMPONENTS
assembly. Dimensional stability can be ensured by inserting a narrow non-magnetic spacer (for example, aluminum wire). In some designs deliberate deformation of the strip can be used to provide limited “tuning” of the ballast performance. The winding of the coil is done in a way to maximize the copper fill factor in the winding window, with the aim being to achieve up to 60%. The achievable fill factor depends on a number of parameters: • wire diameter (fill factors above 55% can not be achieved with wire diameters less than 0.3mm). • wire insulation thickness. • the thickness of the overall coil insulation. • evenness of winding. A poorly wound coil not only increases copper losses, but also shortens ballast life. Clearly the ballast designer has a wide range of variables to consider, depending on whether the priority is lowest cost or lowest losses. In practice the scope for variation is becoming limited because (a) Ballast for compact fluorescent lamps 4-13W 85mm × 42mm × 28mm (b) Ballast for fluorescent lamps 15-40W 150mm × 42mm × 28mm (c) Ballast for fluorescent lamps 58-100W 190230mm × 42mm × 28-29mm
(d) Ballast for 160W UV suntanning lamp 133mm × 66mm × 53mm
Figure 6.41 Some representative reactor ballasts for 230V fluorescent lamps. Photos from Helvar.
of legislation relating to minimum ballast efficiencies. Given a standard ballast cross-section, the principal variations are: • stack length. • quality of transformer steel. • size of copper wire (the bigger the diameter, the lower the copper losses). In the examples shown in Figure 6.35 the considerable variation in stack length and wire diameter can be seen. The “low loss” ballast in the right hand column uses steel with a loss of less than 4W/kg at 1.5T 50Hz, whereas the ballast in the left hand column uses steel with a loss of 7W/kg at 1.5T 50Hz. Electromagnetic ballasts are usually fitted with screw-less terminals suitable only for solid conductors of a specified size range. This matches the requirements of the luminaire manufacturer who may well be using automated assembly/wiring machines. The screwless terminals may be either of the type that requires the wire to be pre-stripped, or of the insulation displacement type. 6.3.4 Transformer fluorescent lamp ballast construction The construction of fluorescent lamp ballasts for the 120V market that use the circuits (d), (e) and (f) of Figure 6.31 is significantly different than that of the 230V reactor ballast. More materials are required, and the final product does not lend itself so well to automated production. Because of the move to electronic ballasts it is unlikely that significant further developments will take place that could improve this situation. A typical construction is shown in Figure 6.43. the primary and secondary of the transformer are wound on separate limbs, and the complete transformer assembly is “potted” into a metal can. The resulting ballast is significantly bigger than its European counterpart, however, the ballast usually has the power factor correction capacitor (and any auxiliary starting capacitor) mounted within the can, and most ballasts serve two tubes. Another significant difference from European practice is that the ballast comes with attached color
201
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
coded leads. In addition, in order to comply with UL requirements, most ballasts are fitted with thermal protection to ensure that they cut out under any overtemperature condition. 6.3.5 Fluorescent lamp ballast ratings Data sheets on electromagnetic ballasts give a number of parameters that indicate ballast performance. Nomenclature for some of these is different in the USA and Europe because of the different supply voltages, technical vocabulary and legislation. International treaties have committed the major economies to reducing carbon dioxide (CO 2 ) emissions. Since electricity generation is responsible for about 30% of CO2 emissions, efforts are being made to improve the efficiency with which we use electricity. In Europe a significant parameter is the Energy Efficiency Index or EEI. Because there are comparatively few manufacturers of electromagnetic ballasts for fluorescent lamps, the ballast became an easy target for legislation. EEC Directives have been
Figure 6.42 Reactor magnetic ballasts are made on highly automated production lines. These can monitor the electrical and magnetic properties of each ballast as it is assembled to ensure that it meets specification. The Helvar production lines have a capacity of one ballast every second.
202
Figure 6.43 Construction of a transformer ballast for fluorescent lamps.
introduced that specify that inefficient ballasts must be phased out of production. The problem then was to define what constituted an efficient ballast. The trade body CELMA (Committee of European Union Luminaire Manufacturers Associations) has a sub-committee devoted to components. In order to protect the interests of both members and customers; and to have an objective and agreed method of measuring whether or not the EEC directives were being complied with, CELMA introduced the EEI. The idea behind the EEI is that it is objective and does not depend on any particular technology. It starts with the parameter Ballast Lumen Factor or BLF. A ballast with BLF = 1 is a ballast that ensures that a lamp gives out its design lumens at an input power that matches the power used by the lamp’s standard reference ballast.
Figure 6.44 Typical magnetic ballasts supplied to the USA market. Examples from Howard Industries.
ELECTROMAGNETIC COMPONENTS
The EEI grades ballasts into categories: D magnetic ballast with very high losses C magnetic ballast with moderate losses B2 magnetic ballast with low losses B1 magnetic ballast with very low losses A3 electronic ballast A2 electronic ballast with reduced losses Each class is defined by the power input needed to achieve the light output corresponding to BLF = 1 for high frequency ballasts and BLF = 0.95 for electromagnetic ballasts. Table 6.3 shows the EEI classification applicable to some common fluorescent lamps. It can be seen from the table that the A2 classification is very close to the reference ballast. There is also an A1 classification; this has now been reserved for dimmable or controllable ballasts. They must have at least the performance of A3 ballasts at 100% light output; but must also use no more than 50% of A3 power when the light output is set at 25%. Further they must be able to reduce the light output to 10% or less of maximum. It is both practically and economically difficult to produce electromagnetic ballasts with an A classification. Within the EEC the situation as at 2002 is as follows: • D classification ballasts may not be sold as from 2002, and C classification ballasts may not be sold after 2005. • B1 and B2 ballasts are permitted, however in 2005 there may be a review of the progress made at
Lamp Types T TC-L TC-F T TC-L TC-F T T TC-LE TC-D TC-DE TC-D TC-DE TC-D TC-T TC-DD
Power at 50Hz 18W 36W 58W 70W 10W 13W 26W 28W
Power at HF 16W 32W 50W 60W 40W 9.5W 12.5W 24W 25W
improving the energy efficiency of ballasts, and this could have implications for these relatively efficient ballasts. In the USA Ballast Factor has essentially the same meaning as BLF. It is expressed either as a percentage or as a decimal figure (e.g. 95% or 0.95). However, federal legislation specifies a minimum Ballast Efficacy Factor or BEF for ballasts; the aim being to achieve a minimum efficacy. The BEF is the Ballast Factor expressed in percent divided by the total circuit input power in watts. For example a ballast for a nominal 40W lamp might have a ballast factor of 93%, but have circuit watts of 49W. Its BEF would, therefore, be 93/49 = 1.898. USA and Canadian standards for common 4ft and 8ft tubes are shown in Table 6.4. In addition to federal standards for BEF, there is pressure from utilities and state authorities for high efficacy. Note that comparisons of BEF are only valid when applied to ballasts operating the same number and type of lamps. New legislation may make the 120V electromagnetic ballast for conventional fluorescent lamps obsolete by 2010. In the USA the equivalent of CELMA is CBM, or Certified Ballast Manufacturers. A CBM certified ballast is one that has been tested by an independent test house to conform with ANSI (American National Standards Institution) standards relating to fluroescent lamps, and which is subject to spot retests in manufacture.
A2
A3
B1
B2
C
D
d19W d36W d55W d68W d45W d11W d14W d27W d29W
d21W d38W d59W d72W d48W d13W d16W d29W d31W
d24W d41W d64W d77W
d26W d43W d67W d80W
d28W d45W d70W d83W
>28W >45W >70W >83W
d14W d17W d32W 34Wd
d16W d19W d34W d36W
d18W 21Wd d36W d38W
>18W >21W >36W 38W
Table 6.3 Examples of the EEI applying to common fluorescent lamps. (N.B. This is only a partial list.) “Power” is the lamp power. See Table 4.2 for illustration of lamp types.
203
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Lamp T12 40W 2 u 40W 2 u 40W 2 u 75W 2 u 110W
Circuit Voltage 120, 277 120 277 120, 277 120, 277
Nominal Power 40W 80W 80W 150W 220W
Minimum BEF 1.805 1.060 1.050 0.570 0.390
Table 6.4 USA minimum Ballast Efficiency Factors for T12 tubes at 2000. By 2005 most T12 lamps will be required to have efficacies that can only be realized with electronic ballasts.
Apart from detailing the lamp(s) for which the ballast is intended to be used, the ballast data sheet will give the following additional information: Power factor (and, if the ballast is a low power factor unit, the required correction capacitor value.) Power loss at a specified temperature (and/or the EEI or BEF). Minimum starting temperature. This is largely determined by the lamp; but the starterless circuits may themselves contribute to a minimum practical starting temperature. Maximum winding temperature tw. This is normally specified as 130°C (see Section 6.1.6.2) for a rated ballast life of 10 years. Temperature rise Δt. This is given as two values. The first is the temperature rise under normal working conditions. The second is the rise under abnormal conditions; and for a reactor ballast these would be the simultaneous short circuit of the starter (so, effectively, the ballast would be across the mains supply) and a 10% rise in line voltage. Typical values might be 35°C normal temperature rise and 95°C fault temperature rise. It is clearly important that the total of ambient temperature and temperature rise do not exceed t w . In some cases a higher “normal” temperature rise is given for a ballast used with a series capacitor. In Europe ballast temperature rise is considered sufficiently important as to be the subject of standards (EN60920). The standard specifies the method of measuring the temperature rise, and calls for the measurements to be made under stable draught-free conditions with the ballast being mounted on wooden blocks.
204
Ballasts can be the cause of audible noise, primarily due to the stray field affecting nearby metalwork, but also due to the ballast construction (especially in the case of transformer ballasts). USA manufacturers ascribe a sound rating to ballasts, but this is a relative rating based on subjective criteria. An “A” rated ballast is suitable for quiet environments.
6.4 Ballasts for HID and arc lamps 6.4.1 Introduction The considerations for the construction of electromagnetic ballasts for fluorescent lamps apply equally to ballasts for HID lamps. The main differences arise for the following reasons. While ballasts for low power HID lamps are similar in size to fluorescent lamp ballasts of the same power rating, the ballasts needed for lamps from 150W up to 1,000W are significantly bigger. The market for luminaires using HID lamps is, in quantity terms, significantly smaller than that for fluorescent luminaires. HID ballasts are, therefore, manufactured in smaller quantities and the methods of manufacture are less automated. Metal halide and high pressure sodium lamps require the use of a high voltage ignitor. Some ignition circuits require that part of the main choke winding is rated to handle the high starting voltage. This feature is, however, only needed when the particular ignition method is being used. There tends to be a wider range of options in respect of fixing methods, cross section, terminal arrangements etc. This is because the application of HID lamps is varied, and luminaire shapes and sizes have to match this variety. Often the HID ballast is fitted in a separate housing which greatly simplifies matters, but there are practical limits on the distance between lamp and ballast because of the ignitor. 6.4.2 European practice With the exception of higher rated low pressure sodium lamps (above 100W), the line voltage in Europe is sufficient to allow the use of simple reactor
ELECTROMAGNETIC COMPONENTS
35W ballast for metal halide lamps
50-70Wballast for metal halide lamps 114mm × 54mm × 46mm
125W ballast for mercury vapour HID lamp 135mm × 70mm × 85mm
150W ballast for HPS lamps 135mm × 70mm × 85mm
400W ballast for HPS lamps 165mm × 70mm × 85mm
high power HID ballast (1,000– 2,000W) for HPS and metal halide lamps 220–330mm × 116mm × 98mm
constant wattage transformer for low pressure sodium SOX lamps
Figure 6.45 Representative HID ballasts available on the European market. Photos from Philips (high power and SOX ballasts) and Helvar.
ballasts for HID lamps. The ballasts for High Pressure Mercury, High Pressure Sodium and Metal Halide discharge lamps are all physically similar, but are designed to have the correct Volt/Ampere characteristic for the lamp concerned. All but the very highest ratings of ballast run on the single phase 230V supply. The highest rating of metal halide lamp used in general applications is 2,000W, and the ballast for this may either only be available, or available as an option, to operate across phases. In practice this means a choke with 380/400/ 415V tappings. Mercury vapor and metal halide lamps have a constant voltage characteristic, which to some extent means that their operation is not affected by line voltage. However, high pressure sodium lamps have an arc voltage that depends on arc power. For this reason lamps should be operated at within 5% of the rated lamp voltage. Within Europe the line voltage is supposed to have been standardized at 230V, but historically the common line voltages were 220V and 240V, and in reality these voltages are often found. Tapped chokes are available for all types of HID lamps, but they are of particular importance for high pressure sodium. An HID lamp circuit takes a varying current, and there are some potentially dangerous fault condi tions. Therefore circuit protection needs some consideration. When the lamp is switched on there may be a significant inrush arising from a low impedance supply feeding the power factor correction capacitors. This means that the circuit protection must not be so fast as to nuisance trip. While the lamp is reaching stable operating conditions it can take a higher current than the running current so both the choke and its protection must be designed to cope with this. Of particular significance is the rectification effect. A discharge lamp can act like a partial rectifier if the current going one way through the lamp is different than that going the other way, and such a condition can arise during warm up when the electrodes may be at different temperatures and may be emitting electrons at a different rate. This can result in some DC in the choke causing saturation and a much higher current. Generally the effect is transitory
205
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
6.4.3 USA practice
Figure 6.46 Safe area of operation for a high pressure sodium lamp. Unlike mercury vapor based lamps, the arc voltage varies considerably with power.
during start up, but a faulty lamp can exhibit a continuous rectification effect and in such cases the circuit must be shut down quickly. Control gear and lamp manufacturers make recommendations as to the protection of lamp circuits. Where it is not practical to fit a fuse or MCB for each lamp, it is sometimes recommended that the ballast be fitted with thermal protection.
As with conventional fluorescent lamps, USA magnetic ballast practice for HID lamps is, in the main, significantly different to European practice. On the 277V supply it is quite practical to use reactor ballasts and these are offered by some manufacturers as “Energy saving” HID ballasts. In fact it is correct that power losses are less in a reactor ballast than in a transformer ballast. However, in the USA the more widespread requirement is for multivoltage ballasts with good regulation to allow for variations in supply voltage. The most popular HID ballast is the constant wattage autotransformer or CWA. Because of the use of autotransformer technique, the circuit results in the smallest and lightest ballast. The combination of leakage reactance construction and the use of a ballasting capacitor results in both good regulation (approximately 5% change in light output for 10% change in line voltage) and high power factor. The constant power characteristic ensures that starting and running currents are similar, and the presence of the series capacitor limits any heating effect arising from rectification effect. The transformer construction lends itself to accommodating different supply voltages, and most
Figure 6.47 Some HID ballast circuits used in the USA. (a) is the standard reactor circuit also used in Europe. (b) is the most popular CWA circuit. (c) is a high reactance autotransformer circuit where the capacitor is for power factor correction only. (d) shows a method of supplying multiple high pressure mercury lamps.
206
ELECTROMAGNETIC COMPONENTS
ballasts sold to the open market are “quad” ballasts suitable for 120, 208, 240 and 277V supplies. Some ballasts are deliberately designed to have a high crest factor (>1.6 compared with 1.4 for a simple reactor) in order to assist lamp starting. This is achieved by cutting slots in the transformer core under the secondary. However, the lamp electrodes are affected by a high crest factor, and the lamp manufacturer does specify a limit. For high pressure sodium lamps, ballasts with better regulation than achieved by the CWA may be offered. One variant uses isolated secondary windings, with a separate winding for the capacitor. This achieves tight regulation, but at the expense of increased losses and greater weight. USA magnetic ballast manufacturers offer “kits” to luminaire manufacturers, with many options. In addition to the standard ballasts mentioned here, other options include multi-lamp ballasts (actually 2-lamp lead-lag configurations) and the series feeding of high pressure mercury lamps through current limiting transformers. As with the fluorescent lamp ballasts, HID ballasts are usually supplied with flying leads.
starting switch, and this is generally provided by the glow starter already described in Section 3.4. The glow starter does have some disadvantages; one being that its life may be no longer than that of the tube. The need for the glow to be established delays the application of the pre-heat current, and the breaking of the circuit is random timed. This means first, that the starting pulse is of indeterminate amplitude, depending on when in the half cycle the break is made, and second, that a start attempt may be made before the cathodes have reached operating temperature. In turn this means that if the starting voltage is too low or early, the tube does not strike, and needless damage may be done to the cathode coating, shortening the life of the tube. A solution to this problem is the use of an electronic starter. This device can do several things: • by being active as soon as power is applied, the starting delay is reduced, since there is no equivalent of waiting for the glow discharge to warm up the bimetallic strip. • the pre-heat time can be precisely timed (usually two seconds).
6.5 Ignitors and starters 6.5.1 Starters for fluorescent lamps Fluorescent lamps using the pre-heat start circuits of Figure 6.31 (a), (b) and (c) require an automatic
Figure 6.48 Examples of CWA (left) and reactor (right) “core and coil” ballasts offered on the USA market. Photo from Howard Industries.
Figure 6.49 Electronic starter circuit using the Power Innovations Fluoractor™. The diode bridge ensures the device works for both half cycles. When power is applied the gate voltage delivered by R1 causes the device to switch on, starting cathode pre-heat. Cathode current causes a voltage to be developed across D6 and D7 which starts to charge the timing capacitor. When peak voltage at TH2 gate is sufficient, TH2 fires and shuts off TH1 creating the starting pulse. In fact a series of pulses is produced until the lamp starts conducting. Once the lamp is conducting the input voltage drops to a point where it is insufficient to trigger TH1. Should the lamp fail to strike, a point is reached where the capacitor is fully charged, and at this point TH2 is held hard on, and, therefore TH1 is held off. Circuit © Power Innovations Ltd
207
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
• the application of the starting pulse can be optimized in relation to the half cycle, so that the starting pulse is of the right energy to start the tube without damaging the cathodes. This also reduces electromagnetic interference. • if a tube fails to start after a set number of attempts (around 10) the luminaire is shut down. This avoids repeated start attempts on a failed tube which can result in ballast overheating. Electronic starters can be triac, thyristor or transistor based. The company Power Innovations have developed a special GTO device for the purpose called the Fluoractor® which combines in one package crowbar protection (and the ability to absorb choke energy when a lamp fails to strike) and thyristor switching with gate turn off control. A circuit showing its use is shown in Figure 6.49. Electronic starters are housed in the same housing as glow starters and are plug-in replacements. See Figure 6.50. While it is recommended that fluorescent lighting that is to be frequently switched (e.g. by occupancy sensors) use electronic ballasts, this cannot always be justified economically, especially if a retro-fit is involved. In such cases the electronic starter is a good investment, its small extra cost being quickly repaid by extended tube life. Its own life is at least that of the ballast. 6.5.2 Ignitors for HID lamps Ignitors for HID lamps are required to generate a high voltage pulse, or series of pulses, to initiate the discharge. In electromagnetic ballast HID circuits
Figure 6.50 An electronic starter for fluorescent lamps; it is a plug-in replacement for the conventional glow starter. Photo from Auralight AB.
208
Figure 6.51 The common ignitor arrangements for HID lamps. Parallel ignitor (top), impulser ignitor (center) and superimposed pulse ignitor (bottom).
there are three main types of ignitor: • parallel • impulser • superimposed pulse The connection arrangements are shown in Figure 6.51. The figure also shows the nature of the starting pulse provided by each arrangement. The parallel starter works by placing a capacitor in series with the ballast inductor, switched into circuit by a thyristor. The high capacitor voltage resulting from the LC combination is then discharged into the lamp. But “high” voltage is relative, and this type of starter is only suitable for small metal halide lamps, and small low pressure sodium lamps using a reactor ballast in place of the more common transformer ballast. The tendency for small metal halide lamps to be operated from electronic ballasts (or to use superimposed pulse start) means that the parallel ignitor is the least used of the three techniques. The impulser ignitor uses a similar capacitor arrangement to generate the initial pulse, but greatly
ELECTROMAGNETIC COMPONENTS
Figure 6.53 Outline circuit of the superimposed pulse ignitor. Figure 6.52 Example of superimposed pulse ignitor. This one is suitable for HPS lamps of 100–400W and metal halide lamps of 70–400W.
increases the voltage by using part of the ballast winding as an autotransformer to boost the pulse to the kilovolts required. The advantages of this system is that the ignitor has virtually no losses when the lamp is running, and the lamp can be at some distance from the ballast. The disadvantage is that the ballast choke must have a suitable tapping and must be able to withstand the high starting voltage – both factors tending to increase the cost and complexity of the ballast. The superimposed pulse ignitor is the most widely used. Its disadvantage is that it is permanently in circuit, so does contribute to circuit losses. Such losses are around 1W for a 70W lamp ignitor, 2.5W for a 250W lamp ignitor and 5W for a 1,000W lamp ignitor. Its great advantage is that it is independent of the ballast construction, and the ballast is not subjected to the starting voltage. As can be seen from the figure, the superimposed pulse ignitor superimposes a number (between two and six) of pulses on each half of the sine wave. These are precisely timed to occur at specified phase angles, for example 60–90° on the positive going cycle and 240–270° on the negative going. The impulse amplitude is between 2kV and 5kV depending on lamp type, and typical pulse width is 1.5μS. Figure 6.53 shows the circuit of a typical ignitor. Here a SIDAC or equivalent device is used to discharge the capacitor. The SIDAC is not unlike a
diac in that it has an avalanche breakdown at a specified voltage, but the voltage, around 130V, is higher than that of a diac. This creates the necessary di/dt in the primary to create the high voltage impulse in the secondary. When an HID lamp reaches end of life, a condition can occur where the arc voltage increases as the discharge is established. This results in the lamp coming on, but quickly going out as the arc voltage exceeds the available instantaneous voltage at each half cycle. The result is a continuous attempt at ignition which is both subjectively irritating and stressful to the lamp circuit. For this reason most superimposed pulse and impulser ignitors are available with an automatic cut-out feature that shuts down the ignitor after a preset period (between 2 and 15 minutes depending on lamp type). 6.5.3 Hot re-strike ignitors; auxiliary switches Where it is possible, hot re-strike of HID lamps requires an ignitor voltage of tens of kilovolts. Xenon (DC) arc lamps require a similar voltage for striking. Ignitors for this duty require an alternative construction to achieve a high enough di/dt in the primary of the pulse circuit. Sometimes a spark gap is used for this purpose. A manually operated hot restrike ignitor circuit is shown in Figure 6.54. However, it is now the case that those applications requiring hot re-strike are also the ones requiring lightweight ballasts. The majority of such applications are now served by electronic ballasts.
209
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 6.54 Manually operated hot re-strike circuit using a spark gap.
There are some applications where the delay in reaching full light output from an HID lamp is unacceptable. For example a secure warehouse with HS lamps might require some immediate lighting when power is applied. This problem is dealt with by having an auxiliary tungsten halogen lamp that is switched on when power is applied to the luminaire, but which switches off after 3–5 minutes when the main lamp has reached full brightness. Automatic auxiliary switches are available for this purpose (Figure 6.55.)
6.6 Lighting control by transformers and ballasts 6.6.1 Summary Variable autotransformers and variable reactors can be used for dimming lighting, and these items are discussed in Section 8.2. Variable autotransformers, multi-tapped autotransformers, and multi-tapped electromagnetic HID ballasts are used to regulate the power available to lighting circuits or luminaires. Therefore they can form the basis of energy management schemes, although care must be taken to ensure that the lighting performance is not compromised to the point at which it is ineffective. These devices are discussed both in Section 14.3 and in the relevant applications sections.
210
Figure 6.55 Auxiliary lamp switches to provide illumination during HID lamp run-up. The one on the left uses hybrid relay and solid state technology, the one on the right is triac based. Photo courtesy of Venture Lighting International.
6.6.2 Dimming of electromagnetic ballasts and transformers Generally the suitability for dimming of particular lamps, ballasts or transformers is covered in the sections dealing with the lamp or the ballast concerned, and is also covered, where relevant, in the applications chapters. Here “suitability for dimming” is taken to mean the connection of the device to conventional thyristor or triac dimmers. As a summary: • electromagnetic ballasts and transformers should not be connected to electronic dimmers unless approved by the transformer/ballast manufacturer. • approved low voltage transformers for incandescent (tungsten, tunsgten halogen) lighting, and high voltage transformers for cold cathode lighting are suitable for dimming. • some HID lamps with reactor or CWA ballasts are suitable for partial dimming. It should be noted that when phase cutting dimmers are used, there is a considerable increase in odd harmonics compared to the harmonics generated at 100% output . For fluorescent lamps today’s trend is towards the use of electronic ballasts when lighting level control is required. However, it is possible to dim T12 fluorescent lamps using electromagnetic ballasts, and the preferred circuit for doing so is shown in Figure 6.56. The circuit is a variation of the rapid start circuit. Instead of having a cathode heating winding on the
ELECTROMAGNETIC COMPONENTS
main ballast, there is a separate cathode heating transformer that maintains full cathode heating at all times the circuit is on. The main tube voltage is varied by the thyristor dimmer, the sharp rising edge of the thyristor switching helping to develop the strike voltage needed. Much lower light levels can be achieved if this is augmented, and one way of doing this is to use a tapped choke with a small capacitor (known to some as the “tickling” capacitor). In any circuit involving the dimming of electromagnetic ballasts it is essential that power factor correction components are installed prior to the dimmer. Connecting power factor correction capacitors to the output of dimmers is likely to destroy both the dimmer and the capacitor. Large installations using dimmed electromagnetic ballasts are candidates for central variable power factor correction (see Section 6.7.3). A possible exception to this rule is where the dimming device has a sine wave output. In practice this means either a variable autotransformer or a sine wave output electronic dimmer (see Section 8.6.1). Specialist suppliers offer systems of this kind for controlling conventional HID lamp circuits to achieve some energy savings when full light output is not required.
6.7 Power factor correction 6.7.1 Introduction All discharge lighting that uses a reactor ballast or leakage reactance transformer to limit the current through the lamp results in a poor power factor, typically as low as cos φ = 0.3–0.5. There are some fluorescent lamp circuits that use a series capacitor to provide a current limiting reactance, and these circuits can have an acceptable power factor. In these circuits the capacitors must have a precise value (within ± 4% or better) to ensure correct lamp current, and because high voltages are developed in these circuits, they need to have a voltage rating considerably higher than line voltage. Apart from this special case, all electromagnetic ballasted lamps above around 30W are fitted with
Figure 6.56 Circuit for the dimming of some types of fluorescent lamps using reactor ballasts. The capacitor is optional; it improves low light level performance.
parallel connected capacitors to provide power factor correction to 0.85 or better. The capacitors are normally fitted within the luminaire alongside the choke; however it is also possible to apply central correction to a large group of lamps. Power factor correction is also referred to as compensation. For electromagnetic ballasts the correction relates to cos φ also known as displacement power factor. With electronic ballasts that generate harmonics the correction requirement is more complicated and is described in the next chapter. 6.7.2 Power factor correction for individual fluorescent and HID lamps. In the standard fluorescent lamp circuit of Figure 6.30 the current i is out of phase with the supply voltage Vs. In Figure 6.57 the circuit is repeated, but this time showing the supply current i having two components, iL going through the lamp, and iC going through a parallel connected power factor correction capacitor. The vector diagram in Figure 6.57 shows as an example φ being 70° (power factor 0.34) for the uncorrected circuit, but only 25° (power factor 0.9) for the corrected circuit. Applying simple trigonometry to the diagram, and using the principles explained in Section 1.4, a formula can be derived for the capacitance in microfarads required to achieve a given power factor. This is also shown in Figure 6.57. The manufacturers of electromagnetic ballasts specify the capacitor to be used to achieve an
211
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
overpressure pushes the layers apart leaving an insulated area, and the capacitor continues normally, with a minute reduction in capacitance as the result of the loss of some electrode area. Depending on the size and the requirements of the luminaire manufacturer, connection to the capacitor is by fixed solid conductor leads, push on terminals or other proprietary connection system. Mounting is by spring clip, molded-on plastic feet, or a stud on the end of the capacitor. Examples of construction are shown in Figure 1.33. The specification of a capacitor for power factor correction in luminaires covers the following: • Line voltage. • Test voltage (typically twice the nominal line voltage for 60 seconds). • Insulation (typically 2kV for 60 seconds for a 250V AC rated capacitor). • Operating frequency (normally only 50/60Hz). • Working temperature range (typical -25°C to 85°C). Lamp Type
Figure 6.57 Power factor correction in a fluorescent lamp circuit. The formula shown here gives the capacitance in microfarads.
acceptable power factor. Table 6.5 lists some examples, and shows clearly the big difference in line current that is involved. This in turn explains why power factor correction is usually applied near the lamp, since central correction would require the conductors to the lamps to be of almost double the current carrying capacity. The comparatively small capacitors needed for correction at the luminaire are normally cylindrical capacitors in metal or self extinguishing plastic cases. The capacitor itself is made from metallized polypropylene film that has self healing characteristics. If there is a voltage breakdown in the capacitor the metal layers around the breakdown channel evaporate due to the high temperature of the arc that is formed. Within microseconds the resulting
212
TC 18W T 18W T8 T 36W T8 T 58W T8 T 85W T12 T 100W T12 T 140W T12 HM 80W HM 250W HS 50W HS 100W HS 150W HS 250W HS 400W HS 1000W HI 35W HI 70W HI 150W HI 400W
Lamp current A
Line current A
Capacitor PF
0.20 0.37 0.43 0.67 0.80 0.96 1.5 0.80 2.15 0.75 1.2 1.8 3.0 4.4 10.3 0.53 1.0 1.8 3.5
0.11 0.16 0.22 0.35 0.43 0.55 0.77 0.45 1.4 0.33 0.56 1.1 1.4 2.2 5.0 0.27 0.43 1.1 2.2
2 4 4 7 8 10 18 8 18 8 12 20 32 50 100 6 12 20 35
Table 6.5 Approximate capacitor values for achieving 0.9 power factor on common fluorescent and HID lamps on 230V 50Hz supply.
ELECTROMAGNETIC COMPONENTS
Figure 6.58 Simplified equivalent circuit of a real capacitor, with losses shown as a parallel resistance. The tangent of angle δ is the loss factor.
• Capacitance tolerance (typical ±10%). • Loss factor. The loss factor of a capacitor is a measure of its imperfection. All capacitors have losses analogous to those in inductors. There is a leakage loss since no solid insulator is perfect, and there is a dielectric heat loss arising from the continual change in current direction – analagous to hysteresis in a magnetic medium. Both together can be considered as a resistance in parallel with the “pure” capacitance, and the loss factor is an expression that relates the resistive to the capacitive impedance. Figure 6.58 shows how a real capacitor does not have the 90° phase shift of a “perfect” capacitor, but has one that is less by a small angle δ. The tangent of this angle is the loss factor, and in a typical lighting capacitor the value is around 0.00003. For clarity the
figure exaggerates δ which is very much less than one degree. There are some safety issues with capacitors. The first is that if power is disconnected from the luminaire, a capacitor may still have a substantial charge, presenting a potentially lethal shock hazard. Where the lamp circuit itself does not provide a discharge path it is usual to fit the capacitor with a resistance that ensures discharge. The resistors are typically 1MΩ for capacitors of less than 20μF and 470kΩ for 20–50μF. The idea is to ensure that the voltage drops to a safe level, e.g. 50V, within 60 seconds. While this arrangement is sufficient for permanently installed luminaires, it is not for luminaires that can be removed by a simple unplugging action (e.g. track mounted). Here the person removing the luminaire could be exposed to shock at the power connector. Under EN60598-1 such luminaires are required to have a discharge method that reduces the voltage to 34V within one second. The resistor required would be such that a lot of heat would be dissipated, and two solutions to the problem are: • either the use of a small reactor choke across the capacitor. This has low dissipation when the AC supply is present, but acts as a low resistance for the capacitor discharge. • or the use of an electronic switch that discharges the capacitor through a low resistance when the AC power is removed. This method has lower losses (<0.5W) and the component is much lighter weight than the reactor alternative. Severe over-voltage, excessive temperature, or excessive rate of change of voltage (dv/dt) can cause a capacitor to explode due to evaporation of the dielectric and failure of the self healing mechanism. Capacitors in aluminum cans can be supplied with an automatic disconnection system that operates in the event of over-pressure. The principle is shown in Figure 6.59. The can is designed with a crimped section that can expand without bursting in the event of excessive internal pressure. One connection to the capacitor electrodes is made with a weakened section that snaps as the can expands.
213
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
6.7.3 Bulk power factor correction It is unusual to install bulk power factor correction equipment for lighting only since, as already mentioned most luminaires fitted with reactor ballasts are also fitted with the necessary power factor correction capacitor. However, in a large installation there can still be a significant overall lagging power factor, especially if a lot of compact fluorescent lamp ballasts without correction capacitors, or low voltage transformers, have been installed. Within a large building there will probably be other devices, especially electric motors, that will contribute to poor power factor. In such cases central correction is used. Often the power factor varies according to load, and it is possible to install equipment that monitors power factor continuously and switches in the needed reactance in steps. In principle the capacitors are larger versions of the ones used in luminaires. The bigger capacitors use metallized paper electrodes and polypropylene dielectric, all vacuum impregnated with a non toxic mineral oil. Capacitor banks for this application are normally specified in terms of their reactive load kVA(r) and are supplied as three-phase assemblies. Losses are remarkably low, typically as little as 0.4W per kVA(r) at the capacitor. The engineering of bulk power factor correction is a job for the specialist because there are some hazards associated with it.
Figure 6.59 Over-pressure safety mechanism in a power factor correction capacitor. When the can expands, the link is broken. Diagram from Electronicon Kondensatoren GmbH.
214
Figure 6.60 Central power factor correction equipment from Comar ranging from 10 to 1,000 kVA(r). The racks are fitted with automatic power factor measurement that switches in the required reactance in up to eight separate steps.
The switching of large capacitors results in huge currents, and massive dv/dt. For this reason the larger capacitors are fitted with a series choke coil to limit dv/dt. There is the possibility of the capacitor setting up a resonance with inductance in the supply network. For this reason de-tuning inductors are added to the circuit. The presence of harmonics poses a real problem. The reduction in amplitude of the fundamental current by power factor correction means that the harmonic content expressed as a percentage actually rises. This can again mean that dv/dt limits are exceeded. While detuning inductors help alleviate the problem, it is often necessary to consider the problems of harmonic rejection and power factor correction together.
Figure 6.61 Components from Electronicon Kondensatoren Gmbh for building central power factor correction equipment. On the right 14.1 kVA(r) three phase capacitors. On the left a three phase de-tuning reactor.
Chapter 7
ELECTRONIC COMPONENTS
Electronic
7.1 Circuit elements The voltage transformation and current regulating tasks of electromagnetic components can now be carried out electronically. This has a number of advantages: • for many applications, especially fluorescent lamp ballasts, the electronic solution results in higher efficiency. • electronic control can provide better regulation and better lamp starting; both can result in longer lamp life and better lumen maintenance throughout life. • besides these economic advantages, the electronic ballast or transfomer can be built in many different shapes or form factors. Whereas magnetic components must conform to certain shapes to allow automated manufacture and to achieve magnetic circuits of practical dimensions, the electronic equivalent is not so restricted. This helps the luminaire designer. • the reason for the flexibility in form factor is that the majority of electronic lighting components operate at high frequency; which in turn means that any transformer or inductor is very small compared to its 50/60Hz counterpart. • in the case of fluorescent lamps the operation of the lamp at high frequency eliminates the stroboscopic effect of lamps running at 50/60Hz (which can result in light pulses at 100/120Hz). Electronic lighting components are “power conversion” devices that turn an incoming AC or DC supply voltage into the voltage required by the lamp. They are based on a number of standard “building blocks”, the principal ones are now reviewed. Rectifier. Unless the power source is itself DC (e.g. a vehicle battery) all electronic transformers
components
and ballasts start by converting the incoming AC supply to DC. This fact is used in some emergency lighting systems, since many electronic ballasts work equally well from AC or DC. Rectification is normally done using a conventional diode bridge rectifier (see Section 2.1.2). From Figures 2.3 and 2.4 it can be seen that if a smoothing capacitor is used the average DC voltage can be quite close to the AC peak. A possible problem can also be identified. The capacitor only receives charge for a small part of the cycle, since it can only charge when the supply voltage exceeds the capacitor voltage. In practice this can mean that the circuit only takes current from the supply for about 30% of the time. Figure 7.1 shows what can happen. While this could be acceptable in low power equipment, the resulting harmonics on the supply for any significant load could be unacceptable. Matters are made worse if there is significant impedance in the supply, because the current peaks can then distort the voltage waveform. Transistor. Throughout this chapter reference is made to the use of transistors as high speed electronic switches. It should be understood that, were this book be an engineering textbook, it would show how some of the circuit elements can also be realized using other power devices – especially thyristors. However, thyristor circuits are more suitable for higher power applications, and most electronic ballasts and transformers for lighting use transistors as the switching devices. DC converter. Rectified AC mains produces a DC voltage dependent on the r.m.s. value of the incoming mains, and on whether or not capacitor smoothing is present. But this may not be the voltage required. By using an inductor and a transistor switch it is possible to convert the rectified DC to a
215
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Rectified AC IN
230VRMS
C
Control circuit
VOUT
LOAD
L
IIN
150 W 220 PF
4
400 Current
0
0
400
2
0
-2
Voltage 0.005
I (A)
U (V)
200
200
Figure 7.2 Principle of the boost converter to derive a higher DC output voltage than can be achieved by the simple rectifier.
0.01 0.015 t (s)
-4 0.02
Figure 7.1 The non linear current waveform arising from the simple bridge rectifier circuit.
higher voltage. The principle of the boost converter is shown in Figure 7.2. The full wave rectified DC (without a smoothing capacitor) is applied to an inductor. If the transistor is switched on, the inductor is directly across the supply, and current flows in it. Then, if the transistor is switched off, the voltage across the inductor is reversed (Lenz’s law, see Section 1.2.5) having the effect of adding its voltage to the supply voltage. The capacitor is thus charged to the higher combined voltage. The diode prevents the storage capacitor discharging back through the transistor. An analysis of what is happening in the boost converter circuit shows that the output voltage can be controlled by the pulse width of the switching. As an approximation: Output Voltage = Supply Voltage × (toff + ton)/toff Where toff and ton are, respectively, the on and off switching times of the transistor. A practical boost converter circuit requires some
216
sophistication in the transistor control circuit. In particular there should be a signal from the supply circuit to ensure that the switching follows the supply voltage waveform. More sophisticated converters also have feedback from the output side to ensure that the output voltage is regulated. There are several other converters. The buck converter, as its name implies, reduces the output voltage. The flyback converter can be configured to boost or buck the supply. All rely on the switching of an inductor and the use of a diode to direct the inductor’s back e.m.f. An Inverter is a device for converting DC to AC. Probably the simplest of these is the DC ChopLoad IL
A
B
DC Supply
I C1
TR 1
C
D1
VDC
I D1
I C2
D2 I D2
Transistor Driver Circuit
Figure 7.3 The push-pull inverter.
TR 2
ELECTRONIC COMPONENTS
Transformer primary voltage
I D1
I C1 TR 1
D1
VDE IL
Transistor Driver Circuit
Load VDE
Load current
D2
TR 2
IL
I D2
I CI
Figure 7.5 The half-bridge inverter.
I C2 Transistor collector current
I DI
I D2 Diode current
t0
t1
t2
I C2
t3
Figure 7.4 Inverter waveforms, showing which element of the circuit is responsible for carrying the load current.
per, also called the Push-Pull Inverter. This uses two switching transistors and a transformer with a center tapped primary. Figure 7.3 shows the arrangement with the load connected to the secondary of the transformer. Figure 7.4 shows the circuit action. At time t0 transistor TR1 switches on; Terminal A of the transformer goes negative with respect to B, and, assuming no switching losses, the full supply voltage V is developed across AC. Due to the inductive load, current is still flowing from the previous cycle and it does so through diode D1 until the current changes direction. At this
time t1 the transistor TR1 takes over conduction as collector current IC1 . At time t2 the transistor drive is inverted. TR1 is switched off and TR2 is switched on. This causes a reversal in the load voltage (since BC now has the full supply voltage across it) but a continuing current through D2 until it reverses at time t3 when TR2 takes over. The effect of one of the transistors switching on is to cause a voltage equal to 2VDC to appear at the
TR 1 G1
TR 3 D1
VDC G2
TR 2
D3
Transistor Driver Circuit
Load D2
G1
D4 TR 4
G2
Figure 7.6 The (full) bridge inverter. The transistors on opposite sides of the bridge conduct in pairs.
217
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
TR1
D2
TR2
D4
TR4
D3
TR3
Instantaneous Load Voltage
Mean Load Voltage
t0 t3
t2
t1
Mean Load Current
TR1
TR1
TR2
TR2
TR4
D3
TR3
D4
Figure 7.7 Using the bridge inverter as a power controller by switching off one transistor in a pair during its half cycle.
collector of the “off” transistor by autotransformer action. Thus transistors used in this circuit must be rated for operation at more than double the supply voltage. It is imperative that the primary switching is symmetrical so that the magnetic flux is zero at the end of each switching cycle; otherwise the transformer could saturate. The secondary voltage is dependent on the turns ratio of the transformer. Circuit diagrams of inverters often omit the diodes (i.e D1 and D2 in Figure 7.3) because the use of MOSFET transistors is as-
218
sumed. Their construction includes a parasitic diode (Figure 2.34) so no external component is needed. However, the diode is essential to the operation of the circuit. The Half Bridge Inverter is shown in Figure 7.5. Assuming the load is inductive, the waveforms are the same as Figure 7.4. In practical half bridge inverters the need for a center tapped power supply is dispensed with by using a pair of capacitors for AC voltage dividing. The (Full) Bridge Inverter of Figure 7.6 does not require a center tapped supply, and in this configuration the transistors are only subjected to the supply voltage. The transistors are switched in pairs – the opposite arms of the bridge switching together. The advantage of bridge converters is that they can easily be controlled to vary the output. While it is clear that on no account should the transistors in the two halves of the bridge be switched on simultaneously (since this would just short circuit the supply and destroy the transistors) there is nothing to say that there must be a continuous voltage applied to the load. Figure 7.7 assumes a bridge inverter with a filtered output and inductive load. The top part of the figure shows the relationship between the instantaneous load voltage, mean (filtered) load voltage and mean load current all at maximum output. It also shows which components in the bridge are conducting.
Lumens as Percentage of 50Hz Value
D1
120 18W 600mm T8 115
110 36W 1200mm T8 105
100 4
8
12
16
20
Operating frequency kHz
Figure 7.8 Graph showing how the output of a fluorescent lamp increases with supply frequency.
ELECTRONIC COMPONENTS
The lower half of the figure shows what happens if TR1 and TR4 are switched on together but, after a time TR4 is switched off. The instantaneous load voltage drops to zero, since the two terminals of the load are effectively short circuited by TR1 and D3. The same then happens with TR2 and TR3 , with TR3 being switched off early. The result is that the mean voltage across the load is reduced. The inverter has become a means of power control.
7.2 Electronic ballasts for fluorescent lamps 7.2.1. Operating frequency and input stage Figure 7.8 shows how the light output of a fluorescent lamp increases with operating frequency. A frequency in the range 25–70kHz is used in fluourescent lighting ballasts for the following reasons: • the lowest frequency is well above the audible range, so any mechanical sound that might be induced by the electromagnetic components is inaudible. • the lowest frequency also corresponds to the point where efficiency gains due to high frequency operation flatten out. • the frequencies used are not so high that there are significant power losses in the components and circuit boards. • components for operating at these frequencies are relatively inexpensive.
AC EMI IN suppressor
Rectifier
Electronic ballasts are made to a wide range of specifications. At the bottom are ballasts with a minimum specification and minimum component count that just meet regulatory requirements. At the top are ballasts with excellent regulation, lamp protection circuitry, sophisticated starting arrangements, remote control and minimum harmonic generation. Figure 7.9 shows a block schematic diagram of a generic electronic ballast. It must be remembered that, although the lamp is now operating at high frequency, the lamp is still a discharge lamp that requires starting and current limiting, so the output stage must act in a manner similar to the conventional electromagnetic component (albeit using physically smaller components because of the high frequency). All electrical equipment must meet the requirements for EMC (electromagnetic compatibility). The requirements are described in more detail in Section 8.5. Here it is sufficient to say that the EMI suppressor at the input of the ballast is intended to: • suppress voltage transients coming from the mains supply. • prevent interference generated by the ballast being sent back on the mains. • assist with limiting or preventing radiated RF. The actual filter design depends on the lamp power, since not unnaturally ballast designers go for the minimum filter that meets the regulations. A standard filter design is the “X-Y” capacitor plus choke design shown in Figure 7.10.
Power storage/ correction
Optional control
HF inverter
Output stage starting circuit
Lamp monitor
Figure 7.9 Block diagram of a generic fluorescent lamp ballast.
219
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
L
L
X1
Y1
X2
Equipment
Mains supply
Y2
N E
N E
Figure 7.10 Input filter for the suppression of electromagnetic interference.
This filter is designed to block both differential and common mode interference components (see Section 2.3.2.5 for definition of modes). Common mode interference is filtered by the choke and the Y capacitors. The choke has two windings on a common core. The core can be toroidal or E-I. Since the two windings are carrying the live and neutral feeds, safety considerations dictate some gap between the windings, but the general idea is 4
I (A)
2
that the differential line-neutral currents largely cancel out. This allows a high inductance value to be achieved in common mode without the problem of saturation from the supply current. While the choke can contribute to differential mode interference suppression through its leakage inductance, the main differential mode suppression is by the X capacitors. The reading list includes a book on the subject of EMC that gives a lot of practical information on the design of this kind of filter. Following the input filter, the rectifier is normally a conventional bridge rectifier. For many applications the output of the bridge, even with capacitor smoothing, is not high enough. When this is the case a boost converter is used; typically this raises the DC output to 400V or so. The boost converter can also be used to provide power factor correction, and this important subject is dealt with in the next section. The end of the power stage is a storage capacitor. This is necessarily an electrolytic component, since a typical required value is 15μF at 450V. This is a critical component that has a significant influence on ballast life, since the working life of an electrolytic capacitor is inversely proportional to the operating temperature. 7.2.2 Power factor correction
0
For electromagnetic ballasts the significant contribution to poor power factor is the displacement between current and voltage cos φ . For electronic
-2 -4 0
0.01
0.02 0.03 t (s)
0.04
f (Hz)
Figure 7.11 The harmonic content (below) of a typical non-linear load (above).
220
Figure 7.12 The first three odd harmonics present in the load of Figure 7.11 and their sum.
ELECTRONIC COMPONENTS
Standard limit
Figure 7.13 The first six odd harmonics present in the load of Figures 7.1 and 7.11 are largely responsible for the resulting waveform.
Figure 7.14 The international agreed limits on harmonics shown against the harmonics present in circuits like that of Figure 7.1.
ballasts the more significant item is distortion power factor associated with harmonics, and particularly with the behavior shown in Figure 7.1. The symbol that embraces all forms of power factor is λ, and it is the ratio of actual power used to the product of the r.m.s. input voltage and r.m.s input current. While the r.m.s VA is easy to measure and understand, the actual power is quite difficult to express when a complex waveform is involved. Mathematically the expression t λ = 1 / T ∫0 V ( t ) × I ( t ) × dt Vrms × Irms
Section 6.1.7.2 explained the problem that odd harmonics present in the supply network. A load with the characteristic of Figure 7.1 has a harmonic content similar to that shown in Figure 7.11. It is interesting to see what happens when the odd harmonics are considered on their own. Figure 7.12 shows the first three odd harmonics of a 50Hz supply, and superimposed on them is their sum. Already the sum waveform has some similarity to Figure 7.1. But Figure 7.13 shows that by simply adding in the next three odd harmonics, we have very nearly reconstructed the original waveform. In the world of electronic ballasts the interna-
is used to define λ. What this is saying is that the real power is found by dividing the voltage and current waveforms into small time slices each of dt and then measuring the voltage and current applying in each slice.
Harmonic order n 2 3 5 7 9 11001f n 001f 39
Maximum permitted harmonic current % 2 30 u O 10 7 5 3
Table 7.1 Harmonic limits for electronic ballasts as defined by IEC (EN) 1000-3-2. The harmonic current is expressed as a percentage of the input current at the fundamental frequency. λ is the circuit power factor.
1.00
0.60
0.20
-0.20
-0.60
-1.00
0.00m
16.00m
32.00m
48.00m
64.00m
80.00m
Figure 7.15 Input current waveform achieved by simple passive filtering.
221
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
tional standards specify a maximum total harmonic distortion or THD. Because in practice even harmonics are not a problem, a simplified formula for this is:
ILpK
THD =
IL
VIN L
I( t )
t
VIN 0010 VOUT L
I( t )
t
on S off ton
toff
T
Figure 7.16 Current waveform in the boost converter choke and the timing of the switch.
I
2 3
+ I I
2 5
+ 0016 I
2 39
RMS
Table 7.1 shows the maximum harmonic current permitted in electronic ballasts and similar equipment. Figure 7.14 shows what this means in relation to the harmonics present in the typical non linear load. The challenge is, therefore, to fit the ballast with circuitry that ensures that the input current waveform is as near sinusoidal as possible. One way is to fit the input with a passive filter consisting of a series inductance and shunt capacitance prior to the rectifier. Such an arrangement has been used in ear-
400
V 200
0
0
t
1.5
1.5
1
1
I
I 0.5
0.5
0 0
1.10
-5
-5 2.10
3.10
t
-5
4.10
-5
5.10
-5
0 0
1.10
-5
-5 2.10
3.10
-5
4.10
-5
-5 5.10
t
Figure 7.17 The rectified AC waveform and above it the comparatively smooth output of the booster circuit. Below two “snapshots” of the current waveform at two points of the mains cycle.
222
ELECTRONIC COMPONENTS
lier generation electronic ballasts, but, in order to achieve the limits set by the standard, the choke is big and heavy. The input current waveform that is achieved this way is shown in Figure 7.15. The passive filter has now given way to active power factor correction. The idea here is to force the current waveform to follow the voltage waveform. There are several ways of doing this, some of which are patented, but a good example, and a widely used method, is the use of the critical conduction mode boost converter. As already described in Section 7.3.1, the boost converter is used to step up the DC voltage. Conveniently it can, at the same time, be used to solve the harmonic/power factor problem. Figure 7.16 shows the current waveform of the choke and the timing of the switch applicable to the boost converter of Figure 7.2. The slope of the triangular current waveform is given by the equations shown in the figure. By controlling the peak current it is possible to ensure that the average input current waveform matches the input voltage waveform. Figure 7.17 shows the rectified AC waveform and, above it the output voltage of the booster circuit. Two “snapshots” of the booster waveforms are shown, each delivering the average current required at a particular point in the mains supply sinewave
Half Wave Rectified Voltage
Inductor Current
Supply Current
Figure 7.18 Diagram illustrating how the switching in a boost converter forces the current to follow the voltage waveform, and how switching occurs at zero current in the critical conduction mode converter.
L
D
S
(b) (a)
C
Lamp circuit
(c)
CONTROL
(d)
Figure 7.19 The boost converter used for power factor correction needs multiple control inputs.
cycle. The result of the process is shown in Figure 7.18. This also demonstrates the critical conduction idea. The ON switching occurs at exactly the point that the current is zero. There are alternative systems that do not allow the current to fall to zero. This has the advantage of reducing ripple currents, but the disadvantage that the inductor cores have residual flux at switching and therefore need to be larger to avoid saturation. In practice critical conduction (also known as discontinuous mode) results in smaller equipment for low power outputs, for example standard electronic ballasts; whereas the alternative continuous mode is more suitable for larger power outputs. Figure 7.19 elaborates Figure 7.2 to show the essential elements of a boost converter also being used for active power factor correction. This shows that the control circuit for the switch (usually a MOSFET) must have at least four input signals: (a) the rectified input voltage. This determines the shape of the current waveform that the circuit is going to control. (b) a signal from an auxiliary winding on the choke to detect zero current in the choke. (c) the current being carried in the switch. (d) the output voltage of the converter.
223
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
+V
C
a
B E Base Drive C B E 0V
+V D
G
b
S L
Gate Drive
Lamp
P D
G
S
C
Snubber
0V
Figure 7.20 Variants of the half bridge inverter used as the basis of fluorescent lamp ballasts.
7.2.3 The inverter The switching transistors used in electronic ballasts can be either bipolar or MOSFET. Power bipolar transistors can be supplied with a built-in diode, so that in bridge circuits there are no more components to mount than would be the case with MOSFET (which has a parasitic diode as part of its construction). A few years ago the argument was that bipolar was less expensive and had significantly lower losses than MOSFET. However this is only true at lower frequencies. The MOSFET can work at higher frequencies and is simpler to drive, but it does have higher on-state losses. The result of recent developments in MOSFET technology has been lower losses and lower price.
224
Because they are easier to drive, MOSFETs are now the most widely used switching components in low power electronic ballasts. The great majority of electronic lighting ballasts use the half bridge inverter as their basis. As a ballast the whole device still needs to have impedance current limiting, but now, because of the high frequency used, the choke is physically very small compared with its 50/60Hz counterpart. Figure 7.20 shows two possible arrangements. In (a) two capacitors make up the other arms of the bridge, and create an AC potential divider. In (b) only a single capacitor is used to provide a return AC current path and to block the DC. The two halves of the half bridge are referred to respectively as the high side driver and the low side driver for obvious reasons. It can be seen that, neglecting any transistor losses, the voltage at the midpoint P is successively switched between the high voltage rail, and the low voltage rail. A snubber capacitor is fitted to the circuit to reduce dv/dt and limit RF radiation. A low resistance may be in series with the snubber capacitor to damp any oscillations arising from a combination of the snubber capacitor and stray circuit inductance. Typical waveforms are shown in Figure 7.21. The high side voltage of above 400V represents the typical output of the PFC boost circuit. The average voltage at the bridge midpoint is half the supply voltage. Switching losses are kept to a minimum because the timing of the “on” signal to a transistor coincides with nearly zero volts across it. This is because the other transistor has switched off and the inductive current is flowing through the transistor’s diode until it changes direction. Figure 7.21 reveals a useful feature of the circuit. The DC blocking capacitor and the choke inductance together form a resonant circuit, and this has the practical result that the lamp voltage waveform is nearly sinusoidal. The half bridge can be driven in various ways. In simple ballasts the whole circuit is made into an oscillator by feeding a signal back from the output to drive the transistors. The idea is shown in Figure 7.23 where a transformer primary is in series with the load. Its two secondary windings are in antiphase to ensure that only one transistor can be on at any
4.00E-05
3.00E-05
2.50E-05
2.00E-05
1.50E-05
1.00E-05
5.00E-06
0.00E+00
450 400 350 300 250 200 150 100 50 0 -50
3.50E-05
Max Power Mean Power Min Power
-5.00E-06
U /V
ELECTRONIC COMPONENTS
t/s
Current through lower FET or half bridge
MaxPower MeanPower MinPower
0.5 0.4 0.3 0.2 0.1 0 -0.1 -0.2 -0.3 -0.4
4.00E-05
3.50E-05
3.00E-05
2.50E-05
2.00E-05
1.50E-05
1.00E-05
5.00E-06
-5.00E-06
0.00E+00
I /A
7.2.4 Lamp starting
t/s
4.00E-05
3.00E-05
2.50E-05
2.00E-05
1.50E-05
1.00E-05
5.00E-06
0.00E+00
3.50E-05
Max Power Mean Power
500 400 300 200 100 0 -100 -200 -300 -400 -500 -5.00E-06
U /V
Lamp voltage
t /s
Lamp current
Figures 7.20 and 7.21 assume that the fluorescent lamp has already been started, and is in its running condition. Several different arrangements are used for starting the lamp. The simplest arrangement is the instant start ballast. In this case no attempt is made to pre-heat the cathodes, and starting relies on a sufficiently high voltage across the tube to cause conduction by field emission. The electronic ballast is able to produce a high enough starting voltage to ensure that lamps with conventional cathodes can be started this way. Figure 7.22 shows a simple instant start circuit, where the addition of the capacitor CR creates a resonant circuit. Since the DC blocking capacitor C is large, the high voltage across CR appears across the lamp tube (typically 400–1,000V depending on the tube). Once the tube has struck, main current flows
MaxPower MeanPower MinPower
0.25 0.2 0.15 0.1 0.05 0 -0.05 -0.1 -0.15 -0.2 -0.25
+V
CR
4.00E-05
3.50E-05
3.00E-05
2.50E-05
2.00E-05
1.50E-05
1.00E-05
5.00E-06
L 0.00E+00
-5.00E-06
U /V
time. An arrangement is needed to get the circuit to start oscillation; one way is to give the low side driver a “kick start” using a diac. In Figure 7.23 RS charges up CS and when the diac’s breakdown voltage is reached CS discharges into the transistor base. Once the main oscillation is under way, the starting diac oscillator is disabled because diode DS discharges the capacitor every time the low side transistor conducts. More sophisticated ballasts use a separate oscillator to provide the transistor drive. This permits better control of the half bridge, allowing the introduction of other premium ballast features.
t/s
Figure 7.21 Waveforms associated with the circuit of Figure 7.20(b) in a typical fluorescent lamp ballast.
C 0V
Figure 7.22 Instant start achieved by the use of a capacitor that resonates with the lamp choke.
225
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
+V PTC RS CI
L
DS CH Diac
CS 0V
C
Figure 7.23 The half bridge based on oscillation by positive feedback, together with a pre-heat starting circuit.
through the tube and the resonant frequency is determined by L and C. Longer tube life is obtained by using pre-heat or rapid starting. The simplest way of achieving this is to wire the starting capacitor through the tube cathodes, so the capacitor current provides heating. In practice the circuit is made more complex as shown in Figure 7.23. The Positive Temperature Coefficient or PTC resistor is a device whose resistance increases dramatically when a certain temperature is reached. When the circuit is switched on the starting resonant frequency is determined by L, C and CH since CI is effectively short circuited by the PTC. When, after a short time, PTC goes high resistance, the resonant frequency goes up because CI joins CH in series. The voltage across CI and CH is sufficient to strike the tube. The value of CH and the characteristics of the PTC are chosen to match the cathode pre-heating requirements of the particular lamp, without producing a sufficient voltage across the tube to start it prematurely. CI is chosen to produce the required starting voltage. Once the tube is conducting there is still a small residual cathode current, representing typically 5% of the circuit power dissipation. In ballast designs using separate oscillator control, other methods of providing pre-heat are used.
226
One method is to use a separate cathode heating transformer, analogous to the conventional rapid start arrangement. This has the potential advantage that it can be taken out of circuit once the tube has struck, thus reducing power losses. Another method is to provide a pre-heat supply from additional windings on the main choke. While this does result in some residual cathode current, the way the circuit operates means that the circuit losses are very small. A typical arrangement is shown in Figure 7.24. Initially the inverter is driven at a high frequency and by transformer action both cathodes are heated by the auxiliary windings on the choke. Each cathode circuit has a series capacitor that prevents any DC flow. The frequency then drops to near the resonant frequency of L and CR creating the required starting voltage. Once the tube has struck the operating frequency drops again because the main current is now through the DC blocking capacitor. The lower operating frequency, compared to the starting frequency, means that the cathode series capacitors are comparatively high impedance when the lamp is running, so that actual cathode heating losses are small. Figure 7.25 shows how the lamp voltage varies with frequency according to the different stages of the starting process.
L Bridge Output
CR 0V
C
Figure 7.24 Cathode heating using auxiliary windings on the choke.
ELECTRONIC COMPONENTS
10000
Log Scales
1000 Voltage
Open circuit characteristic
Running characteristic
100
10
4
Ignition Typical operating Preheat points Running
Frequency
5
10
Figure 7.25 Characteristic of a lamp circuit of the kind shown in Figure 7.24 under different conditions.
In tests results reported by the National Lighting Information Program in the USA, some rather surprising facts emerged. The tests were to determine fluorescent lamp life based on the number of switching cycles. It was not unexpected that lamps equipped with traditional rapid start magnetic ballasts had a significantly shorter life than electronic ballasts, with electronic ballasted lamps being able to withstand at least four times as many starts as the magnetic ballasted. What was unexpected was that many electronic ballasts with pre-heating had no better switching cycle life performance than instant start – in some cases it was actually worse. The tests revealed that unless pre-heating was done correctly, preheat start with its comparatively low striking voltage was actually far worse for the tube than cold starting with a significantly higher voltage. On the other hand, pre-heating done correctly produces a significantly longer life, for example the ability to withstand at least three times as many starts as the instant start circuit. Research revealed that a useful metric is the ratio of the cathode hot resistance RH to its cold resistance RC. Unless the ratio is correct, the benefits of pre-heating are lost. RH / RC of 4.25 for common
lamps corresponds to a cathode temperature of 700°C which is now considered the minimum temperature needed to avoid cathode damage on starting with medium voltages. The optimum temperature is recommended by the lamp manufacturer, and may well be in the range 900–1,200°C. However, with the arrival of new tube types, it has become necessary to have standards for all ballasts that lend themselves to objective measurement. This has resulted in the IEC60929 standard specifying starting performance in terms of the energy delivered to the cathodes. It is up to the lamp manufacturer to define this, but the IEC then specifies objective test methods to demonstrate that the ballast is able to deliver the correct energy. The tests are carried out using resistors as cathode substitutes so that precise measurements can be made. The preheat energy E in Joules is given by an equation of the form: E=Q+P×t where Q is in Joules and represents the energy supplied to the electrodes, and P in watts is the power lost during the pre-heat time, t is the pre-heat time in seconds. The lamp manufacturer gives the maximum and minimum values of P and Q. The manufacturer also gives equivalent cathode heating currents for specified preheat times. Some example values for the T5 High Efficiency series of lamps are as follows: E at 0.5 seconds, minimum 1.30, maximum 2.25J E at 2.0 seconds, minimum 2.40, maximum 4.20J Q minimum 0.90J, maximum 1.60J P minimum 0.75W, maximum 1.30W Preheat current for 0.5s start minimum 290mA Preheat current for 0.5s start maximum 385mA Preheat current for 2.0s start minimum 200mA Preheat current for 2.0s start maximum 265mA While the energy equation itself does not specify ignition times, in practice the manufacturers stipulate a minimum of 0.4s, and a practical maximum of 3.0s.
227
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
the “three hour” test if the actual ON times are significantly longer than three hours. It is clear from the table, however, that it is a bad idea to use instant start circuitry for lamps that are to be frequently switched.
WRONG! Ballast
7.2.5 Ballasts for standard fluorescent lamps Ballast
RIGHT!
Figure 7.26 Serious damage to fluorescent lamp cathodes occurs if an electronic instant start ballast is not connected correctly.
When instant start electronic ballasts are used with fluorecent lamps with conventional cathodes, it is essential that both ends of the cathode are connected to ensure that as far as possible the cathode is at the same potential along its length. See Figure 7.26. International standards specify that the lamp’s glow (non thermionic) discharge time in an instant start circuit must not exceed 0.1s, which is the same as saying that the lamp must strike in less than a tenth of a second. Lamp life is specified in different ways. It is dependent on both the actual number of hours run, and on the number of lamp starts. One test method rates running life on the basis of 3 hour cycles, 165 minutes ON, and 15 minutes OFF. Another rates life on the number of on/off cycles with 3 minutes ON and 3 minutes OFF. Table 7.2 gives the expected life figures for T5 linear lamps on these bases. Note that achievable life will be considerably longer than that shown for
Regime 165 min ON 15 min OFF Instant start 3min ON 3 min OFF Pre-heat start 3 min ON 3 min OFF
Life >16,000 Hours 5,000 starts >100,000 starts
Table 7.2 Life rating for representative T5 linear fluorescent lamps. (Actual figures achieved depend on particular lamp and ballast combination).
228
Electronic ballasts are offered in a wide range of types and physical constructions. In the USA the electronic ballast was at first regarded as a straight substitute for its magnetic forbear, so many USA electronic ballasts are made in similar chunky can sizes to those of USA magnetic ballasts. In Europe the magnetic ballasts were already much smaller (especially in cross section) and, again, electronic ballasts for conventional fluorescent tubes initially followed the form factor of the magnetic components. With the arrival of the slim T5 lamp (which cannot be operated from a 50/60Hz magnetic ballast) a demand for slimmer ballasts emerged, since luminaire designers wanted to take advantage of the lamp’s slimness, and did not want a bulky ballast to take away some of the space gain arising from the thinner lamp. Several important parameters are specified for electronic ballasts. Some of these are international standards, some are manufacturer specific.
Figure 7.27 Examples of electronic ballasts for linear fluorescent tubes. For T8 tubes from Philips/Advance Transformer (USA), and for T5 tubes from Helvar (Finland.)
ELECTRONIC COMPONENTS
Transistors
Input filter Mains connection
Boost converter Main storage capacitor
Half bridge inverter
Output choke Lamp connection
Figure 7.28 A typical standard electronic ballast circuit board with pre-heat start showing the principal components.
Current crest factor should not exceed 1.7 (although in Japan higher crest factors are permitted). In practice meeting the requirements for EMC and harmonic reduction goes a long way to ensuring this requirement is met. Apart from the power quality issues raised by a high crest factor, lamp life can be shortened by it. High quality electronic ballasts usually have a current crest factor of around 1.4, i.e. the same as for a sine wave. Inrush current or I2t. There is a high inrush current on electronic ballasts arising from the charging up of the main DC storage capacitor. When a large group of ballasts is switched on together, the inrush can be great enough to trip the circuit breaker. The ballast data should give an I2t rating to allow a circuit breaker of the right characteristic to be selected. Case temperature or TC. As with magnetic ballasts there is a maximum temperature for the transformer windings to ensure long life. In addition many of the electronic components will also have limits on operating temperature. Ballasts are designed so that internal operating limits are not exceeded provided the case temperature does not exceed a specified case temperature TC. The ballast has a test point marked on it where the temperature should be measured. The ballast specification will indicate a maximum TC usually around 70–80°C. However the ballast life figure may be based on a somewhat lower figure, like 60°C at the test point. A 10°C difference in case temperature can double or halve the expected life.
Lamp starting temperature. A fluorescent lamp will only reliably start if the lamp tube is at a reasonable temperature. The ignition voltage varies considerably with the lamp temperature, being at a minimum at around 10°C for the older T12 tubes and around 45°C for T8 tubes. Either side of the minimum the ignition voltage goes up until a point is reached when the lamp will not strike at all. The starting temperature range will depend on both the lamp and the ballast. A typical temperature range is -15°C to +50°C, but some lamps only reliably start from 0°C or even higher. Leakage current. The Y capacitors in the input filter (see Figure 7.10) are the primary cause of leakage current. Typical ballasts have a leakage current in the range 0.2–0.4mA. Thus if a large number of
Figure 7.29 The business end of a ballast using the circuit of Figure 7.24. In this case the output choke winding is in two halves; and the two cathode windings are positioned between them.
229
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
conditions, and to eliminate lamp flashing when a lamp fails. Fluorescent lamps usually fail through poor cathode emission. If one cathode fails before the other, then more current flows in one direction through the lamp than through the other. There is a partial rectification effect that results in lamp flashing and damage to the ballast because of the DC. Faulty lamps may also fail to start in the first case; this can result in very high circuit voltages (Figure 7.25) which must not be maintained for any length of time. Similarly a faulty lamp holder or lamp
+V
Driver
Converter Control
DC Detect
VCathode
VLamp
Bridge I
+V
0V
Ballast Control Supply Cycle
Figure 7.30 Block diagram showing the function of the ballast monitoring and control circuit.
ballasts are installed in a circuit with RCD protection, there will be a limit to the number that can be connected to a single RCD. Typically a maximum of 30–40 ballasts is recommended for a single 30mA RCD. DC operation. Since the electronic ballast starts by converting incoming AC to DC, it is reasonable to suppose that the ballast can work directly from DC. Most “professional” electronic ballasts can work from DC, and this feature is used for emergency lighting systems based on high voltage batteries. The ballast specification will give the range of AC and DC supply voltages that the ballast can be safely operated from. Any changeover from one supply to another must be on a “break before make” basis. Safety features. The introduction of electronics into the ballast opens up new control possibilities, especially when a controlled oscillator is being used to drive the half bridge. Professional ballasts are fitted with a control circuit that can be used for several purposes including regulating the lamp power. Its prime use, however, is to prevent potentially unsafe
230
Ideal Light Output
Real Light Output
Flicker =
+
x 100%
Area Flicker Index = Area + Area
Figure 7.31 Flicker percentage and flicker index.
ELECTRONIC COMPONENTS
Rating AC line voltage range V Max AC V one hour DC line voltage range V Line frequency range Hz Line current mA Power factor Power loss W Lamp voltage V Lamp current mA Lamp frequency kHz Preheat KH current mA Running KH current mA Start time s O/C Volts t < te O/C Volts t > te Leakage current mA
3u18W T8 1710010264 320 2000010300 49001061 240 0.98 8 55 290 30 480 100 2.0 230 600 0.35
Ballast 2u28W T5 1710010264 320 2000010300 49001061 278 0.98 7 167 170 32 205 70 2.0 185 650 0.35
Type 1u35WT5 1710010264 320 1800010300 49001061 165 0.97 3 205 170 26 210 80 2.0 160 550 0.35
2u36W T8 1700010300 320 1800010400 49001061 190 0.99 6 102 320 25 570 170 1.8 340 550 0.19
1u70W T8 1700010300 320 1800010400 49-61 290 0.99 6.5 129 470 27 950 210 1.35 430 680 0.21
Table 7.3 Typical data for a representative range of rapid start electronic ballasts. In the table KH = cathode heating, and O/C = open circuit. te is the time thermionic emission starts.
connection could result in a high voltage condition. Finally the ballast itself may not operate correctly unless it receives a supply voltage within a specified range. A typical ballast control circuit not only controls the half bridge to operate the starting routine, but also senses the common fault conditions and puts the ballast into a “stand-by” mode if a fault is detected. “Stand by” turns the low side driver on, so the half bridge output is clamped to 0V. It also stops the boost converter, so the DC rail drops to the peak AC supply voltage (or the DC supply voltage). The control circuit might have the following inputs: • sensing of lamp voltage. If this rises above the tube starting voltage, or if a starting voltage does not drop within a few seconds, this denotes tube ignition failure, so stand-by must be invoked. • sensing of half bridge current (this can be done by putting a low value resistance in series with the low side transistor and measuring the voltage across it). If the current is too high, the controller can slightly increase the operating frequency. This will reduce lamp voltage (Figure 7.25).
• sensing of cathode voltage at “cold” end of lamp. If no voltage detected, lamp must be assumed missing or not connected, so stand-by should be invoked. On the other hand when a voltage is once again detected, the ballast can start up again. • sensing of significant DC at the blocking capacitor. This signifies rectification and imminent lamp failure. • sensing of the DC storage capacitor voltage. If this drops low, then stand-by is initiated. • sensing of open circuit filament. It is possible to realize the entire control circuit using discrete components, but more usually a mixture of discrete components and a microprocessor (or ASIC) is used. Figure 7.30 shows a block diagram of the control circuit operation. Cathode failure can lead to severe hotspots, and in the worst case to the lamp “exploding” at the hotspot. The problem is potentially sufficiently serious that for linear and CFL T5 lamps it is mandatory under the IEC standard to have automatic ballast shutdown when an irregular condition is detected. Ballast manufacturers are expected to identify which method of lamp failure detection
231
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
between one make of ballast and another. Obviously parameters that are set by international standards (like lamp voltage) should not, but a parameter like the frequency can change according to the ballast design. 7.2.6 Multiple lamp operation From main ballast
Figure 7.32 Operation of tubes in series.
(assymetric power, assymetric voltage or open circuit filament) they are using. Flicker index/percentage. Flicker arises in discharge lamps because of asymmetry in the discharge. In a 50 or 60 Hz system the light output peaks at 100 or 120Hz. If each half cycle has identical light output, the eye is not disturbed by the variation within cycles. However, if one half cycle is brighter than another, flicker can become noticeable and objectionable. One method of measuring flicker is by the modulation difference between the two half cycles. If measured as a percentage, then some people can find a 1% flicker objectionable and most people find 2% objectionable. The USA uses flicker index as the metric, which compares the amount of light that exceeds the average (as opposed to simple amplitude comparison). Figure 7.31 shows how flicker index and flicker percentage are measured. Flicker index varies from 0– 1, with 0 indicating continuous light output without flicker. Electromagnetic ballasts have a flicker index in the range 0.01 to 0.1, the higher end of which, 0.07–0.1 is perceptible to some people. Electronic ballasts should not in principle suffer from flicker, but even they can have some asymmetry so can have a measurable flicker index below 0.01. A particular problem can be any form of electrical noise, not necessarily related to the line frequency, which can raise the flicker. Table 7.3 gives outline specifications of typical ballasts for linear fluorescent lamps. The values are examples – individual ballasts of a type have variations, and some parameters may vary considerably
232
Electronic ballasts are suitable for operating multiple lamps. Various different circuits are used. Just as for electromagnetic ballasts, series operation of lamps is possible. In instant start circuits starting of the second lamp is assured by the high voltage still present even after the first lamp has ignited. In preheat circuits it is necessary to provide an isolated source of cathode heating for the centre cathode pair, and a simplified circuit for doing this is shown in Figure 7.32. Professional ballasts usually use separate output stages for twin tube operation as in Figure 7.33; but where short tubes are being used, for example in a 4 × 18W ballast, each of the output drivers can drive a series pair. An example of a different circuit topology is shown in Figure 7.34. Here a current fed parallel resonant push-pull inverter is used. The resonance
Bridge Output
0V
Figure 7.33 Professional ballasts have twin output stages. Sometimes a capacitor links the two circuits to balance the light output, and to assist the starting of the second lamp.
ELECTRONIC COMPONENTS
12V Battery
R
Base Drive
C
Base Drive
+
CB C
Figure 7.35 Battery operated single transistor inverter used for low power fluorescent lamps.
+ VDC -
Figure 7.34 Multiple lamp circuit based on a push-pull inverter with isolated output and capacitor ballasting for each lamp.
arises from the capacitor and the primary inductance of the transformer. The current source is the separate inductor, which has the advantage of ensuring a sine wave output from the transformer. In turn this allows the use of ballast capacitors for each tube that can also develop the tube starting voltage. The circuit has the features that the output is isolated and that the failure or disconnection of any one tube does not affect the operation of the remainder. In all electronic ballast circuits the presence of high frequencies makes the realization of EMC standards difficult. The lamp leads themselves can be the source of radiated interference, and the means by which interference can enter the ballast leading to unstable operation. For this reason manufacturers recommend the shortest possible lamp leads, and, in particular, that the “hot” end leads should be as short as possible. Also that on no account should lamp leads run parallel with the ballast supply leads. These factors can determine the position of the ballast within the luminaire, particularly where multiple lamps are being used.
7.2.7 Ballasts for compact fluorescent lamps One of the earliest uses of electronics for lighting was the battery powered inverter for small (4–13W) linear fluorescent lamps as used in vehicle, camper and boat lighting, handlamps and emergency lights. In this application EMC is still an issue, but harmonic generation is generally not an issue. This allows very simple circuitry to be used; the same circuit at high voltages would impose very severe stress on the transistor. Figure 7.35 shows the outline circuit of a single transistor flyback inverter that can be used for this application. An auxiliary winding on the transformer provides feedback to the transistor base to maintain
Figure 7.36 Battery inverters use very few components.
233
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
The monolithic IC in a small SO8 package. But its internal circuit is quite complex
5m
m
Collector
diac Diac Sec
OSC
Figure 7.37 Examples of compact fluorescent lamp ballasts. These are from Philips, Tridonic, Harvard Engineering and Hüco.
oscillation. The frequency of oscillation is determined by the transformer’s primary inductance, R and C. These also determine the mark-space ratio of the switching which must be 1:1 to avoid rectification developing in the lamp. The by-pass capacitor CB provides a low impedance return path for the high frequency. The high voltage needed to operate the lamp is determined by the transformer turns ratio. The transformer is of leakage reactance construction so there is a high open circuit voltage to start the lamp, but current is limited once the lamp is running. The cathodes can be heated by auxiliary windings which, due to the action of the leakage reactance transformer, will deliver more power at start up than when the lamp is running. In general the description “compact” when applied to fluorescent lamps refers to lamps that are single ended; requiring the tube to have been folded along its length one or more times. Some of these are of the same power as the normal tubular lamp, for example TC-L lamps of 36 and 55W rating, and although only half the length of the tubular equivalent are still quite long. They can therefore use electronic ballasts of similar characteristics and form factor to their T equivalents. In luminaires using the shorter lamps (either low power “2-finger” TC-E, or 4 and 6 finger lamps TCDE and TC-TE) the “long thin” ballast is not often
234
+ +
R
Vref
2Vref Source
A complete ballast circuit using two ICs
IC High Side
IC Low Side
Bridge and input filter
Optional additional components for frequency change for pre-heating
AC IN
The completely assembled ballast
goes in to a lamp like this Figure 7.38 An example of the use of small power integrated circuits as the basis of an electronic ballast built in to the lamp. Information courtesy ST Microelectronics.
ELECTRONIC COMPONENTS
convenient, so while the ballast circuitry remains the same or similar, the ballast form factor changes – examples are in Figure 7.37. Now many compact fluorescent lamps are available with built-in ballasts – the ballast becoming “disposable” like the lamp itself. Present international standards allow some relaxation in harmonic performance at the powers at which these lamps operate, so circuitry tends to be simpler. Nonetheless the circuitry must be robust and efficient, as well as being extremely compact. The present trend is to higher circuit integration whereby component count is reduced, but at the same time considerable sophistication can be incorporated. One of the problems at the low powers concerned is that the circuit can start having losses that are a significant proportion of overall circuit power. An example of the kind of integrated component now available is given by the VK05CFL driver from ST Microelectronics. This is a monolithic device that can operate as the complete high side or low side lamp driver in a half bridge circuit. It combines the functions of gate drive and power switching in one device. In order to minimize losses it uses an emitter switch as its power output stage. This is a hybrid arrangement using a low voltage MOSFET in cascade with a high voltage bipolar Darlington pair. The idea is to achieve the advantages of the low ON voltage drop and high OFF breakdown voltage of the bipolar, and the high switching frequency of the MOSFET. Figure 7.38 shows the integrated circuit, its internal block diagram, the circuit of a complete ballast, a completely assembled ballast, and a lamp using this kind of ballast circuit. The intention here is not to suggest that this is the only way of achieving a built-in lamp ballast, but to show the underlying sophistication needed. It can be seen that the circuit embodies: • feedback windings that not only provide oscillation, but also provide power for the IC. • diac initiation of the oscillator. Only one IC needs connection to the external RC network since the other IC will follow. • external capacitors to set oscillator frequency.
• an optional facility for timed pre-heat at a different frequency. 7.2.8 Controllable fluorescent lamp ballasts From Figure 7.30 it is clear that the ballast control circuit could do more. It is already required to ensure stable operation of the lamp at a specified power, so the obvious extension is for it to be able to vary the lamp power in response to an external signal. Controllable ballasts have provision for a low voltage input signal that is electrically isolated from the lamp circuit. The signal can vary the light intensity between set limits. Standards exist for both analog and digital control signals and these are discussed in more detail in Chapter 9. The achievable dimming range with fluorescent lamps is limited by the lamp-ballast combination. Smooth dimming to total extinction is, in practice, impossible. Applications requiring smooth dimming to extinction for esthetic effect, such as cinema auditoria, should not use fluorescent dimming. Some practical points now follow. The relationship between lamp power and light output is not linear. What matters to the user is the light output, so the ballast specification should give the percentage range of light output. For ballasts intended solely for use in energy management schemes, a control range 25–100% (light output) can be sufficient. Where they are to be used in architectural lighting schemes, a range of 10–100% is the minimum specification and 5–100% is preferable. The human eye adapts to low light levels, and 10% light output still seems quite bright. Where a subjectively low level is required, even 5% is too high. Some ballasts are specified for a wider range, such as 3% or even 1%. In practice it may be difficult to get even performance between multiple ballasts at these low levels. Some specifications are only valid at a particular lamp operating temperature. Lamp/ballast combinations should be evaluated under their intended operating conditions if the application is critical. Demand from the residential market is leading to the introduction of “1%” ballasts for selected compact lamps.
235
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
(a)
Lamp L
R CR
C DC
ZL I
ZC V IN
Lamp Voltage VL
VIN R Z2 LCR 000e ZL 000e R
R R
VL
C
Lamp
L
(b)
IL
Lamp Current IL
VIN Z2 LCR 000e ZL 000e R
Figure 7.39 Simplified equivalent circuit of the output stage of an electronic ballast and a fluorescent lamp.
Fluorescent lamps operating at low light levels and high frequencies can develop acoustic resonance (described in Section 7.3.1). With fluorescent lamps the visible result is a series of swirling rings of light of varying intensity along the tube. One way of reducing this effect is to introduce a small amount of DC into the discharge. This can be done by connecting a high resistance between the high side of the DC blocking capacitor C in Figure 7.24 and the main DC rail. In principle there are various ways by which lamp power can be controlled. The obvious method is to simply reduce the voltage on the DC supply rail, by, for example changing the operation of the boost converter. While in principle this method works for a limited range of control, it would be difficult to implement for a wide range of control. Another method is to use pulse width modulation (PWM) of the high frequency output. This is
236
practical, but with conventional fluorescent tubes the situation is complicated by the needs of cathode heating. If PWM is used, a separate circuit is needed to maintain cathode heating. PWM control is widely used for small cold cathode lamps. The practical method that is most widely used is to take advantage of the combination of lamp and circuit characteristics that allow light output to be changed simply by changing the frequency of operation. The details of circuit operation will depend on the circuit topology of the ballast (it must be repeated that this chapter is using examples of recognized practice, and that there are other circuits that can achieve similar results). Taking as an example the output stage of Figure 7.24, and treating the lamp as a resistance, a simplified circuit is Figure 7.39(a). Since the DC blocking capacitor CDC is very large compared with resonant capacitor CR , a further simplification gives Figure 7.39(b), where also a sinewave generator is shown replacing the rounded square wave that is the actual output of the bridge. Here it can be seen that as the frequency of the source changes, the relative voltages across L and C change. The inductance and capacitance represent an AC potential divider where the midpoint voltage changes with frequency; the higher the frequency, the lower the voltage across the capacitor (because at high frequencies the capacitor represents less of an impedance) and the lower the frequency, the higher the voltage (because at low frequencies the inductance represents less of an impedance). The voltage across the lamp is the same as that across the capacitor, so in principle the power delivered to the lamp is inversely proportional to the frequency. In fact the simple principle is complicated by the fact that the lamp resistance is not constant. Figure 7.40 shows an example of the Voltage/Current relationship for a fluorescent lamp, and its variation with operating temperature. Figure 7.39(b) shows the lamp current and lamp voltage for an instantaneous value of lamp resistance. Using these equations and the lamp data of Figure 7.40 it is possible to develop a graph that shows the relationship of lamp power to frequency. In practice the process is somewhat empirical since
ELECTRONIC COMPONENTS
0
10 C 0
Lamp Voltage
-15 C 0
22 C 0
35 C
10
100
Lamp current (log scale)
1000
Figure 7.40 The relationship of lamp voltage and lamp current for a T5 54W fluorescent lamp at different ambient temperatures.
initially the values of C and L are not known. One way to proceed is to choose a practical frequency range; then apply the R value corresponding to full output to the lowest frequency to be used; and the R value corresponding to the lowest output to the highest frequency. A practical capacitor C value can be chosen, and the corresponding L derived. The equations can then be solved for different values of R to derive the power/frequency relationship. In practice this shows that at the high frequency end a small change in frequency makes a significant difference to the power, as shown in Figure 7.41. Table 7.4 shows two examples of controllable ballast characteristics. Once again these are only representative examples. In this case the lamp parameters are related to the control voltage setting, assuming the use of the 1–10V analog control system. The dimming of fluorescent lamps brings potential problems in respect of lamp life. Clearly if the lamp power is reduced, then so is the energy that keeps the cathodes in a thermionic emissive state. It is evidently important to increase cathode heating at low light levels, since the lamp discharge itself will not provide enough energy to keep the cathodes at the right temperature. One method is to provide a separate cathode heating supply that is on at all times; but this is energy
inefficient at high light outputs. Indeed the lamp cathodes are under attack from two different mechanisms: • too much cathode heating leads to cathode loss by evaporation. • too little cathode heating results in cathode loss due to sputtering. This is where the discharge current is concentrated on cathode hotspots. The aim of controllable ballast design is to ensure that the cathodes receive as much additional heating as is necesssary for the particular lamp power level. While the output circuit of Figure 7.24 and others like it have the characteristic that cathode heating current is automatically increased as frequency increases, there is a need to optimize the arrangement, while at the same time keeping the ballast design simple. The concept is that of ensuring that the total energy at the cathode is correct, and the principle is that of extending the starting condition equations to steady state at different dimming levels. The most widely used metric relates to the current in the high current lead. The top part of Figure 7.42 shows how the currents flowing in the two leads of a cathode are different. The high current lead ILH is carrying both the main discharge current ID and the cathode heating current, whereas the low current lead ILL is carrying only the heating current. 100
10
1 10
Frequency kHz
100
Figure 7.41 Multiplying the values of VL and IL using the equations of Figure 7.39, and applying the varying values of R derived from Figure 7.40, gives a lamp power/frequency relationship of the kind shown here.
237
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Control voltage V
10
Line current mA Line power W Power factor Lamp voltage V Lamp current mA Cathode voltage V Lamp frequency kHz Light output %
170 245 39 58 0.98 0.98 100 111 320 455 2.4 1.6 40 43 96 96
8 152 35 0.98 103 270 2.7 44 84
6 225 53 0.98 115 395 2.0 50 87
117 27 0.98 111 180 3.6 51 62
4 167 39 0.98 130 230 3.1 66 60
87 20 0.98 123 95 4.4 55 36
2 108 25 0.94 160 85 4.5 71 27
55 69 12 14 0.92 0.85 130 170 29 20 4.7 4.7 56 72 10 6
1 48 64 11 13 0.87 0.83 120 165 14 14 4.6 4.7 57 72 4 4
Table 7.4 Relationship of control voltage to lamp operation in example lamp-ballast combinations. The red figures relate to 36W T8 lamps, and the black figures relate to 58W T8 lamps. Ballasts are Helvar HFC series.
The bottom part of Figure 7.42 shows the permitted parameters of operation for a T5 high output 54W lamp. Similar diagrams are provided by the lamp manufacturers for each lamp type. The green area is the area of permitted operation; the darker green hatched area is an area in which the discharge current alone is sufficient to maintain cathode temHeating Voltage
ILH ID ILL mA 700 ILH 600 500 400 300 200 100 0
mA ID
Figure 7.42 Limits of operation for a high output 54W T5 fluorescent lamp. See text for description. Figure derived from data provided by Philips and Osram.
238
perature, and where additional cathode heating is optional. In the lighter green area, additional heating is essential. The red line represents one manufacturer’s recommendation for maintaining conditions for maximum tube life. At all times the current in the low current lead ILL must not exceed a value given by the lamp manufacturer (approximately 490mA for the example T5 HO 54W lamp). For low lighting levels, lamps are run at ID currents lower than that implied by the diagram. At the time of writing the IEC standard test procedures are fairly well defined for the equivalent of the green area of the diagram, but performance at low discharge current levels is still the subject of research. Actual performance in the low current area is, therefore, manufacturer specific and depends on the particular lamp/ballast combination. While current recommendations for cathode heating under dimming operation are based on the ILH parameter, research is showing that the interaction between the heating and discharge currents is complex. The two currents may be out of phase, and in most cases there is some discharge current through both lead wires. A new method of defining “target” optimum lamp operating conditions, which may help with the definition of permitted low current operation, is that of setting target values for the sum of the squares of the two lead currents, i.e. ILH2 + ILL2. The aim of this kind of rule is to have a method of measurement that is easy to apply, but which is itself based on both theoretical and experimental data.
ELECTRONIC COMPONENTS
7.2.9 Dimmable fluorescent lamp ballasts For all new installations the most practical method of controlling the light output of fluorescent lamps is by the use of controllable ballasts as described in the previous section. In general these ballasts require a separate control cable that operates at low voltage, which in turn means that the ballast must have some method of isolation for the control signal. It is possible to eliminate the separate control cable by superimposing a control signal on the mains supply. Such a technique is useful in retrofit situations where luminaires are replaced or upgraded with new ballasts, since there is then no need to install the separate control cable. Sometimes there is a requirement that fluorescent lamps be dimmed by standard triac or thyristor dimmers. One method is to use the electronic equivalent of the circuit shown in Figure 6.56, whereby there is a separate maintained live supply to provide cathode heating, and a variable supply for the dimming. This requires additional wiring, so not surprisingly there is a preference for a ballast that can work solely from the variable supply. This is technically possible, but there are two difficulties. In order to meet the requirement to provide sufficient cathode heating at low light levels, the circuit must maintain or increase cathode heating when the power actually reaching the ballast is decreasing. The fact that the incoming waveform is a phasecut sinewave running at 50/60Hz means that con-
Figure 7.43 The Lutron Hi-Lume™ series of electronic ballasts uses a maintained live feed to allow dimming from conventional dimmers. This ensures sufficient cathode heating to allow the attainment of low light levels (1% for some models).
Figure 7.44 On the USA market the Lutron Tu-Wire™ ballast is an example of ballasts that can be dimmed by conventional dimmers without the need for additional control or maintained live wires.
siderable harmonics are developed and it is difficult to design a ballast that meets the standards for harmonics. In the USA some ES (Edison Screw) cap compact fluorescent lamps with built in electronic ballast are offered as being suitable for dimming by triac dimmers. This is convenient since it means they can be used to replace tungsten lamps in table lamps and standard lamps that have built-in triac dimmers. The dimming range is from 5% at best, and maximum lamp power is 23W. In addition some manufacturers offer dimmable electronic ballasts for T8 linear lamps. In Europe thyristor dimmable electronic ballasts for T8 and TC-L fluorescent lamps have been available for many years; however in order to meet EMC standards, the ballasts each use two substantial filter chokes. One choke is fitted in the live feed and one in the neutral. While the combination of two chokes and an electronic ballast seems somewhat clumsy, the convenience of being able to use standard dimmers is helpful. This is particularly the case where standard multi-channel dimmers are being used with predominantly tungsten or transformer fed tungsten halogen lamp loads. For example a prestige conference room might have only one or two fluorescent lamp circuits, but several tungsten halogen circuits. It may
239
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Input voltage Input current Input frequency Output voltage Output current Output frequency Short circuit current Power factor Neon tube 10mm ‡ Argon tube 10mm ‡ Neon tube 20mm ‡ Argon tube 20mm ‡
Nominal 25mA ballast 230V 350mA 50/60Hz 8000V 23mA 25kHz 29mA 0.95 8m 11m 11m 15m
Nominal 50mA ballast 230V 350mA 50/60Hz 4000V 46mA 25kHz 58mA 0.95 5m 7m 7m 12m
Figure 7.45 A European fluorescent luminaire fitted with an electronic ballast suitable for dimming with a thyristor or triac dimmer. Notice the two line filter chokes.
well be easier to use standard dimming equipment throughout, and use dimmable ballasts for the few fluorescent lamps involved. The same argument applies in some entertainment and theater lighting schemes, where it is easier to connect fluorescent lamps direct to the outputs of the dimmer system than to make special control provision for them. 7.2.10 Electronic ballasts for cold cathode and neon lamps Electronic ballasts for cold cathode lamps are similar to ballasts for hot cathode lamps in that they use the half bridge topology, and have active power factor correction. The principal differences are that there is no requirement for cathode heating, and that the open circuit output voltage is significantly higher. The high output voltage is achieved using a transformer at the output, and like its 50/60Hz counterpart it can be of leakage reactance construction to limit the output current. Examples of cold cathode electronic ballasts are shown in Figure 7.46. The ballasts illustrated in Figure 7.46 are not suitable for dimming. It is possible to obtain controllable cold cathode ballasts using the 1–10V control signal, and these work in a similar manner to their hot cathode equivalent, but without the complication of cathode heating. An example is shown in Figure 7.47.
240
Figure 7.46 25mA and 50mA electronic ballasts for neon and cold cathode lighting from Tunewell Transformers Ltd. Tube lengths reduce by 0.5m for each pair of electrodes.
There is a type of neon or cold cathode ballast referred to as “low voltage”. In fact this is something of a misnomer, because the lamp tube itself still operates at high voltage. The ballast device has a high frequency output at around 24V. This is coupled to the lamp using a transformer “boot”. The arrangement has the advantage of confining the high voltage to the immediate vicinity of the electrodes, and therefore simplifies the wiring, especially in multi-lamp systems. This arrangement has an interesting feature. If the high frequency power to the boot is ramped up from zero to full, the discharge in the lamp moves
Figure 7.47 The Coldstar II ballast from AC/DC Lighting Systems Ltd allows control of cold cathode lighting from a 1–10V control signal, bringing it into line with other contollable fluorescent lamp ballasts.
ELECTRONIC COMPONENTS
Figure 7.48 The Smart Neon™ controller from Fluid Light Technologies Inc provides sequence control for four neon tubes; the output is 4 × 24V 36kHz, which is coupled to the tubes using transformer “boots”. Varying the power applied to the boot propagates the discharge through the tube.
down the tube. The speed of propagation is proportional to the rate at which the power is varied, so neon signs can appear to “write” in light. Controllers are available to sequence multiple tubes (see Figure 7.48). Electronic ballasting of neon/cold cathode tubes has another possibility, that of obtaining more than one color from a lamp. The excitation of different gases is affected both by the frequency of excitation, and the waveform. As examples neon is excited better by a short duty cycle, but steep pulse
Figure 7.49 Here a 15inch LCD display is being serviced (left). The inverter for the CCFL backlight is mounted on a separate printed circuit, shown enlarged on the right.
rise-time, and mercury vapor is better excited by a longer duty cycle and slower rise time. If a lamp tube is filled with a mixture of neon and mercury vapor, and if it has a phosphor coating, then it is possible to achieve a range of color. The neon will produce red, but the mercury vapor (which is actually producing ultra-violet) will produce a color that depends on the phosphor. By varying the frequency, rise time and mark-space ratio a tube can be made to change color, for example: • from blue to red • from green to yellow • from turquoise to mauve Small cold cathode lamps as described in Section 3.6.6 present a challenge to the ballast designer since in many applications power and space are scarce. Most ballasts of this kind are supplied as OEM (Original Equipment Manufacturer) items, and are often highly customized to meet the needs of the particular application. For example backlight inverters for LCD screens used in laptop computers, and dashboard and navigation display lighting systems used in automobiles. Figure 7.50 shows some examples. Despite the very low powers involved, the
Figure 7.50 Examples of inverters for cold cathode compact fluorescent lamps. These are the LXM1617 PanelMatch™ series from Microsemi, intended for LCD backlighting. Typical power ratings are in the range 2.5– 6W, operating from 3.3, 5 and 12V supplies.
241
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
starting voltages are considerable, even a 2.5W ballast may need to develop 1,500V from a supply of only a few volts. Lamp brightness can be controlled in various ways, including supply voltage adjustment and frequency control. However the technique that gives the widest dimming range, and the highest efficiency is PWM. Small cold cathode lamps are sometimes required to operate at very low temperatures. Those used in automotive and similar applications may include an arrangement to boost the output at low temperatures before switching over to PWM control. Driving the lamp continuously at 50% above the normal amplitude for a minute or so helps get the lamp up to full light emitting temperature. 7.2.11 Infra red modulation Notwithstanding its general efficiency, the fluorescent lamp radiates a considerable amount of infrared (Figure 3.14). When such a lamp is run from an electronic ballast, the lamp’s radiation is modulated by the high frequency supply and its harmonics. This does result in some practical problems. Most cordless remote controllers for TV sets and video recorders, and also some room controllers, use coded infra-red signals, and operate at frequencies in the range 33–40kHz and at 56kHz. They can be affected by fluorescent lamps running on electronic ballasts. Some efforts have been made to use lamp frequencies that do not interfere, but more significant progress has been made on the controller side, where new modulation schemes have made them less susceptible to interference. Another class of equipment that can be affected is cordless audio equipment; for example cordless headphones for TV and HiFi; and multi-channel simultaneous interpreting equipment, both of which also use modulated infra red. Typically such equipment uses a number of carriers in the 95–250kHz range. It can be affected by the IR from fluorescent lamps, especially those working at the higher frequencies required for dimming. Headphone manufacturers have got round the problem by going to much higher frequencies. Some
242
of the lower frequencies used in the simultaneous interpretation application may prove unuseable in the presence of some luminaire/ballast combinations.
7.3 Electronic ballasts for HID and arc lamps 7.3.1 Introduction Electronic ballasts for fluorescent lamps have been available for many years and it is now the case that for many applications it can be shown that their use has an overall economic benefit – quite apart from any consideration of light quality and energy usage. The situation with HID lamps is different. Electronic ballasts are only used in a minority of mainstream HID applications, although their use is now increasing. The economic advantage is not so clear cut, but as prices come down may become significant. In some applications, for example projection, film lighting and automobile headlamps, electronic ballasts are the only practical method of powering the lamp. At frequencies higher than normal mains supply frequencies, HID lamps suffer from an instability problem referred to as acoustic resonance. Under certain circumstances a near instantaneous change in lamp power causes a rapid fluctuation in plasma temperature. The direct relationship between gas temperature and pressure creates a corresponding pressure wave in the arc. Under these circumstances the arc tube can behave like an organ pipe, since the discharge is between two fixed electrodes. The resulting standing wave can distort the arc; this affects its color, inten-
Figure 7.51 Examples of electronic ballasts for HID lamps.
ELECTRONIC COMPONENTS
sity and electrical characteristics. It may also extinguish the arc if a critical frequency is hit. The acoustic resonance effect sets in at around 1kHz, and is operative at all the “convenient” frequencies used in fluorescent lamp ballast circuits. There are three possible solutions to the problem, but only one is widely used. One possibility is to apply a varying frequency to the lamp, so that there is not time for a resonance to build up. In practice this technique needs considerable knowledge of the particular lamp, and it may not be practical for generic HID lamps. The ballast runs at a frequency of a few tens of kHz, chosen to match a window in the lamp’s performance known to be where acoustic resonance is at a minimum. The basic operating frequency is then frequency modulated at a low frequency (for example 100Hz) to ensure that resonance cannot start. Another possibility is to use a frequency above the acoustic resonance range; but this must be as high as 350kHz, and power components are both lossy and expensive at this frequency. However, transistor technology does now allow designs using this technique to be considered. At such frequencies any ballasting components are very small, so in theory very compact ballasts can be made. The most widely used method is the use of square wave low frequency drive, using a frequency in the range 90–200Hz, and it is this method that is now considered in more detail.
the other does not itself introduce high frequency products. A switching time of around one microsecond proves satisfactory in practice. An outline circuit for an HID electronic ballast is shown in Figure 7.52. The front end is similar to a fluorescent lamp ballast, providing a DC supply and ensuring a good power factor. The lamp circuit is based on a bridge inverter with opposing transistors switching together to ensure a square wave through the load. A capacitor prevents DC going through the load, and an inductor limits the current through the load. The lamp is shunted by an LC filter that ensures that high frequencies bypass the lamp. HID ballasts benefit from being able to integrate the ignitor into the ballast, and one method is to use the suppression choke as the secondary of the ignition transformer. A significant advantage of electronic ballasts for HID is that they can regulate the power to the lamp, making the lamp performance substantially independent of line voltage variations. This regulation
Filter
Rectifier
PFC
Bridge Driver
7.3.2 Ballasts for standard HID lamps The smaller sizes of high pressure sodium and metal halide lamps up to 400W are now well served by a range of electronic ballast products from many manufacturers, all using the square wave drive principle. The idea here is that in a square wave the power is constant, since (during any half cycle) the voltage is constant, so fluctuations are not set up. Two things must then be assured: • first that no high frequency due to active power factor correction or regulation gets into the load. This can be dealt with by simple filtering. • second that the switching from one half cycle to
Ignitor
Choke Filter
Lamp
Figure 7.52 Possible bridge inverter circuit for HID electronic ballast.
243
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
AC In
Boost Choke
Lamp Choke
Filter
Lamp Voltage Sense Lamp
Sidac Ignitor
Lamp Current Sense
Boost Converter PFC Control 50KHz
Regulator PWM Driver 25KHz
Bridge 200Hz Driver
Figure 7.53 Alternative circuit placing the ballast inductor prior to the bridge, showing the principle of power regulation.
can also counter the effect of arc voltage changing with lamp life. In practical MOSFET circuits the regulation is easier to do with a separate transistor, and by placing the ballast choke prior to the bridge as in Figure 7.53. Power control by the bridge itself is not an option since a square wave output is required. In Figure 7.53 the regulating MOSFET is driven by a circuit that receives input signals from the lamp circuit that correspond to lamp current and lamp voltage. It can, therefore, not only regulate the lamp
0V
Half Bridge Drive
Frequency Control 25KHz - 200Hz
+V
C L
Current Feedback
Figure 7.54 A half bridge HID ballast circuit using high frequency start.
244
power under normal conditions, but also shut down the ballast when a lamp failure condition is detected. Power is regulated using PWM, typically at around 25–30kHz. While practical circuits for the larger lamps, and for lamps using conventional ignition circuits, use the full bridge inverter, it is possible to use half bridge inverters, especially if a different method of iginition is used. Figure 7.54 shows an alternative method of ignition suitable for small HID lamps. The circuit operates under two different frequency regimes. When power is first switched on the system operates at high frequency (above 20kHz). The frequency is chosen to match the resonant frequency of the LC combination formed by the choke and capacitor across the lamp. This develops a sufficently high voltage to strike the lamp (the operation is similar to that of the circuit used in fluorescent lamp ballasts). As soon as the lamp strikes, the voltage drops because the circuit goes out of resonance. Once the arc discharge is established the system switches over to low frequency operation. Because the starting voltage is sinusoidal, it is practical to site the lamp some distance from the ballast. The connecting cable must, of course, be suitable for the starting voltage, typically around 1,500V, and of
ELECTRONIC COMPONENTS
low capacitance. Distances of around 25m are practical. All electronic HID ballasts include various safety features, in particular high temperature shutdown, and shutdown for failed lamps. Electronic HID ballasts are not suitable for connecting to normal dimmers. However some ballasts use their regulation circuit as a means of providing lamp dimming, typically over the range 50–100% light output. Electronic ballasts with a remote control input similar to that provided in controllable fluorescent lamp ballasts are becoming available – but such ballasts can not compensate for color temperature shifts arising from lamp dimming. For higher power HID lamps a hybrid circuit can be used. Here the lamp is ballasted by two reactors in series; however the second reactor has a thyristor (or equivalent) regulator across it that varies the amount of current that goes through or by-passes the reactor. This arrangement can be used both to provide superior regulation, overcoming the main disadvantage of the simple reactor, and to vary light level. An example is shown in Figure 7.60. 7.3.3. Benefits of electronic ballasts The electronic HID ballast has small size and light weight compared to the simple electromagnetic equivalent. It has other advantages. There are some savings due to increased circuit efficiency. In percentage terms these are most significant for the low power (30–70W) lamps.
Figure 7.55 Ballasts for low power HID lamps using the principle of high frequency starting are available from Harvard Engineering. Lamps can be up to 25m away from the ballast.
Figure 7.56 Examples of portable ballasts from Powergems Ltd for high powered metal halide lamps used for film lighting. These include a 50–100% dimming facility.
Its excellent regulation means that not only is light output maintained when the line voltage varies, but, just as important, color temperature is also maintained. The light output through life is better maintained. This is probably due to the reduced crest factor arising from the use of square wave drive, and to a more benign starting regime. Some classes of HID lamps suffer from a considerable (~ 40%) drop in light output over life, and the use of electronic ballasts significantly improves the situation. Many OEM electronic ballasts are made for projectors, theatrical luminaires, fiber-optic illumination systems, automobile headlights, film studio lighting equipment and other specialist applications.
7.4 Electronic transformers 7.4.1 Introduction In the lighting world an “Electronic transformer” is a device for providing power to low voltage tungsten halogen lamps. They now command a large part of the market, primarily at low power, serving one to three lamps. In common with their electromagnetic cousins, electronic transformers are designed as SELV devices. In the electrical engineering world: LV or Low Voltage refers to circuits operating at 1,000V or below. ELV or Extra Low Voltage refers to circuits operating at less than 42V r.m.s. FELV or Functional Extra Low Voltage implies
245
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Current Transformer
Output Transformer
Diac
Lamp AC In
Figure 7.57 Outline circuit of an electronic transformer.
that one side of the ELV supply may be grounded or otherwise referenced. SELV or Separated (also Segregated) Extra Low Voltage means that both legs of the ELV supply are “floating”. This is much the safest arrangement since it means that no other electrical circuit will create a current by accidental connection. If, for example, one side was grounded (to the same potential as neutral in a mains system) then any “live” wire touching any part of the ELV system would cause a current to flow. In practice EMC requirements mean that electronic transformers are not quite “pure” SELV
devices because of the presence of suppression capacitors. Because the final load is resistive, the electronic transformer is nowhere near as complex as the fluorescent lamp ballast, although the basic voltage conversion principle is the same. The simplifications arise because: • there is no requirement to pre-heat the lamp or develop high starting voltages. • the load is conventional, and does not need a current limiting choke. • there is no need for active power factor correction, because there is also no need for a high voltage rail that itself is higher than mains voltage. 7.4.2 Electronic transformer circuit The circuit for a typical electronic transformer for feeding nominal 12V tungsten halogen lamps is shown in Figure 7.57, and the photograph Figure 7.58 can be directly related to the circuit. The circuit has much in common with Figure 7.23. The incoming AC is full wave rectified, and it is this waveform that powers the half bridge. Thus the current and voltage taken from the supply are in phase, and in principle power factor and harmonics problems are reduced. Nonetheless a choke must be fitted to prevent high frequencies getting back up the supply, and there are also other EMC
Figure 7.58 An electronic transformer from Kaoyi Electronic Co. using the principles of Figure 7.57. This one is rated at 120W.
