So, February 1984 up-to-date electronics for lab and leisure Rs. 7.50 PLUS : audio embellisher • address decoding • weather vane • universal active filter • audio sleuth • Z80 EPROM programmer • programmable crystal oscillator • digital cassette recording • gyrophone With this unit connected to your stereo system you can produce an effect very like that of a Lesley rotating speaker system. how accurate is your watch? 'Clockwork' watches can be very accurate provided they are adjusted properly. The circuit described here quickly calculates the error in a mechanical ticker so that it can be set correctly. digital cassette recorder Cassette tape is often used as memory storage in personal computers. Unfortunately, the quality of the computer's cassette interface usually leaves a lot to be desired. The present circuit improves matters consider- ably without affecting the audio performance of the cassette recorder. audio signal embellisher A three part modular system that can increase your listening pleasure if you are forced to connect mono and stereo equipment together. universal active filter An 1C that can act as a universal active filter with a minimum of external components is certainly worth having a look at. from thermometer to thermostat Adding a single 1C and a handful of other components to the LCD ther- mometer featured in our October 1982 issue permits it to be used as a thermostat. audio sleuth at work When something goes wrong (and it often does) this article can help you find the root of the problem. wind direction indicator Many lament the passing of the weathercock, but our electronic version has at least one distinct advantage in that you no longer have to see the actual weather vane to know which direction the wind is blowing. Z80 EPROM programmer A small circuit consisting of just a few components is all that is needed to enable any Z80 system to program 2716 EPROMs in situ. home-made low-cost wiring probe Wiring prototype circuits is greatly simplified by keeping the wire tidily on a spool. address decoding One of the least understood aspects of computing is address decoding. This article is intended to throw some light onto the subject. applicator New programmable crystal oscillators in which the oscillator, dividers, and selector circuits ae housed together with the quartz crystal in a 16- pin OIL package. market missing link EPS service switchboard 2-28 2-36 2-38 2-39 2-42 2-50 2-52 2 54 2-59 2-61 2-65 2-65 2-67 2-74 A selection from next month's issue: ■ capacitance meter ■ video combiner ■ video syncbox ■ constant-light source ■ disco control unit ■ tachometer for diesels dia february 1984 2-03 PM 2502 Analogue multimeter. Every feature you need at a down-to-earth price! The PM 2502 is a rugged, compact, general-purpose multi-meter with a 25 p A tautband moving coil movement. Another advanced Philips Multimeter: Autoranging, Digital Multimeter PP9005. For further details contact: Philips India Test & Measuring Instruments Division Plot 80, Bhosari Industrial Estate PUNE 41 1026 Test & Measuring Instruments PHILIPS Philips — the trusted Indian household name for over fifty years. Ceramic capacitors from the Keltron supermarket: KELTRON COMPONENTS THE ELECTRONICS SUPERMARKET Designed to withstand all climatic vicissitudes. With high insulation resistance, low dissipation factor, good dielectric strength and operating stability. In a range of voltages for both temperature compensating and high dielectric constant types. Telex : 0884-273 KEDC Ih Ti 695 001. Tel : 60621 Branch Offices : □ 1 02-A Poonam Chambers, Dr. Annie Besant Road, Worli, Bombay 400 018. Telephone : 893457, 397448 Telex 011- 5139 Telegram : KELTRONBOM □ 75-C Park Street, Calcutta 700 016. Telephone : 245654, 213200 Telex : 021-2207 Telegram • THYRISTOR □ Sudarshan Building, 86, Chamfers Road, Madras 600 018. Telephone : 442310 Telex : 041-7632 Telegram • KELMAD ti Hemkunt Towers, 2nd floor, Nehru Place, New Delhi 110 019. Telephone : 644692. 648493 Telex : 031-3774 Telegram : KELTRON □ Centenary Building, 28, M.G. Road, Bangalore 560 001. Telephone : 564492, 564528 Telex : 0845-746 Telegram : KELTRON □ Syrian Church Road, Near Spencer Junction, Trivandrum 695 010. Telephone : 60241 Telex: 0884-283 Telegram : ELECTRONIC □ 3-4-492 First floor. Barkatpura, Hyderabad 500 027. Telephone : 63786 Telegram : KELTRON □ 67. Pritamnagar. Mangaldas Road, Ellis Bridge, Ahmedabad 380 006. Telephone : 79867, 463664 Telegram : KELTRON. TrikayaKC.il 83 elekeor ind,a lebruary 1 984 2 - 0 7 ELECTRICAL MEASURING INSTRUMENTS & POWER CONTROL EQUIPMENT (AEJ Leading manufacturer of Panel mounting & Portable Electrical Meas- uring Instruments, Moving Iron & Moving Coil type. Dynamometer type. Transducer operated type, for measurement of Voltage, Current , Power, Power Factor , Frequency. ®® DIMMERSTAT* Continuously Variable Auto-Transformers are made since 1950, smallest model is for 0.75Ampand largest is for 200 Amps. Manual & Motor operated operation. (AfpLine Voltage Stabili- zers- First made in 1965 for protecting voltage sensitive Defence equip- ment, they are now used in Industry, where a stable voltage is necessary for efficient working. Single phase models for loads from 1 KVA to 250 KVA & 3 phase models from 5 to 1000 KVA are standard products. Rectifiers for D.C. power required for electroplating & electro-chemical industry are in wide use. For Communication Equipment & Telephone exchanges, they are used as float chargers @D Instruments & Protect ive T ransformers, for use on lines upto 220 KV. 22 KV. & Higher voltage CTs & PTs are oil filled and outdoor type, Lower voltage models are avai- lable as Resin-cast for indoor use. CT/PT measu ring sets are made upto 22 KV. m A El I Automatic Electric I td 1 " RECTIFIER HOUSE ' P.O.BOX NO. 7103. BOMBAY-400 031. ' >K — J PHONE : 8829330 33 TLX. : 11-71546 Back numbers of Elektor India currently available are detailed above, with a brief description of their contents. • Send for your copies now • Price : Any one issue Rs. 7.50. • AUG/SEP Issue Rs. 1 5 /- • Price includes postage • Price subject to change without notice. • NO V.P.P. PLEASE Send payment to Elektor Electronics Private Ltd. 3 Chunam Lane, Dr. Bhadkamkar Marg, Bombay 400 007 Have you Subscribed to elektGf Rs. 75/- FOR ONE YEAR Rs. 140/- FOR TWO YEARS AND Rs. 200/- FOR THREE YEARS. Use the order card in this issue and ensure your copies. pgfCKsrnumDEBrsEruicea 2-10 The new 12" speaker is yet another addition to Luxco’s growing range of Hi Fi speakers. This speaker will complement the existing high performance Iuxco tweeters and midranges very well. Manufactured by : LUX CO Electronics. Allahabad, Telex : 540-286 Sole Se lling AgentrLUXMl & CO. 56, Johnstonganj, Allahabad 211 003 Phone : 54041 Distributors for Gujarat & South India: precious ' Electronics Corporation 3, Chunam Lane, D. Bhadkamkar Marg, Bombay 400007. Phone: 367459/369478 9. Athipattan Street, M ount R 0ae j, Madras 600 002 . Phone : 842718 Distributors for Delhi & Haryana : Railton Electronics Radio Place, Chandni Chowk, Delhi 110 006. Phone : 239944/233187 Wanted Stockists all over India dia february 1984 2 - 1 1 COMPUTERS AND THE C.A. Inaugurating a seminar on ‘expo- sure of computers to chartered accountants’, organised recently by the Computer Society of India (CSI), Delhi Chapter as part of their Computer Day Celebrations, the Union Minister for Steel and Mines. Mr. N.K.P. Salve, called upon chartered accountants to increasingly utilise computers in their work as auditors, tax execu- tives or finance managers; he cau- tioned them, however, to make the machines work for them and not become their slaves. He also dwelt upon the advantages the compu- ters could provide in inventory control and management. Mr. N. Kohli, CSI secretary, obser- ved that computers would have a major impact on the working of finance professionals, adding that computer facilities were now avai- lable at a considerably lower cost. INDO-FRENCH TIE-UP By an agreement initialled at New Delhi, following a three-day meet- ing of Indo-French working groups on computer hardware, software and informations, India and France have agreed to colla- borate in a host of frontline areas of computer technology and elec- tronics. The main feature of the agreement is the vertical transfer of technology by France to India, covering super mini computers, computer peripherals, packaging technologies and sophisticated components, with buy-back arran- gements in certain peripherals like floppy disc drives. Other areas covered by the agreement are joint development of a new multipro- cessor architecture and super- micro-computer based on SM90, setting up of institution based on computerised instruction for train- ing in software, computer net- works and telematics, besides fifth generation computers. The Indian side was led by Dr. N. Seshagiri, director (computer development), Department of Elec- tronics and the French side by Mr. Christian Stoffaes, additional secretary, department of electro- nics and informatics in the French Ministry and Research. ELECTRONICS COMPLEX IN GUJARAT Gujarat State has ambitious plans to develop the electronic industry in the state and has offered seve- ral inducements to the industry to set up units in Gujarat Industrial Development Corporation's (GIDC) exclusive industrial estate for electronics industry at Gandhi- nagar, 22 kms away from Ahmeda- bad. A few of the incentives are: cash subsidy up to Rs. 25 lakhs; sales tax exemption up to Rs. 80 lakhs or sales tax deferment up to Rs. 50 lakhs; a quasi equity, inter- est free loan ranging from 10-20% of the fixed capital investment or Rs. 7.5 crores (whichever is more) to large electronic units under LEEP. The Governemnt has al- ready reduced sales tax on elec- tronic components and end pro- ducts manufactured in the state. The Industrial Training Institute at Gandhinagar is conducting special courses in electronics, so as to provide the industry with a skilled workforce. MICRO-COMPUTER PROJECT IN A.P. A micro-computer project costing Rs. 12.50 lakhs will be set up by the Integrated Data Systems Pvt. Ltd., a company run by techno- crats, and assisted by the Andhra Pradesh Electronic Deve- lopment Corporation (APEDC), for which a promotional agreement has been signed. Products to be manufactured, like video terminals, data entry systems and word pro- cessors of 8 and 16 bit range, find application in business, banking, reservation and process control. The unit is expected to go into commercial production within three months. Bright future for ELECTRONICS Inaugurating a seminar at Bombay on Growth of Electronic Compo- nents' organised by the Institution of Engineers with the Electronic Component Industries Association (ELCINA) and the Institute of Electronic and Telecommunication Engineers (ETE), Dr. P.P. Gupta. Electronics Secretary, said that he expected the output of electronic goods to touch Rs. 10,000 crores by 1990, of which the require- ments of components would be in the order of Rs. 2000 crores. He pointed out that the electronics sector was open to all industries in that there were no restrictions on FERA, MRTP, joint or private sector companies. He added that the government is promoting the concept of independent test labo- ratories where comprehensive facilities are being established; these include various combina- tions of environmental severities, vibration, shock and other mecha- nical endurance test facilities with automated test equipment. COMPUTER SYSTEM INSTALLED A computer system has been in- stalled in the Ministry of Wroks & Housing by the National Informa- tion Centre of the Electronics Commission, to facilitate its work and of its constituent departments and organisations. It is expected to streamline the working of the estate office, Central PWD and other units of the ministry hand- ling data on a large scale. The National Information Centre (NIC) has installed an HP-1000 computer which has stand-along capabili- ties. It is also connected to the computer cyber 170/720 system both through the P & T Lines and the VHF radio link, whereby the Ministry would be able to tap con- siderable computer resources from its location in Nirman Bhavan.The Ministry has also initiated a series of training courses for program- mes organised by the NIC in colla- boration with the National Build- ings Organisation. POLYSILICON PLANT The Union Government has finali- sed agreements for the purchase of foreign know-how for setting up a plant to manufacture polysilicon, the substance used in the produc- tion of photo-voltaic cells and in- tegrated circuits popularly called chips'. It is learnt that a negoti- ation committee constituted for working out technical collabora- tion agreements for the polysilicon plant and confirming it for invest- ment approval, had had detailed discussions with leading polysili- con producers before finalising the agreements. ELECTRONIC COMPLEX ll>4 W.B. The Government of West Bengal has initiated plans to develop the electronics industry in a big way. The committee on electronics, headed by the chief minister, Mr. Jyoti Basu, held its first meet- ing recently at Calcutta, when a number of points on the develop- ment of the industry was discus- sed. The agenda of the meeting also included the clearance for transferring 93 acres of land at Salt Lake to individual parties who will collectively form the electro- nics complex. Also discussed at the meeting was the perspective plan for development of electro- nics in the seventh plan. With a view to decentralising the growth of the industry in the state, offi- cials and experts will visit Darjee- ling, Jalpaiguri, Bankura and Dur- gapur to identify suitable loca- tions. ... to make your stereo wander Figure 1 . Block schematic of the gyrophone. The signal and control paths (the latter in dashed lines) are shown separately to clarify the operation of the gyrophone. Most of us have heard the stereo effect of an express train, a gale force wind, or perhaps an artificially created sound transferring from the right-hand to the left-hand speaker. It's just as impressive when the sound returns from the left- hand to the right-hand speaker, as when, for instance, a train from the opposite direction passes by. The circuit described in this article makes it possible for both effects to happen simultaneously: creating a sound very much like that of a Lesley rotating speaker system. gyrophone... Before we go any further, there is one thing to be borne in mind: the contents of the two stereo channels must be quite distinct from one another if the effect is to be realized. A short listening test will soon show which type of recording is suitable: listen to it and then turn one of the speakers off. If half of the sound just ‘dies’, the recording is usable. Stereo records produced ten years or more ago are particularly suitable. The circuit is not really an electronic version of a Lesley because phase shifts are not catered for, but its action is none the less remarkable. Briefly, the right-hand signal ‘wanders’ to the left-hand channel, and vice versa. Shortly afterwards, the two sounds revert to their original channel. This effect is achieved by periodically inverting the two channels. The block diagram in figure 1 shows that the signals from the two channels are split and applied to four operational transconduc- tance amplifiers (OTAs). However, although both OTA1 and OTA3 are fed with the left- hand signal (and OTA2 and OTA4 with the right-hand signal), they are not controlled by the same sawtooth voltage. The low- frequency oscillator (LFO) drives OTAs 1 and 4 directly and OTAs 2 and 3 via ,.n inverter. This means that OTAs fed with the same stereo signal have opposing control signals. The left-hand information is there- fore amplified in OTA1 but attenuated in OTA3 and consequently appears in the left-hand but not in the right-hand output. From time to time, however, the control signals are such that the left-hand infor- mation appears in the right-hand but not in the left-hand output. The right-hand input signal is treated in an identical manner. The whole process is continuous and therefore causes the characteristic swelling and fading of the loudspeaker outputs. In contrast to a real Lesley, our circuit creates the effect only by differences in volume in each individual channel. A low-frequency oscillator consisting of integrator A1 and trigger A2 (see figure 2) generates a sawtooth voltage. This voltage should not go negative because that would block the OTAs, and a diode, Dl, is there- fore included in the feedback path of A2. The sawtooth voltage is fed to A3 and to inverter IC2. The output of IC2 is applied to the inverting input of A4. Opamps A3 and A4 drive transistors T1 and T2 and these in turn feed the four OTAs. As already explained, the