CONTENTS September 1991 Volume 17 Number 192 In next month's issue Electronic hygrometer AC detector Function generator (1) SOMHz é:bit DAC ‘Areview of coding theory Four-terminal networks. ‘Audio spectrum shift techniques Front cover ‘Maps and charts on a constantly changing video ‘wall—at 25 m (80 ft) the largest in Europe—sgive up-to-the-minute pictures ot how Britain’s wlephone networks are performing, highlighting potential trouble spots, BT's World- wide Network Manage- ment Centre at Oswestry receives every six hours data equivalent to the contents of the Encyclopedia Britannica. BT says that its digital network is more compre- hensive in comparative terms than that of any ‘major operator. The centre currently monitors all the organization's processor-controlled System X exchanges—S7 in the trunk network and 373 Local units. I also ‘monitors the company's three digital intemational exchanges and 1500, routes linking its UK network to 199 countries. Copyright © 1961 Elektuur BV | ABC 38 Product overview 60 Profile of MicroAmps 1 Computers made expert in real time eon 34 Remote contro! ICs MV500 and MV601 Plessey Semiconductors Rea u Peak indicator for loudspeakers 9 an 87C51 programmer by Dr David Kyte 53 PROJECT: Plotter driver by Danielo Sijtsma Se 46 Keyboard circuit by D. Nelson Pees 31 PROJECT: Asymmetrical-to-symmetrical converter by M. Eller 16 PROJECT: Central car locking control by S. Zipp 18 PROJECT: Time-code interface for slide control by A. Rigby 48 PROJECT: A cheap, efficient, strategic fire alarm by C.C. Whitehead PN Se Ong eeu 42 Build the Opticalock by Michael Swartzendruber Eves nr eee cae 24 Fiber optics ~ Part | by Joseph J. Carr Seine wien uenae 30 Understanding the EDIF standard by Tony K.P, Wong rene eee 29 Measurement techniques - Part 7 by RP Zamis, Lee AEN seo Electronics scene 12, 13; Events 13; Corrections 28; Letters $6; New books 57; Readers services 61; Switchboard 62; Terms of business 62; Index of advertisers 74. on ‘Timecode interface for side control pa Plttor driver —p. 53, Fiber optics Part 1~p. 24 ZAP5t: an 87051 programmer ~p. 39 ELEKTOR ELECTRONICS SEPTEMBER 1991COMPUTERS MADE EXPERT IN REAL TIME 'N A large processing plant itis often the cease that no one person can completely ‘comprehend what is happening at any given time. Even where such an expert does exist he or she will not always be on hand, so there must be some dependence on less experi- enced people ‘A particular problem is that the large amount of data available from increasingly complex plant can hinder corrective action in an emergency, because its sheer volume ‘obscures the real problems. ‘Computer-driven expert systems will be used more and more widely as aids to oper- tors to control and monitor process plant safely as it becomes more complex and lo- cated over a larger area, Using data captured automatically from the process by the com puter. the expert system will carry out much ofthe basic decision-making, supporting the superior reasoning of the operator. ‘An expert system is a computer program that contains the distilled knowledge of an expert or experts. This knowledge is in the form of rules the computer uses to mimic the ‘human decision-making process, most com- monly on the lines of the formula 1P-THEN- rast, that is, “if these conditions apply, then {ake this action, or else.” Tn this way. the expert system aids the user in making decisions in a given set of circumstances, bringing both speed and con- fidence to the process. Real time system Iris also conceivable that expert systems, which are a leading-edge application of the futuristic computer technology of artificial intelligence, could make simple decisions unaided, This possibility is closely related to the rend of running some types of process ing plant unmanned. for example, for week- ‘The control room at the British Gas. Midlands Research Station. ‘To make unmanned running feasible, the controlling system would need a primitive form of decision-making intelligence like that of an expert system. For an expert sys~ tem to assist effectively in the on-line moni- toring and control of a continuous process, ithasto operate inreal time. A real-time sys- tem is one that can receive continuously changing data from outside sources and, 10 process that data sufficiently rapidly, is ew pable of influencing those sources. Annew British product, Cogsys, is said t0 be the first generally available expert system for industrial plant monitoring that takes the time factor into account. Its ability to reason inreal ime means it knows” that,iFa certain situation is seen to be developing, the plant operator should be advised within a given interval. Itcanalsobe programmedio fallow trends in monitored plant parameters on a time basis, and is able to predict future de- velopments in the light of continuously changing plant readings. Supervisory role Cogsys comprises an expert system shell — that is, the basic software program, exclud- the process-specific expert knowledge that makes it an expert system — into which is fed information acquired from a human ‘expert about each particular plant where itis used, Ivalsohas a run-time system to monitor the plant, A general-purpose interface lows direct access to plant data as well as communication with devices such as pro grammable logic controllers, ‘Theexpert system does not provide direct digital control of a process plant; rather, it has a supervisory role with or without oper- ‘ator intervention, It is capable of general ‘alarm analysis and fault-finding, data vali- dation and reduction, and itcanuse historical data to maintain product quality and analyse plant performance. An important feature of CCogsys is its ease of configuring, so plant engineers can operate itafter only a brief pe- riod of training, Two British companies representing two very different areas of industrial process control were selected to pioneer the applica Monitoring and control ‘The first was CBM Technology, where the expert system has been applied to a plastics blow moulding machine making laminated plastics battles. Cogsys is used to monitor and control the complex relationship be- tween the variables that control bottle weight Cogsysis interfaced with the blow mould- ing machine's existing control system, based ‘on Intel single-chip microcomputers and ‘multibus equipment linked to a DEC 3100 Fein pub by ELERTOR or 44380200057 eration GERMANY NETHERLANDS ELECTRONICS (Paishng) “eden (059) 20616 han) Exteel Gantt Heswur BY 244 $H)290 616 Ustemainaly Stitertld See 28 Per Treckpetrat 24 {dior pabser- Len Sever 100 AACHEN S101 Vk BEEK “hil Eater J Busing uropean Otc ator: EA Krempetauer For: PEL. Kersenshes ea Owe Fontan 5 ‘cnkhce PORTUGAL Doves (IBOAB BEEK Elie BPE Fra Bono La ‘oi Koad The Netra ara RD. sin ee ONDA Sis.19 “eepoe: 1-4 36644 Tes = ATHENA {en S80 rnd “oer $6617 (leh Editor: E Xan, Feito Jerenias Sequira “Taine 01:57 1588 Nona | Fas 4314637 01 HUNGARY ‘SPAIN orsH4RS7 688 ores Managing Drei: MM, Landman Elo lei ya Resor Eloi Apiada ‘ews pe G) k 1015 Bode Cale Mawes 13 Ea 80159 (avo) Distribution: ‘Baty 13 2a) MADRID ‘SETA9SS Utratonal) SEYMOUR ior Lats Ams ator: Agsin Gonzales Beta 120 Laon Road INDIA SWEDEN Aderting PRB Lind LONDON SWie SoH Eker Elects PVT Lid lcm Pros AB S Woley Terce Ch Biking Box Ss CCHELTENHAMTGLS) (TH Printed a be Neha by ND, SIC, Petr Roa, Gant Rod) aus WUDDINGE “eophone: 4) 810780 Zoaevsute BOMBAY 40 007 Feder Bs Caan Fan OD) 2 Fale: CR. Chania USA & CANADA ISRAEL kur Fecoeies USA Subscriptions: Elektra PO Word Wie Subs Sie Li POR S105 PETERBOROUGH NH 08438-0576 Unit, Gitte Reel na TELAVIV6ISt0 Publisher: Ena TD Paley Road abit Avan “TICEHURST 1S 7H, PS $9480 NIEPPE “Teen (088) 20657 Natal) ELEKTOR ELECTRONICS SEPTEMBER 1991 Fairs: DRS, Meyer GCP Rawdon’(Digital Equipment Corporation) worksta tion. Interfacing with the machine's con- trollers and with a stand-alone workstation allows Cogsys to collect real-time ditt and data input by the operator. ‘The expert system is backed up by statis- tical process control techniques used to pro- vide a graphical representation of product quality, ‘The second pioncer is British Gas which, aitsresearch station inthe Midlands, has also. applied the Cogsys expert system to real- lime process control. Specifically itis using anexpert sysiem to provide decision suppor oprocess engineers operatingasmall gassyn- thesis plant supplying research and develop: ment staff with gas for research activities. The roles of the expert system include advising on optimum control strategy. moni- Joring alarms, and providing a detailed op- crating procedure consistent with the state of the plant At the research sation, the expert system interfaces with 15 control devices within the plant, and with some 100 analogue und igital points through a serial link to a DEC MicroVax computer, enabling the plant to be monitored and controlled by non-expert personnel, Computer language ‘The objective of this application is to show how less-skilled staff, aided by an expert BOOST FOR AUTOMOTIVE, ELECTRONICS SALES? REPORT from Frost & Sullivanpredicts that new legislation in westem Europe governing vehicle exhaust emissions and. passenger safety will cause spi of non-entertainment automotive electron- ies (OE) in coming years, reaching an an- ral revenue of well Over £2 billion by 1995, Since the market for OF in western Europe is still in its infancy compared with North Americaand Japan, penetration levels, which will bring with them a drop in prices. will appear even more impressive, INTELLECTUAL PROPERTY RIGHTS LURING the European Telecommunications Standards Institute's 10th Technical Assembly at Nice in June, concern was ex- pressed to reconcile the needs of standard- ization with the world of Intellectual Property Rights. After a debate lasting a full day, the Assembly agreed on the direction it wants its Intellectual Property Rights Committee to take to overcome the defects of the guide- lines cumently used by the {nferational Starks Organization—ISO. The Assembly decided that the transparency needed oensure its stan dards should not be blocked by patent hold- ersmaking unreasonable demands, These con- Aitions could no longer be guaranteed by a purely voluntary IPR policy. The ETSI Intellectual Property Rights ‘Committee will produce a draft undertaking system, could in future operate a complex plant safely and efficiently. Cogsys gives advice on plant effcieney, and offers alarm handling and analysis. I also provides con- trol loop optimization at a supervisor se Point level and analyses trends for condition ‘monitoring and operational guance. Operating procedure advice is given by text animation, Currently, Cogsys is written to «un on DEC computers using the DEC VMS opet= sting system. Is planned to make the sot ware available to run on other computers. The widely used artifical intelligence lan- ‘guage POP. 0G. originally developed a Sussex University, was selected to develop the ex- pert system shell and interfaces. with ¢ as the language chosen for the run-time sys- tem. Easy input of data into the run-time system by plant operators was a fundame requirement. Joint venture Cogsys isthe result of several years research and development involving 37 companies, including Rolls-Royce. Shell, Courtauldsand British Gas. Tis one of the first commercial joint Ventures 10 result from the govern= ment-sponsored Alvey Project set up in the early 1980s to stimulate joint university and industry research and development projects Cogsys was developed by SD-Scicon, a company working on the application of ar- lificial intelligence to industrial automation that willbe legally binding on its members ‘This will require adelicale bulance between the rights of patent holder and the benefits of widely applied ETSI standards, The aim is for patent holders to have a fair and sea sonable reward their patents are infringed Atthe sametime, other members wouldhave assurances that licences wouldnever be with- held or discriminatory ETSI aims to finalize its policy by the end of the year. NEW DIGITAL LOGIC CIRCUIT SIMULATOR PROGRAM UMBER One Systems Ld have an- nounced a new Digital Logic Circuit Simulator Program, called PULSAR, which is aimed at freeing engineers from the frustra tions of designing with a scope and solder- ing iron, PULSAR enables a design to be tested and proved before a single component is sol dered, and without the need of expensive testequipment.Itwill handle bstables, monos- lables, latches, three-state and open-collec- tor gates in addition to normal types. The program contains a full component library facility that allows the user to add com- ponents, modules and sub-circuits tothe full range of 74LS and 4000 series models sup- plied with it since the early 1980s and with one of the largest artificial intelligence development groups in Europe. The company was re~ cently called on by the Department of Trade and Industry to carry out a study of the uses to which expert systems are being put in Britain ‘To market Cogsys and other industrial automation software based on artificial in telligence, a new venture company, Cogsys Ld hasbeen formed by BritshGas, SD-Scicon ‘and Salford University Business Enterprises CBM Packaging Technology PLC Denchworth Road, Wantage OX12 9BP, Telephone (0235) 772 929. British Gas PLC, Midlands Re Wharf Lane, Solihull BOL 20W, ‘Telephone 021 705 7581 SD-Scicon Ltd, Pembroke Hous Broadway, Camberley, Surrey, Telephone (0276) 686 200. Pembroke Cogsys Lid, Enterprise House, Unit 37 Salford University Business Park Salford M6 6AJ, Telephone 061 745 7604. Salford University Business Enterprises same address as Copsys Lid, above, PULSAR may be operated independent entering and modifying circuit data through an imteractive Netlist Editor, but is also ready to link in with the forthcoming rasy-pc Professional Schematic Capture package and the ANALYSER i Linear Circuit Analyser to form a completely integrated Electronics Engineering Design System, The program is priced at £195.00, com- plete with comprehensive manual. Itis not ‘copy- protected, ‘Number One Systems Ltd, Harding Somersham Road, St. Ives, Huntingdon PEI7 4WR, Telephione (0480) 61778. IBC: CALL FOR PAPERS, CALL for papers has been issued by the orgniversof the Fourteenth International Broadcasting Convention—IBC—tobe held in Amsterdam from 3 10 7 July 1992 TBC has « reputation for high-quality pa- person state-of-the-art developments andthe ‘convention proceedingsareanessential work of reference for all aspects of current and future broadcast technology. To allow for comprehensive coverage of the widest range of topics possible, there will be two parallel streams with time allo- cated for discussion. Among the many top- ies to be covered are advanced, compatible and higher definition television systems: analogue and digital audio recording: studio and network distribution: encryption, sub- nand pay-TV: terrestrial transmission and reception; cable distribution systems::and ELEKTOR ELECTRONICS SEPTEMBER 1991test and measurement technology. Synopses of about 800 words are invited, which mast reach the [BC Convention Office by 11 November 1991. Full papers, which will be vetted independently, will be re- uired by 1 March 1992. IBC Convention Office, Savoy Place, London WC2R OBL. GREEN LIGHT FOR ENGINEERING MERGER NA move that sirenghtens the UK engi: neering profession, members of both the Institution of Electrical Engineers and the Institution of Manufacturing Engineers (for- ‘merly the Institution of Production Engineers) have voted to join forces and create a com- ined institution to mect the challenges pre- sented by rapidly changing technologies. The proposals will come into effect on 1 October 1991, The IEE (110,000 mem- ’bers)and the IMFgE (20,000) will jointly form, the largest of the 46 British engineering in- stitutions. A new Manufacturing Division will be created within the new institution that, with some 40,(000members, will beoneof the most powerful voices for the UK manufacturing, industry NEW BOARDMAKER- PCB CAD PACKAGE ISIEN has announced a new release of itsBoardmaker-2 PCB CAD package for electronics professionals—Version 3.4— building on proven benefits of the system to further boost productivity and speed design cycles, Amongnew facilities forthe PC-based software are more schematic capture inputs. artwork output options and numerous pow- erful CAD functions, including top-down ‘modification, providinganefficientand com- prehensive design-to-manufacture environ- ‘ment for just £295.00. The top-down mo fication makes it simple to introduce mos fications toexisting designs. Schematic changes now automatically modify PCB layout and netlist files, leaving unchanged sections of the PCB layout alone and presenting the de~ signer with ust the modified areas forrework. ‘Tsien (UK) Ltd, Cambridge Research Laboratories, 181A Huntingdon Road, ‘Cambridge CB3 0D, Telephone (0223)277777, ANIMATED CIRCUITS FOR EDUCATION ‘T THE recent Leetronex Electronics Exhibition at Leeds University, Labeenter Electronics introducedtheir Animated Circuits for Education—ACE—soltware package for the IBM PC that combines circuit simu- Tation with animation to produce animated cireuitdiagrams. Colour shows voltage, while dot motion shows current flow. The result is an instantly comprehensible, time-evolvi view of eireuitoperation: something that even banks of test gear connected to areal eireuit can not give you, The program was presented by its author, DrBob Gomersall, Head of Physics, Bradford Grammar School Labeenter Electronies, 14 Marriners Drive. Bradford BD9AJT, Telephone (0274) 42868, Fax (0274) 481078. NEW PORTABLE SURFACE CONDUCTIVITY METER HE measurement of surface conductiv- xy of static dissipative or insulating ma- {erialsis important forthe elimination of static electricity that can destroy sensitive semi ‘conductors and form a hazard in areas where flammable or explosive substances are han- led. Hiltek's new portable surface conduc~ livity test set is designed for use in facto~ ries, warehouses and laboratories, and mea- sures from 107 0 1018 O nl. This is the re sitivity range where static build-up can be problem, butalso whereiitcanbe eliminated Hiltek Ltd, Breedhurst Road, Detling. Maidstone MEI‘ 3EB, Telephone (0622)47472. CDR FILTERING MADE EASY ILICON Systems's new 32F8020., with a Frequency range of 1.5-8.0 MHz and programmable pulse slimming equalization (0-9aB high-frequency boost). greatly eases Constant Density Recording—CDR—filte ing. It also extends the company’s Line of Filter ICsto include audio/video tape. floppy ‘and optical drives nd ultrasound imagingap- plications. ‘The new chip provides an electronically controlled low-pass filter with a separate differentiated low-pass ouput. A seven- pole, 0.05° Equiripple-type linear-phase. low-pass filter comes with a single-pole, zero differentiator. Both outputs have Tustin, CA 92680, USA, Telephone (714) 731-1110, Fax (714) 669 8814. IEE AND IEEIE PROGRAMME 2-5 Sept—Digital processing ot signals in ‘communications sixth international con ference at Loughborough University) 2-6 Sept—Circuit theory and design (lenth European conference at Copenhagen) 9-12Sept—Miewowave (wenty-frst European conference at Stuttgart) 9-12 Sepi—Optical communication (seven- ‘een European conference in Paris) 9-13 Sept—Railway signalling and train ‘ontolsysiems (vacation school at University of Birmingham) 11-13Sept-—Electrical machinesancl drives (fifth international conference at Savoy Place) 11-13 Sept—Algorithms and architectures for realtime control (workshop at the University of Wales, Bangor) 15-20Sepr—Flectromaznetic compatibility (vacation schoo! at University of Kent) 16-18 Sept—Sofwware engineering for real time systems (hird international eonfer- ence at Cirencester) 16-18 Sept—Analogue-to-digital und digi tal-to-analogue conversion (international conferenceat University College Swansea) 17-20 Sept—AC and DC power transmis- sion (fifth international conference at Savoy Place) ELPRTOR ELECTRONICS SEPTEMBER 1991 Sept—Sensors and their applications oP Fifth conference in Edinburgh) 22-27 Sep!