246
ELECTRONIC COMPONENTS
Osram Halotronic™ “Mouse” 70–105VA
IBL IP65 rated 35–105VA
IBL SM range 60–105VA
Relco 60VA
Figure 7.59 Examples of electronic transformers for low voltage tungsten halogen lighting.
components (capacitors) required. The circuit self oscillates, using a current transformer in the output to feed back the transistor base drive signal. Typically the transformer might have 1–2 turns in its primary, and 6 turns in each secondary. The bridge center point has an r.m.s. voltage approximately half that of the AC line voltage; thus for a 12V output the output transformer has a turns ratio of 10:1 on 230V supplies and 5:1 on 115V supplies. In the photograph the secondary has 10 turns, so the transformer has 100 turns on the primary. It is interesting to see that the secondary is wound with four wires in parallel; this is done to simplify the winding (since thick wires are difficult to wind on a small core) and to reduce skin effect losses. In common with their electromagnetic equivalents, electronic transformers are equipped with protective devices to guard against overloads. Since some of these add significantly to cost, it cannot be assumed that all the following will feature in any one transformer. Transient protection. Most electronic transformers have a voltage dependent resistor or VDR at the input. This device carries insignificant current unless a particular voltage is reached, and is used to clamp any voltage spikes on the supply line. Fuse protection. Some electronic transformers
are fitted with fuse protection; but most rely on a thin part of the printed circuit track at the input to blow as a “one time fuse”. This kind of protection is also used in fluorescent lighting ballasts, and would only operate if there was a catastrophic component failure within the ballast or transformer. Over-current protection. By detecting the current flowing in the low side driver, it is possible to shut down the transformer (or to limit the bridge output) in the event of the current exceeding the rated value. This means that suitably equipped electronic transformers are substantially short-circuit proof. Some work on the basis that once the cause of the over-current condition is removed, they start up again automatically. Others require that the power is switched off and then switched on again to restart the oscillator. Thermal protection. Similarly a temperature sensor, usually fixed to the output transformer, can be used to shut down the electronic transformer. Winding temperature limitations are the same as for electromagnetic transformers. Consideration of Figure 7.57 raises the question of what happens when the input voltage to the half bridge drops to zero, as it must do every half cycle. In principle the oscillator stops. In practice this is not a problem; however, it obviously cannot start again until there is sufficient input voltage to kick it into action again. Some electronic transformers have a small capacitor or other arrangement to provide sufficient energy to keep the oscillator going over what would otherwise be the voltage zero. It is a common requirement for electronic transformers to be connected to dimmers. It is essential to check with the manufacturer whether or not the electronic transformer is suitable for dimming. Connecting unsuitable transformers to dimmers can result in the destruction of the dimmer, or the transformer or both. In particular electronic transformers presenting any kind of capacitive load (due to the use of a holdover capacitor or similar) cause trouble on conventional triac or thyristor dimmers. Electronic transformers should include in their specification their suitability for dimming. Some are not suitable at all, some are suitable for trailing edge (transistor) dimming only, and some for leading edge
247
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
arises both because of the high frequency circuit operation limitations, and because of EMC problems.
7.5 Lighting control by electronic transformers and ballasts 7.5.1 Conventional lamps
Figure 7.60 The Philips “Dyna Vision” electronic ballast for 150W high pressure sodium lamps (left) provides 1– 10V analog control of light output (in practice power range 35%–100%; light level range 20%–100%). However, economic considerations mean that for 400W lamps it is at present better to use a hybrid electronic/electromagnetic system to achieve light level range 35%–100% for a power range 50%–100% (right.)
(triac/thyristor) dimming only. Many are suitable for both. When leading or trailing edge phase cutting dimmers are used with electronic transformers of the basic circuit of Figure 7.57, the whole system relies on the oscillator starting every half cycle. This leads to the obvious question of whether a separate dimmer is needed at all. The oscillator start circuit works in a similar way to the gate firing circuit used in simple dimmers (described in Section 8.3.3.1). By deliberately delaying the start, the output power can be regulated. Electronic transformers with a simple built-in dimming facility can be constructed by adding a variable resistor in series with the diac resistor. This is only suitable for local control with very short leads. The variable resistor is several Megohms and operates at mains voltage. Electronic transformers suitable for remote analog or digital control are becoming available. While it might seem attractive to construct such a device with a significant output power (so that the cost of the control element could be spread over a number of lamps) this is not practical because there are restrictions on output lead length. Manufacturers specify the maximum output lead length, usually <2m or even less. The restriction
248
With the exception of the mains voltage incandescent lamp, all other light sources require some kind of “load interface” between the mains supply and the lamp. The interface is required to do one, several, or all of: • providing the right lamp voltage. • limiting the current through the lamp. • providing the correct starting or ignition conditions for the lamp. • providing regulation such that variations in mains supply voltage do not result in significant variations in light output. • providing a facility for varying the light output. In Chapters 6 and 7 we have reviewed the common electromagnetic and electronic devices that provide this interface function. It is clear that the electronic variants offer more possibilities, and, in principle, offer the opportunity to include the lighting level control within the interface device. This topic is developed further in Chapter 8. However, there are some caveats. Some are economic, some technical and some practical. It is unfortunate that the amazing variety of light sources
Figure 7.61 Electronic “ballasts” for LEDs. The Xitanium™ range from Advance Transformer. These provide the required constant current output, and are available in controllable versions.
ELECTRONIC COMPONENTS
that is now available brings with it some uncertainty about compatibility between different items of equipment, and the best way to achieve level control. Technically it may seem attractive to use electronic transformers equipped to receive a separate level control signal; but in practice economics may dictate that it is simpler to use standard electronic transformers, and control the light level by a dimmer on the supply circuit. A lighting scheme may depend on achieving low light levels, but be designed round T5 fluorescent tubes. Technically such tubes are difficult to dim to low levels because their V-I characteristic is highly dependent on temperature. So for practical reasons the scheme may have to include some incandescent lighting to achieve a low light level reliably. At the practical level the simplification of installation may militate in favor of systems that do not require separate control cables, especially in the refurbishment market. A few rules arise out of the topics presented in this chapter: • no electronic ballast or transformer should ever be connected to a separate dimmer device (regardless of its nature) unless it has been designed to be so connected. • even if it has been designed for dimmer connection, there may be a restriction on the type of dimmer. • most self ballasted (i.e. compact lamps with builtin disposable electronic ballast) fluorescent lamps are NOT suitable for any kind of dimming or level control - but inevitably there are exceptions. • all kinds of discharge lamps have restrictions on the range of lighting levels they can be dimmed to, regardless of whether they are connected to electromagnetic or electronic ballasts. It is best to be highly conservative about achievable levels at the low end. Some published specifications are only achievable at specified lamp operating temperatures. The trend is towards more sophisticated electronic ballasts and transformers, providing good remote lighting control possibilities. Practical, technical and economic considerations may mean that
Figure 7.62 The “Power Pipe” from Artistic License provides outputs optimized for high efficency red, green and blue LEDs. The outputs are matched to the most common series-wired groups for each color. Level control is DMX compatible, making the unit ideal for entertainment and exhibit applications.
hybrid control solutions are best for some lamp ratings. 7.5.2 LEDs A whole new generation of lighting products is emerging as a result of the arrival of high output LEDs. The lighting industry is still deciding what to do about them, but it is likely that, as with present light sources, there will be “lamp” manufacturers making the LED devices themselves, and “interface” manufacturers making suitable power supplies. The difference compared to present practice, is that in many cases the “lamps” will be more closely integrated with their power supply. For many luminaire manufacturers it will be convenient to use ready made “ballasts” for multiple LEDs. An example of the kind of product available is shown in Figure 7.61. LEDs are current based devices. Operation from any voltage source requires current limiting resistors, which is wasteful, so electronic power supplies for LED lamps are based on current regulation. In principle they are the same as regulated electronic transformers with a DC output. Products of the kind shown in Figure 7.61 are available with dimming control, based on PWM. A choice of control protocols is offered; in the commercial lighting market these are likely to be 1–10V or DALI. In the entertainment market DMX is preferred, so products similar to the unit shown in Figure7.62 are more suitable.
249
Chapter 8
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Dimmers Part 4 – Dimmers and control systems The next seven chapters describe the technology of lighting control and control systems. Until comparatively recently this would have only needed to cover the technology of dimmers – but it is now the case that while the dimmer remains an important item, it is only an element in the overall concept of lighting control. Computers and micro-electronics have revolutionized the way in which lighting control is used in practice. Furthermore new component technology and new systems architectures are completely changing the way in which systems are configured. Our idea of what constitutes a “dimmer” is also in a state of fundamental change.
The reasons for using lighting control and dimmers in particular are explored in Chapter 10. Here it is sufficient to remark that, no sooner had mankind found ways of making light in the darkness, there was a requirement to control the amount of light. One way is to have a number of different lights which are switched on as more light is required. Indeed, the use of multi-circuit lighting for achieving different levels of lighting is widespread. But many applications require the ability to continuously change lighting levels, or to select any particular level. This can only be done by a dimmer. Whatever technology is used, the concept of a dimmer is simple. Some form of control, be it lever or rotary knob, is moved from a “zero” to “100%” position, and the light output of the luminaire varies accordingly. The light output can be changed either by having a lamp with constant output fitted with some kind of variable filter or shutter; or by varying the power supplied to the lamp. Of course the dimmer designer starts with a knowledge of lumens, and designs the dimmer so that his lever/knob control works in a way that the light output varies linearly with control position. Having checked the performance with a light meter, he is then somewhat puzzled that the performance of the system does not look linear.
250
100
8.1.1 Influence of the eye
If a stage spotlight is directed at a surface, or a projector projecting only white light is directed at a screen, the light distribution can be measured. Obviously one is hoping for uniform illumination, so photometer readings are taken at the screen or beam center, and at various points round the edge of the screen or beam. In practice it is found that quite wide variations of brightness do not seem to matter and, unless the difference in readings is around 50% the
Lumens percent
8.1 Introduction to dimmers
Power input percent
100
Figure 8.1 The relationship between power input and light output for some common light sources. Key: black line is tungsten lighting curve; red is fluorescent; blue is mercury vapor HID and green is metal halide. The dotted region of the blue line indicates a significant color change.
Lumens percent
100
Control setting percent
3200
Color temperature K
To the human factor we have to add other complications which get in the way of our convenient “linear” dimmer. No electric light source displays a completely linear relationship between power input and light output. Figure 8.1 shows the relationship for some common sources. In the case of discharge lamps there is a minimum practical level where arc instability sets in, so these do not go down to zero. Another complication arises in film and television work. Photographic film and TV cameras do NOT have “human eye” characteristics, and are much more fussy about the evenness of illumination. In some cases illumination which might appear quite adequate to us, appears as pitch darkness to the camera. The combination of the human and the technical results in the concept of the dimmer law, relating notional control position to power or light output. In practice there are many applications where a quite pragmatic approach is taken to dimmer laws. Provided the control knob, slider or equivalent is active along its length of travel, and provided the light variation is apparently reasonably linear, without sudden jumps, then it is likely to be acceptable. But in the theater, and in other multi-channel applications, more precision is needed. Besides precision, the relationship needs to be useful to the application. A widely accepted law in theater is the square law shown in Figure 8.2. The top part of the figure shows the relationship between control setting and lumen output. The bottom part of the figure shows what this means for a tungsten lamp (only) in respect of current through the lamp and color temperature. Figure 8.3 shows some other commonly used laws (related to tungsten lamps). The inverted square law has been favored for television because all the ac-
100
8.1.2 Dimmer laws
Current percent
eye hardly notices any difference in brightness! The eye needs quite big changes in light level to notice a real difference. In practice the effect of any dimming is also dependent on field of view, the color concerned, and, when multiple sources are being used, the interaction of such sources.
100
DIMMERS
Control setting percent
100
2200
Figure 8.2 Relationship of dimmer control setting to light output under the square law (top). The bottom graphs show the corresponding current (black line) and color temperature (red line, right hand scale) for a tungsten lamp.
tion is at the top of the curve. The cube law gives more action at the low end, favored by some theater designers. It is also the curve which, for tungsten lighting, most nearly matches a linear “r.m.s. volts to lumens” relationship. The S curve has also been favored, giving extended control at both ends of the range. Coincidentally this curve matches the performance of thyristor dimmers, where a linear change in firing angle gives the S characteristic in light output.
251
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Current percent
100
method where power electronic components, such as thyristors and transistors, are NOT used as the primary means of controlling the light output; but where the source of lighting is electrically powered. It therefore covers mechanical shutters and electrical impedance methods. 8.2.1 Mechanical shutters etc.
Control setting percent
100
Figure 8.3 Some other dimmer laws. Inverted square (black) cube (red) and S curve (blue).
In lighting control systems there can be a problem of where the “law” resides. Should it be in the dimmer? Or should it be in the remote controller if this is a separate device? How should the choice of law be made? Practical systems may also have offsets to match the requirements of particular sources. For example tungsten lamps emit no light at low percentage voltage; on the other hand feeding them a low voltage increases their resistance and improves their initial response. In the case of big lamps it can also reduce the effect of thermal shock when switched on. So some systems allow a “pre-heat” for tungsten lamps (for example 4–5% voltage). Similarly fluorescent and other discharge lamps have a cut-off, so it can make sense to expand the controller range to match only the effective dimming range of the lamp concerned.
8.2 Non-electronic dimming Today most dimming is achieved by electronics. However, there are still some applications where other methods apply. As a matter of semantics, this section regards “non electronic” dimming as any
252
Arc (or discharge) lamps are not suitable for dimming, or at best can be dimmed down to around 50% of output. Even here it is necessary to run the lamp at full output first. However, arc sources, in the form of compact source metal halide lamps, are widely used in entertainment luminaires, and in this case full range dimming may well be required. The only practical way to dim the light is to introduce a method of modulating the light directly. Various methods can be used: • the introduction of a continuously variable filter. This could consist of a film which is transparent at one end, and has a gradually increasing density along its length, through gray to black. • the use of a moving reflecting mirror with similarly graded properties of reflection. • the use of an iris, scissor or venetian blind type shutter in the projected beam. The last method needs some comment. In theatrical lighting fixtures there are usually elements which are deliberately focused, for example beam shaping shutters and gobos which carry a pattern or image. A beam limiting iris may also be included to limit or vary the extent of the beam. This iris does not dim the light, but changes the beam pattern shape. However, it is possible to insert an iris or similar device into another part of the optical path, which is not imaged. In the correct place in the optical path restriction of beam diameter has the effect of restricting illuminance, and, therefore, it can be used for dimming. Different strategies are used for shutter dimming, depending on the lamp design. Venetian blind shutters, bladed shutters and iris shutters are all used, and examples are shown in Figures 8.4 and 8.5. While in theory all the methods of mechanical dimming could be manually operated, in practice they are motor
DIMMERS
Figure 8.4 The “Eclipse” mechanical dowser from Wybron Inc. These examples work on the iris principle, and are suitable for fitting to luminaires and floodlights using HID lamps. All such products are now DMX control compatible.
operated. The motor control system is conformed, so that the whole device looks to the outside world just like a normal stage dimmer, with either analog or serial digital control. Mechanical devices for controlling light output are sometimes referred to as dowsers (alternative spelling dousers). This word is particularly applicable to single blade shutters used to dowse the light in a follow spotlight or cinema film projector.
off the resistance altogether (to ensure no current is carried in the zero light position). • the resistance windings are, in fact, made up using wire of five different resistivities. This is to help achieve a good dimming law. The construction shown was used for dimmers in the 1kW range. Bigger dimmers used a radial contact arrangement where the moving contact moved over a series of studs, between each of which was an incremental resistance. To achieve apparently “stepless” dimming, it was necessary to have at least 100 separate resistance elements. The resistance dimmer suffers from two disadvantages. The obvious one is that it dissipates, and therefore wastes, considerable heat. The heat dissipated is nothing like as much as that of the load, the maximum being around 30% of the “full on” load. In entertainment applications this is an inconvenience, but not a show-stopper; but clearly it rules out the use of resistance dimming for any application requiring continuous running. The second problem is that dimmers are load specific. A 1,000W dimmer is really only suitable for 1,000W lamp load. In order to ensure visible light extinction and continuous control, the resistance needs to be three times the resistance of the lamp (at
8.2.2 Resistance dimmers The first dimmers to be used for lighting control were variable resistance dimmers, and they were still being manufactured into the 1970s. The principle is simple, the variable resistance in series with the load varies the current. A typical construction for a manually operated slider resistance dimmer of around 750W rating is shown in Figure 8.6. Two sets of resistance windings are wound on parallel ceramic formers. A brush contact attached to the control knob, itself on a slide rail, connects the two parts of the resistance together; the whole arrangement being in series with the lamp load. Some less obvious features are: • at the top of the slider travel the brush hits two contacts to ensure that all resistance is shorted out. • at the bottom of the travel, the brush contact comes
Figure 8.5 Earlier versions of the “Eclipse” range used a venetian blind arrangement. The ones shown are intended for 1kW, 2kW and 5kW metal halide floodlights. Photo from Wybron Inc.
253
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 8.6 Example of a Strand Electric resistance dimmer from the 1960s. Notice the varying thickness of the resistance wires. Photo © Victoria and Albert Museum.
full voltage.) In practice it was found possible to use dimmers at around ± 30% of nominal rating, but, of course, when used at anything other than nominal rating, resistance dimmers would give a different, or displaced curve. Small resistance dimmers, like the one in the figure, were directly operated. Larger ones used tracker wires to allow remote (mechanical) control. Motorized resistance dimmers were also widely used (especially for auditorium lighting). Resistance dimmers are now obsolete. However resistance as a means of current limiting may still be found where DC carbon arc lamps are used.
But a variant, called the saturable reactor was widely used because it had no moving parts in itself and was suitable for remote control. The principle is shown in Figure 8.7. The lamp load is in series with the reactor, an inductance with three windings. The two main windings are in series, but wound in opposite sense. This arrangement ensures that, while they represent an inductive impedance, they do not induce a voltage by transformer action in the third winding. The third “control” winding is connected to a variable source of DC. With no control current flowing, the main circuit experiences the full impedance arising from the inductor. As the DC current is increased, however, the iron core saturates. When it is fully saturated, the main coils cease to provide inductive impedance. In the fully saturated position, there is still around 5% volt drop across the device due to the resistance of the windings. A saturable reactor designed for controlling a 3kW 240V load only needs around 250mA of control current. By using a transistor amplifier in the control circuit, it is possible to get the actual control current required down to a few mA – and this technique made it possible to introduce multi-scene preset controls (described in Chapter 11). In practice the saturable reactor was an effective and comparatively inexpensive way of achieving remote controlled dimming. In its standard form it was
230V
Lamp
8.2.3 Reactor dimmers In an AC circuit it is possible to regulate current by means of a variable reactance, in practice a variable inductance, achieved, for example, by a magnetic circuit with variable air gap. This has the advantage of being less wasteful of energy than a variable resistance. However, the mechanical construction of such a device is expensive and, in the days when dimmers were mainly used for entertainment applications, there was no significant advantage in such a device.
254
Low Voltage DC
0 - 250mA
Figure 8.7 Principle of the saturable reactor.
DIMMERS
less load dependent than the resistance dimmer; but when lightly loaded could be slow to respond, and would not always dim to extinction. More elaborate systems used current feedback to improve the load independence of the device. 8.2.4 Transformer dimmers Variable autotransformers as described in Section 6.1.7.4 can be used as dimmers. Their attraction is that the output is sinusoidal, so they do not introduce harmonics. However, they are larger, heavier and more expensive than their electronic counterpart and their use in lighting control is now limited. They may still be found as single channel motorized units for simple auditorium lighting control. They are also widely used in high power voltage regulation systems, where lighting may form only part of the load. Small manually operated low power units are used for lighting control in some laboratory and control room installations where the absence of harmonics is important. Autotransformers are also used in some “Lighting circuit power reducers”, and this application is discussed in Section 14.3.
Figure 8.8 Example of a Strand Electric saturable reactor used in the 1950s and 1960s for stage lighting control. Photo © Victoria and Albert Museum.
Figure 8.9 Strand Electric dual channel motor driven autotransformer used for stage lighting control in the 1950s. Photo © Victoria and Albert Museum.
Prior to the arrival of the thyristor (and its vacuum tube predecessor, the thyratron) variable auto transformers represented the acme of professional lighting control in the theater. A number of special constructions were developed. Such transformers pre-dated toroidal construction; they were made on rectangular cores. This allowed a linear moving contact and, because of the dimming law shown in Figure 8.1, part of the variable transformer (about 14%) did not include the moving contact. It was realized that the core and winding construction represented the greatest cost element, and that it was actually possible to have a single autotransformer coil/core combination with several moving contacts. In Frederick Bentham’s book “The Art of Stage Lighting” of 1968 he refer (nostalgically?) to European practice going back to the 1920s that provided as many as 48 channels of dimming off a single core. This multi-channel transformer system was the Bordoni system. It lent itself well to remote control by mechanical tracker wire, but not to motorization. Strand Electric, for whom Frederick Bentham worked, and who himself was responsible for many of the fundamental concepts of lighting control,
255
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 8.10 A theater dimmer room of the 1950s. Here 200 5kW Strand Electric variable auto transformer dimmers with electromechanical servo drive are installed. Photo © Victoria and Albert Museum.
decided that a two channel autotransformer represented the most practical basis for a remote controllable dimming system. This allowed a simple arrangement for servo operated motor drive. The significance of this was that any dimmer channel not required to move simply stayed where it was; any that did need to move, did so under motor drive. Figure 8.9 shows a close-up of such a dimmer and Figure 8.10 a photograph of a dimmer room with 250 motor controlled dimmer channels. In Figure 8.9 the feedback potentiometer can be clearly seen at the top of the photograph. This was in a Wheatstone Bridge circuit. If the bridge was out of balance a clutch would engage on a rotating shaft to move the autotransformer to the setting that balanced the bridge. In Figure 8.10 the shaft drive motors can be clearly seen. Reminder AWheatstone Bridge is shown in Figure 8.11. The conditions for zero current through the load are that R1/R2 = R3/R4 since, when this is the case the potential difference between the two ends of the load is zero. A servo system suitable for motorized dimmer control can be based on the idea that R1 and R2 are represented by a control potentiometer, and R3 and
256
R4 are represented by a feedback potentiometer. If the bridge is out of balance, the out of balance current can be used to operate a polarized relay that operates clutches that mechanically connect the moving contact to drive it to the required position. It is also possible to construct AC bridges where the resistances are replaced by impedances. The bridge configuration is also fundamental to rectification and power control. Frederick Bentham confirmed that the most difficult part of the design was the transformer brushgear. Too compact, it had a high resistance, too wide, it caused an excess of shorted transformer turns. Nonetheless, the performance of motor-driven transformers as theater dimmers was impressive – indeed if one was to take a time machine back 55–65 years ago one would have seen lighting control performance on a par with today’s (excepting the different requirements of today’s lighting design).
Wheatstone bridge A=0 when
R1 R2
=
R3 R4
R2
R1
A L
R3
R4
Principle of bridge servo Feedback potentiometer
Polarized Relay to give forwards, backwards or stop at bridge balance
Control potentiometer
Figure 8.11 The Wheatstone Bridge. This shows the DC bridge based on resistance. AC bridges based on impedance are also possible.
DIMMERS
Figure 8.12 A 60kW thyristor dimmer set used for cinema auditorium lighting control. One of the first automatic dimmers on the market that did not use electric motors to control the fade. Picture from Electrosonic Ltd. 1964.
8.3 Thyristor and triac dimmers 8.3.1 Introduction See Section 2.4 for definitions, principles of operation etc. that are assumed in this section. Thyristors began to be used for lighting control in the early 1960s and for 40 years they have formed the basis of professional lighting control. When they first appeared they were expensive, and somewhat fragile, so they were only used where their technical advantages outweighed their disadvantages. For example the big cinemas of the 1960s, seating maybe 1,500 patrons, would previously have been fitted with enormous motor-driven “sunset” resistor dimmers that required a whole room to themselves. The thyristor dimmer occupied a fraction of the space (Figure 8.12). The particular example shown in the figure was one of the first to use all electronic control to provide the automatic dimming function – most other automatic thyristor dimmers of the time used motor driven potentiometers. Although it must be said that a patent of 1937 for thyratron dimming anticipated the same principle.
By the late 1960s/early 70s the price of thyristors had come down, and their robustness had improved to the point where thyristor dimming became the norm. Thyristor and triac dimmers are available in a wide range of packaging and price, and at first it is not obvious why a 1kW dimmer should be offered at prices that may differ by a factor of 30 or more. Some of the reasons are as now follow: Dimmers intended for consumer applications are triac based, since this reduces the number of components. However, professional dimmers intended for entertainment applications or for long life permanent architectural installations, more usually employ back to back thyristors. This is because the thyristor pair is inherently more robust, and is better able to withstand high inrush currents and the high currents caused by tungsten lamp filament failure. Many consumer dimmers are “all in one” units, embodying both the power control triac and its control potentiometer. The entire circuit operates at mains supply voltage and uses the minimum practicable number of parts. Professional dimmers are normally remote controlled, requiring safety isolation between the mains voltage power element and the remote control circuit. The construction of professional dimmers is mechanically more complex. Sometimes they have a modular “plug-in” construction; or they may have extra space for connection, socket outlets, rail mounting or other features targeted at a particular application. Dimmers for the consumer market must meet EMC requirements, but otherwise have minimal filtering. Professional dimmers have more extensive filtering, especially for applications like TV studios where tungsten filament “lamp sing” can be a problem. The most critical aspect of the professional dimmer is consistency between units. A single switchplate dimmer for use in a living-room generally has no critical control requirements. Provided the control knob gives a reasonable control law, it is not necessary that one dimmer perform in exactly the same way as another. But when multiple dimmers are used, it is important that 50% on Dimmer A
257
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
is achieved with the same control setting as 50% on Dimmer B. Dimmers with matched control characteristics have more complex control circuitry. Dimmers for pure resistive loads can use simple firing circuits for the triac that only give a firing impulse. Dimmers intended for electromagnetic transformer loads, or for electromagnetic ballasts must use hard firing. Most professional dimmers use hard firing to ensure that they can control all normal types of load, including very low power lamps that might otherwise not provide enough current for the thyristors to latch. Professional dimmers may have to be designed to withstand perturbations on the mains supply that otherwise might introduce flickering of the lights. Such perturbations include: • audio frequencies superimposed on the mains, that can introduce false “zero crossings”. Such superimposed frequencies are used in ripple control, a method of remote controlling street lighting etc. • supply frequency variations. These are especially prevalent on ship-board installations and with other “non grid” electricity supply situations. • distortion of the mains supply due to the presence of multiple dimmers. This can be a problem on high impedance supplies.
main varieties of this rail, an assymetric “G” section rail, and a symmetric “top hat” section rail. The assymetric version is now only used for terminals, so dimmers intended for rail mounting are designed to clip onto the symmetric DIN rail. Individual remote control dimmers would, until recently, have only had one of two types of remote control. A “manual” dimmer would be controlled by an analog voltage of 0–10V or 1–10V. An “automatic” dimmer would be intended for remote push button control, and would have as many control wires as preset levels plus a common wire. A new generaa
b
c
d
8.3.2 Dimmer construction From the above it is clear that dimmers are constructed in many different ways to meet the particular application. Figure 8.13 shows examples of low cost dimmers intended for the consumer market. These are for direct control and are supplied suitable for mounting onto standard wall boxes, or for installing in individual table lamps and similar light fittings. For many customers the finish and styling is as or more important than the functionality as a dimmer. In the case of dimmers on brass plates, the metalwork can cost more than the electronics! Figure 8.14 shows some individual dimmers intended for remote control. These can either be for wall mounting, or for mounting within equipment cabinets. In Europe a favored method of mounting items such as contactors, circuit breakers, timers etc within a cabinet is the use of DIN rail. There are two
258
e
Figure 8.13 Examples of switchplate and similar consumer product dimmers. (a) USA switchplate dimmers match the North American rectangular backbox, like this example from Lutron. (b) In Europe the single gang backboxes are square, as for this switchplate dimmer from MEM 250. Double gang plates may take multiple single dimmers, and be finished in a “traditional” style, like this unit (c) also from MEM 250. Most electrical accessory manufacturers have dimmers in their range, (d) shows dimmer units that fit in to MEM 250’s own grid system, but are also compatible with other manufacturers’ similar modular panel systems. Relco supply many models of dimmers for in-line use with table and floor standing lamps (e).
DIMMERS
tion of digitally controlled unit dimmers is now available that allows for more sophisticated control facilities. Remote control protocols and standards are discussed in Chapter 9. Most professional applications of lighting control are multi-channel. It can, therefore, be more economic to site multiple dimmers in a single case. Often a single control circuit board carries the electronics for several dimmers. Figure 8.15 shows some multi-channel dimmers intended for permanent installation. The special consideration here is ease of installation, so such dimmer packs include generous terminal provision for the outgoing circuits. Some units have built-in circuit breakers; others, especially those that are to be built into cabinets, rely on separately mounted MCBs. Dimmers for touring applications are often designed as multiple units suitable for mounting in 19 a
a
b
c
b d
c
d
Figure 8.14 Examples of remote control dimmers. (a) wall mounting 16A professional remote control dimmer from Helvar. (b) Lutron 48A (3 × 16A) remote control dimmer – 110V supplies mean higher current for a given power. (c) DIN rail mounted 800W remote control dimmer from Helvar. (d) 750W remote control dimmer from Hüco for building in to luminaires, furniture etc.
Figure 8.15 Multi-channel dimmers for permanent installation. (a) Strand DE90 for up to 24 channels. (b) Lutron GrafikEye 6000 for 8–24 channels. (c) Dynalite DTK910 for 12 channels. (d) Helvar Ambience for 4 channels. Dimmers of this kind are usually rated at between 10 and 16A per channel. Some systems (like the Strand DE90) allow for a higher power module to replace a multiple of lower power modules; e.g. a 50A module to replace a 4 × 15A module. Others (like Helvar Ambience) allow channels to be paralleled to achieve a higher single channel rating.
inch racks. (The EIA, Electronic Industries Association, 19 inch rack is now accepted as a worldwide standard, notwithstanding the fact that it is inch based and its increments are 1.75 inches!) Such equipment needs forced air cooling. It is usually fitted with either standard power outlet sockets for each channel or with a large multipole connector for connecting
259
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
b
a
c
Figure 8.16 Multi-channel dimmers for rack mounting. Examples are (above) Rakpak from Pulsar, (below) Micropack from ADB. Channel rating is usually 10A per channel, but 16A and 25A variants are also available from some manufacturers.
all the output circuits simultaneously. Figure 8.16 shows some examples of this kind of equipment. Professional dimmer equipment for the top end of the market is usually constructed on a modular principle. Each module represents one or two dimmer channels. The modules are independent, and are designed to be easily replaced, either by plug connection or by easily re-wirable terminals. It is usually possible to exchange modules while the system is “hot”, although it is always best to isolate the individual module first. Figure 8.17 shows some examples. Today all multi-channel dimmer equipment uses some form of multiplexed control to minimise the control wiring needed. Many systems also offer the option of simple analog control, or of multiple multiplexed control, with either fixed or programmable rules as to which control input takes precedence.
260
Figure 8.17 Professional modular dimmer racks. Racks such as these are assembled to order, and may include many special features. Channel ratings are from 10A to 100A. The examples shown here are the SLD system from Strand Lighting (a), the Sensor installation rack from ETC (b) and the Imagine system from Helvar (c).
8.3.3 Analog dimmer circuits 8.3.3.1 Simple triac dimmer All thyristor and triac dimmer circuits require that the device is triggered at some pre-determined point after the sinewave zero-crossing. The technique is referred to as “phase control” or “phase cutting”. The simplest possible dimmers use a triac as the power control device, and a diac to trigger it. A diac is similar
DIMMERS
L (a)
Lamp
R1
C2
Diac
VR 1 C1
L Lamp
001500110018PH
(b) R1
10k
C2
0.15Pf
VR 1
470k
R2
Diac
10k
C1
47nF
C3
47nF
Figure 8.18 The simple diac-triac dimmer (a). At (b) the additional components reduce the hysteresis effect. Actual component values will vary according to the components used. Values here are typical for a 230V triac dimmer.
to a triac in construction, but has no gate. It is designed to conduct through breakover at a specified voltage. Diacs used for lighting control breakover at around 33V, but must break down symmetrically – usually specified as being no more than ±3V. The diac is rated to deliver the required gate current to the triac it is controlling. Devices known as quadracs are available where the diac and triac are on the same substrate. Figure 8.18(a) shows the extremely simple circuit. The choke L and capacitor C2 are required for EMC purposes. The remaining RC combination provides the phase delay that determines the time to breakover, which happens when C1 has charged up to the breakover voltage. Such a circuit has a hysteresis characteristic that makes it unacceptable for all except the lowest cost applications. If the variable resistance is turned to its maximum value so the lamp goes to extinction, and
Figure 8.19 A 100–300W dimmer from Relco, designed for fitting in the stem of a table lamp. It uses a circuit similar to that of Figure 8.18 with few components.
then turned back again, nothing happens for the first part of the travel. Then the lamp comes on at a brightness greater than would be expected by the setting. The reason for this is that the capacitor C1 has a quite different charging gradient depending on whether the dimmer output is being increased from a lamp off state, or decreased from a lamp on state. (Bearing in mind also that we are looking at an AC circuit where the voltage is changing direction all the time.) The hysteresis problem can be alleviated by adding an additional RC network as in Figure 8.18(b). The phase shift is still provided by R1 VR1 C1 ; but the diac breakover is initiated by the charge voltage of C3 . The presence of R2 prevents significant discharge of C1 through the diac. Figure 8.19 shows a table lamp dimmer using a simple triac circuit. 8.3.3.2 Remote controlled thyristor dimmer Figure 8.20 shows the block diagram of a remote controlled thyristor dimmer. The diagram is the same regardless of whether the final circuit is digital, analog or hybrid. Protection is provided at the input of the dimmer. At this point there may need to be provision for a low voltage power supply to power the control
261
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Optional sub-circuit protection
Loads
Input protection EMC filter
Low voltage control power supply
Zero crossing detector
Thyristor pair or triac
Firing circuit
Optional feedback
From remote control Control circuit Optional remote report back
Figure 8.20 Generic block diagram of a remote controlled dimmer.
circuitry, although in some multi-channel dimmer designs the control circuitry is separately powered. The dimmer is fitted with filter components to meet EMC requirements; depending on the design the filter may be sited entirely before the power devices, or fore and aft of them. The power device(s) itself or themselves will either be a triac or, in professional equipment, back to back thyristors. Usually the input circuit protection also protects the outgoing load circuit, but sometimes there is provision for sub-circuit protection on the dimmer output. The control circuitry consists of three principal items: The zero crossing detector. Also, for reasons that will become clear, known as the “notch” detector. This detects the moment at which the incoming voltage sinewave passes through zero. For phase control this is the essential timing information, since the firing of the thyristors is timed from this point. Sometimes the detector circuit is in a low voltage section which receives its phase information from the secondary of a transformer. In such cases it may be necessary to make an offset to allow for the phase shift that can be introduced by the transformer itself. The firing circuit. This provides the firing signal to the thyristors. For all remote control applica-
262
tions the firing signal must be isolated from the control circuitry. The control circuitry will normally be of SELV or FELV classification. The firing circuit compares the zero crossing point time with an input level signal, and fires the thyristors after the appropriate phase delay. The control circuit. In an analog dimmer this is as simple as a voltage reference circuit, for example allowing the dimmer to be controlled by a 0–10V analog signal. In an analog automatic dimmer, the circuit includes ramped timing circuits to produce automatic fades to preset levels. In digital dimmers the input is a serial digital signal. As an option the dimmer may be fitted with sensors in the output circuitry that modify the control signal. For example these could detect over-current or over-temperature conditions, the signals from which could shut the dimmer down. While the basic dimmer circuit cannot produce an output voltage that is greater than the line voltage, it can provide some regulation at outputs lower than 100% by sensing the supply line voltage. As indicated in the figure, it is also possible for dimmer status information to be reported to a remote location. For example, if the dimmer is fitted with current sensing, the information could be used to
DIMMERS
(a)
V
(b)
Output Pulses
V
Sine wave Clean-up
+
Window comparator
50 or 60Hz Filter
PD
-
+
+ -
+ -
+
Output FET
-
Figure 8.21 A simple zero-crossing detector and notch pulse generator (a). Professional dimmers use a more elaborate three-stage circuit (b).
detect load variations arising from failed lamps. There are several different ways of achieving each of the principal circuit functions. Over the years the details have changed as components have developed; and integrated circuits are available that provide many of the facilities needed. What follows represents one way of achieving the required result. One way of detecting the zero crossing is shown in Figure 8.21(a). Here full wave rectified DC is applied to the base of a transistor, normally holding the transistor hard on. Only at the zero crossing does the transistor switch off, due to the forward voltage drop across the diodes. This results in a sharp pulse appearing at the transistor collector, the timing of which corresponds exactly to the notch in the full wave rectified sinewave. A professional dimmer uses a more sophisticated circuit shown in block diagram form in Figure 8.21(b.) The aim here is to ensure that no “glitches” on the incoming 50 or 60Hz supply cause the generation of spurious or mis-timed notch pulses. The incoming AC is applied to a differential amplifier to produce a pure sinewave output; any interference arriving on both lines cancels out. Distortion arising from the transformer is also eliminated. The sinewave signal is then put through a filter that is designed to pass only 50 or 60Hz, this could be an op-amp filter of the type shown in Figure 2.54. Then, and only then, is the task of notch detection undertaken. A
(a)
(b)
Integrator
Storage capacitor
Input notch pulses
+ -
Timing capacitor
Output ramp
(c) Ramp + -
Comparator Reference
Firing pulse to thyristor(s)
Figure 8.22 (a) shows the ramp that is required. (b) shows a ramp generator circuit that is triggered by the FET output of the circuit shown in Figure 8.21(b). (c) shows the output of the ramp generator fed to a comparator that generates the firing pulse.
263
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Firing pulses In
Live
Live
Main thyristor pair In
Out
Optoisolated thyristor Firing pulses In
Figure 8.23 Firing circuit for a thyristor pair, using optoisolated thyristor firing.
window comparator could be used. This device has the “purified” AC applied to two comparators. The other inputs of the comparators are connected to a voltage divider chain, such that there is a small potential difference between them (the PD determines the width of the notch window). The outputs of the comparators are connected to drive the gate of a FET. Provided the AC drive is at a higher voltage than the “window” PD, there is no nett output. But when the voltage drops below this level (i.e. near the zero crossing) an output is developed that switches on the FET. The firing circuit requires a ramped voltage for determining the delay. This can be a rising or falling voltage; in Figure 8.22(a) a falling ramp waveform of the type required is shown. The start of the ramp is timed to correspond to the zero crossing. One way of generating a ramp is to rapidly charge a capacitor, for example by using the output of the circuit in Figure 8.21; and then allow it to discharge through a resistor, choosing the RC values so that the voltage dropped to zero (or to a known lower voltage) over one half cycle time. A simple RC circuit would give an exponential curve rather than a straight line, so a real ramp generator uses an opamp as an integrator to provide a linear ramp. The firing signal to the thyristors is referenced to
264
a control voltage. For example 10V means 100% output, so there is no firing delay, 5V means 50% output requiring a 90° delay. So if the ramp generator is made to operate over the same range, there is a means of achieving a precise delay by matching the control voltage to the ramp. Figure 8.22(c) shows the input control voltage being fed to one input of a comparator, and the ramp voltage to the other. An output is produced when the ramp voltage falls to the control voltage, and this can be used as the signal to fire the thyristors. The combination of the control input circuit and the ramp generator can be made more sophisticated by the introduction of extra components to introduce: • deliberate non linearity to provide a preferred dimmer “law”. • top and bottom limits to match particular loads. • limitations on the speed of change (for example to introduce a deliberate “soft start” for large tungsten loads). The isolation between the firing circuit and the thyristors is achieved right at the point of firing. Before the advent of reliable high voltage opto-isolating components, isolation was done using pulse transformers. The firing circuit would produce a chain of pulses to ensure reliable firing of the thyristor. Today firing is done using optoisolated thyristors (for firing thyristors) or optoisolated triacs (for firing triacs.) Figure 8.23 shows a typical arrangement. The circuit shown is a regenerative circuit, in that any tendency for the conducting thyristor to switch off, when it should be on, forces the driver thyristor on which in turn keeps the main thyristor forced on – so “hard firing” is inherent in the design. The firing circuit is usually arranged so that the firing photodiodes receive a common signal from the comparator, Figure 8.22(c) every half cycle; but only receive their power signal every other half cycle; ensuring that only the forward conducting thyristor receives a firing pulse. 8.3.4 Digital dimmer circuits There is no merit in introducing digital operation into a thyristor dimmer unless it brings benefits. However, in the professional dimmer, digital operation
DIMMERS
N
N
Thyristor pair and filter
L Voltage & phase Low voltage power supply
Zero crossing detect Sub processor Display
Firing
Load
Temperature
Relay
Current
Micro Processor
Local controls
Analog input
Optional load isolation
Watchdog signal EIA485
UART (S)
Communications
Figure 8.24 Block diagram of a digital thyristor dimmer.
introduces many benefits including: • ensuring that the performance of a dimmer is consistent, so that in multi-channel systems all dimmers perform in the same way. • simplifying the provision of special facilities, for example law changing, temperature and current monitoring, and multiple control inputs.
Imagine thyristor module
Imagine dimmer module open (chokes visible) Figure 8.25 The Helvar IMAGINE™ system is based on twin channel digital thyristor dimmers. Each module is rated 2 × 10A with convection cooling or 2 × 20A with programmed forced cooling.
• Simplifying the use of serial multiplexed control. Some features of the digital dimmer can be achieved entirely in an alternative analog execution; however it is now generally easier to use a digital approach for all timing and control, while retaining analog signals for sensing. Figure 8.24 is a block diagram of a professional digital dimmer. The diagram is representative in showing a range of facilities. Not all dimmers include all the items shown and some dimmers may include features not listed here. It is based on the use of a microprocessor. Since microprocessors can be very powerful devices, some designs use one microprocessor to service several dimmer channels. The dimmer still needs the thyristors and their isolated firing circuits, and still needs a method of zero crossing detection. But it no longer needs a ramp generator because the firing timing can now be done by the microprocessor. Features identified in the diagram include the following: Current sensing. The microprocessor receives as an analog input the output current (via a current transformer). It can therefore detect a short circuit or excess load current and can both modify the dimmer output accordingly and report the condition. Voltage sensing. Part of the control power supply can be arranged to follow the mains voltage, and the signal applied to another analog input to the microprocessor. This can be used to provide regulation; i.e. to counter the effect of mains voltage variation at dimmer outputs that do not require a higher supply
265
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
input than that available. The same input can be used to indicate mains polarity, so that for each half cycle only the correct thyristor is fired. Temperature sensing. A temperature sensor can be fitted to the main heatsink. If an over-temperature condition is detected, the dimmer can reduce its output until it is corrected. An intermediate temperature reading might be used to start auxiliary fan cooling. Analog input. Many “digital” dimmers retain an analog control facility. This can be convenient for multi-purpose systems, for system testing and as a back-up means of control. Usually the analog input takes priority over other inputs, but priority rules can be set as required. Watchdog input. This is a signal input that causes the dimmer to ignore any other control signal inputs and go to a pre-set level. This type of facility is used either when a central control system fails, or when power is restricted. Digital communications. The microprocessor communicates with the outside world through one or more UARTs. In the block diagram shown there are two communication channels, one for outgoing and one for incoming messages. While in theory both could be carried on one line, in practice it is often easier to use two separate lines. The incoming communications could be to a proprietary protocol, or to an open standard like DMX or DALI. Some dimmers accept more than one serial input simultanously, and operate to a pre-programmed priority rule. Both input and output communications are physically carried on EIA485 (but in theory any other digital connection could be used). Local control and indication. It is usually convenient to have some indication of what the dimmer is doing on the unit itself. It may also be useful to have some manual controls on it that allow test operation of the dimmer, or the setting of default levels and dimmer laws. This can be done using additional I/O on the microprocessor, but may require a subprocessor or display controller if a multi-character display is used. In multi-channel systems the local set-up control and indication facility may be provided on a group basis. Auxiliary outputs. Some designs include provision for additional control. In particular there may
266
be a requirement for complete circuit isolation when the dimmer is in the “off” condition, and this is best achieved using a relay. The microprocessor can ensure that the relay only operates when there is no load current. 8.3.5 Filter circuits Because thyristor dimmers operate at 50/60 Hz the physical realization of the filters needed to meet EMC requirements is quite different than that required for electronic ballasts based on transistors working at a much higher frequency. Besides meeting EMC requirements, there is a further practical requirement arising from the sharply rising thyristor waveform. With tungsten lamps there is a tendency for the filaments to “sing” when the lamp is at any intermediate setting, with the problem being particularly noticeable on 110V circuits (due to the comparatively thick filaments) and at 90° conduction. Early efforts at improving dimmer performance in respect of interference and sing were based solely on the use of substantial chokes intended to increase the current rise time. Many USA dimmer specifications give a rise time (for example 400μS) and/or a dI/dt figure (for example 30mA/μS.) However once stricter EMC rules came in, especially in Europe, it became obvious that a more subtle approach is needed. The heavy choke solution did not help, indeed its size could actually make things worse. The aim is to ensure that electromagnetic intereference is limited both on the supply side and on the load side, so a circuit of the type shown in
Figure 8.26 In the ETC Sensor™ system the set-up controls are provided on a per-rack basis, using this Control Electronics Module. The associated dimmer power modules are comparatively “dumb”.
DIMMERS
L1B
L1A L
C1 250V AC
Thyristor Pair
C2 380V AC
N Supply
Load
Figure 8.27 Filter circuit for thyristor dimmer.
Figure 8.27 is needed. The filter choke is wound in two parts, one fore and one aft of the thyristor pair or triac. The suppression capacitor on the load side must be of substantially higher rating than supply voltage, because it has to be able to deal with back e.m.f.s arising from inductive loads. Critical to the success of the choke is the choice of core material. Powdered iron was favored because it did not saturate and helped achieve a long rise time; but it does not have a high enough permeability to deal with the initial edge, and on its own cannot deal with the higher frequencies. A higher permeability material also allows the use of fewer turns; this is helpful since it results in lower self-capacitance which works against effective inductance. In practice a hybrid core may be used. It has been found that if the EMC requirements are met and if, in particular, the initial rise and the “turnover” point are smoothed as shown in Figure 8.29, the problem of lamp sing is also mitigated.
Figure 8.28 110–120V dimmers carry double the current of their 230V equivalents for a given power rating. Suppression chokes are quite substantial as can be seen in this double dimmer module from ETC.
professional” equipment based on triacs sometimes use spring connectors to allow the triacs to be replaced in the field. All this tells us is that the products are not as robust as they could be; today it should not be necessary to sell professional equipment where power device failure is anticipated. The principles of circuit protection given in Section 1.5.2 apply to all dimmers. Dimmer manufacturers have to take a view about the circumstances of installation in deciding what circuit protection to fit within the dimmer, and what to specify or suggest as external protection. Most consumer dimmers are fitted with a fuse for circuit protection, but this may not always protect the triac in the event of a tungsten lamp failure because of the high arc current. Commercial, architectural and multi-channel
“Raw” thyristor waveform
8.3.6 Protection Modern thyristors are remarkably robust. The latest technology results in very compact devices in small packages that are electrically much tougher than their forbears of only a few years ago that were mounted in large “block” packages. Improved diffusion techniques in production mean that current peaks, such as those caused by tungsten lamp failure, are distributed right across the device instead of being concentrated in hot-spots. This in turn means that the devices can be permanently wired in to the circuit, increasing reliability. Here it should be mentioned that low cost “semi-
Soft turnover Filtered waveform
Soft start
Figure 8.29 An effective filter for thyristor dimmers ensures that the initial current rise, and the current peak are rounded. This is more important than the overall rise time.
267
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
dimmers are normally protected by MCBs, either mounted separately or as an integral part of the dimmer pack. The MCBs should be chosen to match the I2t characteristic of the thyristors, and there can be problems when external MCBs are supplied by an electrical contractor who may not be aware of the requirement. Some lower cost multi-channel dimmer units, both those intended for permanent installation and for rack mounting, are fitted with panel fuses for channel protection. Two safety points arise with this arrangement. If the fuses are replaced with the wrong kind (e.g. glass bodied instead of ceramic) the fuseholder is likely to be severely damaged when the fuse blows. Care must also be taken to ensure that the MCB backing up the fuses is of a rating that will protect the fuseholder and its wiring. In most electrical systems using a neutral at earth (ground) potential, dimmers are only protected in the live line; with complete two-pole isolation only being available from the distribution isolator. Dimmers used on electrical systems using floating or center tapped supplies (e.g. in parts of Norway, in some marine installations etc.) require double-pole MCB protection. In the UK and some other countries it is a requirement that dimmers intended for use with portable equipment, in particular stage luminaires as used for school and college stages, be fitted with RCDs for safety reasons. The non-sinusoidal nature of the thyristor waveform, and the resultant pulsating DC component when carrying out a fade, means that RCBOs specified as being suitable for pulsating DC should be used. The usual residual current safety limit of 30mA means in practice one RCBO can only protect 2–4 channels of dimming, depending on the dimmer design and channel capacity. The “per channel” cost of dimmers is now so reasonable that most installations have dimmers and their associated protection rated to match the loads. This gives the greatest flexibility in control, and makes the installation simple. Sometimes there is a requirement for one large dimmer to control a load that itself consists of several circuits, and in this case sub-circuit protection is fitted either within the dimmer or external to it. The sub-circuit protection may
268
be by fuse or MCB. An example of where this arrangement may be required is in a large ballroom chandelier, maybe carrying 4kW of tungsten lamps divided into four 1kW circuits. In very large systems operating from stiff low impedance supplies the consultant electrical engineer may call for additional protection, because the prospective fault currents are so high. All dimmers intended for permanent installation should, ideally, be convection cooled. However, this is not always practical in large multi-channel systems where dimmers are installed in high density racks. Dimmer racks requiring forced air cooling should only be installed at sites where it is known that routine maintenance service is available. The forced air cooling arrangement should be of the “intelligent” kind – i.e. a system that only runs the fan(s) when necessary; and which can automatically shut down or reduce the dimmer power in the event of fan failure or other overheat condition. 8.3.7 Harmonics Because of the steeply rising waveform, the thyristor dimmer is a prolific generator of harmonics. Although in practice this does not present a problem in normal use, there is a general move towards limiting the use of equipment that generates significant harmonics (as has already been described in Chapter 7, see for example Figure 7.14). Within the EEC there is recognition that the professional needs of entertainment and architectural lighting control require the use of thyristor dimmers, and that it would also be impractical to ban the use of small dimmers in the consumer market. The current position is that, provided equipment meets CE requirements in all other respects: • dimmers below 500W are exempt. • individual dimmers using phase control are not permitted unless they have a rating above 16A. • multi-channel systems with an overall rating above 16A are permitted. Within a large power network the harmonic effect of thyristor lighting control is no worse than that of the widespread use of uncorrected “energy efficient” fluorescent lamps with built-in electronic bal-
DIMMERS
lasts. The complex loads presented by motor equipment can have even worse effects. To some extent harmonic generation within a network can cancel out, but to the extent that it does not, it is possible to introduce harmonic reduction devices (see Section 6.1.7.2). 8.3.8 Special applications Reference has been made to the fact that with digital dimmers it is easy to change the dimmer law, either on a fixed basis, or, by using the communication system, on a dynamic basis. This can be done to match specific loads or the preference of the lighting designer. A very common requirement within a lighting control system is where the great majority of the channels are dimmers, but where there are also some “non dim” circuits (controlling, for example HID lamps). It can be convenient to control these through the main dimmer system instead of making separate provision for them. This is easily done by invoking a “switch” dimmer law, which turns the thyristor dimmer into a zero-crossing switch, thus minimizing harmonics.
is why the thyristor still represents the most cost effective solution for many applications. The problems are as follows. The thyristor has been developed to a high pitch of efficiency. Forward volt drop in the latest devices is as little as 0.8V, so even when the volt drop across the suppression choke is taken into account, circuit losses are very small. Transistors have a higher drop in the full on condition, but also in the non saturated switching region have high losses. Thus the more perfect one makes the switch-off curve, the bigger the losses. Finally transistors are delicate creatures compared with thyristors. To form the basis of practical dimmers they need a lot of support circuitry to ensure that they are not destroyed by circuit conditions. There is a choice between the use of IGBT and MOSFET transistors. Until recently MOSFET was not considered for the application because RDS(ON) was so high at high voltages that the circuit losses were too great to allow a practical product to be made. But this situation has changed with the introduction a new generation of MOSFETs with low RDS(ON) (see Section 2.3.3.3) and it is now quite practical to use MOSFET, especially in low power dimmers.
8.4 Transistor dimmers 8.4.1 Choice of transistor At first glance it would seem that the transistor would make a better power controller than the thyristor, because it can eliminate many of the problems arising from the thyristor waveform and its rapid rise time. The transistor can be used to implement trailing edge power control in place of the thyristor’s leading edge control. Whereas the thyristor works on an avalanche principle, the transistor works on a proportional control principle. This means that it is possible to allow the incoming voltage sine wave to start from zero, and then switch it off when required, but, instead of a sudden switch-off, to shape the switch off curve. In turn this means that if the switch-off is done gently enough, there is no need for any suppression choke. There are some snags with this scenario, and this
Figure 8.30 “Silent dimmers” from Relco use IGBTs. This one uses touch button control and is designed for fitting in a modular wall box. It is rated for 300W resistive and 250 VA inductive loads.
269
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
MOSFET has the big advantage of fast switching speed, which could be the key to completely new dimmer architectures. IGBTs have a low forward volt drop and form the basis of most medium power transistor dimmers. It is unfortunate that, in some markets, early IGBT based dimmers gained a poor reputation due to extensive device failure. This seems to have been due to under-estimating the protection needs of the device. In well designed dimmer circuits the IGBT is now a reliable component. Whichever type of transistor is used, the transistor only conducts in one direction. In order to make an AC power controller it is necessary to use two opposing transistors, each with a by-pass or free wheel diode. In the case of MOSFETs this comes free at low frequencies as the body or parasitic diode of the MOSFET (see Section 2.3.3.3). For IGBTs it is necessary to use a separate diode but, because this is such a common requirement, it is possible to get IGBTs with the diode already built in.
(a)
(b)
8.4.2 Examples of transistor dimmers Transistor dimmers are available for a wide range of applications, from consumer market to broadcast studio. They are promoted for their quietness. Because they have no chokes operating at mains frequency, and because the switch off waveform is rounded, they make virtually no audible noise. Most are advertised as being suitable for line voltage tungsten lamps or for low voltage lamps fed by suitable electronic transformers. Unlike their thyristor counterpart, transistor dimmers do not mind capacitive loads. However, many transistor dimmers are not suitable for conventional transformer (inductive) loads. As with all dimmer/load combinations it is best to check with the manufacturers concerned before specifying an untried combination. Figure 8.30 shows a typical low-cost consumer transistor dimmer for use with table lamps etc. It is promoted by its manufacturer for its quietness compared to its triac equivalent. Figure 8.31(a) shows a professional transistor dimmer. This is an exact counterpart of its thyristor equivalent shown in Figure 8.25. It is offered by the
270
Figure 8.31 Examples of professional transistor dimmers. (a) is a two-channel module unit in the same housing as its thyristor “sister” shown in Figure 8.25. (b) is a two channel module for rack mounting from Genlyte.
manufacturer as an alternative for use with electronic transformer loads and for noiseless operation. Its block circuit diagram is almost the same as the thyristor unit (based on the ideas of Figure 8.24) and it uses exactly the same microprocessor and commmunications system. Zero crossing detection is still required, and the same circuit can be used. The difference lies entirely in the power and driving sections. Figure 8.32 shows the IGBT power section of the dimmer. It must be remembered that now the drive circuit is a turn-OFF circuit, not a turn-ON. The control signal to the gate is shaped to provide the required turn-off slope. While it can be seen that both
DIMMERS
gates are operating at similar potentials, it is clear that the whole power section is operating at line voltage. This means that the entire driving section must be isolated from the microprocessor section of the dimmer. This is done by photo transistors working from a driver interface that itself is directly connected to the microprocessor. The box in the diagram marked “IGBT drivers” is a gross simplification. As already mentioned, great care has to be taken with IGBTs to ensure that they are not subjected to conditions that could damage them. The driver circuit includes not only the means of shaping the switch off waveform, but also very fast detection of out of range current and voltage to ensure that the device is switched to a safe operating region in the case of a fault condition. The manufacturer of the professional entertainment dimmers shown in Figure 8.33 mentions that the drive circuits in these dimmers are updated in respect of current and voltage conditions at 40kHz (i.e. 400 times within each half cycle at 50Hz). The dimmers are claimed to be able to detect an overCurrent measurement
Live
Figure 8.33 600 channels of IGBT dimming installed at the Royal National Theatre in London. Dimmers from IES of Holland, photo from Technical Marketing. Load
.005
:
IGBT DRIVERS AND FAST PROTECTION
PWM from microprocessor
OPTO ISOLATION
Typical output waveform
Figure 8.32 The power output section of one half of the double dimmer unit shown in Figure 8.31. The current transformer is used for report back and overall regulation. The very low value resistors in series with the transistors are used for measuring instantaneous current, and form part of the protection circuit (which is necessarily operating at line voltage).
current condition within 2μs and to switch off the transistor within 5μs. An example of a MOSFET based dimmer is shown in Figure 8.34. This is an 800W (230V) dimmer suitable for cabinet mounting or for distributed dimming systems. It is primarily intended for digital remote control using the DALI protocol, although it also includes a simple push button control facility. The control circuits require an isolated low voltage power supply and this is provided by a switched mode power supply. While the main power section of the dimmer does not need interference suppression components, the control circuit does require suppression in order to meet EMC standards. A professional product of this kind is surprisingly sophisticated, both in terms of its microprocessor based control and in the protection circuitry. Its manufacture is not dissimilar to the manufacture of electronic ballasts, being based on the use of surface
271
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
mounted components and automated assembly. The evolution of MOSFETs could well lead to a situation where they challenge thyristors for many dimmer applications.
8.5 Electromagnetic compatibility (EMC) 8.5.1 Introduction This section reviews the requirement for electromagnetic compatibility (introduced in Section 1.6). In principle the information given here applies to all kinds of lighting control equipment. It, therefore, applies just as much to the electronic ballasts described in Chapter 7, and to the control systems to be described in later chapters, as to dimmers. But this is a logical chapter in which to review the requirements because electronic dimmers are, potentially, a
serious source of RFI. Within Europe all electrical equipment is governed by the Low Voltage Directive. This states that electrical equipment cannot be sold unless it meets agreed standards in respect of both electrical safety and electromagnetic compatibility. The standards are European Standards (also known as Euronorms) set by CENELEC (the European Organization for Electrotechnical Standardization) and are classified by EN number. In the USA safety and EMC matters are separated. Most equipment sold to the public or designed for fixed installation in public spaces must be tested for electrical safety according to the codes of Underwriters’ Laboratories (UL) or to local electrical codes. But EMC matters are in the hands of the Federal Communications Commission (FCC.) The requirements for EMC are similar under both EN and FCC. This is partly because both organizations derive their standards from the work of the IEC (International Electrotechnical Commission) and their Special International Committee on Radio Interference (CISPR – the acronym is derived from the French title). The standards vary according to the nature of equipment and the environment it is working in. For simplicity the following information is based on the requirements for residential, commercial and light industrial environments. 8.5.2 Definitions
800W Transistor Dimmer
PWM From Microprocessor
OPTO ISOLATION
MOSFET DRIVERS
S
0.02
S
Figure 8.34 A DIN rail mounting 800W dimmer based MOSFETs. The low value series resistor forms the basis of current measurement in the fast protection circuit.
272
RFI can be radiated or conducted. Radiated interference is in the form of electromagnetic radiation; conducted interference is in the form of electric currents in conductors associated with the equipment concerned. EMC implies limits for emission and immunity. That is to say any item of equipment must not emit interference beyond certain limits, and, just as important, it must be immune to the effects of external interference below certain limits. Immunity is defined in various ways. At best the equipment concerned is not affected at all by the interference; but it is recognized that there may be some disturbances which may temporarily affect the equipment, but from which it quickly recovers.
DIMMERS
Enclosure Port A.C. Power Port
Signal Port Equipment
Control Port
D.C. Power Port Earth (ground) Port
Figure 8.35 The ways in which interference can get into and out of equipment.
In order to define the limits, it is postulated that any equipment has various ports by which interference can either get into it; or from which interference could be emitted. Figure 8.35 shows the idea. When applied to dimmers, and other lighting control equipment, the following apply. The enclosure port refers to the physical boundary of the unit; for example the case in which a dimmer is housed. This is the place where electromagnetic radiation might either be radiated, or where it might impinge. The a.c. power port is self explanatory. The d.c. power port might apply, for example, in control equipment with separate DC supplies, or in equipment with emergency DC power from batteries. Signal and control ports are, again, self explanatory. Earth (or ground) ports can be of two kinds. The obvious safety ground connection; and the less obvious functional earth terminal which may be needed for control or system purposes, but is not intended to carry heavy current. An example might be the ground connection of the screen of a screened cable. The standards documentation refers to disturbance phenomena, and for each of these lays down limits. It also lays down the method of testing for each phenomenon. Equipment manufacturers must either have a laboratory qualified to carry out the tests, or must engage an outside laboratory to do the tests for them. EMC is not limited to radio frequency – it is con-
cerned with all aspects of electrical and electromagnetic behavior. Table 8.1 summarizes the main phenomena covered by the standards. The subject of harmonics on the power supply was discussed in Sections 6.1.7.2, 7.2.2, and 8.3.7 and is not covered further here. . 8.5.3 EMC standards Under the EN system there are generic standards which are intended to apply to all electrical equipment in the absence of any specific standard. However, it was found in practice impossible to have a workable generic standard for everything, so there are a number of standards related to specific classes of equipment. For example EN50 081 Part 1 is the generic standard emission standard for equipment to be used in the residential, commercial and light industrial environment, where no specific standard exists. EN50 082 Part 1 is the corresponding immunity standard. However, EN 55 015 is a specific standard,
Figure 8.36 Measuring the radiated interference from a luminaire. The same procedure must be applied to control equipment. Photo from Helvar.
273
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
EMISSION Phenomenon Radiated RF from complete item Magnetic fields from complete item Conducted RF on AC power port Harmonics on AC power port Voltage fluctuations imposed on AC power port by the equipment Discontinuous emissions (“clicks”) imposed on AC port Inrush currents on AC power port Conducted RF on other ports
IMMUNITY Phenomenon Effect of RF electromagnetic fields on complete item Electrostatic discharge Magnetic fields Effect of fast transients on DC power, signal, control and functional earth ports Effect of conducted amplitude modulated RF on functional earth, signal, control and DC power ports Effect of fast transients on AC input and output ports Effect of voltage dips, interruptions and surges on AC input port Effect of conducted amplitude modulated RF on AC input and output ports
Table 8.1 EMC phenomena.
laying down both the limits and the method of measurement of RF interference of electrical lighting equipment. EN 61 547 is a corresponding immunity standard. Entertainment lighting controls (specifically consoles etc, not dimmers) come under the same classification as audio-visual and video equipment for which the standard is EN55 103. Because classifications and limits are changing, lighting control product manufacturers have to keep abreast of standards requirements. They can do this by joining their local standards organization. Figure 8.38 shows in graphical form the main conducted emission limits. The vertical axis is in dB(μV) measured at the mains terminals. In order to standardize the measurement in respect of mains impedance, the measurements are done using a Line Impedance Stabilizing Network (also known as an artificial mains network) or LISN. The CISPR recommend that this be the equivalent of 50Ω resistance in
274
Figure 8.37 Measuring conducted interference. Many EMC measurements must be carried out in rooms isolated from other sources of radiation. Photo from Helvar.
parallel with 50μH inductance. Class B equipment is that intended for use in domestic environments; Class A for commercial and industrial environments. The graphs also show for comparison the German VDE standards (which go down to lower frequencies) and the FCC standards. Figure 8.39 shows the corresponding radiation limits. This time the vertical axis is in dB(μV/m) normalized to a measuring distance of 10m. Here it should be noted that FCC regulations cover emissions up to 40 GHz. The measurements in Figure 8.38 and 8.39 are
Figure 8.38 Conducted emission limits. Diagram from EMC for Product Designers (see Reading List).
DIMMERS
8.5.4 Meeting EMC standards
Figure 8.39 Radiated emission limits. Diagram from EMC for Product Designers (see Reading List.)
carried out using the CISPR 16 Quasi Peak Detector. This was originally introduced to correlate actual interference to its subjective effect on broadcast transmissions. It applies a time constant to different frequency bands, and distinguishes between annoying impulsive interference, and less disturbing variations. The standards applicable to lighting equipment require that the conducted emissions are also measured with an average detector, and that average readings should be 13dB (Class A) or 10dB (Class B) below the quasi peak emissions. Table 8.1 lists several other phenomena covered by the standards. More details can be found by consulting the standards themselves (and, for example, the book that is the source of Figures 8.38 and 8.39). All the standards include specified methods of measurement. For example, magnetic fields for the enclosure port have different limits close-to (e.g. 0.4 A/m at 10cm) to short distance (e.g. 0.01A/m at 1m). The limit is also frequency dependent (e.g. 4–0.4 A/ m from 50Hz to 500Hz, decreasing as the logarithm of the frequency). In practice ensuring that the measurement conditions are correct is one of the more difficult aspects of EMC compliance. Quite small changes in the measurement set-up can make dramatic differences to the readings. For this reason it is best to have the measurements confirmed by qualified laboratories.
Meeting the EMC standards is not easy. When it became a matter of law that products should meet them, manufacturers encountered a lot of difficulties, but this was because they tended to try and modify existing designs. The hope was that the addition of a capacitor here, or a ferrite bead there, would solve the problem. But this proves to be entirely the wrong way to go about it. The only way to make a product with satisfactory EMC is to incorporate good EMC practice into the design from the outset. There is no doubt that today’s products are significantly better than their predecessors as a direct result of good EMC practice. The layout of printed circuit cards and the routing of cables within a product are both critical to achieving good EMC performance. In lighting control the matter is of considerable importance. Quite apart from the need to meet the legal requirements, failure to ensure good EMC has some very undesirable side effects. In dimmers the result can be instability in light output, and, particularly, interaction between dimmer channels. In control equipment the result can be complete failure of control due, for example, to the “lock up” of a microprocessor caused by interference, or the destruction of an input interface due to electrostatic discharge. A particular difficulty arises in respect of systems, as opposed to individual products. Strictly speaking a system made up by combining a number of separate products must itself meet the EMC standards. For example, a computer consisting of desktop unit, monitor and keyboard sold as a single package should meet the standards when all connected together (this is why computer cables tend to have large ferrite lumps on the connecting cables). But the authorities do recognize the realities of an “installation” where, provided all the components themselves are compliant, the system itself can be considered compliant unless shown to be otherwise. Those who build systems in, for example, 19 inch instrument racks or electrical distribution cabinets are in a difficult position. Strictly speaking the whole
275
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
8.6 New developments in electronic dimming 8.6.1 Sine wave electronic dimmers
Supply Waveform Switching Waveform Filtered Output Waveform
Figure 8.40 Waveforms in the “sine wave” dimmer.
rack or cabinet must conform, even if built as a customized system entirely from compliant components. Since it would be impractical to test every such rack on an individual basis, and indeed it would be difficult to devise valid tests in many cases, system builders need to develop construction methods that not only ensure the use of compliant components, but also ensure assembly practice that anticipates and eliminates the risk of overall non-compliance. Notwithstanding the great efforts that are going in to EMC, there will continue to be unexpected sideeffects arising from technological developments. For example the mass use of electronic ballasts and transformers, all of which individually meet accepted standards, may, overall, produce an unexpected result in a particular installation. L
It is clear that a problem with the present generation of electronic dimmers is the production of harmonics. A “Holy Grail” of dimmer designers is the “sine wave” dimmer that, instead of having a harmonic rich output, has a sine wave output at the same frequency as the supply, with the amplitude being varied to provide the required output, just like the variable autotransformer. Such a device can be created using a chopper circuit in which the variable output is achieved through pulse width modulation. Figure 8.40 shows the input voltage waveform, the switching waveform, and the filtered output waveform. Figure 8.41 shows a block diagram of the requirement. The realization of the switching engine is more complex than inverters used in electronic ballasts because: • for a practical product both input and output must have a common terminal (neutral in the normal AC supply system). • the switching engine is working from an alternating supply, not from DC. • the switching engine has to deal with currents coming back from the load (these arise when the in-
Switching engine
AC in
Input filter
Output filter
N VIN VOUT I OUT
DSP
Load out
Input polarity
PWM
N
Logic and timing Floating gate drives
Control level request
Controller
Figure 8.41 Block diagram of a possible realization of the sine wave dimmer.
276
L
DIMMERS
(a)
(b)
on the principles described. It operates at a switching frequency of 100kHz, and is being used in sensitive applications like lighting control in sound recording studios. A typical actual output waveform is shown in Figure 8.43. It can be seen that the sine wave dimmer needs a lot of small wound components for the filters and power supplies and is, therefore, a complex device to make. Losses (mainly switching losses) are significantly higher than in thyristor dimmers, although this situation will change as better MOSFETs and diodes with faster recovery become available. At full output, switching losses are minimal, so, unlike the thyristor dimmer, maximum heat dissipation is not at full output. 8.6.2 Load interfaces
Figure 8.42 Examples of sine wave dimmers. (a) sine wave dimmer of 10A capacity built in to the same housing as the dimmer modules of Figures 8.25 and 8.31 intended for use in sensitive environments like recording studios. (b) The SVC210 from Dynalite, rated at 2 × 7A and intended for the control of discharge lighting. The sine wave output means that standard HID lamps with electromagnetic ballasts and PFCs can be controlled.
coming voltage is switched off); effectively it is bidirectional. • the gate circuits of the switching devices are all at different potentials and each one requires its own isolated power supply. • the high speed of the current switch-off (typically 100A/μs) causes the diodes to “overshoot”, i.e. they carry some reverse current before recovering to a blocking condition. While the effect is also present in some of the circuits described in Chapter 7, it is critical in this type of circuit. It can lead to inadvertent short circuiting of the supply and to increased switching losses. Figure 8.42(a) shows a sine wave dimmer based
The controllable electronic ballast introduced in Chapter 7 points the way to new dimmer system architectures. While it will remain the case that for many applications the idea of dimmer racks as part of the power distribution will remain valid; there is already an increasing tendency to use “distributed dimming” either in the form of small sub racks, or in the form of “load interfaces”. The latter either being a separate device sited next to the luminaire(s) concerned, or built in to the luminaire. One of the problems to be overcome with this
Figure 8.43 A typical output waveform from the dimmer shown in Figure 8.42(a).
277
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 8.44 4-channel IGBT dimmer unit from IES for attaching to a spotlight bar. Photo from Technical Marketing.
approach is noise. Most thyristor dimmers make a characteristic buzz when at any intermediate setting; on the other hand transistor dimmers generally do not. This indicates that transistor based devices are more suitable for distributed control. Dual control circuit D L
~
FET
Opto triac
S
Load
N
Figure 8.45 Principle of the hybrid transistor-thyristor dimmer (above) and example 500W DIN rail mounting dimmer using the technique (below).
278
Examples of devices providing the basis of distributed remote control dimming are as follows: • controllable ballasts as already introduced in Chapter 7. These are normally built in to luminaires. • controllable electronic transformers. These could either be built in to luminaires, or could be designed to serve a small group of lamps. • individual remote controllable dimmers for installation near luminaires, like that shown in Figure 8.14d. • multiple IGBT dimmer units for entertainment applications, example in Figure 8.44. All these examples use a separate FELV or SELV control line. This can be either analog or digital and represents the most secure method of control. However, both the use of control signals carried on the mains wires, and the use of wireless control are also possible. 8.6.3 Hybrid dimmers One of the most expensive components of a conventional thyristor dimmer is the choke. An interesting technique that is being applied with success to low power dimmers is the use of a transistor to replace it. The idea is to use the controllability of a transistor to achieve the right waveform to maintain EMC, but to use the robustness and power handling capacity of the thyristor to carry the main load (and also to allow the use of inductive loads). A possible realization of the principle is shown in Figure 8.46. In this circuit a single MOSFET transistor is placed across a diode bridge, so that the one transistor can be used for both positive and negative going half cycles; and the whole assembly is in parallel with a thyristor pair, which in this case is controlled by an optoisolated triac. When the required firing angle is reached, the transistor does a slow switch on to achieve the required “rounded” waveform. Once this has been done, the thyristor fires and takes over the load. The transistor is then switched off. While the transistor is required to take a significant momentary current, its duty cycle is very short, and a small component can be used. Another possibility is to replace the bridge/
DIMMERS
Load 1
(a)
Load 2
~ L
~ N
(b) Load 1
Dual control dimmer
Load 2 N
(c)
Figure 8.46 Half wave dimming. (a) the dimmer output waveform showing assymetric half waves. (b) the circuit arrangement. (c) a realization of the principle by ETC with their Dimmer Doubler™.
transistor combination with a transistor pair that is able to carry the full lighting load. With fast current and voltage sensing circuitry, it is then possible to construct a dimmer that can sense whether the load is capacative or inductive. If it senses a capacitive load, only the transistors are used. If it senses an inductive load, the transistors are used for waveform shaping, and the thyristors for the main load. 8.6.4 Half wave dimming A technique that provides additional independent dimming channels is half wave control. It is sometimes referred to as multiplexing, and ETC refer to it as Dimmer Doubling™. The principle is illustrated in Figure 8.47. Each half of the sine wave is controlled independently, i.e. each half can have a different firing angle. The output of the dimmer is then fed to a diode splitter (also known as a “smart twofer”) which feeds two separate lamps. The diodes ensure that each lamp only sees one half cycle of the mains. The technique relies on the thermal inertia of the lamp, and on the use of lamps with a lower continuous voltage rating. In the ETC system 77V lamps are used on suitably controlled 115V dimmers. The technique needs using with care. Only lamps and dimmers designed specifically for the purpose should be used. On no account should any inductive load be connected to either of the outputs.
279
Chapter 9
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Control
signals
9.1 Introduction Early dimmers and switching devices were all manually operated. A lever or knob was used to directly interrupt the supply, or to move a sliding contact. As soon as the idea of remote control became practical there was a need to choose a method of communicating a control signal. It was immediately obvious that any such control signal should itself require minimum power. In some cases, for example the saturable reactor, the nature and power of the control signal was determined by the control process itself. In others, for example the use of motor driven variable resistors or transformers, there was a wide choice of methods depending on application, and on whether ot not the remote control had to operate at low voltage or line voltage. The arrival of electronics completely changed the picture. Its ability to amplify a signal, and the comparative simplicity of incorporating complete electrical isolation, meant that remote control devices could be small. Early electronic lighting control systems simply carried on the traditions of their electromechanical predecessors. Control remained on at least a “one wire per control channel” basis, and any kind of group or mastering control continued to be done by comparatively simple electrical networks. Micro-electronics and microprocessors introduced another step change. First they allowed much greater flexibility and complexity in control, by, for example, providing a memory capability, and second they allowed the practical “multiplexing” of control signals. It is this capability to control many devices via a single cable which is the main subject of this chapter. The rest of this chapter is concerned with today’s practice. While it may seem ideal to have only one standard method of communicating lighting control signals, this is simply not practicable. Present prac-
280
and
protocols
tice has emerged from a mixture of technological progress, tradition, and the widely varying needs of different applications. Note that where protocols and standards are summarized, this is for general information only. Anyone intending to implement the standards must obtain the full standards documentation from the sponsoring authority or organization.
9.2 Analog control 9.2.1 0–10V “standard” Analog control is the simplest method of remote control for dimmers and other controllers that vary voltage or some other measurable parameter. In the early days of electronic dimming many different voltage levels were used. Some of these had historical antecedents (e.g. they matched saturable reactor control voltages); some were technology driven (e.g. the use of negative control voltages because p-n-p transistors were the only practical choice for control circuits). Over the years the entertainment industry (that initially provided the greatest demand for remote control) gravitated towards the use of 0-10V as the “standard”. In fact there is no international (IEC) standard for this application but the USA based Entertainment Services and Technology Association (ESTA) does have a defined 0–10V Analog Control Protocol, reference E1.3. In practice all entertainment dimming equipment offering analog control works in the manner proposed by the protocol. The choice of 10V as the control voltage can be rationalized as follows: • it is a low enough voltage to be safe, and high enough to avoid signal noise problems. • it is easy to express control as a percentage of output, e.g. 5V = 50% output.
CONTROL SIGNALS AND PROTOCOLS
Controlled Device Control range OFF control voltage 100% or ON control voltage Control Safe acceptance range Input impedance
0001010V DC <0V t10V Linear -0.5 to +15V 100K: r 20%
Controlling device Passive control source impedance
<10K:
Active control source impedance Current source capability Stability at constant output Diode blocking capability
<100: >2mA r20mV >15V
Table 9.1 E1.3 standard from ESTA for the use of 0–10V analog control.
• it matches the capability of the electronic components used to implement it (e.g. op-amps and D to A converters.) Table 9.1 outlines the basis of 0–10V control. The need for a low source impedance arises because it is often a requirement that a single controller provides a signal to many controlled devices. The need for diode blocking arises because one controlled device may receive signals from several controllers on a “highest takes precedence” basis. Diodes on the output of controllers prevent any back feed into the controller from any competing source. A possible problem arising from the use of diode blocking is that the zero voltage is offset. It can result in the bottom part of the control range being unavailable, since there is no output signal until the diode forward drop is overcome. In practice the problem is overcome by using an offset voltage in the case of passive controls. In the case of active (op-amp) control it is easy to design the circuit to provide the required offset. When multipole connectors are used to convey the analog control signals to multiple dimmers, the convention is that the pins in the multipole connector are used in number order, corresponding to the
CONTROL 1
0-10V+ DIMMER
CONTROL 2
Common 0V-
Figure 9.1 The use of diode blocking to allow more than one analog controller to control a single dimmer. The highest voltage takes precedence.
channels (e.g. pins 1–14 for Channels 1–14) and that the last two pins in the connector are used for the power. (e.g. in the case of a 16 pole connector, Pin 15 is used for +10V, or higher voltage, and Pin 16 is used as the common 0V). There are no standards as to the type of connector used. The analog control system is intended to be linear; for example 6V control input results in 60% output. But the definition of “ouput” is, conveniently, left flexible. It could be voltage, or it could be light output. If there is any special “law” relating the control signal to the output, then such a law is normally held in the controlled device. The analog control voltage should be SELV, although FELV is permissible if there is a means of disconnecting the signal ground connection. The control signal source should be current limited, so that a short circuit on the control line can do no damage. 9.2.2 1–10V standard IEC60929 The use of an analog signal to control the output of fluorescent lighting ballasts is recognized as an extension to standard IEC60929 that sets out the operating standards of pre-heat start electronic ballasts. The control signal chosen is designed to cover the requirement of a single controller controlling many devices. For example it is quite possible that an installation will require one controller to control
281
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
1
9
10
9
10 11
MAX
LIGHT LEVELS (%)
Within Standard Typical Curve
MIN 0
1
2
3
4
5
6
7
8
12
CONTROL VOLTAGE (VDC)
Figure 9.2 The 1–10V control characteristic for controllable electronic ballasts according to IEC 60929.
282
1.1 1.0
CONTROL CURRENT (mA)
50 or more ballasts simultaneously. While the nominal control voltage range is, once again, 0–10V, the active control range is only 1–10V. This is to ensure that electrical noise does not affect the system performance. Note, however, that in the USA, 0–10V is widely used as the analog remote control signal for electronic ballasts. Figure 9.2 shows the control characteristic. The standard allows for some deviation from the ideal curve. At the bottom end 1.5V is the highest signal that can represent “Minimum Level”, and by 1V the device must be at minimum. At the top end 9V is the lowest signal voltage that can correspond to maximum level. In the recognized execution of the standard, each ballast is a current source, specified as providing a minimum of 0.2mA and a maximum of 1mA. Figure 9.3 shows the specification, with a typical unit sourcing around 0.3mA. This arrangement means that any controller is a current sink, rated for the worst case. From Figure 9.3 a controller for 100 ballasts needs to sink 100mA at <1V and 50mA at 10V. If it is only controlling one ballast, then it must operate at 0.2mA. The standard also specifies that: • there should be double insulation rated 2kV between the control wires and mains supply wires. • polarity reversal of the control signals should cause no damage to the ballasts.
0.9
Within Standard Typical Curve
0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
CONTROL VOLTAGE (VDC)
Figure 9.3 Control current as a function of control voltage for an individual ballast in 1–10V control.
• each controlled device should be able to withstand ±30V on its control wires without damage. It is important to understand that this standard was developed for the control of discharge lighting. The “minimum” and “maximum” levels depend on the type of lamp being controlled and the sophistication of the ballast. Thus for HID lamps, the minimum level would be at least 50%. For fluorescent lamps, the minimum level could lie between 1% and 20% depending on the lamp and ballast combination. The standard does not recognize any “OFF” command, only a “minimum level”. Separate provision must be made for OFF, and for load circuit isolation. Care must be taken in providing circuit isolation because electronic ballasts have a very high inrush current, arising from the charging of the main storage capacitor; and in practice switches, relays and contactors can only switch a limited number of ballasts. There is nothing to prevent the use of the 1–10V “standard” in controlling tungsten and tungsten halogen lamps; especially in architectural lighting control schemes using a mixture of sources. However, in practice it is necessary to use separate channels for the different sources.
CONTROL SIGNALS AND PROTOCOLS
In fact few “true” AMX192 products were made, and AMX192 is unlikely to be encountered today. However the Strand D54 protocol on which it was based (and which differs from it principally in that the clock signal is carried on the same wires as the multiplex signals) was widely used, and products using D54 may still be encountered. There are converters available that convert D54 to the DMX digital signal described in Section 9.4.1, and vice versa. Figure 9.4 The TK4 controller from Helvar can control up to 50 electronic ballasts from its slider control operating at 1–10V. However, the associated power switch can only switch around 20 ballasts at 230V, so a contactor has to be used for power control of more than 20 ballasts.
Because of the different conditions likely to be encountered in real installations, manufacturers usually specify a maximum loop resistance between the control device and the ballasts (for example <8Ω) and may provide a facility on the controller to compensate for volt drop. 9.2.3 Multiplexed analog control Prior to the arrival of full serial digital control some systems used a multiplexed analog control in order to get round the “one control wire per dimmer” problem. In the early 1980s Strand Lighting introduced a system that, in modified form, became the AMX192 standard endorsed by the United States Institute for Theater Technolgy (USITT) in 1986. The main characteristics of the standard were: • one pair of wires carried a synchronizing clock signal. • a second pair of wires carried the analog signal. This was 0–5V. It was sent in 50μs bursts, synchronized to the clock. Each burst corresponding to the level of a particular dimmer. • one control line could carry 192 dimmer signals, with a typical refresh rate of once every 50ms. • a receiver on the line would recover the sequential analog signals and convert them to parallel maintained analog signals. Each receiver would deal with 16 dimmers, so a minimum of 12 receivers were needed on one control line to service a full house of 192 dimmers.
9.2.4 PWM standard IEC 60929 IEC60929 now recognizes three methods of remote controlling electronic ballasts for fluorescent lamps. The analog 1–10V standard described in 9.2.2 is widely used, and the new digital standard that corresponds to DALI described in 9.4.2 is rapidly gaining acceptance. In between IEC60929 recognizes a third method, that of pulse width modulation (PWM.) The PWM signal is specified as having a low value of 0–1.5V, and a high value of 10–25V, then: • full light output is given if the “high” signal is 5% or less of the cycle time. • minimum light output is given if the “high” signal is 95% of the cycle time. • switch off is given if the high signal is more than 95% of the cycle time. • cycle times can be in the range 1ms–10ms. • there is a logarithmic relationship between the pulse width and light output. The PWM method specified by IEC 60929 is now not widely used. However, PWM is an important lighting control technique. It is extensively used for cold cathode fluorescent and for LED sources; in this case the actual PWM control is proprietary and is built in to a control device that itself accepts a standard digital protocol such as DMX or DALI.
9.3 Digital control 9.3.1 Basis of standards A digital lighting control system stores data relating to lighting levels for many channels of lighting. In order to get this information from a controller to a controlled device a link must be provided to convey
283
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
+5V
Single Ended Line Drive
-5V
1
0
0
1
0
1
0
0
1
Differential Line Drive
Here Negative Voltage signifies “1”
“Balanced” Output
Complementary Waveforms
Figure 9.5 Single ended and differential data transmission.
the digital data. While it is open to a manufacturer to design a proprietary method, it obviously makes sense to use a standard method if one exists. Such standards operate at two levels: One defines the electrical signals to be used. It is only concerned with such matters as voltage levels, impedances, data rates etc. and is not concerned at all with the actual data. The second is application specific, and may be referred to as a protocol. For example in lighting the DMX512 protocol, described in detail in Section 9.4.1, describes exactly how lighting level information for 512 dimmer channels is transmitted. It is now rare for a manufacturer to use anything other than a recognized standard for the electrical signalling. There is still a justification for proprietary control protocols, but in general there is a move towards standardization here as well. Such standardization has been well established in the entertainment industry for many years, and is now becoming a factor in architectural and commercial lighting control.
284
Most digital communication by wire is based on using different voltage levels, for example +5V to represent “1” (or “mark”) and 0V to represent “0” (or “space”). In practice 0V is not used in the normal standards, which use a negative voltage to represent “1” and a positive voltage to represent “0”. Data can be transmitted as a single ended (unbalanced) signal or as a differential (balanced) signal as shown in Figure 9.5. The differential drive is more complex and requires two wires referred to signal ground, but is very much more robust since common mode interference cancels out (see also Figure 2.29). The sending and receiving of coded electrical signals was established long before today’s microelectronic products, so some of the terms used in describing communications go back to the invention of the teleprinter. Even then it was necessary to have some form of “hand-shaking” between the sender and receiver of signals to ensure that equipment was ready to receive signals. The terms used then are still in use, and can be encountered in equipment descriptions. Communication can be described as being simplex, half duplex or duplex. Table 9.2 shows the origin of the term, and examples of equivalents in lighting control. Table 9.3 lists some of the common communication signals. The ones shown relate to EIA232, described in Section 9.3.3, but similar signals are needed in any serial communication system. Type
Origin
Lighting example
Simplex
One way telephone
Half duplex
Two way phone, but only one can speak at a time. “Press to talk”. Two-way phone, both parties can speak at once.
DMX signal from lighting console to dimmer rack. DALI lighting protocol that allows ballasts etc to be “polled” for return status information. Professional dimmer racks like those of Figure 8.17, where one data link carries instructions to the dimmers, and a return link sends continuous status data back to the central control.
Duplex
Table 9.2 Types of communication.
CONTROL SIGNALS AND PROTOCOLS
getting additional commands into and out of a computer unit.
Strobe Data Lines
Data Transmitter
1 2 3 4 5 6 7 Here “1” is shown High
9.3.3 Serial Communication, EIA232, EIA485 O
O
O
O
O
1
1
1
1
1
O
O
O
O
O
O
O
O
O
1
1
1
O
1
O
1
O
1
O
1
0
0
1
0
O
0
1
1
0
O
L
I
G H ASCII Characters
T
Data Receiver
0
Figure 9.6 The sending of parallel data using a strobe pulse to ensure correct reading of the data.
9.3.2 Parallel data communication Parallel digital data transmission is essential within a computer device, but is rarely used outside it because of the need for multiple cables and the difficulty of keeping the bit lines in sync. The one commonly encountered exception is a computer’s parallel port that presents data in “byte wide” format with a separate wire for each bit in a byte. The most commonly used configuration is the Centronics Printer Port (and its derivatives) which provides bi-directional 8-bit communication. A separate line carries a series of strobe pulses that indicate to the receiver when to look at the data. The idea is shown in Figure 9.6. The strobe pulses are much narrower than the data signals, ensuring that minor timing differences do not have any effect. Parallel ports are not used much in lighting control – except as printer ports in computer based systems where the printing of cue lists etc. is useful. However, they can sometimes be useful for special applications since they provide a simple means of
Serial data is sent as a continuous stream of data, but clearly there must be a method of timing the data so the receiver knows where bits and bytes start and finish. One method would be to send a synchronizing signal (similar to the strobe signal in the parallel port) but this would involve an extra signal wire and other components. Usually asynchronous data transmission is used, and the receiving device derives its clock from the data itself. It can do this if the data is sent according to pre-defined rules. Figure 9.7 shows the idea. A normal rule is that the data line idles at the (1) state, and that there is a minimum idle time between bytes. Further each byte is preceded by a 0 start bit and followed by a stop bit 1. The transitions at the beginning and end of bytes are always the same, and this enables the Universal Asynchronous Receiver Transmitter or UART to identify the beginning and Term
Stands for
Meaning
DTE
Data Terminal Equipment
DCE
Data Circuitterminating equipment
DTR DSR
Data Terminal Ready Data Set Ready
RTS
Request To Send
CTS
Clear To Send
TXD
Transmitted Data
RXD
Received Data
Equipment originating a data signal (e.g. a computer) Equipment receiving a data signal (e.g. a computer peripheral device) Signal indicating that DTE is ready Signal indicating that DCE is ready Signal asking if it is OK to send information Acknowledgement to the RTS signal “go ahead and send” Data sent from the DTE to the DCE Data sent back from the DCE to the DTE
DCD
Data Carrier Detect Ring Indicator
RI
Table 9.3 Examples of terms used in digital communication that originated in the 1950s but which are still used today in EIA232 serial communication.
285
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Start Bit
Idle
Start Bit
0
0
1
0
0
1
1
0
0
1
0
Data Byte example ASCII L
Start Bit
0
1
Start Bit
0
0
1
0
0
Idle
1
Data Byte example ASCII I
Here “1” is shown HIGH
Figure 9.7 The use of additional bits to provide synchronizing and clock information in serial data communication.
end of each byte and to derive a clock signal. The UART is the device used at each end of a serial data link. Besides looking after the timing, reception and transmission of data, it also acts as a buffer so that the receiving device can take the data when it is ready for it. For many years the most common method of sending serial data from a computer or computerlike device (through its serial port) has been the use of RS232 communication. The name comes from Recommended Standard of the Electronic Industries Alliance, and in fact the standard should now be referred to as EIA232, although many people continue to refer to RS232. Strictly it should also have a letter suffix, such as EIA232C to indicate the particular revision applying. Version A came out in the early sixties, and by the late 90s Version F had arrived. Section 9.5.1 describes the “open systems interconnect” or OSI model for communications, and when referred to this, EIA232 represents the “physical” layer of communication. The EIA232 standard also covers the type of connector used. The fully specified version uses a 25pin “D” connector, but the everyday use version is often based on a 9-pin D-connector with the pin allocations shown in Figure 9.8. EIA232 is simple, generally reliable and widespread. But it does have some disadvantages. It is comparatively slow. The original standard did not envisage a data rate greater than 19.2 kilobaud (in this context baud = bits per second). In practice it is frequently used at 38.4kb, and some users go for 100kb or even 921.6kb over short distances.
286
Because it uses single ended drive it is prone to suffer from noise problems, and can only operate over limited length cables. The standard envisages no more than around 20m, although in practice a distance of 100m is achievable if a suitable cable and a low data rate are used. EIA232 is limited to one transmitter and one receiver. In practice some products are made that allow one transmitter to serve multiple receivers, but such an arrangement does not fully comply with the standard. Most of the limitations are overcome by the use of alternative standards, in particular EIA422 and EIA485. The comparison is shown in Table 9.4, where it is clear that differential drive allows much longer cables and lower signalling voltages. EIA422 is used for applications such as video machine control in broadcasting, and in instrumentation. EIA485 is widely used in industry, and is also the basis of distributing lighting control signals both to the DMX protocol and to other proprietary protocols. DTE uses male connector
DCD
RXD
DSR
DCE uses female connector
TXD
RTS
CTS
DTR
Signal Ground
RI
Figure 9.8 EIA232 signals are often distributed using 9-pin D connectors. The connections relate to Table 9.3.
CONTROL SIGNALS AND PROTOCOLS
EIA232
EIA422
EIA485
Drive single ended Number of drivers 1 Number of receivers 1 Cable length ca 20m Driver impedance 300107k ohm Receiver input 300107k ohm Max data rate 20kb/s Min drive signal ±5V Max drive signal ±15V Data 1 (Mark) -V Data 0 (Space) +V
differential 1 10 1200m 100 ohm 4k ohm 10Mb/s ±2V ±5V -V +V
differential 32 32 1200m 54 ohm 12 k ohm 10Mb/s ±1.5V ±5V -V +V
Table 9.4 Characteristics of some common serial interface standards. The number of drivers and receivers is that permitted on one line. EIA numbers were previously “RS” numbers e.g. RS232.
In any data transmission system there is the danger of data corruption and resulting errors. Within a lighting system single errors are unlikely to lead to any kind of disaster, because in most systems data is updated on a continuous basis, so an error gets corrected almost immediately. The sophistication of error correction depends on the application. Three different methods are commonly used. The idea is that some arithmetic is applied to incoming data that can yield a “good/bad” data signal indpendent of any foreknowledge of the data content. The parity check adds another bit to each byte such that the total of all bits in the byte is always odd or even. In an even parity system, an extra 1 is inserted for all combinations that add up to an odd number, a zero is added if the bits add up to an even number already. Any single bit error within a byte will then cause a parity error to be detected (i.e. if it is an even parity system, all bytes should come out as even parity, and odd ones have an error). Parity checking is normally done in hardware, but the other forms of error correction are done in software. The checksum adds the contents of a number of bytes together, and sends the total as a separate piece of data. The receiver then does the same sum for itself, and checks that its result is the same as the separately transmitted checksum. If it is not, the whole block of data is rejected. The cyclic redundancy check or CRC. While both parity and checksum are easy to implement, they do
bring a data overhead with them, and they are by no means 100% effective. Parity checks only work if there is an odd number of errors in the byte. With checksums multiple errors can result in the same checksum as the correct one. The CRC operates on overlapping long blocks of data, not individual bytes. Each block is divided by a number that yields a remainder that is the check number. Mathematics well beyond this author show that the probability of corrupted data generating the same remainder as that of the correct data is negligible. In a computer system the detection of an error normally results in the sender being asked to send the message again, and the incorrect data being discarded. In lighting control it is only necessary to do this for “one time” signals – for example a single instruction sent to change a lighting scene. In a typical stage lighting system, however, the dimmer levels are updated many times per second, so there is no need to re-send the data. There is a theoretical need to discard the erroneous data, and for the dimmer simply to maintain its previously set level until a new error-free instruction arrives. 9.3.4 Serial communication, USB and IEEE1394 Computer peripheral devices such as digital cameras and color graphic printers require data transfer rates that are far higher than those achievable on EIA232 and EIA485. Two high speed serial links are in common use. Universal Serial Bus or USB was devised by the personal computer manufacturers to provide a simple means of connecting several devices located close to each other. Principal features are: • four wire bus, two for 5V power, two for data. • up to 127 devices on one bus. • hot patching (the ability to make and break bus connections without compromising the activities of other devices on the bus) permitted. • data rate up to 12Mb/s. • cable length limited to 5m. IEEE1394 is based on Apple Computer’s Firewire™. This provides very fast communication in the hundreds, or even thousands, of Mb/s. It is particularly suitable for carrying real time digital
287
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
video and similar bandwidth-greedy signals. A requirement for this type of application is isochronous operation, where time slots (corresponding, for example, to video frames) are guaranteed. IEEE1394 has, since early 2002, been severely challenged by the arrival of USB 2, a new variant of USB that runs at 480 Mb/s. While both these methods of serial communication may be encountered in computer peripheral connections, they do not make any special contribution to lighting control – partly because they have severe restrictions on cable length, but mainly because in practice the high bandwidth needs of lighting control are being met by ethernet LAN (described in Section 9.5.6).
Even when this is done, there will be no guarantee of interoperability. Today’s markets demand such interoperability, and it is having a profound effect on the way in which lighting control is specified. Exactly in the same way that the analog control standards first emerged in theater lighting control, so the theater was first to come forward with a standard multiplex protocol. The result is that, in principle, any entertainment dimmer can work with any entertainment lighting console, regardless of manufacturer. The prime mover was the United States Institute for Theater Technology (USITT). In the 1980s they introduced two multiplexing standards. The first, AMX192 (see Section 9.2.3) is an analog multiplexing arrangement which multiplexes 192 different analog values, and is therefore suitable for the control of 192 dimmers. However, its analog nature means that this standard is now obsolescent. The digital standard, DMX512, however, has gained worldwide acceptance. It has gained such popularity that it is, in fact, being used in ways not originally envisaged. For example, while it was originally intended for dimmer control, it has been pressed into service for the control of many other devices. As its name implies DMX512 is a Digital Multiplex Data Transmission Standard for Dimmers and Controllers, able to control 512 channels of dimming. Its characteristics are summarized in Table 9.5, and its structure is shown in Figure 9.9. Currently the DMX512/1990 standard is being upgraded to
9.4 Standard protocols for lighting control 9.4.1 DMX 9.4.1.1 Description of the protocol It can be expected that, within products, manufacturers will continue to use proprietary control protocols. Within an item of equipment the electrical regime is chosen to optimize the design. However, when a signal must pass from one item of equipment to another, most signals will be sent using one of the commonly used electrical specifications such as EIA-RS485.
Idle
Reset
Min 8Ps
Permitted idle can be zero time
Start Bit
Frame
Frame
Data
Data 513 frames max
First frame all null
0
88Ps
0
0
0
0
0
Mark after reset
0
0
0
1
Stop Bits
Figure 9.9 The DMX 512 data stream.
288
1
0
1
Start Bit
1
0
1
0
0
1
0
0
0
Stop Bits
0
0
1
0
0
1
0
0
0017Ps
1
1
1
CONTROL SIGNALS AND PROTOCOLS
Current version: Electrical specification: Connector:
Connector orientation: Connector pin numbering:
Maximum number of devices controlled: Number of levels per device: Valid levels: Data rate: Bit time: Frame time: Frame format:
Data format in frame: Packet format:
Update time for 512 dimmers: Update rate for 512 dimmers: ASC (Alternate Start Code)
DMX512/1990, now being upgraded to DMX512A EIA-485-A. DMX512A specifies transmitter and receiver protection to r42V. 5 pin XLR (no other XLR type permitted) Variations permitted under special circumstances. e.g. use of RJ45 connector in permanently installed systems based on structured cabling such as CAT5. female on transmitting device male on receiving device 1 signal common 2 drive complement 3 drive true + (Optional second link on Pins 4 & 5) 512 256 (8 bit) 00010255 decimal (00 to FF hex) 250 kilobits/second 4 Psec 44 Psec Bit 1, Start bit Bits 200109, Data bits (least significant first) Bits 10,11, Stop bits Start bit LOW Data bits HIGH = 1, LOW = 0 Stop bits HIGH Reset, LOW (minimum 88 Psec) Mark after reset, HIGH (minimum 8 Ps) Start frame, NULL (all zeros) orASC Data frames (maximum 512) Idle, HIGH (maximum 1 second) 22.67 milliseconds (minimum) 44.11 times per sec (fastest possible) 10010255. Some reserved for specific purposes according to DMX512A
Table 9.5 Principal characteristics of DMX512.
DMX512-A, which in due course will be ratified by ANSI. The differences are small, and relate to the harmonization of some physical aspects of DMX, and to the use of features of the original protocol that were left open in the original standard. The concept is that data is sent in packets. A packet updates all the dimmers, or other devices. The
transmitter (usually a lighting console) is not obliged to send a full packet of 513 frames (one start frame and 512 data frames), but it must not send a larger packet and must obey the timing rules, especially in respect of bit length, and the reset and start conditions. A receiver (normally a dimmer rack) must be able to receive full size packets at the maximum possible rate, but obviously only responds to those data bytes intended for it. The selection of the data is usually done with address switches, but in some systems soft addressing is used from a different data path. Any receiving device is expected to maintain a particular dimmer level until it receives an instruction to change. In DMX the individual data frames do not carry any identification. Their identity is determined solely by the order in which they appear in the packet – so the first frame after the start frame is the data for Dimmer 1, the second for Dimmer 2 and so on. Thus a small self contained system of 48 dimmers and a control console might use only a 48 frame packet. In theory, therefore, a smaller system could have a higher update rate, but in practice the full capacity update rate of 44 times a second is sufficient for most purposes. DMX512/1990 specifies a null start frame where all bits are zero. When this is received, all the subsequent frames provide individual devices with 8-bit level information. DMX512A defines how the other possible 255 start frames may be used. Equipment made to the earlier standard simply rejects anything other than null start frames; new generation equipment can process them if appropriate. The DMX system allocates 8 bits (256 levels) to each channel, but it says nothing about how the bits relate to any particular lighting level. Any “law” relating to how the data equates to a particular lighting level is carried in the dimmer. DMX is essentially a unidirectional (simplex) protocol. It only sends out instructions from the transmitter to one or more receivers. The EIA-485 specification sets out how multiple receivers can be served by one transmitter, and in some systems where a lot of receivers are being used, there may be a need either for signal isolation or signal conditioning.
289
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
9.4.1.2 Alternate Start Code The call for extensions to DMX and for recognizing it as a standard resulted in USITT transferring “ownership” of DMX to the Entertainment Services and Technology Association (ESTA) whose protocols working group is accredited to ANSI (American National Standards Institute.) It is this group that has been refining the original 1990 specification to the new “A” specification. Within this specification the use of data following start frames carrying a value other than null is defined. Some examples are: • the ability to send simple text information that can be displayed on suitably equipped devices. • the ability to send test packets. This is of value in complex installations, especially when, as the standard permits, the tests can be carried out while the system is “live”. • the ability to send “system information packets”, that give details of the installation (for example by identifying the DMX universe currently being used). • the carrying of manufacturer and equipment identification information. • the carrying of additional device specific data. DMX512A specifies some permitted timing variances when sending alternate start codes (ASC) in order to ensure reliable reception of the start code, and also specifies that a null code must be sent at least once per second. Generally it is expected that ASC packets are interleaved with null packets. 9.4.1.3 Limitations of DMX DMX has been and is extremely successful as a standard. It greatly simplifies the lives of rental houses, manufacturers and all who work in entertainment lighting. It does have some limitations, many of which can be overcome by ingenuity. Resolution. 256 levels of light are in principle sufficient for setting the level of lighting in both entertainment and architectural applications. However, when a long slow fade is made the “jump” between one level and another is visible at low levels. For this reason professional lighting controls interpolate additional levels – if you like, they bring back an
290
“analog” characteristic. They can do this either by “slugging” the dimmer input, so any change always takes place as a minimum timed fade, or, if the dimmer is all digital, by inserting extra bits at the dimmer to make the control equivalent to 10-bit or more. Some applications demand the true storage of high resolution data. For example if it is intended to use DMX for moving light control, 8-bit resolution does not even give 1° accuracy. Most moving light control based on DMX allocates two DMX frames (16 bits) to each major movement. On the other hand, some control functions that require only single bit control or low resolution can share a frame. Capacity. When DMX was introduced, 512 dimmers represented a big system. With the advent of moving lights, 512 control channels became a serious limitation – one moving light can alone gobble up 24 channels or data frames. For this reason light-
Device 1 In
Out
Device 1 In
From Console
Out
100: Terminating Resistor
R
In a terminated transmission line all signal power goes into the terminating impedance
In an unterminated line the signal is reflected back down the line
Figure 9.10 DMX equipment is often provided with a “loop through” facility. It is essential that the end of the line is terminated.
CONTROL SIGNALS AND PROTOCOLS
ing consoles often have multiple DMX outputs. Each output is sometimes referred to as a “DMX Universe”, and a typical large console might serve four universes (total of 2048 channels.) DMX is not really optimized for moving light control, but it has nonetheless become a standard for such control because the benefits of standardization outweigh the cost of the “fixes” needed to use it. Lack of error correction. DMX has no provision for error correction. In practice the performance and reliability of EIA485 equipment is so good, and the continuous update of information so effective at masking errors, that its lack is not a problem in its application area of entertainment lighting control. But DMX must not be used for any life-safety related applications such as the control of moving scenery or pyrotechnics. Hard addressing. Each device in a DMX universe must be fitted with an address switch so that it knows which frames in the DMX stream are relevant to it. This can either be a switch (e.g. a set of thumbwheel switches) or a key pad that assign channel addresses through a microprocessor. In the case of multi-channel dimmer racks, neither of these arrangements is inconvenient. The hard switch version sets the address of the first channel in the group, and the others follow sequentially. With the microprocessor/keypad version it is theoretically possible that channels in a multichannel unit do not have to be sequential. This arrangement requiring an address setting on the receiving device becomes less convenient when a system includes devices that are sited in awkward places. Practical signal distribution difficulties. While RS485 is specified to work up to 1,200m, in practice the range is dependent on frequency. DMX is marginal at long cable lengths, and 250m is a more realistic range. In any case it is essential to use the correct cable, a shielded (screened) twisted pair cable of low capacitance, of the correct characteristic impedance (120 ohm) and where the shield is the signal common. DMX is often wired as a “loop through” facility on the equipment it serves, since in many installations a console or other master device is serving many receivers. While convenient, this means that
if the last device in the chain has nothing connected to its DMX “OUT”, there is a danger that the signal will be corrupted. This arises because a data cable acts like a transmission line, resulting in the signals being reflected back along the line from the unterminated end. The problem is solved by placing a terminating resistance at the end of the line, matching its characteristic impedance. This can either be done using a terminating connector or by a switch on the equipment. See Figure 9.10. While 32 devices on a line, as permitted by RS485, sounds a lot, in practice more are often needed, and again it is not wise to work near the limit of the specification. Finally, there is an obvious danger that should any one piece of equipment develop a fault, it could bring down a whole DMX network. In the case of a high voltage fault, it could do serious damage. Sometimes there is a need for dimmers to receive more than one DMX signal, for example from a main console and from a separate houselights controller. Conversely in some small systems there can be a need for two consoles to serve the same DMX universe, for example two 60 channel consoles running a 120 channel dimmer system. This cannot be done without additional electronics. These limitations are got round by using specialized distribution devices. On the safety issue most DMX equipment uses optical isolation. To solve distance and loading problems, distribution amplifiers are available. These have an optically isolated DMX input, and two or more isolated DMX outputs. Such devices “regenerate” and retime the incoming DMX signal, so that any errors do not become cumulative. Multiplexing equipment can combine DMX streams. Where channels overlap, a priority can be set, for example “highest takes precedence” or “DMX stream A takes precedence”. Where they do not overlap, or where it is required that two short DMX streams are joined together, the device creates a new stream. In the case of two small consoles each outputting addresses 1–60, the combiner can make the second consoles signals occupy 61–120 (or any other desired numbers.) While DMX uses the electrical specification of
291
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 9.11 Specialized DMX distribution equipment from Artistic License Ltd. Above, a DMX distribution unit provising multiple isolated outputs. Notice the suffix “RDM” indicating that the device supports bi-directional communication. Below, a protocol converter that can not only convert an incoming data stream to an alternative protocol (e.g. from ADB’s proprietary S20 protocol to DMX) but can also combine (“merge”) two incoming data streams into one.