signals from the two channels are split and the parts are amplified in different OTAs. Output channel L contains a mixture of the signals 1984 2-14 2 gyrophone . . . fed to OTAs 1 and 2 and, similarly, output channel R a mixture from OTAs 3 and 4. The mixing elements are formed by two resistors and a capacitor (for instance, R27/R28/C2). The buffers contained in IC3 and IC4 (pins 7, 8 and 9, 10 respectively) must not be used in this application. Construction and calibration The design has been kept as simple and inexpensive as possible and its construction on a prototyping (Vero)board should not present any trouble to the hobbyist with some experience. Preset PI enables the frequency of the sawtooth generator to be set to your own individual taste. The frequency, f, is given by f = 1/[C1(P1 + Rl)]Hz. With values shown, the frequency can be set anywhere between 0.2 Hz and 4 Hz, corresponding to periods of 5 s and 250 ms respectively. Because IC2 inverts the sawtooth waveform, its output would normally be mostly nega- tive. As stated, this cannot be tolerated as it would block the OTAs. Therefore, the inverted sawtooth voltage is superimposed on a d.c. voltage, the level of which is preset by P2. If an oscilloscope is not available, the presetting can be done by ear. Apply a signal to one of the input channels and set the LFO to a low-frequency output. If P2 has been set correctly, the loudspeaker volume should gradually fade away and then grad- ually swell again. If not, limiting is taking place and this is indicated by an absence of sound for some time followed by a sudden burst of volume. The audio input signals to the circuit may lie between 0.7 V and 10 V. However, when you use inputs of just about 0.7 V and have a powerful main amplifier connected to the output of the gyrophone, it may happen that the maximum and minimum values of the sawtooth voltage become audible in the loudspeakers. This can be prevented by increasing the signal input by, for instance, inserting an additional amplifier between the signal source and the inputs to the gyrophone. We shall be very brief about the required mains supply: the current consumption of the gyrophone is around 50 mA per channel at ± 15 V. M i lebruary 19842- 1 5 Even though quartz watches seem to have alrr\ost completely sup- planted their mechanical counterparts, for many people there is still nothing to compare with the fine mechanical craftsmanship that goes into a clock- work watch. That regular tick, corping from so many carefully made parts, tirelessly assembled to make one whole unit, is some- thing completely different from the invisible, silent shuffling of electrons in a quartz controlled watch. The 'watch tester' described in this article is a crystal controlled circuit that is used as an aid to set a mechanical watch accurately. A crystal is used as a reference to determine, within a few seconds, how much time the watch gains or loses, and this is shown on a display as a certain number of minutes per day. Knowing the error is essential to be Sble to set the watch accurately. how accurate is Y our watch? quartz precision for mechanical watches Man has always tried to measure time in one way or another. Sundials, water clocks, oil lamps, candles and hour glasses are just some of the things that have been used to measure time down through the ages. Then came the mechanical clock. Nobody knows for certain exactly when this first came into existence but they have been made at least since the fourteenth century. Since then, mechanical clocks have been consistently improved and Watches have been made since about the end of the fifteenth century, but it took a long time before the ‘portable clock’ was im- proved enough so that it worked reasonably accurately. The best clocks in the seven- teenth century had an error of about a minute per day. With an average watch an error of a quarter of an hour a week could be expected. Until the beginning of this century watches were normally carried on a chain and it was only around the year 1900 that somebody came up with the idea of a wrist watch. Since then watches developed very quickly. In 1924 the automatic wrist watch arrived and after the second World war the ‘electric’ watch. In 1957 a watch appeared on the • market that used an electromagnetic system to drive the balance weight. Four years later the firm of Bulova produced a much more interesting idea, using an electronically driven tuning fork instead of the balance weight. This tuning fork watch was guaran- teed to be accurate to within one minute 1 2- 1 6 1 The modern watch is the final stage (so far) and uses a quartz crystal as the time base. The accuracy of this design is such that the error per year is neglegible. A mechanical watch always has much more charm than its ‘cold’ electronic counterpart. It is a testament to the skill of the craftsman who made it, and this alone is a great point in its favour. Clockwork watches do have one undeniable advantage, of course: they have no batteries to fail at the most unex- pected and inconvenient moment. There are, of course, still a lot of mechanical watches in circulation and several firms currently sell clockwork watches at the ‘expensive’ end of the market. Mechanical ‘tickers', it seems, are always in fashion. Adjusting a mechanical watch is a lengthy process because changing the effective length of the balance spring does not give an immediately noticeable change. A good watchmaker, certainly, has expensive equip- ment that can measure the error fairly quickly, but anybody else simply could not afford one. With the watch meter here anybody can quickly adjust almost any clockwork watch accurately. The block diagram This circuit uses an optical pick-up. An acoustic pick-up should also be possible but in practice that seemed to be more susceptible to problems with ambient noise. With this optical pick-up we use a small lamp to shine light on the spokes of the balance wheel and the reflections are received by a photo transistor. The pulses given by the photo transistor are processed and compared with a ‘standard’ frequency, and the error is then shown on a display. The block diagram of figure 1 is a bit more complex than our usual circuits, but this simply makes the circuit easier to under- stand. The photo transistor pulses are converted to ‘proper’ digital signals in- the Figure 1. The block diegrem of the arcuit. The pulses picked up et the belence wheel of the wetch elektor india februaiy 1984 2 - 1 7 first block. These pulses then go to a monostable multivibrator. The monostable time can be set to three different values with switch Sla. These values are < 400 ms, < 333 ms and < 200 ms, and they require a short explanation. Almost every mechanical watch falls into one of two standard tick frequencies, namely 18000 ticks per hour (= 5 ticks per second) or 21600 ticks per hour (= 6 ticks per second). The first generally applies to older watches. There are also some clocks that beat with 36000 ticks per hour (10 ticks/s). One complete swing of the balance (from the middle to one side, back to the other side and to the middle again) consists of two ticks. Five ticks then consist of 2.5 swings. Because we want to measure swing times with this circuit the MMV time must be chosen so that only every second tick is registered. In other words the MMV the error in minutes per day. If a time of how accurate is your time must be about 5 . . . 10% less than 20 seconds is used the counter must count watch? the time for two ticks. For 5 ticks per 14400 clock pulses. This means that the second the MMV time must be relative to clock frequency for the counter must be 2 x 200 ms = 400 ms. This drops to 333 ms 1440/2 (or 1400/20) = 720 Hz. This refer- for 6 ticks and 200 ms for 10 ticks. ence frequency is supplied by a crystal and The MMV is followed by a divider that, a few dividers. depending on the position of SI, divides by With a measuring time of 2 seconds the 5, 6 or 10. A signal with a period of preset value of the counter must be - 1440 2 seconds now appears at the wiper of Sib so that the count is exactly zero if the watch (provided that SI is in the correct position is running correctly. The counter can actu- for the watch under test). If the period is ally only count from -99 to +99, so a preset not 2 seconds, this means that the watch is value of -1440 is impossible. Because the not keeping time. A period of less than read out only shows two figures, we set the 2 seconds means that the watch is running preset to -40 (the last two digits of -1440). fast, and more than 2 seconds means it is The counter will then be at zero after two running slow. seconds. This ‘trick’ works here because a This signal then goes to switch S2a which normal watch will never have an error of enables us to select the 2 second signal or more than 99 minutes a day. The counter one ten times as long. The 20 second signal starts by counting from -40 to zero then 'contains' a greater number of ticks and is from zero to 99 and six times from -99 to therefore better than the shorter time for +99 and finally from -99 to zero making measuring the error of a watch. The signal 1440. Note that there is a delay of one clock chosen with S2a then goes to MMV3 and cycle every time the count crosses zero on MMV4, which drive the counter and the its ‘jump’ from +99 to -99. Without this our latch. The latch with a seven segment arithmetic would not be correct. If 20 decoder is driven by a pulse supplied by seconds is used as the measuring time the MMV3, while MMV4 presets the counter counter is preset to zero (the last two digits after the count has been stored in the latch of 14400). (and shown on the display). In practice the counter cannot itself work Finally, the counter. Because we want the out if its count is positive or negative, so the display to show the error in minutes per day, '+’ or sign is stored by a flip-flop. This the counter has to be a bit special. It must flips (or flops) every time the counter is at be able to count positively and negatively as zero, and drives the ± sign in the display, we can have an error in either direction. The Finally there is a reset circuit whereby all clock frequency of the counter must be counters can be reset simply by pressing one carefully chosen to enable the read out to be button. The circuit is then ready to begin in minutes per day. Furthermore the counter measuring anew, must be capable of being preset, so that its output is exactly zero if the watch is work- ing accurately. To enable all this to be done, The practical layout an eight-bit BCD up/down counter is used. As we have spent quite a long time talking Now to the clock frequency. There are about the block diagram, we do not really 1440 minutes in a day (except Monday, need to say much about the actual circuit which has at least twice as many). If a diagram of figure 2. The block diagram also measuring time of two seconds is used, the simplifies matters by stating which corn- counter must receive 1440 clock pulses in ponents make up each block, these two seconds. The error measured by We will have a look at the input stage the counter relative to this 1440 is then separately. The d.c. voltage setting of photo > 2-19 Figure 3. This is the printed circuit boerd transistor T2 is handled by FET Tl. For low voltage goes via voltage divider R9/R10 to design for the measuring | f requenc i es an d d.c., Tl acts as a voltage IC1 where it acts as the trigger-level setting sect.on of the circuit. j source; its drain voltage is then fed back to for this schraitt trigger. The other input of | the gate via R2. The low-pass filter con- the schmitt trigger is fed the voltage changes sisting of R3 and Cl ensures that Tl acts from the photo transistor via C3. This set- as a current source at higher frequencies. up allows the circuit to adapt itself to the Slow variations in the light picked up strength of the input signal. If the photo (from ambient conditions for example) transistor provides a strong input signal then are therefore compensated by the FET, the triggering threshold is high. The strength while fast changes in light cause a large of the input signal is indicated by the meter change in the voltage on the collector of connected parallel to C4. If switch S4 is the photo transistor. This is exactly what closed the output of IC1 is heard through we need to detect the moving spokes of the the buzzer. An LED, D5, at the Q output of balance wheel. These voltage changes are FF1 flashes in time with the tick pulses. The transmitted via C2 to T3 where the pulses measuring time is shown by means of LED are rectified. The voltage on C4 is the same D6 at the output of FF4. as the maximum value of the pulses. This The supply for the whole circuit is handled elektof india lebruary 1 984 2 - 20 by the same 7812 regulator IC. The current consumption is about 250 mA. Constructing the circuit The circuit has been divided between two printed circuit boards that are shown in figures 3 and 4. The ‘measuring’ section is located on the board shown in figure 3 and contains all the components shown in the left half of the circuit diagram, with the exception of R21 and D5. The second board consists of two sections which may be separated if desired. These are the counter section and the read-out (the right half of the circuit diagram with the exception of FF4). The numbered points on the two boards must be connected to each other. The supply for the display must be taken from points 1 and 2. Trying to tap a supply from anywhere else will probably cause problems. It is quite possible that the BS 250 FET may prove difficult for some people to get their hands on. If this is the case, a BC 516 may be substituted for Tl, but R3 must then be 3M9. Fortunately this transistor can be fitted to the board exactly the same as the FET. When all the electronics is assembled we can turn our attention to building the sensor. The photo transistor and the lamp are mounted next to each other, but in such a way that the light from the bulb does not circuit board for the read-out. which can, if enable the display to be mounted away from the r India february 1 984 2-21 fall directly on the photo transistor. This is easily done with a piece of black paper between the two. The emitter of the transis- tor can now be soldered directly to the collar of the lamp. This leaves three connec- tions which can be linked to the printed circuit board with a piece of screened stereo cable. The collar of the lamp (which can be a miniature type) must be connected to the screen. This unit can then be fitted into something like a big felt tip pen. A clip can be made up to hold this 'pen' steady during a measurement. The photos and the front cover show how our prototype was built. A nicer (but also more expensive) possibility is to use a reflection sensor, such as the OPB730, which contains a LED and a photodarlington. If this is done the sensor must be well screened from ambient light, and the value of resistor R1 must be in- creased to 560 n. Adjustment and use Adjustment is very easy. The frequency of the crystal can be set to the exact value required with trimmer C8. To do this a frequency meter with a maximum error of 0.005% is needed. A frequency of 115200 Hz must be measured at test point TP. If you cannot get hold of a good frequency meter then simply put C8 in mid position. In most cases the frequency will then be reasonably accurate. Next, MMV1 must be set, preferably with an oscilloscope. Potentiometer PI is set so that the monostable time is 360 . . . 