—Microwave measurements va cation school at University of Kent) 23-24 Sept—Personal. indoor and mobile ‘dio communications (international sym- Posium at King’s College, London) Information on these, and many other, events may be obtained from the IEE, Savoy Place, London WC2R OBL, Telephone 071 240 1871 or from the IEEIE, Savoy Hill House, Savoy Hill, London WC2R OBS. Telephone O71 $36 3357. Havering College of Further and Higher Education willruna Radio Amateur Exam~ ination preparation classstarting on Monday 9 September and 2 Morse Examination class starting on Thursday 12 September. Formore information wrteto Stuart Woosnam, GONKP, or Chiris Potarzycki, GONIR, at Havering Collegeof F& HE, Quarles Campus, ‘Tring Gardens, Harold Hill, Romford RM39ES, or ring the Department of Electronics and Engineering Technology on 04023 81460, ext. 7131 ‘The International Electronics Design Conference will be held at the National Exhibition Centre, Birmingham. from 24 10 26 September in_ conjunction with the 1991 Electronics Design Show. For more infor ‘mation, contact MGB Exhibitions Lic. Martowe House, 109 Station Road, Sidcup DA |S 7ET, ‘Telephone 081 302 8585. During September, Frost & Sullivan will ‘conduct seminars on, among others, Novell Local Area Networks. Further information from Frost & Sullivan, Sullivan House, 4 Grosvenor Gardens, London SW1W ODH. ‘Telephone 071 730 3438. Vision & Audio "91 will return to London's Earls Court, running from 15 to 18 September in the brand new No. 2 exhibition hall, For ‘more information, contact Denise Pritchard oor Lucindade Freitas at EGA Communications, 1 Jew Street, Brighton BN |UT. Telephone (0273) 23889,designed and marketed as a kit by ELV, provides a visual warning when the maximum drive power to a loudspeaker is approached. The sensitivity of the can be set to power levels between 1 W and 300 W. Connected in parallel with the loudspeaker terminals, the peak indicator is simple to build, and does not require a separate power supply. cuit HE power applicd to a loudspeaker drive unit is disproportional to the vol: ume, which isa subjective quantity. In prac- tice, this means that doubling the volume requires more than double the drive power. Itisfor this, and other, reasons that the maxi ‘mum permissible power applied to a loud: speaker difficult to determinesubjectively, ie, by listening. Unfortunately, this has caused the destruction of many an expensive loudspeaker drive unit, although we must hasten toadd that many drive unit manufac turers and suppliers state peak power ratings that are, to put it mildly, on the high side Adjusted to a sensible indication level, the present peak indicator gives you ease of mind as far as the protection of your costly loudspeakers is concerned. First, however, consider the difference between the follow= ing power specifications of a loudspeaker: peak power and pulse power. In general, the indi- cation level of the peak power monitor should be based on the continuous power rating of the loudspeaker — but only if you continuous power, music power have reason to believe that this value makes sense, in other words, that it is not grossly exaggerated by the manufacturer. [n some es, this means that you have to set an in of the stated by ion level power manufacturer, d Tow as 50 sinuous rating Such a ‘conservative’ setting of the peak power indication may cause the LED on the unit tolightbriefly at times, thatis, when the corresponding volume is at all desired. When the volume is increased further, the LED starts to flash shorter and at a lower rate. Ths isan indication that the maximum, hhas been reached, and that safe power level the volume should be reduced to prevent damage to the loudspeaker (and, even more importantly, your ears), The circuit Two flexible wires are used to connect the input ofthe peak indication circuit in parallel with the loudspeaker terminals. This results in the loudspeaker drive voltage being ap- plied to PCB terminals ST1 and ST2. The po- Iarity of the drive gnal is irrelevant, since the input signal is rectified by Di-D2 (ee Fig. 1), and assumed to be symmetrical. The input voltage (supplied by theampli fier) also serves to power the circuit, This is achieved by rectifying and smoothing the input signal with the aid of diode D: smoothing capacitor C2. Resistor Reserves to increase the internal resistance of the circuit, and so prevents waveform distortion of the input signal The input voltage is also applied to a potential divider, Ri-R2, via diode Di. The signal at junction Ri-R2 is fed to the base of transistor Ti v series resistor, Ro, and a >a -© 909516-11 Fig. 1 Circuit diagram of the peak indicator. Note that the circull is powered by the drive ELEKTOR ELECTRONICS SEPTEMBER 1991Fig. 2. The printed-eircult board for the peak indicator is small and single-sided. zener diode, Ds. The diode serves to set up a well-defined switching threshold. Resistors Rs and Rs are dimensioned such that Ti is kept off in the absence of an input signal When the voltage at junction Ri-Ro ex ceeds 45 V, Th starts to conduct. Conse- quently, T2 is driven via Rv, which causes capacitor C1 to be charged via Rs. Assoon as the charge voltage exceeds about half the supply voltage (measured across C2), tran sistor Ts staris to drive a constant current Table 1. Circuit actuation levels Power (W) ELEKTOR ELECTRONICS SEPTEMBER 1991 "BAK INDICATOR FOR Ce TS ‘Content of it supped by ELV 100K2 preset H Capacitors: 1 2uF263V 1 220uF 63v source, Di-D-Ts-Ris, which in tum sends a current of about 10 mA through the indica tor LED, Ds load condition. ‘At the same time, the constant current causes Ts to start conducting, so that the potential at the emitter of T3 drops to a frac- tion of the previously established half sup- This lights to signal the over ply voltage. This creates an amount of hysteresis sufficient to cause the indicator to light for about half a second, even when the overload condition at the input was pulse like, When no further overload levels are de- tected, C1 is no longer charged, and starts to discharge via Ro, the base-emitter junction of Ts, and Rui-Ris-T4, until the voltage drops below the switching threshold of Ts. When that happens, transistors T3, Ts and Ts are switched off, and the LED, Ds, goes out Preset R1 allows the sensitivity of the ci cuit to be set between | W and 300 W. The corresponding input voltage levels are sum- marized in Table 1 Construction All components are accommodated on the small printed circuit board shown in Fig. 2 (size: approx, 35:48 mm, or 0.14%0.19 inch) Start the construction by fitting and solde ring the low-profile parts shown on the com lay. Then follow the taller components. Take care to fit the polarized components (the diodes, electrolytic capa tors, transistors and the LED) the right way around. The cathode (negative connection) ff the LED is the terminal with the tab near the underside of the plastic body. Fortu- nately, the LED will not normally be dam- aged when fitted the wrong way around, at ponent over! A complete kil of parts for the peak indicator is available fron the d signers’ exclusive worldwide distribu- tors: OUDSPEAKERS ELV France BP. 40 F-57480 Sierck-Les-Bains FRANCE Telephone: +33 82837213 least, so long as the reverse voltage is not ex- Adjustment Run a thorough visual check on the com- pleted printed circuit board before carrying ut the First electrical tests, These require a stabilized, adjustable dic. power supply. De pending on the supply voltage, the « draws few milliamps when the ‘or 10 mA or so when the LED is on. In any case, the current requirement is never higher than 20 mA. Determine the power at which you want the indication to beactuated, and take the re- quired switching threshold from Table 1. For instance, when a level of 100 W is required for a loudspeaker with an impedance of 4.2, the threshold works out at 28.3 V. Similarly for 10W into 16.2, the switching threshold is ov Set the desired threshold voltage on the stabilized PSU, and connect the positive oul- put to STI of the peak indicator, and the ne- ative output to ST2, Set preset Ri to its ‘maximum value, ie, turn it fully clockwise. The LED must remain off at this point. Next, refully turn Ri anti-clockwise until Ds lights. Leave Ri at this setting. Reduce the test voltage. The LED should remain on fora hys When it goes out, increase the test voltage again. Keep an eye on the applied voltage. The LED should light at the pre viously set voltage level. If not, carefully redo the adjustment of Ri Since the LED is supplied from a constant current source, it will ight ata virtually con- stant intensity irrespective of the actually measured loudspeaker voltage. After thead: justment, the circuit may be fitted at a suit: able position in the loudspeaker enclosure, and connected to the input terminals. i while as a result of the relatively la teresis,CENTRAL LOCKING CONTROL by Dip. Eng. S. Zipp Most new cars are now provided with a central locking system, either as standard (most) or as an optional extra, sometimes combined with an anti-theft system. If you are one of the unfortunate many who can not (yet) afford a new car, but would like the many benefits of a central locking system, the project described in this article may well be for you. Not atts of particular modelo are fitted with central locking. It is possible to have such variants modified at a late date, but that is invariably more expen- sive than buying a relevant kit from a car ac ssory dealer, In many countries, there are two DIY kits of parts available: one for the front doors only, and the other for all four doors. We don't know of a kit that also caters for the boot or, in case of a hatchback, the filth door. The two-door kit contains two master ‘modules that must be coupled to the open- ing mechanism in each door. Each module contains an electric motor that via suitable gearing raises or lowersa lock shaftby about 50mm. The lock shafts must be mechani- cally coupled to the existing dootlock mech: anism. When the modules are fitted to both front doors, simultaneous locking or un- lockingofboth doorscan bedone from either doorlock, ‘The four-door kit contains two master modules and two slave modules. All four doorscan be locked and unlocked fromeither front door, but neither of the rear door locks can control the other doors. ‘The main difference between a master module and a slave module is that the for- mer has an additional electrical change-over contact that ascertains the position of the key in the lock ‘Thekitsalsocontainasmall PCB, on which the electronic control circuits are housed These circuits interpret the position of the switchcontactsand controlthe power required by the motors. They can not be modified for the control of say, four or five masters, and, Insome cases, are not infalible. In view of the limitations of the electron- ies part of these kits, it was felt that a far _more sophisticated circuit would not come amiss. Moreover, the proposed circuitalsoof- fers the possibility of being operated by an infra-red remote control, Circuit description , IHis fairly straightfor ward and uses only a handful of compo- nents. tissuitable for thecontrolof up to four master modules and two slave modtes, When thelock ona master-controlled door the evant change-over contacts in the master is operate nsuing signal from the rel- module is applied to one of the inputs of quadruple Schmitt trigger IC; via an RC type low-pass filter. This manitests itself as a voltage jump at one of the outputs of the Schmitt rigger and thisis registered at inputs A0-A3and BO-B3 ofcomparatorlCs, Because ofthe RC-typelow-pass illersat inputs BO-B3, the signal at these inputs is delayed slightly. Consequently, the level at the A=B output of the comparator changes from high to low and this triggers monostable ICs, which is a non-retriggerable type 4538, Duringthemono time, whichisdetermined by time-constant Rj-Cs, the output of the monostable is high, so that transistor T) is switched on and relay R Tong as this condition pertains, and only for so long, the battery supply line to the mo: tors ofall modules is completed via the rest contacts of relays Rey and Res It follows that the mono time must be ong enough to allow the motors operating the lock shaft, thereby locking or unlocking all the doors. On the other hand, the mono time must not be too long, since the motors in their stop position is energized. As rentand overheatquickly. The proposed time of 220 ms has proved ideal in use with a number of different kits. If, nevertheless, are experienced, the time can be longer by increasing the value of Ry and/or Cy, and made shorter by reducing, the value of those components. The direction of rotation of the motors depends on the position of the contacts of Rey and Res. The change-over contacts of thes sysarearrangedan reversing switches Since the motors have fo change direction every time a key i tured, tei atest direc: tion is stored in D-bistable (US: flip-flop) Iq, which fumetions as a binary scaler (US scale-of-two counter). When the Q-oulput Of this stage is high and the monostabe i active (AND gated via Dp), transistor Ts tog- files and relays Re; and Rey are energized When the Q-output of [Cay is low. relays RE and Rez are quiescent and the polarity ofthe voltage across the motors reversed The status of [Cay changes at every lead- ing edge at the Q-output of IC. This has the advantage that the next direction of ror tation ofthe motorsisdeterminedonly when the motors have reached their end-stop, ‘The power supply ofa circuit that is used ELEKTOR ELECTRONICS SEPTEMBER 1991in a car must be designed to overcor problems that can occur in a car's electrical 2 system. The present circuit requires 8 V and this is derived from the car battery via an ELEKTOR ELECTRONICS SEPTEMBER 1991 8-V regulator, ICs, followed by a 1000 uF ea- pacitor, Cys, to earth, Diode Dy prevents the capacitor being discharged via the regulator when the engine is started. The capacitor en sures that the circuit remains operational for at least 10 seconds after the battery voltage fails. If remote control is wanted, one of the four masters must be forgone and the ensu- ing free input at the Schmitt trigger used as the remote control input. Suitable remote control transmitters and receiver units can bbe obtained from most car accessory dealers Construction ‘The circuit i best built on a suitable piece of prototyping board (veroboard) —see Fig. 2. Decoupling capacitors Ci4-Cis must be cated as close as possible tothe relevant pins of IC}-IC; to make certain that any noise on the supply lines does not enter these CMOS devices. The relays can be of the type used for flashing indicators, which are fairly small, yet have contacts that are rated at 20 A. Connections between the board and the modules and the battery are best made with the aid of standard car-type connectors The board may be housed ina simple ABS or other man-made fibre enclosure. Full fitting instructions are enclosed with the DIY central locking kits. The main addi- tional work involvespreparingandlaying the cables tothe various locations. Youshould set aside about 2-21/, hours per door. .TIMECODE INTERFACE FOR SLIDE CONTROL PART 2: CONSTRUCTION As discussed last month, the timecode interface is an ingenious piece of electronics capable of putting time codes on a magnetic tape at 0.1-second intervals. Here, this system is used to control slide projectors. This month we finish the article with details on the construction and practical use of the unit. by A. Rigby Continued from te July you felta little put off by the complexity of the timing diagrams shown last month, rest assured that you need not understand these in detail to be able to build and use the tie mecode interface. The purpose of including the timing diagrams in last month's instal rent was to set out the operation of the cir cuit in sufficient detail whilst avoiding a very long circuit description. Fortunately, the construction and practi- cal use of the fim complex as the actual circuit. The unit is con- trolled by a program supplied on disk through our Readers Services (order code 1611), The program enable: ton the timecode interface with the aid of the Universal I/O Interface for PCs. Also on the disk are a number o files, written in Turbo Pascal, that contain all the routines necessary for the practical use of the ti- mecode interface ‘ode interface are not as joutoruna test Construction The layout of the front panel designed forthe timecode interface is shown in Fig. 7. The self-adhesive fol that gives the front panel a professional look is available ready-made through our Readers Services ‘Motherboard The double-sided, through-plated printed circuit board (Fig, 8) is cut to separate it from the three support pieces at the side of the [| TIMECODE INTERFACE atl MBSA 0 000 BSA 000 Fig. 7. The front panel igned for the timecode interface gives the uni ‘aliractive, professional finish (shown al true size) ELEKTOR ELECTRONICS SEPTEMBER 1991660800081 9909006) $9959599IC1 4 Uy" ‘odode @3 ol TIMECODE INTERFACE — MOTHER. BOARD Resistor 6 4700 Renienia.rer; 4;R3 Ra.R19.RI7:Re7; 33:74:39 Fis Ria. Runie RasArAIOAN; Ri8:R20;R22; | R23-F25, R28, 35-R98iR40 Ra. Ree Re S.RA:AS Ge Ree Ps cr201 ca cscaica.crs: 620 cis 7 co cir carsc2e c162¢3:¢68:¢10; crisct8;c24; 1085 mi 4 3 1 1 Tes: CLEeETTT 100, 10V: axial 100,F: 16V; racial 10, 25; racia AY-3-1015, 3130 7sHOTOO ‘7aHCTS2 7AHCT74 7aHOTeS 74HCT123 7aHO1a2 74HOT139 74HCT2AS 74HCTS41 74HCTS74 74HCT4060 74HCT4068 7605 iscellaneous: PCB-mount ine (RICA) socket Spin SIL pin header ‘Sépin SIL pin header aiD9:010 1176 tea ct ior2 lors Iergio18 ea lers1017 (015 ore in leeuca Ieslcesc7:e10 ca Iezocet io22 tc23. 