EIA485, it is NOT a full EIA485 realization. DMX is a simplex (unidirectional) system; true EIA485 is bi-directional, requiring control methods to prevent conflict between transmitters and receivers. It is also now possible to engineer EIA485 systems to work well over cable lengths of several kilometers, provided the correct cables and terminations are used. High impedance integrated circuit EIA485 drivers are now available that permit the use of up to 256 devices on one line.
of writing the proposed standard is in draft form. The following only describes the principles involved, current status can be obtained from ESTA in the USA or PLASA (Professional Lighting and Sound Association) in the UK. The extension is called Remote Device Management or RDM, and provides half duplex communication over DMX512 networks. The principle is that a system controller uses the main DMX link to request individual devices to return status data on the second link. It does this by using the ASC facility (see 9.4.1.2). Expected features of RDM include: • all RDM enabled devices will have a unique identification code. • the controller can “discover” what devices are on the network by a process of successive polling. Devices report back their identification so that the controller knows what is on the network. • once devices have been identified, they can be asked for status information. This can range from simple status information (like confirming that a color scroller has reached the required color or reporting on the running hours of a lamp) to fault information. Note here, however, that the controlled devices cannot initiate a data transmission, so if fault information is required it must be requested by successive polling. • the flexibility afforded by the introduction of ASC, and the specified response message formats, allow the introduction of some error correction for this additional data. 9.4.2 DALI
9.4.1.4 RDM on DMX networks 9.4.2.1 Introduction Until recently the optional second link on DMX has not been used, to the extent that some manufacturers started using a three-pin XLR connector and two core screened cable. DMX512A forbids this, and mandates that five-pin XLR must be used, and that any DMX device must pass through signals on the second link even if it does not use them. The second link provides the opportunity of a “return data path” from DMX controlled devices. ESTA are proposing an extension to the DMX standard that defines how this should be done. At the time
292
The DMX protocol described in Section 9.4.1 has its origins in the needs of entertainment where high channel capacity and high speed are essential. It requires some expertise in its implementation, and is clumsy when many simple single devices must have individual receivers. DALI, the Digital Addressable Lighting Interface, is at the opposite end of the spectrum. It is targeted at the needs of commercial and architectural lighting; ultimately there is no reason why it cannot be
CONTROL SIGNALS AND PROTOCOLS
ROOM 1
Mains power
ROOM 2
Ballast Lamp Dali load interface
Sensor
Control panel
Cordless control
Sensor
Control panel
Dali control cable
Figure 9.12 An example of a DALI lighting control system. Within the system all the luminaires and their controlling devices share a common control cable.
used in the domestic environment as well. It is proposed as an extension to IEC60929 and, while originally conceived for fluorescent lamp control, is in principle applicable to all light sources. The origin of DALI was the desire of European ballast manufacturers to ensure that digital ballast control was standardized. Osram, Philips and Helvar introduced the concept in 1999, and since then many other manufacturers have joined the DALI “club”. The role of Tridonic should be mentioned since they were the first large manufacturer to introduce digital control of electronic ballasts, using their proprietary DSI protocol. Many of their products now operate on both DSI and DALI. The idea behind DALI is that, within a lighting system, every luminaire is separately controllable, but requires only a single control cable for all devices in the system. Such an arrangement has often been proposed in the past, and is inherently quite possible using some of the bus systems, such as Echelon LON® and EIB, that are described in later sections. The problem has been that such systems have a high cost per node. A node simply being a point on a control system where control signals are either received or transmitted.
In a way, DALI deliberately sets its sights low. It does not, itself, seek to be suitable for building-wide control. Rather its seeks to be the optimum method of controlling lights within a large room or suite of rooms. It is designed to be very easy to install, to have a low cost per node, and to be easy to reconfigure. In its basic form it can meet the needs of, probably, 95% of architectural lighting control applications, but it is also intended to be easy to interface to higher level control systems when necessary. 9.4.2.2 Basic specification The concept of a DALI system is shown in Figure 9.12. Each lamp in the system has a load interface, typically an electronic ballast. The control devices in the system (such as push button panels and presence sensors) are connected to the load interfaces and each other by a simple pair of wires. It is possible for each controlling device to control only specified lamps so that, for example, in an office suite a single DALI system can be used while allowing independent user control in each office area. If partitions get moved, or there are other layout
293
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Maximum number of individually addressable devices in one system: Data rate: Data coding: Signal LOW: Signal HIGH: Maximum volt drop on control line: Maximum control cable length: Signal supply current limited to Nominal signal current per device: Number of levels per device: Bit time: Frame time: Frame format forward:
Frame format backward:
Time between frames:
64 1200 bits/second Manchester (bi-phase) 0V nominal; (-4.5V to +4.5V transmit, -6.5V to +6.5V receive) 16V nominal ( +11.5V to +20.5V transmit, +9.5V to +22.5V receive) 2V 300m 250mA 2mA 255 plus OFF (8 bit) 833.3Ps 15.83ms forward 9.17ms backward Bit 1, Start bit Bits 200109, Addressbits Bits 10001017, Data bits Bits 18,19, Stop bits Bit 1, Start bit Bits 200109, Data bits Bits 10,11, Stop bits Minimum 9.17ms before a forward frame Minimum 2.92ms, max 9.17ms before a backward frame
Table 9.6 Principal characteristics of DALI.
changes, it is easy to re-program the system so that control is re-assigned as required. Table 9.6 summarizes the principal characteristics of DALI. The important points to notice are as follows: Low data rate: in all networks or data distribution systems that work at high speed, there are strict rules about how units are connected together. With DALI the data rate is sufficiently low that there are no rules. The control wires can be looped round devices, or they can be fed star fashion, or a mixture of the two. There is no need for any line termination. Wide signal tolerance: the big difference between the high and low signal voltages, and the wide tolerance at each level, means that the system is largely immune to electrical noise.
294
Galvanic isolation: the control signal paths are completely isolated from the mains supply voltages. However, it is intended that for most systems the control wiring will run next to mains wiring (even sharing the same cable sheath) so the control wiring is not considered SELV. In some systems it is possible to have further isolation in the control network so some devices in it are SELV. An enhanced implementation of DALI allows for the accidental connection of mains voltage to the control circuit without damage, but this is not included in all devices. Control wire isolation ensures that there are no problems with ground loops in the signal paths (as can arise with EIA485 when the cable sheath may get grounded at two different points each at a slightly different potential). Manchester coding. In the simple bit structure used in DMX, the signal line voltage stays at one level if a string of 1s or 0s are sent. More robust data transmission can be achieved using a bi-phase modulation system, of which Manchester coding is an example – and which incidentally has the advantage that the resulting signal does not have a significant DC component, so it can be passed through transformers. In this arrangement there is a voltage transition in the middle of each bit period. If the transition is from low to high, it is read as “1”, if it is high to low, it is read as “0”. Figure 9.13 shows the principle, and Figure 9.14 shows the timing specific to DALI. Bit Interval
0
1
1
0
1
1
0
Manchester bi-phase encoding as used in DALI
Figure 9.13 Manchester bi-phase encoding as used by DALI.
CONTROL SIGNALS AND PROTOCOLS
10μs d trise d 100μs
V
High level: 9.5 to 22.5 V. Typical 16 V. 90%
10μs d tfall d 100μs
10%
Low level : -6.5 to +6.5 V. Typical 0 V
t
833.33 μs ±10 % BI-phase logical ”1” 416.67 μs ±10 %
Figure 9.14 The DALI voltage waveform for Logic “1”.
OFF signal. While the analog 1–10V annex to IEC60929 does not allow for an “OFF” signal, the DALI protocol specifically allows for it. However, it must be appreciated that such an off signal applies only to the controlled device; clearly the feed supply is not isolated when the command is used, so there must be separate provision for total circuit isolation for some kinds of maintenance. Address bits. While the use of address switches on DALI devices is not forbidden, one of the whole points about the system is that they are not needed, so, for example, it is not necessary to go fumbling around in the ceiling trying to change a ballast address number when an office partition is moved. The address byte in the data frame has six bits for device identification, and two other selection bits. In combination with the data bits (which are normally used for level selection) this arrangement allows for: • individual, group or broadcast messages. i.e. it is possible to address an individual luminaire, a group of luminaires, or all the luminaires within a system. • the polling of individual luminaires to request backward data. • the initial assigning of addresses to luminaires. Part of this process is automatic using a discovery routine that identifies the equipment on the network, and then assigns each item with a short address. • the selection of lighting “scenes”. Scene control. The concept of scene control is developed more fully in Chapter 12. Here it is sufficient to say that the use of the scene concept reduces the data load and speeds up response time. In most applications, groups of luminaires are required to move to different level settings together. DALI per-
mits this, so that a single data command can instruct a group of luminaires to go to a “scene”, where, in fact, each luminaire may be at a different level. This results in a very quick response time compared to sending a whole series of commands to individual luminaires. Backward data. An important feature of DALI is the ability to extract status data from the load interface. Since only one transmitter can be active at any time (half duplex operation) this process must be done on a polling basis, where each luminaire is interrogated in turn. Typical data that can be collected are “lamp failure” indication, especially for fluorescent lamps, ON/OFF indication, and “light level” or load current indication. Thus, for example, a Building Management System could poll all luminaires in a large building for lamp failure – it could do this at leisure and simply produce a daily report. Alternatively at night it could poll offices to find which ones still had lights on, indicating that someone might be working late. 9.4.2.3 DALI Components DALI is a comparatively recent introduction, so new product ideas are still emerging. The main items in a DALI system can be classified as: • Load interfaces • Control panels • Sensors • Control interfaces or “gateways” Each device that connects to a DALI control line must be able to observe the signal current/voltage regime listed in Table 9.6. This means that: When it is in the non-active “receive” state, it must accept voltage swings from -6.5V to +22.5V, and that in doing so it must sink no more than 2mA at the HIGH level. When it is in the active “transmit” state, it must output a voltage range -4.5V to +20.5V, and that in doing so it must be able to sink 250mA at the LOW level. Although the system is current limited, this latter requirement does require robust driving transistors, since when signalling, they are effectively short circuiting the line.
295
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Fluorescent lamp ballast (Philips)
Compact fluorescent lamp ballast (Insta)
thyristor dimmers. Therefore, after the fluorescent lamp ballast, the second most important load interface for DALI is a dimmer. • electronic transformers. Of course it is also possible to fit DALI control in to an electronic transformer. At present this is probably only an economic proposition for multi-lamp transformers. • electronic HID ballasts. As these become more widely available, it can be expected that DALI versions will also emerge. They will be needed as low wattage HID lamps invade the interiors market. • blind and curtain controllers. In the office environment the control of daylight from windows, especially when the sun is low in the sky, becomes
Dimmer (Helvar)
Blind controller (Helvar) Push button scene control (Starfield Controls)
Rotary control (Starfield Controls)
Figure 9.15 Examples of DALI load interfaces.
Load interfaces. The first load interfaces to appear with DALI compatibility were fluorescent lamp ballasts. Within DALI compatible fluorescent lamp ballasts the law is logarithmic in an effort to ensure that the perceived light level changes are linear with respect to the control signal. In practice the actual control law does need matching to the particular lamp, especially when the lamp does not permit low light levels. Other load interfaces, either already available or under development, include the following: • dimmers. It remains the case that the control of tungsten and tungsten halogen lighting can often be most conveniently done through transistor or
296
Slider controls (Helvar)
Infa-red cordless control (Philips) Figure 9.16 Examples of DALI control panels.
CONTROL SIGNALS AND PROTOCOLS
Multi-Sensor PIR, Photo-electric and IR (Helvar)
Cordless PIR Sensor using RF (Insta)
els have an IR detector built in which eliminates the need for a separate sensor. The control panel family can also include devices such as timeclocks, although in the larger systems time commands are usually issued by a higher level control system (e.g. BMS) via a gateway. Sensors. Any system that is intended to control light automatically, whether for energy saving or esthetic reasons, requires one or more sensors. The most common requirements are for an occupancy sensor (usually achieved by using a passive infrared or PIR sensor) and for a light level sensor. The subject of sensors is covered in more detail in Chapter 14, some DALI compatible sensors are shown in Figure 9.17. Control interfaces and gateways. No control protocol or system can live in isolation. While many lighting control schemes can be realized entirely in the DALI domain, there is a requirement for gateways to other systems. The most important of these are:
Figure 9.17 Examples of DALI sensors.
an essential part of lighting control. By incorporating the control of venetian blinds and other daylight modifying devices into the DALI world, the user’s life is made easier. Control panels. The control panels within a DALI system can take both conventional and unconventional forms. Lighting levels can be selected by push button with automatic timed fades, or by analog controls such as rotary potentiometers or sliders. A neat feature of such “analog” controls is that it is possible to have several of them controlling one luminaire or group of luminaires. In an all-analog system it is necessary to have a “transfer” switch that determines which control is being used, or to accept that “highest takes precedence”. With DALI all controls are inactive unless operated, in which case the operated control takes command automatically. For many systems cordless control is essential. Both infra-red and RF controls can be used, but both require a receiver device that is connected to the DALI control line. Some wall mounted control pan-
Lonworks gateway
1-10V converter (Osram)
Serial converter for connection to computer Figure 9.18 Examples of DALI control interfaces or “gateways”.
297
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Load interface
Control interface/ gateway
Separate power supply (Insta) Figure 9.19 The alternative ways of providing power for a DALI system.
• 1–10V analog converter. This converts a DALI level command into a 1–10V analog command. It may seem perverse to introduce analog into what would otherwise be an all-digital system, but in the real world 1–10V equipment will be around for a long time yet. This arrangement allows control of groups of lamps, or individual lamps, using the 1–10V interface. It is particularly applicable to fluorescent lamps, cold cathode lamps and HID lamps. • LONWORKS® ® converter. One of the most popular protocols used in Building Management Systems (BMS) is the Echelon LON system. For a large building an economic and practical arrangement is to use DALI for each small group of rooms or offices. In this way each lighting zone works independently, and there is a fast response to locally generated control commands. But then each zone can also have a LON Gateway, so that the BMS can monitor lighting status, and can, if necessary, issue broadcast or group commands. • EIB converter. LON is not the only protocol used in building control. In Europe the European
298
Installation Bus is popular, and there are many other industry or proprietary bus and network systems (e.g BACNET). While LON and EIB represent the first generation of DALI gateway products, it can be expected that several other specialist gateway products will emerge. • EIA232/EIA485 converter. An obvious requirement is to make DALI accessible to other conventional control devices, such as room controllers and personal computers. In its “perfect” form DALI is a distributed control system. The system intelligence is held mainly in the controlled devices, and the impact of individual unit failure is minimal – since there is no central unit whose failure can bring the whole system down. However there is no prohibition of central control units, and some manufacturers retain some form of central unit to simplify installation and to simplify the use of “legacy” products, especially 1– 10V analog controlled devices. 9.4.2.4 Powering and programming DALI If every DALI device had to provide enough power to source a full house of devices, the cost per node would increase considerably. Thus for each system it is necessary to calculate the maximum signal loading, and the loading represented by the control devices, and to provide the necessary power to meet it. This is, in fact, one of the only specialist aspects of DALI systems specification which is otherwise extremely simple. Power can be derived in three ways: • from a load interface • from a control interface • from a separate power supply In general control panels and sensors derive their power from the DALI control line which doubles as the source of power and the signalling medium. The specification of a DALI device will state what its current demand is, for example a control panel may require 10mA, or a fluorescent lamp ballast may require as little as 0.15mA as a receiver; and whether or not it can also provide power. If it is a load interface, it may have two sets of DALI terminals, one for signal and power combined,
CONTROL SIGNALS AND PROTOCOLS
and the other for signal only. In a small system consisting of, for example, a sensor, a control panel and two or three luminaires, all the power required can be derived from the load interface. In practice only the smallest systems derive DALI power from the load interfaces, especially as most ballasts do not provide power. Larger DALI systems derive it either from a control interface, for example a typical 0–10V interface can supply 170mA, or from a dedicated power supply. When larger power supplies are used, only one power connection is usually permitted. When multiple power supplies are connected, polarity must be strictly observed, and most executions of DALI require polarized wiring. However if a single power supply is used, it is possible to have DALI equipment that is polarity insensitive (but this is an optional feature of DALI and can not be assumed). The DALI standard makes no specification as to the type of cable to be used as control cable, in principle the system is intended to work using standard installation cable; however some manufacturers may recommend the use of screened cable for added security. Notice that in this case the screen (shield) is not involved in carrying the signal, unlike the case of EIA485 where the screen is at the centerpoint voltage. The cable must be chosen to ensure that the 2V volt drop rule is not broken, and in practice this can be achieved using conductors of the sizes shown in Table 9.7. The programming of a DALI system varies from the extremely simple to quite complex. It is possible for manufacturers to deliver DALI packages that require no user programming at all. An example might be an office luminaire with a built-in constant light sensor and a user’s cordless control. The next step up is programming using one of the control panels in the system. For example the cordless control shown in Figure 9.18 is designed Cable length < 100m 100m0010150m 150m0010300m
Minimum conductor area 0.5 mm 2 0.75 mm 2 1.5 mm 2
Table 9.7 Recommended conductor sizes for DALI control cables.
No programming; IR control (Helvar)
Programming from simple controller (Insta)
Programming from computer Figure 9.20 Alternative methods of programming a DALI system.
as a small system programmer. A “secret” button converts it from being the normal cordless control to being a programming device. By using an “optical recognition” principle, whereby lamps identify themselves by flashing on, the controller can assign individual lamps to particular groups, and can set individual levels. Dimming speeds can also be selected. While this method of programming is ideal for small system programming, it becomes impractical for larger systems, and it provides no record of what has been programmed. Larger systems are best programmed using a laptop computer. This not only simplifies the programming process, since the computer can show graphically all the individual lighting levels relating to a scene, but also allows a record of the programming to be kept.
299
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
For most sophisticated schemes the lighting designer pre-programs the system on his/her computer prior to arrival on site, so that the whole system can be up and running very quickly; he/she then modifies the pre-programmed levels and dimming speeds to meet the actual requirements of the site.
Signal Signal type Byte format Connector on equipment
9.4.3 MIDI and MSC A control protocol widely used in the entertainment industry is MIDI, or Musical Instrument Digital Interface. It was originally developed to allow the linking of electronic musical instruments such as synthesizers and electronic drum kits, but its low cost and simplicity of execution led to it being used for other purposes. In its basic version it sends musical note information in real time, so, for example, pressing Middle C on an electric piano causes the same note to be emitted by a nearby electronic organ. Some manufacturers have offered equipment that uses the protocol to operate lighting dimmers etc., but this is unsatisfactory and is now discouraged. MIDI is maintained as a standard by the International MIDI Association. While still primarily a music based protocol, there are recognized nonmusic applications known as MIDI Show Control, MIDI Machine Control and MIDI Timecode. MIDI timecode is covered in Section 9.4.4.2, MIDI Show Control or MSC is the variant that is of significance in lighting control. Like DMX, MIDI is a “one way” system without error correction. It has no means of knowing whether anything is connected. A master device emits MIDI signals, and these can be connected to a number of following devices. MIDI devices have a MIDI “OUT” connection if they create MIDI signals, a MIDI “IN” connection if they are to receive MIDI, and possibly a MIDI “THRU” connection for “daisy chaining”. The standard requires that all MIDI ports are optoisolated. More information about the MIDI and MSC protocols is provided in books listed in the reading list. Table 9.8 lists some of the principal characteristics of MSC. The Command Format and Command examples are representative only – a full list
300
Cable
General MSC command format in hex Data is sent in ASCII
Examples of Command Format
Examples of Commands
Asynchronous 31.25 kbps Current loop Current ON = Logic 0 Current OFF = Logic 1 10 bit word START bit, 8-bit data, STOP bit 180° 5-pin DIN female Pin 1 No connection Pin 2 Cable shield MIDI OUT only Pin 3 No connection Pin 4 +5V current source Pin 5 MIDI Data, current sink Twin twisted core shielded cable with 5 pin male DIN connector at both ends. Maximum 50 ft (15m) but line drivers are available for long distance connection F0 7F [Device ID] 02 [Command Format] [Command] [Data] F7 where: F0 = Start of System Exclusive 7F = Real time 02 = Device sub-ID F7 = End of System Exclusive 7F All types 01 Lighting, general 02 Moving lights 03 Color changers 04 Strobes 05 Lasers 06 Chasers 01 Go 02 Stop 03 Resume 04 Timed Go 05 Load 06 Set 07 Fire 08 All Off 09 Restore 0A Reset
Table 9.8 Some characteristics of MSC.
could show up to 127 of each. The MSC command string format arises because the original “musical” MIDI set aside two status bytes that allowed equipment manufacturers to send any kind of data down a MIDI line. These data messages are referred to as “System Exclusive” messages, and all MSC commands fall into this category. They must start with F0 and finish with F7. MIDI also distinguishes between real time messages and messages involving the transfer of data as part of set-up. All MSC messages are real time, but must
CONTROL SIGNALS AND PROTOCOLS
nonetheless include the real time byte 7F to conform with the standard. In Lighting and Show Control MSC is mainly used as a “master” protocol, where it is used to send commands to various sub-system control devices such as lighting consoles, laser scanners, audio mixers, VCRs, pyrotechnic firing systems etc. Of particular importance are “Cue” type commands, where the command data tells a lighting desk, for example, to execute a particular cue sequence from its memory. MSC has been widely used for attractions at theme parks etc. using a combination of “live” and “recorded” effects. A master controller can be used to send out pre-prepared sequences of commands with manual over-ride. As with DMX, there are number of devices on the market such as line drivers, splitters and stream mergers that help solve the practical problems of MIDI. When one controller is intended to operate several devices it is strongly recommended that a splitter is used, giving a separate isolated repeat signal to each device. 9.4.4 Timecode 9.4.4.1 SMPTE timecode There is often a requirement for lighting effects to be synchronized to other devices, to recorded music, or to a video program. A convenient way of doing this is to use a time code that is carried on the master medium, and which other devices can follow. The idea of timecode emerged as videotaping was perfected. The people trying to edit videotape could not help noticing that film makers, with their discipline of film frames and sprockets, actually made a much better job of program editing than they did. Video needed an “electronic sprocket”, and this was provided by timecode. Timecode is an electrical signal that can be recorded in two ways: • as longtitudinal timecode (also called linear timecode), where it is recorded on a standard track of recording tape; for example a track of audiotape, or one of the “audio” tracks on a videotape. • as vertical interval timecode (VITC) where it is
000103 400107 800109 10 11 12001015 16001019 20001023 24001026 27 28001031 32001035 36001039 40001042 43 44001047 48001051 52001055 56001057 58001059 60001063 64001069
Frames Units Binary Word 1 Frames Tens (maximum value 3) Drop Frame Flag Color Frame Flag Binary Word 2 Seconds Units Binary Word 3 Seconds Tens (maximum value 5) Unassigned Binary Word 4 Minutes Units Binary Word 5 Minutes Tens (maximum value 5) Unassigned Binary Word 6 Hours Units Binary Word 7 Hours Tens (maximum value 2) Unassigned Binary Word 8 Synchronizing Word
Table 9.9 The bit sequence in SMPTE/EBU timecode.
recorded within the video signal, onto two of the “spare” lines which are not part of the displayed image. Although originally developed to be recorded on magnetic tape, SMPTE Timecode (Society of Motion Picture and Television Engineers), is also used in computer based audio and video systems. Internationally it is supported by the European Broadcasting Union, so it is sometimes referred to as SMPTE/EBU timecode. The timecode signal is able to define each frame of video in terms of hours, minutes, seconds and frames. All professional video (and many audio) programs are made using timecode as a reference, and it is absolutely fundamental to the computerized editing of videotapes. Mixed media shows requiring the synchronization of lighting, animation and effects with audio or video are, therefore ideally also programmed on a timecode basis. The programming computer or similar device reads timecode from one of the sources, and all special effects are locked to the code which gives a precise definition of where the show is.
301
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
BIT INTERVAL
1
0
1
1
0
0
Figure 9.21 The bi-phase encoding used by SMPTE/EBU timecode.
The standard time codes consist of a stream of data which is issued at a rate corresponding to the “frame rate” concerned. They are: • SMPTE 30 frames per second • SMPTE “drop frame” • EBU 25 frames per second • FILM 24 frames per second The code variants are self-explanatory except for drop frame. An anomaly of the NTSC video signal is that for historical and technical reasons the actual frame rate is 29.97 frames per second. This means that if a 30 fps code is used, there will be a gradual drift away from “time of day” time, which could cause a big problem if two TV stations are trying to synchronize their programs. “Drop Frame” SMPTE code makes the necessary correction by dropping two frames every minute, except on the tenth minute. Drop frame may be encountered in the professional broadcasting environment, but usually it need not be considered for self contained mixed media show programming. Longtitudinal SMPTE/EBU timecode is made up of a data stream which is repeated at the appropriate frame rate. Each packet of data is 80 bits, designated 0–79. The allocation of these bits is as shown in Table 9.9. A similar bit pattern is used in VITC, but here 90 bits are used to allow the inclusion of field sequence information and error correction. The “Binary Words” are optional “user bits” which can be used for recording information such
302
as the date and production number. The Frames, Seconds, Minutes and Hours information is recorded as binary coded decimal (BCD) where 4 bits are used to encode any number 0–9. The Synchronizing Word is a fixed pattern: 0011111111111101 and is used by the decoder to tell it which direction the tape is travelling, and to identify the start of each stream. All 80 bits are recorded, any unwanted “1s” being recorded as “0”. In longtitudinal timecode the 0 and 1 are recorded as “bi-phase marks”, a self clocking technique which allows the decoder to operate over a wide range of tape speeds, and with the data coming in forwards or backwards. The idea is similar to the Manchester coding used by DALI but this time the rule is that: • if there is no transition within a bit time, then Logic 0 applies. • if there is a transition within a bit time, then regardless of direction, Logic 1 applies. The signal thus has no polarity. The principle is shown in Figure 9.21. Most serious show controllers and show control computer programs have the option of working to timecode. Some lighting consoles also have the facility. In automatic shows as used in themed attractions a common arrangment is the use of a SMPTE timecode synchronized DMX playback controller. The principle here is that a show is programmed using a conventional lighting control console. The show is then run from the console. The DMX stream(s) coming from the console is/are simultaneously fed to the controlled devices, and to the playback controller. At the same time the playback controller receives the show timecode (e.g. from a
Figure 9.22 An example of a Lighting Show Playback controller, the Alcorn McBride LightCue. These devices play out a DMX stream synchronized to an incoming SMPTE signal. See also Figure 13.4.
CONTROL SIGNALS AND PROTOCOLS
multi-track hard disc audio recorder). The playback controller “remembers” the whole show as a continuous DMX signal locked to timecode. In the completed installation the conventional console is no longer needed. The playback controller receives timecode from the master playback device, and spits out the DMX corresponding to the timecode. The same principle can be applied to other devices used in the show; for example laser controllers, motor controllers, video processing equipment etc.
sent, so complete frame data is accumulated over the four quarter-frame period. In theory this can allow for higher resolution timecode, but this would only work if the MIDI link concerned is reserved for timecode only. One possible problem is that if a MIDI link is used for other purposes in addition to timecode, there can be timing problems dependent on the MIDI traffic. Converters are available that convert SMPTE timecode to MIDI timecode and vice versa.
9.4.4.2 MIDI timecode
9.5 Networks and buses
Another timecode which may be encountered in mixed media applications is MIDI timecode. This uses the MIDI digital stream to send timecode data. For some users this is more convenient than using SMPTE code. The time data principle is very similar, and the MIDI timecode standard sets out how SMPTE data should be matched to MIDI. The one major difference is, of course, that the MIDI signal cannot be directly recorded to tape. As far as MIDI is concerned the timecode is simply a particular variety of system exclusive message, and it is treated just the same as any other MIDI message. In order to keep MIDI timecode messages short the full time code is only sent at the beginning and end of a sequence; during the sequence run 120 “quarter frame” messages are
9.5.1 General considerations
NODE 1
NODE 2
Application
Application
Presentation
Presentation
Session
Session
Transport
Transport
Network
Network
Data Link
Data Link
Physical
Physical
The open systems interconnect 7 layer model
Figure 9.23 The Open Systems Interconnect 7-layer reference model.
The protocols described in Section 9.4 are comparatively simple. Life gets more complicated when a system must have many devices connected, each one of which can be a generator or receiver of information, and where data rates are potentially much higher than those used in simple “one way” protocols like DMX. In order to be able to evaluate different data communications arrangements, it is useful to have a model of what might be required. The ISO publish the Open Systems Interconnection Model that proposes that a data communications system exists in seven “layers”. Figure 9.23 shows the idea; each layer in the stack provides a service to the the layer above it, while at the same time relying on service from the layer below. Table 9.10 defines the layers in more detail. Data can only pass vertically through the stack, although from a user’s point of view it is as if a layer in one stack communicates directly with its peer in another stack. In the real world not all communication systems match the model; sometimes layers are merged or even omitted, and one of the most important, TCP/ IP, follows a somewhat different model partly because its principles were established before the OSI model was finalized. Nonetheless the OSI model is a useful reference because it shows us what to look out for. Clearly only the bottom layers of the protocol stack are relevant to a protocol like DALI. Even at
303
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Layer
Description
7 Application
W here the user’s application joins the communication system. Includes network applications such as file and message transfer. This is where data is formatted to allow different applications to access the same network. Syntax, special characters etc. are harmonized. Encryption is added here if required. This establishes and manages the link between two applications in a single connection session. It controls the data flow. This layer is responsible for reliable (or unreliable) end to end data delivery. It multiplexes data if required. It manages the re-transmission of lost data, recovery from data collision, and the prevention of the reception of duplicate data. This is a traffic control layer, determining how message packets are sent, and their routing. This packages data for transmission. It frames the data bytes, adds start and stop bits etc, and error correction bits (e.g. CRC). This is the physical and electrical interface (e.g a particular style of connector, a particular electrical regime such as EIA485).
6 Presentation
5 Session 4 Transport
3 Network 2 Data Link 1 Physical
Table 9.10 Brief descriptions of the OSI layers according to ISO7498
this level a number of concepts relating to bus or network communications are revealed. If a large number of devices share a common communication line, there must be some rules about how messages are initiated to avoid bus contention issues. Some methods of doing this are: Polling. Here no device can speak unless spoken to. This requires the concept of “master” and “slave” devices (wheras more sophisticated systems do not make such a distinction and operate on a peer to peer basis). Polling can either be done at regular intervals or only when information is actually needed. DALI uses polling for the return data, as does the RDM extension of DMX. Interrupts. In this case the slave device(s) alert the master to the fact that they have information to send. In practice this is easiest executed by having a separate interrupt signal line. Token passing. This method is appropriate to systems where there is no master/slave distinction. A token data packet is sent from the first node to the second. Assuming that, in fact, the first node has
304
nothing to say to the second node, the second node simply receives the token, and then does one of two things. If it has nothing to say to any other node, it simply passes the token on to the next node. But if it does have something to say, it: • modifies the token to indicate that it is no longer available. • inserts the address(es) of the intended data recipient(s). • inserts the data to be sent. • sends the now fattened data packet on its way. The “unavailable” token passes from node to node until it has reached its intended recipient(s) during which time no other node can initiate a new data packet. Once the data has been received, the once-again “available” token carries on round the network until another node loads it up with data. Collision detection. This is now the most widely used method. It is used by TCP/IP and, at a simpler level, by some implementations of DALI where there may often be more than one “master” device on the line. In this case all devices monitor the data line. If they sense it is free they just send their data – which can be to individual addresses, to a group of addresses, or “broadcast”. Assuming that, for the time that it takes to send the data packet, no other device starts talking, all is well. In principle no other device should start talking
STAR
Master or Hub
RING
BUS
Figure 9.24 Examples of network and bus topologies.
CONTROL SIGNALS AND PROTOCOLS
Medium
Description
Simple pair
A pair of conductors without any specification as to how they are laid together. Prone to pick up interference. Unshielded Twisted Pair. A twisted pair of conductors. A twisted pair is not affected by magnetically induced currents. Shielded Twisted Pair. A twisted pair of conductors with an overall conducting shield or screen connected to ground. Less vulnerable to electrical noise than UTP. A cable with a center conductor, and the return conductor in the form of a braided or metal tape sheath round it. As used for TV Aerial connections. Different sizes and characteristic impedances available. Capable of very high bandwidth. For very high bandwidths. Data is carried in the form of laser diode generated light pulses. Communications data is superimposed on the standard 50/60Hz mains supply. Avoids the need for separate control cables. Limited bandwidth. Requires special interface circuits to isolate the control element from the 110/240V mains. Suitable for cordless control over short distances. For cordless control over short distances. (For special purposes it can also be used over longer distances). UTP cable as used for telephone connections etc. Not standardized. Typical data rate 20kbps UTP cable for data rates up to 4Mbps. Impedance not specified. UTP or STP cable for data rates up to 10Mbps. Impedance 100:. UTP or STP cable for data rates up to 16Mbps. Impedance 100: UTP or STP cable for data rates up to 100Mbps. Impedance 100:. This is the ubiquitous “CAT-5” cable that forms the basis of most local computer networks. CAT 5 also lays down standards for cabling practice and for testing that installations are compliant. The standard cable has four twisted pairs within one sheath, and for portable equipment and for patching uses the RJ-45 connector. Using all four pairs at once, it is possible to transmit data at 1000Mbps (1Gbps) A step up from CAT-5, intended for 250 Mbps, but promoted by some as a possible copper contender for 10Gbps ethernet. Requires completely different test procedures to CAT 5, and as at 2002 was a source of confusion because several CAT 5 compatible devices do not work on CAT 6 cabling.
UTP STP
Coaxial
Fiber optic Power line carrier
Infra-red IR Radio Frequency RF Category 1 Category 2 Category 3 Category 4 Category 5
Category 6
Table 9.11 Examples of media used to carry network signals. The “Categories” are cables with a specified performance in respect of data transmission.
while a message is being sent, because all devices are monitoring the line. However, in fast networks the speed of light becomes an issue. A new message may not be sensed by a node at the far end of the line, which may already have started transmission. Carrier Sense Multiple Access/Collision Detection or CSMA/CD is the principle used. Each node must not only monitor the line (“Carrier Sense”) but must also be able to detect collisions. Any data received as a result of the collision must be discarded, and the sender must re-transmit. Clearly if there has been a collision, there is the immediate possibility of another one, so for this reason any re-transmission is delayed by a random amount. For the system to work, it is also important to limit the size of data packets sent at any one time. In DALI the process is comparatively slow and rarely invoked (it is not often that two people simultaneously press a push button in a small lighting system). However, in computer networks, or bus control systems with a large number of nodes, the process is absolutely fundamental. There are limitations in the way that buses and computer networks can be connected together. Figure 9.24 shows the common topologies. The applicable topology depends on a combination of data speed, data packet length, signal regime (e.g. voltage levels), signal distribution regime, the number of nodes and the carrying medium. In principle the higher the demand for speed and capacity, the greater the topology limitation. For lighting control the ideal is a system that allows a free topology (i.e. a mixture of the standard topologies can be used) since this greatly simplifies installation. Buses and networks can use a variety of media to carry the signals. Some of these are shown in Table 9.11. The following sections review some of the common bus and computer network “standards” that are relevant to lighting control. The examples chosen are all used for multiple applications; in each case lighting control is just one application. The examples help to illustrate aspects of bus and network systems that are relevant to different aspects of lighting control.
305
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
® and LonWorks® ® 9.5.2 LON®
Trademark
Meaning
LON
A control network using Echelon's technology A general descriptor of the technology. e.g. a LonWorks node is any node device on the network. There are also LonWorks accessory devices, such as LonWorks transceivers, routers, interfaces and gateways. The integrated circuit (chip) that each LonWorks device is fitted with. These are made by Toshiba and Cypress. Described in more detail in the main text. The communications protocol that links each LonWorks device. It is a complete 7layer protocol according to OSI. Products carrying the LonMark logo are independently certified as being interoperable with any other products carrying the same mark. A LON network services software tool
LonWorks
Neuron
LonTalk LonMark
LonManager
Table 9.12 Trademarks of the Echelon Corporation relating to LON.
306
LonWorks Transceiver
This example is chosen because the LON system is favored for applications where interoperability is a priority. DMX and DALI are simple examples of inter-operable protocols, because, when correctly implemented, devices from different manufacturers can work together, and, within the limits of the protocol, additional devices can be added to an existing system without having to modify the devices already there. Many other “standards” may provide electrical and signal compatibility without necessarily providing interoperability. Thus a security system might use EIA485 communications, but this does not mean that it can share a data highway with an airconditioning control system that is also using EIA485, because the protocols are likely to be quite different. LON is popular, especially in the building control market, precisely because it does offer full interoperability. A presence sensor in an office fitted with LonWorks connectivity can communicate with any other LonWorks device; so for example the security sub-system can use it to report the presence of someone in an office, and the lighting sys-
Networking medium TP powerline etc
® chip 9.5.2.1 The Neuron®
LonWorks Transceiver
Dimmer Node
Lamp loads
Neuron Chip XTAL
Power
Push Buttons Input Interface
Neuron Chip XTAL
Power
Push Button Node
Figure 9.25 Examples of lighting control nodes in a LON system.
tem can use the same sensor to switch the lights on. LON or Local Operating Network was developed by the Echelon Corporation of the USA. Table 9.12 lists some of the Echelon trademarks relating to LON and their meaning. Each node in a LonWorks® system is fitted with a Neuron® chip. A block diagram of two possible lighting control nodes is shown in Figure 9.25. The Neuron chips are available in a variety of packages, and with different memory capability, but they all work on the same principle. A greatly simplified block diagram of the resources within a Neuron chip is shown in Figure 9.26. The idea is that for most applications the Neuron chip represents the entire communications and processing resource, and that a minimum of additional components are required. The principal items are: Communications port: this receives incoming data and sends out outgoing data. In very small systems the Neuron can communicate directly with other devices, but it more generally works through a transceiver. LON is media independent, so the transceiver can be EIA485, Power Line Carrier, RF
ROM
RAM
EEPROM
Application CPU
Network CPU
Media Access CPU
CONTROL SIGNALS AND PROTOCOLS
16 Bit Address Bus 8 Bit Address Bus
Control Block
Application I/O Block
Application Circuitry
Comms Port
Network (Via transceiver in most cases)
Figure 9.26 Simplified block diagram of an Echelon Neuron® chip.
or whatever suits the project. Application I/O block: this section is fitted with a selection of current sinks and programmable voltage pull-ups. Through firmware these can be associated with registers and counters if required. Associated with this section are counter/timers that can be used to generate or measure I/O waveforms. Control block: this includes the master oscillator and microprocessor control. Memory: depending on the version of the Neuron chip, this offers an array of memory. A typical arrangement is: • 2K of static RAM. • 10K of masked ROM that carries the operating system and the communications protocol. • 2K of EEPROM. This contains a unique identification code for each chip, held in part of the memory that is configured read-only. The writeable part of the memory carries the specific user program and data, and some communications overhead related to the particular installation.
An alternative variant of the Neuron chip allows the use of additional external memory. THREE microprocessors: this is what sets the Neuron chip apart. Most other systems use a single microprocessor at the node. Neuron uses three, albeit sited within the same chip. The arrangement recognizes that, when an application process is to take place at the same time as communication, and when chip area is rationed, it is more efficient to separate out the tasks and carry out concurrent processing. The three processors are: Media Access Processor. This handles Layers 1 and 2 of the protocol stack, and drives the communication port. It carries out the collision avoidance algorithm, and communicates with the second processor through the shared RAM memory. Network Processor. This handles Layers 3–6 of the protocol stack. It handles addressing, network variable processing, authentication, network management, routing and similar functions. It communicates with both the other processors through buffers in the shared memory. Application processor. This communicates with the Network processor, and with the Application Type
Operation
Repeater
A device that simply forwards all data packets between two channels. Similar to a repeater, a bridge forwards all packets that match its domains between two channels. Not part of the LonWorks system itself, but a device that works in a similar way to a bridge to link a LON network to an entirely different kind of network e.g. a DALI to LON gateway. This device monitors network traffic and learns the network topology at the domain/subnet level. With this knowledge it is able to selectively route data packets. This eliminates unnecessary traffic 0010 so, for example, a lighting-only subnet does not receive all the traffic relating to air conditioning. Unnecessary traffic slows down the network. This provides the same function as the learning router, but in this case the network traffic management is determined by preprogrammed routing tables. This are programmed in at installation time using a “network management tool” computer program.
Bridge Gateway
Learning Router
Configured Router
Table 9.13 LonWorks router types. While the descriptions here relate to LON, the principles apply to other network architectures.
307
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
I/O block. It runs the program created by the user. The user can program a Neuron chip using a modifed version of the ANSI C language. 9.5.2.2 LON Communication The communication capabilities of LON are fairly comprehensive, however, many of the concepts introduced here are common to other network systems. The idea is to have a communications system that at one extreme works with a few components (for example a simple lighting control system with a dozen devices on the network) up to one with thousands of components (for example a large building where all the security, air-conditioning and main lighting elements can be accessed.) All LonWorks nodes within a system communicate by means of data packets sent between them on a channel. The channel is simply the physical medium – STP, RF etc. A system can, when necessary, use several channels. This may be a matter of convenience, for example one part of an otherwise STP system might
use a cordless control; or another part might use power-line carrier because of the difficulty of installing a separate control cable. Or it can be a matter of data throughput; one part might use a low data rate to keep cost and power requirements to a minimum. Finally each medium may itself have loading limits that restrict the number of individual devices that can be connected to a channel. A typical system using STP allows 64 devices to be connected. Channels can be linked together using routers. Table 9.13 identifies four different kinds. Some of the terms used in the table relate to the method of addressing nodes. When two channels are linked by a repeater or a bridge, they are said to form the same segment since all devices in the segment receive the same data (whereas devices either side of an intelligent router may not). Figure 9.27 shows how a lighting switching system for a multi-storey building could be arranged using LON components. The addressing arrangement for LON seems, at first, rather complex. It is designed to cope with very large systems and to do so in a way that ensures optimum use of individual channels. It is also de-
Lamp groups
78kb/s Free topology network
Router floor N Lamp switch node
Router floor 2
Light sensor node(s)
Presence sensor node(s)
Push button control node(s)
Typical floor with lamp groups, user controls and sensors
Router floor 1
1250kb/s Network
Keyboard
Computer serial gateway
mo
A A A A A A A A A A A A A A A A A A
u se
Monitor
Monitoring and central control computer A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A
A A A A A
Figure 9.27 A LON lighting switching system for a multi-storey building showing the role of routers.
308
CONTROL SIGNALS AND PROTOCOLS
DOMAIN Logical collection of nodes on one or more channels. Communication can only take place between nodes in the same domain. Different domains can share same channels.
SUBNET Logical collection of up to 127 nodes (in the same domain). Up to 255 subnets within a single domain. All nodes within a subnet must be on the same segment.
NODE Individual device in a subnet. A node can belong to two domains, to do so it must be assigned to a subnet within each domain. A node can be a member of up to 15 groups.
Figure 9.28 The LON addressing hierarchy
signed so that, if an individual node has to be replaced, or a new node is to be added, this can be done without disturbing any other part of the network, and with minimum or no programming effort. This means that, in practice, small, simple installations, can use a simplified addressing scheme. As a definition a name is an identifier that uniquely identifies a single object, and does not change over the object’s life time. Within a Neuron chip, the 48 bit ID is the name because it uniquely distinguishes the chip from all other Neuron chips. This name could be used as the object’s address, but this would be very clumsy since it would only allow one-to-one communication, and the network traffic would be very slow as it would be ferrying large quantities of 48 bit identifiers. In practice, therefore, the use of the unique Neuron ID is normally confined to the installation and configuration phase. So an address is an identifier that identifies an object or group of objects within an object class. Unlike a name, an address may be assigned or changed at any time. In the LON system the address-
Data Rate kb/s 4.883 9.766 19.531 39.063 78.125 156.25 312.5 625.00 1,250.00
Number ofpackets per second 12 bytes per packet 20 35 85 180 320 500 560 560 560
Number of packets per second 64 bytes per packet) 5 10 20 40 80 160 270 470 560
Table 9.14 The sustained channel throughput of the LonTalk protocol at different packet sizes and bit rates.
ing arrangement is hierarchical. Figure 9.28 shows the idea. A single domain can have a maximum of 32,385 nodes (255 × 127). Many LonWorks systems are based on a single domain. However, the idea of domains is that they provide a virtual network. Sometimes it is necessary that two networks share the same channel (for example two neighboring sites may have the same RF channel). In such a case the domain idea prevents interference between nodes in different networks. When domain identification is needed 1,3 or 6 bytes are used. Nodes can be assigned to a group. A group is another logical collection of nodes within a domain, but, unlike membership of a subnet, group membership is independent of physical location, and can span different segments. Up to 256 groups can exist within a domain. Channels are configured for different data rates according to the nature of the channel, distance, required throughput, and power consumption. The available data rates are shown in Table 9.14 which also shows the sustained number of data packets that can be sent per second. Average data packets are Address format
Destination
Size
Domain Domain, subnet Domain, subnet, node Domain, group Domain, Neuron ID
All nodes in the domain All nodes in the subnet Specific node in a subnet All nodes in the group Specific node (using name)
3 3 4 3 9
Table 9.15 LonTalk protocol addressing formats. The size is in bytes.
309
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
1
1
Bit Timing
1
0
1
1
0
0
0
Data +
Data
Bit Sync Byte Sync
Preamble allows receivers to synchronize their clocks.
Data block and 16 bit CRC. Forced Manchester code violation. This indicates the end of the packet.
Delay for receiver start and priority slot.
Fixed delay to allow for oscillator variations, size of network and differences between nodes.
Figure 9.29 The message format used by the LonTalk protocol. The figure shows the differential mode format at the output of the Neuron chip. Single ended working is also possible. Note the use of Differential Manchester (bi-phase) coding.
10–16 bytes long, and the maximum permitted size is 255 bytes. The packet size is affected by the addressing method used and Table 9.15 shows the different addressing methods and their corresponding address size. LonTalk messages can be of different kinds depending on the needs of security, reliability and response time. The message format itself is shown in Figure 9.29 and Table 9.16 describes the different message types in order of reliability. LonTalk also has provision for an optional priority system, whereby critical packets get priority in order to speed response times. Finally a system of authentication is available that allows a receiver to check that the sender was authorized to send the message. This is implemented by distributing 48 bit encryption keys to the nodes at installation time. Clearly the use of authentication should be confined to critical applications that need it (e.g. those involving security) since the communications overhead becomes considerable. In order to achieve inter-operability the LON system has a set of rules about how data is formatted and how variables (e.g. temperature, light levels etc.) that may be required by many devices are de-
310
fined. It does allow for “foreign” data to be carried on the network (like the “system exclusive” messages in MIDI) but such data would be confined to an individual application. This brief description of LON and LonWorks devices is intended to give an insight to both the opportunities and limitations of network based systems. Some particular points that have emerged in the use of LON for lighting control now follow. The system architecture lends itself to forming the basis of large scale systems, especially where inter-operability is required, however the considerable communications overhead introduced by the LonTalk protocol can introduce problems of response times. Thus care must be taken to ensure the correct allocation of subnets and groups. Echelon provide high level software and associated hardware tools for developing LonWorks systems so this kind of problem can be minimized. LonWorks systems are being successfully used for lighting control (e.g. by ECS for lighting management systems, and ETC for architectural dimming systems). Because of the sophistication of the system, the node cost is comparatively high. This is not a prob-
CONTROL SIGNALS AND PROTOCOLS
lem at the level of a dimmer pack or control panel, but makes it difficult to achieve an acceptable node cost down at the individual component level (e.g. switchplate or electronic ballast). For this reason hybrid systems are being widely used. In such cases local control is by a dedicated lighting control package using, for example, DALI, EIB, or a manufacturer’s proprietary protocol. A link is provided to the Building Management System by the installation of a LonWorks “gateway” (see Figure 9.18). 9.5.3 CAN CAN stands for Control Area Network. It was originally developed by Robert Bosch GmbH for the automobile industry, but has found much wider application. Unlike the Echelon LON system, CAN does not seek to be a complete solution. This can lead to some confusion when compatibility issues are discussed. CAN itself is only concerned with the two lowest layers in the OSI reference model, and in ensuring that communications at these levels are compatible. Several CAN user groups have developed protocols at the application level. CiA (CAN in Automation) is an international users and manufacturers group that supports a number of open and proprietary protocols riding above CAN, for example: • DeviceNet, itself supported by the Open DeviceNet Vendors Association. Originated by the Allen Bradley Company. • CAL (CAN Application Layer). • SDS (Smart Distributed System) originated by Honeywell.
• CAN Kingdom. • CANopen. CAN is a multi-master architecture, with a maximum data transfer rate of 1Mb/s. Practical data rates achieved by CAN at different distances are shown in Table 9.17. Unlike traditional networks, CAN does not send messages point to point and does not rely on addressing. It broadcasts all messages, and each message has an identifier which refers only to the content of a message, not its destination address. CAN was developed to provide efficient real time control with a high level of security; for example it is used for life safety-critical applications such as the control of airbags in automobiles. Its automobile origins mean that it is now well supported by chips of varying sophistication. The standard chips simply provide CAN communication, the more sophisticated chips provide other facilities, for example: • Application specific output interfaces (e.g. the control of automobile lighting, seat and window motors, displays etc.). • Application level communication. With the adoption of CAN as the basis of industrial control buses by industry, controller chips carrying the application protocols are available. For example Philips offer chips that can be configured to DeviceNet, CANopen or OSEK. • Support for multiple buses. As with other bus systems, the choice of physical layer is left to the user. For any one network, only one type of physical layer (or “channel”) can be used; for example twisted pair at a specified data rate. This is why controller chips supporting multiple buses can be useful. By confining a CAN
R elia b ility T y p e
D escrip tio n
1=
A ck n o w led g ed
1=
R eq u est/resp o n se
2
R ep eated
3
U n ack n o w led g ed
A n en d -to -en d ack n o w led g ed serv ice. W h en a m essag e is sen t to a n o d e o r a g ro u p o f n o d es an in d iv id u al ack n o w led g em en t is ex p ected fro m each receiv er. If n o ack n o w led g em en t receiv ed , th e sen d er re-tries. T h e n u m b er o f re-tries is co n fig u rab le, an d d u p licate m essag es are ig n o red . F o r ack n o w led g ed serv ice th ere is a m ax im u m o f 6 4 n o d es w ith in a g ro u p . T h is is a sy stem o f p o llin g . In th is case th e receiv in g d ev ice p ro cesses th e in co m in g m essag e b efo re sen d in g an ack n o w led g em en t, an d th e ack n o w led g em en t itself m ay in clu d e d ata. H ere th e m essag e is sen t to a n o d e o r g ro u p o f n o d es sev eral tim es, b u t n o resp o n se is ex p ected . T h is is u sefu l fo r m essag es th at m u st b e b ro ad cast to larg e g ro u p s o f n o d es. T h is g iv es th e fastest co m m u n icatio n , b u t is th e least reliab le. It sh o u ld o n ly b e u sed w h en th e ap p licatio n is n o t sen sitiv e to th e lo ss o f an in d iv id u al m essag e.
Table 9.16 LonTalk message types ranked in order of reliability.
311
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Data Rate
Maximum distance
50 kb/s 125kb/s 250kb/s 500kb/s 1 Mb/s
1,000m 500m 250m 100m 40m
Table 9.17 Typical performance of CAN.
network to a single channel, and by virtue of the CAN implementation itself, some important advantages of simplicity and security are obtained:
• A CAN node requires no knowledge of the system configuration. • Message lengths, including comprehensive methods of error detection, are kept comparatively short. • No node addresses are required. • There is thus no theoretical limit to the number of nodes (but in practice a limit may be determined by the execution of the physical layer). • Nodes can be added to a CAN network without any change in software or hardware to other nodes. • A message is sent to all nodes simultaneously,
Interframe space
Interframe space
Data frame S O Arbitration field F
Control field
CRC field
Data field
Standard format
S O F
R I r 11 bit T D O DLC identifier E R
Ack field
EOF
OR OVERLOAD FRAME
Extended format
S O F
11 bit identifier
S I R D R E
18 bit identifier
R r r T 1 O O
DLC
SOF Start Of Frame. A single dominant bit. 11 bit identifier Most significant bit first. 7 most significant bits must not all be recessive RTR Remote Transmission Request. In Data frames RTR is dominant bit; in Remote frames it is recessive. IDE Identifier Extension bit. It is transmitted recessive in the arbitration field in extended format, and dominant in the control field in standard format. SRR Substitute Remote Request bit. A recessive bit only transmitted in extended format. r1 r0 Reserved bits, transmitted dominant (but receivers “don’t care”). DLC Data Length Code. A 4 bit code that says how many bytes are to follow in the Data Field. Byte range 0-8. CRC The CRC field contains the CRC remainder number, plus a CRC delimiter which is always a single recessive bit. ACK The ACK field consists of the ACK slot, and ACK delimiter. The delimiter is a recessive bit. All stations receiving a matched CRC report this in the ACK slot by overwriting the recessive bit of the transmitter with a dominant bit. EOF End Of Frame. Indicated by a sequence of seven recessive bits. Interframe The transmitter must send an intermission of 3 recessive bits before stopping transmission. In this space time a node can signal an Overload frame. After the intermission the bus goes idle until a new message is transmitted.
Figure 9.30 The data frame format in CAN. At the top, the complete frame, below it the two different versions of the arbitration and control fields.
312
CONTROL SIGNALS AND PROTOCOLS
Frame Type
Meaning
Data Frame Remote Frame
Carries data from a transmitter to the receivers
Error Frame Overload Frame
Transmitted by a bus unit (node) when it wants data. By indicating only the identifier, it alerts the node that has the data to transmit the required data frame. Transmitted by any unit when it detects a bus error. Used to delay the start of new Data or Remote Frames. This may be required by some receivers, especially in the case that the physical layer introduces some latency.
Table 9.18 Frame formats in CAN
its content described by an identifier. The identifier is not concerned with the destination of the data; all nodes receive the data, and decide by message filtering whether or not they should act on it. An important attribute of the identifier is that it determines the priority of the message. • As a consequence of this, all nodes act on a relevant message simultaneously. • Within a CAN network it is guaranteed that a message is simultaneously accepted by all nodes or by no node. The CAN bus can have only one of two complementary logical values, and these are referred to as the dominant (normally Logic 0) and recessive (normally Logic 1). In the case that there is simultaneous transmission of both bit types, the bus assumes the dominant value. A principal feature of CAN is that message arbitration is bitwise. Unlike some other systems using collision detection, it is not necessary for re-transmission of valid messages to be terminated, with consequent loss of time. If two or more units start transmitting messages at the same time, the bus access conflict is resolved by reference to the identifier. During arbitration, every transmitter compares the level of the bit transmitted with the level detected on the bus; if the levels are equal, it continues transmitting, but if a recessive level is sent and a dominant level is detected, it immediately withdraws. CAN is intended to give reliable operation in electrically harsh environments where messages can be lost or corrupted. Table 9.19 summarizes some
aspects of the error correction and detection performance. Table 9.18 lists the four different types of messages carried on the CAN bus. CAN’s evolution since its introduction in the 1980s has resulted in the identifier section of the message being of two different lengths. The standard identifier is 11 bit, and the longer one is 29 bit, resulting in standard and Error detection method/perform ance Monitoring CRC (Cyclic redundancy check)
Bit stuffing Data streams are sent NRZ (Non Return to Zero) meaning that there are no transitions within a bit time. A bit must be either dominant or recessive throughout the bit time. Message frame check
Error detection properties
Residual error probability of an undetected corrupted message
Meaning Bus level continuously monitored by transmitters to check that actual bus level is the same as that transmitted. The CRC is optimized for bit counts <127. The polynomial to be divided is derived from the destuffed data stream prior to the CRC. The generator polynomial is detailed in the CAN specification. Each receiver does a new CRC calculation on the data actually received. If the result does not match the remainder data sent in the message, the whole message is discarded. The majority of a message is coded using the technique of bit stuffing. If a transmitter detects that it should send five successive bits of the same value, it automatically inserts a complementary bit in the transmitted bit stream. At the receiver a STUFF ERROR is flagged if six bit times show the same level.
This looks at the standardized parts of the message and checks there are no illegal bits. It also checks the amount of data received, which should match the number of bytes advised in the control field of the message. All global errors are detected. All local errors at the transmitters are detected. Up to 5 randomly distributed errors in a message are detected. Burst errors <15 bit times in a message are detected. Errors of any odd number in a message are detected. < Message error rate u 4.7 u 10-11
Table 9.19 The error detection performance of CAN.
313
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
extended CAN frames. Figure 9.30 shows the message format for a data frame; the remote frame is of similar format, but without the data field. Error and overload frames are very short. CAN is important for lighting control for two reasons. First it now plays a significant part in the control of vehicle lighting. Second, the wide availability of CAN controller chips has made it a valid choice as the basis of medium sized architectural and commercial lighting control systems. It does not lend itself directly to large scale or multiple network systems (since CAN works on a single network basis) but this limitation can be overcome by the introduction of appropriate bridge and router products. It is important to make the point that just because a system uses CAN does not mean that it provides any kind of standardization or interoperability; this would only be the case if it used a standard application layer protocol.
within a process, and at allowing compliant products from many manufacturers to work together. Apart from LON and the CAN based field buses such as SDS and DeviceNet, there are several others, for example ProfiBus, InterBus and ModBus. Each have different characteristics in respect of data speed, bus utilization efficiency, network size, security, suitability for safety critical applications etc. The result is that most specialist manufacturers of sensors, actuators and logic controllers offer product variants to match the different bus possibilities. Apart from LON and CAN the industrially biased bus systems do not have a part to play in lighting control – but if they ever needed to, it would be a simple matter to devise suitable interface products. At the same time as LON appeared as a viable basis for large scale distributed lighting control, other initiatives were already underway. These included the European Installation Bus, EIB (also known by the name of a proprietary version Instabus) originated in Germany and BatiBus originated in France. At a consumer level the European Home System Association were working along the same lines. In all cases the aim was to come up with system components that at one end would be suitable for installation at home without the need for any
9.5.4 EIB Bus systems used for industrial purposes are often referred to as field buses. They are all aimed at simplifying the interconnection of items of equipment
Connection to other domains Backbone ACU 1
2
3
4
5
6
7
8
9
10
11
12
13
14
ACU 15
LCU 1
2
3
4
5
6
7
8
9
10
11
12
13
14
LCU 15
Line
BAU 1
2
BAU 256
Figure 9.31 The EIB addressing hierarchy, showing one domain. ACU= Area Coupling Unit. LCU = Line coupling unit. BAU = Bus Access Unit.
314
CONTROL SIGNALS AND PROTOCOLS
Maximum frame length 12 x 8 bits 1
2
3
4
Address Field Source Add
Dest Add
5
6
7
Control Information field
8
9
10 11
12 13
Data Field (max)
14 15 16
17 18 19
20 21
22 23 Check
Control Field
0
Figure 9.32 The EIB PDU frame structure. Actual data is equivalent to 14 bytes (with a “long frame” extension to 230).
specialized installation or programming procedures, and at the other could form the basis of systems with thousands of nodes. The further aim was that the components should be suitable for all kinds of technology found in buildings, especially those related to heating, air conditioning, security, power control, energy metering, time – and lighting. In principle these initiatives have merged under the EIB umbrella, so this section describes only EIB. The European Installation Bus Association (EIBA), based in Brussels, is an industry grouping that sets the standard and as at 2001 there were about 110 vendors offering 5,000 interoperable products. Such products display an EIB logo, but are usually marketed under a manufacturer’s own brand. For example the firms Insta, and Siemens, refer to Instabus EIB, and ABB refers to i-bus EIB. EIB uses several techniques and concepts that have already been described in the sections on LON and CAN, so common details are not repeated here. The address architecture is shown in Figure 9.31. The concept is that at the initial level up to 256 devices can be connected to one line. (Actual products offered may limit the number of devices that can be
Medium
EIB specification
Twisted pair
9600 bps. Range 1000m. CSMA with bit-wise collision avoidance. Dominant 0. Power carried on same pair. Fast polling allows up to 14 devices to be polled for single bit status information within 50ms. Typical delay for two simultaneous transmissions 100ms. 1200 bps. Range (max) 600m. Spread spectrum FSK. Randomized re-transmission. Lines have separate carrier frequencies. Range (approx) 300m Separate EIB Anubis standard. See section 9.5.6
Power line RF Ethernet
Table 9.20 Media specifications for EIB.
connected to 64). For many small installations, this would be all that is required, but for larger installations the hierarchy proceeds as: • up to 15 lines can be connected (via routers referred to as line couplers) to a main line to form an area. • up to 15 main lines can be connected to a backbone line via routers referred to as area couplers to form a domain. It is possible to have multi-domain systems if necessary. The architecture is intended to ensure that local traffic does not get bogged down with traffic relevant only to another line, while at the same time allowing area wide information to be disseminated or collected. For example a feature of EIB is that it
Figure 9.33 Example screen from EIB application software from Grässlin. Here the clock application is shown as serving two objects “Date” and “Time”. Another page sets the frequency of transmission for each object, which is typically every day, hour or minute.
315
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Power supply with choke
RS232 Gateway
EIB Line coupler
EIB room thermostat clock
EIB radiator servo drive
EIB master clock (derives standard time from radio tranmission)
EIB slave clock
Figure 9.34 Examples of EIB products, all from Grässlin
316
makes specific provision for a real time signal to operate slave electric clocks; clearly this might be required over the whole system. On the other hand, switching of lights would normally be a local concern, unless, for example, a building management system wanted to poll lighting status, or needed to impose an emergency condition. In addition to the coupler concept, EIB makes provision for other kinds of router similar to those listed in Table 9.13. At the “line” level EIB specifies in detail the way different channel media are to be used. Some of these are still being finalized. Table 9.20 lists the main media used. A detail about the twisted pair medium is that, because it carries power, some provision has to be made for sustaining power when heavy traffic is on the line. A dominant signal on the line is almost a short circuit (whereas the recessive signal is almost an open circuit voltage). In the DALI system, traffic is deliberately limited to ensure that sufficient power remains available, and anyway the Manchester coding ensures that a dominant condition is not maintained for any length of time. EIB similarly ensures continuity, but also allows for additional power supplies on a line. Clearly such an arrangement could have the opposite effect, i.e. a second power supply could hold the line recessive. This is avoided by the use of chokes, and a minimum separation of power supplies of 200m on one line. Such chokes may also be used to permit a single power supply to power more than one line, with the chokes isolating one line from another in respect of the data signals. The EIB communication protocol follows the OSI 7-layer model (although layers 5 and 6 are “reserved” and are transparent to the user). Messages can be point to point, “group” or “broadcast”. The EIB frame or Protocol Data Unit (PDU) is shown in Figure 9.32. The data carried is equivalent to 14 bytes, but the EIB specification allows this data space to be used in many different (but specified) ways. Thus, for example, there are identifiers for most physical quantities like temperature, voltage, current, length, speed, power etc.; and the data can be formatted in many ways including: • individual bit
CONTROL SIGNALS AND PROTOCOLS
• • • •
8, 16, and 32-bit bytes signed and unsigned floating point 16 or 32-bit date expressed as 24 bits time expressed as 24 bits EIB operates at various levels of sophistication. There are many ways in which messages can be used, including that of simply carrying a defined variable (e.g. a light level) which can be used by any device that is interested, in a similar manner to CAN. Small systems can, therefore, be self configuring and require little specialist knowledge on the part of the user. However, it is expected that most installations of any size, requiring unit addressing, will use the services of a specialist installer trained in EIB system configuration. Although many EIB vendors offer their own configuration tools, the EIBA offers a vendor neutral ETS, EIB Tool Software for the design and configuration of projects. EIB has consolidated a strong position over a period of 12 years or so, and is an attractive proposition for some types of building. However, in common with LON, it does carry a comparatively high node cost, and response time can be slow on a heavily loaded network. These considerations are not an issue with many of the applications of EIB (especially heating control etc.) but are an issue with lighting at the individual luminaire level. Several vendors of EIB products also offer DALI lighting control products in the same physical packages and it seems likely that the two protocols will co-exist, especially with the availability of EIB-DALI gateways.
9.5.5 Powerline, X10 and CEBUS 9.5.5.1 Powerline In the protocols already discussed, the possibility of using powerlines as the carrying medium has been mentioned several times. The principle is very simple; since nearly every device in a system is connected to electric power, it should be possible to use the power cable itself as the control cable. If the power is at 50–60Hz, it is possible to superimpose a signal of a much higher frequency, which can be extracted by the use of a simple high pass filter. There are, however, some difficulties: A power supply network is complex, with many alternative paths, dead ends etc. Unless special equipment is used, transmission distance and data capacity are therefore limited. There is a safety issue in isolating the control circuits from the mains. Transformers are still used for some applications; optoisolators help achieve lower costs. The supply network is home to a lot of electrical interference that can easily corrupt data sent over powerlines. 3kHz–9kHz
9kHz–95kHz
95kHz–148.5kHz 95kHz–125kHz and 140kHz–148.5kHz 125kHz–140kHz Access protocol
Injection mode
Figure 9.35 An EIB multi-scene lighting control from Insta. They also offer a DALI lighting controller in the same housing.
Reserved for the use of electricity supply companies (e.g. for load shedding and metering). Max 134dB (PV) Reserved for the use of electricity supply companies and their licensees. Max 134dB (PV) Reserved for consumer use. Max 116dB (PV) Do not require an access protocol Requires the use of an access protocol to allow several systems to co-exist on the same installation. Max transmission time 1s Minimum inter-transmission gap 125ms Transmission in progress indicated by 132.5kHz >80dB (PV) Differential. (Common mode could disturb the operation of RCDs.)
Table 9.21 In Europe signalling on the 230V mains is limited to 3kHz–148.5kHz, according to EN50 065. The table indicates some of the limits that apply. The method of measuring the signal voltage is tightly specified in the standard.
317
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Electronic equipment must meet EMC requirements which determine the amount of mains borne “interference” that can be generated. Clearly the deliberate generation of a signal intended to go back down the mains could breach the limit. In fact the use of powerline signalling is regulated (by CENELEC in Europe, FCC in USA), see Table 9.21. The regime is somewhat less restrictive in the USA. In practice the above limitations lead to two things. In sophisticated networks like LON and EIB, powerline signalling is reserved for those applications where it is cost effective; for example when signals must pass between two buildings on the same site. At the other extreme it is widely used (mainly in the USA) for consumer applications where the occasional error is not considered a problem. Although it is possible to have full error correction across a powerline network, it is generally accepted that powerline signalling is not used for any safety critical applications. Powerline data is usually transmitted by Amplitude Shift Keying ASK, or Frequency Shift Keying FSK. ASK simply modulates a carrier between two different levels, usually zero output and full output, to give Logic 0 and Logic 1. FSK shifts the carrier between two different frequencies to give the two logic values. Sometimes spread spectrum technique is used to improve immunity to interference and lower the peak signal value. In this case the carrier is continuously swept between two different values.
House code
H1
H2
H4
H8
A B C D E F G H I J K L M N O P
0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1
1 1 0 0 0 0 1 1 1 1 0 0 0 0 1 1
1 1 1 1 0 0 0 0 1 1 1 1 0 0 0 0
0 0 0 0 1 1 1 1 1 1 1 1 0 0 0 0
Table 9.22 The 4-bit House Code of X10.
318
9.5.5.2 X10 The X10 system, introduced by BSR in 1978, has the distinction of being the longest established consumer powerline control product. It has sold under many well known consumer brand names, and is still popular today, especially amongst hobbyists. It is a low cost system that is easy to interface to personal computers. X10 products do not, however, meet CENELEC requirements, and, therefore, are not generally available in Europe. X10 uses a 1ms burst of 120kHz to represent Logic 1. The absence of such a burst is received as
Function or number
D1
D2
D4
D8
D16
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 All units OFF All lights ON ON OFF Dim Bright All lights OFF Extended code Hail Request Hail Acknowledge Pre-set Dim
0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 0 0 0 0 0 0 0 1 1 1
1 1 0 0 0 0 1 1 1 1 0 0 0 0 1 1 0 0 0 0 1 1 1 1 0 0 0
1 1 1 1 0 0 0 0 1 1 1 1 0 0 0 0 0 0 1 1 0 0 1 1 0 0 1
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1
Extended data Status = ON Status = OFF Status Request
1 1 1 1
1 1 1 1
0 0 1 1
0 0 0 0 1 1 1 1 1 1 1 1 0 0 0 0 0 1 0 1 0 1 0 1 0 1 0 or 1 0 1 0 1
1 1 1 1
Table 9.23 The function codes of X10. In the special case of “Preset Dim” the most significant bit of the level is D8. The four least significant bits are given by the value of the H bits.
CONTROL SIGNALS AND PROTOCOLS
60Hz Carrier
11 Cycles for a standard message
120Hz bursts at zero crossing
Representing data with complement
1
1
0
Start Code
1
0
1
1
0
0
1
1
0
0
1
1
House code example G
0
1
0
0
1
1
0
Function code example all lights off
Figure 9.36 A typical X10 message. Here the House Code is “G”, and the Function Code is “All Lights OFF”.
Logic 0. The receiver is only operational at the zero crossing point of the mains, so, by definition, the data rate is limited to 120 bps (on 60Hz mains). In fact the data rate is around half this because, with the exception of the start code, each instruction is sent in complement form. Transmitters are arranged to transmit three times within each half cycle, with a phase difference of 60° between transmissions, so that they can service 3-phase supply networks (very common in the home in the USA). Typical transmitter power is 60mW, and the output signal is 5V peak to peak. In order to allow multiple systems to co-exist on the same or neighboring power networks, X10 allows for 16 “House Codes”. It then uses a 5-bit code to select from a choice of 31 numbers or designated commands. Tables 9.22 and 9.23 show the code allocation. Figure 9.36 shows a typical X10 message, occupying 11 power cycles; the start code is always 1110, sent plain. Some other rules concerning X10 are as follows: • Messages are always sent twice without a break. • After each such double message, there must be a break of at least 3 cycles. • The exception to both of the above is if either “Dim” or “Bright” are selected; these are transmitted continuously without any break for as long as the button is pressed.
• Additional data can be sent over an X10 link using the “Extended data” facility. Such data must follow without a break in the form of 8-bit bytes; the first byte indicating how many bytes are to follow. • Additional function codes are available using the “Extended Code” facility. However not all X10 equipment offers either the Extended Code or Extended Data option. X10 equipment can be made more or less sophisticated. The transmission method allows for error detection (by detection of incorrect complements and mismatch of repeated messages), and corrupted messages can be rejected. More advanced realizations can include collision detection. However it inevitably remains a comparatively slow protocol, and its “Extended” facilities do not represent a sound basis for larger systems. Thus while within its limitations X10 is remarkably successful, for more general use something more powerful is needed. 9.5.5.3 CEBus® ® Since this section was written it has become clear that CEBUS as a consumer lighting control protocol has not been widely accepted by the market despite early support by industry heavyweights such as GE (who have since withdrawn from the market). The section is retained to illustrate a number of useful
319
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
AC Line Voltage 25 Cycles 100 us +/- 100 ns
Individual scene selection panel Figure 9.38 The chirp used by CEBus on power lines. It is a 7V p-p swept carrier, consisting of 25 cycles of 100kHz 400kHz lasting 100ms. Remote controlled dimmer unit Multi-room remote control
Panel mounted dimmers with master control Figure 9.37 Examples of X10 equipment. These are from the Lightolier range from Genlyte.
concepts, such as spread spectrum, jabber, and reliability of service. It also serves to illustrate that if protocols are made too complex without any obvious advantage to the job in hand, they are unlikely to be widely adopted. The EIA introduced CEBus as a standard under EIA-600.10, ratified by ANSI, and ultimately intended for consideration by IEC. (Although here local differences in regulations would preclude compatibility at the physical layer.) CEBus, which, by the way, is a registered service mark of the EIA, has the following mission statement (taken direct from the introduction to the standard): • To develop a universal method for devices in the home to communicate, regardless of manufacturer. • To allow the introduction of new products and services to the home with minimum confusion to consumers.
320
• To meet the majority of anticipated home control requirements with a single, multi-media network standard. • To minimize the redundancy of control and operation methods among devices and equipment in the home. CEBus uses four layers (1, 2, 3 and 7) of the OSI model, with most functions of the Transport, Session and Presentation layers being added to Layers 3 and 7. A CEBus feature, not part of the OSI model, is the Layer System Management (or LSM) that applies across all layers. One important task that this carries out is to prevent any one node going into jabber. If for some reason a node sends out a stream of data above a certain length (100ms) the LSM shuts it down for at least 1s to ensure that no node monopolizes the medium. The Physical Layer (1) is divided into two sublayers, one that is medium dependent, and the other, the Symbol Encoding Sublayer, that is the same for all media. CEBus is most widely used with Power Line Carrier, and in this medium functions as follows. The medium is said to be an inferior state when no carrier is present, and superior state when carrier
Symbol times
Logic symbol
1 2 3 4 8
1 0 EOF (End Of Field) EOP (End Of Packet) Preamble EOF
(100Ps) (200Ps) (300Ps) (400Ps) (800Ps)
Table 9.24 Logic symbols in CEBus.
CONTROL SIGNALS AND PROTOCOLS
1
0
1
1
0
100Ps
200Ps
0
0
1
1
1
Figure 9.39 Sending data with CEBus. The red chirps are “Superior Phase 1”, and the green chirps are “Superior Phase 2”. The equivalent data bits are shown below.
EOF
Information field <32 byte
EOF
Source house code <16 bit
EOF
Source address field <16 bit
The Preamble EOF symbol is used for synchronization purposes. The “delivery service” in CEBus is of four kinds, as shown in Table 9.25. This should be compared with Table 9.16. Within the CEBus standard all receivers must support all levels of reception; but the equipment vendor can choose what level(s) of service are provided by the transmitter. CEBus is supported by its object-oriented Common Application Language or CAL. The facilities required are similar to, but on a smaller scale than, those provided by EIB and LON, and similar development tools are available. CEBus uses the context model. Context groups cover such areas as Lighting, Environment Control, Security, Energy Management, Communications, Appliances etc. The Lighting group includes sensors, light level control, scene control etc. Such objects are associated with variables, and it is possible, for example, to relate the light level object to: • default value • current value • minimum value
EOF
Destination address field <16 bit
EOF
Destination address field <16 bit
EOF
Preamble EOF
Control field identifies frame priority and service type
Preamble
is present. (Compare the CAN nomenclature of recessive and dominant.) The inferior state is only adopted between data packets, and in the preamble. The data itself is sent as a continuous signal, divided into symbol times of 100μs. Within each symbol time a chirp is sent. A chirp consists of 25 cycles of a carrier swept between 100kHz and 400kHz, as shown in Figure 9.38. The maximum amplitude is 7V peak-to-peak on 120V lines and 14V on 240V lines. Chirps can be of Superior Phase 1, or Superior Phase 2 state. A Phase 1 chirp has the same phase as the data packet’s first chirp; a Phase 2 chirp has the opposite phase to that of the first chirp. Logic symbols are defined by the number of symbol times as shown in Table 9.24. The start of each symbol is indicated by a change of phase, as shown in Figure 9.39. The result of this method of data transmission is that while the raw data transmission rate is 10K chirps per second, the useable data rate is around 5Kb/s. Data is sent in frames, as shown in Figure 9.40. The Preamble field consists of random inferior and superior states and is used for collision detection.
Figure 9.40 The CEBus data frame. The house code serves exactly the same function as house code in X10.
321
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Reliability Type 1 Addressed Acknowledged
2
Addressed Unacknowledged
3
Acknowledged
4
Unacknowledged
is the basis of Home Plug and Play™, a free specification intended to simplify the installation of networked devices in the home.
Description This is the most reliable service because packets continue to be transmitted until a valid acknowledgement is received. However this service cannot be used for group addresses or broadcast messages. This is the second most reliable service. It depends on sending each packet several times. The receiver discards duplicate information. While more reliable than the unacknowledged service, this service only sends the packet once, and the (unaddressed) acknowledgement is also sent only once. This gives the fastest communication, but is the least reliable. The node transmits a single copy of the data packet. Not recommended on noisy powerlines.
9.5.6 Ethernet and TCP/IP 9.5.6.1 The LAN
Table 9.25 The delivery services offered by CEBus.
• maximum value • required value • transition time to required value • step size etc. The CAL Interoperability Council or CIC supervises interoperability issues. EIA-721 Generic CAL Synchronizes internal clock generator
Indicates type of payload
Preamble
Dest address
Source address
Length of Data
IEEE802.2 Header Optionally with snap extensions
Data 46 - 1500 bytes at 10MHz
C R C
8
6
6
2
3 or 8
Variable
4
Number of bytes
Figure 9.41 The IEEE 802.3 frame format used in LANs.
322
Most of the concepts relating to networking computer-like devices together have been introduced as part of the descriptions of the protocols and bus systems already described. However these have all been related to comparatively low data rates and small data packets. In the late 1970s Xerox Corporation, Intel Corporation and Digital Equipment Corporation defined Ethernet as a way of achieving a Local Area Network or LAN. It was intended as a method of linking a number of computers together so that they could share information, exchange files etc. – and particularly to realize the client-server architecture. In this architecture one of many client computers may require information from the server that stores and keeps up-to-date information required by all the clients. The server should be able to deal with many clients simultaneously. An obvious application for this architecture is an accounts office, where several people may be operating different aspects of the accounting process, yet where all the results must be kept together. Now the same concept is used for systems controlling large scale entertainment systems (including lighting control) and for systems controlling the facilities within large buildings.
CONTROL SIGNALS AND PROTOCOLS
Node device
Node device
Node device
E.g. Computer Optional RJ45 floor plates or similar
ith LE w CAB tor -5 T c A e C conn Rj45
Node device
Total length of cable strictly limited typical 100M
Node device
Node device
Figure 9.42 Simple ethernet hub arrangement. If a switching hub is used there can be simultaneous “conversations” between different pairs of nodes. The hub itself may be a node in a larger network.
The original ethernet was slightly modified and adopted as an IEEE standard (now IEEE 802.3) However it is now this standard that is usually meant when referring to ethernet. The data sent over an ethernet link in the form of a MAC (Media Access Control) frame, of the form shown in Figure 9.41. Notice how the information quantity is huge compared with that sent in the control networks described earlier. The size of the information sector of the frame is shown for ethernet operating at 10Mb/s. Calculations relating the distances between nodes, the speed of light, the collision detection mechanism, and the statistics of network useage have determined that at this data rate optimum results are obtained if the information “payload” is not less than 46 bytes, and not more than 1,500 bytes. Files or other data being transmitted over the LAN have, therefore, to be
“chopped up” into datagrams of 1,500 bytes or less if, as is often the case, they are bigger than this. The first practical versions of ethernet used coaxial cable as the carrying medium. Computers could be linked together using “T” connectors so they shared the same coaxial cable. This method is now not used, and, at the local level, all connection is by twisted pair cable – the CAT-5 cable referred to in Table 9.11. To ensure reliability and to simplify transceiver design a star topology is used, based on the idea of ethernet hubs. The principle is shown in Figure 9.42. Each node has a separate pair of wires for transmission and reception. Over the years ethernet has got faster. Most office networks now use fast ethernet, running at 100Mb/s, and now gigabit ethernet has arrived, running at 1,000Mb/s. In these variations the maximum and minimum number of bytes carried in the information payload within the MAC frame are different. The ethernet hub is available in various forms, just like the routers listed in Table 9.13. A small network might use a single simple hub that simply passes all transmitted data to all nodes. A switched ethernet hub is an intelligent router that knows the identity of each node, and routes traffic accordingly. Large systems may interface to wider networks, and this may necessitate the use of a high capacity backbone, possibly using fiber-optic communication to allow for long distance transmission. The discipline of installing LANs is now fairly well understood. Failure to observe the rules about allowable cable length and permitted topologies results in systems that simply do not work. Lighting control systems that use ethernet as a component are no exception. 9.5.6.2 Internet Protocol In this chapter several different kinds of data frames have been introduced. At base they all have much in common, but vary in complexity according to the data throughput and the complexity of the network involved. It will have already been observed that protocols are often nested or layered one within another.
323
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Applications e.g. SMPT FTP HTTP
Transport TCP UDP
Internetwork
ICMP
IP
ARP RARP
Network Interface and hardware e.g. Ethernet; FDDI; Wireless
Figure 9.43 The TCP/IP protocol stack.
The IEEE 802.3 data frame referred to in the previous section is just one way of ferrying data around a network. In the Internet a particular piece of data may circulate on ethernet (802.3) in an office environment, but then be carried on some entirely different carrier, such as ATM (Asynchronous Transfer Mode) in the big wide world. Wide Area Networks (WANs) use several different carrying protocols, but the actual data carried may be the same. While the OSI 7-layer “protocol stack” is considered the textbook model, in practice the TCP/IP (Transmission Control Protocol/Internet Protocol) protocol stack is more widely used. This is because it was already in use before the OSI model was ratified, and particularly because it was adopted as a mandatory standard for many USA government applications. The TCP/IP stack is shown in Figure 9.43. Here it becomes clear that the 802.3 ethernet lives at the bottom of the stack (along with alternatives) in the Network Interface and Hardware layer, combining the functions of layers 1 and 2 of the 7 layer model. Some of the alternatives to ethernet may themselves
324
be reliable methods of communication, but it is generally assumed that at this level communication is made on the basis of “best effort”. Immediately above this lies the Internetwork Layer. This can be regarded as shielding the upper layers from the physical aspects of communication and creating a standardized “virtual network”. This layer mainly uses Internet Protocol or IP, in the form of IP datagrams, which in ethernet terms represents the data payload shown as “data” in Figure 9.41. IP is a so-called connectionless protocol that does not assume reliable communication, and does not itself provide reliability, message flow control or error recovery. The internetwork layer is also used to carry other protocols concerned with message handling and with addressing, and the most important of these are listed in Table 9.26. The Transport Layer is the layer responsible for the reliable exchange of information. Two different protocols live here and the distinction between them can be of considerable significance when IP is used in lighting control and similar applications. The simplest is User Datagram Protocol or UDP. Like IP itself it is connectionless and does not itself provide reliability, flow control or error recovery. The main additional service it provides over IP itself is its provision of a multiplexing facility for supporting multiple applications. Here the idea of ports is introduced (esentially a method of subaddressing) to allow for several processes to communicate concurrently. In practice UDP can be quite reliable in self contained networks, and has the great advantages of high speed, simplicity and the ability to support multi-casting. Applications where errors can be compensated for by repeat transmissions or judicious use of acknowledgement requests (ACK) are successfully implemented using UDP on its own. However, communications over large networks, over the internet itself, and critical applications all demand something better and this is provided by Transmission Control Protocol or TCP. TCP has the following important attributes: Connection oriented. A direct connection is established between the sender and receiver.
CONTROL SIGNALS AND PROTOCOLS
Protocol
Meaning
Layer in the TCP/IP stack
What it does
FTP SMTP
File Transfer Protocol Simple Mail Transfer Protocol
Applications Applications
Used for high speed file transfers, e.g. disc to disc. Used as an internet mailing system.
Applications
Used for interactive terminal access to remote internet hosts. Used for transferring HTML “documents”. It is the basis of web browsing. Described in more detail in Section 9.6.3. The main reliable method of communication. Described in Section 9.5.6.2. A fast IP protocol that provides multiplexing of applications, but not reliable communication. Described in Section 9.5.6.2. The main protocol that standardizes the data format regardless of the final method of transmission. Described in section 9.5.6.2. This rides in standard IP datagrams and is used as a support service, in particular to report errors (but not, thereby, to make IP reliable) and to support the “Ping” application described in Section 9.5.6.3. A protocol specific to a particular network that converts high level IP addresses to physical network addresses. Similar to ARP, but used when a device does not know its current IP address.
TELNET HTTP
HyperText Transfer Protocol
Applications
TCP
Transport
UDP
Transmission Control Protocol User Datagram Protocol
IP
Internet Protocol
Internetwork
ICMP
Internet Control Message Protocol
Internetwork
ARP
Address Resolution Protocol
Internetwork
RARP
Reverse Address Resolution Protocol
Internetwork
Transport
Table 9.26 Some of the main protocols used in the TCP/IP stack shown in Figure 9.43
Duplex operation. Concurrent data streams in both directions are supported. Standardized data stream. TCP takes the data from the application and converts it into a continuous byte stream divided into TCP segments (so the application itself does not need to prepare datagrams). The TCP segments are in turn passed on to IP for transmission. Multiplexed applications. This is achieved through the use of ports, as in UDP. Reliable communication. This is achieved by a process of continuous acknowledgement by the receiver. Each transmitted byte is assigned a sequence number, and if a positive acknowledgement is not received within a time-out interval, the data is retransmitted. Because the data is transmitted in TCP segments it is in fact sufficient for only the sequence number of the first byte in the segment to be transmitted. The TCP receiver uses the sequence numbers not only to ensure that all data is received, but also to re-arrange segments when they arrive out of order
(which may happen due to deficiencies in the actual transmission method) and to eliminate any duplicated segments. Flow control. In order to make best use of the available bandwidth, the receiver buffers a number of TCP segments so that it can deal with out-oforder data and error corrections without sending needless requests for data to be resent. When the receiver sends an ACK message it also informs the sender how many bytes it can receive beyond the last received segment. This ensures that transmission is as fast as possible without the introduction of delays arising from overloading the receiver. At the top of the protocol stack lives the Applications Layer. This carries the protocol(s) unique to the particular application(s), and it is here that a protocol unique to the needs of lighting control would reside if required. Some common applications layer protocols, used in commerce, are identified in Table 9.26. From this brief description it is clear that, while the general principles of TCP/IP can be readily un-
325
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
derstood, there is an overwhelming amount of detail in the execution, well beyond the scope of this book. Those interested in the details of internet protocols, and their development, should be aware of the Internet Engineering Task Force (IETF), a subsidiary of the Internet Activities Board (IAB). Protocols evolve through a procedure called Request for Comments or RFC. RFC documents are publicly available and can be accessed on line at www.ietf.org. 9.5.6.3 IP addressing The idea of addressing is fundamental to all protocols, and has already been introduced at the simplest level in protocols such as DMX. Here the DMX receiver generally uses a set of switches to set a “hardware” address. In ethernet exactly the same principle is used on a larger scale – each ethernet interface device (often a plug in card, or incorporated in the main circuit board) has a unique hardware address that is stored in EEPROM. Manufacturers are allocated blocks of addresses by the IEEE, but in practice there is nothing to stop users changing ethernet addresses, so long as they ensure there are no clashes within a particular network. However, such an arrangement is only suitable for comparatively small self-contained networks. The Internet has a much bolder ambition, and this is to have an addressing system where every device using IP has a unique address, wherever it is in the world. IP addresses are in the form of a pair of numbers; the network number followed by the host number (the “host” being the device on the network). The network number is unique throughout the internet, and is administered by InterNIC, the Internet Information Center. The IP address is shown in dotted decimal form, for example: 193.76.17.22 could be the address of host number 17.22 on network number 193.76. The actual IP address is a 32-bit binary number considered as four 8-bit bytes, so this example would be expressed as:
326
11000001 1001100 0010001 0010110 In order to make more efficient use of the 32-bit number, IP addresses are divided into five classes and the first few bits are used to identify the class of service. Figure 9.44 shows the idea. In practice IP is evolving all the time, and more complex addressing schemes are being introduced to get round the problem that some classes might otherwise run out of numbers. An IP datagram, carrying TCP, contains both a source IP address and a destination IP address. But, in order to reach a particular physical destination, the IP address must be mapped to the physical address. In the case of ethernet, ARP is used to translate IP addresses to physical MAC addresses. Network administration is complex and dynamic. Part of it can be automatic, in that hosts on a network periodically “look around” to see who else is there. The ping procedure (the name derives from sonar, but, acronym fanciers could not resist also calling it the Packet InterNet Groper) involves sending a message to a host, that just echoes the packet Very Large Networks
Class A 0
7 BIT NET ID
24 BIT HOST ID
Class B 10
Medium Size Networks
14 BIT HOST ID
Class C 110
16 BIT HOST ID Small Networks
21 BIT NET ID
Class D 1110
Multicast 28 BIT MULTICAST Future use
Class E 11110
8 BIT HOST ID
27 BIT FUTURE USE
Figure 9.44 The classes of IP addresses.
CONTROL SIGNALS AND PROTOCOLS
back (if it is there to do so). The sender now knows that it is possible to send data, and, by measuring the time it takes for the packet to come back, can deduce aspects concerning the data path to the destination. Hosts continually update their own knowledge of IP and MAC addresses, and networks often have a DHCP server present. This runs the Dynamic Host Configuration Protocol that can allocate addresses on the fly. This illustrates the point that addresses can be fixed or changing. An Internet Service Provider (ISP) would have a fixed address, but someone using the internet to access them would be allocated an address for a particular transaction. Such an arrangement is known as leasing the address, which can then be re-used after the transaction is complete. The actual configuration of a network, or subnetwork, is carried in devices like routers and DHCP servers. Each element retains the knowledge of its immediate environment in terms of IP addresses and physical addresses. Some of them may introduce security procedures to prevent access to a particular network. IP addresses, being purely numerical, are not easily remembered. On the internet itself, or for email, we are used to seeing addresses like [email protected] So the system has to correlate these hierachical names to actual IP addresses (which may be changing all the time.) The naming system used on the internet is called the Domain Name System. The top level names are abbreviations like .com to indicate a commercial organization, or .org to indicate a non profit organization. At this point the question could well be asked as to whether this is relevant to lighting control. In practice an ethernet network running TCP/IP can exist on its own as a completely private network, and does not need to worry about the additional complexities of the outside world. However, even if there is only one host on the network that may also need to talk to the outside world, there will be a need to be some extra administration to make this possible. Thus an office network may be “private”, but include a separate web server that allows access to the internet and www (probably through a security
filter or firewall). In the lighting control field it is equally likely that a supervisory network might be “private”, but, nonetheless, might have an e-mail facility built in so that it can send service messages or status data to a remote site – for example information concerning running hours that alerts a maintenance company that it is time to replace lamps. 9.5.6.4 The ESTA ACN The huge expansion in networked computing has meant that the cost of network components has fallen, and that they are very well supported. So it is not surprising that the same technology has been pressed in to service to augment the performance of lighting control systems, particularly in the entertainment field. Several companies use TCP or UDP over IP on ethernet to expand the capability of their control systems, both in terms of capacity and in terms of distance. Figure 9.45 shows typical devices that provide multiple DMX outputs, but themselves receive data on fast ethernet. However, this arrangement is proprietary, except at the DMX output when it becomes “standard”. For some years ESTA have sought to solve the problems posed by the limitations of DMX by introducing a new standard based on general network
Figure 9.45 An example of a proprietary IEEE 802.3 network node providing multiple DMX outputs from Strand Lighting (left.) Artistic License offer a royalty free ethernet protocol for carrying DMX data that has been adopted by several manufacturers. They and their licensees provide network nodes like that shown on the right. This example from Artistic License provides hardware selection from multiple DMX universes.
327
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Application Other
DMP
DDL
ACN NMP
SDT
UDP
(TCP)
IP IP
Hardware e.g. IEEE 802.3
Figure 9.46 The ACN protocol stack (with acknowledgements to ESTA).
practice. The idea is to have a standard that allows manufacturers, installers and users to take advantage of the huge range of hardware and software networking products already in the field. The decision was taken to base that standard on the use of the TCP/IP protocol suite. While it would be expected that most installations would then be based on ethernet (802.3) practice, such an approach also allows the use of other carriers such as wireless, ATM and FDDI. However ACN makes the assumption that most ACN installations will only use a single medium at the hardware level. ESTA have called the proposed new standard the ACN or Advanced Control Network. ACN is intended to cover the needs of lighting control in all sizes of systems, and also to be suitable for the control of many other “show” elements. These are expected to include interfacing to audio and video systems, and control of safety critical items such as flying systems and other types of scenery control. In principle it should also be secure enough to run pyrotechnics. At the time of writing ACN is being developed, with the intention that it will first become an ESTA standard, and then ultimately be recognized by ANSI. The final “standard” will represent a massive effort in the form of computer software and docu-
328
mentation. What follows is a very brief description based on information that has been issued publicly by ESTA. Even if, in the event, ACN either fails to appear or appears in modified form, the description introduces or reinforces several concepts that are important to the understanding of complex control systems. The main characteristics are: Multiple controllers, multiple controlled devices, or in computer speak, multiple clients and multiple servers. The protocol must provide for many sources of control information, and many users of that information, to co-exist on the same network. Modern entertainment systems use thousands of devices in a single installation, so this is not a negligible requirement. Interoperability. As with DMX the aim is that all ACN compatible equipment should (where appropriate) communicate usefully. The obvious example is that a lighting console from one manufacturer should communicate with dimmers from another. Less obvious might be that a load monitoring node could ensure that system power demand was kept within a set limit by separately monitoring dimmer loads, and then imposing priorities. A subsidiary point about “useful” communication is that devices communicate to the extent of their capability. For example professional dimmers might have the capability for remote setting of control laws or minimum output. However, they might be controllable by both a sophisticated console and by a simple slider control. The fact that the simple control cannot make use of the extra facilities does not matter. It only affects the dimmer level without bothering itself about the other parameters. Multiple independent use. While only a single network is used, it should be possible to support several independent uses; either by virtue of different functions (e.g. moving scenery as opposed to lighting) or different spaces (e.g different areas within a hotel banqueting suite). The protocol is also intended to allow manufacturer-specific “extensions” whose presence will not affect the performance of the standard protocol. Ease of configuration. For most applications it is intended that ACN is, for the end user, “plug and
CONTROL SIGNALS AND PROTOCOLS
play”. This relies on the process of automatic discovery alluded to in the previous section, whereby devices learn what else on the network, and which other devices may represent a valid connection. In practice it can be expected that the larger installations will have a conventional computer node that will allow for the setting up of installation-specific priorities, but the idea is that such an arrangement will be unnecessary for the smaller installation. How does ACN achieve its objectives? Figure 9.46 shows the ACN protocol stack. Working up from the bottom the layers are: Hardware layer. Identical to that of the TCP/IP stack, giving a choice of media, but with ethernet (IEEE 802.3 – also now an international standard as ISO8802-3) currently preferred. IP layer.This combines the internetwork and transport layers of the TCP/IP stack. ACN requires the IP part to be fully standard, and to provide ARP, ICMP and IGMP (Internet Group Management Protocol). Routers in ACN networks support DVMRP. Significantly ACN is based only on UDP. This is because it is fast and because it provides a multicast facility, essential in many lighting applications. Within the closed or “private” networks envisaged for ACN, it is also reasonably secure. If reliable communication is required, then this is provided by the higher levels. ACN layer. Looked at from the TCP/IP point of view, this is part of the application, but looked at from an ACN application point of view, this is all to do with protocol mechanics and is the heart of ACN. Applications layer. Here resides the actual application, for example moving light controller, or color scroller. The ACN layer itself has many components. These are as follows: NMP (Network Management Protocol). This provides the addressing mechanism, and carries out component discovery. All components in an ACN network have a Component ID, or CID. This is a unique 128 bit identifier. This type of identifier is used in other computing applications, so the means of producing and processing them is already well understood. A component can only have one CID,
but can have several IP addresses, or can share IP addresses with other components. SDT (Session Data Transport). This is the means of transport of all kinds of data that an ACN protocol might require. The aim is to provide flexibility with efficiency. Communication is carried out in sessions, and these can be on any scale. For example the smallest session might involve brief communication between only two components; whereas a big session might involve extended multicast communication involving a thousand components. Sessions can have varying levels of reliability, from unreliable to reliable, depending on the needs of the higher layers. The session manager assists in providing efficient communication by not only matching the method of communication to the task at hand, but also by providing flow control and the appropriate error correction. DMP (Device Management Protocol). Everything described so far has been related to the provision of effective communication, but now the point is reached in the stack at which the data is quite specific to ACN. Any ACN compliant data is based on DMP aided by DDL (Device Description Language). But it can be seen from Figure 9.46 that the system does allow “other” data to be carried. The role of DMP is to allow one component to control or monitor another by setting or acquiring its properties. In addition it organizes multiple devices into a system. It cannot do this, however, without knowing the details of the components concerned. DDL describes devices in terms of those properties (e.g lighting level, color, temperature) whose values represent the state of the device. It goes on to specify how property values can be changed through DMP. ACN uses, wherever possible, established software techniques, since this makes it easier to ensure compliant systems, and simplifies product development. DDL itself is based on the widely used XML (eXtensible Markup Language) but is a very much reduced subset of it. Figure 9.47 gives an example of how DDL models a system. The example is a dimmer cabinet. The
329
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Root Device Dimmer Rack
Group “Supervisory information”
Device variable User name
Device variable Rack temperature Device variable Run-time hours
Child device Dimmer Group “Dimmer”
Device variable Dimmer Level Qualifier Minimum level
Qualifier Minimum setting Qualifier Maximum setting (e.g. 10%)
Figure 9.47 Example of how the DDL of ACN models a sub-system. Only one of many dimmers shown.
main control aspect, or device variable of a dimmer is its level. However, this variable may itself be subject to one or more qualifiers, and the qualifiers themselves may be qualified. The example shows an entertainment dimmer with a minimum output level as a qualifier (this arrangement is called pre-heat and achieves a faster response from high wattage tungsten lamps). The qualifier itself is also qualified by minimum and maximum permitted levels. Another example of a qualifier at this level might be the facility within a dimmer to change level at particular speed (i.e. providing the function of an automatic dimmer). Such a provision might be useful in an architectural application where sudden changes of light output were undesirable. Figure 9.47 identifies the dimmers as forming part of a group, in this case the “Dimmer” group.
330
Sub-groups are also allowed and are used when variables are to some extent connected, for example x, y, z axes of rotation for a moving light, or Red, Green and Blue in a primary color mixing system. Figure 9.47 also identifies examples of variables that might live in a “Supervisory Information Group”, in this case details concerning a rack of dimmers, where the variables could include, but not be limited to: • rack temperature. • elapsed running hours (may be of significance in force ventilated racks where a filter change may be needed at intervals). • user information (rack name etc.). Sometimes qualifiers are cross referenced. For example the dimmer rack as a root device might itself be qualified in respect of dimmer performance. In such a case the child devices within the cabinet (the dimmers) might simply be cross referenced to the root qualification, as opposed to having an individual qualifier for each dimmer channel. A qualifier like the dimmer speed example cited above could fall into this category. Some commentators have said that ACN will revolutionize lighting control. There is no doubt that it has the potential to bring significant benefits to large scale “show” systems, but its adoption at the small system end of the market will depend on a number of factors, particularly: • that it does not add any significant cost. • that it does not introduce additional complexity for the non professional user. • that any benefits are clearly distinguishable. But, in the same way that the gateway router approach is used to link, for example, EIB to DALI, it can be expected that ACN gateway products will become available once the protocol is widely adopted. ™ 9.5.6.5 The ASHRAE BACnet™ An example of a mature network protocol is BACnet, a protocol for Building Automation and Control networks. This evolved in the 1980s under the aegis of ASHRAE, the American Society of Heating, Refrigerating and Air-Conditioning Engineers and became
CONTROL SIGNALS AND PROTOCOLS
BACnet application layer
BACnet network layer
Data link layer IEEE 802.2
IEEE 802.3
Arcnet
MS/TP
PTP
EIA485
EIA232
Figure 9.48 The four layer protocol stack of BACnet™
an ANSI standard in 1995. It was originally introduced to provide interoperability within the realm of heating, ventilating, air-conditioning and refrigeration equipment; however, it was always intended that it should also provide a basis for integrating other building services such as security, fire detection and lighting. By 2001 an addendum had been prepared to facilitate the use of the protocol for fire, life safety and security systems, and a working group had been set up to propose how lighting control would also be covered by the protocol. Previous sections of this chapter have introduced most of the concepts that underpin various communication protocols, so only a brief summary of BACnet’s features is given here. BACnet uses a four layer protocol architecture as shown in Figure 9.48, It corresponds to the Application, Network, Data Link and Physical layers of the OSI model. However, at the data link and physical layers a choice is given. This is for two reasons: • to make maximum use of other well established communication methods. • to ensure that, at the level of simple devices such as sensors and actuators, the protocol needs only a minimum overhead.
It can be seen from Figure 9.48 that ethernet (IEEE 802.3) is one main option, alongside an alternative standard ARCNET. In practice it is now the case that most BACnet communication is done using ethernet 802.3. For connection to simple devices there is a choice of MS/TP (Master-Slave/Token Passing) or of PTP (Point-to-Point) operation. This facilitates the use of the common EIA485 and EIA232 electrical interfaces. In common with EIB, BACnet prescribes how various control devices and physical units are represented, and how the control rules between devices are to be defined. There is provision for manufacturer proprietary communication (like the “other” communication in ACN) to ride on the same network, although here the idea is that such information will follow the same rules as those for the standardized devices. All control devices on BACnet are modelled as a collection of objects, which in turn have properties. The properties may be fixed and unchangeable, such as a name, or readable, such as the varying
Value
Meaning
NO_FAULT_DETECTE D
Present value is reliable; i.e. none of the faults below has been detected. No sensor is connected to the input object. No device is connected to the output object. Connection between defined object and the physical device is giving a value that indicates a short circuit. Connection between defined object and physical device is giving a value that indicates an open circuit. Sensor connected to the input is reading a value higher than the normal operating range. Sensor connected to the input is reading a value lower than the normal operating range. Controller has determined present value is unreliable; but this is for a reason other than those above.
NO_SENSOR NO_OUTPUT SHORTED_LOOP
OPEN_LOOP
OVER_RANGE UNDER_RANGE UNRELIABLE_OTHER
Table 9.27 Values taken by the “Reliability” property in BACnet.
331
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
output of a sensor, or writeable, such as the setting of a preset value of voltage or temperature. Because of its ASHRAE origin, BACnet is strong on scheduling and monitoring. Many object types have a reliability property geared to indicating practical faults in a system. Table 9.27 shows the idea. BACnet offers encryption as an option to provide added security. This is normally only used in widely distributed networks that themselves may include dial-up connections over a public network. Speed of response is not a significant factor when controlling an air conditioning system; however, it can be significant in lighting control. Some early attempts to use BACnet (or similar) for lighting control have foundered for this reason. While BACnet is fine for overall scheduling of, for example, public area lighting, its practical realization may be far too slow if it is also intended as the means of control for end users. The author knows of one case where the reponse time from a manual lighting control on this type of network was 15 seconds. As already observed a response time for this application should be no more than 100ms. From this it is clear that: • if BACnet is to provide full lighting control facilities, any actual execution has to take into account reponse times. • standard BACnet is quite suitable for scheduling, monitoring and load management services living at a level above the local operation of lighting. Thus gateways to DALI, LON, CEBUS or EIB are quite feasible.
9.6 Computers in lighting control 9.6.1 Introduction By now it will be clear that computers in a general sense, including microprocessors, are the basis of many forms of lighting control. This short section is concerned with the idea of using everyday computers (personal computers) as a basis of lighting control. A simple statement is “don’t do it”. Why use a computer to switch a light on, or set a lighting level,
332
when a switch, push button or slider control represents a much easier and more intuitive method of doing so? We can make a number of observations: Simple lighting control should stick to the simplest method of achieving the required result, however, the application of microprocessors to basic lighting control tasks does, if correctly applied, help achieve this simplicity. Nonetheless, personal computers can make the engineer’s and lighting designer’s task easier. It mainly does so by virtue of providing monitoring, design and programming tools – not by being an on-line lighting controller. The role of computers in lighting control can. therefore, be summed up as follows: • as the means by which architectural lighting control schemes can be programmed. The method has the advantage that the computer can, at the same time, provide a database resource for the system. This is particularly important with large systems involving hundred or thousands of circuits. • as the means by which large scale systems are scheduled and monitored. Small lighting control systems working automatically can use simple microprocessor based timers for scheduling; but large scale systems benefit from the flexibility of computers – especially when load monitoring is included. 9.6.2 Operating systems From the above it can be inferred that it is preferable that a lighting control system works on a stand-alone basis and, even better, is based on a distributed architecture where the failure of a single device, for example a central computer, cannot completely disable the system. Where computers are used in a system, it is essential that they are reliable and are not prone to “crashing”. The reliability of computers in this respect is mostly down to the quality of the software, and this includes the computer’s operating system. Current experience is that, for critical applications, the standard Windows™ operating system as used on home personal computers should be avoided, as it will almost certainly crash every few months. It is hoped that this situation will be corrected,
CONTROL SIGNALS AND PROTOCOLS
but at the time of writing the preference for reliable performance is to base the choice of operating system on the use of Windows NT (and its successors), Unix or Linux. 9.6.3 HTML Having said earlier that the idea of using a computer to control one’s desk lights is extravagant, there are occasions when it can make sense. An example could be an open plan office where everyone within it is equipped with a networked personal computer. By virtue of installing a bus based lighting control system, it is theoretically possible for each person to adjust the lighting level in their immediate area – but to do so they need either a wired or wireless control. Such controls become unnecessary if the networked computers can become the local controllers. Here each user would select a page that gave them access only to the control of their own lights; the facility should be as easy to select as looking up a number in the staff telephone directory. The practical and intuitive method of doing this is to provide a control page similar to a web browsing page. This is best done using HyperText Markup Language or HTML. An HTML “document” is a standardized representation of a document with embedded hypertext links. These links or tags describe the basic elements of a document (for example head-
Figure 9.49 An example of an HTML “document” used for office lighting control.