380 ms with Sla in position A. If you do not have an oscilloscope, this MMV can also be adjusted with the aid of a watch that is known to be accurate. Place the watch under the sensor and turn the sensor until the meter shows a strong signal and the buzzer ticks regularly. Turn the preset to maximum, set switch S2 to position A (2 s measuring time) and adjust the preset by turning it backwards a little at a time. After each adjustment wait until the measuring time has passed and see what the read-out shows. At some stage an error of about zero minutes will be dis- played. Turn the preset a little bit further and then leave it at that. A few words about using this circuit will certainly not go astray. First we must know the tick frequency of the watch to be tested. Older gents watches generally have 5 ticks per second, whereas modem gents watches and ladies watches usually have 6. After a bit of practice this can even be heard from the ticking of the watch. Lay the watch under the sensor and point the photo transistor towards the spokes of the balance wheel. Move the watch carefully until the meter reading is as large as possible. If S4 is closed the pulses from the phototransistor can be heard from the buzzer. This should be a regular tick. If it sounds more like ‘sawing’ then the transistor is pointing at the adjusting screws and must be moved slightly. The COUNT LED, D5, should flash regularly to show that the circuit is receiving the pulses. The correct ticking frequency (5, 6, or 10 ticks per second) must be set with SI . A measuring time of 2 seconds is selected using S2. Press the RESET and after 2 seconds LED D6 (GATE TIME) ‘changes’. What we mean is that the LED lights if it was out and it goes out if it was lit. The display now shows the error in minutes per day. Whenever D6 changes the measurement has been taken and the result is shown on the display. If the error of the watch is less than ten minutes, S5 can be moved to position B (20 s measuring time). First press the RESET again and after 20 seconds LED D3 changes and the error is shown on the dis- play in tenths of minutes. With a pocket watch the photo transistor can also be focused on the balance screws and this usually gives good results. In this case, however, it is important to reduce the level of ambient light as much as possible. Incandescent lamps and fluorescent tubes in particular can cause problems. A period counter could also be used in the circuit in place of the counter section and read-out. It is simply connected to the wiper of switch S2a. However, IC2, IC7, XI, C7, C8, C9, C13, R15, R16 and R18 can then be removed and point 4 of the measuring board and pin 1 of IC3 must be connected to earth. The read-out on the meter will not, of course, be in minutes per day any more. It is a simple matter to convert the output to minutes per day using the formula 60 x 24 x (2 - T)/T, where T is the period measured in seconds. If T is 1.986 seconds the error of the watch is 60 x 24 x (2 - 1.986J/1.986 = +10 minutes per day. M incredible accuracy con- tick nearly a half million times per day A mechanical chronometer A = 5 ticks/s B - 6 ticks/s C - 10 ticks/s : A = 2 second B = 20 second 1984 2-22 cassette recorder Cassette recordings are still the most popular memory for home computers because they offer the cheapest method available. Unfortunately, it is not the most reliable method because a cassette recorder is, after all, intended for processing audio rather than digital signals. The present circuit converts a normal cassette recorder into a digital one with vastly improved data transfer capability without the loss of the audio facility. Most home computers have a cassette recorder interface which usually obeys a simple rule: the cheaper and simpler the computer, the worse the data transfer to the recorder. This only becomes evident, of course, when you ‘read’ a newly loaded program and find that all is not what it’s supposed to be. Why is that? And can anything be done about it? In most computers, a signal is delivered to the interface which is not really suitable for an audio cassette recorder. The ampli- tude of the signal is normally limited to prevent the overloading of the recorder, while a transfer speed is chosen which, according to the computer manufacturers, is ‘safe’. In other words, the computer is adapted to the cassette recorder without too much thought to the fact that the recorder was designed for a different purpose. We have tackled the problem from the opposite direction by matching the recorder to the computer. A ‘read’ (playback) and a ‘write’ (recording) amplifier are added which improves the data transfer to the extent that baud rates of 4800 may be used! When you consider that the baud rate in most, if not all, home computers cannot exceed a three- figure number, you realize what a consider- able improvement our circuit offers. Analogue and digital recording The (analogue) recording of audio signals onto magnetic tape requires special circuits to ensure that the playback signal is a faithful reproduction of the original. After all, Dolby and DBX did not come about by accident! One of the important design considerations, for instance, is to prevent saturation of the magnetic tape (as satu- ration would cause distortion). A square-wave pulse, as generated by most computers, consists of a large number of sinusoidal voltages. As the recording/play- back amplifier of a recorder is optimized for audio signals, it will suppress a number of constituents of such a pulse. The result is that what's recorded is no longer a square- wave signal. Further disintegration of the pulse takes place during playback, there is the tape noise, and . . . The consequence of it all is that the Schmitt trigger normally found in the input stages of a cassette interface is not presented with one proper pulse, but several distorted ones. . . . ensures your bits stay on the tape etektor india february 1 984 2-23 The process in a digital recorder is much simpler: the magnetic tape is driven into saturation. This is, without any doubt, the best method for recording data onto tape, particularly if the tape is positively magnet- ized during logic ‘high’ signals and negatively during ‘low’ signals. Before we analyze the circuit diagram, a reassurance about the cassette recorder: it needs only one modification. The screened cable to the tape head needs to be cut and the digital read/write amplifier inserted between the cut ends as shown in figure 2. The audio recording/playback amplifier is not touched at all so that the recorder re- mains fully usable for normal audio oper- The circuit The write/read (recording/playback) ampli- fier consists of two functional units separ- ated by the switch-over unit (see figure 1). The read amplifier is constructed in two parts to which we’ll come back in the circuit description. Other items shown in figure 1 are the write and read indicator LEDs. Write (recording) amplifier As explained in ‘switching’ below, we'll assume that ESI and ES2 (see figure 2) are closed and that contacts Rel and Re2 are open. The square-wave pulses from the computer are applied across preset PI and from there fed to the inverting input of opamp 1C1 via R1 and Cl. Diodes D1 and D2 limit the signal to ± 0.7 V. The gain of IC1 is fixed at about 100 by voltage divider R2/R3. Anti-parallel connected diodes D3 and D4 in the feedback loop limit the output of the opamp to ± 0.7 V. Plus or minus? you may ask. Surely the supply voltage is +12 V only? True, but the non-inverting input of IC1 does not lie at earth potential but at +6 V because of voltage divider R12/R13. The signal output of IC1 is therefore super- imposed onto +6 V. This’ arrangement is also used in other parts of the circuit. Figure 3 shows how a sinusoidal (FSK) input signal is converted by this method: the frequency remains unchanged, but the wave- form becomes rectangular. You can well imagine that if a sine wave is so converted, a distorted rectangular pulse will certainly be fully resorted to its original shape. We have taken an FSK signal as an example because that shows the operation of the circuit most clearly. In general, our digital recorder is not required with computers which have an FSK output, but as this example shows: you never know . . . The rectangular output of IC1 is inverted again by trigger A1 and increased to the maximum possible level of 12 V pp (wave shape 4, figure 3). The output of A 1 is split: one part is applied to terminal 'A' of the tape head via R32 and ESI ; a second is again inverted by trigger A2 and then fed to the earth terminal ‘B’ of the tape head via R33 and ES2. The signal at the tape head is therefore the difference in outputs of the two opamps A1 and A2: note that the tape head is not connected to earth. This method not only saves some coupling capacitors (which might distort the signal slightly) but, what’s far more important, the tape magnetization for a logic low signal is the opposite of that for a high signal. A third part of the output of A 1 is applied to the electronic switching circuit via C3. This circuit consists of electronic switches ESI and ES2, relays Rel and Re2, diodes D7 and D8, and a few resistors and capaci- tors. Itebruary 1984 2-24 The non-inverting input of comparator A3 is at a level of about +6 V via voltage divider R12/R13. Under no-signal conditions, the inverting input is at about +4.4 V via voltage divider R30/R31. The output of A3 is therefore at +12 V and relays Rel and Re2 are actuated. The voltage at the inverting input also exists at the inputs of electronic switches ESI and ES2, but is not sufficient to close the switches: a voltage close to the supply voltage is required to do that. Sum- marizing: under no-signal conditions, ESI and ES2 are open and the contacts of Rel and Re2 closed. The circuit is then in the ‘read’ condition. When a signal arrives from the computer, the output of A 1 is applied to the control inputs of ESI and ES2, and to the inverting input of A3 via C3 and D7. The output of A3 goes low, the relays open, and ESI and ES2 close. The circuit is then in the 'write' condition. Capacitor C4 charges and continues to do so as long as there is a signal coming in from the computer. As the input current of A3, ESI, and ES2 is very small, the charge on C4 is sufficient to keep the switching circuits in the same state even during the pauses between the pulses. When the computer signal ceases, C4 discharges through RIO and the circuit reverts to the ‘read’ condition. Read (playback) amplifier In the ‘read’ condition, Re2 connects the earth terminal of the tape head to the circuit earth (0 V). The tape signal is connected via Rel to the gate of FET T4. This small- signal amplifier is followed by a second consisting of T1 and T2, and a third formed by IC2. To ensure that the maximum signal is available at the output of IC2, its input is ‘raised’ to about 6 V, derived from the voltage divider R12/R13. The total gain of the three stages is around 80 dB, of which half is contributed by IC2. This is ample for many computers and the output of IC2 is therefore available at terminal ‘AN’. The output level can be matched to the com- puter input requirement by preset P3. For those situations where more gain is required, a fourth amplifier, A4, has been provided. The gain of this amplifier can be Figure 2. The new amplifie consists of three parts: a recording (write) and playback (read) amplifier and a switching circuit which separates the two elektot india tebrua ,2-25 -p© condition. It is possible that it continues to light faintly during the ‘read’ condition; if you find this disturbing, the only solution is to replace Dll by a cheaper LED (giving less light). Then there is LED D12. This diode lights during the ‘read’ condition. Capacitor Cl 2 keeps T3 conducting so that this does not switch on and off in time with the input signal. Resistor R25 prevents the indicator circuit affecting the output signal. Finally, diode DIO. This component appears to be located in a somewhat strange pos- ition, but a good look at the circuit will show that it functions as a protection diode for relays Rel and Re2. Construction and calibration Assembling the printed circuit board should not present any difficulties: figure 4 and the parts list give all the information re- quired. One point needs watching, however. Although we are dealing with a double-sided board, the two points ‘B’ must be connected by means of a short length of screened cable. The reason for this is that during 'read' operation the signal from the tape set between 17 dB and 37 dB by preset P2. As A4 is driven into saturation, its output is virtually identical with signal 4 in figure 3. The output is raised to TTL-level via voltage divider R26/D13/D14 and made available at terminal ‘DIG’. To avoid confusion, some aspects of the circuit have been ignored so far. To start with: LED Dll. This lights when the output of A3 is low, that is, during the ‘write’ head is very small (remember the 80 dB gain!). For the same reason, the screened cable between ‘A’ and the head must be kept as short as possible. In contrast to audio circuits, there is no central earth point here, so that the earths at both sides of the cable must be connected with one another. The circuit is very simple to set up. The correct positions of PI ... P3 are dependent upon the type of computer and on the baud rate. If you start at the centre position of these presets and have checked that the d.c. levels shown in the circuit diagram are 2-26 ® • •• • •• • • t V* • •am* • • • • •!••• • • • •• • • • • • • • • • • • • •• • • • • • • Mil • ••• • • • • Milan A • • • • • • |M • • • • •• • • #• *# * W 1*11 • • •Ml • • * • • • • V. • . > • • V iniii # • • iiaiiai. iiaiiii ® - 100 u/16 V electrolytic C9 - 1 ji/6 V electrolytic C11 - 47 (i/16 V Semiconductors: D1 . . . DIO- 1N4148 D1 1.D12- LED D13 = zener diode 2V 7, 400 mW D14 = zener diode 4 400 mW T1 = BF 494 T2.T3 = BC 547B T4 = BF 256C IC1.IC2 = LF 356 IC3 = TL 084 IC4 = 4066B ie 4V7. OK (no-signal conditions), the right settings should soon be apparent. Final tip: load a not-too-small memory region of the tape with a fixed hex-value and program a loop. It is then possible with the aid of an oscilloscope to check the conversion of the signal (with reference to figure 3) at various test points. During ‘write’ operation simply run the tape with this fixed hex-value. It is, by the way, not necessary to press the ‘record’ button during ‘write’ operations to erase any material already present on the tape because the signal now fed to the tape head is considerably stronger than the previous recording. Current consumption of the circuit is around 50 mA and it may therefore just be possible to draw this from the existing recorder power supply. M audio signal embellisher from an idea by J.F. Brange signal restoration with stereo simulation It is often unavoidable to have to connect an item of mono equipment that is rather less than hi-fi to a modern stereo installation. Although this may give some improvement in the resulting sound quality, the reproduction remains monaural (mono) invariably with a level of hum and noise which by present-day standards is unacceptable. We have designed a circuit which by hum suppression, stereo simulation, and dynamic noise limiting (DNL) gives a greatly enhanced performance. The stereo effect is created by splitting the audio spectrum into sixteen frequency bands which are fed alternately to the left and right-hand channels. Ever since the arrival of hi-fi audio equip- ment and the introduction of stereo, our aural senses have been spoilt to the point of addiction. Nowadays when we listen to ordinary monaural music, we soon feel there’s something missing. If in ad- dition the sound is accompanied by hum and noise, this feeling soon becomes one of disappointment or even annoyance. However, sometimes there is no alternative to the poor sound source, if only for the simple reason that we don’t want to throw away perfectly good equipment. This could, for instance, take the form of simple cassette recorders, AM receivers, sound projectors, and TV sets or video recorders. The last three are particularly prone to being neglected by audio designers. While the picture quality is praised (often deservedly so) as hi-bri (high brilliance), more often than not the sound is a disgrace by modern standards. Spatial sound We are aware of depth in sound because we have two ears. As the sound waves reach each ear at a slightly different time and with a slightly different amplitude, the brain receives two separate signals. It is able to deduce the relative position of the sound source from the differences: our ears form a true stereo receiver! The shape of the ear also plays a role: if you want to know more about this, we refer you to 'our remarkable sense of pitch' in the May 1979 (UK) issue of Elektor. What can we do with a mono sound? It is impossible to convert it into true stereo, because the subtle differences between the left and right-hand channels just cannot be added afterwards. What we can do is to create artificial differences by splitting the sound into a number of frequency bands and then feed these selectively to the left or right-hand channel of the stereo instal- lation. This is, by the way, the method k!or india february 1 984 2 “ 2 8 1 used in the TDA3810 stereo-IC featured in 'pseudo stereo' in our November 1983 issue. The present design is rather more radical and effective: the audio spectrum is split into sixteen bands by means of active filters. If the filter outputs are num- bered 1 ... 