INDIA mm m. sin to large electrolytics 'concerns developments in the field of ICs. 6502 tracer A program to help understand or debug other programs. diesel tachometer Operating from a timing signal supplied by the alternator, this rev coun- ter can be used with diesel or petrol engines. programmable disco display lights Designed primarily for use in a disco, this 32-program light display is also suitable for parties at home or for advertising purposes. video combiner Based on a recently introduced 1C, the TEA 1002, we offer a design which combines the various components of a video signal into a com- posite signal. digitester with a difference By lowering the operating frequency of the circuit under test, this meter enables you to test digital circuits under operating conditions. reversing buzzer A small circuit to help prevent you inadvertently shortening your car. memory timing As a follow-up to last month's address decoding we now deal with the all-important control signals. mating logic families We show you how CMOS and TTL devices circuit. A new look for our cover this month which does not, however, mean a new look inside. We will continue to offer you in each issue a host of practical projects chosen from across the whole spectrum of electronics interspersed with articles of a less practical, but none the less interesting, nature in the field of electronics. Above all, we will continue to endeavour to keep up the high standards that readers of Elektor India have come Also new this month is the capacitance meter illustrated. This is the first in a range of test instruments which we will publish at (hopefully) regular intervals. All instru- ments in this range are designed to fit the same family of standard cases and, as our cover shows, they even look quite handsome! We also start a new section this month: 'Chip Selekt’. This section will be published at irregular intervals, alterna- ting with 'Applicator', and is intended to bring to your attention new or improved ICs which have recently been supplied to us by manufacturers or dis- tributors. be used together capacitance meter A relatively inexpensive meter varicaps ' enable you to check capacitors from large electrolytics. Basicode 2 for Junior plus VDU card Many J.C. users will find this combination of Basicode 2 and the VDU card very interesting! constant voltage source Th is cleverly thought-out circuit keeps’ the’ light ’ intensity' of 'a i^nip substantially constant over the life of the battery. CS chip selekt Some new, restyled, and up-dated ICs 'recently received by us’, some of which may not yet be commercially available. video sync box Unlike many similar designs, this one generates not only all sorts of video signals, but also an optional colour bar. automatic battery charger Market switchboard missing link A selection from next month's issue: • real-time analyser • UHF quartz modulator • tape contents controller • electronic labyrinth • triac control board • petrol saver EPS service advertisers index choosing the right friend makes all the difference Two people become friends because they sense in each other a strength and trust that can help them both. In today's jostling realm of microprocessors, we at DMS offer our customers a serious and courteous friendship that stands the test of time. T ake for example, our MICROFRIEND series pP trainer system an ideal combination of friendly firmware & Hardware, designed not only to work for you but with you. All the way. Like a friend should. As an ideal pP training system, our 'MICROFRIEND based on 8085A & Z 80A are the only trainers available today in the market offering lot of outstanding features incorporated on single board like PROM Programmer, RS232C I/F, STD Bus, 46 I/O lines. Audio Cassette I/F backed-up by 4K Powerful system Monitor, 2K CMOS User RAM and onboard expansion with 6 more 24 Pin sockets using 2716/6116 or 2732 & many more. If 16 bit trainer is your requirement, you are just in time. Our MICROFRIEND 86 is now ready to roll into the market with the advanced performance and highly innovative features you've been expecting from DMS. Our creativity inpP based products has set the pace in the industry from the start. If, in this every-man-for himself business environment, our 'MICROFRIEND' sounds too good to be true, try him. Your friend in-the realm of microprocessors. Write or call now for complete information. Dynalog Micro-Systems Head Office : 14, Hanuman Terrace, Lamington Road. Bombay 400 007. Tel : 36 24 21 Cram: EbMADEVICE Telex: 11-5614 Factory: 17 A/2, Karve Road, 30. Prasad Chambers, Pune 411 004. Tel: 35983 Cram: ELMADEVICE Telex : 0145-505-EIMA Bangalore: 15, Hotel Kamadhenu Annexe, Trinity Circle, M.C. Road. Bangalore - 560 008. Marching along the trail in uP based technology. C FOR ACTIVE ELECTRONIC COMPONENTS RECTIFIERS • Ord DIODES • Low ZENERS • Voll THYRISTORS • Ord TRIACS • Pla: DIACS • Gla TRANSISTORS • Pov RELAYS TRIMPOTS HELIPOTS TANTALIUM CAPACITORS A/C- D/C Ir Low Ohms PHILIPS. MELTRON. NAYAR ELECTRONICS. ECIL & HI NO OS TAN 7, KAMAT BUILDING, SHAMRAO VITHAL MARG, LAMINGTON ROAD. BOMBAY-400 007. TELE-385135, 367601. INTIMATE FRIENDS OF AN ELECTRONIC ENGINEER Diamond ADH ESIVE DESOLDERINC Jh reS eal PARVAiL ENTERPRISES IMPORTANT SUBSCRIPTION RENEWALS ARE ON! HAVE YOU RENEWED YOURS? * To use the card in this issue * To quote your subscription number * To mention that it is a renewal * To send your renewal at least one month in advance * Timely renewal ensures continuous receipt of issues without irritating breaks 3-08 elektor in From ECIL India's first miniature metal film resistor at an amazingly low price ECIL now claims another first by Specifications introducing the lowest priced resin Value range coated metal film resistors. Though Standard value inexpensive, these unmoulded resistors meet the most stringent industrial and p ower ratinq defence requirements. Customers will Sj2e also be happy to know that the E-24 stability series is available ex-stock. For further enquiries contact: 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 100 to 1 meg ohm E-24 and E-96 series 200 and 100 ppm/°C 0.25 W at 70° C 6.5 x 2.3 mm NOW THE LATEST BOOKS-KITS ARE AVAILABLE WITH US| DATA BOOKS : signotics SEMICONDUCTORS NATIONAL © iicji _ Texas Instruments MOTOROLA INTERSIL FAIRCHILD ON I.C.'S, TRANSISTORS. TRIACS, SCR'S, DIODES. OPTO-DEVICES KITS : EDUCATIONAL. HOBBY. ENTERTAINMENT, AUDIO, VIDEO, CLOCKS PLEASE WRITE FOR DETAILS ELTEK BOOKS-N-KITS 6, RITCHIE STREET, 1ST FLOOR, MOUNT ROAD, MADRAS-600 002 CALONIX PROFESSIONAL GRADE MINIATURE SWITCH ES 1. TOGGLE 2. ROCKER 3. LEVER TESTED & APPROVED LIFE GUARANTEED AMERICAN KNOW HOW JSS 51201 SPECIFICATION 'kating & Exports Di CALONIX ELECTRONICS PVT. LTD - DEALERS TO BE APPOINTED. _ WE ARE ALSO REPRESENTATIVES FOR QUALITY ROTARY SWITCHES MANUFACTURED BY COLE INSTRUMENTS CORPN. U.S.A. IMPORTS AGAINST AU - LICENCE ZC1A PLOTTCRS 'TifL TECHNICAL TRADE LINKS Ln_J APLAB’S IMPORT REPRESENTATION GROUP 42. Navketan Industrial Estate. Mahakali Caves Road. Andheri (East). Bombay - 400093. Phone : AN 8 COLOR PEN PLOTTER WITH THE SPEED AND PRECISION OF FAR COSTLIER MACHINES. SALIENT FEATURES : 1 . High performance plotters with built in intelligence. 2. Fast Plot turn around with no time outs. 3. Vivid, multicolour business graphics. 4. Scientific and engineering photo without peer. 5. Eight colour, carriage - mounted pens, eliminate time - consuming pen changes. 6. Supported bv the leaders of the software industry. 7. The most flexible - D’ format 8-pen plotter ever. 8. Plotters as sophisticated as the computers, they are designed to support. 9. A plotter anyone can operate. 10. The plotter everyone will want to use. m Nicolet 3-10 Multiple Choice Test from Philips Now you have a wider choice in Philips oscilloscopes, so you can better tailor the scope to your application. Need a 15 MHz/Single time base scope? Choose the PM 3226. A 50 MHz with delayed time base? Choose the PM 321 7. The new 35 MHz scope PM 32 1 8 has a delayed time base, maximum sweep of 10 ns/div. and a trigger hold -off facility that eliminates double triggering on digital signals, making it unnecessary to use the time base in the uncalibrated mode. The PM 3262, 100 MHz dual trace oscillo- scope not only permits the display of high speed current mode logic signals but due to its extensive features, it permits you to take measurements in any electronics environment. Now. more than ever, we want to make your next scope. And your next digital & anolog multimeter. And your Timer/Counter System We are going to give a wider multiple choice to make your selection even For more information contact : Philips India Test & Measuring Instruments Division Plot 80, Bhosari Industrial Estate PUNE 41 1 026 IphiupsI Test & Measuring Instruments PHILIPS Philips — the trusted Indian household name for over fifty years. OB M/8700 R-2 COLOUR TV AWARD The Electronic Component Indus- tries Association (ELCINA) has announced a new award this year for the best Indian designed co- lour television set with the maxi- mum indigenous content. The win- ning design team will be awarded a first prize of Rs. 25000-00 and the firm supporting and funding the project will receive a trophy. There is also a second prize of Rs. 10000-00. Last date for receipt of entries is 30th April 1984. The Association's other annual awards are for R & D, import substitution, export promotion in the field of el- ectronic components and indigeni- sing of capital goods for the manu- facture of electronic components (inclusive of in-house efforts). Since last year these awards nave been thrown open to non-mem- bers also; last date for receipt of entries is 1st March 1984. Distribution ceremony of the awards for 1983 will be held in the last week of March 1984. The add- ress of the Association is: Electronic Component Industries Association 408, Sahyog, 58, Nehru Place New Delhi 110019. Tel.: 681397; cable: ELCINA INDO-JAPANESE DICTATION SYSTEMS Sony Corporation of Japan has joined hands with World Business Machines of New Delhi, to intro- duce microcassette dictation sys- tems into the country. The sys- tems, which are to be marketed under the brand name 'World Memo 120’, include an IC-control- led dictation machine and a speed adjustable microcassetie scriber. According to Mr D.S. Sahaney of the Indian company, the systems will be the first in the country to have additional operation features like 'pause, quick erase and fast forward'. The technical arrange- ment with Sony will allow the Indian company to import the latest microcassette technology from Japan. ONGC USES INSAT -1 B Oil & Natural Gas Commission has started making use of its satellite earth stations at Uran and Bombay High for data communication through INSAT-1B. The earth sta- tions were set up jointly by the Post & Telegraph Dept., and the Radar & Communication Project Office (RPCO) of the Ministry of Defence which is executing pro- ject TITAN for ONGC. With the establishment of the link of the TITAN project, the activities of the BHN platforms and the BHS plat- forms are being monitored by the Uran field control centre. A com- puter installed at the Bombay High South (BHS) platform acquires telemetry data relating to oil and gas extraction and production in the BHS complex. This data is passed through a microwave link to BHN platform from there thro- ugh INSAT IB to Uran and finally through a UHF link to the Bombay Offshore project headquarters in South Bombay EDS SYSTEM AGREEMENT SIGNED The Union Government has signed an agreement with CIT-Alcatel, a French public company, for sett- ing up factories to manufacture five lakh units of electronic digital switching (EDS) system in a year to improve the telecom network in the country. At a recent press con- ference, Mr V.N. Gadgil, Union minister of state for communica- tions, said that one of the factories would be set up in Bangalore and the other in Gonda in U.P., and expressed the hope that produc- tion in both the factories would commence within one and a half PRIORITY FOR ELECTRONICS Keeping in view the immense growth possibilities of the electro- nic industry in the country, the Union Government is to accord top priority for the development of its infrastructure. Its export perfor- mance too was such as to create sanguine expectations for the future. Commenting on these trends. Dr. M.S. Sanjeev Rao, Union deputy minister for electro- nics, told a recent seminar, on electronics held in Ahmedabad, that the potentialities of exports of electronic goods and systems were inexhaustible. In 1982, the country exported electronic goods worth Rs. 89 crores; experts opine that it could be three times more in 1983. The rate of the industry’s growth too has been phenomenal. Starting with a total production of Rs. 65 crores in 1965, it went up to Rs. 365 crores in 1975, Rs. 802 crores in 1980 and reached Rs. 1306 crores in 1982. The growth is likely to gain further mo- mentum, especially in view of the governemnt's decision to extend TV coverage facilities to more than 70 per cent of the country's popu- lation in 1984/85, from the present coverage of only 30 per cent popu- lation, by setting up 140 new trans- mitters and linking them through INSAT satellite. To cope with the problems arising out of the sud- den spurt of grwoth, the govern- ment will lay a firm foundation for this infant industry by way of indi- genous production facilities of basic chemicals, components and other raw materials, besides provi- ding large industrial estates in the states. Further, it has plans to sup- port long-range research prog- rammes both in universities and institutes of technology, and also set up three major computer-aided electronics design institutes for application of chips and develop- ing circuit designs. IMPACT OF CONCESSIONS The fiscal concessions given to the electronic industry by the Government in August last year have started showing encouraging results in that price reductions ranging between 10 to 35, per cent in components have been effected Items of which prices are reduced are picture tubes, semi-conductor devices, magnetic sound heads, plastic film capacitors, etc. These price reductions have, in turn, resulted in appreciable reduction in the prices of certain end pro- ducts like B/W TV sets, cassette recorders, stereo decks, etc. Among the companies which have reduced prices of their products are Teletube Electronics Pvt. Ltd. (picture tubes), Bharat Electronics Ltd. (picture tubes, germanium transistors, silicon signals), Peico Electronics & Electricals India Ltd. (transistor sets, cassette recorders, speaker systems, components, SOFTWARE CENTRE Hindustan Computers Ltd. (HCL) has established an advance soft- ware development centre in Mad- ras, so as to provide complete software and training support to its customers. Equipped with four computers, the centre has pio- neered the use of computers to generate other computer prog- rammes. A special team of Indian experts is responsible for this breakthrough in software techno- logy. The centre provides training in computer operations and langu- ages to users, tests hardware and software items prior to their intro- duction in the field and develops special programming packages to suit particular industries. future developments in ICs Industrial developments are governed by economic feasibility on the one hand and physical possibilities on the other. This is no exception in the 1C industry. As yet, there are no fundamental limits, as far as physics is concerned, to further reductions in the size of ICs. It follows that, for the time being at least, development in the field of ICs is governed totally by ecomic aspects. Four cost factors play a role in determining the price of ICs: those of the crystal, development, testing, and packaging. At this time these four are about equal. There is, however, a tendency for develop- ment costs to take a larger share of the total, unless special attention is paid to this. With no natural limits in sight (as yet), new methods to produce smaller structures are found continu- ally. This is well illustrated by the development of substrates which can be processed, and the width of the tracks which can be realized on them. The diameter of substrates has grown in steps: it is already clear that pro- cessing of 6 in (15 cm) crystals is physically possible. The tracks indus- try can produce are getting narrower and narrower. In the early 1960s track-widths were of the order of 100 pm. Present-day techniques al- low widths of 4 ... 2 pm to be realized (figure 1). The tendency towards larger sub- strates and narrower tracks means, of course, a sharp increase in the capital cost of manufacturing tools. The equipment required for the production of 20 000 chips per month cost between £ 3m and £ 4m in 1980; by 1990 those figures will have trebled. But by then, the chips will be larger and more complex and therefore worth more. Complexity The factor which has the greatest effect on costs is the component density: the number of components which can be housed on a single crystal. As the track -width is reduced by 1 0 per cent each year, the surface area of a transistor decreases by 20 per cent per year. Because the individual connections will occupy relatively more space, the average component density rises less rapidly than one would expect from these figures: about 1 5 per cent per year . On the other hand it is possible to produce larger chips with a satisfac- tory turn-over. Each year the producible chip area increases by about twenty per cent, ement becomes more and more which means that in principle the specific and it therefore becomes complexity of the designs can increase harder to find units in equipment by 37 per cent. This possibility is which have enough applications in fully utilized in certain sectors, such common to guarantee sufficiently as memories. There is no reason why large production runs (figure 2). ever larger building elements should It is for that reason that in practice not be used in the construction of the average complexity of newly memories. In many other applica- introduced building elements in- tions it is not so simple. With in- creases by only twenty per cent creasing complexity, a building el- per year (figure 3). 