THE FUTURE BELONGS TO THE PHOTON Volume-5. Number-2 February 1987 Electronics Technology The future belongs to the photon 2.28 Digital signal processing 2.40 Software for the BBC computer 2.42 Flexicell to beat battery weight 2.66 Projects Indoor unit for satellite TV reception 2.21 Universal control for stepper motors 2.31 Precision power supply 2.44 Computerscope 2 2.51 Information : i Editorial 2.05 News • News • News • 2.17 Meet • • • 2.65 New products 2.70 Licences & letter of intent 2.80 I Corrections 2.82 Guide-lines Switchboard 2.77 Classified ads 2.82 Index of advertisers 2.82 Selex-20 Linear scale ohmmeter 2.58 The Cackling Generator 2.61 Power 2-63 Electronic Switch 2.64 Yabasu PCB Drafting Aids are made fo meet international standards Keeping in mind that the excellence of PCB artwork is determined by the quality and precision ot PCB drafting aids used. Precious offers a wide range of PCB drafting aids which covers: Red, Blue & Black tapes. Donuts 1C Patterns, Connector Patterns, transfer lettering etc. Distributors Dimech Enterprises C/4, Ajanta Flats, Navrongpura, Ahmedabad-380 013 52-C, Proctor Road, Bombay 400 007, Tel: 367459/369478. Telex: (01 1) 76661 ELEKIN J. M. Enterprises Integrated Electronics 382, Lajpatrai Market, Delhi- 1 1 0 006 Instruments „„„„„ e _ w ,„ 8-2-1 74, Red Cross Road, Kaysons Radios Secunderabad-500 003 363, La patrai Market, Delhi- 1 1 0 006 Circuit Aids No. 451 , 64th Cross, V Block, Rajaji Nagar. Bangalore-560 01 0 First time in India YABASU TAPE DISPENSER • Tape end is instantly located • Tape is protected against dust and moisture from hands • Tape aligning 8i positioning is simpler Sole Distributors precious ® Electronics Corporation 2-04 * Sensitivity Superior component tester Serviciability Simplicity. VASAVI OSCILLOSCOPES SYMBOLISE FIVE STARS for details write to: VASAVI ELECTRONICS 630, Alkarim Trade Centre Ranigunj (opposite to Bus Stand) SEGJNDERABAD-500 003. ph: 70995 gms: VELSCOPE Telex: 0155-6834 INDIA’S FIRST Now, Capture Store Display and Analyse signals with HM 205 ANALOG AND DIGITAL STORAGE OSCILLQ3COPE- D9GITRL STORAGE OSCILLOSCOPE FROM scientiFic Scientific now extends the best features of its well- known and time-proven HM series to the Digital technic with the introduction of HM 205. bridging the gap with Analog and Digital Storage operation Features includes:2 Channels-20 MHz 2mV, Separate Memories for each channel, Single and Refresh modes, Battery back-up for memory, Highest TB Speed 20 nS, TV Sync. Separator, Addition and Subtraction of channels, Component-Tester, weighing only 8.5 Kg. OTHER MODELS. "60 MHz. -DT ImV. Max. T B. Speed 5nS "20 MHz -0 T 1 mV. sweep delay "20 MHz. -D T 2mV. Max T B Speed 20nS Manufactured by: *20 MHz, -D.T 5m V. Max TB. Speed 40nS ■ 1 0 MHz, -S.T. Comp. Tester SCIENTIFIC MES-TECHNIK PVT LTD., B-14, Industrial Estate. Pologround, INDORE-452 003 Phone: 31777-78 Cable: SCOPE Telex 0735-267 Designed to lead. Built to last. BRAINSTORM FROM STOVEC . .. ' Stovec Screens India Ltd. in collaboration with Prestwick Circuits of Scotland, bring you state-of- the-art professional grade Printed Circuit Boards. Our manufacturing programme encompasses Double-sided Plated through-hole Boards, Multi- layer Boards, Fine Line Boards, Boards for Computers, Telecommunications, Defence and Medical systems. Quality and reliability are key words in the Stovec lexicon — modes of expression which are given form and meticulously displayed in all phases of the manufacturing programme. We invite you to join the growing list of our satisfied customers. STOVEC SCREENS INDIA LTD. (Circuits Division) Plot No. B-24, G.I.D.C. Electronics Industrial Estate, Sector 25 Gandhinagar 382 025. Tel. (02712)22061 Telex : 121-326 STVC IN Cable : "CIRCUITS" or A.T.E. PRIVATE LTD. 36 SDF II, S.E.E.P2. Andheri East, Bombay-400 096. Tel. . (022) 6349032, (022) 6300195 Telex : 11-71513 DIP IN. 1C SOCKETS HIGHLY RELIABLE CONSISTENT SUPPLY AT AFFORDABLE PRICES I INDOOR UNIT FOR SATELLITE TV RECEPTION — 3 by J & R v. Terborgh This is the final, optional board in the IDU. As promised in the preceding instdlments of the series, it comprises the AFC, scan and remodulator facilities, as well as the LNB theft alarm. The circuit board described in this article is not, strictly speak- ing. indispensable for a fully operative indoor unit, But then, the'optional add-on circuits are relatively simple to build on a single PCB, and may provide you with a number of quite RESET input, pin 4. Electronic switch ESi is closed, while ES. is opened, so that the DC- coupled video signal. CVBS-1 (see Part 2) is routed to TV modulator ICts. The operation of this versatile RF chip will be reverted to. useful extensions. Circuit description The circuit diagram of the op- tional extension board is shown in Fig. 18. The various functions it offers are best discussed by starting from the three possible positions of the front panel mode switch, S4a b. 1 . TUNE: Bia b is set to position 1, as shown in the circuit diagram. Oscillator IC9 is disabled by the low level at its The RF board tuning voltage, ( Vtune, is taken from the output of summing opamp A;. which is driven with the tuning control voltage (terminal T. controls P6-P7). and the output voltage of AFC amplifier Ai. If AFC switch Ss is opened (AFC off), ESs and ES- are off and on, respectively, which means that the voltage at the + input of A- is a fixed level, de- termined With P9. Vtune will, therefore, track the voitage at point T, just as if there were no amplifier of any type in func- Switching on Si. however, causes Bdc, rather than the voltage at the wiper of P9, to be fed to the + input of A-. This creates a feedback loop in the tuning voltage circuit. It will be recalled that Bdc is the smoothed direct voltage com- ponent of the baseband video signal. Tracing its origin will re- veal that Bdc is the proportional equivalent of the PLL-generated tuning voltage across varactor Dj. i.e. it can provide infor- mation about the instantaneous centre frequency of the PLL subcarrier (see Pan 1). Assuming the AFC function to be switched on, and assuming that the selected oscillator, LOl or LOh, starts to deviate from its set frequency— which may well happen owing to thermal ef- fects— the PLL will conse- quently alter the voltage across D?— and hence Bdc— to match its VCO frequency with that of the incoming carrier at about 610 MHz. The AFC circuit next responds to the assumed fluc- tuation of Bdc by correcting Vtune such that the oscillator re- mains at the set frequency, i.e. Bdc also remains constant! The practical limitations of the proposed AFC circuit mainly concern the response speed of the loop, and the AFC hold range. The AFC circuit should be insensitive to the demodu- lated video component, which, of course, is also the PLL action to an FM input signal. This func- tion is taken care of by Cso (see elekto. India februaiy 1987 2-21 Fig. 18. Circuit diagram of the or Part 2), as well as Cei. Feedback resistor Rea defines the AFC hold range, i.e. the span of Vtune that ensures a constant Bdc voltage. The stated value of this resistor fixes the amplification of A. at about 3 KRw + RnVBu]. which will ensure sufficient AFC action in most practical 2. SCAN: Su b is set to position 2. ESi is closed, and ICi oscillates at about 10 Hz. The triangular wave at pins 2 & 6 is amplified to about 30 Vpp by means of Ai, which conse- quently causes the relevant os- cillator, LOl or LOh, to produce a swept output frequency over its entire mixer injection band. The purpose of the SCAN fa- cility is to facilitate the initial dish positioning procedure As soon as the dish "sees" the satellite, there will be a marked change on the TV or monitor screen from stable noise to a rather unsteady flicker, caused by the receiver sweeping across the incoming transpon- der signals. Also, the S-meter will show some deflection and hence can be used to find the initial aerial position. 3. TEST REMODULATOR Si,, b is set to position 3. ESi is opened, causing ICi to oscillate at 1S6.25 kHz. or 10 times the TV ' line frequency. Counter ICio ' supplies two sequential 7 ps pulses; one for use as a line blanking pulse (Qj), and one for j a white vertical bar (00- These | pulses are combined by means ! of ES? and ESi to form what can hardly be referred to as a com- posite video signal, yet is en- tirely satisfactory for the pres- j ent purpose. Resistors Rs, and ; Rss have been dimensioned for a blanking/white ratio of about 1:3. ESi is closed, while ESi is so simple as to obviate the need opened, so that the video test ' for a detailed description. With signal is passed to TV modu- three jumpers installed as lator IC i6 . shown by the dashed lines, LED The remodulator test facility j Du and buzzer Bzt will warn of enables ready tuning of the TV attempts to steal the costly LNB. set to the modulator output fre- The jumper block and the po- quency. thereby slightly allevi- tential-free relay contacts ating the possible difficulty in 1 should enable a straightforward setting up a satellite reception connection of the LNB theft system for the first time. alarm to many types of existing alarm system. Table 3 shows some of the possible alarm con- LNB theft alarm (ICis;Ti3). figurations plus associated The relevant circuit section is ; jumper positions. phone: 101 3481 9420 9425; om EleclroValue; Telephone 41 33803 or 1061 432) 4945. Remodulator (ICie). The Type TDA5660 from Siemens is an all-in-one TV modulator chip which can be configured for a wide variety of TV standards. In this design, it provides a double-sideband, AM vision. FM sound. TV signal at 48 MHz. which is roughly channel 2 (48.2S MHz. Band I). Operation on channel 3 or 4 is also possible by simply using an appropriate crystal in the Xi position. The circuit may also I be modified to output a UHF TV signal (470-790 MHz), but this i rather more complicated than hanging the crystal, and i refore, only recommended j experienced RF construc- . The matter will be reverted j n the section on construc- The audio input signal to the TV modulator chip is passed through a pre-emphasis net- work. Rsj-Css (t =* 30 ps). The modulator chip provides wide- band FM modulation at the audio sub-carrier frequency of 6.0 MHz, as set with W The VHF output signal is available at symmetrical outputs pins 13 and 15. A double pi-filter, Cn-Iai- C96 and C»/-L 2 i-C 98 , precedes 300R-to-75R balun L20, form which the TV signal is taken by C.oo. Trimmer Css is used to set the modulator output filter for optimum balance. The dashed lines around the remodulator circuit denote metal screens elektor mdia febfuary 1987 2-23 the preclude stray which serve radiation. i Construction If you have made it so far in building the IDU, you are not | likely to encounter serious diffi- culties in getting the present extension board up and run- ning. Fig. 19 shows how PC board 1 Type 86082-3 is to be com- j pleted. Only three points re- I quire special attention, namely I making Lis and Iuo, and fitting ! the extension board on top of j the vision-sound-PSU board de- scribed in Part 2 of this series, i In order to avoid unnecessarily j repeating the suggestions for i making one's own inductors, it is recommended to re-read the passage on preparing Lis; this 1 can be found in El'ektor India , j December 1986. j With reterence to Fig. 20 and I Table 4, oscillator coil Ins is i made as follows (note that the white ABS former as part of the Type 7T1S inductor assembly is divided into two equally long | sections by means of a small 1. Starting from f , and observ- [ ing the indicated winding di- I rection, close-wind 11 turns in 1 upward direction onto the I lower section of the former body; doing so will neatly fill this section. Connect to b (not toe!). 2. Starting from e', and once more observing the correct winding direction, close-wind 4 turns upward onto the upper section of the former; the first turn should rest against the rim. Connect to a. 3. Check for any short-circuits between the windings, and verify correct continuity at the pins. 4. If you have a GDO, check I whether the inductor can be j tuned to about 50 MHz with a I 18p capacitor temporarily fitted : across f-b. ; 5. Mount the former plus I screening can onto the PCB. Adjust the yellow-tipped core I until its top is level with the hole in the screening can. As to Iko, the construction of this balun (balanced-to-unbal- anced transformer) is evident from the six-step instruction shown in Fig. 21. Almost any type of small, two-hole ferrite bead rated for at least 100 MHz can be used in this circuit. The inductor is wound with bifilar 2-24 slsklor India Mnusry 1987 20 L18 Neosid 7T1S «.d fioo. unde. eath •e •"W* . •€» •j — ^ • b start ol winding 21 T-V V nip, ;>#c it u u ^A A 2,SWG3 ° 1 n si. in UJ IS 5 51 Fig. 21. Suggested constructior J wire, which is simply made by | twisting two lengths of enam- ! elled copper wire. After wind- ing two times three turns ; through the bead holes, the j string ends are split in order to | identify the four individual wires by means of a resistance meter or a continuity tester (step IV). At this stage, it is a i good idea to check the wires for internal short circuits caused by the insulating enamel coating being damaged as the windings are tightened around the ferrite bead. After making the balun and fit- ting it onto the board, it is time to check whether this is cor- rectly populated. There should ' be six wire links in all, and the of balun Lzo. jumpers in the LNB alarm cir- cuit should be fitted as re- quired. Positions Cn and Cm are vacant as yet. Make sure 1 that all ceramic capacitors in j the remodulator section are | ' mounted with the shortest poss- j ible lead length. The crystal case must not be grounded, j The position of the 12 mm high metal screen around the remodulator circuit, and the lengthwise fitted screen across ICis, is governed by 9 soldering pins. A single strip of 12 mm ( wide thin brass sheet or tin plate is readily cut and bent to ] size. Remember to drill two small (^ 3 mm) holes in the | screen to enable feeding 1 through the shielded wire to the audio input, and, if re- I quired, the $ 3 mm coax cable j from the RF output to Ki on the I enclosure rear panel. The completed extension board is mounted on top of the rear side of the vision-sound- PSU board, i.e. as close as poss- ible to the enclosure rear panel. Remodulator output socket K» can be fitted at a suitable lo- cation in the rear panel, whilst being connected direct to the relevant pin on the PC board, ie. without a length of coax cable. Note, however, that