1 6 in order of ascending centre frequency, all odd-numbered frequency bands are fed to the left-hand channel, and all the even ones to the right-hand channel. The result is truly remarkable: the sound, which at first seemed to come from between the speakers, now seems to 'hang in space’ around the speakers. The block schematic The block schematic in figure 1 clearly shows that the design consists of three distinct main parts: each of these is housed on a separate printed-circuit board. The input of the circuit is a pre-amplifier (with variable sensitivity), followed by a 100 Hz and a 50 Hz band -stop filter (some- times called a 'notch' filter). These filters respectively reject the 100 Hz fundamental frequency of a double-phase rectified voltage and the 50 Hz fundamental of a single-phase rectified voltage. Both filters can be switched out. The next element is a level indicator which is useful when the input sensitivity is set. Nothing sophisticated, just a simple ampli- fier and LED which blinks away quietly when the sensitivity is set correctly. Next, we come to the heart of the design: the sixteen active band-pass filters. The outputs of the odd-numbered filters, and those of the even-numbered ones, are separately combined and are then, in prin- ciple, suitable for processing in a stereo installation. We have, however, added dynamic noise limiting (DNL) stages which, if required, can be switched off or be omitted altogether. Some of you may even use this part of the design only. The circuit diagrams There is a circuit diagram for each of the three mains parts of the design: the pre- amplifier, band-stop filters, and power supply (figure 2), the sixteen-element active band-pass filter (figure 3), and the DNL stages (figure 7). The pre-amplifier, band-stop filter*, and power supply The input sensitivity is preset by means of PI. Pre-amplifier A1 has a gain of about 10 dB and is followed by active band-stop filters A2 (100 Hz) and A3 (50 Hz). The output of A3 is fed to the band-pass filters on the second printed-circuit board (see figure 3), and also to the level indicator stage. After amplification in A4, the signal is applied to the base of T1 via C13. When it exceeds a certain level, T1 conducts to light LED Dl. The power supply for the entire design consists of the customary mains transformer, bridge rectifier, voltage regulators, and smoothing capacitors. The output is sym- metrical: ± 12 V at 85 mA. The band-pass filters The sixteen band-pass filters (see figure 3) are identical in construction. The basic diagram of one of them is shown in figure 4: a common filter circuit with an opamp as the active element and RC combinations to give the required frequency response and Q factor. As you can see from the formulas elector India lebruary 1 984 2 - 2 9 2 OHh BC ) 547B | 12 vQ !> in figure 4, if a fixed value is chosen for R1 and R2, the centre frequency becomes inversely proportional with the value of capacitance C. By appropriate values of C in the sixteen filters, the centre frequencies are varied, but the Q factor and gain A 0 , remain the same. The DNL stages For those of you who are not completely familiar with the operation of a dynamic noise limiter, here is a short description. The simplest noise limiter is a low-pass filter. Unfortunately, its action is somewhat radical and affects the audio signal. A dynamic noise limiter is a low-pass filter with variable cut-off profile which only functions during soft passages (when the noise is most audible) by suppressing those frequencies to which the ear has the highest sensitivity, that is, about 1 . . . 10 kHz. The amount of suppression is therefore dependent upon the level of the input signal. During loud passages, the cut-off frequency is shifted upwards so that the entire audio range is passed, including the noise, but this is then, of course, masked by the audio signal. At lower levels of signal input, the cut-off frequency is lowered, so that a relatively larger amount of noise is suppressed. The action of a DNL is illus- strated by the graphs in figure 5 : for an input signal, Uj, of 2.0 mV, the attenuation with respect to the output level at 1 kHz is 10 dB at 7.5 kHz and 20 dB at 10 kHz. The slope is then approximately - 18 dB/ octave. With input signals above about 8 mV, the response is virtually flat to 20 kHz! The input stage, A, (see figure 6) ensures correct impedance matching between the band-pass filter and the DNL. From here, the signal is fed to two channels: the upper one consists of a high-pass filter (B), ampli- fier (D), variable attenuator (E), and fixed attenuator (G), while the lower one com- prises a phase shifter (C) and a fixed attenu- ator (F). The output of the DNL is the sum of the outputs of the two channels which are, of course, in anti-phase. For low levels of input, Uj, the output, Ui, of the phase shifter is, apart from the phase shift, identical with Uj. The output, U2, of the high-pass filter contains only the high-frequency content of Uj. Signals Uj and U2 are, as already stated, in anti- phase so that if they are summed the high- frequency content of Uj is cancelled out. The net result is therefore that of a low-pass filter. When the level of input signal rises, the variable attenuator in the upper channel comes into operation and reduces the contribution of U2 to the output signal, U 0 . The high-frequency portion of Uj is then no longer (or to a lesser degree) suppressed and U 0 will tend to resemble U j more and more. Turning to the circuit diagram (see figure 7), the input amplifier, transistor T2, in con- 2-30 7 junction with C52 and R70, forms the phase shifter. The output of the phase shifter is taken to the DNL output via fixed attenuator R70/R79. The active high-pass filter, formed by C53, C54, T3, and R72 ... 76, is followed by amplifier T4 and a variable attenuator consisting of T5 and associated components. The collector as well as the emitter of T5 feed a signal to the diode bridge D8 ... Dll. Capacitors C58 and C59 are charged to the emitter voltage via R83/D8 and R84/D11 respectively. If the audio signal level lies below the forward voltage of the diodes, these will not conduct. The signal from T5 is then taken directly to the DNL output where it is summed with the signal from the phase shifter. As the two signals are in anti-phase, the cut-off frequency is about 6 ... 7 kHz and filter action is at a maximum. When the audio signal is greater than the diode forward voltage, the diodes conduct and present a low impedance to audio frequencies. A low-pass filter is then formed by R84, C58, C59, which causes the higher frequencies to be attenuated. The end result will be that fewer (or hardly any) high frequencies are removed from the final output signal, which shows up as a flattening of the overall frequency response. Construction As stated before, the design is built up from three modules: pre-amplifier plus power supply plus band-stop filters, the sixteen- element band-pass filter, and the DNL stages. This type of construction makes it possible for everyone to choose which part(s) of the design he needs: some of you may not want the stereo effect, in which case all you have to do is omit the sixteen- element band-pass filter. If the DNL unit only is built, it is, of course, necessary to add a suitable power supply. When the printed-circuit boards shown in figures 8 ... 10 are used, no particular problems should be encountered in the construction. During the building of the power supply, make sure that one voltage regulator IC is turned 180° with respect to the other. In view of the small current consumption, these ICs do not need heat sinks. The band-pass filter board is best com- menced by wiring in the four wire bridges which are to be located under IC2 . . . IC5: this will make things a lot easier later on. The DNL board consists of two absolutely symmetrical halves: it is possible to cut it into two and have two independent mono DNLs! In contrast to the remainder of the Figure 7. The circuit dia- gram of the DNL: two such circuits are required, one for each channel. Parts list (DNL) Circuit: figure 7 PC board: figure 10 R67.R67' = 270 k R68.R68' “ 150 k R69,R69',R71.R71 ' = 1k5 R70,R70',R80,R80' - 5k6 R72.R72* = 15 k R73.R73' = 2k2 R74.R74' = 180 k R75.R75' ■ 680 k R76.R76' - 3k9 R77.R77' - 330 k R78,R78',R84,R84' - 22 k R79.R79' = 6k8 R81 ,R81 '.R82.R82' = 680 n R83.R83' - 120 k R85.R85' - 220 k P2.P2’ = 47 k (50 k) preset Capacitors: C51,C51 , ,C61,C61 , -4p7/ 16 V C52.C52',C60,C60‘ « 4n7 C53.C53' » 1 n8 C54.C54* - 270 p C55.C55' - 1n5 C56.C56’ = 680 p C57,C57‘ - 2n2 C58.C58',C59,C59' - 22 n C62.C62'- 10 (i/16 V Semiconductors: D8 . . . D11.D8' . . . D1V - 1N4148 T2 . . . T5.T2' . . . T5’ - BC547B Miscellaneous: S4 - DPST switch Figure 8. Layout and component side of the printed-circuit board band-stop filters and power supply. ) (ebruaiy 1 984 2-33 33338 If you have no access to a suitable a.c. voltmeter, adjust the preset(s) by ear. Make sure that with a reasonably large input signal the high frequencies are not cut. If that happens, the input signal is too small and must be adjusted with P2. If this has already been set for maximum sensitivity, adjust PI also. If this still does not give a satisfactory result, the output from the signal source (tuner, record player, tape recorder) is too low, in which case an extra amplifier has to be added. Calibration With the output of a tuner qr record player connected to the input of the pre-amplifier board, adjust the overall sensitivity by means of PI until LED D1 quietly blinks in rhythm with the incoming audio signal. Because the DNL is a variable filter, the action of which is dependent upon the signal level at the base of T2, preset P2 should be adjusted carefully. Connect an a.c. voltmeter (input impedance at least 100 kfl) between the wiper of P2 and earth, and inject a signal of about 1 V into the input terminals of the DNL. Adjust P2 for a reading 775 mV on the voltmeter. If the input signal was derived from a tuner, or record player, it may be necessary to re-adjust PI slightly. The DNL can be inserted almost anywhere into the audio chain, but as its 0 dB input level must correspond to 775 mV it must be located before the volume control. universal active filter Not so very long ago, active-filter ICs would have seemed about as likely as pocket washing machines but today they are, if not exactly commonplace, certainly readily available. With the aid of very few extra components the Reticon R5620 can form the basis of a versatile active filter for use in audio or synthesiser applications - or as an extra piece of test equipment for use in the workshop. All this — and not a single coil in sight! universal adi/e filter five filter modes from one 1C Figure I.The R5620 active filter 1C forms the basis of the universal filter circuit shown here. The binary coding for programming the filter parameters is derived from the two counters IC2 and IC3. The full title of the Reticon R5620 is 'a no calculations! The same is of course true second order switched capacitor filter for the filter centre frequency. As can be seen network'. It is able to implement the five from the table, clock frequency to centre fre- basic filter modes: low pass, band pass, quency ratio (f c /fo) can be varied over two high pass, all pass, and notch. One further, octaves, from 50 to 200, in 32 logarithmi- very useful, function of this 1C is that of cally spaced increments. The Q factor range a programmable sine-wave oscillator. is also in 32 steps from 0.57 to 150 with One could be forgiven for expecting to approximately logarithmic spacing, find all this in a large IC of the LSI variety. The filter mode selection is determined by In fact, it is all contained in an 18-pin routing the AF input to the tree inputs of package thanks to one further feature of the the IC (see table 2) by means of switches. R5620: all functions of the IC are fully All this is illustrated in the circuit diagram programmable. This includes the filter centre of figure 1. frequency and the Q factor both of which are independently programmable by means The circuit diagram of two five-bit binary codes. For example, To make practical use of the R5620, we to program the filter for a given Q factor, have featured the IC in a circuit for a univer- table 1 provides the binary code required sal filter suitable for use as test equipment in no potentiometers, no coils and, best of all, the workshop. r india lebruaty 1 984 2 - 3 6 The AF input signal is fed to the appro- priate inputs of IC1 by wafer switches S3A . . . S3D. The switches also ensure that unused inputs are taken to earth. The five-bit codes for programming the Q factor and centre frequency are presented to IC1 at pins 2 ... 6 (Q) and 13 ... 17 (f Q ) respectively. As a glance at table 1 will show, all that we require to generate the two five-bit codes is a pair of 5-pole 32- way switches! Yes, that's what we thought too, so back to figure 1 ! Both IC2 and IC3 are 7-stage (we only use 5 here) binary ripple counters that will count up (and only up) when presented with a clock input at pin 1. This is provided by the oscillator formed by a 555 (IC4) and its associates. With the component values given the frequency is fairly low and it is possible to step the binary counters along by means of die pushbutton switches SI and S2. The RC networks consisting of R4/C2and R5/C3 are included to ‘debounce' the switches. When the required binary number is arrived at, the switches are re- leased and the R5620 will then be pro- grammed according to table 1. As stated, ICs 2 and 3 are ‘up’ counters only and, therefore, to return to the starting code of 00000, the entire binary code must be run through to the end. This method of operation was chosen simply for the sake of economy (it’s a shade cheaper than 32-way switches anyway!) but the circuit can be modified at will. It is a simple matter for the codes to be made visible by means of driver transistors and LEDs. In the circuit diagram these are T1 . . . T10 and D3 . . . D12. The bases of the transistors are connected to the ter- minal points at the inputs to IC1 marked A . . . J. The connections to pin 2 of ICs 2 and 3 (the 'reset' inputs) enable the two counters to be automatically set to zero when the power supply is first switched on. They also serve a second, slightly moresubtle, function. In the beginning, it was said that the R5620 was also able to operate as a sine-wave oscillator. This is entirely true and for this function the output is switched back (via S3c) to the band-pass (BP) input while the LP and HP inputs are taken to earth: No problem here but there is a strange quirk in the R5620 to be taken care of. To function in the oscillator mode, the Q factor inputs (pins 6 ... 2) must be programmed to 11101. We know, because it says so in the spec, sheet! This is car- ried out by the four Exclusive OR gates, N1 . . . N4, between IC3 and IC1. When the commoned inputs to these gates are taken low, (by switch S3d in positions 1 ... 5), the binary outputs of IC3 are unaffected and pass straight through to IC1. When the oscillator mode is selected (S3 in position 6), the commoned inputs to the gates are taken high by wafer S3d. At the same time, a reset pulse is fed to the reset input of IC3 with the result that all its outputs revert to logic zero. However, the gates now function as inverters and therefore the binary number presented to IC1 will be 11101. The R5620 will now operate as a sine-wave oscillator providing 23.0 28.0 35.0 40.0 80.0 200.0 191.3 182.9 174.9 167.2 | 159.9 152.9 146.2 139.8 133.7 127.9 122.3 106.9 102.3 97.8 pushbutton S2 is not touched! If this should happen inadvertently, simply switch S3 to another position and then back to 6. All that we have left to discuss in the circuit is IC5 and its surrounding components. This is the clock oscillator for IC1 and its frequency is variable by means of potentio- meter P2. We can now clarify the relation- ship between the clock frequency and the binary number that appears on pins 13 ... 