Figure 9.50 This port server provides a simple means of connecting an ethernet LAN to 16 EIA-485 or EIA-232 controlled devices.
ers, paragraphs, text styles) but can also include more sophisticated items such as interactive elements. Most pages seen on the internet use HTML because it has the great advantage that it is independent of the operating system or type of computer being used. Thus the computer on the office desk does not need to know how the lighting control system actually works; it can simply send an HTML description of what it would like to happen. HTTP (see Table 9.26) provides a standard method of transmitting HTML documents. It can be expected that many lighting control systems used for high level consumer and architectural applications will offer an HTML/HTTP option for control. 9.6.4 Port servers A useful item for interfacing computers to real world physical applications is the port server, a specialized router. Figure 9.50 shows an example – in this case a device that lives on an ethernet (802.3) LAN, but which provides a means by which the LAN has access to 16 serial interface ports; in the example they can be configured either EIA232 or EIA485. Port servers are useful for simplifying the realization of systems using mixed equipment, and for systems that are physically widely spread. For example a BACnet user might well standardize on ethernet as the main method of communication and use remote port servers to communicate with local devices (as opposed to using a central computer device having multiple serial ports).
333
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
9.7 Cordless control 9.7.1 Introduction There are many cases where cordless or wireless remote control is required for lighting. The requirements vary from the very simple to the complex: • The very simplest control may only operate a single channel of ON/OFF command. • More usually a control would be multi-channel, for example selecting a number of lighting “scenes” or dimmer levels. • A complex wireless controller might control a large number of channels, or, more likely, might be a wireless node on a data network (e.g. ethernet). There are several different ways of achieving cordless control. For simple on/off operation the most common method is through the use of proximity detectors, and these are described in Chapter 14. Another method, that has not yet been seriously exploited, is the use of voice commands. Speech recognition chips have been developed for the automobile and machine control market, and these hold open the possibility of comparatively low cost speech recognition devices being used for lighting control. However, the most common method is the use of electromagnetic radiation, either in the infra-red part of the spectrum, or at designated frequencies within the RF spectrum. The fact that the same type of controls have application in everything from operating video recorders to automobile locking systems means that the technology is well developed, and that it is possible to realize cordless lighting controls at a sensible cost. 9.7.2 Infra-red control The great majority of simple cordless controls are infra-red controls working on a simplex basis. The ones used for lighting control are based on the same technology as those used for TV and VCR control, although some are designed to work at longer ranges than the consumer TV product. Infra-red has the advantages of low cost, no need for licensing, and being relatively secure in that it
334
Figure 9.51 The “Hello” chip from Philips uses DSP technology and can recognize up to 100 spoken words. The photo shows a development kit suitable for evaluating this technology for lighting control.
only operates in a single room space. It is blocked by any walls or building structure. It can suffer from interference from any device emitting infra-red, including fluorescent lamps, but the worst effects of this are eliminated by using a modulated signal. (See Section 7.2.11.) A typical infra-red controller is based either on a dedicated integrated circuit, or on a low cost microprocessor. The principle is shown in Figure 9.52. The device is normally “asleep”. Pressing any one of the buttons causes it to wake up, and starts the master oscillator. The sensing of which button has been pressed is done on a simple matrix basis. The “drive” lines are sequentially driven low, and by sequentially reading the “sense” lines as each drive line is driven, it is possible to detect the intersection corresponding to the button pressed. According to the button that has been pressed, a code, stored in ROM or EPROM, is generated that appears at the output; simultaneous pressing of two buttons is ignored. The code consists of bursts of the IR modulation frequency. One of the most common schemes used is that of Philips, known as the RC5 protocol. (NB this is a proprietary protocol for which Philips grant a license when their chip sets are used.) This has the following characteristics: • the infra-red LEDs used have a peak radiation around 850nm.
CONTROL SIGNALS AND PROTOCOLS
Drive lines black Sense lines red
Button at each intersection
Keyboard scanning
Timing and control
Master oscillator
ROM
36kHz
Divider
Pulse generator
• the timing of the output signal is done is such a way that the demodulated data is in Manchester biphase code. • the RC5 protocol allows for 2,048 different commands to be transmitted. An IR hand control has only as many buttons as needed for the task; typically less than 10 for lighting control; less than 25 for normal TV and less than 50 for a complex AV controller. The original RC5 transmitted signal is shown in Figure 9.53. There are six command bits (allowing 64 different commands) and five address bits (allowing 32 different groups). These are preceded by start and control bits used by the receiver to synchronize its bit timing. The idea of the different groups is that each group can be applied to a different application, for example TV, DVD, Teletext, VCR, Air Conditioning, Lighting etc. Philips provide a protocol list to licensees, but in practice the RC5 or RC5-like signal is used in many different ways. The selection of address bits may be done: • by hard wiring or programming in the control. Scan time
TIMING FIRST CODE Debounce time
Output driver - may use separate transistor
16
Data transmit 2
14
Total 32 Bit times
IR output LED(s) TIMING SUCCESSIVE CODES
Figure 9.52 Block diagram of an infra-red remote control.
• the transmitted infra-red radiation is modulated at 36kHz. This frequency is derived by dividing down from the master oscillator. A typical oscillator frequency is 4MHz, requiring division by 111. • the modulation scheme involves only a 25% or similar duty cycle; i.e. the 36kHz is not transmitted continuously but in pulses. This allows high peak radiation power (to help good reception) with a low continuous power requirement (to conserve battery life). • the required data in turn modulates the 36kHz signal, with a bit time being equal to 64 pulses.
1st Send
2nd Send
64 Bit times
DATA FORMAT
Start bits
Address bits
Command bits
Control bits Logic 1
Logic 0
Bi phase bit format
Bit time approx 1.8ms
Figure 9.53 The original RC5 IR signal.
335
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Parameter
Value
IR wavelength Sub carrier frequency Start bit length Low bit length High bit length Bit spacing Command spacing Command repetition
910nm 40kHz 108Hz (2700Ps) 18Hz (450Ps) 36Hz (900Ps) 450Ps 900Ps Each command sent 3 times
Table 9.28 The IR signal used by the AddressPro™ digital lighting control protocol.
bus. In protocol-speak, the sensor acts as gateway or protocol converter. The USA company Energy Savings Inc offer AddressPro™ as a digital protocol for commercial Figure 9.54 Examples of IR cordless lighting controls. The lighting control (while it is a proprietary protocol, it one on the right is from Relco. is offered royalty free to other manufacturers provided the trademark is identified.) Products include • by the use of switch selection on the control (usu- fluorescent lighting ballasts, incandescent lamp ally “hidden” DIP switches). dimmers and a range of control panels. Main com• by making the control unit more sophisticated, munication is by a 2-wire bus based on EIA232. requiring sequential operation of buttons, where the Cordless control is provided by IR where the IR data user is able to select a device group. that is transmitted is in exactly the same byte format IR controls of this kind simply transmit the data; as its EIA232 equivalent. The format of the IR transthere is no feedback to indicate successful reception mission is shown in Table 9.28; notice that bits are other than that of the user observing that the expected encoded using PWM. result has occurred. Most controls continue to send AddressPro is primarily a simplex method of a command as long as the button is pressed and any- control at the individual ballast level. When an inway send it twice. stallation is initially configured and programmed, The comparatively simple bit structure of RC5 there is a need to know device identities. Every deand similar codes based on modulated IR allows the vice in an AddressPro system has a unique 27 bit production of learning IR controls. These can copy identity number (a similar concept to that used by a bit pattern from a given control, so, for example, a LONworks™ and RDM). The recovery of these ID home user can transfer the commands from several numbers is achieved by using infra-red as a “back different controllers into one single controller. channel”. This can be done by using the controlled lamp 9.7.3 IR control in bus systems itself as an IR data source. In the case of fluorescent lamps the AddressPro ballast can be instructed to The random nature of cordless control use, and the transmit its ID, and it does this by modulating the possibility of interference, means that if IR controls light output. The modulation is PWM where the are to be used in a bus-based system, some care must space between bits (corresponding to the lamp bebe taken to ensure correct system operation. ing off) is only 150μs and the start, high and low bit A typical strategy is that used in DALI and simi- lengths are respectively 3, 2 and 1ms. lar systems, whereby intelligent sensors are used that Such a method could not work for incandescent buffer the simple IR data, and once validated, re- lamps operating at 60Hz, so here a different method format it to the bus protocol and transmit it onto the is used. The AddressPro dimmer is set to 50% out-
336
CONTROL SIGNALS AND PROTOCOLS
Fluorescent ballast light output when signalling ID 100% Light output
0% Start 3ms
Logic 0 1ms
Logic 1 2ms
1
0 150Ps “off” condition indicates bit boundary
Transistor incandescent dimmer output when signalling ID
Dimmer set at 50% output
60Hz Start
Logic 0
Logic 1
Missing half cycle indicates bit boundary
Figure 9.55 A hand control used for the AddressPro system. It transmits IR commands, but is also a receiver for ID data from individual luminaires.
put, and data bits are separated by omitting half cycles. In this case a single missing half cycle represents the bit boundary, and start, high and low bits are represented by respectively 8, 4 and 2 complete half cycles. Figure 9.55 shows the idea. The procedure with this method is that all luminaires in a system are instructed to transmit their ID (using the wired EIA232 link). The operator then holds a hand control like that of Figure 9.56 next to an individual lamp; by pressing the “GET” button the hand control acquires the ID of the luminaire concerned. The same hand control can then be used to transmit an IR command that assigns the luminaire to the required control grouping.
layer of the protocol stack is subject to variation). • an SIR (Serial Infra Red) variant for serial communication up to 115kbps. • an FIR (Fast Infra Red) variant for communication up to 4Mbps. Higher rates are also envisaged.
9.7.4 IrDA Infra-red is inherently capable of carrying data at significantly higher speeds than those required by simple remote controls. It is now widely used as a means of cordless connection between computers and peripherals; for example printers and digital cameras. The protocols for doing this are developed by the Infrared Data Association or IrDA. Currently the characteristics of IrDA systems include: • IR wavelength usually 910nm (but the physical
Figure 9.56 A hand control from Energy Savings Inc. used for the AddressPro system. It transmits IR commands, but is also a receiver for ID data from individual luminaires.
337
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
• a range of 1m, but ranges up to 10m are proving practical. The IrDA protocol stack is processor intensive, and while originally designed for multipoint communications is, in practice used only for point-topoint. The use of IrDA protocols in lighting control is limited because of the limited range. There is, however, a useful result of its availability. A hand held PDA or computer equipped with IrDA hardware can also be programmed to learn and operate other IR systems including, for example, RC5. This can be useful for system programmers and product developers. The fact that RC5 is normally optimized for 850nm, and IrDA for 910nm is not in practice a problem since most receivers are broadband. 9.7.5 RF control RF (radio frequency) control offers more possibilities, but also more complexity than IR. One of the main differences compared to IR is that it can penetrate partitions, walls etc.; so while it then lacks some security, it can add convenience. Some of the many points to be considered are: Permitted frequencies. Only a few frequencies are allocated internationally for the “telecommand” application. Currently the most useful ones are 433.92MHz (Europe and USA) and 418MHz (UK and USA) because, provided defined limits are not exceeded, these can be used license free. Other frequencies used (which may or may not need a license)
Figure 9.57 For project work and short run manufacture it is easiest to use ready made up RF modules. This 418/ 433.92 MHz transmitter and receiver pair from Radiometrix can implement a data link up to 160kb/s at distances up to 75m indoors or 300m open ground.
338
are 173.225MHz (UK) 173.250MHz (Mainland Europe) 315MHz (USA, Canada) 868MHz (Europe) and 916MHz (USA). Effective radiated power (ERP). This is strictly limited. In Europe EN300 220-3 sets the limits, for example a maximum of 10mW at 433.92MHz. In the USA the FCC Regulations Part 15 sets the limits, and it does so in terms of the electric field strength at 3m from the transmitter. This is set at around 2.8mV/m which means that ERP at this frequency is limited to around 2.5μW. Modulation method. Both AM and FM are used. AM gives the simpler receiver construction, FM gives better range. The data coding rides above the modulation and can be based on various methods including bi-phase (Manchester) coding, DTMF (Dual Tone Multiple Frequency, as used by telephones) and V21 Modem tones. Non Return to Zero (NRZ) signals like standard EIA232 signals are not suitable. Security. Many applications of RF remote control do require security. A particular everyday example is the RF controlled automobile lock. These applications use a code hopping technique. Here receivers “learn” to operate only with designated transmitters. In addition, every transmission subjects the data to an algorithm that means that each time a transmission is made for a particular command, the code changes. The receiver that has “learnt” from its transmitter(s) keeps in synchronization with the code change. It is quite practicable to use codehopping RF devices for lighting control, and their general availability makes them suitable for special project work. However, there are comparatively few applications where the feature is essential. Range. Range is determined by ERP, modulation method, receiver sensitivity and environment. When transmitter and receiver are in line-of-sight it is possible to achieve significant ranges even with low ERP. For example 10mW ERP can easily achieve 1000m at 173MHz, 30m at 433MHz (AM) or 200m at 433MHz (FM). In practice furniture, partitions and people can severely attenuate the signal. Manufacturers can give a “weighting” to all the different attenuating factors to allow a user to calculate the achievable range, but in practice there is
CONTROL SIGNALS AND PROTOCOLS
although there can be the special problem that the receiver may be built in to microprocessor based equipment that is, itself, a source of very localized interference. The second cause is interference arising from other users of the same frequencies. 400MHz telecommand equipment is not supposed to transBit period ranges from 100Ps to 400Ps mit for more than a 10% duty cycle, so provided …1 1 1 1 1 1 null transmissions are random timed, and assuming only occasional operation in lighting control applications, collisions should be rare and easily dealt with by rePreamble to activate and Encrypted data 32 bit transmission. The third is the most troublesome, and settle receiver, and code that changes with synchronize bit time every transmission this is caused by strong transmissions in neighboring parts of the RF spectrum. Fixed data 34 bit code that includes push button status Figure 9.60 shows that the frequencies used in and transmitter serial number the UK are sandwiched between those used for other applications requiring higher power. The low cost 2 status bits receivers used in telecommand can have poor selecOne indicating repeat transmission One indicating battery status tivity making them prone to suffer interference. Problems can arise in the office environment, for example, if the building concerned becomes the site 28 bit serial number for a TETRA base station. There is then a possibility that receivers with moderate selectivity could be 4 bits button saturated by the higher power RF from the base stastatus tion. Local conditions can also be a factor. For examFigure 9.58 The four channel Keeloq™ encoder from Mi- ple in the USA 418MHz is the normal frequency for crochip generates a 66 bit PWM signal that includes a hopRF controls used for lighting. However, in Manhatping code for security. tan, New York, this frequency was licensed for another application and for this reason manufacturers no substitute for simply trying it out. Logic 0
Logic 1
It might seem that all such equipment should be designed to have the maximum possible range consistent with the ERP limitation. However, this may not be the best strategy in places where multiple transmitters are being used over an area. Greater transmitter power and receiver sensitivity can lead to more data collisions and problems due to multipath reception, where a given signal reaches the receiver both by a direct route and by multiple reflection from walls etc. Interference. This is a serious practical problem. It arises from three causes. The most obvious is that arising from general RF disturbance arising from lightning and improperly suppressed equipment. In principle well designed receivers working in the 400MHz region should not be affected by this,
Figure 9.59 In the Lutron RA wireless lighting control system operation is specified as covering a 9.1m (30ft) radius. Where greater range is required, repeaters must be used. Diagram from Lutron.
339
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Selectivity of good 433.92 Mhz Receiver
Power
418MHz
Selectivity of poor 433.92 Mhz Receiver 433.92 Mhz
410-415 415-420 420-425 425-430 Mhz Mhz Mhz Mhz
430-440 Mhz
Current Future Current Future TETRA TETRA TETRA TETRA
Various uses, including radio amateurs
Frequency + existing government and military use over t h e s e frequencies
Figure 9.60 The problem of adjacent spectrum useage in creating interference. The diagram shows the situation in the UK. TETRA is a mobile radio service.
such as Lutron recommend that 434MHz is used in Manhattan. Antennae. Any RF system needs a transmitting antenna and a receiving antenna. In the lighting control application the requirement is usually for very compact devices, and external antennae may not be acceptable. Three main types are used. The external whip antenna is the most effective, but is unacceptable for many products from an appearance or space point of view. More acceptable is the helical antenna (essentially a simple coil of precise specification) that can be mounted on a printed circuit board (PCB). Simplest of all, but least effective, is a loop antenna that is printed directly on to the PCB, but its positioning is critical in relation to other components. If transmitter or receiver are mounted within conducting (metal) enclosures, then only external antennae can be used. Battery life. Many RF installations are similar to their IR counterparts, so the same considerations of battery life apply. The aim is to ensure the transmitter is “asleep” most of the time, so only draws power when transmitting. For intermittent use, battery shelf life can be more important than Ah capacity. Conventional alkaline cells have a shelf life of around 5 years, lithium cells around 10 years. In some cases RF is used on fixed control panels just to simplify system installation and avoid the need to install control cables. It is then sometimes possible
340
to provide transmitter control panels with conventional mains power, but where it is not, long battery life and easy battery replacement are essential. Duplex and half duplex operation. For many applications of RF simplex operation is quite adequate. Where it is being used in a manner similar to that of IR controls any “feedback” is by observation; so the user simply transmits again if a button press does not yield the required result. But in the more sophisticated applications greater relibility of communication is required, and this can only be assured by bi-directional communication. An example is in a multi-scene lighting control where a “scene” may involve the adjustment of lights that the user cannot see – as when setting an “early evening” scene in a house. The scene might involve both interior and exterior lights. The ideal is that each control panel is fitted with a transceiver, so every time a transmitter sends a message the receiving system sends it back again and the original transmitter can check that the right instruction has been received. Clearly such a system is impractical if the remote control panel has a receiver that must be powered up all the time since in any battery powered installation the battery would go flat very quickly. The solution is to arrange that the receiver is only open for business immediately after a transmission
Figure 9.61 It is now quite feasible to program or modify large lighting systems using wireless LAN. In this example a laptop is using 802.11b to communicate to a Helvar Lighting Router, which in turn controls a large system.
CONTROL SIGNALS AND PROTOCOLS
has been made. If a correct acknowledgement is received, the control panel immediately powers down; if no, or an incorrect, acknowledgement is received, it transmits again. Protocol gateway. It would be technically possible to use bi-directional RF to carry some of the common protocols directly. For example the low speed and bi-phase nature of DALI makes it suitable for direct transmission by low power RF transmitters. However, the actual operation and timing of DALI and similar protocols means that in practice it is better to use a gateway device to buffer the signals, in which case it is a matter of choice as to whether the data carried by the RF has the same format as the underlying protocol or not. 9.7.6 RF Wireless networking In the wired world a difference was identified between control protocols and network protocols – one required to carry limited control data at relatively low speed, the other required for the high speed reliable exchange of large quantities of computer data. The same distinction exists in the wireless world. IR data can be used for high performance or network use, but the transmission distance is very short and requires transmitters and receivers to be directional. In principle RF, while being more complex and expensive, gets round this restriction and allows considerable flexibility in the location of devices (nodes) on a network. This can be useful in complex systems, such as large scale entertainment systems using proprietary protocols or ACN over ethernet. An example could be a multi-use space such as a banqueting suite where it would be convenient to have a control console that could be used anywhere in the space without the need for a physical connection. Although it is possible to set up ad hoc wireless LANs that are ethernet-like, the preferred arrangement is to use a base station receiver or Access Point (AP) which itself operates to the IEEE 802.11 wireless networking standard. Wireless nodes communicate with the AP, which is in turn physically connected to an IEEE 802.3 network and thence to other wired nodes. Table 9.29 shows some of the
principal charactersistics of 802.11. Note that CSMA/CA is used in place of CSMA/CD because the nodes operate half duplex; i.e. they cannot simultaneously transmit and receive. Nodes listen to the network all the time, and only transmit when it is clear. However, when they are transmitting, they cannot detect a collision. Acknowledgement (ACK) messages have priority, so when a receiving node sends an ACK, it does so immediately it detects that transmission has stopped. Any other node wanting to transmit has to wait for a random back-off time to allow ACK to work. The DSSS technique is used to achieve the “spread spectrum” and to eliminate the need for high power at a single frequency. The required data is combined on an EX-OR basis with a much faster pseudo random number (PN) sequence; the PN code is then filtered out at the receiver to leave the original data. The result is that the transmitter operates at a continuous low power (although the average power is the same as it would have been with the uncoded data). The performance of such systems is highly dependent on the environment, both in terms of physical obstructions and the presence of other electromagnetic radiation. Obviously it also depends on the number of wireless nodes sharing a base station. Under ideal conditions “wireless ethernet” gives a performance similar to 10Mb/s wired ethernet, with a range of around 30m, but it is recommended that no performance figure is assumed, and that systems are fully tested in their actual environment before being put into service. IEEE802.11 offers two methods of providing security against unauthorized interception. Authentication at its simplest is a means of defining which nodes have authority to communicate with each other in a given coverage area. This is probably quite sufficient for applications like lighting control, where it could be used to ensure that a node did not accidentally take command of some critical lighting circuits – plunging diners into darkness during the soup course was never a good idea. At a higher level the authentication method can be extended to use shared key authentication, whereby the authorization is encrypted, and can only be decoded by nodes
341
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Parameter
Value or comment
Bit rate 802.11 (original standard) Bit rate 802.11b (high rate standard) RF spectrum used
2 Mbps peak, fall back to 1Mbps in noisy environments. 11 Mbps peak, fall back rates of 5.5 2 and 1 Mbps 2.400000102.4835 GHz (USA and most of Europe) This is part of the 2.4GHz Industrial Scientific Medical (ISM) band which has worldwide allocation for unlicensed operation. Approx 20MHz; allows 3 nonoverlapping channels in ISM band. DPSK (Differential Phase Shift Keying). CSMA/CA (Carrier Sense Multiple Access/ Collision Avoidance). DSSS (Direct Sequence Spread Spectrum) Complementary Code Keying (CCK). FHSS (Frequency Hopping Spread Spectrum) also permitted by the original standard at 2Mbps. Not used in 802.11b.
Channel bandwidth Carrier modulation Medium Access Digital modulation method
Table 9.29 Features of IEEE 802.11, the wireless LAN protocol.
having the required decryption key. The next step up is to use full encryption where the data itself is encoded. The aim here is to achieve at least the same level of privacy as can exist on a wired LAN. The RC4 PRNG algorithm from RSA Data Security is used. Compatibility of different wireless ethernet products is monitored by the Wireless Ethernet Compatibility Alliance or WECA (www.wirelessethernet.org). At the time of writing Wireless Ethernet seems likely to be of most interest to lighting control users because of its relevance to ACN, BACnet and similar protocols. However, the wireless protocol that is receiving a lot of press and promotion is Bluetooth (standards information available on www.bluetooth.com). Bluetooth is intended for the wireless connection of computers, mobile phones etc. in an ad hoc way. Communication is usually point-to-point, but can also be point-to-multipoint using piconets. A piconet consists of a master device and up to seven slave devices. It is also
342
possible to link piconets together. Bluetooth is optimized for the real time exchange of voice messages and for the exchange of data within room and small office-sized areas. Bluetooth uses the 2.4GHz ISM band but does so in a quite different way to IEEE 802.11. It uses frequency hopping, whereby each successive data packet is sent on a different channel. Within the ISM band there are, allowing for guard bands, 79 channels each of 1MHz bandwidth. Modulation is by Gaussian Frequency Shift Keying (GFSK) where a binary one is represented by a positive frequency deviation and a binary zero by a negative frequency deviation. Data is transmitted at a symbol rate of 1Ms/s. Bluetooth can be used as a component within a lighting control system, but is most likely to be used when general purpose devices such as mobile phones or laptop computers provide access to a lighting control system as a subsidiary, rather than as a prime, function. There is some concern as to whether Bluetooth and wireless ethernet can coexist, since they use the same ISM band. Fortunately their methods of working are so different that there are few problems in practice. For example IEEE 802.11 only uses the equivalent of 17 channels within the band, so even if it was going flat out, it would still leave space for Bluetooth. Considerable work has been done on the subject, and the worst case seems to be a 20% degradation in service due to mutual interference when both systems are installed in the same area and where both are heavily used. Another band that is available for RF networking is the 5GHz band. This is expected to provide a data bandwidth of 54Mb/s making it suitable for applications such as streaming video. The relevance of this development is that one version of it (Hiswan) is being considered for consumer wireless AV networks in the home, leading to the possibility of interfacing with lighting. There are three contenders. Hiswan from Japan, HiperLAN2 from Europe and IEE 802.11a from the USA. The present success of 802.11b would seem to give 802.11a the edge, but nothing is certain until a market is established.
C h a p t e r 10
WHY LIGHTING CONTROL?
Why
lighting
10.1 The practical role This chapter is an introduction to the rest of the book, and is a “bridge” chapter within this otherwise technology oriented Part 4. When electric light became generally available people were glad enough to have light, without fussing too much about appearances. But it was soon realized that the quality of electric light could be greatly improved by diffusing the light, eliminating glare, and making attractive luminaires. It is still the case that the quality of lighting is mainly determined by the choice and disposition of the luminaires and lamps, and by the nature of the surfaces being illuminated. However, there are many situations where artificial light is used in variable circumstances – variable in that a number of parameters may change, for example: • the contribution of daylight. • the activity in the illuminated area may change. • the activity may itself require variable lighting conditions.
control?
In this book we are primarily concerned with the control of electric lighting using electrical and electronic power control, and within this field it is recognized that applications can be divided into the practical, the esthetic, and the energy saving. Of course they all overlap, but the prime use tends to determine the best approach for any particular application. For example, if you go to the movies it is a practical matter that the auditorium lights are on when you enter the theater so you can see to find your seat, but are off to enable you to see the film properly. The question of whether the cinema owner then uses a contactor to instantaneously douse the lights, or uses a dimmer to fade them down, could be regarded as esthetic – but actually is still practical. Sudden changes in level could cause an accident. Switching the lamps may also shorten lamp life, so there could be an economic argument as well. What we find is that there are a number of activities which take place in, for example, lecture theaters, training rooms, cinemas etc. where, for the
Figure 10.1 Examples of practical, or functional, lighting control. “House” lighting within a cinema auditorium needs automatic dimming (photo of Warner Village Cinemas Auditorium 3 Cheshire Oaks, left). X-Ray and photographic viewing tables may need manual lighting level adjustment as in this example from CPAC (right). The dedicated mammography X-Ray illuminator in the “Perfectview” range provides three levels of illumination, 5,000, 9,000 and 13,000 Cd/m2 that are switch selectable. Motorized masking blanks out the part of the illuminated surface that is not covered by the X-Rays.
343
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
venue to function at all there has to be some form of lighting control and that usually a continuously variable control is most appropriate. There are other activities, such as work, shopping, museum visiting etc. where daylight may make a significant contribution to the lighting for part of the day, but where lighting must be controlled to augment or replace daylight. The question of whether this is done in switched steps, or on a continuously variable basis, may well be determined by esthetic and cost of ownership issues. There are some processes, including some in manufacture, agriculture and horticulture, which require lighting control for practical reasons. There are also applications which are primarily esthetic, but which require additional practical control – for example simply to give enough light for the cleaners to see by. In all these practical applications of lighting control the system design must be as simple as possible. Now we have a huge range of technology to help us, our task is actually to choose the simplest package – to ensure that the practical system really is practical and does not fail to perform because it is too complex for the user.
10.2 The esthetic role It is the esthetic role of lighting control which has developed the lighting control business. Lots of lip service may be paid to energy management, but in fact it is the pleasing effect of changing lights which sells. The essence of this role is that it allows two things: • it allows the balancing of different light sources to ensure that the overall result is pleasing. • it is the mechanism for transition between one lighting state and another. Most modern lighting design depends on the use of a variety of sources, with the variety residing in both the lamp type and the luminaire optical characteristics. Very often the lighting designer needs to set individual lighting channels to different levels to achieve a required lighting balance or to achieve a lighting “scene”. When a particular site needs sev-
344
Figure 10.2 Lighting control for esthetic effect was first used in theatrical and entertainment applications. Here dynamic lighting and synchronized sound bring an exhibit to life at the National Museum of the Civil War Soldier, Petersburg VA. The “Trial by Fire” mixed media presentation was designed and produced by BRC Imagination Arts.
eral scenes, for example to meet the needs of different times of day, or as part of a dynamic lighting scheme, there must be an acceptable method of changing from one scene to another. Chapter 11 develops these points as they apply to theater lighting, and Chapter 12 does the same for architectural lighting. In his book Lighting Modern Buildings Derek Phillips gets to the heart of what good lighting can do. He points out that merely taking the technical tools of lighting – from wall washers to spotlights, downlights to diffusers, does not tell us anything about, or guarantee, a particular look and feel to a space. It is best to start from the emotions – is the space comfortable? dramatic? bland? gloomy? If the aim is a space that is “comfortable”, in the sense of relaxing, pleasant, well modelled and so on, the approach is likely to be different than for a “dramatic” space, which is more likely to evoke adjectives like bright, sparkling, intense. Instead of starting with the tools and technology and working towards a result, the good designer starts with the result and adapts the tools to realize it. Very often this adaptation requires the use of good lighting control. This is especially the case when a particular scheme may be required to have a different feel at different times of day, or for different uses.
WHY LIGHTING CONTROL?
10.3 The energy management role The use of energy is a worldwide concern, both in terms of its cost and on its effect on the environment. Human nature being what it is, here is another case where lip service is paid to the “correct” environmental aspect, but action is taken for cost reasons. Lighting control can make a big contribution to energy saving; but a word of caution is necessary. Where a lighting scheme has complex lighting control for esthetic or practical reasons it should be possible to program its operation so that it uses energy in the most efficient manner. However, when it is proposed to install a special lighting control system (as opposed to a simple conventional manually switched installation) solely on the grounds of saving energy, it is important to check the payback calculations. It has been the case that some complex computer controlled systems have had a poor investment payback (and maybe haven’t saved significant energy either!). Chapter 14 and the relevant applications chapters deal with this subject in practical detail, and identify the role of many different kinds of control devices in energy management. Some important points to emerge are: • In many installations the choice of light source is the biggest single factor in energy consumption. • The light output of some light sources varies during their life. Similarly in some installations the light output from luminaires can vary according to cleaning cycles. Lighting control can help ensure a particular level is maintained, without wasting energy at the start of lamp life or use cycle. • Not all sources are practical for continuously variable control. • It is always necessary to analyze the scope for energy saving by lighting control within an area. Sometimes there isn’t any. • Never forget the “user”; Totally automatic systems which operate without reference to the people in the working area are not a success. • New products and new technologies are increasing the opportunity to save energy by lighting control.
Figure 10.3 Now architectural lighting also makes much use of lighting control for esthetic effect. This exterior example is the Millennium Bridge at Gateshead, England. The striking “eyelid” design “blinks” to let boats past. The main lighting is based on ETC Irideon color changing luminaires. These are controlled by an Irideon Composer which is in turn interfaced to a Lutron controller that provides an override facility from the bridge control system and a scheduling facility. Lighting design by Lighting Architects Group.
An important calculation in any lighting control scheme is the “payback”. If one was to take a holistic view, the calculations would be complex – because not only would it be necessary to know the costs of all the alternatives in cash terms, but also the cost in energy terms. For example there could in theory be an instance where a new style of lamp, or an existing type of lamp with a controller, was shown on paper to save energy over a period of years compared with a conventional arrangement. However, if the energy cost of making the new lamp, or lamp/ controller combination, was greater than the saving in energy over the planned lifetime, there would be no point in making the change. Such a scenario is not wholly impossible. If lamps are used infrequently, for example a cupboard light, it may be impossible to justify the use of an energy efficient lamp. However, wherever lamps are used for a significant proportion of the year, it is possible to devise simple models that give valid payback information. It is easiest to do this in cash terms – but in special cases there may be a case for an energy audit.
345
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Ballast ---> System power in W including ballast losses Power saving in W Annual energy cost for 3,200 hours operation in € or $ Electricity cost saving € or $ Higher price of luminaire € or$ Simple payback time years Payback time at 7% interest
Conventional electromagnetic 142
Low Loss electromagnetic 132
Standard electronic 110
Controllable electronic 60
54.50
10 50.70
32 42.20
82 23.00
-
4.20 10.00 2.38 2.77
12.30 30.00 2.44 2.84
31.50 56.00 1.78 2.07
Table 10.1 Comparison of costs for the ballast alternatives for 258W fluorescent lamps on 230V supply. In this example electricity is assumed to cost €/$0.12 per kWh. The system power for the controllable ballast will vary, the example assumes its effective use. Note that this is for example only. Current actual costs may vary.
In order to demonstrate that there is, in fact, a payback it is necessary to use as a starting point a system which gives the required level of lighting at the lowest initial cost. An example showing the principles involved is as follows. Suppose there is an office area which is to be illuminated by twin 58W fluorescent lamp luminaires. The possible ballast arrangement for this, in ascending order of initial cost, would be: • two conventional electromagnetic ballasts • two low-loss electromagnetic ballasts • one 2-tube standard electronic ballast • one 2-tube controllable electronic ballast Setting aside, for the moment, the fact that today’s modern office would be expected to use at least a conventional electronic ballast anyway, we can do the payback comparisons referred to the conventional electromagnetic ballast. Table 10.1 shows the salient differences between the possible ballast arrangements for a given number of hours operation per year. It is clear that both the low-loss and standard electronic ballast have similar payback times; with the difference that the electronic ballast will give a better quality of light, and will have a lower running cost once the initial payback period is over. The difficult comparison is with the controllable ballast. The luminaire is clearly more expensive, but in many cases the system power will be a lot less because the dimming of the fluorescent lamps can make big savings. The table shows an example based
346
on a constant light system where there is daylight augmentation, and where presence sensors are used to dim lights when nobody is present. The calculations are easy to do, and can take in a number of variables. A neat way of doing a quick estimate for this example is shown in Figure 10.4. Starting with the price of electricity, a line is drawn “South” to intercept the line representing the number of operating hours per year. At the intersection a line is drawn West to intercept the line representing the power saving per luminaire. Next the line goes North to intercept the line representing the increase in cost of the luminaire. The line turns East and when it reaches the vertical axis the simple payback time in years can be read off. A more complete calculation can take account of the cost of the interest to service the additional capital cost – the line travels further East until it hits the appropriate interest rate, and then heads South again to give the payback time allowing for interest. It is clear that, in this example, there could be a very quick payback time when using controllable electronic ballasts. However there are some caveats. • The most obvious is that the figure for system power will depend on how much the dimming capability can be used in practice. An area at the core of the building, which is fully occupied during all building opening hours, is unlikely to make a significant saving above that of the standard electronic ballast.
WHY LIGHTING CONTROL?
Higher price of luminaire ( )
Pay back t0 (years) excluding interest 5
5% 10% 15%
Interest rate p.a.
75 4 50 3
40 30
2
20
III
10 Electricity cost saving ( )
12.5
10
5
7.5
1
2.5
Power saving per luminaire (W)
1 0.05
2 0.1
4
3 0.15
0.2
IV 0.25
Operating hours per year
5
Pay-back time t (years) including interest
Price of electricity ( )/kWh
40 30
II 20
10
I 4,000
2,000
3000
Figure 10.4 Payback diagram for different types of ballast. The reference is the use of conventional electromagnetic ballasts. The red line then demonstrates the payback time, taking into account four main variables. Diagram from the publication “Economical Lighting Comfort with Lighting Electronics” from Fördergemeinschaft Gutes Licht, Germany. Reproduced with their permission.
• However, if sensors and time control are installed, and if the area has some daylight, and is mixed occupancy (with some parts, for example, not fully occupied all day) the savings could be considerable, and actual installations do achieve the figure shown in the example. Clearly the lighting consultant has to make an informed calculation at the design stage. • But for the calculation to be valid the item “increased cost of luminaire” must also cover the cost of any additional control equipment, sensors etc. and the additional system commissioning cost. Here the new generation of digital ballasts reduces the central equipment cost and makes the calculations easier. Similar calculations can be made for all kinds of controlled lighting. They must always be done with care, and with close attention to the circumstances of the particular installation. Examples of points which can be missed are as follows.
In large buildings the air conditioning costs can be considerable. The heat from lighting equipment contributes to the air conditioning load, therefore any reduction in lighting load will also cut air conditioning costs. It can be found that the payback period for installing controlled lighting, designed to ensure that lights are not full on in areas where people are not present, is shortened appreciably compared to the “simple” payback time. When a system includes incandescent and tungsten halogen lamps, it usually does so to achieve the right ambience. Often the lamps are on dimmers to allow the lighting designer to achieve the right balance of light. The energy saving from the use of dimmers will be small (except when coupled to time control etc.) but the running costs could, nonetheless, be substantially reduced by greatly increased lamp life. Here the saving is not just the reduction in
347
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
lamp costs, and in the energy taken to manufacture the lamps, but also the saving in labor. Conversely, one must be careful not to install a system which adversely affects lamp life. Attempts to use certain discharge lamps outside their proper operating regime may result not only in increased lamp costs, but also very marginal energy savings.
10.4 Influence of legislation 10.4.1 USA Lighting has become the subject of legislation at both the national and international level. This legislation is mainly concerned with: • setting standards for certain types of lighting to ensure that it is effective. • setting standards relating to safety and personal security. • ensuring efficient use of energy. • ensuring that the waste products of lighting (especially dud lamps) do not harm the environment. Sometimes the objectives conflict. For example the requirement that unoccupied areas should not have lighting on to save energy conflicts with the requirement that they should be lit for personal safety reasons. Lighting control is affected by legislation in two ways. The first is that there may be a need for control simply to meet the legislation. The second is that a by-product of legislation can be that the marginal cost of more sophisticated control, with some resulting additional benefits, is small. Section 6.3.5 mentioned that in Europe fluorescent lighting ballasts with low EEI are already or are about to be banned. This can only encourage the use of controllable electronic ballasts. In the USA the use of energy within nonresidential buildings is covered by ASHRAE/IESNA (American Society of Heating, Refrigerating and Air conditioning Engineers and the Illuminating Engineering Society of North America) Standard 90.1– 1999. The standard is kept under continuous review, so professionals who are affected by it or its equivalent in other countries need to ensure that they keep up-to-date on its requirements.
348
In respect of lighting, the standard defines two important issues; first the amount of power that may be devoted to lighting, and second the lighting that must have provision for independent and/or automatic control. For each type of building there is a maximum installed lighting power, and there are two methods of calculating this allowance. The first is the Building Area method that applies only to projects involving a whole building. Here the gross lighted floor area is multiplied by a permitted Lighting Power Density (LPD) expressed in W/ft2 or W/m2. Permitted LPDs vary from as little as 3W/m2 for a parking garage to 24W/m2 for a manufacturing facility. The alternative to the Building Area method is the Space-by-Space method. Once again the starting point is the type of building, and the categories include offices, convention centers, hospitals, museums, retail outlets, hotels, theaters, transportation buildings and so on. But now every major area separated by walls or partitions must be calculated separately, with a different LPD applying to specific activities within the space. Enclosed office spaces are typically allowed 16W/m2, whereas specialist spaces such as the emergency reception area in a hospital is allowed 30W/ m2 .
Figure 10.5 Legislation in both Europe and the USA requires that all large area lighting is under control to limit energy usage. Sometimes the requirement conflicts with the need for security. Devices that help meet the requirement include bi-level controls for HID lamps (the one illustrated operates the lamp at two different levels by a dual capacitor system in a CWA circuit) and occupancy sensors that can operate in high bay industrial buildings. Equipment pictures from The Watt Stopper Inc.
WHY LIGHTING CONTROL?
Once the permitted installed lighting power has been calculated, the electrical consultant and/or lighting designer have some freedom as to how they use the available power, although it is in practice impossible to light a building without extensive use of high efficacy sources. Points included in the standard are: • The standard cannot be used to avoid obligations under health, safety and environmental legislation. • Separate allowances are made for exterior lighting (e.g a maximum of 2.7W/m2 for facade lighting. A minimum efficacy of 60 lm/W for lamps used in luminaires sited in the building grounds). • There is an additional allowance for a limited amount of purely decorative lighting, but this must not exceed 11W/m2; for meeting the requirements for visual display terminals in offices (4W/m2) and for highlighting merchandise in retail display (17– 42W/m2) • There are some specific exemptions for display, process and activity lighting. For example display and accent lighting in museums and retail; lighting for horticulture; lighting for medical applications and theater lighting. But, and it is a big but, the exemptions only apply where the lighting is additional to the basic building lighting and is controlled by an independent control device. Within Standard 90.1 (and its equivalents elsewhere) there are mandatory requirements for lighting control, and some of the most important are summarized here: Exterior lighting control. With the exception of lighting required to meet safety legislation or that forms part of permitted signage, all exterior lighting must be under the control of a sensor or astronomical time clock that ensures that the lighting is turned off when there is sufficient daylight. Automatic lighting shut-off. With the exception of lighting intended for 24 hour a day operation, interior lighting within buildings with floor area larger than 5,000ft2 (465m2) must be under the control of either a system that turns the lighting off at scheduled times, or uses an occupant sensor to turn the lighting off within 30 minutes of an occupant leaving a space. One control device may not control
more than one floor or an area greater than 25,000ft2 (2,300m2). Space control. Where a building is divided into separate spaces divided by walls or partitions, each such separate space must have a control device to control the general lighting. Spaces must have separate control for each 2,500ft2 (230m2). Where the control device over-rides the automatic shut-off referred to above, the over-ride is limited to four hours. Additional control. Separate control is mandatory for all lighting that is exempt from LPD requirements. Some specifics: Hotel and motel guest room lighting is required to have a master control device at room entry that controls all permanently installed lighting. Display and accent lighting, display case lighting are required to have separate control. This includes lighting for such things as display refrigeration cabinets, as well as the more obvious museum display cabinet or retail window display. Supplemental task lighting, for example permanently installed undershelf lighting, is required to have separate control – and the control itself must be sited so that it is readily accessible to the user.
Figure 10.6 Hotel guest room lighting should be under master control to reduce unnecessary electricity usage. One method is to use the room key-card to activate a master switch or dimming control.
349
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Process lighting, such as used for plant growth, lighting for food warming and lighting for medical procedures. Advertising and signage. Performance lighting, as used in theaters and auditoria, but also including spaces like athletic areas with facilities for TV broadcasting. The standard does not prescribe the control device to be used in each case, although it can often be inferred. In most cases circuit switching is all that is required, but in many cases dimming will achieve more acceptable and cost effective results. 10.4.2 UK In the United Kingdom the efficiency standards for lighting are laid down as part of the Building Regu-
lations, and have the force of law. The relevant section is “Part L” (see Reading List) itself divided into two, L1 covering residential buildings, and L2 covering commercial and industrial buildings. The latest edition came into effect in April 2002 and has considerable emphasis on high efficiency lighting sources and the need for control. Elements of it tie in with corresponding European legislation, especially in respect of ballast efficiency. The principles involved are exactly the same as those described in Section 10.4.1. However, the measurement method is slightly different. The main metric is efficacy measured as luminaire lumens per circuit watt. This means the effective lumens that come out of the luminaire after any reflection, diffusion etc. The aim is that, with very few exceptions, this figure is greater than 40
Figure 10.7 Norway’s leading telecom, IT and media company, Telenor, relocated its 6,000 Oslo based employees to a new headquarters in 2002. The lighting control requirement was to ensure efficient energy use while at the same time providing individual user choice, and an attractive, safe and practical lighting scheme. This has been achieved by standardizing on DALI compatible luminaires. This allows control of individual luminaires at a sensible cost per node. Where appropriate, lighting zones include control panels and light level sensors for local control. Each zone has a DALI-LON gateway to allow overall control by the LON based BMS. This is able to monitor all the zones as needed, set up start-of-day scenes, and set up out-of-hours scenes. Some of the night time lighting required for both esthetic and security reasons is based on the office lighting run at a low level. While some lighting is switched, nearly all the fluorescent lighting is dimmed. The stepless dimming ensures that scene changes are comfortable or imperceptible. Philips supplied nearly 5,000 DALI ballasts for the luminaires using linear lamps, and Tridonic supplied 3,000 DALI ballasts for CFL luminaires. An idea of the scale of the installation is gained from the fact that it uses nearly 400 separate DALI networks, each of which is controllable locally and independently, but which is also linked to the BMS through a LON gateway. The DALI gateways, power supplies and local control panels were supplied by Helvar. The whole installation is an interesting example of interoperability.
350
WHY LIGHTING CONTROL?
lm/circuit W. The idea behind this metric is that it gives designers some flexibility in the choice of luminaire design, lamp efficacy and the method of control. As can be seen by the formula, the use of control gives a considerable “bonus” in meeting the requirement. η lum =
LOR × Φ lamp I ×∑ P CL
Where: ηlum = the luminaire efficacy in luminaire lumens per circuit watt (the target being >40) P = the total circuit watts for all the luminaires. Σ is the mathematical instruction to add the results of the ensuing expression for all luminaires. LOR = the light output ratio of the luminaire; this means the ratio of the total light output of the luminaire under stated practical conditions to that of the lamp or lamps contained in the luminaire under reference conditions. Typical figures for practical luminaires vary between 0.4 to 0.9. φlamp = the sum of the average initial (after 100 hours use) lumen output of all the lamps in the luminaire. CL = the control factor. This is 0.8 for either photo-electric control or occupancy control; and 0.75 for a combination of both. Other points in the legislation are: • all equipment must meet the CE marking requirements in respect of electrical safety and EMC. • ballasts for fluorescent lamps must meet the requirements of the current European directive (see Section 6.3.5). • the regulations apply to all new buildings, and to all refurbishments of old buildings where the space is 100m2 or more. • there is a (rather mean!) allowance of 500W per building that can be used for lighting that does not meet the standard. • as in the USA there are specific exemptions for display and entertainment lighting; but even here a lower limit of 15 lm/circuit W is specified. • the requirements for residential lighting are less stringent; however, rules must still be followed in new construction or major refurbishments. In principle there is a requirement to provide for the use of
40 lm/circuit W equipment for a specified proportion of the building (approximately one such circuit for every three room spaces) and for all outdoor lighting unless it is under automatic control.
10.5 Lighting design In Richard Pilbrow’s book Stage Lighting Design, David Hersey is quoted as saying “I get asked all the time by students about a career in lighting. I tell them a sense of literature is more important than electricity. You don’t need to be able to mend a switchboard, and you shouldn’t just go the technical route. People tend to think about lighting in just a technical way, and really, you should think about majoring in art history, not technology. You can learn all the technical stuff at any old time”. This book’s role is very much to help the process of technical and applications learning “at any old time”. David Hersey’s point was made from his position as a leading stage lighting designer, but applies just as much to commercial and architectural lighting design. Lighting design, that is to say the design of lighting schemes as opposed to the design of luminaires, is a mixture of engineering and inspiration. It is possible to design a lighting scheme entirely on engineering and photometric principles, and it is necessary to design them so they meet electrical, environmental and safety codes. However, lighting design without insight and flair results in dull schemes which may not even meet their purpose if there is an unusual combination of surfaces and colors. The lighting design profession emerged from understanding this point. Normally it is the architect or interior designer who is responsible for the “look” of the lighting, and the electrical engineering consultant who is responsible for its technical realization. But not all practices have within them people with the understanding of the esthetic and psychological aspects of light, combined with the continuous experience of using many different light sources and luminaires. Lighting designers tend to emerge from other disciplines. They may be engineers with an artistic bent, or interior designers with a special interest in light,
351
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 10.8 The night time image of the Burj Al Arab (Arabian Tower) at the Chicago Beach Resort in Dubai, UAE, is one which is now known all around the world, both because of its architecture (WS Atkins & Partners Overseas) and because of the striking lighting design. Lighting Architects Group of Edinburgh were the principal lighting designers, with Focus Lighting of New York acting as specialist consultants for the exterior lighting controls and effects. The exterior lighting inventory is formidable. 148 Irideon AR500 color changing luminaires light the “sail” facade. 98 High End Systems high power Dataflash strobes are used for special effect, four 7kW Skytracker xenon “searchlights” provide moving pencil beams in the sky. From 300 meters (1,000 ft) away four 7kW PIGI large format filmstrip projectors from E/T/C Audio-visuel in Paris provide giant (40 x 120 meters) projected images on the “sail” for special events and for a night time spectacular. (Note that E/T/C Audio-visuel of Paris and ETC of the USA are separate companies.) Lighting control is based on the use of ETC Concept 2X lighting control computers, one on line, one as hot stand-by. Four universes of DMX are then distributed by fiber-optic link to the remote lighting sites – with a total of 1,200 DMX channels actually used. Most lighting sequences are automatic, but there is a touch screen option for manual operation for special events. Initial sequence programming was greatly simplified by the use of a graphics tablet, illustrated on the right, that allowed the programmer to select an area of the building or a specific lighting fixture, choose a color, and record the change, all with a few taps of a pen.
they may come from a stage lighting background or, following David Hersey’s recommendation, have a PhD in Renaissance Art. This is as it should be. Lighting design as a separate profession is somewhat more developed in the USA than it is in Europe. However, the situation is improving in the sense that all new buildings and building conversions that achieve popular and user acclaim seem to have a lighting designer associated with them as a member of the architect’s team. These buildings’ success derives from the fact that lighting is considered important right from the start, so becomes an
352
integral part of the building design as opposed to being “glued on” at a late stage in the construction process. The cost of lighting as a percentage of construction costs is quite small, but the result of good lighting is both user satisfaction and, possibly, prestige in the community. Far from representing an additional cost, the services of a lighting designer should ensure both cost effective and visually pleasing results. Today’s lighting designers understand the role of lighting control, and are especially careful to specify systems that are practical and intuitive for the end user.
C h a p t e r 11
STAGE AND ENTERTAINMENT LIGHTING CONTROL SYSTEMS
Stage
and
entertainment
control
11.1 Basis of stage lighting control Today architectural and commercial lighting control represents a far bigger market than theater lighting control. However, the theater is where most concepts of lighting control were first developed, so it is logical to consider the principles of theater lighting control first. Many lighting control concepts were well understood before the days of electricity. Ingenious moving shutter systems to dim candlelight, and massive arrays of valves to control gas lighting were developed in the 18th and 19th Century prior to electricity’s arrival in the theater in the 1880s. Electric lighting did, however, greatly simplify the “dimming” of light with many different methods of regulating current being used over the years. Theater lighting quickly identified many of the fundamentals of effective lighting control. Just a few examples: Remote control. It was soon realised that for practical lighting control it is usually an advantage to have the point of control remote from the device actually controlling the power. Many London theaters had resistance dimmers operated by long mechanical tracker wires – this way the dimmers could be placed out of the way at a point optimized for electrical power distribution, whereas the control could be placed where the operator could see the result of his efforts. Multi channel. Few theatrical sets can manage with only one channel of control. Quite apart from the impractability of using one big dimmer, any real lighting system uses several or many light sources. Some are needed simply to provide enough light, some are needed for specific effects. Separate control of each lamp or group of lamps is essential, either to balance the lighting to give the required
lighting
systems
overall “look” to the scene, or to create the specific effect, for example the switching of “real” lamp (referred to as a “practical”) on stage, or the achievement of a sunset effect seen through a stage “window”. Master control. Possibly the most significant early control concept was that of master control. The idea here is that not only does a single control channel have its own control, but that it, in turn, is affected by a master control which controls a whole group of channels. This allows for the obvious requirement of being able to dim a whole stage setting with a single control, instead of having to operate all the individual channel controls simultaneously.
Figure 11.1 Example of a resistance dimmer board from the 1930s. Each dimmer is operated by an individual handle, but can also be controlled by the master wheel. Photo from Strand Lighting / Brian Legge.
353
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Early master controls were mechanical. For example the operation of individual resistance dimmers by a lever could be mastered by locking the levers to a common shaft, and then rotating the shaft by one master lever or wheel (see Figure 11.1). Master control was, from the beginning, seen to need sub-division. Sub-masters could be used to control a particular grouping of lights, for example the footlights. A Grand Master would affect all the lights, typically at curtain call. Chapter 8 reviewed the different kinds of dimmers used for both entertainment and commercial purposes, so information about dimmers will not be repeated in this chapter. It is assumed that, today, any stage or entertainment lighting system consists of the required number of electronic dimmers or dimmer equivalents (for example dimmer devices built in to a luminaire) controlled by a lighting console. It is the operation of the console which is the subject of this chapter. It is further assumed that the control signal(s) from the console to the dimmers and other devices will be by one of the accepted protocols described in Chapter 9, i.e. in very small systems direct analog control, or, in most systems, DMX.
11.2 Simple multichannel lighting controls for entertainment
Figure 11.2 A simple multi-channel lighting control with overall master. The position of the slider control knobs gives a good visual indication of the relative dimmer outputs. Most such panels have analog or DMX output. The one illustrated, however, has an RF output. It is the control for the CCT “Freeway” cordless dimming system (see also Figure 17.2). (a)
+V
MASTER
1
2
3
4
output to dimmers 0-10v
5
Transistor amplifier for all but smallest systems
Microprocessor
A to D converters (b) DMX OUT
EIA 485 Driver
11.2.1 Basic single scene control Figure 11.2 shows the simplest possible multi-channel lighting control console. Here each channel has its own manually operated slider control. It could just as well be a rotary control, but sliders are much easier to operate when more than one channel is to be controlled at once, and it is also easier to see the relative states of several lighting channels by looking at a group of sliders than it is to see the relative positions of several rotary controls. A master slider affects all channels simultaneously, and here the important point is that all channels move relatively when the master is used, i.e. if one channel is set to 40% light, the action of the master is to move it from 0–40% in the same time as a 100% light moves from
354
Figure 11.3 Realization of the simple control shown in Figure 11.2. (a) conventional analog control, and (b) the additional components for digital control.
0–100%; or if the master is set at 50%, the 40% channel will give 20% light output. The realization of such a console is shown in Figure 11.3a. This shows an analog console, which may be nothing more than a simple resistance network. However, it is important that the master voltage is a low impedance source, otherwise there could be interaction between the channel controls. Figure 11.3b shows an alternative execution where the analog slider signal is converted to the multiplexed DMX signal.
STAGE AND ENTERTAINMENT LIGHTING CONTROL SYSTEMS
This simple arrangement is referred to as a single scene control. A scene of lighting is a particular combination of settings of the individual channels. 11.2.2 The two-scene preset control Figure 11.4 shows the two-scene preset control. This introduces the important concept of pre-setting the lighting scene. Now each channel has two slider controls. The A set of controls is used to set the first scene, and the B set to set the second scene. Moving the A master control up to 100% establishes Scene 1. When Scene 2 is required the B master is moved to 100%, and the A master faded down to zero. Once A is down at zero it is now possible to preset Scene 3, by moving the A sliders to new relative settings. Scene 3 is then established by once again cross-fading the A and B masters. The process can be continued for as many scenes as required. The presetting of a scene is blind plotted, that is to say the settings are made when the operator cannot see the effect of operating the sliders. Thus the scene must have previously been rehearsed, so the settings are known. The list of slider settings for successive scenes being the lighting plot. What happens if both masters are set to 100%? In simple terms a new scene is established which combines the attributes of Scenes A and B; but, unless the console has some special rules built in, the normal result will be that for any channel highest takes precedence. That is to say, if on Channel 2, Preset A is 40% and Preset B is 60%, the result will be a 60% output on Channel 2 when both A and B masters are set to 100%. The Highest Takes Precedence principle, abbreviated to HTP, is not the only rule that can apply to multi-channel control systems. 11.2.3 The dipless cross-fade Astute readers will have seen a problem arising from Figure 11.4 if the design of Figure 11.3 is doubled up. Suppose that in a two-scene preset control a particular channel is intended to have the same setting for both scenes. If A is the first scene, there is no problem if B is moved to 100% before A is moved
Figure 11.4 The two-scene preset control. This example is the Artistic License Preset 12. Besides having the two preset masters, it also has an overall grand master.
to zero, but this overlap fade will probably produce an unwanted look to the change. More likely a cross fade will be executed where A goes down at the same or similar speed to B coming up. The channel intended to stay at the same setting will fade down with A until it meets B coming up, in other words there will be a dip in the channel level during the crossfade. In fact it can be seen that dips are just as likely to arise if the second scene level is higher than the first, and that even a move to a lower level may not have the expected appearance. The problem is solved by the dipless cross-fade where, in effect, the system knows the channel destination. This is comparatively easy to achieve in a digital memory based console, since by definition the system does know the destination, and changes to levels can be constrained to take place only in the required “direction”. In analog, non-memory, systems the choice lies between using overlap master fades, or having quite complex mastering circuits. 11.2.4 Multiple preset, multiple group control The two-scene preset principle can be extended. A three-scene preset system has the advantage that a
355
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
scene transition can be taking place while another scene is being plotted. A ten-scene preset system while rather clumsy in terms of the number of controls needed, might be highly practical for a small cabaret stage or similar application, since there might well be only ten or less scenes required overall. The system would then allow all ten scenes to be preset, and the operator could run the show without having to do any plotting at all – simply bringing up the required master control for each act. In fact it is found that a more practical arrangement is to have groups within each preset. Typical consoles in the 1960s and 1970s would have three groups for each preset, and would allow submastering for each group. They would also allow for cross fading between any combination of groups. Figure 11.5 shows the idea. The situation today it is that, with the exception of the simple two-preset control, consoles using multiple controls for each channel are no longer built. While the multiple preset concept works well, it does result in very large consoles for big systems, and requires continual manual operation in a show of any complexity. It is now also the case that the cost of lots of control components, like sliders, becomes an important factor; when the same result can be achieved at lower cost, and in a much smaller space, using electronic memory.
Figure 11.5 The master controls of a three preset/three groups per preset console (left). Close up of the cross fade control (right). This provided manually controlled or automatically timed dipless crossfades between selected combinations of subgroups.
356
11.3 Memory consoles 11.3.1 Background The idea of using “memory” devices in lighting consoles is not new. Frederick Bentham, of the Strand Electric & Engineering Co (now Strand Lighting) harnessed the organ stop grouping ability of the Compton Organ console to theater lighting in the 1930s. IBM computer-punched card readers were extensively used in the USA in the 1960s. In Europe Grosman of Copenhagen offered the “Memocard” system. However, it is fair to say that the era of memory consoles really started when Thorn introduced the “Q File” in 1966 (sold by Kliegl Bros in the USA). Q File was developed for TV studio use, but quickly found its way into the theater. It used the then main method of fast random access storage in computers, the magnetic core memory. In the following years many manufacturers of memory consoles emerged, all of whom sought to take advantage of new memory components as they became available. Whereas the high cost of computer memory in the early days of memory consoles tended to seriously affect their design, today’s computer memory is sufficiently inexpensive that memory cost is no longer a factor. All the early memory consoles, which predated the microprocessor, were based on proprietary hardware designs where, in effect, the memory function was glued on to the hardware controls. With the arrival of the microprocessor and, subsequently, the personal computer, the architecture changed. Today a memory console is invariably based on a standard microcontroller or microprocessor, introducing many benefits, including: • the ability to have the operating software written in a high level language. This in turn allows rapid development and the introduction of new features during the product’s life. • the ability to include operating methods which are already widely familiar – such as the ability to “back up” show data on floppy disk, and to display lighting status on standard displays.
STAGE AND ENTERTAINMENT LIGHTING CONTROL SYSTEMS
Figure 11.6 The complete console of which Figure 11.5 forms a part. The Multiway-2 from Electrosonic (introduced 1976) was available in versions from 36 to 120 channels.
Memory consoles are now available in a range of sizes with many different operating methods. Some are designed to use the minimum number of operating controls, and others are the exact opposite with a large number of “live” controls. Some are general purpose, intended to control all types of lighting device, others are specialized, for example consoles designed only to program moving lights. There are now so many different consoles and methods of working that it would be impossible to describe them all, so the following sections concentrate on the features likely to be found in most consoles, not necessarily using the descriptive terms used here. Some of the terms used in practice are manufacturer specific or market specific, or arise from different use traditions. 11.3.2 Patching When a memory console is connected up, the first thing to be done is to configure it to its controlled devices. Before the advent of memory consoles, the lighting console would be hard wired with one control line to each dimmer. Often there were many fewer dimmers than available lighting circuits, resulting in the need for a high power patch panel, where circuits would be plugged in to the selected
dimmers. While load patching is still used in some applications such as touring, permanent installations are now much more likely to use a one dimmer per load configuration which is permanently wired. This gives much greater flexibility, but introduces some inconvenience and rigidity. The problems that need dealing with are: • each dimmer (or other device) will have one or more DMX (or equivalent) channel numbers. These may be individually set on the dimmer, but are more likely to be allocated in blocks according to the dimmer cabinet construction. • the outgoing loads will have been connected to the dimmers in a manner which is most convenient for installation. This may not result in a logical numbering scheme for control. • whereas, when DMX was introduced, all channels were expected to be used for dimmers, and all channels used 8 bits, this is no longer true. Channels are also used for non-dim circuits, moving lights, scrollers and other devices. Some devices need 16 bits (2 DMX “channels”). These uses may require the channels to be identified in a special way, or allocated to a dedicated part of the console. • the greatly increased demand for “channels” means that a single DMX output is often not sufficient. Many consoles have several DMX outputs, increasing the potential for numbering confusion. Small memory consoles have a single DMX output, and, generally, an arrangement whereby the “start channel” in a DMX stream can be identified, thus a 24 channel console could be set to DMX 1 – 24, or DMX 41–64. All other consoles have a soft patching facility which allows the user to assign any console output channel to any DMX address. Soft patching facilities vary, but in addition to the prime requirement of allocating channels, may include: • the ability to assign many dimmers to a single control channel. This need arises in big event lighting where a single channel might be used to control a large amount of color wash lighting, maybe requiring 20 1kW luminaires, most conveniently served by five 4kW dimmers. • the ability to modify a channel in a specific way. This is most commonly needed for non-dim circuits, where, if a channel is selected, it is essential that it
357
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
only operates at 100% – examples are a fan motor or effects motor not suitable for dimming. • the ability to specify a particular dimming law. • the ability to specify that the channel output is only a proportion of the control output. This is sometimes used where a channel is otherwise linked to one or more other channels for control, but it is found desirable to limit its output, either for balance reasons, or for lamp life reasons. 11.3.3 Displays and back-up The simple action of configuring a console by soft patching indicates a need for some form of information and status display. Within the modern memory console there is a huge amount of information available, and the problem is deciding what to display. When a console includes a significant number of hard controls, such as sliders, wheels, and buttons, it is useful that these are either devoted to a single task, or that each has an associated display to show status. In many cases the display is as simple as an LED indicator to show that the control is “live”, but often it needs to be some form of alphanumeric indicator, especially when, for example, a slider control can be assigned to control an individual channel.
Small consoles include no more than these simple display devices, but consoles of any size invariably use one or more computer monitors. While traditionally these have been CRT monitors, there is an increasing use of compact LCD monitors to save space. Two monitors represent a common configuration, with one monitor used for cue and show status, and the other used for channel status. The arrival of flash memory has allowed the introduction of new console architectures with minimum back-up facilities. However, most consoles have provision for varying types of back up, both for security and to allow show data to be transferred between one console and another. In a small console the entire operating software is carried in EPROM, and the current show data is in battery backed RAM. Short term power failure is not an issue, since the console can recover from it immediately. For long term back-up, and to store shows which may be needed again, some form of exchangeable medium is used. This is either floppy disk or exchangeable non-volatile memory, now generally “flash cards”. Some consoles use the same type of memory to carry the operating program, making field updating of firmware much easier than is the case when EPROMs must be exchanged. Larger consoles work more like a standard computer, sometimes using a conventional operating system. For this reason they may need a UPS. Increasingly, show data and operating software/ firmware updates are being transferred by means of computer network or simple computer-to-console connection. 11.3.4 Controls
Figure 11.7 The Strand Lighting 300 Series memory consoles use two LCD status displays within a compact, transportable, construction.
358
At one extreme a memory console is a personal computer with a standard alphanumeric keyboard, and all data and operating information being entered through the keyboard. At the other extreme it is a descendent of the multi-preset board with a set of controls for each channel for entering the scene information, and the memory feature simply being used to replace the need for manual pre-setting during show playback.
STAGE AND ENTERTAINMENT LIGHTING CONTROL SYSTEMS
While personal computers are sometimes used for pre-programming, all practical memory consoles lie somewhere between the two. The layout and number of controls depends on the application and on user preferences. Modern consoles are often required to control hundreds or even thousands of channels, so it would clearly be impractical to have thousands of slider controls; on the other hand, when setting up a group of lights, it can be much easier to balance them by working a small group of sliders, than having to punch in a series of numbers. Once they have been set, the sliders can be re-assigned to do something else. A generic memory console is shown in Figure 11.8 to identify the controls likely to be encountered. Channel controls. Individual channel controls may not be fitted at all. In this case all access to individual channels is through a numeric keypad. However, consoles used for live events such as concerts (as opposed to repertory theater) do have channel controls and these are simple sliders (slider
potentiometers connected to A/D converters). There may be one slider per channel, but far more likely these will be assignable controls requiring some kind of display to show which channel they are operating and, possibly, channel status if it is active in the current cue. A sophisticated display will not merely refer to a channel number, but will include a short text description. Associated with each slider there may be a push button. This may be used to “flash” or “bump” the channel to full output to aid channel identification. Submaster or group controls are similar, but this time each slider controls a complete group of channels. For some shows one set of submasters may be sufficient, but again these controls can be reassigned. For event lighting the “flash” button is often used for live effect. Wheels, belts, trackballs and mice. A problem with a conventional slider control is that the moveable element denotes a particular “output”. This means that, if a particular control is re-assigned, the
Figure 11.8 The Colortran Innovator 48/96 is shown here to identify the generic controls to be found on memory lighting consoles.
359
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
new “assignee” will “jump” to the existing setting unless the operator has previously moved it to the new assignee’s position. This would result in a sudden change in level for a dimmer or group of dimmers, or a sudden change in speed for a speed control. This problem is avoided if the control is of a “take command” type, whereby the new assignee (whether dimmer channel, group or speed of change) does not change when the control device is connected. It only does so when the control is operated. A commonly encountered control of this type is the computer mouse that does not “grab” anything on the screen unless it is “clicked”, at which point the “drag” action of the ball encoder in the mouse becomes effective. Trackballs and mice are used as components in memory lighting consoles, but more common are the wheel and endless belt. These work in a single dimension only. All these kinds of control are based on shaft encoders that, in response to rotary motion, yield a digital pulse output corresponding to the angular movement made. Playback controls are described in more detail in the next section. In principle they allow manual control of the cue playback process. Playback can
be initiated by the use of “Go” buttons to initiate automatic timed changes, or by manual controls such as sliders. Keypads. There may be one or more of these. They are used for channel and group selection, and for the entry of other numerical data such as dimmer levels and fade times. Most are numeric with simple arithmetic functions. If a full alphanumeric keyboard is needed, it is usually plugged in temporarily at the show configuration stage, but is not used for show operation. Grand master functions. Following theater tradition there may be grand master and blackout controls. Special function keys. These cover all kinds of activity; either part of the show recording process or for operating special effects. When a large number are needed, they operate in the same way as those on a personal computer, in that they can be reassigned or used in combination. All techniques used in computer systems are available to the memory console. So if a designer prefers to use a touch screen in place of function keys, or an “airmouse” in place of a slider, he can,
Figure 11.9 The Strand Lighting 500 Series of memory consoles used for large venues includes support for moving light control. Such consoles can be delivered in configurations able to control up to 6,000 channels. The number of fader controls fitted depends on the application.
360
STAGE AND ENTERTAINMENT LIGHTING CONTROL SYSTEMS
in theory, do so. However while it is the case that console operators welcome innovation in console design, they also look for some consistency and familiarity. For this reason many consoles look similar, with their differences expressed in software detail. 11.3.5 Playback philosophy The way in which memory consoles are used in practice is best understood by examining how shows are “played back”. If we start with the two-scene preset concept introduced in Section 11.2.2 we can imagine that playback could be achieved by having a single cross-fade lever, which is moved up or down at each cue, with the “scenes” automatically advancing. We could imagine further that, instead of using a lever, we could use a simple cue “Go” button to release a cue with automatic timed, dipless, crossfade with its predecessor. If we wanted to modify the timing we could “grab” manual control of the crossfade using the wheel or similar device. This “preset oriented” method of playing back is widely used in theater consoles. It initially proved popular because it matched the method of working of pre-memory consoles, but it is still a valid concept and is widely used. In principle there is a separate memory slot for each scene, so within each scene there is memory space for each channel. More complex consoles allow for the simultaneous operation of more than one preset playback with highest taking precedence. As with the earlier multiple preset consoles, some users prefer to treat their lighting in “group” terms. Each group can consist of any number of selected channels. The groups can either each have their own control, or, more likely be assignable to a number of submaster controls. This approach is favored in applications where the running order may not be predetermined, and in installations with a fixed lighting rig that may be used for multiple purposes with minimum (or no) rehearsal time. Because memory consoles use digital processing and are not dependent on simple electrical networks, it is possible for group and/or submaster controls to undertake some special tasks.
For example, it is possible to have a “negative” or inhibit group or submaster. In most cases dimmers will work on the HTP principle, so if a given channel is assigned to more than one group, the highest signal prevails. However, if an inhibit submaster is operated, all channels in that group will go to zero, regardless of any other signal. A simple example of its use is in a theater where all the “front of house” lighting might be on such a group. Another possibility is to designate a control as solo, so that if it is operated the channel(s) concerned follow its signal only and ignore signals derived from other groups. Group playback consoles have the facility of running many playback sequences; each can consist of a string of cues, and associated with each group move there can be programmed fade-in and fade-out times. It is therefore possible to have several simultaneous lighting moves, all operating at different speeds. A simple dramatic example is a scene with a sunset backgound involving a “sunset group” that is slowly fading down, combined with one or more foreground groups; for example an actor switching on a light requiring a fast fading “room lighting group”. Within the sunset group itself there could be different fade times applied to luminaires within the group to achieve the effect of the deepening red of a sunset. The final playback method is referred to as the tracking method. It derives from the old method of stage lighting control, before the preset era. Then there would be a large number of manually operated dimmers. Any lighting cue would involve only the operation of the dimmers needed for the cue, all others remaining untouched. Tracking sees a show as a continuum. At any moment in it all dimmers will either be at a fixed level, or will have a level determined by a fade actually in progress. With a tracking console the idea is that you can choose a designated point in a show sequence, either by name or time, and the console will immediately adopt the corresponding lighting state. In practice the computer and memory power needed for the three playback methods is roughly the same. Even within a console nominally of one type, features of the other types occur. Many consoles offer a choice of playback methods.
361
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
11.4 “Live” versus “automatic” Stage shows vary enormously in their need for complex lighting control and in the appropriate method of operation. At one extreme is a long running straight play, where the lighting cues are the same for every performance. At the other is the one night concert where even the running order may not be known at the start of the performance, and where the lighting may need to react to the individual performance. The former’s lighting control needs can almost be met by an “automatic” system, where the operator simply has to release pre-programmed lighting cues. The latter requires something much more “hands on” to allow the operator to react quickly to events on stage. The console needed for the “automatic” presentation does not need individual channel controls – it is easier to select channels using a numeric keypad, since this allows groups of channel numbers to be entered with a minimum number of keystrokes. On the contrary the “live” console needs a lot of separate controls that can be manually operated to affect different lighting groups. This difference was understood even before memory consoles became the norm. The advent of the touring “rock and roll” show, pioneered by groups such as The Who, Genesis, Deep Purple, The Rolling Stones, Pink Floyd and others in the 70s re-
Figure 11.10 The Electrosonic “Rockboard” of the 1970s was developed to meet the needs of the “live” touring show.
362
Figure 11.11 Pin matrix used to assign channels to ten matrix master controls on the Rockboard.
vealed the limitations of the conventional multipreset control. The Electrosonic “Rockboard” (Figure 11.10) was one of the first lighting control consoles developed for this market. Although it was basically a two-preset, three groups per preset board, it included the following additional features: • a “flash” button for each channel, allowing individual channels to be momentarily sent to full output. • a “chase” feature that allowed selected groups to be automatically flashed to full output in a high speed sequence. • a set of ten additional group masters. The assignment of channels to these master controls was achieved through a diode pin matrix (Figure 11.11.) • an optional “sound to light” feature that allowed the beat or frequency range of sound to affect the
Figure 11.12 The Celco Gold (1980s) console introduced memory into the touring concert market, but retained manual access to many channels.
STAGE AND ENTERTAINMENT LIGHTING CONTROL SYSTEMS
Figure 11.13 The Avolites Diamond 4 console represents the latest generation of live event consoles. Its 156 faders have individual electronic legends. Each fader can be used to control an individual luminaire, a cue group or a chase sequence. It can serve six DMX lines directly, and additional channels via ethernet. A drawer houses the keyboard used for entering luminaire information and a graphics tablet option for moving light programming.
lighting. It could do this either on assigned groups of lights or could use the beat to trigger steps in a chase sequence. When memory was introduced into this kind of console in the 1980s, by Celco, Avolites and others, the memory feature was used to allow the creation of many more groups and animation sequences that could be called up individually or in parallel. However these consoles, and their successors today, retained the large number of accessible controls. The “rock and roll” or “live” control philosophy had a profound influence on the facilities offered on mainstream theatrical lighting controls. While consoles are still designed to meet the specific needs of the touring market (see Figure 11.13) many other consoles intended for permanent installation offer control features for the single event or music concert either as a standard facility or as an optional extra. Consoles like that of Figure 11.13 use a “Page” philosophy to control large numbers of luminaires. Although the controls will, at any one time, give manual access to, for example, 128 luminaires, operation of the page command gives access to another group of 128. While channel manual access is restricted by the page presentation, actual show play-
back is across all pages simultaneously. The console shown in Figure 11.13 is able to control over 1,000 luminaires if necessary.
11.5 Control of moving lights 11.5.1 Introduction A big change that has emerged in the last 12–15 years has been the widespread introduction of moving lights. Early examples of these were clumsy devices, typically based on a standard theatrical luminaire mounted in a large motorized yoke assembly with its own dedicated control. While the impetus for the use of moving lights came from the “rock and roll” concert, they quickly found a place in more conventional entertainment lighting, and have now entered the architectural field. A wide range of moving light luminaires is now available, with varying degrees of sophistication, precision and light output. So now the situation is that moving lights are accepted as standard weapons in the lighting armory, and this, in turn, means that any serious lighting console must be able to control them, in addition to being able to control standard dimmers.
363
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
convenient to have all lighting control from a single console. Thus the consoles shown in Figures 11.7, 11.9 and 11.13 are all designed to control both conventional dimmers and moving lights. 11.5.2 Principles
Figure 11.14 Close up view of the preset cue masters on the Avolites Diamond 4 showing the programmable electronic display.
When moving lights first appeared, the control systems used were stand-alone; i.e. they only controlled moving lights, and all conventional lighting control (dimming) was done by a separate console. For large scale shows this is still the case (see Section 11.7) but for many applications it is much more
Moving lights work on two basic principles; in both cases the aim is to pan (move horizontally) and tilt (move vertically) the beam of light coming from a luminaire. While any kind of luminaire could be used, it would not make much sense to move a wide angle floodlight, so all moving lights are “spotlights” of one kind or another – i.e. they produce a beam of light that can be focused, or is provided by the nature of the lamp used.
Clay Paky Stage Color 1200.
ETC Irideon AR50 wash luminaire for architectural applications.
MAC 2000 profile moving light from Martin.
Figure 11.15 The Strand Pirouette is a moving spotlight intended for theatrical use. It uses a 2kW tungsten halogen lamp, and is available with fresnel or prism convex optics.
364
Figure 11.16 Examples of moving head lights using discharge lamps.
STAGE AND ENTERTAINMENT LIGHTING CONTROL SYSTEMS
Clay Paky Golden Scan 3.
Roboscan Pro 518 from Martin.
Figure 11.17 Examples of moving mirror luminaires (“scanners”) using discharge lamps.
The most flexible arrangement is to move the whole luminaire. This requires the luminaire to be mounted within a yoke (within which a tilt movement can take place) and the complete assembly is then mounted on a rotating base, to achieve pan. The implication of this arrangement is that fast movement with high precision is difficult to achieve because of the mass of the luminaire. The alternative arrangement is to use a fixed luminaire with a moving mirror arrangement. Moving lights using this principle are sometimes called scanners. The lightweight mirror allows very fast scans to be achieved. However, when a simple single diverter mirror is used, the angular range is limited since outside of certain angles the light beam is obstructed by the luminaire body. In practice a full moving head system achieves 360° or more of pan (typically 440°), and around 300° of tilt (this parameter is limited by the luminaire base dimension, since this represents an obstruction to the beam). On the other hand a scanner may only achieve 180° of pan (although the best get close to 300°) and is limited to around 85° of tilt.
Some early moving lights used simple motor drives and unsophisticated positioning systems, but it was soon realized that, to be more than a gimmick, moving lights had to work with high precision. Fortunately their development coincided with the availability of low cost digital electronics, so all moving lights are now microprocessor controlled. Within the luminaire the priority is to achieve precision positioning of the light beam. In practice this can only be done using stepper motor drive for the moving head or mirror. The stepper motor is a device that, when the correct sequence of electrical pulses is applied to it, “steps” a defined angle. Continuous rotation is achieved with a stream of pulses, but the rotation can be stopped at a precise angular position, and the position is held electromagnetically. A typical stepper motor can stop at any one of 200 positions in one revolution with an accuracy at each position of better than 5% (i.e. each position is 1.8° apart with an accuracy of 0.09°). In a moving light or scanner such motors are used with a combination of microstepping, that involves a more complex pulse sequence to achieve intermediate steps, and gearing to achieve considerable accuracy. Typical performance figures are: Moving light: Pan in 0.013° steps, tilt in 0.007° steps.
Figure 11.18 Wynne Willson Gottelier (WWG) point out that single mirror scanners cannot maintain a symmetrical beam field (except when the mirror is at exactly 45°). They have devised the “Beam Bender”, a dual rotating mirror system that ensures an orthogonal optical path. Here it is shown applied to a 2.5kW theatrical luminaire from Robert Juillat. The resulting product is sold as the Fantôme.
365
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
two are used, the second one is referred to as the “fine” control signal. As an example, the position of a moving light that provides 400° of pan can theoretically be programmed to within .006°, so a performance at least equal to the precision of the stepper motor drive is available. The use of DMX as a control protocol, and the implication that a modified lighting console can be used to control moving lights, introduces the question of what happens when a moving light receives more than one command on a single channel. In Section 11.2.2 the HTP principle was introduced, and this is normally applied to lighting intensity level. However, this does not make sense for lighting movement (or for other moving light attributes described later) so the principle used is latest takes precedence or LTP. 11.5.3 Intensity control Figure 11.19 WWG’s double rotating mirror system also appears as the “Razorhead”, suitable for attaching to 4kW and 7kW xenon arc searchlights. These represent the heavy end of the moving light business! – but the Razorhead gives 360° pan and tilt, far more than can be achieved by the usual method of yoke mounting the searchlight.
Scanner: Pan in 0.028° steps, tilt in 0.056° steps. The incoming control signal defines the required position of the light beam. A single signal causes the beam to move to the required position, a continuous changing signal causes continuous movement. If a luminaire is sent a new position instruction not adjacent to its present position, then it will move at a default speed (or at a separately programmable speed). Otherwise the speed of movement is governed by the control signal, provided it is within the capability of the luminaire. Although some moving lights use proprietary protocols as an option, the “standard” signal for control is DMX, but, as already pointed out in Chapter 9, a single DMX channel can only give 256 positions – an accuracy of worse than 1° for many applications. Most moving lights and scanners allow for both 8-bit and 16-bit resolution. With DMX this is done by providing either one or two DMX channels; when
366
Moving lights are available with tungsten halogen lamps, in which case the lamp is controlled by a conventional dimmer. However, the great majority of moving lights use discharge lamps as the light source. The need to provide a focusable beam of light within a small lightweight housing favors a small point source with high luminous flux. Lamps in the compact source metal halide family are used, providing a life of around 750 hours for the larger lamps and up to 2,000 hours for some smaller ones. System light output is between 1,200 lumens for a scanner fitted with a 200W lamp, up to more than 16,000 lumens for a moving light fitted with a 1,200W lamp. The use of discharge lamps means, in turn, that light intensity has to be controlled by mechanical means. A DMX compatible motorized dimming shutter is used. One of the effects that is demanded of moving lights is the ability to flash or “strobe” the light output. In some cases the dimmer shutter works fast enough to allow the effect to be achieved through the dimming shutter. In other cases a separate strobe shutter is used. The use of discharge lamps implies that the lamps are on all the time, drawing full power and using up lamp life. In shows or applications where it is known that there are long intervals where a group of lights
STAGE AND ENTERTAINMENT LIGHTING CONTROL SYSTEMS
are not being used, or where load management becomes an issue due to limited electrical power available, it is necessary to start the lamp on demand. This must be done in a way that does not produce unexpected visual results, so it is usually coupled with a forced “dimmer zero” command for light intensity. As lamp life is critical, and represents a serious cost of ownership, it is necessary to know how many hours a particular lamp has run, so that timely replacement can be effected. Most moving lights have a local control and readout that gives this information. The moving light business is competitive, and there is great pressure to keep unit cost down. For this reason the ballasts used for the lamp are often electromagnetic; however, the weight reduction and better lamp performance afforded by electronic ballasts is leading to their more widespread adoption. A large moving light rig using lamps that are electromagnetically ballasted may exhibit a poor power factor, and the electrical distribution system must be designed accordingly. Modern electronic ballasts have a near unity power factor, so do not present a problem in this respect. 11.5.4 Moving light attributes
Figure 11.20 Gobos produce simple patterns, and may have a rotation option. Photo from Martin Professional.
The moving light or scanner starts with a focusable optical system and an electronics package that can decode fast streams of digital commands. It is not surprising that this is used as a basis to make the moving light an even more complex instrument. Many of the additions are just as applicable to static luminaires, but, because the moving light industry is so well developed, it can be the case that the most cost effective way of achieving these effects is through the use of a moving light luminaire. All moving lights have a means of selecting color. This subject is dealt with in Section 11.6. In principle color change requires the insertion of one or more filters in the optical path. But color filters are not the only things that can be placed here. Now it is the ingenuity with which extra features can be packed in without significant loss of light that limits the number of effects that can be achieved. In prac-
tice users have to choose between more sophisticated color selection and a greater variety of effects. Some of the additional effects are “out of focus”, and others must be placed in the focal plane of the optical system so that they produce a well focused image. Examples are: Gobos. A gobo is a mask placed in the focal plane of a spotlight. In a slide or scene projector (or video projector) this position is kept as cool as practicable; but within a spotlight it is comparatively hot, and for this reason gobos are often made from etched metal. This generally limits their use to providing simple shapes. In moving lights gobos may be carried on a gobo wheel, providing a choice of patterns. In some designs glass mounted gobos can be used, allowing the creation of almost photographic images. Indeed some special versions of these luminaires allow the use of slides.
367
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
DMX Channel 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Attribute Intensity (dimmer shutter) Pan (first 8 bits) Fine pan (second 8 bits) Tilt (first 8 bits) Fine tilt (second 8 bits) Focus (first 8 bits) Fine focus (second 8 bits) Cyan (for color mixing) Magenta (for color mixing) Yellow (for color mixing) Color wheel (fixed colors) Gobo wheel 1 Gobo wheel 2 Gobo rotation speed Iris Strobe shutter Diffusion (frost) filters Prism effects Zoom Color temperature correction
Table 11.1 Hypothetical moving light luminaire requiring 20 DMX channels. When analyzing the capacity of a control console, the “channel” capacity gets used up very fast when the luminaires have many attributes.
Rotating gobos. A refinement is the ability to produce rotating patterns by having a rotator attached to the gobo holder. Prism effects. By inserting prisms in the optical path, the light can be refracted in various ways, producing multiple light spots. Some luminaires have a single prism, others offer a choice of, for example, three prism sets, with or without rotation. Iris. An iris placed at or near the focal plane can vary the beam diameter (but not the beam illuminance). Iris control can vary the diameter in the range 10–100%. Beam shaping. A more elaborate offering is the ability to shape the beam using motorised X-Y masking shutters. Zoom. Another option for changing beam diameter is the use of a zoom lens. In this case the illumi-
368
nance does increase as diameter decreases. A typical range is to change the beam angle between 15° and 26°. In practice the use of zoom lenses is confined to scanners. Focus. If the throw of the luminaire is changed, there may be a need to re-focus to ensure that the gate image remains sharp (or deliberately soft). Motorized focus control is needed for this. Diffusing filter. If a more diffused light is required, a frost filter is used. This is such a common requirement that it may be a standard feature of the luminaire. The control of moving lights is, therefore, very complex. Each luminaire requires intensity, pan and tilt control, together with control of a variable number of attributes as indicated above. Table 11.1 shows how a hypothetical luminaire could use as many as 20 channels of DMX. Because moving light technology has developed piecemeal, there is no standard method of allocating the use of DMX channels, so in practice lighting consoles are equipped with driver information for as many as 300 or more different luminaires. The user then only has to identify the luminaire being used by name; and the console automatically assigns the first channel in a DMX group as the intensity control; and allocates as many attribute channels as needed. Clearly the way in which DMX is used for controlling attributes can be different than when used for dimmers. Often individual data bits are used to select a single effect. The concept of “mastering” may or may not be relevant. The use of a single
Figure 11.21 Avolites consoles allow push button selection of attributes for programming.
STAGE AND ENTERTAINMENT LIGHTING CONTROL SYSTEMS
master control to initiate and control the speed of a group of lights moving in unison makes sense; the application of a varying 0–100% control for color probably does not. Many attribute control signals have to be suppressed or “masked” except when issuing a single command in order to avoid unwanted movements of color wheels etc. 11.5.5 Programming moving lights The programming of moving lights can be extremely time consuming, and many different methods have been devised to simplify or speed up the process. Consoles come equipped with many macros, programming routines that simplify the entry of repeated functions. Thus if one moving light is set to a particular attitude, it is easy to repeat the instruction to several more luminaires without having to enter numerical data. Many touring shows rely on the setting up of a rehearsal lighting rig that allows the lighting designer to try out all the different effects for a show. Sometimes the rehearsal rig is a scaled down version of the real thing (i.e. all the luminaires are present, but the stage dimensions are smaller). There can be automatic routines that adjust luminaire focus on the “real” rig to match the longer throw. In the same way that dimmers are controlled in groups, moving light activity can be simplified by the use of focus presets and palettes. Focus presets allow the quick recall of a particular luminaire beam position – e.g. center stage. Palettes cover the nonmoving attributes, such as color and optical effects. So a scene requiring ten moving lights to produce a green vertical light curtain effect can be achieved by the selection of a “vertical beam focus preset” and green palette. Many cue sequences are built up by starting from a focus preset. This means that if a position changes, it is only necessary to adjust the preset, and all cues based on it will automatically be adjusted as well. A graphics tablet can be a convenient way of entering moving light information; when coupled to a good display such a device provides a quick way of entering: • beam positions etc.
• selection of color, by having a CIE color triangle on the display and simply selecting the required color from within the triangle. This then automatically generates the required CMY commands. • indication of moving beam pattern. As a simple example setting a moving light to describe a circular rotating pattern. The graphics tablet can be used to indicate the diameter of the circle required and the speed of rotation. A refinement now available from several manufacturers is the use of CAD (Computer Aided Design) principles for live or off line programming. This gives a graphic visualization of the lighting rig, and can show luminaire movements, beam angles etc. By combining this with a graphic tablet, the positioning of luminaires becomes a relatively quick task.
Figure 11.22 WYSIWYG (What you see is what you get) from Cast Lighting is now offered by ETC to assist moving light programming. One of its abilities is to “read” DMX streams and convert them into the simulated lighting shown on the monitor.
369
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
A problem that always arises is that of accurately establishing datum information. One solution is the use of a device that automates the process. Figure 11.24 shows a device that is placed at a focus preset position on stage. The console causes each luminaire in turn to carry out a sweep pattern; when the “Focus Finder” detects light, it radios back to the console, which can then record the lamp position data. A “fine” facility causes the device to look for the center of the light beam. There are some shows where the equivalent of an automated follow spotlight is needed. This can be provided by devices like the Wybron “Autopilot” (Figure 11.25). Here the performer wears a transmitter unit whose signals are picked up by multiple receivers; this allows the calculation of the transmitter position in three dimensions. This information can be passed back to the moving lights. Besides being used for automated following, the same technology can be used for deriving initial set up information.
Figure 11.23 The Visualiser computer program from Avolites allows off-line programming. The center figure and monitor display show an example simulation, and the inset figure shows the actual show in progress. Photos courtesy Bryan Leitch (Lighting Director).
370
Figure 11.24 The Focus Finder from Avolites is placed on stage at a focus reference point. Luminaires scan the stage, and, when they are “found” by the focus finder, the pan and tilt information is radioed back to the console.
11.6 Control of color 11.6.1 Color filters In some architectural applications color is achieved by choosing the appropriate light source (for example a blue metal halide discharge lamp). However, within all entertainment and theatrical lighting, and most architectural lighting, color is achieved by inserting filters into the light path. The important point to bear in mind when using filters is that colors can only be achieved if the required wavelengths are available from the source. When tungsten halogen light sources are used results are reasonably predictable, because the filament lamp has a continuous spectrum. However, when a discharge lamp is used, its spectrum may have several peaks and troughs. Putting a filter in front of it that is calling for a color corresponding to one of the troughs, results in very little light. Colored bulbs are created by coating a lamp with a lacquer film, or some other powder coating. Exceptionally, colored glass may be used. Colored bulbs are mainly used as decorative sources, and not as the basis of any kind of color control. Glass filters, where the glass is colored by adding minerals to the molten glass at the time of manufacture, have the merit of long life, but are not consistent – one batch of a particular color will be
STAGE AND ENTERTAINMENT LIGHTING CONTROL SYSTEMS
11.6.2 Color selection
Figure 11.25 Autopilot from Wybron can track up to four different performers. Each wears a beltpack that sends ultrasonic signals to an array of up to eight receivers. The position of each performer is calculated, and up to 60 lights can be controlled. In use the Autopilot system controller intercepts the DMX stream from the console. It strips out any console pan and tilt information and replaces is with the live tracking data, but leaves in the other luminaire attributes.
subtly different from another. Glass color filters are particularly useful for exterior and underwater lighting. The choice is very limited. Plastic sheet filters. In entertainment the most widely used color filters are made from self-extinguishing dyed plastic sheet. Manufacturers offer hundreds of different colors allowing lighting designers to achieve effects of great subtlety. Polyester filters are flexible and low cost. Polycarbonate filters cost more, but the cost can often be justified because the replacement interval is much longer. Dichroic filters. These use the principle of interference (see Section 1.7) to achieve color filtering. The filters are glass with a multi-layer coating, and can withstand high temperatures. Color temperature control filters (CTC filters). These are “modifier” filters that are sometimes needed when mixed sources are being used. The color temperature of the discharge lamps used in moving lights is much higher than that of tungsten halogen lamps. This may not be a problem in a live theatrical event, but can be a problem if the event is being televised or filmed. The TV camera or film “sees” the difference in color temperature. Using CTC filters to raise or lower color temperatures is a very common requirement (see Section 1.10.4).
When it is required to achieve more than one color from a luminaire or set of luminaires, there is a need to select a color. A simple and practical way of achieving this is to fit the (usually fixed) luminaire with a color scroller. This device has a series of color filters joined together by high temperature adhesive tape into one long roll, providing as many as 20 colors or more. The color roll is motor driven. Scrollers are available in many sizes. They are the only practical way of achieving color change on floodlights. A scroller is normally a DMX controlled device, with different DMX values calling up selected colors. The scroller itself has its own electronic assembly to detect the position of the different colors – for example by simple shaft encoding or by detecting the filter joins. As mentioned in Section 11.5.4, the generation of the color scroller DMX commands has to be done in such a way as to avoid unnecessary scrolling, and to permit blind selection of color (so that the scroll movement is not seen, unless it is wanted as a special effect). Alternative methods of achieving color filter changes are by the use of semaphore color filters, that are used on follow spots, and by color wheels. Color wheels provide multiple circular filters, but because such filters must be small to avoid the wheel becoming impractically large, they are only suitable for focusing spotlights.
Figure 11.26 A color scroller suitable for PARcan lights and small floodlights. The Forerunner from Wybron. It is DMX compatible.
371
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
in practice the subtractive principle does limit light output, and some colors prove impossible to achieve with any useful intensity. For this reason moving lights are often fitted with both a CMY filter set, and a separate “spot color” wheel to achieve particular colors with high intensity.
11.7 Large scale entertainment lighting control
Figure 11.27 The Martin Roboscan CMYR has both CMY color mixing and two color wheels, augmented by two color temperature filters and a UV filter. The black dots show the individual colors that can be achieved using the color wheel dichroic filters and other filters in combination. All colors within the pentangle can be achieved using the CMY subtractive color mixing.
It is possible to achieve a wide range of colors by using the principle of color mixing. Cycloramas are often illuminated by multiple three color floodlights based on the use of the primary colors, red, green and blue. Such additive color mixing allows a wide range of colors to be achieved by simple dimmer control. The same principle is used with cold cathode lighting for ceiling coves, and has been used with fluorescent lighting for exterior floodlighting. It is now also being used in LED luminaires. The actual color gamut attainable in each case does depend on the wavelengths available from the sources concerned. In moving lights the principle of subtractive color mixing (see Section 1.10) is often used to provide a continuous choice of color selection. Cyan, Yellow and Magenta filters can be introduced in varying proportions to achieve many different colors. However,
372
Spectacular entertainment can involve many hundreds or even some thousands of luminaires. It then becomes necessary to decide how to break down the task of control and minimize the risk of system failure. Large systems often use several lighting consoles, and these may each be chosen to optimize particular tasks. The most common division is to use one console for conventional lights, and a separate console for moving lights. VariLite (who can fairly claim to have started the moving light business in entertainment) have long had their own consoles that are rented out as part of a complete moving light package. The Whole Hog consoles (now part of the High
Figure 11.28 Whole Hog consoles, now part of the High End Systems product range, are popular for moving light programming.
STAGE AND ENTERTAINMENT LIGHTING CONTROL SYSTEMS
Figure 11.29 The Virtuoso™ DX control console can be hired from Vari-Lite Europe Ltd. While optimized for VARI*LITE® moving lights, it is also suitable for the control of conventional luminaires and the control of moving lights from other manufacturers. Notice the playback simulation monitor.
End Systems range) are also popular as dedicated moving light consoles. Such an arrangement makes sense as part of a division of labor. It can actually be impractical to program a very large show on a single console if there is limited set up time. If the event is “live”, it may also be impossible for a single operator to run the show. In some cases additional consoles are used for special feature lighting – becoming a “system within a system”. For cues requiring precise synchronization between different consoles, it is possible to use vari-
ous methods of linking control. Most consoles accept MIDI Show commands and/or SMPTE timecode. The major manufacturers also endow their consoles with ethernet LAN facilities. Generally any use of this for linking is proprietary to the manufacturer. In theory an open system will become a reality when ACN (see Section 9.5.6.4) is operational. In the meanwhile some manufacturers are using Art-Net, a public domain protocol available from Artistic License, as a means of linking lighting control subsystems.
373
C h a p t e r 12
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Architectural lighting control systems
12.1 “Setting the scene” Some aspects of architectural lighting control, especially those related to general energy management, are dealt with in Chapter 14. With the exception of practical applications like dimmers for auditorium lighting, lighting control using dimmers was not common outside of the theater until the development of the electronic dimmer. The arrival of the thyristor and the triac then led to two main streams of architectural lighting control: • low cost, directly controlled dimmers made for the consumer market. Here the driver was (and still is) price. • remote control dimmers and dimmer systems of significantly higher cost and complexity for public buildings etc. The low cost units were and are intended only for individual circuit use, such as on a switchplate for living room lighting or within a table lamp. Here linearity of control is not important, and some lack of reliability can be tolerated. Building owners and operators, especially those running 24 hour a day operations like hotels can not accept any unreliability. They generally need remote or automatic control, and also need consistency between units. As the power control devices have become much more robust (in principle overcoming the problem that the failure of a filament in a tungsten lamp could easily destroy a triac even if a fuse was in circuit) and as low cost microprocessors have come on the market, the distinction between markets is not so clear cut. While the extremes still exist, a whole spectrum of middle ground products has emerged. Many of them are entirely suitable for both “high end consumer” and “professional building” use.
374
The spur to using somewhat more complex lighting controls in architectural applications came from both the enthusiasm of certain classes of user, and from the emergence of lighting design as a recognized profession. Lighting designers tend to have, at least in part, an interest in the theater and so it is not surprising that the early lighting designers brought many familiar theater ideas into architectural schemes. One of the first things they did was to import the idea of “scenes” introduced in Section 11.2. An easy example is the dining area of a first class hotel. This could have twenty or more lighting circuits, with each lamp type making a contribution to both the functionality and the appearance of the space. With each circuit on a dimmer, the lighting designer is able to balance the lighting and to change its mood according to the time of day. So at breakfast, for example, the space is bright and cheery. Maybe at lunchtime there is some additional or changed feature, such as a buffet, that requires a change to the lighting. Then in early evening the lighting might
Figure 12.1 In the 1960s and 70s, and even into the 80s, some architectural dimming systems used multiple preset panels fitted with masses of slider or rotary potentiometers for presetting. The photo shows a 36 channel 10 scene control for a hotel installed around 1980. The top section is a time clock.
ARCHITECTURAL LIGHTING CONTROL SYSTEMS
change again to something more intimate for supper, and, later still, change again for a late dinner “scene”. What then must not be forgotten is the need for a “cleaning” scene for use during the night. This may sound complex, and such a system certainly represents an investment. However early adopters quickly found significant benefits: • by making the space look different at different times of day it becomes more attractive. People staying at a hotel may be happy with a bright restaurant for breakfast, but be less inclined to enter such a place for a quiet dinner. Lighting transforms the feel of spaces, and control is the key to frequent transformations. Lighting control is, therefore, helping to bring in customers and brings a direct economic benefit. • in this kind of application the principal lightsource may be tungsten halogen lamps. Lamp life is significantly enhanced by using dimmers, both to slightly under-run the lamps and to ensure “soft” application of power. In installations like hotels that may have thousands of lamps, this is a highly significant economy. • energy use is optimized by ensuring that only the needed level of lighting is provided. Early architectural “scene setting” systems followed their theatrical counterparts exactly. However, in this case there was no one to operate the “console”, so if, for example, six scenes were required, it was necessary to have six sets of controls; a “six preset” control instead of the simple two preset shown in Figure 11.4. The arrangement was that the preset panel itself would be safely locked away and the end user only had access to the master controls. The actual task of setting the scenes in the first place was the job of the lighting designer. Dimmer and control systems for architectural lighting have much in common with their theatrical and entertainment counterparts. However, there are two important differences: • entertainment systems are designed on the basis that the operator of the system knows something about the disposition of the lights being controlled and has some experience in operating lighting controls. Architectural systems must, on the other hand, be entirely intuitive in operation. Controls must be
kept to an absolute minimum, and their operation should always produce a pleasing result. • dimmers made for entertainment purposes can have fast switching times. In other words the minimum fade time is as near instantaneous as possible. This is required especially for entertainment spaces, dance floors etc. However, this arrangement can result in nuisance tripping of circuit breakers, especially at times of lamp failure. Architectural dimmers may have deliberately lengthened minimum fade times (e.g.300ms) in order to eliminate nuisance tripping and to extend lamp life. Indeed the best products are effectively short circuit proof since they have built in current sensing that shuts the dimmer down under fault conditions.
12.2 Manual versus automatic control One thing that was discovered fairly quickly is that head waiters and similar operational staff do not, on the whole, make sensitive users of lighting controls. They may be asked to change a lighting scene at a set time but, even if they remember to do it, they may well move the control very quickly which can be highly disconcerting to diners (or conference attendees, shoppers et al). The other problem with simple manual control systems, whether working a single channel or acting as a master, is that in the basic arrangement there can only be one control point. If more than one control point is required, a transfer switch is needed to transfer control from one panel to another. (Although the new digital protocols can get over this problem, see for example Section 9.4.2.3.) The automatic dimmer or dimming system solves both problems. Here the master control operates over a preset fade speed, and scenes can be selected by push buttons. This allows multiple control points which is an essential feature for all but the smallest architectural lighting controls. Further advantages of automatic dimming systems include the following: • it is easy to add automatic time and calendar control.
375
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
• similarly it is easy to have additional control inputs, such as presence detection and daylight detection. Both these points are developed in more detail in Chapter 14. • installations where the room layout changes, especially conference centers and banqueting suites, can be programmed to change their lighting control configuration automatically.
12.3 Single channel control Outside the residential market most architectural lighting control schemes are multi-channel. However, there remain some applications where the use of single channel control makes sense. An example is a single use space such as a cinema. Many cinemas are comparatively small, especially those in “multiplexes”, and their lighting control requirements are very simple – often needing no more than a couple of automatic dimmers, one for the main auditorium space, and the other for screen or other decorative lighting. In this case robust single channel dimmers, suitable for both push button operation and for control by the cinema automation system are quite sufficient for the task.
Single channel automatic dimmers are offered with varying degrees of complexity. Most allow for the remote selection of “full”, “off” and two or three intermediate “preset” levels. The actual levels and the dimming speed are set by controls on the dimmer itself. Some users, for example in lecture theaters, prefer a facility whereby lighting is raised or lowered by pressing “raise” and “lower” buttons, with the lights fading up or down as long as the button is kept pressed. Such an arrangement can be provided instead of or in addition to the preset level selection. It might be thought that single channel dimming control would be applicable to applications like individual office lighting. While it may well be the case that the user perceives the control in this way, such systems are usually multi-channel, or embody some additional control input, because of energy management requirements. Large single loads can be dealt with using a master automatic dimmer with slave dimmers, the slave dimmers themselves also being suitable for use as remote controlled manual dimmers based on 0–10V analog control. While it is technically possible to use dimmers for controlling fluorescent lamps (using “dimmable” ballasts) it is now better to use “controllable” bal-
Figure 12.2 Examples of single channel dimmers used for auditorium lighting control. The ones illustrated are available in 16A and 25A ratings. The one on the left uses simple remote control push button panels, and the dimming speeds and preset levels are set by preset potentiometers. The one on the right accepts a serial digital control signal (DALI protocol) and programming of levels and speeds can be done with a cordless hand control.
376
ARCHITECTURAL LIGHTING CONTROL SYSTEMS
lasts using separate analog or digital control lines. In this case suitable single channel manual or automatic “ballast controllers” are used. The ballast controller should include an arrangement that provides circuit isolation in the “off” position. Note that when dimmable ballasts are used it is not a good idea to connect both incandescent and fluorescent sources to the same dimmer. While technically feasible, the dimmer law for each source is so different that the arrangement is unlikely to be useable.
12.4 Small multi-channel control systems A good rule to observe when specifying a lighting control system is to use the simplest arrangement that will carry out the required task. Sometimes the task is actually more complex than at first appears, but even then many projects can be broken down into quite small sections, each of which can be treated as a small system in its own right – even if some supervisory control has to be added. Thus it is true to say that 95% or more of multichannel architectural lighting control systems fall into the “small” category. “Small” is intended to cover installations that might typically use 4–12 channels of lighting, and no more than 24 channels. There are now a number of different ways of configuring such systems, the choice of which is most appropriate is made by considering: • the number and type of separate channels, especially if some of them are not incandescent loads. • the preferred electrical arrangement. This will probably be one where the controlling devices are remote from the dimming devices. • the method of programming, and the ease of modification to the programming. In some applications programming may only be done by a specialist for security reasons; in others ease of access for re-programming could be a “must” and anything that made it difficult would be irritating to the user. • the control facilities required. It is easiest to show how these points can be addressed by considering some examples. In review-
Figure 12.3 A self contained 4-scene 4-channel controller. This example is an entry level product in the Grafik Eye® range from Lutron.
ing the examples, it must be remembered that electrical engineering and contracting practice does vary somewhat from country to country, so some options may not be available everywhere. Figure 12.3 shows a popular unit that in its basic form can control four channels of incandescent lighting. It is complete with triac power control for each channel, and is rated such that each channel can control a maximum of 800W, but the whole device is limited to a maximum of 2,000W to allow it to be installed in a flush wall box and run from a single power circuit. The device allows the selection of four scenes and “off”, and the buttons for selecting these are the only controls normally accessible. Under a flap there are individual channel controls, each consisting of a small array of LEDs to indicate channel level and a pair of Up/Down buttons for each channel. Programming of the system is very simple. A scene button is pressed, and then the channel controls are used in turn to set each channel to the desired level for the scene. The system then memorizes these levels as being associated with the scene. The process is repeated for each scene, and the idea is that the cover is closed and only the scene buttons are used from then on. In this particular case the LED indicators are still visible through the dark cover panel. The example product has the following additional features:
377
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Local control panels
4 Flourescent loads
Master Control 12 Incandescent loads
EIA485
1-10V Control
Dimmers
Ballast Controller
Figure 12.4 A heavy duty system suitable for up to 24 channels, programmable to 16 scenes. The Ambience™ system from Helvar. It uses separate “output units” for fluorescent lamp control.
378
that it fits in a standard 2-gang wall box. The master panel is available with a variable number of scene selection buttons (4, 8 or 16) but in all cases it can program 16 scenes over the 24 channels. Additional push button panels can be connected, and these can select from either four or eight of the available scenes. Like the simpler four channel system, it is a feature of this 24 channel system that all programming of the system can be done on the master panel without the aid of additional equipment. However, in this case the programming does require some special procedures since it must be done with the very limited
Partition B
Space A
1 2 3 4
O
Space B
Partition A
• an IR receiver to allow the use of a cordless IR remote control. • the facility to wire in additional remote scene selection panels. • fixed dimming speeds of 3 seconds from OFF to any scene and between scenes; and 10 seconds from a scene to OFF. • individual levels are set in steps of 5%. The product illustrated is an “entry level” product from the manufacturer concerned; higher level products not only deal with more channels and different methods of electrical distribution, but also give more control possibilities (e.g. 1% level setting and programmable dimming speeds). Figure 12.4 illustrates a more heavyweight approach. In this case the dimmers are mounted as part of the electrical distribution system. Each dimmer channel has its own MCB within the dimmer unit; and dimmer units are offered in blocks of 2, 4 or 12. The 12 way unit is suitable for three-phase operation and channels are rated at 10A. Alternative output units are available that provide MCB protection, relay circuit switching and a 1–10V analog control. These are intended for fluorescent lamp control and “non dim” circuits. Each system can consist of up to 24 channels and requires a master control panel. Thanks to the power of microprocessors this is sufficiently small
Control A Scenes 1 - 8
1 2 3 4
O
Control B Scenes 9 - 16
Figure 12.5 When a space can be divided into two, control must be independent or combined according to whether the partition is closed or open.
