CONTENTS February 1990 Volume 16 Number 175 Theme ofthe month in March wil be Components ‘so inthe March issue: Sinewave inverter Bucket brigade memories Surge plug Pause switch for cam- coders + Low-cost Wl display module + SAVE decoder: Part 2 + Temperature compensa tion for LCD madules Front cover New singers and musicians, like Kathy ang lan shown here, can now set up their own home recording studi, using the 16:track console ‘developed by Remix to pro- fessional standards but at a ‘non-professional price. ‘The Remix has proved atrac- tive not ony to bands, com posers and song-writers who ‘want to mix and record at home, but also to smatir, mmigpriced commercial music and fm studios Features include: 6 ausiary sends, separate tape nouts peak and status LEDs, +4 48 fo-108 selectable, 8 bus with {6-track moritring, 8 eects etuns, 100 mm smooth action faders, direct outputs on all channels, send and etum patch points throughout, and monitoring with equalzation. LEADER = 11 Launch of the digital communications era AUDIO & HLFI 23° Calsod now even more versatile Areview 34 PROJECT: Feedback killler by T. Giffard AUTOMOTIVE ELECTRONICS 22 PROJECT: Car thelt deterrent by David Butler COMPONENTS 2 Introducing the OP-series opamps by J. Ruffell 44 AC-DC current sensors COMPUTERS & MICROPROCESSORS 14 PROJECT: Initialization aid for printers by A. Rigby DESIGN IDEAS 32 Waveform modulation ofthe mains voltage by AM. Katie GENERAL INTEREST 53 PROJECT: The digital model tran - Part 11 by T. Wigmore 62 PROJECT: Dark-room clock by A. Rigby ; INTERMEDIATE PROJECT. i 57 Part 8: Reflex MW AM receiver by J. Bareford RADIO & TELEVISION 27 PROJECT: SAVE decoder - Part | by PNP. Wintergreen 38 PROJECT: PC Radio an ELV design 47 PROJECT: by A. Rigby ‘TEST & MEASUREMENT 18 PROJECT: Capacitance meter by D. Folger ‘MISCELLANEOUS INFORMATION? News 12 &13; New books 46; Events 51: Corrections 5 Readers’ Forum (letters & switchboard) 60; Readers services 67 Buyers guide 74; Classified ads 74; Index of advertisers 74 CORRECTION In the description of the pho- tograph on the front cover of the December 1989 issue of Elekior Electronics, we stat- ced that the 4.5 m dia. antenna hhad won a 1989 Queen's Award for Technologi ‘Achievement. This was based on wrong information, In fact, Precision Metal Lid, the designers and manufae- turers of the antenna, received a 1989 Queen's ‘Award for Export Achieve- \ Wii Initialization aid for printers pelt \ = Capacitance meter, p. 18 SAVE decoder, p.27 ELEKTOR ELECTRONICS FEBRUARY 1990NEWS SWITCH OF EUTELSAT SERVICES EUTELSAT, the European Telecommuni- cations Satellite Organization, has switched some services from EUTELSAT L-Fl at 16° E to EUTELSAT I-F2 at 7° E and vice versa, Transferred from Fl are full-time leases while in return telephony, European Broadcasting Union (EBU) transmissions and some occasional-use taflic is wansferred from F2 to Fl. ‘The changes are carried out to guaran- tee complete continuity of all services on BUTELSAT satellites. EUTELSAT LPI is now entering inclined orbit alter almost seven operational years, All telephony and most EBU trafic can be carried by a satel- lite im inclined orbit since transmission and reception are conducted via gateway earth stations that track satellites in space ‘with great precision, ‘THE ATV COMPENDIUM ‘The British Amateur Television Club has recently published the latest issue of its magazine CQ — TV. This includes as usual a varied and interesting selection of arti- cles, ranging from general topics such as “Using oscilloscopes” by Mike Wooding, G6IQM, and “Broadcast band DX-TV re- ception” by Gary Smith and Keith Hamer to more technical articles such as “Camera tubes explained” by Peter Delaney, G8KZG. CQ -TV is a must for all those who are interested in this stimulating branch of amateur radio. You can obiain this issue and much more by joining the British Amateur Television Club. For de- tails write (enclosing a SAE) to Mr D. Lawton, “Grenehurst”, Pinewood Road, High Wycombe HP12 4DD, NEW SWITCHMODE ICS Siliconix has introduced two new switeh- mode power-control ICs, the $i9112 and $i9120, which include the first device ca- pable of operating in. ‘universal-input’ ower supplies from either 110 V or 220 VAC power lines. ‘The new devices allow the design of highly etficient (greater than 80%) power converters with fewer components than other semiconductor solutions, so that de- signers can improve system performance ‘while reducing board space and compo- nent costs, CUSTOMER REVOLUTION IN ‘TELECOMMUNICATIONS British Telecom has implemented one of the world's most challenging computer projects to transform its services 10 cus- tomers. ‘The project, known as Customer Ser- vice Systems, draws together all the main elements of the customer services British Telecom provides. All 23 million cus- tomers of British Telecom are now able to benefit from its advantages. Basically, itis state-of-the-art informa- tion technology in action, ensuring that ‘customers get the best possible response to their needs, FIRST COMPUTER MUSIC LEARN- ING CENTRE IN THE UK Equipped with a wide range of the latest MIDI instruments, video and computers, EMR's Computer Music Learning Centre (CMLC) is based in Southend, Essex, and offers a unique opportunity 10 gain first- band experience of using EMR's extensive music sofware and hardware. The CMLC will intially provide one- ay courses for beginners and more expe- rienced users during holiday periods and at weekends, as well as raining for teach- rs and dealers. The centre has been inthe pipeline for two years, with EMR providing more and more seminars to education and user aroups in the UK and Burope, particularly on the Archimedes, althoug they have pro- duced systems for the Spectrum, Com- modore, Amstrad, MSX and BBC Micro since the company started in 1983, ‘The courses are under the direction of Mike Beecher, tas, Grsw, whose music demonstrations at exhibitions will be well known to many of our readers. His past experience as director of a large Essex ‘musi school and first London synthesizer school, creator of Electronics & Music Maker (now Music Technology) and Home Studio Recording magazines and frequent lecturer in Europe should make your day a worthwhile event—he has also designed the software you'll be using with his team of programmers, so you'll be finding out new ways of making music with your ‘computer, Fall deals from EMR CMLC, Suite 1-3, 50 Hamlet Court Road, WESTCLIFF-ON- SEA SSO 7LX, Telephone (0702) 335747. US TAKE-OVER BOOSTS RACAL'S CAE PRESENCE Britain's Racal Electronics Group, already dominant in computer-aided design (can) systems for printed-circuit boards, has strengthened its presence in computer- aided engineering (cat) with @ £12 million take-over of the American HHB Systems simulation software company. HEB Systems, based in Mahwah, New Jersey, was set up in 1977 to develop and sell high-performance logic and fault sim- ilation software for use in the design and test of complex electronic circuits. Since then, the company has gained over 200 customers worldwide with sales worth $12 million a year and its products have be- come established as industry standards, ‘The two companies have a working re- lationship dating back to 1984, and HHB's products are already fully’ integrated within Racal-Redac’s workstation-based electronic design autopmation (EDa) sys- tem, Visula Plus. At present, the combina- tion of Visula Plus and caDar is the only proven and commercially available systom that can design and simulate application specific integrated circuits (asics) in the context oftheir target systems. POWERFUL WEAPON AGAINST ‘CANCER ‘A new computer developed by the Active Memory Technology (AMT) company of Reading is contributing to what has been described as one of the greatest scientific adventares in human history: to map and identify all human genes. The aim of the project is to decipher the complete set of biological instructions used by nature to make a human being. The resulting “Book of man’ may provide the basis for preventing or treating most hhuman diseases in the next century ts tar- get completion date isthe year 2000. ‘The human genome consists of chemi- cal beads twisted into the famouse double helix of the deoxyribonucleic acid (ONA) molecule. There are four types of bead, IRUARY 1990known scientifically as bases. They are the four chemical letters of the genetic code (A,G,T and C) discovered by James Wat- son and Francis Crick at Cambridge Uni versity in 1953. Watson is one of the lead ers of the human genome project in the United States DUAL-OUTPUT HAL SWITCHES Industry's first family of du effect switches with outputs that are in pendently activated by magnetic fields of opposite polarity is available from the Sprague Semiconductor Group. [nerd ‘The UGN3235K and UGS3235K Hall ‘effect sensors are bipolar ICs designed for ‘commutation of brushless DC motors and ‘other applications that use mult-pole ring magnets in industrial and automotive envi- ronments MORE HEADACHES FOR IBM Shortly after the news that IBM's trade figures were “disappointing” and that the company was consequently laying off 10 000 employees, it was announced that the US Navy is to impose an indefinite ban on doing business with IBM, ‘The Navy decided on the ban after it hhad come to light during a House Govern- ment Operation Committee hearing that IBM had defrauded the US Navy by sell- ing it used computer equipment as new. As if that was not enough, a group of clone makers* have defined their EISA 32 bit bus standard, which is compatible with the ISA—Industry Standard Archi- tecture—system. IBM had hoped to keep the clone makers out of the 32-bit market by keeping their 32-bit (with data path) standard—Microchannel Architecture, ELEKTOR ELECTRONICS FEBRUARY 1990 MCA—to itself. It seems, however, that IBM has seriously underestimated the strength of the clone makers, *AST Research, Compaq, Epson, Hewlett- Packard, NEC, Olivetti, Tandy, Wise, Zenith SATELLITE ACCESS SYSTEM TO BEAT TV PIRATES A new British Telecom satellite manage ment system will allow broadcasters to keep track of their customers and avoid piracy of TV entertainment services. It can also prevent pictures being received in specified areas. The BT Vision system uses transm sion and encryption technology to ensure that only authorized subscribers can re- ceive satellite TV programming, and of- fers programme providers the technology to control transmission of potentially of- fensive satelite broadcast material The system effectively prevents piracy by securely encrypting the programme sig- nal with the aid of a complex algorithm before it is uplinked to the satellite. A stream of authorization messages is sent cout along with the television pictures, au thorizing subscribers’ set-top decoders to pt the programming signal. ‘A very useful feature of the system is its ability to black out pictures in specific locations. This means that TV broadcasts will not be received in countries where the distribution rights have not been agreed or where the broadcasting control body con- siders their content unacceptable (so much for those who thought that the age of cen- sorship was over!) NEW SUPPORT SHIP FOR TRANSATLANTIC CABLE Cable & Wireless Marine's new cable ship, the Sir Bric Sharp, is said to be the most technically advanced vessel of its kkind in the world. It has to be, because itis designed to service and maintain the ‘world’s first private trans-ocean fibre optic link between Britain and the United States. This is the PTAT-1 system, part of Cable & Wireless Marine's global digital high way linking the major financial centres of the world. The vessel will also be used to ‘maintain existing analogue coaxial cables, To enable the vessel to pick up and lay cable in the most efficient manner, i is the first of its type to be equipped perma- rently with a submersible remotely oper- ‘aed vehicle (Rov). EARTH STATION FOR ‘OLYMPUS LINK (One of the two £750,000 mobile earth st tions designed and built by British Aerospace under contract to the European Space Agency has been handed over to fone of Britain's leading academic insttu- tions for communications technology and educational broadcasting, Polytechnic South Westin Plymouth The station, known as TDS 4, will be used for demonstration purposes. It will enable the polytechnic to communicate via Olympus, the world’s most powerful civil communications satelite, which was also built by British Aerospace. The TDS 4 carth station is mounted on 4 12-metre long articulated trailer and in- corporates its own auxiliary diesel genera tor. In operation, it will transmit in the 13.00-13.25 GHz and the 14,00-14.30 GHz bands, and receive signals in the 12.50-12.75 GHz band, The antenna fitted is a 3.5 metre design with hinged sides that cam be folded down for ease of trans- port EEC MEASURES TO LIMIT ADVERTISING ON TELEVISION Britain is supporting the proposed EEC ‘measures to limit the amount of advertis- ing on TV and the number of American programmes shown. The new measures were agreed at a recent mecting of EEC foreign ministers. Under the new rules, no advertisement Will be allowed during news, religious or children's programmes under half an hour Jong. Advertisements will be allowed dur- ing films after the firs 45 minutes. ‘The decision is unlikely to have much effect in the UK where standards arc among the strictest in Europe. :INITIALIZATION AID FOR PRINTERS A. Rigby Many computer users have difficulty in changing printer settings without the help of a manual or handbook. The circuit presented here allows one of sixteen pre-programmed printer settings to be called up at the flick of a switch. Very useful for a good many computer programs that lack a real printer driver, the initialization aid is a low-cost and simple-to-build circui ‘You are working hard on a BASIC pro- gram which is almost finished, and hard copy on paper is required to do the final debugging, Since the program is fairly long, you think it wise to set the printer to condensed characters at 8 lines per inch. ‘The printer manual is consulted and the following command is typed in the direct command mode of GWBASIC: LPRINT CHRS(15);CHRS(Q27), We think you will agree that pushing a small button on the printer initialization is a lot easier than having to look up eight or so ASCII characters, and typing, them into the computer whilst observing the correct order, brackets and delifniters. In the above example, BASIC forta- nately allows the printer to be set to the required mode, albeit in a rather complex way prone to many errors, There are, however, many programs that lack even the most rudimentary means to set the printer to a particular character or paper format. For such programs, the initializa: tion aid isa useful peripheral, provided its setting is not overridden by a printer in- itialization string prefixed to each print- able file by the program in question. Basic operation The flowchart in Fig. 1 illustrates the basic operation of the circuit, which is inserted between the computer's Centronics port and the parallel input of the printer. The diagram shows the functional blocks in the circuit as well as the interrelated tim: ing at which these blocks operate. ‘At power-on, the circuit is reset and forms a straight 36-way connection be- ween the computer's Centronics portand the Centronies input on the printer. No change is made to the previously estab- lished printer setting, and the computer can send printer files as before. When St is pressed, data buffer ICs blocks the datastream from the computer. ‘The circuit actuates the BUSY line to force the computer to stop sending data. Mean while, the data outputs of the EPROM, ICz, are enabled, s0 that the byte at the currently addressed location in the EPROM js sent to the printer. A strobe pulse is generated to signal to the printer that a byte is held ready for transmission Since the start of the first strobe pulse coincides with the moment the EPROM datalines are enabled, the strobe pulse is delayed by about one microsecond to en- sure stable levels on the datalines. After the STROBE output line is actuated, the circuit enters a wait loop to allow the printer to process the databyte. Depend- Ing on the position of jumper JP1, the cir- cuit waits for a negative pulse transition on the BUSY or ACK handshaking line from the printer. When the transition ar- rives, the EPROM address counter is in- cremented by one. The counter starts at the lowest address, nil, by virtue of the power-on reset, and the fact that the cir- ‘uit stops automatically when the counter reaches nil. In the latter case, the "all 16 bytes sent?” loop in the flowchart is left via the "yes! exit, and the cireuit forms a straight connection between the computer and the printer until Sz is pressed again, ‘AI! 16 bytes stored in the EPROM are to the printer under the control of strobe pulses, which are started with the aid of the delayed BUSY or ACK printer signal The delay allows some time for the ad- dress counter to reach thenext higher out- put state, When all bytes have been sent, the same delay is used to reset the circuit to its start configuration, 16x16 in control A part of the circuit drawn in Fig. 2consist of a set of printed-circuit board tracks called the control bus. This bus takes all Centronics control signals not used by the initialization aid from the input: to the output connector. Two lines, CISand C35, form an exception, because they may be used to power the circuit. Whether oF not they can be used for this purpose depends ‘on the printer used. Diodes Da, D and Ds form an OR function that allows the cir- cuit to be powered by different supplies. In case the printer lacks a +5 V output on its Centronics connector, the circuit is powered bya9 VDC mains adapter via Dz and voltage regulator ICs, If the printer supplies #5 V via line C18, the circuit is powered via Da. 145 Visavailableon line C35, Deis used for the same purpose, The current consumption of the circuit does not exceed 50 mA. As already discussed, the circuit is reset automatically the moment the sup. ply voltage is present. Bistable FF: is reset by Ro-C2, and in turn clears address counter ICS. Bistable FF! is configured as ELEKTOR ELECTRONICS FEBRUARY 1990‘meres ae Fig. 1. This flowchart il 2 monostable multivibrator (MMY) and does not requirea power-on reset because it goes to the stand-by state on its own. At this stage, the circuit functions as a 36- way connection between the computer at the input (K») and the printerat the output (Ka). The two devices are connected via three-state inverter/butter ICs, inverters No-Niw and the gates inserted in hand- shake lines BUSY, STROBE and ACK, When Sz is pressed, its contact noise is eliminated by debounce network R7-Cr. ‘The short pulse generated by Cr-Rs pre- vents the initialization sequence being re- started on completion ifthe switch has not ‘been released in the mean time. Actuation of S2 causes Ff to be set. The resultant high level at the Q output of FF2 causes the outputs of three-state buffer IC8 to be switched to thehigh-impedance state,and theBUSY and ACK inputs ofthe computer BLERTOR ELECTRONICS FEBRUARY 199) tes both the Structure of the circuit an the pertinent timing. to be taken high by NINir and No-Nis respectively. The low level at the Q output of FF2 enables the databuffers in EPROM ICryand ends the reset state of counter ICs. ‘The fist strobe pulse is generated by FFi Which receives at its 5 (et) input a short pulse from the Q output of FE: via net- work Riv-Cs. The length of the strobe pulse is deiermined by Ri-Ci. Capacitor Gi slowly discharges after FFi has been set,and resets the bistable. Consequently, output Q goes high, so that Cs is rapidly charged again via Dr. This ensures that FFL is ready to generate the next stzobe pulse The strobe pulse is delayed by about cone microsecond in network RnCr to allow sufficient time for the EPROM data to reach output connector Kz. Gates Ni and Nw feed the strobe puise to the printer Yt — Table 1. DIP switen settings, ‘The circuit waits until the printer is ready to accept a new command by moni- toring either BUSY or ACK. The selection between these handshaking signals is made by the user with jumper JPI; the two are equally suitable, provided the printer supplies the relevant signal. Whatever signal is used, the negative pulse edge signals readiness to accept a new charac- ter. The pulse edge clocks address counter ICS and causes a new strobe pulse to be generated. The start of the strobe pulse is