elector 80 December 1981 U.K.65p. up-to-date electronics for lab and leisure talking board give your pp a voice ! combination a novel approach IPROM a battery RAM ! elektor decer 1981 - UK 03 electronics in focus — competition results! 12-01 selektor 12-02 talking board 12-04 The speech synthesiser featured in this article consists of a single board that can provide a microprocessor system with a vocabulary of several hundred words. We hope that the first words it utters are not - 'take me to your leader' . . . IPROM 12-15 This article describes a 'plug in' instantly programmable read only memory which will fit in a standard 2716 EPROM socket. capacitance meter module 12-18 This add-on module converts a normal frequency counter into an instru- ment capable of accurately measuring capacitor values. NiCad battery monitor I wo. Rothi 12-21 A simple but highly effective method of keeping NiCads permanently 'topped up' is described in this article. lap counter/timer (a. schwain 12-22 This project provides a slot racing track with electronic timing and lap counting facilities bringing it into line with all the well known racing circuits throughout the world. cumulative index 1981 12-26 coming soon . . . perhaps! 12-28 teletext decoder part 3 12-29 The final article of the Teletext decoder series contains the receiver and the video control board. No modification is required to the TV set with these additions. combination lock (R. deBoer) 12-37 A combination lock that used only one rotary control and yet features an infinite number of possible 'combinations'. the new Elektor synthesiser 12-39 This, the third article in the series, begins practical construction with the voltage controlled oscillator. The design and performance of the syn- thesiser depend to a large extent on this module. flashing lights 12-45 These lights can be fitted to an inexpensive toy car to provide an effect very similar to the warning lights seen on ambulances, fire engines and market 12-46 advertisers index UK 24 EDITOR: P. Holmes UK EDITORIAL STAFF T. Day E, Rogans TECHNICAL EDITORIAL STAFF J. Barendrecht G.H.K. Dam E. Krempelsauer G. Nachbar A. Nachtmann K.S.M. Wal raven THE ELEKTOR TALKING BOARD The complete kit of parts for the Talking Board is available £76.00 now The kit for the Talking Board has been produced for the project described in this issue of Elektor. Features of the Talking Board include: single board construction expandable vocabulary low power consumption self contained memory TMS5100 WOODHILL LANE, SHAMLEY GREEN, near GUILDFORD, SURREY Telephone 0483 893236 iber 1981 - 12-01 dedrmks in focus... ...compcddon results! We have come to expect a flood of entries for an Elektor competition - not surprising, with close on a million readersl - and 'Electronics in focus' is no exception. What did come as a very pleasant surprise is the high quality and extreme diversity of the photo's and slides. Single components and complete circuit boards from all angles; close-ups of integrated circuit 'chips', all kinds of photographic and lighting tricks and effects. Some themes proved highly popular: 'components among the flowers' and 'electronics in space', to name just two. The deflection coils for TV picture tubes also turned up at regular intervals. All entries were judged on the following points: - initial 'impact' of the photo; - how clearly does it relate to electronics in general (components or assemblies clearly visible? functional in the photo? etc.); - technical aspects: lighting, focus, etc.; - finally, and mainly to distinguish between almost equally good entries: how 'pure' are the electronics in the photo? A good picture (with high impact etc.) of a few resistors scored marginally higher than an equally 'good' picture of a little man built from a few resistors. Admittedly, we were sorely tempted to add a few further points, like 'how suitable is the entry for printing?' and even 'how easy is it to unpack?! I' . . . But we disregarded these factors, since they were not specified in the original competition rules. Finally, after a lot of soul-searching and long evaluation and comparison sessions — burning the midnight candle at both ends, so to speak — our jury came up with the list of prize-winners shown here. Numbers 1 ... 28 are given in order or merit; 29 ... 78 are in alphabetical order. The next step, obviously, is to publish the prize-winning entries. However, this takes some further organizing: among other things, we must now obtain the original negatives for several of the photo's. If all goes according to plan, we hope to present the winners (in colour!) in the coming February issue. Meanwhile, our congratulations to all prize-winners! The results were astonishingly good. We would like to add that the overall standard of the entries was better than expected — it seems that electronics and photography are hobbies that fit together rather well. First prize of £ 200: 1. P. Gottschalk, GutenbergstraSe 14, 3014 Laatzen 1, Germany. Second prize of £ 100: 2. S. Vernimb, Graumannsweg 46, 2000 Hamburg 76, Germany Third prize of £ 50: 3. D. Campe, Schaubeke 52, 9160 Hamme, Belgium. These readers each receive £ 20: 4. J.W. van Boordt (Holland) 5. D. Reetz (Germany) 6. A. Kwint (Holland) 7. C. Bosch (Switzerland) 8. S. Vernimb (Germany) 9. P. Ekholm (Sweden) 10. K.D. Kromer (Germany) 11. R.M. Smart (England) 12. F. Kolling (Germany) These readers will each receive a free : P. Baas (Holland) G. Bauer (Germany) P.J. Beauprez (France) K. Becker (Germany) K. van der Bent (Holland) F.M. Berden (Holland) B. Bois (France) H. J. Brede (Germany) A. Chaxel (France) J. Drescher (Germany) B. Duranteau (France) J.P. Dzido (France) J. Erker (Germany) H. Feller (Germany) F. Fleer (Germany) S. Fischer (Germany) S. Folliot (Belgium) 13. G. Gorzawski (Germany) 14. D.R. Newell (England) 15. A. Hogeveen (Holland) 16. G. Combe (France) 17. D. Campe (Belgium) 18. J. van den Boom (Holland) 19. A. Kwint (Holland) 20. H. Kottke (Germany) 21. HJ. Figge (Germany) iription for 1982. M. Gerlach (Germany) D. Guillermin (France) B. Haugrund (Germany) M. Held (Germany) A.C. van Hoboken (Holland) F. Jacquot (France) M.G. Jekel (Holland) C. Kohlpaintuer (Germany) V. Kulhanck (Germany) H. van Laarhoven (Holland) J. Laatikainen (Finland) C. Labrut (France) G. Landen (Germany) W. Lehrke (Germany) M. Levy (France) W. Majdic (Germany) Q. Peeters (Belgium) 22. K. Langbehn (Germany) 23. S. Vernimb (Germany) 24. K.D. Kromer (Germany) 25. P. Ekholm (Sweden) 26. R.M. Smart (England) 27. P. Sadonis-Heyse (Belgium) 28. F. Chanet (Belgium) R. Perry (England) E. Peters (Germany) M. Przewloha (Germany) N. Reneberg (Germany) A. Russel (Holland) H. Scholhorn (Germany) R. Slomski (Germany) H. Solter (Germany) R. Thormann (Austria) J.F. Tinot (France) V. Ulle (Germany) L. Veldkamp (Germany) J.A. Walton (England) H.M.F.J. de Wijs (Holland) F. Zwinger (Germany) S. Zywietz (Germany) 12-02 -elektor ar 1981 Solar-hydrogen plants Since the oil-crisis of 1973 the energy problem has become an important issue in our western society. From that point on people started to occupy themselves more intensively with this problem. Energy was saved as much as possible and research into new, not fossil, energy sources sue h as sun, wind and water energy was intensified. These forms of so-called renewable energy can be used to generate electricity, warmth and labour (pumping-engines). Since the oil and gas prices are increasing all over the world, these alternative energy sources are becoming more and more important. During a recently held conference con- cerning photo-voltaic energy (generating electricity from the sun) Reinhard Dahlberg, a leading researcher of AEG- Telefunken, unfolded a plan to cover the world's energy demand up to the year 2040. If Dahlberg's plan is ap- proved, AEG-Telefunken will start the construction of two experimental plants, each having a capacity of 10 MW. These solar-hydrogen plants will convert sunlight into electricity, with which water can be separated into its two components: hydrogen and oxygen by means of electrolysis. Dahlberg is convinced that thousands of solar power stations can be put into use within a few decades. Here we have a closer look at the two processes involved, namely solar electricity and hydrogen (as energy storage). Solar cells The first silicon solar cell used to con- vert solar energy into electricity was demonstrated by assistants of the Bell Telephone laboratory in 1954. This type of cell consists of a wafer of n- doped silicon covered by a p-conducting material. Consequently a p/n-boundary layer comes into being. When the cell is exposed to light an absorbed photon produces two 'holes' in the silicon, which creates a surplus electron. The liberated electrons can't pass the p/n junction because it functions as a barrier. However, they will pass via the metal contacts of an external ring with which the p and n layer are connected. Consequently a voltage of approxi- mately 0.5 V is generated between both electrodes of the solar cell. In the beginning the efficiency of solar cells was very low (about 5%), but due to the improved manufacturing tech- niques it could be raised to 10%. Since fossil fuels were very cheap before 1973 to solar cells were mainly used in astro- nautics. The electricity supply of most satellites is supplied by solar generators. The cost for this application was about £ 65 per watt, but due to the increasing oil prices solar cells suddenly became economical. Nowadays solar electricity costs about £10 per watt. It could be used to generate electricity on vessels, islands or for water pumps in warm, isolated areas. Experts expect the solar cell to become a competitor for the con- ventional energy sources. Hydrogen As stated before, Dahlberg's