20-way PCB mount IDC ‘connector with angled pins ‘and side laches 10,way PCB mount IDC ‘connecter with angled pins and side latches ‘2eway PCB terminal block (Grim piten) ‘lide switch with changeover 1;S2 contact 5.2.MHz quartz orstal enclosure Retex "Re (145017158) print creut board front pane! oi Conia software on disk k10 Ku x 910055 910055-F ESS 1611 * Imhot-Bedco Standard Products Lid, Ash- ley Works, Ashley Road, Uxbridge, Middle | sex UBB 2s0. ELEKTOR ELECTRONICS SEPTEMBER 1991‘TIMECODE INTERFACE FOR SLIDE CONTROL - 2 Although fairly large and densely popu- lated, the motherboard is easily built with the aid of the component overlay and the ‘components list. The PCB accommodates all components, with the exception of switches 1 and 52. Its size is geared to the inside dimensions of the enclosure mentioned in the components list. ‘The construction of the display boards is little more complicated than that of the motherboard, and discussed separately below Note thatthe displaysare optional — they are not strictly required when a com- puter is used to control the unit. They do, however, allow incorrect settings of, for in- stance, switch Si the baudrate or preset Pi, a ae = | to be noticed immediately. The small PCB pieces you have already cut offallow the dis play module to be mounted to the mother- = | board in a simple and secure manner. The supply voltage and the LOAD signalare fed to renee ROE SES eyes cm msm |e Mlsplay vi connector Ks Display Asalready noted, the read-out used is based Fig. 9. All parts required to bulld the display. Note the IC pin strips with tong pins. on an earlier design developed for the Elek- tor Electronics Digital Train System (EEDTS). The construction of this display unit is discussed in Ref. 2 (see Part 1). There are, however, a number of points that deserve your attention. 1 The PCB used for the address display is the triple version, order code 87291-9a, which accommodates six displays. Here, we use only five displays, so that the display section at the extreme right — LD1, IC and RI-R, of the associated SMA section — must not be fitted 2. Before you start the construction ofthe dis- play module, cutoff the display section from the SMA section. Next, mount the wire links, capacitors, SMA ICs and the displays on the boards (note that the bevelled edge of the ICs ‘marks the position of pin 1). The SMA parts are best soldered with thin solder wire of a diameter smaller than 1 mm. The SMA ICs arebest fitted by first soldering two diagonal corner pins, and then aligning the other pins with the copper pads below before they are Fig. 10. In this way the display and the SMA boards can be joined easily. soldered. Thoseof you who wanta clear sep- aration between the minutesand the seconds 7 indications may fit the two 680-2 resistors between pins 6 of the two LD1 displays and ground (see Fig. 13) . 3. Toenabletthe display units to beconnected to the comer pieces, use IC pin strips with Jong pins (1 em; see Fig. 9). These pin strips must be fitted before the two PCBs are joined. 4. On the SMA PCB, all parts are fitted as closely together as possible. If you use the enclosure mentioned in the components list, ‘you must use ultra-miniature resistors (max. length 5 mm). When the display board and the SMA board are joined, these resistors are inserted simultaneously. This is easiest done ‘when the wire ends are cut such that they form an oblique line from one side of the board to the other (see Fig. 10) 5, The supply wires are soldered to the dis- play board. here. 6. Check the complete construction before Fig. 11. The completed display. The IC pin strip is visi ELEKTOR ELECTRONICS SEPTEMBER 1991GENERAL INTEREST My) Tayi bie eon [saeeaeey Fig. 12, Ce you put the two PCBs together. First, solder the four resistors at the edges. Check that the PCBs align, and that there is sufficient room to solder the resistors properly. Also ensure that there remains sufficient space between Fig. 14, Printed circuit board for the S-digit address readout. The surlace-mount assembly (SMA) ICs are fitted at the track side. Fig. 19. ‘These connection must be made at the rear of the SMA PCB, and wired to connector K8. Also shown here is how the two ‘decimal points of two display can be joined. the IC pin strip and the display PCB, so that short circuits can not occur. When the Retex RE? enclosure is used, the distance between the two PCBs is about 8 mm. Finally, solder the remaining resistors and the supply wires, (see Fig. 11). 7. As shown in Fig. 14, the supply and the LOAD connections are interconnected. be- tween the PCB sections. They are connected to wires that allow the display to be driven. from the main beard via connector Ks. Putting it together Cut the clearance in the front panel to size before you mount the completed display unit on to the motherboard, The clearance should be carefully cut (with a jig-saw) and ‘widened (with a small file) so that the dis plays just about pass. The corner pieces are fitted on to the PCB with theaid ofa straight header or short pieces of solid wire. The dashed lines on the corner pieces indicate the Tength to which they must be cut to obtain the correct height for the Retex enclosure. [F you usea different enclosure, determine the ‘Showing how the display and the corner pieces are mounted on to the main board. required size of the comer pieces before cut- ting them. Start the assembly by fitting the headers, (or the wire pieces) on to the motherboard Next, secure the motherboard in the enclo- sure. Insert the display into the clearance in the front panel. Solder each of the three cor- nner pieces at one point, so that their tops are level with the pins of the display. Next, solder the display unit on to the four corner points, so that the front panel is positioned straight. Before soldering the remaining points, run a thorough check on the assembly 10 make sure everything remainsiin place when the enclosure is closed. Notetthat the front panel is designed such that the display unit is fitted as far as ‘possible to the left. As shown in Fig. 14, cach ‘corner piece is atthe left of the header, while the display section is fitted with its left side to the corner pieces, ‘The timecode interface is best powered. bya mainsadaptor with an output voltage of between 9 V and 12 V dc. The current con- ELEKTOR ELECTRONICS SEPTEMBER 1991TIMECODE INTERFACE FOR SLIDE CONTROL Multitrack cassetts recorders are popular with musicians, disk-jockeys and athers who want to mix existing as well as original sound material ‘on tape. It is woll known that many pop bands and ‘budding’ artists Use multirack recorders for the production of demo tapes that are offered as promotion material 1 radio stations and record companies. In the hi-fi aud scone, the multitrack cassette dack is rarely Used as its sound reproduction quality is in principle, not bettor than that of a stereo cassette deck. Because of ths, the multitrack recorder may be Unfamiliar to many of you. whence the following basic description. In its basic form, the muitivack recorder is a standard cassette r ‘corder. It has all the familiar functions like recording, noise, sup- ‘ression and a tape Iansport_ mechanism. ividually, This is arranged with the aid of a number of (ches. and. input and output sockets that en- able you to use any in- ‘ividual track for recording oF playing back. Since the tracks are physically distributed over the full ape Width, the cassette can be used in one direction only. ‘A multitrack recorder alone is not sufficient ta produce, say, a demo } tape for an artist — you will also need a mixing console to control the Jevel of the individual sources (whether new or already recorded ‘ignals). and mix them down to a completa production (‘post mixing’ “Strouph tis torn fe toa usod wit 40y chanel consoles, ‘only), oF to another track. In the days when mixers. and multitrack ‘recorders wers separate units, such a production job meant struggling “witha ot of plugs and cables, with a great risk of errors occurring at all Stages of the production jine. The iniroduction of the so-called Portastucio put an end fo al this trouble by combining a simple mixing ‘console and a multitrack cassette recorderin a single lightweight case, From then an, producing demo tapes with acceptable quality no longer required a sound engineer doing his bit with ots of wires and plugs — 2 great advantage for atisis and pop bands, witness the start of the ‘garage rock’ period. From its introduction, the portastudio was a real uecess, combining ease of use with good sound quality and the freedam to record music, or mix itwith existing Sound matetal, almost THE MULTITRACK SYSTEM The testo Canbelistoned tow rowever pet action xe playing along and ree. tinen we (00k Inside Se Siding. an adional none the recor atin ee restos ONE) ding/playback head is Ss ae em: Besides level and pan rota zack pe, bt veer contol, mos! potas: track ones wie — Satine steel {ules have. loop. ce) ... Uvewsee {or fects equipment tion we find that there eevee wumeters ang | Aeaddphone outputs are a corresponding eee ccciceici | Standard, and. inpuls ruber af vida ooo 7 eset (canbe switched B= heade, The Use!) pan per. Soe _inermdectan | Ieen line amie: tore then wo tec-| Sd POT s aa phone signals. A pitch ordiplayback and ' o pitch controt control or tape speed Srase "heads allows i Switch “sometimes each track lo be te: soot stot provided for the pur: ‘corded and erased in GOS, Timitusses | pose ot playing along anywhere, from the aiticor basement to the garage or local music lub. ‘The term dubbing refers to using an existing recording to make anew ‘one. By making clever use of a portastudio, up ta 10 sources can be recorded white only one dub is required. This is accomplished by first recording three signals, either simultaneously or one after the other, ‘on tracks 1-3. Nox, these three signals are mixed with a fourth signal, and the result is recorded on track 4. This dubbing operation allows tracks 1-3 fo be used again for new recordings. After repeating this peration three times, a total of 10 signals have been recorded. When itis time for the final mix, the panorama (pan’) controls used to give leach track its place in the stereo image. The advantage of this way of ‘working is that the pre- viously recorded signal at low spoed, An in- ‘creasing number of portastudios offer a tape. speed ‘of 9.5 emis, which pro- vides significantly bet- ter reproduction ‘quality. The differences between portastudio units are mainly the number ot Inputs of the mixing console, tne control system (mechanically or electronically), the noise suppression system (Dolby or DBX). and whether or no al racks are accessibo. The timecode interface described inthis articie|s preferably used with a 4-track recorder that has outputs for al four tracks (some multitrack Tecorders have only three outputs). In genoral, noise suppression should not be used because it may affect the frequencies of the control signals recorded on the tape. Problams are likely to occur at relatively high recording levels when the noise suppression is switched on ‘Simply try out the elfect on your recorder — i! problems occur, you wil have fo make do without the noise suppression. Unfortunately, this ‘means that the music programme with yout side presentation has to be recorded without noise suppression. Finally, use good quality cassette tapes, and have your recorder adjusted to these. The litle extra monay spent on a quality cassette is a good investment because it prevents many problems. o sumption of the unit, including the display, on the recorder. thesound programme or theslide series, the is about 250 mA when it is not powered by another circuit or by a computer via connec- tor Ks, The right track ‘You will need at least a stereo tape recorder to be able to use the timecode interface. The sound channel is then recorded in mono, hile the other trackis used for the timecode signal. The sound programme must be ‘mixed prior to recording the timecode ‘The timecode signal is recorded and played back via the normal cinch-, line- or DIN socket of one of the channels available STOR ELECTRONICS SEPTEMBER 1991 Formore sophisticated work, use a mult track tape recorder which enables you to re- cord or play back four tracks either simultaneously or individually. This has the advantage of allowing the sound program me to be recorded and played back in stereo, ‘The timecode is put on one of the remaining, tracks, and this can be done independently of the sound programme. By the way, you cannot use the (optional) special data format provided by the timecode interface unless you have a multitrack recorder. Another advantage of a multitrack re- corder is that it usually features a simple mixing console. After mixing and recording, timecode signals are recorded on track 4. ‘When thisisdone, you havean absolute time reference relative to the music. This time ref- erence is independent of tape stretch and tracking, Track 3 then remains free to record a special data format Finally, a remark on the system timing: if you want to change the frequency of quartz ‘crystal Xi, remember that the time constant (of R0-C7 must be changed accordingly. Ca- pacitor Co will need to be changed, too. Clock frequencies higher than 5.2 MHz are not recommended, however, since they in- crease the error rate, :FIBER OPTICS — Part 1 Fiber optics are the latest transmission media in communications and instrumentation technology. Fibers are showing up in telecommunications, in computer data communications, and all manner of scientific, engineering and medical instruments. Briefly stated, fiber optics is that technology in which light is passed through a plastic or glass fiber so that it can be directed to a specific destination. If the light is encoded (modulated) with an information signal, then that signal is transmitted over the fiber optical path. HERE are many advantages tothe fiber op vical communications or data link, including: + Very high bandwidth (accommodates video signals, many voice channels, or high data rates in computer communications). + Very low weight and small size, + Low loss compared with other media. + Freedom from electromagnetic interference (EMD. 4+ High degree of electrical isolation + Explosion proof. + Good data security + Improved ‘failsafe capability ‘The uility of the high bandwidth capability ofthe fiber optical data link i that it can handle a tremendous amount of eleetronically transmitted ation simultaneously. For example, it can le more than one video signal (which typi- requires $00 kHz (0 10 MHz of andi pending on resolution). Alternatively. it can bhandle a tremendous numberof voice communi cation telephony channels (which is why you see those advertisements on television). High speed ‘computer data communications capabil possible. ithera few channels can be ope ‘exizemely high speeds, or larger number of low- sped parallel data channels are available on the fiber. Fiber optics are so significant, that one ean expact to see them proliferate inthe communica tions industry for years to come. Indeed, some fu turists are calli ive nationwide fiber optic communications network for dats ‘communications that is the info-age equivalent of the interstate highway system, ‘The ight weight and small sizeof fiber optics. ‘coupled with relatively low loss, makes the fiber ‘optic communications links very advantage when large numbers of channels are contemplated, To obvain the same mumber of channels using coaxial cables or “paired wires’ system would require a considerably larg bulkier snd heavier, infrastructure. Electromagnetic interference (EMI) has been a destructive factor in electronics singe Marconi and DeForest interfered with each other in radio vod economic Joseph J. Carr light source Fig. trials for the Newport Yacht Races just prior to the turn of the 20th century. Today. EMI ean be ‘more than merely annoying. and can cause tragic accidents. For example, airliners are operated more and more from digital computers. Indeed, one airline co-pilot recently quipped (about mod: erm ircrat) that one does not need 10 know how to fly anymore, but one does need t0 be able 10 type on a computer keyboard at 80 wonls per ‘minute. While the pilot's comment was meant 10 bbe joke, it points out just how dependent aireratt hhave become on moder digital computers and intercommunication between digital devices. If ‘radio transmiter radar, or lecrical motoris neat ‘one of the itercommunications lines. then itis possible to either introduce false data or corrupt existing data with potentially disastrous results [Because the EMLis eaused by electrical or msg netic fields coupling between electrical cabl fiber optics (being free of such felds) prodices dramatic freedom from EM Electrical isolation is req strumentation systems either forthe safety ofthe user, oF the health of the electronic circuits con- nected othe system. Forexample, in some indus glass rod 9100811 Tight travelling in glass rods was demonstrated more than one hundred years ago. trial processes, high electrical voltages are used, butthe electronic instruments used to monitor the process are both low-voltage and ground refer fenced, As atest, the high voltage can damage the instruments. In fiber optical systems, itis possible to use an electrically floating sensor, and then transmit the data over a fiber link to an elec trical grounded, low-voltage computer, instru ‘ment or conirol system. “The fact that fiber optics uses light beams, and these are generated in non-contseting electronic circuits, make the fiber optic system ideal for use in switehing and control systems around flamm- able gases or fumes. For example, in monitoring ‘gasoline systems, ori cases where natural gas or esthetic ageats such as ether or cyclo propane’ are used, Regular mechanical switche xt.oF when devomact- medical fr elaysarceither on con ing, and those sparks ean create an explosion flammable gases or fumes are present. A numb: ff operating room explosions in. hospitals oc- ‘curred prior to about 1960, and