however, delayed by Ri-Cs to prevent an- other pulse being generated when the Toth (last) byte has been sent to the printer. This delay is used by the circuit to Block FF before it receives a new clock pulse. The end of the printer initialization sequence is marked by output QD of the address counter going low when count 16 4s passed. The change from 16 to ( causes FF: to be clocked, so that FF is blocked — hence, the clock pulse generated in the mean time by Ru-Cs has no effect. The toggling of FF2 also resets the circuit toits initial state The circuit is fully compatible with the Centronics interface on the computer as well as on the printer by virtue of pull-up resistors and open-collector TTL drivers respectively. DIP switch block $1 selects fone of 16 printer initialization strings which have been pre-programmed in the EPROM. The selected address. ranges in the EPROM are listed in Table 1. If fre- {quent changes are expected in the setting OF this switeh, it may be replaced with a more ergonomical type or a set of swit- ches, e.g, four miniature SPST types or a thumbwheel switch, Construction ‘The construction of the circuit on the double-sided, through-plated printed-cir-®®Oo_d eo ice Gs Fig. 2. Circuit diagram of the printer initial , a very useful periph device for advanced computer users. ELEKTOR ELECTRONICS FEBRUARY 1990Fig. 3. Centronics connector pinning, cuit board shown in Fig. 4 is straightfor- ward. The pins of the Centronics eannec- tors are simply pushed over the relevant copper fingers at the edges of the board (ind the position of pin 1 of each connec- tor). Next, solder the fingers to the pins, taking good care to avoid short-circuits between adjacent pins. Bolt the voltage regulator direct on to the board —a heat sink is not required, Printer command strings Upon completion of the board, the EPROM must be loaded with data. The manual supplied with your printer should give ample information on com- mand strings to achieve the settings you require. The circuit allows a maximum of 16 command strings to be programmed, each with a fixed length of 16 bytes. The mini EPROM programmer described in Ref.1is suitable for loading the 256 bytes. If the required initialization sequence is shorter than 16 bytes, the remainder must provide null characters (00), or any other character which is ignored by the printer (again, consult your manual). ‘The sequence of commands to be pro: grammed in the EPROM Type 2764 is first tested ‘on line’ with the ald of, say, GWBASIC. Document the final versions ofthe command strings. Next, compile the INITIALIZATION AID FOR PRINTERS 10 RESTORE 12 OPEN "pea =" 20 FOR N-1'TO 16 30 READ X 440 PRINT #1 ,cHRSCX) 50 NEXT W 5 Lose {60 DATA uiS4 cia {65 DATA G00 Hot 7 Bs POR OUTPUT AS DATA 00, HOC Fig. 5. GWBASIC test program listing databytes to be programmed from databyt data, ‘This inverts each byte to negate the inver sion in the output butters Reference: 1. Mini EPROM programmer. Elektor Electronics January 1990, Fig. 4. Component mounting plan for the double-sided, through: ELEKTOR ELECTRONICS FEBRUARY 1990 Parts list Resistors: Ry = SIL resistor array @10 Ro R= 104 Br =1Mo Ramio.Ris — Fis = 100k G2 470n Ga:Ge = 1n0 Con an7 s=470p GriGarc19~Cre= 100" o= 3300 Semiconductors: Dr = 1Na148 De:On:De = 1Ngo0 Ci = 74HCTOR eeu TaHoTI32 Cais = 74HCTOS ICs = 74HCT93 Ie = 74HCT74 G7 = 2784 or 27084(A) IGa = T4HOTS«O 1G0= 7605 Miscellaneous Ki + 96-way female Contioncs connector wth straight pins Ke= 96-way male Centronis connector with straight pins. Si = 4-way DIP switch block Se = ITW Dataswtch push button S-way pin header. Jumper. PCB Type 900007 (doube-sided: through Plated See Readers Services page).CAPACITANCE METER A capacitance meter is in D. Folger pensable for checking capacitors with illegible or incomprehensible values printed on them, and for matching capacitors in, for instance, higher-order filters. The instrument presented here is based on a simple circuit, has a handy size and five measurement ranges. ‘To own a capacitance meter with a maxt- imum error of 1% and a capacitance range of 1 pF to 1 Fis wishful thinking for many blectronics enthusiasts because such an in strument is not affordable, if it is at all available. Fortunately, not many applica- tions require a capacitance meter with such impressive specifications. Consider, for instance, the selection of capacitors for a higher-order audio filler: here, the relative difference is much more import- fant than the absolute value, and a simple capacitance meter may be used with im- ‘punity to select matching capacitors from an available lot. Measurement principle ‘The measurement principle used ts fairly standard — see Fig. 1. The heart of the circuit is formed by a time-base which triggers a monostable -multivibrator (MMV). The time constant, f, of the time base isset toa value that exceeds the maxi- ‘mum monostable delay, The MMV output changes from low to high on the negative edge of the time-base signal. The time it takes for the MMV to revert to the start state, 4, is proportional to the value of the capacitor under test, C,. Since the dura- tion of the measurement cycle is deter: mined by !, an integrator may be used to providea voltage which is proportional to the value of C, Circuit description Im spite of the relatively simple measure: ment principle described above, the prac tical circuit presents a number of possible problems related to the translation of the unknown capacitance into an accurately determined numerical value. The time-base is formed by a I-MHz quartz crystal and an oscillator/ divider, ICs. Output Q13 supplies a frequency of 1 MHz/16,384 = 61.035 Hz. This signal triggers IC%, one of two MMVs contained in the Type TLC556 LinCMOS dual timer. ‘The choice of the oscillator frequency isa compromise between the stability of the read-out and acceptable current con- sumption when relatively large capacitan- ces are measured. Given a time constant 1=1/61.035 64 ms, t MAIN SPECIFICATIONS Measurement ranges: 2.nf; 20 nF; 200 nF; 2 uF; 20 uF with: overflow indication Max. resolution: pF ‘Max. error: 5% Display: Se digit LCD Scale factor: nF /iF Power supply: 9-V PPS bat- tery ‘Current consumption: 10-20mA a maximum monostable time of 10.5 me and a maximum capacitance of 20 uF, the resistor R in the RC delay network is simple to calculate from fy = LARC R= 105/(1.1020 uF) 70 A practical value of 475 0 (1%) is used since this is available in the E96 series, The ther measurement ranges are created by multiplying the basic value of R by the capacitance range factor, Le., 10: R = 4k75 for the 2 uF range; R = 47K5 for the 200 nF range, R = 475k for the 20 nF range; R 4M75 for the 2 nF range. The output signal of ICw has a duty factor ty/t and requires averaging, to ob- tain a direct voltage proportional to the uty factor and, therefore, to the value of Cyc Integration capacitor Cisis charged via potential divider Ris-Ps-Ris if the MMV ‘outputs high,and discharged ifthe MMV. ‘output is low. This arrangement produces an average voltage, Us = Usavlle/9 Note, however, that because of the ad- justable potential divider the high level, Uys, is lower than the high output level provided by the MMV. The maximum value of Ub is not reached immediately, but alter a delay af about 600 ms, which corresponds roughly to the time constant of the integration network. The 150 pF capacitor connected in par allel with C, is required as @ minimum capacitance in the external RC network of theLinCMOS timer. Without Cs, the MMV may not be triggered reliably if small ca- pacitors (in the pF range) are tested. The resultant off-set is compensated with the aid of a second MMY, ICx, of which the external configuration is almost identical to that of ICx. The low voltage, UL, at the output of the second integrator is used as a refer- fence for Ui. Capacitors Cr and Cs and resistors Ri-Riz determine the time delays set with the MMY, and must, therefore, be close-tolerance types. Spurious triggering and incorrect ca- pacitance indications may occur if ‘measured values exceed the maximum of ‘a particular range. Each measurement cycles, therefore, stopped after 12 ms. As shown in Fig. 3, the remaining 4 ms are cused to discharge the capacitor under test via FET Ts. The timing of the discharge ELEKTOR ELECTRONICS FEBRUARY 1990we, " =~ Tee Fig. 1. period is arranged by the AND diode con- figuration of Ds and Dy at the Q12- and Qi3-outputs of IC These diodes cause the MMV output to remain permanently high when a too large capacitor is con- nected. In this condition, the voltageat the two MMV outputs causes the LC display to indicate ‘1’. The measurement error in- ELEKTOR ELECTRONICS FEBRUARY 1990 Measurement principle and graphs to ilustrate the operation of the A-C integrator. troduced by Te is compensated by its counterpart, Ti, in the other MMV circuit. ‘The display circuit is based on the well- known Type ICL7106 A-D (analogue-to- digital) converter with integrated 3ledigit LCD (liquid crystal display) river. The ICL7106 is used in a standard application circuit with the decimal point capacrrancemerer BR} switching arranged by Su, Dis-Dw and Nn-Nu. The scale factor (nF or UF) indica- tors, Di and D2, are driven by a diode circuit, Dao-Ds ‘The power supply of the eireuit is of a less usual configuration: a Type 79L05 voltage regulator is inserted in the nega tive supply line to create a supply with +5V and -4V outputs, The non-regulated negative voltage is applied as bias to the ADC in the ICL7106, which requires that the voltage at the Y- input {pin 26) is al- ways 1 Vibelow that at INH ‘The circuit around Ts controls the Low baTsymbol on thedisplay. With Ri==220k, this circuit is actuated when the battery voltage drops below about 8 V. Construction and adjustment The complete circuit is built on a printed Gircait Board that fits into an ABS, stand ard-size hand-held enclosure with integral battery compartment. The circuit tothe right ofthe dashed line in the circuit diagram, and diodes Dis-Dis, may be omitted ifthe capacitance meter is used in Conjunction with a digital multimeter that assumes the display function, "The population of the PCH should not present problems The high-value 1% re tors, Read Ri (4.75 MQ), may be dif Fiult fo obtain in small quantities, They aay, however, be selected froma batch ot 127 Ma.5% types with the aid of a DMM Fit two Tou or ivesway. terminal blocks om tothe front panel ofthe enclo- sure for conassting the capacitors to be tested. Be sure to se hight-qualty term- nals toavoid problems with teat and tear of the contacts, A two-way loudspeaker terminal block ior wire insertion i a fine tlterntive to multi-way Black since it provides fairy wide contacts which allow Capacitors with sifferent terminal space sng to be inserted ® Lange capacitors of which the wires can note inserted into the termina blocks on the instrument must be connected with test lends, Do not use these lends for small Capacitors sine the vie capacitance will make the indicated value worthless. Ps iseither a preset with an adjustment spindle ora potentiometer with a 4 mm spindle. It allows test ead capacitance to be compen- sated. ‘The accuracy of the instrument is determined mainly by the quality and tolerance of resistors Rs-I @ te SS o truss Fig. 3. The measurement cycle is stopped 12 ms to discharge too large capaci- tors, ‘Circuit diagram of the capacitance meter. Connect a 1% or 2% polystyrene (sty- roflex) or silver-mica capacitor with a known value in the lowest measurement range (e.g., | nF). Set Sto the 2 nF range, set Pr to the centre of its travel and Pz and Ps to maximum resistance to ground. Con- nect the reference capacitor, and adjust Ps until the correct value is displayed. Dis- connect the reference capacitor and adjust the zero-indication of the meter with P2. Repeat the adjustments of P2 (capacitor connected) and Ps (capacitor discon- nected) until the indicated value and the zero indication are both correct DMM as display ‘The value of the test capacitor must be adjusted with Ps instead of Ps if a DMM. set to the 2 V range is connected to points Land H. The ‘meter-zero' control, Pi, is replaced by a 5.11 kO 1% resistor in this cease because tho adjustments for meter zero and capacitor value interact. Set Ps to maximum resistance to ground, and ad- just P2 until the DMM reads 0 V. Connect the reference capacitor and adjust P3 until the correct value is indicated. Repeat the adjustments of Pzand Ps as detailed above for Pzand Ps. Close-up of the battery compartment. TOR ELECTRONICS FEBRUARY 1990CAPACITANCE METER odeooegogogoooogC 2 | Fig. 4. Component mounting plan of the printed-circult board for the capacitance meter (eft) and suggested front-panel lay-out (right). Parts list Resistors: Bia 10M RePaaiRer Raza 100k ae2Tk Re 47509% ao 475 1% Reo = 4746 1% Br = 475K. Finis = 4M7 59 (508 tox) Ribs 15K 1% iets «106 1% Pref M0 Rien tk5 isa = 220k Ft = 1MB 6% Ra = 200K 18% Bere 22k 1% y= 10k near potontomoter wit 4 ma spindle or 10k preset for horizontal mouni- ing with spindle. Pa = 10% muliium proeat Pe 2k mutitum preset ‘Capacitors: CHOnCei Cre = 2p; 16 V5 raat G2 toon Cats =a Gee 100 70a = 150p polystyrene (siyrctox) Cxci0= ton ‘Cisi@is = 100n MKT Da Dat= Na148 TiFe=B8170 G2 = 74¢TA060 Ics = Theses (Unemos) ies= co40a0 (Gs mici7106 Miscellaneous: Si = mitialur side SPST slide switch Se 2-vole way PCE-mount rolary switch. Xi 1 MHz quart crystal LCD! = general purpose 3Yediglt Lc ole. play, '3:V PP3 baltey wih clip and leads. Hand-held ABS enclosure, e9.. Monacor PLG7BOBN (160%80%75 mm), ‘Loudspeaker terminal Bock, PCB Type 900012 (see Readers Servicus age). ELEKTOR ELECTRONICS FEBRUARY 1990CAR THEFT DETERRENT by David Butler Cars seem to attract thieves, probably because of the rel of entering them. Anyone who has locked their keys inside will testify that a bent coathook can usually secure an early reunion. The deterrent proposed is aimed at discouraging casual joyriders who want not want to risk setting off an alarm — even if, as here, ‘one did not exist. In fact, the deterrent is simply an alarm type indication that, if accompanied by s ble warning stickers, should provide basic protection. It must be stressed, of course, that this deterrent does not act as an alarm and it would be wise to invest in a security system, such as an immobilizer or a Krookloc. ‘The idea of a dummy alarm is not new, but the present design adds more credibility by the use of a single dual function LED. ‘When the ignition is on, the LED radiates steady green light to show that the ‘sys tem’ is ‘disarmed’. When the ignition is switched off, the LED emits a flashing red light to simulate ‘alarm enable’ The success of this idea rests with dis- playing the LED prominently (say, next to the ignition switch) and not telling any- body that itis usta flashing light! The pro- ject is inexpensive, simple to construct and Install, and performs a worthwhile func- tion, Circuit description ‘The circuit is designed around a dual fun tion LED. This device looks like a normal 5 mm diameter clear LED, but has three terminals, The casing actually houses a green LED and a red LED driven by a flasher unit. The cathodes of the LEDs are ‘commoned to the centre pin. When a volt- age in the range 4.75-7.0 V is applied to the ‘R’ terminal, the red LED will flash at a zate of about 1.5 Hz. The green LED is con. nected to the other terminal (indicated by a Features: Deters casual joyriders from stealing your vehicle. Inexpensive, simple design. Compact dimensions. Easy to install (3 wires). flat on the casing) and requires a series current limiting resistor to operate from any supply. ‘The relay used is of extremely small di- mensions, and has a single-pole change- over contact set. The coil is energized by the +12 V ignition cireuit in the car, with back e:m. protection provided by D2 The LED terminals are connected to the relay contact set so that normally the red flashing LED is actuated (ignition off state). The regulator circuit formed by Ri, D1 and Ci provides a fixed voltage for both LEDs, This was chosen to avoid problems ww Cont] COMPONENTS LIST Roster: Fi S609: 025 5% bear satan 22208025 4h orate amon capac (Ce on cae Semiconductors: (SY 400 mis sone ide tee 1001 EDT - cninwnepreatasing rad LED ‘Map Econ ode code YS). iscoaneous: FLA= mice mira spl pce charger cena 1232080 (apn cones or cade BKA7B) Lt = dma ghar PCB ype plo ‘3 vay houing pus eos SrmUEDep. POR vcore Ina hole wih 150 mA 20 rm ce fod) Fig. Circuit diagram (igition switched off. Fig. 2. Printed-circult board ofthe deterrent. The ‘component side is shown tre to scale. BLERTOR EI ECTRONICS FEBRUARY 1990Fig, 3. Pinout of the relay (seen trom the base) swith the car's voltage supply ‘The cireuit consumes about 50 mA when the ignition is on and around 15 mA ‘with the ignition switched off Construction notes FRONICS, Once the LED is mounted, three wires run from the PCB may be connected to it The prototype was constructed on a scrap of vero board and then transferred on to a PCB (see Fig. 2). Although the method of construction is not important, care should be exercised in connecting the polarized ‘components: Di, Dz and the LED. ‘with the aid of Seotchlock type break-in terminals. Usually, the set of wites leading to the radio can be used: +12 V continuous means that +12 V is available even when the ignition is switched off; +12 V ign means that the +12 V line is witehed via the ignition. Installation notes ‘The prototype unit was installed in the au- thor's Metro, which conveniently has a spare blanking plate next to the rear screen heater switch, This is quite close to the ig: nition switch and would be instantly no- ticed by any curious potential thief No doubt, other vehicles will have sim: ilar places 16 mount the LED, which re- quites.a 6.35 mm dia, hole fuse holder with a 150 mA fus used: this rating depends on that of the car fuse fitted to the circuit being used. For extra electrical protection an in-line may be Apart from these brief notes, no other fitting instructions can be given as each vehicle will vary. Generally, the unit should be fitted away from sources of direct heat, with the LED displayed in a prominent position ‘where it does not distract the driver CALSOD NOW EVEN MORE VERSATILE The computing and optimalization of loudspeaker enclosures is nowadays normally effected by computer. One of the programs that enables the complex calculations to be carried out on an IBM PC desktop computer is CALSOD, the first version of which was reviewed in this magazine last year. We have recently received an improved version of this program that offers even more possibilities to the professional designer to approach realistic sound reproduction. Designing loudspeaker enclosures re- quires such extensive measurements and calculations nowadays that it has become virtually impossible without the use of a computer and a suitable pro- ‘gram. Last year we reviewed* CALSOD, 2 unique combination of a simulation and an optimalization program. We then thought that there was very little left to bbe desired. None the less, the designers have succeeded in adding some more facilities to their Iatest version 2.00, which bring the results even more closely to realistic sound reproduction ‘The new version offers the possibil ity of working with a coprocessor. This is an especially welcome addition for XT, since computations on these of three or four-way systems are relatively slow (but still alot faster than with comparable programs). However, this facility is merely to do with speed of processing, To us, the most interesting addition is the RAB sub-module. This makes it possi- ble to caleulate the frequency characteristic of the entire system at a given angle (both horizontal and vertical) with respect to the Histening axis, for instance, =30°, Even in the optimalization of the filters this off-a response may be taken into account. It is thus possibe, for example, to design a filter whose characteristic remains within cer- tain limits for an off-axis response of 210° ELEKTOR ELECTRONICS FEBRUARY 1990 with respect to the listening axis (see illus- tration above. ‘The loudspeaker placement has also bbeon extended. The original version en: ables the loudspeaker location to be caleu= lated in a three-dimensional space: the new version makes it possible to take into account that loudspeakers are placed at an angle (for instance, in case of a backward sloping enclosure front). Also, the effective diameter of the drive units may be taken into account, so that the program may make