plan not only discusses solar energy but also hydrogen as energy storage. Neither electricity nor hydrogen are an energy source, they are just a way to distribute energy. The use of hydrogen has some considerable advantages. Its transport is easier than that of electricity and it is easier to store in large quantities. It could replace natural gas in every appli- cation. Last but not least, hydrogen can generate electricity by means of so- called fuel cells. Cars could run on hydrogen! However, the major problem is that the fuel tank becomes a high explosive bomb, when exposed to oxygen. Hydrogen can be obtained in several ways. When heated enough (3000°C) water can be divided into its two com- ponents: hydrogen and oxygen. Hydro- gen can also be derived by electrolysis. It can be transported in three different — as gas by pipe-lines — as liquid under pressure in containers and tankers — as a solid by combining it with other elements, which evokes the so-called hydrides. elektor december 1981 - 12-03 Solar-hydrogen plants Dahlberg's suggestion combines the advantages of solar electricity and hy- drogen. This suggestion was given a lot of consideration before it was intro- duced. According to his plan, gigantic solar power stations should provide enough electricity to produce hydrogen by electrolysis. As well as the produc- tion of hydrogen, the solar cells would supply electricity to a factory making solar cells. Consequently such a 'hydro- gen plant' would consist of a solar power station, a hydrogen electrolysis installation and a production unit for solar power stations. In this way a hy- drogen plant would eventually produce enough material to 'give birth' to a second plant. A production-unit for a solar power station would contain: - a glass-works (for coating the solar- cells) - a factory to produce silicon which can be used to make solar-cells - a factory to convert silicon into solar- cells and to combine these single cells into panels - a factory to combine panels and dynamos into solar power stations. After building one solar-hydrogen plant enough material would be obtained to build an identical plant in a few years time. According to Dahlberg these enormous plants could be situated in 7 deserts; in Africa, Australia and North and South America. In the year 2040 the plants would cover an area of about 2,000,000 km 2 . The hydrogen could be transported to the 'civilised' world via pipe-lines or as hydrides. Dahlberg divided his plan into three phases: 1979-1989 (first phase); 1990- 2000 (second phase); 2001-2040 (third phase). During the first phase the main concern is to build a factory for the production of solar cells, using con- ventional energy sources. Dahlberg told his audience that AEG-Telefunken is negotiating with several firms in Japan, Europe and the United States in order to form an international consortium. Its task will be to build a factory for pro- ducing solar cells having a total capacity of 1.5 Gigawatt in 1989. In the second phase (1984) AEG would start the construction of two prototype plants having a capacity of 1 Gigawatt each and covering an area of 10 square kilometres. They should start produc- tion at the beginning of the nineties. They will provide hydrogen and solar If the initial plants appear to be success- ful, 10 plant families with 10 plants each will be built between 1990 and 2000. They will cover an area of 200 square kilometres. The generators for the first family (family A) will be sup- plied by the factories of the inter- national consortium. Thereupon AEG will begin the construction of families B and C in 1996. These factories will be built from parts supplied by the consor- tium factories. By then the members of family A will reach the last stage of their self-replication towards the end of the century. During the last stage of Dahlberg's plan all members of families B and C will be completed towards 2005. Together with family A they will supply the parts required for families D, E, F, G, H, I, J. Ambitious but practicable The greatest investments will have to be made in the period between 2000 and 2040, due to the fact that the number of plants is raised to a square. By the year 2005 the plants will only be used to produce hydrogen. The main solar cell production will take place in the second generation. The third and fourth generation of family A will start produc- tion in 2020 and 2030 respectively. The fourth generation of the families B and C will be completed in 2035. The 10 families, producing 10,000 copies during a period of 50 years, will provide the world with 100,000 plants. Their total production will be equivalent to 15 bil- lion tons of oil (15,000,000,000,000), which is four times the world energy consumption. The whole operation will cost about 14 billion pounds (£14,000,000,000). Dahlberg admits this amount to be a 'little' steep, but he points out that the same amount of money will be needed for oil as fuel in the conventional power stations during the next 20 years. Until 1989 'only' £6,000,000,000 will be needed. The largest investments have to be made towards the end. Dahlberg admits that his plan is ambitious. How- ever, he sees no reason why his plan shouldn't give considerable thought. After all, hydrogen as an energy source fulfils all the conditions to become the fuel of the future. It can match other energy sources. There is no shortage of raw materials when hydro- gen is produced by means of solar energy. Dahlberg's plan makes it pos- sible to produce enough hydrogen to replace the fossil fuels before they are exhausted. 3 Photo 3. The advantages of solar energy are 1 2-04 - In an earlier article ('Chattering chips', Elektor September 1981), several speech synthesis systems were discussed. For various reasons, the Texas Instruments 'Solid State Speech' system seemed the best bet — certainly for microprocessor enthusiasts. In the first place, it can produce an output that is something like a human voice coming over a telephone line: not hi-fi, admittedly, but good enough to notice traces of an american accent coming through! Furthermore, the coding system used is fairly 'logical '. which means that it is quite feasible to work out codes for new words — without having to resort to a huge computer. talking board a solid-state voice I n the early days of science fiction, robots could walk and talk like human beings. Later on, as authors learned of the possibilities and limitations of computers, it became more realistic to reserve the power of speech for huge, 'space-ship filling' electronic brains. Now, in this project, we can proceed to science fact: a single board that can provide a vocabulary of several hundred words for a microprocessor system! Having decided to use the Texas Instru- ments system, the next step is to make a choice between the two versions: the older TMS5100, intended for talking games and the like, or the new TMS 5200 that is intended for use in microprocessor systems. Surprisingly enough, we decided to use the 51 00, for two good reasons: there is a much larger vocabulary available for this chip, as well as a good circuit in the Tl application note! With only a few further modifications and additions, this system can be interfaced to almost any microprocessor system. The basic principle of the actual speech synthesis process will be discussed later. For the moment, the only important thing to know is that a serial bit stream must be fed into the 'VSP' (Voice Synthesis Processor) in order to make it talk. For the word 'help', say, a total of 534 bits are required: just less than 67 bytes. Since this is a fairly short word, it will be obvious that a consider- able memory range is required for a total vocabulary of several hundred words. To avoid wasting memory range in the 'host' microprocessor system, the 'speech memory' is included on the speech board — complete with a local address counter and associated control circuits. The block diagram of the 'talking board' is given in figure 1. The lower half of this diagram shows the memory and control circuits. Initially, the first address for a given word must be loaded into the address buffer/counter. Since 16-bit addressing is used, the first address is loaded in two bytes (8 bit) : first the low by te is placed on the data bus and LDA 1 is toggled briefly, after which t he high byte is loaded by pulsing LDA 0 . The 'bit counter' is reset when LDA 1 is toggled. Once the first address is loaded, the unit can be given the 'talk' command. Each I/O clock pulse from the VSP incre- ments the bit counter, causing the 'parallel-to-serial bit stream converter' to select the next bit in the selected speech memory byte. The same I/O pulse clocks each bit in turn into a flip-flop, which passes the bit stream to the speech processor. When the bit counter has scanned all eight bits, it increments the address buffer/counter to select the next memory byte. As illustrated in the block diagram, the connection between the bit stream converter and the following flip-flop can be interrupted, and both sides brought out to the 'host' processor. Data from the speech memory can be read into the host's RAM area via the Y output; after modification, to obtain a new word or sentence, it can be fed back in via the D input. Admittedly, this will often require a little interface — but we intend to publish a suitable circuit in the near future. The upper part of the block diagram shows the word processor proper (the 'VSP'). Two control inputs, C0 and Cl, come in at the left. These give the commands 'reset', 'talk' and 'test busy' as shown in table 1 . The t est busy command refers to the 'busy' output: when enabled, this goes high at the end of a speech sequence. The VSP chip contains a clock oscillator — among