17 of IC1. When the code is 00000, the centre frequency of the filter is l/200th of the clock frequency as can be seen in table 1. It will now be apparent that the code sets the centre frequency to a ratio of the clock frequency. This gives a very wide filter response range. Some final points worthy of note! It is of course possible to do away with the switches and counters and simply 'hard wire' the R5620 inputs to whatever function and parameters that are required. Bear in mind that 10 V can be considered as a maximum for the power supply voltage and some protection from tum-on transients must be included. The clock frequency range is fairly wide and can be anywhere between 10 Hz and 1.25 MHz. In conclusion, the R5620 uses NMOS technology and its chances of instant death due to mishandling are inversely proportional to the quantity you have of them at that The R5620 is available from: EG and G Reticon, 34/35 Market Place, Wokingham, Berkshire. H The LCD thermometer featured in the October 1982 (UK) issue was originally intended as an ambient temperature indicator. We don't, of course, know what you're using it for, but from the many ietters we have received asking for a switched output extension, it would appear that many of you would like to use it as a thermostat. We wouldn't dream of disappointing you! from thermometer to thermostat At first glance, the circuit does not look too (pin 2) of IC1 . exciting: a preset and a comparator. Yet If the voltage at pin 3 is greater than that there’s more to it than meets the eye: after at pin 2 (that is, measured temperature is all, it has to work reliably for very long higher than reference temperature), die periods. Tests conducted in our own lab- voltage at the output (pin 6) of IC1 is oratories over a long period of time have high (nearly Ub). A current will then flow proved the extension to be entirely trouble- through R3 and R4 which is sufficient to free. cause a drop of about 1 .5 V across R4. Operation is simple: if the ambient tempera- This is more than enough to make T1 ture rises above a value preset with PI , the conduct. The consequent collector current relay is actuated. The relay contacts can, of flows through the relay, Re, which is then course, be connected to whatever you actuated. An optical indication can also be wish: an alarm, the contacts of a room provided by the LED. thermostat, and the like. It is also possible The supply voltage for the extension can be to have an optical warning of rising tern- obtained from terminal B (+Ub) on the peratures by connecting an LED and suitable printed-circuit board of the thermometer, series resistor (R v ) as shown in dotted lines Pin 3 of IC1 can be soldered directly to in figure 1. In this case, the relay may not be junction R10/R11, while R22 in the exten- required and R3 and R4 can be replaced by sion should be soldered to the junction of a single resistor of 10 k. And, of course, Rll and P2 (suitable soldering points are there are many other possibilities as a little already provided on the printed-circuit thought will show. board). Don’t forget to connect the two The non-inverting input (pin 3) of the earths together! opamp, IC1, is connected to the junction If the thermometer is powered from a of R10/R11 in the LCD thermometer. The primary battery, it would be wise to provide voltage at this point is proportional to the the power for the relay from a separate measured temperature. A reference voltage, source - a low-current relay is, of course, representing the set temperature, is preset ideal. M by PI and applied to the inverting input switched out- put for LCD thermometer Figure 1 . The circuit o< the switched extension shows that it takes only a preset, a comparator, and a switching stage to thermometer into a ther- I ! i i audio sleuth at work faultfinding in audio installations First of all, we are not going to suggest that you open up each item of your installation, heat your soldering iron, and prepare your- self for 'surgery'. On the contrary, the hints in this article deal with fault-finding without special tools and without expensive test equipment. As a rule, start your fault-finding with a list of questions. How did the system behave before the fault? Was everything all right? Was there any noise, hum, or crackle? Has it ever worked satisfactorily? Such a list often points to the most likely area of the fault. You then carry out a quick check of whether this is indeed so. If so, all well and good; if not, a more systematic check has to be made. One of the quickest methods is the so-called 'halving method’. Let us assume that the fault lies in an unknown part of a chain of units or circuits. Such a chain may consist of any number of items: figure 1 shows a typical 'audio chain’. The finding of a fault in an audio system would have been very much to Sir Arthur Conan Doyle's liking. Like Sherlock Holmes, you should sit down and calmly reason out what's wrong. Take the symptoms one by one, put them in logical order and then try to find the solution by deduction. If a sig- nal is applied the inpi of the chain and something is wrong with the output of the pre-amplifier, you know that the fault lies some- where in that unit. Then ‘halve’ the possi- bilities, and check the signal at the tape output: if this is all right, the fault lies between there and the final output. If, however, the signal at the tape ‘OUT’ is faulty, the fault lies in the pre-amplifier before the tape output. Never start with the more complicated checks but rather with the simple ones; only when these give negative results, bring in the big guns. The possibilities vary from checking whether the mains plug is securely in the socket to ‘open heart surgery’ where .2-39 audio sleuth at work the main amplifier with the various printed- the fault lies before the point where the circuit boards temporarily removed is channels were interchanged. If the signs of surrounded by an array of test instruments disorder continue in the same channel, the like a de luxe sine/square-wave generator, fault exists after the cross-over point. Take a double-beam oscilloscope, spectrum ana- care to make only one interchange at a lyzer, and so on. time! Checking the mains plug may sound ridicu- Restore the crossed-over point and make lous, but in practice many problems can be a similar check elsewhere in the chain. Such traced back to this sort of simple cause. a check may also be combined with the Check therefore whether somewhere in the ‘halving’ check. It is true that the number chain there are no controls in the wrong of possible interchange points in figure 1 position, and whether all fuses are OK. is not great, but we felt it better not to show all the intermediate ones. If the amplifier uses DIN connectors, an The ‘interchange trick' adapter as shown in figure 2 may have to be A check which is very suitable as an indi- made up to enable cross-overs to be made, cator is the so-called ‘interchange trick’ in If ‘phono’ connectors are used, making which the left and right-hand channels an interchange is, of course, simplicity are crossed over somewhere in the chain. itself. Figure 1 shows which inputs and outputs If the checks described so far fail to give of an amplifier can be used in such a check. the right result, the time has come to bring If we assume that the symptom is the non- in the big guns! Get the temporary use of satisfactory operation of one channel, a second, soundly functioning audio system change left to right and vice versa. If now and replace one or more of the units from the other channel shows the symptom, the malfunctioning chain by the corre- eleklor india lebruary 1984 2-40 4 3 sponding ones from the auxiliary system. The interchange points indicated in figure 1 can be used for connecting the replacement Balance check If a loudspeaker is connected between the 'hot' terminals of a stereo amplifier (the two earth terminals thus remain ‘open’), sound will come from the speaker even if only one channel is working properly. If no sound at all is audible, neither channel is operating. With the loudspeaker connected as above, apply a mono signal to both channels and set the mono/stereo selector to mono. With the balance control in its mid position, no sound will come from the loudspeaker, while increasing sound should be heard when the balance control is turned left or right. The sound-null will often coincide with the popular ‘12 o’clock' position of the balance control. Because only one loudspeaker is used, the coincidence is not the result of acoustical imbalance (that is, incorrect positioning of the loudspeakers), but rather of electronic imbalance of the two channels (it could also be faulty positioning of the knob of the balance control onto its spindle). Signal generator Before getting out the tone generator (if you have one), remember that you yourself are an excellent hum generator. Take a piece of bare wire between thumb and index finger and insert it into the relevant input. Before you do, turn down the volume control! A better, but still inexpensive, alternative is the test circuit shown in figure 3 which, believe it or not, enables you to even check the high-frequency control! It uses a small transformer (for instance, a bell trans- former) of which the secondary voltage is rectified and from which the d.c. component is removed by Cl. The result is an alternating voltage with a fundamental frequency of 100 Hz and a large number of harmonics (primarily caused by the characteristic of diodes D1 . . . D4). When S2 is switched from position 1 to 2, the unit to which the circuit is connected should produce more hum. If it does not, a fault is indicated. Open circuits and dirty contacts Is the sound weak and shrill, in other words, does the output consist mainly of high frequencies? That could indicate an open circuit, like a break in a cable (the high fre- quencies still come through, albeit attenu- ated, via the capacitance caused by the break). Any crackling or loud clicks when a switch is turned? That may be caused by leaking coupling capacitors. Just behind each output coupling capacitor, and just before an input coupling capacitor, a resistor connected to earth is required to keep the d.c. across the capacitor constant. If d.c. appears across the resistor, the capacitor leaks and should be replaced. This sort of check requires the amplifier to be on: using a multimeter (lowest d.c. voltage range), measure the d.c. voltage across the relevant resistors. Often the cause for the crackling and clickling is far simpler and can be cured by the following ‘shock therapy'. Switch off the amplifier and turn each switch a couple of times from one to the other extreme positions: this normally ‘cleans' the switch contacts. This sort of remedy is also very useful for the connections at the back of the amplifier. Remove and re-insert each plug a couple of times. Phono connectors should be turned around their axis so that the contact areas are moved. Loudspeaker connections should be given a ‘fresh’ start by renewing the bare ends. Do NOT tin the new ends! It does, of course, no harm to carry out this sort of ‘shock’ treatment once in a while even if there is no fault. Phase check If the sound is all rightish, but not really ‘stereo’, the betting is that the phasing of the loudspeaker connections is not right. The most reliable check for this is still the battery check. Take a 1.5 V battery and remove the cloth from the loudspeakers so that the cones become visible. Remove the speaker leads from the rear of the amplifier. Connect one of these leads to the + terminal of the battery and with the other touch the - terminal briefly. The cone of the loud- speaker will make a forward or a backward movement. Repeat this with the second loudspeaker. Both cones should move in the same direction if the speaker leads are connected to the battery with identical polarity. If not, the connections of one of the loudspeakers to the amplifier should be reversed. H r India february 1 984 2 - 4 1 wind direchon indicator R. Bakx "revolving pointer often in shape of The article featuring the wind speed meter (anemometer) published in our November 1983 issue prompted us to expand the 'Elektor weather station' by adding an electronic wind direction meter. This instrument consists of a 'pick- up' and a read-out, connected together by means of two wires. The read-out indicates the wind direction with 16 LEDs. This could also be expanded so that the read-out is shown on an alphanumeric display. In this electronic wind direction indicator Before going on to look at the circuit dia- the position of a wind vane is first translated gram, we must first see how the power and into a code, which is sent below to display the wind direction information are carried the wind direction on a wind compass card on the same line. This will then make the made up of 16 LEDs. The great advantage of layout of the circuit much easier to under - cock mounted in high place esp. on church spire to show whence wind blows," (OED) the set-up used here is that only two wires are needed for interconnection between the pick-up section (at the wind vane) and the read-out section (with the wind compass). These two wires are used to provide the power for both sections and at the same time to carry the wind direction information to the read-out. The principle Because a simple connection between the two sections was considered important in this design, an easy method had to be found to allow both the measurement signal and the supply voltage to be transmitted over a single line. As we will see later, we solved this problem in a very unusual way. The direction of the wind is translated into a four bit code by means of a coding disc fixed to the wind vane and four reflection sensors mounted below the disc. This code must now be sent in serial form to the receiver. There the signal is reconverted into a four bit code that is used to drive the 16 LEDs of the wind compass. The block diagram of figure la shows the main parts of the circuit. stand. The diagram of figure 1 b shows how this two-wire ‘traffic’. is achieved. In prin- ciple the supply transformer is situated between the pick-up and the read-out sections. Each section has its own supply buffer consisting of a diode and an electro- lytic capacitor. Data is transferred between the two sections by means of a transistor in the ‘transmitter’ end and an opto coupler in the ‘receiver’ (display) end. The trans- former is linked to the connecting cable via a diode and a resistor as shown. Positive half-cycles of the mains frequency are now treated differently from the nega- tive. What happens during a positive half- cycle is shown in figure lc. The trans- former voltage is half-wave rectified by a diode so that the two electrolytic capaci- tors are charged and the two sections of the circuit are provided with a d.c. voltage. The diodes prevent the capacitors from dis- charging during negative half-cycles. As we have said, the negative half-cycles are treated differently, and this is illustrated in figure Id. If transistor T conducts the two wires are short circuited. If T is not conducting a current will flow through the LED in the opto coupler of the read-out section, so that 2-42 the opto transistor will give a pulse. The operation of the whole circuit is as easy as it is clever; when T is conducting no pulse appears at the output of the opto coupler, but when T is not conducting the opto coupler gives one pulse for each negative half-cycle. In this way signals can be trans- mitted during the time when there are no supply pulses on the line. The lines therefore carry positive pulses with a frequency of 50 Hz and negative pulses ‘supplied’ by T. The result is shown in figure Id. We use the number of 50 Hz pulses between two negative pulses as information relating to the wind direction. As far as logic is concerned, the circuit for the wind direction indicator is also split into -two sections; the pick-up (figure 2) and the readout (figure 3). We will begin with the pick-up circuit, which will later be fixed to the wind vane. The power supply for this section is handled by D5, C2, C3 and regulator IC3. The 50 Hz pulses appearing at point P are formed into a square wave by N3. High frequency inter- ference on the lines is suppressed by RC network R18/C4. Negative signals on the line are blocked by diode D6. 