provision for the radiation pattern of each individual drive unit, Apart fom optimalization of the filter components to obtain a given characteris- all the facilities we have disc files are compatible with Version 2.00 ($A349.00) so that transfer to the profes- sional version ata later date is facilitated. tic, itis now also possible to optimalize the impedance of an individual loud- speaker or of the entire system. To that end, compensation networks are added that are calculated by the computer in a manner that keeps the impedance in a predetermined frequeney range within a given percentage of a certain valu. TI is a very useful facility for compensating the bohaviour of individual loudspeakers and for straightening out the impedance characteristic of a complete system, to ensure that the power amplifier is con- nected to a truly resistive load ‘There is also a ‘student’ version (1.20) of the program available at a sharply re= ‘duced price ($99.00). This version offers ssed. Its data CALSOD is av lable from the design- ers, Audiosoft, 128 Oriel Road, West Hei- delberg 3081, MELBOURNE, Australia. It may also be obtained from Old Colony Sound Lab, PO Box 243, PETERBOR- OUGH NH 03458, USA, or, in Furope, from Audio Speci 22, 1000 BERLIN, Federal Germany. lists, Weichselstrasse *Elekor Electronics, January 1989, p. 62INTRODUCING OP-SERIES OPAMPS J. Ruffell Dozens of new, improved operational amplifiers are introduced every month by leading IC manufacturers around the world. A number of high-performance opamps from the ‘young’ OP-series are described in relation with construction projects featured in this magazine over the past year or so. The reasons for preferring these new devices over, say, a Type 741 are manifold and call for an introduction to opamp selection criteria. Designing a wide variety of clever elec- tronic circuits on the basis of ideal oper- ational amplifiers i fairly easy. In net 8 me Ee ee oe few cases, however, the cori fematn sain fo mw | paper desipoa tiny Hoe sgppcintg ama ale] [ofel a Eulfs obtained with practic construc: [sie soa | =| tions. The problem fo clear: the ideal com - * pimp does not exist The ideal det BS aed does have its uses, however, because it ao * ve allows the operation of complex circuits Seoraie [RENEE Tre 5 a based on opamps to be analysed and sanction | in proctc, the Heal opamp ie only oe tefl for the design of low pertormance Somer cirenits, and for function analyses. In all masecies (emo ‘other cases, the most important non-ideal voarenioe T= ae v= ae] characteristics must be taken into account. [Taunt Toaan . The relative importance of all the design ae et parameters involved depends on the type = faire in hich theopamp isto beused inanalternating-vollage anrpifir for in Stance, offset drifts fess important than the aldw rete. Conversely, the perfor mance of a DC amplifies inited by off fet dit rather than the slew rte Distinguishing between the important and the not so important opamp charac loristics on the basis of the function of the circuit not only raises the quality of the final product, but also avoids needlessly Fig, 4, complex calculations, Clearly, a wvell- Main specifications of the OFO7 and OPO7A compared to those of the 741A/EIC. ELEKTOR ELECTRON IRUARY 1990founded choice can only be made on the basis of knowledge and experience, Infinite amplification and input im- pedance and zero output impedance are probably familiar terms in relation to opamps. The real thing, however, starts with off-set voltages, input bias currents, noise, slew rate, and many other factors. Opamp technology Several integration technologies are used im the production of opamips In ther Quest forthe esl opamp, manafacturers are faced witha real dilemma: improve: tment of one opamp characleistic resus in degradation of another, Inevitably, a particular integration technology is Tinked with nent-optimam performance im one or a few respects ohly. A clear Cxample ofthis vicious cle tha an increase of the slew rate iv inevitably Coupled to Sncreased current consump: toa Although not all opamp, charac teristics can be optimized at the same time, today’s production techniques do allow improvements in the performance tobe achieved if theaimisclesry defined ‘This results ina particular opamp perfor ring better thas another in the same ap plication, The use of he mast favourable Production technology for-n particular Eharacterstichas resulted in many opamp Tams with any membors. Currently, the four major technologies are hipaa BIFET, BMS and LinCMOS (Rel 1 Off-set and drift Bipolar opamps such as the ubiquitous 7AL consist of npn and pnp transistors. Bipolar amplifiers have the est specifica- tions as regards stability of the input off- set voltage. This characteristic is often referred to as input voltage drift, or the temperature co-eificient, dUl./dT, of the Input off-set voltage. Field-effect transistors are less easy to match than bipolar transistors, and as a result BIMOS and BIFET opamps gener- ally suffer from the resultant drift effects, Modern bipolar opamps such as the Type OP07 are specified for a maximum off-set voltage drift of 1.3 pV /K, although values better than 0.2 wV/K are no exception, Fig. 2. Standard bipolar opamp input. BLEKTOR ELECTRONICS FEBRUARY 1990 Fig. 3. Opamp input with external offs ‘compensation preset, ‘This means that the drift specification of the OPO? is roughly 15 times better than that of the 741, Extracts from the datasheets of the OPO7 and the 741 are given in Fig. 1 Clearly, the OPO7 has a smaller drift, and the absolute value of the off-set voltage is alsolower, Apparently, the drift specifica tion may be improved by keeping the ab solute off-set voltage, Uy., small. Zener-zap trimming “This technique is used to reduce {ofthe OPO? toa value smaller than 200 iW with cout the use of external components, The function and operation of zener-zap tciming is beat explained by studying the cause of the offset voltage ‘The basic siructure ofa standard bipo- lat opamp input creuit & shown in Fig 2 The onotput voltage, Ly, is the potential diferente between the collectors Ts and Ts Us=A aire = ketasRal I the collector resistors Re are equal in Doth branches, Uy = 0 if Lory equals J, ‘Also, assuming that Ty and T2are perfect- ly matched, ie, identical, equal collector Currents result in equal threshold voltages Ue) and Up ois» From this it follows that the output voltage, Ui, can only become zeroiif the input voltage, Li, equal to Ulserr1)~ Usacray {s also zero. In practice, Ti and Tare never identical, so that the threshold voltages al yao = far are always (slightly) different, Hence, U; must be made equal to the off- set voltage toachieve Uy =0V. In most opamp circuits, the output volt- age is made zero by applying the required bias voltage to the input terminals. Some ‘opamps, however, have separate termi- nals that provide access to internal com- ponents. This allows an external trimmer potentiometer to compensate the off-set voltage without the need of a zero-adjust circuit at the sensitive inputs of the ‘opamp. The principle is illustrated in the cireuil diagram in Fig, 3 ‘The designers of the OPO7 have gone a Iittle further, however, as shown in Fig. 4 the collector resistor consists of a number Of series-connected individual resistors, of which two have a zener diode in paral lol. In the production process, the off-set voltage is reduced to the minimum value immediately after the chip is ready. Ac- curately controlled current pulses are ap. plied to blow (‘zap’) one or more zener Jiodes, which then tuzn into shorting This automatic ‘zener-zap' process allows off-set voltages to be reduced to 50 V (typical). In some cases, provision is made to compensate even that level with the aid of an external trimmer preset, which is usually included in a resistive bridge circuit. This arrangement is used because the temperature co-efficlents of the components in the bridge havea much smaller effect than those in the circuit in Fig. 3. ‘All these goodies may lead you to start replacing all 741s by OPO7s just like that. Remember, however, that the resultant improvement in the circuit performance, ifatall required, does not come cheap: the OPO7 is about 10 times as expensive as the 7A Input bias current At room temperature, bipolar transistors have a much higher input bias current than JFETs Gunction FETs), although this current is virewally temperature-inde- pendent. The input bias current of BiFET ‘opamps roughly doubles for every ten de- Fig. 4. Basic structure of OPO7 input with zener-z0p collector resistors,COMPONENTS ‘grees of temperature rise. In general, the input bias current (at normal operating temperatures) of a commonly used BIFET opamp such as the TLO74 is often higher than that of a good bipolar type. This should not be taken to mean that BIFET ‘opamps do not have advantages over bi- polar types, since they offer much higher slew rate values in many cases. The BIFET Type OPI6, for instance, hasa slew rate of 25 V/us, which makes the device about 75 times faster than the OPO? and the 741 BIMOS opamps such as the Type €A3140 have MOSFET inputs and a bipo- lar output circuit. Their input bias current is strongly temperature-dependent owing to the presence of integrated input protec tion diodes. The use of current mirrors in the OPO7 has resulted in a drastic reduction of the input bias current with respect to the 741 designers should allow for about 50 nA for the OPO7 compared with 500 nA for the 741 Opamp selection Summarizing the above, the choice of an opamp is governed by the type of circuit tis used in, Broadly speaking, there are five application areas for opamps: general purpose low-power micro-power high-speed high accuracy Each of these applications requires a par- ticular type of opamp for optimum perfor- ‘mance. Each opamp family, in tum, has members that are tailored to give opti- mum performance in one respect only The overview In Table 1 may prove useful as a first guide to opamp sélection. ‘The right-hand column gives the best known types in a particular series Reference: 1. LinCMOS circuits, Elektor Electronics July/August 1989, p. 20 ff Application area Primary characteristics ‘Opamp type general-purpose fow-cust Lurat, P02 (6744 upgrade) low supply current (fet ma) ‘TLo2T1; OP21 vory low supply current (e100 uA) Tice71;0R20 high speed high gain-bandwicth F956: A340; Opts, 0F18 high accuracy ‘oPo07, 0F77: (0°27 (low- noi : rte i —_| I, 1, i. Los | I. Fig. 5. Introducing the OPI77 ultra-precision opamp (ilustrations courtesy Precision Monolithics inc.) KTOR ELECTRONICS FEBRUARY 1990P.N.P. Wintergreen The scrambling system employed for the BBC-TV Europe and the now defunct Premiere channel on the Intelsat-VF11 TV satellite is generally classified as a low-level video encoding scheme. Based on controlled interference in the video spectrum it is, however, not so simple to defeat with traditional filter designs that have already caused much hair-pulling and teeth-gnashing among technically inclined dish owners. The decoder discussed here leaves traditional filters for what they are, yet offers a perfectly decoded picture at a ‘The Intelsat-VF11 TV satellite is positioned in geostationary orbit at 27.5° west, and can be received in large parts of western Europe if a dish of 1.2 m or larger is used. The BBC-TV Europe programme is transmitted at 10.987 GHz with vertical polarization, The channel is encoded ac- cording to the SAVE (sound and vision encryption) system, with occasional changes in the interference frequency used. Tackling SAVE ‘The SAVE scrambling system, looks relatively simple at first glance, because the encoding 15 effected entirely in the fre: quency domain. This is in contrast to many other analogue encoding systems, Which use signal inversion in combination with amplitudeshifting and frame and/or line polarity control (Ref. 1. In the SAVE system, the video signal from the TV studio has its amplitude re- duced by 50% before it is inverted. The original amplitude is then restored by the addition of a pure sine-wave with a fre- quency of about 94 kHz. This encoding scheme has prompted many technically inclined dish owners to design and build a wide variety of filters, only to find that the interfering signal is much harder to get rid of than expected. This is mainly because any residual interfering signal, however small, gives an irritating effect ‘on the TV screen. As found out the hard way, the 94 kHz notch is simply not ca- pable of ensuring a perfectly stable pic- ture. Without a suitable decoder, a SAVE- encoded video signal produces a totally unintelligible picture. In principle, the signal becomes at least visible by invert FEBRUARY 1990) cost of £25 or so. ing it and doubling the amplitude, Al- though these measures result in a picture that can be recognized on many TV sets the quality is very poor owing to the inter- fering carrier that is still present at its full SAVE DECODER || + low-cost design: no cuttororder quart crystals) ‘+ complete suppression of SAVE inter ference | + L-Coseltator with amplitude and pase cenit loops fer optimum stabity ‘+ simple-to-connecto any indoor unit + standard clamoed video output: 1Vewit5.0 * auiomatioswitch-over between non-en: coded and SAVE-encoded channels + aulomatic sarch forintererence tre- quency within 1.5 kHe band level. The suppression, in particular, of the inter- ference poses considerable problems if conventional fil- ter techniques are used. The phaseand amplitude recovery of the 94 KH. inter- ference signal is affected by the video signal, and de- signers must take into ac- count that this can have an infinite number of instanta- neous amplitude-frequency configurations because the spectrum is a function of the current picture content, which changes in real-time. ‘A narrow 94 kFz filter, however, will give reason~ able results in many cases. A crystal filter as used in com- munications ideal for this function. Unfortunately, crystal filters are notoriously expensive and hard to obtain, especially for a non- standard frequency like 94 kHz. These problems may be resolved, however, by a compensation circuit with a PLL (pihise- locked! loop), and a VXO (owriable erystat ‘oseillator) that functions as a VCO (voltage controlled oscillator). An advantage of this arrangement is that it can be designed to operate at a multiple of 94 kHz. The block diagram of Fig, 1 shows the basic structure of a VXO-PLL (Ref. 2). AF though not used for the present design, the principle merits a short discussion. The crysial is cut to order for one particu- Jar interference frequency around 94 kHz In practice, however, the interference fre- ‘quency is changed on an irregular basis, so that a number of crystals are required when BBC-TV Europe is to be decoded. Clearly, this is a relatively expensive sol- tution ‘since the station operators can ‘change the interference frequency within a certain range, requiring new crystals to bbe cut. The present decoder is much moreRADIO & TELEVISION flexible in respect of frequency control be- cause itis based on an L-C oscillator that uses inexpensive components, and works at 94 Kriz. A number of special tech- niques, combined with ‘all-analogue’ sig- nal processing have resulted in a stop-band filter (notch) whose perfor- ‘mance in respect of selectivity comes close to that of a crystal filter. The low-cost decoder presented here features automatic processing of any SAVE-encoded video signal that contains fan interfering carrier between about 93 kHz and 94:5 kHz, so that BBC-TV Eu- rope (fi293,275 Hz) is roliably recognized and decoded, irrespective of the current Interference frequency. Spectral analysis ‘The frequency spectra of Fig. 2 illustrate the basic operation of pre-and de-emphasis techniques used in the present decoder. The spectrum af Fig. 2a shows the typical amplitude distribution, within the fre- quency range from 0 to 500 kHz, of a video signal, It should be noted that the drawing is purposely simplified: the spee- trum applies fo a completely black pic- ture, while the effect of the raster synchronization pulses is not shown. The simplified spectrum with line syne com: ponents and the interference signal is, however, still useful for a basic analysis ‘The frequency range between 90 and 95 kHz is shown enlarged in the lower spectrum of Fig, 2a. Clearly, the inter ference carriers are quite close to the sixth harmonic of the line frequency. The fre- quency differences are small — only 475 Haz in thecase of BBC-TV Europe, and Fig. 1. The: approach toa SAVE decod nearly always a VCXO design. The t disadvantages are, however, relatively high cost and the ‘one channel, one crystal’ rule (source: Ret. 2). 575 Hz in the case of Premiere — and make selective suppression of the inter- ference frequency fairly difficult because the phase and amplitude of the sixth har- monic of the line syne must be left com- pletely unaffected. If this condition is not met, the picture is visibly distorted Hence, a simple notch is unsuitable for building a reliable SAVE-decoder. For those less familiar with video tech- niques, it may come asa surprise that the interference signal must be amplified be- fore it can be suppressed. Pre- en de-em- phasis are used to achieve this. These Fig. 2. techniques are applied in FM communica- tions systems to improve the signal-to- noise ratio. At the transmitter side, a certain part of the frequency spectrum of the modulation signal is given a greater amplitude, which is reduced again in the receiver by a matched filter to restore the original signal level. Ideally, the fre- quency response between the input of the transmitter and the output of the receiver i flat because the pre- and de-emphasis filters are closely matched by means of complementary curves ‘A two-section L-C filler in the present Frequency spectrum of a SAVE-encoded video signal with no picture content (2a); theoretical pass-band of the pre-emphasis filter (2b); and the result of the pre-emphasis operation on the original spectrum (2c). It should be noted that the spectra shown are theoretical ‘and purposely simpli ELEKTOR ELECTRONICS FEBRUARY 1990SAVE decoder raises frequency compo- nents around 94 kHy and 188 kHz about 20 dB with respect to the rest of the video spectrum. The theoretical _pass-band curve of this filter is shown in Fig. 2b, and its effect on the spectrum of Fig. 2a is ap- parent from Fig. 2c. The interference nal has the highest amplitude in the latter spectrum, The control principle adopted for the SAVE decoder requires the second pass- band at 188 kHz. The analogue control circuit that is to supply the regenerated 94 kHz signal does not have filters to en- sure a well-defined phase response, and ‘multiplication of the 94 kHz signal inevit ably produces a 188 kHz component. (Owing to cross-talk, a part of this compo- nent ends up in the video signal, where it produces interference. This is prevented, however, with the aid of a de-emphasis filter, of which one section is tuned to 188 kHz. Since the interference frequency is al- waysan odd multiple of the interlace fro- quency (25 Hz), residual levels of the 94 kEiz signal that remain after decoding are virtually unnoticed owing to the opt cal averaging function of the human eye. After removal of the 94 kHz compo- nent, the signal amplitude must be re- stored (0 the level before pre-emphasis. This is achieved with a de-emphasis filter whose pass-band curve is accurately dimensioned to form the inverse, of the pre-emphasis curve. Block diagram ‘The basie operation ofthe SAVE decoder is best deseribed slong the fines Of the block diagram shovn i Fg. 3, The heart ofthe circuit is formed by an {-C oscillator, which s controled by loop circuits for amplitude and phase. it prin Ciple, the regenerated 31 kHzsine-wave fudded tothe scrambled video signal. The amplitude ofthe regenerated sine-wave is identical to thatoftheinterforencecompo- nent, But the phase is opposite. This ts Achieved with the aid of a control circuit that fs eapable oF continously monitor ing, and sf necessary, correcting, the phase and the amplitae of the eenet- fed 94 KHL signal The oscillator used is 2 voltage-controed Colpits type based on an LC tuned circuit with 3 felatvely high Q-factor In combination with a double loop filter that forms part of PLL, seenwtres good phase ability of the re generated sine-wave when the circuit is Tocked The block diagram shows that the un- clamped video signal supplied by the sat elteeTV receiver is ampliied before ts passed through the preemphasis ter