other things, this determines the pitch of the spoken output. To synchronise the external CCLK (control 1 clock) input to this on-chip clock, the two signals are fed through a flip-flop. The result goes back into the PDC (processor data clock) input. The VSP indicates that it needs the next speech data bit by toggling its I/O output; as described earlier, this clocks the next bit into the flip-flop and updates the bit counter. When entering speech data from external RAM, the (70 output must be used for correct synchronis- ation. Finally, the two differential speech outputs are passed through a low-pass filter and power amplifier to the loudspeaker. Timing Obviously, the various control signals must be applied to the board in the correct sequence. This is illustrated in figure 2. After power-up, the circuit Table 1. The three commands which are initiated via the control inputs C0 and Cl. must be initialised. This is done by applying a logic 1 level to C0 and Cl (corresponding to 'reset') and toggling the CCLK input three times; then, C0 and Cl are set to logic 0 (test busy) and CCLK is toggled a further three times. The unit is now 'ready to go'. To output a word, the low ad dress by te is put onto the data bus and LDA 1 is pulled low briefly; then the high address byte is loaded from the data bus by toggling LDA 0. Cl is now set to logic 1 (C0 remains low), corresponding to the 'talk' command, and the CCLK input is toggled. This initiates the speech out- put. Meanwhile, Cl is returned to logic 0 and the CCLK input is toggled twice. This enables the 'busy' output, so that 12-96 - elektor d T- T ROMCLK-6.25«» Table 2. The timing requirements for the it will go high at the end of the word. At that point, a further CCLK pulse will reeet the VSP in readiness for the next word. All control signals must meet the timing requirements shown in figure 3 and table 2. Figure 3a corresponds to the initialisation procedure; the main point here is that the CCLK pulses must be sufficiently long for guaran- teed synchronisation with the VSPs 'ROMCLK' oscillator. This means that both T(j own and T U p must be at least 6.25 /js, in most practical applications. Figure 3b shows the situation for 'talking'. The T w period, for loading the lower and upper address bytes, must be long enough for the address buffer/counter to latch: 20 ns or more. The shaded portions on the C0/C1 lines and data bus indicate that the logic levels are unimportant at that time. The circuit The general layout of the circuit dia- gram (figure 4) corresponds to that of the block diagram given in figure 1. Starting at the top, for a change: T1 . . . T3 convert the C0/C1 inputs into the actual control signals required by the processor, and N2 buffers the Busy output. PI sets the frequency of the on-chip oscillator: the correct set- ting corresponds to 160 kHz at pin 3 of IC1. No frequency counter is required, however: the output signal should sound like a normal male voice — not Donald Duck or 'infra-lwan-Rebroff'l Normally, the mid-position of PI should be fairly accurate. Note that this ad- justment does effect the minimum length of the CCLK pulses — the 6.25 ns mentioned above corresponds to 160 kHz! The CCLK input is synchronised to the ROMCLK output at pin 3 by means of FF1; via T4, this signal goes back to the PDC input of the VSP, IC1. The other flip-flop and T5 are used to clock the bit stream into the ADD8 of IC1, under the control of the I/O output. In between these two, the speech outputs (SPK1 and SPK2) are passed to the low- pass filter (A1 and A2) and the power amplifier (A3, T6 . . . T9). The output level is set by means of P2. The lower section of the circuit is the memory with its associated control circuits. IC4 . . . IC7 are the address buffer/counter. When the parallel load inputs (pi n 11) are pulled low, via LDA 0 or LDA 1, the byte on the data bus is transferred to the corresponding pair of ICs. The outputs from these ICs drive the address inputs of IC12. . . IC19 (the actual EPROMs) and the EPROM selector, IC9. 4 Table 3 K3 K4 K8 K9 0000 0100 0111 1101 1101 1110 1101 1101 1101 1011 1010 1001 1001 0010 0000 0000 0000 0 0 0 0 0 0 0 0 P K1 K2 ooooo iooii oiiio 00000 10010 10000 10100 10011 10011 10100 01101 01111 ioioo oiiio oioii 10011 10001 01010 11010 10010 01101 00111 10001 OHIO 01101 OHIO 00101 00101 OOOOO 10100 01011 OOOOO 10001 01011 ooooo OOOOO 10011 00111 OOOOO 10010 00101 K5 1001 0111 1000 0110 0111 1010 1010 1001 1000 1100 1101 0110 1001 1111 1000 1100 1111 1101 1001 1110 1011 1000 1011 0110 1010 0110 1011 0101 K6 K7 1000 1010 100 101 0111 1000 100 101 1000 0100 100 011 1011 0101 010 000 0111 0010 001 010 0101 0111 001 011 K10 FRAME TYPE SILENCE UV - REPEAT 010 V V - REPEAT V - REPEAT 101 V 101 V 110 V V -REPEAT 110 V V - REPEAT V- REPEAT V - REPEAT 011 V SILENCE SILENCE SILENCE UV UV UV - REPEAT UV UV SILENCE STOP CODE V = VOICED UV = UNVOICED E = ENERGY R = REPEAT P = PITCH K1 . . . K10 = FILTER PARAMETERS Table 3 This sequence of digital code words will make the Texas Instruments chip shout for help! Figure 5. The power supply fot the talking board can be derived from that of the host computer (figure Sal. Alternatively, a separate power supply can be constructed quite simply (figure 5b). The confusing array of wire links are included so that different types of EPROM can be used. For 2716s, links 2, 6, 7 and 9 should be used; the EPROMs are then addressed in the following sequence: IC12, IC13, IC16, IC17, IC14, IC15, IC18, IC19— corresponding to the address range from 0000 to 3FFF in 2 kbyte chunks. For 2732s, as in the Talking Board kit, links 1, 6, 8 and 10 should be mounted, as shown. The EPROMs are now selected in sequence, from IC12 to IC19, to cover the address range from 0000 to 7FFF. Finally, links 1, 4, 8 and 11 are provi- ded for 2764s; these cover the complete address range from 0000 to FFFF in the following sequence: IC12, IC14, 1C 1 6, IC18, IC13, IC15, 1C 1 7, IC19. It should be noted that the board layout and pinning is given for 2764s; the other types are slightly shorter, as indicated by dotted lines on the board. This means that pin 1 of a 2716 or 2732 is inserted in the pin 3 position, and so on down. Finally, the lower right-hand corner of talking board iber 1981 - 12-09 0000 0048 0084 0000 01CE 0222 02*C 02CC 0324 036E 0388 03CE 03E4 041 A 0484 0484 0400 04F2 0522 05A0 05FC 0634 EPROM 2 0000 004C 008A OOC2 OOEA 0114 01 4C 0178 01 AO 01 FO 021 E 0250 02BE ADDRESS WORD ADDRESS WORD (HEX) (HEX) AGAIN 06DA DOWN 0724 HELLO 0760 MESSAGE 079C MISTAKE 07B4 NAME 0800 NEED 082A PLEASE 0856 PUT 0890 REPEAT 08C4 RIGHT 0906 THANK 0946 UP 0970 WANT 099A •S 09D6 ALL 0A08 AN 0A44 DO DOES FOR FROM 0A9E 0AF6 0820 086C 0BB4 0C06 0C5A 0C94 0CC6 0CF8 0D2E HOW IS IT ME MUCH NO NOT NOW OF ON OR OUT THE THERE THIS USE WHAT WOULD YES YOU 0D6C 8 0DA8 C ODEC D 0E36 E 0E60 F 0E94 G 0EC4 H OEFE I 0F34 J OF 80 K ONE TWO THREE SEVEN EIGHT TEN ELEVEN TWELVE THIRTEEN 0774 0800 0864 08C2 08FC 0952 0986 0BC2 0C3E 0C94 0D54 0094 OOF4 0E62 OECE 0F18 OF 66 FOURTEEN SEVENTEEN EIGHTEEN NINETEEN TWENTY THIRTY FORTY SIXTY SEVENTY HUNDRED THOUSAND EQUAL NUMBER PERCENT DEGREES FARAD FREQUENCY HENRY HERTZ MEGA MICRO MINUS PLUS POINT POWER SECONDS TEMPERATURE TIME READY SWITCH CONTROL WARNING OFF CHECK BUTTON TELEPHONE BUSY INVALID MONDAY TUESDAY WEDNESDAY THURSDAY FRIDAY SATURDAY SUNDAY 0A98 OADE 0B5C 0BA6 ocoo 0C48 0C9C OCEA 0046 009A 0E10 0E7E OEBE 0F4C DATE LEFT CHANGE DIRECTION ENTER SLOW GO STOP HIGH LOW MOVE RANGE EXIT CARDS ATTACK DESTROY Table 4. The vocabulary of the talking board is contained in EPROM. Note that the first digit in the word address must correspond to the position of the EPROM. If these are mounted in sequence in the IC12 . . . IC14 position, the first address in EPROM 2 will be 1000; EPROM 3 then starts at 2000. figure 4. IC8 is the bit counter: the incoming 170 (clock) signal is divided by 8, to select the eight bits in each byte in sequence. Actually. IC8 is a 4-bit counter, but the fourth bit (QD) is fed back to the 'load' input so that 0000 is loaded as soon as it goes high. The three lower bits, Qa • • • Qc, control the data multiplexer (IC10) that selects the correct bit from the memory output byte. After each group of eight bits has been scanned, a pulse is fed from IC8, via N4, to the count input of IC7. This causes the address counter to increment to the following address. Power supply Very little needs to be said on this subject. The main board contains a sufficient number of smoothing capaci- tors, as shown in figure 5a, and an 1C that derives the -5 V supply from the incoming —12 V rail. The board therefore requires an ad- equately smoothed +12V/+5V/0V/ —12 V input. This can be provided by the 'host' microprocessor, or derived from an additional supply circuit as shown in figure 5b. The 5 V supply must be capable of delivering 300 mA. The quiescent current consumption of the ±12V supply is 50 mA, but this will increase at high audio output levels. How it talks Having dealt with the basic hardware, it is time to take a closer look at the software — in particular, how a given word is coded. Basically, the processor is an electronic analagon of the human speech tract. In plain language, it simu- lates the lungs ('energy'), the vocal cords ('pitch') and the shape of mouth and lips ('filters'); when the vocal cords are not resonating ('unvoiced' sounds, like S and F) a noise generator is used instead of a tone generator. All this information, for a given word, is con- tained in a succession of digital bits. A practical example will help to make this clear. Table 3 gives the complete code for the word 'help'. The first group of bits is 0000: silence. Then, 0100 sets the initial energy; the repeat bit is zero (we'll come to this later) and the 'pitch' is 0000 - corresponding to 'unvoiced'. For unvoiced sounds, the next 18 bits set four filter parameters as shown. The next line starts with a higher 'energy' setting (0111), followed by the repeat bit at logic high: un- modified filter settings. The pitch remains 0000, for unvoiced. Since the filter settings remain unchanged, we can proceed to the next line. A higher energy is defined, no repeat, and a non-zero pitch: 10010, defining the desired tone generator frequency. For voiced sounds, more precise filtering is required. This results in a total of 39 bits to determine the settings of all ten filters. Fortunately, the filter settings can remain unaltered for the next two lines (repeat bit one), although lOOOOftdoOpOj l oo o oooogg mmW 0 p p p pp pp gppp p W 0 OPOPOOUI W ppp p < beat monitor (E63/64, No. 101) . . . 