! 50 June 1979 U.K.55p. U.S.A Can. $1.75 up-to-dqte^electronics for lab and leisure elekdoorbell program your own signature tune or seasonal^ melody ! i wm mvZ i V w 3 A 'UjL ajjjr* 0 S WP nicads to charge or not to charge? i.f. strip and stereo decoder for a top-notch vhf-fm receiver. Australia S 1.50- Germany DM. 4.20 Austria S 33 Netherlands OFL. 3.50 Belgium F, 59 New Zealand $ 1 50 Danmark Kr. 10 Norway Kr. 10 Franca F. 8 Sweden Kr. 14 • r»comm«nd*d Switzerland F 4 40 Atf*.- - St ) 1 .yVS Measure Resistance to 0.01 A . . . At a Price that has no resistance at all l\| ew ^gSSSiSEBBl Digital Multimeter M1200B USA ONLY £55 (+£3 p&p + VAT£4.64 = £62.64) * FULLY GUARANTEED FOR 2 YEARS * METAL CASE ELENCo «"«a7£ '5.00 200 ■ = -'Z s*** i 0 *EX STOCK DELIVERY ( Subject to availability ) P.Q Q JhcnoT THE ULTIMATE IN PERFORMANCE - MEASURES RESISTANCE TO 0.01 OHMS, VOLTAGE TO 100 MICROVOLTS, CURRENT TO 1 MICROAMPS AT LOWEST EVER PRICE! FEATURES • 3’/ 2 digits 0.56" high LED for easy reading • 100/i V, 1 /j A, 0.0112 resolution • High input impedance 10 Megohm • High accuracy achieved with precision resistors, not unstable trimpots • Input overload protected to 1000V (except 200mV scale to 600V) • Auto zeroing, autopolarity • Mains (with adaptors not supplieo) or battery operation-built-in charging circuitry for NiCads • Overrange indication • Hi Low power ohms, Lo for resistors in circuit, Hi for diodes At £55, M1200B is the best buy among DMM's currently available. Its 0.01 ohms resolution allows you to detect shorted wind- ings in coils, transformers or motors. It is also useful in checking low contact resistance in switches, relays or connectors. Poor solder connections can also be spotted. The low power ohms function permits accurate measurements of in circuit resistance without forward biasing semiconductor junctions. DC Volts SPECIFICATIONS Range 200mV. 2V, 20V, 200V, 1000V AC Volts Accuracy 1% + 1 digit. Resolution ImV Overload protection 1,000 volts max Range 200mV. 2 V, 20V, 200V, 1000V (R«pon*a 45Hz to 5KHz) DC Current Accuracy 1.5% ± 2 digits, Resolution ImV Overload protection 1000V max, 200m V scale 600V Range 2mA, 20mA, 200mA, 2amp. AC Current Accuracy 1% ± 1 digit, Resolution 1 Microamp Overload protection - 2 amp fuse and diodes Range 2mA, 20mA, 200mA, 2 amp Resistance Accuracy 1.5% ± 2 digits. Resolution 1 Microamp Overload protection - 2 amp fuse and diodes Range 20. 200, 2K, 200K, 2 Meg. 20 Meg. Environmental Accuracy 1% ± 1 digit. Resolution .01 ohms Temp coefficient 0‘to 30“ C ± .025% °C General Operating Temp 0 C to 50° C Storage - 20° to 60 u C Mains adaptor: 6 9 Volts @ 200mA (not supplied) 4C size batteries (not supplied) Size 8% x 5% x 2% Weight Th lbs. You have been waiting a long time for a digital multimeter with all these features at a price like this. Now its yours. Also available from retail shops: Audio Electronics, 301 Edgware Rd, London W2 Z & I Aero Servic s, 85 Tottenham Court Road London W.1 * AGENTS W _0 ELENC0 PRECISION Sole UK Distributor Maclin-Zand Electronics Ltd 38 Mount Pleasant, London WC1XOAP Tel. 01-837 1165 Telex. 8953684 MACLIN G © N Zand To: Maclin-Zand Electronics Ltd elektor 6 1st Floor, Unit 10, East Block 38 Mount Pleasant, London WC1X OAP Please send me DMM M1200B @ £62.64 inc. p & p + VAT (overseas £60). I enclose cheque/P. O. /Bank Draft for £ Name _ Address (BLOCK LETTERS PLEASE) ^ decoder elektor june 1979 — UK 3 elektor 50 decoder Volume 5 Number 6 Elektor Publishers Ltd., Elektor House, 10 Longport, Canterbury CT1 1PE, Kent, U.K. Tel.: Canterbury (0227) 54430. Telex: 965504. Office hours: 8.30 - 1Z45 and 13.30 - 16.45. Bank: 1. Midland Bank Ltd., Canterbury, A/C no. 11014587 Sorting code 40-16-11, Giro no. 315.42.54 2. U.S.A. only: Bank of America, c/o World Way Postal Center, P.O. Box 80689. Los Angeles, CA 90080, A/C no. 12350-04207. 3. Canada only: The Royal Bank of Canada, c/o Lockbox 1969, Postal Station A, Toronto, Ontario, M5W 1W9. A/C no. 160-269-7. Please make all cheques payable to Elektor Publishers Ltd. at the above address. Elektor is published monthly. Number 51/52 (July/August) is a double issue. SUBSCRIPTIONS: Mrs. S. Barber Subscription 1979, January to December incl.: U.K. U.S.A./Can. other countries surface mail airmail surface mail airmail £8.50 $21.00 $31.00 £8.50 £14.00 Subscriptions normally run to December incl. Subscriptions from July/August issue: U.S.A./Can. other countries U K - surface mail airmail surface mail airmail £4.25 $11.00 $15.00 £4.25 £7.00 Back issues are available at original cover price. Change of address: Please allow at least six weeks for change of address. Include your old address, enclosing, if possible, an address label from a recent issue. ADVERTISING MANAGER: N.M. Willis National advertising rates for the English-language edition of Elektor and international rates for advertising in the Dutch, French and German issues are available on request. ^ EDITOR | AL STAFF EDITOR W. van der Horst J. Barendrecht G.H.K. Dam P. Holmes E. Krempelsauer G. Nachbar T. Day I. Meiklejohn P. Williams TECHNICAL EDITORIAL STAFF A. Nachtmann J. Oudelaar A.C. Pauptit K. S.M. Walraven P. de Winter Technical telephone query service, Mondays only, 13.30- 16.45. For written queries, letters should be addressed to dept. TQ. Please enclose a stamped, addressed anvelope or a self-addressed envelope plus an I RC. ART EDITOR: F. v. Rooij Letters should be addressed to the department concerned: TQ = Technical Queries ADV = Advertisements ED = Editorial (articles sub- ADM = Administration mitted for publications etc.) EPS = Elektor printed circuit SUB — Subscriptions board service The circuits published are for domestic use only. The submission of designs or articles to Elektor implies permission to the publishers to alter and translate the text and design, and to use the contents in other Elektor publications and activities. The publishers cannot guarantee to return any material submitted to them. All drawings, photographs, printed circuit boards and articles published in Elektor are copyright and may not be reproduced or imitated in whole or part without prior written permission of the publishers. Patent protection may exist in respect of circuits, devices, components etc. described in this magazine. The publishers do not accept responsibility for failing to identify such patent or other protection. Dutch edition: Elektuur B.V., Postbus 75, 6190 AB Beek (L), the Netherlands. German edition: Elektor Verlag GmbH, 5133 Gangelt, W-Germany French edition: Elektor Sari, Le Doulieu, 59940 Estaires, France. Distribution in U.K.: Seymour Press Ltd., 334 Brixton Road, London SW9 7AG. Distribution in CANADA: Fordon and Gotch (Can.) Ltd., 55 York Street, Toronto, Ontario M5J 1S4. Copyright © 1979 Elektor publishers Ltd. — Canterbury. Printed in the UK. MEHBIN 0# 1H| buqit •umau of circulations What is a TUN? What is 10 n? What is the EPS service? What is the TQ service? What is a missing link? Semiconductor types Very often, a large number of equivalent semiconductors exist with different type numbers. For this reason, 'abbreviated' type numbers are used in Elektor wherever possible: • '741 ' stand for pA741 , LM741, MC641, MIC741, RM741 , SN72741 , etc. • 'TUP' or 'TUN' (Transistor, Universal, PNPor NPN respect- ively) stand for any low fre- quency silicon transistor that meets the following specifi- " cations: UCEO. max 20V • C, max 100 mA hfe, min 100 Ptot, max 100 mW fT, min 100 MHz Some 'TUN's are: BC107, BC108 and BC109 families; 2N3856A. 2N3859, 2N3860, 2N3904, 2N3947, 2N4124. Some 'TUP's are: BC177 and BC178 families; BC1 79 family with the possible exeption of BC1 59 and BC179; 2N2412, 2N3251 , 2N3906, 2N4126, 2N4291. e 'DUS' or 'DUG' vDiode Univer- sal, Silicon or Germanium respectively) stands for any diode that meets the following specifications: DUS l&UG UR, max IF, max lR, max Ptot, max CD, max 25V 100mA IpA 250mW 5pF 20V 35mA 100 pA 250mW IQpF Some DUS's are: BA127, BA217 BA218, BA221 , BA222, BA317, BA318, BAX13, BAY61 , 1N914. 1N4148. Some 'DUG's are: OA85. OA91 , OA95, AA116. e BC107B', BC237B', BC547B' all refer to the same 'family' of almost identical better-quality silicon transistors. In general, any other member of the same family can be used instead. BC107 (-8, -9) families: BC107 (-8, -9). BC147 (-8, -9) BC207 (-8, 9), BC237 (-8, -9) BC31 7 (-8. -9), BC347 (-8, -9) BC547 (-8, -9), BC171 (-2, -3). BC182 (-3, -4), BC382 (-3, -4), BC437 (-8, -9). BC414 BC177 (-8, -9) families: BC177 (-8, -9), BC1 57 (-8,-9). BC204 (-5. -6), BC307 (-8, -9). BC320 (-1, -2), BC350 (-1,-2), BC557 (-8, -9), BC251 (-2. -31, BC212 (-3, A), BC512 (-3. -4). BC261 (-2, -3), BC416. Resistor and capacitor values When giving component values, decimal points and large numbers of zeros are avoided wherever possible. The decimal point is usually replaced by one of the following abbreviations: P (pico-) = 10 n (nano-) = 10' M (micro-) = 10" m (milli-) « 10' k (kilo-) = 10 J M (mega-) = 10‘ G (giga-) - 10” A few examples: Resistance value 2k7: 2700 n. Resistance value 470: 470 Q. Capacitance value 4p7: 4.7 pF, or 0.000 000 000 004 7 F . . Capacitance value lOn: this is the international way of writing 10,000 pF or .01 pF, since 1 n is 10'” farads or 1000 pF. Resistors are % Watt 5% carbon types, unless otherwise specified. The DC working voltage of capacitors (other than electro- lytics) is normally assumed to be at least 60 V. As a rule of thumb, a safe value is usually approxi- mately twice the DC supply voltage. Test voltages The DC test voltages shown are measured with a 20 kO/V instru- ment, unless otherwise specified. U, not V The international letter symbol 'U' for voltage is often used instead of the ambiguous 'V'. 'V' is normally reserved for Volts'. For instance: U b = 10 V, not V b = 10 V. Mains voltages No mains (power line) voltages are listed in Elektor circuits. It is assumed that our readers know what voltage is standard in their part of the world! Readers in countries that use 60 Hz should note that Elektor circuits are designed for 50 Hz operation. This will not normally be a problem; however, in cases where the mains frequency is used for synchronisation some modifi- cation may be required. Technical services to readers • EPS service. Many Elektor articles include a lay-out for a printed circuit board. Some - but not all — of these boards are avail- able ready-etched and predrilled. The 'EPS print service list' in the current issue always gives a com- plete list of available boards. • Technical queries. Members of the technical staff are available to answer technical queries (relating to articles published in Elektor) by telephone on Mondays from 13.30 to 16.45 Letters with technical queries should be addressed to: Dept. TQ. Please enclose a stamped, self addressed envelope; readers outside U.K. please enclose an IRC instead of stamps. • Missing link. Any important modifications to, additions to, improvements on or corrections in Elektor circuits are generally listed under the heading 'Missing Link' at the earliest opportunity. UK 4 — elektor june 1979 For all their undoubted advantages, batteries have one major disadvantage: they will eventually run out and almost always at the most inconvenient time. There is a remedy to this problem: rechargeable nicads. p. 6-04 contents contents Basically, an FM stereo receiver consists of three sections: front-end, IF strip and stereo decoder. Several ready-built front-ends are available, and these can be used in conjunction with the MF IF strip and stereo decoder described in this issue. p. 6-16 and p. 6-22 The monoselektor is the basis of a remote control system with up to 1 5 separate channels. It was originally intended as a simple-to-operate remote control unit for the handicapped, but it should also prove useful in many other applications. p. 6-32 The elekdoorbell is the subject of this month's cover: a programmable doorbell that can be built for around a fiver. selektor 6-01 nicads 6-04 This article takes a look at the type of applications for which nicads are particularly suitable — i.e. when they will prove more economical than primary cells — and explains how their effective life can be prolonged by careful use. elekdoorbell 6-12 Gone are the days of the simple 'ding-dong' doorbell ('Avon calling . . .'). Many modern householders seem to want door- bells that play anything from the 1812 Overture to a slightly abridged version of Handel's Messiah. The 'Elekdoorbell' is capable of playing almost any desired tune up to 1 28 notes in length. stereo decoder 6-16 This article describes a practical circuit for a stereo decoder incorporating the TCA4500A, for which (as ever!) a printed circuit board is also available to facilitate construction. The circuit is also compatible with the IF strip published in this issue, so that, with the addition of a suitable quality front- end, one has all the ingredients for a high performance FM tuner. VHF stereo test transmissions 6-20 variable logic gate ij.c. Knapp) 6-21 FM IF strip 6-22 For some years now the 3089 limiter/demodulator 1C has been an industry standard for use in the IF stages of FM receivers. Recently, however, an improved version of the chip, the CA3189E has appeared, and the new 1C boasts superior specifications and several additional facilities. This article takes a look at the new device and presents a circuit design and printed circuit board for a high-quality IF strip. goodbye E300/E310, hello J300/J310 6-27 aquarium thermostat (w. v. Dreumei) 6-28 Constant water temperature is vital for tropical fish. The electronic thermostat described in this article will maintain the temperature within 1°F (’/4°C). It is fail-safe, in the sense that it will give loud and clear warning if anything goes wrong. monoselektor 6-32 Remote control of up to a dozen separate systems can lead to some confusion in the hardware. The main feature of the project described in this article is that only one operating switch is used to control up to 15 outputs with either 'digital' or 'analogue' capabilities (together with a visual indication of the condition of each channel). right-up and left-down (w.v. Rooyen) 6-40 missing link UK-13 market UK-13 advertiser's index UK-22 Supplement: BASIC (part 4 ), an introduction to a simple computer language. advertisement elektor June 1979-UK7 DIGIBOOK Logically laid out to accept both 0.3" and 0.6" pitch DIL packages as well as Capacitors, Resistors, LED's, Transistors and components with leads up to ,85mm dia. 500 individual connections in the central breadboarding area, spaced to accept all sizes of DIL package without running out of connection points, plus 4 Integral Power Bus Strips around all edges for minimum inter -connection lengths. All connection rows and columns are now numbered or lettered enabling exact location indexing. Double-sided nickel silver contacts for long life ( 1 0K insertions) and low contact resistance ( £6.80 D or 1 EuroBread Board <® £5 80 D or 2 EuroBreadBoards £1 1 .00 □ (All prices include VAT and P P .but add 15% for overseas orders). Company. Address. . Please make cheque/P .0 s payable to David George Sales Elektor book service The following books are available direct from the publishers, Elektor Publishers Ltd. EUROSOLDERSUCKER ROBREADBOARD DIGIBOOK This brand new book from Elektor, provides a simple step- by-step introduction to the basic theory and application of digital electronics. Written in Elektor's typical style, there is no need to memorise dry, abstract formulae, instead you will find clear explanations of the fundamentals of digital circuitry, backed up by experiments designed to reinforce this newly acquired knowledge. For this reason DIGIBOOK is accompanied by an experimenter's printed circuit board which will faciltate practical circuit construction. When ordering any of these books please use the Elektor readers service order form in this issue. When ordering any of these books please use the Elektor readers service order form in this issue. Elektor Publishers Ltd., Elektor House, 10 Longport, Canterbury, Kent CT1 1PE price £4.50 (DIGIBOOK PLUS PCB) post & pack 25p extra USA & Canada S9.50 A selection of some of the most interesting and popular construction projects that were originally published in Elektor magazine issues 1 to 8. 30 projects are contained in this book, plus a 'DATA' section which includes a chart of pin connections and performance for common-anode LED displays, valuable information on MOSand TTL-ICs, opamps, transistors and our tup- tun-dug-dus code system for transistors and diodes. With over 1 00 pages, the wide variety of projects in this stimulates the professional designer to up-date his knowledge and even the beginning amateur should be able to build most of the projects. price £3.00 (post & pack 30p extra) USA & Canada S6.50 (sent by Airmail) ELEKTOR BOOK l A ) selektor elaktor June 1979 — 6-01 ELEKTD Questions of colour Dr. M. R. Pointer* We seem to have become more aware, in recent years, of the colourful world in which we live. Colour television is now as common as black-and-white was ten years ago; people almost invariably use colour film in their cameras; colour supplements and coloured advertise- ments appear more often in our national newspapers. Why are things coloured? How do we see colour? And how do we measure it? Out of the whole of the electromagnetic spectrum, stretching from gamma rays to radio waves, only a small band of radiation is capable of stimulating the eye. The limits of this visible spectrum are not well defined but we may think of it as occupying a wavelength band stretching from 380 to 780 nano- metres (1 nanometre = 10 -9 metre); the shorter wavelength represents the blue end of the spectrum and the longer wavelength the red end. Before we can see a coloured sample we must have a source of light; it is this that provides the electro-magnetic en- ergy or power to which our eye re- sponds. Most light sources can be con- sidered white, but if their spectral power distributions were measured they would show that different lamps pro- vide different amounts of energy at different wavelengths. Even the most basic of light sources, daylight, comes in many forms; generally, however, day- light has more power at the blue end of the spectrum, whereas incandescent or tungsten light tends to be strong in energy at the red end. Consider now the object that is being illuminated by this white light : if it is coloured, it usually contains a pigment or dye that absorbs light selectively from some parts of the spectrum, so a piece of fabric that has been dyed with a red dye looks red, even though it is being illuminated with white light, because it is absorbing light at the short (blue) and middle (green and yellow) wavelengths but reflecting light at the long (red) wavelength. Similarly, a piece of purple glass must absorb green light and allow the red and blue light to be transmitted to give the visual impression of purple. Research Division, Kodak Ltd, Harrow, Middlesex Of course, different white lights render coloured objects in different ways: the predominance of red energy in tungsten light should make red objects appear more red than if they were viewed in daylight, which has less red energy. To some extent the eye allows for these differences in the light sources, an ability known as chromatic adaptation. Colour rendering is a subject of continuous research, specifically by M. B. Halstead in the UK and by other workers in Holland and the USA. One area of particular interest is that of colour preference. For example, do people prefer apples to appear a certain colour regardless of the light source used to illuminate them? Obtaining this sort of information is important to the lamp manufacturer