some gasoline stations have exploded because of acing in elec trical switebes. ‘System security is enhanced because fiber op- se ony ry wl ay ting sidered io angie eb 1970 ne ww lamb sub acre wily sl “TOR ELECTRONICS SEPTEMBER 1991ro ey BIBER OPTICS —Part | Fig. 2, Basie refraction phenomenon, lies is ifficult ro tap. An actual physical eonnec lion must be made tothe system. In wire systems, ‘capacitive or inductive pick-ups can acquire sig nals with less than total physical connection, i. no splice is needed. Similarly, a system is more secure in another sense of the word because the fiber optical transmitters and receivers ean be de signed to failsafe so that one fault does not take down te system. Lrevall a hospital coronary care tnt data system that used parallel wire connec tions between the data output ports on bedside ‘monitors and the cenial monitoring computer at the nurses station, single short circuit in paral- lel data lines would reduce the system to chaos! That is less likely to happen in a fiber optical sys Fiber optics: history and practical applications ‘Thebasic fact of fiber optics i. the propagation of light beams in transparent glass conductor (Fig. 1), was noted in the early 1870s when John ‘Tyndall introduced members of The Royal So- ciety this experimental apparatus. An early. but rot very practical, colour television system patented by J.L. Baird used glass rods to carry the colour information. By 1966, G. Hockham and C. Kao (Great Britain) demonstrated a system in whieh light beams euried data communications via glass fibers, The significant fact that made the Hockham/Kao system work was the reduction of lossin the glass dielectric material to a reasonable level. By 1970, practical fiber optic communica tions was possible. Medicine has made use of fiber opties For more than two decades. Fiber optic endoscopes can be passed imo various etices of the body, either natural or surgically made, to inspect the Ierior of a patient's body. Typically there are two bundles, one for viewing and one for passing alight froma (misnamed) ‘cold light source into the body. For example, gynaccologists can in- spect and operate on certain internal organs in fe ‘males using a laprascope introduced through a ELEKTOR ELECTRONICS EMBER 1991 Fig. 3. Retraction involved In Total Internal reflection’ = Ne mater Fig. @ Waveguide a jogy to fiber optics. band-aid” incision in the ablomen. Knee sur _geons can use a iber optic arthroscope to perform nearly miraculous operations on the human knee with far Tess trauma than previous procedures (Other physicians use fiber optic endoscopes i in- spect the stomach and gastric wack. A probe is passed through the mouth or nose, down the eso: ‘hgus into the stomach so that tumors and ulcers ‘ean be inspected without resort to surgery. In more recent times, miniature TV cameras using CCD arrays have been made available, with the fiber optics carrying the light into the stomach, Fiber optic inspection is used elsewhere than Jn medicine. For example, I recall an advertise ‘ment for a septic tank service company that used. fiber optics and television 10 inspect the tank: similarly for plumbers. Other industrial and residential services also use fiber optics inspect areas thal are ither inaccessible or too dangerous for direct viewing, Before examining fiber optic technology, itis useful todiscuss some ofthe basis of optical sys tems as applied to the fiber opti system. Review of some basics ‘The index of refraction (n), or efractive index of ‘4 material, isthe ratio of the speed of the Tight wave in a vacuum to the speed of the light wave in the material (e.g. glass plasti, water). For practical purposes, the speed of light in ai is lose enough to the speed in a vacuum to be con sidered the same. Mathematically, the index of relia uw where is the speed of Hight in a vacuum (approx. 3x10" mA) v8 the speed of Tight inthe mectum, Refraction s the phenomenon in whicha light ray changes direction asit passes across the boundary surface (or “interface’) between two mediums of differing indices of refraction njzus). Consider Fig. 2 in which two materials, with indices of re fraction n and m respectively. Consider incident Tightray A, approaching the interface from the less dense side (7972). AS it crosses the iner- face it changes direction towards a line normal {ce at right angles) to the surface. Conversely, Tightray (B) approaches the interface from theGENERAL INTEREST snore dense side (n>) bn this e38e te light ray is similarly refracted from is original path, but the direction of refraction is away from the oral ine Inrefractive systems the angle of refraction is. function of the ratio ofthe owo indices of refrac tion, ie, it obeys Snell’ Law ‘nysin®,= m2 sin, el ny _sin@,. al sin, ‘The panicular case which concerns fiber optics is where the light ray passes from a more dense me dium toa less dense medium. We ean use either 2 water to air system, ora system in whieh wo diferent glasses, with dissimilar indices of re- fraction, are interfaced. This type of system was addressed in Fig. 2 by ray B. Figure 3 shows a similar system with three different lightrays (ray ‘A, ray B and ray C) approach the same point on the interface from three different angles (ivan respectively). Ray A approaches af suber cal angle, soit wil spit into two portions (A" and A"). The reflected portion (A") contains a rela- tively small amount of the original light energy, and may indeed be nearly indiscernible. The ‘major portion of the light energy is transmitted across the boundary, and refrats at an angle 7 in the usual manner. LLighiray B, on the other hand, approaches the interface a the eritical angle, ry’, and is refracted along a line that is orthogonal to the normal Fine, ie, it travels along the interface boundary su face. This angle is normally labelled Cin optical textbooks. rally, ray C approaches the interface a an angle greater than the critical angle, ie, a super- critical angle. None of this ray is transmitted across the boundary. but rather itis turned back ito the original media; i. itis subject to total inerna reflection (TIR). tis the phenomenon of ‘otal internal reflection that allows fiber optics to work. Fiber optics “The fiber optic is somewhat similar to micro- ‘wave waveguide, and an understanding of wave~ fuide action is useful in understanding. fiber plies. A schematic model of a fiber optic is shown in Fig. 4. A slab of denser material (1) sandwiched between two slabs of a Tess dense ‘material (72), Lightrays that approach from a supercritical angle are totally internally reflected from the wo interfaces (nny and yn) Al- though only one “bounce is shown in our illus- tration, the ray will be subjected 10 successive ‘TAR reflections as it propagates through the 1, miteral, The amount of light energy that iste ected through the TIR mechanist sof the order ‘of 99,9 per cent, which compares quite favour- ably with the 85.96 per eent typically Found in planar mirrors. Fiber optic lines are not rectangular, but rather ace cylindrical s shown in Fig. 5. These compo- rents are called clad fer optics because the den: caging lading (N2)_ core (1) rye er ctical angle optical Fig. 5. Light propagates in fiber optios by repetitive total internal reflection. ser inner core i surrounded by a less dense layer called cladding, Shown in Fig. $ are two rays, ‘each of which is propagated into the system such thatthe critical angles are exceeded, These rays will propagate dawn the eylindical optical fiber ‘with very lite loss of energy. There are actually ‘two forms of propagation. The minority tor (Fig. 62), caled meridional rays, ae easier to un- derstand and mathematically model in textbooks Ipocause all rays Hie in a plane with the optical avis. The mote numerous skew rays (Fig. 6b) fol- low a helical path soare somewhat more difficult to discuss (Ref. 1), “The light acceptance ofthe liber optic (Fig. 7) is a cone shaped region centred on the optical axis. The acceptance angle ais the critical angle For the transition from air (n =n, to the core ma terial (= ng). The ability to collect Tight Is dh recily related to the sizeof the acceptance cone, and is expressed in tems of the numerical per ture. NA, which is NA=sin®, (ar “The refraction angle ofthe rays internally, across the airs, interface, is given by Snell's law rae {In terms of the relative indices of refraction be- tween the ambient environment outside the fiber the core ofthe fiber and the cladding material he humeral apenture is given by % is) NA=sin@,= | Yin? =e (61 If the ambient material is air. then the numerical aperture equation reduces to m optical _| ‘meridional transmission skew transmission Fig. 6 ') meridional propagation; b) skew propagation. ELEKTOR ELECTRONICS SEPTEMBER 1991Fig. 7. Cone of acceptance of @ Fig, 6, Transmission modes In fiber optics. Intemally. the angles of reflection (a) and az), at the critical angle, are determined by the relation ship between the indices of refraction of the 1wo ‘materials, n) and O. (si Fig. 8. a) input ight pulse; b) output pulse Is dispersed in time. TOR ELECTRONICS SEPTEMBER 1991, ‘Typical fiber optic components have numerical fpertures of 0.1 to 0.5: typical fibers have & diameter D of 25 um 10 680 yim, The ability of ‘o collect light is proportional to the square of the numerical aperture: G=(NAxD} io} Intermodal dispersion ‘When alight ray is launched ina fiberoptic it can {ake any of a numberof different paths, depend- ing in parton its angle of arival (Fig. 8). These paths are known as transmission modes, and vary From very low order mades parallel the optical axis of the fiber (ray A in Fig. 8), tothe highest, ‘order mode close tothe critical angle (ray C); in ‘dition, there are very large numberof ray in between these two limits. An important feature of the different modes is that the respective path lengths vary tremendously, being shortest with the low order modes and longest with high order ‘modes. Ia fiber optic has only a single core and single layer of cladding. itis called a step index fiber because the index of refraction changes ab- tupily Srom the coe to the cladding. The number ‘of modes, N,that can be supported are given by: Ne tEDINAL2D (10) ‘Any fiber with a core diameter, D, greater than shout ten wavelengths (10R) will support a very large number of modes, so is typically called a ‘multimode fiber. typical light beam launched into such a step index fiber optic will simulta neously find a large number of modes available {ot This may ormay notaffect analogue signals, s but has a deleterious effect on digital signals called intermodal dispersion. Figure 9 illustrates the effect of intermodal dispersion ona digital signal. When a short dura- tion light pulse (Fig. 94) i applied toa fiber optic that exhibits a high degree of intermodal disper sion, the received signal (9b) is smeared, or s+ persed’ overa wider area, At slow data rates this mg>m5> 14> E> <2 13) “The overall index of retraction determines the mu ‘merical aperture, ond is taken as an average of the ifferent layers With graded fibers, the velocity of propage- tion ofthe light ray in the material is faster in the layers away from the optical axis than im the lower layers. AS a result, a higher order mode ‘wave will travel Faster than a wave in a lower order. The numberof index fiber are: wes available vo the graded _ (ED INAJ/2)" Mi 4 ha Some cables operate jn a critical mode, desig- ‘nated HE, (to borrow from microwave termine- logy) in which the cable is very thin compared with malinodal cables. As the diameter of the core decreases, so does the number of availableFig. 10. Effects of dispersion on digital signal bendwidth: a) original data signal; b) Tight pulse inputto fiber system: c) dispersed light pulses overlapped NID>N2>NS>NQ>N5,..Ni 91009221 Fig. 11. Graded index fiber. smodes and eventually the cable becomes mono modal; Ifthe core gets down to 310 5 microns, then only the HE, mode becomes available. The critical diameter required for monomoxal oper 2ah . nINAl 15) Because the monomodal cable potentially re ddaces the number of available modes, it also re ddaces intermodal dispersion. Thus. the ‘moomode fiber is capable of extremely high data ates or analogue bandwidths. Next month. In the second and final instalment of this article we will take a look at losses in fiberoptic sy tems, fiber optic communications and some ofthe basic driver and receiver circuits neeied to make fiber optics work. a Reference: 1. “Optical-fibve communication” Elekior Etec- tronics February 1991 CORRECTIONS Wattmeter April 1991, p. 32 ‘With reference the circuit diggram, Fig. 1, the righthand terminal of the lower section of switch S> should be connected to the circuit ‘ground. This point is indicated by a dot. In the adjustment procedure given on page 35. the references to presets Pasnd Ps have beer transposed. Contrary to what is sated, Pa sets the VY offset, nd Ps the x offset. The functions ofthe presets are shown corr diagram, Fi. | “To improve the accuracy ofthe instrument connect Rs direct tothe circuit ground instead of junction Re-R7, Finally, all ctcuit board tracks tarrying mains current most be strengthened With 2.5-mm? cross-sectional area solid copper wire if currents higher than about A are measured Ss 80C32/8052 Single-board computer May 1991, p. 17-23 ‘When a CPU type 8031 or S052AH-BASIC is used, ICI, 1C2, IC3, and ICHICI2 must be ‘T4HCT types. Jumper B is erroneously reffered t0-as Bro in the text under "On-board EPROM programmer’. Contrary to what is stated, this jumper must be Fited only when an EPROM is to be programmed — for all other use of the SSBC itmust be removed. Also note that jumper 1B may only be fited when the programming LED is out Sequential control July/August 1991, p.61 ‘Motor M should bea d.c type, not an a. type as shown in the creult diagram. Digital phase meter June 1991, p.32-39 In Fig. 5, the switch between input ‘A’ and [C1 shouldbe identified $1", and hat between input Band IC2 $2" Switch Ssisan on/off type.not ‘4 push-button as shown in the diagram. Capac tors C2 anid Cs sre shown with the wrong po- lanty. The component overlay of the relevant printed-cincuit board (Fig. 8) i all right, Universal NiCd battery charger June 1991, p. M19 ‘The parts list on page 19 should be corrected to read (C7 = 2200, 25V ‘When difficult 1 obtain, the BYW20/100 (Ds) may be replaced by the BY229, which is rated a6 A. ‘The lext under the heading ‘Calibration should be replaced by. 4. Connect a multimeter between points G and H on the board, and adjust Ps until the measured voltage is I V lower than the voltage ‘on the battery terminal. MIDI program changer April 1991, p. 14-17 ‘The contents ofthe EPROM should be modified a8 follows: address data ‘008C ES 0c7 so 0c: cB oc 75 00CA 7B 0CB 2 nce 0 ocd D2 00CE a 00CF 2 0D0 80 opt a Readers who have obtained the EPROM ready programmed through the Readers Services may return it to abiain an update Electronic exposure timer March 1991, p. 31-35 Please add tothe pars list on p Clo=33 pF 2 Augmented A-matrices May 1991, p. 42-43 "The drawing below was erroneously omitted in the lef-hand bottom comer of page 3. igi vl ve 1058-14 ELEKTOR ELECTRONICS SEPTEMBER 1991,MEASUREMENT TECHNIQUES — PART 7 High-frequency measurements HIE test methods described so far inthis series apply only to low-frequency (LF) equipment. In circuits carrying high-fre- ‘quency (HF) signals, the effects of the mea suring equipment become an important fac tor. First of all, of course. the test instru- ments used must be designed for HF mea: surements. Also, unless suitable measures are laken, tests may result in reflections and ra dation of HF energy, which will invalida the test result Measuring HF voltages The bandwidth of virtually all inexpensive voltmeters makes them unsuitable for me suring HF voltages. Moreover. the capaci tance of the test leads and that of the input circuit of the instrument will adversely af- fect the measurement. Also, in case of in- struments with a not very high input in ppedance, the circuit Q wil be derated. For accurate measurements of HF valt= ages, special HF peak rectifiersareused, After rectification of the voltage. a standard DC instrument indicates the peak value of the ‘measurand. It is then not possible to me sure the rams. value of non-sinusoidal volt- ages. The rectified voltage should be fed to the test point in a direct way 10 ensure that the leads from the rest point co the display sec- by EP. Zantis tion have no detrimental effect on the mest surement, In. many cases, therefore asshown in Fig. 61 isused. The rectifier should have the smallest possible junctionca- pacitance and this is met by germanium lypes. The wide-band properties of hisdiode determine the overall bandwidth in the first instance. The coupling between the tip of the probe and the diode is capacitive, so that any direct voltageat the tips blocked. Enrors in measurement are normally caused by in correct earthing. The earth point ofthe probe should be as close as possible to the test point and be connected to the earth of the ‘equipment on test If the frequency is higher than about 1 GHz, even these measures are not suificient, since part of the HF energy is then radiated by various components and the wiring, which. of course, makes the mea- surement invalid or even impossible, For voltage measurements at frequencies up t about 2 GHz, coaxial insertion probes are used whose construction obviates the unwanted radiation, Alignment of tuned circuits Relative measurements suffice jn the aligs ment of tuned circuits —see Fig. 62. Therefore if a probe is not available, the measurement may be carried out by connecting a low- 61 value capacitor, C, in series with the instru: ment. This reduces the effect of the input capacitance of the test instrument on the tuned circuit. The circuit is aligned by justing C2 for maximum or minimum ection of the voltmeter. In direct voltage ‘measurements, the effect of the meter is te- duced by connecting a high-value carbon resistor (which has a high parasitic induc~ tance) in series with the meter. Asin a probe, it is advisable to connect the series capaci- tor or resistor close fo the test point twill be realized that these simple meth- ‘ods can only be used for alignment and nor for absolute measurements. ‘The effect of the instrument on the cir ‘cuit on test can be reduced further by the absorption method, for which a circuit as shown in Fig. 63 is used. Coil Ly is brought close to the inductor of the circuit on (est. Mutual inductance ‘causes a potential across L, that may be fed to an oscilloscope or, after