this mounting method requires making a suitably sized hole in the previously mentioned screen, allowing for the passing of the socket. The lowest possible mounting height of the present board above the vision-sound-PSU board is determined mainly by the height of the fuseholder on the latter PCB. Sufficient stab- ility of the "sandwich" con- struction is ensured by using two conventional 15-20 mm long PCB spacers in the two rear positions. It goes without saying that the overall height of the two-board unit should enable the IDU to be closed properly. Also, the vision-sound-PSU board should be fully operative and correctly aligned, since many of its ad- justment controls are no longer accessible with the extension board fitted on top. The wiring of the boards should be fairly straightforward, requir- ing no further remarks other than that the audio, Bd’c and V tune connections should be made in conventional shielded microphone cable, while the CVBS-1 connection is made in if 3 mm coax. In all cases, ground the cable shield at the lower board only. Finally, the external loop con- nection can be made with whatever type of socket or ter- minal strip is thought most con- venient; a 3- or 5-way DIN socket is satisfactory. Setting up Before detailing a suggested setting up procedure for the present board, it must be made expressly clear that attempting to use the completed extension PCB along with as yet un-oper- ative RF and vision-sound-PSU boards needlessly complicates getting the IDU to function cor- rectly. Therefore always build up the receiver as detailed in Part 2, and familiarize yourself with the various adjustment points and their typical re- sponse, before adding the pres- 1. Set S« to TUNE, and switch off the AFC (Ss). Turn Pi (coarse tuning) to check whether Vtune varies from about 1-30 V. Tune to a satellite programme and check the presence of com- posite video at pin 10 of ICia. Do the same for the audio at Measure Bdc, note the value, and adjust Pa for an identical voltage at its wiper. Switch on the AFC and check its hold range by turning P 7 ; reception should remain unaltered over a certain portion of the tuning control travel, then suddenly be lost. 2. Set S. to SCAN, and switch off the AFC. Use a scope to check measuring points (Q) and ©. Vtune should be an undistorted triangular wave, ie. it should have clearly defined points of inflection, and no clipped tops or appreciable offset. If necessary, Rat and R$s may be redimensioned to achieve the correct wave-form and ampli- tude respectively. Set Pa to the centre of its travel and observe the monitor screen to see the effect of the SCAN mode when a satellite is re- ceived. You may want to experi- ment a little with the value of Cai to obtain the best noticeable effect on the screen. Try to remember what it looks like! 3. Set Su to TEST REMOD.. and connect a TV set to K<. Tune the TV to channel 2. Adjust the core in L 6 . Depending on the degree of crystal activity, it may be worth while to fit a damping re- sistor (1K0-10K) across pins f and bofL,.. Run a quick check on the oper- ation of the LNB theft alarm by disconnecting the downlead cable at K> Please note that the alarm circuit is fed from the unswitched + 12 V supply. Therefore the +Bzi terminal on the PCB should be wired to the buzzer as well as the ap- propriate connection of Sa (see Pan 2). Finally, if the setting of Pa fails to give a satisfactory compromise between the operation of the SCAN function and that of the internal test pattern generator, try fitting a number of small capacitors in the Cn position. Remodulator on UHF The circuit diagram of Fig. 22 shows how to modify the on- board. TDA5660-based, TV modulator for operation in the UHF TV band (470-790 MHz). As this modification is not sup- ported by the PCB layout, alter- ing the circuit is recommended for experienced RF construc- tors only. Preset P is used to set the desired output frequency, which must be well removed from the PLL VCO frequency to avoid carrier interference. Therefore do not tune ICia to the generally used modulator channel 36. The small ceramic NP0 capaci- tors can be fitted in a three-di- mensional construction, along with oscillator inductor Luhf which can be spaced or com- pressed slightly to set the initial output frequency. The lp5 capacitors are, of course, fitted direct across the relevant IC pins at the PCB track side. The modulator output filter must also be altered as shown to allow for the higher fre- quency. Use a suitably rated bead for L 20 , and wind two turns through each hole, rather than three as in the VHF circuit. The data for Luhf, Lx and Ly can be found in Table 4. Aerial positioning unit The circuit diagram of Fig. 23a and the photograph of Fig. 23b show a simple, yet indispens- ible accessory unit for the IDU. It is a hand-held remote meter circuit which is connected to the IDU over a length of 6- or 7-way cable, enabling the user to monitor the S-meter indi- cation while lining up the aerial for optimum reception. It should be noted that the cir- cuit diagram and practical re- alization are but suggestions; other configurations, as well as more sophisticated controls are perfectly feasible, and con- structors should have little diffi- culty in tailoring the aerial Fig. 22. Modified circuit for the remodulator, if this is to operate ir HH Test set up to examine the performance of the BFG65 prestage in the IDU. Display indications, left to right: frequency (MHz) as- sociated gain idB); noise figure IdB). Courtesy of SSB Elec Fig. 23. Circuit diagram (23a) and practical outlook (23b) of the aerial positioning unit. positioning unit to their specific 1 requirement. With reference to Fig. 23a, the meter should be a more sensi- tive type than that incorporated in the IDU Either a switch, mounted onto the IDU rear panel, or a socket contact, is used to break the S meter driver output to the front-panel mounted meter, and route the signal to the aerial positioning unit. A buzzer is fitted to enable the person remaining at the IDU to notify the other person at the aerial that the IDU is switched from SCAN to TUNE following the slightest sign of reception on the TV or monitor screen. In practice, the aerial position- ing unit may be used as follows (note that a detailed aerial posi- tioning method will be dis- cussed in next month’s final instalment of this series): 1. Set the IDU to SCAN. LOl or LOh depending on the satellite to be received; connect the positioning unit cable, and, if possible, install a helper at the IDU. 2. Take the positioning unit to the aerial site (on the roof, in the garden, or wherever reception is thought feasible). 3. Set the unit to maximum meter sensitivity and line up the dish until some deflection is seen. Hopefully, the person in- side has noted the SCAN effect on the screen, and, via the buzzer, notified you that the meter indication will be lost for an instant as he tunes to some transponder. If no help is available, leave the dish roughly positioned and go inside to switch from SCAN to TUNE yourself. Reception of the satellite may still be weak at this stage, but you have at least managed to find a stable signal. 4. Co outside again and line up j the aerial for highest meter deflection, turning down the sensitivity any time the meter reaches its fsd indication. Threshold extension The following is a necessarily brief examination of a number of experiments with the PLL demodulator. IC2, on the RF board. As these experiments are not supported by the PCB layout, their being carried out is only recommended for ex- perienced RF constructors. Also, since the objective of the proposed modifications is to further lower the PLL noise threshold so as to improve upon I reception with relatively low C/n ratios (8-10 dB), there is no | point in altering the PLL circuit if your specific outdoor unit en- sures a C/n output of more than , about 12 dB. When the C/n ratio at the input of the PLL demodulator ap- proaches the noise threshold, the received picture is more or less impaired owing to noise j spikes occurring primarily in j the saturated colour areas. This effect is mainly due to insuf- ficient open loop gain of the PLL at the chroma subcarrier, 4.433 MHz (PAL system). Incorporating a chrominance filter in the secondary PLL loop may improve reception to some extent, but it should be noted that the effect depends on the transponder deviation and bandwidth. For instance, the signal from Teleclub Switzer- land could be slightly improved by peaking the chroma filter whilst observing the few re- maining sparklies in the ochre rectangle at the lower right of the test chart. Correct tuning of the series filter will enable the sharp white-to-black transitions in the chart to appear with a clearly improved definition. The practical circuit of the chroma filter extension is shown in Fig. 24a. It will be recalled that C20 and Czi define the secondary loop response and hence the oper- ation of the PLL at a specific portant to realize that, at pres- ent, there is no single standard for the peak-to-peak deviation of transponders, not even if these are part of one and the same satellite. Research carried out by the EBU and the CCIR has provided evidence for the proposition that, given a specific C/n ratio, S/N rises with increasing deviation. It is, therefore, arguable that future satellites will hold transponders with larger output bandwidth; after all, a number of the pres- ent generation of TV satellites were originally designed to j operate in data communication networks. It may be interesting to experi- ment with the values of C20 and j C2i while observing the signal | from a relatively weak transpon- der. The range of values that can be fitted in the stated ca- j pacitor positions is quite large j —see the small inset table in ' Fig. 24a. Fig. 24b shows how the secondary loop differential amplifier is converted into a single-sided type by decoup- ling the IiFB? input and the V output with lOOn ceramic ca- pacitors. This modification is called for when receiving satel- lite signals with a peak-to-peak deviation of the order of 25 MHz. It should be noted that such a high deviation value does not necessarily mean a higher bandwidth; in next month’s article we will examine the exact relationship between Finally, interested constructors are advised that Plessey have recently introduced the Type SL1455 quadrature FM TV de- modulator. which is stated to achieve a noise threshold of about 7.5 d& is. it is some 1 dB bener than the SL1451 con- figured for optimum operation given a specific deviation. RF board measurements The IF amplifier chain on the RF board was studied with respect to its frequency vs amplitude characteristic. Use was made of a 0-1800 MHz spectrum analyzer plus associated sweep unit. Fig. 25a shows the curve of a wrongly adjusted IF chain; one of the four bandfilter trimmers has obviously been set at too low a frequency, causing a marked peak outside the re- quisite pass band. While adjusting the bandfilters to obtain a satisfactory filter response, it was found possible to locate the pass band any- where in the 450-650 MHz band, while the bandwidth was never less than about 35 MHz. There- fore, constructors not in pos- session of an RF sweep gener- ator or other sophisticated equipment to measure the IF bandwidth need not worry too much about the overall pass- band of the RF board. As long as all trimmers can be tuned for stable noise output, the initial alignment is satisfactory. Fig. 25b shows the band pass curve obtained after very care- fully peaking the trimmers for optimum reception of the test chart broadcast by Teleclub Switzerland on ECS-1. The curve thus obtained may be compared to the theoretically required one shown in Fig. 25c. The latter is used by the EBU to specify the minimum require- ment for Eutelsat-1 receiving stations. Next time Next month's concluding article in this series will tackle a wide variety of questions raised in connection with satellite TV reception. So. should any aspect of the present subject matter still puzzle you, see whether it is among the sub- jects qualified for closer exam- ination in Part 4. RGK;Bu 25c (dB) (dB) (dB) I (dB) (dB) 2.5 (I0.0)" (25)' ment for out-of-channel filtering in the transmit equipmei amended that out-of-channel filtering be provided in the 86082-3-25c 2-27 THE FUTURE BELONGS TO THE PHOTON Electronics has been the main engine ot innovation since the invention ot the transistor 40 years ago. Most of tomorrow's interesting technologies will work by manipulating light, not electricity. The electronics revolution is young. The electron was ident- ified less than a century ago and the microchip, on which today's information-technology industry utterly depends, has been around for fewer than 20 years. The successes crammed into these two hectic decades have created the impression that electronics is a technology capable of limitless improve- It is not. Electronics will give way to a superior technology based not on electricity but on light. Physicists did not realize until early in this century that light came in the separate packets they now call photons. But science has made startling progress in manipulation photons. A photonics revolution is already in the making. The first shot of the electronics revolution was the transistor. Photonics' first shot was the in- vention, in 1960, of the laser. Un- til then, those trying to do tricks with light had to make do with a jumble of disorderly wave- lengths. Lasers create a source of light with a uniform wave- length and with each wave mov- ing in step with its companions. This is a tool of immense power. Lasers can— or so President Reagan hopes— destroy ballis- tic missiles thousands of miles away. They can cut metal in fac- tories and repair blood vessels in human eyes. Hospitals use laser beams guided through optical fibres to shatter people's kidney stones. A French inven- tor has replaced the strings of a harp with laser beams. Like transistors, lasers have shrunk: they can now be generated by a chip the size of a grain of sugar. This is paving the way for a wholesale switch from elec- trons to photons. Why is the switch worth mak- ing? Because photons travel faster than electrons; because J they have no mass; because (un- like electrons, which interfere with each other) photons can be made to pass through each other unperturbed; because light behaves both as a particle and as an electromagnetic j wave-which means that optical devices could be based on much the same operating prin- ciples as those already used in electronics. Moreover, electronics