2-44 LED audio level meter (E75/76, No. 23) 10-07 protecting dynamic RAMs (E75/76, No. 11) ... 9-48 scoreboard (E75/76, No. 1) 10-07 TV games extended (E77) 10-07 70cm transverter part 1 (E74) 10-07 s / f N cumulaihv index coming soon.. • perhaps New developments in consumer electronics In recent years a number of new developments in the L.S.I. integrated circuit field have revolutionised consumer elec- tronics. Described here are some exciting new ideas, even now nearing fruition in the engineering department of Silicon Hollow International Technology Inc. These new lead balloons for the 1981 Christmas season include the Cuckoo Clock chip, Smoke Detector with snooze feature and the digital L.C.D. Sundial with frontlight and melody alarm. The new nine-pin dual-in-line-and-a-bit package outline of the CUK 100A cuckoo clock chip (figure 1) allows for direct connection of the pendulum without the need for external conditioning circuitry. The audio output (pin 5) drives a piezo- electric transducer to produce an authentic 'cuckoo' sound encoded by Time Domain Linear Predictive Coded Formant Synthesis. The synthesis is based on an actual recording of a cuckoo made by the company's director of research. The high pass filter in the audio output circuit of the chip is necessary to remove a low frequency spurious signal in the cuckoo recording caused by the research director's inordinate love of baked beans. The device will be available in evaluation kit form, including chip, crystal, audio annunciator with beak and printed circuit board. There are three grades of kit available: CUK 100 A EV/KIT/S - With dead sparrow CUK 100 A EV/KIT/P- With dead pidgeon CUK 100 A EV/KIT/C - With dead cuckoo Silicon Hollow's projected smoke detector with snooze borrows technology from a number of areas. The smoke detector head uses the well known 'coughing canary' system and is supplied complete with sand tray and a one year supply of bird seed. The programming of the snooze timer is taken from the company's successful microwave oven controller, with settings for RARE. MEDIUM RARE. WELL DONE and TOTALLY INCINERATED. The digital L.C.D. sundial chip uses quadrature Hall Effect sensors to determine magnetic declination and hence latitude. By comparing local Time of day (entered via a calculator ! keyboard by the user) with G.M.T. (derived from a quartz | crystal controlled master clock on the chip), the chip is also able to determine longitude. Knowing these factors, the chip I adjusts the angle of the frontlight relative to the display, so that the shadow cast falls on an array of photoreceptors the output of which is digitised to provide the display. Thus the user receives the impression of using an authentic sundial combined with the convenience of digital L.C.D. display. Figure 1. CUK 100A pir part 3 The decoder section described in part two does have its limitations: no simul- taneous reproduction of Teletext infor- mation and the television programme: no time indication; no subtitles, etc. The video control board makes all this possible, but as indicated by figure 9, there are still a number of items to be considered. The output of this circuit can once again be connected to the video input (if present) of a television set. The amplitude of the video output signal can be adjusted to a maximum of twice the input level (by means of P8). teletext decoder . . . that does not require modifications to the TV set The final article concerning the Elektor Teletext decoder sets out to prove the subheading (at last!). By adding the video control board and the receiver section, both described in this article, a fully independent Teletext decoder can be constructed. At the same time, the circuit produces a UHF output which makes it possible to supply the Teletext signal directly via the aerial connecter. Consequently, it is no longer necessary to work inside the TV set itself, but it is necessary to construct a separate receiver section to produce the required video signal. How- ever, before we continue this discussion any further, we will first provide some more information about the video control board. The signals involved in t he sw itching are the After Hours Sync (AHS), Blanking (BL), Picture On (PO) and Display Enable (DE). These signals determine whether the television screen (PO) or Teletext picture (DE) becomes visible. A combination of these two 'pictures' is als o pos sible. The AHS and BL signals serve to avoid erroneous signal combinations by means of a number of wired 'OR' gates. These control four analogue switches that determine which signals are to be fed to elektor december 1981 - 12-29 1C 24. This 1C (LM 1889N) is a complete colour modulator which is able to produce a colour video signal from the luminance signal (Y, pin 13) and the R-Y and B-Y signals. The three signals needed for this purpose are not supplied by the TROM ( I C 1 0) in the Teletext decoder, this 1C only produces the basic signals R(ed), G(reen) and B(lue), therefore IC20 is used as a converter. Besides a matrix to generate the signals Y, R-Y and B-Y this 1C (LM 1886N) contains a number of inputs to provide colour modulation in accordance with the PAL system. The signals neces sary for these inputs are derived from AHS with the aid of IC21 . . . IC23. We are virtually dealing with a pocket size colour TV transmit- ter having only one drawback; it can only be used with digital signals. The colour possibilities of this design are only partly utilised by the Teletext decoder, since the LM 1886N has three inputs per colour and this results in a total of nine bits for the colour infor- mation. In addition to the eight basic Teletext colours (including black and white), this 1C is capable of generating a set of eight shades of each colour and a large number of additional combinations The Y input (from ICIO-TROM) and the Video I input are used solely for special functions. Video I is the normal TV programme signal and it is from this that the T eletext information is derived. This original programme signal is fed to IC24 via ES2 (see figure 7) and appears again at the Video II output and the VHF/UHF output. This occurs when the Teletext decoder is switched off from the keyboard. The Y signal can be mixed with the programme signal via ESI. This gives a clearly contrasting reproduction of the Teletext page which is superimposed on the TV picture. The Y signal consists of digital information, the amplitude of which can be adjusted very precisely with the aid of preset potentiometers P3 and P4. During this 'mixing' mode the colour carrier wave from IC24 must be switched off. This is accomplished by means of ES4. It is now time to discuss the TV sound signal, as this has not been mentioned so far. The LM 1889N 1C includes a separate oscillator which is capable of generating a sound carrier wave. For this purpose it must be mixed with the video signal via pin 12 of IC24. In this design the oscillator is not used, since the separation between picture and sound has already taken place inside the TV. A 6 MHz sound carrier wave is also avail- able when the receiver design of figure 1 1 is used, so a separate oscillator is also superfluous in that instance. A final remark about the video amplifier IC25. This amplifier is only needed when the amplitude of the Video I signal is greater than 3 V DD and has to be attenuated. Consequently IC25 could well prove to be superfluous and can then be replaced by a simple emitter R50.R56 = 4k7 R53 - 3k3 R54.R55 = 270 SI R57 = 2k2 R58 = 82 S7 R59 = 10 k P3 . . . P5,P8 = 4k7-preset P6 = 1 -k-preset P7 = 10-k -preset C46 = 150 p C50 = 4 . . . 40-p-trimmer C55 = 10 . . . 