rectification, 10a voltmeter. If, owing to sereening of the in- ductorin the tuned circuit, inductivecoupling is impossible. wiring that carries the HF voltage may be touched with the tip of the probe. Capacitor Cis adjusted formaximum voltage across the tuned circuit, after which the frequency can be read on the oscitlo~ scope of, if this is provided, the scale of C} 63 ELEKTOR ELECTRONICS SEPTEMBER 199164 65 (A) (B) To really minimize theeffect on the tuned circuit, itis advisable to amplify the voltage obtained by inductive coupling. A suitable circuit is shown in Fig. 64. Coil Ly may be 4 plug-in type to make alignment over a ‘wide range of frequencies possible. The tap ‘on the coil ensures matching to the low-im- pedance transistor input, Dip meter ‘The dip meter is still the most poputar in- strumeat fr the radioand television amateur. Although tis basically asmall HF tansmiter, the meter functions in'@ manner similar to the absorption method described earlier. “This active instrument makes the effect on thecircuitontestevensmaller,and, ofcourse, greater sensitivity is obtained. ‘As inthe absorption meter, the frequency is read on the scale ofthe variable capacitor, while the meter serves merely as a maxic ‘mum oF minimum indicator. ‘Apart fromits uses resonant-frequency ‘measuring instrument.the dipmetermay also be used as a signal source in signal tracing in HF circuits The values of the coil and variable ea- pacitor are readily calculated with the fol- lowing formulas if the value of one and th frequency (in MHz) are known. (why (pF L=25400/C/2 AOU A suitable circuit based on a field-effect, [is shown in Fig. 65. The ignal is generated across the gain-drain ‘on, The display instrument is con- nected either in series with the gate bias re- sistor as in A or across the source as in B. ‘The sensitivity of the circuit is set with the potentiometer. ‘When the circuit is used as an absorption meter and the display instrument is con- nected asin B, the 1002 potentiometer must be disconnected to ensure that the full sen- sitivity of the meter is available. Winding data forcoilsare giv in Table 1. Measurements on transmission lines The characteristic impedance of a transmis- sion Tine is the value of load resistance that ‘enablesmaximum powertransfer lromsource to load. The characteristic impedance of a transmission line, Z<, is given by Ze=ViLulCa. where isthe distributed inductance perunit, length and Cy is the capacitance per unit length, Practical values of Z, vary from 100 2t0 1000 £2 for parallel wites and 10 2 0 180 Q for coaxial cable, ‘When a transmission line, suchas that con- necting.a transmitter to its antenna, i termi nated by an impedance different from itsehiar~ acteristicimpedance, some ofthe forward sig- nal wave is reflected back. The reflected wave mixes with the forward wave, and the resultant amplitude at any point of the trans mission line is the algebraic sum of the arn plitudes of the two waves. The nodes and antinodes do not move relative to the transmission line, that i, the are stationary and the waves are called stand- ing waves. An important consideration in transmission line and antenna design is the standing wave ratio—SWR. ‘The SWR may be ascertained with the aid of an SWR meter as shown in Fig, 66. ‘The pick-up consists of three conductors of which the centre one isthe core of the trans mission line, while the outer two, called re flected pick-up wire (top) and forward pick up wire (bottom), are spaced equally from this. Usually, the pick-up unit is made as a pattern on a printed-cireuit board: a strip line. The $1 @ resistors are, in this ease. equal to the characteristic impedance of the transmission line (if this were 75 Q the re- sistors would also have to be 75.Q). In op- eration, any RF signal on one of the pick-up wires will be rectified by the relevant diode and then applied as DC to the associated in- dicator, “To ensure that the indicators can be used independently of heactual power. theirscales are calibrated in relative values. That in the forward pick-up circuit, F, shows the rela livepower asa percentage, whereas the other shows the SWR. The ratio is given by the formuls SWR L+U pI, where U isthereflected voltage. lisassumed 66 67 ELEKTOR ELECTRONICS SEPTEMBER 1991thatthe forward voltage has been set to 100% of the relative power. Optimum performance is, of course, established when U, =0. A ‘good antenna system has an SWR that varies between I:1 and 1:1.2, Poor systems have ratios greater than 1:1.5. Most SWR meters do not give a reading beyond 1:3. Measurement of field strength For tuning a transmitter or testing an an- tenna system, a field strength meter is al- most essential, In its simplest form, such an instrument consists of abasic receiverin which the headphones have been replaced by a sen. sitive moving-coil meter. A possible circuit is shown in Fig. 67. ‘The meter is zeroed with the 50 kG po- tentiometer. For relative field strength mea~ surements, the antenna circuit need not be tuned: setting the tuned circuit to the centre of its bandwidth is sufficient. Other numberof different inductors may beselected with the aid of switch S,, which makes it possible to tune and test transmitters over a Wide range of frequencies. {MEASUREMENT Table 1 Range Frequency Inductance —_—No.. of Wire dia. (MHz) (uH) turns (mm) 1 12-24 300 76 os 2 28-47 82 36 08 3 46-92 a1 18 0s 4 82.166 65 8 10 5 162-326 17 5 10 6 222-448 09 3 20 7 384-776 08 a 20 8 745-150.0 0.08 1 20 ‘The coll diameter for ranges 1-8 must be about 38 mm (1.5 in.) and that for range 7, 25 mm (1 In), ‘The col for range 8 is shaped like a hairpin, 0 mm (2 in.) iong; wire separation is 10 mm (0.4 in}. Finally Instrumentsare usually multi-purpose: fr i stance, its quite common foran SWR meter to be combined with a field strength meter and a power output meter. Next month's instalment will deal with “Measurements in digital circutts ASYMMETRICAL-TO-SYMMETRICAL CONVERTER T OFTEN happens in electrophonics(elec- tronic music) that hum-and-noise loops occur when two or more different instru ‘ments areintercoupled asin Fig, 1a.Some mu sicians play with their lives by covering the ‘earth pins with insulating tape to get rid of the hum. This is, of course, not only very stupid, but also highly dangerous. A safe and certain method of getting rid of these loops is offered by the converter Whose circuitisshowninFig.2.Theconverter is connected between two instruments as shown in Fig. 1b to provide electrical sepa- ration of the instruments. Make ure that the ‘isolated’ instruments are not drawn into new hum-and-noise loops througha common enclosure: each and every instrument must be isolated from the enclosure. Theconverter resolves another problem also. On cost grounds, many commercial instruments have only a relatively high im pedance output, Even pick-ups often suf- {er from this, When long connecting leads are used, or the signal is divided, or a fol- lowing instrument has a low impedance input, noiseand hum are the result and the quality of the music suffers, The converter has a high impedance input, and two low impedance outputs (one not isolated). ‘Circuit description ‘The input signal arrives at Ky and from there itis applied via C; to IC;, which is arranged ELEKTOR ELECTRONICS SEPTEMBER 1991 by M. Eller la 4 wena WS {( mm aise ) } 1b ning eoRsDIeASYMMETRICAL-T0-SYMMETRICAL € /ONVERTER, y as an impedance converter. The TL. 061 was designed specially for use in battery operated Circuits. Its input impedance is determined by Rrand Ry ‘The output of ICy is taken via C6 to K: which is the non-isolated output that may be used, as in Fig. 3, as the input to an am- plifier. The output of IC2 is also taken via Ry 10 the heart of the converter: transformer Te which must * operate from a 600.0 source: ‘© have a transformation ratio of 1:1 * haveallinear frequency response ov 3 stag Converter ising Console range 20 Hz to 20 kHy; # beableto handlesignal levels of 600 mV + ‘These characteristics are met by a number of available commercial types; final choice de pends on the operational requirements. The size of the transformer core determines the undistorted signal level at low frequencies: the larger the core, the better the perfor- ‘mance, but, unfortunately, also the higher the price and the size of the converter Low-frequency distortion maybe reduced by incr e value of Rs, but this will be to the detriment of the signal level, Normally, the LED is connected to the supply line viaa series resistor. Itthen draws a current of about 10-20 mA, which is too tery operation. T contained in IC;, reduces this t these values, When S; isclosed, Cyischarged slowly via Ry until Ds conducts, Thereis then ‘high at the input of the gate-chain, which is repeated through the fourgates, whereupon it switches on transistor TI, resulting in the LED lighting, Since the value of R> is too high for the LED current, the diode draws its energy from Cy After a while, the vollage across Cy be comes too low to energize the LED, where- upon zenerdiodeDswitches offand thelogic levels atthe gates are reversed. The circuit is ‘once again in the output state, until C4 has been charged anew. The LED will flash in a rhythm determined by time constant Rr-Cy, Resistor Ry also serves to conserve en- egy by reducing the current drawn by IC; and changing the switching threshold from ow to high. These measures ensure that a 9 V battery(PP3) gives roughly 400 hours operation. The average current drain at full drive is about 1 mA. The battery should be replaced when its voltage under load drops sing t 1 gate-chain, out 1% of ELEKTOR ELECTRONICS SEPTEMBER 1991below 6 V Construction of the converter is straight forward if the printed-circuit board shown in Fig. 4is used. As mentioned earlier, make sure that the converter is fully isolated when PARTS LIST R2 = 33k R3= 109 R4, R6, R10 = 1 ko. R5=1 MQ. R7, RB = 470 k2 Ro = 680 0 Rit =1.5k0 R12 =10k2 Capacitors: C1 = 1 pF, 16 V, radial C2= 100 nF €3, C6 = 22 uF, 16 V, radial (C4 = 47 uF, 16 V, radial C5 = 470 nF Semiconductors’ D1 = LED for chassis mounting D2 = zener, 2.7 V, 400 mW Tt = 805478 101 =4011 IC2 = TLoe1 K1, K2 = 6.3 mm audio socket for panel mounting, mono, insulated (LR socket for panel mounting 9-V battery (UK: PP3) 600 @ transformer — see text Enclosure to personal requirements PCB 910072 Front panel foil 910072-F K3 al 20) 20-05 20-9 3 no) ae Bee “|S? | oo ‘SYMM. OUT 8 ASYMMETRICAL - TO - SYMMETRICAL CONVERTER ourpur D POWER ®D input o. ELEKTOR ELECTRONICS SEPTEMBER 1991REMOTE CONTROL ICs MV500 AND MV601 (Plessey Semiconductors) (O integrated circuits from Plessey, the M300 transmitter and the MV6OT re- ceiver, allow remote control systems to be built from a minimum of components. The two ICs have been used already ina number of projets carried in this magazine (Ref. 1, 2), The infa-red remote controls described there are reliable, marked by low cost, alow component count, and the absence of quartz cxystals and adjustable inductors. Note, however, that the MV601 receiver requires Cone IC, an SL486, to be added when used to buile an infra-red remote control system. Data coding and transmission ‘The inputs of the transmitter are supplied ‘with the information to be conveyed to the receiver. This information is provided in parallel form by, for instance, a keyboard. As shown in Fig. 1, the parallel data is trans- mitted serially. The transmit operation is in- itiated by a start’ pulse with a synchronization pause, after the key de- bounce time, fy. Then follow the five da- tabits. The data packet is transmitted as long as the key is pressed. When the key is re leased, the data packet is always completed, CLAY TE 4 1024 CLOCK C¥ELES “I = mite ices os L0G 0 TE tg = 31 £ oo UNI Lu. lJ PULSE WOTH = 8CLOCK CYCLES oe IL sun rows) eat ras Se re sss rose rae Fig. 1, Pulse/pause modulation applied in the MV500 transmitter. ELECTRICAL CHAnACTERIsTiCS Vigt costar ec oternae ste) MOTH EST ne a fae sev wracteristic Pin sae Conditions Cy Seeeerae ‘in| typ [mor _| Yt gs es@0" | operaung susp current a os 2 ma [ercutigs rf SEE | sunray eae . cay] 2 | ww [umwenone, ose sweet y | se | 0 | 200 | mn |seo 26. ¥on =v a wo | a | | ma |v =a vo = MPI | heyboad contact esstarce ae asym vousesy ven 100 = | we S| oscnaa Froasency 1817 | a0 v0 | ane Fig. 2. Main data and pinning of the MVE00 transmitter. ELEKTOR ELECTRONICS SEPTEMBER 1991REMOTE CONTROL ICs MVS00 AND MY601 =| om ates | ts I. T. Fig. 3, Internal diagram of the MV5O0. and a stop pulse is affixed. ‘The modulation of the information car- rier (je, infra-red light) is based on pulse/pause (mark/space) modulation ‘This means that the five databits and synchronization bits are encoded by means of their length (‘mark’) and the position of the pauses (‘space’). A short pause of two times the bit transmission rate, T, indicates a logic 1; a slightly longer pause of 37 indi- cates a logic 0; and a pause of 6T indicates synchronization (see Fig. 1) In this system, the logic levels are thus determined by the length and the position of the pauses in the serial datastream. The pulses serve no other purpose than to set the pauses apart, and ‘can, therefore, be relatively short (fy=17 us) and of afixed length. \s compared with pulsewidth modula- tion, the pulse/pause encoding system is less susceptible to interference, and more ef ficient in regard of battery power —theaver- age current drawn by the actuated transmitter is smaller than 10 mA. ‘The length of the pauses is not constant, not even when they have the same logic con- tent. Thisis caused by the three transmission rates, A, B, or A+B, that can be set on the transmitter and the receiver. The use of dif- ferent transmission rates on transmitter/re- ceiver sets allows up to three remote control systems with 32 channels each to be used in ‘one room. This gives a total of no fewer than 96 remotely controlled channels. Table 1. Transmission rate settings AB clocks t to ts 0 outputs disabled 1 2048 4096 6144 12208 ° 6144. 1 1024 1536 9072 024 2048 3072 st2 Transmitter MV500 ‘The most essential information on the MYS00 remote control transmitter isgivenin Figs, 2and 3, IC pins 2to 12 form the channel selection inputs, which allow up to 32 chan- nels tobe used with the aid of a(switch-) ma- trix of 8 rows and 3 columns. A shift register converts the parallel information intoa serial datastream. A sequence control block in the MVS00 operates together with the output control to ensure the correct pause position and timing, The required transmission rateis set at pins 14 and 15. When both inputs are ‘made logic low, the ICis disabled. The clock frequency is furnished by a oscillator that operates with an external ceramic resonator (fundamental resonance frequency between 400 kHz. and MHz; max. tolerance 5%). As shown in the diagrams, the power control block is connected to the row decoder. In quiescent mode, the associated inputs are held at ground potential. The os- cillator is then disabled, and the IC is switched to its power-down state in which the current consumption is reduced to 2 uA oF $0, On detection of a high level at one of the column decoders, the power controller actuates the entire IC, and the transmitter starts to operate (provided, of course, one of Input cireits| Main cireuit ae ersoet NS Output circuits Fig. @ ELEKTOR ELECTRONIC SEPTEMBER 1991 ‘Options for the input and output circuitry of the MVEOO.APPLICATION NOTES: ELECTRICAL CHARAC! Tana = C10 70 TERISTICS, Test condition unis otherwise states) Yoon ea Sv t0 = 5 8 characteristic Value Units onaitions yp | we INPUTS OSCIN RATE R RATE 8, MOM "UAT, 08% now mg vtage a Inet ow tage Inpa gh wage ‘ABSOLUTE MAXIMUM RATINGS Supply voltage Voo Inputvorage Operating Temperature Storage Temperature Output sink and Source Gren Hum CLEAR RATER, RATES v Veo #0 3V to ¥55:03¥ ero = 10 sste1o + 125%| Soma Outou ng te ou tpt ahaa cu ‘rests Yorge abn) 10 5 ws Vi = Vis 03¥ wap +039 a0 n= Vie 03V10¥g0 +03 ort * Von 030 Fig, 5. Main data and pinning of the MV6OI receiver. the transmission rate inputs is made logic high; otherwise, the IC remains off). “The main application circuits of the trans mitter are given in Fig, 4. As already men- tioned, very few external parts are required, — the basic circuit works with an inexpens- ive ceramic resonator and two capacitors. Inprinciple, there isa choice between two types of input circuit. Of these, the most fre «quently applies is probably a keyboard ma trix (option B). Alternatively, the code to be transmitted may be supplied by a computer ‘or another digital circuit (option A). When the inputs are connected to an 8x by thediodecan be made narrower by fitting the device with a small reflector of, to achieve even greater directivity, small ens. ‘The current limiting resistor may be omitted when the full transmit power is called for. Note, however, that this inereases the cur- rent drain, and shifts the operating point of the IR diode beyond that of maximum effi- ciency. Remember, the peak current through the IR diode can be as high as 2 A, although this current flows for a couple of microsec- ‘onds only. The IR diode shown in the circuit diagram is a type with a peak pulse current ‘of 25 A. Options D and E show suggested circuits for use with a 9-V battery block and a5-V power supply respectively. Receiver MV601 Figures 5 and 6 fill you in on the receiver IC, the MV601. The serial data applied to the input of the IC are converted into parallel to give an exact copy of the encoded data at the transmitter side. The block marked ‘Noise ‘matrix (option B), the 5-bit parallel code is composed from bits A, Band C supplied by the row decoder, and bits D and E supplied by the column decoder. Note that bits D and E must be inverted because the relevant 1C inputs, pins 11 and 12, are active low. The other data inputs, pins 2to 9, areactive high patra feel | J ¢ [Bowne ‘At the output of the MV300, the encoded 12] ssncouras [a4 (L Soestow mon || § Lomo serial data is modulated on toa carrier, In its Seas es ae Shaplet fom this carter many be 2 die at Ee pane salage cava wae weno foi As en 4 Terms oh oid ee lesen 7 Elmers | 1a ts Me fete alnpe wie remoe contol agen med [Lee H feplonch Note however, fat his range a dl ent repines te tnomive we operest | gM few) fy SV beret thasamaspiy vollazcan thare- a ee ceiver. For wireless control, the choice of surrivott—e vw arene counren [™] perecros | a {ni Se ghtan hearer bcbvioun Since Guate soe the output ofthe MVSQ) isnot capable of an drtorgon nbaredeahte nde ned, we seals nett power dives Totctesoctie Fngeofthetagonter telR beamemtted Fig 6 Block diagram ofthe WV601 ELEKTOR ELECTRONICS US