is dis- covering its limits. One is the speed at which electrons travel through semiconductor ma- terials. So long as electrons re- main the information carriers ol computers, this sets an absolute limit on the speed— and hence power— of computing. Elec ironies has not reached that limit yet, but it is drawing close enough to worry engineers. The customary way to make computers cheaper and fastei is to squeeze electronic compo- nents closer together. The number that can be fitted on a single chip has grown from about a dozen 20 years ago to 2m today. But miniaturization, too, is bumping against limits. Engineers are running out ol ways to etch into chips ever- smaller paths along which elec- trons can run. And when com- ponents get too close, the chips are plagued by "cross talk— the leakage of charges from one component to another. If computers are to work faster still, a new approach is needed. The best bet is "parallel pro- cessing'— the notion that com- puters ought to be able to per- form a lot of operations simul- I taneously, instead of channel- ling all their calculations through one bottlenecked cen- tral processing unit. Here, too, the case for a photonic solution is compelling. Sending several electric currents through one chip at the same time risks cross-talk and disaster. Not so with beams of light: a chip could process several at once without their interfering with each other. | Still sceptical? Consider how j rapidly light has nudged elec- I tronics out of two pillars of in- formation technology: telecom- munications and the storage of information. In communications, telephone companies are tearing out their j copper cables as quickly as they can afford to and replacing them with hair-thin optical | fibres made of glass. Light is a better messenger than elec- tricity. It wastes less heat and is | immune to electromagnetic in- 1 terference. Better still is light’s enormous bandwidth. Because it spans so many frequencies, light can squeeze in far more in- | formation than electricity can. j The quality of the optical fibres themselves has improved dra- matically. In early (circa 1970s) j fibres, light ran in a disorgan- ized zig-zag through a relatively large core within the fibre. The resulting collisions with the fibre's cladding absorbed much of the light, requiring frequent repeaters to refresh the signals. In 1977. experimental fibres transmitted up to 140 megabits of data a second, and needed a repeater every six miles or so. Today, one experimental fibre network installed in Britain carries telephone traffic at 1200 megabits a second, with 30 miles between repeaters. The first transAtlantic fibres will be carrying data and telephone conversations between Europe and America in 1988. Yet the technology is on the threshold of another luminous leap. This will not come from changes in the fibre itself, but from the devices used to send and receive the optical signals. The first step is to combine in a single device all the parapher- nalia that optical fibres re- quire— lasers to send signals, detectors for receiving them, and a rag-bag of lenses, mirrors and electronic controls. I The second step is to transmit light beams “coherently— ie. in tightly-defined wavelengths— [ into a receiver that can be tuned to select the required wavelengths and sort out the separate streams of data. In principle, coherent trans- mission enables a single fibre to carry 10m telephone conver- | sations or 10 000 digital tele- [ vision channels at once. The optical assault on data storage— that other pillar of in- | formation technology— has j been as impressive. Music j lovers were in the van with their compact discs. The music is I turned into digital signals, burned on the disc as a series of minute pits and then de- coded for playback by a low- power laser. Audio discs like these are only the first big success of a technology restlessly seeking new applications and markets. Optical discs are beginning to replace magnetic ones as a way to store computer archives. Because they are tough, the discs can be stored inside specially-constructed juke- boxes. One 4.7-inch disc can store about 550m bytes of data— the equivalent of 1500 floppy discs or about 250 000 printed pages. Which means a jukebox can store the archives of an entire government depart- Optical discs suffer from one drawback: erasing them or writing new information on j them is difficult. This has im- | peded their marriage with com- puters. but has also prompted an imaginative hunt for appli- cations in which data must be stored permanently without I alteration. Discs sold under a standard I format known as compact- disc read-only memory (CD- ROM) are enabling data-base I companies to sell archival infor- mation to subscribers cheaply by post instead of expensively by telephone. Grolier, an | American publisher, has put its j Academic American Encyclo- I paedia (30000 articles, 10000 pages) on one-tenth of one disc, which it sells for less than $200. | A new generation of discs J called WORMs (write-once- read-many-times) is half-way | there. These are sold blank, so the end user can store whatever j data he likes on them, although the information, once stored, is I there to stay. But the technology | for a fully-erasable disc will | probably be perfected by the end of the decade. Two ideas for making them are. already showing particular promise. One is based on a magneto- optical process. The disc's re- cording layer is an alloy of ter- [ bium, iron and cobalt. To store information, a laser heats up a j tiny spot on this layer, creating a | ( vertical magnetic field. The in- J [ formation is read by another laser: whenever it encounters a j magnetised spot, the light’s plane of polarization is rotated. The information can be erased 1 j by reheating the spot, j The other approach is chemi- cal. Here, a laser is used to I switch the structure of a tel- j lurium alloy back and forth be- j tween amorphous and crystal- , ■ line phases, which reflect light differently. Impressive as they are. the progress made by optical discs and fibres do not amount to a revolution. Photonics will not j come fully of age until it equals and then surpasses, the central triumph of the electronics revol- ution: the computer. At the heart of the computer | sits the transistor. A transis- tor, remember, is a switch, a \ device that can flip backwards and forwards between two states. Computers are chains of switches. They treat sequences of ons and offs to denote num- bers (in which case ons and offs 1 are read as the ones and zeros of binary counting) or to denote "true or false" (in which case ■ chains of switches can be used as the building blocks of algebraic logic). The challenge for photonics is to invent a device that does for light what the transistor does for elec- Into the heart of the ! computer It has virtually happened. At ! AT&T's Bell Laboratories and Britain's Heriot-Watt University in Edinburgh, small and primi- j tive circuits of the kind that could one day grow into com- puters are already running on i light. The switches they use- known variously as bistable op- tical devices (BODs) or trans- phasors— are essentially optical transistors. Light emerges from them as a strong beam (on) or a weak one (off). Put a bunch of transphasors together, shine laser beams through them, and you have the basic ingredients of an optical computer, lb understand how a transphasor works, think of it as two partially-reflecting minors facing each other. If a beam of light is shone through them some of it gets trapped, bounc- ing backwards and forwards between the mirrored surfaces (see diagram on next page). As these waves cross each other they can either interfere with and weaken the beam or align with it and reinforce it. This phenomenon is the basis of a simple instrument— used to measure wavelenghts— in- vented by two French scientists, Charles Fabry and Alfred Perot, in 1896. The Fabry-Perot interferometer emits a strong beam or a weak beam depending on whether the waves are being reinforced inside the cavity. On its own, however, it is not a switch: a useful switch needs to be ob- viously on or obviously off. Common sense says that a gradual change in the intensity of the beam shining in will pro- duce a gradual change in the beam getting out, not the ab- rupt change that is needed. In ordinary circumstances, com- mon sense would be right. In the case of the transphasor. it is To make the Fabry-Perot in- terferometer into a switch, physicists hit on the idea of mar- rying it with a phenomenon known as optical bistability, first observed at Bell Labora- tories in 1976. The secret is in the cavity between the mirrors. I If this were filled with an or- dinary medium— air, say, or I most solids— the intensity of the beam passing out of the minor would, indeed, change in pro- portion to changes in the inten- sity of the beam shining in. Transphasors, however, use a family of materials (such as in- dium antinomide and zinc sel- enide) that are "non linear". If a laser beam shines into these materials, a slight change in its intensity can trigger the wave- reinforcement and make the beam coming out of the trans- phasor suddenly brighter— and make it stay that way until the j trigger is released. Bell Laboratories and Heriot- | Watt have made different sorts of transphasors, but they both work. Heriot-Watt's are entirely optical: the laser beams are shone into bistable plates made of zinc selenide. Bell is trying a hybrid approach. Its devices, made of gallium arsenide, use | electro-optical interference | within the cavity to trigger the I reinforcement effect. In an op- j tical computer, these devices [ would be the "chips", and the "wires" would consist of laser I To make a computer, it is not I enough to be able to turn just j one switch on or off. Computers : are complex arrays of switches, | each of which feeds signals into the next. So optical switches must be “cascadable— the beams of light emerging from one transphasor must be able to flip the next, and so on. They must also be able to receive and send several signals at the same time (properties known respectively as "fan-in" and “fan-out"). These obstacles are tumbling fast. Last year, for example, the team at Heriot-Watt University showed that its zinc-selenide transphasors could be kept near their threshold by a holding laser, then switched by turning on a small extra beam. Earlier this year, the team an- nounced that it had placed several transphasors in a cyc- ling loop. Optical switches should, in theory, be able to operate 1000 times faster than electronic ones. But do not throw your electronic computer away just yet. For the present, trans- phasors are primitive. They still have to be pumped by too much light, and they are still bulky, separate devices— they have not yet been squeezed together on chips in the way electronics switches have. Even so, optical switching works. Hybrid vigour Laboratories everywhere are rushing to bring optical and electronic switches together. One motive is to make even bet- ter use of optical fibres. Exist- ing optical networks do not work at the speed of light, because the messages the fibres carry are shuttled be- tween machines such as tele- phones and computers that run— for now— on electricity, not light. So at each end of even the niftiest optical fibre sits a cumbersome device whose job is to transform optical pulses into electronic ones and vice To speed this procedure, engineers are creating op- toelectronic chips. To do so. they have had to conquer a dis- advantage of the photon— its in- ability to carry an electrical | charge. Picking signals off the j end of an optical fibre demands some way to sort out waves of light and send them to different destinations. Electrons can be shunted by the application of an electric field; chargeless photons are impervious to such methods. The answer has been to chan- | nel the light through "wave- j guides" etched into chips made of materials with unusual | optical properties. These materials change their ability to conduct light when an electric field is applied to them. Using lithium niobate, engineers have | been able to make a wide range of optoelectronic modu- lators, switches and other de- j But there is another reason for 1 wanting to bring the photon and ' the electron together: parallel , processing. Britain's Plessey j has developed a BOD in which j the bistability comes from in- | serting a photochromic ma- ill | 87009-2 terial— one whose chemical form changes when exposed to different wavelenghts of light— into the cavity. Plessey believes the device could be used for parallel processing. The idea is to squeeze an array of BODs on a single two-dimensional plate. Each then becomes an in- dependent switching centre that can be addressed simul- taneously by an incoming laser beam (see diagram below). This approach comes into its own in applications such as image-processing, in which the value of thousands of picture elements (pixels) must be in- dividually calculated to build up a whole picture. Plessey aims to get around this data- processing bottleneck by using light to process all the pixels at once. The optical switches are not yet as fast as elec- tronic ones, but that hardly mat- ters when they work simul- taneously. Plessey reckons that with its photochromic BOD, a device the size of a finger-nail could process 4m pixels in one ten-thousandth of a second. Photonics has come a long way in the quarter century since the arrival of the laser. But entirely new ideas for manipulating and exploiting light are still pop- j ping up. These range from the mundane (mechanical and bio- ] logical sensors based on op- j ticai fibres) to the frankly I quixotic (travelling to the stars j by giving spacecraft sails that catch photons). Physicists have begun to use laser beams to trap individual atoms so they can be observed in detail. Engineers envisage massive computer memories with data encoded within the light-waves of a hologram. Why this sudden flowering? In the 1970s, physics made a wealth of discoveries about the ways in which light interacted with matter. These discoveries are now finding applications. The properties of non-linear materials— which made the transphasor possible— are one example, but there are others. In some circumstances, light travelling through a material sets up internal sound waves that contour themselves like a deformable minor, sending the light backwards out of the substance on the path along which it entered. In 1972, Dr Boris Zeldovich and colleagues at the PN Lebedev