3 . [ Applicability | The required area per transistor, including connections, is reduced by about 13 per cent per year. In spite of the higher capital costs mentioned earlier, the production costs per unit of area fall by about ten per cent per year. This is because larger substrates are being used. The combination of these two factors means that the cost of the transistor itself drops by around twenty per cent per year. The other cost factors should show a similar tendency. The problem is attacked in two ways: firstly, by increasing the productivity of designers and, secondly, by aiming at an increase in production per design. The latter means that more effort is required towards standardization and the use of universal components instead of specific ones. Design productivity can be increased in various ways: in the first instance by the use of Computer Aided Design (CAD) in which the computer is used in practically all phases of the design process. Planning Furthermore, by changes in the design planning. In principle, the 1C designer has a large degree of free- dom in virtually every detail of the design. Making profitable use of this freedom takes time, however. Limiting the possibilities of the design increases efficiency, but will, of course, often mean that the design is realized at a higher abstract level. Freedom in the design can be further limited by planning a chip on. the basis of a pattern of a large number of standard basic functions, where the designer can only determine which circuits are connected and On the other hand, it is necessary that the productivity of the designer increases. Greater complexity renders the designs more and more specific to certain applications. This results in a larger need for diversity: in other words, a growing number of more complex circuits will have to be designed. Designs It is to be expected that the design community — which is independent of the 1C producers — will grow rapidly. Design firms will come into being to design ICs which can be produced simultaneously by a number of manufacturers. If, for instance', the design task is the development of a new family of ICs, the complexity may be such that in practice it can- not be realized by a single company in an acceptable period of time. Joint developments by a number of companies will therefore become commonplace. Miscellaneous cost considerations The increasing component density will also be a factor in the cost of testing a transistor. The complexity of ICs can become so great that it becomes extremely difficult to test individual transistors. The result may be an explosion in the cost of testing. This problem is tackled in several ways: by modifying test strategy, by making more allowance for testing during the design stages, and by introducing special test circuits onto the chips. The last two measures generally mean that test costs are exchanged for higher crystal costs. It is expected, how- ever, that by adopting these measures the cost of testing will not rise, and may actually show a slight fall. The price of packaging is virtually constant. As the contents of the chip become more complex, the number of pins will inevitably in- crease which means a rise in costs. Intrinsic material costs will, of course, rise as well. However, auto- mation of the packaging processes in manufacture can compensate completely for these cost increases. All the same, the apportioned cost of the package in the price per transistor will decrease by about twenty per cent per year, because the number of transistors per package increases. Summary Undoubtedly there are many chal- lenging tasks in store for the re- searchers. It would be regrettable, if only from the point of view of available resources, if national govern- ments in Europe would engender local, subsidized 1C industries which would all develop similar products. In this field also, European co- operation is of the utmost import- Extract from Philips press notice 3-15 J. Ruppert Being able to see what a processor does as it runs a machine code program is a great aid in understanding the program, in fault finding, in testing, and in fact in everything a programmer does when developing some new software. The program given here makes it possible to do this automatically. At each step the contents of the CPU registers, the stack and its pointer are displayed for the corresponding instruction. 6502 tracer program analysis software for Junior Computer and other 6502-based systems Table 1. 6502 TRACER is an analysis program that must run in RAM. but you from storing it in some other kind of transferring it to RAM to This program is aimed not only at users of the Junior Computer but also at the owners of any 6502-based system. It oc- cupies about Vi K of memory and uses two bytes in page zero. Very few changes are needed to adapt it to a system other than the Junior. How is it used? The program operates as a son of ‘step by step monitor'. This means in effect that any program the user wishes to analyse, or debug, is executed instruction by instruc- tion with the contents of registers A, X, and Y, the status register flags (NV DIZC) and the stack pointer being displayed each time. It is notable from the list of flags (NV DIZC) that the 'break' flag is not included; the reason is that the '6502 TRACER’ program accepts all in- structions except those which are the result of, or which result in, an interrupt (BRK, IRQ and NMI). As table 3 shows, it is much easier to analyse a program (the example here con- tains a lot of register and flag manipu- lations) with the aid of the information displayed by the tracer program in the three right hand columns. The first, at the extreme right, refers to the stack: $FF is the least significant byte of the pointer (the most significant byte is $01). Near the end of the listing there are a few ad- dresses stacked during JSR or RTS instruc- tions. The next column gives the logic levels of the status register flags NV DIZC. Finally, beside this the contents of the A, X, Y and processor registers are to be found. The step by step tracing of the program in these columns is followed in the first two columns by the disassembled listing of the addresses and instructions. The fact that all jumps and branches are included explains why the program returns from address $020D (DO/FA) to address $0209 but the Z flag remains low. How does it work? The length of this article does not give us the scope to provide a complete source listing of this tracer program, so we will have to be content with the hex dump shown in table 1. It is, however, quite im- portant to have some pointers about how to use the software. Before a run the start address of the pro- gram to be tested must be stored at ad- dresses $00ED and $00EE which act as a pseudo program counter. The program under test may be in back-up memory but the tracer program must be in RAM: as shown here it starts at address $0500 Between addresses $0500 and $0523 3-16 several buffer bytes acting as a pseudo stack that starts at $0713 (we will return to this later) are initialized, the column headings are displayed and the IRQ vec- tor is positioned (the IRQ routine begins at address $0526). The tracing proper starts at $05A2, by displaying the program counter address, loading the op-code, filling the op-field with 00s, and calculating the length of the instruction (the routine used begins at $06A8 and is quite similar to the LENACC routine in the Junior Computer). The op- field is a four-byte zone ($0619. . .$061C) where the analysis program places in turn each of the instructions of the program under test in order to execute them. As these instructions never contain more than three bytes they are always followed by at least one 00 and this functions as a BRK. Immediately after executing an instruction of the program under test, therefore, this BRK causes the IRQ routine at $0526 to be The pseudo program counter ($00ED and $00EE) is incremented at $05DB. This in- crementation depends on the format of the preceding instruction, with the number of bytes making up the instruction being stored in address $071E. Any jump instructions in the program must be filtered out to be dealt with separately and this begins at $05E6. From $060B onwards stacking of registers A, X and Y for the program under test starts. The op-field, located at $0619, contains the instruction to be analysed and because every instruc- tion is always followed by at least one BRK it is also followed immediately by the IRQ routine. As could be expected, this begins by storing the conditions of the processor registers. Then it displays their contents and proceeds to the next instruction. The special instructions for executing jump commands are located at $06 ID The addresses for relative jumps are calculated at $0672 and $068A. The ad- dresses of the Junior Computer’s PRBYT and PRCHA routines are contained in $06A1, $06 A2, $06A6 and $06A7, so these must be changed if the program is to be used with a different 6502 system. The commands for printing the headings of the columns are at $06CC to $0702. The format of each instruction that is to be run is determined by comparing it to the values contained in the look-up table located from $0703 to $0712. There are a number of buffers between $0713 and $0721 that are used by the tracer program to store the stack pointer, the contents of the top of the stack, the op-code under test, the number of bytes in the instruc- These were the most important points about this program and the rest is easily deciphered with the aid of a dis- assembler. M 3-17 A tachometer is probably the single most important 'meter' in a car's dash- board (unless you're in the habit of running out of fuel). It informs the driver how hard the engine is working and, when used correctly, it is an aid to economy, performance and engine longevity (to name but a few). These things are, of course, no less im- portant to the driver of a diesel car than to his petrol-powered counter- part, but most electronic tachometers cannot be used with diesel engines. The reason for this is that they take their 'timing' from the ignition circuit, which diesel cars do not have. The engine speed of a diesel car can, however, be measured by 'picking the brains' of another part of the electrical system, namely the alter- nator. diesel tachometer connected to the alternator measures r.p.m. in virtually any (12 V) diesel or petrol engined car The difficulty with fitting a tachometer to a diesel engined car has not escaped the car manufacturers’ notice. Many diesel cars sold today have an extra (so-called ‘W’) connection available on the outside of the alternator, and the purpose of this is to enable the engine speed to be measured without unnecessary complication or cost. Petrol cars are, of course, not a problem as the tachometer timing is conventionally taken from the ignition system (the contact breaker pants). The diesel engine, however, does not use spark plugs to ignite the fuel/air mixture and this is the root of the problem with fitting a tachometer to a diesel car. Some different value must be found, there- fore, that is directly proportional to the engine speed. This should preferably be an electrical value to make connection to the electronics easier. The ever-present alternator looks like a good possibility. Be- cause it is driven from the crankshaft via the fan belt it turns at a speed which is directly proportional to the engine speed. The ‘circuit’ of an alternator is shown in figure lb, and this is typical of the layout used in the vast majority of modern cars. The diagram shows that a 'pick-up' for the engine speed need only be connected to one of the points U, V or W. Most manu- facturers select W and feed this to the outside of the alternator. When a car is available with either petrol or diesel engine the alternator is generally the same for both versions, so even petrol cars often have the W connection available at the outside of the alternator. If your car lacks this W connection don’t panic, under the section ‘W connection’ we will return to this problem to show how such a connection could, if needed, be made. At the input of the circuit diagram of figure la we see roughly what the signal taken from the alternator looks like. The actual form of the signal is unimportant; what is of interest is that the frequency of the signal depends on the engine speed. This frequency is from about 125 Hz to 1250 Hz depending on the type of car but variations can be taken care of by our circuit. Having got a signal at the input, all that remains is to convert the frequency variations at the input into voltage variations, which brings us to the circuit of our tach- ometer. The circuit As the circuit diagram of figure la shows, this tachometer contains nothing compli- cated as regards electronics. The supply is taken from the car battery via R1 and protection diode Dl. The input resistance and input current (1.5 mA maximum) are defined by resistors R2 and R3. The pulse signal coming from the W connection is limited to 12 V by means of zener diode D2. Any high frequency noise that might be present is shorted to earth by C2. The signal is then fed to the inverting input of op-amp IC1 which operates as a schmitt trigger. The hysteresis of this schmitt trigger is about 6 V, and the signal at its output (pin 6) is a rectangular waveform with an amplitude of 6 Vpp and a frequency corre- sponding to that of the input signal. The signal oscillates about the 6 V line. Differentiating network C3/R8 converts the rectangular waveform into the ’peaked’ signal shown at the junction of these two components. The positive peaks are limited to about 0.65 V by D3. The negative peaks, on the other hand, are used to trigger MMV IC2. The width of the output pulse from this 555 can be varied with PI between 1 50 and 450 ns. The output signal from IC2 is limited to 5.6 V by zener D4 and is then integrated by R1 1 and C6 before being fed to the moving coil meter Ml. As a result of this integration, and also to a certain extent because of the inertia of the meter, Ml gives a stable read-out of the engine speed. 3-18 eleklor in Figure 1b. Most modern configuration shown here of six diodes. Construction and calibration The printed circuit board layout for this circuit is shown in figure 2. The connection points from the circuit have purposely been made big so that the normal automotive type connectors and push-on clips can be used. No holes have been drilled in the board for the meter connections, but large copper areas have been left for this purpose. Holes can be drilled to suit the type of meter used, and the board can then be fixed directly to the rear of the moving coil meter. It goes without saying that the meter must be connected with the right polarity. The meter needs a suitable scale, of course; this could be made by using one of the dry transfer systems available. There are three possible methods of cali- brating the circuit (no, we don't mean do-it-yourself, get somebody else to do it, or don’t do it at all). The handiest method is to use a hand-held tachometer, which can probably be borrowed from a garage (if you grease the right palms). If you also enlist the help of an assistant things will be considerably •-peeded up. Run the engine at about 2/3 of its maximum speed. Your helper measures the engine speed at the crankshaft with the borrowed tachometer and tells you what the reading is. You then adjust the Elektor tachometer to this value with PL The second method of calibration involves a bit of arithmetic but in this case no refer- ence tachometer is needed. Knowing the rpm/mph ratios of the car in various gears enables you to calculate the engine speed corresponding to a certain road speed in a certain gear. Then find a straight level road and drive at a steady speed for which you have calculated the engine speed. Your (indispensible) helper now adjusts the tachometer to the appropriate reading. The disadvantage of this method is that you are using the car’s speedometer as a reference so the reading is almost certainly going to be a few percent incorrect. The third calibration method involves measuring the diameters of the pulleys on crankshaft and alternator carefully and thereby calculating the ratio of engine speed to alternator speed. An example of this is given in figure 3. From the technical Table 1 data about the alternator, the ratio between the rotary speed and the frequency of the ‘W’ signal can be worked out. If, for example, it is a 12-pole type then the frequency is exactly six times as high as the speed. An example of such a calculation is given in table 1. The tachometer can now, on the basis of this information, quite simply be adjusted by feeding in a signal from a sine wave generator with an amplitude of about 14 V. The 'W' connection Alternators that do not have a W connection as standard can often be modified using special adapter sets. (Bosch, for instance, market a kit, no. ET-1 127 011 062, for VW and Audi diesels.) The best thing to do is look at the make and type of alternator and then ask at the appropriate garage if an adapter kit exists. An adapter kit is not, however, strictly necessary. The rectifier in the alternator generally conasts of a six -diode bridge as shown in figure lb. Points U, V and W are all located at the anode-cathode junction of two diodes. For our purposes there is no