Siseussed shove, The output signal ofthe twostage fC filter fcompared to that of the 94 Kee oscillator. The phase compara- tor san analogue maltipiter available in IC Type XR2208 from Beat. The result of the multiplication operation isa dik ference frequency that is fed tthe frst, loop liter. This filter is realized with the ELEKTOR ELECTRONICS FEBRUARY 1990 el LSS 5 Fig. 3. The block diagram of the SAVE de ult is based on a 94 kH2 L-CVCO surrouns opamp available in the XR2208, When the Glreuil i not locked, the frst lop filer is given a bandwidth of few hundred Hi (enable the PLL to lock onto suliciently strong signals in a frequency. band of about 1.3 Kriz around 94 kHz, The PLL does not lock om to signal blow a certain threshold Tevel so that, for instance, the sixth harmonic ofthe line frequency can not switch the cleuit to the decoding mode. Without special metsures, how ever, the sixth harmonic of the line fre ‘quency is likely fo cause problems once the PLL has locked om fo, ay, the SAVE fncaded BBCTV Europe signal. In that onaition, the phase comparator supplies {73 the frequency dference bebween the sith harmonic and the Interference Signal, Without a correclly dimensioned loop titer, this difference frequency gives @ to phase modulation of the ¥CO, Which i turn, ends to incomplete com’ raation ofthe interference, The results fn annoying” interference ‘on the TV ‘The solution to ths problem has been found in attomati bandwith reduction of the loop file. Afterthe PLL haslocked, ‘coder is faitly complex. The operation of the wed by control loops for amplitude and phase. the loop filter is given a pass-band of about 15 Hz rather than a few hundred Hz, This arrangement requires high VCO stability, however, since frequency devia- ions can only be corrected relatively slowly. Fortunately, the L-C oscillator vused meets this requirement, so that a phase-locked reference signal is obtained. ‘What remains is control of the amplitude, anda shift of phase by 90°, Both functions are realized by an operational transcon- ductance amplifier (OTA), a Type LM13700 from National Semiconductor ‘A useful background on OTAs can be found in Ref. 3 ‘The VCO reference signal is passed through an amplitude control circuit and an integrator before it is added to the input signal alter pre-emphasis, The re- sult of the addition is phase-shifted by 90° and fed to a multiplier thatdetermines the frequency difference between its input signals. Evidently, the difference is 0 Hz only when the VCO frequency equals the interference frequency. As shown in the “Theoretical background’ inset, complete compensation is achieved when the am- plitude of the 94 kHz VCO signal equalsthe amplitude of the interference signal ‘The control loop performs automatic cor- rections to the amplitude of the VCO sig- nal until a stable condition is reached. ‘Once aligned, the decoder is, therefore, insensitive to small amplitude variations of the input signal, and other forms of instability ‘The integrator that shifts the phase of the VCO signal by 90° gives a frequency- independent phase shift, and is located after the amplitude control circuit. An ad- ditional benefit of the integrator is its ability to suppress harmonics generated by light distortion. These harmonics are suppressed at 6 dB per octave. “The control voltage for the amplitude controller is used to light a LED that indi- cates the decoding of a SAVE signal. When the LED lights, the loop filter is automatically switched to reduced band- width, and the output signal of the de-em- phasis network is fed to the video buffer With associated clamping circuit. The decoded video signal has an amplitude of about 1 Vee ata load impedance of 75 0, and is suitable for direct connection to a monitor or remodulator. From theory to practice ‘The previously discussed functions are easily found back inthe circuit diagram of Fig. 4. The unclamped video signal from the satellite-TV receiver must havean am- plitude of at least 0.5 Vpp. Components P1, Ci,Rrand ESi take the non-encoded video signal to amplifier Ts, From there, it is passed through clamping circuit Di-Ds-Ts before itisapplied to output buffer Ts. The output amplitude is set to 1 Vpp into 75 0 with the aid of preset Ps ‘The pre-emphasis L- filter in the col- lector line of Tr supplies a video signal with a 94 kHz component of about 3 Vpp when the receiver is tuned to a SAVE-en- coded TV channel. The transistor worksas an inverting amplifier. Capacitor Cr feeds the signal with the 94 kHz component to the phase comparator in ICs, where it is multiplied with the VCO signal applied via Cie Phase control ‘The 94 kHz Colpitts VCO built around FET sis tuned by means of a direct volt- age applied to dual variable capacitance diode Dv. The circuit around De and Ty is ‘an amplitude stabilizer. When the ampli tude of the oscillator signal excoeds a cer- tain level, Ts conducts via Ds, and draws current through Ru. This causes the drain ‘current of the oscillator FET to drop to a level at which the feedback gain in the 1. Pre. 1nd de-emphasis Current source Tt uses pre-emphasis impedance Zformad by Li-Le | Ce-Co-Rs to generate a signal voltage at TPI Theoretical background When pre-emphasis network Le {very high impedance, Ris alone then detarmines the impadance of 2. With Bra = 27000: 4 2 |Z) resonates at 94 kHz, it forms Ue= 1.28 Under 0.08 Uses | Uam=u | Ua =U Correctly aligned, the pre- and de-emphasis networks yield more than 20 dB suppression of the 94 kHz companent, without distorting the | lanys-tary=-U/ As video signal [Ure stein (2+ 2) =-U (ZHAI 2, Amplitude control circuit The frequency of the amplitude control voltage, Urey, is 0 Hz it the | eurtanindstequey musewbe tmpereye! Waltorieoeae | te Since Ts offers high current ampilication, Zis hardly loaded, ‘The regenerated 94 KHz sine-wava is added in antl-phase to the video signal. Emitter follower Ts forms a virtually Ideal vattaga source for Ae eomeheeirpeseroe. Ung [sit U= Ung) 8nto) 48 Usa = ar == 2+ ANIA ipsin@ ts 2 * * | Gage [5050 U-Ung)tt-cosi2ornae are | 294000 [rads] “Tniu aor gs vu complet spppresin ot he 180 component, so that: The decoded video signal across Ris may be written as Uni Us ue Fis/( i842) Unre, Urog" J05(U- Ung) at ‘This equation can be solved it plete compensation. rep, which corresponds to com: “CR 16+ 2) o [WEED sapsina trol 3. 80° phase shit The input voltage, U;, may be writin as Pin 5 of OTA1 supplies a sinusoidal current, fo, whichis integrated U= Unieeot upsinie t+ 4) ines | Owing to noise inte input sgna, te resicual 94 kHz component. | — cost) Urn has 19 be akon fe account iu vas=1/6f bat Us= R18 /( B15 +Z) (-Uvideol Z4+ AT) /FS + Uses) z Ue 16f, peoswtyat Us=-Unaag (227), As F5(2+ Ais) ZA Jeay= Pe sino ) Ueae= (Fe sinte ty With Fe = Rs were | a= 2194000 [ rads} Us=-R5/ Rs Usieot M8 /(Z+ R18) Us “The phase is shifted 90° independently of frequency. 4.28 Usseet 18/( Z+FS) Use FLEKTOR ELECTRONICS FEBRUARY 1990saveneconen EY] Fig. 4. Circuit diagram of the SAVE decoder. oscillators unity. The result isa clean and stable oscillator signal that is coupled out inductively via Ls. Transistors Tio and Ts amplify the oscillator signal and convert it into a symmetrical current that is fed to the pair of diodes at each of the parallel- connected OTA inputs (ICS). ‘The OTA inputs are also connected to Tiand Tito forma goir-cell. In principle, a gain-cell provides linear conversion of an input signal into an output signal. The {gain of the cell is determined by a current supplied by an external source. In the present case, this control current ema- nates from Rez, while the basic gain of the cell is defined by the ratio Re-Res. ‘The 94 KHz signal is taken asymmetri- cally from the collector of Tiz and fed to the phase comparator via Cis, The dif- ference frequency is available as a bal- anced signal between pins 1 and 2 of the XR2208. The balanced difference signal is passed to the opamp in the XR2208, by means of networks Ree-Ro-Car and Roe Ru-Coz, Depending on whether the PLL is ELEKTOR ELECTRONICS FEBRUARY 1990 locked or not, the bandwidth of the loop Filter is changed by FS2 connecting Rr in parallel with Ras, and E53 selecting a dif- ferent R-C network at the output of the ‘opamp in the XR2208. The control voltage is passed to the varicap via Rn and Ree This closes the phase-locked loop. Amplitude control ‘The amplitude of the regenerated 94 kHz sine-wave is determined with the aid of a current sent into pin 1 of OTA\. This cur- rent is supplied by OTA2, which works as 4 multiplier. The resull is that pin 5 of OTAr supplies a sinusoidal current to in- tegrator Cx. The mathematical deduction in the Theoretical background! inset dem onstrates that the phase is shifted 90° in- dependently of frequency. The voltage developed across Cav is fed back into the IC via pin 7, and reappears buffered at pin 8. The regenerated, amplitude-con- trolled and 90" phase-shifted 94 kHe sig- nal is fed to T via Cs, and from there to summation point T=Ri2. This forms the nucleus of the circuit: the interference dis- appears against its regenerated counter- part, which has the same amplitude but the opposite phase. ‘Thesummation signal is taken through the de-emphasis network in the emitter line of Ts. The first of the two L-C sections removes residual 94 kHz levels, and the second ensures sufficient suppression of the 188 kHz component. The decoded sig- nal is fed to the clamping circuit via ES! ‘The amplitude contro} circuit receives its error signal from the summation point, via Res. Components Cas, Ris, Res, Cov and Rw shift the phase of the error signal by 90° so that it can be used for driving multi- plier OTA2, This supplies a current that is integrated by Cx, The resulting voltage on this capacitor is buffered by the darling- ton transistor in the OTA, and serves to supply the control current to the pre- viously mentioned gain cell (OTA) via Rev. This closes the amplitude-contralled loop. The construction and alignment of the decoder will be deseribed in Part 2 of this article| 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. WAVEFORM MODULATION OF THE MAINS VOLTAGE by A.M. Karailiev In order to simplify the transmitter and to improve the noise immunity of the remote control receivers described in his earlier article*, Mr Karailiev here offers another method of modulating and demodulating the mains voltage. A thyristor, rated at not less than 25 A, is connected directly across the mains sup- ply. Itis controlled by a circuit as in Fig, 1 in a manner to make it conductive for a time Ar is at the end of every positive half cycle of the mains voltage, . The programming unit is an iniegrated circuit Plessey Type SL120, or SGS Type L120, or equivalent ‘Assume that the thyristor will short-cit~ cuit the mains at the instant this reaches a value of +24 V — see diagram in Fig, 2. Then, sin or eo Sin 200 T ) 24 = 310ssin2nxS0A, Since, for a small angle, sind = 6, 24 = 310x2mS0Ar, and AV (24108) / 3105250 = 246 ps, ‘Thus, for a period of 246 is, the mains voltage is practically zero and this mani- fests itself, at minimal power dissipation, in a distortion signal on the mains voltage. {is interesting to note that sinee the signal is caused by a lack of voltage, it will not be attenuated by the power lines, so that it ble over fairly large dis Type of modulation The effet ofthe momentary short circuit ofthe mains voltage will bea change ofits sinusoidal waveform, so that we may ‘Mains Signalling”, Elektor Electronies, November 1988, p. 27, Readers are advised that modulating the ‘mains voltage in the United Kingdom is subject to the provisions of British Stan- dard BS6899. Further information on the subject may be obtained from BIMSA (BEAMA interactive and Mains Systems ‘Association, Leicester House, 8 Leicester Street, LONDON WC2H 7BN, Telephone 01-437 0678. speik of waveform modulation. However, as the depth of modulation is defined by the angle @ during which the short circuit ‘ecuts, or by the time duration of the short circuit, other definitions of the type of ‘modulation may be applied Demodulation The information on the mains voltage is extracted from it by means of a circuit as shown in Fig. 6. The instantaneous volt- ‘age, w, at the secondary of the input trans- former, Tr, is N= Uy, Sinn VT 9, and this is represented graphically. in 3 If the full-wave rectified voltage (Usa 17 V) is applied to the inverting input of an operational amplifier acting as 8 zero crossing detector, and a direct volt- age of 0.25 V is applied to the non-invert ing input, the output of the detector will msist of 93.66 pis wide pulses at a pulse rate equal 10 twice the mains frequency, When the positive voltage applied to the non-inverting input is higher then the voltage caused by the modulation, the pulse spacing (pulse repetition period) de- creases. When, however, the voltage caused by the modulation is higher than that applied to the non-inverting input, the pulse spacing increases ~ see Fig. 4 In the absence of modulation, the zero crossing detector should be adjusted to make the width of the ouput pulses as Fig. 1. Basic control circuit for the modulator; the value of A dapends on the type of thyristor. IR ELECTRONICS FEBRUARY 199)ne 96 3 ee ey [NZ |e ft. 1 LL, J) bo OP TT of PLT 1 BY ai ef Fig, 5, Pulse diagrams t selected points in the circuit of Fig 6. The numbers Fig at the lett correspond with the circled numbers in Fig. 6. small as possible. For example, if the de- lector produces pulses 100 jis wide in the absence of modulation, and a modulated signal with a depth of modulqtion of 250 pis is applied, the width of the output pulses will increase to 300 us. Ifthe dete, {or is made more sophisticated, it may pro- duce pulses only when there is modulation present and no pulses at all in the absence of modulation, Receiver In the circuit of the receiver shown in Fig, 6, T! serves as the zero crossing de- tector. It should preferably be a germa- Fig. 6. Circuit diagram ofthe proposed receiver. ELEKTOR ELECTRONICS FEBRUARY 1990 In diagrams a and b thore is no modulation; in diagrams ¢ and d, ‘modulation is present, nium type to ensure minimum-width pulses in the absence of modulation, Network Rs-Ri-Cs forms a charge/dis- charge circuit that converts the pulse width modulation into pulse-amplitude modulation. The pulse height can be ad- Justed by Rs, so that monostable [C1 may be triggered only by pulses above a certain height. The monostable is retriggerable and produces 1 long output pulse at its pin 12 ‘when there is modulation present, Her the length of the pulses is predetermined at 500 is. Shorter pulses will be suppressed by a noise protection circuit consisting of Ro-Ri-Ce and the two Schmitt triggers ‘contained in IC. The output pulses of IC) are counted by ICs. When the pulse train ceases, re- tiggerable monostable IC2 emits a pulse that is applied to the clock input of D-type bistable ICs. This causes the contents of fone stage of the counter to be transferred to ICs, which acts as a latch, Since ICs is a four-stage counter, it is thus possible to control four different devices or machines. Diagrams of the pulses at selected parts, of the cireuit are shown in Fig. 5 The higher noise immunity of the re ceiver as compared with that in my earlier article is obtained because: I. its input is open to the signal and closed to noise, since the thyristor short- circuits all noise: 2. of the noise protection cireuit; 3. of the special form of the signal (ab- sence of voltage). .FEEDBACK KILLER 7. Giffard « This circuit helps to prevent feedback in public address (PA) systems. Feedback occurs at relatively high volume settings when a microphone is held too close to a loudspeaker. The effect is characterized by whistling or howling sounds that are annoying to the audience, the sound engineer and, of course, the speaker on the stage. Feedback isa constant source of worry to mixing desk oper- ators in the backstage line, since the now shrill then howling sounds always seem to occur unexpectedly, and tend to irritate both the per- formers on the stage and the audience, who respond with boos, catcalls and plugging of ears, No good this equipment! ‘Ata certain frequency, or frequencies, the phase shift in the closed loop formed by the microphone, the amplifier, the loudspeaker and the lis- tening room is 0 degrees, 360 dogrees or a multiple thereot. If, m this situation, the ampli- fier gain is greater than the room attenuation, the feed back signal rises to a level at Which if tops all other sound This is the well-known high-pitch whistle: the PA system oscillates, Figure 1 illustrates what happens, The sound produced by the loudspeakers is, reflected by the room or by objects or per- sons in the room, and is subsequently picked up by the microphone. The phase shift and attenuation depend on the acoustic parameters of the listening room (note that the audience forms a moving ' part of the room and must be taken into account as regards the loudspeaker posi- tions), Evidently, the best way to prevent feedback is to place the microphones in positions where they are unlikely to pick up reflections from the loudspeakers Sound technicians know a lot about microphone placement and offen spend hours on small adjustments of the monitor loudspeakers. However, many artists are inclined to take the microphone fro: stand the moment they are on stage. This requires great attention on part of the sound technician, since there remains a large risk of feedback occurring, "The circuit described here raises the sound level at which the feedback starts by afew decibels. This should notbe taken to mean that it makes feedback impossible: only the point at which the effect starts is shifted toa much higher level The risk of feedback occur ring is reduced by creating a small frequency difference of 1 Hz to 30 Hz between the am- plifier input-and output-signal. ‘This operation causes the fre quency of a (potential) feedback signal to beraised a little so that the condition for oscillation is no longer satisfied Block diagram A number of modulation tech- niques may Be applied t obtain 2 xe regency shit between an amplifier inpat-and output Signal In practice, virtually al these techniques make use of amples, ters and mixers "he block diagram ofthe feedback kil ter is shown in Fig. 2. The principle of operation is fairly simple. The input sg- ais paaed through two all-pase filters hose output signals have a phase di forence of 90°. These signals re mull plied with two clock signals that are also So" out of phase, The resulting signals are added. AS shown by equation (A) In Fig. 2, the multiplication and. addition operations produce a nev signal ofa fre Nn x G ere od ~ LH} >< ELEKTOR ELECTRONICS FEBRUARY 1990Fig. 3. Circuit diagram of the feedback killer. ‘quency equal to the sum of the frequency oof the input signal and the frequency of the carrier with which it is mixed, The upshot is that the input signal is shifted ‘over a small range in the frequency spec- tram. Although the equation shows that the sum signal isa single frequency com- ponent only, it should be noted that this is only valid if the input signal is shifted exactly 90°. Unfortunately, this is only possible in theory. Any practical cizcuit produces spurious signals that cause am- plitude-modulation (AM) of the mixer ‘output signal. The all-pass filter used in the feedback killer ensures a phase shift of vory nearly 90° over a frequency range of 50 He to 7 Fz ELEKTOR ELECTRONICS FEBRUARY 1990 The output signal of the mixer is taken through an active low-pass filter. This ‘operation is required because the input signal is multiplied by 90°-shifted rectan- gular waves, rather than a sine-wave and a cosine-wave which would have given rise to a considerably more complex cir- cuit. From Fourier analysis, a rectangular wave is composed of a fundamental fre- quency and an infinite number of odd- numbered sine-wave harmonics as -Ain(cos{ot)-1/3e08| et) +1/5c0s( St) ‘The amplitude of the harmonics decreases with frequency. The use of tw090"-shifted rectangular waves results in a series of sine-wave products and a series of cosine- wave products, The harmonics are removed in the previously mentioned low-pass filter, so that only the wanted mixer product remains There are two ways of shifting the input signal over a small range (I Hz~ 30 Hz jin the frequency spectrum: 1) usea low-frequency modulation signal or 2) mix the signal a second time with a signal of which the frequency is almost the same as that of the first rectangular wave. Next, remove the undesirable sidebands with the aid of a filter. ‘The second principle is adopted in the feedback killer. Two carriers of a relative- ly high, but slightly different, frequencyAUDIO AND HIF are used. The sum signal, A, obtained from the first mixer is filtered’ before it is mixed with a carrier, sin(a.