Physical In- stitute in Moscow used this pro- perty to make something called a phase-conjugate mirror. This is no ordinary minor: it can take an image that has been dis- torted and then straighten out the jumbled-up waves to recon- stitute the original image. Like so many technologies, the mir- ror was treated as a laboratory curiosity at first. It is now being pressed into service by astron- omers to take the twinkle out of stars, and by star-wars generals to shoot laser beams through the turbulent atmosphere. The minors can also be used to pro- ject three-dimensional images through optical fibres and to etch tiny components on micro- chips. One way or another, light looks like the wave of the future. Reproduced with permission from The Economist UNIVERSAL CONTROL FOR STEPPER MOTORS With good quality stepper motors widely available at reasonable cost, this flexible, computer-driven, control board will make it rather hard to hold on to the belief that stepper motors are the exclusive realm of industrial electronics. If you are suspicious about "universal", just glance at the speficications Table below; if you are into industrial electronics, well. . . Stepper motors come in an as- tounding variety of types and sizes, and they are frequently spotted items in electronic surplus stores and on hobby venues. Sheer curiosity has I prompted many a home con- structor to purchase one at a fraction of its original price. However the number of wires | coming from the device, and I the fact that it is often found far more difficult to get going than often than not causes the per- plexed owner to carefully put his price possession in the junkbox, together with other "possibly useful" materials. In Stepping Motors, Elektor India , May 1985, the general methods were examined for the the driving of stepper motors. Also that article provides a use- ) ful discussion of stepper motor I terminology, used further on in I this article. ' The main specifications of the proposed control board are summarized in the shaded Table on this page. The board is readily tailored to suit the user's requirement, but it should be made quite clear at the onset that each of the following sec- tions is to be read closely to be able to decide on the most favourable circuit configuration for a specific application. A de- tailed discussion of each of the technical features is, therefore. I indispensable to a good under- standing of the operation of this I fully user-configurable inter- face board between computer and, for instance, robot limbs, a I pantograph, or a plotter. eleklor mdia lebruary 1 987 2 -3 1 : rismm I L293E fitted: 1 A phase. L298 fitted: 2 A phase Software-controlled polarity 10. 35 V with L293E fitted. 10. 45 V with L298 fitted. Stepper motors: some problems The following is a necessarily brief discussion of the main dif- ficulties to be overcome when using stepper motors. Limited speed range: the stator windings constitute an induc- tive load, which limits the com- mutation speed of the coil current. Also, the revolving, permanent magnet rotor causes an inductive voltage which further worsens the commuta- tion. These effects limit the maximum attainable step rate (also: pull-out rate), but can be overcome by utilising current drive control. Resonance the undamped character of a stepper motor operating at a relatively low step rate causes its movement to be rather halting. The upper oscilloscope trace in Fig. 1 shows the considerable over- shoot after each step. Should the step frequency equal that of the underdamped oscillations, resonance inevitably occurs, causing a powerful, jerky move- ment of the spindle. Mechan- ical damping devices have been developed to ensure a smoother spindle movement, but these permanent loads typically cause the already low efficiency of the stepper motor to fall below the acceptable The lower oscilloscope trace shown in Fig. 1 provides evidence for the proposition that micro-step operation can provide a marked improvement in linear spindle movement, thus enabling the direct transfer of motor power via a set of Low efficiency: an energized stepper motor dissipates an amount of energy in the re- sistive load formed by its stator windings. When the spindle is held stationary, this resistance is the sole current limiting fac- tor; also the stall torque is often needlessly high. Current drive systems may enhance the dy- namic characteristic of the step- per motor to some degree, but linearly controlled current sources, unfortunately, exhibit a very low efficiency. The present design is based on the use of high efficiency, switch-mode current sources, thereby going round the prob- Overshoot j£>| a i \ “ ' a Fig. 2. Block diagram of the stepper motor control board lems associated with the previously mentioned systems. Also, the proposed current driver has the advantage of be- ing uncritical of its input supply voltage; extensive regulation and smoothing circuits are, therefore, not required— an im- portant fact in view of the poss- ibly high currents involved in operating the stepper motor. As the current through the wind- ings is fully programmable, the user can arrange for the overall dissipation of the stalled motor to be significantly reduced. Limited resolution: stepper motors are classified according to the number of steps per spindle revolution. Using the micro-step mode, this specifi- cation becomes less important, and a specific type of motor can, therefore, be tailored far better to the task it is to per- form. Block diagram After these preliminary con- siderations, it is time to have a look at the block diagram of the stepper motor control board- see Fig. 2. This design is in essence a quad bipolar power driver. Each driver consists of a bridge circuit and can supply both negative and positive out- put current from a single supply. Starting at the input side, it is seen that each driver comprises a latch and a D/A converter to enable program- ming the level and the polarity of the current fed to each in- dividual stator in the stepper The switch-mode current sources are essentially voltage- controlled pulsewidth modu- lators (PWMs), driven with the difference between the object amount of stator current and the actually measured current. These two values are obtained from the D/A converter and a DC current sense amplifier, re- spectively. The four driving PWMs are synchronized via a common 40 kHz oscillator sig- nal. which ensures a favour- able switching frequency— the switch losses are still accept- able, and the signal is inaud- ible— as well as the absence of beat signals. At the top of the block diagram, there are some more circuit functions common to the four drivers. An address decoder uses the two MS (most signifi- cant) bits to discriminate be- tween the control data sent to each of the four driver circuits. Provision has been made to use handshaking with the computer for optimum reliability of the of data transfer to the board. A ref- erence voltage source makes it possible to use D/A converters without an interna! reference circuit. Finally, a S V supply powers all logic circuits on the Depending on the application you have in mind for the step- per motor control board, this need not incorporate all of the previously introduced circuits. For instance, the relatively ex- pensive D/A converters may be omitted if you do not envisage using the micro-step facility, but would still want to be able to program semi-step operation. The proposed board makes it possible to drive a four-stator system, even with two separate two-stator motors. It is possible to operate one motor in the micro-step mode, while the other one is controlled in the standard way, i.e. by means of a "stripped down" driver circuit. The user is offered a choice of two possible types of driver IC, which can be fitted as required by the expected output current. As you can see, our use of the word "universal" in the title of the present article is fully justi- fied. Circuit description It is not very difficult to spot the I various functional blocks in the circuit diagram, Fig. 3. As to the j aforementioned common cir- cuits on the board, ICj is the 5 V regulator, IC* the 40 kHz oscil- lator, ICe the one-of-four driver | decoder, and zener diodes Du and Du may be used to provide DACs IC1.-IC.4 with a highly stable 2.5 V reference. i On rece ipt o f a computer-gen- erated STB or STB (strobe) pulse. ICe decodes De and D> in the sent dataword and enables the corresponding sextuple latch, ICi. ..ICio, to clock the 6-bit value which determines the output current level sup- plied by the driver (Do. . .D4) as | well as the polarity (Do), j Therefore, only five bits of the J six or eight-bit DACs are used to translate the latch output into a voltage between 0 and 2.5 V in [ 32 increments (2 5 ). Each of the DAC output voltages is used to drive the inverting (+) input of opamps A2, A3, As and A7. How these in turn are capable of | determining the stator output i current is detailed in the next section. Returning to the handshake cir- cuit composed of ICe. N • and N2, it is seen that both positive and negative-going strobe , pulses can be used by fitting the appropriate wire jumper, a | (STB) or b (STB). Note, however, ; that in many Z80-based systems ! STB is an input signal, and RDY (ready) is an output signal, i.e. the signals are reversed as com- pared with the Centronics stan- dard. Jumper a is to be fitted when driving the stepper motor board with either a Z80 PIO, or a 6522 VIA, while jumper b ac- comodates the use of a Cen- , ironies port. More information on the handshaking circuit can be found in Table 4, while Z80 PIO users may consult MSX ex- tensions - 4, elsewhere in this issue. PWMs and current drive In order to make clear the oper- ation of the switch-mode cur- rent driver circuits in this design, it is necessary to study Fig. 4. From a functional point of view, the Types L298 and L293E from SGS Ates are largely identical; these devices merely differ in respect of the maxi- mum available output current. The L298 is twice as powerful as the L2993E and is. therefore, housed in a Multiwatt -15 SIL enclosure, rather than a 20-pin DIL package as is the L293E. Each IC holds two indepen- dently controllable bridge cir- cuits plus associated logic drivers. Since these ICs are to be driven with logic voltages only, there would seem to be no way of controlling the bridge currents with a linear regulating system. However in each driver the emitters of the lower bridge transistors are brought out to pins, enabling the connection of an external current sense re- sistor which provides a voltage drop proportional to the stator current. Fig. 5 further illustrates this principle, which forms the basis of the negative feedback controlled switch-mode current Any duty cycle of the current drive system starts with IC* generating a 1 fis negative reset pulse for all four monostable multvibra.tors MMVi. . . MMV* . Taking MMV. and the upper section of ICi as an example, the reset pulse causes C.2 to be discharged to the zener voltage of Du. Simultaneously, MMVi is triggered, and provides an out- put period determined with network Ru-Cu as well as the DC level applied to the control voltage input, pin 3. This level is internally compared with the voltage across C12 and hence determines the length of the output period. Since the com- parator internal to the Type 556 MMV is incapable of linear op- eration with input control volt- ages below 1.5 V, Du leaves sufficient residual charge in Cu for the MMV to produce suffi- ciently short output periods. From this it is seen that the MMVs in the circuit essentially function as voltage-controlled pulsewidth modulators, en- abling the power output stages contained in IC> and IC2 for the duration of their output periods. I Therefore, current sense re- sistor Re carries the stator cur- j rent and hence produces a | proportional voltage drop, | which is averaged in network Ci6-Ru and raised in amplifier Ai. J Opamp A 2 compares the measured current (— input) with the object current (+ in- put), and corrects its output j voltage to MMV. until these two values equal. Simple as this may seem at a first glance, there is, however, a snag in the measur- ing of the stator current. As long as the bridge is enabled, stator current Is flows through Rscnse, and its voltage drop is simply IsRsenso volts— see Fig. 5, line a. The disabling of the bridge im- mediately breaks the current through R se nse, but not that through the stator winding, whose inductance causes it to supply a lagging current, which is driven into the supply via free-wheeling diodes-see Fig. S, dashed line b. In es- sence. the self-inductance of the stator winding has a smoothing effect upon the stator current. Therefore, the average value of Urs«i»o is not a direct measure for the stator current, since it does not com- j prise the free-wheeling current. I With most types of stepper | motors, the period L/R of the I stator winding is long as com- pared to that supplied by the PWM drivers (T--- 1/40 kHz = 25 ns). In practice, the variation | in free-wheeling current in be- [ tween driver pulses hardly causes any ripple, and the error incurred by only measuring the current through the sense re- sistor is, therefore, caused by the duty factor variation. In general, a relatively small duty ) factor variation suffices to give a considerable stator current span. As soon as the duty factor rises above some 50%, and the free-wheeling period starts to overlap the bridge on-time, Is rises relatively quickly. The re- quired duty cycle giving maxi- mum stator current is a function of the ohmic resistance of the stator winding and the supply voltage level. The higher that | voltage, or the lower that resist- ance, the stronger the tendency [ to large variations in Is around a 50% duty factor. The foregoing considerations J can not but lead to the conclu- sion that the output signal of Ai [ need not be exactly pro- portional to the stator current. 