designing new lamps. Seeing Colour Light entering the eye is focused on to the retina, the nervous tissue that lines the inner wall of the eyeball. The light receptors, or sensors, known as the rods and cones, are in the layer furthest from the front of the eye, so the light has to pass through several other layers of cells to reach them. Reaction at the receptors is photo-chemical and for colour it is the cones that are important. The rods are for low intensity (scotopic) vision, while the cones operate at high (photopic) intensities. Scotopic vision is in monochrome but photopic vision is in colour. It is generally supposed that there are three types of cone receptor, some sensitive mainly to long-wavelength light, some to short-wavelength light and some with a peak sensitivity between the two. The chemical substances in the three types of cone that are responsible for the photo- chemical reactions have yet to be isolated from any one retina, and possible sensitivity curves have been found only indirectly. The idea of three types of cone receptors was first postulated by Thomas Young and supported by Helmholtz, who showed that a colour can be reproduced by only three radiations, such as a red, a green, and a blue, mixed in correct proportions. This trichormatic theory provides the basis for many colour measuring instruments and colour reproduction systems. Signals generated in the individual receptors are combined and coded in the various cells that come after the rods and cones in the visual system. They are then transmitted out of the eye, along in the optic nerve to the brain. Coding is necessary because it is estimated that there are six million cones in a human retina and only one million nerve fibres leaving the eye. Moreover, those nerve fibres also have to transmit signals from the rods, estimated to number about 100 million. Recent research has suggested that the signals are not really ’’colour” signals but a luminance signal, composed of the combined red, green and blue signals, and two colour-difference signals. This bears a remarkable resemblance to the transmission of a luminance signal and two chrominance signals that works successfully in colour television. selektor 6-02 - elektor June 1979 hijiAA ’ACiLi 3 Figure 1. Relative spectral power distributions of a typical phase of daylight (continuous line) and incandescent or tungsten light (broken line). Figure 2. Principle of the flexible-optic tintometer for measuring colour. Light from a quartz halogen lamp is taken to the sample by two paths, one direct and the other by way of a set of colour filters. Light fallin on the sample from one path is viewed side by side with that from the other, allowing a colour match to be made by selecting filters. Brightness also can be matched by means of a variable aperture. From the known properties of the filters it is possible to derive "red", "green" and "blue" values which can be converted into a set of definitive co-ordinates. Figure 3. Spectral transmittance of a green glass (continuous line) and a purple one (broken line). The green glass transmits mostly green light; the purple one absorbs green but transmits red and blue. Measuring Colour There are three main ways of measuring colour. The first obvious one is by eye. An example of a visual colorimeter is the flexible optic tintometer, designed and developed in the UK, which has been used with great success in quality control laboratories. The instrument contains a quartz halogen lamp to illuminate the sample to be measured (if necessary the light can be taken to the sample by a fibre-optic pipe). It is also used to illuminate filters, com- prising sets of red-absorbing, green- absorbing and blue-absorbing glasses of various densities, which are remarkably stable and can be manufactured to a high degree of accuracy and reproduci- bility. An optical system arranges for the light from the sample and from another beam that passes directly through the filters to be viewed side by side, so that a colour match can be obtained by adjusting the filters. Bright- ness is also matched, by means of a variable aperture. The known properties of the filters can be used to derive a ’’red” value, a ’’green” value and a ’’blue” value, and these numbers can be converted into a set of coordinates that enable the colour to be uniquely characterized. This co-ordinate system is adopted by the CIE (Commission Internationale de L’Eclairage) and is used internationally. The second important way of measuring colour is with the photo-electric colori- meter. It attempts to copy the way the human eye works by using three photo- cells, with suitable filters over them, to reproduce the spectral responses of the eye’s three colour-sensitive channels. An exact match in responses is very difficult to achieve but many modem instruments get very near to it. Modem photo-electric colorimeters, usually from the USA, are often popular because of the speed with which they work. They incorporate microprocessors that enable the results to be displayed in a number of different co-ordinate systems. The difference between the colour being measured and a standard colour stored in the memory of the microprocessor can also be displayed. Spectrophotometry The third way of measuring colour is with the spectrophotometer. This does not measure the ’’red”, ’’green” and ’’blue” values directly, but the relative amount of energy reflected from or transmitted by a sample in terms of wavelength. A typical instrument uses a white light source, usually a quartz halogen lamp, to illuminate a grating which produces a spectmm of light in the wavelength range 220 to 900 nm. A deuterium lamp is used for the ultra- violet or short-wavelength end of the spectrum because the quartz halogen lamp is not a good source of ultra-violet energy. A relatively new British spectro- photometer uses a holographic grating instead of the more traditional ruled- glass grating, thereby reducing the amount of stray light in the instrument by an order of magnitude - a significant improvement. Older instruments used one or two glass prisms to produce the spectrum but these are bulky and expensive to make and disperse red wavelengths less than blue. The spectrum produced by the grating is scanned past a slit, which is variable and is used to define the bandwidth. The beam of light is then split into two beams by, for example, a rotating mirror system. When transmission is being measured, one of the beams goes through the sample and the other straight to the detector. The energy in the two beams is then compared by detecting the light with a photomulti- plier, giving a measurement of the attenuation that has taken place in the sample. By scanning the spectrum past the slit it is possible to measure the whole spectral response curve and trace it out on a chart recorder. The latest spectrophotometers incorporate micro- processors and are able to store the data needed to control a particular scan, such as the wavelength at which the scan starts, the finishing wavelength, the scan speed and so on. These data can also be recorded in a computer coupled to the instrument and the spectral data can be sslektor elektor june 1979 — 6-03 5ELEKTD M a a a a m m a stored while they are being generated by the instrument. It is possible to control the spectrophotometer directly from the computer, too, by sending it the necessary scanning information. The CIE have been able to specify the colour matching characteristics of a ’’standard observer”, based on the results of about 20 real observers. The CIE has also standardized several light sources by tabulating their spectral power distributions. Two of these are Standard Illuminant Sa, representing incandescent illumination, and Standard Illuminant D6s representing a phase of daylight which has a correlated colour temperature of about 6500K. By integrating the spectral transmittance or reflectance data, obtained from the spectrophotometer, with a specified illuminant and each of the three spectral response curves of the standard observer in turn, ’’red”, ’’green” and ’’blue” values can be calculated. In this way it is possible to measure, or calculate, a unique specification for any colour in terms of three numbers. The accuracy varies from one instrument to another. For example, if the precision of a spectrophotometer is to match that of the human eye, the spectral transmittance, or reflectance, has to be measured to within 0.4 per cent of the correct value for a random error or to within 0.2 per cent for an error varying systematically with wavelength. Appearance of Colour Measurements by colorimetry tell us only approximately what the colour looks like in particular viewing con- ditions. The appearance of a colour is affected by many external influences such as the level and spectral compo- sition of the illumination used to view it, the luminance and chromaticity of the area surrounding the colour, famili- arity with the object and knowledge of its particular colour, and whatever colour was seen by the viewer immedi- ately beforehand (the eye retains an after-image for a short time). We have developed a technique in our laboratory to assess the appearance of a colour in the conditions in which it is viewed. Dr R. W. G. Hunt, also of the Kodak Research Laboratories, has suggested and defined a framework of terms for the observers to use, some of them new to colour science. Basic Attributes Colour sensations have three basic attri- butes, namely hue, brightness and colourfulness. The brightness of an object is often judged with respect to a white one in the same surroundings and light, a measure referred to as light- ness. Colourfulness is a new term that is the subjective measure of the chro- Figure 4. Assessment of colour represented geometrically as a cylindrical "colour solid". First, hue is scaled by assessing which two of the four primary colours (red, yellow, green and blue) are present; the four hues are imagined as points equally spaced on a circle, after which the dominant hue is estimated as a percentage (for example, an orange may be scaled as 60 per cent yellow, 40 per cent red). Lightness is a measure of the brightness of an object judged with respect to a white one in the same surroundings and light, and is represented on a scale with one end representing black, or zero, and the other end white, or 100. If the hue circle is just one slice from the cylinder, lightness is the distance of the slice from the bottom of the cylinder. Colourfulness is a subjective measure of chromatic content: judged in proportion to the brightness of a colour, it is the degree of saturation; in proportion to the average brightness of the surroundings it is called "perceived chroma". mafic content of a colour. If this is judged in proportion to the brightness of the colour, the measure is expressed as the degree of saturation; if judged in proportion to the average brightness of the surroundings the term used is perceived chroma. If a coloured cushion is seen partly lit by direct lighting and partly in shadow, its colourfulness in the two parts is different because the level of illumination is different, but the saturation is the same and the perceived chroma is, too, for although the illumi- nation level is different, the brightness of the cushion and the brightness of the surroundings are lower in the shadow area, and these factors are taken into account. Several experiments have shown that observers can scale hue, colourfulness and lightness of a colour on a purely psychological basis. Hue of a test colour can be scaled by first assessing which two of the four primary colours, red, yellow, green and blue, are present. These four hues can be imagined as points equally spaced on a hue circle; hues at opposite ends of a diameter of the circle cannot be experienced simul- taneously, so the initial decision fixes the hue in a particular quadrant of the circle. In this way the dominant hue is determined and its amount is estimated as a percentage. Following this, the secondary hue is assessed: an orange colour, for example, may be scaled 60 per cent yellow and 40 per cent red. The observer is assessing on an interval scale, that is, a scale with fixed points at both ends. Experiment has shown that observers seem to have an almost uniform idea of pure red, yellow, green and blue. Colourfulness is plotted as a ratio. For a given set of test colours, the observer is shown a ’’neutral” colour and asked to give the first test colour a number that he feels fairly represents its colourful- ness on a scale referred to the neutral. He is then asked to scale all the other colours using the same scale. The scale is open-ended, so the first test colour merely provides an anchor somewhere on it. Each observer uses numbers in a distinct way, but it does not really matter what number he gives the first colour provided he puts all the colours he sees subsequently on the same scale. Geometric Terms The difference between observers means that ordinary arithmetical statistics cannot have any meaning, so the geometric mean must be taken to find an average. In geometrical terms we can think of colourfulness as the radial distance of the colour from the centre of the hue circle, the centre representing the neutral point. Lightness is scaled on an interval scale. One end repre- sents black, or zero, and the other end represents white, or 100. Geometrically, if the hue circle is thought of as being just one slice from a cylinder, lightness is the distance of the slice from the bottom of the cylinder. Experience shows that these ideas can be easily accepted by observers and that they can scale colours easily, not only consistently on their own but with good general agreement with one another. Work is now going on to try to relate these subjective assessments to objective measures, with proper allowance for the other parameters that define a particular scene and influence the appearance of a colour in that scene. Colour science has many facets. It depends on instruments that give objective measures and on human observers, who can provide reliable, subjective measures. But those of us who have an interest in colour science may share the thoughts of Alice in the book by Lewis Carroll: ”It’s all very pretty,” said Alice, ’’but it seems rather hard to understand.” Spectrum No. 160. (478 S) 6-04 — elektor june 1979 N icads Nicads To charge or not to charge? For all their undoubted advantages, batteries have one major disadvantage, they will eventually run out and almost always at the most inconvenient time. The pain in our wallet is a familiar feeling as the time for replacing them arrives since they, like everything else today, do not get any cheaper. If you are a consistent 'sufferer', there is a remedy for this problem, namely rechargeable nicads. Although initially more expensive than conventional batteries they offer considerable savings in both money and convenience. Assuming one observes a few elementary rules when using nicads, they should have an extremely long life. The following article takes a look at the type of applications for which nicads are particularly suitable — i.e. when they will prove more economical than primary cells — and explains how their effective life can be prolonged by careful use. There are a number of factors to con- sider when faced with the choice of using conventional ’dry’ batteries or rechargeable nicads. Although nicads could cost as much as three times the price of dry batteries, the fact that they are rechargeable offers a potential saving in the long term. They are mechanically robust, less prone to electrolyte-seepage than dry cells, have an extremely low internal resistance, and a virtually constant voltage for over 90% of the discharge period. The latter two features are particularly useful. The low internal resistance means that it is possible to connect larger numbers of nicad cells in series and allows higher flash frequencies in battery-operated flash units, whilst the stable working voltage means that there is less chance of pocket calculators becoming unreliable as the batteries begin to run down. Similarly, there is less risk of cassette recorders or cine- cameras slowing down during recording or filming - a problem encountered with dry batteries which are exhausted but can ‘feign’ a normally charged state when switched on after periods of disuse. Nicads are much more honest: they either provide a current at a steady voltage or nothing at all. However nicads also have one or two disadvantages: the average discharge voltage is only 1.2 V (as opposed to 1.5 V for dry batteries), so that for example, 4 ‘D’ cells (HP 2) would provide only 4.8 V, instead of 6 V. Not every piece of equipment will operate satisfactorily with a 20% re- duction in supply voltage. This, in turn, may lead to over-frequent recharging, since only the first third of the dis- charge cycle (when the working voltage may exceed 1.2 V) is then used. Generally speaking the capacity of nicads and of conventional primary cells is roughly the same, although there are special manganese-alkali cells with greater capabilities than those of nicads. However it must be borne in mind that the capacity of a cell, and in particular zinc-carbon primary cells, is to a greater or lesser extent dependent upon the discharge rate. Nicads are particularly suited to high discharge rates and continuous use, whereas primary dry cells are intended for intermittent service at lower discharge currents. Thus a nicad will have a Nicads elektor june 1979 — 6-05 Table la I.E.C. no. R03 KR 15/51 KR 27/50 KR 35/62 6F 22 U.S.A.S.I. AAA AA C D - Berec ceils U16 HP7 HP1 1 HP2 PP3 v— a? 18RS 501 RS RS 1.8 K RS 4 K TR 7/8 cafT n ' ca d bAh 1 cells - VR 0.5 VR 2 c VR 40 — Berec nicad Berec cells RX6 RX14 RX20 RX22 Table 1b HP7 HP1 1 HP2 PP3 AA C D — Nominal voltage dry cell 1.5 V 1.5 V 1.5 V 9 V Nominal voltage nicad cell/battery 1.2 V 1.2 V 1.2 V 9.0 V capacity (Ah) dry cell/battery 0.15-2 0.4-6 1-10 0.05-5 Capacity (Ah) 1.8 0.09 standard (sinterelectrode) 0.5 2.0 + 0.11 nicad cell/battery 0.45 1.65 All capacities stated at 5 hour discharge rate, temp. = 20° C greater effective capacity in such applications as a cassette recorder being operated for long periods. The capacity of nicad batteries of identical dimensions often differs from manufacturer to manufacturer. However in contrast to dry batteries, the nominal capacity of nicads is always indicated on the case, and these figures can be compared directly with one another. Of course the single greatest advantage of nicads, namely the fact that they are rechargeable, inevitably entails a practical inconvenience for the user — the need for a charger. Choosing the right battery — shape, capacity, charge time and price Having decided to opt for the use of nicads there comes the question of which type of cell to buy. If the nicads are being used to replace conventional dry batteries, then the choice of size is already decided. Table la shows the more commonly available nicad equiv- alents for popular dry cells, while table 1 b lists their basic differences. In the case of dry cells, the effective capacity can vary by as much as a factor of 10, depending upon the type of cell and the discharge rate. The lower figure as a rule refers to standard zinc-carbon cells operating under unfavourable dis- charge conditions (5 hour continuous discharge), whilst the higher figure refers to manganese-alkali cells. The nominal capacity of nicad cells is usually expressed for a five hour dis- charge C/5. That is to say, the capacity which is obtained when fully charged cells are discharged at such a rate as to bring them to an end point of 1 .0 V in 5 hours. The figures obtained for a one hour discharge are approximately 10 to 15% lower, whilst a longer dis- charge period of 1 0 hours results in a negligible increase in capacity. The influence of the discharge current on the cell voltage and on the cell capacity (as a result of increased intern- al resistance) is slightly greater in the case of smaller cells (capacity less than 100 mAh), but is still virtually negligible when compared to conventional dry cells. A number of manufacturers produce cells of different capacity for a given size, thus for example it is possible to buy A A cells with a 0.5 and Table la. This table lists the type numbers of nicad equivalents for the most commonly- used sizes of dry cells. A number of manufac- tures produce a range of intermediate sizes. (1/2 A, 1/3 AA, RR, F, etc.l. Table 1b. Comparison of main characteristics of nicads and equivalent dry cells/batteries. The considerable variations in the stated capacity of dry cells is the result of variations in discharge rate and in the type of cell construction. The highest capacities are ob- tained with manganese-alkali cells, which in general have four times the energy content of standard dry cells. 