SEPTEMBER 1991 ELEKTOR ELECTRONICS SEPTEMBER 1991Input circuits | Main circuit Output circuits option option a Loscioeosest sue, Fig, 7. Application options of the aveot mveot OQ0G0 0 PPM INPUT FROM SLA36 008 18 sv D1 C1 a1 ar Ea 2 cz )B2 a2 DATA READY Fig. 8 96-channel remote control extension. ELEKTOR ELECTRONICS SEPTEMBER 1991 Detector’ is capable of recognizing and sup- pressing interference. Similar to that in the transmitter, the clock is provided by an oscil: lator. After being processed in a number of latches, counters and timers, the data are fed tothe output latch, and from there to the IC ‘output pins, A-D, pins 11-15. Each latch out- put can source 45 mA and sink 26 mA. The internal timers and the output latch may be reset with theaid of alow pulseat the PowER CLEAR input (pin 2). This input is connected toan on-chip 150-K2 resistor, so thata single capacitor is sufficient to achieve automatic resetting of the IC on power-up. ‘The logic level applied to the NOMEN: TARY/CATCH Input, pin 5, determines whether the data are kept in the latch (level =D), oF erased (level = 1) when there is no valid code at pin 1. When the OUTPUT ENABLE pin is made low, the dataword is fed to the IC outputs. A high level switches pins 11 to 15 to the high-impedance state. This allows the outputs of two or more MV DIS to be connected toa databus, The DATAREADY out ‘put goes low when a valid word is present. The main application options of the MVo01 are shown in Fig. 7. Like the trans mitter, the receiver contains a small number ‘of components only — the simplest appli tion circuit works with a single ceramic res- ‘nator and two capacitors. A resistor may be connected in series with the resonator to en sure resonance at the fundamental fre- quency Depending on the transmitter configura tion, thereareseveral options for the receiver input circuit. No additional components are required for the tvo-wire remote control system, option F, where pin 1 ofthe receiver is simply connected to pin 1 ofthe transmit- ter. The wireless remote control, option G, is more complex because the MV601 requires fn external infra-red preamplifier such as the SL486 (see Refs. 1 and 2), ‘The outputs of the M601 may be comAPPLICATION NOTES DATAREADY( veot O nected to driver transistors (option H), or solid-state relays (option 1). It should be noted, though, that these output configura- tions allow only 3of the 32 possible channels ‘910028 -20 Fig, 10, Keyboard select extension for the ansmitter. DATAREADY2 PPM INPUT FROM SL486 DATAREADY3 aon ro Fig. 8, S6-channel remote control receiver Implemented on an 6-bil microprocessor bus. tobeswitched, since decoding not applied. When more than 5 channels are required, simply add one or two L-of-16 decoders, eg, the CD4514 (option). Receiver option K works with the pre= viously mentioned computer-controlled ver- sion of the transmitter, option B. The output register is sot to latching (pin 5 held logic Jow), and the outputs of the MV601 are con- nected direct toa microprocessor input port A simple handshaking arrangement may be implemented between the IR receiver and the microprocessor with the aid of the DATA READY and the OUTPUT ENABLE lines. ‘A single receiver has a maximum cecod- ing capacity of 32 channels. This figure can be trebled by using three sets of transmit ter/receiver, each wired to operateatits own, transmission rate. As shown in Fig, 8, the complexity of a 96-channel remote control systemis still within reason, mainly because the clock forall three receiversis supplied by a single oscillator. Also, a single reset line is used for the complete receiver system. Figure 9 shows how to implement a 96- channel receiver when an 8-bit microproces- ELEKTOR ELECTRONICS USA SEPTEMBER 1991 sor bus is available. Here, the DATA READY ‘output is connected to the OUTPUT ENABLE iver is not too complex either when it is extended. The two transmission rate inputs are held logic high with pull-up resistors, When a key is pressed, one of the inputs goes low, and a different MV6OL is addressed. It is almost impossible to tamper With this system since the transmitter re- mains off when two keys are pressed simul: taneously. Finally, the circuit in Fig. 10 shows an input configuration that allows you to switch between three keyboards, i References: 1. “Infra-red remote control", Elektor Elec- tronics September 1990, 2. “Dimmer for halogen lights", Eletor Elec tronics April 1991 Source: Satellite & Cable TV IC Handbook, Publica- tion P-S.2020, October 1988 (Plessey Semi- conductors). PLESSEY SEMICONDUCTORS: Uk head office: Plessey Semiconductors Ltd., Cheney Manor, Swindon, Wiltshire SN2 2QW,, United Kingdom. Te 0739) 36251, fax: (0795) 616765. North American head office: Plessey Semiconductors, Sequoia Re: szarchPark, 1500 Gree ils Roa Sats California 95066, US.A. Tele- hone 48) 48 250, fa (48) 438 623. ELEPKTOR ELECTRONICS SEPTEMBER 1991ZAP51: AN 87C51 PROGRAMMER N my first article (Ref. 1) 1 described the roblems with developing 87CSI single-chip microcontroller applica tions, especially the software. In the second article (Ref. 2), I described a low-cost in-cr cuit emulator (ICE) that provides the hob: byist with a route into simulation conventionally the corpor- ate/ professional developers only. One final problem remains for the hobbyist. Having developed the software in situ in the targe environment, the single-chip microcontrol ler needs to be programmed. Most low-cost EPROM programmers have 28-pin sockets and can not, therefore, program the 40-pin 87C51 The ZAPSI described here isa minimum cost programmer that uses the ICES1 hard province of wareas the program controller. The diagram in Fig, 1 describes the hardware. ‘The ZAP51 consists of a socket, Ps, to plug the ICES1 into; a voltage converter cir cuit to generate 12.5 V from the 15-V power supply; and two zero-insertion force (ZIF) sockeis: one 28pin type, Ul, to program EPROM devices, and a 40-pin type, Us, to program the microcontrollers. The ZAPS1 can program CMOS devices only: hence, it will program the 87C51, but not the NMOS part, the 8751, The 8751 requires a 21-V pro: gramming voltage that, on cost consider ations, was rejected on the programmer. Software The software supplied with the ZAPS1 is splitinto two parts, The first part executes on the PC; the second part runs on the ICESI The ICES1 program extends the function- ality of the standard ICESI monitor program to inchade programming code. The PC sof ware includes menu-driven code to control the ICE51. A brief de dure to program a presented below. This is followed by a de. Scription of each menu option. . Note that some of these components are polarized electrolytic types; make sure that correct polarity is ob- served. When the capacitors are fitted cor rectly on the board, solder them into place. Prepare the following from solid 22 AWG (0 mm dia.) eireuit wires two bin. (15 em) lengths, stripped !/ in. (6 mm) at each end for the alarm reset switch, Sy and three 8-i, (20 cm) lengths for lock enable switch Ss, Solder these wires to theappropriate locations fon the boare and then attach them to-the rel evant lugs on the switches. ‘Test the board by applying 24 V dc. and 5Vd.c totheappropriatepointsonthe board. (observe polarity) and make the following. measurements: pins 7 and 8 of IC>and ICs should be at 24 V Le; all others pins on the two sockets should be OV Thsert Sy into its socket with the ON side located towards Ros-Riz Closealltheswitches, whereupon pins 2,4, 6, 8, 11, 13,15,and 17, of ICs should be at 5 V dc. Open all the switches, whereupon these pins should be at OV. The level at pin 1 may be high or low, depending on the position of change-over switch Ss, connected to points X, Y and Z. Pin 20 should be at 5 V de. all the time. All other pins should be at 0 V all the time. Pins 2, 4, 5,7,9, 12, and 16, of ICy should be at 5 V dc. All other pins should be at OV all the time, Pin 14 of ICsand IC, should be at 5 V de. all other pins on these sockets should be at ov, If you measure any levels other than the ones stated, stop and find the reason for the Incorrect reading, Once this has been found, retest the board as outlined Next, remove the power from the circuit. Attach the ribbon cable between the sensor array board (completed earlier) andthe main logic board. The letters on the main logic board correspond tothechannel letter output ‘of the sensor detector /amplifier array Prepare the cable for assembly by carefully stripping the insulation from the ends of th ‘wires. Be carofil not to nick the fine-strand conductors. Twistthe conductorstogetherand lightly tin the wires: take care not to damage the cable insulation with heat from the sol- dering iron. Separate the cablealong the cen- tre conductors inthe cable toa length of 3 in (75cm) fromtheend.Next,separate each con ductor in the ribbon over about 1 in, (2.5em) from the end. Insert the wires from each channel into the corresponding pad of the ‘main logic board, Note that channels 7 and 8 are not in onder and must be given a ‘half twist Attach a relay or solenoid coil, or any ther suitable type of load, to the load points fn the logic board. Mouint the sensor detec tor/amplifier array, the main logic board and the output loads in an appropriate en- ‘Closure Set switchSss0 that pin 1 ofCyislow to enable the lock driver. Install ICy-IC, into their respective sock. cts: make sure that they are correctly ori ented and that all pins insert correctly be- fore pushing them home. Note that ICs, IC ICz and ICy, are sensitive to electrostatic en- ergysotaketheappropriate precautions when handling these devices, . Next month's final part ofthis article wit! deat with programming and testing the opticalock. DESIGN IDEAS The contents of this column are based solely on information supplied by the author and do not imply practical experience by Elektor Electronics Keyboard circuit HERE are many chipson the marker that may be used ina keyboard circuit, but many ofthese are fixed and do not provide all the keys required. For example. RCA'S CDP 1871 is an excellent device, CMOS logic and low power requirement, bit Joes not provide forthe backspace key io be coded as hex 08. Instead. it uses the delete key, coded as hex 7F. Although itis possible to change the software to recognize the dilfer- ent code, this is not always easy. The aim of the design offered here is « keyboard that can produce every code but is also able 10 be hooked up as a standard keyboard. To provide a matrix that will encode the 128 combinations of the ASCH code, 16.8 Hines are needed. If all 128 keys are to be used, 4 strobe output to signal that @ key by D. Nelson has been operated is also required. In the diagram, ICs, ICg, and ICy are all CD40SI single 8-channel multiplexers. The strobe is generated by stopping the clock that drives seven-stage ripple counter IC via an AND gate, because the battery on pin 9 is earthed via the matrix and 1wo of the multiplexers. The key code is contained in the output of the ripple counter and the strobe is pro- vided by the Schmitt rigger NAND gate that will oscillate with an unequal mark-space ratio, since the diode used is a germanium lype. A silicon diode would, of course, do as well, provided itis shunted by ahigh value resistor. The oscillation provides a repeat function of the key that is pressed. Since the strobe is in the wrong sense to that re quired by the following circuit, it is in- verted by an additional transistor. To give the conventional keyboard the functions of shift and control keys, extra gat- ing is required, and that is the purpose of the rest ofthe circuit. Control has precedence and, by means of wo AND gates, forces its 6 and 7 low. The control key 1s non-lock- y ed toconvert lower case letters to upper case so that it is nec- cessary to wire the alpha keys in the Lo ‘case matrix positions. This key is conven- tional non-locking and inverts bits 6 and S of the output code, The shifi-lock is cou pled capacitively to the same AND gate as the shift key, so that the gate is toggled by the shift-lock key and will stay with the‘output low when the shifi-lock is released. The shift-lock is released by pressing the shift key and releasing it again. ‘One other feature has been included to make the keyboard friendly to computers: the alpha key. This changes lower case to upper case without shifting the keys with, an ASCII code between 20 and 3F hex, that is, the numbers row on a standard ki board. This key islocking andimplemented in the same way as the shift-lock key. The alpha key is released by operating the shift- lock. ‘Only one key now remains an enigma: the space bar. Without any extra gating, this gives a 0 ifthe shift or shift-lock is op- crated, and this is clearly not acceptable Tocorrect this, five diodes have been added, fone each from the four least significant lines and one between the two inputs of the XOR gate in bit 5, so that a space re- mains a space, shifted or not Two further features ofthe cireuit require a mention. One is the transistor feeding the most significant bit8. this isconnected to its own key, it can be used in conjunc- tion with other keys to give hex codes from 80 to FF, The other istied up with the clock and strobe. A diode is fed back from the output of the repeat strobe oscillator to the main clock. This is necessary to prevent a false output if'a second key is pressed be Fore the first one is released. . KEYBOARD CIRC * hate 16 x 8 ELEKTOR ELECTRONICS SEPTEMBER 1991at A CHEAP, EFFICIENT, STRATEGIC FIRE ALARM by C.C. Whitehead A fire may be started by an unstubbed cigarette-end, a faulty electrical appliance carelessly left on, or lighted paper shoved through the letter-box by vandals. You may, like a great many people have furniture stuffed with deadly polyurethane foam, which catches alight easily, burns rapidly and produces vast quantities of dense and highly toxic smoke. In most cases, smoke is the killer. People asleep are poisoned and rendered helpless long before the heat gets to them to finish the job. And all this happens in a surprisingly short time: NEIGHBOUR of mine showed me the fire alarm she had had installed in her living oom. Itwasnot particularly cheap, but nevertheless reasonable in price. It was effl- ent, but certainly not strategic. In effect, it protected one room only, orat most one floor of a two-storey house if all doors on that floor were left open. ‘Theexperts—professional firemen—have shown us clearly how domestic fires nearly always start, what happens in the first few ‘minutes from the start, and what the main danger to life i Inall cases, smoke is the main hazard and that is why most fire alarms are smoke de tectors. Fortunately, these can be made very olfcient at little cost ‘Once the fire has started, hot air carrying smoke quickly makes its way to the highest point itcan roach in the building, forming a layer just below the ceiling, and then filling first the room or passage and subsequently the whole building from the ceiling down- wards. That is why firemen, and those who are trying to escape from the fire, keep as low down near the floor as they can, even crawl if necessary. It is clear that a smoke detector should be installed on or near the ceiling, certainly high up in a room. a few minutes at most. Fig. 1, Correct position of smoke detector. A smoke detector/alarm can be installed in every room, but that is quite an expensive business. So isthe installation of smoke de- ecto alarm, The in every room connected to a central ch involves wiring costs apest method, without sacrific: ing too much in the way of efficiency is to installa single smoke detector/alarm at the most strategic pointin the building, Ina two- storey house that isat the top of the staircase leading to the second floor, closeto or on the celling—see Fig. 1 Dealers may, of course—and quiterightly— point out that this system loses a few sec ‘onds of time in indicating the start of the fire, and that, ifthe fire starts in a room of which the door is closed, considerable dam- age may occur before the alarm goes off. This istrue, butthealarmstil givestimetoawaken sleeping people and allows them time to make their escape; in this respect itis as el fective as any other system: itis all a matter fof seconds after the start of a fire Circuit diagrams ofa battery-operated ver sion and of a mains-operated version of the unitareshownin Fig. 2and Fig, respectively, To avoid duplication of the description, the designation ofthe essential componentsisthe same in the two diagrams, The relay is a miniature type available from many suppliers: it has an operating ‘on’ voltage of 5-7 V at 10 mA, The coil re sistance is 600 Q. In a simplified version of the battery unit, Tandy Type 275-232 may be used: this has only one set of contacts. The i, | wt Fig 2 Battery-operated version of smoke detector. Fig, 3. Mains-operated version of smoke detector. ELEKTOR ELECTRONICS SEPTEMBER 1991ACHEAP, € i= Fig. 4 Detall of mounting of uit (see Fig. 1). Fig. 5. Smoke hole at underside of uni. — Fig. 6. General view of front of uit Fig. 8, Smoke escape holes and holes for con- necting leads to mains, escape lamp and ac cess to Py Fig. 7. General view of right-hand side of unit. objective isto keep the power consumption of the relay (in the non-alarm condition of thewnit) as low as possible toextend the bat- tery life There are two ways in which the unit can ‘be made to function, depending on the posi tion of infra-red ([Riemuitting diode Ds and IR photodetector T>. As shown in the dia- ‘grams, when there is no interruption of the Tight path between Ds and T, the relay is energized and thesupply tothebuzzerisheld off, A slight obstruction of the light path causes the relay contacts to open, where upon the buzzer (and the ‘escape light’, if filted) is energized In the battery-operated version, the post Fig. 8. Location of components in mains-operated version. In the battery-operated version, removal ‘of the transformer, zener diode and R makes space forthe battery. ELEKTOR ELECTRONICS SEPTEMBER 1991 [FICIENT. STRATEGIC FIRE ALARM tionsof eversed and thecon nectionsoftherelay contactsasshownto give a slightly reduced current drain on the bat tery. The circuit works well in either case Capacitor C; in the mains-operated ver- sion has two functions: theusualone of reser- voir and that of holding a charge so that when theunitis disconnected orswitched off, thebuzzer givesaloud bleep. Thisshowsthat theunitis functioning correctly andisthe re: son that the capacitor is retained in the bat tery-operated version In the mains-operated version, the mini ture transformer may have a secondary volt age of 12-20 V and be capable of providir avload current of 150-200 mA, The value of R; is 15U, where U,is the secondary voltage of thetransformer, its tol- ferance may be £20 Theunitmay beconstructed ona 10x7 Sem (4 in.) piece of prototyping board (vero- board or perfboard). The location of the com: ponents, other than Ds, T2, P; and the tr former, is not important. The positions of and Ds are obvious and should be about 25mm (1 in.) above the board. The dimensions of the enclosure are 911.52 cm (3.5x4.553.5 in.) (LxWxD). The positions of the smoke holes and the hole for screwdriver access to adjust Py can beas- certained from the photographs. Ik will be found that there are two points on P, at which the buzzer can be switched ‘on and off: only one of these is correct. The adjustment may be made on a workbench, since; isnot sensitiveto visiblelight.Tosind, thecorrectadjustment point, wagglea match stick between Ds and T>. Once this has been. found, carefully adjust P; again until the buzzerjust goes off. The buzzershould make aloud and unpleasant noisetomakesurethat it wakes healthy sleepers. 