+8)t, where 8 is the required shift in radians. The resulting signal, B, may be expressed as shown in Fig, 2. In the equation, -cos(o,-2u+6)t is an unwanted sideband, and cosie.6 the modulated input signal shifted by an amount Bin the frequency spectrum. Ifthe cartier {requency is relatively high lap- prox. 125 kHz}, suppressing the un- wanted sideband is relatively easy with the aid of a band filter, which doubles as a filter for the sidebands that are gener- ated by theuse of rectangular waves in the second clock generator. Circuit diagram Figure 3 shows the circuit diagram of the feedback killer. The essential blocks in the circuit are an alk-pass iter around IC>- IG, an output filter around Atsand Ave and toclock generatorsaround IC7-ICin ‘The audio signal is butfered by Ai before itis applied to the all-pass filter which introduces a phase shift of 9020.5" for sig- nals between 50 Hz and’7 Kitz, The filter, components have a tolerance of 17 to en Sure that the deviation from 90° remains as small as possible (remember that the ‘numberof unwanted sidelnds rises with the deviation from 9X). The frequency Tange of the allspass flier restricts the practical use of the circuit to speech. The feedback killer, must, therefore, not be used with singers or musicians, since the Absolute frequency shift upsets harmonic relations and so causes the performers to play or sing out of tune. “The all-pass iter is followed bya butf- wt, Arad Which nteodiicen a phase shi of 180° raqaired for the first multiplica tion with the rectangular signal supplied by the first clock generator’ The mullipli cation proper is effected by electronic Switches ES-ES: as illustrate in Fig. 4 The non-ineerted and the inverted sine- wave is passed during the positive and regative halfeycle of the lock signa re spectively. ‘Opamp Azadds theta mixed signals, and supplies, in principle, the compo: Tents of which the recangular wave con- Sisis. The frequencies of the components are, however, valued by the frequency of the inpst signal. Opamp Aw supprecs all unwanted higher harmonics generated by the use of rectangtlar-wave lock (ear- rier) signals. Ava Testilt, the sum signal contains only the mixing product of the input signal and the fundamental f= gutency of the clock signal. The ‘digital mixer” approach uses fewer components than an’ equivalent all-analogue.tsine wave/cosine-wave) circuit, and has the Additional benefits of lower cost and the absence of adjustments The filtered signal is available in in verted form after opamp AIS, and is mived a second time with a clock signal Like the first mixer, the second one uses slectronic switches. ‘The frequency dife Fig. 4. Principle of frequency multiplica- tion with the ald of electronic switches. ference between the two clock signals determines the effective frequency shitt of the input signal. The output signal of the second mixer is taken through a second- order high-pass filter, Avo, and a second= order low-pass filter, Ais, to remove all unwanted harmonics. The AF output is a virtual copy of the input signal: the only difference is thatit covers a slightly differ- ent range in the frequency spectrum The clock generators around ICs and IC> are identical with a Type CD4060 os- cillator/divider and an 8 MHz quarte crystal as the frequency determining ele- ment. The Q3-output of each CDS060 sup- plies the oscillator frequency’ divided by 16, Le, 500 KHz. The tivo bistables (FFi- FF2 and FFs-FPd) secure the required phase shift of 90° between the two clock signals. In this operation, the clock frequency is dlivicled by four, So that the carriers have a frequency of 125 kHz. The trimmer ca- pacitors in the crystal oscillators allow the frequency shift applied to the AF input signal to be adjusted to requirement. Components The capacitors in the all-pass filter are preferably 1%-tolerance polystyrene ‘sty roflex') or silver-mica types, although the board allows MKT (plastic encapsulated multi-layer polytherephtelate) types to be fitted also, These capacitors from Siemens generally have a tolerance of 5%, Note that when MKT capacitors are used, there is ittle pointin using 1% resistors. The use of 5¥-tolerance components in the all- pass filler increases the amplitude modu- Tation of the output signal ‘Two types of PTFE foil trimmer, 105 mm raster and 5x7.5 mm raster, may be used in positions Csr and Cx, Do not use ceramic types: they are not partly transparent as PTFE foil types and do not therefore allow the set capacitance to be deduced from the position of the rotor relative to the stator blades, Be sure to use quartz crystals and ‘CD4060's from one and the same manu- facturer and, if possible, from a single batch (look at the production date indica- tion, the type and batch number), The circuit is a combination of anal- ogue and digital electronics and has, therefore, relatively many decoupling ca- ppacitors to keep cross-talk and inter ference to a minimum. Fortunately, there is virtually no tolerance or quality re quirement for the decoupling capacitors, so that inexpensive types may be used with impunity Construction and adjustment The circuit is preferably constructed on the single-sided PCB shown in Fig. 5. The construction is entirely straightforward and best started with fitting the wire inks, Note that a fair number of resistors is ‘mounted upright, The adjustment of the circuit is simple, As already discussed, the frequency shift of the audio signal is determined by the frequency difference between the’ two crystal oscillators. This frequency dif ference is set with the respective trim- mers, Cx and C: The function of the cireuit is relatively ‘easy to check. Set both trimmers to about one-third of their travel. Connect a two channel oscilloscope to the input and the output of the feedback killer, Apply a sine-wave to the input and synchronize the top channel of the scope. The circuit worksif adjustment of oneof the two trim- mers causes the (output) signal on the lower channel to start moving horizon- tally across the scope screen, The frequeney shift required for opti mum suppression of feedback depends largely on the equipment used, the power ratings and anticipated sound levels, and, of course, the acoustic characteristics of the listening room, Finally, it will be clear that the effect of the circuit is less noticed al relatively small frequency shifts. ‘The circuit draws less than 50 mA from. a regulated 45 V to #8 V power supply. ELEKTOR ELECTRONICS FEBRUARY 1990FEEDBACK KILLER, 3 oe Fa1l2:3 eegooee S8o- 0 G- 8-0 |)} 15S =O Son10 eitov"2=|°| |" coootto bd Sette fozam0 zi Fig. §. Track layout and component mounting plan of the printed-circuit board for the feedback killer Part list Feo = 19%6 19% cu. Fis =931K0 1% Ft = 20k3 1% Fs = 1M5, FusPer Fite = 100k 1% Fu Rcon Fs sa = 47k Foss = 12 esse ~ 470k Capacitors: Pe =9k3t 1%: r= 6800 ‘Semiconductors: Ry = 819k 1% 2110 19% tes L074 Rio= 17k 1% G3= 407 1% Riv = 209K 1% Cs Ban 1% Fire 19k7 19% Ge= t00n 1% GEC» 4073 Fis» 750k 1% Ge 2n2 1% Few 22Kt 196 Gr= ion 1% Miscellaneous: Pie = 806K 1% C= 47n 1% XsiXe= qualia crystal 8 MHz Fea = 15k4 1% r= 3300 1% PCB Type 900002 (not avaiable ready- Pes = iM Ginga = B20 ‘made through the Readers Services) ec 18k2 1% GieCis= 10 Far = 22 Cua 118 ELEKTOR ELECTRONICS FEBRUARY 1990PC RADIO This FM radio, designed by ELV GmbH, consists of an insertion card for IBM PC-XTs, ATs and compatibles and is available as a kit or a ready-built and aligned unit. The radio has an on-board AF power amplifier for driving a loudspeaker or a headphone set, and is powered by the computer. A menu-driven program is supplied to control the radio settings. Take a look around many modern offices and you will notice PC users or terminal operators wearing headphones while sit ting in front of their screens. In most cases, a portable radio or cassette recorder is placed on the desk or near the PC to Tighten the work with a little music with- ‘out annoying colleagues, This PC radio with integrated power amplifier is constructed on an insertion ‘ard that can be plugged into any free bus extension slot on the motherboard of an IBM PC-XT, PC-AT or compatible com- puter running under MSDOS control. Control program and cursor keys The control software for the radio is sup- plied ona 5¥irinch MSDOS formatted dis- Kotte with 360 KByte capacity. After the computer is switched on, this disk is in serled into one of the disk drives. Use the DOS to select the relevant disk drive and type PCRADIO to load and run the control program, which is automatically cone figured for the graphics adapter card used in the computer. The cursor keys are used to select the menus areas shown on the introductory photograph. The adjust- ments that can be made include bass, treble, volume, manual or automatic tur ing and a signai-level- depencent muting, ‘The volume area is selecied as a default when the program is started. The left and right keys on the cursor keypad allow the volume to be reduced or increased respec- tively, while a coloured horizontal bar on the screen indicates the relative setting. A further volume indication is provided by a large two-digit number to the right of the bar, The volume setting can’ take values between 0 and 63 The up and down cursor keys select the menu areas. Pressing 1, for instance, moves the PC radio setting from the vole ‘ume to the treble area, which uses a col- ured bar and a two-digit indication like the volume setting, Pressing J again se- lects the bass setting. Here, again, the — and — keys may be used to change the setting ‘The next area reached by pressing 4 from the bass area is the tuning control Press the > and = keys to fune up and down respectively. The frequency step size in the VHF FM band is 10 kHz. The tuning rate is increased automatically to The PCRADIO menu screen (control pro gram version 1.2). about 2 MEiz per second if the < or > keys are depressed longer than necessary for a single keystroke. The fast tuning fea- ture allows the full VHE FM broadcast band to be covered in about 10 seconds Manual tuning isaccomplished by typ- ing the required frequency on the number Keys. Automatic station scan is selected on pressing the T or L key. In this mode, the ‘and — keys select between scanning up ‘or down from the current frequency. The scanner stops when a station is received If the band is empty (hardly likely these days) or if the stations are too weak, the radio automatically reduces its detection threshold and starts a new scan, The high sensitivity mode may also be entered by pressing the scan cursor keys twice within ‘one second at the start of the station scan, ‘The relative field strength of the re- ceived station Is indicated in the lower left-hand corner of the screen. The ‘mute’ function allows the level to be defined below which the AF amplifier of the radio is automatically muted. This function climinates noise between stations as the scan mode is in operation. Iti selected via the T and L keys, while the — and — keys allow the mute level — displayed in a different colour — to be adjusted to indi- idual requirement. Reception of ve weak signals is made possible by moving the mute bar to the extreme left of the sereen. ‘The menu area in the lower right cor- ner of the screen is reserved for the station name, provided this is among the ones listed to the left of the screen. Any station can be pre-programmed. When itis listed, it may be selected rapidly by pressing the associated function key Fl through F10, Stations are pre-programmed as fol- lows. First, tune to the desired station, then hold the siFT key depressed before pressing the function key to define the station number in the list. Next, type the station name on the keyboard and press ‘The PCRADIO control program is left ton pressing FSc (escape). Next, the user is asked whether or not the previously made changes to the settings and the station list are to be saved or cancelled. After going through this menu, the user is returned to the DOS prompt. The radio remains on, however, using the current seltings, al- though the control program is removed from the computer's memory. The com- puter is ready to lond and run other soft- ware at this stage. For advanced PC users ‘The following information is aimed at ad- vanced PC users, allowing them to use a number of special features of the PC radio, The PC radio kit is supplied with a diskette that, apart from the previously described user software, contains a do. cumentation file, READ.ME,and an installa: tion program, INSTALL-EXE. Hard copy of the (ASCII) README fileis obtained in the usual way by typing CONTROL? followed by TYPE READAIE and a carriage return. Turn off the printer again by typing CoN’ ‘The PCRADIO program may be made memory-resident by calling it up with a switch PCRADO /1 Tam OD ° treble D o oO D oO : : Fa Te te 5 5 i Sa amplitude of about 80 mVpp at pin 14. Tone control ‘The Type TDALS24A preamplifier/tone control chip receives its input signal (Cia. The second input of the chip, pin. decoupled by Cis Because the stereo mode is not used in this application. Direct volt ages applied to pins 1, 9and 10enable the toneand volume parameters to be set the aid of the internal electronic poten- tiometers. Capacitors Cis and Cis and re- sistor R07 are used for the bass com while treble control is effected with Resistor Res is optional, and selects tween preamplifier. If chip output pin 17 plies a current between 1.5 mA 10 mA, the preamplifier works linearly. If the current is smaller than about 0.3 the lodness/contour correction is actu ated, ‘The bass response of the TDAIS24A also depends on external components. may be omitted, and Cis or Cis may be replaced by a wire link to limit the frequency range. The chip receives its supply voltage via pin 3. The chip supply voltage is cleaned And buffered by Cs and Cir respectively AE power amplifier ‘The AF output signal at pin 11 of TDAIS24A is coupled out with Ce taken to input pin 1 of an integrated AF power amplifier, ICie, a Type TDA2030. ‘This chip provides a voltage amplification of about § as well as the necessary current amplification in its power output stage. Since the chip output is at about the supply voltage, a coupling capacitor, Css, is required fo connect’ the loud- speaker. Components Rss and Cs press any tendency of the AF amplifier to oscillate, year and contour operation of the Digital section ‘The description of the digital section of the PC radio card is best started at the address decoder. This consists of an &-bit comparator, ICs, for the base address, and twa L-ofd decoders in Ci for the ad- dresses within the selected block, Address lines A2 through A9 are applied to the comparator to achieve block decoding. The base address is set with jumpers Bri through Bri. Omitting jumpers Bri and Br, for instance, selects base address 30W0n, the default Setting of the card (eon- sult Table | for the address assignment) When the CPU in the PC addresses 1/0. location 300 (AS and AG are high), output P2Q of ICs goes low. Assuming that the 1/0 access is a write operation (Le,, pind oF IC2 is low), pin 15 of IC goes low also. Depencling on the bit combinationson ad- dress lines AO and AL, one of four outputs YO through Y3 changes to low. If, for in- stance, address 302: is selected, pin 11 of 7 ICs» goes low and provides a latch pulse for ICs, The 8-bit word on the databus is latched into D-type register IC, which holds the bass setting. The four 6-bit D-A converters are dis- crete types set up around four 8-bit latch- ing registers, ICs through IC, and associated R-2R ladder networks. The voltage at the output of each ladder is programmable between 170 mV and 3.2V in 64 steps. Cireuit ICs, an 8-bit latch, is used for driving the 12-bit D-A converter, IC, as well as for field strength measurement. ‘The 12-bit DAC receives an 8-bit word from ICs, and 2-bit words from IC> and ICs. The DAC reads these words from ad- dress nast-1. The reference voltage for the DAC is provided by voltage divider Rex Revat pin 19, Opamp ICiis supplies a volt- age between 0 V and —1 V in 4,096 10") steps. This voltage is amplified and in- verted by a further opamp, ICuis, Resistors ais with ro Cee ee sup- and mA, the and hale sup-RADIO AND TELEVISION seal, oy HE [aad Id Oa ca, yltloler el b Wo] ay Fig. 2. Circuit diagram of the RF and AF sections of the PC radio. Rt and Ros feed the tuning voltage to the varicap in the FM tuner. The gain of the second opamp is adjustable over a rela- tively wide range to allow the tuning range required to cover the FM band tobe set with preset Re The field strength voltage provided by the TDA7O2IT is’ applied to pin 10 of ‘opamp [Cite A field strength indication of 0. on ‘the screen corresponds to about 1.35 V, and one of 63 to about 245 mV. The ‘opamp raises the fiold strength voltage to level at which pin 12 of comparator [Cia ishele ata voltage of between 0 Vand 4V. The DAC built around ICs and the R-2R network allow the comparator threshold at pin 13 of [Cis to be changed. The out- ut state of the comparator can be re- quested at 1/0 address wase~1 via butler ICiay, Since the output voltage of the ‘opamp can take values between -5 V and +12 V, Ree, Roo, Ds and Ds are required to limit the voltage to0 Vand 5 V. The remaining three drivers in ICia are used to check the addressing of DACs ICs, IC and IC. The circuit is reset at power-on by a 100-ms long low pulse at pins 1 and 3 of ICs. This causes bistable ICs-ICss to be reset via gates IC and ICs, Transistor Ti short-circuits the voltage at pin 1 of the TDAIS244 to ground to reset the internal alectronic volume potentiometer. After 100 ms, pin 4 of ICs reverts to logic high, and it can be set, ie, pin 6 can be made low, by a read operation to address BAse-2 via [Cis. Depending on the position of jumper IPs, the radio may also be switched Off-set address read operation write operation 0 (base) D5-D0~ S-biLDAG for volume control 00 = comparator output for fielsirength ‘measurement D6 = not used Dt « control bit D7 =control bit 1 2 = control bi 2 Ba = contol bit 3 4 DS~D0=s-bit DAC for bass contiol__—tatoh data into 12-bit DAC D7: D6 = D8; Ds for 12-bit tuning OAC 07 =control bit 2 2 D5—D0=S:bitDAC for treble control turn on AF D7: DG = D11:010 of 12-bit tuning DAC D7 control bit 9 “ D5~D0=6-bit DAC for fieldsirength tun off AF D7-=D0 = D7~ Da for 12-bit tuning DAC Table 1. Input/output address assignment on the PC radio card. EKTOR ELECTRONICS FEBRUARY 1990to mute when the PC is reset, Another possibility to control the mute function exists ina read operation to address wastes (Cx pin 7). All supply voltages for the cireuit are taken from the PC expansion bus. Note, however, that the 5 V supply for the anal gue circuits is provided by voltage regu: lator IC ‘The complete circuit is built on a double- sided through-plated printed circuit board supplied with thekit. The sizeofthe Construction ‘ae seh a ggg ede Lal SePres a; oe ae ae ae ey al poogges gaa92a9 Fig. 3. ELEKTOR ELECTRONICS FEBRUARY 1990) Cireult diagram of the digital section and the PC-interface of the FM radio card.RADIO AND TELEVISION 42 Fig, 4. Cutting and drilling details of the ‘aluminium support bracket. oard is 230x108 mm. Start with fitting the low-profile com- ponents as per the indications on the com ponent overlay (Fig. 5) and the parts list. Proceed with the taller components. Since the board is through-plated, soldering at the component side is not required. Leave the wire terminals of Css and C=s longer than usual to enable these capacitors to be bent towards the board as shown in the photograph. The TDA2030 and its heat- sink are also mounted horizontally and secured on to the board with an M3 bolt ‘The TDA7021T FM receiver chip is an SMA Gurface-mount assembly) type Which requires great care and precision in handling as well as soldering. First place the IC on to the board. Next, position it accurately with the aid of a pair of twee- zers so that the terminals align with the solder areas. Use a low-power solder iron with a fine tip and a power rating of not more than 16 W to solder one of the corner pins of TDA7021T to the associated track. Next, use a sharp device to align the other IC pins with the tracks below. Heat the pins briefly in rapid