1 (DOr "*n Fortunately, the overall linearity is still acceptable, and occa- sional deviations can be com- pensated by suitable software. Returning to the circuit dia- gram, Fig. 3, the remainder of the circuit functions are quite conventional designs. Timer IC. provides the nega- tive-going 40 kHz synchron- ization signai for the R and T inputs of the MMVs. In the absence of a common sync signal, the input supply would be corrupted by a good many inductive voltage peaks, which would readily lead to the MMVs being triggered in error and the j entire circuit operation being j upset in consequence. Network R,-D« prevents 5 V regulator ICi from being dam- aged by too high an input volt- age. As the maximum input volt- age for IC3 is 35 V, the use of the Type L298 stepper motor driver (Vs:wx>=45 Vpe a k) necessitates fitting the voitage limiting net- work. But even with the L293E fitted in the circuit, it is still a good idea to use Ri and D22, as they also afford protection against inductive voltage peaks on the unregulated supply rail. The use of the 2.5 V reference diodes Du and Du' is not obligatory, and their use will be reverted to in the section on construction. The logic sections of the circuit are composed of CMOS ICs only. This means that the logic drive to the board must be capable of supplying CMOS- compatible signals. Should you want to drive the board with TTL signals from a Centronics port, the stated CMOS ICs must be replaced by the suggested HCMOS versions. Construction Before embarking on the con- struction of the present board. IC. 555 or 7555 ICi 4069B or 74HCT04 IC. 4556B or 74HCT139 ICr. . .IC 10 incl. -40174B or 74HCT174 IC.. IC.* incl. -ZN436 or busconnector (It required! Heatsink for ICt;ICr as requirec PCB Type 87003 (see Readers Services! the type and the number of economic Type L293E. which motor current is fully program- D« and ICi. then fit a wire link stepper motors must be con- can be fitted in the 1C and mable, but in order to attain op- in the holes provided for the sidered in order to be able to ICt' positions on the PCB. In timum resolution in the micro- two outer pins of the regulator, decide on the most favourable most cases, the copper surface step mode, the maximum value As to the D/A converters, there as well as the most economical soldered to pins 5,6. 15 and 16 • ' of Is must be defined by means are a number of types to choose realization of the circuit. of these chips provides suf- of selecting appropriate re- j from. In principle, the Type To begin with, there are the ficient cooling, while ICj is best sistors in the Rt and Rs, as well ZN436E gives satisfactory per- L293E and the L298 to choose fitted with an insulated, stan- as in the Rat . . . Rr; positions ! formance for most applications, between. The latter should be dard U-shaped vane radiator, -consult Table 2. As Isimaxi is ; Note, however, that it comes used with currents in in excess ; Should you decide to use a L298 also related to the self-induct- ; without an internal reference, of 1 A per phase. Two L298s can for two stator windings, and a ance of the windings, it is ad- so that Di; (Dw 1 ) must be fitted, be bolted onto a common heat- L293E for the other two, do not visable to actually measure the and R« must be a 1K2 type, sink, together with regulator forget to limit the input voltage current consumption of the while Rs must be omitted ICi. As all conductive surfaces in accordance with the maxi- motor. —consult Table 3. Jumpers c of these ICs are at ground po- ! mum specification of the latter. The +5 V suppiy rail is made and d are not used, and jumper tential, there is no need for in- Depending on the type of out- available at a separate pin of the e is fitted to pass the refer- sulating washers and the like, put driver fitted, dimension R- I/O connector. When feeding ence voltage to the REF IN pins Relatively low ouput currents as per Table 1 the stepper control board from of IC» and IC14. The Type can be handled by the more As already stated, the stepper ! an external 5 V supply, omit R., i ZN426-X (the suffix indicates the Fig. 6. Track layout and component mounting plan for the motor control board. 2-36 eletoor tndia february 1987 Fig 7 Where micro-step operation is not required, each of the DACs in the circuit may be replaced by this resistor combination. number of bits: 6, 7, or 8) is also usable but is expected to be somewhat more expensive, as it holds an internal reference circuit, which can be used by fitting jumper c or d, depend- ing on the position of the DAC on the board, and using a 390R resistor in the R« or R* position, whichever is appro- priate. Should you want to do without the micro-step facility altogether, mount two 10K re- sistors as shown in Fig. 7. Com- pleting the stepper motor con- trol board is very straightfor- ward indeed when using ready- made, through-plated PCB Type 87003 (see Fig. 6) available from our Readers Services. When using the L293E driver chip, solder it straight onto the board to effect sufficient cooling by the large copper surfaces at the track side of the PCB. Connections In general, the connection of bipolar stepper motors is fairly simple. A two-phase motor re- quires to be driven with one half of the control board cir- cuitry. The actual connection of the stator windings is largely uncritical. Reversing the po- larity of one stator winding, or interchanging both windings simply causes the motor to run in reverse. A bipolar four-phase motor requires to be driven with the whole of the control board. When using such a motor, observe the correct phase relationship between the stator windings, else the spindle will merely oscillate be- tween two positions, rather than revolve. Basically, unipolar motors can be connected in three ways, as shown in Fig. 8. The first method, shown in Fig. 8a, re- quires passing less than normal current through the series con- nected windings to preclude overheating and/or saturation effects in the stator. Also, the increased stator inductance causes a considerably lower pull-in rate. The second method involves creating a centre-tapped wind- ing-see Fig. 8b. In principle, this arrangement always results ir. one half of the winding being short-circuited to the positive supply rail. As compared with the above method, there is the advantage of the lower overall inductance, but the short-cir- cuited half-winding gives rise to an increased motor dissipation, owing to the inevitably high in- duced current, which is only advantageous m that it ensures good damping characteristics and hence a relatively smooth spindle movement. The last alternative is shown in Fig. 8c This method of connect- ing a unipolar motor is based upon the use of the individual windings as if these were of the bipolar type. In case the two windings of a stator are not con- nected internal to the motor, anti-parallel connection is pref- erable. A normal, parallel con- nection immediately results in the magnetic fields counterac- ting, causing the spindle to re- main stalled. Provision has been made on the PCB to fit a 64-way, a & c row busconnector. Kj. Its connec- tions are left vacant to enable users to configure the bus wir- ing as required. At the other side of the board is K<. a 20-way angled plug which is used for the Centronics signals. De- pending on the set-up of the computer system in which the present board is to be incor- porated, wires may have to be run from Ki to K>, or Ki may be used for mechanical support only. Those users intending to make a stand-alone peripheral device of the stepper motor control may want to cut off the PCB section provided for Kj altogether. The power supply As already stated, the present board is rather uncritical of its input supply voltage. Extensive regulation and smoothing of the 12 ... 35 (45) V input rail is not recommended in view of the overall system efficiency. When designing the power supply in question, merely observe that the ripple voltage does not ex- ceed 10 to 15% of the output voltage. It must be reiterated that the maximum permissible peak in- put voltage for the baord de- pends on the type of bridge driver IC fitted; for the L298, V,n = 45V P oak; for the L293E, Vin=36 Vpeak. In practice, it is recommended to keep the in- put voltage a few volts below these values to allow for the in- duced peaks caused by the free-wheeling current. A second factor to be con- sidered in the establishing of the supply voltage is the ohmic resistance of the stator wind- ings in the stepper motor. As a rule of thumb, the supply volt- age for the board must be at least two times the typical op- erating voltage of the motor op- erated ' with voltage drive. In principle, therefore, most com- monly available 5 V stepper elcktoi india (ebruaiy 1987 2-37 | motors should work all right I with a board supply of 10-12 V, [ but a higher supply is prefer- | able for improved current drive I characteristics and hence a | higher pull-in rate. { The total current consumption | of the system goes mainly on | account of the stepper motor(s). ! Due account should be taken of j the fact that the total current ; drain may amount to 8 A when | using the board to drive 4 off 2 A stator windings. Obviously, 1 the mains supply should be de- j signed to reliably cater for j possibly high current peaks, and the same goes for the supply wiring. Also observe the 2 times 4 contacts on K<, re- j served for the connection of the input supply; keep the total current drain in mind and, if necessary, use soldering pins to avoid overloading the rela- tively thin connecting posts in Kj. Driving stepper motors As the stepper motor control board is essentially only a per- ipheral device, the computer- or more precisely the soft- ware— determines the move- ments of the stepper motor spindle. The key to the driving of the motor(s) is the 8-bit control word sent to the board via the computer’s parallel output port. Fig. 9 shows the bit assignment for that control word. The two MS bits-D« and D;— are used to address one of four stator driver circuits. Bit Ds provides the polarization control, while Do- Da determine the stator current in 32 (2 5 ) increments. Note that some Centronics output ports are open-collector types, re- quiring the data input lines and the STB line to be pulled high to + 5 V with 470R-1K0 resistors. Quite essential to the operation of the stepper motor is the stator current timing sequence. Fig. 10a shows the timing for full step operation, in which the stator current is arranged to re- verse with every step. Semi-step operation is illustrated in Fig. 10b; during the reversal of the stator current, this is held at nought. This basic method is further exploited in the quarter- step mode shown in Fig. 10c, while extrapolation of this prin- ciple leads to the stator current being reversed linear with time. as shown in Fig. lOd. In prac- I tice, however, the linear com- | mutation is slightly problematic, since the sub-steps at the cur- rent cross-over point are in- evitably larger than those dur- ing the start and the end of the commutation cycle. Moreover, the available torque will vary considerably during the sub- steps, as the total stator current is not constant. During the current reversal, a permanent load fitted to the spindle will cause the rotor to deviate more from the object j position than during moments of maximum current, resulting in irregularity of the sub-step size. This effect is generally found to be rather more mani- fest with dual-stator motors than with four-stator types. Up to and including quarter-step oper- ation, dual-stator motors have an adequate performance, but four-stator types are clearly to be preferred for all applications mentioned so far. The reason for this is the more constant j average stator current of the lat- ter motors. In conclusion, dual- i stator motors are best operated j with a constant total stator cur- rent, as shown in Fig. lOe. The commutation characteristic required for equal step size is mainly determined by the specific type of motor to hand, and some trial-and-error pro- gramming may be required to attain optimum performance. j Sending bits to the board The simplest method of driving the stepper motor is probably the writing of a array which holds all data for a full com- mutation cycle. Such a cycle essentially involves once re- versing the current, and revers- ing it again to return to the original polarity. In a four-stator motor, this corresponds to 8 full steps. A programmed pointer is used to send the datawords to the board, and can be read, in- cremented or decremented to control the direction of the spindle rotation. To get the motor to run as required, the pointer is programmed to ad- dress the individual array en- tries in a closed loop. Table 5a is a data dump of an ar- ray to control a four-stator motor j according to the timing dia- I gram of Fig. lOd. Note es- ' pecially the toggling of the 8>885288 | SS>SS2S8 | data M 1 data IF ID 80 3F 3D IE 19 82 3E 39 17 15 84 37 35 S 13 33 31 T OF 0D 88 2F 2D A OB 8A 2B 29 T 07 05 8C 27 25 O 03 8E 23 21 21 23 01 03 25 27 92 05 07 29 2B 94 09 0B 2D 2F 96 31 33 98 1 1 35 37 9A 15 17 39 3B 9C 19 IB 3D 3F 9E ID IF 5F 50 A0 7F 7D 5B 59 7B 79 57 55 A4 77 75 s 53 51 Ab 73 71 T 4D A8 6F 6D A 4B 49 6B 69 T 47 45 AC 65 o 43 41 AE 63 61 R 61 B0 43 65 B2 4b 47 2 bB B4 49 4B 6D 6F B6 4D 4F /I 73 B8 51 53 75 77 BA 55 57 79 7B BC 59 7D 7F BE 5D 5F 9F 9D CO BF BD 9B C2 BB B9 97 95 C4 B7 B5 S 93 91 C6 B3 B1 T 8F 8D C8 AF AD A 8B 89 CA AB T 87 85 CC A7 A5 0 83 81 CE A3 A1 A3 DO 81 83 A5 A7 D2 85 87 A9 AB D4 89 8B AD AF 8D 8F B1 B3 D8 91 93 B5 B7 DA 95 97 B9 BB DC 99 9B BD BF DE 9D 9F DF DD E0 FF FD DB D9 E2 : B F9 D7 D5 E4 =7 F5 s D3 D1 E6 =3 CF CD E8 EF ED V CB L9 EB E9 r C7 C5 :C E7 E5 > C3 Cl EE E3 El i El E3 : 0 :i C3 E5 E7 ■2 :5 C7 i E9 EB : 4 :9 CB ED EF ■6 :d CF FI F3 ; 8 )i D3 F5 F7 )5 F9 FB C 19 DB FL> ff LFE_! DD DF ion cycle. Table 5a is for a four-stator motor operating or type operating as per Fig. lOe. stator address bits and the cur- rent polarity bit. Table 5b is a similar dump intended as a guide in controlling a dual- stator motor according to the timing diagram of Fig. lOe. For both applications it it advisable to provide for an interrupt- based synchronization facility, as offered by, for instance, the Type 6S22 VIA. Unfortunately, the fairly large number of sub-steps often makes it impossible for the motor to attain its maximum speed. In this context, there is no doubt about the advantage of machine language subrou- tines over BASIC programs. Should the need arise to have the motor run at a relatively high speed, it is possible to program for more than one step at a time. At high switching fre- quencies, the stator inductance limits the current to such an ex- tent, that accurate current drive, and hence micro-stepping, is unattainable anyhow, However this is of little consequence, since the motor will nonethe- less run smoothly with the step rate well in excess of the res- onance frequency. Micro-step- ping is, therefore, primarily of use either for relatively low motor speeds, or for accurate spindle positioning. When skipping array entries to realize sufficient motor speed, care should be taken to finish with the last byte of the rel- evant stator phase. Large steps should, therefore, always com- prise sub-steps which are powers of two (2, 4, 8. 