1 Figure 1. A rough block ~7 1 JUUULTL ?i| Figure 2. This is the d The wind vane is fixed to a four bit Gray erode disc, by means of which 16 wind directions are coded into a four bit code. The disc contains opaque and translucent sections, and its layout is shown in figure 5. A digital signal is supplied by four reflection sensors, IC11 ... IC14, mounted below the disc. Alternatively, four LEDs and four photo transistors could be substituted, with the diodes shining through the disc onto the transistors. These are indicated in the parts list as D1 . . . D4 and T1 . . . T4, which are simply four red LEDs and four ordinary photo transistors. The signal from each sensor is amplified by a transistor stage (T5 . . . T8), so that the output of each stage is logic zero if no light is falling on the photo transistor and logic one if the opposite is the case. The four bit wind direction information is now available at points PO . . . P3. This code is fed to the preset inputs of counter IC1. This counter is arranged so that it counts down from the preset value to zero. When it reaches zero the counter automatically presets itself via the monostable multivibrator consisting of N1 and N2. The clock signal (50 Hz) is supplied by N3. The pulse given by N2 lasts about 5 ms and is used to transmit the wind direction information to the 'receiver'. The appear- ance of the pulse causes the LED (and therefore the photo transistor) in the opto coupler to be switched off via T9, and this in turn means that T10 is turned off. The moment at which N2 gives the pulse is defined by the preset value of the counter. Because IC1 is clocked at the mains fre- quency, the number of mains pulses between two successive N2 pulses is exactly equal to the binary code at the preset inputs. Assume, for example, that the binary code is 1001 (= 9), then N2 will give an ‘infor- mation pulse’ after every 9 mains pulses. Because transistor T10 and the photo transistor in IC4 need to be protected against positive mains pulses, two extra diodes, D7 and D8, have been added. The circuit for the read-out section is shown in figure 3. Here we see the mains trans- former with the diode (Dll) and resistor (R19), just as they appeared in the block diagram. The supply section (D12, C6, C7 and IC6) and clock pulse circuitry (R20, R21 , C5, D9 and N4) are identical to these parts of the pick-up section. When an information pulse from N2 is received, the LED in opto coupler IC7 will light, causing the photo transistor to conduct and short the input of N5 to ground. In this section diode DIO is used as a protection against positive voltage pulses on the line. The serial information is recon- verted to a four bit code by IC8 and IC9. IC8 is a four bit counter that counts up The mechanical layout from 0000 at the clock frequency. Whenever All the electronics we have just been de- the circuit receives an information pulse scribing is located on the four printed the counter is reset via the monostable circuit boards shown in figure 4. The two multivibrator of N5 and N6. Just before IC8 circular boards contain the pick-up section, is reset the count is read into latch IC9 with an d the read-out section is on the other a latch pulse from N5. The latch stores this two boards. These four boards are supplied count until a new information pulse arrives. a s one unit through the EPS service and have The outputs of the latch therefore show the to be separated. The two read-out boards same four bit information that was supplied could also be left together, depending on to the preset inputs of IC1. The code then the amount of room available, goes to IC10, which acts as a 4 to 16 line The mechanical construction for the pick-up decoder. The 1 6 outputs drive the LEDs that section with the wind vane is fairly straight- indicate the wind direction. forward. There are various details that must The current through the LEDs is limited to be considered, however. One thing that about 20 mA by resistor R24. The table must be decided is whether to use LEDs and beside the diagram shows the conditions photo transistors or reflection sensors. The for indicating each wind direction. latter are recommended due to the fact that r indie febnjaiy 1 984 2 - 45 shielding from stray light can be a major problem when discrete LEDs are used. The layout of the coding disc is shown in figure 5, and also (full size) on the layout pages at the centre of this issue. A disc is made up with either the shape of figure 5a or 5b. If reflection sensors are used then 5a is needed, otherwise 5b is used with LEDs mounted above the disc and phototransistors below them on the printed board. The two pick-up boards are cut into a circular disc shape and the components can then be mounted. Capacitor C3 must be soldered to the track side of the board, ideally with some form of insulation between it and the copper. Six points on the two boards (PO, PI, P2 P3, +8 V and i) must be connected by means of wires or some ribbon cable. The boards can then be fixed together ‘sandwich fashion’ held in place by a 5 mm diameter rod that is fixed to the base of the ‘transmit- ter’ casing. The coding disc is fitted in such a way that it is allowed to rotate freely about 1 mm above the reflection sensors. A further plastic disc with two strong mag- nets glued diametrically opposite each other ® & rindiafebiuacy 19842-46 is fixed above the coding disc such that the two discs rotate together. The whole con- struction must fit into the (inverted) trans- parent jar so that the disc with the magnets can rotate freely. The connecting cable is passed through a hole drilled in the lid and soldered to the lower printed circuit board. The opening is then sealed well. The form of construction is illustrated in figure 6 but, as usual, individual ideas will probably change this significantly. Now all the electronics is protected in a watertight package, but, if the light sensitive components are not to be affected by ambient light, it must also be made light- proof. This can easily be done by painting the outside of the jar black. Looking at the mechanical construction it is obvious why again we recommend using reflection sensors if possible. If LEDs and photo transistors are used the LEDs must somehow be fixed above the coding disc and they must also be provided with their own power supply. The construction of the outer casing is very dependent upon what material is available. It could, for example, be made using PVC tubing. This outer casing ideally should have bearings for the shaft of the wind vane and some sort of cap is needed to prevent rainwater from getting at these bearings. Remember to provide a hole at the bottom of the casing to prevent condensation building up. Another plastic disc (or simply a strip of plastic) with two strong magnets is mounted at the lower end of the wind vane shaft . Be sure to get the 'polarity' of the magnets correct as their purpose is to induce the magnets inside the jar to rotate ‘in sym- pathy’ with them. It may be necessary to experiment with the value of resistor R 1 . In reflection sensors the sensitivity is often so good that the current through the LEDs can easily be reduced and so help to prevent false' reflections. With normal LEDs the current could be increased a little. Trial and error is probably the best method to use here until a value is found that enables all wind directions to be cor- rectly indicated. Constructing the read-out is very straight- forward. Depending on the case used, the two boards can either be left joined or separated, but in this latter case points AO . . . A3, +8 V and X must be linked on both boatds. To keep this section as small as possible the two boards can again be mounted sandwich fashion. The transformer is connected to the read-out 1 2-47 Figure 6. This drawing and -case - . The electronics are protected from water by sealing them inside a jam jar. Magnetic coupling section, but if desired it can be moved to some other point on the cable. In this case, of course. Dll and R19 stay with the transformer and are not mounted on the printed circuit board. This unit is then connected to the cable as shown in figure 7. Finally the electronic weathercock must still be calibrated. With the aid of a compass the wind vane is pointed North, and then the whole ‘case’ is rotated until the read-out shows 'North'. If the pick-up section is already fixed in position on the roof, it could also be calibrated by turning the magnet mounting disc on the shaft of the enabling the 16 wind directions to be shown on three dot matrix displays. The circuit for this ‘extra’ is given in figure 8. This is connected to the data outputs AO ... A3 of the read- out section (the outputs of IC9). The 'data' for driving the displays is con- tained in a 2 Kbyte EPROM, IC1. The hexdump for the contents of this EPROM is shown in table 1, and this chip is also available from Technomatic Ltd. The displays are multiplexed by counter/oscil- lator IC3 and 4 to 16 line decoder IC4. The outputs of IC4 drive the 15 LED columns of the displays via transistors T8 . . . T22. The multiplexing frequency is about 3.5 kHz. The LED rows of the displays are driven by the data outputs DO ... D6 of the EPROM. The output signals are amplified by transistors T1 . . . T7, and the current through the LEDs is defined by the values of resistors R3 . . . R9. The maximum current through the LEDs is about 75 mA. This current is needed because each LED is only driven for 1/16 of the time. The four outputs of IC4 are also connected to the address inputs AO ... A3 of IC1, so that when a certain LED column is being driven the appropriate 'switching' data appears at the output. Address inputs A4 . . . A7 receive their data from the latch in the read-out section so that, de- pending on the wind direction, a specific 16 byte address of the EPROM is selected that contains the information needed to give the correct display. Voltage dividers R12 . . . R15/R16 . . . R19 are included to reduce the 8 V signals of the read-out circuit to the 5 V used by the display. Finally, a link must be connected between pins 12 and 21 of the 2716. This is necessary to select the correct section of the EPROM. The power supply for this section is handled by a separate 5 V stabilizer (IC2). The current consumption of this circuit is about 150 mA. H Z 80 EPROM prograr B. Barink any Z 80 system with static RAM can be used to program 2716 EPROMs Figure 1. This is the timing diagram for the Z80 control signals during read of MRE Q and WR. whereas MREQ and RD appear circuit is used to set the WAIT line 'low' as soon as the EPROM is addressed. even during a write cycle. Carefu l mani pulation of the WAIT input of the Z 80 is what enables this little circuit to fulfil the particular conditions that have to be met to pro- gram an EPROM in situ given the unusual timing of the control signals of this processor. programmer In order to program a 2716 EPROM there are several conditions that have to be met. The OE (Output Enable) pin must be ‘high’, the levels on the address and data lines must be stable, the potential on pin Vpp must rise from 5 V to the programming voltage of 25 V and finally, the CE (Chip Enable) pin must go 'high' for 50 ms. There is nothing really unusual there but a certain amount of care is needed as the speed of the processor must be slowed down and the peculiarities of the timing of the control signals must be taken into account. It is notable, looking at figure 1, that the RD (read) signal appears at the sa me tim e as the memory validation signal MREQ (memory request), whereas during a write operation there is a delay of one clock cycle between the appearance of MRfiQ and the transition to Tow’ of the WR signal (write). This is important for us as the programming consists of a prolonged write operation. However, to be able to access the EPROM it must be located somewhere in the addressable area. An address decoding (not represented here) is needed to supply a validation signal for the memory zone occupied by the EPROM. The circuit and its timing The address decoding signal must set point ‘A’ in figure 2 logic Tow'. If this signal has been generated with out com bining the address lines with the MREQ line, they can still be combined using OR gate N7. If these signals have already been combined, the 1 * ^njTjarirLTLrL 2-50 Z 80 EPROM progra decoding signal, called ADDRESS here can be applied directly to point ‘A’. We will return later to the PE (program enable) signal which could, in certain applications, take the place of a validation signal. When the EPROM is addressed, the logic level applied to point ‘A’ of the programmer produces a falling edge at the output of N3, which triggers monostable MMV1. A calibrated 50 ms pulse then appears at pin 8 of this JC and is used as a programming pulse at the CE inp ut of th e EPROM. This same pulse sets the WAIT input of the Z 80 ‘low’ via N1 and N5 so that the address word and the data word on the buses remain stable. As the RD line is ‘high', input OE of the EPROM is also 'high'. At the same time T1 is turned off, T2 saturates and the potential at pin V pp of the EPROM goes from 5 V to 25 V. None of this will happen, however, if the WR signal is not delayed, as we mentioned at the beginning of this article. In fact the output of OR gate N3 cannot go 'low' unless the WR line is also ‘low’. Also the delay introduced by monostable MM VI must be taken into account. This is the reason for adding a circuit to introduce a ‘wait’ of several cycles. It consists of a series of flip-fl ops FF1 . . . FF4, which hold the WAIT pin of the Z 80 ‘low’ im- mediately after point 'A' goes ‘low’. The max imum delay between the time that the WAIT input should go ‘low’ (making the address and data words on the buses stable) and the time when the 'low' appears on the WR line is about 150 ns. A few dozen ns delay introduced by MMV1 must be added to this. With the four flip-flops we gain three wait cycles, or 750 ns with a 4 MHz clock. As the tim ing diagram of figure 1 s hows, t he WAIT input goes ’low’ just after MREQ, even though the WR line is still ‘high’. As soon as the 50 ms CE pulse arrives, the address and data buses are fixed and remain so for the duration of the programming. Read cycle The wait circuit is triggered by the address decoding signal, so it also works during the read cycles of the EPROM. This gets over the problem of EPROMs whose access time is normally too long (450 ns). The monostable, on_ the other hand, is not activated, so CE remains ‘low’, as the first part of the timing diagram shows. OE, however, goes Tow’ as soon as RD does. Then all the conditions required for the EPROM to put data onto the bus are met. In order to retain the normal reading speed the wait cycles must be cancelled. This is easily done by linking pin 6 of N4 (OE) with pin 4 (PR) of flip-flop FF1, which will then no longer be connected to +5 V. Programming in situ This is not a totally autonomous EPROM programmer. It is, in fact, an auxiliary circuit in which the EPROM socket has wire wrapping terminals. There are, of I course, a few links that have t o be wired in: PHI/EX (the clock), WAIT, R.D, WE, the address decoding signal (or PE) and finally the programming potential of 26 V (not 25 V as there will be some voltage dropped across D3 and T2). M ake sure th at the address decodin g sig nal (ADDRESS) does not contain the RD signal as its pres- ence would prevent any writing, and there- fore programming, from taking place. The programming unit of the polyphonic synthesizer is a nice example of program- ming in situ. If you look at the circuit diagram in the relevant article you will see what we mean. In this case there is no need even to fit a special socket for the EPROM as it takes the place of RAM IC9. The 4071 (IC6) is removed from its socket and the signals for the EPROM are then applied to the pins as follows: pin 10 (IC6): OE (pin 20 of the EPROM) pin 1 1 (IC6): V pp (pin 21 of the EPROM) pin 4 (IC6): CE (pin 18 of the EPROM) The clock signal PHI/EX is available at pin 27a of the jrP bus^as are RD (at 31c) and WR (31a). Signa l PE is available at the output of N10. The WAIT signal is applied to pin 5c of the 64-way connector. Then, whenever the potential of 26 V is present, every operation to write to memory (store enable) causes the EPROM to be pro- grammed. * Figure 2. The circuit dia- gram for the Z 80 2716 EPROM programmer consists of a monostable that generates a calibrated 50 ms programming pulse. the WAIT line ’low' even before the WR signal appears. By mounting this circuit on a piece of veroboard fitted with 24 wire wrap pins, this pro- tuted for the EPROM to 2-5 course, the breaking of tracks has become superfluous. When this type of board used, all connections must therefore be made with suitable wire for which an appropriate technique has envolved. In this, use is made of thin enamelled copper wire. When a connec- tion is to be made, the enamel is removed from one end of the wire with a hot soldering iron. As the wire is very thin, it can be inserted without too much trouble into the relevant hole beside the connecting wire of the component. To prevent the copper wire jumping from the hole, it is wound several times round the component terminal. In this way it is possible to make multiple connections before they are soldered. The insertion can, of course, be done very well by hand, but there is a simpler way: with a wring p'obe. How to make this practical aid is described below. printed f circuit board I 1 is ideal for con- structing reliable circuits. Not everyone, however, has the necessary material and tools to produce such boards. Apart from that, it is often not worth the trouble and expense to design, photograph and etch a print layout for one printed circuit board. There are however more ways which lead to Rome. How to make it A propelling pencil with a lead diameter of 0.5 mm, a cotton reel and a strip of alu- minium (about 90 x 20 mm) are required. If a propelling pencil is not available, take a ball-pen and hypodermic needle (also with an opening of 0.5 mm). Remove the top of the propelling pencil so that is becomes open-ended. When a ball-pen is used, remove the ink reservoir and operating pin or button; the hypodeimic needle is then placed in the pen such that it protrudes about 5 mm from the normal writing end. At the centre of the strip of aluminium drill a hole of suitable diameter into which the top end of the pencil or ball-pen is to be inserted. Two smaller holes are then drilled at either side of, and equidistant to, the centre hole. The aluminium is then bent into a U-shape so that the cotton reel fits between the two home-made low-cost wiring probe time-saving device for the wiring of circuit boards from a contribution by H. MeBmer There are two main alternative prototyping circuit boards which differ principally in the method of wiring. The first is one with continuous copper tracks: when this is used, only a few additional connections have to be made - provided, of course, that the component layout has been so well thought out that the final product has as few wire connections as possible. Readers who like solving puzzles are well away with these boards! However, particularly in the case of digital circuits, these boards can give problems: depending on the position of IC's, it is often necessary to break the copper track between the connecting pins. Even with the right tools this can prove to be a tiresome and time-consuming job. The second alternative is better suited to such circuits: boards containing only solder pads. Because no account needs to be taken of copper tracks, components can be placed rather more freely on such boards and, of vertical sides as shown in figure 1. To ensure that the reel can rotate freely, use a 2 BA screw and nut as spindle. All that remains to be done is to wind a suitable length of enamelled copper wire onto the reel. Home-constructed circuits should present no problems Material Prototyping circuit boards are usually available from an electronics retailer in so- called Eurocard sizes. The most suitable material is epoxy board which is appreciably more stable than pertinax. The wire to be used is common enamelled copper wire of 0.25 . . . 