difference between the points and you can feed any one of them to the outside. Using the tachometer We are not, of course, going to tell you how you should drive, but nonetheless it may be no harm to see how a tachometer (any tachometer) can be put to its best use. A lot of information about a car’s engine can be gleaned by looking at graphs such as those shown in figure 4. These show the relationship between engine speed and both power and torque for a common diesel car, the Volkswagen Golf. The engine speed ranges from about 1000 to 5000 rpm. As one of the curves shows, the power rises almost linearly with (engine) speed up to about 4000 rpm. After this the power does not rise at the same rate so acceleration will be less. This is very important to know for overtaking, for example. Torque is also dependent upon engine speed, but in this case maximum torque does not correspond to maximum engine speed. The engine is at its most efficient, and most economical, at maximum torque. This fact is used every day by people who wish to drive economically. It is a common fallacy that only racing drivers need a tachometer. Certainly those for whom high speed driving is a profession do place a great importance on the informa- tion they get from the tachometer, but so can every driver on the road. Mechanical suffering is becoming more and more diffi- cult to hear in today’s well sound proofed cars, or at least that is the plea of the (ob- viously cloth-eared) driver who has his car ticking over at much too high a level and insists on thrashing it before it is fully warmed up. If you see him give him the message - don’t put the pedal to the metal when the car is still cold. After all, can you work properly before you're well awake in the morning? M 3-20 elektor in The essential ingredients for any self-respecting disco are well known, plenty of the right type of music, an abundance of coloured lights and gently top up the remaining space with people. Any hi-fi will provide the music but the light display is rather more specialized. There are of course many variations on the theme ranging from the mediocre to 'way-out' and invariably classified by price. The disco light display described in this article is a very advanced design with many desirable features but can still be built at a very reasonable cost. programmable disco light ^ display Light displays are very feZE* popul. for many appli- cations other than discos. They are excellent in the ; home, for instance, as a means of providing ‘atmosphere’ at \ ^ parties or social gatherings. They Tm* are also very useful for advertising purposes for the enterprising businessman. It cannot be denied that the more ’ interesting a disco light display is, the more complex the electronics tends to be. This is mainly due to the fact that each light source, in most cases a mains powered lamp, must be controlled seperately, resulting in a ‘channel’ usually consisting of some form of logic decoding, a mains interface, and a firing circuit. This channel must then be duplicated for however many lamps are involved. Regretfully, we have not been able to do away with this problem. Ironically, however, it may be seen as an advantage simply because it allows for the easy expan- sion of the overall system - especially if the control electronics are designed with this in mind! It will become apparent that the circuit in this article can be as large as your imagination or your wallet allows! A major disadvantage of the average disco light display is that the available light patterns are an integral part of the control circuit, possibly the contents of a memory IC, which must be purchased. This means that the pattern cannot be changed very easily, if at all. At this point we can start to sing the praises of the circuit here because this disco display is fully programmable. Furthermore, program changes can be made at any time by simply operating switches (no IC changes). The circuit also contains its own memory allowing up to 32 different programs to be stored. There are numerous other highly desirable features of the circuit that put this disco H. Theunissen with up to 32 programs in memory different level from the OjVjJr average - including most commercial units. This list of do’s and don’ts explains all . . . ■ Entirely user -programmable at any time. ■ Up to 30 channels can be accomodated. ■ Program selection can be run fully automatically or manually. ■ 8 switched program run times available. ■ Internal memory divided into: 16 (2 ‘banks’ of 8) programs of 128 steps, or 32 (4 ‘banks’ of 8) programs of 64 steps. ■ Overall size of memory optional. ■ Battery back-up for memory. ■ Programs, banks and current memory address indicated by LED displays. ■ Opto isolation from mains. ■ All lamps switched at zero-crossing point of mains to reduce interference. ■ Personal choice of display configuration (a matrix configuration makes possible a display with 225 lamps! ) So much for what the circuit can do, now how about what it doesn’t do! ■ It doesn’t cost an arm and a leg. ■ It does not require any programming skill. ■ It does not require a great deal of practi- cal ability to build it. ■ It doesn't play the Hokey Cokey (although some may not consider this to be a major disadvantage) yet! To sum up then, the circuit contains all of the desirable features (that we could think 3-21 of at least) and yet it can be operated without 'computer' experience. The com- plete display’ can be as small or as large as desired; it may even be expanded at a later date. Basic principles Those readers who have already sneaked a quick look at figure 3 (this of course in- cludes everybody) may be getting somewhat alarmed at what most articles would refer to as ‘a slightly complex' circuit diagram. This impression is just a figment of the imagin- ation as can be proved with the aid of the block diagram of figure 1 . Since it is the memory that holds all the information, this forms the heart of the circuit and all other 'blocks’ either feed to or from it. The structure of the contents of the memory is illustrated in figure 2. It will be seen that it is divided into 'banks’ (two or four depending on desired memory size) each of which in turn is divided into 8 pro- grams. This simple method allows the total memory to be divided into reasonable program lengths and provides an excellent means of finding any given program quickly especially if the program and bank coun- ters are given 7-segment display read-outs! The address counter, as its name suggests, determines the address of that part of the program which is being displayed at any one time. Obviously the same can be said of the bank and program counters. The block with the elegant title of 'mains sync' is a shade more subtle in both its activities and its purpose in life. Basically it provides a synchronization signal for the circuit at the frequency of the mains supply. Simple, you say - but wait. It also ensures that the clock signal is synchronized to the zero-crossing point of the mains frequency and, by so doing, it eliminates the need for all those zero-crossing detectors that usually accompany each lamp-switching triac in the mains interface of the display. The answer to the next question is that, since the clock is synced to the zero-crossing point of the mains, all data changes at the output of the memory will always occur at the same point. The lamps will therefore always switch on (or off) at the zero-crossing point! One further point before we leave the block diagram. The printed-circuit board designs for the display drive circuits do not appear in this article but they should grace the next issue. The circuit diagram The mains zero-crossing point detector is formed by IC1 (gates N1 . . . N3) in the circuit diagram of figure 3. The mains supply is present between the X and Z terminals and is applied to N1 via a voltage divider consist- ing of resistors R1 . . . R3. The inputs of N1 contain two diodes which chop the wave- form of the mains supply to provide a square wave with an amplitude that is equal to the supply voltage of IC1. The output of IC1 is differentiated by means of C1/R5 and C2/R6 and fed to the two inputs of N3. The resulting output of N3 is a pulse of about 200 ps at every zero-crossing point of the mains frequency. This pulse train is then fed, via a driver transistor, Tl, and an opto-coupler, IC2, to the clock input of FF1. This ensures complete isolation between the mains supply in the zero- crossing detector stage and the rest of the circuit. It is for this reason also that the power supply - connected between X and Y - for the detector stage is derived from the triac control board. The memory address counter is IC7 which will increment the address data by one at every clock pulse received at its clock input at pin 10. The clock signal for this purpose is generated by means of the variable fre- 2 by the user. Function table SI: A - RUN MODE 8 - PROGRAM & STEP MODE S2: STEP (increment address counter) S3: BANK increment (+1 ) S4: BANK automatic increment ON/OFF S5: PROGRAM RUN TIMES in minutes S6: MANUAL PRO- GRAM increment (+1) S7: DATA WRITE S8: WRITE PROTECT (key switch) S9: Mains switch S10: RESET switch S1 1 . . . S40: DATA switches PI: RUN speed control 3-22 ekHuorin programmable disco light quency oscillator formed by gate N4. If, for S8. Switch S8 is a safety ‘lock-out’ key dis P |a v instance, the display pattern were a running switch which, although not absolutely light, the speed at which the lights run could necessary, is strongly recommended to be increased or reduced by PI. However, the prevent accidental damage to a program, clock signal is not fed directly to the address How and when to use S7 will be covered a counter but via FF1 which, you will remem- little later on. ber, is itself clocked by the zero-crossing The power supply for the memory ICs point detector. The end result is that any is taken from the 5 V line via diode D2. change in the address counter is directly should the 5 V supply fail (when the equip- synced to the mains zero-crossing point. m ent is not in use for instance), the 4.5 V Switch SI is included to allow the address battery will preserve the contents of the counter to be ‘stepped’ by means of push- memory via D3. At the same time, the button S2. This is of course necessary during absence of the 5 V supply will switch off programming. transistor T2 and inhibit the memory One half of IC8 (IC8a) forms the program outputs by causing the CE inputs of the counter which has a continuous count-up memory ICs to go to logic 1 via resistor cycle from 0 ... 7. That is, it counts up R24. In short, the memory will be in- 8 steps (8 programs) and then resets to 0 operative (in the low power mode) but the only to begin the cycle again. The program contents of the memory will be preserved, counter is clocked by the program timer, The current consumption is so low in this IC9, which provides 8 different program condition that the battery will last quite run times ranging from 7.5 seconds to 16 literally for years but should still be changed minutes, selected by switch S5. every 12 months or so. A NiCd (three cells The program counter can also be incremen- of 1-2 V) can also be used in which case ted by one step at a time by means of switch resistor R18 (270 fl) is required to provide a S6 which, incidentally, overides the timer charging current. This resistor is not needed output. It should be realized that if S5 is with a dry battery. switched to one of its off positions, any Each of the data lines of the memory ICs is given program will run indefinitely until it fed to the LED in an opto-coupler on the is changed manually by S6. triac board via a driver, N15 . . . N45, and an The remaining half of IC8 (IC8b) forms the indicator LED. The indicator LED provides bank counter which, depending on the pro- a direct read-out of the data at that particu- gram size, continually counts up in either 2 lar address. This is of course essential during or 4 steps. This counter can also be stepped programming. The data lines are also fed to manually by means of pushbutton S3. To the programming switches S10 . . . S40 via . obtain fully automatic operation, that is, resistors. When S7 is pressed, and S8 is a continuous cycle through all the programs switched on, the data set by these switches in the memory, switch S4 can be closed is written in the memory at the address and at the highest program count, the indicated. bank counter will be incremented by one. One final detail before we leave the circuit It will be seen that the counters for program diagram of figure 3. The DO output of IC10, and bank are interconnected via an OR gate, switch S10, and the associated LED (via N8. This ensures that each time either the driver N15) all have a particular significance, bank or program counters are updated the it will be seen in ‘programming’ below that address counter is reset to zero; after all, it the length of a program (or sequence) can be is only reasonable that a new program a maximum of 128 or 64 steps. However, should begin at the beginning! For those this may be more than required and there- who are wondering what that strange little fore some means of programming the end of thing perched on the line to S4 is, it is a sequence and returning to the beginning simply a redundant gate. must be provided. This is where data line DO We now come to the memory itself, of of IC10 comes in. In the normal course of I which the full complement of four 2K- programming DO will be logic low until the CMOS-RAMS are shown in the circuit end of a sequence when a 1 will be entered diagram (IC10 . . . IC13). In normal oper- at this location (by S10). When the display is ation these are of course in the ‘READ’ up and running, a 1 appearing at DO will be mode and the contents of the address, synced with the address oscillator by FF2 determined by the address, program, and and used to reset the address counter to zero bank counters, are used to switch on (or via N8. The display sequence will then start off as appropriate) the output to the display from the beginning again. LED D8 serves to itself. Normally then, the R/W pins of each indicate this ‘reset’ pulse when it occurs, memory are held high by resistor R23. This The reset bit (DO of IC10) is not synced line must therefore be taken low whenever with the zero-crossing-point pulses. How- a program is to be entered or modified and ever, as the reset only occurs at the end of a this is carried out by switch S7 via switch program, this will cause negligible inter- 3-24 elektor india ference. The circuit diagram for the four LED displays, LD1 . . . LD4, is shown in figure 4. The printed-circuit board layout for this circuit is shown in figure 6. The address reference AO . . . A10 refer to those at the right of the main circuit diagram of figure 3. An appropriate link must be made to set the program step-length at the input to the decoder for LD3. If the program step-length is 128 steps, transistor T1 switches on the decimal point of LD1 for address counts above 63. Construction If the printed-circuit boards illustrated in figures 5 and 6 are used, construction of the electronics section of the disco display should prove no problem. However, before assembly can begin, the final design format must be decided upon. This refers in particu- lar to the LED read-out display board which, it will be noticed, can be divided into three separate sections. This has been done to allow the maximum flexibility of the design as it was considered that many readers may wish to build the display controller into an existing piece of equipment. The complete printed-circuit board as shown in figure 6 will match the suggested front panel design illustrated in figure 7. After assembly has been completed, not forgetting the two links (64 or 128 step program length), the two boards should be interconnected. This can be carried out by short lengths of wire or, if preferred, ribbon cable can be used. All the address lines as marked on the two boards, with the ex- ception of A6, must be connected. For a 64- step program this is taken to the point marked A6 on the board containing display LD3. If a 128-step program has been chosen, it must be connected to A6 on the board containing displays LD1 and LD2. There are three + terminals and three 1 ter- minals on the display board. These are separ- ately interconnected: one + and one i is connected to the + and 0 terminals respect- ively on the main board near Cl 2. If the display board is separate, each + and 1 should be connected with the + and 0 on the main board. The common point for the anodes of the indicator LEDs should be taken to the + terminal near Cl 2. The cathodes are connected to channel outputs 1 ... 