0.6 Ah capacity. The time taken to fully recharge a standard sintered nicad cell is approxi- mately 1 4 hours, assuming the maxi- mum normal charging current is used (i.e. the cell is not being fast-charged with the aid of a charger which will automatically switch off once the fully-charged state is reached). Charging at the normal rate can be done with a very simple circuit (e.g. a half-wave rectifier plus series resistor). -f shorter charge times are desired, then there are specially designed quick-charge cells which will accept three times the normal charging current. With these cells the charge cycle is completed in 4 to 6 hours. In addition, it is possible to use automatic quick-chargers, which are capable of fast-charging the cell in 1 5 to 60 minutes. Depending upon the manufacturer, these ‘superchargers’ can be used with special quick-charge sintered nicad cells or ordinary standard sintered nicads. Generally speaking, however, successful operation is only guaranteed if both the charger and cells are from the same manufacturer. Naturally, these quick-charge units are fairly expensive. The use of fast chargers 6-06 — elektor june 1979 Nicads is discussed in more detail later in the article. Points to check before buying a nicad cell 1 ) Make sure that the equipment or circuit in question will operate satisfactorily at the lower supply voltage provided by the nicad(s). In the case of battery-operated devices this should not normally present a problem. To obtain reason- able use from the battery the equip- ment should continue to function properly with a 1 V cell voltage. 2) Check to see whether the equiv- alent nicad cell will actually fit into the battery compartment of the equipment for which it is intended. In the case of cylindrical cells there should be no difficulty, however it can happen that the 9 V battery packs will not fit, even though they conform to IEC 6F22 dimensions. Using Nicads Charge cycle Nicads are normally charged with a constant current of roughly one tenth of the cell’s nominal capacity. It should be stressed that the charging current must be fairly constant (or limited to a safe value). Due to the very low internal resistance of nicad cells, constant voltage charging is to be avoided, since there would then be the possibility of very high currents being drawn by the cell, leading to overheating or thermal runaway. The cell does not actually store the charging current for the entire duration of the charge cycle. At the start and end of the charge period energy provided by the charging current is used to reconstitute the electrodes and in the production of gas. For this reason, the amount of current supplied during the charge cycle must be greater than the nominal capacity of the cell. A charge factor of 1.4 is used, i.e. the cell must be provided with 1.4 of its nominal capacity to reach the fully charged state. Thus with a normal charging current of C5/10 we arrive at a charge time of 14 hours. If the nominal charge time is exceeded, the cell starts to overcharge, i.e. it ceases to store the charge current, but rather converts it into heat. Providing this occurs with the normal charging current and the cell temperature remains within the limits prescribed by the manufac- turer, nicad cells are capable of with- standing long term overcharging. De- pending upon the manufacturer and type of cell, the minimum temperature for normal charging is between 0 and 10°C, whilst the maximum is between 40 and 60°C. Optimum performance is obtained between 10 and 25 C. Within this range it is possible to charge the cell to 1 00% of its nominal capacity. At lower temperatures the capacity of the cell starts to fall so that it takes less time to become fully charged, and the cell is less able to withstand over- charging. At higher temperatures, the 1 t 1 a ! , Vx 1 6 © © ©^ 1 Capacity input I 79074 • 3 1 ! Charge tima in hours - charge rata 0.1 C % capacity input Nicads elektor June 1979 — 6-07 4 79074 1 charge acceptance of the cel] sinks even further; at around 55°C it falls by some 30%. The more often it is charged at high temperatures, the greater the danger of permanent damage to the cell. It is important to check the manufac- turer’s specifications regarding over- charging of cells with the normal charging current, since the maximum permissible length of time varies. For example one manufacturer quotes 20,000 hours as being acceptable, a second draws the line at 500 hours, whilst a third is more cautious and states that occasionally overcharging for up to several days should have no untoward consequences. The last remark is valid for all types of nicads, although no harm is done if the charge cycle is ended once the cell is fully charged. Charging the cell with a current lower than the normal charge current is possible, but there are a couple of drawbacks. The charge factor of the cell rises to around 2, whilst repeated charging at lower than recommended currents can cause a temporary re- duction in the effective capacity of the cell. A charge current of less than half the prescribed value should not be used for regular recharging of cells which are completely exhausted. It is possible to use smaller currents to recharge partially discharged cells, al- though again there is a loss of capacity due to the so-called ‘memory’effect. The cell as it were notices that it is not being fully charged and adapts accord- ingly. This effect is also caused by repeated charging at high temperatures. The lost capacity can normally be restored by several charge cycles with ‘full’ current, the cell being discharged fairly rapidly between each cycle. Overcharge If a cell is required to be kept fully Figure 1. Charge acceptance of a nicad cell. At the start and end of the charge cycle only a portion of the charge current is stored by the cell (sections 1 and 3). During over- charge almost all of the energy input to the cell is either consumed in the production of gas or converted into heat. Figure 2. The greater the temperature, the lower the charge acceptance of a nicad cell, hence the need for slightly higher charge rates. Care should, however be taken to avoid overcharging. Figure 3. Charge acceptance is better at high rates of charge. As the charge current de- creases, so the amount of overcharge required to regain full capacity increases. Figure 4. Comparisons of the discharge voltage characteristics of nicads and conven- tional dry cells. The discharge conditions represented in this graph (i.e. continuous discharge at a fairly high constant current) are extremely disadvantageous for normal zinc- carbon cells. With intermittent discharge (i.e. the cell is given time to recuperate between discharge periods) and a smaller discharge current, standard zinc-carbon cells will provide roughly the same capacity as nicad cells of equivalent size, whilst special high power zinc-carbon cells can achieve approx- imately twice the capacity of nicads. The capacity of manganese-alkali cells on the other hand are less dependent upon the dis- charge conditions. charged (and is regularly discharged completely) then it is possible to leave the cell on continuous charge at a current equal to or just a little below the normal value. If the cell is used in a standby mode, however, so that it discharges only rarely and then not completely, it is recommended that, once fully charged, the current be reduced to a holding value of between 20 and 50% of the normal charging current. Fast charging Fast charging, using a current above the prescribed normal charging current, is possible with most types of nicads, however once again opinions differ among manufacturers. Generally speaking, it is acceptable to fast charge fully discharged cells at rates of two to three times the normal charging current for periods of several hours — with the proviso that the temperature is greater than 20°C and not more than roughly 45° C. At tem- peratures lower than this the normal charging current should under no circumstances be exceeded. Fast charging should not be undertaken without a charger with a timer or temperature sensor to automatically terminate charge or reduce it to safe levels. Obviously the use of a timer as control element requires that the initial state of the cell is known (in practice this usually means complete discharge). To take an example, fast charging a standard nicad cell at double the normal charging current is only permissible providing the temperature is not less than 20 C and the charge cycle does not exceed 6 hours. An exception to the rule, however, are the standard nicad cells from SAFT (all the cells of the VR-series, with the exception of the VR 10), which can be charged at three times the normal current for up to 10 days. 6-08 — elektor june 1979 Nicads 1 ] C Most sintered nicad cells can be fast charged at rates even greater than those mentioned above (depending upon the type and manufacturer), however cells cannot normally be fully charged under these conditions and the same pre- cautions must be taken to prevent overcharging and overheating. For example, an AA-cell (nominal capacity 450 mAJi) which has been completely discharged, is fast charged for 45 minutes at 10 times the nominal charg- ing current (450 mA) and then topped up at the normal current (45 mA). Over- charging at the normal rate is then possible. Discharge cycle Again depending upon the manufac- turer, nicads can be discharged over a temperature range of between roughly -40 to — 20°C and +45 to +60°C. The nominal capacity of a cell is stated for a temperature of 20 C. At higher temperatures the effective capacity of the cell rises marginally, whilst at lower temperatures there is a significant reduction. At 0°C the capacity may be between 5 and 25" down on the nominal value. The internal resistance of the cell exhibits a negative temperature coef- ficient. As the temperature falls the inter- nal resistance rises considerably (e.g. by as much as 75% over a temperature range of +60°C to -20°C). The increased internal resistance naturally causes a drop in the cell voltage under low temperature discharge conditions. Because of internal resistance, the capacity of the cell is also affected by Nicads elektor june 1979 — 6-09 Dtschaqjtd at 2(T C i 5“ C Max. ptrmiBibla continuous discharge current: 14.4 A Ceil voltage Voltage and capacity variation for different discharge rates Discharged capacity in Ah the discharge rate - the greater the discharge current the smaller the ca- pacity. The relationship between ca- pacity and discharge rate (i.e. how much the capacity falls for a given increase in discharge current) is determined by the cell’s internal resistance, which in turn is largely determined by the size and physical construction of the cell. As far as construction is concerned, a basic distinction can be made between sintered and non-sintered cells. Almost all nicad cells intended to replace conventional cylindrical batteries are of the sintered-electrode type, whilst in the case of so-called ‘massplate’ button cells and 9 V battery packs (which in fact are built up of button cells) non-sintered cells are the norm. The internal resistance of sintered cells is on average something between a quarter and a tenth ’that of non- sintered types, and remains constant over virtually the entire discharge cycle. The internal resistance of non- sintered cells, however, rises to between 3 and 5 times its original value. Non- sintered types do have one slight advan- Figure 5. Variation of internal resistance for changes in dischargeable capacity. The inter- nal resistance of cylindrical cells with sintered electrodes is virtually constant for almost 90% of the discharge cycle, with the result that the discharge voltage is also constant for this period. As can be seen, however, the internal resistance of massplate button cells varies considerably with capacity. Figure 6a. Variations of internal resistance with temperature. Figure 6b. The greater the discharge current, the greater the effect of internal resistance at low temperatures. Figure 7. The effect of discharge current on cell voltage and dischargeable capacity for a size C nicad cell (SAFT VR 2 C). tage, in that they have a lower self discharge rate — more of this later. Irrespective of the construction of cell, it is a general rule that the larger the cell (i.e. the greater the capacity of the cell), the smaller the internal resistance. However different manufacturers often state different internal resistances for cells of identical dimensions and ca- pacity! Thus for size AA cells, one finds internal resistances of between 1 5 and 35 mfi (milliohms), for C cells values between 10 and 20 mf2, and for D type cells resistances around 5 to 6-10 — elektor june 1979 Nicads Cutaway vi aw of VR 4 D call 1 . Pontiv* connection* 2. Cowar 3. Central button (poartiva) 4. Safety Vent 5. Positive plate 8. Nickel -plated steel ease 9. Ne ga tive connection 79074 2 1 5 mfl. A good quality manganese- alkali cell will have an internal resistance of approximately 300 m£2 when new, rising to around 900 mfl when 20% discharged and several ohms when almost exhausted. For the majority of applications the internal resistance of nicad cells can to all intents and purposes be dis- regarded, since it is several hundred percent smaller than that of conven- tional dry cells. However in cases where extremely high discharge rates are required, such as the powering of radio controlled model planes and ships, the effect of the internal resistance on discharge current and cell capacity must be taken into account. For example, if a cell is discharged at a rate of four times its nominal capacity (e.g. at 4 A in the case of a 1 Ah cell), the effective capacity of the cell is reduced by approx. 30%. The maximum permissible continuous discharge cur- rent is determined by the amount of heat generated in the cell, and, depend- ing upon the type of cell and the manufacturer, is somewhere between 4 and 10 times the nominal capacity. With adequate ventilation or if the cell is being operated in a pulsed mode, considerably higher discharge rates are possible - up to 1 50 times the nominal capacity in the case of certain button cells. For such specialised applications, however, one should always carefully observe the manufacturer’s specifi- cations on the cell, and if in any doubt seek advice from the company’s engin- eers. For all applications the behaviour of nicads when fully discharged is worth attention. When several nicads are connected in series (i.e. whenever a supply voltage of greater than 1 .2 V is required), it can happen that one cell discharges before the rest. In such a case it is possible that the polarity of the cell is reversed and is then effec- tively charged the ‘wrong way round’ by the remaining cells. This can lead to the production of gas, a build-up of internal pressure, gas venting and the lose of electrolyte. Repeated or sub- stantial loss of electrolyte results in permanent deterioration in the cell’s capacity and cycle life. For this reason care should be taken to avoid excessive discharching. The risk of cell reversal and consequent damage is greater under heavy load conditions and also when larger numbers of cells are connec- ted in series. As a result of both heat production and polarity reversal (in series connections), cells can be permanently damaged through short circuits. If the short occurs externally, i.e. in the circuit which the cell is powering, the resulting extremely high current may cause considerable damage on sensitive components or on the tracks p.c.b., not to mention the risk of personal injury. It should be remembered that the top cover and shell of the case are connected directly to the electrodes of the cell and it is almost Figure 8. Construction of a cylindrical nickel- cadmium cell with sintered electrodes. The case is hermetically sealed, however a resealing safety vent on the top cover of the cell enables gas which has built up due to ex- cessive charge or discharge rates to be released, thereby preventing the cell from rupturing. Repeated venting will result in a loss of electrolyte and permanent degradation in the cell's performance. Figure 9. Cell voltage characteristic during cell reversal. Excessive reverse charge (where e.g. large numbers of cells are connected in series) can result in permanent damage to the cell. Figure 10. Self discharge rates of massplate button cells (top) and sintered cells (bottom) for different temperatures. As can be seen, the nonsintered cells have the better charge retention characteristics. always advisable to take the precaution of including a safety fuse in the live lead of the supply. Shelf life and charge retention Nicad cells can be stored virtually indefinitely in any state of charge at ambient temperatures between approx- imately —40 C and +50°C. Fully or partially charged cells gradually lose capacity as a result of self-discharge however; sintered cells more so than non-sintered cells. The rate of self-dis- charge varies with temperature, charge retention being improved at lower temperatures. At 40°C the cell will have completely discharged within a few weeks, whilst at room temperature a fully charged cell with still deliver 60 to 80% of its capacity after 3 months. At temperatures below 0°C capacities of 80 to 90% can be obtained even after several months. Because of the effect of different storage temperatures and times it is normally not possible for the buyer of nicad cells to know what state of charge new cells will be in. For this reason one should always fully re- charge new cells before they are put into service. Expected life As long as they are not subjected to abusive treatment, nicad cells can be charged and discharged upwards of 500 times. As the number of cycles increases, Nicads elektor june 1 979 — 6-1 1 «• Discharge* bit o Reverse capacity charge 79074 9 there is a gradual loss of capacity. It is considered that the useful life of the cell has ended when its capacity has drop- ped to around 70 to 80% of its nominal capacity. When this point is reached will depend upon the average ambient operating temperature (high tempera- tures shorten a cell’s life) and the average depth of discharge. Although incomplete discharging in- creases the cycle life of the cell, it should not significantly affect the total energy output of the cell during its useful life. In practice, however, it is often possible, with a little luck, to reach 1500 or even 2000 cycles if the cell is never discharged below 50% capacity. Thus recharging the batteries in one’s pocket calculator every week, although it may only be necessary to do so every fortnight, could mean that a single set of batteries will last not five, but seven to ten years. On the other hand, radio modellers using nicads at high rates of discharge and then fast recharging them from a car battery cannot reckon with cycle lives much over 100. In other words, the useful life of nicad cells varies considerably with the operating conditions and recharge rates. Precautions to ensure a long operating life 1) Never solder direct to the terminals of the battery (unless solder tags are provided) 2) Take care when charging cells at low IQa lOb 79074 ■ 10b temperatures. Cells which are used out of doors during the winter should first be warmed to room temperature before being recharged. Avoid charging and discharging at high temperatures, and under no circumstances exceed the prescribed temperature range. 