0.35 mm diameter. A special type of wire is available which, although it is a little dearer, is more easily tinned and soldered. Moreover, it is available in different colours, which is useful for complicated circuits. Whatever wire is used, however, there is one golden rule: tin first, solder 2-52 afterwards! Readers who are thinking of using the wire of a transformer or choke will find that the enamel on such wire is very difficult to remove. A further disadvantage is that the enamel has often become so hard that it crumbles during removal of the wire from the transformer or during rewinding onto the cotton reel: the possibility of a short then becomes very real! The most important tool, the soldering iron, is required to have a tip temperature of 350 . . . 400 degrees centigrade, otherwise it will not be possible to remove the enamel with it. An iron with adjustable temperature is ideal, but if this is not available, try to remove the enamel with the one that is to hand. More tools are not really required, although a pair of small pliers and a pair of tweezers are very useful. Preparation and construction It is advisable at all times (and not just with this method of construction) to use 1C sockets, as soldering direct onto IC pins often ruins the component. It may also be worthwhile, especially for beginners, to take sufficient time to consider the best 1 are for decoupling. location for the IC’s. A mirror image sketch or drawing of the IC connections obviates a lot of turning over of the board. First place the socket onto the board and solder the diagonally opposite pins (for instance, the + and - of the 1C) to the board. After all other components, screws, pins, and so on, have been placed in their respective positions on the board, a start can be made with the wiring. The supply lines should be done first (see figure 2). The 0 V (earth) line is best done in bare copper wire and the + line in insulated copper wire, somewhat thicker than is used for the remainder of the connections. In most digital circuits a diameter of 0.4 mm for the supply lines is adequate. A hint: mark pin 1 of all IC's on both sides of the board: this will simplify finding one’s way in the tangle of wires appreciably ! With careful work, it is possible to con- struct even a 16 or 64 k RAM-card in this way, which shows that prototyping circuit boards are not necessarily inferior to printed circuit boards! M 2-53 Memory in a computer is a hardware combination of logic elements which is totally independent of the software but which the software must take into account. The structure and organization of the addressable area is far more than simply a matter of getting the appearance right. This is one of the least understood characteristics of computers, and yet it plays an essential role in the operation of the machine, in the layout of the software, and even in adding memory extensions or peripherals, such as input/output modules. address decoding why and how an addressable area is organized The memory of a computer could be com- pared to a.large library: the information, or data if you prefer, is the books and their contents, which we will only mention briefly here. What interests us in this library is its filing system, and especially the way that it is laid out, with its groups, categories, sub-groups and so on. In other words, it is the reference system that we are interested The value of the information Imagine a catalogue of several billion works dealing with the most varied and different subjects. Our library, of course, contains books on electronics. These are gathered under the reference 'E'. Books about digital electronics are located under the reference of ‘ED’, whereas those concerning analogue subjects are classified under ‘EA’. In data terms we would call the letter ‘E’ the most significant bit of the references 'ED' and ‘EA’, and ‘D’ and ‘A’ are less significant bits. This distinction is easily seen as the letter 'E' here signifies all works dealing with electronics in our imaginary library, whereas the letters 'D' and ‘A’ refer only to a certain number of these books. If we continue to make our references even more detailed, the next character (which is less significant again than the previous two) could, for example, be used to distinguish between works in English and those that are not. So a book filed under ‘EDE’ is in English and deals with digital electronics, while a book with the reference ‘EAF’ deals with analogue electronics and is written in French. This last character ( English or not) is less significant than its predecessor ( digital or analogue): within the category of ‘elec- tronic works’, the distinction between ‘digital’ works and ‘analogue’ works is more important than between works written in English and those written in French. To finish with this attempt to clarify the idea of the significance (or importance) of information, here is a little example. It has to do with the prices displayed by shopkeepers on their merchandise. They would much rather ask £ 9999.99 than £ 10000.00 for a product. Why is that? The most significant information (the number of thousands of pounds seems cheaper between one price and the other, but in fact the difference is insignificant as it only involves a very slight change in the least significant information character. Subdivision and double addressing Let us now turn to computer memories. These appear as a stack of compartments (called memory cells), each containing 8 irreducible units in the systems most familiar to us, that is 8-bit microcomputers. These discrete units, the bits, are not separately accessible: they constitute an eight-bit word called a byte, and their logic values make up the data. This word travels to the interior of the system via the data bus, which con- sists of eight lines numbered D7 . . . D0, each corresponding to one data bit. The words in the memory are accessed by the processor via an address bus, consisting of 16 lines numbered A15 . . . A0, along which our compartments are arranged. This organization could be compared to that of the library in the preceding example. In figure 1 we have represented the six least significant address bits (A5 . . . A0) as corridors with successive branches as it could be imagined in a library. Whether a left or right turn is taken in these corridors, the end is reached little by little. The de- cision to go ‘left or right’ in an address line is indicated by its high or low logic level (indicated as '1' or ‘0’), which are the only two states possible. The more the binary 'weight' of an address bit is increased, the more important the zone covered by it becomes. Because bits 5 and 4 in figure 1 are both 'O’, a ‘0’ at bit 3 means that the area from 00 to 07 is selected, whereas if bit 3 is ‘1’ the zone from 08 to 0F is ac- cessed. If bit 4 then changes to ‘1’ with 5 still being ‘O’, the decision of bit 3 selects between zone 10 ... 17 and 18 . . . IF. Assume that in a specific application the logic level of bit 3 is not defined while bits 4 and 5 are both ‘O', then the result is that the zones mentioned before are no longer differentiated. Zone 00 ... 07 will be confused with zone 08 . . . 0F. This is called double addressing. Depending on the binary weight of the undefined bit, the range of the doubly addressed zones will be more or less important. 2 “ = 65536 The six most significant address lines are shown in figure 2, which also indicates their contribution to splitting up the addressable area. Quantities indicated by the sign ‘K’ are always multiples of 1024 (not 1000), which is the number of memory cells ac- 2-54 cessible with the first ten address lines (A9 . . . A 0; 2 10 = 1024). Consequently, when talking about memory, the sign 'K' designates 1024 bytes and not 1024 bits. Depending on whether address line A15 is at a high or low logic level, one of the two 32768-word halves of the total memory addressable with 16 lines (2 16 = 65536) is selected. Within each of these blocks, line A14 differentiates between two blocks of 16384 words . . . and so on until line A10 which allows two blocks of 1024 words to be selected within a block of 2048 words decoded by A1 1. As we mentioned before, if the logic level of one of the address lines is undefined two normally distinct blocks are confused. So if the logic level of A15 is not specified, address 0 and address 32768 are mixed up. The same applies for address 1 and address 32769, and so on. Don’t forget that for addressing, no matter what the base (binary, decimal or hexadeci- mal), the count always starts from 0. This leads us to table 1 , which shows the 16 address lines, their 65536 possible combinations and the corresponding ad- dresses. Despite the apparent linearity of the progression of this table, the weight of the address lines increases from right to left, and in line with this increase the range of the zones covered by the decision of an address bit becomes more important. This is shown at the extreme left of the table where the ranges of the zones decoded are indicated. Generating enable signals So far we have considered the problem of addressing purely as a matter of topography. Looking at the integrated circuits that we must manipulate, we see that the most common ones do not have 16 address lines but a lesser number, proportional to their capacity. As can be deduced from figure 2, a chip containing 4 K (such as a 2732 EPROM) must have 12 address lines (A1 1 . . . A0). Addressing each of the 4096 words is achieved by means of an internal address decoder incorporated in the IC. In the same way an IC containing 2 K of memory (for example the still common 6116 RAM) will have 11 address lines (A10 . . . A0) which will enable the internal decoder to distinguish between the 2048 memory cells. What is called address decod- ing is not, strictly speaking, this internal i tebruary 19842-55 a certain number of control signals are also decoding in the block of memory contained in an IC, but rather the location of this block in the area addressable by the CPU. For our examples we will concentrate on the 6502 and Z80, both of which have 16 address lines and can therefore decode up to 64 K of memory. Every memory 1C has, in addition to the address lines we have just mentioned, one or more enable inputs. These have to be brought to a certain logic level (generally low, which is indicated by a negation bar above the ‘name’ of the corresponding pin) to make the chip active. This means that the internal addressing only takes place when the enable signal is present, and the data words are not placed on the data bus until this condition is fulfilled. This enable signal is obtained using the most significant address lines, combined with certain control signals that are essential for the timing of the operations (see figure 3). These control signals are different for each system; for the 6502 they are: ■ clock signal <1>2 which only permits reading and writing operations during the second half of each clock cycle of the processor, and ■ the R/W signal which distinguishes between read oper ations (Read) and write operations (Write). The corresponding signals in the Z 80 are: ■ WE and RE to distin guish between writing (W rite Enable) and reading (Read Enable), and ■ MREQ and IOREQ to distinguish be- tween operations carried out with the memory and those dealing with the input/ output module for which the Z80 has specific instructions. The differences be- tween the two processors are clarified by figures 4a and 4b. The validation signals, obtained from the most significant address signals and the con t rol signals, a re all re- ferred to here as CS (Chip Select). Just for the sake of making things easier to follow, we will assume that they are always active at the low logic level. However, depending on the system and the manufacturer, it is possible to find some signals, including the enable signal, which are active high. Before getting on to the logic combinations which will allow these enable signals to be generated it is no harm to emphasize the importance of the hexadecimal base. We have sixteen address lines grouped as 4 x 4 lines. There is a hexadecimal figure (0 . . . F; 0 ... 15 in decimal) corresponding to each group of four lines. In address 4A2F, for example, the 4 corresponds to the binary word for lines A15, A14, A13 and A12 (0100), the A corresponds to the binary word on lines All, A10, A9 and A8 (1010), the 2 to the word on lines A7, A6, A5 and A4 (0010) and the F to that on A3, A2, A1 and A0 (1111). This simple conversion allows the configuration of the 16 address lines, corresponding to an address given in hexadecimal, to be easily found. Fixed logic combinations Now we will start looking at the address decoding proper, achieved by means of more or less complex logic combinations. Imagine a memory circuit to be enabled between addresses 2000 and 2FFF. Lines All ... A0 decode 4098 memory cells between X000 and XFFF. Combining the A15 . . . A12 lines as shown in figure 5a provides a CS signal active (at logic zero) only when the configuration of the lines is '0010', that is the number 2. The example of figure 5b shows more precise decoding. The enable signal CS, obtained by com- bining lines A15 . . . All logically, is only active when the configuration of these lines gives the values E0 . . . E7. The other address lines allow each of the 2048 addresses between E000 and E7FF to be addressed. The decoding obtained with the combination shown in figure 5c is even more precise: CS is only at logic zero when A3 ... A1 5 give the hexadecimal value C10; while the three remaining lines are used for addressing the eight bytes between C100 and C107. 