30. Another set of + and 1 terminals will be found on the main board: these are for the switches. The switch connections should preferably be commoned after the switches have been mounted on the front panel as this requires only two wires to be returned to the main board. Normally, the channel indicator LEDs are connected in series with the LEDs in the opto-couplers on the triac board. To enable the circuit to be tested at this stage, some form of current limiting must therefore be included as a temporary measure. Two diodes type 1N4001 are therefore connected in series with the 5 V supply and the com- mon anode connection of the indicator LEDs. The LEDs should have a forward voltage of about 1.6 V. If the indicator LEDs are dispensed with and only the opto- coupler LED is used, the display pattern is, of course, shown by the display itself: re- sistors R58 . . . R87 should then be 330 S2. It will be remembered that the supply for the mains zero-crossing point detector is derived from the triac control board: this 3-25 A final construction point: IC19 must be fitted on a small heat sink. Programming When the circuit is first switched on (prior to any programming), the memory ICs will contain garbage, but you knew that, of Display sequences will obviously depend on the contents of the memory and the chosen program format (64 or 128 steps). Further- 3-27 programmable disco light display Resistors: R1 » 10 k/1/8 W R2 , . . R30- 330 Sll 1/8 W Capacitors: Cl “ 10(i/16 V Semiconductors: T1 - BC 547B IC1 - 74185 IC2 . . . IC5 = 74LS247 LD1 . . . LD4- 7750 Printed-circuit board 84007-2 more, a full memory will allow a total of 30 channels (lamps) and these can be ar- ranged in any number of pleasing designs including a dot-matrix for alpha-numerics (that means letters and numbers, sir!). Having decided upon the display format and the sort of programs that are to be used, the links at the A6 address line must be fitted as described under ‘construction’. Before starting the programming, it is advisable to commit the desired sequences or patterns to paper as even 64 steps can get decidedly confusing. To adjust preset P2, set switch S5 in position 'A (minute) and adjust preset P2 so that the program display jumps on one every 30 seconds. Off we go then. Set switch SI to position B (step mode), S4 to OFF, and S5 to off to prevent the program from jumping on during loading. Switch on the switch-key S8 and press S6 and S3 to get the right program and bank. The address display should read 00; if not, press S3 or S6 until the right program and bank are indicated on the display. The program data is set by switches S10 . . . S40 (or whatever number you have decided upon). Any one of these switches set to 5 V denotes a logic high and causes the appro- priate lamp to light. A switch set the other way gives a logic 0 and the corresponding lamp does not light. Are you still with us? Set the program data and press switch S7. The data lines are now inputs and the memory ICs will be fed with a write pulse and accept the data set with the switches. When S7 is released, the data lines revert to being outputs and the set pattern will be indicated by the channel LEDs. Now press S2 once (to increment the address by one), set the data switches, and again press S7. If an error was made during the entering of 3-28 programmable disco light display Figure 7. A suggested de- sign for a 19 in (483 mm) front panel for the pro- grammable disco lights. This panel matches the complete printed-circuit board of figure 6. the data, simply set the correct data and press S7 again. This works, however, only before S2 has been pressed. If S2 has been operated, press S6 until the same program is indicated on the display. Then go to the faulty address by means of S2, alter the data, press S7, and proceed to the next address by pressing S2. As mentioned previously, data line DO will remain logic low until the end of a pattern. On the address following the last line of the sequence, set S10 to 5 V (logic high); this can also be done on the last line of the sequence itself (together with the program data) if preferred. With some display pat- terns (especially running light patterns) it improves the display continuity, but it really is a matter of choice. Try some simple patterns to see the effect. And that’s all there is to it apart from a few pointers. At the end of the programming, do not forget to switch off the key-switch otherwise (in the case of disco’s) you might find your- self arriving at a booking with this terrific new display you have been raving about only to find that you have a completely garbled memory. Not good for the old ego, chaps! Don’t be too worried about making a false entry during programming as mistakes can be easily rectified. You do not have to reprogram the entire memory, just the line containing the error. Unless of course, you have a major disaster on your hands. In this case, turn the telly off and lock the door before starting! It is possible to include delays and acceler- ation in your program by repeating the same data in several addresses. This makes a very effective display when properly done but it does require careful planning with due regard to program length (64 or 128 steps). Remember not to be caught out by the address counter read-out. This just indicates from 0 to 63. If a program length of 128 steps is opted for, the decimal point of LD1 signifies the ’upper’ 64 step range. Set SI to position A when the program should run; adjust the run time with PI. It may happen that when the run time is increased (that is, smaller resistance of PI), the pattern on the display does not run smoothly (stutters) or even stops altogether. This is caused by the frequency of N4 being too high in relation to that of the zero- crossing pulses. Because the trigger levels of different makes of 4093 show wide variations, this erratic running may or may not occur. The adjustment range of PI should be set by means of R8 and/or C3 so that stuttering or stopping of the pattern just does not occur. A point worth noting! The disco display is completed, programmed, and ready to go to work . . . However, when it is switched on, nothing happens: no lights, no LEDs, just panic! Fear not, gentle DJ, all will be as it should be if you just press the program step switch, S6, and a program will start from the beginning. To end, we are sure you don’t need re- minding that the mains supply is a little ’conspicuous’ in this circuit. Please take care, as we have no desire to reduce our circu- lation by stopping yours! M 3-29 The video combiner is a circuit which 'welds' the various components of a video signal, such as the synchronizing pulses, blanking signal, colour information, and so on, into a composite video signal. Although this is a fairly complex matter, a recently introduced integrated circuit, the TEA 1002, makes it possible to keep the combiner reasonably simple. video combiner together. In that case, neither the chrominance subcarrier oscillator (pins 13, 14) nor a signal at the CBF (colour burst flag) input (pin 15) is required. The circuit diagram The circuit may be divided into three parts (see figure 2): the PAL switch (FF1), the combiner proper (IC2), and a buffer stage (Tl). The PAL switch, flip-flop FF1, is controlled The TEA 1002 is a PAL colour encoder by ,h e line synchronizing pulses at its with video combining stages. It converts a clock input (pin 3). (See also 'video sync number of appropriate input signals into a box’ elsewhere in this issue.) complete video signal, that is, one contain- The TEA 1002 (IC2) contains a chromi- ing synchronizing pulses for the line and nance (chroma) and a luminance encoder, field scans, luminance and chrominance The luminance is dependent on the signals, blanking pulses, and a colour voltage level at pin 9, which is preset by burst signal. PI. If this voltage is greater than 4 volts, a The required input signals are derived 75 pe r cent colour signal (as defined by from the ‘video sync box’ described the EBU — European Broadcasting Union) elsewhere in this issue. The printed-circuit j s generated. When the voltage falls boards for that box and the present circuit below 3 volts, the brightness is increased are of the same dimensions so that they to 95 per cent, which, no doubt, will nor- can be built conveniently into one unit. mally be preferred as it gives a clearer picture. It should be noted that the voltage The TEA 1002 at pin 9 should not rise above 5 volts to The ’innards' of the TEA 1002 are shown in prevent saturation of the buffer, Tl. schematic form in figure 1. The logic The TEA 1002 also contains a divider decoder generates colours according to which produces a 3.54 MHz clock signal the logic levels at pins 1 . . . 4 (see table 1). from the 8.86 MHz subcarrier oscillator. If only black and white signals are re- The clock (pin 17) may be used to syn- quired, pins 2. . .4 are simply strapped chronize other circuits. The oscillator can 3-30 be pulled to its correct frequency by ad- justing capacitor Cl for minimum interference (least ragged image fringe). Setting Cl to its mid-position will in prac- tice normally be sufficient. The output level of the buffer stage, emit- ter follower Tl, is preset by means of P2. With values shown, the output impedance is around 75 ohms. The output level is nor- mally adjusted to give 1 V pp across 75 Q, that is, 2 V pp emf. Construction and application The printed-circuit board, shown in figure 3, has the same dimensions as that for the video sync box, so that the two boards can conveniently be built into one compact unit. The various terminals on the boards are located such that the length of the interconnecting wires is kept to a minimum. The circuits should, of course, be preset before making the interconnec- tions. The board has provision for an optional wire bridge. If this is used, the logic levels at pins 2 ... 4 ('O') produce standard colours and the chrominance signal is at normal level. If the wire bridge is omitted, the colours are inverted (see table 1) and the chrominance signal is reduced by 6dR Power supply requirements are 12 V at 100 mA maximum. The combiner lends itself to a variety of applications. For example, when used with a personal computer with video interface which has colour information available (in the form of Red, Green, and Blue signals) it makes possible the production of a composite video signal. In combination with the video sync box, the combiner can produce a colour bar which is suitable for use as a test signal, as a space marker for video recorders, or with local cable systems. For these uses, the R, Q and B pins on one printed-circuit board must be linked to the correspond- ing ones on the other board. Finally, the combination may in appropriate cases form the link between electronic equipment and a colour TV. H 3-31 digitester with a difference The testing of digital circuits can be quite a headache. Our old faithful, the multimeter, is quite useless because of the operating frequencies: the logic levels change so rapidly _ thousands or millions of times per second that even a digital multimeter is unable to cope. This problem can be solved in two ways: buy a higher quality test instrument or lower the operating frequency of the circuit under test. If you opt for the last, you will find our digitester just the thing! digitester with a difference universal test aid for digital circuits Digital circuits normally operate at speeds which make it impossible for normal test instruments to be used for checking or fault finding. For instance, in an analogue multimeter, the inertia of the pointer prevents the reading of the level of a pulse train. The normal digitester does not help here either; it may well give an op- tical indication of the logic level at the pin of an IC, but only as regards a static or slowly changing situation. When rapid changes occur, the digitester is also quite useless. This is, however, not because of inadequacies in the instrument, but rather because of the slowness of our eyes. When an LED blinks at only 20 Hz, few of us see this as a series of light pulses: most will just see a continuous light. It is by now evident that to test digital cir- cuits or to be able to experiment with them it is necessary to slow down the speed of operation. The easiest way to do this is to disable the internal clock o' the circuit and replace it by an external one operating at a much lower frequency. In some cases, it is even better to work with just one pulse at a time, instead of with a pulse train. bounce switches SI and S2 which ensures that only one pulse is present at their out- puts. This pulse can be used in the circuit under test as clock, counter, reset, and so Apart from single pulses, there is, of course, a need for low-frequency pulse trains. The generator required for this is formed by NAND Schmitt trigger N9, resistor R13, and capacitor Cl. With values shown, the frequency is around 50 Hz. A second pulse-train oscillator, formed by N10. R14, and C2, operates at the much lower frequency of 2 Hz, which is optically indicated by LED D5. So much for the description of the re- quired generators. But what if you want to apply a single pulse followed by a pulse train to a circuit? It would not do to have to change from one output to another. No, for this purpose we have added an elec- tronic switching circuit consisting of S3, NOR gates N5 . . . N7, NAND Schmitt Trig- gers N8, Nil, and NAND gates N12/N13 and N16/N17. The output, pin 8 of N17, de- pending on the setting of S3, is either the 02, the 2 Hz, or the 50 Hz signal. The logic output level is optically indicated by LED D6: The circuit Fewer ICs than shown in figure 1 are re- quired to generate a single pulse or pulse train of low frequency. None the less, the additional ICs used here make for a more ’comfortable' circuit. So, let's see. . The generation of a single pulse is ef- fected primarily by NAND gates N1 . . . N4 Gates Nl, N2 form a flip-flop of which the logic level at outputs Q1 and Q1 is depen- dent on the position of switch SI. As drawn, Q1 is logic low (0) while Oi is logic high (1). The high signal at 01 is indicated optically by LED D1 via gate N18. A se- cond single pulse is generated by gates N3. N4: their logic levels at outputs Q2 and 02 are optically indicated by LEDs D3 and D4. The two bistables, N1/N2 and N3/N4, de- ■ if D6 lights continuously and indepen- dent of the setting of S2, the output is a 50 Hz pulse train; ■ if it blinks rhythmically, the output is a 2 Hz pulse train; ■ if it lights depending upon the setting of S2, the output is the logic level of 02. All outputs are buffered, which enables up to 30 TTL circuits to be connected to Finally, the function of switch S4. When this switch is open, the output of N17 is open, that is, it contains the signal selected by S3. If S4 is closed, however, the output of N17 is logic low and the signal selected by S3 is therefore not available at pin 8. 3-32 elektor in Construction As you have seen from the circuit dia- gram, the digitester needs a supply of 5 V. This is best obtained by means of a 5 V voltage regulator: a 500 mA type will do nicely. The supply voltage to each of the ICs should be individually decoupled by a 100 nF capacitor. The power supply is best built into a case together with the digitester so that you have an independent, self-contained test instrument for digital circuits. Figui If you want to use the digitester with circi CMOS circuits, it will be necessary to gene adapt the (TTL) outputs to the CMOS logic ,rain under test. This is relatively simple and is elecl described in some detail in the article 'mating logic families' elsewhere in this Almost every new car manufactured in the world today is fitted with reversing lights. Great idea! Not only do they help you see where you are reversing in the dark, but they also make your intentions clear to anybody behind the car. In some Asian countries it is even a legal obligation for every car to have an externally audible reversing indicator. The one problem with these ideas is that the car driver does not directly benefit from them. reversing buzzer 'clunk, click, buzz . . .' It is an undeniable fact of human nature that we often forget or neglect the care and caution instilled into us while learning a new skill. Nowhere is this more obvious than in driving a car. We frequently tend to do what is convenient rather than what is correct. Just one small, but common, ‘fault’ is starting the car in gear with the clutch depressed. Then you only have to release the clutch ar.d away you go . . . But in which direction? It can prove very ‘surprising’, to say the least, when you expect to move smoothly forwards but instead find the driver of the ‘slightly shortened’ car behind you tapping on your window to express his opinion of your character in a somewhat heated manner. The circuit The circuit here also gets excited when you start the car in reverse gear but all it does is buzz at you in displeasure. When the ignition is switched on the car battery voltage is applied to the circuit and the oscillator around N2 starts. This provides one of the inputs to N3. If the car is in reverse gear, the second input of N3 is taken high via R7, and this causes the buzzer to Figure 1. The circuit, es shown here, uses normal components thet most people will probebly hove lying around. It can easily be built on a small piece of veroboard and needs only three external connections. +12 V, ground and a friendly neighbourhood mechanic may help to find a suitable place to tap this 2 Figure 2. This 'circuit' can be used in place of the buzzer, and so get rid of ponent in the whole Simultaneously, pin 12 of the CD 4060 is taken high and this chip is reset. This IC is a 14-stage binary counter and oscillator, the frequency of which is set by external components (C2, R3 and R4). After a certain time (about six seconds), the Q13 output (pin 3) of IC1 goes high and stops oscillator N2 by taking its input (pin 5) low via N1 . This, of course, stops the buzzer and ensures that it does not sound every time the car is put into reverse gear, which would be very annoying. An alternative to using the buzzer is the small circuit shown in figure 2, consisting of a loudspeaker driven by a darlington pair. Transistors T1 and T2 may also be replaced by a single-package darlington such as a BC516. M 3-34 The problems of address decoding in a microprocessor system are generally memory timing summarised in the question 'where, when and how is the memory accessed?'