60-p-trimmer Semiconductors: D1 ...D7 = 1N4148 T2 = BC557B > oHho _> r 1981 - 12-35 R63 = 470 Si R64 - 47 n R65.R89 = 4k7 R66.R68 = 1k2 R67 = 22 17 R69 - 68 Si R70 - 47 k R71 - 1M2 R72 = 10 k R73 - 270 k R74 = 68 k R75.R76 ■ 1k5 R77 = 100 k R78 - 220 n R79 = 6k8 R80 = 3k3 R81 = 680 n R82.R85 = 100 n R83 = 1 k R84 = 2k7 R86 = 270 n R87 - 18 n R88 - 27 n P9.P10 = 100 km P11 = 100 k prese strength. The receiver section (printed circuit board 4) can be tuned to a TV station as soon as a correct setting is found. Most of the time the strongest transmitter in the neighbourhood can be received, but it will have poor picture quality since the demodulator is not yet calibrated. In the worst case we should already be satisfied with a ghost picture of the test pattern. Now that we have completed this step we can move on to the real calibration. Receiver board Since the tuners are already calibrated in the factory, the reader only has to adjust L9 and L10 to achieve maximum picture quality. Calibration must be started with L9 and can be improved by L10 (AFC coil). This adjustment needs to be repeated several times before the circuit is optimally calibrated. A word of warning: be very careful because the cores are fairly fragile. Use only a well fitting plastic trimming tool since the metal of an ordinary screwdriver will influence the circuit, thereby making it extremely difficult to find the optimal adjustment point. It is possible that the picture will remain dark in spite of correct adjustment of the demodulator, in which case a humming will be heard from the speaker. This problem can be solved by turning back P5 on the video control board, provided that the Video I signal is not Altering the AGC setting, with the aid of P11, is only meaningful when the tuners are provided with very strong aerial signals, therefore P1 1 is not that important and can remain untouched. After this simple calibration of the receiver, a usable input signal is available for the decoder. However, before we can proceed to calibrate the Teletext decoder the video control board has to be calibrated first. Video control board All the video signals are superimposed on a d.c. voltage level of 5 V, due to the fact that the Y signal of IC20 also alternates around this voltage. Of course, 5 V is an agreeable choice in TTL sur- 12-36 - elektor december 1981 roundings. Before the actual calibration we first have to measure some fixed voltage levels. The lowest level of the video signal is reached during reception of the sy nc p ulses. In order to set this level the AHS input of the video control board has to be grounded. The voltage level at pin 6 of IC20 should then be about 4.25 V. The measured value must be noted for later use. The highest level can also be measured at pin 6. The connect ions to the R, G and B inputs and the AHS signal must be disconnected for this. Now there should be a voltage of about 6.75 V present, therefore the V signal has a peak-to-peak amplitude of 2.5 V. The amplitude of the video signal from the receiver will be at least 2.6 V with a good aerial signal. This video signal has to be matched to the amplitude and the d.c. voltage level of the Y signal from IC20 by means of P5 and P6. The amplitude levels will be very similar and therefore it will probably be sufficient to adjust P5 in such a way (the Video I signal has to be disconnected) that the wiper of P6 has a d.c. voltage of about 4.25 V (the noted voltage level at pin 6 of IC20). When the signal from the receiver turns out to be too strong, in other words, when the picture still remains dark after P5 has been re-adjusted, a lower ampli- tude can be set with P6 and the cali- bration of P5 must then be repeated. It may be found that P6 can be turned to a certain point, beyond which the reception of a powerful transmitter will cause the test pattern to 'grey'. After this calibration is completed, and all the connections have been put in their correct place again, it should be possible to switch between Teletext and the normal programme by pushing the TXT-nor and TXT-off keys. The synchronisation of the TV should not be influenced during the switching. The mixing of the Teletext page and the program picture requires adjustment of the amplitude of the Y signal of IC10 with the aid of P3 and P4. The lower threshold is adjusted by P4 when the Teletext signal is switched off (TXT-off). The d.c. voltage at the emitter of T3 should then be 5 V. The upper threshold is adjusted by P3, according to the requirements. In order to do this, the 'mix' key must be depressed, so that the Teletext page becomes super- imposed on the test pattern. This Teletext page will probably consist of some incoherent words, letters or just the characters 'PI 00'. P3 is then adjusted in such a way that the 'white' of the characters can be clearly seen on the picture. After calibration of the decoder this adjustment should be repeated. This completes the calibration of the video control board, for the time being at least. If a VHF tuner is employed, the unit can be calibrated slightly more accurately (on channel 3) with the aid of capacitor C55. In practice, however. 14 Figure 14. The printed circuit board and component layout for the VHF-UHF converter and this adjustment will usually prove unnecessary due to the continuous tuning capability of modern channel selectors. The adjustment of capacitor C50 is only meaningful after the decoder has been calibrated, because clearly differentiated colour information is only available at that time. The frequency of the colour carrier wave is derived from a crystal and therefore the influence of C50 will be very slight. The TV receiver will only reproduce Teletext information in full colour inside a certain capacity range of C50. The trimmer must, therefore, be set in the centre position of this range. Decoder board Readers who intend to build the Teletext decoder ‘without all the trimmings' come in please! The next passage also refers to you. It is very important for the receiver to be tuned correctly, in other words, that it has a sharp, clear colour picture. After depressing the keys 'reset' and 'mix', the page header, or at least the letters 'PI 00', should appear on the screen. The core of L2 should then be turned until the time indication becomes visible in the top right-hand corner of the TV screen. This will be indicated very clearly within a limited adjustment range and the correct setting for L2 is in the centre of this range. It is best to press the 'reset' key repeatedly during this last cali- bration in order to wipe out the non- sense that the decoder displays on the screen. The decoder should now be ready for use. Note: The Home Office has recently granted permission for the broadcasting companies to transmit teletext infor- mation on four TV lines instead of two. This means that the access time has now been halved to approximately 12.5 seconds per page. Directions for use Since the directions for use were already described in part 2 (November 1981), we will now discuss the meaning of the key legends very briefly. TXT-off. When this key is pressed the Teletext display disappears. TXT-nor. This key calls the selected Teletext page onto the screen. Mix. The Teletext page is superimposed on the normal program picture. Numeric keys. Page selection is per- formed with these keys. RESET The 'reset' key returns the page number to 100 and simultaneously erases the displayed page. Timed page. This key allows presetting the moment of display of a previously selected page. Full page/half page. The 'half page' key selects either the upper or lower half of the page and doubles the character height The 'full page' key resets the page to its normal proportions. Reveal. Hidden information, for instance for video games, can be made visible by pressing the 'reveal' key. HOLD. This key freezes the present page. Time/B7. Time is displayed for 5 seconds when the keys ’B7' and ’time' are pressed simultaneously. For detailed information refer to Teletext part 2. iber 1981 - 12-37 The block diagram in figure 1 shows that the device consists of two basic sections, namely a window discrimi- nator and a so-called 'voltage generator'. Initially, output '0' of the voltage gener- ator has a high logic level, in contrast to the other nine outputs. By adjusting preset potentiometer P2, any voltage between 0 V ... 1 2 V can be fed to the first input of the window discriminator via diode D2. The window discriminator then checks to see whether this voltage is the same as that present at the wiper of potentiometer PI (the code 'switch'). If so, the window discriminator transmits a clock pulse to the voltage generator The window discriminator is con- structed around two opamps (IC2 and IC3) having a high open-loop gain. Therefore, the outputs of these two ICs can only be logic one or logic zero. The voltage at the inverting input of IC3 is approximately 0.6 V less than that at the non-inverting input of IC2. The so-called window voltage depends on the setting of the preset potentiometers P2...P10. The voltage presented to the other input of the window dis- criminator depends on the setting of potentiometer PI. Both opamp outputs will be high if the inverting input of IC2 and the non-inverting input of IC3 are combination lock A combination lock can be constructed by connecting a number of switches in series with a solenoid or other form of electric door unlocking device. However, this procedure has two main disadvantages: quite a few (costly) switches are required to obtain a reasonably safe lock, if any person forgets to reset the switches after the lock has been opened the complete code remains visible. To be on the safe side, this design has only one rotary 'switch'. Therefore, only part of the code remains visible. What is more, the number of possible combinations is virtually infinite. when the 'enter' switch, SI , is depressed. This means that the voltage generator will then supply a second (and success- ive) voltage code(s). By repeating this procedure a total of nine times, output 9 will eventually go high and the relay will be activated: the lock opens. If the voltage supplied by potentiometer PI is 'outside the window', the window discriminator will send' a reset pulse to the voltage generator when SI is de- pressed. The entire circuit is then reset and the procedure will have to be repeated from the beginning. The circuit The complete circuit diagram for the combination lock is given in figure 2. The heart of the voltage generator is the well known counter 1C, the 4017. Each output of this 1C is connected to a preset potentiometer and it is these presets which are used to set up the desired code. The combination voltages are applied to one input of the window discriminator one after the other via diodes D2. . . DIO. 