succession while ap- plying very little solder Interconnect the points marked A and those marked B on the PCB with an insi lated wire of 0.4 mm’ cross-sectional area or larger. The position of the tin-plate screen around the FM tuner section on the PCB is shown in the top view of the completed printed-circuit board in Fig. 5. Bend the screen in the form of a rectangle and join its ends by soldering. Place the screen so made vertically on to the component side of the PCB, and align it with the PCB ‘edges. This is necessary for the PCB sup- port bracket which is sécured to the screen Tater, Soltier the screen to the PCB, ap- plying a generous amount of solder tin from the inside, but taking care not to damage components by overheating. Be sure to cover the entire length of the sereen, i,, do not solder it to the PCE in just a few locations, Next, bend and fit a similarly shaped, but lower, screen on to the component side of the PCB, opposite thescreen just mounted, Drilla hole in the taller screen for the aerial input socket (4 mm dia). Use an M3x6 mm bolt, a washer and nut to secure a small aluminium bracket to the lower side of the PCB, Next, prepare a standard PCB support bracket as shown Po = too Fee = 2200 Fea = 6800 Rea = 8200 Fas « 1k9 a = 18 Ree = 212 Reson sea oe = 4k7 Aa — RoReoPet-Res Aro re — Ran = 10k Air Rar iso ~ Re ss ~ Pur; Pse— Bs = 12k Fis = 15k Rr ~ Rie-Ran Ros — Ronis ir ReziRaafeo~ Ps = 24k Far © 3% Rrra = 47h es = 68k RoR Rex ftsesn Ps © AMT ee 50k proset V 100% Capacitors: C= $8 a= 33p Gan = 580 Gzr= 6p CsnGx: = t80p n= 2209p Ces 270, Coe 8209 Cae ta5 Cuaans Gu=an7 CarCsrCs1 = 10" Geen 151 | GseCesrCee CeCue = 220 coramic CeiGrCieon:Ce5:05: =47n Cages = 560 (CasCarGx2 = 100 ‘C10 = 200 C1= Cs 0o:Ciice0= 1y0; 16 CaS 22: 16 V Gre = CekiCao = 10u: 18 V Csi0 = 474; 16 V Car 1001.16 (CsniGes= 22004: 18 V (G5= 40p trimmer ‘Semiconductors: ICio= AD7545 IG = TDATO2IT (SMD) les = 7aLs00 ioe= 7aLs32 IG = 74.8125, IGs=7aLsi39 IC) = 7aLs285 ICs Ito 74.8374 Ien G19 = 7805, Tha acsa KO; = 85809 Deby = 1Na001 Ds= Daw 1NGTA8 Miseotianoou: But » loudspeaker sock SKI heat-cink ‘coax socket, Sway pin header. jumper. aluminium bracket. ‘un plate fr screening screw Ma.6, ‘screw M8, net M3, ‘etal cover, Solder pins. 90:mm insulated wire, min. 0.4 me £40 mm siver-patad wir, 93m coaxial cable, in Fig. 4. Secure thisbracket and the aerial input socket to the screen on the PCB. Solder the socket to screen, and secure the support bracket to it with two M3x6 mi bolts and nuts. Connect a short piece of insulated wire between the signal pin of the socket and point sti on the PCB, Note the position of Cs on the PCB and drill a hole in the metal cover to be mounted on to thesereen at the component side of the A complete kit of parts for the PC is available from the designers’ exclusive worldwide distbutors (re- grettably not in the USA and Canada): ELV France BP. 40 57480 Sierek-les-Bains FRANCE Telephone: +33 82857213 Fax 433 82838180 ‘Also see ELV France's advertisement ‘elsewhere in this issue, board. This hole serves to access Cs with a trimming tool. Solder the cavers to the two screens, Install the PC radio card into the com- puter, and screw the support bracket to the back panel. Do not forget to check that this panel is connected to ground of the PC — this is essential for the FM tuner on the PC radio card, Finally, connect an EM aerial via 75 Q coaxial cable. Adjustment Start the PC radio program from floppy- fo hard disk, Use the menu to set the fre- quency to that of a station in the lower range of the FM band, e.., to 88.5 MFlz. Align Cs with an insulated trimming tool ‘until this station is received (use a second radio as a reference). Next, set a receive frequency near the top of the band, e.g, 106 MHz, and adjust Rev until a pre. viously selected station is received. If the adjustment of Cs is correct, the frequency shown on the menu screen corresponds to the transmit frequency of the station, BLEKTOR ELECTRONICS FEBRUARY 1990rerapio PRY os Fig. 5. Top view of the completed PC radio card with the screening installed (Itt) and the component mounting plan (right. ELEKTOR ELECTRONICS FEURUARY 1990AC-DC CURRENT SENSORS Current may be measured without introducing losses in its path with the aid of the well-known series current transformers with clip-on probes or reed sensorsin which a contact is closed by the presence of a magnetic field. These devices suffer from a serious drawback, however: they can not measure direct currents. The sensors described here depend on the Hall effect and are suitable for measuring both alternating and direct currents. The Hall effect ‘The Hall effect is observed in thin strips of metal and in some semiconductors. When «a strip carrying current is placed in a mag netic ficld that is perpendicular 10 the strip’s plane, a voltage appears between ‘opposite edges of the strip that, although it is weak, will force a current through an external circuit The force due to the magnetic flux den- sity, B, on a charge carrier of charge e and drift velocity, v, is equal to evB. This force displaces the carriers and sets up a non- uniform charge distribution that gives rise to an electric field, £. At equilibrium, the force on the charge earriers, cE. due to this fleld just balances that due to the flux den- sity cE + evB =O, Fig. 1. Movement of electrons in:(a) an electric: field; (b) 2 homogeneous electro-magnetic field ia vacuum; and (ina degenerate sericon- ductor material The drift velocity is related to the cur- rent density, j, by jeney where 1 is the number of charge carriers per unit volume. Thus, the electric field is related to the vector product of the mag netic flux density and the current density by: ‘Ry(iB) where Ry is the Hall coefficient that is equal to I/e. The electric field results in a potential difference. Uy. the Hall volt- age, across the mat In some materials, the direction of the field is reversed, which means that these ‘materials have a positive Hall coefficient, This indicates that in these materials the current is carried by positively charged ‘carriers, that is, by holes, Electrons in an electric field move in suraight lines as shown in Fig, La, whereas in an homegeneous electro-magnetic field they follow eycloidal paths (Fig. 1b). In degenerate semiconductor materials, they can not follow such paths owing to fre- quent collisions with atoms of the semi- conductor material, Because of these col sions, the electrons lose velocity and this results in a waning of the magnetic field, since the strength of that field is directly proportional to the velocity of the charge cartiers. The electrons are consequently attracted more strongly by the electric field as shown ia Fig. le When a wafer oF semiconductor mate- rial is connected as shown in Fig. 2 and placed in a magnetic field, an e.m-f., the Hall voltage, Uy, is generated across the longer sides of the wafer. Because of this, the wafer is called a Hall generator. The level of the e.m.. is directly proportional to the strength of the magnetic field. If the direction of the field is reversed, the direc- tion of the voltage also changes by 180°, Within certain limits, the e,m.F is Uy = RyiB Ud where d is the thickness of the wafer. Note that the Hall voltage is inversely propor- tional 10 the thickness of the material snd = Fig. 2. The Hall voltage, U, is generated across the longitudinal sides ofa Hall generator Hall generators have come more into vogue with the arrival of semiconductor materials such as indium antimonide (InSb) ancl indium arsenide (InAs), which have a high electron mobility, resulting in a large Hall coeificient, and a low specific resistance. Table | gives a comparison of the electron mobility (average drift veloc- ity per unit electric field) v, in copper and ‘number of semiconductor materials. somes Om sx Fig, 3. With the Hall gonerator inthe ei gap of the iron yoke ofa current sensor, the Hal volt ‘age is directly proportional tothe curent flow= ing ina conductor that is at right anges to the cross-sectional plane of the yoke. ELEKTOR ELECTRONICS FEBRUARY 1999)AC-DC CURRENT SENSORS. Material Type Drit velocity Parameter (om vs) copper metal 27-43 Current rating silicon group 4 semiconductor 1950 Nominal outut germanium group S semiconductor 3900, Nominal oufut inks ‘AWB semiconductor 23.000 impedance InSb AW'BY semiconductor 65.000 Supply curent “able Eleton mobityn copper anda numberof semi Supply voltage conductor materials. Ambion temperature Measuring accuracy Linearity lie Dr wth empraue ! Max frequency i Response time >i. | ei Output signal ie Otset U__., vit inte Hysteresis rit wth temperature 4 Base circuit of a direct reading current sensor. Fe-elet dba asstontia men Dit with supply voltage Direct-eading Compensating 150-9000 A 100-250 A, 10 400 mA soko 502 15m BmA H8V HY o-80°c ~25t0470°C a <0.5% M% 0.1% 250 ppmi® C +550 ppm’? C 1OkHz 75-300 kH2 2804s us 20m simA 20-50 mY 0.05-0.4 mA <5 mV <02mA simvec sta pArc 0.10 mv <0.01-0.03 mAV Note that in spite of the high value of in indium antimonide this material is not used much be ‘cause of the fairly large change in ‘with temperature: over the tem= perature range 20-120" C, for ex ample, the value of v changes by a factor 5 Practical sensors Fig. 5. na compensating current sensor, the Ha In practice, the wafer thickness is generates a current in 2 secondary winding of the order of 1-5 um, which not only gives a larger Hall voltage, but also enables the water to be Fig, 6. The operating parameters may be determined from 3 four-quadrant diagram. ELEKTOR ELECTRONICS FEBRUARY 1990 ‘ablo2. Comparison of parameters of direct-reading and compensating sensors. used in very small air gaps necessitated by ‘weak magnetic fields. This may be seen in the schematic representation of a sensor in Fig. 3,in which the Hall voltage is directly proportional to the current flowing through a conductor that is perpendicular to the cross-sectional plane of the yoke of the sensor: ‘There are two types of Hall-effect eur- rent sensor: direct-reading and compensat- ing—see Fig. 4 and Fig. 5. ‘The main advantage of the direct-read- ing type lies in the minimal losses it causes, which is, of course, particularly important when large currents are mea- sured, The amplified output voltage, tf. is directly proportional to current i The compensating type is particularly useful at higher frequencies—see Table 2 ‘The current to be measured is coupled magnetically to a secondary winding on the yoke. The Hall element serves as a de- tector for the magnetic “O° which is in duced with the aid of an auxiliary (com pensating) current fed into the secondary winding. This compensating current is a very precise measure of the primary eur- rent. This type of sensor is eminently suit- able for measuring square-wave currents, Equivalent parameters of the compen- sating sensor may be determined from a four-quadrant diagram as in Fig. 6. If, for ‘example, the rms. value of the current to be measured is 100 A. its peak value is 150 A, and the maximum ambient temper- ature is 70° C. Drawing lines at right an- gles to the three corresponding axes gives 2 load resistance of 40-50 0, .VIDEO MIXER PART 2: MODULATION BOARD A. Rigby & G. Dam Last month's introductory article described the general setup of the video mixer, and discussed the operation and construction of the video switching board. This month we turn our attention to the modulation board which supplies two sets of modulation waveforms that together put many attractive picture mixing, wipe, fade-in and fade-out effects at your disposal. As already discussed in Part 1 of this ar- ticle, the video mixer consists of three blocks: a video switching board, a modu- lation board and a keyboard. These blocks form the basic configuration of the mixer as shown in Fig. 1 in last month's instal- ment. The modulation board and the key- board receive a number of synchronization signals and the supply voltage from the video switching board The Keyboard circuit supplies all the necessary control signals to the video switching board and the modjlation board. tore Effects waveforms All picture mixing effects provided by the video mixer, and all combinations thereof solected by the user, are based on three elementary waveforms: the triangle, the ramp (linearly rising sawtooth) and the parabola. All three are available in the horizontal (ine-synehronous) as well as the vertical (raster-synchronous) picture plane. All picture mixing and combination ef- fects rely on the switching between {vo too eventen video sources, A left-to-ight wipe (cur- DO ws | seucr k tain) effect, for instance, requires a circuit that switches between two video sources at accurately defined instants in the pic vance bey Laser LL varpose. Similarly, the VSW signal is used went the wipe effect is required vertically DO me — (op-te-bottom). Both HSW and VSW are ture line. The HSW signal is used for this iedongular signalewincealduty Extiooie IA h controlled to obtain the wipe effect. The von of required duty factor variation is secured 0 LA] ‘seusct ae with the aid of a comparator that com- > pares a ramp voltage to a reference volt- [Al age set with the wipe control (@ slide potentiometer on the front panel of the vermicat ideo mice. The other plctre mlxing QO Nae and effects are obtained by combining dif- 0 ferent patterns, horizontally as well as vertically. Block diagram ‘The waveform generators for the mixing Fig, 6, Blook diagram of the modulation board. effects are shown to the left in the block ELEKTOR FLECTRONICS FEBRUARY 1990RADIO AND TELEVISION diagram in Fig, 6, Since the same wave- forms are in principle required horizon- tally and vertically (only the frequency is different almont identical generators are used. In both cases, the waveforms are synchronized with the video signal. The parabola- and triangle wavelorms. are Slerived from the ramp voltage. Al three svaveformsare applied tocomparators via tlectronie switches, Depending on theit function, the comparators supply either a horizontal {HSW) or 2 vertical (SIV) switching signal AA separate modulation input allows additional conttol over the selected effect The modulation signal applied #0 this input may. be. synchronous or. mon synchronous in elation tothe picture, De- pending on the available video material tnd the applied wavevorm, remarkable snd sometines quite unpredictable cifets may be ablained in addition to the ones normally provided by the video mixer. The Rovizontal and vertical keying in puts, finally, enable “home-made” pat- Ferns othe aie! ta the "A bulfer and an inverter feed the se lected vertical waveform fo the horizontal comparator, which compares ito the waveforms’ generated in. synchronise with the horizontal ine pulses. The inver= tet may be switched on and off by a con- trol on the keyboard. Depending on the levels applied to the comparator. an HSW signal is supplied that switches between two video sources ata particular stant in the picture line. This combination of vertical and horizontal waveforms allows fairly complex mining effects to be rea. ined The vertical waveforms are available in non-inverted as wel a inverted form, Theselection s made by the user with the aid ofa keyboard contol and allows two Video ‘signals to be transposed on the screen. Taking the previously mentioned ‘erical wipe effect as an example, the n= ‘ort contol allows the top and the “bot tom pietures to change. places in the Circuit descripti ‘The circuit diagram of the modulation board is given in Fig. 7. The three-stage horizontal waveform generator is shown in the top left hand corner. The almost identical vertical waveform generator is found in the lower left-hand comer. The following description of the operation of these circuits refers to the horizontal waveform generator, ICas-ICs-ICae-ICos, Teand ICuo ‘The HSYNC signal is applied to the waveform generator to ensure that this is synchronized with each picture line, Dar- ing the HSYNC pulse, capacitor Cin is dis charged via Nero that integrator IC2s.can start from zero at the end of the pulse. The result of the integration is a line- synchronous ramp voltage at the output of ICs, To produce a triangle voltage, the ramp is frst converted into a rectangular wave by comparator ICz, which com- pares the instantaneous amplitude of the ramp to the reference voltage at the wiper of preset Ps. The rectangular signal 50 0b- tained isapplied tointegrator IC, which charges or discharges Cx as the output of ICxr goes high or low. The result of this second integration is a triangle voltage at the output of the opamp, ‘The horizontal synchronization pulst HSYNC, causes the integrator capacitor, Cr, tobe completely discharged via Nat at the start of each picture line, Preset Ps determines the switching level of the com= paralor and with it the symmetry of the friangle voltage, while preset Pr sets the amplitude. ‘Transistor Ta converts the triangle volt ‘age applied to its base via preset Ps into a parabolic voltage, which is subsequently amplified by [C2s. The amplitudeand the off-set voltage of the parabolic waveform are adjusted with presets Psand Pe respec tively ‘The vertical waveiormsare obtained in ‘a manner similar to the horizontal ones, However, the frequency is 50 Hz. instead of 15,625 Hz, and VSYNC is used to en- sure vertical synchronization to the mixed iceure. Pipemulliplexers ICay and IC together determine the waveform selec tion for the mixing effects. This selection is accomplished in conjunction with elec- tronic switches Nio-Nssand Ns-N=, De. pending on the logic levels on control lines MCI-MC2 (IC) and MC3MC4 (ICow),either the samp-, triangle-, parabo- la- or KBY- signal is used. The selected signal is applied to opamp ICs for com- paring tothe voltageat the inverting input Which takes either a steady voltage or one of the vertical waveforms, The signal at the non-inverting input of ICus is selected by electronic switches Nui ‘or Nas, To ensure that the effects selection remains in synchronism with the raster, these switehes are controlled incom plementary fashion by bistable IChw. The ‘other bistable in the 74HCT74 package, ICn,, functions as a monostable multi brator. It is locked with VSYNC and sup- plies short pulses at its Q output. These palses clock [Cow and time the instant it Iatches the logic level of cantrol line MC3 connected to its D (data) input. The latched level is subsequently transterred to the Qand Q outputs The direct voltage. supplied by the horizontal FADING contro), P's, may be ap- plied to the input of ICs, by switching on No. The toggling of the FISW signal sup- plied by inverter Nis determines the bor- der betiveen the mixed pictures — in other swords, the location (or instant) inthe plc ture line-at which the switching. between the two video sources takes place. Presets Pisand Pie serve to give potentiometer Ps its maximum usable range, ie, the width of the picture. Capaciter Ca allows an alternating voltage ta be superimposed on to the direct voltage at the wiper of Pis, This alternating voltage is taken from the external effects input and first buffered by ICy. If the external input is not used, the negative terminal of Cais taken to ground by switei Now. This is done to prevent noise being picked up, which would cause interference inthe mixed picture The modulation signal supplied by Cy is also applied to. comparator IC which adds it to the direct voltage sup: plied by vertical FADING control Pas, The selected vertical waveform is applied to the positive input of the comparator. The output signal is VSW or VSW which indi cates the vertical (raster-synchronous) switching instant Scries-connected opamps ICys-ICur- Cis form a precision inverter or bulfer for the vertical effects waveform selected by the user via Ns ~ Nes, Presets Pir, Pit and Par are adjusted to cancel off-set voltages ELEKTOR ELECTRONICS FEBRUARY 1990VIDEO MIXER. PART 2: MODULATION BOARD | 4 4 g 204 anae & 1.8 = Pk: = % 4 Fig. 7. Circuit diagram of the second unit in the video mixer, the modulation board, ELEKTOR ELECTRONICS FEBRUARY 1990qa ce L6 é See \@ o——0e ogsamo i[,> p Sibe oumiabateld eco eine Sine Sie =i ‘pl loegh hyytater epee iicey? MEG e040 colfo-ofaso wes. The fairly extensive compensation is required to the output of C36 supplies tI verse of the signal applied to the effects input. Control lin select between the inverted and non-in verted version of the vertical effects volt age. Construction ‘The single-sided printed-circuit board for the circuit is shown in Fig. 8. Construction should not present problems to the experi- enced electronics constructor at whom this project is aimed. Start the construction ofthe board with fitting all the wire links. Next, mount the solder pins, connectors and IC socket used. Check your work so far and proc: with the passive parts and the single diode. Be sare to of bol them straight on tothe sinks are not requized, Lastly, plugthe Ics in their sockets or, if sockets are not used solder them direct on to the board. The slide potentiometers are mounted on to the keyboard PCB to be discussed in Part 3 of this article. Set all presets on the mod Jation board to the centre of their travel To be continued next month Par 1 of this apticte appented the January 108) ELEKTOR ELECTRONICS FEBRUARY 1990Ray Pee Pas:Aae ros Pierro; rit = 10% ‘Ra = 6800 Ase Res:Roe Roe Rize = ho Foe» 47K ‘Ray = 82k | Ras = 470% Foehiesioai ie Ayisir Au Fiz « 220K Rire= 180k ‘Biet- 550k 57 = 10k preset H PsP ioP17 = 50k preset H PePigPisPiePaxPa = 100k preset H PoPis = 500k preset H PrajPs9~ 5k preset H PussPa0 = 10k linear Side potentiometer {not on PCB) ‘Piao -2k5 ereset H P= 50k presot H ‘Capacitors: Cai = 3900, CeziGes CeriGer Cee GreiGrasornCre, (Grain CasiCas ~ CniCar. Cee (Gar = CiusGroriCris~ Cive= 1000 CenGes= 270 GeeiGe0 = 159 Gre tn0 Gn =tep Crea 68n ‘GesiOeniOan;Oes = 105; 16 V racial Gas;Ca0= 100 GreiGice = 3300 Cior=279 Giron tr ‘Semiconductors: Be = Nats Iss = 74HCTOS Ges = 7805 Gxe4C20C2eiCar = TL08t IGzrcs: = CASIO NGaelCanicss « 7.082 (Cro = 74H0299 lem = 74H0T74 IG LMaTON IG3e— Ces » 4066 1G 7905, Te=B0s16 = BC8878 Miscellaneous: KMC1 = 20-way PCB header. PCB Type 8704-2 (cee Roaders Sorvioes pagel IEE MEETINGS 5 Feb — Evaluating student performance: key issues and strategies 6 Feb — Integrating control system design with the analysis of flexible structures 7 Feb — The role of venture capital in the UK electronics industry. 