. Our first article on address decoding did not answer the 'when' part of this question. For this reason we decided a second article was needed to deal with the timing of operations and signals. We also decided to have a look at an example of modifying an existing decoding system. memory timing A logic combination of most significant address lines can be used to provide an enable signal that is only active for certain configurations of the lines used. As we have already seen in our first article, this signal is applied to one or a number of memory ICs accessed by the least significant address lines, which, in fact, control the chip’s internal address decoder. Data is transferred via the data bus. No matter how high the clock frequency of the processor, the address and data signals do not appear either instantaneously or simultaneously. On the one hand there is always what is known as a set-up time for the signals and on the other hand even the clock signal itself takes a finite time to appear. This is an added complication, but fortunately the difficulties are somewhat reduced by the presence of control signals supplied by the processor. These signals are used to synchronize the address decoding and read or write operations. Timing of Z80 and 6502 signals As the timing di agram of figure 1 shows, the MREQ, RD and WR signals of the Z80 do not appear at the begin ning of read or write operations. When the MREQ and RD signals are not Tow’ during a read operation (left half of the timing diagram), the address signals A0 . . . A15 cannot be considered stable. The same applies to a write cycle when MREQ is not act ive. The upshot of this is that the MREQ signal and address decoding signal must always be combined before being applied to a memory IC. As the right half of figure 1 shows, the WR sig nal is fo llowed after a significant delay by MREQ and the beginning of the phase to establish the data signals. These latter can only be considered stable after WR appears. It should be noted that the WR line becomes inactive again a half clock cycle before the address and data words change (T3 of the write cycle). The WR signal could also be used as is to change the memory from read to write mode and vice versa (R/W). The timing diagram for Z80 signals corre- sponding to an input/output instruction is shown in figure 2. Note in passing the presence of a spontaneous wait cycle gene- rated by the processor itself to allow the - generally s low - input/output circuits to produce a WAIT signal if necessary. Here again, the address and data signals can only be considered stable after the appearance of control signals. In the 6502 timing diagram shown in figure 3 the essential enable signal is 2. As soon as this signal is 'high’ the address signals, and, immediately after them, the data signals, can be considered stable. The same is true of the signal to change between read and write modes (R/W ). As this pro- cessor has no specific I/O instructions, it also lacks any particular control signals for this type ofcircuit. A RAM-R/W signal is often found on 6502 systems, and this is obtained by combining the 4>2 and R/W signals. This can then be applied as and when desired to memory ICs to change between reading and writing. For EPROM chips the 2 is combined with the address decoding signal (this is shown as gates N41 and N44 on the interface card of the Junior Computer). For inputs/outputs, various combinations of 4>2, R/W and the address decoding signal are possible. The R/W signal (and possibly 2) could also be used for switching the bi- direction al data buffers (the READ and WRITE mode- switching signals on the Junior Computer interface card_are obtained from, among others, the R/W signal). We must stress the importance here, for the designer, of carefully noting the timing of the control signals which must be catered for by the logic for decoding and enabling memory ICs. Modifying an existing decoding system After so much theory we will now deal with a practical example of using an existing system: the interface card of the Junior Computer, in fact. The aim of the modifica- tion is to reduce the importance of the doubly addressed zone between F800 and F9FF (or 1800 . . . 19FF in the DOS ver- microprocessor control signals and their sequence Figure 1. The timing diagram of the Z80 shows that the address ^nd data information is only usable Note that the timing diagrams are taken fror the Synertec data book and may be different 3-35 2 sion), and to be able to address a new input/output circuit there. The 6522 VIA (IC1 on the interface card) occupies addresses F800 . . . F9FF (1800 . . - 19FF), but this is a bit waste- ful as 16 addresses are sufficient to ad- dress all the registers of this IC. The K6 signal is active between F800 and FBFF (1800 . . . 1BFF). Address line A9 permits the zone of F800 . . . F9FF occupied by the VIA to be distinguished from the FA00 . . . FBFF (1A00 . . . 1BFF) area oc- cupied by the 6532 on the main board. It would be nice to ‘regain’ the unused ad- dresses for a new input/output circuit, as long as there are not too many modifications required. Looking at the ‘circuit’ of figure 4, we recognize it as a section of the circuit for the interface card containing the VIA, gate N35 _and PROM IC17. Signal K6 applied to the CS2 inppt is active between F800 and FBFF (1800 . . . 1BFF), while the CS1 input receives a signal called VIA (active at logic ‘high’) obtained from K6 and address line AB9, between F800 and F9FF ; this same signal is applied to the PROM and thus enables the buffers in read or write mode while the address signals are present on the bus. Every little detail must be taken into account! The same components are seen in figure 5, along with a new 6520 P1A and a slight modification to the address decoding. The VIA signal is unchanged; it is still applied to the CS1 input of the 6522 and to the PROM (if this signal were modified then so also would the enable signal for the bf_ directional buffers be changed). The CS2 signal for the 6522 is now supplied by AB8, and this means that the VIA no longer occupies addresses F800 . . . F8FF (1800 . . . 18FF). Line AB8 is also connec- ted to the CS0 input of the 6520 PIA for which our VIA signal (still obtained from K6 and AB9) provides the CS1 signal (active logic high, just like CS0). The third enable input of the 6520, CS2, is activated by the AB9 signal, so that the PIA is addressed between F900 and F9FF. This IC can be put anywhere as long as it is after the bi- directional data buffers (IC11 and IC12 of the interface board). Table 1 shows a sum- mary of the operation of the new configura- tion in the form of a truth table. Instead of mounting the new PIA on the bus, it could also be soldered directly on top of the 6522 on the interface board. This operation, relatively perilous in itself, has the advantage that is makes things much simpler. The lines common to b oth IC s are DB0 . . . DB7 (pins 33 . . . 26), RES (pin 34), 2 (ENABLE; pin 25), +5 V (pin 20), earth (pin 1), R/W (pin 22 of the 6522 - pin 21 of the 6520), RS0 (A0; pin 38 of the 6522 - pin 36 of the 6520), RSI (Al; pin 37 of the 6522 - pin 35 of the 6520) and (pin 21 of the 6522 - pins 37 and 38 of the 6520). The connection between K6 and pin 23 of the 6522 (CS2) must be broken; this pin is then connected to AB8. Pin 23 of the 6520 (CS2) must be connected to AB9, pin 24 (CS1) to the VIA line (pin 24 of the 6522), and pin 22 (C30) to AB8 (pin 23 of the 6522). 3-36 elektor in Finally, we must give some indication about how to access the registers of the 6520 PIA. The addressing is as follows: $F900 : PAD or PADD (data or direction register A) $F901 : CRA (command register port A) $F902 : PBD or PBDD (data or direction register B) $F903- : CRB (command register port B) When the CRA bit is high the register addressed at $F900 is PAD, the date register. If this bit is low the register addressed is PADD, the date direction register. The same applies for CRB, with PBD and PBDD. In spite of this somewhat peculiar addressing method, the operation of the 6520 ports K6 A9 A8 || VIA | zone addressed XXXX SF A00 . . . SFBFF (S1A00 . . . $1BFF) SF80C . . . $F8FF (S1800 . . S18FFI SF900 . . . SF9FF ($1900. . . S19FF) is the same as those of the 6522 except for some details (essential for some applications) which we will not go into here. Complete information on the 6522 is, in any case, contained in the Elektor book about this IC - Ml Figure 5. By applying the AB8 signal (instead of K6) to the CS2 input of the 6522 VIA the doubly addressed zone ond half can then be d to address a new I circuit. A 6520 PIA is divided between 1984 3-37 mating logic families Connecting digital ICs within the same logic family will rarely cause any problems, as long as such things as fan-out and parasitic line and input capacitance are taken into account. It is quite a different matter, however, to try to use the different logic families, TTL (standard, LS and ALS) and CMOS, together. There is a great temptation to do just this as the possibilities for combinations becomes even greater with logic families continually being expanded. The new family of high speed CMOS (HCMOS) recently released onto the market prompts even more questions about the compatibility of its two variations with existing logic circuits. Maybe a lot of problems will be solved if we simply answer the question 'How digital is digital?'. mating logic families it was all very easy when there was only TTL The popularity of digital electronics is very easy to understand. What could be simpler than a system in which there are only two values, T or ‘O’? Certainly this makes design and fault finding much simpler, but there are also some other considerations. As long as the elements of a design are kept ‘in the family’, with only TTL or only CMOS, for example, the manufacturers have already sorted out the problem of matching different gates. The logic levels are well defined and the input and output currents are virtually the same. Combining different logic families, however, is a completely different kettle of fish. Then our old friend Murphy appears with a vengeance and seems to have taken a personal dislike to your design, whatever it is. With a bit of determination, however, even Murphy can be defeated (temporarily at least). How the families compare There are. of course, some advantages to having different logic families. It becomes easier, for instance, to combine speed with economy. The appearance of dif- ferent logic families is largely based on the attempts to achieve ever shorter switching times and smaller power con- sumption. Within the scope of a short article we cannot deal with all logic families, and this would, in any case be overdoing it a bit as we are only in- terested in logic elements that are already, or will soon be, readily available. A sum- mary of the families to be dealt with and their most notable characteristics is shown in table 1. These data should only be con- sidered as an indication and not as exact values. The only function of this table is to aid the general comparison so the values can vary depending on the circuit and even the manufacturer. The values of logic T and ’O’ have to be specified as a certain voltage and the symbols we will use to make this defi- nition are shown in figure 1. In order for the circuit to operate under the least favourable conditions (the worst case) UOH must always be larger than UIH and UOL must be smaller than UlL- A summary of the voltages needed for the various logic families is given in table 2. Table 1 Figure 1. Here we define voltages. Between 'high' The CMOS levels are only given at 5 V because we want to ensure compatibility with TTL. Possibilities for combination First we will see which families are matched purely on the basis of input and output levels. Most notable is the fact that interconnecting elements within the TTL group causes no problems. In one case the noise margin is even improved; this happens if LS or ALS is used in place of standard TTL. Connecting TTL to HCTMOS is no prob- lem either as this version of high speed CMOS is TTL compatible. Furthermore the user does not even need to know he is working with CMOS because its gates appear to be extra-efficient LSTTL devices. The supply voltage tolerance with HCTMOS is larger than with TTL (10% in- stead of 5%), which simply means that the 3-38 I TTL supply can be used for HCTMOS but the reverse is not necessarily true. It is not quite so easy to connect TTL to CMOS. The UOH in TTL is lower than the UlH in CMOS with a supply of 5 V. This means that a logic 1 at a TTL output will not be interpreted as 'high' by a CMOS The same applies if we want to use HCMOS and TTL with a supply of S V. In this case Uih (for HCMOS) is a minimum of 3.15 V, which is much too high for TTL. All is not lost, however, as the supply voltage for HCMOS can be anywhere be- tween 2 and 6 V. If the HCMOS section of a circuit is operating on a supply of 3 V then UIH is 2.1 V (70% of 3 V). Now TTL can provide a logic 1, with a margin of 0.3 V. One situation that can arise here, however, is that the TTL output level can be higher than the HCMOS supply voltage. In this case the current flowing through the 150 S input resistor and the in- put protection diode is limited by the resistor and by the collector resistance in the output circuit of the TTL gate. As long as the input current does not exceed 20 mA nothing untoward should happen. What logic 0 means in this case is three things: UIL is a maximum of 0.6 V (20% of 3 V), while the UOL for TTL is 0.5 V which leaves a margin of 0.1 V. Fortunately it is not such a problem to drive TTL from CMOS or HCMOS, as long as the supply is 5 V. The input levels for TTL do not have to be very precise; UlL is relatively high and UlH fairly low. The CMOS output voltages are therefore quite suitable for the TTL inputs, but care must be taken to ensure that the CMOS can handle the relatively high TTL input current. This ap- plies particularly to driving standard TTL from ordinary CMOS, we will deal with this further under 'fan-out'. If the CMOS is operating at a higher supply level then obviously a level adapter circuit is needed at the connection to TTL or HCTMOS Fan out As regards input current a distinction must be made between TTL and CMOS. The in- put of a TTL gate consists of a (multi- emitter) transistor whose base is con- nected to VCC via a resistor, as shown in figure 3. Consequently a floating input is always seen as a logic 1. The output is logic 0 if the input is earthed, then a cur- rent, the sink current, flows from the input. The sink current is 1.6 mA in standard TTL, 0.4 mA for LSTTL and 0.2 mA with ALSTTL. These values are also stated in table 2. The output of the driving gate must be able to handle this current. This, of course, presents no problems for TTL as the outputs are designed with this in mind, but CMOS is a different matter. Within the CMOS family the outputs are not expected to deliver large currents. The only current that will flow is the charging current for the input capacitance (and otherwise only the input leakage cur- rent) which has a value of a few nA. As a general rule the fan out, even bet- ween different families, can be calculated by dividing the maximum output current by the required input current. These cur- rents are defined for both logic levels (see table 2 again). Because of the set-up of the Figure i.. Input protectior circuits tor HCMOS and CMOS circuits. 3-39 TTL input circuit IlL is considerably greater than IIH- As a consequence of this lack of symmetry the fan out at both logic levels must be calculated and the smaller value taken as the limit. Using the data irt table 2, the fan out for various different combinations is easily found. A separate table (table 3) has been drawn up to show a summary of the results. This fan out is only shown for the combinations whose logic levels adapt directly to each other, as indicated by table 4. Because of its large sink current, TTL can be heavily driven. The fan out, for driving normal TTL, is, as a rule, low. Not even CMOS can handle the sink current of 1.6 mA. As a result of this it is not possible to connect CMOS directly to TTL, even though their voltage levels are similar. There are, however some CMOS ICs available with buffered outputs that can sink a current of 1.6 mA. Otherwise the CMOS outputs could also be connected in parallel until the IlL required is reached. The data sheets from the relevant manufacturer should be studied for more details about this. There are less problems with driving LSTTL and ALSTTL because of their smaller sink current. CMOS can also drive LSTTL and ALSTTL directly. The input requirements of all the MOS families are so modest that the fan out is theoretically very high (several thousand). In practice this is limited by the input capacitance and the lead capacitance. If the maximum frequency stated by the manufacturer is to be attained (generally given by CL = 10, IS, 50, or 100 pF), then the fan out is defined by dividing CL by the input capacitance. As a general rule 10 pF per input can be taken as the norm. Remember, of course, that the input capacitance is very dependent on the technology used; CMOS ICs manufactured by metal gate techniques have a larger in- put capacitance than those made using silicon gate technology. Also, a length of ribbon cable or a track on a printed cir- 3-40 elektor i cuit board can form a considerable capacitance. In all these cases it is up to the user to set an acceptable delay time and therefore the fan out. Adapter circuits In order to connect TTL (standard, LS and ALS) to 5 V supplied CMOS and HCMOS, the TTL must be able to provide the logic 1 level needed by the CMOS (at least 3.5 V). This is simply done using a pull up resistor, as figure 4 shows. A small value results in a high speed as parasitic capacitances are then charged more quickly. The minimum resistance possible is decided by the maximum load for the output. In theory the number of inputs driven by this output must also be taken into account, but if they are MOS inputs, which have negligible input current, that can simply be ignored. The minimum value of the pull up resistor is defined by: R(min.) = [VCC(min.) - UOH]/[lOL - ZIIL] The second term in the denominator, the sum of the input currents, can be neglected if we are talking about MOS inputs. There is also a maximum permissable value for the pull up resistor. Because of leakage currents at the output (if, for ex- ample, several open collector outputs are connected together) and the input, there is a voltage drop across the pull up resistor at logic 1. As the output voltage may never be less than UlH. the maximum value of the pull up resistor is defined by: R(max.) = [VCC(min,)-UOH]/ [ZIOH + ZIIH] Here again the second term of the denominator can be ignored for MOS inputs. What all this means, in effect, is that the pull up resistor must have a value of 1 ... 