'inside' this window. If not, the output of one of the opamps will be high and the output of the other will be low. The two output signals from the opamps are NANDed together by N1. This means that when the voltage set by PI is inside the window voltage, the output of N1 will be low and the output of the inverter, N3, will be high. This enables gate N4 so that when the 'enter' switch, SI, is depressed, a clock pulse is trans- mitted to the voltage generator, 1C 1 . This causes the next output of IC1 to go high. Correct adjustment of PI for the voltage supplied by this output, and again depressing SI, will generate another clock pulse. However, if the voltage code is set incorrectly, the outputs of N1 and N3 will be high and low, or vice versa. In this instance, a pulse will be applied to the reset input of IC1 via gate N2 and the differentiation network C2/R5 when switch SI is depressed. This means that the whole procedure has to be started from scratch. If the complete combination has been 1 entered correctly, output 9 of IC1 will go high causing transistor T1 to conduct and energise the relay < Re 1 ). Conse- quently, the lock is now open. If SI is depressed once more, IC1 will be reset and the lock will be closed. Components Cl and R4, together with the Schmitt trigger inputs of gates N2 and N4, are responsible for contact bounce suppression. Resistor R2 is included so that the wiper voltage of PI does not exceed the common mode input voltage of the opamps. Construction It is recommended to choose a large type of potentiometer for PI. A scale should be made up with a graduation from 0 ... 9. The digits 0 and 9 should be placed at an angle of 30° from the start and end positions of the scale, respectively. This is necessary so that PI can be adjusted to give an output voltage which is less than the wiper voltage of the preset in question. The remaining digits are then distributed evenly in the space between these limits. A ten digit scale leads to 10 s = 1 billion different combinations! In principle, this figure can be further increased by selecting a finer scale division. However, I there is no real point in having more I than 15 numbers, as the window Tab la: comparator will no longer be able to distinguish between consecutive digits. If a door lock having a drive voltage of 12 V is utilised, relay Rel may be omitted and the door lock may be directly controlled by transistor T1. The current consumption of the door lock should not exceed 400 mA in this instance. The lock may be connected to a different drive voltage level. The connection between point A and positive supply line is then broken (see figure 2). Point A can then be connected to an unstabilised d.c. supply up to a maxi- mum of 30 V/400 mA. If the door lock is to be driven by a voltage greater than 30 V d.c., or with an a.c. voltage relay Rel will have to be incorporated. The circuit itself must be powered from a stabilised supply. The current con- sumption of the circuit largely depends on the pull-in current of the relay or door lock used. It is not advisable to power the circuit by batteries, for if and when they run down, the lock will have to be forced open! If required, an on/off switch can be connected in series with the positive supply line. The voltage generator, 1C 1 , will automatically be reset when the power is switched on. Calibration First, select a suitable nine digit code (for example, your date of birth and one other figure). Reset the circuit by depressing SI (output '0' of IC1 will now be high). Set the pointer of poten- tiometer PI to the first digit of the desired code and connect a multimeter to the test point TP (d.c. measurement range > 12V). Adjust P2 until the output of N1 (TP) becomes low. This will be true for a specific range of adjustment. Set P2 in the centre pos- ition of this range. Depress SI (output 'T of IC1 goes high) and adjust the setting of P3 for the second digit of the secret code, and so on, until preset P 10 has been adjusted. If the code has to be divulged for any reason, it can be modified quickly by altering the setting of presets P2 ...P10. H Elektor synthesis the NEW >r december 1981 - (he VKW Eleklor synthesiser A first glance at the circuit diagram in figure 1 will raise some doubts as to how 'simple' this simplified synthesiser really is! The VCO 1C (CEM 3340), already de- scribed in the October issue of Elektor, forms the heart of the circuit. Together with six opamps it performs as well as the complete VCO module of the Formant synthesiser. The remaining space is used for the control logic which is necessary for the 'preset' and 'poly- phonic' modes of operation. Therefore, it is certainly a simplification in the long run, since all the required components can be mounted on one printed circuit board, thereby saving both time and expense. The VCO module This is the third article in the series relating to the Curtis ICs and the new Elektor synthesiser. Whereas the previous articles were mainly concerned with the 'theory' of the project, we now continue with the practical side of the story. The voltage controlled oscillator (VCO) is essentially the heart of any synthesiser and the stability of the completed instrument depends to a large extent on the design and performance of this module. For this reason it is important that all the information contained in this article is followed very carefully, especially during the calibration procedure. The circuit The first item to consider is the power supply. In contrast to the Formant synthesiser, the VCO described here only requires a symmetrical + and —15 V power supply. The current consumption of the basic version of the instrument (without polophony) is less than 200 mA per supply line. The positive supply voltage is fed to pins 11 and 12 of the 723 adjustable voltage regulator, IC2. The (11.05 V) output voltage at pin 10 of this 1C is fed to pin 16 of 1C 1 . Besides this positive supply, the CEM 3340 requires two further voltages which are gener- ated by opamps A1 and IC5. These provide output voltages of +5 V and I 5hJZT 1 m ■e NEW Elektor synthesise O0o EEJn —5 V respectively. The output voltage of A1 is also used to adjust the fre- quency range and pitch of the VCO. The output voltage of IC5 provides the negative supply requirement for IC1 and is fed to pins 1 ... 3 of this 1C. The audio signals (squarewave, sawtooth and triangle) are fed from pins 4, 8 and 10 of IC2 via the buffer stages A2. A3 and A4 and a select switch (S2) before reaching the outside world. Control voltages Pin 15 of IC1 is the input for the various control voltages which deter- mine the actual VCO frequency. A bias voltage is applied via a potential divider network (see figure 2). The values of resistors used determine the volt/octave characteristics of the corresponding control voltage source. The control logic for the 'preset' and 'polyphonic' modes Although the three 4066 CMOS switches and the 4001 are not required tor the 6 construction of an 'ordinary' synthesiser, the relevant copper tracks are already on the printed circuit board. Therefore, a future extension will not require the addition of another printed circuit board. Thus, the associated resistors and integrated circuits can be omitted for the time being. This means that the wire links B1.B2 and B3 should be mounted in the 1C socket instead of IC4 and IC7...IC9. Links B1 and B2 supply the VCO with the control voltage from the keyboard, the range switch (SI) and the tune potentiometer PI 0. Link B3 provides a connection between the wiper of S2 and the output socket (see also figure 3). The wire links must be placed into the following positions: - link B1; pins 8 and 9 of IC7 - link B2, pins 1 and 2 of IC7 - link B3; pins 10 and 1 1 of IC8. A precise description of the function of the CMOS switches and the inverters Figure 6. a : will be dealt with in a future article uggened method of mechanical construction of each module for the synthesizer. 12-44 - elektor decamber 1981 NEW Elektor synthesiser ised modules without any problems, however, this may not hold for every card housing (see figure 7). Operation The power supply voltage should be connected and checked at the various IC-pins before the ICs are mounted. This avoids the possibility of damage to the expensive ICs if there does happen to be a wiring error or component fault somewhere. The voltages at the various pins of the 1C sockets should then be tested and should correspond to the values given in table 1. If this is the case, you can be sure that the circuit has been con- structed correctly. After disconnecting the supply voltage, IC2 (723) can be placed in its socket. The power supply is then re-connected and the voltage at pin 10 is adjusted to exactly 11.05 V by means of preset potentiometer P2. The voltage at the output of A1 is then adjusted by means of P3 and should be set to exactly 5 V. As opamp IC5 is connected as an in- verter, the output of this device will automatically be — 5 V. Subsequently, you should check that the voltages of +1 1.05 V, +5 V and -5 V are present at the corresponding pins of the socket for IC1 (see table 2). The voltage level at the output of IC6 should change by one volt for each position of the range switch SI. This voltage change can be measured with a digital volt meter (DVM). The voltage at pin 5 of IC1 should be adjustable between 0 ... 