Space A
Partition A,B
ARCHITECTURAL LIGHTING CONTROL SYSTEMS
Space B Partition B,C
Space D
1 2 3
Scenes
4
O
Control A 1 2 3 4
1 2 3 4
O
Control B 5 6 7 8
Partition C,D
Partition A,D
1 2 3 4
O
Space C
Control C 9 10 11 12
1 2 3 4
O
Control D 13 14 15 16
Figure 12.6 Many banqueting suites have more than two sections. This diagram shows that a 16-scene controller can only just deal with a space than can divide into four.
number of buttons on the panel. Programming is helped by the compact display, and is based on using four programming buttons in conjunction with the channel buttons. The system can be programmed as follows: • 16 separate scenes. Each scene can control any combination of the 24 channels. • each channel can be set 0–100% in 1% increments. • programmed dimming speed associated with each scene. • possibility of automatic “looping” of scenes. • partition programming. The last two points require further explanation. In some applications, for example retail display and bar or restaurant lighting, there can be a need for simple automatic cycling of lighting. This could be as simple as two dimmers continuously cross fading to animate a display, or a group of dimmers providing a continuous color change effect. While a lighting controller of the type illustrated cannot substitute for an entertainment lighting control, it can be quite suitable for simple architectural or display effect lighting. The “partition problem” is common to hotel banqueting suites and similar places. Figure 12.5 shows
a space that can be divided into two. Clearly when this is the case the lighting control panels in each half must only affect the lighting in their own space; but when the partition is opened, the panels in both halves must now operate all the lighting. In cases like this each room section is usually lit in a similar manner, and the possible scenes are the same for each space; for example: • Conference full • Dining • Conference presentation • Late • Plus an “OFF”control. In the system illustrated it is possible to use the 16 scene capacity in a way that allocates eight scenes to one half, and eight to the other. Each half works independently when the rooms are separated, but control is linked when the space is being used as one. In this example selecting Scene 1 also selects Scene 9; Scene 2 is linked with Scene 10 and so on. Two points arise. First, in exactly the same way that theater lighting controls found the need for groups, here the architectural control does too. In this arrangement scene selection only affects certain channels. In simple scene setting systems only one scene is selected at a time, and the scene controls all channels; but in this example two scenes can be selected at a time over two channel groups. When the room space is divided the scene control of one group cannot affect channels in the other group.
Figure 12.7 A control panel offering a “scene modifying” facility. Here the scene lighting levels are preset, but the user can raise or lower the selected scene without affecting the memorized levels.
379
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 12.8 A small lighting control system using a “distributed” architecture. No central controller is used, and the components of the system may be from different manufacturers. In this hypothetical example the control panels and dimmers are from Helvar, but the fluorescent lamp ballasts are from Philips.
Second, there must be an easy way to invoke the linked control. This can be done using the existing panel controls, but since it is an unobvious command, it is better to have an arrangement that does it automatically. The best method is to fit a microswitch or sensor to the partition itself. (See Section 14.2.3 for details.) The 16 scene 24 channel system illustrated is able to deal with a room space that divides into four; but this is is on its limit since it only allows 4 scenes (including OFF) per space. With this arrangement there is a need for four partition sensor inputs to cater for any possible combination of the spaces. While a system of this kind meets most practical requirements there may be cases where the end user (who will know nothing about how the system was programmed in the first place) finds that none of the pre-programmed scenes gives the light level required. This problem can be eliminated by installing remote controls with scene modifier buttons (Figure 12.7). These do not affect the levels stored in memory, but simply allow the scene in use to be raised or lowered. This is equivalent to providing a manual master control for the group of circuits used for the scene. The “up” or “down” button is pressed
380
for as long as it takes to reach the required level. Both systems described so far rely on some form of central controller, and for load wiring to be channel specific i.e. each dimmer load circuit must be wired back individually. For many installations in the residential and hospitality fields this may not present a problem. However, in the office and commercial field it complicates the initial installation and restricts flexibility in making subsequent changes. Figure 12.8 shows a quite different architecture. Here power is distributed to the lighting in a way that is most convenient to the building main structure, and minimizes the number of separate circuits – for example in an office by linear circuit runs that are reasonably well loaded. Each luminaire (or small group of luminaires) is fitted with a load interface. Controls, such as entry door control panels or light sensors, are sited wherever they are needed. They and the lighting interfaces are linked together by a control signal path. This could theoretically be the mains wiring, but in practice it is more reliable to use a separate signal path. In this case a 2-wire cable is used, and because the data rate involved is very low, the routing of the control cable is not critical.
ARCHITECTURAL LIGHTING CONTROL SYSTEMS
The system shown can have a maximum of 64 nodes on the control line, where a node is either a load interface or a control device. It uses the DALI protocol (described in detail in Section 9.4.2) which includes the following practical features: • no central controller is needed. • each system can have 16 light groups; for example the lighting in a number of separate offices. • lighting can be controlled on an individual channel basis or on a scene basis. • it is possible to use devices from different manufacturers, for example digital electronic ballasts for fluorescent lighting from one manufacturer and control panels from another. • small systems can be programmed using an IR cordless controller, which is also used for control in normal use. Larger systems are programmed using a laptop computer. • control options include manual (rotary or slider) controls, push button controls and sensor control. The three example “small multi-channel” architectural lighting control systems described above are chosen to illustrate the principles involved and the facilities required. There are comparable products of each type made by many manufacturers.
12.5 Large multi-channel control systems Some large scale multi-channel architectural lighting control systems using dimming are often realized by simply using a lot of “small” systems. However, this does not meet some professional needs arising from: • high power loads (say above 10A). • long length control cables, for example a DALI sub system is limited to 300m, 1,000ft, cable length whereas large buildings may well need 1,000m. • the requirement for quick and easy exchange of dimmer modules. Such exchange is rarely required, but in a major public space the failure of a particular lighting circuit may present a real problem so quick exchange may be essential. • the requirement for more complex programming, with many more channels and scenes available. In the previous section it was shown that a “small” 16 scene system quickly runs out of capacity in a multiuse space. Some big banqueting suites may need hundreds of possible space/scene combinations. • the requirement for system monitoring. Such monitoring can include reporting on circuit
KEY DALI 1
EIA485
Router
TCP/IP ETHERNET NETWORK
Control Panels Dimmer Cabinets 12 Router RF-Link or Ethernet Card
Ethernet Hub
Control Panels
Figure 12.9 An example of a large multi-channel architectural lighting control system. This is the Helvar Imagine™ system. It uses three different protocols to link the system components, as described in the main text.
381
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
conditions, equipment temperature etc. and can in some cases be used to report lamp failure. • the requirement for different operating regimes and methods of control. A particular example is dealing with an area that normally works on an automatic scene selection basis, but which at times is also required to work as part of a theatrical system. This may require equipment that can accept multiple control protocols. • the requirement for off-site pre-programming, and fast on-site final programming. • the requirement for simple back up of memory so that in the event of malfunction of the scene memory device it can be exchanged without the need for complete re-programming. While the systems described in Section 12.4 use distributed dimmers or compact multiple dimmer units like those shown in Figure 8.15, the larger professional systems may well use dimmer racks like those in Figure 8.17. Figure 12.9 shows an example of a large scale system. It might use many dimmer racks of different kinds in different locations, and multiple linked controllers. The scene controller (or “lighting router”) is in this case responsible for a maximum of 255 channels. This is not in principle a technical limitation, but a practical one. In an architectural system it is a good idea to divide the system into sections so no one controller is reponsible for too large an area. In the example system each router can provide control signals (both level and dimming speed) for dimmers up to 1,000m away. The scenes stored in the scene controller are selected by remote control panels using serial control. Up to 20 control panels can be connected to each router, and, in order to keep control panel cost to a minimum, the protocol used at this level is as simple as possible (the “DIGIDIM” link shown in the figure is DALI compatible). The low level control bus may also have other devices connected to it, such as sensors. When it is necessary to consider a large number of channels as a single system, the scene controllers can be linked; in this example up to twelve units can operate together, creating a single 3,060 channel sys-
382
tem. In practice the decision as to whether to operate as a single system or multiple smaller systems depends on the application. Scene information is stored in non-volatile memory, in this case flash memory. Provision is made for storing up to 256 scenes per router, plus another 1,024 system wide scenes. This may sound a lot, but scenes are quickly used up when room layouts are complex, or when automatic lighting sequences are required. It is necessary not only to store the scene information, but also configuration information arising from external modifier signals. Modifiers can include: • signals from sensors detecting changes in room layout. • hard control signals providing an over-ride in the case of emergency. • signals from sensors detecting ambient or external lighting conditions. • time signals that change the available lighting scenes at different times of day or year. The example system uses a standard TCP/IP ethernet network to link the routers. This arrangement has the advantage that it is then easy to add a supervisory computer to provide load scheduling and monitoring if required. It also allows the use of wireless ethernet as an aid to programming the system. The network computer(s) also facilitate the creation of removable back-up files, so if a router needs to be replaced settings can be quickly transferred. Whenever a system is installed that relies on a single central controller, it is necessary to consider what happens should that controller fail. A number of possibilities exist, either separately or in combination. One idea is simply to duplicate the controller; installing a “hot stand-by” unit. In the live entertainment world this is frequently done, by using two identical consoles with an output changeover arrangement. In the architectural world it is less common. In a big installation a spare unit should be available that can be exchanged in a matter of minutes - but this does assume that the memory medium can also be exchanged, or that there is a quick procedure for downloading the scene information to the replacement module.
ARCHITECTURAL LIGHTING CONTROL SYSTEMS
Dimmers may be fitted with manual over-ride controls, so that they can be operated directly at the rack position. Often the controller is fitted with a watch dog circuit. This checks that the controller is working by monitoring a part of the circuit that should have constant activity. If activity ceases, the watchdog signal can be used to send all dimmers to a preset level. Another, more common, failure mode to consider is what to do when power fails. The subject of emergency lighting is reviewed in Section 12.8. Dimmer systems may well have additional cabinet wiring and switching devices to allow partial use of the system on an emergency supply. Generally, however, all channels will go off when power fails, and then the question is what happens when power is restored. Most users expect lights to come back on to the state they were in when power failed. This can be achieved if the scene controller has non-volatile memory of its last state. Sometimes considerations of inrush current mean that it is better to restore channels in sequence. Large multi-channel architectural lighting control systems tend to use EIA485 serial control with medium data rates (around 115kb/s) at the device level to allow the use of long control cables. Control protocols are often proprietary, for example in Figure 12.9 the “S-Dim” link from the router to the dimmers; but most systems use a mixture of proprietary and standard protocols. For example DMX is sometimes used as an auxiliary protocol, where part of a system is used with an entertainment lighting console instead of or in addition to the main automatic control. The lighting router shown in Figure 12.9 is able to receive DMX from a standard entertainment console as an additional input. LON™ may be required (but note that in some cases LON hardware is used without certified interoperability) and again a device like the lighting router can be made to interface with LON. Distributed systems using EIB or DALI protocols may also form part of the system, so a product like the router must also be able to service this type of network.
The programming of large systems is done in one of two ways – and this applies just as much to the switching systems described in Section 12.6 as to the dimming systems described in this section. Either dedicated microprocessor based programming devices are used, or a computer is used. (see Figure 12.18). When a computer is used the program usually provides a good graphic user interface. Large building-wide systems may have a permanently installed desktop computer, but most systems are programmed using a laptop computer. When modern lighting control systems first appeared, the personal computer was a mystery to most people and still comparatively expensive. Now computers are regarded as commodity products and everyone has at least some experience of them. For this reason the trend is towards the use of computers for programming any medium to large scale architectural lighting control system.
12.6 Switching systems 12.6.1 Introduction The systems described in Sections 12.4 and 12.5 are based on the idea that the majority of lighting circuits controlled require level control (dimming). The systems do allow for non dim circuits which usually relate to HID lamps. These are catered for either by dedicated relay/contactor units or, where very few circuits are in use, by configuring the dimmer as a switch. Not all dimmers are suitable, and it is essential to check that the dimmer concerned is designed for this duty. Many lighting schemes do not use dimmers, but do require programmed switching. They may use the “scene” concept, but based only on the selection of switched circuits. Such systems are widely used in institutions, schools, public buildings of all kinds, offices etc. The practical method of achieving this switching varies according to local installation practice. Electrical consultants may also strongly influence the method used. The emphasis is on ease of installation while still allowing some flexibility for later changes.
383
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 12.10 Standard USA practice for relay switched lighting control. Standard circuit breaker panels are followed by relay units. In an office environment the main lighting runs from 277V. Diagram from GE.
12.6.2 Examples of USA practice Typical of USA practice is the arrangement shown in Figure 12.10. Here all the protection of the lighting circuits is done using standard MCB panels. The breaker panels are augmented by separate relay units that provide the remote control facility.
The relay unit itself is supplied in a backbox or tub that can be supplied with a choice of hinged front cover suitable for either surface or flush mounting. The tub itself may be in two or three compartments; one or two outer compartments that contain the line voltage side of the relays (and any other line voltage components such as transformers for control power supplies) and a center compartment that contains the low voltage control circuitry. Systems are sometimes delivered to site in a stripped down form, minus their low voltage electronics, to facilitate early first fix wiring when the site may not be clean enough for the installation of electronic equipment. Many, if not most, systems use relays of a type and form factor originally developed by GE. These relays are latching relays, the principle of which is shown in Figure 12.12. On the controlled side there is a single contact set rated at 20A and suitable for line voltages up to 277V (or even 347V). The contacts are designed to withstand both moderately high inrush and high breaking currents to match real life lighting loads, however, the electrical life of the relay is finite. A typical figure is 50,000 operations at 80% load, but if this is likely to represent a problem, higher specification relays are available, or it
Figure 12.11 Relay units are designed to be assembled into a back box on site, or to be factory assembled in advance. Note the separate high voltage compartments. Examples shown here are from GE (left) and Intelligent Lighting Controls Inc (ILC, right).
384
ARCHITECTURAL LIGHTING CONTROL SYSTEMS
Figure 12.12 The principle of the latching relay. A split coil is used, one half for ON, the other for OFF. Pulsed control signals are used to make the armature move in one direction or the other. Latching is maintained by permanent magnets. Diagram from GE.
may be better to ensure all circuits operate at a lower current. The same type of relay is available with an auxiliary contact that can be used for independent status feedback, or for the operation of indicator lamps etc. Most systems restrict lighting circuits to 20A rating to permit standardization in the use of wire sizes. The standard relay is marginal for switching fully loaded circuits of HID lamps at high voltage, so for exterior lighting, warehouse lighting and similar applications compatible contactors may be used. Alternatively the standard relay is used to operate separately mounted conventional contactors with AC coil.
Figure 12.13 The RR relay from GE is the most widely used lighting control relay. The version on the right has auxiliary low voltage contacts. Other manufacturers offer relays with similar form factor.
Figure 12.14 When fully loaded HID lamp circuits are required it is possible to use latching contactors. This one, from ILC, is a 2 pole 20A device rated 480V able to accept a 2,000A inrush current or 1,500A short circuit.
The latching relay is a bistable device. It uses small permanent magnets and a twin coil arrangement. Operation requires only a momentary control signal, typically of a minimum of 50ms. A momentary current in the ON coil causes the relay to switch on, and the condition is maintained by a permanent magnet. If a momentary signal is then applied to the OFF coil, the magnetic circuit is biased in the opposite direction, causing the relay to go off, the condition then being maintained by a second permanent magnet. The use of latching relays brings some advantages: • neglible power is needed for control. • very simple control schemes can be used. • the relay status is maintained regardless of main power failure, so when power is restored any lighting controlled by such relays returns to its status prior to the failure. The latter point can also be a disadvantage if system inrush current is then excessive. The standard relay requires a low voltage supply for control. This is provided by a transformer sited in the high voltage part of the relay cabinet, with the secondary appearing on the low voltage side. The relays are designed to work from DC or rectified AC. An alternative arrangement, originally developed by 3M, is the use of the transformer relay. This
385
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 12.15 A device that does away with the need for separate low voltage power supplies, and requires only two wires for control, is the transformer relay. This example is from ILC.
device does away with the need for a separate control power supply, but does mean the relay takes a small continuous load. Figure 12.15 shows an example. Only two wires are needed for control, which can be exercised by short circuiting the secondary of the transformer with a diode. Depending on the orientation of the diode the secondary biases the magnetic circuit either ON or OFF. The control circuitry prior to the relays is of varying sophistication. The very simplest systems allow individual control of each relay from remote controls that are center-off, momentary action, single pole double throw switches. This simple direct control is often fitted in more complex systems to allow local control override of individual circuits. More usually the relay unit is fitted with a simple programmable I/O (input-output) facility. This permits a number of input controls that are in turn operated by suitable remote control panels, individual push buttons, or momentary switches. The inputs are then each assigned to operate any combination of the output relays. The programming procedure is very simple and can be carried out at the relay cabinet itself (see Figure 12.16). It is easy to implement this arrangement using a microcontroller with non-volatile memory, and as these now have plenty of capacity, it is also easy to
386
add other practical features; examples include: • a “sweep” function that sets all circuits to OFF in sequence. • a “blink” function that switches ON circuits to OFF for a second or so. This is to warn occupants that an OFF sweep is about to take place. If they operate their local override at this time, then the OFF sweep does not affect their circuit. • an interface connection to a building wide control system. At its simplest the building wide system may be no more than a programmable time clock, augmented by daylight sensors. At its most complex it may provide many additional facilities, for example many systems are either solely based on LonWorks™, or offer it as an option. This facilitates integrating the control of lighting with other aspects of building management. Most vendors offer programming software that runs on standard computers. This allows the importation of building schematics and photographs that
Figure 12.16 The GE Softwired Contactor System is representative of simple programmable relay lighting controls. This example has eight inputs and 24 outputs. Holding down a channel (input) button for a few seconds causes its associated LED to flash, as well as the LEDs associated with relays already assigned to it. At this point relays can be added or deleted by pressing the individual relay button. Once the new arrangement has been completed, the channel button is pressed again to record the new arrangement.
ARCHITECTURAL LIGHTING CONTROL SYSTEMS
AREA BUS
RELAY PANEL #1
RELAY PANEL #2
RELAY PANEL #N
LOCAL CONTROL PANELS ON SMALL DATA BUS
GATEWAY (PORT SERVER)
SWITCHING HUB
OTHER AREA BUSES
OTHER COMPUTERS ON NETWORK
TELEPHONE SYSTEM
LOCAL HARD WIRED INPUTS
MONITORING & SCHEDULING COMPUTER CONTROL BY TELEPHONE
Figure 12.17 Block diagram of a typical large lighting switching system. Local control can be hardwired or through a slow speed serial bus (usually proprietary). Area control uses higher speed serial communication (often LON on EIA485). Overall supervision and scheduling may involve connection to a computer network. This allows facilities such as control from the desktop, or control by telephone keypad.
facilitate both the initial programming and subsequent changes. Similar software is offered to monitor system performance and actual usage. This helps with maintenance planning. In offices and similar environments the system may also have a connection to the telephone system. This allows users to control lighting in their immediate area using the touch buttons on their telephone. Figure 12.17 shows a block diagram of a generic large building lighting switching system.
based on the use of contactors. However, the more sophisticated systems are built using distributed controllers. While simple contactors are suitable for large load, single use circuits – for example warehouse lighting, they are not suitable for applications where individual circuits are small loads, such as the lighting for a small office area. In the previous section it was shown that common USA practice requires a
12.6.3 Examples of UK practice The objectives of lighting control are the same both sides of the Atlantic Ocean, so switching systems installed in Europe are similar in principle to those described in the previous section. However, there are differences in detail that arise from different installation practices. Differences also arise because USA practice, while often using 277V for lighting, is based on 115V and its higher current for a given lamp rating. In the UK many switching systems are custom built extensions to MCB distribution boards, and are
Figure 12.18 The programming of large architectural lighting control systems can be done using dedicated controllers. Examples shown here are the Scenemaker™ from Helvar for large dimming systems; and the ProSys™ soft wired clock from GE for their relay switching systems.
387
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
“home run” for each lighting circuit back to a relay unit; and similarly individual override controls also require a control wire back to the relay unit. The alternative approach is to use distributed lighting controllers that help reduce the amount of cabling required. Leading manufacturers in this field are ECS Philips and Delmatic, both of whom base their products on the use of LON™. The aim of both systems is to be able to control lighting down to the individual luminaire level. Figure 12.20 Examples of distributed lighting controllers. An eight luminaire Lighting Control Module used in the Philips ECS Lightmaster 100 system, left, and a ten output control module used in the Delmatic ZMC system, right.
Figure 12.19 Most large lighting control systems are now programmed using computers that are either temporarily connected to the system for programming, or are a permanent component of the finished system. They make full use of graphical user interfaces (GUI) to represent the lighting layout and to facilitate changes. Illustrations from ECS Philips Lighting Control.
388
Figure 12.20 shows some representative controllers. These vary in complexity and the in the number of luminaires they serve, but have the following common features: • as with the USA systems, circuit protection is conventional using MCB distribution boards sited to match the building electrical layout. Each controller is on its own circuit. • each luminaire is fitted with a short flexible tail with multipin plug. With good planning these are factory fitted, so site installation time is greatly reduced. • each controller is equipped with sockets to allow the connection of a number of luminaires; typically six, eight or ten. • within each controller there are a number of relays. This may be as many relays as luminaires, or one relay for every two luminaires. The relays are conventional, since any memory or latching function is carried out within the electronics. • associated with each relay there is provision for a local control input. Inputs can be sensors or simple switches. All inputs operate at ELV. • within each controller there is an electronics unit that provides the local control and connection to the (LON) control bus. It is a feature of building sites that they are not electronics friendly. For this reason standard installation practice is to have a first fix when all conduit, trunking etc. is put in place, and wires are drawn through. A second fix follows later when the site is cleaner and free of the wet trades. This is when
ARCHITECTURAL LIGHTING CONTROL SYSTEMS
luminaires are actually fitted into place. The lighting controllers described here are supplied in such a way that there is an option of single or two fix installation. The back box with mains and control terminations can be supplied in advance. The front panel, carrying the electronic assembly can be plugged in later as part of the second fix. Programming of the controllers may be entirely through the control bus, or include some local hardware programming. This programming includes the identification of emergency lighting circuits (see Section 12.8) and the “power up” rules where there are no local control switches. i.e. whether on application or restoration of power a circuit remains OFF or switches ON. Clearly if there is a local control switch which was ON when power failed, the circuit would come on again when power is restored. The introduction of controllable electronic ballasts for fluorescent lighting has resulted in a demand for systems of this kind to include a dimming option. This is realized by providing a 1–10V control output on the luminaire connector. However, the advent of the DALI protocol may result in a change in the system architecture at the luminaire level. Apart from the local input, the lighting controllers described here are usually under the control of an area controller. This might serve the whole of one floor in an office building, or some other logical grouping. The area controller might provide such facilities as:
Figure 12.21 Systems originally designed with switching in mind have been extended to include dimming. The ECS Wiremaster Plus provides the option of a 1–10V analog control signal for each luminaire. The Delmatic DELi unit provides DALI digital compatibility at the luminaire level. Each DELi unit (which itself is controlled by Delmatic’s LonWorks™ compatible ZMC system) can control 16 DALI compatible luminaires.
Figure 12.22 The ZMC system from Delmatic allows lighting to be controlled from a user’s personal computer or telephone. An example installation is at Globe House, HQ of BAT Industries.
• time control (although this might be provided at a higher level). • group control, especially that related to corridor lighting and other access or exit lighting. When individuals work late they must have an easy method of ensuring they can leave the building without being plunged into darkness. • testing of emergency lighting. For many systems control at the area level may be sufficient in itself. Larger systems link the area controllers together to provide building-wide control and to provide system monitoring. There may also be links to other aspects of the building management.
389
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Type
Description
DIN 46277-1 EN 50.035
Assymmetric rail or “G” profile. The original DIN rail, now not so widely used as the symmetric rails. 32mm. Miniature symmetric rail, or “top hat” profile. 15mm u 5.5mm Symmetric rail, “top hat” profile. Now the most widely used rail. 35mm u 7.5mm Deep top hat. A variant of the standard symmetrical rail with a deeper profile. 35mm u 15mm
DIN 46277-2 EN 50.045 DIN 46277-3 EN 50.022 EN 50.022
Table 12.1 The commonly used DIN mounting rails.
12.6.4 Continental European practice Here there is a greater tendency to the use of custom built control systems. In Germany in particular there are many local “panel builders” who assemble electrical control systems into off-the-shelf cabinets. Lighting control is seen as a simple extension of the electrical distribution system. Many of the components used are designed to fit on “DIN Rail”. DIN is Deutsches Institut für Normung – the German Standards Institute. DIN rail is simply a standardized method of mounting electrical components that “clip on” to the rail. In fact there are several different DIN rails (Table 12.1), most of which, in common with many other DIN originated standards, have been accepted by other standards bodies such as CENELEC (for European Standards, “EN” or EuroNorm) and BSI. Within continental Europe LON™ based systems are used, but EIB based systems are more widespread. EIB now also embraces Batibus (from France) see Section 9.5.4.
12.7 Centralized versus distributed systems Although it may appear that any large building would benefit from a uniform approach to lighting control, there are a number of factors which mean that each installation has to be considered individually. If a building is multi-tenanted, it may or may not be possible to apply a uniform solution across the building. While in theory the remote control systems allow easy re-configuration of systems as room
390
Figure 12.23 Standard symmetrical DIN rail is suitable for mounting terminals and many switching and control components.
layouts change, in practice such re-configuration is only easily carried out if the day to day management of the building is the responsibility of a single management company. Two points that are repeated throughout this book always apply. First, each individual sub-system must be able to work independently, so it is not reliant on central control as such. Clearly if systems were to be wholly dependent on any kind of central control, the failure of this control could be catastrophic. Second, systems should be as simple as possible to meet the requirement. “Might be nice” features should be avoided, since anything that adds complexity to day-to-day operation ends up costing money. In practice this means that the role of any kind of central system is to: • provide broadcast control signals for load shedding and for dealing with emergencies. • provide time control (although this may also be done on an area basis). • monitor the system, and then derive management information from the monitoring. This subject is covered further in Section 14.8. Note that functional monitoring (e.g. detecting occupancy and then ensuring that exit routes are illuminated) should be done on a local basis, although it can be reported centrally. • provide a link to other systems. Some systems provide a local lighting control facility to individuals at their desks through either their telephone keyset, or through their computer. This necessarily involves a gateway from the light-
ARCHITECTURAL LIGHTING CONTROL SYSTEMS
Figure 12.24 An example of a lighting control cabinet built to common continental European practice.
ing control network into either the telephone system or the computer LAN. In such cases there must always be another way of operating the lights, at least on a local area basis.
• assuming that any such emergency lighting or signage is powered by battery or other auxiliary supply, such supply must be able to maintain lighting for a specified period after main power failure. The only exceptions to the requirement are private houses and small workplaces employing less than five people, but the detailed requirements are different in different countries. Standards vary slightly, but are typically as follows: • open areas without any hazard should have an average illuminance of 1 lux. • hazardous areas, for example those with machinery and/or obstructions, should have an illuminance of 10% normal lighting level, but not less than 15 lux. • the provision of the emergency/escape lighting should occur within 0.5s of the failure of the mains supply. • whatever power source is used to sustain the emergency lighting must be able to do so for a period of between one and three hours, depending on type of installation and local regulations. Emergency lighting is defined as being “non maintained” if it is provided by a luminaire that is normally off, but only comes on if the mains power
12.8 Emergency and safety 12.8.1 Introduction The provision of emergency lighting can have a significant impact on the way that lighting control systems are configured. It is, therefore, important to understand the principles involved so that the most suitable configuration is used. Emergency lighting is required to operate whenever there is a mains power failure. The general principle is that in all buildings, including offices, workshops, factories, schools etc.: • escape exits are clearly marked, and there must be enough light for the exit sign to be visible, or in many circumstances, the exit sign itself must be self illuminated. • there must be sufficient lighting to enable people to see their way to an exit.
Figure 12.25 Example of a non-maintained emergency lighting luminaire from IBL, shown with its cover removed. The lamps are 12V 20W tungsten halogen and the batteries are sealed lead acid 14Ah to provide 3 hour operation. The battery charging inverter takes 24 hours to restore full charge, and then keeps the batteries on trickle charge. Power to the lamps is switched by the large relay, and the circuit prevents total discharge of the batteries.
391
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
fails. Figure 12.25 shows an example of a nonmaintained emergency lighting luminaire of a type that is widely used in industrial and service areas. In office and similar environments “maintained” emergency lighting is used. Here some luminaires within the lighting system (or lamps within a multilamp luminaire) are designated for emergency lighting. They operate from the mains when mains power is available, but switch automatically to back-up power whene the mains fails. A variation is the “sustained” lamp or luminaire that derives its power from a battery which itself is on continuous charge – a similar philosophy to that used in the Uninterruptable Power Supplies (UPS) used by computers and similar devices. 12.8.2 Provision of emergency power There is a choice as to how the back-up power is provided. Battery in dedicated luminaire. Here a rechargeable battery with its charger is fitted within a dedicated emergency lighting luminaire, which may be intended for either maintained or non maintained operation. Non maintained luminaires based on the use of tungsten halogen lamps have batteries rated to match the lamps. Most emergency luminaires use fluorescent lamps and are fitted with an inverter ballast. Battery in converted luminaire. In offices and similar environments it can be unsightly to use separate emergency luminaires, and preferred practice is to designate some of the luminaires (or lamps within luminaires) in a lighting scheme for emergency duty. This requires the luminaire concerned to be modified, so that the lamp operates normally from the mains but switches over to local battery/ inverter operation when the mains power fails. Because of the need to operate for up to three hours after mains failure most battery/inverter kits used for this application give a reduced light output, for example 30–40% of normal. This factor must be taken into account when calculating whether the lighting installation will achieve the required minimum illuminance. Battery life is a major issue in any arrangement
392
Figure 12.26 Example of a conversion kit to allow one lamp in a maintained luminaire to be used for emergency lighting. The batteries are in the tube. The electronics unit houses the battery charger, the mains failure detection, 230V inverter and the changeover relay. This unit from Ventilux is intended for use with fluorescent lamps from 7W to 70W fitted with standard electromagnetic or electronic ballasts. The red LED is used to indicate that the battery is trickle charging correctly.
where small batteries are installed in many luminaires. Nickel Cadmium (NiCd) batteries are widely used; but Nickel Metal Hydride (NiMH) is claimed to give longer life. All batteries used for this purpose must meet IEC standards which include the ability to achieve over 500 charge/discharge cycles. In practice a life of six to ten years is achieved by batteries fitted within luminaires. DC central system. Here there are no batteries in the luminaires. Batteries are sited centrally, al-
Figure 12.27 It is possible to combine the inverter/charger unit with the mains voltage lamp ballast in a single housing, as is done in the TridonicPC Combo series.
ARCHITECTURAL LIGHTING CONTROL SYSTEMS
Figure 12.28 Example of a central battery unit with static AC inverter intended for emergency lighting. This particular system is modular, with each module able to support 1,500VA for one hour. Valve regulated lead acid batteries with a design life of 10 years are used. The complete system has a microprocessor based monitoring unit which can form an integral part of the automatic testing system. Photo from IBL.
lowing the use of high capacity, lower cost leadacid batteries. Automobile batteries are not suitable for this duty. DC central systems can directly power halogen lamps, or can power inverters within the luminaires. Electronic ballasts for fluorescent lamps are generally suitable for DC operation, but this does require that the DC supply is of similar voltage to the rms AC voltage of the mains supply. AC central system. Much more common is the use of central batteries fitted with a high power inverter to provide an AC supply. Some lower cost units provide a square wave output; but this is not recommended since there are some loads that can be damaged by a square wave supply. A sine wave output is preferred since then the loads can be completely standard. Central systems have the advantage that they are easy to maintain and test, have a long life, and they can be designed for longer power outages. They have the disadvantage of requiring additional wiring. There is also the possibility that some types of emergency can arise where the centrally located supply is damaged, denying lighting to other areas that need escape lighting. Generator. Another possibility is to have a generator on site that provides an alternative supply.
Such an arrangement is used in hospitals, critical processes and other places where power must be available all the time; and where a substantial proportion of the normal supply must be maintained (as opposed to a small fraction for escape lighting). In practice this arrangement is considered part of the normal supply, since unless the generator is kept running all the time in a standby mode, the minimum time taken to get a generator going is around five seconds. Emergency and escape lighting is most likely to be still treated separately. The detection of mains power failure can be done by a simple watchdog relay or its electronic equivalent. In a non maintained luminaire the relay is permanently held in while power is present. When power fails the relay’s NC contacts switch the battery direct to the halogen lamp(s) or inverter(s). In a maintained system it is necessary to provide a changeover action so that the lamp and its ballast switch over from the normal mains to the alternative supply. If electronic ballasts are being used for fluorescent lamps, there may be some timing rules about how this changeover is made. Under certain circumstances the ballast may falsely detect a tube failure as the new supply is connected, and shut down. 12.8.3 Testing emergency lighting systems The whole point of emergency lighting is that it does, in fact, operate when power fails. In most countries there are strict regulations covering the inspection and testing of emergency lighting; and there is a duty
Figure 12.29 Testing of emergency lighting. (left) A test switch from Ventilux using a key card (token) that can only be operated by authorized users. This arrangement is completely manual. (right) A self test module from IBL that can be added to an emergency luminaire. It carries out the tests automatically.
393
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 12.30 A self testing emergency luminaire from Ventilux. Testing is automatic, but can also be invoked by the red push button. The LED (green in the photo) is twocolor. Steady green means self test is in operation; flashing green indicates lamp failure; flashing red indicates charger/ inverter failure and alternate flashing green/red indicates battery failure.
to keep written records of the tests and any remedial action taken. The requirement is in the form: • frequent check that all maintained emergency luminaires are working. • frequent check of any diagnostic indication (e.g. LEDs) given by individual luminaires or central control equipment.
Figure 12.31 The DATACHECK hand control from Präzisa Industrieelektronik can be used to initiate tests in the company’s self testing luminaires; it can also recover data (time and result) of the last 26 tests carried out by the luminaire.
394
Figure 12.32 The Ventilux NETCOM 5 checks both central battery systems and individual luminaires. A separate control bus (based on CAN) is used, and up to 1,000 luminaires can be connected. The control panel (left) can be connected to a computer or BMS via EIA232. Each luminaire must be equipped with an “intelligent node” (right.) The amber LED indicates unit on test, network idle or luminaire failure; the red LED indicates mains present and, for luminaires with batteries fitted, luminaire charging.
• regular functional check, requiring a short simulated mains failure. This must be done at least monthly, and may be specified to be done weekly. Such a test might last for a period of one to ten minutes. • a six monthly functional check with a simulated mains failure of at least one quarter of the rated time. For three hour rated systems this test is done for one hour. • an annual functional check with a simulated mains failure of the full rated time. e.g three hours. A commercial or institutional building may have tens, hundreds or even thousands of luminaires providing emergency/escape lighting. Meeting the test requirements would be very tedious if each test was carried out manually on the individual luminaires. Some of the methods used to simplify the procedure are as follows: Simple group testing. Here the supply to a group of luminaires (e.g. exit signs) is interrupted for a preset time. During the test the luminaires are inspected for correct operation. Self testing luminaires. Here each luminaire is fitted with a microcontroller that carries out the test routines automatically. The controller should be “intelligent” when applied to converted luminaires, and
ARCHITECTURAL LIGHTING CONTROL SYSTEMS
• they can produce the required printed record of tests and their results. • they are much easier to monitor. As pointed out in Section 12.6 the testing of emergency lighting can be carried out as a function within the overall lighting control system. This arrangement is preferred, because tests can then be done when the system “knows” that the area concerned is unoccupied. 12.8.4 Emergency lighting with dimming systems Figure 12.33 The CentralSystem from Beghelli can control and record the testing of up to 1024 luminaires and 10 central battery units. It uses power line carrier for communication, so no extra wiring is necessary. It includes a small strip printer, so printed records are available from the unit itself. Communication with the luminaires is done on a polling basis and includes automatic discovery of devices.
carry out its tests when the lighting is little used (e.g. evenings or weekends). Such self testing luminaires have LED indication of status, so must still be regularly inspected. Cordless tester. Either as an addition to the automatic self tester, or as a means of invoking self testing, it is possible to use an IR control. If this is fitted with bi-directional communication it is possible to “read” data from the tested fitting to assist with record keeping. Central testing. Large systems use central testing. This is done either using a separate data bus, or by using power line carrier signals to avoid the need for additional control wiring. The big advantages of central testing systems are: • they can be linked to the BMS system to ensure that testing is only carried out when areas are unoccupied.
When scene-setting dimming systems are used, the connection of emergency lighting can present some problems. Some considerations are as follows. When a separate mains voltage emergency supply is available, the dimmer or relay cabinet is wired so that designated circuits transfer to the emergency supply in the event of mains failure. It may be necessary for at least part of the control system to be supported by a UPS to ensure that correct control voltages are applied. Where controllable ballasts are used in maintained luminaires, it may be necessary to force a control voltage to give the required level of illumination. In some architectural applications where lighting levels change over a wide range depending on application, it may be necessary to use exit lights that are dimmable, but which still conform to standards. An example is a multi-purpose auditorium where exit lights can produce annoying glare in the dark, and where it is both aesthetically more pleasing and more functional to operate them at a lower level than when the same space is being used with a high lighting level.
395
C h a p t e r 13
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
The merging of architectural and entertainment lighting control systems
13.1 User demands and the influence of designers This short chapter links together some of the themes developed in the previous two chapters, looked at from a technical point of view. Most of the points discussed are developed in the later applications chapters. While there is a very clear difference between the lighting control needs of an individual office and a theater stage, and in general it is true that quite different products are needed for such disparate applications, there are cases where the needs of architectural and entertainment lighting control come together. A wide range of activities now have an element of “show” about them. For example: • hotel banqueting suites get used for product launches or “industrial theater”, in addition to their regular use for meetings, wedding receptions and Rotary Lunches. • shopping malls, and even shops themselves, are now places of entertainment, with special events, fashion shows and automated displays. • pubs and bars may offer a quiet drink Monday to Thursday, but run live entertainment on Friday and Saturday. • museums and visitor centers often include dioramas and settings that include complex dynamic lighting sequences that run continuously or at intervals. Some of these activities are supported by rental companies and event staging specialists, but very often the permanent lighting installation is expected to do double duty. The problem then is that the user expects to get the showbiz result without providing show operating staff. Everything must run automatically.
396
A second problem is that those who design the show elements like to have the programming facilities they are familiar with for live events; they want minimum concession to the needs of unattended operation. In practice it is possible to design lighting control systems to meet both needs. In fact, the control side is not so much of a problem; it is the regular maintenance of the luminaires that often gets neglected, and there is nothing sadder than a once exciting installation suffering from wrongly focused luminaires, burnt out color filters, and incorrect lamp replacement. There are some simple rules that users and their designers should follow when they require dynamic “entertainment” type lighting in permanent installations that operate continuously and without specialist operation.
Figure 13.1 The IBM e-Business Centre in London uses theatrical techniques in an office block environment. In this case DMX is used as the protocol to control both architectural and theatrical lighting, including moving lights. A presenter is equipped with an ultrasonic transponder that allows the lighting control system to know where the presenter is, and to change the lighting in accordance with the progress of the presentation. The lighting control is by a device of the kind illustrated in Figure 13.4. and the transponder system is a Martin “Lighting Director”. Lighting design was by Lighting Architects Group.
THE MERGING OF ARCHITECTURAL AND ENTERTAINMENT LIGHTING CONTROL SYSTEMS
“DMX relay”. Operates 24 independent relays from three frames of DMX.
“DMX Demux”. Available as a 48 or 72 channel unit. Each DMX frame converted to analog output.
Figure 13.2 Examples of “Gateway” products (from Artistic License) that allow a lighting console outputting DMX to interface to non-DMX lighting control systems. Such products can be of assistance for special events.
Cost of ownership. Designers sometimes forget about this. Lighting suitable for a stage production, where operation is a few hours a day, and (it is hoped) a paying audience is directly covering the running cost, may be wholly unaffordable for a fixed installation. Here lamp cost is the big issue. Luminaires using lamps costing $250 each with a life of only 750 hours (just over two months at 12 hours a day operation) become a serious liability. Whereas the use of long life discharge lamps represent an economy for general lighting, their short life equivalents used in moving lights (which have to be on even when no light is being used) represent both a continuous power cost and a heavy consumable cost. Thus “show” lighting incorporated into permanent installation must be considered as a whole and its running cost (including maintenance) properly budgeted. Some of the lower power moving lights are now available with long life lamp options. Theatrical luminaires using tungsten halogen lamps are less of a problem. Long life lamps are available, and life is greatly extended by only slight under-run. In addition they do not consume electricity or use lamp life when light is not required. From a control point of view this leads to the need to include proper power and lamp life management in the control programming. Simplicity. The user rules are the same as those for architectural lighting. Use must be simple and intuitive. This means that control for general lighting should be separated from that of the show ele-
ment in order to ensure that there is no possibility that the general lighting might behave in an unpredictable manner. While this should not happen, unfortunately it can, due to human error. A typical example is where an entertainment control system is used as the main control for a combined architectural and “show” installation; and where a lighting designer or technician comes in to do some quick re-programming either for a special event or as part of a regular “refreshment” of the show element. They can, and do, leave the control system in a state where access to the general lighting is either denied without specialist knowledge of the control, or produces unexpected and inappropriate results.
Figure 13.3 The Strand 510i controller can be used as a stand-alone controller, or as a back up to a console. Used as a playback device, it can play back multiple events in parallel either from simple GO commands or from timecode. When fitted with a keyboard and monitor it can also be used to program shows.
397
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Thus all general lighting should retain the architectural “scene” concept with simple push-button or automatic access. The “show” subsystem should, as far as the user is concerned, appear as a simple extension to this – even if its invocation initiates a complex sequence. Gateways. Just as annoying is the inverse situation. Lighting a presentation within a hotel ballroom may well require detailed access to the normal room lighting system. Often the situation arises where sections of the lighting (for example an individual chandelier, or a group of wall lights) require separate control, or inhibition, for the special event. This can result in the desperate and surreptitious removal of lamps! Obviously a much better solution is for the event lighting operator to have circuit access to the general room lighting. One solution is the provision of a control panel providing extended control of the room for special events. This allows the temporary event access to individual circuit control while not disturbing the normal preset scene controls. An alternative is the use of a gateway that allows temporary control access to the room lighting system. A simple example is the use of dimmer racks with an alternative DMX control input. A temporary console controlling the special event controls both the temporary event lighting and the room lighting. Where DMX dimmers are not in use, but where the lighting director for the event wants all cues to run from a single console, there are other possibilities. There are DMX to analog, and DMX to switched command interfaces available that may provide the
Figure 13.4 The ETC “Expression” is an example of a DMX “recorder”. These devices can record one or more DMX streams in real time and form the basis of automatic show playback systems. Such devices allow lighting designers to create shows on their favorite or most suitable console, and eliminate the problems and cost of using a console as part of the permanent installation.
398
Figure 13.5 The Avolites Azure console provides the functions of a full entertainment console, but is optimized for the playback of multiple pre-programmed sequences in night clubs, discotheques and themed environments.
needed gateway. The point to remember for the designers of the original space is that some form of interface may be needed, and that it should be easily available and well documented (preferably by a short laminated document firmly fastened to the dimmer rack). By providing a control connection in the first place, it can be made electrically isolated and remove any temptation to open up the equipment!
13.2 Automatic lighting control in public shows and public areas Complex lighting schemes involving many scenes and dynamic changes can be controlled by standard architectural lighting control systems, but, especially when moving lights are involved, there are schemes where entertainment control systems are not only
Figure 13.6 Martin offer the 2510 controller to operate a specific package of lighting based on moving lights. It comes with a whole range of pre-programmed scenes and dynamic sequences that can be selected by number.
THE MERGING OF ARCHITECTURAL AND ENTERTAINMENT LIGHTING CONTROL SYSTEMS
the preference of the lighting designer, but also represent the practical solution. Standardization on DMX as the protocol acceptable by both dimmers and moving lights etc. has, however, simplified matters. In many cases it is neither appropriate nor economic to install a full entertainment lighting control console. Non specialist staff cannot operate it, and misuse can cause serious problems. The solution is to use a console, easily hired, for initial programming and a suitable “playback” device for the permenent installation. These can take various forms. The most sophisticated are “cut down” versions of their fully equipped counterparts. These provide full access to all devices and allow simplifed programming. They can store any show that could be stored by the console they are based on.
The simplest are so-called “DMX recorders”. These devices can record the activity on a DMX control line. Thus a show can be created on a sophisticated console, and then played back in real time while the DMX recorder records the control stream(s). Once recorded the device can play back the show as often as required. These playback devices normally allow access to more than one show, and random access to points within a show. For shows linked to sound they are equipped with the facility for synchronization to SMPTE or MIDI timecode. Computer memory is now so inexpensive that the devices can be solid state with no moving parts, based on flash memory. Some units have exchangeable memory modules, and others use hard disc drives as the storage medium. An intermediate solution is required by the
Figure 13.7 When Marks & Spencer opened a new store in Fenchurch Street in the City of London, they celebrated the completion of the building with a spectacular lightshow based on both exterior and interior lighting. The exterior lighting used luminaires sited across the road. Overall control was by an Avolites Pearl 2000 console; its link to the remote luminaires was via wireless ethernet, using the Artistic License Art-Net protocol. Lighting design by Darren Parker, lighting equipment by AC Lighting, event production by hotcakes. Photos from Artistic License.
399
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
cause most such lighting is static, control is normally by one of the switching systems already described in Chapter 12. Occasionally, however, there is a requirement for dynamic lighting – in terms of scene change, color change, use of moving lights or any combination of the three. If the requirement is temporary, there is no problem. Experienced rental and staging companies Figure 13.8 Moving lights used for outdoor lighting must either be weatherproof, like the Irideon AR500 from ETC (left) or must use a weatherproof housing, such as the example from Martin (right).
application that may have a skilled operator/ programmer available some of the time; but is normally operated by an unskilled operator. An obvious example is a discotheque or a bar with occasional live entertainment. Here the requirement is for equipment capable of running many preprogrammed sequences by push button selection. Examples of the equipment available to meet these different requirements are shown in the accompanying figures.
13.3 Control of exterior lighting Increasing attention is being given to the lighting of the exteriors of buildings. Because there is usually a limit on the permitted level of illumination, and be-
Figure 13.9 Peter Fink of Art2Architecture designed the color change lighting for the facade of the Art and Design Faculty of Wolverhampton University. Neolec Lighting installed the 98 cold cathode lamps, each 3m long, needed to create the effect. Control is by standard architectural dimmers from iLight.
400
Figure 13.10 The 42 story Entel Tower in Santiago is the tallest building in Chile. It is an example of how color and control are changing the night-time face of public buildings throughout the world. The lighting scheme for the Entel Tower was designed by Focus Lighting Inc of New York, and is based on three main elements. Each face of the tower is illuminated by ETC AR500 architectural fixtures that provide programmable color and intensity. The upper part of the tower has five rings of light, each ring made up of 10 white, 10 red, 10 green and 20 blue metal halide fixtures. Finally, at the top of the tower there are six 7kW xenon searchlights. These can either conduct automatic “sweeps” or adopt fixed positions. The lighting is under the control of an ETC Expression playback unit that carries a number of automatically programmed dynamic sequences. The architectural fixtures receive DMX directly, and the switched HID lamps and the searchlights respond to DMX through appropriate interfaces. The playback unit is equipped with a monitor to show lighting status, and a custom graphics tablet laid out to match the tower lighting. This facilitates the setting up of fixed lighting schemes for seasonal or special events.
THE MERGING OF ARCHITECTURAL AND ENTERTAINMENT LIGHTING CONTROL SYSTEMS
choose appropriate equipment and methods of operation, and have the technical staff to ensure all goes well. Figure 13.7 shows an example of this kind of special event and the spectacular results that can be achieved. But, again, any permanent system requires careful consideration of running cost, suitability of equipment and simplicity of control. Moving lights represent a particular problem. They may not be permitted at all by local planning regulations. If they are, it is easiest to use architectural versions of moving lights within weatherproof housings, see Figure 13.8. Color changing in exterior lighting is generally reserved for buildings related to entertainment, signage, and structures such as bridges and transmission towers. Where the structure outline is de-
lineated, cold cathode lighting has been the standard method, but now LED illuminated lightguides are also being used and these give very good color mixing performance. Control for exterior color and scene change lighting can be conventional provided the equipment is accommodated indoors. If control equipment must be sited outdoors, then attention must be paid to the way it is housed. The temperature inside an unsheltered outdoor weatherproof equipment cabinet could vary in the range -40°C to +65°C, not even taking into account any heat dissipated by the equipment itself. Thus it may be necessary to include a thermostatically controlled cabinet heater and air conditioner within the cabinet to maintain a more benign operating temperature range (e.g. 5°C to 35°C).
Figure 13.11 When Park Avenue Productions produced the Queen’s Jubilee events for British Airways London Eye, they turned to Lighting Designer Paul Cook and Vari-Lite Europe Ltd to realize the complex moving light design. The short installation time, limited rehearsal time and an 8 meter tidal range meant that all lighting cues had to be pre-programmed. This is possible using a simulation program that allows the designer to “see” the complete lighting sequence in advance of installation. The center picture (© Miguel Ribeiro, Vari-Lite Europe Ltd) shows simulation of a lighting cue as created on a “what you see is what you get” or WISIWYG basis. The left and right pictures (© Louise Stickland) shows the London Eye in real life, lit as planned by the simulation, and a close up of the Vari-Lite™ luminaires.
401
C h a p t e r 14
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Energy
management control
14.1 Principles The emphasis on ecological and “green” issues is having a profound effect on how people think about lighting control. Since “artificial” lighting is responsible for using a very large proportion of electricity generated, it is natural to think that much energy, and greenhouse gases, could be saved if unwanted lighting could be switched off automatically. Industry jargon refers to “energy management”, but what is really meant is “energy saving”, whether it be on altruistic grounds or, more usually, on pure monetary grounds where energy saving equals money saving. The principles of energy management as applied to lighting control can be summed up as: • choosing efficient light sources and associated power interface (ballast or transformer). • automatically switching off lighting which is not required, either because there is nobody in the area to use the light, or because there is more than sufficient daylight. • controlling light sources so they give the required level of light and no more. • where appropriate, balancing artificial light and daylight to achieve a required level of lighting. In practice there are a number of factors which make “automatic” lighting control systems, intended as the basis of an energy management or energy saving scheme more difficult to achieve than might be expected. Examples of the problems encountered are: • systems which attempt to balance daylight with artificial light do not always work in the way expected because of the way the eye works. Users disable such systems if they behave in an apparently unacceptable way.
402
and
building
systems
• some lighting source and interface combinations incorporating control do not actually save significant energy when set to a low light output. • individual users can have different preferences, and get upset if they are not allowed for. A “uniform” approach to automatic lighting control may be unacceptable. This point is discussed in more detail in Section 15.3. • arrangements which control lighting automatically in the interests of energy saving may well have negative results in respect of security and personal safety. In practice it is found that very simple methods of control, which work in a wholly predictable manner, are effective and achieve the intended result. However, the more sophisticated the system becomes, the more difficult it is to ensure acceptable results. Much work is currently being done to improve the acceptability of automatic control systems. The basis of energy management in lighting control can be summed up as: • having systems which encourage the actual users to set lighting to meet their requirements, and yet not be wasteful. • incorporating light sensors into the system which can measure the achieved lighting level. The measurement is then used to adjust the level to a preplanned value. This applies in two ways; one is to compensate for the fact that some lightsources have a declining light output during life, so are actually giving too much light (and using too much energy) at the beginning of life. The other is to take account of daylight, so artificial light is not used unnecessarily. • incorporating presence sensors (also known as occupancy sensors) into the system which can de-
ENERGY MANAGEMENT AND BUILDING CONTROL SYSTEMS
tect the presence of human beings; thus ensuring that lighting is switched off in areas where nobody is present. • similarly using timers or programmable controllers to switch off lighting when it is known that it is not going to be required. Modern commercial and public buildings are equipped with Building Management Systems or BMS (USA usage also refers to BAS, Building Automation Systems). Originally these were simply concerned with the functional operation of the heating and air conditioning within the building, but today their function can include lighting control, and functions such as security control and sub-systems supervision. It then becomes possible for the system to take a holistic view of energy use within a building. Figure 14.2 A motor-driven timer switch with integral battery back-up. This one provides 48 settings of 30 minutes over 24 hours. Time settings are made by tappets. Photo from Grässlin.
14.2 Sensors and timers 14.2.1 Timers
Figure 14.1 Time delay switches are used for hall and passageway lighting. This DIN rail mounting unit from Grässlin responds to remote push buttons to switch lights on for an adjustable period from 30 seconds to 20 minutes.
The “time switch” has existed since the start of the electric light era. For lighting control it now exists in many forms which, in ascending order of complexity, can be summarized as follows. The time delay switch. A device which, when actuated, will switch on a circuit for a pre-set time. When this “delay time” expires, the switch automatically turns off. Typical applications include common area and stair lighting in apartment blocks (especially in Europe) although safety considerations now favor a hybrid system also including presence sensors. The programmable time-of-day switch. A device to switch a circuit on or off at preset times of day, for example lighting in public places where daylight is present. The programmable controller, incorporating time control. This may either be a dedicated lighting controller or a general purpose programmable logic controller or PLC. Such devices have
403
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 14.3 A multi-circuit electronic timer from Grässlin. Devices like these provide great flexibility, for example by allowing the control of several independent circuits. Outputs can be programmed to be “pulsed” or “latched”. Events can be programmed to occur daily, weekly or on specific dates over a one year cycle. Automatic adjustment for daylight saving in summer can be included.
additional conditional logic functions in addition to the pure time control. For example the circuits to be switched may be determined not only by time, but by other conditions, such as whether particular doors are open etc. The computer based control system. This is where a large scale lighting system is under some form of central computer control or BMS control, and where the time scheduling of lighting is determined centrally. The computer then uses its own internal real time clock to provide the timing function. Simple time switches and delay timers are of two kinds; motor driven or electronic. The very simplest delay timers, known as mechanical run-back timers use a clockwork mechanism, and the time is set (and the spring “wound up”) by the user turning the timer
404
knob to the required setting. Such devices are widely used in, for example, American hotel bathrooms for operating infra-red lamps. Motor driven timers use a small electric motor to drive a clock mechanism. Programming of such devices is usually by tappets, and is fairly coarse. For example it may only allow time definition to the nearest 30 minutes, and a limited number of on and off times in a 24 hour period. The accuracy of motor driven timers depends on the motor used. Simple timeswitches are based on mains operated synchronous motors, so their accuracy depends on how well the electricity company maintains the frequency at exactly 50 or 60Hz. In general they aim to compensate for any inaccuracy over a 24 hour period. Many motor driven timers have a back up in case of supply failure. This used to consist of a clockwork mechanism, but now the most common arrangement is to use a battery back-up. In this case the motor speed is quartz crystal controlled, so accuracy is typically around ± 5 minutes per year. For many applications the motor driven timer has given way to the electronic timer. When first introduced the electronic timer was not always reliable, and could suffer from interference; but today’s electronic timers are very reliable and are much more flexible in respect of programming possibilities. Indeed they can be so flexible that the programming may not be at all obvious (just like a VCR control!). The more sophisticated electronic timeswitches not only allow very precise time settings, with no practical limit on their number, but may also have additional facilities. For example the control of multiple circuits and the ability to accept sensor inputs for limited logic control. These devices are microprocessor based. An advantage of the electronic timer is that it is easy to provide with back-up in case of supply failure. Most use so little power that a capacitor can be used to provide back-up power which is obviously much more satisfactory than using a battery which will ultimately need replacing. A typical specification is that an electronic timeswitch will maintain correct timing and all programmed settings for at least 72 hours after mains failure.
ENERGY MANAGEMENT AND BUILDING CONTROL SYSTEMS
14.2.2 Calendar, season and latitude adjustment to timers Simple delay timers and 24 hour timeswitches meet many requirements in lighting control. However, many timed functions are dependent on calendar or seasonal influences. These apply just as much to computer and controller based systems as they do to timeswitches and are as follows. Calendar adjustment. The simplest is the sevenday week cycle, requiring, for example, different switching times at weekends. More complex systems provide full calendar control, allowing the preprogramming of weekday holidays etc. Up-to-date computer based sytems can drive themselves round the problems of leap-years, but older systems may need manual adjustment to cope. Seasonal adjustment. Where timeswitches are being used to control street lighting etc., their operating times, to match dawn and dusk, are seasonally dependent. The old electromechanical timeswitches used for the purpose included a solar dial, effectively a cam to move the timing actuator, to change the switching time according to month. Today it is more usual practice to use a light sensor for individual circuits. However, computer based multi-circuit systems must include the seasonal data. Latitude adjustment. The times of dawn and
Figure 14.4 Dedicated lighting controller for selecting lighting “scenes” on a timed basis. This one includes automatic adjustment for dawn and dusk based on latitude. Photo from Helvar.
Figure 14.5 Examples of microswitches with different actuator styles. These can be used as the basis of limit switches. Industrial versions are in heavy duty housings. The left photo shows a plunger actuator. More practical for applications like partition sensing is either the roller actuator (center) or leaf actuator (right). The moving element to be sensed is arranged to slide past the switch, avoiding over-travel.
dusk depend not only on the date in the year, but also on the geographical position of the site. Lighting control systems intended for installation anywhere in the world have a latitude setting to ensure the correct dawn and dusk times are applied. Vendors sometimes refer to an “astronomical clock” (or, on more than one occasion, an “astrological clock”!) as providing this feature. 14.2.3 Mechanical sensors Automatic lighting control systems require a variety of sensors to detect external conditions. The most obvious of these are light sensors and people sensors. Sometimes it is necessary to know the position of mechanical or structural items, the most common examples being those of detecting the opening of doors and the position of movable partitions. Microswitches form the basis of the simplest mechanical sensors. A microswitch is simply a switch using a bi-stable mechanically actuated spring contact. At the point of operation, the “snap” switch action is absolutely positive. Actuation can be by a simple push button or switch toggle, but in sensing operations it is by the direct action of the item to be detected. To avoid smashing the switch by the over-travel of the door etc., the microswitch is fitted with a suitable actuator, and mounted in such a way that the over-travel is not in the direction of the switch itself
405
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
(see examples in Figure 14.5). Heavy duty switches of this kind are referred to as limit switches. For small, well fitted, doors, reed switches are used as sensors, especially for home security applications. These consist of two parts. A sealed relay contact carried on a soft iron wire (or reed) is mounted on the fixed door frame or equivalent. The moving door is fitted with a sealed permanent magnet. When the permanent magnet is close to the reed switch, the switch operates. The operating distance is short, and quite critical, so while the method is ideal for simple doors; it is not suitable for the detection of large partitions etc., where the microswitch method should be used. For lighting control applications it is usually sufficient to use mechanical sensors for simple “on/off” information. The sensor switch can either operate the light directly (as, for example, with a cupboard door light) or can give a signal to a lighting control system; as, for example, detecting the position of partitions in a conference suite. In special display applications there may be a requirement for a sensor to give an output signal which corresponds to the precise position of a moving element. For example the linear position of a sliding panel, or the angle of rotation of a turntable. The signal might be a simple analog signal, where a varying DC voltage represented distance or angle of rotation; or it could be a more complex analog signal, where a varying AC frequency indicated the position; or it could be a digital signal giving an exact numeric value.
0.5
Sensitivity A/W
0.4 0.3 0.2 0.1
200
400
600
800
1000
1200
Wavelength nM
Typical sensitivity curve of “raw” diode.
Sensitivity A/W 0.3 0.2 0.1
200
400
600 800 1000 Wavelength nM
1200
Sensitivity curve when fitted with filter
Characteristics at 25°C, when illuminated with 2,850K tungsten lamp at 1,000 lux. Open circuit voltage Short circuit current Dark current Response range Peak response at
400mV 4μA 100pA 340-720nm 580nm
Figure 14.7 Characteristics of a typical silicon photodiode.
Figure 14.6 A reed switch assembly suitable for detecting door closure. Suitable for security applications and for providing sensing signals to lighting control systems.
406
For this purpose there are many types of transducer (device to convert a physical quantity to an electrical quantity) of varying sophistication used in industrial applications.
ENERGY MANAGEMENT AND BUILDING CONTROL SYSTEMS
14.2.4 Light sensors Prior to the arrival of silicon based electronic components, a number of chemical compounds were used which exhibit a photo-electric characteristic. This can be photo-voltaic, where the device develops a voltage in proportion to the incident light, or photo-conductive where the device has a voltage across it, and the current through it varies according to the incident light. One of these compounds, cadmium sulfide, is still in use as the basis of photoconductive cells. But it is safe to say that today the overwhelming majority of light sensors used in lighting control applications are based on the silicon photo-diode described in Section 2.2.6. The silicon photodiode is usually used in a photovoltaic mode, coupled to an amplifier which then provides either a proportional output, or a signal for operating a relay. The characteristics of a typical device are shown in Figure 14.7. In its standard form it is very sensitive in the infra-red region, but for lighting control purposes it is fitted with filters to match its response more closely to the V(λ) function (see section 1.9). The open circuit voltage and resistive short circuit current vary linearly with the incident light over a very wide range (e.g. from 0.1 lux to 100,000 lux). However, there is a variation in performance with temperature. Over the range -25°C to 90°C the dark current changes by a factor of 1,000, and this has to be taken into account in any practical circuit. The photodiode itself is fairly directional. Typically its response has dropped to 50% of on-axis value at 45° solid angle. In any practical detector a simple diffusing integrator is used to eliminate directional effects (but, of course, some designs are deliberately directional). Many applications are served by simple light sensitive switches where the photodiode controls a relay. A simple adjustment sets the required lighting level for operation. The adjustment can be electrical (i.e. by adjusting a potentiometer control) or mechanical, where a simple shutter arrangement restricts the amount of light reaching the photodiode. Such switches have a differential switching arrange-
Figure 14.8 Light sensitive switch. The NightMatic from Steinel is intended for residential exterior lighting. It can switch 500W inductive or 1,000W resistive load, and switches on at an adjustable “twilight setting” of between 2 and 10 lux. One variant includes a timer so that the light goes out after midnight, but on again in the early morning if it is still dark.
ment to prevent nuisance switching around the setpoint. For example a typical unit might switch on when the ambient light level dropped to 35 lux; but not switch off again until the light level reached 70 lux. Simpler units use a time delay to prevent intermittent switching. Another simple application calls for the unit to produce a varying output. Figure 14.9 shows a sensor used in fluorescent luminaires. It provides a 1–10V analog output to directly control an electronic ballast. 14.2.5 Presence sensors The most common presence sensors used in lighting control are beam detectors and passive detectors.
Figure 14.9 Variable light sensor from Helvar for installation in fluorescent luminaires. Level setting is by a simple rotating shutter; output is 1-10V to match controllable electronic ballasts.
407
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Top view Drain
IR Receiver
A C
B
9mm
IR Transmitter
D
A
Source
+ +
B
Combined receiver + transmitter
C
Reflector
Filter
Element
To5 can
5mm
D
+ + Ground
Circuit
Leads
Figure 14.11 Basic construction of the PIR. Figure 14.10 Principle of the beam detector, and examples of a typical reflex units. These are from Pepperl+Fuchs. The MLV12 (left) has a range of 5m, and the RL28 (right) a range of up to 12m. They use modulated IR, and can withstand high levels of local visible lighting (>10,000 lux of daylight). They are available with an adjustable delay timer.
Beam detectors are photo-electric. In their earliest form a beam of light shone at a photocell, and if the beam was broken by someone passing through it, a relay would operate. In today’s equivalent an infra-red (invisible) beam is used, with a matching infra-red detector. The infra red is generated by LEDs designed for emission in the IR region, and the detection is by silicon photodiodes, which, as pointed out in 14.2.4, are, in their basic form, highly sensitive to near infra-red. Suitable IR optical components can be used to produce a narrow beam at the transmitter, and a narrow acceptance angle at the receiver, ensuring that the system is not affected by other sources of light or IR. Modulated IR may be used to provide further discrimination. Beam detectors can work over quite long distances; outdoor units are available with 100m range, and typical indoor ranges are 10–20m. They are ideal for detecting people passing through a defined entrance area, for example a doorway, especially when
408
multiple beam systems are used. However, unless a complex multiple beam system is installed (which may be practical for a large space) a beam detector is not good for volumetric people detection, and this application is best served by the passive infra-red sensor. This device works on a pyrometric principle. This the measurement of radiation, in this case heat. The underlying principle of the passive infra-red sensor is that the device “looks” at the total radiated heat emanating from the area it is monitoring. If there is no movement in the area it sees a steady heat picture, and produces no output signal. If, however, any object above a certain size, and with a difSection
Top view
Element
Figure 14.12 Typical lens arrangement on a PIR sensor. In this case 16 lenses are shown, all of them focused on the detector element.
ENERGY MANAGEMENT AND BUILDING CONTROL SYSTEMS
(m) 4
30o
20
o
-20
o
3 2 1 -30 o
0 1
2
3
5
4
(m)
Vertical view pattern
(m) 3
50
o
40o 30o
2
20 1
10
o
o
Typical maximum 7m
0 -10
1
-20 2
o
-30o -50o
3
Figure 14.14 X-Y cross section showing how the beams from Figure 14.13 translate into detection zones.
o
1
2
-40 o 3
4
Horizontal view pattern
5 (m)
Figure 14.13 Detection area diagram for the PIR showing a typical beam configuration, corresponding to the 16 lens arrangement. Note that the lenses may have different maximum detection distances.
ferent temperature to its surroundings, moves in the area, the device detects the temperature difference and produces an output signal. Note that it can only detect a moving object; if a “new” object moves into the space, but then stays stock still, it will initially be detected, but once it becomes part of the “still heat picture” it will no longer give rise to a signal. Passive Infra-Red, or PIR sensors are designed specifically for the detection of the human body, and typical parameters quoted are the ability to detect movement of between 0.3 and 2.0 m/sec when the “body” has a temperature difference of at least 2°C from its immediate background. The PIR sensor is constructed in various ways. Figure 14.11 identifies the main components of a typical device. In this case it is housed in a TO-5 transistor can with a transparent top. This includes a filter to filter out visible light. The sensor itself consists of four devices (once again similar to
photodiodes, detecting IR) connected series opposed as in the diagram, and in turn forming part of the gate circuit of a unipolar transistor. Provided all devices “see” the same IR signal, no current flows in the gate circuit. But if the “AC” pair receives more radiation than the “BD” pair, a current flows which is amplified by the transistor. If the sensor simply looked at the whole scene, it would be very difficult for it to sense movement. The solution to the problem is to focus the movement onto the differential detector, so that a moving image of the sensed object causes the difference signal required. The focusing is done by a “fly eye” multiple lens as shown in Figure 14.12. The lens construction has the effect of directing multiple beams onto the sensor, as shown in the detection area diagram Figure 14.13. Taking a cross section through the beams, as in Figure 14.14, shows that the motion detection is done in multiple zones – not uniformly through the space being monitored. However, there are so many beams that for practical purposes it is possible to detect movement throughout the space. The performance of a complete PIR sensor depends on the detection device, the lensing arrangement, and the sophistication of the supporting electronics. Simple indoor sensors have a range of 5–10m.
409
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 14.15 PIR detector unit of a kind used for both security and for lighting control. Photo from The Watt Stopper Inc.
There are other kinds of presence sensors, in particular microwave and ultrasonic, but their use in lighting control is comparatively limited. Ultrasonic sensors use the Doppler Effect for detection. This effect is that of a moving source of sound. A police car with a siren makes a higher pitch sound as it comes towards you, and a lower pitch as it goes away from you. The sound waves are “bunched”, or raised in frequency with an approaching source, and “stretched”, or lowered in frequency for a receding source. In a similar way, if a person moves in an area flooded with ultrasonic waves (sound waves of a frequency out of the human hearing range, typically from 30–200kHz) the reflections
Figure 14.17 Sensor from The Watt Stopper Inc based on dual technology. This one combines ultrasonic and PIR detection.
off the person can be detected as being of a different frequency. Ultrasonic sensors can be useful as people detectors in places like partitioned washrooms, where detection still works with signals reflected off partition walls. Microwave detectors also use the Doppler Effect, but this time with microwave electromagnetic radiation at around 20GHz. They are used as part of security systems, especially in areas of high ambient light and other cases where PIR sensors might be affected by the environment. For some applications (especially those related to security) the use of dual technology sensors is recommended. Combined microwave/PIR and ultrasonic/PIR sensors are available, and these work on the basis that they only give an output if both technologies agree. 14.2.6 The positioning of sensors
Figure 14.16 Typical security luminaire fitted with a PIR detector. The unit is also fitted with a CdS photocell to prevent operation in daylight, and a timer to keep the lamp on for a set time after switch-on. Photo from The Watt Stopper Inc.
410
Sophisticated control systems sometimes have the unfortunate habit of failing for the simplest of reasons. The incorrect placing of sensors in automatic systems can lead to annoyance, inconvenience and even danger. The fitting of mechanical (microswitch etc.) sen-
ENERGY MANAGEMENT AND BUILDING CONTROL SYSTEMS
sors for partition movement and similar applications is a matter of common sense. The main criteria are: • to ensure that the switch cannot be damaged by the moving element being detected. • to ensure that the arrangement has enough tolerance to allow for the in-service “loosening up” of the mechanical elements. Both these points are addressed by the choice of switch actuator and housing. • to place the sensing switch where it cannot be operated other than in the designed manner - either accidentally or deliberately (and yet is still accessible for maintenance). • to have the switching element always on the “fixed” side of the movement, to avoid the need for any flexible cable connections. Some specifiers prefer to use photo-electric, reed switch or other non-mechanical methods of door and partition movement detection; but these tend to need more expert fitting to ensure trouble free operation over a long time. The placing of light sensors is probably the most critical in the sense that this is where most mistakes are made. In lighting control many interior applications are “closed loop”, in the sense that the intention is that the light sensor measures the illumination in the area of light use. Therefore in such cases it is important to ensure that: • the sensor is oriented towards the working area to be illuminated. • it must NOT receive direct light from the luminaire(s) under control, or from the windows – it must only “see” the result of the combination. To this extent it should see as wide an area as practicable. • however, the surface which the sensor “sees” must itself be free of glare induced by either the luminaires or daylight. • the sensor is sited in such a way that it is not possible to interrupt the light reaching it accidentally. This can be a factor in low-ceilinged offices. • if they are fitted with any kind of sensitivity control, sensors must be installed in such a way that any adjustment can be made easily. Some light sensors are installed to monitor outside daylight as part of open loop control systems.
Figure 14.18 A multidirectional sensor from Luxmate. This device uses eight photocells with V(λ) correction and an infra red sensor to report on overall sky luminance and sunlight direction. It forms the basis of open loop control for large buildings in conjunction with a suitable computer program. Such a system can then issue different control signals to various parts of the building according to which direction they face.
Examples are sensors sited on the exterior of buildings. A knowledge of the strength of exterior daylight may well be required to optimize internal lighting, especially in respect of countering the effect of glare. To be effective, multiple sensors may be needed in order to detect the direction of lighting. The effects of daylight are both seasonal and weather dependent. Another example relates to sensors sited outside road tunnels. Here the entrance to the tunnel must be much more brightly lit at entrance and exit when there is bright daylight. At night, both the entrance and exit lighting, and even the main tunnel lighting, can be considerably reduced. One of the problems of external sensors is ensuring that they are not adversely affected by the environment. In extreme cases the sensor can be covered by dust or other pollution that affects its reading. If such problems are anticipated, it is necessary both to have a cleaning maintenance program, and to use a more sophisticated measurement technique. The placing of occupancy sensors for reliable operation within a lighting control system also needs care. Some ground rules here are as follows: • sensors must be chosen to have the correct
411
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
lensing arrangement for the area to be covered. Wall mounting and ceiling mounted devices may have quite different characteristics. Units are available optimized to cover rooms or spaces of different heights. • in places like individual offices, where it is necessary to detect comparatively small movements by the occupier, it is important to place the sensor where it is best able to detect such movements. PIR detectors are available in “short range, small movement” variants, for example a range of only 2m, but the ability to detect a 20cm movement of a human hand. • PIR sensors must not be placed where they could be subject to direct sunlight or other bright sources of light. • PIR sensors should not be placed where a heater lies in the detection range, nor where they are in the airstream of airconditioning equipment. • Glass does not transmit infra-red, so PIR detectors must not be placed behind glass.
Figure 14.19 Examples of sensors with different outputs. On the left a combined photoelectric and PIR sensor with DALI serial digital output from Helvar; on the right a light sensor with analog output intended for daylight monitoring from Luxmate.
14.3 Switching versus dimming, control algorithms
14.2.7 Sensor outputs
14.3.1 Introduction
Sensors are usually low voltage, low current devices. In order to be of use they must be fitted with an output interface to match the application. For switch control, microswitches can switch power directly, and devices like PIR sensors can be fitted with a relay. In both cases it is important to pay attention to the contact rating, and, if frequent operation is envisaged, the contact life. Frequent operation is not an issue if an electronic (triac or similar) output is used. Tungsten lamps have high inrush currents, resulting in the need to specify switch contacts of a much higher rating than the running current. Inductive loads may give rise to high back e.m.f.s, requiring suppression components. Where the sensor is only required to give a low voltage signal, as an input to a control system, this may now be of several different kinds: • volt free contact. • DC voltage (e.g. 0V “off”, 12V “on”). • DC analog voltage, for example corresponding to light level or position (e.g. 0–10V). • serial digital (e.g. LON or DALI).
There are many lighting control applications where dimming control is essential. However, in the case of energy management, there may well be a choice between using simple switching and dimming. The natural, and usually correct, instinct is to always choose the simplest solution to any control problem, but with lighting there are some subtle points which need taking into account. New developments in controllable components are improving the payback times for dimming and widening the options. On the face of it, a properly designed fixed lighting system, not requiring dimming for any practical or esthetic reasons, is best served by simple switching from an energy management point of view. The big savings are made using the correct lamp/ballast combinations and switching off unwanted circuits. The savings from dimming may seem marginal, especially when dimmable or controllable luminaires (and their dimmers or controllers) add to initial cost. However, a more detailed analysis taking into account other cost factors and user acceptability, shows that there is a good case for dimming in a wide range of applications.
412
ENERGY MANAGEMENT AND BUILDING CONTROL SYSTEMS
14.3.2 Practical studies There is always a suspicion that, when claims are made for products or techniques, the promoter is grinding his own axe, or simply working from “gut feel”. It is reasonable to ask if any independent studies have been done to check whether energy saving systems actually do save energy, and whether or not they prove reliable and acceptable to users. It is good, therefore, to be able to report that, over a period of years, several significant studies have been carried out by, for example: • The Building Research Establishment, UK. • The National School of State Public Works, France. • Helsinki University of Technology, Finland. • Lighting Research Center at Rensselaer Polytechnic Institute, Troy, NY. • Building Technologies Department at the Ernest Lawrence Berkeley National Laboratory, University of California, Berkeley CA.
Table 14.1 lists some of the relevant papers produced by these, and other, organizations. Many of the points made in this chapter (and in some of the applications chapters – especially Section 15.3) draw on the conclusions of these and similar studies; but readers needing details of the studies themselves, and the test methods used, should refer to the original papers. The papers cited give references to many additional studies. 14.3.3 Control algorithms Energy is only one aspect of the cost of artificial light. Clearly if saving energy costs puts up other costs, very little is gained. Other factors that have to be considered include the cost of lamp replacement (both the cost of the lamp itself and the labor), the cost of luminaire maintenance, and the cost (in the form of reduced productivity) of inadequate lighting.
Title
Authors
Publication
Summary
A comparison of photo-sensor controlled electronic dimming systems in a small office.
R Mistrick C-H Chen A Bierman D Felts
Journal of the IES Volume 29 No 1. 2000
Comparsion of control options in private offices in an advanced lighting testbed.
JD Jennings FM Rubinstein D DiBartolomeo S Blanc PJ Littlefair
Journal of the IES Volume 29 No 2. 2000.
D Maniccia A Tweed A Bierman W Von Nieda
Journal of the IES Volume 30 No 2. 2001
8 different photo-sensor arrangements are compared. Important conclusions about the need for sensor calibration, correct positioning and wide sensor range are reached. A total of 85,000 ft2 (8500 m 2) of offices in 175 zones measured over a period of 7 months. 5 different control scenarios compared. Recommendations for optimizing the use of photocells in daylight integrated systems. Based on 3 years continuous measurement of the contribution of daylight illuminances. An analysis of the actual occupancy of 158 rooms of 5 different types, showing the cost and energy saving effect of different time outs.
W Von Nieda D Maniccia A Tweed
Journal of the IES Volume 30 No 2. 2001
T Moore DJ Carter AI Slater
Lighting Research & Technology Volume 34 No 3. 2002
Photoelectric control: the effectiveness of techniques to reduce switching frequency. The effects of changing occupancy sensor time-out setting on energy savings, lamp cycling and maintenance costs. An analysis of the energy and cost savings potential of occupancy sensors for commercial lighting systems. A field study of occupant controlled lighting in offices. User attitudes toward occupant controlled office lighting.
Lighting Research & Technology Volume 33 No 1. 2001
A companion study to the above. It demonstrates that there are considerable energy savings to be achieved during normal working hours. Two complementary studies based on 14 open plan office buildings, 400 users. Demonstrates the benefits of providing individual user control. Indicates that many users prefer a low lighting level, pointing the way to additional savings.
Table 14.1 Examples of independent studies into the effectiveness of lighting controls and sensors in saving energy and meeting user requirements. The examples are chosen as being easily accessible, and represent only a small sample of those that have been published.
413
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
“Reduced productivity” may not only be expressed in simple piece-work terms, but also in employee dissatisfaction resulting in staff turnover. The lamp replacement issue becomes complex if the life of the lamp is dependent on the number of switching cycles it is subjected to. The need for more frequent lamp changing due to an inappropriate ballast/switching arrangement can result in costs higher than the energy costs saved. Early studies showed that, if the behavior of automatic lighting control systems was unexpected, and/or did not allow for some local user adjustment, users simply tried to defeat the automatic element – thus eliminating any savings altogether. The good news is that the studies cited in the table confirm that there are valuable energy savings to be made, so the problem becomes one of using control algorithms and system designs that are acceptable to users, and do not introduce other costs. This in practice means minimizing the number of switching cycles, or, where dimming is used, ensuring that level changes are smooth. It is valid to use occupancy detection as a means of switching lights on in a space that is naturally dark, e.g. a store room. However, in any space which is not continually occupied, and where there is some daylight or low level ambient light available, it is best that the switching on of lights is done manually. The automatic switching (or dimming) of lights to off by occupancy detection requires a delay after the space is detected as being vacant in order to eliminate nuisance switching. This delay is typically of the order of 5–20 minutes. It has been found that different applications require different time-outs, for example infrequently occupied spaces can use much shorter time-outs without the likelihood of causing user dissatisfaction. Systems based on daylight linking with photoelectric sensors should use differential switching to avoid unacceptable drops in light level. This means that the lights switch off at a significantly higher level of daylight augmentation than they switch on. If dimming is being used, there is less of a problem because there should be no sudden changes, however it can be advantageous to vary the target level
414
according to whether exterior lighting is increasing or decreasing. All lighting using automatic photoelectric control should incorporate time delay to prevent changes arising from short term daylight changes, like a short burst of sunlight through clouds. 14.3.4 Commissioning A point that many studies are agreed on is that the energy savings actually achieved by lighting control systems can be disappointing if the system is not correctly commissioned. Unfortunately the way that many building contracts are let does not recognize this problem. The electrical contractor thinks it is sufficient simply to install the equipment; the manufacturer has only been asked to supply equipment, not to attend site. When a lighting designer is involved, there is less of a problem. The designer knows that commissioning is part of the process, and will insist on the necessary contractor or manufacturer attendance being included in the lighting offer. With large scale systems involving dimming this procedure (and its cost) is usually accepted; the problem comes with many small systems, for example in offices, where there may be many occupancy sensors. Considering the long term importance of the effectiveness of lighting control in saving energy and providing user satisfaction, the ideal situation is that commissioning is recognized as a separate task that should be carried out in conjunction with the building facilities management. This has the advantage that the people who are going to be responsible for the running of the building and its servicing get a full understanding of how the systems work. It is easy to under-estimate the work needed for effective commissioning, or even to overlook the need for certain tasks to be completed. Some of the tasks that may be needed are listed in Table 14.2. Note that some of them require a light level meter. The commissioning of lots of closed loop photo electrically controlled systems, designed to maintain a particular illuminance level, can present particular difficulties unless the sensors provide a fine adjustment over a wide range. The person making the setting can sometimes partially obstruct the sen-
ENERGY MANAGEMENT AND BUILDING CONTROL SYSTEMS
Item
Procedure
Occupancy sensor.
Check that placement of sensor has unobstructed field of view. Check operation and sensitivity of sensor in relation to expected position of occupants. Adjust time delay if the delay function is in the sensor (as opposed to being in a local control subsystem). If PIR, check sensor is not in an air stream or close to a heat source. Check operation. Set maximum and minimum levels if appropriate. If push button “raise lower” operation is being used, set speed of dimming. If the dimming/switching function is operated through a control system (e.g. access through telephone keypad) the test must be through the entire control chain. If multiple access methods are available (e.g. cordless control and door switchplate) all methods must be tested. (Ideally the system will have a fast response for setting up, that can be slowed down for normal operation.) Check position and orientation of sensor. Adjust position or add partial shielding if necessary. Use light meter to check illuminance on the working surface, adjust sensor to achieve required illuminance, taking care not to obstruct sensor while making adjustment. Set response speed for normal operation (if applicable). Here the sensor monitors external daylight conditions instead of the working area. For small systems the procedure is similar to above. For large systems using multidirectional sensors and computer control, the actual lighting level in each area must be measured and weighted adjustments made to the system. Full commissioning of such systems may require fine tuning over a period of months. It is a legal requirement that all emergency and escape access lighting be regularly tested when the building is in operation. Commissioning of this part of the system, therefore, requires following precisely the procedures laid down by local legislation and by the manufacturers of the emergency lighting equipment fitted. Clearly commissioning also requires detailed observation of the signage luminaires for conformity, and may require light level readings to be taken to ensure that escape access illuminance is sufficient. Where automatic switch-off is controlled by a local sub-system, the necessary scheduling information must be entered. If the local sub-system derives its clock information from a central source, then the whole chain must be tested. Where provision is made for illuminating egress routes in a partially occupied building, all likely exit scenarios must be checked. This should involve “walking” the routes and checking that manual access to lighting control is available. Where the BMS provides monitoring functions or provides linkage between the lighting control and other functions such as security, there will be a need for entering operating data, monitoring programs etc. into the central system and then carrying out comprehensive testing and commissioning. This work is intensive and may extend over several weeks. It cannot be completed until the building has been in full operation for a period. Prestige offices and meeting rooms may use multiple lighting circuits that are controlled on a “scene selection” basis. Here each individual circuit must be tested first. The lighting /interior designer should set the levels applying to the different use scenes, and set the speed of change between scenes. Should be reviewed after a short period of occupation. (E.g. for large atrium or public spaces within a building.) This should always involve a lighting designer, who will first want to check the look and operation of the luminaires (including focusing and fitting of color media if appropriate), and only then will set up lighting levels for different “scenes”. If these are to be subsequently invoked manually or automatically, the complete control chain must be tested. Scenes may include dynamic sequences, and full commisioning may take several days and nights. Check manual or automatic operation as appropriate. Check operation in BMS and/or open loop systems as above. Requires night work.
Circuit on local switch or manual dimmer (single channel). Closed loop photoelectric control.
Open loop photoelectric control. Emergency lighting provision.
Timing control/BMS/BAS.
Monitoring, linking to other systems.
Small multi-scene control. Large multi-scene control.
Exterior lighting.
Table 14.2 Examples of commissioning procedures for lighting control systems in commercial buildings.
sor if it is poorly sited. In this respect large areas may be better served by an open loop system providing a supervisory level of control, while individual areas have a manual override facility. Particular care needs taking in programming large systems based on switching. Not only must the operation of exit route lighting be assured, but also it is important that realistic times for automatic sweep switch-off are entered. In a multi-tenanted building, times may differ for different tenants. A review of the functioning of the lighting control should take
place after the building has been occupied for a period in order to address user concerns, and to ensure that timing data are accurate. 14.3.5 Power reducers There are products on the market known as “power reducers” sold with the claim of saving energy. They are placed at the lighting power distribution point and produce the saving by simply reducing the voltage delivered to the circuit. They are available in
415
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 14.20 In the Intelux range of controllers for HID lighting from Merloni Progetti, the current waveform is controlled to ensure that there there is always some current flowing, and that the peak voltage is maintained. PFC capacitors must be placed prior to the controller. The controller illustrated uses thyristor control, but the latest generation of Intelux controller uses IGBT control.
both electromagnetic (autotransformer) and electronic form. One claim is that they can be applied to existing conventional installations, thus reducing installation costs. There is a problem with the concept since any reduction in line voltage necessarily results in lower light output. An obvious question is “if a lower lighting level is satisfactory, why wasn’t this installed in the first place?” This is particularly relevant since the lighting is likely to operate at a lower efficacy when the voltage is reduced. If a variable light output is required as an operational feature for fluorescent lamp installations, it is better to use controllable ballasts for any new installations. These will save far more energy than would be saved by putting a power reducer on the front of a conventional ballast system. It is now almost certainly the case that in any retrofit situation the replacement of electromagnetic ballasts by elec-
416
tronic ballasts would pay for itself very quickly and produce greater savings than could be produced by power reducers, without compromising the lighting performance. A different argument applies to HID lighting. Here there are two factors that may warrant the introduction of a device that lowers line voltage. The first is that many HID lamps running on electromagnetic ballasts exhibit poor lumen maintenance through life. There is, therefore, an argument for under-running them in the early part of the life cycle and increasing the power as light output decreases. For such a system to work it is necessary that all lamps on a circuit are exchanged at the same time. Systems normally start at full voltage to ensure reliable starting, and then check back to a programmed lower voltage depending on the life state of the lamps. The arrangement has the added advantage that systems of this kind usually present a stabilized output, so line variations do not affect lamp life. The second is that there are applications of HID lighting that do need two levels of light output, and where it is better to lower the voltage to all luminaires, rather than selectively switching, in order to achieve even lighting. One example is large warehouses. Many warehouses must keep a level of lighting on for security, but only need the working level when personnel are in the particular area. Valid savings can, therefore, be made if lighting is reduced to 50% output (the practical minimum). Another example is street lighting, an application described in more detail in Chapter 16. In many areas it was considered permissible to switch street lighting off in the middle of the night because there was little traffic, but now most authorities require lighting throughout the hours of darkness while allowing a reduction in light level “off peak” to save energy and reduce light pollution. The majority of warehouse and street lighting uses electromagnetic ballasts. Any power reduction device in the line must not be affected by the complex load that may include capacitors, and must not itself damage the ballast or lamp. This means that the output waveform of the controller must be matched to the task, or must be sinewave.
ENERGY MANAGEMENT AND BUILDING CONTROL SYSTEMS
14.4 Local versus central control A recurring theme in this book is “keep it simple”. If systems behave in an unpredictable manner, they prove unacceptable to users. Looked at at purely from the point of view of energy saving, it might be thought that centralized control would yield the best results. But for the reasons given in the above section, this is not always true. For applications like public area lighting, exterior lighting etc. central control is fine – but anywhere where people are working requires local control. There is also an advantage in having a diverse system, since this is less likely to result in a wide area blackout due to control malfunction. The best system architecture is based on autonomous local subsystems whose individual performance is not critical. Central control should, therefore, be reserved for “management” instructions for the local subsystems. Such instructions modify the local lighting scenes for such reasons as: • need to reduce overall load. • switchover to emergency power. • end of day switch-off in unoccupied areas (although even this can be done by a local subsystem, with the central system being used to provide a standard clock throughout the site).
14.5 Impact of lighting on HVAC With the exception of the very small amount of visible radiation that actually “escapes” from a building though the windows, all energy used for artificial lighting within a building ends up as heat, however efficient the lightsources (since the visible radiation is itself ultimately absorbed). This means that, whatever the external climate, artificial lighting is heating the building. Airconditioning engineers know this, and take it into account when designing the heating and airconditioning system for a building. In the winter the heat from the lighting can be used to augment other methods of heating.
In the summer the lighting heat simply aggravates the air cooling problem. In energy terms it can take around 40% as much energy again to remove the heat generated by lighting. Therefore the use of lighting control systems to minimize energy use can actually save more energy than the “headline” figure indicated by the lighting source energy alone, by reducing the demand for airconditioning. There is, however, a caveat. Very often the return airpath is through the luminaires to facilitate the removal of heat. The efficacy of fluorescent lamps is highly dependent on lamp wall temperature. Some designs of air return through lamp compartments can be so effective that the tube wall may not reach optimum temperature and efficacy is reduced. Apart from reducing the overall efficiency of the lighting system, this factor can cause a problem when fluorescent lamps are dimmed. At low light output the heat coming from the lamp cannot maintain adequate lampwall temperature and instability can result. The problem is particularly severe on T5 tubes that are designed for higher operating temperatures than T8.
14.6 Power quality Power users pay for electricity under several different tariff schemes. These are designed to help the utility companies predict demand, to maximize the use of their plant, and to discourage users from placing unexpected demands on the supply system. This can mean that the calculation of energy cost savings can be more complex than at first appears. Generally a large user pays a flat fee to allow him access to an agreed maximum demand; he then pays a relatively low rate per kWh. However, if he exceeds the agreed maximum demand, he pays a penal rate for the excess. Quite clearly lighting control can play a significant part in avoiding the penalty zone. Usually this aspect of lighting control is dealt with by the BMS, since this can be organized to take into account the building’s entire usage of electricity. It can then send scene modifying commands to local subsystems to limit the amount of energy used.
417
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Perversely this charging arrangement can, under some circumstances, show a poorer payback for lighting control than would be expected. In prolonged periods of not getting near maximum demand the marginal cost of electricity is artificially low. However, this should not be used as an argument against lighting control for energy saving, rather one for ensuring the use of the right tariff. Complex loads, as may be represented by lighting ballasts, can introduce harmonics on to the electricity supply, or can have a poor displacement power factor. Either of these requires the supply system to have a larger capacity than would otherwise be needed, and can have other serious side effects. Both imply a needless waste of energy, and must be corrected. See Section 6.7 relating to power factor correction, especially in respect of electromagnetic ballasts. See Section 7.2.2. in relation to harmonic reduction and power factor correction in electronic ballasts. See Section 6.1.7.2 for a description of harmonic rejection transformers for three phase supplies.
14.7 Integrated versus separate lighting control The arrival of interoperable bus systems, like LON™, and industry standard protocols like BACnet™, make it superficially attractive to link all aspects of building control such as heating, airconditioning, lighting control and security onto a single network, and to use a single control system. The fact that, for example, a single PIR sensor can do multiple duty, serving the needs of the lighting control, the energy management and the security systems would seem to reduce installation complexity. The truth is somewhat more complicated. It is unfortunately the case that the writing and maintenance of reliable control software intended to work across many disciplines is both difficult and expensive – and costs seem to go up exponentially as the systems get bigger. The special problem of lighting
418
control is that long response times are quite unacceptable. The question of responsibility also arises. Those responsible for lighting are not keen to find that they have some poorly defined responsibility for security, so long term support for systems seems to work better if the different disciplines are kept separate. As already suggested in Section 14.4 above, and in Chapter 12, it is in practice best to keep the lighting control separate, and base it on comparatively small subsystems. For energy management purposes it is then quite acceptable to use gateways to link the lighting control into other systems for operations that do not depend on a fast response time.
14.8 Monitoring systems The monitoring of lighting control systems can bring advantages in operational efficiency. However, monitoring can turn out to be a “might be nice” feature – i.e. one that the engineer fancies, but which in practice the user does not use or derive any benefit from. For this reason its use is confined to large systems where the benefits can be quantified, or where its introduction is an essential element of the building operation. Status monitoring simply reports on whether circuits are active. It is an optional feature of all large architectural lighting control systems, and can easily be added to DALI and similar sub-systems. It is typically used for security purposes, so it is known when parts of a building are in use. Fault monitoring, in particular lamp failure, can be added to the basic status monitoring. This is easy in, for example, DALI compatible ballasts, since the facility should be built in. It is also comparatively easy to include in single lamp dimmer loads. With multi-lamp dimmer loads it is rather more difficult, since it requires that the system is able to detect current changes due to an individual lamp failure. Other faults that can easily be centrally reported are items like overheat or fan failure conditions within dimmer cabinets, or even individual dimmers (or other components). Electrical load monitoring. A common requirement is to be able to measure or calculate the
ENERGY MANAGEMENT AND BUILDING CONTROL SYSTEMS
Figure 14.21 Example of a lighting control monitoring screen. The Helvar Workshop program allows access to all the parameters in a system. Some of them shown here are scene data, system layout, event log and channel log. One example of the use of such a program is on board a large cruise liner. The 24 hour a day operation, frequent changes of crew, frequent changes of time zone, and wide dispersion of thousands of lighting circuits justifies a system that can report lighting status and monitor fault conditions.
electrical load being taken by lighting in order to implement energy management policies. In principle this can be done using the status information, in conjunction with a database carrying details of each circuit load. This approach gives detailed information about different parts of the system, which could not be gained by simply measuring the overall power used (this would imply that all lighting circuits are fed from a separate metered supply; not usually practical).
Lamp life monitoring. Lamp status monitoring can also be used as the basis of measuring lamp usage. The information can be used as a basis for planning lamp replacement. Central monitoring systems that use features available in the lighting control system are not a complete solution to the problem of ensuring continued good system performance. This is ensured by demanding users who notice something amiss, and who have procedures to ensure that it gets fixed quickly.
419
C h a p t e r 15
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Architectural
applications
Part 5 − Applications The last three chapters are devoted to the applications of lighting control. Until the advent of electronic lighting control, the main applications lay in the theater and entertainment. Now commercial, industrial and architectural applications represent a significantly greater market. Often the way in which lighting control makes a contribution is not obvious. Part 5 does not attempt an exhaustive study of each application, but rather gives examples, and points out some of the special considerations which affect each application.
15.1 The home 15.1.1 Introduction Lighting is an indispensable part of the modern home that we all take for granted. For most people simple switch control is adequate and until the advent of electronic dimmers in the 1960s, this was all that was available. Now a combination of aspiration and a desire to be seen to be saving energy has created a considerable demand for sophisticated lighting control in the home. The problem is that the home demands total reliability –or at least a guarantee that the failure of a lighting control element cannot plunge the whole
Figure 15.1 Even small living spaces benefit from the use of dimming controls. Controlling light level by dimming instead of circuit switching eliminates glare and is esthetically more attractive.
420
household into darkness, and yet, with the exception of the very top of the market, the home has the smallest budgets. For this reason the smaller home is best served by stand-alone devices. The failure of an individual device at a weekend is not likely to require the immediate attention of a service engineer; and replacement may well be within the householder’s own capability – or is not expensive. Individual control devices include items such as PIR detectors operating exterior lighting and switchplate dimmers. However, there are now many medium and large size homes and apartments that have multiple lighting circuits in each room, and the “simple” approach then has many practical disadvantages. Sophisticated homeowners now look for lighting control facilities that provide the same functionality as is to be found in the most advanced professionally designed lighting schemes.
Figure 15.2 A plug-in dimmer adaptor for free standing lamps from Lutron.
ARCHITECTURAL APPLICATIONS
15.1.2 Small homes, retrofits Most home occupiers have to take the electrical wiring of their home “as is”. This usually makes no concessions to the needs of lighting control so unless a re-wiring project is undertaken, the only possibility is to insert additional components into the existing system. Fortunately switchplate dimmers can be fitted on to existing switchplates with, at worst, the need to expand the backbox size to allow the fitting of more than one dimmer. For floor standard and other free standing lamps it is possible to fit simple in-line dimmers (see Figure 8.19) or use dimmer socket adaptors, as in Figure 15.2. Very often the choice of dimmer used depends on the esthetics of the panel. Apart from the need to fit the size of backbox prevailing in a particular market, dimmer manufacturers serve a wide variation in taste, from “Scandinavian Cool White” to “British Empire Brass”. (See Figure 8.13.) The addition of only one or two dimmers to dining rooms and living rooms can transform the spaces concerned, and make them much more friendly and pleasing places to be − whether for a bright lunch party or for an intimate dinner; reading the paper or watching late night TV.
Figure 15.3 The Leviton SureSlide™ dimmer is available in several different versions to match different user requirements and different loads. The load versions are incandescent line voltage, electromagnetic transformer, and a version for some nominated dimmable fluorescent ballasts. It is available in five colors. When it has a power switch fitted the slider can be used as a preset; the power switch can be single pole or 3-way, allowing power control from other positions. Illuminated versions are also available.
Figure 15.4 Simple dimmers can interact with each other, especially on three-phase supplies. It is best to wire circuit neutrals individually back to the bus-bar, and not to link them in a way that introduces a common neutral impedance.
Until recently it could be assumed that all domestic lighting in main rooms would be tungsten or tungsten halogen. The arrival of “energy efficient” compact fluorescent lamps intended as direct replacements for filament lamps is causing confusion. The use of such lamps in places where level control is not required, and where the lamp is on for extended hours, is to be encouraged (provided lamps with the right color temperature are chosen) but in general they are not suitable for dimming without additional wiring. Even then, most fluorescent lamps cannot be dimmed to near extinction, so they should not be used in applications like home theater. The different characteristics of magnetic and electronic transformers and ballasts mean that it in all cases other than the line voltage filament lamp it is necessary to check that the dimmer/load combination will work. This information should be available from the dimmer manufacturer or re-seller. A problem that can occur with retrofit multiple dimmer installations is that dimmers interact with each other. This arises from the fact they see the neutral through the load, and disturbances on the
421
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
neutral can cause mistiming of firing the triac. The problem is particularly prevalent on 3-phase mains (not likely to be encountered in a UK domestic installation, but common in the USA) because of unbalanced loads. Sometimes the problem can be cured simply by running separate neutrals.
Master scene programming panel
15.1.3. Larger homes, new installations Anyone planning a new electrical installation for a home should pay attention to how the lighting is to be controlled, and make provision for suitable power and control wiring. System architecture varies greatly; from simple schemes using individual switchplate dimmers up to large multi-channel systems with full remote scene control. The larger schemes almost invariably call on the services of a lighting designer working with an electrical contractor familiar with the installation of professional lighting control equipment; the smaller ones may be realized by a local electrician working to the undocumented wishes of the householder. When one room has several lighting circuits, each requiring its own dimmer, practical use problems quickly emerge. The problems are solved by the provision of a scene setting facility, and the ability to control the lighting from more than one place. How can this be implemented without the expense of a lot of extra wiring, which in some cases may be impractical? For reasons explained in Section 9.5.5 powerline carrier control is not widely used in domestic installations in Europe; however it is widely used in the USA. So there the simplest multi-channel lighting systems in the home use X10 or similar protocols, and require no extra wiring. An example of a range of such equipment is shown in Figure 15.5 (see also Figure 9.36). In practice electrical interference on the powerline can lead to malfunction, so the vendors of such equipment include line filters and other signal enhancing devices in their product range, and examples are shown in Figure 15.6. An alternative approach is to use cordless telecommand equipment, the simplest of which is infra-red (IR). Figure 15.7 shows a range of prod-
422
Dimmer with remote scene selection (also works manually)
Scene selection control
Programmable power outlet
Plug in lamp dimmer module
Figure 15.5 Scene setting control is possible without separate control wiring if power line carrier control is used. Leviton have a range of components including scene selection panels, dimmers that can respond to scene commands and switching devices. Notice the code switches that select device number and house number.
ucts that includes switchplate dimmers which work independently or can respond to a scene selection command. Scene selection can either be from a wall
ARCHITECTURAL APPLICATIONS
Coupler to ensure control signals operate on all three phases.
Blocking coupler to filter external electrical noise (e.g from other appartments in the same builing).
Electrical noise filter for individual appliance.
lighting from an automobile. This is obviously easily included in an all-RF control system, but in principle can be added to other scene setting lighting systems using a suitable interface. This arrangement is useful for country properties where the ability to switch on exterior lighting and a basic “welcome home” scene is useful. While they may well use IR or RF as an auxiliary means of cordless control, the most reliable forms of multichannel control do rely on a separate control cable. At their simplest, such systems still use distributed dimmers, but require a separate control wire, for example the arrangement shown in Figure 15.9. However, better practice is to install the dimmers (and relays if applicable) as part of the electrical distribution system. Such an arrangement has the characteristics of a professional lighting control system, and can embody much more sophisticated programming if required. Because of the introduction of separate control cables, and more complex programming procedures, these systems are usually commissioned by specialists.
Signal strength indicator used by installing contractor.
Figure 15.6 Examples of devices that may be needed to ensure that powerline carrier control systems work satisfactorily. (From Leviton.)
mounted master panel mounted in the same box as the dimmers, or by a cordless hand control. While having the merit of simplicity, such a system is only suitable for single room spaces. For larger spaces, and where multi-space control is required, the practical alternative is the use of RF control. Even here range may be limited to ensure reliability and to comply with regulations (see Figure 9.59) but range can be extended by the installation of repeaters. Figure 15.8 shows examples of RF lighting control products suitable for home lighting control, including “whole home” control. RF is also a convenient way of controlling home
Figure 15.7 The Spacer® System from Lutron is based on switchplate dimmers able to accept scene selection commands by an IR hand control (above) or by a matching wall mounted scene selector (below.)
423
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Tabletop master control
Wall mounted master control
Repeater
Plug-in dimmer for low voltage halogen lamp
Dimmer
• interface to control motorized blinds (shades). • control by telephone interface. Figure 15.11 shows the components of a typical system. A conventional distribution panel is installed, and next to it a cabinet containing dimmers, relays (if applicable) and a central controller. The system is then controlled by as many remote controls as required, providing both simple circuit control and scene selection. Wiring to the remote controls is by a low voltage two-wire cable, and wiring topology is not critical (i.e the control wiring can loop round the control stations, or each can have a “home run”). Manufacturers offer many variations on this theme, with a wide choice of cabinet construction etc. depending on the loads involved. Figures 15.2–15.11 have featured USA products as examples. European regulations mean that some of these product concepts are not available in Europe, however the principles remain the same. Because powerline carrier is difficult to use, systems tend to be divided between simple switchplatedimmer based installations and remote control systems. While the DALI protocol was originally devised for fluorescent lighting, it has great potential as the basis for home lighting control systems. One of its most important attributes is that all intelligence is distributed, so there is no central controller whose failure could result in complete system or subsystem failure.
Figure 15.8 Components from Lutron’s RadioRA® system.
The systems can include many attributes normally associated with the systems used in public buildings. For example: • exterior lighting controlled by combination of photo-electric and astronomical timeclock. • conditional programming using “if..then” commands. • interface to other control systems (e.g in home theater).
424
Figure 15.9 The larger home benefits from the use of remote controlled dimming. Photo from Genlyte.
ARCHITECTURAL APPLICATIONS
Figure 15.10 Leviton’s scene controller in conjunction with their Monet™ or Mural™ dimmers uses a single looped control wire. It must run in conduit with the other high voltage wiring.
Figure 15.11 Components of a remote lighting control system for the home. This example is from Vantage.
425
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 15.12 Remote controlled lighting control system suitable for home use, based on the DALI protocol.
Figure 15.13 This villa in Norway was one of the first houses to use the DALI protocol as the basis of home lighting control. The majority of the light sources are low voltage tungsten halogen. System supplied by Vanpee AS. Photos by Mats Wernberg.
426
Figure 15.12 shows the basis of a possible home lighting control system based on DALI. One (or more) cabinets are installed that combine circuit MCB protection with DALI dimmers and relay switching units. Where fluorescent lighting is installed there are three choices: • install dimmable fluorescent ballasts and use a thyristor dimmer. • use 1–10V controllable ballasts and a DALI converter. The converter would be sited in the central cabinet (or where convenient). • (best) use DALI compatible digital ballasts, whose control terminals are simply looped in with those of other DALI items. Control panels are installed as required throughout the house. DALI control requires a free topology two-wire control bus (using standard electrical installation cable) that can be installed alongside the 220V wiring or in separate conduit according to convenience. One point that needs consideration in home systems using remote dimmers is how to deal with free standing luminaires that are normally plugged in to a wall or floor outlet. Obviously in these systems the outlets must be dedicated to lighting. If conventional mains connectors are used there is the danger that appliances could inadvertently be plugged in to the output of an electronic dimmer which could adversely affect the appliance, the dimmer or both. The outlets need some form of identification, preferably the use of an alternative approved connector.
ARCHITECTURAL APPLICATIONS
is also the “living room”. The “study/internet room” is also the “office”. Good lighting with the appropriate control makes this possible, and the developer who can demonstrate this to potential occupiers gets the sale. Unless the apartment is rented fully furnished, apartment owners and occupiers normally want to express their own identity by the way in which they furnish or decorate the room spaces. Therefore any permanently installed lighting control system must allow at least some flexibility for later modification. This means the ability to add, subtract or relocate luminaires, and possibly modify the control system and its programming. The interior/lighting designer should specify the installation of conduit etc. that allows the later addition of outlets at positions that could be relevant for a modified lighting design. Central control and distribution cabinets should have space to allow later additions. Figure 15.14 When providing dimmed circuits (with remote dimmers) for floor standard lamps and table lamps with plugs, it is best, where permitted, to use a different style of mains connector to prevent appliances being connected to dimmers by mistake.