10 kQ. Generally these formulae can be applied to the pull up resistors of open collector outputs whether they drive CMOS or HCMOS or not. The situation is quite different if one logic device operates at a different supply level, which also means different logic levels. A single 4009, 4010, 4049 or 4050 buffer can be used as a high-low adapter, for exam- ple from 15 V CMOS to 5 V TTL. Each package contains six buffers, and in the case of the 4009 and 4049 they are also in- verters. These buffers, which can drive up to 2 TTL inputs or 9 LSTTL inputs, can also be used, for example, to drive standard TTL from CMOS. And so to the last combination possibility: from 5 V TTL to CMOS working at a higher level, or HCMOS at 6 V. This is also fairly straightforward if we are working with open collector outputs. In some cases the output transistor's UCE is higher than VCC- Examples of this are the 7406 and 7407 with 30 V open collector outputs and the 7416 and 7417 with 15 V open col- lector outputs. The value of the pull up resistor must be carefully chosen so that the sink current does not become too The fan out of the 74XX buffers listed above is three times the standard fan out, so the pull up resistor is unlikely to be too small. The disadvantage of this is that un- necessarily small pull up resistors can draw a much larger current for a negligi- ble increase in speed. A discrete buffer stage could, of course, be built using a transistor and two resistors to drive CMOS from TTL. This ef- fectively creates an open collector output. Two possibilities for this are shown in figure 4d, the first having the advantage of a faster switching time. Finally It is not a good idea to leave unused TTL inputs floating, even though they normally act as if they were logic 1. If, for example, a certain LSTTL IC with a floating input is changed for its HCTMOS equivalent this causes problems. The very high input im- pedance means that the logic level will not be defined and the circuit will not work (properly). The moral of this is that unused inputs should always be tied to one logic level: for TTL use a pull up resistor (1 ... 10 kQ) to Vcc> connect directly to earth or to an input that is used (LS inputs can be connected directly to + 5 V). With all MOS veisions connect unused inputs to VCC. earth or an input that is used. It is virtually impossible to list how to in- terconnect all the various logic families from different manufacturers who specify different testing conditions for their gates. However, that was never our intention, and we think that this summary of the present possibilities should at least enlighten the hobbyist as to what can be done with what is available today. M ECL Emitter Coupled Logic, fast ' unsaturated logic. TTL (7400 series) Transistor Transistor Logic, saturation. SloweVt'tan ECL. HTTL (74H seriesl High speed TTL. LTTL I74L series) Low power TTL. STTL I74S series) Schottky TTL. The use of Schottky diodes prevents saturation. This increases the switching speed. LSTTL (74LS seriesl Low power Schottky TTL. A LSTTL I74ALS series) Advanced Low power Schottky TTL.These are the fastest and most economical TTL devices. HCMOS (74HC series) High speed CMOS, CMOS with LSTTL switching times. HCTMOS (74HCT seriesl High speed TTL compatible CMOS, low power 1984 3-41 The measurement of capacitance In the early days of electronics, the values of capacitors and inductors were determined by impedance measurement in bridge circuits. Such measuring bridges contained, apart from an oscillator, power supply, and sensitive meter amplifier, also very precise and therefore very expensive reference capacitors or inductors. Furthermore, oper- ating these bridges correctly was not a simple matter. None the less, there can be no doubt about the superiority of them. For instance, they make possible the quick determination of factors other than the value, such as the Q factor and inherent losses, which are equally important for the calculation of the impedance of a circuit. However, these factors are not normally of great importance to us. Capacitors are used mainly as blocking, smoothing, or decoupling elements, and also as frequency determining com- ponents in HF and AF en- gineering. If capacitors are to be used in filters, they should be as close as possible to the cal- culated value. That nor- mally means the use of high stability ca- pacitors, but the pre- cise value can of course, be determined by some sort of measuring in- strument, and this is where our capacitance meter comes in! It will enable you to determine the exact value of the capacitors, easily and conveniently. The capacitance meter will, of course, also be to tell you whether a suspect capacitor needs replacing or not. The meter is a precision instrument with a 3'/^ digit liquid crystal display which enables the measurement of capacitances from 0. 1 pf to 20 mF in six ranges. capacitance meter... Simple and easy-to-operate capacitance meters usually require the unknown capaci- tor, C x , to be inserted into an oscillator circuit. The frequency of the resulting signal is measured with a frequency counter or a voltmeter (after conversion to a propor- tional voltage). An appropriately calibrated scale on these instruments makes it possible to read off the value of the capacitor directly (see, for instance, Elektor, December 1981: ‘capacitance meter module’ page 12-18). A different method of measurement is illustrated in figure 1. The point of this method is that the unknown capacitance, Cx, after differentiation of the input signal (by Cx/Rs). is determined by a voltage measurement. By making the value of R s much smaller than the impedance X c , ... to find those elusive farads! Measuring ranges: 200 pF; 20 nF; 2 u F; 200 pF; 2000 pF; 20 mF (all f.s.d). Accuracy: 1 per cent (if calibrated with a 1% reference capacitor - otherwise larger); 10 ... 15 per cent in the 20 mF range. Read-out on 354-digit liquid crystal display (LCD). Capacitor leakage current does not affect 3-42 elek.br in the value of C ? 1 x can be calculated from C x = Ui/27rf 0 R s U in which n, f Q , R s , and U are known con- stants, so that only the value of the measured voltage, Ui, needs to be inserted. You don't, of course, want to be bothered with pen, paper, and pocket calculator every time you measure a capacitor, but want to read off its value directly. The diagram of figure 1 is therefore extended into that of figure 2. The triangular output of the generator is passed to C x which has been connected in a differentiating circuit. The output of this circuit is a square wave of which the ampli- tude is proportional to the value of C x (like Ui in figure 1). The square wave is rectified in a phase-selective synchronous rectifier: the level of the resulting voltage is measured by a digital voltmeter. The phase-selective rectifier operates as follows. The square-wave output of the differentiator is applied to electronic switch ES5 in phase with the rectangular output of the generator, and to electronic switch ES6 in antiphase with the rectangular output of the generator. The switches are synchronized with the triangular waveform and only pass the positive portions of the square waves. The two resulting square waves are added to provide a d.c. voltage. The relationship between the waveforms is illustrated in figure 3. The “rooftops' on the rectangular waveshapes are caused by the leakage current through C x . This current, which is caused by the triangular output of the generator, does not affect the measure- ment. Firstly, it largely disappears in the build-up of the average level (figure 3B), and, secondly, it is not accepted by the o phase-selective rectifier because it is 90° out of phase with the triangular current. In an ideal circuit, the triangular signal superimposed onto the d.c. voltage (figure 3C) does not rise at all. The circuit diagram The waveform generator is built up from two opamps: a Schmitt trigger (IC1) and an integrator (IC2). When the output of the integrator reaches the upper trigger level of the Schmitt trigger, the input to the inte- grator is inverted. The output level of IC2 then decays until the lower trigger level of the Schmitt trigger is reached. In this way, IC1 produces a rectangular signal and IC2 a triangular one. The output voltage of IC2 is the test signal for C x and is connected to the inverting input of the differentiator IC3. The output of the differentiator istherefore a rectangular voltage, the level of which is proportional to the value of C x . The phase-selective rectifier consists of electronic switches ES5 and ES6 which obtain their signals direct from IC3 and inverted from IC4. The control signal for the switches is taken from IC1 and fed direct to ES5 and inverted (by ES4) to ES6. The output signals of ES5 and ES6 are added and taken to the digital voltmeter via R20 (see figure 5). 2 The low-pass filter formed by PI, R6, and C2 derives a small triangular signal from the square-wave output of IC1, which is applied to the input of IC3 via C3. As the test signal is in antiphase with this voltage, the unavoidable parasitic capacitance at the test terminals is simply ‘spirited away'. In practice this means: adjust PI with open test terminals so that the DVM reads ‘O’. If the wrong measuring range has been selected, IC5 switches on electronic switch ES7 at a certain input level. When that happens, a large d.c. voltage is applied to 4 R,R 7 (C4+C5+C6) the DVM via R21 and the meter shows an overflow. When the value of C x is too high for the selected measuring range, IC3 no longer functions as differentiator but rather as comparator for the triangular signal at its input. The result is that a rectangular signal appears at the output of IC3 which is 90° out of phase with respect to the signal which would have appeared under correct conditions. The rectifier will then not have an output, and the DVM reads ‘O'. Some more points about the measuring ranges and the test signals. Switch SI is the range selector. For capacitors between 0 and 2 pF, the amplitude of the triangular signal is about 1.8 V pp at a frequency of around 1000 Hz. Switches ESI and ES2 are then closed. This enables the measure- ment of all non-electrolytic capacitors in three ranges: the test conditions conform to the manufacturers' specifications. Three ranges are also available for the measurement of electrolytic capacitors. These measure- ments are carried out at lower frequency and voltage (f = 100 Hz, and Uj = 18 mV pp ) and are also in accordance with manufactu- rers’ conditions of test. In the T range, the frequency is reduced to 10 Hz (ES3 closed), because the current at 100 Hz would be about 72 mA which is too much for the opamp. The consequence of this is that the accuracy in this range is only 10 ... 1 5 per cent. Fortunately, this is not so bad, because the exact value of electrolytics in this range is normally not very important. If it is required to measure an electrolytic capacitor in range 'c', switch S2 raises the test signal by about 1.5 V to ensure that the test voltage in this range is always positive. In the other ranges, the very small negative voltage of about 9 mV pp causes no harm. The circuit of figure 5 is basically that of the ‘LCD panel meter’ featured in the October 1981 issue of Elektor. However, in the present circuit the decimal point is switched by Sib and associated diode matrix. Moreover, the selected range is indicated by LEDs D3 . . . D7. Construction First of all, mount (but do not solder) all resistors up to and including Rll and all capacitors up to and including C9 onto the metering board shown in figure 6. It’s best to use soldering pins for this to simplify the soldering after calibration. Next, fit all components (except R1 and R7) to the display board shown in figure 7. The display and LEDs must be located on the track side: solder the LEDs so that they are well separated from the display. For 3-44 eleklor-india the time being, substitute wire bridges for R1 and R7. Do not yet solder diodes D1 and D3. Lastly, fit wire bridge B. The mechanical construction is best carried out with an eye on the sketch in figure 8. We have used a Vero case into which the aluminium mount' .g tray can simply be inserted after the calibration. Both printed- circuit boards are mounted onto this tray: the display board at the front and the meter board at the rear. This method also ensures the screening of these circuits from one another. Terminals with identical markings on the two boards should be interconnected with short lengths of wire, but keep terminals T, ‘CDp’, and ‘Z’ on the display board free. The terminals for connecting C x should be connected to the meter board by twin screened cable. The screen should be soldered ONLY to the common earth terminal (1) near the C x pins. Finally, the time has come to connect S2 to the meter board and the earth terminals on the front panel and mounting tray to earth. Then mount the mains transformer, mains on/off switch, and the fuse carrier and fuse in the case. Keep the transformer as far away as possible from the meter board. After sticking the transfer onto the front panel, this and the mounting tray may be inserted into the grooves provided on the case. Calibration First, set the range selector, SI, to position *f’ and adjust preset P3 for zero reading of the display. Next, set SI to ‘a’ and adjust preset PI on the meter board for zero reading of the display. Switch off the mains supply to the capaci- tance meter and solder a high stability (1%) resistor of 330 kfl in the R12 position and a capacitor of 150 pF in the CIO position (both on the meter board). Then connect a 1.5 pF (not electrolytic!) capacitor to the C x terminals. Set SI to 'd', switch on the mains supply, and note the indicated value. Then set SI to ‘c’, and adjust P2 so that the display indicates the same value as just noted. The position of the decimal point is irrelevant. Remove the 330 kft resistor and 150 pF capacitor and solder a 3.3 kfi resistor and 15 nF capacitor in their place. Finally, connect a 10 nF, 1% tolerance, capacitor across the C x terminals, set SI to position 'b', and adjust PI on the display board so that the display reads exactly 10.00 nF. If the 10 nF capacitor used has a larger tolerance, measuring results will also have a larger tolerance. This completes the calibration; all components should now be soldered into place. Applications The capacitance meter can also be used as interface for a digital voltmeter: the display board is then, of course, not required. Resistor R20 should be 100 kfl instead of 1 MS2 and a multi-turn preset of 1 Mil should be connected between terminals HI and LO. The wiper of this preset be- comes the output of the interface. The new preset will be used instead of PI (on the display board) for calibrating the circuit. There is only one (minor) snag: the decimal point is not in the right position! So, re- member this! It is also possible to use the capacitance meter for the measurement of varicaps, but it will then have to be provided with a variable voltage source. A design for 3-45 IC1 = ICL7106 IC2 = 4070 LCD = liquid crystal display type NDP530- 035A-S-RF-P1C \p*__ L *} ? e Summarizing .... . . . some of the outstanding points of the capacitance meter: ■ All capacitances are measured at the correct frequency. ■ Leakage currents have negligible influence on the measurement results. ■ The effect of wiring capacitances has been reduced to such an extent that capacitance values smaller than 1 pF may be measured. ■ After the capacitor under test has been connected, the display indication will appear in less than one second: this remains true for values up to 1 000 /tF ! M Sources: ‘Capacitance-to-voltage converter', W.B. de Ruyter, Wireless World, June 1983, page 68. ‘LCD panel meter', Elektor, October 1981, page 10-32. UK. Basicode-2 for junior plus Two of our recent projects, the VDU card and the Basicode-2 VDU rard interface, can both be used individually with the Junior Computer. However, there are bound to be some JC users who are interested in using these two 'extra's' together. The program given here was designed to do just this and thus provide the best of both worlds. Two versions of the software have been developed, for the extended Junior and for the DOS Junior. Basicode-2 for Junior plus VDU card Junior Computer + VDU card + Basicode-2 = the best of all worlds The description of Basicode-2 and the adaption to use Basicode-2 with the Junior Computer in particular have already been dealt with in Elektor U-K-102, October 1983. All details of the hardware and soft- ware needed are given there so we will not go into that again here. The only change needed to use Basicode-2 with the Junior Computer and VDU card is to modify the standard subroutines. Two tables of these subroutines are given in this article: one for the extended Junior with VDU card and the other for DOS Junior with VDU card. A few changes Some modifications or additions to the 'old' subroutines are needed. Subroutine 110 is changed. We