16 or 32 steps at a time). TW 2-39 DIGITAL SIGNAL PROCESSING Compact disc players have been with us for some time. Digital television receivers are becoming commonplace. These, and other apparatus, have an important aspect in common: digital signal processing. But what is really involved in this? Digital circuits only respond to discrete values of input voltage and produce discrete values of output voltage. Usually, these circuits operate between two discrete voltage levels, ie„ high and low (logic) levels. It is therefore clear that before such a circuit can operate the analogue signals have to be converted into digital (= binary) signals. Some fundamentals Fig. 1 shows the basic set-up of a digital processing circuit. The incoming analogue signals at X are digitized, in an analogue-to- digital (A-D) converter, pro- cessed in a (digital) signal pro- cessor, and then reconverted into analogue signals in a D-A circuit. The A-D converter produces a stream of binary values by quantization. In this method, the incoming waveform is divided into a finite number of subranges each of which is represented by an assigned binary value within the subrange. In a compact disc player, a 16-bit A-D converter is perfectly adequate, while in video circuits 8-bit converters are satisfactory. Since the signal processor operates by computation, it can handle only a finite number of pulses in unit time. It is the task of the A-D converter to ensure that the input capacity of the processor is not exceeded, and I this in turn determines the sam- pling rate. Sampling is a technique in which only some portions of the (analogue) input are used to produce the set of binary values to represent the infor- mation contained in the whole signal. To ensure that the output values represent the input signal without significant loss of information, Nyquist’s Sampling Theorem states that the rate of Signal processors As already mentioned, virtually all requirements are met by the basic operations of multipli- cation and addition. Also, it was shown that the signal processor does not have all that much time left for each computation. Sig- nal processors have, therefore, microprocessors with typical instruction codes: they are rela- tively small but, none the less, quite fast. Sequences such as: "fetch value 1; fetch value 2; multiply values 1 and 2; add value 1 to the result; load the ac- cumulator at the position of value 1 and increase the ad- dress counter" as a rule have only one oper- ational code. Moreover, while an instruction is being pro- cessed, the next instruction and the next two values are retracted from the memory (pipelining). This means that such an instruction takes three clock pulses from start to finish. With a 10 MHz clock, a 16-bit multiplication and addition lasts only 300 ns. Even faster are signal pro- cessors that use the Harvard in- stead of the von Neumann architecture. In the latter, data and instructions are stored in a common memory, whereas in the former separate memories are used (see Fig. 2). In Harvard-type processors, in- structions and data (in some even two sets of 16-bit data) are fetched from the memory simul- taneously. This means that two to three times as many oper- ations can be carried out per second as compared with a von Neumann device. The software for the required function is first computed and loaded into a normal computer, with which the run of the pro- cessing cycle is simulated before the PROM of the signal processor is loaded. To conclude, and specially for Fig. 3. Basic recursive filter. Output signal yltl is stored in intermediate memory and used as input signal ylt-1) for t next computing cycle. 4. Basic non-recursive filter. Output signal yltl is buil im a succession of inputs: xlt). . .y(t-n). Secondary m< s required for each of the inputs. Dim X1%(64l Dim Y1%(64) Dim Z1%<64) For 1% = 1 To 64 X1%(l%) = l%*10-5 Next 1% Polyline 64,X1*.0,Y1%() Offset O.V% Do Mouse X%. Y%. T% If T% = 1 If Tl% = 1 X% = lnt(X% / 10) + 1 Y% = lnt(Y%/10)*10 + 5 Y1%(X%) = Y%-V% Polyline 64,X1%(|,Y1%() Offset 0,V% Else CIS Input "Select filter order: For I ^ 1 To Ord For 1% * 1 To 64 Z1%II%I = B*Z1%(I%U II -BI*Z1*>(I%-1) Next 1% Polyline 64.X1%(I.Y1%<) Offset Alert 1, "Change filter?'', 2, "New filter; end". Z BASIC for an RC low-pass filter of the n ,h order and it: graphical representation. (Fig. 5). those readers who want to design a digital filter and are not too familiar with Fourier or filter theory, a sample design for a personal computer. Basically, there are two types of filter: recursive and non- recursive. Figure 3 shows an example of the simplest type of recursive filter, where the out- put signal is available for further use a computation cycle T later. This type of filter can be used | for high- or low-pass purposes. Non-recursive filters are formed by inserting the input signal(s) into two or more successive filter sections as shown in Fig. 4. Each section must, of course, have a secondary memory. This type of filter is suitable for use as a transverse or comb filter. A low-pass filter is easily com- ] puted from the following | mathematical relation yft)=ax(t)+by(t-\) where yfl) is the output signal I resulting from an input signal x ' y(t- 1) is the output signal one computing cycle before /(/); A=l— a; 0 until the direct voltage at the output of opamp A2 (pin 6) is 0.00 V. Next, inject a 1 kHz square-wave signal into the drive unit, and set the input sensitivity (lines Vo-Vi) to 0000 Adjust trimmers On and Cjo to obtain a true square-wave signal at the out- put (pin 6) of Aj. Set the input sensitivity (lines Vo-Vi) to 0001 and adjust Cn to regain a proper square wave at pin 6 of As. Repeat this procedure with sen- sitivities of 0010 0100 and 1000 when C24, C?s and Cj«, and Cu and C2« re- spectively are adjusted. At all times, adapt the level of the square-wave input but take care to avoid overloading the circuit (the level at pin 6 of A? should not exceed 5 Vpp). Redo all the adjustments mentioned to make sure that all settings are correct. With the aid of a voltage divider (made from 1% resistors: e.g. 22 0 and 2k7, or 82 8 and 10k) derive a voltage of 40.0 mV from the power supply, and apply this to the junction R23-R24. Set the signal on lines OFo-OFe to Software First, the PIA (if used) is initialized. Make the RESET line low, which results in all the PIA registers to be set to nought. The adapter occupies four addresses: I/O to 1/0+3 incl. (see Fig. 14). Two of the locations have consecutive registers, and these are selected by making bit b2 :n the associated control register 1 (data register) or 0 (data direction register). Select DDRA as shown above and write FF in this register: all A ports are then set as outputs. Then write 06 in CRA which results in input CAi reacting to a leading edge, as well as data register DRA being selected. It is then possible to write into this register, for instance, 10 which pulls the PA« line high. The B pons are arranged as outputs by making control register B logic low, and writing FF in DDRB. They are set as inputs by making bit b; in CRB 0, and writing a 0 into 1/0+2 and a 4 into 1/0 + 3. Arrange the A and B pons as out- puts; disable the interrupt; and set the interrupt flag (bit 7 of data register A) to a leading edge at CAi. A timing diagram of all important control signals is given in Fig. 15: this gives a good idea how communi- cation between drive unit and com- puter takes place. All PA lines are made 0, after which the data for setting the interface can be written into the latches via the PB 50 ?ddrb = 6FF 60 ?drb = oFF + 64+ 128'ING:?dra = &14 70 ?drb = NIV a 64 - 128"TH:?dra = &12 80 ?drb -TB + 16’AM:?dra = &11 90 ?ddrb = 0:?dra = 0:?dra = &40:?dra = &10 100 HOLD = TIME + (T8 + iriO:REPEATUNTILTIME>HOLD 110 IFTRIG = 0THEN?dra = &30 120 IFTRIG = 1THEN?dra = &38 130 IFTRIG*2THEN140ELSEIFINKEY-99THEN?dra = &90ELSE130 140 REPEATUNTIL?ifr< >0 160 FORI -0TO255:PLOT69,2*I,4’ ■ 170 'dra -- &20 180 FORI 256T0511:PL0T69.2'I .4 H 190 ?dra = &10 200 END &40:?dra = O:NEXT a &60:?dra - &20:N£XT 10 MODEO [ 20 dra = &FCB0:ddra dratcra &FCBl:drb &FC62:ddrb drb:crb &FCB3 I 30 ?cra = 0:?ddra - &FF:?cra - 6:?dra &10 40 oFF : 0: ING 0NIV - 0:TH 0 TB 8: AM 10:TRIG - 0 50 ?crb-0:?ddrb &FF:?crb 4 60 ?drb = oFF * 64 + 128*ING:?dra &14 70 ?drb = NIV * 64 * 128'TH;’dra = 612 80 ?dlto=TB + 16‘AM;?dra = &11 90 ?crb = 0:?ddrb = 0:’crb - 4.?dra = 0:’dra = 640:?dia = 610 100 HOLD - TIME + (TB + If 10:REPEATUNTILTIME>HOLD j 110 IFTRIG - 0THEN?dra - 630 120 IFTRIG - ITHEN?dia - 638 130 IFTRIG«THEN140ELSEIFINKEY 99THEN'dta &90ELSE130 140 REPEATUNTIL?cra>127 150 P = ?dra:?dra = 0:’dra = &20:?dra 0 160 FORI = 0TO255:PLOT69.2'l,4*?drb:?dra = 640:?dra - 0:NEXT 170 ?dra = 620 180 FORI =256T0511:PL0T69.2*l.4'?dib:?dra - 660:?dia = 620:NEXT 190 ?dra = &10 200 END Fig. 12. The printed-circuit board for the PIA Ct - 100 n 1C. = 6821 PCB 86100 For Electron: le 74LS04 ie 74LS133 Table 1. Test program for the BBC Micro. Table 2. Test program for the Acorn Electron. ports. These data relate to the time base; the off-set; the trigger level; leading- or trailing-edge triggering; selection of input sensitivity; and selection of AC or DC inputs. Note that PAo to PAj incl. are used here as clock signals. See also under Control signals in Part 1. Tables 3-8 show the correlation between data and selec- ted settings. The PB ports axe then set as inputs; PA* is made logic 0; and PA* is briefly made logic 1. This results in the off-set data in the D-A converter being read. Next, make the PA* line high, which creates a waiting period of at least 256 times the selected time base. This ensures that the first memory page no longer contains old data. Make the PAs line (INH) logic high, which results in the digitized input signal being compared with the set trigger level. As soon as these levels are equal, the highest data bit in the RAMs is made 1 (which makes it possible later to determine exactly where triggering took place); the RAM counter is reset; writing is discontinued; and the circuit pulls the READY line (CAt) high to indi- cate to the computer that the two RAM pages are full. The computer then makes lines PA* and PAs logic low, which results in the READY line being pulled low. The computer can then read the RAMs. First, however, the PAs line is briefly made 1 to reset the RAM counter to nought, so that the first memory lo- cation can be read immediately. After this, CPUL pulses on PA« enable the data of successive ad- dresses to be read at each leading CPUL edge. After the first memory page (255 bytes) has been read, make PAs (INH) high: this serves as the eighth address bit for the memory. Subsequently, the second page of 256 bits is read in a similar manner. All data can be stored or processed immediately, depending upon the available memory. Finally, new data may be written (with the PB lines arranged as in- puts). A pulse on the PA6 will cause the off-set data in the D-A converter to be clocked. Making the PA* line high will cause the PIA to start again with writing into the first memory page. After an interval of not less than 256 time-base clock pulses, the trigger may be enabled again. As stated in Part 1, complete pro- grams for the Acorn Electron, the BBC Micro, the Commodore C64, and MSX machines are supplied with printed-circuit board 86083. To enable owners of other makes of computer to compile their own pro- gram, a flow diagram of the program T6 TS T4 T3 T2 T1 OF6 OF5 OF4 OF3 OF2 OF1 OFO Table 3. Corre- lation between time bases and data on TBo-TBa lines. Table 4. Corre- lation between sensitivity set- tings. maximum input voltage and data on Vo-Va lines. Table 5. Corre- lation between trigger levels and data on To-Ts lines. Table 6. Corre- lation between off-set voltages and data on OFo- OFt lines. Table 7. Corre- lation between type of input and data on AC /DC line. Table 8. Corre- lation between type of trig- gering and data on +/— line. is given in Fig. 16. The quality of the screen image will depend largely on the resolution of the computer. First of all, the location on the first memory page where the trigger bit (D7b) went high must be determined. The next location is the first for a complete picture, from which the whole page may be read. The sec- ond page can be started at the first location, since all data there are in correct sequence. ■2-57 selex-20 Linear Scale Ohmmeter Why should anyone construct an ohmmeter, when every multimeter has several resistance ranges? True, the multimeter has higher end of the resistance scale reads the values very poorly. Due to very closely spaced while are widely spread. ohmmeter, the divisions are equispaced over the entire I scale. This advantage is due to the small I.C. OpAmp The Circuit A simplified schematic diagram of the linear scale ohmmeter is shown in figure 2 The actual circuit is shown in figure 3, which looks much more complex than it really is. First, let us concentrate on the circuit shown in figure 2. The main component of the circuit is the Op Amp IC1. IC1 contains a multistage differential amplifier circuit. A differential amplifier amplifies the voltage difference between its two inputs. The voltage on the non-inverting input (*) increases the output voltage, whereas a voltage on the inverting input (-) reduces the output voltage. The gain of such amplifiers is a few hundred thousands. Gain of one hundred thousand means that a difference of 10 microvolts Volt at the output. A voltage divider made of Rx and R is connected across the output and the voltage at the interconnection of Rx and R is fed back to the amplifier at its inverting input. This is feedback. This has an effect on the circuit which makes the voltage on the inverting input practically equal to that on the non-inverting input. To understand exactly what happens, let us 2-58 selex from 5.6 fo 6.6 V. i.e. 1 V. The output voltage will try to increase by 100000 V. The voltage on the inverting simultaneously try to rise depending on the ratio of Rx and R. This in turn will try to bring down the output voltage. The result of this will be that the voltage on the inverting input will also rise to almost the same voltage which is on the non-inverting input. In case of the linear scale ohmmeter circuit, the input voltage on the non-inverting input remains constant at 5.6 V. The voltage across R is thus fixed at 5.6 V as we above experiment. This means that the output voltage at the output of the amplifier depends