7 Lee Resistors: PI =22kO preset Capacitor C1 = 1000 uF, 35 V DC working Semiconductors: Di-D4 = IN4001 DS = infra-red emitting diode D6 = IN914 D7 =zener diode, 12 V,1W TI =BC149 or similar T2 = infra-red photodetector Miscellaneous: S1=SPST switch Rel =miniature relay resistance 600: Baitl = 12 V battery Trl = miniature mains transformer, TV, 10 mA, coilSCIENCE & TECHNOLOGY Understanding the EDIF standard by Tony K.P. Wong EDIF is a standard data format for electronic design information transfer between CAD systems, and supports gate array and semicustom IC design. It plays an important role in the VLSI world. Some of the basic structure of the EDIF and an application of its netlist are presented in this article. “T MAY be recognized that the VLSI inte- arated cireuit design process relies heay- ily on communication between all parties inyolved in the design, manufacture and testingof VLSI products, thatis, “Joint Ventures” scheme. This has to be so irrespective of the different CAD systems that might be used at cach end of the data exchange. However. problems arise since each CAD system em ploys its own system-specific data structure for storing product data, This data may not be transferred from one system to another or be read by other systems, as they have no. ‘common data structure. Hence, to obtain the maximum benefit from CAD systems, data exchange becomes essential 10 the manu facturers Many national and international organi- zations have developed a public domain standard for data transfer, which is called Electronic Design Interchange Format, or EDIF. This standard enables the designs of semicustom and custom ICs to be commu- nicated unambiguously to device makers, EDIF structure EDIFisaneutral data format that may be used for transferring netlists, schematics, IC ay outs and other electronic data between com: puter workstations. A translator program Will be required to translate the information from the sender's data base and to create a sequential file in the format defined by the EDIF specification. Another translator pro- ‘gram may be used 10 toad the design infor- mation into the receiver's data base. The basic format of the EDIF syntax is an ASCII file. An EDIF file consists of an hierarchical set of structures based on the LISP programming language, Each con- Structofthe EDIF languageisa parenthesized list oF items of information. Each distinet EDIF construct is identified by a keyword, Thissyntaxis very simple,easily parsed. and extensible, but its power for representing design information comes trom its struc- lure. The hierarchical structure of EDIFmeans that itis abstract at the highest levels and becomes progressively more detailed as one sdescends withinthe hierarchy. Figure I shows portion of the EDIF siructure model and the path of the netlist view is specified, Further details of the structure of EDIF may be obtained from Ref. | In general, an EDIF file may contain one cor more libraries of cell data, A library con: sists of cells grouped according to the com- mon characteristics, 1 Cells in the library may contain instances of other cells from an external library. Each cell currently supporis seven view types behaviour; schematic; symbolic; netlist; mask Jayout; document; and stranger. Each viewof the cell contains information particular to 3 specific use of the cell, The “library” and ‘view’ structures ean make a possible change in the transfer of partial or incomplete design information, so thal the data transfer time may be reduced, View cells can be divided into two se tions: the interface section defines the cells communication with other cells, and the contents section defines the detailed im- plementation of the cell In the top level, the Design block con. tains the name of a design and provides a path for finding data, The Status block spec ifies the file information such as the pro- ‘gram version, creation time, and so on. The User Data block contains the user specific data, for instance, local user extensions. All Of the data deseri design is collected in the Library block. The Library may make reference to the External Library, that is, a {brary is known to the sender and receiver byname, butnodetails are transferred. Within the Technology construct, information of pays ical design rules, global constants, and so con, may be included. For the View level, the Netlist view, the netlist description contains a list of cell in- stances and a collection of nets that specify signal connections among the ports on the instances and the external ports of the cell itself. IL is possible to display the property ame andits valuein any context, Properties ‘may include optional ‘owner’ and “unit” val- example, (port AAA (property BBB (number (e X-Y)) (unit CCC) (owner DDD))) describes that the port named AAA of a cell has. property BBB (e-g.. Cin). The value of the property isexpX —Y (eg., 110-!2) and it has the unit of CCC (e.s., capacitance) ‘The owner of this property is DDD, which is user-defined to avoid the same name of properties from different data bases. I should be noted that, when the file is translated into the receiver's data base, any scaling of values for the unit CCC, which is inthe technology section ofthe EDIF applied to the property ‘The Mask Layout view also contains an interface and a contents section. In the con- tents of the view there is a set of shapes on different mask layers, called figureGroups in EDIF, and an instance of another cell The figureGroup will refer to a figure group defined in the technology library ELEKTOR Fl TRONICS SEPTEMBER 1991UNDERSTANDING THE EDIF STANDARD 2 irseten Gnstance (quality DEVICE LIB x_10uF) x s0uF_NET C1) (Tinestanp 1991 29 14 5 40) (sesnea (accounting Program *METLIST.EXE") cauatity R21) external aNnLoe 18) (quanity 62 <2) (external ASSENBLY_L1B) cquatity R212) (external INTEL_LIBD > (ecternal SHAFES_L1B) (auatity 1) (external sPICe_L18) > (externat THLE) Goines (oess9n OPANP_SCH (qualify 14 root? Cauatity Rt) (status (quaisty ut pus neuT) (comment “The ABOVE TimeStanp 4s local tine") : (vie NETLIST root NET cauatity RS) (aetine input port 1RARP Iquatsty Ra x2) (Gnstance (quality PSPICE LIP x_108) X.A0K NET Fo) scos.12 Ginstance (qualify PSPICE LIB X 10K) 408 NET 2) ELEKTOR ELECTRONICS SEPTEMBER 19913 = =| ee] Fe] pay CONTENTS ais a = ng Only in : =. ‘The Symbolic view contains the descrip tion on elements of connectivity and geo- metric layout. It ean be used to describe placement androuteing problems, and totrans- rmitsymbolic layouts for technology re-map- ping. Under the Symbolic view, the con- struct elements specified in the Interface and Contents sections may be implemented. ‘The Schematic view is used for the trans mission of logic data and schematic dia- gram plusconnectivity. The cllsinclude logic elements and symbols. An EDIF netlist application ‘Netlist is list of components and connee- tions that describes the connectivity of a cir cuit (Ref. 2). Figure 2 shows an EDIF file that describes the cireuit of the band-pass net work shown in Fig. 3. The schematic cap- ture was produced by OrCAD/SDT Ill soft- ‘ware, The EDIF netlist file is one of the for mats that the software can support. The de- tails of the file related to the circuit are de- scribed below, OPAMP.SCH is the name of the il Under the Status construct, the EDIF ver- sionandlevel are specified; the local timeand. the program (NETLIST.EXE) trom which the EDIF file was generated are described in the Written block Following these, the external library files are referred to; they are named with the ex- tension, LIB, as shown and, in this case. they may be found in the library directory (of Schematic capture softwa ‘A root Cell must reside in the Design en- tity and be cell in one of the libraries, ‘The Status block following the Cell root contains information about this particular cell. Lithen shows the Netlist view of the cel At the beginning of the view, all input and ‘ouiput ports are defined under the Interface block. The Rename construct shows the origi nal port name, The designation of compo- ents and their electrical values are speci fied in the Instance construct. Each compo- nent value and its related library file are also mentioned with a Qualify construct. For instance, in this ease, Rj has a value of 10 k® and belongs to the external PSPICE library file nally, the connectivity of the circuit is described in the Joined blocks. Bach Joined block specifies the connections at one node ofthe cireuit-Inpractice,anumberisassigned to the terminals of each discrete compo- nent: the left-hand terminal is | and the right-hand one is 2. For instance, as shown in the first Joined block, terminal 1 of Ca and Ry is joined to terminal 2 of Ry and the ve input of Up. Conclusion EDIF Version 1.00, published in 1985, has provided a solution o the data base compat ibility problem in the gate array and semi- custom IC designs. Although itis not com- plete, users have processed a number of suc ‘cessful data base conversion programs, Many CCAD sofware manufacturers have included the format in their products, for instance, OrCAD, HiWIRE, Schema Ill, and so on. EDIF Version 2,00 has just been pub- lished. This can address many of the prob. lems and difficulties of the earlier version, and meets the needs of real data transfer in the design and manulacture of printed cir- cuit boards. Moreover, it contains provision forbehavioural modellingand includes some new constructs forinstance, Netand Property. Uniortunately,eventhelatest version cannot yet do everything: features that are still under evelopment include: transfer of design changes: PCE layout transfer between CAD sys- tems: transfer of data for board testing: transfer of data from CAD systems to board manufacture electrical design rules. ‘These features may form part of the next version. However, there will be a signifi- ‘cant performance penalty to using EDIF for data exchange. because the format tends to require more storage space than the CADdata, and may slow down the data base conver- sions. Simplification and more advanced eonstructs may. therefore, be required. Hl References: EDIF Specification, Version 1.00, March 1985) "Shaping up the Netlist”, by Tony K.P. Wong, Electronics World, Vol. 96, No. 1656, Oct. 1990. 3. “Focus Report: Engineering Software”, IEEE Spectrum, Nol. 27, No. M1 Nov. 1990, pp. 60-85, ELEKTOR ELECTRONICS SEPTEMBER 1991PLOTTER DRIVER re requires some additional computer activity be 's can be put on to paper. I presented that tran 1OR those of you who have not followed the story so far: the plotter referred to above is a mechanical as well as electronic construction project described in Ref. 1, with some useful mechanical alterations brought together in a follow-up article, Ref. 2. The second article also describes a partly HPGL- compatible plotter driver, which is less soph: isticated than the one discussed here. It may be useful to know that the modified version of the plotter is available asa kit from some regular advertisers in Elektor Electronics. It muist be noted, though, that the contents of their kits may deviate from the modified mechanical parts list included in Ref. 2 The driver program CCAD programs — or more generally, draw ing programs —suchas AutoCAD, OrCAD, DrawPerfect, Ultboard and Smartwork, to mention but afew, provide drivers for plot- ters that work with a certain industry-stand- ard ‘language’ for instance, HPGL (Hewlett Packard Graphics Language). Since the Elek- tor Electronics plotter is not compatible with any graphics language, it requires a conver sion operation in which the commands con- tained in the computer output file are werted into appropriate bit combinations that can beapplied to the interface board that belongs with the plotter In practice this con- version operation is carried out by the pro- ELEKTOR ELECTRONICS SEPTEMBER 1991 gram described here. This program accepts HPGL input data, translates this intoa series of plotter commands, and drives the plotter vvia the Centronics port on the PC and the in- terface board attached to the plotter. The plotter driver program is written in Turbo Pascal 6.0, and comes on on ONDRIAAN PLOTTER DRIVER Plot irr Ara Pies Pee ae ree aT Pete Peer east erie ad Sea Poneoge eet Dre ror aCe CHAT ATA ros 360 KByte MS DOS formatted floppy disk ‘The programis suitable for MS-DOS compu- ters only. Itcan be modified to drive plotters other than the Elektor Electronics one by ‘adapting a number of procedures in one of the source files. Evidently, to beable tomake such modifications you must be conversant roa rat ie og ed 724 pd H Copyright c) Danielo Sijtsmawith the Turbo Pascal 60.compiler. The basic information on the modifications is contai- ned in the source file tse. The plotter driver program, PL.EXE, translates HPGL filesinto control commands forthe Elektor Electronics plotter, and inaddi- tionallowsa number of plotter-specific para meters to be defined. Before the plotter driver can be used, the drawing made on the PC must be converted into an HPGL file. Most CAD programs can do this without problems when an HPGL- compatible plotter (eg., the HP7475A) is se- lected in the output menu, and the plot fileis, printed to disk. When the HPGL file is avai- lable on the system, the plotter driver, PLEEXE, may be called up to complete the sion into plotter commands. Comple- tely menu-driven, the converter program is, uncomplicated in practical use Hardware recap and tuning The interlace board with the Elektor Elec tronics plotter is connected to the Centronics port of the PC as shown in Fig. 1. Note that a number of pins on the Centronics connector are tied to #5 V to make them permanently logic high. The +5 V can be taken from the plotter interface via an unused wire in the cable to the Centronics port on the PC. For connector pin functions, look in the file PLPROC-PAS on the diskette, Before you can start plottinga drawing, a number of parameters must be set. This is done in the ‘settings’ mem, and in particu- Jar, in the sub-menu ‘mechanical charac- teristics. Wait time between steps ‘This parameter determines the plotting speed. The larger the wait time, the slower theplotting, but the more accurate the result Since the wait time is set by means of a soft ware counter, it may have to be adapted in accordance with the clock speed of the com- puter. Half step The half step parameter determines the final resolution of the plotter. When you switch the motors from half step mode to full step ‘mode, the size ofthe plotted drawing will r= main the same, ie, the co-ordinates are re scaled internally Wait time between pen up/down This isa constant that must be determined empirically for best results. It provides a cer tain delay before a pen is lifted or put down, ‘on the paper. Y-compensation factor ‘This factor is a constant that allows dif- ferences in the resolution between the X and Y driving systems to be compensated. It is programmed by writing the following line of SCI text, and sending it to the plotter: PR1000,0; PRO,1000 °Z Make sure that the clipping area (see ‘paper size’)is not effective, and if necessary adapt theenlarge factor. The ¥factoris obtained by dividing the distance travelled in the X dire tion by the distance travelled in the Y direc COMPUTERS AND MICROPROCESSORS THE HPGL COMMAND SET — AN INTRODUCTION — HPGL (Hewlett-Packard Graphics Language) commands are abbreviations ot plotter control instructions. An HPGL file may be wnitten with any ASCli-compatible word processor, and consists of HPGL commands and co-ordinates. Tha best known com: mands from the HPGL set are probably PA (plot absolute), SP (select pen), Cl (circle) and LB (label). ‘The number of HPGL commands supported determines whether or not a certain plotter ‘can be used with a drawing program. Most drawing programs use only a few af the 50 commands avalabe nthe HPGL. The wellknown AUIOGAD package is|2 good ‘example, because it uses the HPGL basic commands only, such as PA (plot absolute) and SP ‘select pen). Texts generated with AutoCAD afe plotted as series of PA ‘commands rather than with the aid of the LB (label) command. The same goes for circles ‘and arcs, which are also plotted with PA command strings. There are, however, other programs that use a much larger subset of HPGL commands. It you want to know if the Elektor Electronics plotter can be used with your drawing program, itis best to check ‘the commands contained in a generated plot fle against the commands supported by the driver program presented in this article, Supported HPGL commands ‘The plotter criver program discussed here supports all he HPGL commands necessary to work with most commonly used drawing programs. In pracice, the drive has been {ested successtully in conjunction with AutoCAD, OrCAD. Utibaard, Smartwork and DrawPertect. ‘The HPGL commands supported by the author's Turbo Pascal 6.0 program are: PU (X1,Y1)(,.) Pen Up PD IKI YIIC..) Pen Down PAIKIYI)(..) Plot Absolute PRIXIYI)(..) Plot Relative Cl radius(.chord angie) Circle ‘AA.X.Y,ate angio(.chord angle): Arc Absolute ‘AR X.Yiate anglo(.chord angle): Arc Relative LB ASCII string (c) LaBel ASCII string Dre: Define label Terminator (c) Slwidth, height: absolute character Size ‘SR width, height: Relative character Size Dicos.sin: absolute Direction SP Select Pon ‘SC Xmin. Xmax, Ymin, Ymax __SCale info user units, IP P1x,Piy(,P2s,P2y): InPut Pt (,P2) WW: INitialize OF: Set DeFault values Optional paramators in brackets (): parameters are deimited with a comma. HPGL ‘commands in a pot file may be separated by a