have writ- ten a small machine-code program to speed up the positioning of the cursor (to HO, VE). Whenever a jump to line 20 is made in a Basicode-2 program (as always happens), then a piece of machine-code is first written into RAM. If the program then comes to subroutine 110 at any stage this machine-code program is called and the cursor is brought very quickly to the posi- tion defined by HO and VE. Subroutine 120, requesting the position of the cursor on the screen, is possible with this combination even though it did not work with the Junior/Elekterminal combination. The only routine that is still unworkable is subroutine 200. The Junior simply cannot determine if a key is pressed at a par- ticular moment. A GOSUB 200 in a pro- gram must therefore be changed. Actually there are two routines that do not work, the second being subroutine 250. However, the bleep that should be generated by a GOSUB 250 can hardly be considered essential for the correct operation of a program. One further important note. If the Basicode-2 translation program is used with the DOS Junior, great care should be exercised when using the DISK!" ..." command. If, for instance, a BASIC pro- gram is loaded from the floppy with the command DISK!"LO ..." and this program is then to be saved on tape in Basicode-2 format, the 'save' may not work because DISK!" ..." causes page zero to be 'swapped'. The result of this is that the pointers needed in the Basicode-2 trans- lation program are not longer correct. There is a very simple solution for this. After removing anything from (or storing anything on) the floppy disk the number 1 is typed and a (CARRIAGE) RETURN given. A dummy line has then been in- cluded and the pointers are again correct. Everything will the operate conectly pro- vided there is nothing on line 1, otherwise a different (blank) line number must be used. M 3-48 Elektor staff are a versatile lot! The present design came into being because one of our designers is a fervent speleologist, more popularly known as a caver. Regularly he risks life and limb in all sorts of dark caves, only to emerge hours later into daylight, covered in mud, sweating, and dead-tired, but happy and content. A good, reliable light source is, of course, indispensable for those treks in the dark. Many of the caver's lamp units in use today are powered by rechargeable (lead-acid or NiCd) batteries. Such batteries are inexpensive over their life — provided they are used often and regularly — and provide a near-constant output voltage. Dry batteries are relatively inexpensive to buy, offer small volume and low weight, and can easily be carried as spares. The last three points are, of course, of inestimable value during caving and in many other applications! Unfortunately, dry batteries have a serious draw-back: their output voltage falls linearly with time, so that at the beginning of their life the lamp burns brightly, while long before they are exhausted, the lamp begins to resemble a glowworm! Not only is this highly undesirable from a safety point of view, but it also makes for low efficiency. Our versatile designer decided, therefore, to design a voltage source for battery-operated lamps which offers a substantially constant output at high efficiency. constant voltage source. . . The design is basically a dc/ac converter based on a cleverly thought-out circuit which keeps the power supplied to the lamp, and therefore the light intensity, vir- tually constant over the normal life of the battery. The circuit itself has very low power consumption so that the efficiency of the whole is high. The principle lb control power at high efficiency, it is best to make use of pulse-width control. As the power supplied to the lamp must remain constant, the control should work so that the pulse width increases as the battery voltage decreases. Tb be sure, it is quite simple to design a pulse-width con- trol whereby the pulse width is inversely proportional to the supply voltage. That is, however, not the solution to the require- ment, because the power to the lamp is given by P = UfaVR, where Ub is the bat- tery voltage and R is the resistance of the lamp. What is required is compensation of Ub 2 , and this is achieved by using two pulse-width controls, operating at different frequencies, but with identical duty fac- tors (see figure 1). One reference voltage determines the pulse-width setting of both controls (the pulse width remains inversely proportional to the battery voltage). The outputs of the controls are multiplied in at AND gate, resulting in a signal of which the pulse width is inversely proportional to Ub 2 ! The circuit The constant voltage source is based on one IC — a quad comparator type LM 339 — and a couple of transistors (see . . for battery operated lamps Figure 1. Simplified block power fed to the lam| 3-49 Some arithmetic In the following, Ub = battery voltage Ue = effective value of pulse voltage D = duty factor of BOTH pulse-width controls P = power supplied to lamp R = resistance of lamp The duty factor, D is inversely pro- portional to Ub Each PWC delivers a (pulse) voltage of which U e = Ub^ D Multiplier A4, an AND gate which only recognizes logic levels, multiplies pulse-widths but NOT voltages: its out- put is, therefore U = Ub^ D^D = UbD The power supplied to the lamp is therefore P = Ub'DVR As both Ub and D are expressed as second-order quantities which are in- versely proportional, and R is constant, it is evident that P is independent of Ub- lamps versus Ub is given in figure 3. The circuit is suitable for use with input voltages, Ub, of 3.5. . .15 V. The average current consumed is about 15 mA. Connect an oscilloscope to pin 2 of IC1 and adjust PI until A1 just commences to oscillate. If no instruments other than a multimeter are available, the voltage source may be calibrated as follows. Connect a suitable lamp to the lamp terminals, and the multimeter (resistance range) between pin 6 of IC1 and the junction of Pl-Rl. Adjust PI for minimum resistance. Remove the multimeter and connect a suitable battery to the battery terminals. Adjust PI for good brightness of the lamp. H Calibration Calibrating the voltage source is fairly simple. Connect a suitable lamp to the lamp terminals and a variable, stabilized power supply to the battery terminals. Set the output of the power supply to the nominal voltage of the lamp used. Figure 3. Characteristics showing the efficiency of three different lamps ver- sus the battery voltage: the efficiency rises with larger lamp currents. 3-51 TELEPHONE BELL IC TYPE MC 34012 (Motorola Limited) The MC 34012 is intendend primarily as replacement for the usual telephone bell and is therefore of particular interest to Elektor readers who want a second telephone bell. The MC 34012 presents a load to the telephone line which is smaller than that of a "real" se- cond telephone bell. The input of the chip is connected to the usual input wires to the telephone bell, and its output to a piezo-electric buzzer (eg. a Toko type). As soon as the ringing signal (intermittent ac.) on the line exceeds 35 V, the IC is turned on and the buzzer emits a pleasant tone. Note that the IC needs no power supply as the necessary energy is drawn from the ringing signal! The chip does not respond to d.c. voltages such as the speech signals after the telephone receiver has been lifted. The quiescent current is therefore zero! CS - chip selekt DIGITAL CLINICAL THERMOMETER IC TYPE ZN 412 (Ferranti Electronics Limited) The recently announced ZN 412 contains all the necessary linear and digital functions to enable a clinical thermometer to be con- structed with a minimum of external com- ponents. The multiplexed data outputs of the chip are capable of directly driving a 3-digit seven-segment LED display. These outputs are controlled by an integral A/D processor which converts the output of an external probe into a digital number. A temperature range of 35.0 . . . 47.6°C can be displayed with an ac- curacy of 0.1°C and a response time of 5 seconds. The ZN 412 includes a self testing facility, battery status indication, reset, and display hold. Supply requirements are 4.5 V at 14 mA. Shown at the far right is a prototype of a digital clinical thermometer based on the ZN 412. PRECISION CENTIGRADE TEMPERATURE SENSORS SERIES LM 35 (National Semiconductor Corporation) The series LM 35 temperature sensors are precision ICs which have two important advan- tages over the usual sensors in that they are already calibrated and start at 0°C. Their output voltage is directly proportional to the temperature measured in degrees centigrade (10 mV/°C). The usual sensor must invariably be calibrated to obtain the required voltage/ temperature slope and starts at 0 K (— 273°C). The low output impedance of the LM 35 series (0.1 ohm for a 2 mA load), linear output, and precise, inherent calibration make interfacing these sensors to read-out or control circuits very easy. Power supply may be single or sym- metrical: the operating voltage may lie bet- ween 4 and 30 volts. Accuracy is typically 0.5°G Owing to the low current consumption of 60 hA, the internal heat dissipation amounts typically to only 0.08°C in still air. 3-52 COMPLEX SOUND GENERATOR ICS TYPES SN 76488 & SN 76495 (Texas Instruments Inc.) These circuits are up-dated versions of the SN 76477 which was featured in our March 1981 issue. The SN 76495 is a simplified version in a 16-pin housing, while the SN 76488 retains a 28-pin package. The main advantage of the pre- sent circuits over the SN 76477 is the on-chip audio amplifier which can deliver up to 125 mW into an 8-ohm load. Like the SN 76477, both circuits are compatible with computer systems. Unlike their predecessor, however, they operate from a 7.5 ... 10 V supply. An on- chip regulator provides a stabilized 5 V supply for driving external circuits, or for use as high logic level. Shown at the left is a typical demonstration cir- cuit of the SN 76495. POWER SWITCHING REGULATOR IC TYPE L296 (SCS-ATES) The L 296 is a high-power monolithic switching regulator (the first in the world according to the makers) which can supply a current of 4 A at voltages of 5.1 ... 40 V. As the IC is capable of operating at switching frequencies of up to 200 kHz, external components such as induc- tors and capacitors can be kept small and therefore relatively inexpensive. Features in- clude soft start (which slows down the rise time of the output voltage when the power is switched on), programmable current limiting (with the load current sensing resistor on the chip), reset output (intended primarily for microprocessors), and thermal shutdown at junction temperatures above 150°C. m ii ' i MAINS-CARRIER TRANSCEIVER IC TYPE LM 1893 (National Semiconductor Corporation) As the name implies, mains-carrier trans- ceivers use the mains supply lines to trans- fer information between remote locations. The LM 1893 bipolar chip performs as a mains inter- face for bi-directional (semi-duplex) com- munication of serial bit streams of virtually any coding. During transmission, a sinusoidal car- rier is FSK modulated and superimposed on the mains voltage via an on-chip driver stage. During reception, a PLL type demodulator ex- tracts the information from the mains. Some of the features of- the LM 1893 are: transmission rate of up to 4800 baud, choice of carrier fre- quency between 50 and 300 kHz, TTL and CMOS compatible logic levels, and regulated voltage to power logic. eleklor India march 1984 3-53 The syncbox is a circuit for use with a video-audio modulator (VAM) or a video combiner. It provides all kinds of signals that are needed to build up a complete video signal. A syncbox can be used, for example, to fill up the space between two recordings on a video tape. The 'noise' that would then normally appear can thus be replaced by a black image or a colour bar. video syncbox with colour bar This syncbox is an independent video signal source using an oscillator signal of 125 kHz to produce a number of basic signals that can be used for all kinds of video equipment and circuits. An external crystal-controlled signal can be used to clock the circuit if very high stability is demanded. Using the signals from the syncbox a black image (for video recorders) or even good quality colour bar can be produced. The circuit In the circuit diagram shown in figure 1 all signals are formed from the output signals of the 4040 (IC1). This IC, together with gates N4, N8 and N9, function as a 'divide by 2496’ circuit. A simple clock oscillator (N2, N3) supplying a frequency of 125 kHz, feeds the input of the divider. Using this signal the divider provides a raster fre- quency of 50.08 Hz. The raster frequency is normally 50 Hz, but because we want a non-interlaced image (that does not shake) we have chosen a raster time that is 32 ps shorter than normal. The number of lines per raster is then 312 instead of the normal 312V*. Without a lot of extra electronics in- terlacing is not possible with this circuit. The line frequency is the normal value of 15625 Hz, and this is necessary as the PAL delay line in colour television sets is tuned exactly to this value of 64 ps, Longer or shorter line times give rise to colour faults on the screen as colours run into each other. The line frequency (horizontal synchronization, HS) is derived by adding the oscillator signal from N2/N3 and out- puts Q0, Q1 and Q2 of IC1. Because of the fairly symmetrical block of the 125 kHz clock a synchronization pulse of about Figure 1. The circuit diegrem cleerly shows the simplicity ot the cir- cuit. Only a few CMOS ICs are needed to generate the necessary 3-54 4 fis width appears at the output of N7. The raster synchronization pulse is de- rived directly from the line synchroni- zation signal by inverting the latter during the raster synchronization time. The advan- tage of this is that the line synchronization signal remains safe during the raster syn- chronization. The electronics in the TV set automatically ensures that the inverted line synchronization is recognized as the raster synchronization. Switching between line and raster syn- chronization is handled by FF1 which is clocked at the line frequency from output 02 of IC1. The positive edge of this clock signal occurs in the middle of the line time so the raster synchronization, which lasts for eight line times, always starts and ends with a half line. The outputs of FF1 are connected to N10 and Nil, which in turn feed N12, thus combining the line and raster synchronization. Note that the raster synchronization signal eventually has the same polarity as the HS signal (as a quick glance at the timing diagram of figure 2a will confirm). The HS signal is also fed to the outside world, where it is used in colour video systems to control the PAL switch. The CS signal (Composite Synchronization) is not suitable for this because it contains an extra positive and negative edge during the raster synchronization (see figure 2a). One of these two edges will trigger the PAL flip-flop (in the VAM or video combiner) one extra time which is enough to con- fuse the receiver and activate the colour killer. The problem is avoided by using the HS signal. The carrier for the colour information must be regularly synchronized in order to keep the colours reproducible. This happens directly after the line synchron- ization by means of a burst signal (con- sisting of a number of periods of colour carrier with a fixed phase). The BE (Burst Enable) signal (or BE) is used to activate this burst. This signal is generated with the aid of two monostable multivibrators formed from FF3 and FF4. The inverted trailing edge of the HS signal triggers FF3, and the output of this flip-flop gives a pulse of 1.6 fis (set with P2). The same trail- ing edge of HS triggers FF4 and this in turn gives a BE pulse of 2.2S p s (set with P3). This is shown by the small timing chart in figure 2b. Small deviations from these times are rarely a difficulty as neither a shorter delay between synchron- ization and burst nor a longer burst pulse are likely to cause any problems. A blanking signal is not absolutely necessary but it is often handy. In our cir- cuit this signal is produced by FF2, which, again, operates as an MMV. The pulse width is set to about 12 ps with R2 and C2. During the raster synchronization the pulse widths of FF2, FF3 and FF4 are defined by FF1 as this then disables the set inputs of the three flip-flops. A false burst pulse is then given by the BE out- put, but this causes no adverse effects as J LT be — r ITT IT be_ii n_r t n_ 2b it appears in the middle of the line time. Simultaneously the signal at the set input of FF2 causes a raster blanking signal CBLK (composite blank) to be generated. Construction The printed circuit board layout for this circuit is shown in figure 3, and if this is used construction should give absolutely no problems. We do, however, rec- ommend that sockets be used for the ICs. 3-55 4 Figure 4. This is how the video syncbox is con- nected to the VAM end video combiner respectively. R1 = 100 k R2 = 18 k PI = 100 k preset P2.P3 = 10 k preset Capacitors: Cl = 82 p C2 = 820 p C3 = 560 p IC4 = 4011 IC5. IC6 = 4013 The supply for the circuit can be between 5 and 12 V, and current consumption is only a few milliamps. To use this circuit with the video combiner elsewhere in this issue we suggest that that article be read before building the syncbox. An oscillo- scope is needed for adjusting the three presets. Failing this they will simply have to be set up 'by eye'. Use The syncbox is only useful when com- bined with some other suitable circuit. It could, for example, provide the control signal for a simple pattern generator, or it could be used with the video/audio modulator (VAM) from Elektor, February 1983, or the video combiner in this