entirely on the value of Rx. The relation between these values can be calculated as follows: Voltage on R = 5.6V Voltage on Rx and R = Uout ^»5.6V Which clearly shows that Uout is directly proportional to Rx if the constant value of 5.6V is taken care of during calibration with Rx = Oil To take care of this, the meter is placed on the non-inverting input in the actual circuit, so that the voltage of 5.6V does not play any part in the measurement. The zener diode produces the stable input voltage current for D1 is supplied by R5. The output voltage is measured through the combination R6 - PI - Ml. Diode 02 protects the meter Ml from very high voltages, which can occur when the ohmmeter is connected without a test resistance. Construction As usual, the mechanical work involved in construction is much more compared to the job of soldering the electronic components together on the PCB. The mechanical work can be simplified by using a plastic enclosure, which is easier to handle than metal enclosures. Suitable holes must be drilled in the lid for sockets, switches SI, S2 and the meter Ml. A large cutout of 50 mm diameter must be carefully made for the meter body. '2-59 A standard Selex PCB can accomodate all the circuit component. Layout and wiring is quite simple and is shown in figure 4. Pin | D1. D2 must be properly observed. Three different Op Amps | have been specified in the | component list for IC1 . These are all pin compatible. The commonly used Op Amp 741 will not work in this circuit. The resistances used must be of very close tolerances, typically 2.5% or less, for R1 to R4. This ensures that the scale is unformly divided. The construction details are shown in figure 5. The Selex PCB is fixed on the bottom of the enclosure and the batteries are clamped using an aluminium clamp. After wiring and assembly, the potentiometer PI is adjusted such that the meter shows full scale deflection for a 1 K!! resistance in Range 4 (0 to 1 K). The test resistance of IK used here Table 1 shows the setting of switch S2 and the range covered by the setting. Also indicated is the current through the test resistance for each range. The scale of our prototype is suitable for range 2 (0-1 00K) For other ranges, the reading must be multiplied by 10 (Range 1 ), 0.1 (Range 3)0.01 (Range 4). selex The Cackling Generator Functional description The audio frequency noise I signal generated by the third oscillator can be adjusted by the potentiometer PI to set the desired sound level. cackling quality, four short and then a long cackling noise with rising sound level must be generated. This is achieved by the connection of first two oscillator via a RC network | consisting of R4. R8. RIO, j | C3, C4 and D2. D3, D4. I The sound of cackle can be j modified by changing C7 by i i two cackling cycles. the second oscillator (using N3 & N4| provides the envelopes for the four different cackling sounds full cackling cycle The selex charged. When RIO has more positve voltage than that on R4. 03 blocks and C4 is charged 04 prevents the voltage on R8 from becoming negative The audio signal finally amplified and fed lo the loudspeaker. Observe the shape of the transistor T1 ir the circuit diagram. Construction This circuit has many components to be accommodated and requires a double size SELEX PCB The layout is showsn in capacitors require the maximum PCB space. While soldering their polarity must be correctly observed The current consumption is between 5 to 1 5 mA, and a small 9V battery pack is adequate to supply this If one wants to combine this circuit with a kitchen timer, the relay contacts of the timer c.l?eulf can be suitably connected into this circuit starts when the set time If you want to pack this circuit nicely in shape of a hen as shown in the photograph (5). it should be assembled on two small SELEX PCBs and interconnections should be made with wires. The loudspeakers and battery can be filed as shown in the photograph. Potentiometer PI can be fitted in front as shown, so that sound level can be conveniently adjusted. 2-62 elettc. selex Electrical Power characterises the use or supply ol electricity. In the abbreviated form it is represented by the letter P and the units for measuring electrical power are Watts (W|. Th higher the power of a drilling machine, the generated by it. The higher the Wattage of a bulb, the brighter is its glow. A water heater with 3000 W rating gives more heat than a 1000 W heater. The higher the power of a stereo amplifier, the louder is the However, all the previous examples are not identical In case of the drilling specified power from mains [ supply when it is drilling a hole in a hard material. The power consumption is much running state This is not true in case of the bulb or tne water heater, because 'hey draw the specified as they are switched on. Also the example of the amplifier is still different, e power drawn by the POWER amplifier can be controlled externally by the setting of the volume control knob, between a minimum and maximum amount The specified power of the amplifier generally refers to the maximum power. The useful power is much less than the power drawn from mams When the amplifier draws 30 W from the mains, it does not supply 30 W to the loudspeakers. Even the bulb with 100 W rating does not convert all the 1 00 W of power into light, most a part of it is given as light There can be two meanings to the power specification of any electrical appliance It can be the actual power drawn by the appliance from mains or it can be the maximum power the appliance is capable of drawing from the mains supply Another distinction W from vei. only a few en out as light lining power is o heat So if lying the power the bulb to be jt the stereo kers It generally twice as much as ul Thus an which is capable >g 30 W from II not deliver more ut 16 W to the loudspeakers. converted from one form to power loss, what we really the power is not put to any useless form of energy, as in case of the bulb where converted into heat. However, if for any reason, we were using a 100 W bulb to heat something, then we would say that most of the 100 W of power is converted to useful heat and some part is lost as light! It all depends on which form of energy the appliance is expected to deliver The stereo amplifier draws electrical power at 50 Hz from the mains It gives out electrical power at the audio frequencies to the loudspeaker The loudspeaker in turn takes up the electrical power from amplifier and converts a part of it into sound energy, of the loudspeaker. Though 2 DJ selex Electronic Switch 4 Described here is the construction of a simple electronic switch which is electrically isolated from t electrically conductive voltage has no effect on the switching mechanism is directly placed in mplest way to connect and disconnect an appliance Irom the mains supply. However, the disadvantage of such type of switching is that the full supply voltage is always present on one terminal of the switch. This may not always be acceptable, especially in case of switching to be activated by sensitive circuits like computers. In such cases one can also relay, but the relay contacts can create problems when they get worn out. Even during normal operation, the closing and opening of relay contacts can produce electrical disturbances which may in turn affect the actuating circuits of the computer. The better way is to use an electronic switch similar to the one described here. Even though the practical circuit of an electronic switch used for controlling computer is not as simple as this, the principle remains same. What is described here is a simple battery operated version. The circuit still ensures full electrical isolation from I The Circuit | The circuit of the eletronic The Control Unit. a) 16-Bit microprocessor (INTEL 8086) b) Present on card memory of 1 6K (EPROM) and 8K (RAM c) Provides communication between MARS and user witf appropriate displays on monitor d) Easily expandable to control large number of MARS systems simultaneously. el Speed control by simple command from user f) Uses a * 1 2. -1 2 and *5 volt for motor and control care g) User has three modes of operation to choose 1 1 TEST mode 2i MANUAL mode 3) TEACH mode The Mechanical Unit An omnidirectional ground transporting robot on four wheels are powered by a pair of stepper motors. Each of Ibese motors are capable of independent motion thus very easily MARS can turn about any point The ARM unit is capable of handling loads of upto 500gm held at the gripper High degree of accuracy and low mechanical power input is achieved by the use of gears. ARM unit has 90 degree freedom of movement up and down making it capable of lifting objects from the floor level. The BASE unit has 360 degree freedom of movement. The very cost effective design of gripper achieves a high degree of compliance to suit any application. Driven by high speed DC motors this gripper makes hold' and release' action almost instantaneous. selex Construction As always, all the rules for construction of a circuit which connects to the mains, must be observed strictly. The circuit can be inside a suitable plastic enclosure Standard plug/ socket combinations connected to points 1 . 2 and 3 shown in the circuit diagram of figure 1 This gives us three alternatives for switching on the load. 1 Directly by switching SI 2. Through an external switch or relay contact qonneted across sockets 1 and 2 3. By applying an external voltage of 4.5 to 6V across sockets 2 and 3. Triac TIC 206M, or equivalent, can handle loads upto 200W Triac TIC 226M. loads upto 300 W An important point to remember here is that the lamp takes a little time to extinguish when disconnected from battery, and this will introduce a short delay between turning off switch SI and switching off the load from the mains. FLEXICELLS TO BEAT BATTERY WEIGHT by Dr Alan Hooper, Materials Developments Division, Harwell Engineers designing electrical and electronics equipment, from electric traction vehicles to port- able radios for domestic or military use, have always been frustrated by the weight and size of batteries that have to be carried. Now under development at Britain's largest laboratories, in collaboration with other scientists in the UK and in Denmark, all-solid-state rechargeable lithium batteries bring pollution-free driving a great deal nearer and may trigger many new and exciting ideas for battery-powered equipment. | Battery-powered electric vehicles (EVs) are already | in use in many countries. I One example, in the UK, is | the humble milkdelivery | wagon, or 'milk float' It is i successful because to do its job it needs to work | over only a short range and a low speed is acceptable in built-up I areas, where it has the added advantage over ! the internal combustion engine of not causing pollution. It is efficient and j convenient tor continual | stop-start operation and a | commercial fleet of such vehicles is easy to maintain. On the other hand, its restricted performance j causes considerable frustration to motorists who meet it on the open road, for it cannot travel at the speed of the rest of the traffic Across the Atlantic, the golf-cart would hardly be welcomed on the freeway. So the view of the general public is that electric vehicles have a poor performance but are I acceptable tor specialist ! duties. .' 66 r . ! 987 It is the source of power, the battery, which lies at the heart ot the problem. To put it simply, traction batteries are too heavy and too large for the amount of energy they store or the power they can provide: a large frac- tion ot the energy stored in a typical traction bat- tery is needed just to pro- pel the battery itself Aqueous electrolytes For practical purposes, the present choice of batteries for EV traction is between two systems, each employ- ing an aqueous elec- trolyte. which is either leadacid or nickel'iron. This situation has remained essentially unchanged since the beginning of the 20th cen- tury despite many attempts, especially over the last 25 years, to develop new systems. Over that period, stimulation by the appearance of poten- tial rivals has led to signifi- cant improvements in the performance of existing > systems and of vehicles with good short-range, traffic-compatible i capabilities. Most of the vehicles now available are urban delivery vans but one ot the latest is a version of the popular Peugeot 205 car. powered by a nickel'iron battery. There are certain practical drawbacks specific to individual systems, but the main, general problem is still that of limited range I EVs are still, in general, economically uncompeti- tive with their internal combustion engined counterparts. The performance offered by the enormous energy density of petroleum, with more than 10 000 Wh kg (watt hours per kilogram) compared with 20-40 Wh kg for leadacid ! traction batteries and a i highrate recharge j capability (two minutes at the pump in contrast to a battery charge ot several . hours), will never be j matched by that ot any I battery system, in spite ot I an on-board energy con- version efficiency that Is | five times better. However, i if a battery were available 1 ! with high energy density I (100 to 200 Wh'kg) it would j significantly affect the j practical value of EVs in a \ wide variety of appli- | cations from wheelchairs and bicycles to commuter I I cars, taxis and delivery I vehicles. 