3) At all costs avoid short circuits and over-discharging at high current rates. In general completely discharging cells should be avoided. 4) Never leave cells on continuous overcharge at greater than their nor- mal charging current (1/10 nominal capacity). 5) Recharge new cells before use. 6) When charging a number of cells, always connect them in series, never in parallel. Alternatively, charge them one at a time. 7) Avoid connecting the cells the wrong way round (reversed polarity) when charging or discharging. 8) Never open or throw nicad cells into the fire. Take care if cells are physi- cally damaged — cadmium is toxic and the electrolyte in the cell is highly corrosive. 9) Take note of the manufacturer’s recommendations. Hopefully, the above article has suc- ceeded in shedding a little light on the topic of nicad cells, and has shown that there ii'e many applications where their high rates of charge and discharge, allied to their ability to be used again and again, make them a highly attractive proposition. Literature: Ever Ready: ‘ Nickel Cadmium, Engin- eering Data’ Ever Ready (Special Batteries) Limited, Hockley , Essex, Dec. 76. General Electric: ‘Nickel-Cadmium Akkumulatoren ’, Anwendungstechni- sches Handbuch, 2. Ausgabe 1978, General Electric Plastics Ltd., Batteries Division, Eisenstr. 5, 609 Riisselsheim. SAFT: ‘Gasdichte Nickel-Cadmium Akkumulatoren’, published by F. Briick and F. Putois, S.A.F.T. Accumulators and Batteries Ltd., 605 Offenbach/ Main Kaiserleistr. 44. Sanyo: Rechargeable CADNICA Battery Engineering Handbook, Sanyo Electric Trading Co. Ltd., Osaka, Japan 1976. Varta: ‘Gasdichte Nickel-Cadmium- Akkumulatoren, Hinweise fiir die Pro- jektierung’, Publ. Nr. 5 40 07, Varta Batterie AG, Am Leineufer 51, 3000 Hannover 21. M 6-12 — elektor June 1979 elekdoorbell elekdoorbell Programmable doorbell Gone are the days of the simple 'ding-dong' doorbell ('Avon calling . . Many modern house- holders seem to want doorbells that play anything from the 1812 Overture to a slightly abridged version of Handel's Messiah. Unfortunately, most commercially available programmable doorbells, and in particular those which are microprocessor-controlled, are anything but cheap (£ 20 is not uncommon). For this reason we decided it would be nice to design a programmable doorbell which could be built for around a fiver. The result of our efforts — the 'Elekdoorbell' — is capable of playing almost any desired tune up to 128 notes in length. The circuit can be mounted on a single board, and the current consumption is extremely low, so that an existing bell transformer can be used for the power supply. The block diagram of the Elekdoorbell is shown in figure 1 . At the heart of the circuit is an 8-key keyboard, by means of which the notes of each tune are selected. The Read/Write and Store switches are used to program the notes into the system’s memory. With the Read/Write switch in the ‘Write’ pos- ition, depressing the Store switch writes whichever note has been set up on the keyboard into memory. The length of time for which the Store switch is depressed determines the length of the note. In principle, up to 128 notes can be stored, each note occupying 8 memory locations. Holding the Store switch down for longer causes the note to occupy successive multiples of 8 memory locations. The number of possible notes which can be stored is of course then reduced accordingly. Assuming that a tune has already been stored in the system’s memory and that the Read/Write switch is in the ‘Read’ position, depressing the Start button (which will normally be mounted outside the front door) causes a flip-flop to change state. This resets a 12-bit counter which serves as an address decoder. The clock generator now clocks the address decoder through the various memory locations; the contents are transferred to a serial-parallel converter. The parallel information is fed to a digital-analogue converter which provides a current proportional to the binary value of the contents of each memory location. This current in turn determines the frequency of the audio oscillator. The last output of the address decoder is used to reset the flip- flop, preventing the tune from being repeated. The output of the oscillator must be fed to a separate audio ampli- fier, preferably provided with a volume control. The flip-flop also controls an LED, which lights whenever the Start button is depressed. In this way it is possible to turn ‘off the doorbell (by turning down the volume of the audio amplifier) and still retain an indication when someone is at the door. Figure 1. Block diagram of the Elekdoorbell. Figure 2. Complete circuit diagram of the Elekdoorbell. Although the circuit may at first sight appear complicated, in fact it employs relatively few components and should cost little to build. elekdoorbell eiektor June 1979 — 6-13 6-14 — elektor june 1979 alekdoorbell ^0 t>|n^] Parts list. Resistors: R1,R5= 10k R2 - 270 R3= 15k R4 = 4k7 R6 = 18 k R7 . . . R15 = 10 R16 = 47 k R17 = 22 k PI .. ,P9= 1 M Semiconductors : IC1 = 11 ... 16 = CD 4049 IC2 = N1 . . . N4 = CD 4093 IC3 = N5,N6 = CD 4012 IC4.IC5 = CD 4042 IC6 = CD 4040 IC7 = CD 4034 IC8 = 2102 IC9 = FF1 ,FF2 = CD 4013 D1 . . . D16.D18 = DUS D17 = LED T1 = BC547B Capacitors: C1,C3 = 100 n C2 = 68 n C4= 180 n Miscellaneous: S1 1 = Pushbutton switch, push- to-make, momentary action (bellpush). SlOab = DP changeover switch SI . . . S9 = 'digitast' (Schadow) pushbutton switch. elekdoorbell elektor june 1979 — 6-15 To store a tune in memory, the Read/Write switch is set to the ‘Write’ position and the Start button depressed. The initial note of the melody is then selected from the keyboard. If no one key provides the desired pitch, combi- nations of keys can be tried (the pitch produced by each key can be varied to suit individual requirements — see the setting up procedure later in the article). Once the correct note has been found, it is programmed into memory by depress- ing the Store switch. This has the effect of starting the clock oscillator, which in turn enables the address decoder. The keyboard data representing the selected note is fed, via a parallel-to-serial converter, to the memory locations whose addresses are provided by the • address decoder. Complete circuit The complete circuit diagram of the Elekdoorbell is shown in figure 2. The clock generator is formed by the circuit round N1 . The frequency can be varied between approximately 30 and 100 Hz by means of PI. The clock pulses are fed to one input of N2, the other input of this gate being connected to the output of N3. The output state of N3 is determined by the position of the Read/Write switch, S2, and by the Store switch, S3. When S2 is in the ‘Write’ position the number of clock pulses fed out via N2 is limited to 16 via N5 and 13. When the output of IC6 (a 12-bit binary counter) reaches 16, the output of N5 goes low, taking the output of 13 high. Eight memory locations, and hence eight addresses are required for each note. This corresponds to 1 6 clock pulses, since two are required for each address. If the Store switch is depressed for a longer period (with a clock fre- quency of 30 Hz, holding the switch down for greater than 0.5 sec.) more than one clock cycle will be generated, lengthening the note by a corresponding amount. The clock signal is also fed via a divide- by-two counter in IC6 to N4, whilst the direct clock signal from N2 is fed to N6. The pulse train which is generated at the output of N6, and which is used to * clock an 8-bit bus register (IC7), is shifted in time with respect to the clock signal fed to IC6. The pulse train, which is asymmetrical, actually has half the frequency of the original clock signal (output of Nl). N1 also controls the CE input of IC8, the system memory. The memory is enabled by a logic ‘0’ appearing at the output of N 1 or N4. The outputs of IC6 which form the 10 address lines are connected to the address inputs of IC8. Data is either written into or read out of these address locations depending upon the state of the Read/Write line, which in turn is of course determined by the state of S2. Data is read out of IC8 via pin 12, and into IC8 via pin 11. Since this can only occur when the clock Figure 3. Track pattern and component layout of the printed circuit board for the Elekdoorbell (EPS 79095). Figure 4. Two examples of suitable power supply circuits. signal is low, data can only be written to (or read from) one address at a time. This explains the function of the second clock signal derived from N4/N6. IC8 can only receive or output data when the address on its address lines cannot change. Data being read out of memory is fed serially to pin 10 of IC7. Since S2 is in the ‘Read’ position, the data is con- verted into a parallel format and fed via pins 16 ... 23 to ICs 4 and 5. As was already mentioned, the clock signal fed to IC7 is half the frequency of the original clock signal from N 1 . The reason for this is that IC6 requires two clock pulses to generate each new address. Since the divided-by-two clock signal is also delayed with respect to the first ‘address’ clock pulse, there is no possi- bility of an undefined address existing when the first bit of data is read out or in. ICs 4 and 5 form a ‘scratch-pad’ memory, on which the data from IC7 can be temporarily stored. The outputs of IC4 and 5 are normally low, except of course for the output corresponding to the note stored in memory (if several keys were depressed in order to obtain a special note, then more than one output will be high). A logic ‘1’ on the output of IC4 or 5 causes a current to flow, via the corresponding diode and poten- tiometer, into the base of Tl. This transistor, together with D12 . . . D15, 14 and 15 forms a current controlled oscillator, the frequency being deter- mined by the base current of Tl . Poten- tiometers P2 . . . P9 are adjusted to pass different current levels to Tl to produce a different note for each output. A tune is programmed into memory by first switching S2 to ‘Write’, and pressing SI to reset IC6. Once the desired note has been selected, S3 is depressed to enable the clock oscillator. Pin 13 of IC7 is taken high each time the clock oscillator delivers a cycle of 1 6 clock pulses. The parallel data on the outputs of IC7 is then shifted in serial form to pin 23, transferred via FF2 to the data input of IC8, and stored in the memory location whose address is 6-16 — elektor june 1979 elekdoorbell Stereo Decoder Stereo Decoder provided by IC6. As already mentioned, to produce a longer note, S3 should be held down for a period in excess of 16 clock pulses. Construction As can be seen from figure 3, the entire circuit of the Elekdoorbell, including the CCO, can be accommodated on a relatively small board. The keyboard is formed by a number of ‘Digit ast’ key switches or similar. As always, the use of IC sockets is recommended. The Start button should be connected to the rest of the circuit via a length of twin-core cable. If desired, three-core cable can be used and the LED mounted on or next to the pushbutton so that the caller is aware that the button has activated the bell, even if he/she cannot hear it. In that case two of the conductors are used to connect the board, LED and pushbutton, whilst the third conductor provides the earth return lead. SI and S3 should be push-to-make momentary action switches, i.e. their contacts are closed only when held down. The current consumption of the circuit is sufficiently small that an existing bell transformer can be used for the power supply. Otherwise an 8 V/500 mA trans- former can be used (assuming that the audio amplifier is powered separately). Figure 4 gives two examples of suitable power supply circuits. The four decoup- ling capacitors are not always strictly necessary, but serve to prevent inter- ference from mains-switched equipment (fridges etc.) causing unnecessary trips to the door. Setting up/Programming the Elekdoorbell Once the circuit is constructed, the keyboard can be ‘tuned’ to the desired scale and a tune programmed as follows: 1 . Switch S2 to the ‘Write’ position 2. Press the Start button (once). 3. Depress S4 and gradually adjust P2 until a tone is obtained which corresponds to the lowest note in the tune/scale. 4. Repeat the above procedure for switches S5 . . . S 1 1 and poten- tiometers P3 . . . P9, until the de- sired scale of notes is obtained (e.g. the tempered tonic scale of doh, re, mi, fa, etc.). 5. Turn PI fully anticlockwise (mini- mum clock frequency). 6. Play through all the notes of the tune to be programmed, pressing S3 between each note. The longer S3 is held down, the longer the note. 7. Switch S2 to ‘Read’. 8. Press the Start button. 9. Check whether the melody is correct, and adjust the clock fre- quency to obtain the desired tempo. 10. Ensure that the supply voltage is not interrupted, otherwise the tune will have to be reprogrammed. H Since its appearance in 1972, the 1310 stereo decoder IC has been something of a standard device for FM receivers, and indeed has featured more than once in Elektor designs (see e.g. Vario- meter Tuner, Elektor 24). This reliable and well-proven IC is, however, no longer representative of the current state of technology, and more recently a number of new stereo decoder ICs have appeared on the market. One such device is the TCA 4500 A, which in addition to outperforming the 1310 on virtually every count, offers one or two novel features such as continuously adjustable channel separation. This article describes a practical circuit for a stereo decoder incorporating the TCA 4500A, for which (as ever!) a printed circuit board is also available to facilitate construction. The circuit is also compatible with the IF strip published in this issue, so that, with the addition of a suitable quality front-end, one has all the ingredients for a high performance FM tuner. The block diagram of the decoder circuit is shown in figure 1 . The multi- plex encoded (MPX) stereo signal is fed via an input buffer amplifier to the TCA 4500A, which forms the heart of the circuit. The IC provides the left and right channel audio signals, which are amplified and fed to a twin-notch filter which eliminates residual traces of the 1 9 kHz pilot tone and the 38 kHz subcarrier. Although these fre- quencies are themselves inaudible, they can give rise to intermodulation dis- tortion in the amplifier, or worse still, when used in conjunction with a tape recorder, the 38 kHz subcarrier can beat with the bias oscillator, producing distinctly unpleasant results. The pilot tone filter is followed by a low im- pedance output buffer amplifier, which ensures that the audio signals are of acceptable level. TCA 4500A The internal block diagram of the TCA 45 00 A is shown in figure 2. The multiplexed stereo signal is first ampli- fied, then fed to a 19 kHz phase-locked loop detector (1) which controls the frequency of the VCO. The detector is essentially a phase comparator, the output voltage of which is zero if the phase shift between the transmitted pilot frequency and the internally generated 19 kHz signal is 90°. In all other cases the discriminator provides an output with a DC component which is filtered out by the lowpass filter and used to control the frequency of the VCO such that the phase shift difference will tend to 90°. n _r "L. >* TCA 4500A » JBkHi -H Sv \, n JL ^^j^separation k^ 79087 7 79082-1 Stereo Decoder elektor june 1979 — 6-17 Figure 1. Block diagram of the stereo decoder circuit. Figure 2. Internal block diagram of the TCA 4500A stereo decoder 1C. Figure 3. The 19 kHz pilot tone and 38 kHz subcarrier regenerated in the stereo decoder are not, as is usually the case, symmetrical squarewaves, but 'two-step' staircase wave- forms, as shown here. The advantage of this type of waveform is that it contains a relative- ly small proportion of harmonics. The VCO frequency is in fact locked to the 12th harmonic of the pilot tone, i.e. 228 kHz. The frequency of the VCO output is divided by a factor of 6 to obtain the 38 kHz subcarrier (which was suppressed at the transmitter).This is required to demodulate the original audio information. Subsequent fre- quency division by a factor of two provides the 19 kHz PLL feedback signal to detector 1 and a second 1 9 kHz signal which is fed to the pilot tone detector (2). The latter determines the presence or absence of the pilot tone in the MPX stereo signal. The 19 kHz signal is not, as is usually the case, a symmetrical squarewave, but rather a ‘two-step’ staircase, which represents an approximation to a sine- wave (see figure 3). The advantage of such a waveform is that it contains fewer harmonics. If the 1 9 kHz pilot tone is present in the MPX signal, the output of detector 2 provides a drive voltage for the Schmitt trigger which controls the stereo switch (and stereo indicator lamp). The switch then feeds the re- generated 38 kHz subcarrier to detector 3, which demodulates the left and right channel audio signals. These are buffered before being summed with the MPX input signal. The summing ratio is determined by the position of potentio- meter P, which thus determines the channel separation of the resultant stereo signal. Finally, both the left and right channel signals are also buffered to provide a low output impedance (100 J2). The main specifications of the TCA 4500A are summarised in table 1. 