15.1.4 Apartments It is now common practice to include provision for lighting control within newly built or refurbished prestige apartments and condominiums. Particularly where lighting is concealed or is in the form of multiple ceiling recessed halogen lamps the installation of a comprehensive control system quickly pays for itself. First, by ensuring that all lamps are on dimmer control, lamp life is greatly extended both by “soft start” and by slight under-running. Second, by providing the facility to “balance” the lighting, the occupier is less likely to demand changes arising from lighting being too bright or dim. This point is particularly important if there is a choice of wall and ceiling finishes. In many homes, but particularly in apartments, room spaces must do double duty. The “kitchen” is often also the “dining room”. The “television room”
Figure 15.15 In the USA a popular choice for apartment lighting control is the Lutron Grafik Eye™ system. For this application the basic controller can have built-in low power dimmers or remotely sited dimmers. Systems can be configured for up to 24 control points, 16 scenes, and 48 channels (or “zones”). The same manufacturer’s RadioRA® system is also used in this application when it is anticipated that the lighting layout may be subject to frequent change.
427
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
15.2 Integrated home control systems 15.2.1 Environment and house audio control The remote control lighting systems described in the previous section are in principle able to control other devices besides lighting. An obvious candidate for control is the motorized blind or shade, since this in practice controls an important source of light. Most vendors offer standard interfaces to do this. If blinds can be controlled, why not the airconditioning and the security system? No reason in principle, and some vendors make them a part of their offering. The reasons it is not so widely done are threefold. First, unless the project is “new build”, the cost of adapting existing systems is high in relation to the benefit achieved. Second, very often security systems are backed by specialist firms offering a guaranteed response time, and they may prefer to have “ownership” of the security system so there are no disputes if anything goes wrong with it. And third, it is not always possible to get a single contractor able to cover all three disciplines. To sim-
Figure 15.16 This apartment block, 15-18 Lancaster Gate in London, England, uses a Helvar Digidim™ DALI compatible lighting control system in each apartment. Such a system can be easily extended, and the fact that there is no central controller makes for a robust system. Because the building is a conversion, the apartments are of different sizes. The modular nature of the control system meant that it was easy to tailor the control packages accordingly.
Figure 15.17 Motorized blinds are suitable for remote control and can be controlled independently or as part of lighting “scenes”. Photo of remote blind controller from Lutron.
428
Figure 15.18 The Linn Knekt audio server based multiroom sound system is at the top of this specialist market. Each room has a simple controller able to select from 16 sources and independently control sound in up to 128 rooms. In theory such systems can have an auxiliary serial output that can be used for lighting control, but in practice Linn offer alternative control panels with space for integrating other manufacturers’ lighting controls.
ARCHITECTURAL APPLICATIONS
plify commissioning and to avoid “finger pointing” when service is needed, there are benefits in keeping the systems separate. However, exactly as is the case in systems used on commercial premises, there can be “gateways” between systems if needed. More realistically there can be links between the lighting control and home entertainment systems. A popular arrangement is to have permanently installed loudspeaker systems in each room and a facility whereby a central music source system can be routed to any room. While in any room the occupier can adjust the sound level, and in many cases can select the source. Different rooms can listen to different sources (subject to there being sufficient available). At the very top of this market audio servers are installed that can, as an extreme example, contain 1,400 hours of uncompressed CD quality audio and can simultaneously play eight different stereo programs with an access time of less than one second to any chosen music track. The server is equipped with a high speed CD reader that transfers the user’s CD albums direct on to hard disc. More usually these kind of systems are based on conventional multiple disc CD players and other conventional audio sources (radio, tape etc). Not surprisingly these systems also need a multiplexed control in each room, usually with a display to indicate what program material is playing or to assist user control. It can be argued that the same type of control could also operate the lighting. This does happen when a specialist home systems installer is responsible for both sound and lighting; but again for reasons of responsibility and because the installations may not be carried out simultanously, they are more often separate.
Figure 15.19 Living Control Ltd. offer a multi-room control system, primarily intend for audio but also able to control lighting. Like the manufacturers of residential lighting systems, they must offer their control panels in an assortment of finishes.
or a plasma screen “hanging on the wall”. Where space permits, a projector is used to give images that match the room space; screens in the range 6ft−10ft (1.8m−3m) wide are common. • a “surround sound” system. Home theater buffs look to replicate the movie experience, so good lighting control is essential. Most people simply install suitable lighting for the
15.2.2 Home theater Home theater now represents a significant market. A home theater system is characterized by: • a separate area of the home dedicated to showing movies (or a living room area with a “movies” scene). • a means of showing a big image. At a minimum this is a back projection TV set with a 45 inch screen,
Figure 15.20 Home theater’s large screens require good lighting control to ensure high image contrast. Most users want a different lighting scene for watching movies and watching other programs such as sport. Photo from Barco.
429
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
area and use scene selection to call up the “show” scene − in practice there will be another scene for other uses of the system, such as watching football. Home theater specialists often provide a simple link between the video controls and the lighting control; but in practice this can quickly get very complicated. For this reason top of the line systems use dedicated room controllers of the same type that are used in corporate boardrooms and similar environments. This type of system is described in Section 15.4.1 below; they integrate the control of video, audio and lighting to allow single point control, often using a “touch screen”. 15.2.3 Computer control There is no technical problem whatever in linking any of the sophisticated multichannel home lighting control systems into a computer. The DALI based systems and the top European and USA based systems offer two main options, an EIA232 interface or an ethernet LAN interface. The problem is, having got computer control, what are you going to do with it? Home computer nerds can enjoy themselves by making their favorite computer game flash the lights every time they zap the villain. Real virtuosos can use the internet from
Figure 15.21 A computer is a convenient way of setting up lighting scenes on a big home lighting system. The computer is not needed once programming is complete. However, there are benefits in keeping the lighting control system as a stand-alone system, and systems like those offered by Vantage (see Figure 15.11) are well able to cope with large houses like these. Photo from Vantage.
430
Figure 15.22 Remote control of home lighting via the internet or using a cellphone is quite practical. Photo from Lutron.
the other side of the world to plunge their family into darkness when they are least expecting it. To be serious, however, there are reasons for linking to computer. The principal one is that, for the most sophisticated systems, this may be the way they are programmed in the first place. For example simple DALI systems can be programmed using a hand control, but as soon as multiple rooms are involved, the practical method is by using a laptop computer. However, in these cases, once the programming has been done, the computer is not needed, since the operating information is now stored within the system. There is now, indeed there has been for many years, a call for the “wired home” where every item within it is networked together. The call has come from computer manufacturers desperate to foist their wares on the public rather than from any objective benefit for the majority of homeowners, so the test for the average home-owner is relevance. It may well be relevant for a home worker, student, sports fanatic or hobby scholar to have access to broadband communications, therefore making this facility available in every room in the house may be valid. Whether this is done by CAT-6 cable plant or second generation wireless ethernet (first generation is simply not fast enough for big screen video, but is fine for normal work) is a matter of choice. The ques-
ARCHITECTURAL APPLICATIONS
tion that remains is “does it really confer benefit to put the lighting control through the same system?”. The objective answer must be “very little”, least of all if the lighting thereby becomes a potential victim of all-too-frequent computer “crashes”. However, if a vacationer wants to spoil his holiday by spending time at his laptop looking lovingly at pictures of rooms in his home and fiddling with the lighting and airconditioning from 3,000 miles away, there is no technical barrier to him doing so. In a very large home the problem of system performance monitoring becomes no different from that applicable to commercial installations. If a computer is being used to monitor and schedule other home systems, then clearly it can also carry out these tasks for the lighting as well. This in turn raises another possibility. In the same way that in large buildings such systems are used to initiate load shedding, it is possible to imagine that electricity uitilities may soon take a more sophisticated approach to demand management than they do today. Instead of half California being blacked out because the only way to limit demand is to switch off completely, it is theoretically possible for the utility to simply turn off everyone’s airconditioning while still leaving them with their lights. But would this really mean that every appliance would come fitted with electronics so it became a node in a giant network? Many industry groups with vested interests think so. LonWorks™ have introduced Ma [email protected] for the internet enabled home, Sun Microsystems and 70 partners are sponsoring Open Service Gateway (OSGi), and many other more entertainment oriented networking “standards” are being promoted. More realistically it may well become a requirement that homes have separate controllable supplies for airconditioning and lighting. It would then be realistic for utility companies to use powerline carrier (which in Europe at least is largely reserved for power company use − see Table 9.21) to manage demand. However, in the end, it will be a combination of regulatory and commercial pressures that determines the outcome and the actual method used. In the meanwhile there are many ways of integrating lighting control into other systems for those who
want to. For the cautious, and for the installer who does not want frequent call-outs, it will remain best practice to install a robust, self contained lighting control system not in itself subject to the vagaries of Microsoft’s frequent tinkering with operating systems. By all means then, if there are measurable benefits, allow gateway access to the system for alternative means of control or monitoring.
15.3 The workplace 15.3.1 Workplace spaces For the purposes of this section the “workplace” is taken to mean the modern office or commercial building. It is probably the most important application of lighting control, and is also one where individuals have strong opinions. It is unfortunately the case that some lighting control systems installed in
Figure 15.23 Office lighting control can combine the control of luminaires and window blinds to optimize workplace lighting according to external conditions and user preference. Photos from Luxmate.
431
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 15.24 The private office is an example of an “owned” space. The occupier appreciates local lighting control – but such spaces represent a significant opportunity for energy saving using occupancy sensing.
offices with the best of intentions, i.e. to provide a comfortable working environment and to save energy, have been unsuccessful to the extent that users have disconnected the automatic element, or even the entire lighting control system. Advances in lighting controls, sensors and control algorithms; and a better understanding of the problems, backed by objective studies, mean that there is now no reason why lighting control should not be highly effective in the office environment. But an understanding of the potential problems is essential if mistakes are not to be made. As far as lighting control is concerned, it is useful to divide the workplace into different kinds of spaces. Information papers issued by the Building Research Establishment (UK) have made the following useful distinctions. Owned spaces are small spaces occupied by only one or two people. These could be private offices,
432
or cellular offices within a larger area. In such spaces it is expected that users would require direct control of their lighting. Shared spaces are spaces occupied by a comparatively large number of people, for example open plan offices, workshops and laboratories. Here individuals may have no say in setting lighting conditions, but at best may be able to control the lighting affecting their immediate workplace (or desk). Temporarily owned spaces are spaces like meeting rooms, where the temporary user would expect to be able to control lighting while in the space. Occasionally visited spaces are spaces where visiting users would not necessarily expect to be able to control the lighting, but would expect satisfactory lighting to be there already. Examples are toilets and storerooms. Unowned spaces are spaces like corridors, elevator lobbies and other circulation areas. Again visiting users expect their way to be lit, but do not expect to operate the lighting. Managed spaces are functional public areas, including entrance halls, atria, and staff restaurants. Considerations here are the same as they would be in other public buildings, hotels etc., in that lighting in the areas may be designed for multiple scenes. Individual user/visitors would not expect to control the lighting, but there may well be someone in charge who does. 15.3.2 Factors affecting the effectiveness of workplace lighting control Human factors, and the routine of the particular business concerned, can have a significant effect on the effectiveness of lighting control in the workplace. Some observations which have been made both informally and in formal studies include: • People behave differently if they know that lighting is under central control. For example if they know that their private office lighting will go off if they leave the office, they will not bother to switch it off – on the other hand, if they know it is not centrally controlled, most people are quite careful to turn lights out.
ARCHITECTURAL APPLICATIONS
In addition to the human factors, there are many factors related to the space itself, and to the way in which control is implemented. The most significant of these are:
Figure 15.25 The largest proportion of office space is “shared”. Even here, however, the provision of individual control of task lighting, or even lighting zones, is appreciated by staff and is now practical with digital lighting control.
• People of different ages and visual acuity have different requirements and preferences. Those who have the facility to adjust lighting at their workplace to suit their own requirements both appreciate the facility and make use of it. • The occupancy of owned spaces varies very widely. For sole occupancy the “occupied” time expressed as a percentage of the business day can average as little as 50% or even less, making owned spaces an obvious target for energy management. • The one thing which everyone dislikes is any kind of unpredictability in lighting which is not under their own control. Photo-cell daylight-linked lighting systems are disliked if there are sudden changes of light level. Timed systems which have no obvious local over-ride are equally unpopular. • The human eye cannot deal with very large differences in illuminance, and has a problem with quick changes in lighting level. This can mean that some occasionally visited spaces may need to be lit to a higher level than expected, because at some times of day visitors to them will come from much brighter surroundings than at others. Similarly in systems which use electric lighting to “top up” daylighting, the switching off of the electric lighting when a target level of daylighting is reached can result in a perceived unacceptably low lighting level.
Figure 15.26 Entrance lobbies are examples of managed spaces where lighting can make a statement about the organization. Control is usually needed to match time of day and the season. Lucent use lighting control to great effect at their Nürnberg facility in Germany. Lighting design by, and photos from, DHA Design.
433
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
correct positioning of sensors can yield unsatisfactory results, either unhappy users or failure to achieve expected energy savings. The point is discussed in Section 14.2.6. • The occupancy of the space, i.e whether the space is “owned” or “shared”. This could also be considered a “human” factor. 15.3.3 Recommended practices
Figure 15.27 The control of window blinds is an integral part of workplace lighting control. Modern lighting control systems have interface units able to operate standard blinds. Luxmate® offer alternative interfaces suitable for blinds using 230V AC motors or 24V DC motors (left). A DALI compatible interface for blind control (right).
• The extent to which the space is lit or partially lit by daylight. The main problem is that in most cases the daylighting is not uniform, for example in an open plan office where daylight may be significant near the windows, but less so nearer the building core. • The orientation of the building (when daylight is present). The daylight contribution takes on a quite different nature according to whether the windows are north or south facing, or are affected by early and late sun. In temperate and cold latitudes the latter problem is highly seasonal. • The use of blinds. Wherever workspaces have outside windows it is usually necessary to have some form of window blind. On bright sunny days, or especially when the sun is low in the sky, the glare resulting from daylight can be uncomfortable or even make work impossible. Translucent or “venetian” blinds are the only practical remedy. However, their introduction does affect the way in which any automatic lighting control system works. • The placing of light and presence sensors. In-
434
There are many ways by which lighting can be controlled in the workplace (and in other architectural applications). The most common are as follows. Manual by door. The most obvious, and most widely used. In the simplest possible arrangement this is just a light switch, directly controlling the power to the lamp − but in most systems it is a switch or button sending a signal to a remotely controlled power interface (which may provide dimming or switching control). Flexible manual. This gives greater flexibility in the location of the controls − for example by providing control at desk positions. This would also include the use of cordless remote controllers. Time switching. Where lights are turned on and off automatically under time clock control. Also applicable to dimming systems. Timed off/Manual on. In many cases time switching is uneconomic because it switches on lights where they may not be needed. So this arrangement assumes that lights are turned on manually, but turned off under timeswitch control. Light sensor switching where lighting is controlled by a photocell. Usually to ensure lights are switched off when daylight is present. Light sensor dimming. This is a more sophisticated arrangement where lights are dimmed in response to a photocell signal. Presence detection. Or one could just as well say “absence detection”. The use of people sensors to ensure that lights are switched off when nobody is in the room. Centralized manual. In large buildings it is common to have a central manual control panel operated by, for example, security staff for controlling the lighting of managed spaces.
ARCHITECTURAL APPLICATIONS
Space
Daylit high occupancy
Daylit low occupancy
Non-daylit high occupancy
Non-daylit low occupancy
Owned
Manual at door Flexible manual Timed off/manual on Light sensor dimming Flexible manual Timed off/manual on Light sensor dimming
Manual at door Flexible manual Timed off/manual on Presence detection Flexible manual Timed off/manual on Light sensor dimming Presence detection Local manual Presence detection Flexible manual Timed off/manual on Key control Presence detection Full occupancy link Local manual Timed off/manual on Key control Full occupancy link Presence detection Timed off/manual on Light sensor dimming Light sensor switching Light sensor dimming Time switching Centralized manual Light sensor switching Programmed scene setting Full occupancy link
Manual at door Flexible manual
Manual at door Flexible manual Presence detection
Flexible manual Time switching
Flexible manual Presence detection
Local manual Presence detection
Local manual Presence detection Flexible manual Timed off/manual on Key control Presence detection Full occupancy link Local manual Timed off/manual on Key control Full occupancy link Presence detection Timed off/manual on
Shared
Temporarily owned
Occasionally visited
Local manual Flexible manual Presence detection Timed off/manual on Light sensor dimming Not applicable
Unowned
Light sensor dimming Light sensor switching
Managed
Light sensor dimming Time switching Centralized manual Light sensor switching Programmed scene setting
Not applicable
Time switching Presence detection
Time switching Centralized manual Programmed scene setting
Time switching Centralized manual Programmed scene setting Full occupancy link
Table 15.1 Lighting control for the work space. The green items are “recommended”, the blue items are “assess for the particular installation”. Adapted from BRE Information Paper 2/99 “Photoelectric control of lighting; design, setup and installation issues”. © Building Research Establishment Ltd 1999.
Key control. Some lighting control may only be accessible to qualified keyholders or passcard holders. Full occupancy link. Where a building has a system which accounts for people’s presence in the building (e.g. through a “swipe card” system) then there can be a signal available to indicate whether or not the building or space is occupied. Programmed scene setting. Push button control of multi-circuit lighting to set different “scenes” for different applications. See Section 12.1. Using the above definitions, Table 15.1 gives the recommended method of control for different spaces. Clearly there will be many cases where methods are combined in order to achieve the objectives of efficient, comfortable lighting and best energy saving. Table 15.1 can, with little modification, also be
applied to applications other than office workspaces, for example schools and college premises. For some applications the emphasis might vary. For example under “Temporarily owned” spaces it would be fair to add programmed scene setting as a recommended method if the space is a meeting room used for presentations; but in such a case the rules for conference centers etc. would begin to apply anyway. 15.3.4 Other workplaces The emphasis has, so far, been on office and similar lighting where control can be seen to bring user satisfaction and economic benefits. But workplaces range from offices to factories, warehouses to institutions. Clearly a warehouse has significantly different requirements to an office. Warehouse
435
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
lighting is concerned with safety and with providing sufficient illuminance to carry out the storage and retrieval task. Even here, however, there can be scope for some control. A warehouse or similar building operating 24 hours a day, with all parts of it active − such as a freight operation or electronic parts distribution business, does not benefit from a programmable control system unless there is some daylight available. But many warehouses do not operate “24/7”, and some, such as storage warehouses, only require partial access on any day. Premises of this type benefit from a dual level lighting system providing a low background level for safety and egress, complemented by a controllable higher illuminance system for access to a specific part. In order to have a quick response, such a system would be based on fluorescent (or even, in a very lightly used area, tungsten halogen) lighting. Ideally the system is operated by the warehouse staff using local or cordless controls, augmented by presence sensors to ensure automatic switch-off. Heavy industry uses high bay lighting based on HID lamps where there is limited opportunity for control other than a heavy duty version of the remote controlled relay switching systems described in Chapter 12. Where HID lamps are used for work lighting it must be remembered HID lighting operating at mains frequency (i.e. using electromagnetic ballasts) does exhibit some stroboscopic effect. For this reason neighboring luminaires should be on different phases to minimize the effect overall. Also most HID lighting used in industrial workplaces does not instantly re-strike. This means that if there is a power failure, there is necessarily an interval before lighting can be restored. For this reason some installations have separately controlled auxiliary tungsten halogen lighting for security. Some installations benefit from the use of “power reducers” or dimming devices, such as those described in Section 14.3.5, to provide a lower level of light with corresponding energy saving when full light is not required. In order not to compromise lamp life, these must be of a kind that start lamps at full output and only make level changes slowly. Ideally they should ensure that some current is always flow-
436
ing, and this favors the use of variable autotransformer, IGBT or specialized thyristor equipment (for example that of Figure 14.20). Much modern manufacturing is carried out in clean areas that are lit to the same standards as offices, so the same rules and ideas apply. Clearly, if a critical process is involved it must not be possible for any central system to modify lighting levels without local authorization.
15.4 Meeting rooms, conference centers, and auditoria 15.4.1 Room control A characteristic of modern meeting rooms (and there are said to be 25 million of them in the world!) is that there is more to control than just lighting. Audio-visual systems of varying complexity require control of both the program sources and the display devices. In recent years “traditional” methods of AV presentation such as slide projection and the use of the OHP (overhead projector) have given way to elec-
Figure 15.28 A meeting room with some daylight. Because front projection is being used, the window blind system must be controlled as part of the lighting scene.
ARCHITECTURAL APPLICATIONS
Meetings with visual display. In this scene it is essential that direct light on the front of the screen (or plasma display) is minimized to improve contrast. Rear projection makes good contrast easier to achieve.
Figure 15.29 This photo shows the importance of lighting in videoconferencing. On the left the bright surface behind the participants causes the camera to react in a way that makes the foreground figures dark. On the right some foreground illumination of the participants, especially some reflected light from a light colored table, combined with a moderately dark background, ensures they can be seen properly.
tronic methods. Small rooms are using plasma display panels, and all others are using electronic projectors based on several different technologies such as LCD (Liquid Crystal Display), LCOS (Liquid Crystal On Silicon) and DLP™ (Digital Light Processing, a reflective technology from Texas Instruments). The new generation of projectors produce images that are significantly brighter than those given by the earlier video projectors and by slide projectors. As a result the need for lighting control has changed, but it has definitely not disappeared. Meeting rooms take many forms; a simple training room is like a classroom, a boardroom may be luxuriously furnished, an executive briefing center may be elaborately fitted out with specialist display systems − but all have similar lighting control requirements. Typical lighting scenes that are required areas follows. Meetings without visual display. Good, bright general lighting.
Figure 15.30 Rear projection allows high levels of ambient light to be used, including some daylight, while ensuring a high contrast image. This facilitates participatory meetings (above). However, when a pre-recorded “show” is to be given, a quite different lighting scene is more appropriate (below).
437
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 15.31 Principle of the room controller. Figures 15.28–15.31 are courtesy Electrosonic Ltd.
Videoconferencing (sometimes). Here the participants must be correctly illuminated, see Figure 15.29. Presentation. Here again direct front lighting on the screen must be minimized. Maybe the general lighting is reduced while maintaining sufficient light for note-taking. Additional lighting is provided for the presenter(s). Show. If complete video programs (“films”) are to be run, they have more impact if the lighting level is reduced, see Figure 15.30. A typical meeting room may have anything between two and twenty lighting circuits to achieve the scenes envisaged by the interior/lighting designer. Scenes may be somewhat different according to whether daylight is available, and the control of window blinds becomes an integral part of the scene selection. However, it is clearly difficult for an occasional user of the room to operate a complex lighting system at the same time as operating the AV system. So for all but the simplest rooms it is best to use a room controller. These devices are like special purpose programmable logic controllers. Standard room controllers typically consist of a rack mounting unit that is fitted with a central processor (usually PC based) and with a range of plug-in cards that provide input and output interfacing for all the different kinds of equipment to be found in a meeting room
438
(or home theater, boardroom, briefing center etc.). Inputs can be serial or switch contact; outputs can be relay contact or many kinds of serial data. The whole device can be loaded with drivers corresponding to the serial commands needed to operate many items of standard equipment, including projectors, videocassette recorders, DVD players, video switchers/routers and programmable audio mixers. Control of the room controller can be by push button panel or touch screen. The whole assembly is programmed using a computer program that allows the user to configure the control in any way required. One command might be to select the VCR. This single selection may be required to initiate a complete sequence of commands:
AMX Axcent 3 room controller
AMX AXT touch screen control
Rear view of Creston NET 2 rack room controller
Creston touchscreen control
Figure 15.32 Room control devices from Crestron and AMX.
ARCHITECTURAL APPLICATIONS
Until recently room controllers were stand alone devices. However, versions are now available that include an ethernet LAN facility. In due course room control architecture may well change so that the controller provides a local server function for media – and/or acts as a client device for remote media sources. At present (2003) the networking facility is of most use in multi-room installations. 15.4.2 Conference centers Figure 15.33 A meeting room at Lucent Technologies Center of Excellence, Warren NJ. Here lighting (designed by DHA Design) makes a major contribution to the room’s appearance. Notice the room controller on the table.
• switch on the projector if it is not already on. • close the window blinds if not already closed. • lower a motorized screen if not already lowered. • adjust the lighting to the “show scene”. • instruct the video/audio switcher to route the VCR output to the projector and to the audio system. • instruct the programmable audio mixer to adopt the preprogrammed equalization for VCR. • either start the VCR, or, more likely, present the VCR operating controls on the touch screen controller. Room control systems can be customized so that the control panel matches the installed equipment exactly. If the system is modified later, the programming can be modified accordingly. However, making this modification quickly does require familiarity with the equipment and the control program, and for this reason the installation and subsequent servicing and updating of room control systems is done by specialist AV installers. Some of these specialize in the needs of the corporate market, and others in the needs of home theater and similar installations. The best lighting control strategy is to use a standard multi-scene lighting controller that can operate independently, but which has the facility for interfacing with the room controller. Some simple systems use individual contact closures to select a scene, but serial control is more flexible and professional lighting control equipment offers EIA232 or EIA485 interfacing.
For the purposes of this section “conference centers” are several different things, for example. Within a large corporate building there may be many meeting rooms of the type described in the previous section. Major financial and IT companies have as many as 50 or more meeting rooms in one building. The aggregation of such spaces can be likened to a conference center.
Figure 15.34 Principle of networking multiple meeting rooms. In this example the room controller also acts as a local media server. The central supervising computer can also be used to run a room booking service. Figure from Electrosonic.
439
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Purpose built convention centers provide a mixture of exhibit space, auditoria and meeting rooms. A large convention center may have 100 or more separate meeting spaces. Hotels have banqueting suites that at their largest are similar to convention centers. The difference here is that hotels tend to cater for a wider range of events. As well as corporate and association events, they cover personal events such as wedding receptions and parties. The lighting control needs of the corporate installation are simply an extension of those outlined in the previous section. Each room should be independently equipped. However, there are two possibilities for overall supervision. If lighting control is considered on its own, each subsystem can have a gateway to an overall BMS, so that room lighting status can be centrally monitored. With the advent of networked room controllers, an alternative method is to monitor lighting, and other equipment, status through the room control system. This can also be used to monitor projector lamp life and give early warning as to when lamp
Figure 15.35 Convention centers have many meeting rooms, often with partitioned spaces. Lighting control must be intuitive to use, and must automatically match the controls to the space layout. Photo of one of the meeting spaces at Orange County Convention Center, Orlando FL, which uses Strand Lighting’s Premiere architectural lighting control system. Notice that at the control position there is a push-button scene selection panel. This is linked to partition sensors. The separate connector panel on the right carries a 5 pin XLR connector to allow the lighting to be DMX controlled for special events. The lower control panel is the audio control.
440
replacement is due. A common arrangement is to have a central controller that combines the function of system monitoring with that of providing a room booking service. In convention centers and hotels the situation is somewhat different. Generally the meeting spaces within them are not permanently equipped with projection equipment, and even the audio equipment is limited to loudspeakers and amplifiers. Sometimes motorized screens are fitted. When AV is required, the equipment is brought in on an as-needed basis. Simple events are dealt with by a local company franchised to provide the service. Large scale events like product launches bring in their own staging companies. This means that all meeting spaces should be equipped with multi-scene lighting control that is intuitive to use. For example, spaces that can be subdivided require that individual control panels affect only the divided space when a partition is closed, but affect the whole space when it is open. (See Section 12.4.) In large spaces there may be a need to allow alternative control access to the lighting. i.e. it is normally under push button multi-scene control, but there is also a facility for manual control or to connect a lighting console using a DMX gateway for special events. (See Section 13.1.) In hotel banqueting/conference suites there may be some spaces that have additional lighting for small stages, most of which is entertainment oriented. In such a case the main architectural control system needs to be able to impose suitable scene control on the entertainment sub-system. There is usually a requirement for the provision of power for temporary lighting rigs. This needs careful consideration in respect of the protection provided. Staging companies bring in dimmer racks and other lighting distribution equipment that is complete with sub-circuit protection but requires input protection matched to the rack size. (See Section 17.4.) 15.4.3 Auditoria An auditorium differs from the meeting spaces described above in that seating is fixed, and has a sin-
ARCHITECTURAL APPLICATIONS
gle focus of attention. Examples are lecture theaters, cinemas, main auditoria within congress centers and concert halls (covered in Section 15.4.4). Lecture theaters are a feature of research centers, corporate training establishments, universities and other institutions of higher education. They have raked seating to ensure that everyone has a good view of the large “teaching surface”, that may consist of a combination of writing boards and projection screens. Lecture theaters for scientific subjects also feature demonstration benches. Lecture theaters benefit from multi-scene lighting control with example scenes being: “Normal lecture”. This provides good all round illumination so that the writing surfaces are well illuminated and students have good lighting for note taking. “Special lecture”. In a superior facility there can be an arrangement that when a “feature” or special occasion lecture is given, the audience lighting is at a lower level, and the lecturer is lit at a single position. This does not apply to a normal lecture because lecturers otherwise want to move around. “Projection”. There may be more than one projection scene, depending on the type of projection in use. However, as with business applications, the need for lower level lighting when using projection has reduced with the advent of high light output electronic projectors. There is still a need to keep stray light off the projection screen(s) to ensure reasonable contrast. As with the business application, the showing of “films” as complete audio visual programmes may justify a lower lighting level. A factor arising with lecture theaters is that for many of them the overhead projector (OHP) and the slide projector still remain important methods of image presentation. “Demonstration”. A separate lighting scene covering the use of the demonstration bench may be justified. Many lecture theaters have daylight as a source of lighting. Ideally a photosensor is then used to modify the lighting scenes to account for the daylight contribution. Scenes requiring low lighting levels then require window blind control. Control panels for selecting lighting scenes
should be sited at the lectern (or lecturer’s position at the demonstration bench) at the entrance door(s) and at the projection position, which may or may not be in a separate projection room. Most lecture theater lighting is fluorescent, possibly augmented by incandescent lighting for features such as demonstration areas and writing surfaces. It should be remembered that, in practice, fluorescent lighting cannot be dimmed to subjectively low levels. For teaching the dimming levels that can be achieved are perfectly adequate, so fluorescent lighting with a dimming facility is fine for most lecture theater work. However, if “special” lectures or events are to take place that require sustained low levels of lighting or smooth dimming to extinction, it is necessary to include incandescent lighting in the mix.
Figure 15.36 A lecture theater at KTH (Royal Technical Highschool) in Stockholm. Advantage is taken of daylight when available, but when projection is required daylight must be eliminated. User control of the lighting and AV facilities is from a touchscreen operated room controller.
441
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 15.37 The re-furbished lecture theaters at the Institution of Electrical Engineers in London have lighting control arranged to ensure that both the lecturer and the projection screen can be seen at their best. Interior design by Renton Howard Wood Levin Partnership; AV systems installation by Quest Technical Systems.
Depending on their size, lecture theaters only require a few channels of dimming, and their needs can be served using the required number of single channel dimmers, or a small multichannel unit. In Europe the DALI option is a good one, since this allows the addressing of individual luminaires, so minor “tweaking” of scenes, for example to reduce light level near a screen, is easily accomplished. Cinemas. Most modern cinema theaters have very simple lighting control requirements. In all but the largest cinemas only one or two dimmer channels are needed. Simple single channel dimmers with automatic push button control are favored since these
442
interface easily with cinema automation systems. Smooth dimming to extinction is an absolute requirement, so most lighting is incandescent − this does not have significant energy use repercussions since, by definition, cinema auditoria are dark for most of the time. (See Section 12.3 and Figure 12.2.) In cinemas a “cleaning” scene based on fluorescent working lights may be appropriate in order to give a high illuminance for cleaning. These can be contactor controlled. Lighting control points are usually provided in the projection room and at “secret” locations at the entrance doors. The latter may be arranged so that during performances it is not possible to select full lighting (and least of all the cleaning scene.) However, and conversely, many local authorites require a “Panic” button to be provided that is easily accessible and which brings up full lighting in the auditorium. In some lighting designs continuous cove lighting is required. The problem is achieving an uninterrupted, even, spread of lighting using a source that is dimmable to extinction. Cold cathode lighting gives acceptable results, and provides color options. Festoons of closely spaced low voltage incandescent lamps also work, but only those with a guaranteed life of tens of thousands of hours should be considered. LED sources may well take over in this application, using PWM control.
Figure 15.38 The Warner Village Cinemas Screen 2 at Star City uses automatic dimmers like those shown in Figure 12.2 for auditorium lighting control.
ARCHITECTURAL APPLICATIONS
Figure 15.39 Most cinemas run fully automated presentations. The Kineton range of 35mm and 70mm projectors have the option of a built-in logic controller that controls all the projector functions and the auditorium house lighting.
All public auditoria, including cinemas, are required by regulations (that differ slightly between counries and local jurisdictions, but which are the same in principle everywhere) that specify a minimum lighting level at all times the space is occupied. The same regulations require emergency exit signage to be illuminated. The low maintained level may be less than 0.1 lux, and anyway must be directed so that it does not fall on the screen. There can be a problem with exit signs appearing too bright in the dark, giving rise to unacceptable glare. In cinemas it is possible to set the signage luminance to a level that is acceptable for both showing films and for when the houselights are on. However, in some mixed use auditoria that for some events have much higher lighting levels, it may be necessary to use exit lighting that can operate at a different luminance for showing films in the dark. The selection of this lower level should be automatically invoked when the “film” scene is selected. Large auditoria within congress centers. These have more complex lighting control systems for both architectural and practical use reasons. A prestige auditorium may well have extensive architectural detail that is best exposed by lighting. The lighting designer may require separate control
of many lighting circuits in order to achieve the lighting balance he requires for different uses of the space. Nonetheless, overall control for most uses of the space can be reduced to the selection of a number of scenes. Most of these can be anticipated at the time of commissioning the system, but it is a good idea to allow for a the addition of extra scenes and the fine tuning of existing ones after the auditorium has been in use for a few months. The type of event that takes place in large auditoria varies sufficiently that for some of them it is necessary to have additional lighting for special events. This lighting is akin to theatrical lighting, so may well have a small theatrical control console. Even if such a facility is provided, it may be necessary for the additional lighting to be used as part of easily selectable scenes; and this means that either it can be controlled by both the main scene selection system and by the theatrical console, or that the console has a default state that enables it to be controlled by the main house system. Large public auditoria are often the venue for televised events. The additional lighting that is
Figure 15.40 The Aula Magna at Stockholm University. This leading Swedish conference venue has a large auditorium where the main lighting control is operated on a scene selection basis.
443
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
15.4.4 Concert halls Some of the points made in this section apply equally to theater auditoria, especially opera houses and straight drama theaters.
Concert halls have many features in common with the large auditorium. Main lighting is scene selectable, and the concert stage lighting is both scene selectable and controlled by a theatrical console. The major consideration for classical music concert hall lighting is that the entire system must be silent. When consideration is given to the influence of background noise within a building space, acousticians refer to background noise design criteria. They examine every possible source of sound and aim to ensure that, for the activity concerned, background noise does not exceed a certain level. Various different metrics have been used, but the most common is the use of perceived noise criteria curves. PNC curves show the sound pressure level within each octave band as in Figure 15.41. An area like an open plan office might be at PNC35, a private house living room at PNC30, and a retail store at PNC45. Within a concert hall the aim is the very demanding PNC15. Acoustic consultants Arup Acoustics make the embarrassing (for the lighting community) comment that lighting equipment of one kind or another is almost entirely responsible for the difficulties in meeting PNC15. They say that the really difficult problem of combating noise from air conditioning systems has been solved, and can be implemented at the building design stage. Unfortunately too few people realise that luminaires and lighting control systems in combination are potent sources of acoustic noise − so, just as the conductor raises his baton for that magic moment of silence, the effect is ruined by the whirring of fans or the resonant buzz of a thousand partially dimmed filament lamps. Examples of the problems that must be antici-
444
pated are as follows (not all relate to lighting control as such, but are included to give a more complete picture). Thermal contraction. Lights get hot, and in doing so their components expand. In expanding, but even more so in contracting when the light is switched off or dimmed, they can make intermittent popping or creaking noises. Well designed theatrical luminaires should not suffer from the problem, but it can affect even simple things like color frames. Problems sometimes arise with specially designed luminaires that are customized for the venue. All such items should be acoustically tested before they are installed. Lamp noise. Another reason for testing all luminaires for noise is filament sing caused by thyristor or transistor dimming. The luminaire design may be such that the effect is amplified or made directional. The new generation of theatrical luminaires exemplifed by ETC’s Source Four® range use lamps with a filament design that is “quiet”; older luminaires with long filaments may
80 70 PNC
60
Octave band SPL dB
required for television can now be accommodated as part of the “special event” lighting, but may require some extension to the control facility for the TV producer.
65 60
50
55 50
40
45 40
30
35 30
20
25 20
10 0 31.5
15 63
125
250
500
1000
Frequency Hz
2000
4000
8000
Figure 15.41 Perceived noise criteria or PNC curves. The dB figures for SPL are related to the threshold of hearing of 20mPa (see Table 2.2). SPLs are shown in “octave bands”. The demanding PNC15 calls for only 15dB at 1kHz. A dBA figure for total noise is dependent on the spectral composition of the noise, so can only be used as a rough guide.
ARCHITECTURAL APPLICATIONS
Figure 15.42 The Bridgewater Hall in Manchester, England. Acoustic design by Arup Acoustics, auditorium lighting design by LDP (lighting designer Mark Sutton Vane). The splendid chandelier, based on the use of tungsten halogen lamp capsules fed by remote transformers, presented an interesting problem. The original layout of the open feed conductors and conventional transformers resulted in the high current inducing currents in metalwork some distance away. This would have presented both acoustic and electrical noise problems, especially when dimming. The solution was to use HF electronic transformers with DC output, and with voltage measurement at the lamps to regulate the output. Photo by P. Mackinven. © Arup.
be troublesome. Any testing must be done with the actual dimmers to be used − it has even been found that different makes of nominally the same lamp type give different results on dimmers with nominally the same filter specification. Concert hall dimmers are specified with long rise times to minimize the problem, but there is a trend towards the use of transistor dimmers (and ultimately sinewave dimmers) for this application. Dimmer noise. Dimmers fitted with chokes to increase rise time and to meet EMC requirements do make a buzz. Dimmer rooms must, therefore, be completely isolated from the auditorium. Even mounting dimmers on the wall of a dimmer room can transmit sound through the building structure. The propensity for dimmers to create acoustic noise, and/or to induce audible noise voltages in to sound systems can be objectively measured, see Section 17.3.2. Fiber optic light engine noise. It might be thought that fiber optic lighting would help the acoustic problem, but in practice it has proved one of the worst offenders. Most of the light engines used
for the application use noisy fans, and it is very difficult to prevent sound transmission round or even through the light pipe, and to provide total isolation for the concealed engine. Moving lights. These make a noise when moving, but movement cues can in practice be masked by wanted sound. Luminaires based on tungsten halogen lamps may be practical, but usually designers want to use HID sources that introduce both luminaire and power supply fan noise. Few moving lights can meet PNC15. Color scrollers. It is practically impossible to meet rigorous noise criteria when color scrollers actually scroll; so the only solution is to ensure that scrolling cues are invoked when wanted sound (orchestra at fff!!) can mask the sound of the scrollers. A subsidiary problem is that scrollers on larger luminaires use fans for cooling. These need to be intelligently operated, so they only run when necessary, and run at low speed when light colored filters are in place. Console noise. The practice of siting lighting consoles (and sound mixing systems) within the auditorium space or in un-isolated control rooms is
Figure 15.43 Vari-Lite have developed moving lights to meet the stringent noise requirements of concert halls and similar venues. This is the VL1000™ ERS luminaire.
445
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
another unexpected source of noise. All computer devices, including laptop computers, have fans that are more or less noisy, and modern lighting consoles, based on standard computer components are no exception. 15.4.5 Assemblies A special class of meeting space, that combines the requirements of a conference room, an auditorium and (regrettably, sometimes) a place of worship, is the assembly. This can range from a local government council chamber to a government legislature. Courts of law and professional tribunals are in the same category. Such places are all characterized by the need for the building to express an air of authority. Good lighting enhances the architectural statement and is an essential part of the building. The need for lighting control has developed as the use of these spaces has become more sophisticated. In particular, many now require the proceedings to be televised, either continuously or for particular debates or proceedings. This may be done for record purposes, for public broadcasting or as means of relaying proceedings to another site.
Figure 15.44 Completed in the late 1980s, the Civic Centre council chamber of Mississauga, Ontario, Canada makes use of multi-scene lighting control. The chamber is used for both local government and community use. The normal meeting scene ensures light for reading and brings out the architectural features of the building. For televised events additional lighting can be selected. The chamber is equipped with full AV facilities based on concealed motorized screens, and this requires low level “AV scenes”.
446
Figure 15.45 The Reichstag, the new German parliament in Berlin, uses multi-scene lighting control both in the legislative chambers and in the public circulation areas. This photograph of the plenary chamber shows some of the lighting, the cupola, and its associated light sculpture. The slender inverted conical structure is covered with 360 glass mirrors, and is 2.5m (8ft) wide at the point it punctures the chamber ceiling, widening to 16m (53ft). Associated with it is a sun-following movable shield to block penetration of direct sunlight, with its associated problems of glare and solar heat. The shield is powered by photo-voltaic cells. The architects were Foster and Partners, and the lighting consultants were Claude & Danielle Engle. The photograph is by Nigel Young of Foster and Partners.
Assemblies may also require the display of shared information; for example in the form of electronic voting displays or projected images, and tribunals and court rooms require participants to use computer terminals and similar displays for the examination of photographic and documentary evidence. Legislative assemblies have ceremonial events, where the focus of attention may vary and security issues dictate the need for a high level of stand-by power in the event of main power supply failure. Such considerations require at the least a multiscene control system. When broadcast television is involved, there may be a need for additional control operated by the television crew. This would be for lighting that augmented the permanent building lighting. A special issue that arises with this kind of installation is that of who has authority to operate the lighting control system. Careful consideration must be given to the location of the control and the means of authorizing its use.
ARCHITECTURAL APPLICATIONS
Figure 15.46 Court rooms and tribunals are now required to cater for videoconferencing to allow remote witnesses, and for electronically presented documents and evidence. The General Dental Council chamber (architects Bryant Harvey) in London operates as a court of law for disciplinary proceedings. Lighting and AV facilities are controlled using a standard room control system. Lighting control is by a system similar to that shown in Figure 12.4 with four fluorescent circuits, three incandescent downlight circuits, and a switched fiber optic circuit as loads. There is an EIA232 interface between the stand-alone lighting control system and the room control system.
15.5 Places of worship 15.5.1 The architecture Buildings that are used as places of worship for religious observance display immense variety, from the flamboyance of a baroque cathedral to the simplicity of a small chapel. The proceedings that take place within them also vary in complexity according to the religion, or sect, concerned. The central part that religions have played in the civilizations of the last two millennia has meant that their buildings have correspondingly been at the center of communities. Their construction from durable materials means that there are buildings of hundreds, or even thousands, of years old still in use; but at the same time the development of new communities means that there are also many modern churches, temples, synagogues or other examples of religious buildings. The older buildings make subtle use of daylight
to create dramatic lighting effects; for example by the use of carefully sited windows or slots and by the use of stained glass. Architects of the more modern buildings often pay homage to their forbears by incorporating similar tricks in to their designs. Now that all such buildings use electric lighting, it can be seen that in any new building the role of lighting is taken fully into account at the building design stage. The lighting is an integral part of the building, and it is a matter of design choice as to whether the lighting sources are concealed or are architectural features in their own right. In the older buildings, modern lighting has had to be “tacked on” to the existing structure, usually based on the conversion of pre-electric lighting based on flame based sources; augmented by largely concealed lighting intended to emphasize the architectural features of the building. 15.5.2 The role of control There are many different activities that can take place within a place of worship, not all of them necessarily religious in nature. A large building, such as a cathedral, may have several active areas such as side chapels that allow either for multiple events, or for a given event (such as an early morning or late evening service) to take place on a different scale. The activities and areas associated with religious services can be classified. A congregation area. The congregation may take an active part in parts of the service or ritual (for example by singing) or may be passive observers. Reading or preaching locations. Where individuals read from texts or preach to the congregation. A leadership area. An area from which the leaders of the worship conduct the service. This may be augmented by a space for a choir. The leadership area itself may be centered around a table or altar. The lighting of these main areas must be sufficient for the task, especially in respect of reading, but should also provide dramatic emphasis. In an elaborate building the prime area lighting is supported by architectural and accent lighting that brings
447
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 15.47 St Peter’s Collegiate Church, Wolverhampton, England has medieval origins with large Victorian additions. The two side chapels and the choir are visually remote from the main part of the church, and some services use only part of the church. A scene-setting control system, providing 16 scenes, is used to match the many different types of service, and to make the most of the interesting architectural and sculpted features. Three control points are provided, one in the vestry, one at the rear of the church and one in the choir. Lighting design and photographs by Sutton Vane Associates, Lighting Design, London.
out the details of the building structure and highlights works of art or memorial. In all but the smallest building the lighting requires many circuits, and certainly requires a multiscene control. Table 15.2 suggests some of the scenes that may be required. Scene selection must be possible from several different places in the building, and the rules for its use are, first, that only those scenes relevant to the local activity should be available from a particular control position. Second, that controls are only operated by an authorized person. This means that they are only accessible via a concealed panel or similar arrangement.
448
The full lighting of a large building is expensive to run. In some European cathedrals and churches the normal visiting lighting is, for this reason, quite limited. Sometimes visitors are given the option of paying to have extra lighting put on for a short period, for example to illuminate a high gothic ceiling. The dramatic nature of these buildings favors the use of theatrical techniques, so tungsten halogen lighting is widely used. This simplifies control from both an optical and light level point of view. In order to reduce running costs in “visiting” hours, HID lighting (metal halide) may also be appropriate for large area lighting, for example the illumination of high ceilings. The incorporation of this into “scenes” needs careful consideration, since smooth dimming is not always possible. Chandeliers should use incandescent lighting, ideal for control. Full lighting may only be needed for a few hours a week, so reasonable economy and long lamp life can be achieved by using a low level outside service hours. Some churches have made the dreadful mistake of retrofitting chandeliers with compact fluorescent lamps with internal ballasts, often with an inappropriate color temperature. The result is awful, and the energy saving is probably insignificant. Unless luminaires have been specifically designed for dimmable fluorescent sources (of
Figure 15.48 In the 21st Century places of worship take many forms. In the USA many churches have huge congregations, and use the latest in video and audio technology, like this example of The People’s Church in Franklin TN. Their needs for lighting control are more theatrical and “hands on” for services, but they still need an easy means of selecting the lighting scenes required.
ARCHITECTURAL APPLICATIONS
Scene
Com m ent
Pre-service M ain service Readings Serm on
Congregation area em phasized. Leadership area em phasized. Lectern em phasized. Pulpit or equivalent em phasized; leadership area reduced. M ay require additional lighting or em phasis in different parts of the building. Scene em phasizing architectural detail. Used when no m ain event is in progress. (If applicable) M ay require alternative em phasis lighting at locations not used in norm al services. Often forgotten. M ay som etim es use alternative sources, and m ay be zoned. This m ight cover the needs of concerts or sim ilar secular or sem i-secular events.
Processions or other ritual Visiting W eddings, funerals etc. Cleaning Special event
energy. Indeed in old buildings the way in which daylight acts creates a whole new set of “scenes”. In the USA there are many very large new churches that have congregations numbering thousands of people. Such churches often include video screens and use TV cameras to help with “image reinforcement” for the preacher(s) and leaders. This in turn means that the lighting system, and its control, is more akin to those used for large convention auditoria.
15.6 Museums, art galleries and libraries
Table 15.2 Examples of lighting scenes in places of worship.
15.6.1 Museums
an appropriate color temperature) fluorescent should be avoided. Many churches have a significant daylight contribution, in which case “visiting” scenes in particular can be modified to take advantage of this to save
This section considers “traditional” museums in the sense of places that conserve and display historical artifacts. Museums now span a wide range, from entirely scholarly collections, through to highly interactive displays based more on ideas than on objects. Many museums, for example the Science Museum in London, or the American Museum of Natural History in New York do both. So in practice many museums combine the techniques described in this section with those described in Section 15.7. The lighting of museum displays raises a number of problems: • many museums have an extensive daylight lighting contribution, especially in the “classic” museum galleries. • museums want to make the objects on display as striking as possible, requiring carefully directed light. • but often the objects to be displayed are light sensitive, and will deteriorate if exposed to continuous high levels of light. • museums, in common with other energy users, want to keep energy consumption to a minimum. • but if there is free circulation, all galleries must be illuminated in opening hours. Some simple precautions can be taken. Galleries with a lot of daylight are simply not used for sensitive objects; and anyway may be fitted with motorized louvers to eliminate the effects of direct sunlight. All lighting sources must be “low UV”, either
Figure 15.49 Many church buildings are floodlit; the control requirements are usually very simple being based on simple time and photo-electric controls. Sometimes dynamic lighting is justified. The ruins of St John’s Church, Chester, UK are a listed monument. Dimmer controlled floodlighting allows the building to be revealed in stages. Lighting design and photographs by Sutton Vane Associates, Lighting Design, London.
449
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Fluorescent tube
Dichrioc spotlight
T.H. Flood to light front wall and graphics
Diffusing film Top glass Vase lit to high light levels and modelled well by the spotlight
Viewing position
Figure 15.50 The “Genius of Wedgwood” exhibition at the Victoria and Albert Museum, London displayed both pottery, that has no light level restrictions, and documents, which must be limited to 50 lux. Individually controlled light sources ensure that light levels are safe, and that there is an interesting and dramatic lighting balance, with good modelling of three dimensional objects, and clearly visible captions without glare. Lighting design, drawing, and photograph by Sutton Vane Associates, Lighting Design, London.
inherently, or more usually by the fitting of filters, with the aim of eliminating any wavelength shorter than 400nm. Conservation experts will determine what any particular item can withstand. In practice this results in two recommendations. First, a maximum illuminance. This is typically 50 lux for highly sensitive items like manuscripts, natural fibers, textiles and watercolors; and 200 lux for medium sensitive materials like wood finishes, leather and stable dyes. Second, a maximum annual exposure; assuming all radiation <400nm has been filtered out, this can be
450
expressed as “lux hours per year”. For example 50,000 lux hours for highly sensitive material, and 500,000 lux hours for medium sensitive material. So materials with a recommended exposure limit of 50 lux should not be illuminated for more than 1,000 hours a year (or they could be illuminated at 30 lux for 1,600 hours). These limitations present the lighting designer with some problems. If the lighting is to highlight the object, the designer will want the lowest practicable ambient light. On the other hand practical and legislative pressures will at the same time be demanding that all captions are clearly legible, and can be easily read by people with moderate vision, and that all walking areas and means of egress are lit to a minimum level. Adam Grater of DHA Design says that the museum lighting designer has an almost impossible task unless he “owns the space”. In practice it is simply not possible to meet the demands of conservation and interesting display unless the designer can start with a “black box”. At first sight a lot of museum lighting seems simple in principle. For general lighting and the lighting of separate items, track lighting may be used. In display cases fiber-optic lighting is extensively used because it is inherently UV free, and gives great flexibility in illuminating small objects. A technique advocated by DHA Design is, using “TV lighting” terms, to use fiber optic for key lighting and to use UV filtered 3,000°K fluorescent for softlighting. Particular care must be taken with caption lighting to ensure that visibility needs do not result in unacceptable glare. But, and this is a big but, to get a satisfactory balance, and to achieve the required illuminance levels, some form of lighting control is needed on every circuit. Most galleries only have a single “scene” (although there may also be some special provision for cleaning lights) so a cost effective method of providing the necessary control is to use luminaires with individual dimmers that can be manually set to the required level, for example by using: • controllable fluorescent ballasts with local control potentiometer.
ARCHITECTURAL APPLICATIONS
Figure 15.51 The main galleries at the J. Paul Getty Museum, The Getty Center, Los Angeles, CA, are designed to be primarily illuminated by daylight. Electric lighting is only used to augment daylight on overcast days and for evening openings. A Central Lighting Controller (actually a BMS from Landis Staefa) monitors exterior lighting levels, interior lighting levels and the position of the motorized louvers used to control the amount of daylight. The central lighting controller uses the feedback information to activate the louver system to hold the illumination within prescribed limits, and activates the subsystems that operate the electric lighting when required. The controller also responds to scheduling information and special event over-ride. A Lutron dimming system is used to provide two circuits of supplemental art light in each gallery. One is run at a cool daylight color temperature, and the other at 3,000K. The supplemental light is faded in or out over the course of several minutes as required. Maximum output is 90% to extend lamp life. A GE relay based lighting control sub-system (see Section 12.6) is used for controlling working, cleaning and security lighting. It also switches in fluorescent lamps that are used as uplighters on the closed louvers when the museum is open in the evening. In the lower photograph the louver system and supplementary lighting can be clearly seen. Notice also the high truncated pyramid ceiling that helps diffuse the daylight. Lighting design by Fisher Marantz Stone. Photos by Tom Bonner © The J. Paul Getty Trust.
• track luminaires with built-in dimmers. • electronic transformers with built in dimming facility and local potentiometer to source the tungsten halogen lamps used for fiber optic illumination. Tungsten halogen is preferred to discharge lighting because of its continuous spectrum and better color rendering. Where particularly sensitive objects are being displayed, and where visitor numbers are low or variable, it is possible to augment the control with presence sensor control (PIR or ultrasonic) so that exposure is minimized. In larger galleries the use of multichannel dimming with multi-scene control may be recommended, especially if the galleries ever have alternative uses, for example for evening receptions, or where the gallery has daylight. An intermediate solution can be provided by the use of DALI protocol compatible devices. This brings the benefits of individual luminaire control combined with proper scene programming at comparatively low cost. It avoids the need to physically access individual luminaires when adjusting levels, and provides a means of properly recording all level settings.
Figure 15.52 The Wellcome Wing at the Science Museum in London makes striking use of lighting. The exterior and open space lighting design was by Hollands Licht, and the exhibit lighting design, shown here, by DHA Design. All lighting is under programmed scene control using a mixture of switching and dimming. Control system by Delmatic.
451
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
ing the latest effusions from art’s enfants terribles. The latter can use the same techniques as are used in retail display; but mainstream art galleries have to be more careful. Again the issue is conservation. As an extreme the display of some of Toulouse-Lautrec’s work (although recent by most standards) at the Museé d’Orsay in Paris require very low light levels. But many need to follow at least the 50 lux rule. Formal studies have been done on the subject of color fastness, and there is an ISO rating system based on a comparative test called the Blue Wool Test. Colors are rated: • permanent, unaffected by light exposure. • rating 7−8 durable; 100−300 years at 1,000 lux. • rating 4−6 intermediate; 3−30 years at 1,000 lux. • rating 1−3 fugitive; 2−20 years at 50 lux. In the above each “year” is considered as 3,000 hours, and the year rating is the time to just perceptible color change. While modern pigments have been developed
Figure 15.53 The American Museum of Natural History in New York has a rolling program of opening new or refurbished galleries that make subtle use of audio-visual and lighting techniques. In the Hall of Biodiversity (above) visitors see big screen back projected footage of rainforest activity through a foreground of foliage with dynamic lighting. In the Gottesman Hall of Planet Earth (below) exhibit lighting is controlled so that the spectacular set pieces can be viewed to best advantage. Exhibit design by Ralph Appelbaum Associates. Bottom photo© AMNH.
Whichever method is chosen it may well be the case in a large museum that the lighting control for each gallery is itself under the over-riding control of a BMS. 15.6.2 Art galleries Art galleries range from the great national galleries like the Louvre to trendy boutique galleries promot-
452
Figure 15.54 The National Portrait Gallery in London uses a variety of lighting techniques. This section, used for 20th Century and contemporary portraits, makes extensive use of fluorescent lighting, both in coves and from track mounted fixtures. The tracks also accommodate tungsten halogen luminaires. All lighting is under dimmer control. Dimmable ballasts are used for the fluorescent lighting to simplify the circuit wiring. Consultants Merz Orchard.
ARCHITECTURAL APPLICATIONS
that are all in the high durable or permanent categories, many important paintings use pigments that are at best in the intermediate category. This means that any lighting must be controlled to ensure that annual exposure limits are not exceeded, and here the lighting level “out of hours” is very important. It is no good meeting the public hours objective if the gallery cleaning team goes and leaves the cleaning lights on all night. 15.6.3 Libraries Libraries can exhibit the characteristics of a monument, an office, a study room, a museum, an art gallery, a conference center or a warehouse. To take one of the world’s most celebrated libraries as an example, the British Library in London embraces: • a four level basement bookstore holding 12 million books on 340km (211 miles) of shelving. An automated handling system delivers over 40,000 boxes of books a month from the storage areas to the reading areas. • eleven reading areas with seats for 1,200 readers. Most of these have provision for readers to use a computer. Some have gallery or wall spaces for exhibiting paintings or documents from the collections. • a conference center with an auditorium and a suite of four meeting rooms. • the unique King’s Library, a 60,000 volume collection of King George III, displayed in a 17m (56ft) high glass walled tower. • three exhibition areas; a permanent exhibition of the “treasures” of the library; a temporary exhibit area, and a sponsored exhibit area. • a large public entrance area, including a restaurant and other facilities. • a huge (34,000m2, 340,000ft2) suite of plant rooms to provide air conditioning, fire protection, energy management and security. The air in the storage and reading areas is filtered to the same standards as a hospital operating theater; and is maintained within ± 1°C of 21°C for reading and 17°C for storage. Humidity is kept within ± 5% of 50%. • office accommodation, in the form of both open plan and individual offices, for academic, adminis-
trative and operations staff (around 1,000 people). • a striking exterior with a large piazza. Clearly few libraries operate on such a massive scale, but most will contain the same elements. Lighting of the bookstacks in the storage area at the British Library is designed to deliver 40 lux at the base of the stack, and this requires a total of 6,000 luminaires, although at any time no more than 10% of them are in use. Because the need for lighting is unpredictable, the switching on of these luminaires is done manually by the book-pickers using pullcord switches. Switching off is automatic after a preset time. The permanent exhibition areas and the lighting of the book spines of the King’s Library require lighting at conservation levels, so here customized luminaires and dedicated control systems are used to ensure limits are not exceeded. The lighting for the whole of the above ground
Figure 15.55 The reading rooms at the British Library have individual reading light control, but the rooms as a whole are under the main lighting control system. Many of the reading areas have a daylight contribution.
453
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
parts of the building, parts of the basements and all external lighting is under distributed control with local manual over-ride. In all there are a total of 114 intelligent outstations. These receive local inputs from user switches, photo-cells, timers and sensors and issue the appropiate output commands via isolating relays to contactors installed within the main power distribution system. The outstations are all linked to a central computer. This monitors the outstations and also issues scheduling commands based on security, calendar and time of day parameters. (See accompanying figures.) Some areas, such as the restaurant, entrance area and conference suite have the addition of full scenesetting capabilities and dimmer control. Programming of the scenes is done locally, but their recall can be achieved either from local control panels or from the central computer.
Figure 15.57 All lighting in the main areas at the British Library can be scheduled and monitored from a central position.
15.7 Visitor centers and exhibitions 15.7.1 Definitions There are now countless sites that span culture and entertainment that are covered by the general description “visitor centers”. They include some kinds of museum (or parts of museum), science centers, aquaria, countryside and parkland visitor centers, historic buildings, corporate visitor centers and some
Figure 15.56 There are 114 intelligent outstations controlling the lighting at the British Library. As shown here they are installed in the electrical risers.
454
Figure 15.58 The Golf Hall of Fame in St Augustine, FL, is an example of a visitor center. It combines entertainment with some attributes of a specialist museum. Visitors see an introduction show “Passion to play” that combines archival and contemporary footage on a multi-screen format. The lighting in the area is dynamically controlled as part of the show. Exhibit design by Ralph Appelbaum Associates.
ARCHITECTURAL APPLICATIONS
commercial attractions. Most make an admission charge, but few rely solely on admission charges for funding. Most provide a mixture of education and entertainment. The exhibition areas of these places offer opportunities for good lighting design. Often they provide areas that are “black box” or close to it, which gives free rein to the designer. For static exhibition displays the lighting and its associated control follows the same principles as those used for museums, stage sets and retail display, depending on the nature of the exhibit. Standard multi-scene architectural lighting control is ideal for this. Often, however, visitor centers require dynamic lighting, and this introduces the need for automatic show control. 15.7.2 Show control Many exhibits require a sequence of lighting changes that either runs continuously, or is initiated by a presence sensor, timeswitch or push button. If the sequence is very simple, like the continous crossfading of two dimmers to operate the ever popular “Pepper’s Ghost” illusion exhibit, it can be achieved by standard architectural lighting controllers that have built-in linked-scene capability. But usually something more elaborate is required, so normal practice is to separate the basic lighting control task from the programming. “Show control” can be applied to small self contained exhibits, right up to major multi-media extravaganzas. Most such exhibits require the lighting to operate in synchronization with audio-visual elements. In meeting rooms (Section 15.4.1) the concept of “Room control” was introduced. Show control does something similar for exhibits, but here the emphasis is on the reliable replication of an automatic sequence that is precise in its timing. Figure 15.59 shows examples of show control devices. One architecture is to have a central controller that itself can control a small exhibit in its entirety, but which can also support several additional output interfaces. An example specification is that the controller: • can itself operate eight relays with isolated contacts.
• has four serial ports, configurable EIA232/ EIA485. • has four opto isolated inputs (for program start signals etc.). • is entirely solid state with no moving parts. • stores its operating program in flash memory (allowing for different applications and program updates). • stores its show program in exchangeable flash memory on a smart card. • has an SMPTE timecode option, allowing it to synchronize to an external source of timecode. • can work with additional interface units to provide any reasonable number of analog, digital and serial outputs. The “analog” outputs can also be in the form of DMX lighting control signals. One “special” output interface that such equipment can deal with is the automatic slide projector. Only a few years ago, almost to the late 1990s, “multi-image” based on slide projection was the main way of presenting spectacular product launches and exhibits, and had been since the 1960s. It has almost entirely given way to electronic
ESLINX Show controller from Electrosonic
Smart PAX controller from Dataton
FrEND network compatible controller from Mediasonic
Figure 15.59 Examples of show control devices.
455
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 15.60 Slide projection is a special branch of show control. Electrosonic France offer a special version of the Kodak Ektapro™ projector with built-in lamp dimmer to provide the dissolve slide change facility and full EIA232 control.
projection, but because of its very high resolution, flexibility, “film look” and because the slide projector is akin to a standard theatrical luminaire, it is still favored for some exhibit work. On a larger scale giant “scene projectors”, based on huge slides or filmstrips, are used to produce images that are well beyond the capability of current electronic projectors. When show control is in use, there is a choice as to how it interfaces with the lighting control. The method depends on the scale of the exhibit and the preferences of the show producer, exhibit designer and lighting designer (who may all be one person, or may be representatives of three different organizations).
Figure 15.61 Guiness Storehouse in Dublin, Ireland, is a popular visitor attraction. It includes the Arthur Guinness Show which makes extensive use of programmed slide projection. Complete exhibit design by Imagination.
456
Figure 15.62 Magna, a science adventure attraction in Rotherham, UK, is inside an old steel works 100ft (30m) high. The introductory “Face of Steel” show uses multiscreen video projection, but in the background huge scrolling images are projected directly onto the inside of the building and the exhibit structure. Exhibit design is by Event Communications and lighting design by DHA Design. Scene projection uses Pani projectors and E/T/C PIGI scrollers. Show control is by Electrosonic.
At its simplest a show controller simply issues contact closure commands at the appropriate points in the show sequence to operate either a multi-scene architectural lighting controller, or a few individual automatic dimmers. Lighting levels and dimming speeds are set up on the controlled devices. The next step up is that the controller itself issues the level and speed information either as analog signals or as DMX signals. For big and complex shows, current practice is to use a separate “lighting playback controller” or “DMX recorder” of the type shown in Figure 13.4 and described in Section 13.2. There are choices in the method of programming. Today all show controllers are initially programmed on a personal computer, with the program then being downloaded into the device (unless it is a big show, in which case a computer may be used on line). Simple shows can be put together using simple command line instructions, but most show producers hate this approach and look for a more flexible and producer friendly arrangement. Show control programs such as Dataton Trax™ and Electrosonic EASY™ allow programming to be done on line; this allows the programmer to see the results as commands are entered, and allows “rock
ARCHITECTURAL APPLICATIONS
changes at any time. They include a vital lock-out feature that prevents use of the programming controls, so the end user can only select preset sequences. 15.7.3 Aquaria Aquarium lighting is a specialized subject that combines showmanship with husbandry. Most public aquaria display exotic fish that occupy tropical or sub-tropical waters, so often the aim is to simulate the lighting conditions of such waters. This results in the need for a high surface illuminance, replicating that of sunlight. Daylight
Figure 15.63 For medium scale exhibition areas and shows based mainly on dynamic lighting, an economic approach is to use a controller that allows on site programming, but which can then have the controls “locked out” so the end user can only select specific scenes and automatic sequences. The Pulsar Masterpiece range, originally developed for live entertainment use, is an example.
and roll” – the ability to repeatedly view small sections of the show as fine tuning of timing and effects is carried out. All elements of the show, including images, sound and effects stay in sync as show segments are run back and forth (although some compressed video files may not allow full tracking). For big shows where a DMX recorder is used, there is usually a lighting designer/director. Again he or she will want to review the lighting programming directly, and will probably feel most comfortable using a console they are familiar with, particularly if moving lights are involved. So the simplest procedure is to carry out the programming of the lighting using an entertainment console, and then transfer the DMX sequence into the playback device. For intermediate sized installations an alternative is to use a complete controller of the kind shown in Figure 15.63. Such devices are reasonably priced and allow on-site programming and programme
Figure 15.64 The “Endless Oceans” tank at The Deep, Hull, UK holds 2.3million liters of salt water and at the time of completion was the deepest aquarium tank in Europe. The specialist tank lighting was specified by International Concept Management of Grand Junction, CO, and realized by Lighting Technology Group of the UK. The complete system uses 60 metal halide luminaires (70–700W) and two 1kW xenon arc searchlights on a total of 38 circuits. Sunrise and sunset are simulated by switching the lighting in circuit sequence over a 90 minute period, using an ETC Unison lighting controller.
457
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
local lighting control backed up by a BMS as suggested in Section 15.6.1. However, when a site uses a lot of AV and interactive displays in conjunction with lighting, it can make sense to use one system to manage both – even if there is also input from some overall management system. Since modern network (ethernet LAN) technology is now well established, one possibility is to put all interactive computers, video servers and show
Figure 15.65 The Deep also features extensive exhibition areas. Here the lighting is controlled by a conventional multi-scene architectural lighting system (Helvar Ambience). Exhibit design by John Csáky Associates, lighting design by DHA Design.
illuminances can reach 100,000 lux or more which would be somewhat impractical to maintain. More realistically surface illuminances of 2,000–5,000 lux are used, possibly augmented by “searchlight” type luminaires to achieve a “shafts of sunlight” effect. A high color temperature is required, for example 10,000K. Depending on the type of lamps being used, and on the nature of the fish being displayed, there may also be a need to augment the UV content. This is both to match the native habitat of the species, and to achieve the correct color rendering. There must be a means of simulating a daylight cycle. In small aquaria this may be achieved by simple timeswitches or manual control. A large public aquarium should use an automatic system. A sudden change in light level can stress the animals, so lighting should be reduced or increased in stages. A PLC or dedicated switching lighting controller is suitable for the task. 15.7.4 Networks in visitor centers Large museums and visitor centers may have hundreds of separate exhibits requiring audio visual and lighting control facilities. There is then a problem of scheduling and monitoring their operation. A conventional museum might well use simple
458
Figure 15.66 At both The Deep (Figure 15.65) and ThinkTank (above) in Birmingham, England, all audio visual and computer-based interactive exhibits are on an ethernet LAN and are under the control of ESCAN, a control program from Electrosonic that schedules and monitors the operation of the complete exhibition. The control extends to the exhibit lighting using appropriate gateway interfaces.
ARCHITECTURAL APPLICATIONS
controllers on to a network. The network is then used solely for supervisory and scheduling purposes. A system of this kind can be expected to schedule the operation of all AV and interactive exhibits, all exhibit lighting and all show control. Such scheduling includes power control and, in particular, the proper shut down of projectors which are usually required to go into a stand-by mode for lamp cooling before being switched off completely. The system may itself provide show control for simple exhibits requiring only a few cues and will monitor the operation of all equipment; e.g. by checking that all interactive computers are active and running the correct program; checking that lamps are OK and warning of the impending need for lamp change. It will also provide over-ride facilities that, for example, vary the overall level of audio depending on visitor numbers, or even run shorter shows when the venue is crowded. This arrangement also caters for special events outside normal opening hours. A system of this kind is illustrated in Figure 15.66. As the network approach takes hold, it may well be that system architectures will move much more to distributed client devices. However, these will only be accepted by creative producers if the programming facilities they are now familiar with are not compromised.
Because of these factors, this application group has tended to be in the lead in the use of new lighting control techniques and new lighting sources. However, there can, unfortunately, be some disparity between what the designer or consultant intends should happen, and what actually happens. Such disparities arise for two main reasons. One is a failure to understand the nature of the lightsources being used, and the other is a failure to understand how users behave in reality. An example is the use of compact fluorescent lamps with built in electronic ballasts for corridor lighting, with the worthy aim of saving energy; but which, because they are not easily controlled in respect of light output, may result in either over-bright or patchy lighting at night. Another example is the provision of manual dimming controls that are operated by staff who do not understand how the system is supposed to work, and, as a result, introduce either or both of inappropriate lighting levels and sudden lighting level changes. Lighting control is of great importance to all kinds of “24/7” accommodation applications, and should be considered both from the total cost of ownership and human interaction points of view. The lowest capital-cost lighting arrangement is unlikely to give the most cost-effective or highest usersatisfaction result.
15.8 Hotels, hospitals and institutions
15.8.2 Corridor lighting
15.8.1 Introduction It may seem odd to bring together under one heading places as diverse as prisons and five star hotels. But in practice all the places covered by this section share a number of things in common. • they operate 24 hours a day, seven days a week. • their occupants are not in their own homes, but are to a greater or lesser degree transitory. • lighting operating and maintenance costs are a significant factor. • security and personal safety are important issues, implying the need for great reliability.
The lighting of corridors in large hotels and hospitals is a serious issue. For security and safety, lighting must be on 24 hours a day, but usually nighttime levels must be significantly lower than daytime levels both for comfort and economy reasons. Examples of how not to deal with the problem are given by those installations that switch off three out of every four lights down a corridor. This gives the worst possible result, as the glare from the remaining lights is uncomfortable, and the large patches of comparative darkness between them are potentially unsafe. The actual method used depends on the site. A first class hotel uses different lighting to a hospital, but the required functional result is similar. An ideal
459
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
arrangement is to have three lighting “scenes” for corridors: • daytime; full lighting level. • evening; a possible intermediate lighting level, while there is still considerable activity. • night; a low level without any “glare spots” for safe movement without the need to switch on lights. The majority of such lighting uses fluorescent sources, and the ideal is that they are fitted with controllable electronic ballasts. Any scene changes must be slow; a sudden change from night level to day level is disconcerting. Most such systems can run completely automatically from a BMS or timeclock arrangement. In some cases local over-ride must be available, for example in some areas of hospitals. Where there is a daylight contribution, a photo-sensor can be fitted to achieve “constant light” control on those luminaires that are installed in the daylit parts of the corridor.
Figure 15.67 Hotel and hospital corridor lighting should be under scene control. Fagerhult manufacture luminaires for hospital lighting that use controllable ballasts to provide lighting that is suitable for both day and night.
460
Where very low light levels are preferred, and where activity is intermittent, presence sensors can be used to modify the night scene. In a long corridor several sensors can be fitted down the length of the corridor that introduce a temporary increase in lighting level. Some hotel corridor lighting includes decorative elements, and may even use some incandescent lighting. Here maintenance costs can be an issue, so it is best to under-run the lamps slightly even at “full” to ensure maximum lamp life. The actual configuration is influenced by the emergency lighting arrangements (see Section 12.8). 15.8.3 Hotel rooms The lighting of a hotel bedroom or suite is no different than lighting rooms at home. It can vary from the utilitarian to the luxurious. Curiously it is only recently that the merits of good lighting control in this application have been appreciated. Many rooms still have multiple, individually switched, lamps where the occupant has to go round to each lamp individually to switch it on or off. At the very least, for both energy saving and convenience reasons, there should be a master switch at the door. Some hotels require the insertion of the room key into the master switch unit to help ensure that lights are switched off when the room is unoccupied (see Figure 10.6.) The same arrangement can save additional energy by being used to automatically turn down the heating or air conditioning. Occupancy sensors are not really practical for lighting control in hotel rooms due to the likelihood of nuisance operation; however they are sometimes used as a component in the safety/security system. First class hotels are now installing proper multiscene lighting controllers into suites and superior rooms. This has a number of advantages. It optimizes the use of energy and reduces the cost of ownership when incandescent sources are used, by increasing lamp life. Further, it allows the realization of good multicircuit lighting designs, and enhances the appearance of the room. Finally, it avoids the problem of the occupant having to work out for him or herself
ARCHITECTURAL APPLICATIONS
Figure 15.68 Guest suites in first class hotels, top private hospitals and luxury cruise ships have multi-scene lighting. Photo of a part of a suite on the Imagination from Kvaerner Masa Yards, Finland.
what each lighting circuit does. Hotel visitors often stay for only one night; they don’t have the time or inclination to fiddle around with separate switches and controls. 15.8.4 Hotel public areas Hotels are one of the most significant users of lighting control, since lighting control helps them make multiple use of their various spaces. Some of their main uses are covered in Sections 15.4 (meeting rooms) and 15.9 (restaurants and bars). The entrance lobby of a hotel makes a strong statement about the kind of establishment that it is. It may be a huge atrium, or a discreet panelled space indicating exclusivity. It includes public seating areas and lounges. Most have at least some element of daylight. Lighting control in these areas runs on automatic timeclock control, with the possibility of manual over-ride for special events and photo-electric input signal if there is a significant daylight contribution. This simply modifies the otherwise automatically selected scene. The idea of the main control is to make the hotel look as inviting as possible at all times of day. The main lobby area may not need to change, other than due to daylight contribution, but lounge areas benefit from a bright cheerful look in the main part of
Figure 15.69 The spectacular atrium lobby (above) and the lounge bar area (below) at the Fairmont Hotel in Dubai. Both use mulit-scene lighting control to match the time of day. Lighting design by DHA Design.
the day, and more subdued lighting in the evening. Transitions between scenes should be programmed to be imperceptible, using dimming speeds of the order of several minutes. 15.8.5 Hospital wards See also Section 16.4 on Healthcare. Hospital ward lighting has the following characteristics: • Where possible there is a considerable daylight contribution. Daylight is considered beneficial to patients’ wellbeing. • During the daytime and visiting hours ward lighting should be bright.
461
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
• At night a safety level of lighting is required throughout the ward. This may be provided centrally or as part of the individual bed lighting arrangement. • There must be no flicker. All fluorescent lighting should be based on the use of high frequency electronic ballasts. • Patients must have individual reading lights; while these must give sufficient light for reading for elderly patients, they must not be a source of discomfort glare to other patients in the same ward. • Individual patient lighting must be easy to operate, even by patients with some inability to move. • Depending on the type of ward, there may be a need for patient examination lighting at each bed. • Usually there is a nurse station associated with each ward, either within the ward itself or next to it. Ideally the station provides a good view of the ward. This introduces a conflict at night, since the nurse requires a working light level, but this can introduce uncomfortable glare for the patients. The above statements indicate that all forms of lighting within a ward should have individual variable control. Events within a ward are determined by some unpredictable factors, so any form of auto-
Figure 15.71 For those who cannot reach the wall control, possibilities include a wired or cordless “pad” control (left), or a different style of reading light with pendant switch (right). The switch has an LED indicator for location in the dark. Photos from Derungs Licht.
matic lighting programming is inappropriate. Lighting levels are adjusted either by the nursing staff or by the patients themselves. Depending on the room orientation and latitude, daylight may need controlling by curtains or louver blinds. Ideally the lighting around an individual bed should be controllable by the patient; the ward design may require additional general lighting which is controlled from the nurse’s station. Simple examination lighting can be provided by the reading light, especially if it has a variable output; more specialist arrangements are described in Section 16.4. Nurse stations require different daytime and night lighting which should be on separate dimmer control, operated by the nurse. 15.8.6 Secure institutions
Figure 15.70 The Amadea® luminaire from Derungs Licht uses both fluorescent and incandescent sources. It is intended for remote control. In its standard configuration it provides switch selection of general lighting, lighting for reading, lighting for examination and night lighting. Dimming control of the main lamp is an option. The accompanying bedside control provides lighting control through a backlit keypad. The unit also connects to the nurse call system, and provides both power and communication connections.
462
Secure institutions include prisons, jails and some types of accommodation for patients with severe mental illness. Lighting for the cells or rooms is designed to be comfortable and restful. However, the luminaires have to be designed to be tamper proof, for example by fully recessing, and this does place limitations
ARCHITECTURAL APPLICATIONS
Figure 15.72 In hospital wards the general lighting must be separately controllable from the individual bed lighting. Photo from Fagerhult.
on what can be achieved. The control of the lighting varies according to the degree of security: In high and medium security facilities control of the cell/room lighting is not available to the occupant. Control may be available by key operated switches outside the cell/room, and from a central security control point. In low security facilities the occupant has control, but central over-ride may also be fitted. Lighting control for corridors and other common areas has to take into account the need for surveillance 24 hours a day, so although night time levels are lower than daytime, the range cannot be as wide as is the case with hospitals and hotels. Switched lighting systems are used, controlled from a central security station. (See for example Section 12.6.1.) The system also controls exterior floodlighting when appropriate.
Figure 15.73 Correctional facilities require lighting control systems that can only be operated by authorized users, generally from a central security room. The company ILC (see Figure 12.11) has supplied control systems to many such facilities. Photos from Digital Vision.
as opposed to going out. However, it can be difficult to make a restaurant space inviting for both a cheerful breakfast and a late night supper. Lighting changes can do just this, and in the case of evening lighting can help make a space inviting regardless of the level of occupancy. (See also Section 12.1.) The prime requirement for lighting control in restaurants and bars is providing scene lighting for the different meals or times of day. Modern lighting design can result in the installation of many lighting circuits, and the manual adjustment of these is impractical. An automatic multi-scene controller is required. In most cases this can run on a simple timeclock basis with manual push button over-ride that allows either the modification of an existing
15.9 Restaurants, bars and pubs The lighting of restaurants, bars and pubs varies from the utilitarian to the flamboyant, and the subject could just as well be covered in the “entertainment” chapter. It was the hotels that took the lead in the use of lighting control in restaurants, because they were the first to realize that good lighting meant increased revenue. Hotels make real money if their guests can be persuaded to eat in a hotel restaurant
Figure 15.74 As a minimum restaurants and bars need scene lighting control to give the right look for different times of day. The Berns Salony in Stockholm has a fourscene, push button controlled system with long fade times. Two scenes are shown here.
463
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
scene, or the selection of an alternative scene. A common failing with this type of installation is the incorrect setting of dimming speeds. Unless the venue is some kind of “show” restaurant or bar, where the lighting is operated as part of an event, scene transitions should be programmed to be virtually imperceptible. Many installations correctly change the lighting level at, for example, 9 p.m., but do so suddenly with a dimming speed of no more than a few seconds. This is disconcerting for the occupants; it is much better to program changes to take place over a few minutes. Many restaurants and bars use lighting to make a statement about the kind of place they are. These may well need lighting control that is additional to the functional “time of day” scene selection. Examples are where color wash or feature lighting is installed or where other multi channel lighting is required to operate through a continuous cycle.
Figure 15.76 The Bar Room Bar, Harbourside, Bristol UK owned by Punch Devco is designed to appeal to the student market. Here the lighting is intended to be fun and transient, with a lot of color. 60 cold cathode lamps are used to edge light three inflateables; other lighting is by low voltage tungsten halogen and Encapsulite filtered fluorescent lamps. The programmed lighting sequence runs over the general lighting presets. Lighting design by Kate Wilkins for architect Buckley Gray. Lighting control based on a 30 channel Dynalite Dimtek system. Photo © Chris Gascoigne/VIEW.
15.10 Illuminated signs Figure 15.75 Sophie’s Steakhouse owned by Ruso Ltd has a lighting scheme designed around the bare light bulb, and depends entirely on sophisticated, but simple-to-use, lighting control to achieve the designed effect. The control allows the 40W long-life GLS lamps to appear like droplets of water, and/or to support a late evening lighting scene balanced around candle light. Even in a brighter daytime scene lamps are kept at around 40% to give the right appearance, and also to ensure that lamps are maintained at a low temperature in case they are accidentally touched. T5 fluorescent lighting and low voltage tungsten halogen lighting is used to support the background lighting level. Lighting design by Kate Wilkins for architect Michaelis Boyd Associates. Lighting control based on a Helvar Ambience 24-channel control system.
464
15.10.1 Static signs For the most part the control of outdoor electrically illuminated signs is simple. The majority are static and are controlled by standard devices such as timeswitches and photosensors. Where fluorescent lamps are used as backlights, it is important to ensure that the lamp/ballast combination is able to operate under all expected temperature conditions. It is for this reason that cold cathode lighting is often used as a light source. Cold cathode is also com-
ARCHITECTURAL APPLICATIONS
paratively easy to make into customized lengths and shapes. “Neon” or colored cold cathode lighting is widely used, but for many applications requiring color it is expected that it will give way to the use of LED strips mounted in diffusers. The wide application of electronic ballasts for fluorescent lamps and the introduction of LED based lighting, which necessarily requires electronic power conversion, means that it is now practical and cost effective to include a dimming facility on signs. While many signs are only illuminated at night, there are others that are illuminated in daylight hours. This can result in the sign being too bright at night; dimming ensures that the sign is not an uncomfortable source of glare, and that energy is conserved. 15.10.2 Animated signs
Figure 15.77 The Morimoto Restaurant in Philadelphia makes use of color and relief lighting in a scheme notable for the absence of visible fixtures, designed by Focus Lighting Inc of New York. A main feature is the use of LED strip lighting to illuminate the frosted glass booths. In order to avoid the strong LED colors adversely affecting the appearance of the patrons and their food, each table is also served by a table lamp and a tightly focused MR16 spot. The formed plaster walls are illuminated by closely spaced PAR30 lamps sited at the top and bottom of the wall. The middle of the wall is illuminated by ceiling and floor recessed MR16 fixtures. The lighting control is divided into two. The LED strips and their control are from Color Kinetics, and together they provide a slowly changing color scheme. While the programming system would allow for fast multicolor sequences, the cycle at Morimoto is based on slow transformations between six main colors on a two hour cycle. The photographs above show two of the color “scenes”. The incandescent lighting is run from a flexible 40 zone 4 preset control system from ALM Systems. This is a “traditional” system with manual slider preset controls and a single master slider for each zone. This allows the user to modify individual zones on an as-needed basis, although in practice daily operation is by push button scene selection from control panels at the Bar and at the Maitre’D position.
Animated electric signs have been popular since the advent of electric light, but their use in public spaces is tightly controlled. In Europe they may only be used in designated places and are generally forbidden where there is moving traffic. For many years simple animated signs used electromechanical switching devices – for example a motor driven camshaft actuating a set of microswitches or mercury switches. Such devices can still be found in use today, but in new installations they have been superseded by solid state switching devices. Sign manufacturers adopt different strategies for the control of animated signs. Many have developed their own hardware, others use off-the-shelf devices such as PLCs (Programmable Logic Controllers). Proprietary hardware is based on a central microprocessor with flash-EPROM programming. In a small sign the output drivers may also be centrally sited, but in a large sign the control signals are multiplexed to the output drivers (e.g. optically isolated triacs) sited near the loads. Text and image signs based on matrices of lamps are a special case. The simplest of these are arrays of incandescent lamps designed to show successive or scrolling text messages (and in the USA are known as message centers). Such displays have a long history, having been used in Times Square since the
465
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
1960s. When solid state switching was introduced, an economical method was to treat each line of lamps as a shift register, and simply feed an “on” or “off” signal as appropriate to the end of the line at the required interval. Now matrix switching is used whereby the output drivers are addressed on a row and column basis. The “ultimate” animated sign is the display that can show a full moving image. Prior to the arrival of the computer this was achieved by having an array of photocells, with each photocell arranged to switch a corresponding lamp. A 35mm movie film with the required moving image was then projected onto the photocell array. Such arrangements have given way completely to the idea that any moving image display should
Figure 15.78 Traditional “message centers” use arrays of incandescent lamps. Young Electric Sign Co (YESCO) now uses automobile style wedgebase lamps of 12W. Previously 30W lamps were widely used. Message sequences and their timing can be downloaded into the sign by modem. Where it is not practical to provide a wired control signal, RF is used.
early 20th Century. One way of achieving the “ticker tape” effect of sideways moving text was by moving a wide perforated tape, of reinforced paper or similar insulating material, across a bath of mercury. An array of spring wires, one for each lamp on the display, rested on the tape; when a hole passed under the wire, it made contact with the mercury, lighting the corresponding lamp. Such hazardous methods were still in use in the
466
Figure 15.79 Incandescent lamps are rapidly giving way to LED. YESCO is promoting the idea of LED “clusters” to replace incandescent lamps, both in new signs and as retrofits. Besides using a fraction of the energy of incandescent lamps, LEDs give new possibilities in respect of color and control and much greater life expectancy.
ARCHITECTURAL APPLICATIONS
Today’s full motion signs are, in effect, large TV screens. Each pixel (picture element) is a cluster of LEDs, red, green and blue. Modern LEDs allow the achievement of very high brightness. 5,000 nit is considered the minimum for outdoor displays, whereas 1,000 nit is sufficient for indoor applications such as covered malls. LED display screens are complex. Modulation of the light output is by PWM, with the display being refreshed at a multiple of the image frame rate (e.g. at around 180Hz instead of the original computer image’s 60Hz). High specification displays use 12-bit resolution to achieve satisfactory black level performance. Individual LEDs have widely varying performance, so it is necessary that each LED in the
Figure 15.80 A flamboyant example of the electric sign maker’s art from YESCO. This 90ft (27m) high sign for The Peppermill Hotel in Reno, Nevada, uses two 30ft × 20ft (9m × 6m) LED video screens. These are run from a dedicated control system that is compatible with standard video and graphics computer software. The rest of the sign includes several thousand incandescent lamps and no less than two miles (3km) of neon tubing on 100 circuits. These are controlled by YESCO’s proprietary microprocessor based controllers.
itself be constructed in a way that it is directly compatible with a standard computer image; for example VGA (Video Graphics Array) and its successors as used by personal computers. Displays using incandescent lamps based on this principle have been widely available, however, the thermal inertia of the filaments results in a limitation on the speed of animation.
Figure 15.81 The Fremont Street Experience in Las Vegas uses around two million incandescent lamps in a 1,400ft (427m) long canopy display. The dynamic show requires 121 standard computers to provide the images. The display was built by YESCO.
467
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
display is calibrated. Any required offset is then stored in non volatile memory so that the overall performance is uniform. Large LED based video screens are now an established part of special events, sports venues etc. Two different styles of construction are used. The products that have emerged from the professional
468
video industry are modular in construction, allowing for quick assembly and dismantling. On the other hand the products developed by the sign industry tend to be complete, individually fabricated, displays. The argument is that this not only results in a less costly display, but is actually more suited to long term permanent outdoor installation.
C h a p t e r 16
FUNCTIONAL APPLICATIONS
Functional
16.1 Retail spaces At this stage of the book the main concepts of lighting control have been explored, so in order to avoid repetition, this chapter does not seek to give an exhaustive description of lighting control in each application area, but rather to point up examples where the application introduces the need for special techniques. For many people “shopping” is now something quite different from obtaining the necessities of life.
applications
Shopping is now a social and entertainment activity, so shops and their environs compete for attention, often using lighting to do so. (See also Section 17.8). Retail display has always been a leader in the use of good lighting and in the exploitation of new light sources and luminaire designs. In many retail applications light levels are very high, operating hours are long and many luminaires are used. This leads to questions like: • how to provide flexibility in the lighting with minimum complexity for the user.
Figure 16.1 Charles Schwab at 300 Park Avenue is a retail banking space in New York City. Focus Lighting Inc were responsible for a lighting design that provides a clean bright look for daytime banking transactions, but also a striking statement “after hours”. This is achieved using two-circuit cove lighting based on high output T5 fluorescent lamps. One circuit uses lamps with color medium applied as “sleeves” (see Figure 3.20) for the night time effect. The lighting is controlled by a Powerlink G3 power panel control system. This provides relay control with overall intelligence. Because the retail activities involve variable opening and closing times, it is necessary to have a system that can be easily operated by staff, yet will not invoke the inappropriate night time scene by mistake. User input to the system is by color coded controls that are operated in a prescribed manner by staff who are “first in” and “last out”. The control system invokes the correct lighting look according to occupancy, season and time of day. After 11 p.m the system selects a security scene. The pictures show the side entrance lighting in “daytime” (left) and “out of hours” (right).
469
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
• how to reduce cost of ownership. • how to meet energy use targets. The figures accompanying this section give some examples of how lighting control can contribute to solving these problems. Figure 16.1 shows how, for many retail operations, it is important that the site looks good even when it is not open to the public. With variable hours and no staff permanently charged with operating the lighting, it is important in a case like this that the control is entirely intuitive in use. Figure 16.3 The Disney Store in Florence, Italy, is sited in an historic building where a frescoed ceiling prohibited the ceiling mounting of luminaires and demanded a special lighting solution. Lighting is all based on 12V 50W tungsten halogen with the lamps carried on tensioned wires fixed to the walls 8m (25ft) apart. In order to achieve long lamp life, all lighting is “soft started”, and run at 10.8V from electronically stabilized Multiload Voltmasters. The lighting consultants were Pinniger & Partners.
Figure 16.2 At the Grenville Library Shop at the British Museum merchandise is color coded by the color of the display case which, in turn, corresponds to the appropriate collection in the museum itself. A touch screen controller is used by staff to select the required color for each display. Color selection is based on primary color mixing. The lightsources are T5 fluorescent lamps with color filters and analog controlled ballasts. Lighting design by the Lighting Architects Group, control system by Dynalite.
470
Color is clearly understood by retail designers and their lighting consultants as making a great impression on customers. For some applications, such as the display of meat in chiller cabinets, the choice of lightsource is crucial, but, once made, does not require any special control consideration. However, when fashion, cultural or leisure goods are involved, there may well be a need to change color schemes on a seasonal or other periodic basis. Figure 16.2 is an example where shifting stock patterns require the lighting color scheme to change. Again the control is made intuitive for the staff. The cost of running lighting is not simply an energy cost. The cost of replacement lamps, and, more important, the labor cost of carrying out the replacement, can be as high or higher than the energy cost. Anything that can reduce the frequency and associated inconvenience of lamp replacement is to be welcomed. For some applications the arrival of CDM lamps has been welcomed as providing high efficacy with good life; however, they are not suitable for all applications and tungsten halogen lamps are the most widely used source when modelling and effect is required. Figure 16.3 shows an example of where control contributes to extended lamp life and reduced cost of ownership.
FUNCTIONAL APPLICATIONS
Figure 16.4 Bluewater, near Dartford, England is (at 2002) the largest shopping complex in Europe. For lighting control purposes it is divided into several zones, each of which has a separate control system for interior and exterior lighting. Lighting scenes are selected on the basis of time of day (morning, noon, evening, night) as modified by both the season and the exterior daylight lighting conditions. Touchscreens within each zone allow scene modification for special events and maintenance. For the most part the zones operate independently, but they are networked together for supervisory purposes. It was originally intended that the lighting system would in turn be under the control of an overall BMS, but this was found to bring no additional benefit, while needlessly adding to complexity. So in this case the lighting control operates as an independent system. Lighting design was by the Lighting Architects Group, with the control system being supplied by Dynalite.
A development of the latter part of the twentieth century was the mega shopping mall; a direct descendent of the city center covered arcades of the nineteenth century but on a much larger scale. Within them, each retailer will make their own effort to be noticed, but the mall operator will have an overriding interest in keeping visitors in the mall. Every extra minute that people spend in such places can be measured in money. One element that helps is the striking architectural environment and its lighting − which must necessarily change according to the time of day. Special consideration has to be given to the appearance when, for example, many of the retail operations are closed for the evening, but entertainment and dining locations are still open. Figure 16.4 shows an example of lighting control applied to large malls. Energy usage is of prime importance to very large stores, and stores that operate 24 hours a day. An example of how a team of client, architect, electrical
consultant and lighting consultant can work together to achieve significant savings is given by the Sainsbury’s superstore in North Greenwich, England, shown in Figure 16.5. This flagship store was built on reclaimed land, and right from the outset the aim was to build a store that would use approximately half the energy per square meter compared to conventional stores. This was to be done without compromising the merchandise lighting (or any other energy related environmental factor such as temperature). Lighting in supermarkets is intended to give a bright and attractive ambience, with high horizontal and vertical illuminances. Sainsbury’s specified a minimum 1,000 lux for the vertical lighting of the merchandise gondolas. In practice vertical illuminances throughout the store are between 600 and 2,000 lux in opening hours.
Figure 16.5 The Sainsbury’s Millennium Superstore, built on reclaimed land on the North Greenwich Peninsula in London, uses lighting control to help achieve low energy consumption. It is described in the main text. The Lighting Consultants, Pinniger & Partners worked closely with the Mechanical and Electrical Consultants, Oscar Faber & Partners and the architects, Chetwood Associates.
471
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
One of the ways of optimizing vertical illuminance was to make sure that the building itself contributed to it by using reflective surfaces where possible. For example the floor is terrazo with 73% reflectance. Maximum practical use is made of daylight, using northlights. These provide a horizontal illuminance of 2,000 lux when the sky is overcast. At night time the northlights would act as light absorbers, and would create considerable external light pollution, so they are fitted with reflective shutters that operate automatically so that, after dark, internal light is reflected back into the store. When daylight is not sufficient for general lighting, continuous runs of fluorescent lighting come on under photo-electric control. Over the majority of the store these are based on the use of the 28W high efficiency T5 lamp, but over the entrance areas and where financial transactions are carried out, the high output 54W lamps are used. The high efficiency T5 lamps are also used as the main basis of the gondola lighting, with the luminaires being mounted on the gondolas themselves. These are under programmed switch control, and are turned off in store closing hours. However, even then the store requires a minimum horizontal illuminance of 200 lux throughout for cleaning, maintenance and re-stocking and this is provide by the overhead lighting. At the time the store was built, the conventional supermarket in the UK used approximately 45−50W/ m 2 for lighting. The Sainsbury’s Millennium Superstore uses only 24W/m2.
ticulture, especially in cold climate areas where day lengths vary. The reaction to light is not confined to the direct response to the radiation, but is also time dependent. Most species are subject to biological rhythms that may either be inherent in the organism or are determined, or partly determined, by geophysical cycles. The most important of these is the 24 hour day-night cycle which affects the circadian rhythms. Tests on animals and human beings have demonstrated that we all have an inbuilt rythm that, left free running, is not exactly based on a 24 hour cycle. However, in the presence of the daylight/night sequence we “lock on” to the 24 hour cycle, largely owing to light. Transferred to agriculture and horticulture, this fact leads to some interesting consequences, that are best illustrated by example. Fundamental to plant life is photosynthesis. This is the process by which light provides the energy necessary for the conversion of carbon dioxide and water into carbohydrates by chlorophyll. The carbohydrates
16.2 Agriculture and horticulture Light affects all living flora and fauna. More strictly, electromagnetic radiation does so, because the response of various species varies quite significantly with wavelength. This applies just as much to sight, where some species can “see” well into the IR region, as to other responses, where plants or animals are affected by both visible radiation and UV and IR. A consequence of this is that lighting plays an important part in some aspects of agriculture and hor-
472
Figure 16.6 Example of lighting being used to promote photosynthesis in commercial tomato growing. Such lighting is under automatic control to make the best use of daylight, and to optimize plant maturity to match market conditions. In this case the luminaires are Hortilux DEEP reflectors with remote ballasts sited next to the plants. The system results in less shadow, and any ballast heat helps promote plant growth. Photo courtesy Hortilux Schréder.
FUNCTIONAL APPLICATIONS
are essential to the plant’s growth. Many plants require a minimum light level to grow at all, and most can be stimulated to greater or faster growth by additional light. Many growing houses use lighting to promote photosynthesis, for example to grow salad vegetables in northern countries in the winter months. Here, as in most agricultural and horticultural applications, the choice of the light source is critical. For this reason many different lighting combinations are used, in order to achieve the spectrum required for the particular application. Tungsten, fluorescent and HID lamps are all used, and special versions of these lamps are produced to match particular crops. In intensive commercial or research growth rooms high radiation levels are required, equivalent to 5,000 to 100,000 lux or more. The preferred metric is photosynthetic photon flux density or PPFD which is a measure of the photon flux in the wavelength range 400−700nm. Where a precise flux is required the installation is made up of uniformly spaced lamps on separately switched or dimmed circuits. Manual or automatic control is used to maintain a desired flux density and to compensate for flux reduction through lamp aging. Plants are characterized by photoperiodism. The ratio of light period to dark period determines the time needed for flower plant to set buds, and then to bloom. Commercial growers are able to get flowers to bloom out-of-season by using artificial light (not necessarily in a greenhouse, it can also be done in the field) to vary the ratio. Further, by fine tuning the lighting timing, it is possible to delay or speed up the blooming process to match market demand. The case of poultry is particularly interesting. The trigger for the growth of ova is light of specific wavelength received by the chicken’s eyes. It is found that maximum egg production is achieved when the illuminance of the chickens is around 10 lux − additional lighting is counter productive, since then activities like pecking each other set in. Chickens require a light cycle, and optimum egg production is achieved by a 24 hour cycle with 14 hour daylength. For growing chickens it has been found that a daylength of 11 to 12 hours is optimal. This has the effect of slowing down the time to maturity, which in turn leads to the chicken producing larger eggs
without affecting the number of eggs it goes on to produce. Other branches of animal husbandry, including pig rearing, terraria and aquaria make use of lighting that, for the benefit of the animals concerned, must provide correct day/night cycles (see also Section 15.7.3). Nearly all such systems require automatic control, mainly based on time control. Where daylight is present photo-electric control is also used. Level control is usually by circuit switching, although some installations use dimming.
16.3 Manufacturing processes Most manufacturing processes only require light for the human participants to see by. Therefore lighting control has the same part to play as in other workplaces (See Section 15.3). There are some industrial processes that use electromagnetic radiation as part of the process, and these are primarily concerned with drying, heating and curing. The main radiation used is IR. Many processes now use IR lamps as described in Section 5.8. Process examples include paint baking, ink drying, preheating of plastic for forming, heating of metal parts for shrink fitting, and materials dehydrating. The attraction of using IR lamps for these applications is their controllability. Because they can be controlled like a tungsten lamp, they can use the same control techniques, and because of their rapid
Figure 16.7 IR is used in many industrial processes. The left photo shows an IR paint drying system. Most such systems use thyristor control equipment to regulate the output, an example of which is shown in the right hand photo. Photos courtesy of Advance Infra Red Systems Ltd.
473
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
response, can respond to sensors to maintain accurate temperature control. The use of controlled radiant heat is also energy efficient, so that, for example, a paint drying oven is only “on” when there are parts to be dried. Photochemical processes generally use UV sources (see Sections 3.6.2 and 5.7). Some of these, like plastics polymerization and water purification, require continuous radiation, others, like UV lithography require time controlled exposure. Control of the UV source lamps is the same as that for their “visible” counterparts.
16.4 Healthcare See also Section 15.8 that refers to lighting control in hospitals. Special types of lighting are used in healthcare for functional, therapy and cosmetic reasons. Most require some form of control to ensure that the task can be carried out or that safety limits are not exceeded. Patient examination lighting requires higher than normal light levels. Within hospital wards it can be provided as an adjunct to the normal bed lighting, but is separately controlled. In critical care rooms a wide range of lighting levels is required. Main lighting should be dimmable; but each patient bed may be equipped with separately controlled surgical task lights for emergency procedures. Surgical task lights are highly specialized luminaires. They are required to provide high and uniform illuminance and a typical specification is: • counterbalanced ceiling mounting, providing complete flexibility in the positioning of the lamp. • illuminance up to 135,000 lux at the working area which is typically 150–300mm diameter. • minimum practicable infra-red output, especially in the range 800–1,000nm. • continuous adjustment of light pattern area. • 50cm depth of field, so that it is not necessary to refocus the lamp when working at different levels within a surgical cavity. • wide diameter multifaceted reflector system that tolerates a large proportion of the projected light being blocked before any shadow is seen.
474
• suitable for use within the sterile field. This last requirement means that any control adjustments available to the surgeon must be sterile. These include the positioning and focusing adjustments and, sometimes, a means of varying the intensity. It may also be possible to make the intensity adjustment remotely. Light is used for therapy in two ways. Phototherapy is the technique of using light of specific wavelengths to treat disease. Examples are: • treatment of jaundice in new-born children, using visible radiation in the 450nm region. • treatment of skin diseases using UV-B. • treatment of tumors using radiation around 630nm. In all cases the irradiance dosage is critical in terms of effectiveness and safety. Control is semiautomatic using timers and sensors. SAD Therapy is light therapy for ameliorating Seasonal Affective Disorder, a depressive condition that affects a significant proportion of the population in latitudes with short winter days. Although research has still some way to go in quantifying the effect, it has been established that illuminance of
Figure 16.8 Surgical task lights must provide very high illuminance, and must be easy to control by the surgical team. The CHROMOPHARE® range from BERCHTOLD has a sterile handgrip to provide control of focusing and intensity. The luminaires include automatic lamp changing; if the primary lamp fails, its replacement moves into place instantly.
FUNCTIONAL APPLICATIONS
16.5 Simulation
1.6 1.4
2
2,500 to 10,000 lux for periods of 30 minutes to four hours provides a therapeutic effect. It seems that it is the light received by the eye that triggers the response. This is leading to the introduction of additional lighting in private homes, offices and hotels to provide the lighting required for SAD Therapy. Such lighting may be under simple automatic time control, or may be more elaborate, such as automatic dimmer based “dawn simulation”. Light may also be used for cosmetic reasons, the most common application being the tanning lamp (See Section 3.6.2). The most important control aspect of these is ensuring that sessions are automatically timed so that users do not receive an overdose of radiation.
Spectral irradiance W/(m .nm
Figure 16.9 Additional lighting for treating seasonal affective disorder may now be found in locations such as hotels, in addition to specialized clinics.
do this in a cockpit that is a re-creation of the real thing. The simulator cockpit lighting and its control can also be real (See Section 16.7.1). Another type of simulation is solar simulation. This is where there is a requirement to create radiation that corresponds to sunlight. Such simulation is widely used in the defense industry to check the performance of military hardware under laboratory conditions. It is obviously much easier to make measurements all over a tank or an aircraft if it is stationery, rather than being in the middle of a desert or flying at 50,000 feet. In order to create a valid simulation, it is essential that the spectral power distribution and the intensity of the radiation matches that of the sun at the required altitude. At the edge of the earth’s atmosphere, energy is received at an average rate of 1,350W/m2, at a color temperature of 6,500K. By the time the radiation reaches sea level, 25% of the energy has been lost, and the spectrum has been modified by absorption, principally by oxygen, water, ozone and carbon dioxide, as shown in Figure 16.10. A solar simulator must not only create the same visible effect as sunlight (which can exceed 100,000 lux) but must also provide the same spectral power distribution. In practice this can be achieved by tungsten halogen lamps augmented by UV lamps (see, for example, Section 5.7). Another method is to use special HID lamps that have a spectrum that matches
1.2 1.0 0.8 0.6 0.4 0.2
400
Simulation is used when it is not practical, safe or economic to carry out a procedure “for real”. The best known example is that of flight simulation, where pilots learn how to fly civil and military aircraft. They
800
1200
1600
2000
Wavelength in nm
Figure 16.10 The spectral power distribution of solar radiation at sea level. The irregularities in the curve, which would otherwise be like a blackbody curve, arise due to the absorption bands of molecules in the atmosphere.
475
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
16.6.1 Small boats
Figure 16.11 A dimming system supplied in the 1980s for controlling simulated solar radiation in two military vehicle test chambers. For “military vehicle” read anything up to the size of a tank. The total load of the system is 1.44MW of tungsten halogen and IR lamps. Photo from Electrosonic Ltd.
the UV and visible requirements, augmented by some IR lamps. There is usually a need for control with solar simulators − items under test may require a radiation cycle to check the effects of heating and cooling, and it would be unrealistic to run the full output continuously. With the tungsten halogen system, control is easy to achieve using high power dimmers. With the discharge lamp system, control is less precise because of the difficulties of dimming HID lamps. It can be achieved by a combination of circuit switching and power reduction. While the main example given here is military, the same principles are used on a smaller scale by many industries that need to test their products’ sensitivity to solar radiation.
16.6 On water In this section two extreme topics have been chosen to illustrate how the control of lighting on ships and boats differs from that on dry land.
476
Anyone who owns or who has borrowed or hired a boat, becomes aware that boat “electrics” are a mysterious branch of technology, and the source of much despair. Many of the problems encountered are due to the way in which the wiring is installed, often done without labelled and color coded wiring and without benefit of a proper wiring diagram. These problems can be solved by discipline and by following the recommendations and regulations of the American Boat & Yacht Council (ABYC) but there remain some issues that are raised by the boat environment – water and electricity are a troublesome combination. Most small boat electrical systems are based on 12V battery power because this allows the use of automobile components and readily available low voltage luminaires. It is easy to allow a battery to go flat, so many boats use a dual battery system, with one battery reserved for engine starting duty only. Multiple battery installations use diode blocks to prevent the discharge of one battery into another.