4b 5a & 5b, Examples d address decoding, and 2 K bytes. As ne addressed Figure 4b. The internal structure of a Z 80 system is quite similar to that of a 6502. except that it the problems associat 1984 2-57 Figure 5c. Another example of fixed address Figure 6. The 74LS138 decoder allows an 8 K block (decoded using A13 . . . A15) to be easily 1 K, each with its own C5 signal. The second enable input is treated differently depending on whether it is used with the Z 80 or the 6502. These three examples show how the decod- ing is narrowed down by using a larger number of significant address lines to generate the enable signal, and how this reduces the range of the zone addressed. For the sake of simplification, these examples have completely ignored the command signals that are needed to put all this into practice. A multiple address decoding circuit is shown in figure 6. It contains a commonly used decoder IC, the 74LS138, which has three binary data inputs a nd tw o enable inputs (G2A, G2B). Signal G2A, which is ob- tained from a combination of A13 . . . A15, is only active betwee n 00 00 and DFFF, a block of 8 K. Input G26 picks up the MREQ signal from a Z 80, or is tied to earth (logic zero) if used with a 6502 processor. The three bit binary word created by com- bining A10 . . . A12 allows eight successive blocks of 1 K to be decoded. The eight 03 signals thus produced could be applied to the memory, in conjunction with com- mand signals WE, KD or R/W. Variable logic combinations The decoding examples examined so far have one thing in common, that they are invariable, but variable address decoding is also possible, as illustrated by figure 7. The main part of this diagram is the four bit magnitude comparator, a 74LS85. A binary word A0 . . . A3 is provided by address lines A12 . . . A 15. This is compared by the 74LS85 with the binary word sup- plied by four switches connected to earth and four polarizing resistors to the high logic level. When binary word A0 . . . A3 is the same as binary word B0 . . . B3 pin 3 (A = B) goes logic high. The output of this pin is then inverted and becomes the C3 signal for a 4 K memory block (X000 . . . XFFF, where X is the hexadeci- mal value corresponding to binary word B0 . . . B3). The same sort of programmable address decoding could be achieved using EXNOR gates, as shown in figure 7b. The open collector outputs of the 74LS266 are all logic high only when the two inputs of each gate are at the same logic level. Each gate compares one bit of the address word formed by A12 . . . A15 with the corresponding bit of the binary word pro- grammed using the switches and polarizing resistors. This procedure has the advantage that it adds flexibility to the address decod- ing. Furthermore, as the dotted lines of figure 7b suggest, it is quite easy to narrow the programmable decoding by increasing the number of significant address lines used, and thus reducmg the range of the block enabled by the CS signal. With that we will finish this article on address decoding, and, while we realize that there is much that has not been said about the subject, we hope that at least some light has been thrown on the address bus and how it works. M that determines when the binary word formed by lines A12 . . . A15 is the puts of the 74LS266 are , all high only when the are at the same logic level. klor india lebruaiy 1 984 2-58 A programmable crystal oscillator Programmable crystal oscillators (PXOs) are not new. They normally consist of a discrete stabilized oscillator, quartz crystal, and one or more dividers which are controlled by logic levels. What is new about the range of PXOs recently introduced by Statek Corporation, one of the largest oscillator manufacturers in the USA, is that the oscillator, dividers, and selector circuits are constructed as a CMOS- 1C which is housed together with the quartz crystal in a standard 16-pin DIL package. Statek has already brought eight of these PXO units onto the market: the only difference between them is the fundamental quartz frequency. This frequency is indicated by the number in the type-coding on the unit: for instance, in a PXO-600 it is 600 kHz. Standard crystal fre- quencies at this moment are: 192 kHz, 327.68 kHz, 600 kHz, 768 kHz, 983 kHz, 1 MHz, 1.3 MHz, 1.6 MHz, and 1.97 MHz. Statek can meet individual customer's requirements for non-standard frequencies. The internal construction and pin- out are shown in figure 1. The direct output of the internal oscillator (OSC) is amplified and then available at pin 11 (F ou t). The oscillator is also connected to the selection logic (SEL) which is controlled from pin 13 (CSEL). When this pin is logic high (TTL-level), the selector connects an external clock (EXC — pin 12) instead of the internal oscillator to the first divider. The divide ratios of the two dividers are determined by three inputs each (PROG 1 ... 3 and 4 ... 6 respect- ively): table 1 correlates the inputs and the ratios. A little arithmetic will show that 57 different fre- quencies are available from a single crystal. The output of the second divider is amplified and then available at pin 9 (OUT). A logic 0 at the RESET input (pin 14) sets the dividers to 1/1 and the OUTput (pin 9) to logic low. A somewhat unfortunate designation has been given to pin 10: TEST. When this pin is logic high, the output frequency is multiplied by 1000, provided the overall divide ratio is not lower than 1/1000. Internal pull-down resistors in the dividers, and a pull-up resistor in the reset circuit, ensure a non- Table 2 Table 2. Output frequencies of the PXO-768 model for various logic levels at the PROGram pins (Unit shown: Hz.) ambiguous logic level, even if the relevant pins are not connected. Pins 1 and 15 are not used. Other important technical parameters are: • high calibration tolerance — stan- dard ± 1 00 ppm • low ageing - maximum 10 ppm in first 12 months • high frequency stability — maxi- mum drift ±0.015% over the temperature range — 10°C . . . +75°C (not including the calibration toler- ance) • low current consumption (CMOS), yet fully TTL compatible • very short rise and decay times (in the PXO-600, for instance, typically 70 ns and 30 ns respect- ively) A typical application is shown in figure 2 where a PXO-768 is connec- ted as a baud rate generator. Table 2 shows typical rates available from this unit. The baud rate is obtained by dividing the output frequency by 16: the extreme values of 0.0004 and 48,000 baud/sec are, of course, hardly ever used. It is, unfortunately, not possible to obtain all baud rates encountered in practice from each PXO unit: a rate of 75, for instance, cannot be derived from a PXO-768 (although it can from a PXO-600). The PXOs can also be used for a variety of other applications, such as a square-wave generator, a rec- tangular-wave generator with variable duty-cycle, or a monostable multi- vibrator. M 2 Table 3. Some baud rates — in baud per : - available from the generator in Further information from: I.Q.D. Limited 29 Market Street Crewkerne Somerset TA 18 7JU Telephone: (0460)74433 Literature: Statek Corporation data sheet 'Programmable Crystal Oscil- lator' er India februaiy 1 984 2 - 60 i mM MINI PCB CONNECTOR Om Lila's, Bombay, have introduced mini HCB connectors used in colour and B/W TV's, tape recorders, electronic instruments, computers, etc. Range of availability is: 1 to 8 pin & 10, 12. 14, 16, 18, 20 & 22 pin. The male part is directly soldered onto the PCB and wires on the female part of the connector are soldered. Wire connections can beeasily connected/disconnected during servi- cing or testing. OM LILA S Plot No. 84-A/B Govt. Ind. Estate Kandivlee (W) Bombay - 400 067. DIL COMPONENT CARRIERS These carriers are ideal tor mounting discrete components to form a plug-in modular circuit. The contacts are 1C spaced ie. 2.54 mm between contacts and 5.08 mm between the two rows of contacts. The body is moulded from glass filled nylon to withstand soldering tempera- tures. Top covers are available for 8.14 and 16 pin carriers. For further details write to Instrument Control Devices, 14, Manorama Niwas, Datar Colony, Bhandup, Bombay - 400 078. (INDIA). COOLING FAN The £lectra fan can be used to cool a wide range of instruments and equipment, viz. computers, power supplies, copying machines, medical insruments, etc. The manufactures claim that it is very rugged and reliable with a long life trouble free motor and silent polycar- bonate blade. The usual model operates on 230 VAC. Models for other voltages can be made available. Power consumption is 25 VA with a 90 CFM delivery. ELECTRA Gopi Building, Behind Vivek Appt.. Dandia Bazar, Baroda 390 001. TITRATION POTENTIOMETER Rashmi Electronics have developed a titration potentiometer, useful in chemical, pharmaceutical, sugar and distillstnes for neutralisation, oxidisa- tion, reduction, complex formation, etc. Range: 0-400 mV continuously variable, calibrated dial with a six- position switch extending the range in steps of 400 mV to total 2.4 V. Operates on 230 V. AC. For further information, contact Rashmi Electronics 2-15-34, Kadrabad, (Polas Lane Corner) Jalna 431 203 MINIATURE RELAY Electronic Relays (India) Pvt. Ltd., have introduced a miniature type dual in line solid state relay in Dip configuration compatible with standard 1C bases. The relay offers phtoisolation between input and output terminals with inputs compatible to TTL sources. The manufacturers state that it has high noise immunity, vibration and dust immunity as it is completely potted in a special potting compound. Overall size: 19 mmx 12 mmx 11 mm. Available in two types: DC/DC relay-handling DC load up to 0.5A/40, 60,80V and DC/AC relay-handling AC load upto 0.5A/220, 330V. For more details, write to Electronic Relays (India) Pvt. Ltd. 10/2, Lalbagh Road, Richmond Circle Bangalore 560 027 DIP EXTRACTOR Instrument Control Devices have introduced a dip extractor with which ICs and similar devices can be safely removed from their sockets. Also useful in removing or handling components like transistors, capaci- tors, resistors and hardware, etc., the extractor is particularly useful in high density packaging situation, according to the manufacturers. It is made of one piece spring steel: a lock-nut adjusts movement and controls pressure. The top portion of the device is insulated to prevent electric shocks. Instrument Control Devices 14 Manorama Niwas, Datar Colony Bhandup, Bombay 400 078 TWILIGHT SWITCH Barathronics manufacture a twilight switch which can switch on/off any electrical load, working on mains 230 V, 50 Hz, up to 500 W., Suggested uses: domestic, remote control, street lighting, counting machines, etc. Specifications are: For more details, write to Barathtronics 53, Temple Street, Malleswaram Bangalore 560 003 1 2-6 1 PCB Stripboard and Cutter Fiji Electronics have placed on the market a general purpose PCB stripboard and cutter to help wire circuits conveniently. Two types of stripboards are currently offered: with 0,15 in pitch suitable for circuits using descrete devices and with 0.1 in pitch for circuits using DIL type ICs as also descrete devices. such as oil rigs and their supply vessels as well as in the petrochemicals industry. The switch has a maximum electrical rating of 15A 480 VAC. Specifications conform to BS 4683, Part 4 and SFA3909. The switches can be imported against actual user's import licence. microprocessor or microprocessor- based systems in laboratories or industry. The tool can be used for the development and debugging of the hardware and software of the system and later in the phase of system integration. Its modular configuration allows upto 7 users to develop software and test hardware/software integration simultaneously. Demands of memory capacity are met by the built-in Winchester drive with upto 20M byte capacity. Add-on disc extension units allow bulk memory to be increased upto 147M bytes on line. Fiji Electronics Mail Order Sales (PCB) Puthencurichy, Trivandrum 695 303 NUMERIC PRINTER Further information from Jost's Engineering Co. Ltd. 60. Sir Phirozshah Mehta Road Bombay 400 001 For further particulars, contact Peico Electronics & Electricals Ltd. (Corporate Relations), Band Box Building, 254-D, Dr. Annie Besant Road, . Bombay 400 025 Process & Control Elements otter a numeric printer, PACE NU-18, which accepts BCD data and gives a print out. It prints 18 columns. 15/16 of which are numerals and 3/2 symbols depending upon user applications. Symbols can be used to identify the variables. Applications: Printing DPMs; Multipoint digital scanner/recorder; El cheapo data loggers; ticketing machines; cash registers. LED HOLDERS Jain Electronics have develop) holders in a wide range for 5 mr mm LEDs. Made in brass and c plated, they are available in r< type, both round and squares LH-43 and LH-49 dome type, as in other specifications. DIGITAL OHMMETER Thermal Sensors have designed a compact digital ohmmeter. Salient features of the device are: resistance readings to accuracies of *_ 0.1% of F.S.D. taken in seconds; compact (6" x 2" x 3%"); lightweight (600 gms); elimination of high battery consump- tion by use of liquid crystal display; readings can be taken from a distance of 20 feet. For long-term accuracy, self test feature is provided. Further details can be had from Process & Control Elements 111-A, Hind Saurashtra Ind. Estate Andheri-Kurla Road, Marol Naka Andheri. Bombay 400 059 EXPLOSION PROOF SWITCHES For details, contact Jain Electronics F-37, Nand Dham Industrial Estate Marol, Bombay 400 059 UNIVERSAL MICROCOMPUTER DEVELOPMENT SYSTEM Burgess Micro Switch Co. Ltd., of UK have introduced a stainless steel version of explosion proof switches with aluminium castings. Main features are: resistance to corrosion from sea water and immunity from incendive sparking. The manufacturers recom- mend its use in marine applications More particulars can be had from. Thermal Sensors 37 A, Electronics Complex, Kushaiguda Hyderabad 5 00 762 From Philips comes the Universal Microcomputer Development System type PM 4422, designed as an effective aid in the software design of f india february 1 984 2-62 Grendon Underwood. Aylesbury. Bucks, HP 1 8 OS U Telephone: 029 677503 Toroids from STC transformers froi Cotswold Electronics L Kingsditch Trading Esti frequency rating range :ondary vol' nominal in ORDERING INFORMATION: 1) All payments in advance by M.O., P.O. or D.D. only. NO V.P.P. PLEASE. 2) boards will be sent by registered post parcel only. 3) Please add 15% to order value for packing and postage. Minimum Hs. 5/-. PRINTED CIRCUIT BOARDS E: 1 MAY 1983 PCB PRELUDE-MC/MM Phono Preamp. (MC PCB) 83022-2 PRELUDE-MC/MM Phono Preamp (MM-PCB) 83022-3 INTERLUDE-Remote Control Preamp 83022-4 7-0ay Timer/Controller 83041 PRELUDE - Tone Control 83022-5 Programmable Oarkroom Timer 82048 Rs. P. 16.50 21.50 14.75 16.50 15.75 15.00 E: 2 JUNE 1983 Watt Meter 83052 9.00 ASCII Keyboard 83058 76.50 MAESTRO- Remote Control 83051-1 5.50 Morse Converter 83054 8.75 78 L Voltage Regulators and 79 1 78052 2.50 E: 3 JULY 1983 Energy Meter s— 83067 9.25 RTTY Decoder 83044 8.75 Spectrum Display - Display board 83071-1 11.00 Spectrum Display-Filter board 83071-3 11.00 Spectrum Display - Control board 83071-2 11.00 PCB Flashing Running light 83503 Prelude Bullet 83562 Video Pattern generator 83551 R C Generator 83561 Heat Sink Thermometer 83410 Microprocessor Aid 83515 Simple D/A Converter 83558 Microphone Amplifier with Preset tone control 83552 Thermal Indicator for heat sinks 83563 6.50 2.50 E: 6 oct. 1983 Autotest 83083 15.25 Personal F.M. 83087 5.50 Alarm extension Transmittet83069-1 V 9.50 Alarm extension Receiver 83069-2 9.50 Stamp 80543 2.50 E: 7 NOVEMBER 1983 Basicode Interface for the Junior Computer 83101-2 Battery Eliminators— 83098 Music Quantisuer 83095 Cosmetics for FSK Signals 831 Uti Voltage Regulator 83088 2.50 4.00 10.75 9.50 5 2b SIMPLE ANEMOMETER Memory Board Measuring Board E:8 DECEMBER 1983 Ooorbell-or telephone -operated flashlight Power controller for metronome Power flasher PCB. Rs. P. 83103-1 12.00 83103-2 2.00 83104 6.75 83110 11.25 83107-1 8.25 83107-2 3.25 83ll4 2.75 78003 2.25 E : 9 JANUARY 1984 64 way bus board 83102 MF/HFVSB Marine Receiver 83024 Symmetrical Power Supply 83121 Video amplifier 83113 Frost warning device 83123 0ISC0 PHASER Delay line 83120-1 Oscillator and control 83120-2 49.50 35.25 13-75 11.50 5-75 13-75 8.25 E: 4/5 Aug./Sept Constant light source 83553 7.25 Car PDM Amplifier IS* 83584 9.25 eIeIctor eIectroni'cs PVT lid. 3, chunam lane dr d bhadkamkar marg, bombay-400 007. phones: 367459:369478. llebruary 1984 2-65 When reliability and stability are your prime criteria for selection... EClL’s Tantalum Capacitors. Today sophisticated electronic instruments and circuits require more compact, highly reliable and stable capacitors with a wider temperature range of operation and less parametric changes due to temperature variation EClL’s Tantalum Capacitors are designed to meet these exacting standards. Highly reliable, they find application in several areas including: * Defence equipment * Computers * Professional electronic equipment * On-board space modules * Process and industrial control equipment EClL's Tantalum Capacitor Type HST is approved by LCSO, as per JSS - 50205. Electronics Corporation of India Limited (A Govt, of India Enterprise) Passive Components Marketing ECIL Post, Hyderabad-500 762 Phone: 852231-258, 852569 Telex: 0155-254 Grams: ECIL HD MH/BY WEST-228 CDsmic India's / J I I 1 1 bi / India's pioneering electronics enterprise, ■ ■■ f ,ythe house of cosmic, has continuosly kept India S J' abreast of the latest technology to establish ^ the widest range in entertainment electronics . OV with Sales and Service centres all over India. / Amplifiers from 30 watts to 600 watts, with matching •) Hi-Fi Speaker Systems, Tape Decks, Tape Recorders, $ Turntables, Headphones, Pro-Power Cassettes, • • % Black & White TV, Colour TV and Videos. • | Quality, service and long-term value have satisfied • # audiophiles and encouraged them to re-invest in cosmic • Authorised Sales & Service O International quality created for India by cosmic