6-18 — etektor june 1979 Stereo Decoder Complete circuit The complete circuit diagram of the stereo decoder is shown in figure 4. The MPX stereo signal is first amplified by Tl, and fed via capacitor C3 to the input of the IC (pin 1). The free- running frequency of the VCO is determined by R5, R6, P2 and CIO. The 228 kHz VCO signal is available externally (at pin 15) for test and alignment purposes. The manual mono/ stereo switch is connected to pin 9 of the 1C, and the stereo indicator LED is connected via R8 to pin 7. When the decoder is switched to mono operation, the oscillator is disabled. This elimina- tes the possibility of interference due to the oscillator signal if the decoder is used in a combined AM/FM receiver. The channel separation of the decoder is determined by the potential at pin 1 1 (point B). The relationship between channel separation and the voltage at this point is illustrated in figure 5. The voltage can be varied over a range between approximately 0.5 and 2 V by potentiometer P4. The advantage of continuously variable channel separation is that it then becomes possible to reach a compromise between channel separa- tion and signal-to-noise ratio of the audio signal. In the case of a weak transmitter and a noisy stereo signal one can reduce the channel separation to the point where one obtains relatively noise-free reception. The components for the lowpass filter which provides the DC control voltage for the VCO are connected between pins 13 and 14. The lowpass filter for the second detector is somewhat simpler and requires only a single external component, C6. The de-emphasis net- works for the left and right channel signals are formed by C8/R12 and C9/R11 respectively. T2 and T3 are buffer amplifiers which provide approximately 6 dB of gain. The pilot tone filter, Toko type BLR 3107N, contains two identical but separate LC networks for the rejection of the 1 9 kHz pilot tone frequency and the 38 kHz subcarrier. The response of the filter is shown in figure 6. This filter is pre-tuned at the factory and requires no further adjustment. Finally, each channel is provided with an output buffer in the form of a JFET op-amp (IC2 and IC3). Printed circuit board The component overlay and track pattern of the printed circuit board for the decoder circuit are shown in figure 7. Construction should not present any particular problems, the only point to watch being the length of the wiring to the mono/stereo switch. Since the capacitance between pin 9 of the IC and earth must not exceed 100 pF, the wiring should be kept as short as possible. Of course, SI can also be omitted, if so desired, although parti- culary in fringe reception areas it is often useful to be able to switch to mono in order to obtain less noisy reception. It is also possible to omit P4, in which case the channel separation will be permanently at maximum. Alignment 1. First of all the gain of Tl is in- creased to maximum by setting PI for zero resistance. 2. Assuming that the IF stage of the tuner has already been aligned, tune to stereo transmission and adjust P2 until the stereo indicator lamp (Dl) lights up. Generally speaking, the lamp should remain lit over a fairly wide range of setting for P2. With P2 set to the middle of this range, reduce the gain of Tl by increasing the resistance of PI slightly, and R11 O jR-io f o Q- | RIO | ho OH ho oOH> o{ png ♦ C3 o o V O BLR 3107 N o o Figure 4. Complete circuit diagram of the stereo decoder. Figure 5. Relationship between channel separation and DC voltage, Uq, at point B (pin 11 of theTCA4500A). Figure 6. Frequency response of the twin- notch pilot tone filter, BLR 3107N. Parts list Resistors: R1 = 8k2 R2,R9,R10 = 56 k R3= 270 « R4 = 2k7 R5 = 10 k R6= 100 n R7,R21 ,R22 = 1 k R8 = 680 n R11,R12 = 5k6 R13,R15 = 2k2 R14.R16 = 4k7 R17 . . . R20 = 10 k PI = preset potentiometer, 2k2 (2k5) P2 = preset potentiometer, 4k7 (5 k) P3 - preset potentiometer, 10 k P4 - linear potentiometer, 100 k Capacitors: Cl = 1 p C2 = 470 p C3 = Ip/40 V C4.C6 = 220 n C5 = 470 n C7 = 6n8 C8,C9 = 10 n CIO = 220 p Cl 1, Cl 4 = 2p2/40 V C12.C13 = 22 p/16 V C15 “ 10p/16 V (Tantalum) Semiconductors: T1. . . T3 = BC 109B, BC 549B or equ D1 = LED IC1 = TCA4500A (Motorola) IC2.IC3 = LF 356 Miscellaneous: FI = BLR 3107N (Toko) SI = SP switch Printed circuit board EPS 79082. Figure 7. Track pattern and component overlay of the printed circuit board for the stereo decoder (EPS 79082). 6-20 — elektor June 1979 Stereo Decoder once again adjust P2 until it is roughly at the centre of the range over which D1 remains lit. 3. The above step is repeated several times, gradually increasing the resis- tance of P 1 . The range of settings of P2 over which the lamp remains lit will become increasingly smaller and may even disappear. When P 1 can no longer be usefully adjusted, its resistance is reduced to the point where the AC voltage at point A is not greater than 2.5 Vpp. In the case of a normal FM stereo trans- mission, the average value of the voltage at this point will be in the region of 1 V p p, i.e. 350 mV RMS. In the absence of an AC millivolt- meter, it is still possible to determine the presence of excessively large signal levels at this point by the increase in distortion in the audio signal. In case of doubt PI can simply be set to the mid -position. Of course if one possesses a frequency meter P2 can be adjusted directly with the aid of the test output (P2 is simply adjusted until a reading of 228 kHz is obtained — the receiver is not tuned to a transmission). The channel separation can be optimised with the aid of potentiometer P3. The adjustment is carried out with P4 set for maximum resistance. Adjusting P3 affects the crosstalk from both right to left and left to right channels, and the optimum position represents a compro- mise situation where there is equal crosstalk in both directions. For this adjustment in particular the test tone transmissions broadcast by the BBC — which are discussed below — are ex- tremely useful. H VHF Stereo Test Transmissions Each Monday and Saturday evening, at the end of normal programmes. Radio 3 VHF transmitters broadcast a series of test tones which can be used to help set up and check FM receivers. A number of these test signals are useful when aligning both the IF strip and the stereo decoder published in this issue. The various test signals broadcasted are listed in table 1 . All signals are sinus- oidal, and are transmitted with pre- emphasis. The fact that most of the test signals are broadcast at a constant level for at least a minute provides an excellent opportunity to accurately align the quadrature tuned circuit of the IF demodulator. As was described in the alignment procedure for the IF strip, the core of the phase quadrature coil, L2 (see figure 4), is adjusted for maximum (audio) output signal, where- upon the core of L3 (if present) is adjusted for minimum output signal level. This procedure is obviously rendered more difficult if there are variations in the amplitude of the input signal. The first test signal is intended to identify the left and right channels and to establish the reference signal level for the subsequent test signals. In practice the 0 dB level corresponds roughly to the average signal level of a normal FM transmission. The second test signal can be used to adjust the VCO frequency of the stereo decoder by increasing the difference signal of the two channels to a maximum. When the amplifier to which the stereo decoder is connected is switched to mono, there should in theory be no audible output. Both this signal and the next can also be used to set the input sensitivity of the stereo decoder. PI of the decoder circuit is adjusted until the AC voltage at point A (see figure 4) is approximately 2.5 V pp . If a distortion meter is available, it is also possible to check the distortion of the tuner (one should bear in mind that the distortion level of the signal before it is processed by the tuner is already some 0.2%). Test signals 4 and 5 enable the crosstalk between both channels to be monitored and then minimised by adjusting P3 of the decoder circuit. Tests 6 and 7 are intended to check the frequency response of the two channels, and to determine the amount of crosstalk between channels at low and high frequencies. Finally, test 8, during which only the 19 kHz pilot tone is transmitted, allows the noise level to be checked (since one is hearing nothing but noise!). It can also be used to tune the fre- quency of the VCO in the stereo de- coder, using the procedure described in that article. Note that the 19 kHz pilot tone is transmitted during all the above-mentioned test signals. Although the full schedule of test signals is broadcast only on Mondays and Saturdays, on other days a 250 Hz tone is transmitted in the left channel only for approximately 20 minutes after close of programmes on Radio 3. This signal can be used for channel identification and to check cross-talk. Further information on the BBC’s VHF test transmissions and FM recep- tion can be obtained from the Engineering Information Department. BBC, Broadcasting House, London W1 A 1AA. Table 1. Test No. I Time Left channel Right channel 1 T* 250 Hz at 0 dB 440 Hz at 0 dB 2 T + 2 min 900 Hz at +7 dB 900 Hz at +7 dB in in antiphase to left channel 3 T + 6 min 900 Hz at +7 dB 900 Hz at +7 dB in phase with left channel 4 T + 7 min 900 Hz at +7 dB — 5 T + 8 min — 900 Hz at +7 dB 6 T + 9 min 20 sec Tone sequence at - 4 dB: 60 Hz, 900 Hz, 5 kHz. 10 kHz. Sequence is repeated 7 T + 10 min 20 sec Tone sequence as for L channel on test 6 8 T + 1 1 min 20 sec *T is approximately 4 minutes after close of Radio 3 programmes (see Radio Times) J elektor june 1979 — 6-21 variable logic gate J. C. Knapp The logic gate type MC14530 (and its equivalents) is not used very much. This ‘dual, five-input majority gate’ is, however, intriguing. Amongst other things, it can be used as a ‘variable logic gate’ as described here: by inter- connecting some of its inputs, a logic gate is obtained that will give various logic functions depending on the logic level at two ‘control inputs’. This can be useful in education, as well as in micro-computers! Figure 1 is a simplified block diagram of the MCI 45 30. As can be seen, each of the two ‘majority gates’ is actually followed by an EXNOR. The output M of the majority gate is determined in a most democratic way: it is equal to the majority vote (input levels)! If three or more of the inputs A ... E are at logic ‘1’, the output will also be ‘high’; if three or more inputs are at logic ‘0’ the output will also be ‘low’. This corresponds to the Boolean equation: M = ABC + ABD + ABE + ACD + ACE + ADE + BCD + BCE + BDE + CDE. The function of the EXNOR at the output is simply to invert the output level, if required. As can be seen from the truth table (table 1), the output Z is equal to M if control input W is at logic ‘ 1’; otherwise it is inverted (Z = M). The ‘variable logic gate’ is obtained by linking two inputs A and B of the majority gate to form a common input X, and similarly Unking C and D into one input Y (see figure 2). The remaining input E is used as one of the control inputs. The complete unit now functions as shown in table 2. As can be seen, four different logic functions are determined by the logic levels on the control inputs E and W. Furthermore, if X and Y are Unked to form a single input the complete gate will operate as an inverter as long as W is ‘low’; with W at logic ‘F a non-inverting buffer is obtained. Finally, with X and Y once again Unked (forming a common input V), W and V can be used as two inputs of an EXNOR. Of course. 1 2 r wo {— “1 xo^j: A B pE>h 2 C M D J i i i EO 1- E i 1 1 1/2 MC 14530 | 79083 2 Table 1. W M 2 comments 0 0 1 Z = M 0 1 0 1 0 0 2 = M 1 1 1 Table 2. control logic combined logic inputs inputs output function E W X Y 2 0 0 1 0 0 0 1 1 0 1 1 NAND: Z = X • Y 1 1 0 0 0 0 0 1 0 1 1 0 0 0 AND: Z = X • Y 1 1 1 0 0 1 1 0 0 1 1 0 0 0 NOR: Z = X + Y 1 1 0 0 0 0 1 1 0 1 1 0 1 1 OR: Z = X + Y 1 1 1 V X 0 0 1 1 0 NOT: Z = V 1 • 0 0 1 1 BUFFER: Z = V X 0 0 1 X 0 1 0 X 1 0 0 EXNOR: Z = V ® W X 1 1 1 X = Don't care. 6-22 — elektor june 1979 FM IF strip PM IF Strip using the CA 3189E For some years now the 3089 limiter/demodulator 1C has been an industry standard for use in the IF stages of FM receivers. This 1C has featured in numerous tuner designs and been proven to combine high quality performance with reliability. Recently, however, an improved version of the chip, the CA 3189E has appeared, and the new 1C boasts superior specifications and several additional facilities. This article takes a look at the new device and presents a circuit design and printed circuit board for a high- quality IF strip incorporating this 'state-of-the-art' 1C. The circuit is also compatible with the stereo decoder published in this issue. The basic design of an FM stereo re- ceiver is shown in the block diagram of figure 1. The incoming RF signal is amplified and converted into the 10.7 MHz intermediate frequency (IF) by the front-end. The function of the IF stages is to limit any variations in the amplitude of the IF signal, thereby eliminating AM interference, and de- modulate the limited signal, i.e. convert it into the multiplex encoded (MPX) stereo signal. The latter is then decoded to obtain the left and right channel audio information. Two signals are fed back from the IF stages to the front-end, the automatic frequency control (AFC) signal, and the automatic gain control (AGC) signal (shown dotted). The AFC signal ensures that the front-end remains exactly tuned to the received signal fre- quency, whilst the AGC signal varies the gain of the front-end to compensate for fluctuations in signal strength. The reason why the AGC connection is shown as a dotted line is that in many situations, in particular where there are two transmitters close together broad- casting signals of approximately the same strength, automatic gain control can do more harm than good. AGC is really only useful with very high per- formance front-ends, and even then only in certain cases (i.e. when the front-end uses a PIN diode attenuator). Improved performance The CA 3 1 89E is basically similar to the 3089, however it possesses a number of advantages over its predecessor. It has a greater signal-to-noise ratio (72 dB as opposed to 67 dB) and the bandwidth has been reduced from 25 to 15 MHz, thereby improving the stability of the circuit. The CA3189E also offers a number of additional facilities not provided on the 3089. By providing for the use of an external audio load resistor rather than fabricating the load resistor on-chip, as is the case with the 3089, the audio output level can be varied by selecting different value resistors. Many FM tuners incorporate an audio muting circuit, which elimin- ates annoying interstation noise by ‘squelching’ the audio signal when no transmission is being received. The squelch circuit is triggered when the audio signal falls below a certain threshold level. In addition to this type of noise muting circuit, the CA 3189E also provides ‘deviation muting’, the audio signal being squelched when the receiver is mis-tuned. This facility prevents those loud thumps which occur when one tunes rapidly through a station. Finally the new IC offers one or two extra ‘luxuries’, such as an on-channel indicator (i.e. an output which goes low when the receiver is correctly tuned into a station), adjustable AGC, and sup- pression of the tuning meter voltage at very low signal levels, when the meter reading is no longer significant. The main specifications of the CA 3189E are listed in table 1. The block diagram of an IF circuit incorporating the CA 3189E is shown in figure 2. As can be seen, the basic design conforms to the now almost universally adopted approach of a single IF filter before the amplifier and demodulator stages. This is in contrast to the older method of interposing several filters between successive amplifiers. ILF FM IF strip elektor june 1979 — 6-23 Figure 1. Block diagram of a typical FM tuner. Figure 2. The design of the IF circuit con- forms to the now commonly adopted approach of a single 10.7 MHz filter before the amplifier/limiter stage. Figure 3. Internal block diagram of the CA 3189E. The actual circuit diagram of the 1C is rather difficult to follow, comprising as it does some 100 transistors, almost 100 re- sistors, a dozen diodes and 15 capacitors! Inside the CA 3189E The operation of the CA3189E can best be explained with reference to the internal block diagram of the device, which is shown in figure 3. The input IF signal is fed to three lim- iting amplifiers connected in cascade. The gain of these stages is such that limiting occurs at an input voltage of 12/iV. If one takes into account the gain of the front-end, then this is equiv- alent to an antenna sensitivity of 1 to 2 /iV or even less. The output of each of the amplifier/lim- iters is also fed to a peak level detector, the outputs of which are in turn summed to provide the drive voltage for the signal strength meter and the control voltage for the AGC circuit. The latter provides an output signal which goes high (i.e. approx. 9.5 V) when the input signal to the IC falls below a certain threshold value which is set by PI. Once amplified and limited, the IF signal is fed to the demodulator, which employs a quadrature-tuned circuit connected between pins 8, 9 and 10 of the IC. In addition to the MPX audio signal, the demodulator provides a control signal for the AFC circuit. As already mentioned, the audio mute (squelch) is triggered both by inadequate audio signal level (this is determined by the level detector connected to the demodulator) and by frequency devi- ation. The AFC circuit provides a DC control voltage which varies accord- ing to the extent to which the tuner deviates from the station frequency. This is fed to the ‘deviation mute’ circuit, which in turn (via the mute drive and external mute circuits) controls the squelch signal. The level at which the noise muting cuts in can be varied by means of the external mute circuit. The voltage level at pin 12 of the IC can also be used as an on-channel indicator. When an FM transmission is being received the voltage at this pin is 0 V, whilst if the receiver is not tuned to a station it will be roughly 5.6 V. A complete amplifier/limiter/demodulator The circuit diagram of a complete IF amplifier/limiter/demodulator incor- porating the CA 3189E is shown in figure 4. T1 and associated components form the input buffer amplifier of figure 2. There are various possibilities for the 10.7 MHz filter, and in principle either ceramic filters or LC tuned circuits can be employed. Generally speaking, however, ceramic filters are preferable since they do not require alignment. Several possible types are listed in table 2, along with the values of the corresponding resistors and capacitors in each case. If two filters connected in series are used, they must of course have the same resonant frequency (which in the case of the SFE 10.7 MA means that they should both have the red colour code). The bandwidth-limited IF signal is then fed to the input of the IC. The quadra- ture tuned circuit is formed by a double- tuned LC network. It is also possible to use a single tuned circuit, in which case L3 is omitted and a through connection made along the dotted line. R9 is also omitted and the value of R7 altered to 2k7. In the event of a single tuned circuit being employed, an IF trans- former, type 33733 (Toko), can be used instead of L2. The advantage of a single tuned circuit is that alignment becomes much simpler, however it also inevitably 6-24 — elektor june 1979 FM IF strip Table 1. Main characteristics of the CA 3189E (U b = 12 V;T amb = 25°C) DC conditions (no input signal) min. typ. max. supply current 20 31 44 mA DC voltages pin 1 1.2 1.9 2.4 V pin 2 1.2 1.9 2.4 V pin 3 1.2 1.9 2.4 V pin 15 (AGC) 7.5 9.5 11 V pin 10 (Uref) AC conditions 5 5.6 6 V input sensitivity (limiting threshold) - 12 25 MV AM rejection 45 55 — dB output voltage (pin 6) harmonic distortion 325 500 650 mV single tuned circuit — 0.5 1 % double tuned circuit — 0.1 — % signal-to-noise ratio 65 72 - dB Table 2. Some possible IF filters type bandwidth (kHz) R4,R5,C3,C4 SFJ 10.7 MA (red)* 280 SFW 10.7 MA (red)* 220 R4 = 330 n R5 = 330 n two filters 280 SFE 10.7 MA (red)* replaced by two filters** wire links CFS 10.7 A 300 BBR 3132 A** 240 R4 = 1 k * Murata ** Toko R5 = 560 n C3 = 100 n C4 = 100 n entails an increase in harmonic distor- tion from around 0.1% to 0.5%. The MPX audio signal is fed out via pin 6 of the IC. R 1 1 and Cll form the de-emphasis network. If the circuit is followed by a stereo decoder, Cll should have a value of 56 p. If, on the other hand, one is only interested in mono reception C 1 1 is increased to 6n8 and the MPX signal represents the audio output signal. The AFC voltage U a f c , is brought out to pin 7. The relationship between U a f c and the average value of input fre- quency, fi, is illustrated by the graph in figure 5 a. Since, strictly speaking, the CA3189E provides an AFC current instead of a