4 V with the aid of potentiometer P11 (PWM). If all the supply voltages for IC1 are correct, this 1C can be inserted into its socket. If you are the owner of a variable power supply, it is advisable to increase the supply voltage slowly. The current consumption can then be monitored to ensure that there is no short circuit. After having taken all the necessary precautions the calibration can be carried out. Calibration The curve of the control voltage/ frequency characteristic of the VCO is relatively linear. Consequently, the adjustment to the correct voltage level per octave is limited. A DVM is required to check that the voltages at the output of IC6 are exactly 0, 1,2, 3 V etc. For an acoustic check of the circuit, the output of the VCO (connection point 8 on the printed circuit board) should be connected to the input of an audio amplifier. Where- upon the setting of preset poten- tiometer P9 can be altered very slowly until the VCO frequency changes by an octave for each successive range switch position. Readers who do not possess a frequency counter can use an audio oscillator or a tuning fork. A word of warning: do not depend entirely on your sense of hearing, as it is not precise. (Even Elektor readers are only human!). The tune potentiometer (P10) can be used to adjust the VCO frequency to give 'zero beat' when an 'auxiliary' sound source (such as a quartz tuning fork) is employed. A clear discord can be heard if the VCO frequency does not alter by exactly one octave. After a little practice, this adjustment procedure becomes very simple. It is wise to bear in mind that if P9 is altered the frequency of the VCO changes. The latter must then be re-adjusted each time (using the tune potentiometer. A linearity correction in the upper fre- quency ranges of the VCO can be performed with the aid of preset poten- tiometer P7. The effect of this preset is very slight; with experimental set-ups the effect was nominal when the wiper of P7 was turned towards ground. An aural adjustment is very difficult to perform when the keyboard is discon- nected, due to the very low VCO frequency. For this purpose PI should be adjusted so that the lowest octaves can be heard. Connection of the keyboard The control voltage output from the keyboard is to be connected to contact 1 0 (potentiometer P5) of the printed circuit board. This potentiometer is adjusted so that the VCO frequency alters by one octave when two keys having a difference of one octave are pressed one after the other. To be absolutely sure, this procedure should be repeated several times with other keys and different settings of PI and SI. The final adjustment of PI is ac- complished as follows: Select the highest octave with the aid of the range switch. T urn the tune poten- tiometer, which has an adjustment range slightly greater than one octave, to the mid position. Turn off the 'coarse octave' switch on the Formant key- board and depress the highest key. Using the tuning fork mentioned previously, the VCO frequency is adjusted by means of PI until the key producing tone A corresponds to the frequency of the tuning fork. The overall octave position is a matter of taste; PI can be adjusted so that the highest note on the keyboard is placed just within the threshold of audibility. Whether this is meaningful or not is another question. Thecoarse octave switch on the Formant keyboard enables the VCO frequency to be shifted into other ranges. Setting the signal amplitudes Once construction of the circuit is complete, the output waveform from the VCO can be selected by the three position switch, S2. The triangular signal will sound lower in volume than a sawtooth waveform of the same ampli- Resistors: R1.R11 = 2k2 R2 . . . R8 = 4k7 (metal film) R9.R10.R14 . . . R18. R33 . . . R37 = 100 k R1 2.R22.R39 = 4k7 R13 = 470 k R19.R26 - 470 R20.R21 = 560 k R23 - 22 k R24 = 5k6 R25- 1k8 R27 - 1M5 R28.R29.R31 - 10 k R30 * 15 k R38 = 100 k (metal film) PI ■ 100 k multiturn preset P2 = 1 k preset P3 - 10 k multiturn preset P4.P7.P8 = 10 k preset P5.P6 - 200 k multiturn preset P9 - 20 k multiturn preset P10.P11 - 10 k lin potentiometer Capacitors: Cl ,C2 = 330 n C3= 10/25 V C4 - 470 p C5.C7.C8 = 10 n C6.C10 = 0 1 C9 = 1 n polystyrene C11 =1 n Semiconductors: IC1 = CEM 3340 IC2 = 723 IC3 = LM 324 (TL 084) IC4 “ 4001 IC5.IC6 = LM 741 IC7 . . . IC9 = 4066 51 - 6 pole rotary switch 52 - dual ganged 3 pole rotary switch tude; due to the smaller number of harmonics. When adjusting the preset potentiometer P8 and P4 the following items should be borne in mind: P8 has to be adjusted so that the amplitude of the triangular signal reaches a maximum without becoming trapezoidal. Sub- sequently, P4 should be adjusted so that the audible volume of the sawtooth signal correspond to the volume of the triangular signal. The duty cycle of the squarewave signal can be adjusted between 0 and 100% by means of potentiometer P1 1. Both edges of the triangle waveform and the leading edge of the sawtooth waveform are ex- tremely linear. The trailing edges of the squarewave and sawtooth waveforms are very steep and can therefore hardly be distinguished on an oscilloscope. If desired, preset P8 can be mounted on the front panel (as a potentiometer) so that the triangular signal can be made trapezoidal for various 'effects'. K flashing lights 1981 - 12-45 Now that the evenings are long and dark and the price of a pint makes you think twice before venturing out of the house, why not save a little money and con- struct your own Christmas presents? This is an ideal opportunity to indulge in electronics as a hobby for the benefit of the children. Toy cars are always appreciated and, provided they are not too small, can usually accommodate a small circuit board and a couple of batteries. This particular circuit adds a special touch to the 'common or garden’ toy car. As mentioned earlier, the flashing lights are very similar to those found on police cars etc. What is more, the effect is so well simulated that there is no need to include any moving parts. flashing lights With Christmas just around the corner, why not 'brighten' the festivities by constructing the flashing lights described here? These lights can be fitted to an inexpensive (plastic?) toy car to provide an effect very similar to the warning lights seen on ambulances, fire engines and police vehicles. When used with the Hi-Fi siren it will, at minimal cost, add new dimensions to a toy, which any child will find fascinating. tor value and the total resistance of R1 and PI. The frequency can be altered by adjusting PI. The RC network C2/R3 connected to the output of N1 acts as a differentiator. Since R3 is connected to the positive supply rail, the network is only sensitive to the negative-going edges of the squarewave signal. These short 'spikes' are then converted to usable pulses by gate N2 to drive the darlington transis- tor T1. In turn, this transistor switches the lamp connected to its collector on for a short period of time. A resistor (R5) has been included across the emitter and collector of the transistor to ensure that the lamp remains at the correct temperature. This has the ad- vantage that the initial current through the lamp is much less than normal and therefore the lamp will have a much longer life span. To make the lamp light up brightly, a 6 V type can be used with a (recommended) supply voltage of 9 V. The only difference between the first and second sections of the circuit is the fact that the second one can be 'pro- grammed' to perform in one of three different ways. This is accomplished with the aid of a wire link on the board. By linking points 3 and M, two com- pletely independant flashing lights are obtained. By linking points 2 and M, the lamps light alternately. The frequency can then be adjusted by means of PI. Finally, if points 1 and M are linked, the two lamps will light simultaneously. Again, the frequency is determined by means of PI . Straightforward and to the point As the circuit diagram in figure 1 shows, it is still possible to design all sorts of amusing and 'fun' circuits with a minimum of components. The entire unit consists of two identical low frequency oscillator circuits each con- trolling a small lamp. The principle of operation can be described quite briefly. As both circuits are identical, only one need be described. The oscillator (astable multivibrator) is constructed around the Schmitt trigger N1. Capacitor Cl is connected between the inputs of the gate and ground. The output of N1 is fed back to the input via resistor R1 and potentiometer PI. The capacitor is either charged or discharged by way of these resistors, depending on the logic level at the output of N1. Whenever the voltage across the capacitor reaches one of the trigger levels, the output of the gate 'toggles'. Thus, the multivibrator produces a squarewave output signal, the frequency of which is determined by the relationship between the capaci- R1 ,R6 = 47 k R2,R7 = 10k R3.R8 - 470 k R4.R9 - 22 k R5.R10-470S7 (see text) rii - ioon P1.P2- 1 M preset All resistors % W Capacitors: C1.C3 ■ 820 n C2.C4 = 100 n C5- 10 m/16 V Semiconductors: T1.T2-BC517 IC1 = 4093 Miscellaneous: Lai ,La2 = 6 V . 50 mA bulb The printed circuit board The two oscillator circuits can both be mounted on the printed circuit board shown in figure 2. The frequency controls, PI and P2, can be either normal potentiometers or preset types. Do not forget to make the link between point M and one of the points 1 ... 3. The supply voltage for the circuit can be anything between 3 ... 