issue. Outputs BE, CBLK and CS of the video syncbox must be connected to the BE, BE and sync inputs of the VAM. Links V-W and X-Y on the VAM board must be re- moved. If the VAM is only used in combi- nation with the video syncbox, IC4 and IC5 of the VAM may be removed. A blank image (for example to fill up a vi- deo tape) can be obtained by connecting •he BL input (or the R G B inputs) to ground. For a colour bar there are three extra connections between syncbox and VAM needed. Points R, G and B of the syncbox must be connected to the R, G and B inputs of the VAM. The three inputs per colour of the VAM can be connected together. The resulting colour bar has the following colours in this sequence (from left to right), blue-red-magenta-green-cyan- yellow. White and black do not appear on the screen. Different colour combinations and patterns can be made by using dif- ferent outputs of the 4040. If the syncbox and video combiner are used together the 'common' points on both boards must be linked. In this case points R, G and B need only be connected if a colour bar is required with this combination. K 3-56 elek.or in automatic battery charger Recharging lead-acid batteries is often assumed to be an extremely straightforward matter. And that is indeed the case, assuming that no special demands are being made on the life of the battery. On the other hand, if one wishes to ensure that the battery lasts as long as possible, then certain constraints are placed upon the charge cycle. Figure 1 illustrates the ideal charge current characteristic for a normal 12 V lead-acid battery which is com- pletely discharged. During the first phase (A— B), a limited charging current is used, until the battery volt- age reaches approximately 10 V. This restriction on the charging current is necessary to ensure that the charger is not overloaded (excessive dissi- pation). For the next phase (C-D), the battery is charged with the '5-hour charging current'. The size of this current is determined by dividing the nominal capacity of the battery in ampere-hours Ah) by 5. At the end of this period the battery should be charged to 14.4 V, whereupon the final phase (E— F) starts. The battery is charged with a much smaller 'top- up' current, which gradually would decrease to zero if the battery volt- age were to reach 16.5 V. The circuit described here (see figure 2) is intended to provide a charge cycle which follows that described above. If the battery is completely discharged (volt- age <10V), so little current flows through D3 that T1 is turned off. The output of IC1 will be low, so that the base currents of T2 and T3, and hence the charging current, are determined solely by the position of PI. If the battery voltage is between 10 and 14 V, D3 is forward biased and T1 is turned on. The output of IC1 still remains low. so that the charging current is now determined by both PI and P2. If the wiper voltage of P3 exceeds the zener voltage of D1. then due to the positive feedback via R4, the output voltage of IC1 will swing up to a value determined by the zener voltage of D1 and the forward voltage drop of D2. As a result T1 is turned off and the charge current is once again determined by the pos- ition of PI. In contrast to phase A— B, however, the higher output voltage of IC1 means that current through PI, and hence the charging current, is reduced accordingly. Since D2 is forward biased, the effect of resistors R2 and R3 will be to gradually reduce the charging current still further, as the battery voltage continues to rise. To calibrate the circuit, P3 is adjusted so that the output of IC1 swings high when the output (i.e. battery) volt- age is 14.4 V. By means of PI the 'top-up' charge current is set to the 20-hour value (capacity of the battery in Ah divided 3-57 R1 R2 R3 12k 10 k 82 k ,R6 = 8k2 = 100JI - 3k9 P2 P3 100 k preset 220 k ... 250 k preset 10 k preset Capacitors: Cla ■ Clb - 4700 u/40 V T1 - TUN T2 a BD138. BD140 T3 - TIP2955 D1 - 6V8. 400 mW zener diode D2 - DUS D3 - 5V6, 400 mW zener diode IC1 - 741 by 20) for voltages between 14.5 and 15 V. Finally, with a battery voltage of between 11 and 14 V, P2 is ad- justed for the nominal (5-hour) charging current. The initial charging current (phase A-B) is set by the value of the 'top-up' current, and depending upon the characteristics of the transistors, will be approximately 30 to 100% greater. Miscellaneous: Tr - 16 V, 8 A mains transformer B - B80C10000 bridge rectifier fuse 1 0.5 A slo-blo Siemens Components Report Volume XIII. No. 1 March 1978. W. Ferdinand When starting a car journey after dark it is useful to have a device which will keep the interior lighting on for a while after the doors have been closed, and so make it easier for the occupants to fasten safety belts and insert the ignition key. This can be done with the simple time-switch circuit shown. afterburner maim THE VI AUTOMATIC’ light control , SWITCH _ i ANTENNAS POCKET TRANSISTOR TABLE LAMP-CUM-CLOCK Maxtronix are handlinq the line of antennas manufactured by Centurin International, Inc., of USA, which come in six styles of construction, viz: high band, low band, UHF Whip, Stubby VHF, mini VHF and new UHF Y4 wave. They are capable of handling greater than 50 watts of power continuously at 50 OHMS GP impedance. Manufac- turers also offer telescoping antennas and the traditional Rubber Ducky antennas available in 25 different connector styles, besides adapter cable assembles, right angle adapters and scanner antennas. Navabharat Radio Agencies have come out with a new transistor in their Jetking range claimed by them as the slimmest medium wave pocket transis- tor. It has 1 diode and 6 transistors and operates on two penlite cells. It is provided with an earphone and a wrist strap for single-hand operation . Electronics Hobby Centre have developed a new product, viz: table lamp-cum-clock. The clock has a quartz movement and the table lamp is provided with a dimmer system. Approx, base size is 23 x 1 1 x 4 cms and height of lamp 29 cms. For further details, write to Electronics Hobby Centre F-37, Nand-dham Industrial Estate Marol, Bombay 400 059 More particulars Irom Navbharat Radio Agencies 350 Lamington road Bombay 400 007 PCB ASSEMBLY JIG PCS Assembly Jig (Universal), introduced by Time Engineers, is an aid for PCB assembly, deisgned to increase productivity with ease. The device has adjustable grooved sliding rails to accomodate PCBs of different sizes. After the PCB is mounted, components of various sizes are loaded and the foam pad clamped, the foam exerting sufficient pressure all over the PCB to secure components in their respective positions. The jig is then turned to obtain convenient working angles for soldering. CABLE TIES More intormation Irom Maxtronix Box 332 GPO Hyderabad 500 Ont Novoflex Cable Industries have developed cable ties for use in all sophisticated electronic equipment like computers, communication equip- ments, nuclear and process control instrumentation, inverter, power supplies, etc., as also in consumer electronic equipments such as televi- sions, stereo systems, PA systems, intercoms, VCRs etc. The manufac- turers claim that the cable ties are of unique design, possess high tensile strength and flexibility. They are available for bundling diameter of 1.6 mm to 51 mm. AUTO SWITCH Guiarat Electronics & Controls have developed an automatic light control switch designed to put on the lights at dusk and switch them off at dawn, irrespective of sun set/rise timings! Using ICs and working on the principle of light dependent resistor, it is provided with starting and final delavs to overcome the effect of cloud and lighting respectively. Has auto off/ manual switch and neon indication and comes in single phase supply models in 10A and 25A. More details can be had Irom Novoflex Cable Industries Block A-14, 7th Floor, Chatterjee International Centre 33-A, Chowringhee Road, Calcutta 700 071 Further aprticulars may be had from Time Engineers P. Box No. 58 MIDC Rly Station, Satara Village Road Aurangabad 431 001 For further details, contact Guiarat Electronibs & Controls 9, Advani Market, Outside Delhi Gate Ahmedabad 380 001 maim lODOO MOTOR PROTECTOR A METER ELECTRONIC IGNITION introduced a CVH Lektronics have deloped an with four electronic ignition system for cars ind 0.1 mill- which, according to them, does away e accuracy of W ith starting troubles prevalent in the iment. Other conventional system, particularly on y/mains ope- cold mornings or when battery is weak, andy. has 3V4 Some of the advantages listed by the display. makers are: easy starting even at very low temperature less current consum- ptiorr: eliminates burning of contact Wi. breaker point. The system also comes Product Promoters have promoted Miller Electro Protector with advanced electronic circuitry and sensing, the main function of the device being to protect electric motors against excess or low current, high, low or unbalanced voltages, excess load, short circuit and absence of one phase. When current or voltage are beyond safe limits ind rated values, the device will cut off the mains supply and restart automatically after proper supply is restored. It has been approved by National Physical Laboratory, etc., and models are available for all kinds of electrical equipments and upto 100 HP. i milli OHM TYPE VRM22M More information CVP Lektronics 6, Second Main R Annamalaipuram Madras 600 028. Further details can be had from Vasavi Instruments Corporation 162 Vasavinagar Secunderabad 500 003 Detailed information Product Promoters Post Box 3577 New Delhi 1 10 024 DIAL CLAMP Instrument Control Devices have introduced the dial clamp, designed to prevent detuning of the dial when locked by the captive knurled nut. The single- hand operation consists of halt turn rotation of the thumb nut either clockwise or anti-clockwise to lock or unlock the dial respectively. The manufacturers recommend its use in temperature controllers and process controllers to prevent accidental- rotation of the dial. THERMISTORS Electronics Unlimited manufacture a range of thermistors, the type EUBC incorporates a directly heated head of semiconductor material in a solid glass pellet, connection being by means of two dummert wires. The whole unit is encapsulated in brass and sealed in silicone with silicone-coated fibre glass sleeve. As this type of thermistors is intended to operate at temperatures upto 300° C, the dummet wire is untinned to render it suitable for welding or brazing. Normally available in a tolerance of *20%. they can also be supplied in +1%. DIGITAL WATTMETER Norma Messtechnik GmbH, of Austria, manufacture a precision digital watt- meter, model D 4155. for measurement of active power of single and polyphase electrical circuit at any load. Accuracy of the instrument is monitored through automatic reference test. Salient : features are: input voltage - 110. 220, 440, 550 V; input current - 1 .2,5, 10 A upto 200 A visa shunt; power range - 220 W... 11 kW; frequency - 15 Hz ... 1 kHz; accuracy - *0.1% of range; display -4% digit. Further particulars can be hi Instrument Control Devices 14 Manorama Niwas Datar Colony, Bhandup Bombay 400078 For further information, contact Inde Associates FREE Dept., 202, Vikram Tower. Rajendra Pis New Delhi 1 10008 More details from Electronics Unlimited 70, Krishna Kunj, Gen. Thimmayya Road Pune 411001 V including batteries. Thandar Electronics Limited, London Road, Cambridgeshire, PEI 7 4HJ Telephone: 0480 64646 Electronic tag for motor c, has many roles Several European motor car manuf, company car only when authorised. If there is a pre-set limit for the vehicle, the pump will give just that amount. Otherwise the pump will automatically shut off once delivered may instantly be debited, through the Eureka system, direct to the vehicle's owner. (LPS). lainst theft after _ which has been paid for construction in which the base and top company, are nominally 2 mm and 1.6 mm thick Eureka's system is based respectively, the standard colour scheme operated tag. about the si is brown base and beige top panels, with stamp, which continues the whole units held rigidly together by own identity code to ele an a battery- 13 amp socket tester ! of a postage For the professional and the housi transmit its this pocket sized tester can save ironic sensors time. Just plug in and switch a of the completed vehicle can be logged as ini it leaves the factory for storage parks. Ca Even when sold to its eventual owner. SCi the tagged car may continue to be moni- Lit tored by sensors which can be placed at No 1 st ructions i International , Telephone: 0492 641298 0638 716101 (28371 LCD digital multimeter Low cost multimeter 3-62 THE CURRENT * CHOICE FOR! When you own a computer system, electro- medical equipment, laboratory equipment, paper copiers, photography equipment. ..or if you plan to buy one. you know how imperative and essential it is that your equipment gets protection that's absolutely reliable. Your search for protection that's totally invincible has ended. KEURON SERVO-CONTROliED VOLW3E S1ABUSERS • Backed by years of electronic expertise • Easily available in the market. • Wide range from single-phase 1 KVA to 3-phase 45 KVA. Specifications made to order to suit your needs. • Wide spread after-sales service network) all over the country. £f ,t r. on SCVS-although you pay a little moreO Ml Ilf I- f-f|P you h ave the satisfaction of knowing ■ I l r -r. P r Vm9 for superior quality and A I l-r-t— fh» h h^ reliability.For a product that is made with AK\I ll I | I uZ the highest grade components in the market. “UVVLU I L hfilS"? manufactures and institutions ha& r— 1_ - 1 ™ Ptoof of Keltron's unsurpassed quality. r ■ So, why hesitate? y q | L\^ | IV JN five VO. ur equipment, absolute protection ^ and optimum voltage, with a Keltron SCVS. 1 ™ ptoof of Keltron's unsurpassed quality. ppnTPrTirMvi s ° why hes,,a,e? o ■ I LVm I l\^/| N y ? Ur equipment, absolute protection ^ and optimum voltage, with a Keltron SCVS. KELTRON SERVO-CONTROLLED VOLTAGE STABILISERS elekt©? switchboard I FOR SALE - circuits with servic- WAN I tu - vna eie. ina tiDS for various Japanese tuner, pushbutton pushbutton telephone/diallers, dialler, circuit. UHF/VI Subhankar Dutta. P-94. Lake ter. Taiwan make Natic Road. Calcutta 700 029. pocket radio PS . , . . 2284, Sector 35-C. C REQUEST - full details about , 60 036 Te | : 32670. Ham radio operator licence. WANTED - two osc. coils for GUIDANCE -requested from ex- l» Cosmic tapedeck model perienced ham operators, prefer- | l lyf C06500D or any other osc. coils ably local, for becoming ham “ 1 with circuit diagram. Eric Fer- operator. P.V. Mohamed Rafi. ■ -1 nandes. Rose Blossom. 1 st Floor. Quick Tempo Service, llnd Main ItAtlrtl C-26 Shitaladevi. Mahim. Bom- Road. R.C. Puram. Bangalore WUIUU bay 400 01 6. 560021. WANTED - Harmon & Kardon FOR SALE - 1 5-tunes musical Nn 7155 for US make ster- door-bell kit. Rs. 65-00, postage WANTED - VHS electronic TV Ra)U . T exIra . Remit Rs 10-00. balance tuner, pushbutton telephone ~ ar " pl '™' \ M “J^ n t Park by VPP. Post-free if full value is dialler, circuit. UHF/VHF conver- Nagar. no. o ' v" ’ • „ I Somasunda- ter. Taiwan make National MW 1C Street agar. 439/4 | C F South Colony, pocket 'adio. P5. Chalokia 600 017 Madras 600 038. 2284, Sector 35-C. Chandigarh F0R SALE - unused latest 160 036. Tel: 32670. Canon A1 electronic earners WANTED - FM wireless mic. cir- UC1 _ . r „ inn . v equipment, (automatic flash etc.) cult and good quality recording Ham radio operator licence. eouioment (automatic flash etc.) cuit and good quality recording Prabir Debnath. C/o : H.S. HELP! - Can anyone supply p R p' ansa ,e. 6-D. Madhav playback tape pre-amp circuit Debnath, (Dutt Kumar's House), simple circuit diagram or iviw , 15-A Bhavanishankar with pcb. Vivek, "Ambala .51, tmnsiner receiver 7 Harish V Nagar, 115-A. Bhavams Shetty. MT C' type quart Buildg Road. Bombay 400 028. No.1 48-2/4. Sadda. Marmagoa WANTED - basic manual for ^ ....... WANTED - tried simple circuit Ha , bour 403 804 Casio PB 1 00. original/xerox and WANTED - 1C 41 5E and full kit of TV games in exchange of other software for same also Sinclair of personal FM receiver, except circuits or against payment. C. FOR SALE -sparingly used word Soectru'm Cash pur- headphones, published in Elektor Sanjay, 11. Balaram Street, processor upgradab e computer chaseSSN ^ pa | 20 5, Ash,ana. Oct.'83. Rammder Singh C/o : Adyar, Madras 600 020. Epson MX80 printer Instalment Pratap Society Off J.P. Industrial Hardwares, Guru Bazar. WANTED - 1C TDA 7000 and ‘‘To" Dh“ **■ Andlwi ITO Bomb,, Cbo»* Sod.n, J.I.odh,. coils for personal FM project Hanuman Roa d, Vile Parle (East). 400 058 (Dec. 83): also tried circuitry for Bombay 400 057. FOR SALE - secondhand valve WANTED - ZX Spectrum mm DMM. against cash. Vijay N. type PA amplifier and matching 16K ram with manuals printer Phadke, 63 Ganesh Baug. 208 Dr WANTED - walkie-talkie circuit with Iape m i C . rec programmes, etc. Paresh Sheth Ambedkar Road. Matunga. and circuit for small MW moots Output 60W. Anand 4B, Navyug Nivas, 1 67 Laming Bombay 400 019. ^L^.hi^TaT^haseTl Sh.rali. Flat A. Bldg 6, Dhake ton Road, Bombay 400 007 WANTED - information and cir- gAg N M 0hali. Ropar Colony, Road. Andhen W. Tel 89032 cuits of walkie-talkie using indi- , 60055 Bombay 400 058. FREE EXCHANGE -of software FOR SALE - secondhand valve WANTED - ZX Spectrum min. type PA amplifier and matching 16K ram with manuals printer. Ambedkar Road. Matunga. and circuit ror Bombay 400 019. nutter with g< Dhiman. Kothi WANTED - information and cir- gAS Nagar cuits of walkie-talkie using indi- 1 &Q Q55 qenous components and rets photo linear. Subhash P. Singh. INTERESTED 1 83/7, Sheel Kunj. University of shortwave lis Roorkee, Roorkee247 667. with amateur 208 Dr WANTED - walkie-talkie circuit