voltage, the relationship shown in figure 5 a is only valid if the AFC output is not loaded too heavily. At the IF frequency of 10.7 MHz the AFC voltage will be 5.6 V, and equal to the AFC reference voltage provided by the IC. Any slight differences can be eliminated by adjusting preset P2. The relationship between the AFC current, I a f c , (flowing into the IC) and the input frequency, fj, is illustrated in figure 5b. The response shown is virtually inde- pendent of load conditions. In addition to providing the feedback signal to the front-end, the AFC output can be used to drive a centre zero tuning meter, which is connected in series with R8. The mute drive signal is fed out to pin 12, and the external mute circuit which, with the aid of P3, allows the muting threshold to be varied, is con- nected between pins 12 and 5. The deviation mute threshold is determined by the value of R8. With the value shown muting occurs when the average input frequency deviates by more than roughly 35 kHz from the required 10.7 MHz. The signal strength meter is connected to pin 13 of the IC. The meter response is approximately logarithmic (i.e. the voltage at pin 13 increases as the logarithm of the input voltage), thereby permitting input signals of widely differing strength to be displayed. Thus, for example, an input voltage of 10/iV will cause a current of approximately 1 /uA to flow through the meter, whilst an input voltage of 100 mV will produce a meter current of roughly 100 /lA. The components shown in dotted lines are part of the AGC circuit. As was mentioned earlier, such a facility is not always desirable. The voltage at pin 16 controls the AGC threshold, i.e. the signal level at which gain reduction occurs. This can therefore be varied by means of preset PI. If AGC is not desired, the components in question are simply omitted and pin 16 connected to ground. Construction The circuit can be mounted on the printed circuit board of figure 6. To improve stability the board is extensively clad with copper earth planes. The board is designed to accom- modate several types of IF filter. The SFW 10.7 MA is mounted on the smaller rectangular contour, whilst the SFJ 10.7 MA is shown as the dotted oval. The CFS 10.7 A and SFE 10.7 MA, two of which must be connected in series, have three terminals and are shown as small ovals. Although the input and output connections of ceramic filters are sometimes marked on the case (a red dot marks the output), they are constructed symmetrically, so that in principle it is possible to mount them two ways. If ceramic filters are used, then capacitors C3 and C4 should be replaced by wire links. By far and away the best results are obtained if linear phase filters, type 3132A from Toko, are used. This filter contains six coupled tuned LC circuits which are pre-aligned at the Figure 4. Complete -circuit diagram of the IF amplifier/limiter/demodulator. Most types of IF filter can be used for FI; several possibi- lities are listed in table 2. If AGC is not required, PI and R14 are omitted. Figure 5. AFC voltage (a) and AFC cur- rent (b) as a function of input frequency. The response of figure (a) is strongly influenced by load conditions, whilst the second is independent of load. factory, and has a group delay of less than 0.5 /us over the frequency range 10.525 to 10.875 MHz. The linear phase filter is mounted in the large rectangle shown on the p.c.b. The input and out- put connections are indicated on the underside of the filter, whilst the input is also marked on the case of the filter by a dot. As is apparent from table 2, capacitors C3 and C4 are required if the linear phase filter is used. Finally, it is recommended that the circuit be screened, using for example copper clad board or tin plate. Alignment The alignment procedure is comparative- ly simple. With the AFC disconnected, one tunes to an FM transmission, where- upon L2 is adjusted for maximum out- put signal. If L3 is used, the next step is Parts list Capacitors: C1,C2,C5,C6,C7 = lOOn R1 =47 k C3,C4 — see table 2 R2,R8 = 10k C8 = 220 m/16 V R3.R13 = 470 C9,C10,C14 = 10m/16 V R4.R5 — see table 2 Cl 1 = 56p (stereo), 6n8 (mono)* R6 = 100 n C12 = 1 m/16 V R7 = 18 k* C13,C15,C16 = 10 n R9 = 2k7* R10 = 68 k Semiconductors: R11 = 5k6 T1 = BF 494 R12 = 22 k IC1 = CA3189E (RCA) R14 = 10 k* R 1 5 = 33 k Miscellaneous: PI = preset potentiometer, 47 k (50 k)* LI = miniature choke, 22 mH L2 = 34343 (Toko)* P2 = preset potentiometer, 10 k L3 = 34342 (Toko)* P3 = linear potentiometer, 10 k FI = IF filter — see table 2 M = moving coil meter, * see text 150 .. . 250 mA to adjust L3 for minimum output volt- age; distortion should then be minimal. It goes without saying that the align- ment procedure is greatly facilitated if one has suitable test gear — oscilloscope, high impedance millivoltmeter, distor- tion meter, FM test generator (note that a circuit for an FM test generator was published in Elektor 45). In the absence of a suitable millivoltmeter, one way of monitoring the output voltage of the circuit is to use the VU meter on a tape recorder. However it is also possible to obtain quite satisfactory results tuning by ear. The use of the BBC VHF stereo test transmissions as an aid to alignment of the IF strip is discussed in a separate section with the article on the stereo decoder published in this issue. The AFC control voltage is adjusted by means of P2. With the AFC inoperative one tunes to a fairly weak transmission and then switches in the AFC. In most cases the strength of the reception will change (this will be apparent audibly or Figure 6. Printed circuit board and com- ponent layout for the IF strip (EPS 78087). as a different reading on the signal strength meter). The adjustment proce- dure basically consists of restoring the original reception. Once P2 has been correctly set, switching the AFC in and out should have no effect upon the strength of a received signal. The AGC threshold voltage is set by means of PI, and will depend upon the gain of the front-end which is used. The input level at which AGC occurs can be varied between roughly 200 /iV and 200 mV. M elektor june 1979 — 6-27 Intermodulation distortion The reasons why automatic gain control is not always desirable become clearer if we examine the phenomenon of intermodulation distortion. This type of distortion ‘occurs particularly when several strong transmissions on adjacent frequencies reach the tuner front-end. As a result of various non-linearities in — especially — the input stage of the front-end, second, third, etc harmonics of the input signals are generated. In themselves, these are not necessarily significant, since their frequencies are considerably removed from the original input signals. However in addition to the simple harmonics, all sorts of sum and difference products are produced. These intermodulation products are often at frequencies close to that of the input signals, and can considerably distort the desired audio information which we wish to receive. Figure 1 shows the spectrum of a signal afflicted by intermodulation distortion. In addi- tion to the two original transmitter frequencies fi and ft , a number of sum and difference products are present. The degree of intermodulation, and thus the level of distortion, is often directly proportional to the DC current of the input transistor, and increases considerably as the collector current of the transistor decreases. By way of an example, figure 2 shows the relation- ship between intermodulation and the collector current of a BFT 66/BFT 67. Many AGC circuits are arranged so that the gain of the input stage is reduced (and with it the current through the stage) to compensate for the presence of large input signals. However, as we have seen, this has the effect of increas- ing the intermodulation distortion. Such AGC circuits will function satisfac- torily if only one reasonably strong transmitter signal is received, but not if two such signals on adjacent fre- quencies are picked up. An exception to this rule is formed by AGC circuits which employ PIN-diode attenuators, since the current through the input transistor then remains unaffected. 78087A 2 goodbye E300/E310, hello J300/310 The E300 and E310 are well-known types of JFET, which have often been used in Elektor projects. Recently however, these two old ‘workhorses’ have been withdrawn by the manufac- turers, and replaced by two new equi- valents which have the same number, but prefixed by a different letter — J300 and J310. The two new types can be used wherever the E300 and E310 were specified previously. Fair enough, why all the fuss then? Unfor- tunately there is a slight snag, due to the fact that the J300 and J310 are housed in a different type of package. The new case has the advantage that the connec- tions can be identified from above, however it also means that the pin-out has been completely revised. The differences between the old and new cases are shown below: When using Elektor printed circuit boards, it is therefore important to check whether they are intended to accommodate the E300/E310 or the J300/J310. A further point worth noting is that the J300 comes in four different versions, A, B, C and D respectively. There was a considerable leeway in the specifications of the E300. The J300 types on the other hand, whilst covering the same ground as the E300, are much more accurately specified. The main differ- ences are listed in table 1 . K Table 1. J300A J300B J300C J300D Characteristics Min Max Min Max Min Max Min Max Unit Test Conditions •OSS Saturation Drain (Note 2) Current 4 9 7 15 12 25 21 45 mA V DS = 10 V VqS = 0 Gate Source V D S = io V Ip » 1 nA v GS(off) Cutoff Voltage -1.5 -3.0 -2.0 -4.0 q ID 1 CM 1 -3.5 -7.0 V 6-28 — elektor june 1979 Aquarium thermostat Aquarium thermostat (W. v. Dreumel) Constant water temperature is vital for tropical fish. The electronic thermostat described in this article will maintain the temperature within 1°F (!6°C). It is fail-safe, in the sense that it will give loud and clear warning if anything goes wrong. When keeping tropical fish in an aquarium, a good heating system is required. It should meet the following specifications: • It must be possible to set the desired temperature accurately. • The circuit must maintain the tem- perature within ± 1°F (± Vi°C). • The actual temperature must be clearly displayed. • A clear, audible warning must be given if the temperature falls too low or rises too high; furthermore if the temperature rises above the preset limit the heating must certainly be switched off. A block diagram of the complete system is given in figure 1. A temperature sensor provides an output voltage that varies linearly with temperature. An offset compensation is used to set the desired measuring range; the resulting voltage is amplified and used to control a LED display, providing a clear replace- ment for the old mercury thermometer. The control voltage for the heater switch is taken from the desired ‘tap’ on the LED display. As shown in figure 2, an LM3911 is used as a temperature sensor. This IC gives an output voltage that rises with temperature at the rate of 10 mV/°K. At 295°K (72° Fahrenheit, or 22°C) the output voltage is 2.95 V. If this tempe- rature is taken as the low end of the scale, an offset of 2.95 V is required. This is achieved with the voltage divider chain R2 . . . R5 and PI ; with PI correctly set, the voltage at the R3/R4 junction rises from 0 V (at 72°F) at the rate of 10 mV/°K. With the gain of IC2 set at x 100, this results in an output voltage that rises at 1 V/°C. Alterna- tively, if a scale calibrated in degrees Fahrenheit is required, a gain of x90 can be obtained by replacing R8 by two 1 8 k resistors connected in parallel (giving an equivalent resistance of 9 k). In this case, the output voltage will increase at 0.5 V/°F. For the UAA 180, ‘full scale’ corres- ponds to an input voltage of 6 V. Each of the twelve LEDs D4 . . . D15 therefore corresponds to a temperature step of 1°F (or 0.5°C, for R8 = j0k); the total measuring range is 12°F or 6°C. The ‘low end’ of the range can be set by PI. If four green LEDs are used 1 79106 1 l Aquarium thermostat elektor june 1979 — 6-29 top view 3911 top view 3911 * Fahrenheit scale : R8 - 9K (18K//18K) Centigrade scale : R8 = 10K (see text) £/V~ pin 4 r\ connected i / to case in the middle of the range and yellow and/or red LEDs towards the end, a clear indication of temperature is obtained. A multi-position switch is used to select the cathode of the LED that corresponds Table 1. This 'truth table' shows how the various parts of the total circuit (figures 2 and 3) react as the temperature rises over the full range. It is assumed that the switch is set in position 6; normally, of course, the heater would be switched off at this point so that the temperature would not rise any further. Figure 1. This simplified block diagram illustrates the basic principle of the aquarium thermostat. Figure 2. This section of the circuit works as a thermometer with high accuracy over a limited range of temperatures. Table 1. switch position i 6 >J 00 CD o 77 78 1 79 I 80 I 81 i 1 nnnm T5 1 a buzzer !_ T6 T7 0 i — relay i to the desired temperature. The voltage at this point is used to control the heater switch shown in figure 3. As long as the water temperature is lower than the selected value, the voltage at the base of T1 will be high (see figure 4) so that this transistor is cut off. T2 will turn on, lighting the LED in the opto- coupler (IC4). This, in turn, causes T3 to turn off; the voltage across D21 and D22 rises above their zener voltage so that gate current flows into the triac. The heater is turned on. When the desired temperature is reached, the voltage at the base of T1 will drop so that this transistor is turned on. Via the steps outlined above, this causes the triac to extinguish — turning off the heater. As the water temperature swings up and down around the required value, the heater will be switched off and on - maintaining a fairly constant tem- perature. If the heater fails, the temperature will drop until even the lowest LED in the scale goes out. At this point, T4 will turn off; T5 will then conduct so that the buzzer sounds. On the other hand, if the heater remains on for some reason (if the triac fails, for instance) the temperature will rise towards the upper limit of the range. At that point, T6 will turn on. Via D24, T4 and T5 this operates the buzzer; simul- taneously, T7 is turned off so that the relay drops out — disconnecting the supply to the heater. Construction and calibration For safety reasons, the heater must be extremely well insulated. This applies, of course, to any heater used in an aquarium .... Although the block diagram shows the temperature sensor dangling in the water, this is by no means the best place. It is more practical to clamp or glue it to the outside of the glass. The best position (relative to the position of the heater) can be determined by watching the temperature fluctuation on the LED scale with the circuit in operation. There is only one calibration point in the circuit: PI. This is used to set the desired measuring range — using a normal thermometer as a reference. Figure 3. The second part of the circuit is the heater switch with additional fail-safe circuits. Figure 4. This diagram illustrates the way in which the two parts of the circuit are inter- connected. A multi-position switch is used to select the desired operating point for the heater switch. Figure 5. A suitable power supply circuit for the aquarium thermostat. l/ASl Aquarium thermostat elektor june 1 979 — 6-31 Initially, PI should be set to minimum resistance (this corresponds to the lowest ‘minimum temperature’ setting). If even this ‘minimum temperature’ proves too high, due to component tolerances, R4 should be decreased to 22 k. On the other hand, if the ‘minimum’ setting remains too low even with PI set to maximum, R4 can be increased to 33 k. It is not a good idea to increase the value of PI — adjustment is critical enough as it is: the nominal adjustment range for the minimum temperature is from 57°F to 153°F! It might even be a good idea to use a multi-turn potentiometer . . . M 6-32 - elektor june 1979 monoselektor monoselektor Remote control of up to a dozen separate systems can lead to some confusion in the hardware. The main feature of the project described in this article is that only one operating s-.vitch is used to control up to 15 outputs with either 'digital' or 'analogue' capabilities (together with a visual indication of the condition of each channel). One other highly desirable (and unusual) feature has been included. The Monoselektor will allow you to check that you really are turning the (remote controlled) central heating up before inadvertently closing the (remote controlled) garage door during the wife's car parking operations. Figure 1. Front panel fascia for the Mono- selektor. The fascia is a contact-adhesive plastic sheet, which is available through the EPS service (EPS 79039-F). The Monoselektor is the basis of a remote control system with up to 15 separate channels. The outputs of channels 1 to 1 1 function as on/off switches while the remaining four (12-15) each have two complementary outputs (i.e. when one is on the other is off) and can be used as a form of ‘analogue’ control. For a clearer understanding of the operation of the Monoselektor it is useful to refer to the illustration of the front panel in figure 1 where the two rows of LEDS can be seen. The upper row contains the channel condition indicators and will display the state of the channel outputs. It will be noted that channels 12-15 each have two LEDS since these are the complemen- tary (or analogue) channels and there- fore have two outputs. Channel operation The LEDs forming the lower row are used as the means by which each channel is selected and do in fact ‘run’ (i.e. light in sequence from left to right). The ‘flow chart’ in figure 2 gives a clear indication of the methods of control for the different channels of the Mono- selektor. The selection of a channel is carried out by pressing the ‘selektor’ button when the LED in the lower row of the required channel is lit. This will stop the running LEDS at that point but the channel output state will not be altered until the selektor is pressed a second time. There is a ‘finite’ time (up to ten seconds) after the LEDS have stopped in which the button must be operated to effect an output change. This time period is termed the ‘wait’ time and is indicated by the ‘wait’ LED on the front panel (figure 1). After each further operation of the button (effect- ing an output change) a further ‘wait’ time is initiated, during which the ‘command’ can be cancelled by again pressing the button. If no action is taken within the ‘wait’ time, the run- ning LEDS resume. The 'stand-by' channel The ‘stand-by’ mode is exactly what the term implies. The Monoselektor will be on stand-by when (a) mains power is first applied (i.e. on switch on) and (b) when channel ‘0’ is selected. The following conditions will then be in effect: (1) channel ‘0’ upper and lower LEDs will be on; (2) with the exception of channel ‘O’, the lower row of LEDs will be off and will remain off; (3) the outputs of channels 1-15 will remain as they were with a visual indication given by the upper row of LEDs. Block diagram Before looking too deeply at the circuit itself, a description of the block diagram in figure 3 will simplify matters. The decoder, controlled by the oscillator and counter, enables each of the chan- nels in turn, thus producing the running light effect on the bottom row of LEDs. The ‘on’ time of each LED can be varied by altering the frequency of the oscillator. The section of the circuit labelled ‘automatic reset’ ensures that, when the circuit is initially switched on, channel 0 is activated and channels 1 ... 