comma, The plot file may contain carriage returns (CRs) and line feeds (LFS) which are ignored by the driver. These control codes, may however help to improve the readability of the plat file. Co-ordinate system ‘A number of plotter commands are followed by co-ordinates, which can be absolute (i.e, with respect to the origin), or relative (i.e., with respect to the current pen position) Gapending on the preceding command ‘The unit of distance used in the co-ordinate system is a plotter unit rather than a SI unit ‘such as the millimetre. The values assigned to the co-ordinates depend more on the resolution uses by the drawing program than on the size at the drawing. The plotter units used by the Elektor Electronics plotter are determined by the resolution of the stepper motors and the mechanical drive Systems. In half-step mode, 2 resolution of about 0.1mm is achieved, The HPGL provides the SC command for plotter-unit to user-unit conversion. The SC command also allows the origin to be moved, for instance, from the centre of the paper to the lower lefi-hand corner. Another, related, scaling command, IP, enables two reference points to be detined, Pt ‘and P2, which are also used during rescaling to user units. The SR command, which Sets the relative character size, Is also relatod to the reference points P1 and P2. All the juggling with co-ordinates is of little interest to the user, who is only affer the final size of the drawing. This size will depend on the drawing program used. A scaling manipulation feature must be built into the plotier driver or the drawing program to ‘enable the output of different programs to be drawn on one (maximum) paper size {nthe drawing program, the scale is entered as a parameter together with the plot ‘command, of as a global setting in the program Itself. This allows the co-ordinates ta, be converted directly to the co-ordinates that belong with the new scale, Aliernatively, the SC (scale into user units) command may be placed at the beginning of the plot file to enable the plotter driver to do the co-ordinate conversion tse. ‘When the driver is used fo scale co-ordinates, this is best done with the ald ol the ‘educefentarge option of which the parameters can be set via the ‘paper menu. ELEKTOR ELECTRONICS SEPTEMBER 1991Dae Path bag quit et ean Pace a) aed aay Ses rena) Cao eet Yes Gone eee Po rc) 0,130 Cry cea eet ee pen ee eres Sen Position Fast Print LOiCan TA tio: ¥ factor = dX/4Y. The larger the dis tances travelled, the more accurate the Y fac tor X-and Y-compensation pen 1/2/3 ‘These parameters serve to compensate the distance between the pens in the carriage. Enter the following text in an ASCII word processor, and send it to the plotter: PU;SPI; PD; PU; SP2; PD; PU; SP3; PD; PU %Z, ‘This should produce a single spot on the paper without off setbetween the individual dots made by the pens. Number of pens Optionally, you may work with fewer than three pens. The pen numbers in the SP com mands are counted in a cyclic manner. When, forinstance, thedriver encounters the command ‘SP3' (Gelect pen 3), while the ‘maximum numberof pens is set to2, pen Tis selected, Paper size ‘This parameter speaks for itself. The X and ao Parone ace) co-ordinates are not the HPGL co-ordinates n the plot file, but the maximum numbe steps to be made by the stepper motors on the plotter. The limits that belong with acer: tain paper size setting can be calculated on thebasis of the diameter of the platenand the string wheel for the carriage drive, and the number of steps per spindle revolution of the stepper motors used. Next, the enlarge/re duce factor is used to fit the drawing on the selected papersize. When the co-ordinates of the drawing fall outside these borders, the relevant part of the drawing is not plotted, and a warning is given when the plotting is finished Trouble shooting and user hints Ieis very well possible that the first results ‘obtained with the plotter and the driver pro- ‘gram will look like drawing attempts of a two-yearold. Any one with someexperience in setting up electromechanical drawing Ploter diver Fig. 7 plotter driver board. ELEKTOR ELECTRONICS SEPTEMBER 1991 ce i = é Z| | o —— = 4 = i 4) | s hee ao E EF Hardware recap: suggested connections between the PC Centronics port and the SOFTWARE ON DISKETTE ‘The plotter driver program and con- figuration utlities described in this ar- ticle are available on a 5¥4-inch ‘360 KByte MSDOS formatted floppy disk under order number ESS1541. Details on cost and ordering are given ‘on the Readers Services page else- where in this issue. equipment will confirm this without shame, In the worst case, nothing is plotted. First check whether manual pen positioning works (goto the sub-menu ‘plotting’. If this works, you have probably loaded a non HPGL compatible plot file (the name of the file processed is displayed in the lower left hand comer of the screen). A problem reported by many construc: tors ofthe Elektor Flectromies plotter has todo with the lifting and lowering of the pens. First, check that the pens can move up and down without friction, Marker pensin part cular are prone to remaining stuck. This hap. pens because the lever that lifts the pen remains stuck to the core in the solenoid (lift magnet). The problem is simple to eliminate by sticking a small piece of self-adhesive tape between the core and the lever. This creates a small air gap between these two metal parts, and rediuces the risk of sticking (caused by permanent magnetism) toa mini= Real’ plotter pens are much heavier than marker pens, and are much more likely to re main down on the paper because the sole- noid in the carriage is incapable of lifting the ‘weight. There are two possible solutions to this problem: reduce the tension of the solenoid springs, or remove them altogether; increase the value of capacitors C2, C and Cs on the interface board from 470 UF to 1,000 pF ‘The distance between the lever and the pen is critical. When the lever is too close to the pen, it can not move freely. Conversely, ‘when the distance is too great, the pen can bounce up when itis lifted. The critical dis- tance can be ‘tuned’ by bending the ‘rear leg’ of the solenoid a little forward or to the rea. If the plotted drawing is mirrored with respect to that produced with the CAD pro- gram, simply reverse the connections to one of the stator windings on one of the stepper motors. Paper slip may be caused by insufficient sure on the platen. The correct amount Pre Of pressure must be determined empirically by fitting springs with different tensions, Another type of slip (in the Xas well asin the ¥ direction) may occur when the platen or the string wheel for the carriage move ment is not fitted securely on the spindle of the relevant stepper motor, . References: 1. *Plotter’, Elektor Electronics May and June 1988.‘Would you be an inventor?* Sir—In your issue of July/August 1991 ‘you include an article with the above title ‘by a Mr Whitehead. Unfortunately, the ar- licle is wrong in several important re- spects, and I arn concerned in case inven tors try (0 take guidance from i For example, Mr Whitehead bases much of his antcle on the contention that ‘itis an established principle that employees’ in ‘entions belong to the employer”. This is not the case. The United Kingdom law, for ‘example. lays down very strict conditions {for an employce’s invention to belong 19 the employer, and in all other respects the invention belongs to the employee. Furthermore, the law makes it impossible for the employer to extend his rights to the inventions of an employee by contract of employment, or by other means. Mr Whitehead also suggest that the date ‘when the patent is accepted is important and sets up ‘an expensive brawl between lawyers to decide which specification was filed frst, and thus oust the other’. Again, this is totally incorrect. Any patent applica. tion filed any here in the warld has a fil- ‘mg date from the outset und this estab lishes the precedence of all patents. is also distressing that Mr Whitehead suggests that individuals and smaller firms cannot ke action for infringement. In this cease. it is now possible to insure fo cover the costs of infringement actions. There is also a riew Court. the Patents County Court, which is designed to cut the cost of infringement actions by several orders of magnitude, Jacqueline Needle Chairman, Press & Public Relations Committee ‘The Chartered Institute of Patent Agents London Satelite receiver conversion Sir—Further to my telephone conversation with your Technical Editor, T enclose the information on how to convert a 16-chan- nel Alba/Bushjeve set ro a 20-channel set. Jn principe, this conversion works in all satelite receivers that are based on the SGS-Thomson Type M494 processor ‘The conversion has the advantage that the remote control of the set remains fully funetional. because the same cades are now merely interpreted differently. In fact some keys on the remote control transta ter are no longer used. 5‘ LETTER ‘The conversion is effected by soldering two Type IN4L48 diodes between the op- tion select line (pin 6) and data lines DO and D3 (pins 2 and 5) of the processor. It ‘costs only a few pence and, in any case, ‘many amateurs have the diodes in their ‘workshop. How they are soldered on to the ‘board in the Alba receiver js shown in the sketeh: R. van Zon, ‘Nunspeet, The Netherlands ‘Wattmeter Sir—In the article 'Wattmeter’ in your ‘April 1991 issue, the author elaimns that his device will give reasonably accurate readings when dealing with a reactive wad, and that the device aceounts for the phase angle difference between voltage and current. I'm not convinced, Power is equal to Elcos®, when cos@ is the cosine of the difference angle between, the voltage across the load and the cur~ rent through it. Lsee no circuit in his: device to sense this difference. and no circuit fo multiply a *cos@" by the El product jene May Belvoir, USA You are rigiu! The watrmeter measures the apparent power (EL) anal not the re active power (Elcas®). Our apologies for this error Ed. Hall-effect sensor / Gaussmeter Sir—I have a present need to be able to sense if there he 3 magnetic field around a conducting element, most particularly. an Insulated water pipe. 1 would like to use a Hall-effect sensor, but do not know who manufactures such devices, where to buy them and, above all, what the component designations may be. Could you help me with the appropriate data? Perhaps m some future issue you might care o give us a circuit for a gausmeter. for one, would enjoy building such an in- strument and adding ito my electronic equipment. Harold Willis Milnes Lubbock, USA We have asked our cotleagues at Peterborough. NH, if they can help with the ferst part of your letter. Your interest in a guussmeter has been passed 0710 our design deparnnent for investigation. Ed Plotter Mark 1 Sir—You may recall my letter to you of 14 December 1990 in which Texpressed con- ‘cern about some aspects of the software you supply on diskette EL17—the plotter driver software MONDRIAN®. You re~ ferred the letter to your office in Holland who gave me the address of the author in Germany. After putting my difficulties to him, he replied 10 say these had been pointed to him previously by German reader, and that he had expanded the basi ‘commands so that PUX.y; and PDx,y; are ‘now correctly interpreted In my case. I was particularly concemed with the HPGL files produced by the EASYPC software supplied by Number (One Systems. The author very kindly suy plied me with a corrected version of the Software and all is now well, He states that [should fee! free to help other users with similar problems. Therefore, should any of your readers encounter similar problems, 1 ‘would be glad to assist them. G.H. Olsen, 5 Fieldhouse Close, Hepseott, Morpeth, Northumberland NE61 6LU, * Pier Mondrian (or Mondrian) (1872-1944) was a Duteh artist who, with AL van Doesberg (1883-1931), founded the De Stijl movement in architecture and painting. His rectilinear abstracts in black, white and primary colours have had con siderable influence and he is considered the leader of neo-Plasticisn. Eu, ELEKTOR ELECTRONICS SEPTEMBER 1991PROFILE OF MICRO AMPS ei by Bernard Hubbard COMPANY, opened some time ago by two former University of Surrey students, is breaking new ground in the field of microprocessor hardware and software development. David Kyteand Richard Marriott es- tablished the company only a few miles from the university andre busily, becoming the leading technical edge invarious areas of development under the name Micro AMPS. For the former students. it was Surrey University that gave thema lly ing start (o business when, two years ago, a spin-off company placed an order for delivery just two months later, David told Elektor Electronics: “We had to produce a DSP based Vocoder for a satellite system for SPEKA and ‘wehad untilearly February to produce the board that was to be dispatched toIndonesia shortly after that date, We met the deadline and Micro AMPS was in business Ironically. the original specifica~ ton did not entirely satisfy the Indonesian client's requirements and so further development was necessary. The fol- lowing April, 10 boards, incorporat ing the additional development, were shipped to the customer Havingeutits teeth on this contract the fledgling company was promised a flow of further orders. However. to date this has failed to materialize and instead the young company has been busy spreadingits expertise toa wider cusiomer base. “Our object.” says David, “is to in crease the profitability per individual and we see licensing coupled with sub-contracting manufacture as the way forward. Weare, in fact, acontract research and development unit, We have formulated our strategy. rightly or wrongly. We aim to be involved in the explosive personal satellite com- munications growth in the next few Initially, Micro AMPS began in David Ryte's spare bedroom. Unfortunately, the company outgrew this space just at the same time as Mrs Kyte was ex- pecting her first baby and the company moved into a converted brick kiln on the outskirts of Cranleigh—see pho- tograph. The family business link has been retained, however, for Mrs Kyte keeps the books after re-training as a bookkeeper /administrator. Today, the company employs 3!/2 people (Mrs Kyte is part-time), and the company’s customer list is be: ning toread likea Who's Whoof British blue-chip electronics business in the south of England. In its first year, the young busi ness had a turnover of over £125,000 and hope to increase this to £280,000 to £300,00 for the current year. Says David: "Half our enquiries are now coming from overseas, from coun- triesas far alield as Hong Kong. Thailand and Finland. Usually, we deal with the business direct, but recently we appointed a distributor in Italy. We have regularenquiries irom people who wish to become our distributors, but we arc looking for the right people to handle our business. We would rather keep on looking than appoint just anybody". Members of the Micro AMPS team were enjoying good business growth until theautumn oflast yearwhen two of their major customers went into liquidation. Although thisdid notmean a direct financial loss. it did remove partofthe flow of regular project work. Locality has been very important to the business for, as David explains: “The very fact that the University of Surrey {5 on the doorstep has led to the situation where customers drop equipment off at 6 p.m. on their way home and call back at 9 a.m. the fol- lowing morning to collect the modi- fied boards. “We arecurrently working very hard, often until late at night to ensure we carry out project work on time for cu tomers”, Says David. Despite the present general eco- nomic difficulties, the company is makinga modest profit. which should increase as more and more of the li- censing fees begin to build up. Recent work lor Rank Xerox should alsostart sending more money intothe Micro AM PScoffers.as sales of Rank's new printer begin (o mount. “They are forecasting substantial sales in the first year, and we receive a royalty on each unit sold”, says David. ‘The extension of the contract de- velopment customer base comes from recommendations from satisfied cus- tomers. However, Micro AMPS is not content to leave it there, but is in- creasingly concentrating on the mar- keting of its products, Amongthe ever -growinglistofhard- ware and software products the com- pany offers are ICE 751 - a low-cost development and programming platform for the Philips 87C751 microcontroller. The ICE 751 uses a bond-out version of the 87C751 processor connected to 1) a dual port memory which emu- lates the on-chip EPROM and 2) an emulation cable that can be plugged into the user's target board. The ICE 751 hasan RS232 bullered serial port to pass data to, and accept com- mands from, a host PC. Once the application is developed, the emu- lation cable isremoved and an 870751 device can be programmed using the on-board programmer. ICES1- alow-cost In-Circuit Emulator for the 8031 family of processors, The board offers many features normally only foundonexpensive ICE systems. Its low power requirements enables the board to be powered from the target system. 32 Kbyte of RAMis pro- vided internally within the proces- sor which may be used as internal program. external data memory or a combination of the two. A small monitor program resides in the ICE which is driven by a PC serial port. ‘The PC software provides powerful debugging facilities. All of the inter- nal RAM is available to the user for application software. The host pro- vides tools to single-step through a program, disassemble code, set break points, down- and up-load Intel hex. format files, interrogate and assign registers, internal and external, and code memory. PEB 552 — a low-cost, easily cus tomized single extended eurocard for the Philips high performance PCB80C522 microcontroller. The PEB 552 can be programmed in high evel languages such as Forth, Basi C, PLM, or conventional assemblers, ‘A major design feature Is its large 1138 wire-wrap prototypingarea on a 0.1" grid. All of the mieroproces- sor signals. 1/0 decoding and an uncommitied 96-pin connector are conveniently routed to the edge ofthe prototypingarea, The fully expanded memory offers 124 bytes of RAM and EPROM and 4 kbytes of recontig- urable decoded I/O space. The PEB 552 has been designed to, program EPROMs or the 87C552 processor via an on-board adaptor Micro AMPS, 66 Smithbrook Kilns, Cranleigh, Surrey GU68JJ; Telephone (0483) 268999: Fax (0483) 268397 ELEKTOR ELECTRONICS SEPTEMBER 1991