Figure 16.12 The solar simulator installed at the Boscombe Down laboratory of the Defence Evaluation Research Agency (UK) being used to test the outer surfaces of the Eurofighter. Main lighting is by 60 Thorn CSI Series floodlights fitted with Tridonic HID ballasts and special 1kW compact source HID lamps from GE. These are augmented by 250 IR lamps. Control is by a system supplied by Hasset Industries, and is based on a combination of ballast switching and variable autotransformer control. The system achieves an average horizontal illuminance at the canopy level of 194,000 lux and a total irradiance level of 1,200W/ m2 with a uniformity of >0.85. Photo courtesy Thorn Lighting Ltd and Tridonic Ltd.
FUNCTIONAL APPLICATIONS
Boat lighting should have its own distribution board, and should use DC circuit breakers for protection, although in practice automobile fuses are also used. Navigation lighting should have its own circuit(s) and switch panel. Small fluorescent lamps with built-in inverters are widely used for general lighting because of their high efficacy. They are suitable for illuminating most areas, especially galley, heads and engine space. They are also suitable for general saloon and cabin lighting if the right color temperature lamp is used, but for appearance (and lower capital cost) many boaters prefer incandescent sources, mainly using tungsten halogen lamps. At 12V there is significant volt drop on even low powered lamps unless thick cables are used. While fluorescent inverters can compensate for this, the directly fed incandescent lamps cannot. For lighting it is accepted that there may be a drop of up to 10%, but this is inefficient. On the other hand a fully charged “12V” battery actually gives around 14V, so some drop, though wasteful, is necessary to avoid short lamp life. A better, but significantly more expensive, solution is to provide dimmer control for the lighting, arranged to ensure that the lamp’s nominal operating voltage cannot be exceeded. Because the supply is DC, the practical method of electronic control is to use PWM.
Figure 16.13 Example of a galvanic isolator based on the diode principle that meets ABYC requirements, from Sterling Power Products.
Many boats now have dual wiring systems; one at 12v (occasionally 24V) DC, and the other at mains voltage AC. The AC power distribution is fed from an inverter working from the boat batteries, or from shore power. Some inverter designs are combined inverter/charger units, so when shore power is connected, the batteries are charged from it rather than from the engine. In small boats the inverter is normally used for heavy “domestic” loads such as microwave ovens and TV sets, but there is no reason in principle why lighting should not also be run from AC – however there is little point since operation direct from the batteries represents a more efficient use of scarce battery power. The combination of boats with metal hulls or metal fittings, water, especially sea water, and electricity is dangerous and potentially damaging to the boat. Even when no conventional electricity is present, there is a problem of electrochemical corrosion; the problem is much worse when batteries and shore power are involved. The corrosion problem arises because of electrochemical potential differences between different metals. A “battery” can be formed by placing two dissimilar metals in an electrolyte. Sea water makes a good electrolyte, but so does any water other than distilled water, so the problem exists just as much for canal and river craft. Dissimilar metals arise in many different ways, including steel hulls, bronze propellers, stainless steel fittings, brass, copper and aluminum. If the two metals are connected, either deliberately or by accident, a current flows and this results in the depletion or corrosion of one of the “electrodes”. Metals and alloys can be placed in a galvanic series to show their propensity to corrosion. As a practical guide the ABYC list a range of nautical metals and alloys and give them each a corrosion potential measured in volts. This is done with reference to a silver/silver chloride standard electrode, given the value 0V, and a sea water electrolyte. In this series mild steel stands at -0.6V, Aluminum bronze at -0.35V, brass at -0.32V and zinc at -1V. When several different metals are immersed in the same electrolyte, there is depletion only on the one with the highest (negative) potential. Boats are, therefore,
477
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
fitted with sacrificial anodes. These are zinc blocks, plates or annular rings which protect metal hulls, propellers, and metal fittings on GRP boats from corrosion. Corrosion due to dissimilar metals is slow because the potential difference involved is small, usually no more than 0.6V. However, when another source of electricity is around, unexpected stray currents can arise. Such currents can cause corrosion even when the current is between “electrodes” of the same metal. A submerged bilge pump with a faulty terminal could apply 12V to the bilgewater and thence to another metal fitting or to the metal hull. A voltage this high creates a high current and fast corrosion. Within boats a technique of bonding all metal parts to a common ground plane may be used to short circuit currents of this kind. The connection of shore power, and the presence of neighboring boats, creates more opportunities for stray currents and galvanic corrosion. In particular the presence of two shore power cables on neighboring boats can lead to corrosion on one of them. The solution here is to provide galvanic isolation between the boat and its shore supply. This is best done with an isolating transformer, but such an item is big and heavy, so often a pair of back to back diodes are used. These can pass the AC line voltage, but because diodes have a forward voltage drop, they can block the small DC voltage. Usually two diodes are used in each direction to ensure a block of at least 1.2V. 16.6.2 Cruise liners 16.6.2.1 Problems of ship power The cruise industry has undergone a phenomenal expansion, resulting in the building of huge ships of strange construction. The modern cruise liner combines the attributes of a luxury hotel, theme park, shopping mall, casino, and show lounge − and that is just what the passengers see. While, in principle, each aspect of the ship can use lighting control of the same type as used on dry land, there are some problems peculiar to this environment. Modern cruise liners use electric propulsion. This
478
is highly flexible, since it provides precise control of speed and propulsion power; and also allows “thrusters” to be installed round the hull to assist manouverability. When electric propulsion was first used, the practice was to have separately powered generators for the propulsion. Now the arrangement is to have a single “power station” within the ship that provides power for all the ship’s services as well as the power for propulsion. A typical arrangement is to have a number of diesel engines each driving a synchronous generator. A big ship may have as many as six or more of these, each with an electrical output of around 8.5MVA at 60Hz. When in port only one may be used, but when cruising four or more may be on line. A dual 6.6kV main bus bar system may be used to provide diversity of supply. The propulsion motor control converters are connected directly to the main busbars, as are 440V transformers that distribute main power to services throughout the ship. However, in order to provide a more stable supply for those circuits providing power for computers, navigation equipment, show equipment, outlets in the passenger cabins and dimmer controlled lighting, motor generator sets are used to provide a 220V 60Hz supply. “Stability”, is, however, only relative. Table 16.1 compares permitted variations on land and sea. Overall there is a nasty cocktail of adverse conditions that apply: • wide voltage variations • wide frequency variations • high impedance supplies
Item
Land
Sea
Voltage, permanent Voltage, transient Frequency, permanent Frequency, Transient THD
+10% -6%
+6% -10% +20% -15% +5% -5%
+1% -1%
+10% -5% 5%
8%
Table 16.1 Comparison of permitted supply variations. The “land” figures relate to the UK domestic supply, the “sea” figures relate to Lloyds Register of Shipping Rules and Regulations for the Classification of Ships. Unfortunately the duration of “transient” changes is not defined, and in practice they can last for several seconds. In both cases the Total Harmonic Distortion (THD) specification limits the contribution that any one harmonic can make.
FUNCTIONAL APPLICATIONS
0.2 Light level (lux)
0.19 0.18 0.17 0.16 0.15 0.14 0.13 0.12
0
0.05
0.1
0.15
0.2 0.25 Time (s)
0.3
0.35
0.4
0.45
Figure 16.14 Measurements of light level in the dining room of a major cruise liner, with the incandescent lights at a low dimmed setting. The high frequency (120 Hz) variation is to be expected, would apply however good the supply, and is not visible to the eye. The problem lies in the superimposed variation at around 4.25Hz which is visible. It arises because of a corresponding variation in the supply frequency, created by oscillations in the flexible coupling between the diesel engine and the generator. In this and similar cases it was necessary to replace the original dimmers with dimmers that could eliminate the effect of the rapidly varying frequency.
• high harmonic content The last condition is actually worse than implied by the table, since many of the loads, including dimmers and some power supplies, create harmonics themselves. Unfortunately any substantial variation in the main supply does get transmitted through the motor generator sets. For example, if the ship is running on a constant power regime, the propeller speed can vary according to its depth in the water, which varies if the sea is at all rough; this in turn causes the generator frequency to vary. In one particular case, Figure 16.14, it was found that even when the ship was in port there was a cyclic variation in supply frequency. This was found to correspond to the frequency of torsional oscillation within the flexible couplings between the diesel engines and the generators. The combination of variable frequency, high harmonic content, noise and high impedance supply leads to difficulty in determining the zero crossing − the combination of adverse factors can even lead to the creation of multiple zero crossings. In practice this can result in many conventional dimmers, especially low cost multi-channel entertainment dimmers, being unsuitable for shipboard use.
The aim of a dimmer is to ensure the energy delivered to the load is the same for each half cycle. Two methods are used to achieve this in the presence of adverse conditions like those on board ship. In dimmer systems using “dumb” dimmer modules and a sophisticated multi-channel central firing pulse generator, the output of a filter/comparator is subject to fast digital processing to integrate over the half cycle, and thence to determine the zero crossing. The possible problem with this method is that the measurement is made at one position within the rack, and may be referenced to only one phase. Where more sophisticated individual dimmer units are used, it is possible to include active analog filters, followed by digital processing, on a per channel basis. In practice this method has been found to overcome most of the problems encountered on board ship. The extent to which the power systems on board a ship may affect lighting control depends very much on the electrical system of the ship, and these vary considerably as advances are made in power control. New vessels are being equipped with reactors to minimize the effect of propulsion changes, but it seems that all important public area and entertainment lighting will continue to need special dimmers. Even at the individual cabin level, dimmers with sophisticated zero crossing detection are required. Where fluorescent lighting with controllable electronic ballasts is being used, there is less of a problem because the boost converters used have excellent regulation and are generally tolerant of frequency changes. 16.6.2.2 Automatic control A cruise liner has thousands of public area lighting circuits, and the only practical way of controlling them is through an automatic control system. Crews are constantly changing, so any system must be intuitive to use and easy to maintain. Such systems may provide for automatic scene change at set times, with provision for local override. The problem here is knowing what time of day it is. The ship may be crossing lines of longtitude, and its latitude will determine dawn and dusk times
479
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 16.15 All public area lighting on large cruise liners is under programmed control. Ideally the system also monitors each circuit. (See also Figure 14.21.) This is the pool deck on the huge “Brilliance of the Seas” seen at night. Photo courtesy Project International.
depending on the time of year. Thus any clock system not only has to take into account seasonal variations, but also the position of the ship. This can be done either by having a GPS receiver as part of the lighting control system, or by deriving the required information from other shipboard systems. An essential feature of large shipboard systems is that of monitoring. The system monitor must be
Figure 16.16 The scale of large cruise liners is enormous. This is the central arcade on the “Adventure of the Seas”. The lighting is programmed to achieve a different appearance at different times of day – here an evening setting is using neon to achieve a color wash. Photo courtesy Project International.
480
able to display the entire lighting configuration; a newly arrived engineering officer will not know the details of the system layout, but when he needs to know it, he will be in a hurry. The monitor should also be able to report on the status of all circuits. This can be extended to include fault reporting, such as dimmer over-temperature, and dimmer load changes such as may arise due to failed lamps. Commissioning and programming these systems has to be carried out very quickly, and often at sea, such is the financial pressure to get large ships into service. The process is aided if the lighting in each area can be programmed from a laptop computer that is connected to the lighting control network; and which includes a graphic visualization of the space and its lighting provision. The program used must allow for the creation of accurate records of both system configuration and actual scene settings. The general lighting on a ship is governed by regulation. So while most of it can be controlled by an extension of the main system, using relays and contactors, there are special “fail safe” considerations to do with both normal and emergency lighting. The method of wiring these varies according to the authority concerned, and generally differ from dry land practice. Some spaces on a ship have to do multiple duty. For example a lounge space might be used as a relaxing area by day, but become a show or dance area
Figure 16.17 Programming the lighting on board ship is speeded up if there is a graphic display of the space and lighting circuit arrangement. Picture courtesy Quo Vadis Ltd.
FUNCTIONAL APPLICATIONS
at night. This in turn might mean that for most of the time it is best served by an architectural lighting control system, but at night needs an entertainment system. Until recently the tendency would be to use a separate system for each application, but the arrival of lighting control routers able to deal with multiple protocols and carrying media (e.g. DMX 512a, DALI, 802.3, 802.11b, and proprietary protocols on EIA485) may lead to a greater integration of the entertainment and architectural lighting control.
16.7 In the air As in the previous section two examples are chosen to illustrate how lighting connected with flying uses the same lighting control concepts as conventional applications, but introduces a number of special techniques. 16.7.1 Aircraft interior lighting 16.7.1.1 Aircraft power Large passenger carrying aircraft use a mixture of power supplies. The main supply, used for the majority of lighting, is nominally 115V 400Hz derived directly from three-phase generators powered by the main engines and by the Auxiliary Power Unit (APU) a small auxiliary jet engine that is used for providing power when the main engines are off. In practice the power varies between 97−134V and 360−440Hz, and this must be taken into account when designing lighting control systems. In addition there is a 28V DC supply, itself derived from the 115V AC through transformerrectifier units. This may also be used for some lighting and lighting control electronics. Finally emergency lighting operates at 6V DC. A number of 6V charger/supply units with battery back-up are located throughout the aircraft. The power input and output wiring and location of these is specified by regulation to ensure that a minimum of emergency lighting is affected in the event of any individual wire break or equipment failure. Lighting within the aircraft is zoned, to correspond to the class of seating, to match galley posi-
tions and to provide sensible loads for protection. Protection is by circuit breakers rated at a maximum of 15A. As an example a “standard” Boeing 747 aircraft might have five zones, each with separate 115V ceiling wash, 115V sidewall wash, 28V DC reading lights, 28V DC utility lights in the toilets and galleys and 28V DC lights for door areas etc. 16.7.1.2 Light sources The primary cabin lighting is by linear fluorescent lamps. In commercial passenger aircraft both T8 and T5 tubes are used, and the aim is to use standard tube types that can be easily sourced at a reasonable cost. Tri-phosphor lamps are preferred, and if colored lamps are required, filters are used. Cold cathode lighting is also used, but today the emphasis is on using standard light sources to minimize cost of ownership. Reading lights and other spotlights are tungsten halogen, usually 28V DC 11.5W. Sometimes 12V AC (derived from the 115V supply) is used for halogen lamps. Higher wattage lamps up to a maximum of 50W may be used for feature or door area lighting. LEDs are finding increasing acceptance in aircraft lighting, but there is still some doubt about “real” LED life. They are being used for color wash lighting, for close reading lights and escape path lighting. Low power HID lamps are used as fiber-optic illuminators (LED and tungsten halogen are also used for this purpose) with the fiber optic lighting being used for some reading lights and for decorative lighting. Electroluminescent lighting is used for escape path lighting and for indicating the location of emergency escape hatches. It is evident that the majority these sources need electronic ballasts or transformers to interface to the available power. While these are similar in principle to their earth bound equivalents, the differences are: • the need to match aircraft power supplies. • the need to meet stringent EMC requirements (see Section 16.7.1.4). • the need to be as compact and light weight as
481
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
possible, while at the same time conforming to industry standards in respect to method of construction and materials. • in most cases the need to incorporate the means of light level control. In respect of the last item, fluorescent ballasts use the principles described in Chapter 7, and usually offer a dimming facility. Many have logic control
Figure 16.19 Premium class cabin layouts call for more ambitious lighting schemes. This is the bar area in “upper class” on a Virgin Atlantic plane. Picture from AIM Aviation.
(see next section) but some use analog 0−10V control. Lamps supplied by DC (and LEDs) use PWM for level control. 16.7.1.3 Control
Figure 16.18 Ceiling wash lighting in the premium classes of passenger aircraft may provide color options as shown in the above two pictures. This is achieved using triphosphor fluorescent lamps with colored sleeves as shown in the lower photo. Pictures from AIM Aviation.
482
Large passenger aircraft were one of the first users of multiplexed lighting control. It was introduced with the 747 because of the distance between passenger seats and the lights. It was no longer practical for passengers to simply reach up to switch a reading light − so while on smaller aircraft it remains possible for the reading switch to be sited next to the light it operates, on large aircraft it is not. The weight penalty of using separate wiring for each lamp would be huge. In large aircraft all cabin services including cabin lights, reading lights, signs, attendant call, and oxygen mask monitoring and release are under electronic control by the cabin management system or CMS. A shielded EIA485 data bus runs through the aircraft and links together the cabin controllers and the requisite number of interfaces. Modern cabin control units are based on touch panel displays of one kind
FUNCTIONAL APPLICATIONS
Figure 16.20 The “Mood lighting” control panel in Virgin Atlantic aicraft selects lighting according to the flight phase and exterior lighting conditions. Picture from AIM Aviation.
or another, for example membrane switches or LCD touch panels. The load interface units are referred to as overhead electronic units or OEUs. The two major manufacturers are Boeing and Airbus. They each have their own bus which they keep proprietary. However, the specifications of inputs and outputs to the OEUs and similar node equipment is open, so provided independent vendors match the specification, they can supply the actual lighting devices and some input devices. Output ports on the OEUs may be switched AC, switched DC or at logic level. (“Ground logic” may be used where logic 1 is at ground potential and logic 0 is floating.) In many aircraft the CMS control panel offers BRIGHT, DIM 1, DIM 2 and NIGHT for each lighting zone. These instructions are sent to the OEUs and in turn, via the logic control, to the lighting ballasts. The ballasts used for the lighting have three logic level connections that allow for eight binary control states to call up preset levels or instruct the ballast to conduct a self test. A ballast reports the test result by setting a fault status bit on one of the OEU inputs. The move from one level to another is somewhat abrupt. Long haul planes are fitted with personal or cabin in-flight entertainment (IFE) systems. In many cases the hand control or seat mounted control combines the IFE controls with the lighting and cabin service controls. However, they are two quite separate systems. The cabin service controls are multiplexed to
the CMS, and the IFE data and signal streams run separately. However the two systems may share a seat disconnect. Recently there have been some major developments in cabin lighting control. They have arisen partly because airlines have been upgrading their premium class seating, and partly because there is now a better understanding as to how lighting can contribute to a more comfortable and stress-free journey. This has led to the introduction of stand-alone cabin lighting control systems that allow much greater flexibility in control (while retaining interfacing to conventional CMS to ensure compliance with mandatory control related to flight operations or emergencies). AIM is a specialist group of companies providing complete aircraft interiors and their components. Within their portfolio they offer an independent lighting control system with the following significant characteristics: • ability to control any reasonable number of lighting channels (typically 30−50). • each channel can be individually programmed in respect of level, fade time, and color (if appropriate) for each scene. • level control 0−100%, but source dependent. Significantly special fluorescent ballasts have been developed to give reliable dimming down to 0.5%, such low levels are essential in establishing night time
Figure 16.21 Modular multichannel lighting controller from AIM Aviation. This controls the lighting ballasts; and itself is controlled by a remote scene controller. The data link between the two is designed to produce a minimum of interference by using modified sinewave data signals instead of square wave.
483
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
scenes or in providing a smooth fade to black. • fade time control from 1 second to 20 minutes. It is now appreciated that sudden changes in light level are disturbing, and are particularly inimical to sleep. In this connection the sudden appearance of high brightness sources in an otherwise darkened area also prevents sleep − such as might arise when a door to a brightly lit toilet is opened, or when a nearby reading light is switched on. For this reason work is continuing on ensuring that lighting in other spaces (such as toilets and galleys) follows the main scene requirement. But of course if a door is then shut, it is acceptable to raise the lighting level in the space. However, ideally, it should be to a lower level than in daytime, to maintain dark adaptation. The AIM system is, therefore, similar to sophisticated ground based architectural lighting controls. A lighting designer can prepare the basic lighting programming in advance on a computer; tweak the settings in situ as part of the initial commissioning process, and then have the same program down loaded into the systems on other aircraft in the same fleet. 16.7.1.4 EMC considerations Most readers will be familiar with the injunction to “switch off all electronic devices” at take-off and landing. Modern passenger aircraft are totally dependent on electronics, especially with the arrival of “fly by wire”, and it is vital that electrical equipment causes no significant interference. This applies to the lighting and all its associated ballasts and control equipment. In principle the EMC requirements are similar to those reviewed in Section 8.5.3, but they are considerably more stringent. An industry standard used is referred to as RTCA DO-160D, and its equivalent of Figure 8.38 sets the limit above 2MHz to 40 dBμV and extends it through 100MHz up into the GHz region. There are also some specific frequency bands where lower levels are specified. This leads to a particular problem when using digital signals (which have a high harmonic content) to control remote devices, and in practice means that
484
screened control cables must be used. Fiber optic communication is also used to get round the problem. Screening adds weight. In an effort to eliminate this weight problem, and to allow the use of simple UTP cable, AIM Aviation have developed a patented method of converting their digital lighting control signals into modified 50kHz sine wave signals. These are converted back to square wave at each receiver. The system produces negligible interference. Cockpit lighting is a separate subject. Lighting levels must be matched to the exterior daylight, or lack of it. The main issue is the lighting of instrumentation, which must be dimmable over a very wide range. Today most instrumentation is light emissive, using LED, CRT or CCFL backlit LCD. It has to be viewable when incident light levels from daylight may be as high as 100,000 lux. Control is manual. 16.7.1.5 Military and small aircraft Most military and small aircraft have comparatively simple lighting control. This results in the use of electronic ballasts and transformers that provide simple dimming facilities by direct remote control (multiplexing is not used). In military aircraft the main emphasis is on cockpit lighting. “People carrying” and military cargo aircraft use simple switched control in the main fuselage, with the exception of command aircraft. Here
Figure 16.22 Examples of military aircraft and helicopter cockpit lighting controllers from Ultra Electronics Ltd Electrics Division. Both devices use a 5kΩ remote control potentiometer, are short circuit proof, and have an operating temperature range of -40°C to +70°C.
FUNCTIONAL APPLICATIONS
working crew do need full light level control, and for EMC and simplicity reasons may use ceramic potentiometers to directly control individual ballasts by varying the 115V supply. In private “executive” aircraft there is more use of cold cathode lighting for cabin lighting. This allows more flexibility in matching lighting to ceiling coves; this market is less concerned about the need for the use of standard sources than the passenger aircraft market. 16.7.2 Aviation Ground Lighting (AGL) control 16.7.2.1 General description of AGL The lighting of airfields plays a critical part in aviation safety. The performance of the lighting system and its associated control equipment must be of a very high order. This factor, combined with the unusual electrical arrangements used, means that conventional lighting control equipment is unsuitable, and only a limited number of manufacturers are involved in this specialist market. Lighting at airports is regulated by the International Civil Aviation Organization (ICAO), by corresponding national organizations such as the FAA (Federal Aviation Administration) and, to a lesser extent, by international standards bodies (e.g. EN). Other associations, such as those of the airlines and the pilots also provide input through the ICAO. The aim is to ensure that all airports have the same rec-
Sequenced Flashing Lights
Approach Bars
300M Bar
Runway Threshold
150M Bar
Approach Lighting
ognizable pattern of lighting, and that it is 100% reliable. The extent of the lighting is determined by both the size of the airport and the permitted landing categories. For example CAT I requires the runway lights to be visible at a minimum range of 550m, and the cloud ceiling to be at least 60m. CAT III is itself divided into three parts. IIIA requires visibilty of 200m, or a 30m cloud ceiling; IIIC is “blind landing” requiring no minimum visual range and no minimum cloud ceiling. A generic AGL layout is shown in Figure 16.23. In the description that follows, ratings, numbers and disposition may vary according to the size and type of airport, and the country concerned. The principal components are: Elevated approach lights. These are arranged in a sequence of “bars” of five lamps. The bars are 30m apart, and the lamps within the bars 3.5m apart. The lamps are prefocused halogen lamps of 150− 200W to achieve an average intensity of 6,000− 23,000 cd within the narrow beam (typically 6−10°). There is a wide bar of 22 lamps at 300m from the runway threshold. For CAT II and III installations an additional 150m crossbar of 14 lamps is installed, augmented by nine pairs of side row “barette” sets of three red lights. Precision approach path indicator (PAPI). This device is a requirement for CAT I installations. It consists of two sets of four luminaires set either side of the runway, 200−600m from the runway
Touchdown Zone
PAPI
Runway Centerline
Runway End
Runway Edge
Stop bar Taxiway Edge
Taxiway Centerline
Figure 16.23 Aviation Ground Lighting. This diagram shows the principal elements of approach, runway and taxiway lighting.
485
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
threshold. The luminaires are arranged so that when viewed from different vertical angles, they appear as white or red. The glideslope for aircraft is normally 3°, and the PAPI is arranged so that if the plane is exactly on the glideslope the pilot sees two red and two white lights in each group of four. If the plane is above 3°30', all lights appear white, if below 2°30' all lights apear red. Three lights of one color indicate an error between 10' and 30'. PAPI devices use two or three prefocused halogen lamps to achieve the required intensity and security. Flashing approach lights. The main approach lighting is sometimes augmented by a series of flashing lights installed on the centerline of the main approach lights. These are xenon flash lamps that flash in sequence; they are used as a poor weather landing aid, and, sometimes in good visibility as a runway location aid in urban areas. Threshold lights. These may be either elevated or inset. A bar of green lights, the full width of the runway, indicates the runway landing threshold; and, at the other end, a bar of red lights indicates runway end. Touchdown zone lights, centerline lights. The full length of the runway is identified by a series of lights down the centerline. The lights are white for most of the runway, changing to an alternating red and white pattern 900m from the runway end, and to red only 300m from the runway end. These are set into the runway, and must not protrude more than 12.7mm (0.5 inch) above the runway surface. To achieve this the prefocus halogen lamp is augmented by a prismatic optical system. In the touchdown zone pairs of three inset lights are installed that continue the pattern of the red barettes of the approach (see Figure 16.23). An obvious complication of all lighting concerned with a runway is that, as so far described, it assumes a single landing direction. Clearly if the runway is to be used in both directions, it must be equipped with a dual, reversible lighting system. Thus inset runway lights are available in bidirectional versions. Runway edge lighting. The extent of the runway is indicated by edge lighting which depends on the type of airport and landing procedures used. The lights may be elevated or inset. They are white for
486
the main runway, but for instrument landings are yellow for the last 600m. The elevated lights are available with an omnidirectional component that assists with runway location when a plane is circling. Taxiway lighting. Taxiway edge lighting is by omnidirectional blue lamps of either elevated or inset construction. Taxiway centerline lighting is used at busy airports, both to assist the pilot and as a means
Elevated approach light
PAPI projector
BI-directional inset runway light
Elevated taxiway edge light
Inset light for centerline of curved taxiways
Figure 16.24 Examples of ground lighting luminaires. Illustrations from ADB-Siemens.
FUNCTIONAL APPLICATIONS
of indicating which taxiway to use. Special inset lights are needed for curved taxiways and for “rapid exit” taxiways off the runway. Rows of inset lights are used to indicate “stop bars” to prevent an aicraft moving on to a runway until authorized. Other lights. Several other lighting devices and variants of the main forms of lighting are also used. Thes include signage (both mandatory and informational) hazard warning lights, apron floodlighting, airfield beacons etc.
Two single pole primary connections
16.7.2.2 AGL lighting control From the description of the lighting used at airports, it is clear that various control elements are required: • a range of intensity control to deal with with different daylight and weather conditions. • switching control to set up runway direction, taxiway use, signage etc. • overall control to ensure optimum lamp usage. Most of the lamps used in aviation ground lighting are tungsten halogen lamps, with a rated life of between 1,000 and 2,000 hours at full brightness. In practice a life of 3,000−4,000 hours is achieved because full output is rarely required. Many of the lamps used are supplied with attached high temperature leads, so that connections are low temperature. Repeated lamp changing using conventional lampholders could otherwise result in lampholder connection failure due to the high temperature. Apart from the need for reliability, AGL presents a practical problem of power distribution. The dis-
Figure 16.25 Pilot’s eye view of approach lighting. Photo from ADB-Siemens.
Secondary lamp connection
Figure 16.26 Typical transformer and connections used in AGL. From ADB-Siemens.
tances are considerable (several kilometers) so conventional parallel feed at, for example, 230V, presents problems in terms of cable sizing and ensuring that lamps receive the correct power. Since the advent of civil aviation, techniques to solve the problem have changed in response to need and technological development. Today’s solution is to use a constant current feeder. All lamps run at 6.6A for full output, regardless of watt rating (although for higher power lamps – 300W and above – there is an alternative of 20A). Each lamp is fed by a transformer that takes a primary current of 6.6A, and the primaries of the transformers are wired in series. The transformers are completely encapsulated units that, in the worst case, can simply be buried in the ground. More usually there is a small chamber beneath or beside the luminaire. Connecting cables are attached to the transformer and are complete with shrouded connectors, see Figure 16.26. The secondary connection is two-pole connecting to the luminaire, and the primary consists of two single pole connectors. The transformer design is such that the transformer is self protected. It can operate indefinitely at full load, with an open circuit due to lamp failure, or even with a short circuited output. The primary cable need only be rated for 6.6A (and is typically 6mm2 or 8 AWG) but must be insulated for 5kV. It is usually screened with copper or
487
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Constant current controller + step-up transformer
Figure 16.27 Aviation Ground Lighting. The arrangement for power distribution to the luminaires is based on a constant current feeder.
brass tape, and the screening is grounded at various points along its length. The screening minimizes RF radiation, and helps to equalize both the potential gradient along the cable and the potential pressure across the insulation. A tough metal shield also protects against termite attack. A complete lighting circuit has the configuration shown in Figure 16.27. The key to the operation of the system is the constant current regulator. This ensures that the current flowing through the primary circuit is kept at a constant value, even if the load or Circuit Breaker
the incoming supply fluctuates. Thus if one or more lamps fail, reducing the load taken, the regulator ensures that the current does not change. The regulator is obviously the key to varying the brightness of the lights. Modern regulators offer up to eight fixed current steps in the range 2.2A to 6.6A to achieve eight brightness levels, or a continuous variation using 8-bit control to achieve 255 levels. A block diagram of a regulator is shown in Figure 16.28. Regulation is done at line voltage (typically 220 or 380V single phase) using a conventional thyristor pair. The output of the power regulator is fed, via a choke, to a high voltage (HV) transformer. The HV transformer is multi-tapped so that, for a given installation, the output is optimized to the load − i.e. at full load the thyristors are fully conducting. Regulators are available in a wide range of ratings from 2kVA to 30kVA; anything higher would present a problem in terms of output voltage rating. Regulators are all convection cooled, the use of cooling fans not being permitted. The regulator controller varies the thyristor firing angle according to the required output current; it measures the output current on the HV side by a current transformer. The complete unit may also include, as part of the basic unit or as optional extras, the following:
Contactor Choke
Thyristor pair
HV transformer
Load taps
Low voltage power supply
Local control
Voltage measurement
Thyristor controller and control logic
Indication V, A, lamp + grand fault
Current transformer
Remote control interface
Ground fault detect
EIA485
Figure 16.28 Block diagram of a constant current regulator used for AGL.
488
Lightning arresters
To series lighting circuit
FUNCTIONAL APPLICATIONS
ule between the lamp and its isolating transformer. The module derives power from the transformer and is able to both switch the lamp and detect lamp failure. It receives instructions and reports back lamp failure through the powerline using a powerline carrier (FM in the case of the example shown in Figure 16.30.) The system requires a high voltage modem to be fitted on the output of the constant current regulator. The constant current regulator equipment is sited in a vault that is located to minimize the HV cable runs. Remote control from the tower is of varying sophistication depending on the size and age of the airport. However, to comply with regulations requiring traceable records and realtime update of lamp status, modern systems are based on bi-directional serial control using an industry standard fieldbus. In the tower, control is by a GUI, using as many display screens as needed, that displays the lighting status, reports system faults, and allows the Ground
Figure 16.29 Typical constant current regulator for AGL. Photo from ADB-Siemens.
• means of measuring circuit parameters such as output voltage, primary current and HV insulation to ground. • means of remote control (typically either simple hard wired or serial EIA485). • means of indicating that one or more lamps have failed. This can be done by detecting the difference in output voltage level that arises when a lamp fails. The regulator shown in Figure 16.29 can detect up to 31 failed lamps in a single circuit. Regulations covering busy airports and airports offering CAT III require the control tower to know when critical individual lamps have failed. In addition there may be a need to separately control certain luminaires, for example information signs and stop bars. Both facilities could be expensive if additional individual circuit wiring was required. One way of solving the problem is the insertion of a control mod-
Figure 16.30 The BRITE system from ADB-Siemens provides an individual lamp control facility for AGL. Each lamp is fitted with a device that can control an individual lamp and can report lamp failure. It derives power from the constant current feed and communicates with the central control system, illustrated here, using an FM power line carrier signal. Photo from ADB-Siemens.
489
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
16.8 On the road 16.8.1 Automobile lighting control 16.8.1.1 Automobile electric power
Figure 16.31 View from the control tower. The ground lighting is controlled from a computer that gives a graphical image of the current lighting status. Photo of Oslo Airport AGL control from ADB-Siemens.
Controller to select different lighting scenes depending on runway direction, taxiway usage etc. The computer “front end” control may itself be part of an independent or general computer network based on ethernet. The long distances involved may result in the use of fiber-optic as the communications medium. The flashing lights that can form part of the approach lighting have individual controllers that can flash the 2,000V xenon flash tubes at up to two flashes per second (60J per flash − up to 8,000cd in a 15° beam). Three brightness levels are provided, typically 2%, 10% and 100%. One controller unit acts as master to ensure the lights flash in sequence. Because the controllers must be sited near the lamps, they have to be able to withstand extremes of temperature. Cabinet heaters are fitted to prevent internal condensation in cold weather. This brief description of AGL control is based on the requirements of large airports. Small airfields, which generally do not offer precision approach or blind landing, have simpler requirements, but the principles remain the same. One interesting requirement (mainly applicable to the USA) is that for small fields that do not have 24 hour operation of their control tower, it is possible for the pilot of an incoming aircraft to switch on the airfield lighting by using the mcirophone switch of the aircraft’s radio.
490
We are familiar with the lead acid battery that is used in cars and trucks, and refer to it as a “12V” battery, or 24V for big trucks. However, the voltage seen by loads in a vehicle depends on several factors: • there is a slight variation in open circuit voltage depending on the state of charge. A fully charged 12V lead acid battery has an open circuit voltage of around 13.2V. • when the engine is started using a conventional starter motor, the voltage may drop to as little as half the nominal voltage. • when the engine is running and the alternator’s rectified output is charging the battery, the system voltage is around 14V. • under certain conditions the system voltage can be higher, either on a continuous basis, or as a transient. In conventional automobiles the “load dump” condition arising when a significant load such as the air conditioning compressor switches off can result in a 0.5s transient of around 85V. These considerations have a considerable effect on any electronic system within the vehicle; for example requiring semi-conductors able to withstand a wide voltage range. Modern cars have an average electrical load of over 500W, and larger cars use 1,500W or more. New
Parameter Minimum voltage for engine start Minimum voltage engine off Maximum voltage engine running Maximum continuous overvoltage Maximum dynamic overvoltage (“load dump”)
14V system 9V
42V system 25V
11V
33V
14.3V
43V
16V
52V
20V
55V
Table 16.2 Proposed standards for vehicle power systems. Note that these are intended to be independent of the battery technology used.
FUNCTIONAL APPLICATIONS
developments, such as electrically assisted power steering (replacing hydraulic) are placing still more demands on the system. The wiring associated with such big loads is very heavy. There is now a move to a “42V” system. This allows much smaller cables and semiconductors to be used for power distribution and control since the current is greatly reduced, and losses are related to the square of the current. In practice it is expected that vehicles will first use a dual 14/42V system allowing traditional car electrical items to be used for low power items, and reserving 42V for the high power items. The introduction of such a system is also intended to standardize the voltage regime, making it somewhat more “electronics friendly”. Table 16.2 shows proposed parameters for new car electrical systems.
tors and Darlington pairs have all been used, but advances in FET technology mean that FET devices are now preferred. They can accept the high inrush currents associated with tungsten lamps, and have simpler drive requirements. Switching can be low side or high side. The choice is determined both by the switching technology and by consideration of what happens under fault conditions given the use of the vehicle chassis as a common. Low side switching, where the lamp is permanently connected to the battery, and switching is to common, allows the use of low cost NPN transistors or N-channel FETs. However, the permanent presence of full battery voltage at the lamp, with +12V
LAMPS 1
16.8.1.2 Vehicle lighting
3
4
D S G M
K
MICROCONTROLLER
INDICATOR OUTPUTS
CONTROL INPUTS
+12V LED DISPLAY LAMP FAILURE
REMOTE CONTROL SWITCHES
A/D COMMON
A/D IN
0V
CONTROL OUT
The variable nature of the supply means that any light source must be robust enough to withstand the variation. Traditionally lamps have been rugged duty tungsten and tungsten halogen, and these will continue to be the principal light sources. The nominal 12V lamps receive their rated voltage at the lamp by virtue of volt drop in the wiring and switching. The higher power lights such as headlights have traditionally been controlled by relays; but there is a move towards transistor switching for all main vehicle lighting. The cost of the electronics required has come down, reliability has gone up, and the technique introduces a number of advantages: • a more flexible arrangement of power distribution is possible. • detection of lamp failure can be incorporated for little extra cost. • lamp power regulation is possible; for example permitting dimming of interior lights, or ramped switching to extend lamp life. • the technique suits multiplexed control, described in the next section, better. The design of automobile electronic systems is cost driven, and this has resulted in the topology of switching circuits changing as semiconductor technology has advanced. PNP and NPN bipolar transis-
2
SERIAL LINK TO OTHER DEVICES
Figure 16.32 A single chip microcontroller is well able to form the basis of a “lamp control module”. Here is a hypothetical four-lamp module equipped with four low side switching FET devices. The FETs are fitted with extra electrodes providing a “current mirror” facility; the M electrode provides a voltage that is an analog of the current flowing through the main device (for example 40mV per Amp of main current) and this can be used by an analog input on the microcontroller.
491
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
possible leakage currents in wet conditions, means that low side switching is not often used for exterior lighting. High side switching is generally safer, except that in the event of a short circuit the transistor can be destroyed. However, over-current detection can be used to force an off condition. Higher cost P-channel FETs give the simplest circuit. N-channel FETs used high side require a gate bias that is higher than the source voltage; but they can still be used, either by the provision of a separate bias supply or by the use of a local voltage multiplier circuit. FETs with built-in current sensing are available for the application. The current sensing signal can be used both as a means of limiting current in the case of faults, and as a means of detecting lamp failure. Electronic control of lamps means that there is always a dimming option. This is provided by PWM operating at 100Hz. Very short on or off pulses are avoided, so a typical strategy is to offer 5−95% dimming plus OFF. Figure 16.32 shows how a single chip microcontroller can be used to control four lamps. Although the diagram indicates discrete inputs and outputs, the obvious alternative is that the information is multiplexed to another part of the vehicle. The use of a software based microcontroller allows sophistication in the control to eliminate practical problems. For example, the current sense signal from the FETs is used to both prevent overload and to indi-
Figure 16.33 The sensor system for the automatic headlight control on the Citroen C5 automobile. Forward and upward photosensors are combined with a rain sensor (that also provides an input to the automatic windscreen wiper control).
492
cate faulty lamps. However, it can also be used to “soft start” lamps to minimize inrush current, or, where the lamp (e.g. a stop or indicator lamp) must provide instant on, to let through an inrush of preprogrammed duration. Obviously the FETs must, in that case, be rated for high inrush. Other examples are that the lamp fault indicators can be programmed to give a different signal for a lamp failure or short circuit condition (e.g. one could be “flashing”) and that intensity control can be added with no extra component cost at the lamp control module. If only basic on/off switching is required, it is still the case that relay control of vehicle lighting has a lower cost than electronic control, although this would cease to be true if lighting were to change to 42V where electronic switching would be cheaper. However, as soon as there is any sophistication in the control, the only option is electronic. A trend is towards automatic lighting control. Figure 16.33 shows the sensor system for an automatic headlight control. This one automatically switches on the headlights whenever external lighting conditions require them. This application is a prime example of where the task is much more complex than at first appears, and which can only practically be carried out by a microprocessor because of the complexity of the algorithms involved. • external lighting must be evaluated both according to the luminance of the road ahead, and according to the luminance of the sky above. • sudden changes of level must be evaluated; is the car going under a bridge (not requiring headlight switch-on), or is it entering a tunnel (requiring headlights on)? • vehicle speed is a factor in determining the setpoint. • if it is raining, then the setpoint is raised. If the wipers are running continuously instead of intermittently, the lights may be required to be on all the time above a certain speed, regardless of actual daylight levels. • hysteresis must be introduced (as is the case with photo-electrically controlled building lighting) to minimize nuisance switching. • additional timing routines may be required when
FUNCTIONAL APPLICATIONS
All lighting systems in vehicles have to withstand harsh conditions and meet stringent regulations. Reliable operation over a temperature range of -30° to +70°C is required, and EMC is important. Not only must the equipment conform to the regulations, but there can be no mutual interference between the various electronic sub-systems within the vehicle. This has led to tight restrictions on switching frequencies, data rates and data waveforms. 16.8.1.3 Multiplex control
Figure 16.34 Compact HID lamps are now widely used as the light source in dipped beam headlights. Their efficacy is about three times that of the tungsten halogen lamp. The lamps are supplied complete with a sealed hot restrike ballast. This example is a 35W Xenarc® lamp from Osram that gives 3,200 lumens.
HID headlights are being used, to avoid multiple lamp starts. This last point demonstrates the emergence of many new lightsources in vehicles. All sources other than incandescent require appropriate electronic ballasts or drivers, and in general these can be designed to provide a constant current to the lamp itself, regardless of fluctuations in the vehicle power supply. LEDs are being introduced as alternatives for brake, stop, indicator and rear lights. Fluorescent lamps are widely used in large vehicles such as minibuses, campers and buses. Cold cathode fluorescent lamps are used for instrumentation lighting. Top-ofthe-range cars are now mostly fitted with so-called “Xenon” headlights. The latter introduce a new regime to car servicing − motor mechanics are not used to finding 25,000Volts under the hood except as part of the ignition system; now the volts are needed to start the headlights. “Xenon” headlights are a variation on the compact metal halide theme, where the lamp fill is augmented by xenon to improve the spectrum and starting performance.
Chapter 9 showed that the ideal of a single control protocol for lighting was unachievable because of technical and commercial realities, and because the speed of developments in electronics outpaces efforts at standardization. The same has been true in automobile electronics. Here lighting control is only one of many items where some form of multiplexing can bring advantages. Engine management, instrumentation, braking control, safety, anti-theft, entertainment, window motors and other items are all potential beneficiaries. The questions that are raised are: does multiplexed control bring real benefits? Should there be a multiple network architecture, with each main subsystem having its own network, or should there be one network only so that sensors etc can be shared between applications? And where does lighting fit in? Many years ago the Society of Automotive Engineers (SAE) proposed three classes of vehicle communication network. Class A is where the network nodes would not have existed in a conventionally wired vehicle. Vehicle wiring is reduced by sending multiple signals over a single bus, where otherwise separate wires would have been used. If the device shown in Figure 16.32 used multiplexed remote control, without any other inputs, it would form part of a Class A network. Class B is where parametric data (e.g. speed) is shared between nodes to eliminate redundant sensors. If the device in Figure 16.32 was also able to receive other data, then it could be on a Class B network. An example would be the automatic headlight switching already referred to where the lighting
493
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Logic 0 00150017Ps
Dominant
sitions per bit. The low speed version of J1850 runs at 10.4 kb/s, runs on a single wire and uses “variable” PWM or VPWM, where there is only a single transition per bit. While full PWM permits a higher data rate, VPWM is more suited to the “one wire” approach as it can use longer rise times thus reducing EMC problems.
Logic 1
Logic 0
16.8.2 Road tunnel lighting control
OR Recessive
16.8.2.1 Principles
Logic 1 Dominant OR Recessive
128Ps
64Ps
Figure 16.35 The two modulation schemes of J1850. In the faster PWM version (above) the bit time is always the same and there are two transitions per bit. The slower VPWM version (below) has only a single phase change per bit, so bit timing is variable – however rise times can be much longer for better EMC.
system shares data with the windscreen wiper system. Class C defines high speed networks for real time control systems, such as engine controls and antilock brakes. Because of the safety and cost issues involved, vehicle manufacturers have been cautious about the introduction of networks. Many different proprietary systems have been introduced, and the tendency has been to keep the real time engine/transmission control separate. Also because multiplexing only becomes economic when Class B features are required, Class A has dropped out of the picture. The two most widely used bus systems used are CAN and SAE J1850. CAN is described in Section 9.5.3. It is suitable for both Class B and Class C networks. CAN chips (inluding chips able to service multiple CAN buses) are widely available and well supported. Whereas CAN is suitable for data rates up to 1 Mb/s, J1850 is a simpler system intended only for Class B applications. J1850 comes in two flavors. The higher speed version operates at 41.6 kb/s and runs on a twisted pair. Modulation is PWM, where there are two tran-
494
“Tunnel” lighting as applied to roads means the lighting of any covered roadway. Thus it embraces everything from comparatively short city underpasses to long transalpine tunnels. It introduces some interesting control concepts, not all of which are intuitive. The lighting of road tunnels is specified by national standards (for example those recommended by the American Association of State Highway and Transportation Officials, AASHTO) but these are all mainly based on many long term experiments that resulted in the CIE’s publication CIE 88 “Guide for the lighting of road tunnels and underpasses”. The key metric is the Safe Stopping Sight Distance. This is the time taken by a driver to stop a car on seeing an obstruction ahead (including reaction time). Examples are shown in Table 16.3. When driving towards a tunnel entrance the big danger is the “black hole” effect, whereby a driver cannot see an obstruction in the tunnel entrance because of the huge difference in exterior and tunnel interior light levels. The aim of tunnel lighting is, therefore, to ensure that any black hole is beyond the SSSD, and this in turn requires very high lighting levels at tunnel entrances.
Speed kph
Speed mph 48 64 80 88 96
SSSD m 30 40 50 55 60
60 90 140 165 200
SSSD ft 200 300 450 540 650
Table 16.3 Safe Stopping Sight Distances as recommended by AASHTO.
FUNCTIONAL APPLICATIONS
Access zone
Threshold zone
Transition zone
Interior zone
Entrance portal
Figure 16.36 The zone concept for tunnel lighting.
Tunnel lighting is divided into zones as shown in Figure 16.36. The idea is that the level of lighting changes gradually from the high outside level to the low level of the tunnel interior, so that drivers have
100
time to adapt. The CIE recommend that the lighting level varies in accordance with the curve shown in Figure 16.37. This curve applies when daylight conditions apply outside. Clearly at night the lighting level within a tunnel should be of the same order as that of the exterior road lighting. The figure shows that with short tunnels a high level of lighting is required throughout, but that with long tunnels the lighting level inside may only be 1% of that outside. However, the length of the transition zones depends on the traffic speed. The figure is based on transition time, but also shows the distance involved for traffic flowing at 50 mph (80 kph), but if the speed increased to 75 mph (120kph) the transition zones would have to span 800m (half a mile). The level of lighting required at threshold (which in turn determines the subsequent transition levels) depends on the external lighting conditions. The
Lighting level % log scale
80 60 40
20
10 8 6 4
2
1
Threshold zone Stopping distance in threshold zone 0.5SD 0.5SD
0
2
4
100 325
6
8
10
12
14
Transition zone time elapsed in seconds 200
16
18
300
400m
650 975 Distance at 80 kph 50mph
1300ft
20
Figure 16.37 The CIE recommended variation of lighting level with time on entering a tunnel. Clearly this results in transition zone lighting of a length that depends on traffic speed. The length equivalent is shown for 50 mph (80 kph).
495
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
old lighting system, and is normally in the range 0.05 to 0.1 (typically 0.07.) CIE and national standards recommend k factors according to traffic density, traffic speed, portal orientation and the type of lighting systems to be installed. Some lighting systems provide higher contrast, by having a high ratio of horizontal luminance to vertical illuminance, and these may be permitted a lower k factor. 16.8.2.2 The control problem
Figure 16.38 One bore of the Cheung Ching tunnel in Hong Kong. The 2.5km twin bore tunnel uses over 6,200 luminaires. This photo from Thorlux Lighting shows the additional lighting needed in the threshold and transition zones. Notice the tunnel wall finish; an aim of tunnel lighting is that the tunnel walls up to 2m from road level should have the same luminance as the road surface.
relevant paramater is luminance, since drivers see obstacles by virtue of their contrast to the road background. The metrics are: L20 the access zone luminance, which is measured as the average of all luminances contained in a conical field of view, subtending an angle of 20° at the driver’s eye, and centered on a point at a height of one quarter of the tunnel mouth. It is measured at the SSSD in front of the tunnel mouth. The maximum value of L20 depends very much on how much sky can be seen within the viewing cone, which can be as little as 2% in an urban area to as much as 40% in open countryside. Other factors are the orientation of the tunnel and the nature of the surroundings. CIE and national standards recommend how the maximum average value of L20 can be calculated in advance, but it is normally in the range 1,500−6,000 cd/m2. Lth is the average road surface luminance at any location within the threshold zone. k is the ratio Lth/L20. It is used to design the thresh-
496
For all but the shortest tunnels, the control of lighting assumes great importance for both economic and safety reasons. The control needs are as follows: • a continuous and reliable method of measuring L20. • adjustment of Lth in accordance with changes in L20 applying the relevant k factor. • monitoring of lamp running hours to determine the optimum replacement schedule. • balancing the running hours of lamps where possible.
Figure 16.39 The Meyer Model TS-101 luminance photometer for road tunnel lighting control. It is normally supplied for a luminance range 0–6,500 Cd/m2 for which it produces a linearly proportional current loop output 4–20mA. The output is within ± 1% over a temperature range -15 to +50°C. A heating element is fitted within the housing to eliminate condensation, and the wiper/washer unit can be under automatic or manual control. Measurement is by a silicon photodiode with V(λ) correction (see Section 14.2.4.)
FUNCTIONAL APPLICATIONS
• monitoring lumen maintenance. • detecting faulty luminaires. • dealing with power failure. Short tunnels can be treated as extensions of normal street lighting, with simple external illuminance measurement being used to provide two- or threestage lighting. But any long tunnels require control to address most or all of the above issues. Figure 16.39 shows an example of a photometer used for measurement of L20. It is installed at a height of around 4m (14ft) above road level and at a distance of 120−200m (400−650ft) from the tunnel entrance depending on traffic speed. The varying of the threshold and transition lighting to achieve the results implied by Figure 16.37 requires a large complement of separately controllable luminaires. The threshold zone may itself be divided into two sections, and to provide a more gradual change in level there may be as many as three transition zones. Typically there is a requirement for lighting to operate at six different levels according to the measured value of L20. Bi-directional tunnels must be fitted with symmetrical systems at entry and exit. Uni-directional tunnels can have a simpler exit zone system, but still require augmented lighting at the exit. Most twin-bore installations operate unidirectionally but are equipped for bi-directional operation in case of a bore closure. With computer control it is comparatively easy to monitor running hours so maintenance schedules can be planned. Within the threshold and transition zones it may be possible to equalize lamp running times by selecting different lamps to achieve intermediate light settings, but this depends on the lighting design. Lumen maintenance depends on both the running hours of the lamps and on the cleanliness or otherwise of the luminaires. Some systems incorporate illuminance measurement within the tunnel to check that design light output is being achieved and to provide an additional input into the control system. Until recently the only economic and practical method of identifying faulty luminaires was by observation. It is now becoming economic to provide individual control and monitoring for each lamp. The provision of emergency lighting varies. Most
long tunnel systems are fed from two independent supplies to lessen the likelihood of blackout; and the general requirement is that any blackout should be less than 0.5s, and that lighting is restored to a level at least equal to the night time level in the main tunnel. This implies the use of auxiliary generators with battery inverter support to cover the run-up time. 16.8.2.3 Tunnel lighting control solutions Tunnel lighting uses a variety of lightsources, principally fluorescent, HPS and metal halide HID. Fluorescent lighting is favored for the interior tunnel lighting for a number of reasons. • it is easier to achieve the required uniformity of lighting. • successive point sources can create a subjective flicker, depending on luminaire spacing and vehicle speed. A near continuous line of fluorescent lighting eliminates this problem. • fluorescent has virtually instant start and re-start, ensuring a quick response to control commands and
Figure 16.40 A tunnel lighting controller from P. Ducker Systems Ltd. In a typical installation two of these would be used, one sited at each end of the tunnel, and each working from independent supplies. It operates automatically but can also receive commands from a host computer, or from the touch screen mounted on the front door. It receives the analog L20 data from the photometer, and communicates with the luminaires through multiple CANbus outputs. Each luminaire has an integral addressable controller.
497
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 16.41 Example of computer control of a short tunnel. This is a sample computer screen designed for the remote monitoring of the Birmingham, England, Bull Ring Tunnel. It provides control and monitoring of each individual lamp via the lighting controller shown in Figure 16.40. Screen image from P. Ducker Systems Ltd.
recovery from power failure. Economical HID lamps have warm up times and/or re-strike delays. • fluorescent lamps lend themselves more easily to dimming control if required. Many tunnels use hybrid lighting systems where the threshold and transition zones use HID lamps to help achieve the high level of lighting required in the daytime. Lighting level adjustment is mainly by selective circuit switching; but increasingly dimming is being used to assist in achieving lower lighting levels while maintaining evenness of illumination and as part of lumen maintenance programs. Here lamps are run at less than full output when the luminaires are clean, but are then ramped up as the luminaires get dirty.
498
In Germany, Switzerland, Austria and the Netherlands most new and refurbished tunnels are equipped with dimming systems. Where HID dimming is used this has been based on the use of standard electromagnetic ballasts either with tapped autotransformer dimming or with thyristor equipment of the type shown in Figure 14.20. However, the increasing cost effectiveness of electronic ballasts may see a change in technique. Control systems tend to be proprietary, but are based on distributed switching and control devices. The advent of DALI fluorescent ballasts may introduce an element of standardization at the luminaire, not only for controlling the light level, but particularly for lamp failure detection.
FUNCTIONAL APPLICATIONS
The advantage of ripple control is that it requires no additional wiring and can be applied to large areas of the supply network. The principle is that of superimposing an audio frequency on to the mains supply. Clearly the frequency must be relatively low R
3 phase medium voltage network
Figure 16.42 The Lion Rock twin 1.4km tunnel in Hong Kong uses 3,300 twin fluorescent (2 × 58W) luminaires with controllable (1–10V) ballasts. Six lighting levels or “stages” are available (including a “night” stage), and the dimming facility is used to achieve the lowest stages, Stage 1 at 5 Cd/m2 and Stage 2 at 10 Cd/m2. Dimming achieves more uniform illumination than selective switching, and assists with lumen maintenance. The threshold and transition zones are augmented with over 500 twin lamp HPS luminaires. Photo from Thorlux Lighting.
S
T
Ripple control transmitter
Tunnel lighting control systems can run fully automatically, but in any large tunnel there is a tunnel control room that allows the tunnel supervisors to monitor lighting status at all times and take over manual control when required. 16.8.3 Road lighting control Road and street lighting control has evolved over many years, with the result that systems vary in sophistication. This section does not attempt to be comprehensive but examines three topics that illustrate the special problems and ingenious solutions that can apply to this type of lighting control. 16.8.3.1 Ripple control The oldest form of power line carrier transmission of control signals is popularly called ripple control. It is extensively used in continental Europe for street lighting control and for giving metering systems notice of tariff changes. This latter can also be used for load shedding and for switching off-peak loads.
Figure 16.43 Serial injection of ripple control signals is used for frequencies below 200Hz. Diagram of principle (above) and photo of 20kV serial coupling for 40MVA load (below) from repas AEG.
499
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
R
Medium voltage network
S T
Coupling capacitors Coupling inductance
Ripple control transmitter
Figure 16.44 Principle of parallel ripple control injection (top). Example of three-phase coupling inductance with voltage transformer (center) and a complete 10kV 40MVA parallel coupling (bottom). Diagram and photos from repas AEG.
500
for the signal to get through the power network, and typical frequencies are in the low hundreds of Hz. Early ripple control systems used rotary frequency converters to generate the required AF signal, but present day equipment is based on solid state inverters. The method of injecting the signal into a network depends on the signal frequency and the size of the network. Injection is usually into the medium voltage (10−20kV) sub-network. Figure 16.43 illustrates a method of serial coupling, used for frequencies below 200Hz. The ripple control transmitter produces a three-phase output that is usually locked to the timing of the 50Hz main supply. The alternative parallel coupling is used for frequencies above 200Hz. To prevent the injection system loading the network, it uses a resonant LC circuit. Figure 16.44 shows that the size of the coupling system is proportional to the network load. The ripple control transmitter, Figure 16.45, is suitable for either kind of coupling. Transmitters are available in various power ratings to match different network sizes, and are programmable in respect of frequency, output current, phase offset, and type of output signal. The signal injected into the network is, necessarily, simple and slow. Short bursts of AF are used to create “bits” of information. Several different “standards” exist, the most common being SEMAGYR, originated by the Landis & Gyr division of Siemens Metering, and Pulsadis, originated by Electricité de France. As an example, Pulsadis uses a carrier of 175Hz. The amplitude is around 1% of the nominal network voltage, say 2.3V on a 230V supply. A burst of one second, followed by 2.75 seconds silence indicates the start of data frame. The receiver recognizes the start signal, and times the succeeding information from it. The data frame consists of 40 bits, each of which occupies 2.5 seconds. During the 2.5 seconds the first one second may include the 175Hz carrier, indicating data 1, or may not have it, indicating data 0. There is always a 1.5 second silence separating the data slots. The meaning of the data “telegram” is specific to the operator, but in France EDF use Pulsadis for tariff management and, as an example bit 5 on its own is “wake up”, bits 5 and 15 together
FUNCTIONAL APPLICATIONS
A recent European development is “radio ripple control”. Here the control signals are transmitted as long wave RF, and the powerline is not used. Many ripple control receivers are now available in both RF and powerline versions. Since 1995 the company Europäische Funk-Rundsteurung GmbH (EFR) has run the service based in Germany. The transmitting station, callsign DCF49, is based at Mainflingen and operates at 129.1 and 139.0kHz. Radiated power is 60kW from a 324m high antenna, giving an effective range of over 750km. Modulation is FSK, with a frequency shift of 340Hz. Data transmission is at 200 bps, and is in the form of standard ASCII characters. Typical data “telegrams” are only a few bytes long.
Figure 16.45 The repas AEG Gearic II ripple control transmitter can be configured with outputs up to 210kVA. The inverter frequency range is 100–1,500Hz, and output voltage is in the range 0–415V.
means “start of peak demand” and bit 15 on its own is “end of peak demand”. Ripple control has served its market well, with thousands of local authorities using it as a means of controlling road lighting (apart from the main use of the system as a means of controlling metering and switching tariff sensitive loads). However, it does suffer from some disadvantages. It requires power to operate which, if aggregated over many networks, represents a significant electrical load. The ripple signal itself can also have some undesirable side effects. Under some circumstances it can affect the zero crossing of the mains sine wave, which in turn can produce undesirable flickering in phase controlled thyristor dimmers − although professional dimmers include filter circuits designed to eliminate the problem. Studies have also shown that it can introduce undesirable flicker into fluorescent lamp loads using electromagnetic ballasts.
Figure 16.46 Example of a ripple control receiver suitable for street lighting control. The Landis & Gyr RCR131 is a combination of timeswitch and ripple control receiver. Devices of this kind have programmable filters to detect the carrier frequency (110–2,000Hz) and have a sensitivity of between 0.3 and 2.5% of the network voltage. They can operate over a temperature range -20° to +60°C.
501
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Subscribers to the EFR service (for example local authorities) are able to send their switching requirements to EFR over the telephone network. For most users this task is only required occasionally; seasonal variations can be accommodated automatically without extra input from the subscriber. A central computer marshalls and prioritizes the required messages and sends them out as required. While most “telegrams” are addressed to individual users, there are also broadcast messages, such as time and date information, available to all. 16.8.3.2 Road lighting level control Street lighting is usually switched, with the switching being determined by a combination of timed and photo-electric control. Considerable economies can, however, be effected if the lighting is regulated. Apart from saving energy, regulation can extend lamp life by preventing over-voltages reaching the lamp, and can assist lumen maintenance as lamps age. In the late 1990s the Dutch Ministry of Transport carried out an interesting test on a 14km stretch of six lane highway. The idea was to match the lighting level to prevailing traffic and weather conditions, and also to reduce carbon emissions by reducing the amount of energy used by the system. The system was called DYNO 20−100−200. This stands for Dynamic Public Lighting, operating at 20%, 100% and 200% of “normal” level. A normal level in this context is a road surface luminance of 1−2 Cd/m2. The 200% level was reserved for exceptional conditions, such as fog, a combination of rain and high traffic levels, and accidents. The 100% normal level was used for high traffic density, and the 20% level used for low traffic levels late at night. From a safety point of view it had been found that generous lighting when traffic was limited resulted in higher average speeds, which to some extent negated the safety benefits of having lighting at all. In practice, provided traffic was light and the weather fine, the study showed that the 20% level was sufficient for light traffic and did not have any negative safety consequences. The 200% level was found not to be justified, the increase in cost was considerable, and any safety ben-
502
efits marginal or unmeasurable. However, the simpler DYNO 20−100 regime was found to be a good investment with a rapid payback, in addition to meeting the government’s “green” objectives. The system used to regulate street lighting level depends on the lighting design and the range of control required. It is essential that dimming, as opposed to, for example, switching alternate luminaires, is used in order to ensure lighting uniformity. Assuming that HID lamps are used, the achievement of a 20% level needs electronic control equipment like that of Figure 14.20 or the use of controllable HID electronic ballasts. For most installations the achievement of a 50% dimming range yields a good combination of energy saving and reasonable capital cost, in which case electromagnetic regulators like that of Figure 16.47 can be used. The particular system shown in Figure 16.47 is interesting because it aims to give a similar performance to variable autotransformers, without the main-
Figure 16.47 A 3-phase regulator for street lighting control, prior to its fitting into a weatherproof cabinet, from the Reverberi range of Merloni Progetti. Each section has a microprocessor controller that programs the tap changing of an autotransformer. The resulting voltage is used to vary the current in the control winding of a main inductance in series with the lamp load. Models are available in the range 14–200A per phase.
FUNCTIONAL APPLICATIONS
Vehicles per hour >3,000 3,00000101,500 <1,500
Lighting level 100% 75% 50%
Table 16.4 Lighting level settings for different traffic flows on the M65 motorway.
Figure 16.48 The regulator of Figure 16.47 can communicate with a central control and monitoring computer using telephone, GSM or radio modem. Image from the Reverberi division of Merloni Progetti.
tenance problems arising from the use of moving contacts. Rather than using a simple autotransformer with switched taps, this system (Reverberi patent) uses a variable reactance with the current to the control winding being provided by a multi-tapped autotransformer, much smaller than would be needed for carrying the whole load. Tap changing is done with relays operating at the zero crossing. The result is equipment that is much lighter than its full transformer equivalent, and which is able to give reasonably smooth voltage changes because of the series inductance. Shortly after the Dutch study was completed. Lancashire County Council in the UK took the concept of dimmable motorway lighting further in a seven mile (11km) long installation on the M65 motorway.
They used “Elgadi” electronic ballasts from Royce Thompson that combine in one package an HPS dimmable ballast (range 30−100%) and powerline carrier control communications. Lighting level is controlled according to traffic flow as shown in Table 16.4. It is expected that, when all factors are taken into account, the system may deliver a saving in running costs of as much as 50% over a conventional installation. The Department of Optometry and Neuroscience at the University of Manchester Institute of Science and Technology (UMIST) carried out studies on the M65 installation relating to the eye’s response to the glare from bright overhead lights. As mentioned in Section 16.8.2.3 the flicker effect of passing under a series of overhead lights of high brightness at speed is uncomfortable, and, for some, possibly dangerous. UMIST’s study measured this dynamic discomfort glare and its resulting ocular stress. This showed that under light traffic conditions, driving is significantly more comfortable if the lighting level is reduced. Dimming of motorway lighting would, therefore, appear to have the potential to produce both significant savings in energy and safer driving. 16.8.3.3 Lamp monitoring Road lighting systems are expensive to run by virtue of both energy costs and maintenance costs. A significant practical problem is that of monitoring luminaire performance, usually requiring the patrolling of the roads and streets concerned in the hours of darkness, followed by a repair visit in the daytime. The aims of lower cost and better service could be served by automated monitoring. Systems like that of Figure 16.48 can report circuit failure centrally, but need some enhancement if the reporting is to be down to the luminaire level.
503
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 16.49 Example of a street lighting control and monitoring system using a communications unit at each luminaire. This is the TSS “Scout” system from WRTL Exterior Lighting Ltd.
This is now achieved in several different ways. At its most sophisticated, an electronic ballast can be combined with a communications unit in a single package. More usually a separate communications device is used, with options as to the method of communication. Figure 16.49 shows the principle of one system. The manufacturer concerned offers a choice of powerline carrier (especially suitable for retrofit installations) and EIA485 (offering a longer range) as the means of communication. Each luminaire is fitted with a communications unit designed to work with both conventional and electronic ballasts. The unit monitors lamp status and measures lamp running
504
Figure 16.50 The principal components of the system shown in Figure 16.49. On the left the “Scout” unit installed in each luminaire, on the right the “Scoutmaster” that controls and monitors up to 255 luminaires. Photos from TSS division of WRTL Exterior Lighting Ltd.
FUNCTIONAL APPLICATIONS
hours, besides providing control of the lamp (which includes selection of two light levels if suitable ballasts are being used). All units report back to a central controller. This device issues the control commands and logs data received back from the luminaires. Either automatically or on request it sends the logged data back to a central office, using either the conventional telephone network or, more usually, the GSM mobile phone network. The advances in wireless communications are making the idea of using wireless control of individual street lamps a viable and economic proposition. The aim of the system is the same, but instead of using powerline or separately wired communications, each luminaire is fitted with an RF transceiver. An example system is that of Telemics Inc. Here each luminaire is fitted with their CheckPoint™ device shown in Figure 16.51. Several hundred such devices work together as an intelligent mesh network (like a cellular phone network on a small scale). The idea is that they then all report to an “Access Point” that forwards that data to a Telemics Data Center. The Data Center is based on a secure server system that can control and monitor many networks. End customers can communicate with, and receive reports from, the Data Center via a secure website. Figure 16.52 The CheckPoint™ device is designed to be fitted directly on to the luminaire. Photo from Telemics Inc.
16.9 On railways 16.9.1 Platform lighting control
Figure 16.51 The CheckPoint™ from Telemics Inc can control individual street lights fitted with magnetic ballasts. It can also detect lamp, ballast and photo-controller failures, and it measures lamp operating hours. The transceiver operates in the 902–928MHz band.
Many of the principles that have been described in Section 16.8 also apply to the lighting of railway platforms and concourses, whether above or below ground. The long operating hours and the high cost of lamp replacement are incentives towards using control systems to reduce energy use and prolong lamp service life. Wherever there is a daylight contribution to station concourse lighting, there is a case for daylight linked control. In transition zones there will be a need for more electric lighting in conditions of bright outside lighting, which needs reducing at night. Care
505
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
must be taken that any lighting required for safety reasons is always kept at the right level. Figure 16.53 illustrates both these points. One of the main enemies of effective exterior lighting is dirt, and this is especially true of subway and railway platform lighting. As with road tunnel lighting a case can be made for electronic control and monitoring of luminaires so that the light output is reasonably constant. For example when the luminaires are clean the system might only require 70% electrical input, building up to 100% as the luminaires get dirty. Indeed such a system can be used for both predictive cleaning maintenance and predictive lamp replacement. Some subway platforms are fitted with lighting below platform level to provide lighting at foot level at the entrance doors. This is especially the case when the subway is fitted with platform doors (as in Hong Kong and parts of London’s Jubilee line) where there
Figure 16.54 The Central Terminal station of the Heathrow Express train service from London’s Heathrow airport. The luminaires are hermetically sealed to facilitate fast cleaning. Their electronic ballasts are mounted remotely within the cable management system above. This allows the ballasts to run at a lower temperature and facilitates replacement should it be necessary. Notice the lighting at below platform level. Lighting design by Pinniger and Partners.
is a greater gap. There is no reason for this lighting to be on all the time, and it should be controlled in conjunction with the door opening mechanism. The duty cycle is sufficiently low that the use of dimmer controlled tungsten halogen lighting is cost effective for this application. 16.9.2 Train interior lighting control
Figure 16.53 At the Canary Wharf underground (subway) station on the Jubilee Line in London, there is a considerable daylight contribution. Daylight linked control is provided both to save energy, and to provide a comfortable lighting level at night. Compact fluorescent lamps with electronic ballasts are fitted to the escalator balustrades to ensure safe lighting on the treads at all times. Lighting design by, and photograph from, Pinniger and Partners.
506
Train interior lighting control has advanced considerably in recent years because of the increase in reliability of electronic components. The factors that distinguish this application are: • the need for long service life e.g. >25 years. • the ability to operate over a temperature range that may be as wide as -40° to +70°C . • the ability to withstand continuous vibration. • possibly a wide range of humidity. • a tough electrical regime. • the need to meet specific railway standards in respect of safety and EMC. For example EN50121-32. For historical engineering reasons the lighting electric supply regime can vary considerably. Intercity trains might start with a 25kV supply. Typically this is then converted to 700V DC for the inverters that control the traction motors. A supply for other
FUNCTIONAL APPLICATIONS
Figure 16.55 Excil Electronics Ltd make a range of electronic ballasts for train lighting. Customers include London Underground, suburban rail and intercity train companies. The Excil units are mounted in aluminum extrusions with end plates. These are sealed with gaskets to IP65. Equipment of this kind is supplied through the major train builders such as Bombardier and Alsthom.
services is derived from the intermediate traction voltage, typically ending up with a 110V DC supply that is used to power the lighting inverters and to charge the batteries. Light rail systems might start with a considerably lower voltage, such as 1.5kV, but still end up with 110V DC for lighting. London Underground presents a peculiar problem. Many lines still use a dedicated 110V 850Hz AC supply for main lighting with a separate 50V DC battery powered system for emergency lighting. The 110V is derived from the 600V DC traction supply, and arises because earlier systems were based on rotary converters, and the high frequency allowed the weight of the electromagnetic components to be kept down.
Main lights
LINET nodes and ballast
LINET controller
Seat lights Light switch
Wireless connection ) (GSMR) Extension bus to the train management system Host Twisted computer pair cable
Figure 16.56 A possible control system for train interior lighting from Teknoware of Finland. Described in the main text.
Figure 16.57 Typical modern rolling stock fitted with the latest generation of fluorescent lighting. This is the “double deck” suburban stock of Finnish Railways. Photo provided by Teknoware.
The great majority of train lighting is now fluorescent. The inverters must withstand the somewhat erratic supply regime, with frequent switching cycles and transients, but must not themselves produce any intereference that would make matters worse. The requirements are sufficiently stringent that the market is mainly served by specialist manufacturers. Traditionally control has been very simple, with contactors in each carriage being under the control of a master switch in the driver’s cab, and automatic changeover to battery power when the main supply is interrupted. However, the electronic ballast offers many additional possibilities, and the trend now is for train interior lighting to be similar to aircraft interior lighting. The trend is helped by the fact that modern trains use computers or PLCs as services controllers, and lighting control is easily integrated. An example of what is possible is shown in Figure 16.56. Here there is main saloon lighting and individual reading lights. All control is by a separate control bus, and every lamp can be dimmed. In this case a proprietary bus from Linet is used. The Linet system results in a very simple installation. Only two wires are needed for both powering the network and distributing the control signal. The 20kHz signal and power waveform is near sinusoidal so interference is minimal; the negative going half cycle is used for power, and the positive going for data. 0 and 1 bits
507
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 16.58 The ballast/control interface needed to realize the scheme of Figure 16.56. Each inverter is accompanied by a compact Linet control interface. Photo from Teknoware.
are indicated by the shape of the data part of the waveform. Up to 255 nodes are permitted in a system, and as each node has its own alloted time slot, collision detection is not required. Nodes can be input nodes, for example passenger light switches, and output nodes. The output nodes can be configured to provide an analog output for ballast control, a PWM output for DC lamp control, and a thyristor trigger pulse output for AC phase control. Each network requires a master controller that provides power, timing and gateway facility. In the example shown in Figure 16.56 the Linet network is shown as being connected to the train host computer. This would be used to set lighting scenes in accordance with operating conditions. Lighting settings would depend on time of day, train speed, daylight and other factors. The diagram also indicates that system monitoring could be undertaken remotely, using the GSM network.
taken. Control rooms are used by the military, utilities, financial centers, and logistics organizations. Control rooms are characterized by having many “operator positions”, each of which has one or more computer terminals or similar devices with display screens. In addition there may be a large “overview” display that can show common information, or a collage of images derived from the individual workstations. An example might be a transport monitoring room that monitors traffic flows and signalling on a large highway network, and that uses a big central display to monitor real time traffic images and help control accident situations. The lighting rules are an extension of those found in other workplaces: • the operators are viewing their screens for prolonged periods, they must be able to adjust the brightness of their display for comfort. • some display types are low brightness, so there must be control over ambient lighting. • when a central display is used, the lighting conditions for it must be suitable, and it must not compromise the use of the individual displays. In the civilian world computer displays are now bright enough to work in reasonable ambient light, but in the military world many displays, such as ra-
16.10 Control rooms A product of the electrical, and then electronic, age, the control room is a place where large systems or networks are monitored, and where decisions are
508
Figure 16.59 The displays used in modern control rooms are bright, so there is no serious restriction on the lighting levels. Individual operators still like to have local lighting control. The photo is of the control room of GE Power in Atlanta, GA.
FUNCTIONAL APPLICATIONS
Figure 16.60 Military operations rooms often work at low ambient light levels. Again operators need to be able to set levels for comfortable working over long watches. The photo is of an Operations Room on the aircraft carrier USS Carl Vinson, courtesy SPAWAR of Charleston, SC.
dar displays, are low brightness. There may also be a preference or necessity to work in low lighting levels. For this reason personnel like to have control over the ambient lighting. Often in these applications compact, rugged variable transformers are used for this purpose. Big multi image displays or “videowalls” are the centerpiece of many control rooms. The first generation of these were based on CRT projection displays. These were of comparatively low brightness when showing detailed graphic images, and required a low ambient light level to be effective. All new control rooms use projection based on DLP™ or LCD technique, and these are sufficiently bright to work in normal office conditions. In all cases some form of scene control is recommended to ensure comfortable working under all conditions (and even to allow the users to have a change every now-and-again!).
509
C h a p t e r 17
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Entertainment
17.1 Small stages 17.1.1 Introduction The development of low cost, but sophisticated, lighting controls, combined with a parallel advance in luminaire design and much greater aspirations on the part of users means that even the smallest stage may have extensive control facilities. Schools and amateur dramatic groups now demand facilities that are the near equal, or are in some cases superior, to their professional counterparts. Small stages take many forms. Often they are part of a multi-purpose facility, where the hall concerned is also used as an assembly hall, and even sports hall. Sometimes they are purpose built within small theaters or studios, and in the commercial world they can form part of a large restaurant or bar complex. In most of these applications the specialized “performance” lighting is a mixture of permanent and temporary luminaires. There may be a limited set of luminaires to provide basic stage lighting that is augmented by hired-in equipment whenever a “production” is put on. In other cases the basic set-up may be augmented by incremental purchases during the lifetime of the installation. This raises a problem in respect of lighting control. Unlike the touring show or product launch show where professional staging staff are in charge, and expensive, heavy duty, touring equipment is used, the small stage is usually “staffed” by amateurs, maybe children. As far as possible the use of temporary power distribution and long flexible power cables must be avoided. This means that the priority in making provision for lighting on this kind of stage is the infrastructure. Luminaires can always be added later or hired; but it is difficult to extend permanent wiring. In his book Stage Lighting Design, Richard Pilbrow says “There are no hard and fast rules as to how to
510
applications
lay out circuits around a theater...the permanent installation should be chosen bearing in mind the average use of the facility and with an eye to a healthy balance between economy and lighting enthusiasm”. 17.1.2 Circuit arrangement for small stages Even the smallest stage should be equipped with as many separate circuits as affordable or appropriate with permanent outlets sited at all likely luminaire positions. This is usually done in association with the mounting of the luminaires − for example by providing a multicore connection to a pre-wired spotlight bar already fitted with outlet sockets; or providing sockets next to front of house (FOH) lighting positions. In an ideal world all circuits are connected to permanently installed multi-channel dimmer racks that include power distribution and circuit protection. Local legislation may require that all circuits are RCD protected, especially for school installations. However, some installations are designed with fewer dimmers than circuits, and in this case a load patch panel must be installed. If possible this should be avoided, especially as the cost of patch panels is not insignificant, and in some cases it can be shown that it is actually more expensive to install a patch panel than it is to install the full complement of dimmers. Often provision is made for hiring in additional dimmers for “the big show” − in the UK this might be the annual pantomime. If this is likely to be done, then ideally there should be protected power sockets available that match the temporary dimmers (that are usually in the form of portable dimmer packs containing 6 or 12 dimmers) and the additional circuits should be available to plug in at the dimmer position. To facilitate the use of standard packs, this requires flexible cable connection. The extra circuits may either be distributed throughout the system (e.g. six FOH sockets as part of the permanent system,
ENTERTAINMENT APPLICATIONS
Figure 17.1 When budgets are limited, or when the space is very flexible, load patch panels may be permanently installed. Photo from Stagetec.
and another six that are only used when extra dimmers are connected) or represent an additional concentration of resources such as an extra spot bar. Although very small stages can use a manual multi-channel single or two-preset analog control, any system intended for drama should use DMX control throughout, since this greatly simplifies the control wiring and facilitates temporary additions. If there is provision for installing temporary dimmers, it is recommended that the DMX control signals to them should be optically isolated to eliminate the possibility that a temporary control connection could damage or disable the permanent system. One way of solving the problem of adding extra dimmers and their associated DMX control wiring for temporary installations is to use luminaires with built-in or “clipped on” dimmers with RF control. An example is shown in Figure 17.2.
tive, some users have a simple permanently installed console with basic facilities, and hire in a more sophisticated unit for drama productions. When electric lighting arrived on stage, the lighting control position was back stage close to the equipment. With remote lighting control, this is no longer necessary and all theaters now have FOH lighting control positions so that the operator has a proper view of the stage. On a small, occasionally used stage it may be necessary to site the control backstage, but if drama is involved it is much better to have an FOH position designated for it. A common practice is to have DMX connectors sited at the back of the auditorium, allowing the console to be plugged in for drama events. This leads on to the needs of houselights and, in a multi-purpose space, the needs of other lighting. For drama, small scale concerts, meetings and similar events it is best to have a separate lighting control for the houselights. This can be as simple as a single automatic dimmer with push button control to give access to different lighting levels. For a larger space, a multi-channel, multi-scene unit should be used, but again using push-button control so that it can be accessed from entrance door positions and from the stage. It must also be accessible from the designated stage lighting control position.
17.1.3 Control console and houselights The control console should be intuitive to use, at least for its basic functions, since many different people may use it on an occasional basis. In order to avoid the problems arising from a fixed load wiring installation, possibly augmented by additional dimmer packs, it is best if the console has a soft-patching facility for the control, since this simplifies the logical arrangement of lighting groups. As an alterna-
Figure 17.2 The “Freeway” radio dimmer system from CCT Lighting sites the dimmer at or within the luminaire. Control is by RF operating at 433MHz. The remote controller (shown in Figure 11.2) can be interfaced to DMX.
511
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
If other lighting is needed for alternative uses of the space, it should be considered separately. Houselights for drama should be incandescent to ensure good fade to black, but there is no reason why the space should not have controllable fluorescent lights that can be called up as part of non-drama “scenes”. In some cases a practical alternative is to have conventionally switched circuits for the discharge lighting with an arrangement that disables it when drama events are taking place. Occasionally it is a requirement that some of the “stage” lighting is used as part of the house lighting when the space is being used for events other than drama. When this happens it is inconvenient and impractical to have to use the stage lighting console. There are two simple options: • some dimmer systems include a facility for storing back-up scenes that can be called up independent of the DMX. The push button selection of these can be coupled in to the houselight controls. • simple controllers are available that store prerecorded DMX commands. These can be connected to the dimmers either using a DMX changeover switch, or, where dimmers can accept dual DMX inputs, directly. Again the controllers can be linked to the houselights push-button control.
Figure 17.3 The control console should be sited where it has a good view of the stage. The permanently installed console should be one that is intuitive to use. Photo of control position at the Mick Jagger Arts Centre from Carr & Angier.
512
Figure 17.4 Multi-purpose halls, such as the main hall at the Mick Jagger Arts Centre, Dartford, UK, require additional lighting for non-drama activities. Also certain events require some of the “stage” lighting to be used as “houselights”. Photo© Grant Smith/VIEW.
17.1.4 Power plugs, sockets and connectors The comments in this section are equally applicable to large stages. There is considerable variation in practice in the provision of sockets for the connection of luminaires, whether these be part of the infrastructure or fitted to temporary dimmer packs. Local practice is often determined just as much by the rental business as by regulation. Because of the different voltage regimes, connectors used in the USA and Canada have to carry double the current of their European counterparts, requiring the accommodation of much thicker cables. The UK is nearly unique in having a domestic/ office small power distribution system that uses a fused plug as the means of connection. Such a device is a practical nuisance in theater lighting since the fuses used blow on lamp failure when a lamp of any size is used, and it is much better practice to have MCBs on the dimmer racks for protection. There is, therefore, a need to use an unfused connector. For many years the connector of choice has been the BESA 15A connector (BS546) based on the plugs used for electric fires and similar high current domestic loads prior to the introduction of the BS1363 13A plug system. Some small systems used the cor-
ENTERTAINMENT APPLICATIONS
Figure 17.5 Examples of plugs and sockets used for load connection in theater applications. Individual connector photos all from Union Connector.
responding 5A connector, but this is no longer recommended. The modern BS546 socket is shuttered, and the plug used in theater applications is usually the molded rubber version. The 15A rating makes the connector suitable for loads up to 3kW, but it is generally used on 2 or 2.5kW dimmers. The most widespread mains connector used in mainland Europe is the Schuko connector. It can be used on circuits rated at 10A or 16A, but when used on dimmers and stage lighting circuits the luminaire rating is usually limited to 2kW. In this application the socket is often supplied in the industrial version with a spring loaded protective flap that covers the outlet when not in use. A complete range of power connectors covering a wide range of current ratings, suitable for both single and three-phase circuits, and available in a wide range of versions for different industrial applications is defined by IEC309 (BS4343). The 16A and 32A, often referred to as CEE16 and CEE32, single phase versions are now widely used as lighting circuit connectors throughout Europe, with some consultants and users specifying them in place of the old 15A connector in the UK. High current 3-phase versions (e.g. 63A) are often used as the power input connectors to portable dimmer packs. In the USA small dimmer packs use domestic power outlets rated at 15A; this is the standard USA straight blade plug with a ground pin. For ratings above 15A one choice is the Twist-lock® connector which is commonly used in a 20A rating for 2kW loads. For higher ratings the “stage pin” connector is widely used. At the dimmer pack level there is wide local variation. Naturally countries like Switzerland, Denmark, Australia, South Africa, France and others may require local power connectors, and it may even be the case that different areas of a single country adopt a different practice. As a result theatrical dimmer manufacturers offer many connector options. The most widely used connector for appliance power connection worldwide is the IEC320/CE22 connector. It is rated at 6A or 10A. In stage work it is normally only used for connecting electronic devices such as DMX splitters and its use for lighting loads, especially dimmer circuits, is discouraged. There is
513
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
an alternative 16A version with a different pin arrangement that is sometimes used for lighting load connection in confined spaces where the alternative CEE16 would be impractical because of its size. Where multipole load plug and socket connections are required within a dimmer system a current popular choice in Europe is the IEC664 range, generally known as Harting connectors. Connectors have pin ratings of 10A, 16A, 35A, 80A and 100A and are available in many pin configurations. In the USA and for the international touring market the Socapex range of multipole connectors is widely used.
Until recently this had approximately 1,000 circuits, with a ratio of dim to non-dim of 20:1. This has been changed to a new arrangement providing a ratio of 6:1 (but without a significant difference in the total number of permanently installed circuits). The greater provision of non-dim circuits arises for several reasons: • the widespread use of moving lights that use an HID source. • the greatly increased use of accessory lighting devices such as color scrollers. • the need for power to operate DMX distribution units and similar additional control devices.
17.2 Large stages 17.2.1 Circuit arrangement The provision of control for professional stages has been gradually changing, so there are now significant differences from even only ten years ago. Some mature installations are being modified to include these new features. The widespread introduction of moving lights, color scrollers and other devices has changed the requirement for power distribution and for control. Until recently the aim was to have the great majority of available circuits fitted with dimmers, with a limited provision for “non dim” circuits. Non dim circuits are used for: • powering luminaires fitted with non-dimmable lamps (e.g. a UV discharge lamp.) • powering items that should not be operated through dimmers (e.g. motorized effects discs, fans). The circuits can either be fed through the dimmer racks, or through separate distribution. In the former case the circuit within the rack is served: • by a dimmer designed to be used as a “non dim” switch. • in the case of modular plug-in systems, by replacing the dimmer module with a non-dim contactor module. • by a separately wired section of the rack devoted to non-dim operation. An example of a professional stage is the Lyttleton Theatre at the Royal National Theatre in London.
514
Figure 17.6 The lighting control position (above) in the Lyttleton Theatre (Royal National Theatre) is at the back of the auditorium and has a clear view of the stage. It has a motorized front glass window. Next to it is the stage manager’s position (below). The main technical features of the latter are a good intercom system, a cue light control panel, and a closed circuit video link to back stage. This is a dual link, one being standard video, and the other using an infra-red camera, so that an image is still obtained even when there is a full blackout.
ENTERTAINMENT APPLICATIONS
• the realistic possibility of using distributed dimmers should additional channels be needed. Some thought must be given to how the non-dim circuits are controlled, and the arrangement varies according to local preferences and the type of theater. At one extreme all circuits are under DMX control and can be controlled from the main console, at the other all circuits are under the control of an extension to the working lights control system (see Section 17.2.4). 17.2.2 Dimmer ratings and distributed dimmers In all but the very largest theaters the lamps used are limited to 2kW. This means that the great majority of dimmers are rated at around 10A in 230V areas and 20A in 110V areas. This allows either one 2kW lamp or two 1,000W lamps to be connected to each circuit. For larger stages there may be a reasonable provision of 20/25A circuits (230V) and 50A circuits (110V) to allow for higher power lamps or to allow two 2kW lamps on one circuit. Lighting designers prefer a one lamp per circuit arrangement where possible, since there is otherwise always the danger that a lamp failure could cause an MCB to trip, thus taking out both lamps, even though only one has actually failed. A big change in theatrical practice has arisen with a new generation of luminaires based on new lamps and improved optics, first exemplified by ETC’s Source 4® luminaire. (Source 4 lamp construction is patented and the lamps are exclusive to ETC products. The filament arrangement is cylindrical, producing a better “point source”.) These 575W/600W luminaires can give as much or more light than their 1,000W “conventional” equivalents. The extensive use of such lamps could justify the specification of lower rated dimmers, for example 1.2kW. This, in turn, raises the question of whether the arrangement of central dimmer rack is the best option. With the arrival of moving lights, users have become accustomed to the idea that the dimming device is actually located within the luminaire. There is no reason in theory why this should not become a more general practice for conventional lights, but in practice its application is initially most likely to
apply when lamps operating at other than line voltage are being used. More likely is the introduction of distributed dimming devices. Typically these have 1−6 outputs, and can operate from a suitable non-dim outlet (which now needs to be of high current rating). Any distributed dimmers must be acoustically silent in operation, which favors the use of transistor technology (either trailing edge or sine wave) with convection cooling only. Although such dimmers may be considerably more expensive than their rack mounted thyristor equivalent, they do bring great flexibility, a reduction in the space required by equipment, and a reduction in the complexity of the permanent infrastructure. It is probable that new installations will use a hybrid arrangement, with a “conventional” backbone based on central dimmers augmented by an extensive non-dim power installation for both moving lights and distributed dimming. The latter is likely to be used in those applications where concentrations of dimmers are needed for a particular production, and where space for central equipment is at a
Figure 17.7 Distributed dimmers in use at the Olivier Theatre (Royal National Theatre). Each pack (from IES) is 6 × 2.5kW and uses IGBTs.
515
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
premium. One application that lends itself well is the flown spot bar, where it can be easier to provide a bar with a limited number of high current circuits, avoiding the complexity of multicore power cables. Distributed dimming is suitable for stages that have a highly qualified full time technical staff. It is argued by some consultants that accessibility is an issue, and that rapid and safe service is easier to achieve on central dimmers. 17.2.3 Control wiring The advent of moving lights, color scrollers, distributed dimming and other devices requiring control has meant that DMX is no longer confined to linking the console to the dimmer racks. It must now be available at many positions in the theater. In addition there is now a demand for other signal cabling, especially for ethernet LAN. New facilities are equipped for both ethernet and DMX. Some practioners keep the two separate, retaining the XLR connector and conventional cabling for DMX, and using CAT 5 cabling for the LANs. Figure 17.9 The extensive use of color scrollers requires DMX distribution throughout the main luminaire positions. Photo from the RNT.
Figure 17.8 The patching of CAT 5 for multiple applications is best done with color coded patch cords, as here at the Royal National Theatre (RNT) in London.
516
However, the installation of CAT 5 is now so all pervasive in the commercial world that many facilities are standardizing on CAT 5 for all signalling applications. This does not mean that they are using ethernet to distribute DMX as such (although this is an option for multiplexing several DMX universes) but that the CAT 5 cable plant is used for carrying the standard DMX signal. The result of advances in technology now means that within a theatre the CAT 5 cable plant may be used for applications as diverse as: • Lighting control (DMX). • Ethernet LAN (linking consoles direct to dimmer racks without the use of DMX). • Ethernet LAN for conventional computer networks. • TCP/IP ethernet audio distribution. • Digital intercomm and stage management-cueing systems.
ENTERTAINMENT APPLICATIONS
• Telephone. Using the Royal National Theatre, once again, as an example, the Cottesloe Theatre within it has 400 CAT 5 lines. The mixed usage of the cable plant does require care. RJ45 patch panels are used, but the cables and connectors are color coded to indicate the different useage. The planning of DMX distribution also requires care. The RNT’s Olivier Theatre has nine universes of DMX and many distributed devices including 260 color scrollers and 40 distributed dimmer packs (sited FOH where there was previously inadequate provision) each 6 × 2.5kW. DMX splitters with optical isolation are used (see Figure 9.11 for example) to ensure that the remote devices receive an adequate signal, and that there is no danger of faulty equipment compromising the DMX signals. 17.2.4 Working lights Within a theater complex there are many activities that take place between performances. This means that the stage area must have good working lighting. Furthermore, even during a performance, there must be some backstage lighting for access and safety. Working lights are a permanent part of the theatre infrastructure so that, as a rule, there is no danger that they become ineffective due to over-enthusastic modification for that “special” production. They may be based on a mixture of lightsources including fluorescent and tungsten halogen. Even here the “scene” concept prevails, distinguishing between lighting needed during set construction, during rehearsal and during performance. The low level lighting needed for performance is referred to as “blue” light, because some of it is indeed blue filtered lighting, used when the stage is blacked out or operating at a low lighting level when any backstage lighting could ruin the effect. Working lights are usually operated through a separate relay/contactor system using a PLC or Room Controller “front end”. The use of a PLC or similar device allows for several different combinations, and for the creation of “scenes” specific to a production. Control can be from many different positions, some of which will be locked out by the PLC during per-
Figure 17.10 At the RNT a touchscreen controller is used as the user interface for operating the working lights. It is protected by a hinged transparent cover.
formance. Main control is from the prompt corner (stage manager’s) position, although some practitioners prefer the responsibility for working light control to pass to the main lighting control position during performance. When performances are not in progress, the working light system also controls lighting in the auditorium, either using separate “cleaning” lights, or through the houselights control system. 17.2.5 Rigging and rehearsal controls The complexity of modern stage lighting rigs means that there are some severe practical problems in setting up a lighting rig, even before rehearsals can begin. The lighting designer will have designed his lighting for a production in advance, using a computer program that allows him or her to visualize the
517
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
trol console itself must be moved to the rehearsal position, or that an alternative console is used for rehearsal. The designer starts by checking each individual luminaire as seen from the audience position, and gets any obvious mistakes corrected. Then he can start creating the lighting cues needed for the show – but again, most of these will have already been worked out in preliminary form, so the process is not one of starting from scratch, but of fine tuning the already designed concept. A problem that has emerged with the introduction of moving lights and similar devices is that of unwanted light from indicator lights. Most of these devices have LED displays or indicators on them to show status and addressing information. This is essential at setting up time, but the high point brightness is highly distracting when luminaires are visible to the audience in dark scenes. A recent introduction to these devices is the facility to switch off all indicator displays remotely. 17.2.6 Choice of console
Figure 17.11 Examples of hand held controls that can be used for testing sub-systems and as riggers’ controls to facilitate the installation and focusing of luminaires.
outcome, and that creates a full schedule of all the luminiares required. The circuit number, position and focusing requirement for each luminaire is also shown either in diagram or schedule form (or both.) This allows the lighting riggers to get ahead of the game. They can get everything in position before the designer starts work on site. However, to do this, they must have a simple facility for controlling luminaires on an individual basis. This is achieved by a rigger’s control a compact hand held device that gives DMX access to individual luminaires for focussing. Wireless versions of these are available. Once the rig is complete, the lighting designer moves into the auditorium to complete his or her work. This means that either the main lighting con-
518
See also Chapter 11. Andy Collier, now of Technical Marketing Ltd, but with many years experience with Strand Lighting and others, remarks that when he started in stage lighting a memory console cost around 100 times his annual salary. Today a console with better facilities costs maybe one third of the annual salary of a senior theater technician. The downside of this spectacular change is that, because the stage lighting market is not that big, and the actual number of consoles sold is quite small, there is now limited resource to develop the software used within consoles. Some types of console have left the market, not because they were no good, but simply because not enough were being sold to allow the manufacturers to keep the software up-to-date, and/or because they could not afford to update the hardware in the face of obsolescent components. While lighting designers have preferences for setting up a show, they are not the people who operate the console during the run of a show; and often are not even the people who actually enter in the show cues and do the detailed programming.
ENTERTAINMENT APPLICATIONS
Following on from this, it is common practice to have a back up console permanently available. At best this tracks the live console and can be switched in almost instantaneously. It is not for this book to say what constitutes a “leader”, since fashion and geography both play their part in the choice. Currently well known international names in the field are (in alphabetical order) ADB, Compulite, ETC, NSI/Colortran, and Strand. When moving lights are required to be treated separately, consoles from MA Lighting, Varilite and Whole Hog are popular, but this group are rarely installed as part of a permanent installation. Figure 17.12 Possibly the first system that was designed from the outset to provide remote control from an FOH position, and which allowed all control to be carried out by one seated person, was the Light Console, invented by Frederick Bentham of Strand Electric in 1934/5. Photo© Victoria and Albert Museum.
Operators who can work fast under pressure during set-up and rehearsal, and who are capable of comprehending large lighting rigs are few and far between. Most large facilities will, therefore, be equipped with a “standard” console that is able to control the installation well; but will also allow for the installation of additional or alternative consoles for individual productions in order to meet the needs or preferences of the operator and designer. The most common split arises when a production uses a lot of moving lights, in which case a separate console may be used for them, while retaining the house console for conventional lighting. In fact the leading consoles have good facilities for moving lights used in theatrical productions (as opposed to “rock and roll” productions) so such a split is not needed often. The installed console will tend to be one from one of the industry leaders for several good reasons: • there will be more operators around who know how to use it. • the manufacturer has the resources to provide a back up console at a few hours notice. • it is easy to hire in the same or similar model to extend the system when needed. • touring productions may well have their lighting already plotted to run on it.
17.2.7 Houselights Houselights within the large stage facility require many of the same considerations as set out in Sections 17.1.3, 15.4.3 and (especially) 15.4.4. However, there is a greater tendency to view the houselights as part of the overall show lighting resource, and for this reason they may well use the same type of dimmers as the stage lighting system and, at show time, be operated through the main console.
Figure 17.13 There is usually a requirement that houselights can be switched to full on in emergency. In London the facility is known as the Lord Chamberlain’s switch, in Scandinavia as the Panik button. The RNT system is required to transfer the houselights to an alternative supply, by-passing the dimmers. Photo from the RNT.
519
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Out of show time a separate scene store within the dimmer rack is used to allow manual operation from several points within the auditorium; and this operation may be combined with the ability to select additional cleaning lights. The operation of the houselights other than from the main console position is inhibited during a show. However local legislation usually requires the provision of a panic lighting facility that is operated by attendant staff using a concealed switch. In London the facility is quaintly still known as the “Lord Chamberlain’s Switch”. (The Lord Chamberlain is in charge of the Monarch’s Household, with the exception, you will be glad to hear, of the bedchamber. In the past he was also responsible for the morals and safety of Londoners in that he both censored plays and made the regulations for theaters.) This requires not only that the lights go to full brightness, but that they do so on a changeover system that by-passes the dimmer. These days the facility becomes part of the emergency lighting system.
17.3 Television 17.3.1 Introduction In the context of this section, lighting control for television refers to television studios. It has already been mentioned that there are other venues that require augmented lighting for TV, such as convention auditoria and large places of worship.
Television studios were, at one time, highly specialized places requiring very high lighting levels for the then insensitive cameras. They also required a lot of qualified staff to run them. With the advent of a new generation of cameras, the widening of the application of TV, the need to keep staffing to a practical minimum, and the enormous growth in demand for TV programs, there are now more “studios” than ever. However, they vary in complexity. Table 17.1 gives an idea of the variety of spaces used and their range of requirements in respect of lighting channels. However, what constitutes a “studio” is changing. They are as diverse as: • Professional broadcast studios permanently equipped with a high density of luminaires to cater for many kinds of programs, and quick changeover from one program to another. • Small automated (even announcer operated) TV news studios with a permanent, comparatively simple, lighting rig. • Commercial general purpose studios for a wide range of productions, including TV commercials, generally with a limited permanent provision, but arranged to accept additional equipment on a rental basis. • Studios used by educational establishments. • “Media centers” where a fixed lighting rig supports an informal “studio” space. As used by governments, financial organizations, and others who have frequent need to brief the media. • “Virtual reality” studios where the presenters work entirely or partly in front a blue or green cyclo-
Type
Typical area
Lighting power
Educational studio; small broadcast news studio Small general purpose rental studio; small fixed application broadcast studio Medium broadcast studio; general purpose rental studio Large broadcast studio
60–100 sq m 650–1,100 sq ft 100–200 sq m 1,300–2,200 sq ft
200–300W per sq m 20–30W per sq ft 250–350W per sq m 25–35W per sq ft
200–300 sq m 1,900–3,300 sq ft 600–800 sq m 6,400–8,500 sq ft
350–450W per sq m 35–45W per sq ft 400–500W per sq m 40001050W per sq ft
Lighting (dimmer) circuits 0.5 per sq m (10 sq ft) 0.5–0.6 per sq m (10 sq ft) 0.65–0.75 per sq m(10 sq ft) 1 per sq m (10 sq ft)
Table 17.1 Examples of TV studio sizes and the provision made for studio lighting. The lighting power is the actual maximum load used. Dimmer capacity is much greater.
520
ENTERTAINMENT APPLICATIONS
rama and where the background is inserted electronically. The majority of studios are fitted with some kind of lighting support “grid” with an associated power distribution system that allows luminaires to be sited anywhere in the space. Broadcast studios use time saving devices such as pole adjustable luminaires and motorized hoists. In order to achieve flexibility the power installation is much bigger than actually required. For example a big studio may have several hundred dimmer channels each rated at 5kW. While the distribution system must be fully rated, in the sense that a single dimmer or group of dimmers may well be required to run at full load, considerable diversity is allowed, so in reality the studio might never draw more power than 20−40% of the circuit capacity. Because the lighting layout may change frequently, and it may be difficult to make the necessary calculations in the heat of the job, large studios are fitted with power monitoring equipment that can warn of potential overload. 17.3.2 Dimmers for TV The type of dimmer used in a TV studio depends on both how big and how “professional” the studio is. In order to avoid problems of color rendering, and to allow continuous level control, incandescent (tungsten halogen) is the most widely used source. Here the maximum lamp rating used has come down. While in the past TV large studios might have been equipped with many 5kW and 10kW dimmer channels, now the largest individual lamp used is often
Figure 17.14 In large TV studios it is necessary to monitor the power being used by the lighting. Photo from LSI Projects Ltd.
only 2kW; so dimmer channel ratings have come down accordingly. Many smaller studios have dimmers rated at only 2.5–3kW, and even larger studios may have a majority at 3kW with a sprinkling of 5kW. An important consideration in TV work is noise, both acoustic and electrical. This means that if conventional thyristor dimmers are used, it is best to site them in a separate dimmer room. If there are to be dimmers in the studio itself, they need to be of a kind that make no significant acoustic noise. As far as electrical noise is concerned, the received wisdom is that dimmers should have as high a rise time as practicable. Rise time is specified in various ways, an
Dimmer type
TV Application
Multi channel dimmer packs; portable or wall mounting. Racked multi-channel dimmer packs, wall mounting dimmer racks, or modular racked dimmers. Modular racked digital dimmers with system monitoring. Modular racked dimmers with additional filter chokes.
School or college TV studio.
Rise time Ps 800010100
Commercial video studio. Small TV news studio. Media center.
2000010250
Broadcast TV studio. Television theater. Main TV news studio. Recording studio. Concert hall.
3800010450 6000010800
Table 17.2 Examples of dimmer rise times for centrally installed phase cutting dimmers.
521
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
accepted one being the time in microseconds for the full load output current to rise from 10% to 90% at 90° conduction angle. Table 17.2 shows some examples. However, as observed in Section 8.3.5, rise time on its own is an inadequate metric (partly because of lack of rigorous methods of measurement). In practice the biggest problem with dimmers in TV studios is their potential to interfere with sound systems. This is especially the case when floor level luminaires with flexible power cables are used, when it may be difficult to separate them from audio and video flexible cables. Back in the 1960s the BBC and others realized that if one could be sure that the dimmer would not interfere with audio, then it probably would not cause any other problem. Interference with audio is only a problem if it is audible, so an objective test is to measure the audible noise signal. The BBC specify dimmers in this way. Figure 17.15 shows a measurement circuit. Within the box “weighted network” is an R-C filter network weighted to the response of the human ear in respect of audible frequencies. The measurement is made using a true r.m.s. voltmeter, and must be made at all settings of the dimmer. A typical specification calls for the maximum r.m.s voltage measured this way to be 15mV for a 5kW dimmer operating with a full load. In the case of thyristor
L
Dimmer Load 0.33:
N Weighted Network 0.1P
150 15K 15K 0.01P
0.005P V RMS
Figure 17.15 Circuit used for measuring the noise produced by dimmers. It uses a network weighted to the response of the human ear. The measurement circuit is connected by screened cables of specified length.
522
Figure 17.16 Example of professional TV studio dimmer rack with modular plug in dimmers. The dimmers have 450ms rise time. Photo from ADB.
dimmers the maximum occurs at 90° conduction angle. Apart from its objectivity, this test has the advantage that it does not make any assumptions about the actual method of dimming. Large professional broadcast studios use rack mounted dimmers of modular construction. Digital dimmers are used to ensure an absolutely matched performance between units, and to facilitate dimmer monitoring. Commercial, educational and small broadcast studios use wall mounted dimmer racks. 17.3.3 Distributed dimming, control TV studios are being affected by the same developments as stage lighting. New light sources require new methods of control, and some “studios” warrant a complete re-think in terms of their control architecture.
ENTERTAINMENT APPLICATIONS
Fluorescent lamps are being used for fill or “soft” lighting. Some studio fluorescent luminaires are delivered with dimmable fluorescent ballasts, so they can be connected to a conventional dimmer system. Others are delivered with a controllable ballast with an accompanying DMX decoder. This in turn requires DMX to be available at the lighting grid. Some studios (especially those related to music shows) are now using moving lights, and they generally require a non-dim supply and DMX control. The arrival of silent IGBT and MOSFET dimmers means that it is now practical to use distributed dimming. In practice the technique is most suitable for special purpose studios such as news studios and media centers where the lighting rig is not changed often. The extra cost of the dimmers is offset by the reduction in cost arising from not needing a dimmer room and from the simplification of the infra-structure – for example by needing fewer power circuits. The emphasis in TV lighting control is on the image seen by the camera. The lighting director makes his decisions on what he sees on the production monitor, not what can be seen on the studio floor. The need is to be able to control lighting in groups, but at the same time to be able to make a fine adjustment to an individual luminaire. The TV camera has a difFigure 17.18 At the MTV Studio in London a conventional system based on high power centrally mounted dimmers was replaced by a distributed dimmer system using 2.5kW IGBT dimmers from IES. Self climbing hoists are fed by a 10kW power feed, that is then split into one switched and three dimmed circuits. Photo from Technical Marketing.
ferent response to that of the human eye – the action of a dimmer has most effect at the high end of the control range, unlike in stage lighting. The control consoles used in TV studios are, today, the same as those used in stage work. Very small studios use simple manual controls, all others use memory consoles of varying capacity and sophistication. Figure 17.17 Fluorescent soft lights, such as this example from ADB, use lamps with a color temperature of 3,200K in order to match the tungsten halogen luminaires. 5,600K is an optional alternative. They are available with dimmable ballasts, allowing connection to conventional dimmers, or with DMX compatibility.
17.3.4 Outdoor TV, Film lighting In TV studios the standard color temperature is 3,200K because this matches the tungsten halogen lamps used. However, any outdoor work, and the
523
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
17.4 Touring shows
Figure 17.19 The BBC multi-media news center at the Aldwych in London is an example of how unconventional some “studio” facilities have become. It uses distributed dimming both in the studio area (IES with DMX control) and in the waiting area (Helvar/Philips with DALI control). Note the presence of daylight.
majority of film (as opposed to TV) studio work is done at a nominal daylight 5,600K. There are also now TV news studios that include a high daylight contribution, and some of these work at the higher color temperature. It is inefficient to use filters on tungsten halogen luminaires to achieve the higher color temperature, so arc sources are used. The most widely used HID sources are metal halide lamps tailored to the needs of the industry. Conventional electromagnetic ballasts have been used, but these present serious practical problems. First they are very heavy, second the light output on standard mains is variable because the arc extinguishes each half cycle. Unless the camera is itself synchronized to the lighting, the result is flicker as seen by the camera. The problem is particularly severe with film. For this reason electronic ballasts, that provide a square wave drive to the lamp, are now nearly universally used. (See example in Figure 7.56.) Dimming of HID lamps can be done by using mechanical shutters. Some electronic ballasts provide a dimming facility, but this is of limited range and may affect the color temperature. (See Section 7.3.) It can be expected, however, that as HID sources continue to improve, and controllable electronic ballasts become standard, most studio luminaires will have a dimming option.