8 Feb — Broadcast TV recording. 8 Feb — Technical visit and discussion meeting at Colchester Technology Park. 12 Feb —The radio data system. 13 Feb — Safety critical software in vehi- cle and traffic control. 13 Feb — Data storage technology. 14 Feb — Stereo sound for television: im plementation, 15 Feb — Interactive computing: a revolu- tionary medium for teaching and de- sign. 20 Feb — Definition of traceability. 21 Feb — The interaction of radiowaves with the sea surface. 28 Feb — Safety regulations and codes for power and lighting in public areas. 28 Feb — Nucleonic instrumentation. Information on these, and many oth events may be obtained from the TEE, Savoy Place, LONDON WC2R OBL, ‘Telephone 01-240 1871, ‘The UKCMG conference entitled “Ne dimensions in application development” ELEKTOR ELECTRONICS FEBRUARY 1990 EVENTS organized by the UKCMG in association with Blenheim Online will be held on 7-8 February at the Metropole Hotel, Brighton ‘The conference is designed to assist in making sense of the critical developments in the IT industry today. Application de- velopers need to understand the issues in- herent in the move towards high produc~ tivity development environments, the de- sire for open systems and application portability and the continuing trend to- wards more flexible and adaptable database management systems. Further de- tails from Blenheim Online, Blenheim House, Ash Hill Drive, PINNER HAS 2AE, Telephone 01-868 4466. Frost & Sullivan have organized a number of seminars for this month on {nforma- tion Technology; Telecommunications & Data Communications; and Electronic Engineering. Details from Frost & Sulli- van, Sullivan House, 4 Grosvenor G: dens, LONDON SWIW ODH, Tele- phone 01-730 3438, CALL FOR PAPERS Papers are invited for the Second Intern: tional Conference on Rural Communi cations, which will be held at the Institue tion of Electrical Engineers, London, on 29-31 October this year, and for the Fourth Conference on Television Mea- surements, which will be held in Mon- treux, Switzerland, on 20-22 June 1991 Details from IEE, Savoy Place, LON- DON WC2R OBL, Telephone 01-240 1871. ‘A number of courses are conducted by Learning ‘Tree International between now and May in London, Stockholm, Paris, Los Angeles, Washington, Toronto and Ottawa. There are courses for “Man- agers in technical environments”, on “Networks, Datacom and Telecom”, ancl on “Software Development and Sys- tems”. For further information phone (0800 282 353 or 01-748 6667 in London (fax 01-748 5005) or 08-18 99 00 in Stockholm (fax 08-18 99 07). Saudicom 90, Riyadh's third electronics, communications and computer show takes place from 28 January to | February. All the giants of the industry such A T & T, Siemens, Ericson, British Telecom, Telet- tra, Plessey, France Telecom, Aerospatial, Alcatel, Telecom Australia International, and others, will be there, Further informa- tion from Overseas Exhibition Services Ltd, I Manchester Square, LONDON WIM SAB, Telephone 01-486 1951.THE DIGITAL MODEL TRAIN PART 11 - THE MONITOR UNIT by T. Wigmore The monitor unit enables the system to read signals from, for in- stance, the rails. This makes it possible to determine the position of any train at any one instant. Monitor units are indispensable where track sections are protected via a host computer or if the running schedule is controlled by the computer. ‘The monitor units have eight digital inputs ‘and 62 of them may be used, so that a total ff 496 inputs can currently be provided in the system, but this number may be dou- bled in the future, Note that the monitor units are not ex- changeable with decoder Type s88 from Gappinger. Those decoders have 16 inputs and are interconnected via a 6-way instead ofa 5-way connector, The circuit of the monitor unit is fairly simple as may be seen in Fig, 70. Itis basest fon an B-bit shift register Type 4014 (ICD) that reads the parallel data at the eight in- puts serially. The inputs are not connected irectto the parallel load pins of the regis- ter but via eight R-S bistables (IC2 and IC), This is done in this manner because the serial reading of the monitor units can only take place when the host computer has issued a status request instruction via the RS232 bus, The bistables in the monitor units also ‘ensure that even very short input signals to the system are acted upon, Resistors Ri-Rs form some protection for the inputs and also, in conjunction with C1-CS, suppress noise pulses Self diagnosis When requesting the status of the monitor tits, the eystet loads the position ofthe tight bstables in IC2 and ICs into the shift register IC, vin a pulse at pin lof Ki. The bistobles are reset automatically by the AAAREAEE Ef wate alal alabels "1 ye ons ae Fig. 7. Circuit aiagram ofthe monitor unt PTRONICS FEBRUARY 1990, trailing edge of this pulse. After the data have been loaded, the system reads them serially from the shift register. If more than fone monitor unit is used, the shift registers are cascaded so that they may be read as fone large shift register. ‘The number of data bits that the system has to read does, of course, depend on the number of monitor units in use, To prevent time being wasted in reading data of non- connected monitors units, the system car- ries out a self-diagnosis at power-up and reset, when the number of connected mon- itor units is determined automatically. All Distables are then reset and read immedi- ately. “The serial input of the last shift register is shorted to the positive supply line via Ris, which causes the system to read eight times ‘1’ immediately after the data of the last monitor unit have been received. [n this way, the system knows that the last unit has been read. ‘At each subsequent status request, the system takes account af the number of ac tually connected units. This self-diagnosis causes a small limitation: on power-up, it is not permissible for all eight inputs of a monitor unit to be active simultaneously, Decause that would indicate that the unit ‘was the last in the rove Vf the status of a non-connected unit is requested, the error LED lights. The num- Der of monitor units may also be deter- mined via a separate RS232 instruction {sce Table 8 in last month's instalment) Construction ‘The construction of the monitor units is simple and is best carried out on the PCB shown in Fig. 71 Before the board is fitted in the enclo- sure specified in the components list, two corners (indicated on the board) should be ccut away as shown in Fig. 73. The monitor units are interconnected via fairly inexpensive 5-way DIN connec- tors. If they are going to be used in fixed locations, Soldered connections, with or without the use of soldering pins, are, of course, perfectly all right Connexions to the inputs may also beConnecting up and actuating the inputs Monitor units are interconnected via Ki and K2, The arrow shown near these con nectors in Fig 70 mnt slveys point to the mother board. If this isnot observed, the system will not work, although nathing more disastrous wil happen, ‘All connector pins nus be connected to the same numbered pin on the connec: tors in other units, preferably by the use of rendy-male non-ivisted, Score DIN cable ‘The fot length of cable between the units Js of no importance to the correct operation of the system In principle, a monitor unit is actuated when the brown wize or ral ofthe system is connected to earth. Fortis purpose, two corshing points are provised on tne PCB. It also possible to actuate it by connecting am input to the red wire of the system, but not ifthe system is in the STOP mode Electronic train detection Signalling to and from locomotives and coaches is often performed with the aid of reed relays. Bath from a technical and an aesthetic point of view, this is not a good ‘way of doing it electronic means are invis- ible and seldom fail (which can not be sail of reed relays). If Marklin rails are used, M-sections are available in which one rail is isolated from the ‘metal gravel’ base. This isolated sec- tion is simply connected to one of the in- puts of a monitor unit Where K-rails (which are isolated from each other) are used, a section of one rail ‘may be isolated from that rail by two saw ‘cuts as shown in Fig. 72. The length of this section should be at least as long as the largest wheel base of locomotives and coaches used: this results in one long signal instead of a series of shorter ones. Cut the rail at an angle rather than straight across: in that way trains will have a smoother ride over the section. Where a standard rail system is used, the detection methods shown in Fig. 4 or Fig. 75 should be used. ‘The isolated section of rail in Fig. 74 is powered via two anti-paral- Jel-connected diodes. It dovs not mat- tor whether the diodes are connected to the brown or the red terminal Rails sections from which no signal is roquired may be connected direct to the booster unit ‘A disadvantage of the method in Fig. 74 is that only powered rolling stock can be detected. Ifa train loses fone or more coaches, these will be ‘in visible’ to the system. Therefore, the circuit in Fig, 75 may be preferred. coed er pecans . pisceuegod ‘The output of the comparator in that Circuit is actuated when a relatively high resistance exists between the two rails. By providing coaches with a 10k _resistor as shown in Fig, 76, they can also be detected by monitor units. Status requests The status of monitor units can be re- ted only via the RS232 interface. After Power-up or a reset, one or more groups of two bytes from two successive monitor “units (that is, groups of 16 inputs) are sent in response to each status request instruc- tion. This is done in this way to ensure full compatibility with the Marklin Digital Sys- tem. However, as explained in Part 10, if desined, the monitor units may be made to react individually, Fig. 71. Printed ciruit board fr the monitor unt. (COMPONENTS LIST Captor: oie ciate Semleonductors: icreatte Ioz tea sou Siseaoneous Krk Suay ON cman PCB rearing Ecosse OK Tipe aa71 1 aks of 2 rm pgs and sekes (etx) POR Type 8729-8 FARE ARATE A Fig. 72, Hw to connect an isolated section ofthe track toa monitor unit (lack arrow points to input ‘on monitor unit). Inthe ilustation, use is made of Makin rails. Fig. 73. The finished board fitted inthe specified ‘enclosure; note the cut-away corners. ELEKTOR ELECTRON FEBRUARY 19903 ina $ fit Fig. 74. This crcult for electronic tr detection can locate only powered raling stock (locomotives and coaches with lighting). Fig. 78. Providing non-powered rolling stock with @ 10k resistor enables It to be detected with the al ofthe clrcult in Fig 75. (2n—1) unit 2nd answer byte only with M (default at power-up and reset) 2n unit 'klin emulation Fig. 77. Composition of bytes sent by the system in response toa status request vi the S282 interface. Fig. 75 Circuit ofa sensitive detector system tha is actuated by a rela tively high resistance between the two rails. System miscellany Locomotive decoders (Part 2). Itappears that loc decosers can be affected by switching instructions intended for points and signals. This can be prevented by increasing the vale of Ri from 12 KO to 33 KO or even 39 KO, ‘Switching decoders (Part ). An ertor has occurred in the circuit dia- gram and components list. Circuits ICs and ICs should be Type ULN2003 and not ULN2001 as stated. The decoder will work ELEKTOR ELECTRONICS FEBRUARY 1990 with the ULN2001 but not entirely to spee- ification Using more than one booster (Part 6). If the size of the track makes it necessary for several booster units to be used (when ‘each boaster powers a section of the track), a small addition is necessary to provide adquate overload protection, The circuit for this is shown in Fig. 78. Ibis basically a ‘composite OR gate that detects overloads in up to five booster units. If any one booster unit becomes averloaded, the en- tire system is placed in the STOP mode. Fig. 78. Additonal circuit for obtaining proper overload protection when ‘more than one booster is used, Parts published so far. Part Part 2, Part 3 Part 4 Marklin systom (description), Loe decoder (description). Loc decoder (construction). Universal signal and switching decoder. Description of Elektor Electronics system. Booster unit Mother board (description), Part 8. Mother board (construction), Part. Keyboards, Part 10, RS232 Interface. Part 5. Part 6, Part7,8. Reflex MW AM receiver Radio techniques even at the most fun damental level will continue to fescinate electronics enthusiasts because a handful Of components ivork the miracle of wire lees reception. This month we save a once popular cicuit from oblivion. The radio ecination catches on with thiseflex AM receiver, sporting one tuned circll for medium-wave ceeption The scaled direct Conversion veiver forms the basis of all ado receivers Inits most rudimentary form, it provides an aerial input, a tuned crcat fa select the wanted RE signal from many others trans: triting in the same band, andl a detector to extoct the modulation signal tie, the information) from the earrer. In a more tlaborate design, an RF- and an AF- ame plifier may be used before and after the detector respectively. The dizect conver slon recelver differs from the super-he terodyne receiver in that it doet not use frequency conversion of the received RE gieliat any stage Althoph the direct Conversion receiver can not stand com- ;LEKTOR ELECTRONICS FEBRUARY 1990 _ INTERMEDIATE PROJECT sss of projects for the not-so-experienced constructor. Although each article __ will describe in detail the operation, use, construction and, where relevant, the _ underlying theory of the project, constructors will, none the less, require an mentary knowledge of electronic engineering. Each project in the series will be based on inexpensive and commonly available parts. J. Bareford parison with the selectivity of the superhet, it is none the less interesting to build as a first ever electronics project. Again making the comparison with the superheterodyne receiver, the direct con- version type is inexpensive, uncritical, and free af interference effects. What is mare, it provides quite acceptable sound quality. Flashback or start The circuit diagram of the simplest of re ceivers, the direct conversion type with- out RF or AF amplification, is shown in Fig. 1. The combination of an inductor in parallel with a variable capacitor forms a tuned circuit that resonates at the transmit frequency of the station to be received. One side if the tuned circuit is connected to ground, the other side to a small capa- itor that takes the RF signals from a long- wire aerial. The detector, a semiconductor diode, is connected to a tap on the induc tor. This is done to reduce the loading (damping) effect on the tuned circuit, which would degrade the selectivity, The RF component, i, the carrier, is shorted to ground by a capacitor at the anode of the diode. ‘The resulting AF component can be heard in the high-impedance ear- piece connected across the capacitor. The one advantage of the circuit in Fig. 1 is that it does not require a supply voltage. The disadvantages, however, are more important: sensitivity is poor, and the AF output signal is too small in many cases. ‘The addition ofa FET (field-effect tran- sistor) and a resistor as shown in Fig. 2 results in a better receiver, The FET pro- vides considerable AF amplification and at the same time, by virtue of its high input impedance, allows the detector to be connected to the ‘top’ of the tuned circuit. ‘This makes this type of receiver fairly sen- sitive, It is definitely worth while to build the circuit of Fig. 2. A few hints: use a wire aerial of at least 5 metres, and connect ground of the circuit to the water supply Fig. 1. The one advantage of the diode re- Ccelver Is that it dows not need a power sup- ply. The circuit is, however, useless without 4 large aerial and an earth connection. Fig. 2. Theamplifirin this modified diode receiver raises the sensitivity of the basic design considerably.Fig. 3. Extending the direct-conversion receiver with an amplifier for RF signals only (8a) and one that provides gain for RF as well 3 AF signals (3b). or the central heating system. The indue- toris made by winding 85 turns of 0.2 mm diameter (6WG36) enamelled copper wire fon a 10-cm long ferrite rod. The receiver draws about 4 mA from the 9 V battery. RF amplification It will be clear that an RF amplifier as drawn in Fig. 3 is required if the receiver in Fig. 2 is to work with a much smaller (shorter) aerial, and without the connec. Hon to earth via the water supply or the domestic CH system. Also note that the germanium diode has a threshold voltage Of about 200 mV, which makes reception of signals below this level impossible — weaker signals are simply not detected. IF the RF amplification is high enough, the tuned circuit may double as an aerial, ob- viating the long wire. A second tuned cir cuit, identical to the one already used, nay be addled to improve the selectivity. A further improvement that requires no extra components is shown in Fig. 3b. A, choke is connected to the output af the RF amplifier, and the detecior is not con- nected to the headphones, but fed back to the ‘cold’ side of the tuned circuit. The result of this configuration is that the am- plified RF signal is ‘forced’ towards the detector because it is blocked by the choke, After detection, itis applied to the input of the amplifier. Since the choke does not present a high reactance to AF frequencies, the amplifier raises the de- modulated signal to a level suitable for driving earphones. In other words: RF and AF amplification are provided by a single amplifier. This forms the principle of the reflex receiver. Enter two MOSFETs ‘Transistors are available today that im- prove the performance of the good old reflex receiver considerably. Figure 4 shows the circuit diagram of a receiver based on the reflex principle and de- veloped as a result of many experiments Dual-gate MOSFETs Type BF9S1 are used One functions as an AF amplifier that ob- viates passive components, and the other as an AF/RF amplifier with very high input impedance. ‘The heart of the receiver is formed by Ty. Resistors Rs and Rs provide the gate: bias voltage, which is decoupled by Ci Capacitor Cs similarly decouples. the drain voltage. The tuned circuit of the re- ceiver is formed by Li-C). The inductor is wound on a small ferrite rod which forms the aerial. The RF signal at the ‘top’ of the tuned circuit is fed to gate-1 of the MOS- FET via Ci, The extremely high gate im: pedance of the transistor eliminates all likelihood of damping, and obviates a tap connection or a coupled winding. All this simplifies