1 1 are disabled, i.e. the Monoselektor is in the standby mode. When the selector switch is depressed, the pulse shaper (which serves to eliminate the effects of contact bounce) supplies a trigger pulse, (c), to the ‘wait’ circuit. This initial operation of the selector switch causes the ‘wait’ cjnrnit to do two things: A signal at Cv stops the oscil- lator, which remains inhibited for a short time (presettable) and the ‘wait’ LED on the front panel is lit to indicate that a ‘wait’ period has been initiated. If the ‘wait’ time of e.g. 5 seconds is allowed to elapse without the selector switch being operated a second time, then — again via (E) — the inhibit on the oscillator is removed and the ‘wait’ LED is extinguished. If, however, during this period the selector switch is pressed, then the oscillator is inhibited for a further 5 seconds and the ‘wait’ circuit provides a pulse output at 15 ). This activates the output of the channel at which the running light was stopped (if that channel was already activated the pulse will have the opposite effect, i.e. disable the output in question). 6-34 - etektor june 1979 monoselektor 79039 2 The oscillator can also be inhibited by means of a signal at (H), which is in fact the output of channel 0 (the standby channel). Output ( 5 ) of the ‘wait’ circuit is nor- mally held low. A single operation of the selector switch has no effect upon the state of this signal. If, however, the selector switch is pressed twice, the second operation occurring inside the ‘wait’ period of the first, then (D) is taken high for the duration of the period that SI is held closed. Thus it is signal (D) which provides the control pulses which enable or disable channel outputs. The oscillator, and thus the running light sequence on the bottom row of LEDs, is inhibited when the level at (E) (normally low) is taken high. The output logic states of channels 1 ... 1 1 are switched by a positive going pulse on the (D) line. These out- puts are active low. In the case of out- puts 12... 1 5,they are taken low for as long as the (D) line remains high, i.e. these outputs stay low for as long as the selector button is held down. Circuit diagram The complete circuit diagram of the Monoselektor is shown in figure 4. Since channels 1 ... 1 1 are identical and channels 1 2 ... 1 5 are likewise all the same, the circuit diagram shows only channels 1 and 12. Thus R15, for example, occurs 1 1 times, and in order to distinguish the various com- ponents a suffix is added, which indi- cates which channel they belong to. The suffix ‘a’ refers to channel 1, ‘b’ to channel 2, ‘c’ to channel 3, and so on. Pulse shaper The pulse shaper circuit consists of a monostable multivibrator and Schmitt trigger (MMV1 and Nl). When the selector switch is depressed, (g) is taken low. However, as a result of the inevitable contact bounce of the switch, there is considerable ‘jitter’ on the leading edge of this pulse. The fact that the Q output of the monostable is taken low by the first negative-going edge of (8) eliminates the effects of contact bounce, ensuring that a ‘clean’ pulse is available at the output of the Schmitt trigger Nl. This ensures that (C)in figure 4 remains high for the period of time the switch is depressed. The 'wait' period circuit The ‘wait’ circuit (MMV2, FF1, N2, N3, N5 . . . N7) is required to perform three basic functions. 1 ) When the selector switch is depressed (and (£) goes high) the ‘wait’ LED should light and the oscillator must be inhibited. 2) As soon as the selector switch is released, a preset ‘wait’ period is initiated. 3) If, within the ‘wait’ period, the selector switch is pressed again, a control pulse must be supplied to the appropriate channel. As was apparent from figure 3, this means (B) going high and remaining in that state for as long as the selector switch remains depressed. The actual operation of the ‘wait’ circuit is as follows: When the selector switch is pressed for the first time, fcV goes high, taking the output of N7 ((?)) low. This in turn causes the output of N3 ((E)) to go high. Figure 2. This flow chart illustrates the 'operating principle' of the Monoselektor. Figure 3. Block diagram of the Monoselektor. Figure 4. Complete circuit diagram of the Monoselektor. monoselektor elektor june 1979 — 6-35 6-36 - elektor june 1979 monoselektor J ® 79039 9 a thereby lighting LED D 1 , andinhibiting the oscillator. As long as © remains low, the reset input of FF1 is also held low, allowing the flip-flop to be trig- gered. The RC-network comprising R4 and C4, which provides a delay of approx 250 ms, is included in order to suppress propagation delay ‘spikes’. When the selector switch is released, X) goes low, triggering MMV2 (via the (B) input) and taking the Q output high. The output of N7 therefore remains low. The time for which the Q output remains high (i.e. the ‘wait’ time) is determined by the RC-constant of the monostable and can be varied by means of PI between 1 and 10 seconds. As a result of © going low, the clock input of FF1 is taken high (via N6) with the result that the Q output of this flip-flop goes low, enabling gate N5. If the selector switch is now pressed within the ‘wait’ time, © will go high, taking the outputs of N2 low and N5 (©) high. When the selector switch is released FF1 will not change state immediately, since the Q output of the flip-flop, which is high, will hold the clock input of FF1 low via N6. Releas- ing the switch will however retrigger MMV2, initiating a new ‘wait’ period. If during the first ‘wait’ period the selector switch is not pressed again, then when this period has elapsed the Q output of MMV2 will be returned low, resetting FF1 and taking© low. The oscillator is restarted and the LEDs on the bottom row will once more begin to light up in succession. Oscillator and counter circuit The oscillator is formed by Schmitt trigger N4 in conjunction with C5, R7 and P2. The frequency can be altered by means of P2, allowing the rate at which the channels are scanned to be varied between 0.3 and 3 seconds per channel. The oscillator is only enabled if the input of N4 (pin 12) is high, which is only the case if both inputs of N8 are low. Thus, taking either © or © high will cause the oscillator to be inhibited. The oscillator provides the clock to a 4-bit binary counter, IC4. The binary output of the counter is decoded by 1C5, a 4-16 line converter. The result is that the 16 outputs of ICS (Qo . . . Qjs ) are taken high in succession. Channel 0 — the stand-by channel The principal components of this channel are FF2, T3 and T4. When the Qo output of IC5 goes high, T3 turns on and D2 lights. Pressing the selector switch stops the oscillator and holds the counter at channel 0. If the selector switch is then pressed within the ‘wait’ time, a positive pulse from N5 will take the output of N9 high and trigger FF2, a_ ‘D-type’ flip-flop. By connecting the Q output to the D input, the output of the flip-flop will ‘toggle’ each time a pulse is applied to the clock input. Once FF2 has been triggered, the oscillator will remain inhibited, since © (the Q output of the flip-flop) is held high. T4 will be turned on, light- ing the stand-by indicator D3. The Monoselektor is taken out of the stand-by mode by pressing the selector switch twice in succession. This takes the © line high once more, with the result that the output of N9 goes high, FF2 receives a clock pulse and the Q output of the flip-flop (©) goes low. Automatic reset When the unit is first switched on, the automatic reset circuit (formed by T2, R8, R9, R10 and C6) generates a positive pulse of approximately 500 ms. This resets IC4, disables channels 1 . . . 1 1 and selects the stand-by mode by setting FF2 which in turn (via N8) inhibits the oscillator. As the Mono- selektor is now in the stand-by mode LED D3 will be on. Channels 1 ... 11 These channels function in virtually the same way as channel 0 with only minor differences. The outputs of channels 1 ... 1 1 are brought out via a buffer transistor. The main component of monoselektor elektor june 1979 — 6-37 channel 1 is FF3 a . When its Q output goes high, T6 a is turned on, LED D5 a lights to indicate that the channel is switched on, and the channel output is pulled low. Resistor R 1 l a limits the current through T6 a to a safe value, whilst diode D6 a protects the LED against large reverse voltages. D7 a protects T6 a against inductively -caused back-EMF’s. Channel 1 has been used here as an illustration, but the descrip- tion applies also to channels 2 . . . 11. Since the channel outputs are active- low, i.e. they are at logic ‘0’ when switched on, relays should be connected between positive supply and the channel outputs. Channels 12 . . . 15 Channels 12... 15 each have two out- puts and can be used as ‘analogue’ channels. Since they are identical, channel 12 will serve as an example for the rest. When the Q12 output of 1C5 goes high, LED D8/ lights — indicating that this channel has been enabled. If the selector switch is now pressed, the D line goes high and the positive-going edge of this pulse will trigger FF12/. If the Q output of this flip-flop was originally low, it is now taken high, with the result that both inputs of N13/ will be high and T9/ turned on. Thus, as long as (S) remains high, the red LED, D12/, will be lit and the R output of channel 12 will be low. If the selector switch is released and pressed a second time within the ‘ wait ’ period, the Q and (j outputs of FF12; will change state, so that now both inputs of N 1 2/ are high. As long as the selector switch remains depressed, T8/ is turned on, the green LED, D9/ is lit and the L output of channel 12 is held low. If desired, the current-limit resistors R22/ and R27/ can be reduced in value or shorted out, provided that the specifications of the transistors (T8/, T9/) are not exceeded (current < 100 mA). By connecting the two outputs to an electric motor (via a relay) such that it turns clockwise when the R output is low and anticlockwise when the L output is low it is possible to operate ‘analogue’ controls, such as the tuning dial on a radio etc. Power supply The circuit diagram of the power supply is shown in figure 5. The power supply is mounted on the same board as the circuit of figure 4. As can be seen, there are in fact two supply voltages: one for the LEDs (U 0 ) and one for the rest of the circuit. While the circuit supply is ‘unspectacular’, the LED power supply is somewhat more sophisticated. With the aid of a light dependent resistor the brightness of the LEDs is automatically varied to suit the ambient lighting conditions. For example, in strong sunlight the resistance of the LDR (R32) will be extremely low, with the result that T10 will turn off and the base voltage of the Darlington pair T11/T12 increases, causing U 0 to rise. If there is little ambient light, however, the resistance of the LDR will increase, turning on T10 and causing the base voltage of T1 1 to fall. This in turn pulls down U Q , with the result that the LEDs become dimmer. Resistor R29, which is con- nected in parallel with the LDR, ensures that the LEDs are never completely extinguished. If the LEDs are not sufficiently bright in the dark, the value of this resistor should be reduced. T13 ensures that the U Q supply rail is short-circuit proof. It has been found that, in some excep- tional cases, using an unregulated supply for the main circuit can give rise to problems. If this situation should occur there is an easy answer: Construct a 12 volt 1 amp regulated power supply (using a 7812 for instance; see Elektor March 1979, E 47, p. 3-36) and connect its output across C7, taking care to ensure correct polarity. Diodes D15 . . . D 1 8 can be omitted in this case. Construction The track pattern and component lay- out of the printed circuit board for the Monoselektor are shown in figures 6 and 7 respectively. Board assembly involves Figure 5. Power supply for the Monoselektor. With the aid of an LDR the brightness of the LEDs is automatically varied to suit the ambient lighting conditions. Photo A. Example of board with all the components mounted in place. The use of 1C sockets is recommended. Photo B. A close-up of the wire-link connec- tions which have to be made on the com- ponent side of the board. Figure 6. Copper side of the p.c.b. for the Monoselektor. The LEDs and LDR are mounted on this side of the board. Figure 7. Component overlay of the Mono- selektor p.c.b. (EPS 790391. (see overleaf) a slightly unusual method of construc- tion. The components printed in red on the component overlay, namely the LEDs and the LDR, are intended to be set into the front panel and are there- fore mounted on the copper side of the board, not the component side. The anode connections of LEDs 4, 8, 9 and 15 are allowed to protrude through the board by about % inch and are then commoned by a ‘busbar’ or wire link connections. An example of a finished board is shown in photo A, whilst the close-up (photo B) illustrates how the connections are made. The points where these connections are required are clearly indicated on the component overlay and should present no problem. The ‘wait’ LED, Dl, and the on/off LED, D19, are mounted on the front of the case, and connected to the circuit via insulated wire. The light dependent resistor should of course ‘see’ the ambient lighting con- ditions, but must not be mounted such that it is exposed to direct light from the LEDs. The best solution is to provide a small opening in the front panel into which the LDR can be recessed. Transistor T12 should be cooled and the simplest solution is therefore to mount it (making sure it is insulated by a mica washer!) on the back of the case. A self adhesive front panel fascia, as shown in figure 1, is available through the Elektor print service (EPS). It would be desirable to have external access to preset potentiometers P 1 and P2 so that the ‘wait’ period and the ‘speed’ of the running light can be varied. This can be made possible by drilling two holes in the front of the case at the points indi- cated on the front panel fascia. Other constructors may prefer to mount PI and P2 on the back of the printed circuit board. In conclusion It will be apparent that the Mono- selektor is essentially a control unit and therefore requires some form of inter- face between the channel outputs and the ‘outside world’. This can take many forms depending on the user’s require- ments, although so called solid state switches would suit many applications. Further suggestions are given in the article — ‘Switching mains powered equipment’ (Elektor 49) which will be of some interest for this project (it could almost have been written for it . . .). M The Monoselektor was the result of a design project by a study group at the Technical University of Eindhoven in Holland. The initial design requirement was for a simple-to-operate remote control unit for the handicapped and as such the Monoselektor has proved to be ideal. A number of systems have been in use in Holland for some time ( saying a lot for its reliability). 6-38 — elektor june 1979 monoselaktor M 79039 Parts list Resistors: R1 = 10k R2 = 6k8 R3,R7 = 4k7 R4 = 22k R5, R11, R13, R15 a ... R15 k , R18 a ... R18 k , R20| ... R20 o , R23| ... R23 0 , R26| ... R26 0 , R33 = 120 n R6, R8, R12, R14, R16 a ... R16 k , R19 a ... R19 k , R21| R21 0 , R24| ... R24 0 , R25| . R25 0 = 3k3 R9, RIO = 100k R17 a ... R17 k , R22| ... R22 0 , R27| ... R27 0 = 150 fl R28 = Ik R29 = 12k R30 = 1 n R31 = 27k R32 = LDR PI , P2 = 50k preset Capacitors: Cl = 1 p/16 V C2 = 22 m/16 V C3 = 470 m/16 V C4 = 10 m/16 V C5,C6= 100 m/16 V C7 = 2200 m/16 V C8 = 47 m/16 V C9 = lOn Semiconductors: D8| ... D8 0 , 01 2| ... D12 0 , D19 = LED red D2, D3, D4f, D5f. D9| ... D9 0 = LED green D6 a ... D6 k , D10| ... D10 o , D 1 3| ... D13 0 = DUS D7 a ... D7 k , D 1 1 1 ... D11 0 . D 1 4| ... D14 0 , D15, D16.D17, D18 = 1N4001 T1 ,T2,T3,T4, T5 a k , T6 a | T7| o. T8, o, T9| 0 ,fi( T1 1 , T13 = TUN T12 = 2N3055 (heatsink!) IC1 = 4528, 4098 (Motorola MC 14528, RCA CD 4098) IC2 = 4093 IC3 = 4001 IC4 = 4520 IC5 = 4514 IC6 ... IC14 = 4013 IC15 ... IC20 = 4081 Miscellaneous: 51 = SP switch 52 = DP switch FI = 100 mA Trl = trafo 6 ... 9 V, min. 0.5A (or 12 V, 0.5 ... 1 A, see note) .a 6-40 — elektor june 1979 Right up and left down RightAip and left-down W. v. Rooyen Don’t ask us what the following circuit Anyway we are sure some inventive the switch is rotated, a number of pulses is intended to be used for — although reader will come up with other possible are delivered to either the ‘up’- or we have no doubt that there must be applications. ‘down’ output. When the switch is left some application for it. The author Basically the circuit uses a 12-way in one position, both outputs are held himself said: ‘The circuit can be used to switch (or any other number of ways, high. The switch must be of a break- set a digital clock. . .’ - which is quite provided it is a multiple of three), and before-make type, and the end-stop at true, as long as the clock uses TTL. depending upon the direction in which the 1 2th position must be removed. K market elektor june 1979 — UK 13 missing link Modifications to Additions to Improvements on Corrections in Circuits published in Elektor VHF/UHF TV modulator Elektor 42, October 1978, and TV games computer Elektor 48, April 1979. Some readers have commented that the value given for Cl in the VHF/UHF modulator is 22p in the original article and 33p in the TV-games article. Either value may be used, although 22p is preferred. Elekterminal Elektor 44, December 1978. There are a few inconsistencies between the circuit diagram (figure 6) and the p.c. board (figure 7). These do not affect operation of the circuit, but may be misleading if the circuit is to be modified or extended: • In figure 6, pin numbers 1 and 3 of N22 are transposed; • the same applies to pins 12 and 13 of N23. Furthermore, the ‘b’ and ‘c’ indi- cations beside T3 on the com- ponent overlay are transposed; the copper layout is correct, so the transistor should be mounted as shown (ignoring the ‘b’ and ‘c’ indications). BASIC microcomputer Elektor 49, May 1979. For special applications (multi-processor sys- tems and the like), the NHOLD and CONT connections to the SC/MP are brought out to the connector. Normally these will not be used, in which case 10k pull-up resistors should be in- cluded: one between pin 5c and pins la, b on the connector and one between pin 11c and pins la, b. In figure 3, the indications X and Y for the two possible wire links to the NENIN input are trans- posed. For single-processor sys- tems, wire link X should therefore be mounted on the board. The output of N2 (pin 11) is also brought out to the connector: connector pin 31a. NIBL-E Elektor 49, May 1979. In Table 1, in line 1 280 the second byte should be 0E instead of IE. Otherwise, the indication AeRROR AT...’ will be given instead of ‘BRK AT . . .’ (at end of program or after operating the ‘Break’ key). liLLiitLi/. Single board microcomputer A new tutorial microcomputer system has been announced by Texas Instruments Limited. Intended primarily for colleges and universities the TM 990/ 189 M utilises a 16 bit CPU and is available fully assembled. The board is self-contained with IK bytes of RAM (expandable on board to 2K) and 4K bytes of ROM (expandable on board to 6K). The 4K of ROM contains the system monitor ‘unibug’ and a symbolic assembler. The TM 990/ 1 89 M has its own cassette interface and is provided with a 45 key alphanumeric keyboard and a 10 digit, seven segment display. The display can be shifted right or left to display any 10 characters in its 32 charac- ter buffer. Facilities to externally add a standard EIA terminal or TTY interface are also provided. Other features of this single board system include a series of addressable LEDs and a piezoelec- tric speaker. When the microcomputer is switched on, a group of four LEDs flash, the speaker beeps and the display then signals, “CPU READY”. To indicate when the CPU is idle, when the keyboard is in the ‘shifted mode’ and the status of the cassette, other LEDs are used. A comprehensive applications textbook is also provided with the TM 990/189 M which gives instruction in microcomputer fundamentals, assembly and machine language and micro- computer interfacing. Microprocessor Product Marketing Department, Texas Instruments Limited, Manton Lane, Bedford, MK41 7PA. (Tel. 0234 - 67466) (11 95 M ) Relays and more relays A new range of relays have been recently introduced by B and R Relays Limited. Called the E series, these low cost, general purpose, compact four-in-line relays offer a wide range of features and options. They are available in 2 or 4 pole double throw configurations with ratings from low level using bifurcated contacts to 3 Amps or 5 Amps per pole. These relays occupy only 1.2 cu inches of space and come in dust-cover or hermeti- cally sealed versions. There is also a washable vented type suitable for flow soldering assem- bly which features an adhesive tab that is removed after board cleaning to allow correct venting. Both solder and printed circuit mounting versions are offered with optional features such as a test button and mounting plate for top or side mounting. Sockets include snap-in as well as printed circuit versions. The small size and high reliability of 1.5 million operations (at 3A, 28V d.c.) should make the E series relays suitable for general automation, office equipment, copiers, computers etc. Specifications in- clude coil voltages from 5 to 110 V d.c. and 6 to 240 V a.c. Maximum coil dissipation is 2.0 Watts and contact operate time is 13 ms. Operating temperature range is -40 to +70°C and shock and vibration resistance are to lOg and 5g respectively. B and R Relays Limited, Templefields, Harlow, CM20 2BG, Essex. (Tel. 0279-34561) (1189 Ml Solid-State software! A plug-in module for the T158/59 programmable calculators con- taining business decisions soft- ware has been launched by Texas Instruments Limited. The software library contains a range of programs designed to ease the basic tasks involved in business decisions, and enables managers to have instant access to, and awareness of, processes that are normally handled by computers. In addition the library provides a number of programs which can be incorporated in specific routines for a particular company without the need to produce programs involving detailed analysis of specific techniques such as demand forecasting or economic ordering. Texas Instruments Limited, European Consumer Division, Manton Lane, Bedford, MK42 7PA. (1188 M) Combined oscilloscope and digital multimeter A new service instrument from Tektronix, the model 305, combines a dual trace oscilloscope with a full-function autoranging digital multimeter. The oscilloscope functions of the model 305 combine 5 MHz band- width with 5 mV sensitivity, sweep speed from 1 ps to 50 ms per division (with an X10 magni- fier providing 0. 1 #rs per division), full X/Y capability and a range of channel modes and trigger facilities have been incorporated. The multimeter section covers 1 2 voltage and resistance ranges up to a maximum of 1 kV DC or 700 V AC and a maximum resistance of 2 Mn. Accuracy on the DC voltage range is ± 0.1%. The whole package measures only 1 1.2 x 23.7 x 36.0 cm and weighs only 4.8 kg including a built-in rechargeable battery pack. The instrument’s portability and wide range of measurement functions make the 305 ideally suited to field servicing. It is equipped with a carrying strap which allows the user to keep both hands free, and can be operated easily on cat- walks, ladders and in confined spaces normally inaccessible for testing purposes. The model 305 can be operated from AC or DC supplies as well as from the internal rechargeable battery pack. Typical power consumption is 15 W, and battery life 5 hours for oscilloscope operation only and 3Vt hours with both oscilloscope and multimeter in operation. Tektronix U.K. Ltd., Beaverton House, P.O. Box 69, Harpenden, Herts market UK 14 — elektor june 1979 ULMii: UK launches first slot- machine TV newspaper The world’s first coin-operated computerised public information service has been launched in Britain. Slot-machine television sets which will begin to appear in hotels, shops, clubs, airports and other places will bring Britain’s advanced teletext system View- data within everyone’s reach. Until now this latest Post Office communications service has been available only to those who rent special TV sets. A year ago, the Post Office joined with the British Cherry Leisure company to investigate the possibility of producing a coin-operated Viewdata TV terminal. Cherry Leisure is a subsidiary of one of the world’s largest operators of fruit, vending and video machines. By last September, the first prototypes had received Post Office approval and coin-operated sets have been in production in the UK since October. The first installation, formally handed over on 24 January, is in the foyer of a London hotel and other sets are about to be set up in a Central London post office and a passen- ger waiting area at London’s Heathrow Airport. It is likely that up to 50 sets will be established in a chain of hotels. Once a coin has been fed into the TV set a call is made automati- cally to the Post Office computer store which provides the user with a contents sheet of information available. The user can then select a subject and question the com- puter. As well as information on such matters as stock market prices, sports results and weather forecast, the system can offer a series of games and will even help the user to write a poem by offering rhyming words. (469S) Versatile connector The new Intra-Connector from Lektrokit permits quick testing of previously inaccessible inter- connect lines. The single-piece unit comprises a male and female connector that can be simply and quickly interposed between a plug and socket. A pin connection for every contact is brought out to a probe pin on the top of the Intra- Connector. This offers a fast and simple access for electrical moni- toring of individual lines without physical disturbance. Daisy chaining from a single connector cable end is also possible. There are five Intra-Connect units in the Lektrokit range with 20, 26, 34, 40, and 50 contacts respectively. Pins are on a standard 0. lin. pitch as are the test pins that protrude at right angles from the Intra-Connect body to a height of 0.24in. (typ.). Contacts are non-corrosive alloy 770 while the body material is a high grade glass loaded polyester. The Intra-Connector will mate with any standard 0.1 x 0. lin. dual row connectors. Lektrokit Ltd., Sutton Industrial Park, London Road, Earley, Reading RG6 1AZ, Berks. (Tel. 0734-66 91 16/7) (1192 Ml Something new in C.R.T. systems A totally modular C.R.T. electronics system called Visionpack has been announced by Digivision. The system is comprised of five plug together modules which are based upon proven circuitry and components. By dispensing with the problems of dealing with high voltage C.R.T. circuits, Visionpack enables the designer with little experience of TV type circuitry to produce a design for an overall piece of data display equipment. The main interface unit, the Horizontal Scanning Module (1), is connec- ted to the Line Output Transformer Module (2) which produces the auxiliary voltages required by the system. This module is connected to the Vertical Timebase Module (3). A Video Amplifier Module (4) is mounted with the C.R.T. base connector on a small round printed circuit board, while the Scanning Yoke Assembly (5) is fitted to the neck of the C.R.T. The Scanning Yoke Assembly will fit most 20 mm C.R.T. ’s in the size range of 7 to 14”. The five fibreglass printed circuit boards are electrically interconnected via the Pressac 200 connector system. Low cost snap on-off locking pillars can be used to mount modules 1 , 2 and 3 to a conventional base plate or chassis. The modules can be replaced quickly and simply, and the cost of each module is low enough that they can be discarded rather than repaired. Visionpack gives flexibility to the design of data display equipment and its rugged construction gives it reliability. Digivision, 82, Cannock Road, Leicester LE4 7HR. (1197 Ml Miniature cassette tape deck For all you do-it-yourself (micro) computer builders thinking about how and where to mount your tape recorder system take a look at the CM 600miniature cassette transport system now available from BFI Electronics Limited. The complete transport is self- contained (76 x 76 x 64 mm), light (230 g) and includes all necessary Read/Write amplifiers and control circuitry. The CM600 has a two-track recording head which produces a recording V density of 800 Bits/inch (maximum) and a data rate of 2400 Baud. It has a reel-to-reel drive system developing an average forward search speed of 5 i.p.s. and rewind speed of 15 i.p.s. The data capacity of a standard 100 ft miniature tape cassette is 1.6 MBits while the total Read/Write time is 5 minutes 40 seconds. Operation of the CM600 is controlled entirely by external electronic logic signals. Typically these represent tape direction (forward/release), tape motion (stop/go), tape speed (fast/ slow), select Read/Write and data input. Output lines carry data, and indicate the cassette side being used, cassette presence, clear leader, etc. The transport can be operated at any angle between horizontal and vertical while power consumption is typically less than 1 Watt from a single +5 volt supply. BFI Electronics Limited, 516 Walton Road, West Molesey, Surrey, KT8 OQF. (Tel. 01-941 4066). (1191 M) Magnitude comparator for control applications A new 4-bit magnitude comparator is available from RCA Solid State. Designated the CD4585B, this device was designed for servo- motor controls, process control- lers, and other logic applications that require the comparison of two 4-bit words. The circuit can determine whether one 4-bit word (binary or binary-coded-decimal) is 'less than’, 'equal to’, or 'greater than’ a second 4-bit word. The CD 4585B has eight comparing inputs, three outputs and three cascading inputs which allow expansion of the comparator functions in multiples of four bits. Speed of operation is typi- cally 1 80 ns at 10 V, maximum input current is 1 >iA at 1 0 V and 100 nA at 18 V. The integrated circuit is available in 16-lead plastic or hermetic ceramic dual-in-line packages or in chip form. RCA Solid State — Europe, Su n bury -on- Thames, Middlesex, TW16 7HW. (1199 M) market elektor june 1979 - UK 15 Intelligent alphanumeric display Litronix have introduced two new alphanumeric displays to their range of optoelectronic devices which incorporate no less than 17 segments per character! Is this a record? Both new devices, the DL-1414 and the DL-2416, consists of four charcters and have their own decoding, memory, multiplexing, driving and control circuitry. Additional features are computer optimised lens for magnification with minimum distortion and end-to-end stacka- bility. Their inputs are TTL compatible, as are their power supply requirements and each device contains ASCII coding to segment conversion and input static RAM buffering (no refresh). The DL-1414 display hasaO.112 in. character height, requires low power supply and can be battery operated, as a result being ideally suited for hand-held devices. The DL-2416 has a 0.16in. character height, wide viewing angle, fast access time and is a small rugged package in standard DIP form. These ‘intelligent’ displays afford considerable ease of use by connecting data inputs directly (buffered if required) connecting address inputs to address bus, using CE to expand displays, generating single write pulse and having cursor facility if required. Litronix Inc., 23 Churchgate, Hitchin, Herts, SG5 1DN. <1190 M) versatile and purpose-built packages for a host of electronic or electro-mechanical instruments. They are made up from single components - top, bottom, sides and ends - are dust and splashproof when assembled and can be supplied in beige, black or blue textured colours requiring no extra furnishing. Standard width is 212 mm, depth is 232 mm with height ranging from 62 mm to 88 mm in 6 mm increments. In addition a Mini Series is also offered with dimensions starting at 37 mm high x 1 30 mm wide x 144 mm deep. Vertical circuit board guides are fitted and options include rail and card slide adaptors, standard or special front and rear panels, RFI/EMI shielding, handles and tilt stands. OK Machine and Tool (UK) Ltd., 48a, The Avenue, Southampton, Hants. SOI 2SY (Tel. 0703 - 38966/7) will flash until manually reset). Two control outputs are provided, one operating at set point plus hysteresis and the other at set point minus hysteresis. Accuracy of temperature sensing is ± 1° C, and hysteresis can be set to various values between 0 and 8 degrees. In addition, control output and display jitter has been avoided by the inclusion of 0.05° display hysteresis. With minor changes to the pheripheral circuitry, (change of thermistor type) the AY-3-1270 can be used to measure different temperature ranges. Furthermore, the chip can be used as the heart of a 3% digit digital voltmeter with automatic zero. Power requirements are not critical, a single 9 V, 40 mA supply being all that is needed. The package is a 40 pin dual-in- line type suitable for use in ambient temperatures from -25° C to +70° C. General Instrument Microelectronics Ltd., Regency House, 1-4 Warwick Street, London W1R 5 WB, England (Tel 01-4391891) (1 198 M) is DC to 50 MHz over the same voltage range. A ‘pulse stretching’ memory retains short pulses for 50 ms, allowing reliable observation, then resets automati- cally. The probe also indicates an open circuit or high impedance condition by extinguishing both LEDs. The PRB-1 is protected from overloads up to ± 70 VDC and has an impedance of 1 20 kohms making it virtually ‘invisible’ to the circuit under test. The probe itself draws only 0.5 mA at 2.5 V with both LEDs driven, and only 15 mA on standby. OK Machine and Tool (UK) Ltd., 48a, The Avenue, Southampton, Hants, SOI 2SY. (Tel. 0703-38966/7) (1194 M) Temperature controller chip A universal digital thermometer/ controller microcircuit has been announced by General Instrument Microelectronics Limited. This device can be used in many applications including the control of home heating, cooling and air conditioning systems. This new chip, designated type AY-3-1270, will measure the temperatures found in domestic and commercial equipment and display them on LED or LDC display panels. The device will accept inputs direct from a thermistor temperature sensor, and will drive the display without interface circuitry. The AY-3-1 270 includes an on-chip power failure detector, which provides warning if power has been removed for longer than a specified time. It also provides warning should temperatures vary outside normal levels (the display Digital logic probe Until recently digital electronics servicing had to rely on the oscilloscope for logic level analysis. Although accurate and sensitive, the oscilloscope is generally expensive and not very portable. However, this new probe from OK Machine and Tool (UK) Ltd. is said to rival the best oscilloscopes in performance, is completely portable and is lower in cost than most comparable units. This pen-sized probe, the PRB-1, is powered by the circuit under test and is fully compatible with all logic ‘families’ thereby simpli- fying the task of trouble-shooting in even the most sophisticated circuits. To ensure reliable detection of logic signals the thresholds are set at 60 percent (high) and 15 percent (low) of supply voltage and the unit is permanently adjusted so that no recalibration is required. Further- more, no switch resetting or manual adjustments are needed to go from one IC family to another. The probe body is both impact, and solvent resistant. The light weight power cord is coiled for convenience, detachable, and extends to 6ft (1.8 m) if necessary, terminating in mini-alligator clips. The constant brightness LEDs are situated for maximum visibility, and a logic truth table is printed above them. The probe is sensitive enough to detect a pulse of less than 10 ns over a voltage range of 4 - 15 VDC. Frequency response Transformers A new range of transformers has been announced by Verospeed. The transformers are provided with two 120 volt primary windings, which may be connec- ted in series or parallel for 50 and 60 Hz operation, and two secon- dary windings with output voltages ranging from 0-3 V to 0-20 V rated at between 1.2 VA Pretty boxes, little boxes — No, these boxes are not made from ‘ticky-tacky’, but arc moulded in ABS. They are part of the new Pac Tec *C’ Series range of enclosures from OK Machine and Tool (UK) Limited. These cases are available in over 25 sizes and have been designed to provide and 50 VA. All transformers in the range are fitted with full shrouds and are varnish protected. Verospeed, Barton Park Industrial Estate, Eastleigh, Hampshire, SOS 5RR. (Tel 0703-618525) (1193 Ml advertisement elektor June 1979-UK21 UNBEATABLE LOW PRICES WE STOCK PARTS TO BOILD MUSICAL PROJECTS As published by leading Magazines, send large S.A.E. for lists: — DALSTON ELECTRONICS, 40A Dalston Lane, Dalston Junction, London. E8 2AZ Tel: 01 249 5624 Positive light sensitive Aerosol Lacquer Enables YOU to produce perfect printed circuits in minutes! Method; Spray cleaned board with lacquer. When dry, place positive master of required circuit on now sensitised surface. Expose to daylight, develop and etch. Any number of exact copies can of course be made from one master. Widely used in industry for prototype work. FOTOLAK £1.50 Pre-coated 1/16" Fibre-glass Developer 30 board: 204 mm x 114 mm £1.50 Ferric Chloride 40 204 mm x 228 mm £3.00 408 mm x 22 8 mm £6.00 467 mm x 305 mm £9.00 Plain Copper-clad Fibre-glass. Single-sided. Double sided. Approx. 3.18 mm thick. Sq. ft £1.25 £1.50 Aporox. 2.00 mm thick. Sq. ft £2.00 ’ ’ " C2.25 Approx. 1.00 mm thick. Sq. ft £1.50 ' £^75 Plain Copper-clad Fibre-glass. Single-sided. Double sided. Approx. 3.18 mm thick. Sq. ft £1.25 £1.50 Approx. 2.00 mm thick. Sq. ft £2.00 ’ ’ " C2.25 Approx. 1.00 mm thick. Sq. ft £1.50 " £^75 Single-sided Copper-clad paxolin. 10 sheets 245 mm x 150 mm . £2.50 Clear Acrylic Sheet for making master 12 Postage & packing 60p per order. VAT 8 % on total. G.F. Milward Electronic Components Ltd., 369, Alurr Rock Road, Birmingham B8 3DR Telephone: 021-327-2339 QUALITY ITEMS Compare performance and specification with units costing 3 times as much! DIGITAL! 24 HOUR CLOCK • silewJU Sated numerals OOC699^7^ THREE FOR £13.50 THREE FOR £20 POST 6. VAT INCLUSIVE Send cheque. P.O. ' M O. for the correct amount which includes VAT and P8. P or pay bv Access/Barclaycard. Send name/card number ( i f applicable) and address to: HENRY'S RADIO 404 EOGWARE ROAD, LONDON W2 1ED ENGLAND AVAILABLE ONLY FROM HeHR *sLo THE ELEKTOR SOFTWARE SERVICE NEW ESS 003 T.V. GAMES Id? PROGRAMMES The first disc for this project is now available and contains 5 programmes for the following games. Four-in-a-row Surround Music Box Fun and Games Clock ESS 003 Price £1.30/82.80 wraps®