1 5 V, but, as mentioned before, a 9 V battery supply (PP3) would be ideal. For optimum performance, the voltage rating of the lamps should be about 2/3 of the supply voltage, while the current rating should not exceed 400 mA. The values of resistors R5 and RIO should be chosen empirically so that the lamps are just on the verge of lighting. We do not intend to give any details about installing the finished article into the model car. This is very much depen- dent on the particular model chosen. Usually, all that is required is a couple of holes for the lamps and some simple method of mounting the bits and pieces. OK Machine & Tool (UK! L Dutton Lane. EASTLEIGH. Hants S05 AAA. Telephone: 0703 610944. Available with electrical travel of 80 mm II Wl and 120 mm (2 Wl at 10K nominal resistance, both models incorporate specially designed stand-off terminals to allow ease of PCB mounting. Cetronic Limited, Hoddesdon Road, Slanstaad Abbots, Herts SG128EJ. Telephone: Ware 0920.871077 (2107 Ml Dual standard fuel computer Quality products from Bopla includes a comprehensive range of compuier terminal housings 111. These enclosures are injection moulded in foam plastic with interchangeable front mouldings. There are many variations and several custom options on offer. To complete the terminal housing, a range of option, and some can be supplied with handles. West Hyde Developments Ltd., Unit 9, Park Street Industrial Estate, Aylesbury, Buckinghamshire, HP20 1 ET, Telephone: Aylesbury 10296 J 20441 . (2130 Ml Rabans Lane. Bucks HP19 3RG. Telephone: 1 0296 ) 32881. LE40 soldering iron Telephone: 10296) 32881 Litesold's LE40 Soldering Iron, with in- handle electronic control, now provides the facility for users to adjust the temperature without dismantling. An access hole is provided in the handle to permit adjustment of the setting poten Rechargeable battery system tiometer to vary the bit temperature step- lessly from approximately 300°C to 400° C. The Gould 'Again & Again Irons are normally set at the factory to battery system, offers a low- Gould Portable Battery Division. iety Raynham Road, s to Bishop's Stortford, Hertfordshire CM23 SPF, England Telephone: (0279) 55155. Low cost chart recorder The CR450 series ol J.J. Lloyd Instrume signed to provide simple to operate ir counting during output. Counts can be stored in the battery powered memory for up to supply. The H7A-4 also provides a 12VDC supply at 40 mA to drive external sensing devices to provide a complete counting sys- tem. Operating from 110/240 VAC, the H7A-4 gives a voltage output and a simul- tanous single pole contact output rated at 2A @ 240 VAC variable via a potentiometer between 0. 1 sec and 1.0 sec. The integral relay has bifurcated contacts to ensure reliable switching of both high and low This compact panel mounted digital counter has a 72 DIN facia with a 4 digit LED dislay that includes count in and count output indicators. Connection is via screw terminals and the H7A-4 can accept contact or voltage inputs of up to 300 count per second. The H7A-4 is available as a Totalizer (H7A-4TM) or Batch Counter (H7A-DM). IMO Precision Controls Ltd., 349 Edgware Road. London W2 IBS, Tel: 01 723 2231/4 end 01 402 7333/6, Telex: 28514 Cables: Omrontrols Ldn. (2175 M) 'Super Beep' Super Beep’ is Barkway Electronics' new, low-cost, V.H.F. ’mini’ radio paging system for a whole host of applications, both per- manent and temporary, including offices. The receivers, which are feather-light and easily carried in the pocket, on a lapel, or in a handbag, use two mercury batteries lasting 1,000 hours or six months with eight hour Barkway Electronics Limited, Barkway. Royston, Hertfordshire SG8 8EE. England, Telephone: Bark way I STD 0763 841 666. Telex: 817651/BARCOM G, Cables: BARCOM Royston Hertfordshire. (2193 Ml Citizens Band transceiver R.F. Technology Ltd. have developed and are commencing production of a Citizens Band (ex-Open Channel) transceiver. This 20 channel set operates on the U.H.F. band at 934 megahertz. Although more expensive than the corre- sponding 27 megahertz sets, it has the advan- Security cases Imhof-Bedco Standard Products Ltd. has launched a new range of "camera craft" security cases which feature an aluminium frame and facing on rigid wooden panels, with light weight. Lockable toggle catches, robust hinges and riveted corner re- inforcements add to the protection offered. knife provided to accommodate equipment of various shapes and sizes. A briefcase ver- wallet in the lid is also offered. Imhof-Bedco Standard Products Ltd.. Ashley Road, Uxbridge, Middlesex UB82SQ, Telephone: Uxbridge 10895) 37123. (2186 M) UK 14 - elektor december 1981 advertisement ELEKTOR BOOK SERVICE JUNIOR COMPUTER BOOK 1 - for anyone wishing to become familiar with (micro)computers, this book gives the opportunity to build and program a personal computer at a very reasonable cost. Price — UK £4.25 Overseas £4.50 JUNIOR COMPUTER BOOK 2 - follows in a logical continuation of Bookl, and contains a detailed app- raisal of the software. Three major programming tools, the monitor, an assembler and an editor, are dis- cussed together with practical proposals for input output and peripherals. Price - UK £4.75 Overseas £5.00 300 CIRCUITS for the home constructor - 300 projects ranging from the basic to the very sophisticated. Price - UK £3.75 Overseas £4.00 DIGIBOOK - provides a simple step-by-step introduction to the basic theory and application of digital electronics and gives clear explanations of the fundamentals of digital circuitry, backed up by experiments designed to reinforce this newly acquired knowledge. Supplied with an experimenter's PCB. Price — UK £5.00 Overseas £5.25 FORMANT - complete constructional details of the Elektor Formant Synthesiser - comes with a FREE cassette of sounds that the Formant is capable of producing together with advice on how to achieve them. Price - UK £4.75 Overseas £5.00 SC/MPUTER (1) - describes how to build and operate your own microprocessor system - the first book of a series - further books will show how the system may be extended to meet various requirements. Price - UK £3.95 Overseas £4.20 SC/MPUTER (2) - the second book in the series. An updated version of the monitor program (Elbug II) is introduced together with a number of expansion possibilities. By adding the Elekterminal to the system described in Book 1 the microcomputer becomes even more versatile. Price - UK £4.25 Overseas £4.50 BOOK 75 - a selection of some of the most interesting and popular construction projects that were originally published in Elektor issues 1 to 8. Price - UK £3.75 Overseas £4.00 When ordering please use the Elektor Readers' Order Card in this issue (the above prices include p. & p.) (j^restWciy) '“ELECTRONICS WOODHILL LANE, SHAMLEY GREEN, near GUILDFORD, SURREY Send cheque or postal order to Crestway Electronics Ltd., write or ring with your Access/Barclaycard acc- ount no. (do not send your card). Please add 40p to all U.K. orders for postage and packing. All our prices include VAT. 24 hour answering service on 0483 893236. Crestway Electronics Ltd., Woodhill Lane, Shamley Green, » a |u«am>u| near Guildford, Surrey. r \ ^ | vk* | TV Games extension (81143) Disco light controller (81155) DFM + DVM (81156) RF test generator (81150) EPROM programmer (81594) Frequency counter (82026) Teletext decoder (82025) The ELEKTOR MINI ORGAN (82020) write or phone for details of these kits. THE ELEKTOR METAL DETECTOR Kit no. 82021. An advanced metal detector available for the first time to the home constructor. Pro- fessional quality ABS vacuum formed case, complete with 10" ready set up and foamed search head. This kit can be constructed easily in a few hours. High performance specification includes phase locked loop oscillator and discriminator circuitry giving four reject modes woth ground effect elimination. For full details and specification see the article in this issue of Elektor. complete kit now available £89.95 p&p £2.00 CRESTLEK KITS Ioniser (9823) Negative ion generator £10.50 Talk Funny (80052) Ring Modulator £10.00 Sound Effects Unit (81 112) Guns, trains, etc. £ 8.30 Elektornado (9874) 100 W power Amp . . . £19.50 Top-preamp (80023) Hi-fi preamp £34.40 Guitar Preamp (77020) £ 6.50 2% Digit DVM (81105) £23.35 AM Receiver (81111) Easy to build £1 0.25 Touch Switch (81008) 12 positions £10.00 200 W Disco Power Amp (81082) £20.85 STAMP (80543) Mini amplifier £ 3.75 Pest Pester (80130) Insect repellent £ 2.35 Aerial Booster (80022) £ 5.65 LATEST KITS.... ATARI The World -beating ATARI PERSONAL COMPUTERS 3 consoles available Atari 400 with 16K RAM (AF36P) £345 Atari 400 with 32K RAM (AF37S) £395 Atari 800 with 16K RAM (AF02C) £645 (expandable to 48KI All consoles when connected to a standard UK colour (or black and white) TV set can generate the most amazing graphics you've ever seen. Look at what vou get: 1 ■ MORE HARDWARE * Background colour, plotting colour, text ■ Atari 410 Cassette Recorder (AF28F) £50 colour and border colour settable to any ■ Atan 810 Disk Drive IAFO 6 G 1 £345 one Of 16 colours with 8 levels of ■ Atan 822 40 column Thermal Printer (AF04B £265 Atari 850 Interlace (AF29GI £135 Joystick ControHers (AC37SJ Paddle Controllers (AC29GI 16K RAM Memory Module IAF08JI £65 I MUCH MORE FOR ATARI COMING SOON be played individually or together and each has 1785 possible sounds playable at any one of eight volume settings, for game output to TV. * BASIC cartridge and 10K ROM operating system and full documentation. de NOW ♦ Word Processor * VIS) CALC ♦ AOVENTURE GAMES ♦Arcade Games ♦ Trek Games ♦ ASSEMBLER & DISASSEMBLER ♦FORTH ♦Teaching ♦ 30 GRAPHICS ♦ Character Set | LE STICK For Atari Computer or Video Game Replaces standard yoystick. but much SPECIAL PACKAGE OFFER m (or £725 with LeStick Everything in "Look at what you get" list Can any other computer on the market ofter all this at anything like this price? __ se hand movements. Large pushbutton on top ol Stick. Squeeze Stick to freeze motion A MUST for SPACE INyAOERS. STAR RAIDERS 6 ASTEROIDS. ONLY £24.95 IAC45YI VERSAWRITER 1214 » Bin. drawing board. Drawing on board is reproduced on TV via Atari with 32K RAM and Disk Drive Closed areas may be filled in with one ol 3 colours. Text may be added in any one of 4 fonts Paint brush mode: select size of brush and paint away Air brush mode: shade in your drawing - colour and density is up to you. Plus many more features. S.a e. for price and further details. Maplin Electronic Supplies Ltd P.O. Box 3, Rayleigh, Essex. Tel: Southend (0702) 55291 1 /554155