524
For the puposes of this section a “touring show” is taken to mean one where no permanent lighting rig exists at the venue. This excludes live theater productions that simply move from one fully equipped permanent venue to another, but includes things like rock concerts, outdoor theater, special events, and any show or presentation that uses a temporary lighting rig. This can include product launches, awards festivals and similar events that take place in hotel banqueting suites and large auditoria that are only partially equipped. The temporary nature of the events means that lighting directors/designers take a hands-on approach, and generally like to use control consoles of the kind reviewed in Sections 11.4−11.7. Moving lights are
Figure 17.20 Touring shows need robust and safe methods of distributing power. The 270H distribution unit from the Rubber Box Co is based on the use of different sizes of CEE connectors and is made from rubber.
ENTERTAINMENT APPLICATIONS
Figure 17.21 A typical portable dimmer pack used in small touring shows. This one is 6 × 2.5kW and is from Strand Lighting. See also Figure 8.16.
widely used because of the nature of the events. Nonetheless there is always a requirement for conventional dimmers. The special problem of these kinds of event is the safe and practical distribution of lighting power. Power is likely to be available from MCCBs or heavy duty switchgear in multiples of 3 × 63A, 3 × 100A, 3 × 200A or higher. Connection to power is by heavy duty cables fitted with high current connectors. Distribution is by temporary distribution boards fitted with circuit breakers and connectors for outgoing circuits. For moving lights and other non-dim luminaires this is all that is needed; but all dimmed circuits receive their power through “touring” dimmer racks. For small theatrical events these are quite compact, and may be a repackaged version of a product otherwise used for permanent installation. For most touring applications, however, the tendency is to somewhat higher channel ratings, and to special constructions. Touring dimmer racks have developed into a specialist product form, with customers often requiring bespoke arrangements in terms of protection and connection. While small systems can be built up using a number of portable “packs” like that of Figure 17.21, larger systems rack up a number of multi-channel units as shown in Figure 17.22. Special features can include: • Heavy current input connectors allowing the use of flexible power input cables. • Means of monitoring supply voltage and load currents – selectable to the different phases.
• Input circuit breakers. • Auxiliary non dim circuit breakers. • Standard output connectors (e.g. CEE 32A in Europe; Hubble Twistlock® or stage pin in USA). • (As an alternative) multipole output connectors allowing the use of multicore load cables. These go to lighting bars carrying multiple standard outlet sockets for connection of luminaires. Socapex® connectors are a favorite for this application. • Load patch panel. The use of multi-core load cables can result in unacceptable restrictions in terms of control. Therefore touring racks with multipole connectors are usually also fitted with comprehensive load patching. Patch connectors vary, but include Wieland 18/3 for currents up to 16A, and “Multi Lam” jack style connectors for up to 32A. • Indication of load status and circuit breaker tripping.
Figure 17.22 A touring dimmer rack from Avolites. This one is for 48 channels at 16A. It is fitted with ammeter, voltmeter and a local controller for dimmer configuration, DMX patching, and backup control.
525
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
• MCB options including the fitting of double pole breakers. • RCD protection option. • Local microprocessor control, with status indication. This is used to set DMX addresses and dimmer laws, and provides a back-up means of operating the dimmers. It can include individual channel controls to set up back-up scenes for use when the main console is not available. Temporary installations can suffer from accidents and errors in set-up. This means that touring dimmer racks must be proof against foreseeable hazards. Principal among these is the possibility of a loose neutral, which results in the dimmers or control equipment seeing phase-to-phase voltage. Where cooling fans are used, these must be of the intelligent variety that only operate when needed. The sensor used must also restrict dimmer output in the event that heatsink temperature limits are exceeded, whether or not this is due to fan failure.
Figure 17.24 Load patch panels can be incorporated into the dimmer rack. Avolites use Wieland connectors for 16A patch panels (top left) and Multi Lam jack connectors for 32A panels (top right). A touring rack with patch facility from ETC (bottom).
17.5 Outdoor shows, son et lumière, pyrotechnics 17.5.1 Spectaculars
Figure 17.23 The protection and connection arrangements of touring racks are customized. The Avolites rack can have individual channel output connectors (left) or multipole connectors (right). Both these examples show the use of Camlock™ power input connectors at the bottom of the rack.
526
Outdoor theater and concerts are staged events that use equipment similar in principle to their indoor counterparts, but either protected from or adapted to the weather. The show principles are the same, even if audiences may be much larger. Another class of event is the “spectacular”. This can range from a single event, to a seasonal event and includes such things as pageants and parades. International EXPOs and theme parks put on special shows in the evening to keep visitors in their park
ENTERTAINMENT APPLICATIONS
Pyrotechnics
Moving lights
Conventional lights
Scene projector(s)
Video projectors or L.E.D. displays
Fountains and special effects
Loud speakers
Control Interfaces
Amplifier racks
Remote Dimmers
Remote interfaces DMX
DMX recorder
Scene projection controller
Video server
Fountain (etc) control computer
Audio mixer and router
EIA485 or fiber equivalent Pyrotechnic Controller
Multi-track source e.g. hard disc Ethernet Hub and/or Port Server
Master Timecode Generator Safety hold-down
Main show computer
(may be derived from audio or video source)
Figure 17.25 Block diagram of the control system for a hypothetical outdoor spectacular.
until late at night. Many of these shows use the water and firework effects described in Sections 17.5.2 and 17.5.4, and all of them can be considered outsize versions of Son et Lumière (Section 17.5.3) since they are invariably accompanied by sound. While the one day or short season event takes advantage of the services of a staging company, the theme park spectacular tends to be highly automated in order to minimize the number of staff needed to run the show, and, particularly, to ensure that the show is consistent for every performance. The best approach to controlling spectaculars is to break them down into manageable chunks, so that each subsystem can be programmed separately. Figure 17.25 shows the idea. This arrangement also makes trouble-shooting much easier. The synchronization of the different elements of the show is not a problem if timecode is used. The degree of supervising automation varies because some human control will still be required in most cases, and there is no point in automating something which can be done
Figure 17.26 “Expo Noche” was the night time show at EXPO 92, Seville, Spain. The picture shows one of six water screens, each of which was served by a 70mm movie projector, a giant filmstrip projector, and a slide (scene) projector all using 7kW xenon lamps. The show also used lighting, searchlights, lasers, pyrotechnics and fountains. Each component had its own control sub-system, synchronized by time code. Control cable distances were up to 23km (1.2-1.8 miles) so fiber-optic (EIA485 equivalent) was used to distribute control signals. Show production by Resorte Communicacion. Photo from Lighting & Sound International.
527
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
just as well by a human being who is there anyway. Pyrotechnics, and any other show element involving life safety, must have human supervision. This in turn indicates the need for excellent communication between all those supervising the running of the show. Where long distances are involved there is a need to decide on the level of control protocol used, and the means of transmission. For example if there is a dimmer cabinet 1,000m (3,280 ft) away from the main control position, is it best to: • send the actual dimmer data as DMX? • or just send the timecode, and have a local DMX playback unit? Then, because “raw” DMX and, to a lesser extent, timecode are not suitable for long distance transmission, what kind of carrier should be used for the data? Can EIA485 be “souped up”, for example by using a fiber-optic link, to do it? Or is it better to convert to ethernet LAN? Today’s practice is definitely in the direction of using ethernet LAN in its wired, fiber and wireless manifestations; but “older” methods should not be discounted. The aim of the systems designer is a reliable and safe show, it is not to show how clever he or she is with computer technology.
Figure 17.27 ECA2 of France specialize in the production of spectaculars. This example is the Acqua Matrix show at EXPO 98 in Lisbon, Portugal. The show used moving metal structures, 20m high jets of flame, pyrotechnics and extensive lighting. The main feature is a 30m high inflateable “egg”. This 9,000 cubic meter structure used a blowing capacity of 150 cubic meters per second for rapid inflation. Inside it, six synchronized large format filmstrip projectors presented the giant images.
528
Figure 17.28 Examples of submersible luminaires. These ones from PEM Fountain Co are made in cast bronze with stainless steel fittings. They accept the PAR 56 300/500W lamp in 12, 24, 120 or 240V versions, and can be fitted with dichroic color filters. When these are used a shade is used to block stray color (scattered from the dichroic filter, and not the same wavelength as the required color) from the edge of the luminaire.
17.5.2 Fountains Illuminated fountains are a popular night time entertainment, whether they are stand-alone, part of a spectacular, or part of a Son et Lumière production. The luminaires used for the underlighting of fountains have to be completely watertight and corrosion resistant, and their color filters are made of tempered glass. Best color results are obtained with dichroic filters, but the range of colors is limited. Both high voltage (120/240V) and low voltage (12/24V) lamps may be used. The latter give a problem with volt drop, and the former may give problems with permissible residual currents (only 5mA in the USA and Canada) so in either case the electrical installation must be planned with great care. Control of fountain lighting can be based on standard dimmers, but the dimmer arrangement must be matched to the loads; for example in respect of transformer type and protection. Very often the control of the lighting of fountains is related to the control of the fountain jets themselves. There is a wide range of options in respect of fountain spray or jet control.
ENTERTAINMENT APPLICATIONS
Figure 17.29 Special jet for the creation of water screens. If a water presssure of 12.4 Bar is applied to this one, a fan radius of 18m (60ft) is produced. This requires over 3,000 l/m (800 US gallons/minute). Photo from PEM Fountain Co.
Choice of jets and sprays. A fountain array may have several different types of jet, such as straight jets, fan jets, dandelion spray etc. Control of pressure. For each jet type the control may be simple on/off or may require proportional control to achieve variable heights. Choice of direction. Jets are usually fixed, but moving or swinging jets are also available. These require special consideration for lighting. The control of water pressure is done using several different methods. One method is to have a separate electric pump for each main fountain effect; with the possibility of varying the speed of the pump to produce different jet heights. While it is possible to install solenoid operated valves after the pump to switch individual jets on and off, this can only be done where an individual jet represents a small part of the pump load. Big installations use one or more big pumps (similar to fire pumps) that build up pressure in a manifold. Excess
Figure 17.30 Proprietary fountain and light “packages” are offered by Waltzing Waters Inc. The top picture shows a complete system for indoor use, and the middle picture is an outdoor show, about 200ft (61m) wide. The systems are factory built, and can be placed into position in one piece, as shown in the bottom picture. The method of control is interesting. All “Waltzing Waters™” shows are synchronized to music that is played from a conventional audio CD. The data for controlling the lighting and fountains is carried in the least significant bit of the audio signal, which in the circumstances of show playback is effectively in the noise floor of the audio signal. The audio is played back using the analog outputs of the CD player. The digital output is fed to a decoder that extracts the show data from the data stream.
529
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
specialists who offer a “package deal” that includes equipment and changing shows. Many permanently installed public fountain systems use Programmable Logic Controllers to run simple sequences. A device that has become popular as a component in spectaculars is the water screen. Figure 17.26 shows an example. While it is possible to use a “drop” screen, where a curtain is produced by falling water, this implies some kind of superstructure. Therefore a system relying on a jet sited at water level is usually preferred. The problem is to achieve a thin “screen” that will diffuse the light and create an image, but is not so disturbed, by wind or the water falling back on itself, that the image is blurred. The standard way of doing this is force water to strike a Figure 17.31 The Bellagio Hotel in Las Vegas has a spectacular sound, light and fountain show that takes place at intervals during the evening. Lighting design by DHA Design.
pressure is released by a dump valve. Such an arrangement can then feed many jets simultaneously, with each set of jets (or individual jet) having its own solenoid operated or proportional control. Suitable proportional valves may themselves be compressed air operated. An automated fountain show system must, therefore, be able to issue: • DMX commands for lighting control. • Analog signals for the control of proportional valves (and possibly pump motor speed). • “Digital” or on/off switch commands for solenoid valves. • (Possibly) serial control commands for special controllers or associated devices. Because programmed lighting control is so well developed, one option is to use a lighting console for all the control, and adapt the fountain control to work from DMX. The proportional control of jets is not dissimilar to controlling moving lights. Show systems can then easily run preset sequences either direct from a console or from a DMX recorder, and can easily be synchronized to SMPTE timecode for synchronizing to a sound track. There are also dedicated computer programs for fountain control. These are often used by fountain
530
Figure 17.32 The “Big Melt” show at Magna, Rotherham, UK, is about as spectacular as Son et Lumière gets. Every half an hour a 1960s vintage electric arc furnace comes to life. A fully operational overhead gantry crane carries a DHA light curtain as one of many dramatic lighting effects. Real (propane fuelled) flames are augmented by 8m high showers of sparks. It would be impractical to use stage fireworks for this, so the effect is available on demand by melting lengths of metal wire (using arc welding technology) and blowing the molten mass out in a jet of air at high pressure. Exhibit design by Event Communications, Lighting design by DHA Design, special effects by Howard Eaton Lighting Ltd, Audio design by Peter Key. The main control is based on the use of DMX, run from a Whole Hog playback system, and using 84 channels of dimmers in additional to the special effects. Photo by Richard Handley.
ENTERTAINMENT APPLICATIONS
metal plate to produce a 180° arc spray, using a special water screen jet like that shown in Figure 17.29. When water screens first appeared the projection on to them was done by scene projectors and 35 or 70mm film projectors; however the latest generation of high output video projectors now give enough light to make them viable sources for the application. 17.5.3 Son et Lumière As can be inferred from the name, Son et Lumière as an entertainment had its origins in France. Although the basic idea can be traced back to the beginning of the 20th century, it came into vogue in the 1950s as a way of encouraging evening visitors to French chateaux. Early productions used a recorded sound track with the lighting being operated manually, but very soon the systems used various methods of automation so that shows could be run consistently without the need for significant operating staff. If the truth be known many of the early Son et Lumière productions were rather dull. The late
Figure 17.33 In the French town of Bourges they have a “walk round” Son et Lumière show in the summer months. Visitors walk about 2km (1.2 miles) in a 90 minute period during which they see six separate shows related to different buildings in the town. Some of them use image projection like the one shown here at the Augustin Cloisters. Show designed by Philippe Noir and Christine de Vichet; Lighting Design by Pierre Boudeau and Vladimir Lyszczynski, sound by Daniel Deshays, images by Serge Fouillet. Main contractor AEB. Each show segment is played back from CD, and the lighting and projection show data is also carried on the CD. Control system by Electrosonic France.
Figure 17.34 St Michael’s Cave Gibraltar has had a Son et Lumière show for nearly 40 years. It has recently been updated. During the day the cave has “architectural” lighting with ambient sound for walk-round visitors. In the evening the main show takes place, using a substantial multi-channel sound system and programmed lighting. The lighting is based on 12 moving lights and 48 dimmer controlled channels of conventional luminaires. Control is by a DMX recorder of the type shown in Figure 9.22. Lighting design by, and photo from, David Atkinson.
Frederick Bentham writing in 1968 (see Reading List) referred scathingly to “..in Son et Lumière a kind of steam-radio documentary of the worst type tied to such visual changes as the present façade, often not there at that particular moment of history, can be persuaded by floodlighting to take”. Here he summed up the main problem. A single building façade cannot support a 50−60 minute show on its own − even with a limitless equipment budget there are only so many ways it can be illuminated, and even then the relationship between the illuminated building and the recorded script is likely to be tenuous. Today’s Son et Lumière productions are more exciting because they recognize the limitations of the medium. Show sequences are may be comparatively short, for example 10−15 minutes, with the opportunity for the audience to move round so they see several “mini-shows” in sequence, each associated with a different building or different view of a large building. Alternatively or in addition the shows introduce large scale projection to introduce images that relate directly to the story. Some examples are shown in the accompanying figures. Productions are not confined to castles and fine houses, they can be seen
531
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
applied to caves and disused steel works as well. The control technology depends on the scale of the show, and on whether it is a permanent or temporary installation. Temporary Son et Lumière productions can easily be mounted by staging companies, who will likely use standard “touring” lighting control systems (see Section 17.4) with a control console of their choice set to run an automatic sequence linked to SMPTE timecode. Small permanent installations can use show controllers of the kind described in Section 15.7.2. If small scale slide projection is used, the controller can operate the slide projectors directly. If big scene projection is used the projectors may be controlled by a lighting controller, but are more likely to have a dedicated controller of their own that can follow timecode. Large permanent installations use DMX recorders as show playback devices (see Section 13.2). 17.5.4 Pyrotechnics The control of pyrotechnics can be considered as being on the edge of “lighting control”, but is nonetheless of relevance because pyrotechnics often form part of outdoor spectaculars. Special low power pyrotechnic devices are also used for indoor shows. Technically the control of pyrotechnics is not difficult; however there are major safety issues involved, and the advice is to leave the handling of pyrotechnic devices to experts. In many locations their use is controlled by legislation. All public shows use electrical firing. The fireworks themselves, whether they be shells fired from mortars, or free-standing “fountain” type, have a fast fuse fitted. This can be ignited by a simple electrical igniter (also known as electric matches or squibs) of the type illustrated in Figure 17.35, which is attached to the fuse. The characteristics of a typical igniter are shown in the figure. It can be seen that the critical parameter that sets off the reaction is the current flowing through the igniter. An example of a manual firing control is shown in Figure 17.36. It has the following characteristics: • A key operated master switch. • Each firing circuit has a three-position switch. Center position is OFF (safe). A momentary up posi-
532
tion is TEST. This passes a small current through the firing circuit which is displayed on the main meter. This allows for the safe testing of the complete igniter circuit, and is a vital feature both for ensuring show integrity and improving safety (by helping to eliminate the problem of unfired shells.) The down position of the switch is ARM. This does not fire the igniter, but makes the circuit active. • This particular control is for 60 circuits, divided into four groups of 15. At the left hand end of each row of switches there is a circuit breaker that arms the row. • A FIRE button fires the armed circuit(s). In practice any circuit that has already fired goes open circuit, so there is no need to switch off already fired circuits. • Four heavy duty multipole connectors for the cables that are run out to the firing positions. The control shown is operated from a 12V automobile battery, for which there is a battery condition indicator. For shows involving long cable runs higher voltages are often used to ensure sufficient firing current. Although the rated firing current for the device Resistance: approx 2 Ohms Highest current for non ignition: 200mA Minimum current guaranteed to fire: 370mA Recommended firing current: 1A Recommended maximum test current: 10mA Firing time at 1A: 2ms
Figure 17.35 Example of electric igniters used for pyrotechnic shows. The igniters are protected by a sleeve (one of them can be seen with the sleeve withdrawn). The yellow plastic cap at the other end of the wire is to protect the copper wire before connection. A typical specification is given above. Photo from The Firework Shop.
ENTERTAINMENT APPLICATIONS
Figure 17.36 Example of a manual firing control used for outdoor shows. Photo from The Firework Shop (UK).
shown in Figure 17.35 is only 200mA, it is recommended that at least 1A be available to ensure reliable firing − the resistance of very long cables, combined with the resistance of the device itself may mean that 12V is not sufficient. Often one circuit fires several igniters simultaneously. This is usually done by paralleling the igniters, on the grounds that if a connection to one fails, the others will still work. However some practioners advocate series firing for safety reasons, because less energy is required, and because when the circuit is tested it gives assurance that all igniters are in circuit (testing a circuit with a number of igniters in parallel cannot indicate an individual igniter connection failure). Big shows, and all automatic shows that run nightly at theme parks, are pre-programmed and are often choreographed to accompany music. The equipment used is similar to that used for other sound and light programming, using programmed switch commands to fire the igniters. For safety reasons, however, it is usually separate from other parts of the show system. Figure 17.37 shows a computer controlled firing system that has the following features: • It can be synchronized to SMPTE timecode, so it can follow a master music track. • For security it has a “hold down” control. An operator who has a full view of the show, and is in communication with other show direction and safety monitoring staff, must hold this down during the firing sequence − if he lets go, all firing stops (similar
idea to the “dead man’s handle” on electric trains). • It does not directly connect to the launch points, but has one or more serial outputs (EIA485.) These are fed to local decoders or firing modules that in turn have connections to the individual igniters. The outputs may have an additional safety feature which is a provision to short circuit all outputs until “armed”. • It provides a complete circuit test facility, using the bi-directional serial communication. Very big shows may have many launch points, and use multiple firing control systems linked by timecode. Sometimes individual devices may be remote from the control position, so even on small shows RF remote control may be used. This is done using secure RF control communication; for example a small group of devices might be controlled using the signal shown in Figure 9.58. Some companies offer RF control of complete systems as in Figure 17.38. However when RF is used, great care must be taken with arming procedures. Small scale pyrotechnics are also used indoors. The most widely used stage fireworks are referred to as gerbs,which produce a “fountain” effect. They are classified by the height of the effect and by their burn time; for example a 3.6m (12ft) height with 6 second burn. Other devices available include: Mines, which propel an effect up into the air, like a miniature mortar. Maroons, which make a very loud bang, and which must be let off in a bomb tank.
Figure 17.37 Computer controlled pyrotechnic firing system. The Pyrodigital Phase III system from Pyrodigital (USA).
533
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Flame projectors, which propel a big jet of flame for a stated distance for a few seconds. Waterfall. Wire guided rockets. All stage fireworks must be fired electrically, and simple control panels are available for this purpose, as shown in Figure 17.39. These may use capacitors to provide the firing current. This arrangement (which is also used in some outdoor systems) can have the following advantages: • It can deliver a high peak current, ensuring reliable firing. • There is no risk that if a large number of devices are to be fired simultaneously, there is a firing fail-
Firing controller
Gerbs mounted in an array holder
Bomb tank for the firing of theatrical maroons
Figure 17.39 Indoor firework components from Le Maitre (UK and Canada).
ure due to lack of power. • It allows the use of smaller power supplies. • After firing the output is “dead”, and cannot give a second firing unless rearmed.
17.6 Stadia, arenas, sporting facilities 17.6.1 Introduction Figure 17.38 RF firing system from Parente Fireworks (Italy). Each outstation (above) can control 24 firing circuits, and will only respond to security-coded commands from the master controller (below).
534
Sports venues require lighting to different levels according to the activity, and recommendations for these are well documented (for example in the IESNA Recommended Practice booklet listed in the Read-
ENTERTAINMENT APPLICATIONS
ing List). The great majority of lighting used is HID based, but there are contributions from fluorescent lighting, and for venues catering for special events even some contribution from incandescent. The majority of control is by relay/contactor systems of the type described in Section 12.6. There are several running cost, security, and practical issues associated with the lighting of big sporting venues that are best illustrated by example. Acknowledgement is given that the information concerning the Manchester (UK) installations has been provided by the Manchester office of Arup. 17.6.2 An Aquatics Center
Figure 17.40 The Manchester Aquatics Centre, site of aquatic events at the 2002 Commonwealth Games, has an Olympic standard main pool (top) a diving pool (center) and a training pool (bottom). The lighting was designed by Arup & Partners, and the control system was supplied by Home Automation MEM250. See the main text for description. Photos from Arup.
The Manchester Aquatics Centre has three main facilities. An international standard swimming pool with seating for 1,600, a diving pool and an underground training pool. The whole complex is used for competition events, training, medical research and leisure. The main pool is lit by a combination of 40 1kW metal halide floodlights and 24 high pressure sodium floodlights. These are used to achieve 600 lux (horizontal) for competition, and 300 lux (horizontal) for training. Competition lighting is provided by the metal halide lamps; lighting for leisure at 300 lux (horizontal) is by a mixture of HPS and metal halide. The diving pool is illuminated to 600 lux (horizontal) at one meter above the water level using metal halide lamps. The training pool, which has a low ceiling, is lit by fluorescent lighting to 300 lux for training and to 600 lux for underwater photography and medical research. In order to provide complete flexibility in setting up lighting scenes, and to allow monitoring of lamp life so that lamp replacement can be planned, every luminaire is switched separately via relay cabinets sited around the center. The relay cabinets are fitted with manual over-ride switches for testing and as a back-up facility, but normally the relays are operated from a central computer using a bi-directional control link. The computer is equipped with a GUI that presents the user with a number of pre-set options
535
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 17.41 The control position at the Manchester Aquatics Centre. Photo John Waite, Arup.
for the lighting of the different types of event. It also provides: • Mimic diagrams showing the status of the lighting. The mimic diagrams can be “zoomed” to allow more detailed information about individual luminaires to be displayed. • The ability to set up new “scenes” as required, or to modify existing scenes. • System monitoring information, in particular the lamp life monitoring already referred to. • A modem connection to the system manufacturer that provides remote diagnostics. Day-to-day operation of the system is from the Operations Manager’s office; but with a system of this kind there is no reason why additional control points can not be added at a later date. Significantly the GUI was developed in conjunction with the user. This once again makes the point that the best control systems result from having both user input and enough time to allow system fine tuning in the light of actual use experience.
A brief description of the GUI not only gives an idea of how the whole system operates, but also reveals the practical requirements of this kind of system. Figure 17.42 shows the main menu page. This allows the user: • Direct control of the main groups of lighting, password authorized. • Access to several “pages” of status information and control, corresponding to the main levels of the stadium. • Access to the alarm history log. • To view the current system global status, for example “Daytime”, “Evening”, “Event”, “Post event”, etc. The figure also shows that this page can be made to display a choice of control mode. All except the “View only” mode require password access. Any mode other than “View only” reverts to “View only” after 15 minutes if no use is being made of the system. The modes available are: View only. Security, allowing manual switching of groups of luminaires in the stadium. Event, this allows access to the event menu, where the timing and lighting requirements of specific events, taking place within the next seven days, are set up within the constraints of preset lighting level combinations.
17.6.3 A Stadium Like the Manchester Aquatics Centre, the City of Manchester Stadium was built to be ready for the 2002 Commonwealth Games. The control system works on similar principles, but, because it is much bigger, many more features are required.
536
Figure 17.42 The main menu page of the lighting control system at the Manchester Stadium. This figure and the ones following are courtesy Arup.
ENTERTAINMENT APPLICATIONS
Supervisor, this allows for setting up the preset lighting arrangements for each area. Administration, this level is used to set passwords for each user. It is also used to enter certain system wide parameters, such as the lighting level at which light level sensors trigger a lighting change. Figure 17.43 shows three of the “level” pages. When selected in “view only” mode, these pages simply display the lighting status, but if the user has the necessary authority, he or she can use the selection “buttons” to switch the circuits on or off. Selection is by mouse “point and click” (although in principle there is no reason why touch screen could not be used). Figure 17.44 shows an example of an event setup page. This shows that events use different lighting levels; for example an evening training session would use a much lower level than a televised event. The nature of discharge lighting, and the large loads involved, mean that there is a measure of automatic control underlying the simple manual or timed automatic operation described so far. Figure 17.45 shows one of the “Floods overview” pages. These can give details of the operation of each individual lamp. They also indicate the availability of lamps, and the ramp up process needed to achieve a required light level. Two factors must be taken into account. First, it should not be possible to switch on large groups of lamps all at once; the warm up current would be too high. The floodlights used in this installation reach a normal running current in approximately 2.5 minutes after switch-on. Thus lamp groups must be switched on in sequence in order to stay within maximum current limits. There may also be operational reasons that determine a maximum running load. Second, the lamps do not hot restrike. Thus if a lamp is switched off, it is not available again for about 15 minutes. In order to deal with these practical points, the “buttons” providing individual lamp control indicate not only “on” and “off”, but also “available”. Thus lamps that are “off” may not be “available” to turn on, either because of maximum current restrictions, or because they are cooling down.
Figure 17.43 Examples of the “Level” status pages. Two of these show the lighting status at basement and main concourse level, the third shows the status of the tower lighting. The “buttons” change color from grey (off) to yellow (on).
537
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 17.44 Example of an event set-up page at the Manchester Stadium.
The “Floods overview” pages allow the user to see individual lamp running hours, number of starts and other maintenance related data. When individual lamps near the end of their life, the system triggers an alarm that must be acknowledged. At the City of Manchester Stadium the two main positions for lighting control are in the Event Control Room and the 24 hour manned Security office. The computers sited in each are identical, their use being solely determined by the authorization status of the user. The system also provides for:
Figure 17.45 An example of the “Floods overview” page. These pages have sub-menus that allow control and status viewing of individual lamps.
538
Figure 17.46 The Event Control Room at Manchester Stadium showing the lighting control computer terminal. The control system was specified by Arup, and designed and built by enLIGHTen as part of the Hayden Young electrical installation contract works.
• A fire alarm input that automatically brings on designated lighting circuits. • Manual control of emergency lighting circuits in the event of power failure (the lighting control computers themselves are fitted with uninterruptable power supplies.) • Manual control of additional lighting during an event. If daylight conditions are unexpectedly dark, the event manager can incrementally increase the floodlighting level. However, lights can only be switched ON during an event. They can then only be switched off when “Post event” mode is selected. The system computers are on an ethernet network, but communication to and from the lighting control relay units is by C-Bus, a proprietary protocol of Clipsal Integrated Systems. The Clipsal system requires the installation of C-Gate™ server software on the main computers and then uses a proprietary portserver, the Etherlite Terminal Server, to communicate to the remote equipment. At the Manchester Stadium this device supports seven separate C-Bus networks on EIA232.
ENTERTAINMENT APPLICATIONS
17.6.4 Arenas Covered arenas for sports such as basketball, ice hockey, tennis and boxing use lighting control systems based on exactly the same principles described in Sections 17.6.2 and 17.6.3 above. However, many of these venues are multi-purpose and for this reason may require more flexible lighting systems to cater for the different events. The needs of television are taken into account from the beginning. Often major sporting events, like international tennis tournaments and boxing championships, have elaborate “pre-match build-ups” to hype up the crowd and entertain television viewers. The additional lighting needed for this is usually provided by a staging company working for the event promoters. Also, many of these venues are used for concert events. In both cases the staging company will bring in a complete show lighting rig with its own control system. Arenas must, therefore, have provision for powering large temporary lighting systems. This in turn may mean that load monitoring is required to prevent both the sports TV lighting and the show lighting being switched on at the same time.
Figure 17.47 Indoor arenas use lighting controls similar to those described for stadiums, however, flexibility may be required to deal with the needs of different sports (and associated television lighting levels). Photo of the Target Center, Minneapolis.
17.7 Theme parks 17.7.1 Attractions Lighting control in theme parks brings together many different applications, and the approach taken depends very much on the procedures favored by the particular park operator. The differentiator here is the need for “park wide” thinking. It is, as with so many applications of lighting control, best to think in terms of self-contained subsystems. Looked at this way, the theme park simply becomes a collection of many of the other applications already discussed – including visitor centers, exhibitions, retail outlets, restaurants, theatrical stages, spectaculars etc. Components of theme parks that have special lighting control implications are rides, night time spectaculars and parades. Rides can be considered in a similar way to themed exhibits, the difference being that the audience is completely controlled in terms of their posi-
tion in relation to the display. Most dark rides use comparatively static lighting, sometimes augmented by dynamic effects that are triggered by the ride vehicle and/or the ride controller. Lighting control for the show lighting may use a combination of architectural and theatrical technique, depending on the nature of the ride. Rides themselves are controlled by PLCs (Programmable Logic Controllers). The main issue is safety, and being sure that if anything fails it does so in a predictable manner. For this reason the major park operators have very tight specifications as to what may and may not be used. It is permissible for lighting sub-systems to operate from the PLC, but some vendors provide lighting control cards (putting out DMX) that fit in to the PLC frame itself. If there is a ride fault or emergency, the PLC calls up the lighting required. A temporary fault, arising, for example from someone having difficulty getting on or off the ride, may make no adjustment or may
539
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Figure 17.48 The Spider-Man® ride at Universal Studios Islands of Adventure in Orlando, Florida combines big screen 3D moving image projection with huge scenic sets. The audience are unaware of where hard “scenery” finishes and projection begins. The audience travel on moving vehicles that themselves incorporate a computer controlled motion base. The illusion is helped by the sophisticated lighting design (by Anne Millitello of Vortex Lighting), and the careful matching of scenic lighting levels with projected lighting levels. Changes in lighting scenes and spot lighting effects are triggered by trackside sensors that detect the passing of the ride vehicle. Picture courtesy of and © Universal Studios.
simply modify the existing lighting scene. But any other fault will result in working lights or emergency lights being switched on as a quite separate function from the show lighting. Night time spectaculars. Some principles involved with these are reviewed in Section 17.5. The exterior lighting is a very important part of the theme park experience – it must make a good overall impression and help persuade visitors to visit a particular attraction. When a night time spectacular is to take place, it is necessary to douse all park lighting that would otherwise detract from the show, while at the same time retaining sufficient low level lighting for safety. Most of the exterior building lighting will be HID,
540
so the only practical method is to sequentially switch the lighting off. Sometimes, however, lighting near where the audience are standing is fluorescent or incandescent, allowing dimming to be used for the last stage of blackout. In either case the equipment doing the dimming or switching is reasonably close to the luminaires. There is, therefore, a requirement to have remote control of this from a central position. Such control can be based on standard architectural lighting control systems, or may be proprietary. Parades. Parades are part of the fun of theme parks. Most run during the day time where lighting is not an issue, but some run at night, and can introduce an interesting lighting control problem.
ENTERTAINMENT APPLICATIONS
Fixed Loudspeakers
802.11b
802.11b
802.11b
Moving Floats
802.11b
PARADE ROUTE
Local lighting controllers To Amplifiers
Ethernet Hub 802.11b
Main computer lighting control
Audio source
Audio mixer and router
Figure 17.49 Basis of controlling lighting and sound automatically on a theme park parade route. The main moving floats are all on a wireless LAN, and they report their position to a central computer that triggers the appropriate lighting cues, and also keeps the “fixed” sound playback in sync with sound coming from the moving vehicles.
When a parade takes place, there may be a requirement to modify the lighting along the parade route to match that of the float vehicle that is passing by at a particular moment. One way of doing this is to manually select the lighting on the basis of closed circuit TV surveillance of the parade. An alternative is to automate the process, but this can only be done by having a system that automatically reports the position of each float. A method that works is as follows. Each float is equipped with a controller that itself is wireless LAN (e.g. 802.11b) enabled. This does require that the LAN architecture itself is carefully planned, but assuming that the parade routes are fixed, it is possible to ensure the required coverage by a thorough commissioning process. Each float is also equipped with a method of determining its position. One way is to “sow” the parade route with passive RF tags (the same kind of device that is used to recognize individual automobiles at highway toll booths) buried under the road surface. A tag reader mounted under the chassis of the float reads the tag identity, and the information is transmitted back to the central controller over the LAN. An alternative position sensing method is to equip the float with a GPS (Global Positioning Sat-
ellite) reader. This then gives a continuous update of the float position. However, in order to get sufficient accuracy (to within about 10ft, 3m) it is necessary that the base station is also GPS equipped and for it to return a correcting signal back to the float. The same position reporting method can be used as the basis of keeping float-based sound in sync with the sound coming from fixed loudspeakers on the parade route.
Figure 17.50 The daytime parade at Universal Studios Japan uses the GPS system to locate the floats. Photo courtesy of Universal Studios Japan.
541
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
PARADE ROUTE WITH PASSIVE TAGS
Tag reader
Micro
Float equipment
802.11b
TC
GPS
802.11b
Solid state Sound store
Computer
only succeeds if the park management and operational culture is right. Much maintenance is done by practical hands-on people who are good at observing and fixing; but are not necessarily enamored of computer screens. Systems designers therefore need to think not only of the initial installation and the problems of programming shows and ambient lighting in the first instance, but also the realities of dayto-day operation by staff who may well change several times during the life of the system.
17.8 Entertainment within retail As already observed in Section 16.1, shopping has become a social activity. When not associated with the necessities of life it can become an entertainment for many people. Taken to an extreme a shopping location can introduce pure entertainment shows and features. Such features are usually free, and are designed to attract people to the location. They have the result that casual visitor traffic increases so circulation spaces have to be increased accordingly. The problem is not so acute when the attraction is in the circu-
Figure 17.51 Two methods of determining the position of a moving float. The use of passive tags buried in the parade route (above) and the use of GPS position location (below). The GPS system can also be used to generate timecode to synchronize sound sources or effects carried on the float.
17.7.2 Park wide systems The major parks have all been in place for some time, and one of the problems they face is the need for continuous development and improvement. The other is the need to have a very high standard of maintenance to ensure that attractions with a design life of 10 years or more look as good as new throughout their operational life. On the lighting control side one approach being taken is to think in terms of park wide lighting control, so that all the separate sub-systems are networked together and can be centrally monitored. This really
542
Figure 17.52 Strand Lighting offers the ParkNet facility management system for theme park “park wide” lighting control. In effect it is a supervisory system that rides on top of the required mixture of localized entertainment and architectural lighting control sub-systems.
ENTERTAINMENT APPLICATIONS
lation space of a large mall, but may become so when it is sited within a store. Depending on the nature of the attraction, the lighting control system can be similar to those used for other visitor attractions such as visitor centers and theme parks. However, there is a need to provide greater flexibility: • It must be easy to re-program the control of individual areas. Retail spaces are changed frequently. • It must be easy for store or operating staff to select lighting scenes for special events. • There may be a need for completely different lighting sequences for different times of day or different levels of visitor traffic. An example demonstrating all these principles is the new Toys “R” Us in Times Square, New York. This 100,000 sq ft (9,300 sq m) store is built on three levels, but includes a big atrium space that accom-
Figure 17.54 At NikeTown in London circulation spaces are animated by multi-media shows that mix big screen video sequences of athletic and sporting achievement with dynamic lighting. The same spaces are used for special attractor events, such as personality appearances. This requires the use of a separate “event” lighting control. Original lighting design by Natasha Katz, lighting installation by AC Lighting. Mixed media show control by Electrosonic.
Figure 17.53 The Ferris wheel at the Toys “R” Us store in Times Square is equipped with 200 neon chevrons, 420 ft (128m) of LED strip and 14 strobe lights. An Entertainment Technology Horizon playback controller rides on the wheel to provide a selection of lighting sequences. Lighting design by Focus Lighting Inc.
modates a 60ft (18m) high Ferris wheel. The whole store has 39 different areas, some pure retail, some entertainment and mostly a mixture of both. Many of these areas are sponsored by major brand owners, who may well require them to be changed on a seasonal basis, or when new products are introduced. Focus Lighting Inc had the task of providing a lighting scheme that worked as a coherent whole, but provided the necessary localized flexibility. For control they standardized on DMX protocol, and the whole scheme needed eight universes of DMX. Show playback uses four Entertainment Technology “Horizon” playback controllers. These are supplied with a monitor screen that allows the store staff access to a number of “virtual faders” that can change lighting scenes or vary the sequence speed.
543
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
Focus Lighting cite the need to vary the lighting sequence speed on the major attractions to match visitor traffic, and the need to select spot effects to match a mood or to interact with visitors. When visitor interaction is not required the system runs either automatically from a master time clock, or under the direction of the main building management system.
17.9 Discotheques, dancefloors and clubs Discotheques and clubs vary greatly in scale, so the lighting equipment used varies accordingly both in power and quantity. Similarly the budgets available determine the nature of the equipment used. At the top end of the market professional equipment of the same kind as used for touring shows and theater is found; but at the low end of the market some of the
Figure 17.55 One of the dance spaces at the Ministry of Sound in London. Such spaces require a wide range of specialist lighting devices, that in turn implies the need for many channels of control. Photo from Pulsar Light of Cambridge.
544
Figure 17.56 The “LJ” must have a clear view of the floor, and easy access to all the controlling devices. Photo from Pulsar Light of Cambridge.
equipment borders on the “cheap and cheerful”. In most markets, however, the expectations of customers has raised standards, and the large worldwide demand has resulted in equipment that is amazing value for money. The essence of disco and club lighting is that it is dynamic, and that there is a conscious effort to match the lighting to the music being played. A venue of any size requires a team of operators to achieve the required result. If the music is live, there is someone to mix the sound, possibly a member of the band. But if the music is “disco” then the disc jockey takes charge. But the DJ has become a performer in his own right and is not physically able to do much more than present the discs. Also, if the music is live, there may or may not be a separate lighting operator; but again if it is “disco” there is almost always an LJ (Light Jockey) who works alongside the DJ to produce the overall effect. In large venues there may be more than one LJ; a new breed is the VJ (Video Jockey) who operates video source equipment and routes it to multiple displays. In some cases video screens are used as lightsources. Club and disco lighting consists of all kinds of fixed and moving luminaires and a variety of special
ENTERTAINMENT APPLICATIONS
Figure 17.57 Many LJs like to play lighting effects like playing a keyboard. This programmable touch panel from Pulsar Light of Cambridge allows them to do so.
effects, for example: Stroboscopes and multiple flash devices (e.g. flexible plastic tubes fitted with a series of flash tubes that can flash in sequence). Mirror effects, ranging from the classic rotating mirror ball to multiple mirror devices with several moving elements. Projected effects, not confined to standard gobos and slides, but including rotating images and images of immiscible liquids. Multiple beam luminaires. These have a single light source that is imaged by many lenses to produce multiple pencil beams. The whole optical assembly may rotate at variable speed. Lasers. These are effective, but as with “spectaculars”, great care must be taken in their installation and use. Because of the required minimum beam height (i.e. the level at which the laser beams must be kept to avoid the possibility that anyone can see into the laser beam) the use of lasers is confined to the larger and higher venues. “Smoke” or “fog” machines. These are used as an effect in their own right, but also, as in theatrical and spectacular events, to make laser beams and other forms of beam lighting more visible. Club and disco spaces can be small relative to large stages, and usually lower light levels are used. This means that most of the luminaires can be based on tungsten halogen sources. This reduces cost and greatly simplifies control since conventional dimmers can be used. Clubs often have multiple spaces, including bars and “quiet” spaces that can be served by conventional architectural lighting control. However, in most cases there is a demand for systems that provide compara-
Figure 17.58 Examples of specialist lighting devices used in club and disco lighting. The “Flexiflash” from Pulsar contains a series of xenon flash tubes that can be programmed to flash in sequence (top). The Astroaggi “Power” (center) and the “Atlas” (bottom) both from Clay Paky are examples of rotating multi-beam luminaires.
tively static lighting for early evening, but which become more dynamic as the evening progresses. Thus the favored control is one that provides maximum “hands on” control for the wilder sessions, but which can also be used to set up complex scenes and sequences that can operate fully automatically at other
545
LIGHTING CONTROL – TECHNOLOGY AND APPLICATIONS
times. The enormous number of circuits used, and the fact that many of the devices used are outside the scope of conventional architectural lighting, has led to the development of special controllers such as those shown in Figures 13.5, 13.6 and 17.56. An interesting consequence of this is that some of the controllers originally developed for the club and disco market are now also used for more conventional architectural lighting control, because of their flexibility and value for money.
17.10 Conclusion Lighting affects our lives in many ways, and the enormous technical advances made through the twentieth century opened up new ways of both meeting
546
basic lighting needs and creating lighting that is attractive. Lighting control is now an integral part of good lighting schemes, but its role is not always fully understood. It is hoped that this book provides both a basis for understanding the principles involved, and a source of background information for its many applications. The principles of good lighting design and sensible control are largely independent of technology. There is, therefore, no need to be afraid of technology, but the flexibility afforded by new technology now makes many previously expensive and complex applications easier to realize. However, the guiding motto remains “keep it simple”. Unless a lighting control system is easy to use, it cannot achieve its objective, whether this objective is functional, esthetic or “green”.
SOME SUGGESTIONS FOR FURTHER READING
Some suggestions for further reading
In no way does this list claim to be comprehensive. Nonetheless it does provide a basis for further study on both the technical and esthetic aspects of lighting. An effort has been made to include books that are readily available in either the UK or the USA or, in most cases, both.
Books Lamps and Lighting. J.R Coaton and A.M. Marsden (editors) 4th Edition 1997. Butterworth - Heinemann. (Originally published by Arnold in the UK and John Wiley in USA.) An invaluable textbook, which has been kept up-to-date through several editions since its first appearance in 1966. Written by many expert contributors with direct experience of both the lamp making and luminaire industries. Lighting Historic Buildings. Derek Phillips. Architectural Press 1997. A “must” for everyone concerned with lighting design. Phillips’ architectural background brings great insight, and the examples have lessons for contemporary lighting design, not only for old buildings. Lighting Modern Buildings. Derek Phillips. Architectural Press 2000. A companion to the “Historic” book. Phillips’ 40 year career in lighting design and his real feel for the subject makes this a useful review. Includes 59 well illustrated case studies. Lighting Design. Carl Gardner and Barry Hannaford. The Design Council (UK) 1993. Subtitled “An Introductory Guide for Professionals” this is an excellent, well illustrated, review of the field. Particularly good comments on the role of the lighting designer.
Lighting. D.C. Pritchard. Addison Wesley Longman, 5th Edition 1995. A useful textbook for the engineering student or technician. Includes the basics of lighting, lighting design, lamps and luminaire design. Lighting Handbook. Mark S. Rea, Editor in Chief. Illuminating Engineering Society of North America (IESNA), 9th Edition 2000. An invaluable (if heavy and expensive) 1,000 page reference and applications review. Should be in every lighting company’s technical library. Lighting Technology. Brian Fitt and Joe Thornley. Focal Press 2nd Edition 2001. Has the subtitle “A guide for the entertainment industry”. Strongest on TV and Film lighting. A practical review of professional lighting equipment and practice. Lighting Systems in TV Studios. Nick Mobsby. Entertainment Technology Press 2001. Strong on the nuts and bolts of TV studio installations, but with an emphasis on UK practice. Stage Lighting Controls. Ulf Sandström. Focal Press 1997. A clearly written primer on the operation of stage lighting controls, invaluable for anyone using stage lighting. Control Systems for Live Entertainment. John Huntington. 2nd Edition Focal Press 2000. A useful handbook that reviews the practical methods of show control. Covers lighting, sound, video and special effects. Good source of information on MIDI Show Control. Stage Lighting Design. Richard Pilbrow. Design Press (USA) Nick Hern Books Ltd (UK) 1997. Subtitled “The Art, The Craft, The Life”, this book,
547
SOME SUGGESTIONS FOR FURTHER READING
from a (if not the) leader of the profession, is a delight to read. Full of know-how, practical advice, insight, anecdote, and contributions from many leading entertainment lighting designers. The Art of Stage Lighting. Frederick Bentham. Sir Isaac Pitman & Sons Ltd (UK) Pitman Publishing Corporation (USA) 1968. Written at the time that electronics was taking over from electromechanical systems, it includes extensive chapters on the lighting control systems of the time. A reminder that most aspects of complex stage lighting control have been well understood for many years. Sixty Years of Light Work – an autobiography. Fred Bentham. Entertainment Technology Press 2001. Originally published in 1991. Fred (when Pitman published his books, he was recorded as “Frederick”, but was known to thousands as “Fred”) was hugely influential in the evolution of lighting controls, but, by the time this book was written, could also see where he got some things wrong. A very personal memoire with some revealing insights as to how the theater lighting industry developed. This edition includes a facsimile edition of Strand Electric’s 1936 catalog, with some fascinating pictures of the big multi-channel control systems of the time.
transformers. A classic that first appeared in 1947, it is one of the few books devoted to transformer theory and design. Betriebsgeräte und Schaltungen für elektrische Lampen. C.H.Sturm and E. Klein. Siemens (Germany). 1992 in German. The subtitle is “control gear, transformers, starters, lamps and lighting standards”. The serious industry textbook for those involved in the design of lighting ballast and interface products.
Booklets Fördergemeinschaft Gutes Licht of Frankfurt, an association within the German Electrical Industries Federation, ZVEI, promotes good lighting practice. It publishes a series of informative booklets on different aspects and applications of lighting. 15 booklets are available, all but four available in English as well as German. Of particular relevance is No. 12 Economical Lighting Comfort with Lighting Electronics (1996.)
EMC for Product Designers.Tim Williams. Newnes 3rd Edition 2001. An excellent guide to the realities of EMC.
Lamp Designation System – LBS – Lampenbezeichnungssystem another booklet from ZVEI (Zentralverband Elektrotechnik- und Elektonikindustrie; Postfach 70 12 61 60591 Frankfurt am Main, Germany). This one explains the letter codes used to describe lamps for general lighting.
The New Let there be Neon. Rudi Stern. ST Publications. 1988-1996. A colorful review of the art of neon; from Claude’s fundamental work in France in the early 1900’s to neon sculpture in the 1980s.
DALI Manual. A booklet from the DALI Activity Group of the Luminaires Division of ZVEI. A useful introduction to the DALI protocol published in September 2001. Includes a list of the companies that are members of the group.
Electronic Transformers and Circuits. Reuben Lee, Leo Wilson and Charles E, Carter. 3rd Edition. John Wiley & Sons Inc. 1988. Should now be titled “Transformers for electronic circuits” since its subject is low power conventional and high frequency
Specifier Reports and Lighting Answers are published by the National Lighting Product Information Program (NLPIP) in the USA. NLPIP is administered by the Lighting Research Center at Rensselaer Polytechnic Institute, Troy, NY.
548
SOME SUGGESTIONS FOR FURTHER READING
IESNA Recommended Procedure for Determining Interior and Exterior Lighting Power Allowances. Publication Reference LEM-1-99. Published by IESNA. Summarizes the provisions of the standard ASHRAE/IESNA 90.1-99 in respect of lighting power allowances for buildings and mandatory lighting control. IESNA also publish a whole series of Recommended Practice booklets. These are well illustrated, and are particularly strong on the subject of lighting levels required for different activities. Some of them embody the relevant ANSI standards. Examples are Sports and Recreational Area Lighting (IESNA RP6-01) Lighting for Houses of Worship (IESNA RP25-91) and Museum and Gallery Lighting (IESNA RP-30-96). There are currently 25 booklets in the series. ILE (Institution of Lighting Engineers, UK) also publish a series of reports. Examples are Brightness of illuminated advertisements (ILE TR05) and Interior high intensity discharge lighting (ILE TR21). The ILE reports are strong on exterior lighting. The Building Regulations 2000 – Conservation of fuel and power. Approved documents L1 and L2. Published by The Stationery Office and the Department for Transport, Local Government and the Regions (UK). These documents correspond to the IESNA “Power allowances” document referred to above, but apply to the UK. They mandate lighting efficiency performance for residential buildings (L1) and commercial-industrial buildings (L2). Precise color communication. A 60 page booklet with the subtitle “Color control from perception to instrumentation” published by Minolta, Japan (1998). An excellent introduction to the subject of color and how it can be measured.
Society Journals Lighting Research & Technology. Published quarterly by The Society of Light and Lighting, part of the Chartered Institution of Building Services Engineers (UK). Journal of the Illuminating Engineering Society. Published twice yearly by the Illuminating Engineering Society of North America (USA). The Lighting Journal. Published bimonthly by the Institution of Lighting Engineers (UK).
Trade Magazines There are several trade magazines and magazines devoted to specific applications which report on the use of lighting control. Examples of trade magazines are: Lighting Equipment News. Published monthly by EMAP Trenton (UK). Covers all kinds of commercial lighting equipment. A good source of business news and new product information. LD+A, Lighting Design + Application. Published monthly by IESNA (USA). While published by the society, this is not the society journal. It is good for case studies and news of new products in the architectural lighting field. Within each application area there are specialist publications for which lighting and lighting control are important components in a wider mix. A list of such publications would be extensive. In the professional entertainment world mention should be made of Lighting and Sound International, published in the UK, and Entertainment Design published in the USA.
549
TABLE O INDEX F ACRONYMS
Table of acronyms In order to simplify the index, the meaning of common acronyms is listed here. Often the organization, device, unit or protocol is better known by its acronym than by its meaning, so in some cases the main index only refers to the acronym.
AASHTO ABYC AC ACN ACTFEL ADC AGL ALU AM AMX ANSI AP APU ARCNET ARP ASC ASHRAE ASCII ASIC ASK ATM AV BACnet BAS BCD BEF BLF BMS CAD CAN CBM CCT CCFL CD CDM CD-ROM CEBUS CELMA CENELEC CFL CIE CISPR
550
American Association of State Highway and Transportation Officials American Boat & Yacht Council Alternating Current Advanced Control Network ACThin Film ElectroLuminescent (Display) Analog to Digital Converter Aviation Ground Lighting Arithmetic/Logic Unit Amplitude Modulation Analog MultipleX American National Standards Institute Access Point Auxiliary Power Unit Attached Resource Computer NETwoRrk Address Resolution Protocol Alternate Start Code American Society of Heating, Refrigeration and Air-conditioning Engineers American Standard Code for Information Interchange Application Specific Integrated Circuit Amplitude Shift Keying Asynchronous Transfer Mode Audio Visual Building Automation Control networks Building Automation System Binary Coded Decimal Ballast Efficacy Factor Ballast Lumen Factor Building Management System Computer Aided Design Control Area Network Certified Ballast Manufacturers Correlated Color Temperature Cold Cathode Fluorescent Lamp Compact Disc Ceramic Discharge Metal halide Compact Disc Read Only Memory Consumer Electronic BUS Committee of European union Luminaire Manufacturers Associations Comité Européen de Normalisation ELECtrotechnique Compact Fluorescent Lamp Commission Internationale de l’Éclairage Comité Internationale des Perturbations Radiotechniques
CMOS CMRR CMY CMYK CPU CRC CRI CRT CSMA/CA CSMA/CD CTC CWA dB DAC DALI DC DHCP DIN DJ DLP™ DMD DMX DPSK DRAM DSI™ DSP DSSS DTMF DVD DVMRP EBU ECL EEI EEPROM EFR EIA EIB ELV EMC EMF EMI EPROM ERP ES ESTA FAA FCC FDDI FELV FET FHSS
Complementary Metal Oxide Silicon Common Mode Rejection Ratio Cyan Magenta Yellow Cyan Magenta Yellow black Central Processing Unit Cyclic Redundancy Check Color Rendering Index Cathode Ray Tube Carrier Sense Multiple Access/ Collision Avoidance Carrier Sense Multiple Access/ Collision Detection Color Temperature Control Constant Wattage Autotransformer deciBel Digital to Analog Converter Digita Addressable Lighting Interface Direct Current Dynamic Host Configuration Protocol Deutsches Institut für Normung Disc Jockey Digital Light Processing Digital Micromirror Device Digital MultipleX Differential Phase Shift Keying Dynamic Random Access Memory Digital Serial Interface Digital Signal Processor Direct Sequence Spread Spectrum Dual Tone Multiple Frequency Digital Versatile Disc Distance Vector Multicast Routing Protocol European Broadcasting Union Emitter Coupled Logic Energy Efficiency Index Electrically Eraseable Programmable Read Only Memory Europäische Funk-Rundsteurung Electronic Industries Alliance (or Association) European Installation Bus Extra Low Voltage Electro-Magnetic Compatibility Electro-Motive Force Electro-Magnetic Interference Eraseable Programable Read Only Memory Effective Radiated Power Edison Screw Entertainment Services and Technology Association Federal Aviation Administration Federal Communications Commission Fiber Distributed Data Interface Functional Extra Low Voltage Field Effect Transistor Frequency Hopping Spread Spectrum
TABLE OINDEX F ACRONYMS
FIFO FM FOH FPGA FSK GFSK GLS GPS GSM GTO GUI HID HPS HTML HTP HTTP HVAC IC ICAO ICMP IEC IEEE IESNA IGFET IGMP I/O IP IR IrDA ISM ISO JFET LAN LASER LCD LED LEP LISN LJ LON® LSI LTP LV MAC MCB MCCB MIDI MIRED MKS MMF MOSFET MR MSC NRE NRZ OEM OLED
First In First Out Frequency Modulation Front Of House Field Programmable Gate Array Frequency Shift Keying Gaussian Frequency Shift Keying General Lighting Service Global Positioning System Global System for Mobile communications Gate Turn-Off (thyristor or switch) Graphic User Interface High Intensity Discharge High Pressure Sodium HyperText Mark-up Language Highest Takes Precedence HyperText Transfer Protocol Heating, Ventilation, Air Conditioning Integrated Circuit International Civil Aviation Organization Internet Control Message Protocol International Electrotechnical Commission Institute of Electrical and Electronic Engineers Illuminating Engineering Society of North America Insulated Gate Field Effect Transistor Internet Group Management Protocol Input/Ouput Internet Protocol Infra-Red Infra-Red Data Association Industrial Scientific and Medical (band) International Standards Organization Junction Field Effect Transistor Local Area Network Light Amplification by Stimulated Emission of Radiation Liquid Crystal Display Light Emitting Diode Light Emitting Polymer Line Impedance Stabilizing Network Light Jockey Local Operating Network Large Scale Integration Latest Takes Precedence Low Voltage Media Access Control Miniature Circuit Breaker Moulded Case Circuit Breaker Musical Instrument Digital Interface MIcro REciprocal Degree Meter Kilogram Second Magneto-Motive Force Metal-Oxide Silicon Field Effect Transistor Multifaceted Reflector M(IDI) Show Control Non Recurring Engineering (charges) Non Return to Zero Original Equipment Manufacturer Organic Light Emitting Diode
OSI OTP PAPI PAR PCM PD PFC PLC PLL PNC PPFD PROM PTC PWM RAM RARP RCB RCBO RCD RDM RF RGB SAE SCR S-DRAM SELV SI SMPTE SMTP SPL SRAM STP TCP/IP THD TTL UART UDP UGR UL UPS USB USITT UTP UV VA VCA VCO VCP VDR VITC VJ VLSI VRAM WECA XML
Open Systems Interconnection One Time Programmable Precision Approach Path Indicator Parabolic Aluminized Reflector Pulse Code Modulation Potential Difference Power Factor Correction Programmable Logic Controller Phase Locked Loop Perceived Noise Criteria Photosynthetic Photon Flux Density Programmable Read Only Memory Positive Temperature Coefficient Pulse Width Modulation Random Access Memory Reverse Address Resolution Protocol Residual Current Breaker Residual Current Breaker with Overload protection Residual Current Device Remote Device Management Radio Frequency Red Green Blue Society of Automotive Engineers Silicon Controlled Rectifier Synchronous Dynamic Random Access Memory Separated (also Segregated) Extra Low Voltage Système International (d’Unités) Society of Motion Picture and Television Engineers Simple Mail Transfer Protocol Sound Pressure Level Static Random Access Memory Shielded Twisted Pair Transmission Control Protocol/ Internet Protocol Total Harmonic Distortion Transistor Transistor Logic Universal Asynchronous Receiver Transmitter User Datagram Protocol Unified Glare Rating Underwriters’ Laboratories Uninterruptable Power Supply Universal Serial Bus United States Institute for Theater Technology Unshielded Twisted Pair Ultra-Violet VoltAmpere Voltage Controlled Amplifier Voltage Controlled Oscillator Visual Comfort Probability Voltage Dependent Resistor Vertical Interval Time Code Video Jockey Very Large Scale Integration Video Random Access Memory Wireless Ethernet Compatibility Alliance eXtensible Mark-up Language
551
INDEX
Index
A weighting, 92 AASHTO, 494 ABYC, 476, 477 AC, 17−19, 215 three-phase, 19−22, 181−184, 422 Access point, 341 Accumulator, 105 ACN, 327−330, 341, 342, 373 child device, 330 component ID, 329 device description language, 329 device management protocol, 329 group, 330 network management protocol, 329 protocol stack, 328 qualifier, 330 root device 330 session data transport, 329 ACTFEL, 153 ADC, 88 flash, 89 Adder, four bit with carry, 98 full, 97 half, 97 Address, bus, 103 DALI, 295 DMX, 291 dotted decimal, 326 EIB, 317 IP, 326 LON, 309 memory location, 103 AddressPro™, 336 Agriculture, 472, 473 Airconditioning, 417 Aircraft, APU, 481 cabin management system, 482 interior lighting, 481−485 military, 484 power, 481 Airgap, 196, 254 Alternating current, See AC Alternator, 17−19 single phase, 19 three phase, 20 ALU, 105 Ampere (Amp) 2, 7
552
Ampere-hour, 15 Ampere-turn, 13 Amplifier, 59, 86 class A, 72 class B, 73 class C, 74 class D, 74, 90 differential, 75 operational, 94, 281 voltage, 71 AMX192, 283, 288 Analog (control or signal) 85, 280−283, 354 Angle, conduction, 84 delay, 84 phase, 26, 28 trigger, 84 Anode, lamp, 119 sacrificial, 478 thyristor, 81 ANSI, 289, 320, 328, 331 Antenna, 340 Anti-fuse, 110 Anti-reflection coating, 45 AP, 341 Aquaria, 457 Aquatic sports, 535, 536 Arc, carbon, 112, 130 mercury vapor, 119, 130 sodium vapor, 133,134 ARCNET, 331 Arenas, 539 ARP, 325, 329 Art-Net, 373 Art galleries, 451, 452 ASC, 290, 292 Assemblies, 446 ASHRAE, 330 ASHRAE/IESNA standard 90.1, 348 ASCII, 88, 501 ASK, 318 Atom, 1 Audio-visual (AV) 436, 440, 446, 447, 458 Auditoria, cinema, 343, 442 concert hall, 444 lecture theater, 441
Aurora Borealis, 119 Authentication, 341 Automobile lighting, 490–494 Autotransformer, 24, 184, 255, 503 constant wattage, 206 Avalanche effect, 65, 66 Aviation ground lighting (AGL) 485− 490 approach, 485 control, 487−490 runway, 486 transformer, 487 BACnet, 330−332, 342, 418 objects, 331 properties, 331 protocol stack, 331 reliability property, 331, 332 Balanced signal, 76 Ballast, 121, 172 cold cathode lighting, 191, 240−242 controller, 377 core and coil, 207 CWA, 206 DALI, 296 efficacy factor, 203 electromagnetic, see Ballast, electromagnetic electronic, see Ballast, electronic energy efficiency index, 202 HID lighting, 204−207, 242−245 IEC reference, standard, 196, 197 low pressure sodium HID, 134 lumen factor, 202 mercury vapor HID, 132, 204 reactor, 193, 195 self, 132 Ballast, electromagnetic, 121, 192−207, 346 airgap, 196 construction, 195−204 copper fill factor, 201 efficiency, 197, 198 fluorescent lamp, 192−204 HID lamp, 204−207 instant start, 194 losses, 199 low loss, 201 manufacture, 199 pre-heat start, 193 rapid start, 193
INDEX
temperature rise, 204 thermal protection, 202 Ballast, electronic, 122, 219−245, 346 automatic shutdown, 231 compact fluorescent lamp, 233−235 controllable, 235−238, 240, 346, 347, 377, 450, 479 DC operation, 230 dimmable, 239, 377, 452, 523 fluorescent lamp, 228−232 frequency, 219 generic, 219 HID, 242−245 lamp failure detection, 231 lamp starting, 225−228 monitoring, 230, 231 multiple lamp operation, 232 power factor correction, 220−223 Barrel shifter, 111 BAS, 403 Base, 69 Batibus, 314, 390 Battery, 4, 15, 340, 392, 477, 490, 532 BCD, 87 BEF, 203 Bel, 91 Belt (control) 359 Binary (data) 87 Binary coded decimal, 87 BIOS, 106 Bit, 87, 285 dominant, 313, 316 recessive, 313, 316 stuffing, 313 Blackbody, 41 locus, 54 radiator, 41 radiation, 41, 113 BLF, 202 Blind, motorized, 428, 431, 434, 438 Bluetooth, 342 Blue wool test, 452 BMS, 394, 403, 404, 415, 440, 451, 460, 471 Boats, 476−478 Bombtank, 533, 534 Boolean algebra, 98 Boost, 185 converter, 216 Bordoni system, 255 Bridge, data, 307 rectifier, 59, 215, 220 Wheatstone, 256 Buck, 185 Building regulations Part L, 350 Bus, address, 103
control, 103 data, 103 field, 314 parallel, 87,103 results data, 111 topology, 305 Byte, 87 CAD, 102, 369 CAN, 311−314, 394, 494, 497 data frame, 312 error detection, 313 frame format, 313 Candela, 46, 47, 155 Capacitance, 7 Capacitor, 9−11 DC blocking, 224, 225 electrolytic, 11 loss factor, 213 metallized paper, 11 metallized polyester, 11 over-pressure, 214 parallel, 11 power factor correction, 29, 211 safety, 213, 214 series, 11 storage, 220 X, 220 Y, 220 Carrier, semiconductor, 64 Carrier sense, 305 Category, data cables (CAT) 305, 516, 517 aircraft landing, 485, 489 Cathode (lamp), cold, 127 fall, 127 heating, 226, 236−238 hot, 119 lamp, 119 resistance, 227 sputtering, 237 temperature, 227 Cathode (thyristor) 81 CBM, 203 C-Bus, 538 CCT, 51, 120, 137, 138 CD, 86, 87, 89, 104, 423, 529, 531 CDM, 137, 470 CEBus®, 319, 332 common application language, 321 data frame, 321 CELMA, 202 CE mark, 351 CENELEC, 272, 390 Cell, primary, 15 library, 102
secondary, 15 Centi-, 5 Channel, conduction, 77 control, 353, 359 start, 357 Charge (electric) 1, 7 Chase, 362 Checksum, 287 Chirp, 321 Choke, 172, 213 Chromaticity, 51 co-ordinates, 54 Chromaticness, 57 CIE, 49, 494 chromaticity diagram, 54, 369 color rendering index, 53 LAB color space, 57 LUV color space, 57 Cinema auditorium, 257, 376 Circadian rhythm, 472 Circuit breaker, 31 magnetic, 31 nuisance tripping, 31 thermal, 31 CISPR, 272 Client-server (architecture) 322 Clock, 87 CMMR, 75 CMOS, logic gate construction, 99 transmission gate, 96 CMY, 51, 369, 372 CMYK, 51 Code, ASCII, 88 binary, 87 bi-phase, 294, 302, 310 binary coded decimal, 87 hexadecimal, 88 hopping, 338 machine, 105 Manchester, 294, 310, 338 operation, 109 source, 105 Coercive force, 14 Collector, 69 Collision detection, 304 Color, 51−57, 470 chroma, 51 fastness, 452 hue, 51 model, 57 primary, 50 secondary, 50 space, 57 triangle, 53 tristimulus value, 53, 55
553
INDEX
Color (continued) tristimulus meter, 56 value, 51 Color filter, 50, 370 semaphore, 371 Color matching, 56 metameric, 56 spectral, 56 Color mixing/programming, 371 additive, 50, 372 exterior lighting, 401 neon lamp, 241 subtractive, 50, 372 Color rendering, 53 index, 53 Color scroller, 357, 371, 516, 517 Color temperature, 41, 51, 251, 458 change, 114 correlated, 51 Communication, duplex, 284, 340 half duplex, 284, 340 parallel, 285 reliability, 311, 322, 325 serial, 285 simplex, 284, 289, 340 synchronization, 284 Comparator, 89 Compensation, 211 Compiler, 106 Component, electronic active, 62 discrete, 93 integrated, 93 opto-,68 passive, 62 Compound, 1 Computer, 102 lighting control, 332, 430 operating systems, 332 personal, 103, 359 program, 105 programmable, 102 Concert hall, 438 Conductor, 2 Conduction heat, 163 intrinsic, 65 Conductivity, 6 Cones, 45 Conference center, 439 Connectors, 512–514, 524, 525, 526 Console (lighting) 511, 518, 519 ADB, 519 Avolites Diamond, 363 back-up, 358 Celco Gold, 362
554
Colortran Innovator, 359 Compulite, 519 controls, 358−361 displays, 358 ETC, 369, 519 generic, 359 Light console, 519 live show, 362 MA lighting, 519 memory 356−361 multiple preset, 355 Multiway, 357 NSI Colortran, 519 playback, 361 group, 361 page, 363 preset, 361 tracking, 361 rock and roll, 362 Rockboard, 362 Strand 300 series, 358 Strand 500 series, 360, 519 Virtuoso, 373 Whole Hog, 372, 519 Contactor, 35, 385, 387 Control rooms, 508 Convention center, 440 Converter, buck, 216 boost, 216, 220, 222 critical conduction mode, 223 DC, 215 flyback, 216 Cornea, 45 Correctional facilities, 463 Corridors, 459, 463 Corrosion potential, 477 Cosine, 16 φ 28, 211, 212 Coulomb, 2, 7 Counter, 99 program, 105 Courtrooms, 447 CPU, 105 CRC, 287, 313 Crest factor, 21, 207 CRI, 53, 120, 135, 137 Critical angle, 43 CRT, 54, 104, 171, 358, 509 Crossfade, 355 dipless, 355 CSMA/CA, 341 CSMA/CD, 305, 341 CTC filter, 371 Curie point, 174 Current, constant, 94, 487, 488
crest factor, 229 eddy, 23, 176 exciting, 176 fault, 183 gain, 69 holding, 82 inrush, 114, 229, 257, 383, 492 leakage, 229 limiting transformer, 184 magnetizing, 174 peak forward surge, 83 ratio in dB, 92 reverse leakage, 65 ripple, 59, 223 sensing, 265 waveform in reactor ballast, 193 CWA, 206 Cycle, regenerative, 115 tungsten halogen, 116 DAC, 88 DALI, 249, 271, 283, 292−300, 332, 341, 350, 381, 382, 389, 412, 418, 430, 434, 451, 481, 498, 524 address, 295 backward data, 295 coding, 294 components, 295 conductor size, 299 data rate, 294 isolation, 294 power, 298 programming, 299 scene control, 295 signal tolerance, 294 Dance floor, 544–546 Darlington pair, 74, 235, 491 Data, analog, 85 binary, 87 decimal, 87 differential transmission, 284 frame, 289 null, 289 packet, 289 parallel, 87, 285 port, 103, 286 serial, 87, 285 single ended transmission, 284 Datagram, 323, 324, 326 Daylight, 402, 411, 414, 433−435, 441, 449, 461, 472 dB, 91 dBA, 92, 444 DC, 15 component in transformers, 189, 190
INDEX
converter, 215 Debugging, 105 Deci-, 5 deciBel, 91 Deka-, 5 Delay, propagation, 96 Delta connection (AC) 21, 22 Depletion layer, 64 DHCP, 327 Diac, 225, 235, 260 Dielectric, 9 constant, 11 Digital (control or signal) 86 Dimmer, 250−279 aircraft, 484 analog, 260 consumer, 374 digital, 264, 522 distributed, 277, 515, 517, 523 fade time, 375 filter, 266 firing circuit, 262 flickering, 258 flood lighting control, 449 half wave, 279 harmonics, 268 hybrid, 278 laws, 251, 377 leading edge switching, 189 modular, 260 multi-channel, 259, 451 multiple autotransformer, 255 noise, 445, 521, 522 plug-in, 420, 421, 522 professional building, 374 protection, 267 remote control, 258, 261, 374 resistance, 253, 353, 354 retrofit (home) 421 saturable reactor, 254 ship, 479 shutter, 252, 524 sine wave, 211, 276, 445, 515 slider, 415 switchplate, 258, 420 table lamp, 261 television studio, 521, 522 touring, 259, 525, 526 transistor, 269−271, 445, 515, 523 triac, 257−269, 260 thyratron, 255, 257 thyristor, 189, 257−269 trailing edge switching, 189, 515 triac, 189 variable autotransformer, 189,211, 254 zero-crossing detector, 262
Dimming, compact fluorescent lamps, 124 compact source metal halide lamps, 139 electroluminescent lamps, 153 electromagnetic ballasts, 210 electronic ballasts, 239, 249 electronic transformers, 247, 248 flat lamps, 149 fluorescent lamps, 122, 282 half wave, 279 HID lamps, 248, 249, 282, 498, 502, 524 HPS lamps, 135, 248 incandescent lamps, 114, 282 low voltage lighting transformers, 189, 210 metal halide lamps, 138 non-electronic, 252 LEDs, 157, 249 miniature cold cathode lamps, 129 pre-heat, 252 shutter, 138, 252 speed, 376, 378, 461, 464 transformers, 210 tungsten halogen lamps, 116 DIN rail, 258, 390 Diode, 63 block, 281 bridge, 215 free wheel, 270 light emitting, see LED parasitic, 79, 218, 270 photo, 67, 407 PIN, 67 point contact, 63 semiconductor junction, 63−66 voltage reference, 66 zener, 66 Direct current, See DC Disc (disk) compact, 104 floppy, 104, 356 hard, 104 operating system, 106 Discotheque, 544–546 Distortion, 73, 74, 86 crossover, 73 harmonic, 222, 478, 479 Diversity, 521 DLP™, 437, 509 DMD, 140 DMX, 249, 284, 288−292, 352, 354, 357, 383, 396, 440, 455, 481, 511, 512, 515, 517, 518, 523, 530, 539 addressing, 291 alternate start code, 290
capacity, 290 color selection by, 371 data stream, 288 error correction, 291 gateway, 397, 398 moving light attributes by, 368 moving light control by, 366 network node, 327 patching, 357, 517 playback controller, 302, 397, 399, 400, 456, 543 recorder, 398, 399, 456, 457, 532 resolution, 290 signal distribution, 291 universe, 291 DMX512-A, 289 DMX512/1990, 288 Domain name system, 327 Doppler effect, 410 DOS, 106 DPSK, 342 Drain, 77 DRAM, 108 Driver, high side, 224, 235 low side, 224, 235 DSI, 293 DSP, 110 DSSS, 341, 342 DTMF, 338 Dumet wire, 113 Duplex (communication) 284 DVMRP, 329 Earth, functional, 273 leakage trip, 32 safety, 273 EBU, 301 ECL logic gate construction, 99, 101 EEI, 202, 348 EEPROM, 107 Efficacy, see Luminous efficacy EFR, 501 EIA, 259, 320 EIA232, 103, 284, 285, 298, 331, 333, 338, 430, 447, 455, 456, 538 EIA422, 286 EIA485, 286, 288, 298, 331, 333, 383, 387, 455, 482, 489, 504, 528 EIB, 293, 298, 314−317, 332, 383, 390 addressing, 315 configuration, 317 products, 316 protocol data unit, 315, 316 EIBA, 315 Electric
555
INDEX
charge, 1, 7 current, 7 energy, 7 field strength, 11 force, 11 potential, 7 power, 7 Eddy current, 23, 176 Electrical distribution, 15−24 Electricity, 1−4 supplies, 33 Electroluminescence, 151 Electroluminescent, lamp, 151, 481 sheet, 152 thin film display, 153 wire, 153 Electromagnetic compatibility, see EMC Electromagnetic radiation, 37, 39 wavelength, 37 velocity, 37 Electromotive force, 2 Electron, 1 free, 1, 3 orbit, 1, 2, 40 shell, 1, 2 valency, 63 Electron-volt, 40 Element, 1 pentavalent, 63 tetravalent, 63 trivalent, 63 ELV, 245, 368, 388 EMC, 36, 145, 219, 220, 233, 239, 248, 261, 266, 272−276, 351, 445, 481, 484, 506 disturbance phenomena, 273, 274 emission limits, 274 standards, 273 Emergency lighting, 389, 391–395, 415, 460, 481, 497, 538 testing, 389, 393 power, 392 EMF, 2 back, 13 EMI suppressor, 219 Emitter, 69 follower, 74 Encryption, 339, 341 Endura lamp, 146 Energy, 6, 7 management, 345−347, 402−419, 432, 433, 460, 471 state, 41 EPROM, 107, 358, 465 EEPROM, 107, 326 ERP, 338
556
ESTA E1.3, 280 Ethernet, 322, 328, 331, 333, 341, 430, 458, 528, 538, 541 backbone, 323 fast, 323 frame format, 322 gigabit, 323 hub, 323 switched hub, 323 wireless, 342, 430 Eye, 45, 250 Fade time, 375 Fan-out, 96 Farad, 7 FCC, 272 FDDI, 328 Feedback, negative, 73 FELV, 245, 278, 281 Ferrite,174, 181 Ferromagnetism, 174 FET, 76, 491, 492 FHSS, 342 Fiber optics, 166−170, 450 attenuation, 167 communication, 323, 484, 490 configuration, 167 end optics, 169 harness, 167 illumination, 138, 168, 245 leaky, 169 packing fraction loss, 168 principle 166 transmission, 167 Field bus, 314 Field effect, 76 Filament, carbon, 113 coiled coil, 114 sing, 115 tungsten, 113 Filter, optical dichroic, 44, 117 high pass, 44 low pass, 44 interference, 44 Filter, color, 50 CTC, 371 dichroic, 371 fluorescent lamp, 125 glass, 370 lacquer, 370 plastic sheet, 371 Filter, electrical active, 95 dimmer, 239, 266 high pass, 62
input for electronic ballast, 220 low pass, 62 passive, 221, 223 powerline carrier, 422, 423 Firewall, 327 Firewire™, 103, 287 Firing controller (pyro) 532–534 Firmware, 106, 107 Flash, 362 Flash tube, 164 airport runway, 166, 486, 490 photographic, 165 stroboscopic, 165, 166 Flicker, 91, 232 index, 230, 232 percentage, 230, 232 Flip-flop, D, 99, 101 JK, 99 SR, 98 Floppy disk, 104 Fluoractor, 207 Fluorescence, 112, 119 Focus, 368 preset, 369 FOH, 510, 511, 519 Footcandle, 48 Foot Lambert, 48 Force, 6 Form factor electronic ballast, 215, 235 sine wave, 21 Fountains, 528–530 Fovea, 45 FPGA, 110 Frequency, 6 resonant, 62 sampling, 88 FSK, 318, 501 FTP, 325 Fuse, 30, 267, 512 anti-surge, 187 arcing time, 30 clearing time, 30 discrimination, 31 HRC, 30 one-time, 247 pre-arcing time, 30 thermal, 187 Fuseholder, industrial, 30 panel mounted, 30, 268 Gain DC current, 70 closed loop, 73 common mode, 75
INDEX
open loop, 73 small signal, 70 static current, 70 Galvanic isolation, 477, 478 Galvanic series, 477 Gas, filling (for lamps) 114 lighting, 112 mantle, 112 Gate, logic AND, 97, 98 array, 102 astable, 98 bistable, 98 construction, 99 EX-NOR, 97 EX-OR, 97 FET, 77 logic, 96 monostable, 98 NAND, 97, 100 NOR, 97 NOT, 97, 100 OR, 97, 98 sea of, 102 Gate, semiconductor floating, 107 IGBT, 81 JFET, 77 MOSFET, 107 thyristor, 81 Gateway, 297, 307, 336, 341, 431 Genura lamp, 144 Gerb, 533, 534 GFSK, 342 Giga-, 5 Glare, 49, 395 disability, 49 discomfort, 49, 503 index, 49 unified rating, 49 GLS (lamp) 113, 123 Gobo, 367 GPS, 480, 541, 542 Graphics controller, 104 Graphics tablet, 369 Ground, functional, 273 safety, 273 Group (control) 356, 359, 362 inhibit, 361 GSM, 503, 505 GTO, 85, 208 GUI, 388, 489, 535 Half duplex (communication) 284 Hard disk, 104
Hardware, 106 Harmonic, 37, 183, 220, 221 dimmer, 268 distortion, 222, 478, 479 power factor correction, 214 rejection transformer, 183 Healthcare, 474 Heisenberg, uncertainty principle, 40 Hekto-, 5 Henry, 7 Hertz, 6 Hexadecimal code, 88 HiperLAN2, 342 HiSwan, 342 Hole, 63 Home theater, 429 Horsepower, 5 Horticulture, 472, 473 Hospital, 459 corridor, 459 nurse station, 462 ward, 461 Hotel, 374, 459−461 banqueting suites, 379, 440 corridor, 459 guest room lighting, 349, 460 public areas, 461, 462 Houselights, 343, 511, 519, 520 HTML, 333 HTP, 355, 361, 366 HTTP, 325, 333 Hub, 323 Hue, 51 HVAC, 417 Hysteresis, 15 effect, 261 loop, 14, 175 loss, 176 IC see Integrated Circuit ICMP, 325, 329 IEC, 231, 272, 392 ballast standard 60929, 227, 281, 283, 293 connectors, 513, 514 harmonic limit 1000-3-2, 221 standard ballast, 196, 197 IEEE 1394, 103, 287 802.3, 322, 323, 327, 328, 331, 333, 341, 481 802.11, 341, 342 802.11a, 342 802.11b, 342, 481 IESNA, 348, 534 IGBT, 80, 269, 523 drive circuit, 270
IGFET, 77 IGMP, 329 Ignitor, auxiliary switch, 210 hot re-strike, 209 impulser, 208 parallel, 208 superimposed pulse, 135, 208 Illuminance, 47, 48 Illuminant standard, 53 Impedance, 26 output, 74 Incandescence, 112 Inductance, 7 mutual, 22 Induction lamp, 144−147 Inductor, 11−13 Infra-red, 44,115, 163, 476, 514 beam detector, 408 control, 297, 334−337, 378, 381, 395, 422 heat source, 164, 473 modulation, 242 Input-output, 103 Instabus™, 314 Instruction set, 109 Insulator, 6 Integrated circuit, application specific, 101 cell, 102 custom, 102 diffusion, 77 digital, 93, 96 epitaxial growth, 77 four bit adder with carry, 98 gate array, 102 hybrid, 93 ion implantation, 101 large scale, 101 linear, 93, 94 manufacture, 77 mask, 102 metal formation, 77 monolithic, 93 oxide growth, 77 quad two input AND, 98 sea of gates, 102 semi-custom, 101 very large scale, 101 via, 102 Integrated home system, 428 Interface, control, 297 load, 277, 296, 380 Interference, 339 common mode, 284
557
INDEX
Interface (continued) conducted, 36 filter, 44 light waves, 44 radiated, 36 Interoperability, 288, 310, 328, 350 Interpreter, 105 Interrupt, 105, 304 Inverter, battery, 233, 477 bridge, 217, 218, 243, 244 flyback, 233 half bridge, 217, 218, 224, 226, 244 logic, 100 push-pull, 217, 233 waveforms, 217, 225 I/O, 103 IP, 323−327, 329 addressing, 326 IR, see Infra-red IrDA, 337, 338 serial infra-red, 337 fast infra-red, 337 Iris, 45 Irradiance, 47 ISM band, 342 ISO, 303 I2t, 31 JFET, 76 n channel, 76 p channel, 76 Joule, 6, 7 Kelvin (temperature) 5 Keyboard, computer, 103 Kilo-, 5 Kilogram, 5 Kilowatt, 5 Kirchhoff’s law, 69 Laminations, reactor ballast, 199, 200 transformer, 23, 178, 179 Lamp, black light, 125 classification, 141−143 disinfectant, 126 electroluminescent, 151−153 Endura™, 146 EPROM erasure, 126 flat, 148 fluorescent, see Lamp, fluorescent general service tungsten filament, 113 Genura™, 144 GLS, 113, 123
558
halogen infra-red, 117 HID, see Lamp, HID incandescent, 113−118, 465, 466 induction, 144−147 infra-red, 115, 163, 476 insect attracting, 126 life of fluorescent, 122 life of incandescent, 114, 118 microwave, 147 MR, 118 neon, 128, 150, 190 neon glow, 149 PAR, 118 Planon™, 148 QL™, 145 sing, 115, 267 special purpose fluorescent, 124 sulfur, 147, 170 sun tan, 125 tungsten halogen, 115, 117, 486, 487 UHP, 139 ultra-violet, 125, 162 xenon arc, 140 Lamp, fluorescent, 119−129, 192−204, 416 AC characteristic, 192 circular, 123 cold cathode, 127, 128, 190, 240, 241, 400, 442, 493 compact, 123 DC characteristic, 192 electrodeless, 129, failure detection, 231 high frequency operation, 215, 218, 219, 227, 236 hot cathode, 119 instant start, 126 life, 228 long life, 129 miniature, 128 reflector, 126 U-shaped, 123 UV, 125,162 voltage/current curve, 236 Lamp, HID, 130−140, 385, 416 compact source metal halide, 138, 493 high pressure mercury, 139 high pressure sodium, 134, 205 low pressure sodium, 133 mercury free HPS, 135 mercury vapor, 130, 162, 205 metal halide, 136, 162, 205, 400 rectification effect, 205 safe area of operation, 205, 206 LAN, 322 see also Ethernet
Language, assembler, 105 high level, 105, 356 Laser, 159–162, 545 construction, 159 diode, 161 gas, 159 Helium-Neon, 160 ion, 160 optical bench, 160 pulsed, 159 ruby, 161 semiconductor, 161 Law, cube, 251 dimmer, 251 inverted square, 251 S-curve, 251 switch, 269 square, 251 LCD, 104, 140, 148, 241, 358, 509 LED, 67, 90, 153−159, 249, 465, 466, 467, 481, 481, 493, 518 blue, 154, 155, 171 dimming, 249 dominant wavelength, 155 engine, 155 green, 154, 155 life, 156 packaging, 154 power supply, 249 principle, 153 red, 154 spectral half bandwidth, 155 structure, 154 video screen, 110 viewing cone angle, 155 white, 156 Lens, 45 Lenz’s law, 12 LEP, 158 Let-through energy, 31, 83 Libraries, 453 Light, 39−49 velocity, 37 Light emitting polymer, 158 Light guide, 169 Lighting control, agriculture, 472, 473 aircraft interior, 481−485 algorithms, 413, 492 analog, 280−283, 376, 412, 482 apartment, 421 architectural, 374−395 art galleries, 452 assemblies, 446 auditorium, 343, 376
INDEX
automatic, 375 automatic cycle, 379 automobile, 490–494 aviation, ground, 487−490 closed loop, 411, 414, 415 commissioning, 415 control rooms, 508 cordless remote, 297, 334−342, 378 digital 283−299 group, 356, 361, 362 home (residential) 420−427 integrated versus separate, 418 horticulture, 472, 473 hospital, 459, 461 hotel, 459−461 key, 435, 460, 463 large scale entertainment, 373 local versus central, 417 manual, 375, 434, 453 manufacturing processes, 473 master, 353, 354 grand, 354, 360 sub, 354, 359 meeting room, 436 memory, 356−361 moving lights, 363−370 multi-channel, 353, 354 large, 381 small, 377, 422 multiplex, 280, 491, 493 multiplexed analog, 283 multi-scene preset, 254, 355, 421 museum, 449−452 open loop, 411, 415 page, 363 preset, 355, 361 PWM, 283 railway, 505 remote, 353 restaurant, 463 retail spaces, 469−471 road, 494–505 router, 382, 481 single channel, 376 scene, 295, 317, 354, 405, 435 linking, 380 looping, 379 modifier, 379, 380 setting, 374, 435, 448, 449 single, 354 two preset, 355, 511 three preset, 355 secure institutions, 462 sound-to-light, 363 stage, 353−373, 510–520 switched, 308, 383–389, 434, 451, 458, 463
telephone, 387, 390 tracking, 361 train interior, 506–508 tunnel, 494–499 workplace, 431−436 worship, places of, 447 Lighting design, 351, 374, 396, 450, 455, 463, 510 Lighting power density, 348 Light pipe, 169 Line impedance stabilizing network, 274 Linet bus, 507 Litz wire, 176 Live connection (AC) 20 Logic gate, 96 Lord Chamberlain’s switch, 519 LON® Lonworks, 293, 306−310, 332, 336, 350, 383, 386–388, 390, 418, 431 addressing, 308 communications, 308 gateway, 297, 298 Neuron chip, 306 Loss, copper, 176 core, 176 eddy current, 23, 176 flux leakage, 176 hysteresis, 176 Long tail pair, 75 Loss factor, 213 LPD, 348 LTP, 366 Lumen, 46, 47, 155 Luminaire, AGL, 486 discotheque, 545 emergency, 392, 394 fiber optic, 169 fluorescent, 44 increased safety, 127 HID, 204 LED, 155, 372 submersible, 528 surgical task, 474 television studio, 523 theatrical, 138, 139, 515 UV, 163 Luminance, 46, 47 photometer, 49, 496 Luminescence, 113 Luminous efficacy of radiant flux, 46 Luminous efficacy of source, 46, 113, 117, 123, 132, 133, 135, 141, 155, 157 Luminous flux, 46, 47
Luminous intensity, 46, 47 distribution, 118, 121 Lux, 47, 48 Lux-hour, 450 LV, 245 MAC, 323 Macro, 369 Magnetic flux, 7, 195 density, 7, 14, 174, 198 remanent, 14 distribution in ballast, 198 leakage, 172, 180, 184, 200 mutual, 172, 174 Magnetic field, 200, 275 Magnetic saturation, 14, 15, 195, 196, 223, 254 Magnetizing current, 175 Magnetizing force, 13, 195 Magnetomotive force, 13, 196 Manchester coding, 294, 310, 338 Manufacturing processes, 473, 474 Maxwell corkscrew rule, 13 MCB, 32, 259, 267, 384, 387, 378, 512, 515, 526 double pole, 268 MCCB, 32, 525 Meetings, 436 Meeting room, 439 Mega-, 5 Memocard, 356 Memory, cell, 108 dynamic, 108 flash, 108, 358, 465 magnetic core, 356 mask programmed, 107 non-volatile, 106 one time programmable, 107 programmable, 107 random access, 108 read-enabled, 107 read only, 107 static, 108 video, 108 volatile, 106 write-enabled, 107 Mesh connection (AC) 22 Message centers, 466 Meter, 5 Micro-, 5 Microcontroller, 109, 356, 386, 491 Microprocessor, 105, 280, 356, 404 Microwave lamp, 147 MIDI, 300 Milli-, 5 MIRED, 52, 55
559
INDEX
Mirror cold, 44 dynamic beam diverter, 161 hot, 44 static beam diverter, 161 MMF, 13 Modulation, amplitude, 90, 338 delta, 91 frequency, 90, 338, 489 infra-red, 242 pulse, 90 pulse amplitude, 91 pulse code, 91 pulse width, 90 Monitor, computer, 104 CRT, 104 LCD, 104 Monitoring (of lighting control) electrical load, 418, 480, 521 fault, 418, 480 lamp life, 419, 536, 538 road lighting, 503–505 status, 418, 480 MOSFET, 77, 223, 224, 235, 269, 277, 523 depletion mode, 77 double diffused, 79 enhancement mode, 77 floating gate, 107 memory cell, 107 power, 78 Motor, electric, 35 stepper, 365 Mouse, 103, 359 Moving light, 357, 363−370 attributes, 367 automated follow spot, 370 automated focusing, 370 beam shaping, 368 color selection, 369 diffusing filter, 368 focus, 368 focus preset, 369 gobos, 367 intensity control, 366 iris, 368 moving head, 365 outdoor lighting, 400, 401 palette, 369 pan, 364 prism effect, 368 programming, 369, 401 protocols, 366 scanner, 365
560
tilt, 364 zoom, 368 MSC, 300, 373 Multiplexer, 96 Multiplexing, 280 Munsell system, 51, 52 Mutual conductance, 70 Mutual inductance, 22 Nano-, 5 Neon, glow lamp, 149 sign lamps, 128, 150, 190, 465 Network, see also Ethernet node, 293, 309, 317, 323, 327 topology, 305 Neuron® chip, 306 Neutral, connection (AC) 20, 22 trap, 183 Newton, 6 Neutron, 1 Nit, 47 Noise, 86, 521 background design criteria, 444 console, 445 dimmer, 445, 521 lamp, 444 luminaire, 445 perceived criteria, 444 Non-dim, 356, 383, 514 NRE charges, 102 NRZ, 338 Nucleus, 1 Nyquist frequency, 89 Offices, 431−435 Ohm, 3, 7 Ohm’s law, 3, 7 OHP, 436, 441 OLED, 158 Op-amp, 94, 281 Opcode, 109 Optic nerve, 45 Opto-coupler, 67 Oscillator, 60, 62, 225, 226, 235, 247 OSI, 303, 311, 316, 320, 331 Page (control) 363 Palette, 369 Panic button, 442, 519, 520 PAPI, 485 PAR lamp, 118 Parallel (data) 87 Parity, 287 Partition, 379, 380 Pascal, 6
Patching, soft, 357 Patch panel, CAT-5, 516, 517 electronic, 96 power, 510, 511, 525, 526 PCM, 91 PD, 3, 7 Pepper’s ghost, 455 Permeability, 13 of free space, 14, 197 relative, 14, 175, 195, 197 Permittivity of free space, 11 relative, 11 Phase, control, 260 cutting, 260 locked loop, 95 Phon, 92 Phosphor, 41, 119, 151 halophosphate, 120 multi-band, 120 tri-phosphor, 120 standard white, 120 Phosphorescence, 120 Photoperiodism, 473 Photon, 39 energy, 40 wavelength, 40 Photometer, 48, 497 Photopic spectral luminous efficiency, 46, 48 Photosynthesis, 473 Phototherapy, 474 Pico-, 5 Piconet, 342 Ping, 326 PIR, 297, 408, 409, 420 Pixel, 104, 467 Planck’s constant, 40 Planckian locus, 51, 54, 55 Plane angle, 5 Planon lamp, 148 PLC (programmable logic controller) 109, 403, 458, 465, 507, 517, 539 PLL, 95 Plot, blind, 355 lighting, 355 PN junction, 64 PNC, 444 Polling, 304 Port, Centronics, 285 interference, 273 parallel, 103, 109, 285 serial, 103, 109, 286, 333, 455
INDEX
printer, 285 server, 333 Potential difference, 3, 7 barrier, 64 Potentiometer, 8 Power, 2, 6, 7 aircraft, 481 automobile, 490 control, 35, 83 conversion, 215 dissipation, 66 emergency, 392 failure, 383 quality, 417, 478 ratio in dB, 92 ship, 478 television studio, 520 train, 506 Power factor, 27, 367 bulk/central correction, 214 correction, 28, 121, 124, 172, 194, 195, 211−213, 220−223 active, 223 displacement, 211 lagging, 28 leading, 28, 194 Powerline carrier, 317−322, 395, 422, 489, 503 Power reducer, 415, 416, 436 PPFD, 473 Preset, 355, 356 Pressure, 6 Primary cell, 15 Primary coil (winding), 172, 178 Program, applications, 106 computer, 105 counter, 105 Programming, architectural controls, 377, 378, 382, 383, 387, 480 moving lights, 369 partition, 379 Projector, projection 3D, 540 film, 140, 443, 466, 531 overhead, 436, 441 scene, 456, 532 slide, 436, 441, 455, 456 video/data, 138, 139, 141, 437, 456, 531 water screen, 529, 530 PROM, 107 Protection, dimmer, 259, 261, 267 discrimination, 31 electronic transformer, 247
life safety, 30 over-current, 247 overload, 30 short circuit, 29 sub-circuit, 268 thermal, 187, 247 transformer, 187 transient, 247 Protocol, 288−303, 322−332 connectionless, 324 Proton, 1 PTC, 226 Pulsadis, 500 Push-pull, 73 PWM, 90, 110, 129, 157, 236, 242, 244, 249, 276, 283, 336, 467, 477, 482, 492, 494, 508 Pyroluminescence, 112 Pyrotechnics, 532–534 firing control, 533 igniters, 532 indoor, 533 Q file, 356 QL lamp, 145 Quadrac, 261 Quadrant, triac, 84 Quantization, 88 Quantum light, 39 mechanical tunnelling, 108 theory, 39 Quartz, 116 Quasi peak detector, 275 Rack, 19 inch, 259, 275 Radian, 5 Radiance, 47 Radiant flux, 46, 47 Radiant intensity, 47 Radiation, IR, 163 Railway, 505–508 platforms, 505 train interiors, 506 RAM, 103, 104, 106, 108, 358 Ramp generator, 264 RARP, 325 Ratio, common mode rejection, 75 hot/cold cathode resistance, 227 mark-to-space, 74 RC5, 334, 338 RCB, 32 RCBO, 32, 268 RCD, 32, 230, 268, 510, 526 RDM, 292, 336
RDS(ON), 79, 269 Reactance, 24−26 capacative, 25, 172 de-tuning, 214 inductive, 25, 172 leakage, 184, 190 Reciprocal megakelvin, 52 Rectification effect, 205, 206, 230, 234 Rectifier, 59, 215, 220 bridge, 59, 215, 256 copper oxide, 63 full wave, 59 half wave, 59 mercury arc, 63, 205 selenium, 63 Reflectance, specular, 42 Reflection, diffuse, 43 laws of, 43 specular,42 Reflector, cool, 117 dichroic, 117 hot, 117 Refraction, 43 critical angle, 43 Refractive index, 43 Register, 99 instruction, 105 Relay, 33, 393 bistable, 34 coils, 34 contacts, 34 contact suppression, 34 latching, 384 reed, 34 transformer, 385 Reluctance, 196 Remanent flux density, 14 Repeater, 307 Resistance, 3, 7 cathode (hot) 227 cathode (cold) 227 internal, 4 terminating, 291 thermal, 66 Resistor, 6 parallel, 8 PTC, 226 series, 8 surface mounted, 8 thin film, 8 variable, 8 wire wound, 8 Resolution, 87 Resonance, 62, 214, 232 acoustic, 236, 244
561
INDEX
Resonant circuit, 61, 224, 244 Restaurant, 379, 463−465 Retail, 379, 469−472 entertainment, 543, 544 Retina, 45 RF, 334, 338−342, 423, 466, 501, 505, 511, 533 antennae, 340 ERP, 338 modulation, 338 permitted frequencies, 338 radiated, 219 range, 338 security, 339 wireless network, 341, 342 RFI, 339 conducted, 272 emission limits,272, 274 immunity limits, 272, 274 radiated, 272 suppression, 38 RGB, 51, 54 Rigger’s control, 518 Ripple control, 258, 499–502 Ripple current, 59, 223 RISC, 109 Risetime, 266, 521 Road (lighting control) 416, 494–505 Rock and roll, 363 Rockboard, 362 Rods, 45 ROM, 104 Room control, 436−439 Room controller, 430, 438, 517 Root mean square, 20 Router, configured, 307 learning, 307 RS232, 286, 287 RS422, 287 RS485, 287 SAD therapy, 474, 475 SAE, 493 J1850, 494 Sampling, 88 frequency, 88 Sawtooth wave, 38 Scalar quantity, 17 Schmitt trigger circuit, 95 Schools, 435 SCR, 81 S-DRAM, 106 Seasonal Affective Disorder, 474 Second, 5 Secondary cell, 15 Secondary coil (winding), 172, 178
562
Secure institutions, 462 SELV, 246, 278, 281, 294 SEMAGYR, 500 Semi-conductor, 6 Sensor, 347, 405−412 beam detector, 408 light, 402, 407, 434 mechanical, 405, 410 microwave, 410 occupancy, 402, 414, 415, 460 outputs, 412 photo-conductive, 407 photo-electric, 407, 435, 460, 492 multi-directional, 411 reflex, 408 photo-voltaic, 407 PIR, 297, 408, 409, 412, 451 position, 406 positioning of, 410 presence, 402, 407, 434 pyrometric, 408 rain, 492 ultrasonic, 410, 451 Serial (data) 87 Shift register, 99 Ships, 478−481 Shunt, magnetic, 184 Show control, 455 Show controller, 455 Show control program, 465 SI, 5 SIDAC, 209 Signs, illuminated, 464−468 animated, 465 static, 464 Silicon, 63 controlled rectifier, 81 diode characteristic, 65 n type, 64 photodiode, 407 p type, 64 Simplex (communication) 284 Simulation, 475 sunlight, 163, 475 Sine, 16 Sine wave, 17 dimmer, 276 half wave average value, 21 harmonic, 37 peak value, 21 r.m.s. value, 20, 21 Skin effect, 176, 247 SMPTE, 301 SMTP, 325 Snell’s laws, 43 Snubber, 34, 224 Software, 106
Solar cell, 68 Solar dial, 405 Solar simulation, 475 Solid angle, 5 Son et Lumière, 531–533 Sound, intensity, 91 Sound pressure level, 91 Sound-to-light, 363 Source (FET), 77 Source code, 105 Spectaculars, 526–528, 540 Spectral emissivity, 41 Spectral line, 41 Spectral power distribution, 116, 121, 124, 131, 133, 134, 140, 157, 475, 476 Spectroscope, 41 Spectrum absorption, 41 electromagnetic, 39 emission, 41 locus, 54 Sports lighting, 534–539 Square wave, 38, 243 SRAM, 108 Stadium lighting, 536–538 Stage lighting control, 353–373, 510– 520 large stages, 514–520 small stages, 510–514 Stage management, 514 Starter (lamp) electronic, 122, 207 fluorescent lamp, 121, 123, 207 glow, 122, 207 thermal, 122 Starting (lamp) auxiliary electrode, 132, 136 cold, 127 fluorescent lamp, 121, 192, 207 electronic ballast, 225−228 gas, 134 hot-restrike, 138, 139 induction lamp, 145 instant, 127, 194, 225, 228 low pressure sodium lamp, 134 mercury vapor lamp, 130 metal halide lamp, 136 pre-heat, 127, 193, 226 rapid, 193, 226 sequence, 195 UHP lamp, 139 xenon arc lamp, 140 Stefan-Boltzmann law, 52 Steradian, 5 Street lighting control see Road lighting control
INDEX
Strobe pulse, 285 Stroboscope, 165, 166 Stroboscopic effect, 215, 436 STP, 305 Switch, 29 electronic, 58 light sensitive, 407 limit, 406 micro-, 405 reed, 406, 411 time, 403−405 transistor, 70 Tangent, 16 δ , 213 TCP/IP, 323−330, 382 addressing, 326 protocol stack, 324 Television lighting, 443, 446, 449, 520– 524 Temperature, case, 229 coefficient of resistance, 9 lamp starting, 204, 229 winding, 179, 204 Tera-, 5 Tesla, 7 THD, 222 Theme parks, 539–542 parades, 540, 541 parkwide control, 542 rides, 539, 540 Thermal drift, 65 Thermal fuse, 187 Thermal inertia, 35 Thermal resistance, 66 Thermionic emission, 127, 231 Thyratron, 255, 257 Thyristor, 81, 257 characteristic, 83 firing circuit, 262 gate turn-off, 85 insulated gate, 85 latching, 82 photo, 85, 264 Timecode, EBU, 301 drop frame, 302 film, 302 linear, 301 longtitudinal, 301 MIDI, 303, 399 SMPTE, 301, 373, 399, 455, 530, 532, 533 vertical interval, 301 Timer, see also Timeswitch watchdog, 109
Timeswitch, 403−405, 458, 460, 461 calendar adjust, 405 delay, 403 electronic, 404 latitude adjust, 405 mechanical, 404 motor driven, 403, 404 programmable, 403 season adjust, 405 Token passing, 304 Touch screen, 438, 441, 470 Touring shows, 524–526 Trackball, 359 Transducer, 59, 86, 406 Transformer, 22−24 action, 172 AGL, 487 auto, 24, 184 motorized variable, 186, 189, variable, 185, 503 cold cathode lighting, 190 constant wattage auto, 206 construction, 178 core materials, 174, 175 delta-delta, 182 delta-star, 182, 183 delta-wye, 182, 183 delta-zig-zag, 182 distribution, 181 efficiency, 177 electronic, 245−248, 451 equivalent circuit, 177 ferrite cored, 181 harmonic rejection, 182 impregnation, 179 leakage reactance, 184, 190, 195 laminations, 23, 178, 179 lighting, 186−192 losses, 176, 187 perfect, 172 power factor, 178 protection, 187, 191, 247 regulation, 177, 187, 206 relay, 385 series, 192, 487 star-delta, 182 star-star, 182 temperature rise, 179 three phase, 181 toroidal, 180 construction, 180, 188 lighting, 188 mounting, 181 wire, 178 class A, 179 class B, 179, 189 Transistor, 68−81, 215
bipolar, 68−72, 491 field effect, 76, 491 insulated gate bipolar, 80 n channel, 76,100, 492 NPN, 68 p channel, 76, 100, 492 photo, 81 PNP, 68 safe operating area, 71 saturation, 72 turn-off time, 75 turn-on time, 75 unipolar, 76−81 Triac, 84, Trigonometry, 16 TTL logic gate construction, 99 Tungsten filament, 113 radiation, 42 Tunnel (lighting control), 411, 494–499 UART, 285 UDP, 324, 325 UGR, 49 UHP lamp, 139 UL, 272 UPS, 358, 392, 395 USB, 103, 287 USITT, 288 UTP, 305, 484 UV, 41, 116, 123, 137, 139, 162, 449, 458, 476, 474 -A, 125, 132 -B, 125, 132, 474 -C, 125 Variable, 105 VCA, 86 VCP, 49 VDR, 247 Vector quantity, 17 Video conferencing, 437, 438 Video displays, 436, 467 as lightsources, 171 Vision persistence of, 91 photopic, 45 scotopic, 45 Visitor centers, 454−459 VITC, 301 Volt, 2, 7 Voltage dependent resistor, 34 Voltage, adder, 94 average, 21, 84 breakdown, 65 controlled amplifier, 86
563
INDEX
Voltage (continued) controlled oscillator, 96 constant, 120 dependent resistor, 247 follower, 94 forward blocking, 81 forward drop, 65 input offset, 94 peak, 21, 84 pinch-off, 77 rate of change, 25, 83, 213, 214, 224 ratio in dB, 92 regulation, 186 reverse, 65 r.m.s., 21, 84 sensing, 265 subtractor, 94 surge suppression, 67 waveform in reactor ballast, 193 VoltAmpere, 23
564
VRAM, 108 Warehouse, 435 Watchdog, 109, 266, 383, 393 Watt, 4, 6, 7 Wavelength conversion, 40 particle, 40 transverse wave, 37 Wave motion, 37 transverse, 37 Weber, 7 per square meter, 7 WECA, 342 Wheatstone Bridge, 256 Wheel, color, 371 lighting control 359 Wien displacement law, 52 Window comparator, 263 Woods Glass, 132
Work, 6 Working lights, 517 Worship, places of, 447−449 X10, 318, 422 X capacitor, 220 Xenon, automobile headlights, 493 arc lamp, 140 ceramic reflector lamp, 141 exchangeable arc lamp module, 141 flash tube, 164, 486, 490 Planon discharge lamp, 148 XML, 329 X-Ray illuminator, 343 Y capacitor, 220 Y connection (AC) 21, 182, 183 Zener diode, 66 Zero-crossing detector, 262, 478