1 Not only would longer i ranges and greater load- | carrying capabilities be 1 realised, but the improve- ments in gravimetric en- ergy density would open j j up considerable scope for ; innovative engineering in | vehicle design, using lighter and cheaper materials. It is this, rather I than cheaper batteries, which would lead to a j cost-competitive electric | vehicle. Portable electronics Similar problems are to be found in other technologi- j cally important areas. The vast demand for portable ' I electronics equipment in ! the computing and com- munications fields bring with it a need for small, lightweight, rechargeable power sources. Both the business executive and the infantryman in the field would benefit from a lighter load to carry. It is not only important to achieve lower absolute weights and smaller volumes, to avoid the hand-held cellular radio- telephone or 'wrist-watch' device having a suitcase- size battery, but to be able to provide batteries that are suitably shaped, too. For example, a flat-screen television ideally requires a flat battery pack. There are also growing markets in the telecom- munications and other industries tor standby power sources. Here, too, there is a trend towards smaller electronics packages and corre- spondingly small power sources. NiCd batteries have been used traditionally in these markets and. more re- cently, NifHi batteries too for space applications such as power sources for satellites where cycle life and reliability are also ot prime importance; but the low energy densities so far achieved have restricted the electrical load capabilities ot missions. Space stations and deep space probes will require power sources with higher energy densities. Much better energy den- sities are theoretically available from alkali-metal couples, but materials problems have restricted their use mainly to pri- mary battery systems and to secondary batteries operating at high tem- perature. Of the latter, the sodiumsulphur battery is the best developed. It uses an Na‘-ion conduc- ting solid, sodium-beta- alumina, as a solid elec- trolyte and has to be operated at 350°C Predicted energy densities are more than 100 Wh/kg; more prototype traction batteries have been made and vehicle demon- strations carried out in several countries. However, sodiumsulphur batteries are still not com- mercially available, even after some 17 years' research and develop- ment by large teams of scientists around the world. Remaining prob- lems include the repro- ducibility of manufacture and reliability in use of beta-alumina ceramic tubes, and the thermal control and safety ot large batteries High tempera- ture systems of this kind will, even it successful, be useful only where large batteries are needed I A small, room-temperoture. . rechargeable lithium bat- l tery with a liquid organic electrolyte has recently , become commercially j available in Canada. Its ! cathode material (M 0 S 2 ) I leads to a low open- circuit voltage and I moderate energy density, i A useful life of more than 100 charge discharge cycles is quoted but little information is yet available from field trials j Applications under con- j sideration include photographic flashguns and electric wheelchairs j Radical departure Rechargeable all-solid- 1 state lithium batteries now | being developed at Harwell constitute what is j perhaps the most radical ' new departure in battery technology tor decades. They also promise very exciting commercial pro- spects Based on thick-film polymer technology, with no liquid components, they offer very high energy density, mechanical flexi- bility and variable geometry as well as being j robust and safe. This work has evolved from j a programme begun here in 1978 to investigate materials for advanced alkali-metal rechargeable batteries It was shared between Harwell, univer- sities in the UK and research and develop- ment establishments in Denmark. The Anglo- Danish Battery Programme as it became known was jointly sponsored at Harwell by the UK Depart- I ment of Trade and Industry (DTI) and the European Community. The aim of the programme was to examine the pro- | perties and behaviour of 1 several promising solid electrolytes and electrode materials described in the literature, to obtain a sound idea of their pro- perties. to define the prob- lems to do with their use in batteries and to assess their compatibility with other materials in cells. Such work would enable us to find out reliably which materials might be j technologically useful for electric vehicle batteries in the future. It was hoped j to obtain a fairly hard- headed assessment of | whether alkali metal bat- teries could be developed | thal would achieve their potential energy density | advantages and to ident- ify which materials could best be chosen for future cell development studies. A working temperature range of 100°C to 200°C was considered accep- table for a first generation EV battery. All-solid-state Because of persisting diffi- culties with organic liquid electrolyte batteries, all- | solid-state cells were seen as the only practical way forward tor operation at ambient and moderate temperatures. The cells developed in the pro- gramme have lithium anodes and a so-called intercalation or insertion compound as reversible cathodes. Examples are V 6 0.3 and TiSz. Although the early stages of the programme studied ,‘l«nd,a tetwusiv 1987 2-67 in depth the very interesting crystalline inorganic lithium-ion- conducting electrolytes LiiN and Lil(AlzOa), the choice of this type of cell was made more realistic by the discovery ot polymer-based solid elec- trolytes by Michel Armand and fellow workers in France. Certain polar organic materials such as polyethylene oxide) will dissolve alkali metal salts and manifest rapid olkali- ion conductivity. The absolute conduc- tivities of such polymer- based materials are not in general as high as those of crystalline solid elec- trolytes, but they may be made into thin, pinhole- free plastic sheets with good enough conduc- tance for use in cells and batteries. Equally import- ant is that the plasticity of the polymers overcomes the other big problem of solid-state battery systems, namely how to maintain good contact between faces Harwell staff have concen- trated over the last four years on developing the technology for making the polymer-electrolyte plastic battery* and have built and tested cells. Tech- niques for continuous pro- duction of the electrolyte and cathode components in the form of thin films have been developed and their dimensions can be scaled-up when required. The thickness of iplete ceil is only including printed and packaging materials and photographic film. It has been shown that in laboratory-scale cells, operating at around 120°G there is a high utilization of the active cell materials at dis- charge rates of a few hours and with lives of over 100 deep discharge cycles. Larger cells, of up to 500 cm 2 area, and series-connecled multi-cell stacks have also been successfully mode and tested. From these results we predict usable-energy densities tor solid-stale traction batteries that would make them one-fifth of the weight and one- third of the size of lead acid batteries now in service Temperature range At present the cells, which are polyfethylene oxide)- based, operate most effec- tively at 100°C or just above, so they are quite suitable in that respect for vehicle traction service and for use in satellites. Earliest specialist appli- cations may also be found where the environ- ment is hostile with tem- peratures of up to 150°C, a region where most con- ventional batteries fail. They may include down- hole instrumentation in the oil industry and certain standby power sources Furthermore, lower- temperature performance can be achieved with existing materials and cells when the power requirements are low, as tor many micro-electronics jobs. One attractive possibility in this field is the inte- gration of the battery with the circuit it powers: the thin-film planar tech- nology is compatible with conventional printed cir- cuit board and hybrid electronic circuitry. For example, the technology lends itself to the develop- ment of a self-powered intelligent credit card incorporating a micropro- But for many other pro- spective uses, operation af room temperature and below is required, at high power levels. This will mean developing new cell materials, especially new polymer electrolytes. Work is now going on in many countries and a research and develop- ment programme here is j being sponsored by an industrial group or 'club' of j battery users, manufac- turers and materials specialists. Supported by the DTI, our Solid-State Bat- tery Working Parly aims to provide the basic tech- nology to make all-solid- state lithium batteries, based on polymeric elec- trolytes, for as many ap- plications as possible. Studies will concentrate at first on developing better electrolytes but expand as membership of the group grows Success in this area will open up many new uses in the military, industrial and domestic sectors. It might well lead to cord- less' vacuum cleaners, lawnmowers and power tools, and to new flashlights, toys and elec- tronics and communi- cations equipment. The idea of batteries based on an all-solid-state ' polymer electrolyte, perhaps using various i materials and construction I technologies tor different applications, holds out one of the most versatile and exciting prospects for battery development this century. Always a move ahead Resistance Measurement ill For further details write to: r— THE MOTWANE I MANUFACTURING COMPANY MOTWANE R 61084 Telex 752-247 MMPL IN Grams: MOT- WANE or Gyan Ghar. Plot. 434 A. 14th Road. t Khar, Bombay-400 052. Grams: MOTESTEM L_ :w PRODUCTS • NEW PRODUCTS • NE\ IEC Strip Connectors are available in wide range, from 5 Amps to 30 Amps in 12 ways, moulded in Bakelite & P VC. The metal parts are made of brass and screws of M S. duly plated to prevent corrosion. The strip connectors are tested to withstand High Voltage for 2 Lock Switch provides add safety to electrical and electronic equipments an prevents unauthorised us Panel Projection 6 mm Action 3 Pcs SPOT Contact Resistance 20 Milliohm Rating 125 VAC 5A, 251 For further information, please contact: - ASIA ELECTRIC COMPANY Kataru Mansion 132 A, Dr A B Road. Worli Naha. Bombay 400 018 PLA introduces Series 101 Miniature relay in a slim style design with overall dimensions 26(1} x 1 2 5(W! x 24.5(H) mm Available with t changeover contact rated for 6 amps, at 240V Ac 28V DC It is ideal for high density PCB applications in the field of communication and Industrial Control Systems as well as house-hold electrical appliances Excel have come out with 'Henry Flat Cables'. Satisfying UL and CSA standards, these cables are generally available in 6 to 1 2 ways either in soft Copper alloy, with silver or gold plating and rated at 300V, 5A Capable of operation in a temperature range of -45°C to 100°C. the cables can be used as jumpers for interconnections in electronics instruments, communication equipments, business machines and computers BETA TESTER This Transistor Beta Tester measures static gain (Beta) upto 300, at collector currents upto 10 Amps and base currents upto 1 Amp, at VCE of 4 Volts, as per international specifications The currents are pulsed at 2% duty cycle at 50 For further inform .. ELCOM 103. J ay y opal Indt. M S SAI ELECTRONICS Thakor Estate Kurla Kirol Road Bombay 400 086 Phone. 5131219 5136601 For details contact M/S. EXCEL ELECTRICALS. C 4 Raj Mahal Apartment. Coves Road. Jogeshwari IE). Bombay -400 060 TEMPERATURE DATA LOGGER SCR's Temp Data Logger is a sw PRODUCTS » NEW PRODUCTS « NEW 1 — J DIGITAL MULTIMETER MECO has just introduced the nodt I Mil . ( 1 Digits Multimeter wlncti features a single knob operation for all functions. It measures AC & DC currents Tom 200 uA to 10 A with a min resolution of 0 I mA. AC voltage upto 750 V and DC voltage upto 1000V; Resistance from 200 ohms to 20 megohms; diode checks and continuity tests It has an accuracy of 0.5”'. • 1 dgt for DC Amps Volt & Resistance measurements and 1% • 3 dgt for AC Volt /Amp measurments. It temperature f Wli 1 & & - -- 1 .**' It is over load protected on all the ranges except the 10A range It has facilities for low battery indication and overload indications. It operates on a 9V cell with battery life in excess of 800 for further information. MECO INSTRUMENTS PVT LTD Bharat Industrial Estate. T.J. Road. Sewree. Bombay 400 OIS Phones: 413-7423. 413-2435. 413-0747. construction m coolants, v.brat AC and DC ver« for further inform, itn INDIAN ENGINEERING COMPANY Katara Mansion. PB NO 16551 rt •'! Bombay 400 018 POWER PUSH BUTTON SWITCH Rajkumar engineers offer a new power push button switch incorporating the latest advances in switching The switch is rated at 5 amperes continuous load and has double pole single throw switching configuration bimetal sandwitclied contacts A tor fail safe operations it preventing electrical cross | for further information I Rajkumar Engineers j 106. Ba/sons Ind Estate Chakafa Road. i Andheri lEastl | Bombay 400 099 Indio I DIGITAL FREQUENCY | COUNTER VDC18 VDC18 is the smallest sire ever made in India. Features include I BATTERY OPERATION cum I mains operation through [ adaptor. 7digit 0.5 inch LEO I display. 30 MHz frequency | range, light weight, resolution selection etc.etc. VDC18 incorporate latest L.S I circuitry Model VDC19 has frequency range upto 500 MHz and PERIOD. FREQUENCY for details contact: VASA VI ELEC TRONICS (Marketing Division J 630.Alkarim Trade Centre Rantganj Secunderabad 500 003 Phone 70995 PLASTIC INSTRUMENT BOX FOR BACK MOUNTING Comtech T-77' is an elegantly designed plastic moulded instrument box suitable for back mounted instruments such as Timers. & various other control instruments, having overall dimensions of 110 mm. I x 77mm W x 100 mm B It consists of a moulded box. a cover. & a M S plate for back mounting The box has an mm for various components. A six way terminal strip fixed at the top & bottom, infront of the box provides a n easy access for the terminals. The cover can accommodate a PCB of 77 min X 72 mm from inside & has a I 1 .2 mm deep recess in front to take an Aluminium plate of 65 mm x 66 mm. for control indications. The box offered in Black & Grey colour with most suitable for small instruments to be mounted side by side from the back, like eg. counters, controllers & timers PROXIMITY SWITCHES IEC offers a new line of I inductive Proximity switches which are basically contactless limit switches. In | addition, they feature for further Details contact ADVANCE INDUSTRIES 1 1 . finer ala Bldg Tribhuvan Road Near Dreamland Cinema Bombay 400 004. for further details contact COMPONENT TECHNIQUE 8, Orion Appartment 29 -A La/lubhai Park Road. Andheri /West) Bombay 400 058 2-72 R.N. No. 3988 1/83 MH/BY WEST -228 LIC No 91 Don’t miss the BUS More and more industries are catching the Microprocessor BUS. From Mixers and Toys to highly sophisticated CNC Machines and Industrial Robots, a wide range of products can take advantage of the Microprocessor Technology. > If you too are switching over to Microprocessors, come to Dynalog Micro-Systems. We have the widest range of Microprocessor Development Systems and standard building block type circuit boards which can go into your products/systems as OEM parts. Write or call today and find out more about these products and their applications. Dynalog Micro-Systems 14, Hanuman Terrace, Tara Temple Lane, Lamington Road, Bombay 400 007 Tel: 362421, 353029 Telex: 011-71801 DYNA IN Gram: ELMADEVICE Primer Publisher - C.R. Chandarana. 2. Koumari. 14th A tied at Trupti Offset. 103 Vasan Udyog Bhavan. TuLsi Pipe R< Road. Khar. Boi .mbay 400 052. I. Bombay 400 013.