the construction of the inductor and increases the sensitivity of the re ceiver because at the same (high) load re- sistance, the voltageat the top ofthe tuned Fig. 4. circuit at the input doubl Circuit diagram of the retlex receiver based on dual 88 compact aerial circuit is higher than at a tap towards the earthy side. The RF signal amplified by Ti is blocked by Le and consequently arrives at detector Di-D2-C+-Ro via coupling eapa- itor Ci, The detector also functions as a voltage doubler circuit. Resistor Ri feeds the rectified signal back to gate-1 of the MOSFET, which functionsasan AFampli- fier, The amplified AF signal passes through L2and is coupled out via C7 to the volume control, P Transistor T2raises the AF signal to a level suitable for driving lightweight headphones. Compact ‘The low number of components and the Use of a small ferrite rod enables the reflex riraseraen le MOSFETs. The L-Ctuned ‘medium-wave reception, ELEKTOR ELECTRONICS FEBRUARY 1990receiver and the 9 Vbattery tobe built into 4 miniature enclosure of the size of, say, two match boxes. If you consider size less important and at the same time want to go round con- structional problems, build the circuit on Universal Prototyping Board Size-1 as shown in Fig. 5. Pay attention te the orien- fation of the MOSFETs: when in doubt look at the pinning given in the circuit diagram (transistor viewed from the top. ‘Choke La isa ready-made type. The aerial is made from 70 turns of 0.2 mm diameter (3eSWG) enamelled copper wire wound on a 4em long, 10-mm diameter ferrite rod. The rod is secured on tothe PCB with the aid of plastic ties as shown in the photograph. The two diodes may be re- laced by almost any other smallesignal germanium type such as the OA9O or GA95. Do not use silicon types: thei thre- shold voltage is too high for this applica- sion. “The receiver is connected to a pair of headphones with a minimum impedance ff 64 Q. This perhaps less usual Value is simple fo achieve by connecting the two 3202 earpieces of the headphones in REFLEX AM MW RECEIVER EX) Fig. 6. Suggested construction of the re- celver on ready-made PCB Type UPBS-1 Parts ist Resistors: eRe = IMO Raa d7e Fu =4700 Fe 56 Fea tho Pr = 100k logarithmic potentiometer Capacitors: Gi 500p mica fll tuning capacitor + Inductors: t= wound on ferie rod» (28 fox) Lb AmH7 choke, eg, Tok 181LY-472 + Misoollaneous: ‘Sr = miniature ow ot switch, Headphones impedance 2 640 (2320) PCB Type UPAS-* (See Roaders Services age) + Suggested supolier: Cirle Park Lano, ‘BROXBOURNE EN10 7NO. Telephone (0992) 441306, Fax (0982) 464857, E Hard disk monitor December 1989, p. 54 ‘The parts list is notin agreement with the circuit diagram, and should be modified to read: Fi —RrsFisRis— AisFeo = 100% Cine = 220 Simple transmission line experiments September 1989, p. 38, ‘The value of Ra in the experiment should be 225 0, not 43 0 Logic analyser with Atari ST October 1988, p. 30 Please add to the parts list: Pa Res Rie = 100k Gs= 1000 ‘The control program (ESS111) is suitable for monochrome systems only. LF/HF signal tracer December 1989, p. 22 The resistor next to preset Ps on the printed-circuit board should be labelled ELEKTOR ELECTRONICS FEBRUARY 1990 CORRECTIONS Ru, not Roe. 1/0 extension for Archimedes November 1989, p. 14-15 ‘The parts list should be modified to read: Is= 6522 ‘The circuit diagram in Fig, 2 contains an error: Re should be connected between ground and the +input of IC. The printed- circuit board is all right and requires no modification, Intruder alarm November 1989, p. 45-47 ‘The references t@ Csin lines 6 and 7 of the last full paragraph on page 45 should read Gs, ‘The reference to Rs in the third para- graph under Alarm sensing should read Ro ‘The frst full sentence in the centre col umn on page 47 should read ‘Two assurance bleeps are generated: q one at the instant of switch-on, and oneas the exit door is opened for departure be- fore the end of the 15-s switch-on delay.” In the circuit diagram, Fig. 2, the diode across Bz: should be numbered Dw, not Diz. Capacitor Cs is missing: it isan 10 uF electrolytic type connected between the +12 V and ground rail. Finally, the inset text ‘D1...D8 = 1N4148" should read D1, 1D2,D3,D5,D6,D7,D8,D10,D12=1Na148. 314-digit SMD voltmeter November 1989, p. 37-41 The obsolete half-digit LED display Type HD1108 from Siemens may be replaced by ‘Telefunken’s Type D29xPK, where ‘x’ In- dicates the colour: 0 for red, 1 for amber, 2 for green and 3 for yellow. These dis” plays are also available under new type Indications as TDSR3120, TDSO3120, ‘TDSG3120 and TDSY3120 respectively. Telefunken also manufactures. equival- ents for the HD1105: the type numbers are TDSR3150, TDSO3150, TDSG3150 and Y3150. Distributor information on Telefunken components from AEG (UK) Ltd. + 217 Bath Road + SLOUGH SLI 4AW. Tele- phone: (0753) 872101, Fax: (0753) 872176,READERS' FORUM LETTERS Dear Sir — With reference to the Decem- ber 1989 issue of Elekror Elecironics, T would like to comment on two of the pro: ects. I. “Solid-state preamplifier”. This project is very good as far as it goes. The virtual earth CMOS switches have been used for some years on audio mixing consoles to great effect. [Suparbar [CMOS swatches An enhancement that I would person- ally like to see is the inclusion of active, symmetrical balanced inputs and outputs — Thave for some time been toying with a theoretical design based on a circuit ele- ‘ment called a ‘superbal Tean lay no claim to originating this cireuit (in fact, it came from Messrs Stan- coil, the manufacturers of Alice mixers, ‘who are no longer in business), but I have several circuits roughed out that incorpo- rate this element, [enclose the following sketch of the el- ‘ement together with the proposed CMOS switch eireuit 2. “AC tester for the IBM PC™. | have re- cently constructed this project with no problems at all. A few hardware enhance ‘ments were incorporated as follows; (i) the address selection links were re- placed by a four-pole DIP switch to facili- tate later changing of the base addres (ii) the ZIF socket and LED were built into a small plastic box along with another -pin D-connector to tidy up the test PCB: ii) the ribbon cable from the main board to the test PCB was discarded in favour of a standard 25-way D-conneetor computer cable, While the hardware cannot be faulted. the software supplied is not quite so bug fre. As far as I can tell, there are no prob= Jems with the interpreter, but the library files contain several errors. | have since hhad two further copies of the sofware: though these correct some omissions. ‘ther errors still exist Perhaps you might consider an occa- sional article on listings for test routines for ICs other than those contained in my list attached, AGG. Crane, Kings Lynn, Your sketches are reproduced here, since I feel that these will of interest to a number of other readers (they certainly were to me). Also, aif your comments have been forwarded 10 the designer of the “All- solid-state preamplifier”. Any comments Jrom him will be published in a furure Issue, ‘As regards the software buigs, your list hhas been forwarded to ELV, the de- signers of the IC tester. As soon as [This gives several 10s of dB headroom! we have their comments, we will publish your list and their re- [-o 9} ‘sponse, i any. [Ed) [-o |CMOS powered from #6 V; opamp powered from 420 V Dear Sir — I have just received the October 1989 issue of Elektor Electronics and am surprised to note that you are charging £2.00 for ‘paper copy of a PCB pattern to addresses outside Europe. This is rather unfair on readers like me. I like to keep records of PCBs for future use and have tried to do so from issue | of Blektor Electronics: 1988 was a bad year for PCB patterns, Many articles did not have a PCB pattern, only @ component layout diagram. If 1 have to pay out £2.00 for every PCB pat- tem, I may have to dish out £40-£60! Even copies of articles are £2.50 per anicle, not per page. Why can't a similar standard be used for PCBs? Most PCBs don't even fill half a page! Again, what you are doing is unfair and painful — please be more reasonable. K.G. Vergis, Likas, Malaysia, First, it appears as if you think that the basie prices charged to customers outside Europe are different from those to cus- tomers in the UK or Europe: this #s nov so ~ the differences in prices are caused by the different postage rates for the The processing of a PCB layout on paper costs about £1.20 (internal han- dling of order, placing order on central computer store. packing, and so an). Sur= Jace mail postage is the same the world ‘over and amionuts (on average) 10 £0.35: airmail charges outside Europe amount (om average) to £0.75. The processing of a PCB fayout on film costs about £5.00 (making of the film is around £4.00, the rest és as for the paper layout, although the packing is slightly ore expensive). Postage rates ave as for she paper layouts. We have no desire of making a profit on this service to our readers and other customers. The reason that in generat we do not publish the PCB track layout in the maga- sine is that it serves no purpose to the eonsiructor who is only interested in the component layout, 1 is impossible to pro- duce a PCB from the track layout put lished in the magazine (because of the printing on the reverse page) 1 do synipathize with your feetings, but unfortunately we all live in a harsh com merical world: we wish we could supply these track layouts free of charge, but it's Just not feasible without increasing the cost of the magazine. None the less, in some special cases, such as schools and colleges, I may be able so make special arrangements, {Ed} Dear Sir — Regarding “DC-DC power converter (November 1989, p. $4), don't you think it is irresponsible to publish a power supply circuit that is not short cuit protected. A car battery will easily ELUKTOR ELECTRONICS FEBRUARY 1990provide a hundred or more amps. quite suificient to destroy this circuit and most of anything that had the misfortune to be connected to it ‘The author claims that power rises ex- poneatially with current, When I went to school, power was directly proportional to the square of the current. I do not believe this has changed. Is skin effect really significant at the audio frequencies which the author sug- gests are used? At 20 kHz, the skin depth might of the order of 0.5 mm. If this is sig- nificant, the author should justify his claim. B.A. Jones, Cheltenham, Mr Jones has further pointed out that the point of his construetive criticism was that rany of those attempting to constrwet pro~ jects publisked in Elekior Electronics are heginners who weed everything explained in detail, a sentiment with which F agree wholeheartedly, and which will certainly he borne in mind én future The author comments: “The article states clearly that the converter is nos pro- tected against short-circuits and thar a 16.A fuse must be inserted beeween it and the bavtery. On reading Mr Jones’ com- ‘ments, it might perhaps have been, better to have shown this fuse fm the circuit dia- gram but, as @ rule, parts that are not on the PCB are not shown (10 maintain the proper relation between parts list, board ‘and circuit), Moreover, the provision of short-circuit protection would have increased the cost and complexity of the converter” ‘Iris true that a car battery can supply very high currents, but long before these Ihave risen to unacceptable levels, the 16.A fuse, and possibly resistors R2 and R3. ‘choke LI and diode D1, would have given, up the ghost. Moreover. ‘anything which had the misfortune 10 be connected to it {the converter) will not be damaged by a short-circuit at the ouput terminats of the converter, since its input voltage is then What you have learned at school about the relationship between current ‘and power and what 1 say about it are not in disagreement. I feet you may have cow fused the adjectival and noun uses of the ‘word ‘exponential’. Any parameter that varies in accordance with the power of a quamity does so ‘exponentially’, whether that power is 2,5 or ¢ (2.72). Only when the power is \ do we speak of direct (lin- ear) proportionatity”. (Note: | agree with ‘Mr Jones (in a follow up letter) that the ‘words ‘exponent’ and ‘exponentially’ are well defined in mathematics and should not be changed ‘at the drop of a hat’. Ed). ELEKTOR ELECTRONICS FEBRUARY 1990 “As explained in the article, the pulse repetition rate is a fimetion of the current demand of the load and has a maximum value of 30-50 kH>. The reason ihar the skin effect is significant is that the charge pulses have a fixed duration of 100 4s or 0, which means that they give rise to a large number of harmonies. These har: monics miuist be passed by the inductor wire without attenuation 10 mainsain the pulse rise and fall times. Fourier analysis is required to establish the actual skin depth, but this would be a superfluous tux uury in this kind of article, The recommen- dation 10 use a number of thin parallel connected wires rather than one thick one ensues from a series of practical experi i) ‘ments on my prototypes” SWITCHBOARD ‘Swltehiboard allows all PRIVATE READERS of Elokor Blecrunier one FREE advertisement of ‘wp ta 108 characters, including spaces, com Sera pera ot pr oath ‘Write the advertisement, which must re Tate tw electrons, 0 the eoupan an thi pager it MUS? INCLUDE a private tlephone sum ‘er or name and addres: post office boxes are rat acceptable BlcktorElestrnios Publishing can nota cept reaporaiility for any corrempondence oF Iraiaetion aa result ofa fee advertisement fof any inaccuracy in the txt of such aa ad Advertizoment wil be placed inthe order in which they are reoeina Blektor Blestrones (Publishing) reserve the right refuse advertisment witht ge ing enone o without returning tet, PEN FRIENDS. | am an olacronies engineer and ‘aminteested in pen ends on electrics, comput ts and other alle subjects. Mohammed Grantar, Damavand Ave, 15 Metr th, 8 Mei Gth, No. 10 Post code 17318 Tehran, Ire, WANTED. Sweap function generator upto 150 ki to include triangle to ramp wave culout. Phone (0267) 275497. FOR SALE. Now 12 inch black and tito vidoo ‘monitor with sound input: akes 1 V composte; £50 Phone (0704) 214594, FOR SALE, Tektronix 545 esciloscope plugin unt syle in good working order. £90. Phone (0732) 350691 any time. FOR SALE Eekor Electonics os 28 3, 45.4, une 81-Des 83 (oxcept Ai 82:31 copes £20, pi post. J. Brown, 74 Humber Ave, South Ock- fendon, Essex, FMS 5. FOR SALE. Compuiors IBM XT £490 ono up to -ATSB6 £1290 ono, Many accessories also avalable, Phone (08444) $853 for details. WANTED. Gieuit diagram for Telequgment Servi- scope, HL, Earnshaw, 88 Prices Lane, REIGATE RH? BAY, Telephone (0797) 221480. WANTED. Elokor Eloctonics uP Frequency meter, Capacitance metr and inductance mete. Phone. 01-594 7754 ater 8 pm. WANTED. Engish manval fr NEG PCE0Ot sans also exchange software for IBM PSI2 on diks, Hamidreza Madjdabadi, No 30 Western 194 St, TTehranpars, TEHRAN 16538, ran. FOR SALE. Components owing to giving up hobby. ‘Too numerous to ist 3 ng for deals, Ores in- site for he whole lot, Gordon Brown, 72 Bury Hil, WOODBRIDGE IF12 15D, Telephone (0394) 385652, FOR SALE. Tektronix $75 curve rer, Hatley soope 13 A, Creed 444 silent cover (ree 444), Bast ‘fers. 13A spare manuals £2.50 inc. R.A. Lord, ‘22 Elizabeth Crescent, EAST GRINSTEAD RHI JA, DESIGN SERVICE, Low-cost PCB CAD and manu: facture, single o” double-sided, to your design, Small or large quaities. Phone Oave at (0463) 790008. WANTED. Manual or cir dlagram for Tektronix scllosoope Type 817. Phone Howard at (0272) 658191 PEN FRIENDS. | am interested in pen ends rom allover West Germany, Mohammed Etminan, Shi- raz University, Schoo! of Engineering, SHIRAZ, Iran. ‘Send this coupon to: Blektor Electronies {Publishing}, Down House, Broomhill Road, LONDON SW18.4JQ. ‘Black capi paso one character o each box ELEKTOR ELECTRONICS FEBRUARY 1090, None a ates MUST be pen eeDARK-ROOM CLOCK A. Rigby Timing is an essential factor in any dark-room. Keeping a constant eye on the clock while the photograph is being developed, however, is a nuisance, and makes it impossible to attend to other activities. The low-cost timer described here can be programmed to give accurately timed audible signals as a reminder to development tank, add the fixative, remove PE paper, and so on. The quality of photographic reproduc- tions can be kept constant only if the de- velopment tanks are flipped, and baths are changed or stirred in good time, A timing error of a couple of seconds can be tolerated in some cases, but longer delays, caused by other activities in the dark- room, may result in an unusable print The present clock offers a number of selectable alarm functions and a maxi mum time lapse of 30 minutes, which is ample for most dark-room activities. The clock isa compact unit to prevent it taking up too much space in the dark-room. ‘The block schematic diagram in Fig. 1 shows the general structure of the dark- room clock. The 2 Hz clock generator af- fords simple selection of a number of differently timed alarm sounds. The 2 Hz clock signal is applied to a binary counter ‘whose (non-used) least-significant output bit counts half-seconds. The 11 other out- put lines of the counter are connected to the address inputs of an EPROM that con- verts the 11-bit counter value into control signals for a 4-digit multiplexed LED dis- play. This function requires 7 of the 8 nary Fig. 1 Block diagram of the dark-room clock. ELEKTOR ELECTRONICS FEBRUARY 1990GENERAL INTEREST EPROM data outputs. The remaining bit is used for the control of an alarm sounder, EPROM as a decoder ‘The EPROM used in the dark-room clock isessentially a programmablelogic device (PLD), whose combined functions allow the remainder of the circuit to be kept relatively simple. The 8 output lines of the EPROM are used for display segment con- trol and sounder control. Each address location in the EPROM is programmed 10 supply a bit pattern that determines whether a display segment is lit or not, and whether the sounder is actuated or not, Table 1 lists the bit patterns for num- bers 0 through 9. The sounder is actuated by adding 80x (bit 7 goes high) to the values shown. Table 2 shows how these bit patterns are stored in the EPROM, and illustrates the functions of its address in- puts, which, given the application, may also be considered decoder inputs. The most-significant (MS) address imput, A13, divides the EPROM address space into two blocks. One of these en- sures that the sounder is actuated every 30 seconds, which is useful during the de- ‘velopment of films, The other block is pro- Table time-lapse buzzer. The databits loaded into the EPROM control the lay segments and the Fig. 2. Circuit diagram of the dar BLBKTOR ELECTRONICS FEBRUARY 1990 yoom clock. The heart of the eircul formed by EPROM ICs.buzzer lapsed time igtt Arc Gan Seconds unis) cou | oF f Seeonds ons) | 0001 | Gminutos, Oseconds | ecrsedien, | goal | BF rites (tes) coos | 80 et | Ominutes, + cocond | Seeonds {tens} | 0008 | ar mires (tens) 20 | | | Seconds ui) | J05 | ‘seconds tens) 1089 V7 minites, 80 secon’) nutes unis) | 1084 | Br fecee ee rinvtes tne) 1088 | 6 / fcondsunisy | tcig | Ze seconds (tens) iis | 6 Peminaes, $8 seconds| inates (unis) | AG1A | BF mines ens) | $18 | 88 secontetnis) “| gic | er | . Seconds tens ii | 6D | Pomminaes, BO seconde) nutes ants) | iciE | 6F rinutos (lens) | | 248 bytes, not pro- orammed i seconds (unite) 2000| BF seconds(iens) | 2001 | BE | Oris Sconce | inaeume) | ama | aE minutos (tens) 2008 | | “seconds (unite) zee | ar epnraee ‘seconds (tens) F {minute $0'seconds | minutos (unis) 86 |. minutes Gens) 80 after $0 s ; feconds unis) gore ar seconds tens) | aGID | 29 mites, 8 seconds| nutes unis) | Scie | SF | | nates tons} acre | 88 | | sco. | FF | i 248 bytes, not pro: | | oremied | | | 2 I [sere | Fe Lae ano Table 1, EPROM organization. DARK-ROOM CLOCK grammed to supply a single alarm after ‘90 seconds, which is useful for the devel- ‘opment of PE (poly-etiylene) paper. The development time of PE paper is normally stated as about 60 seconds, but 90 seconds is often found to give better darkening and quality of reproduction, ‘The time in seconds is applied as an 11-bit binary number to EPROM inputs A2 through A12. This arrangement results in four bytes that hold the segment pat- tems for the LED displays to be selected every second. Address lines AO and AT are driven by two signals that control the multiplexing of the 4 displays. Summariz- ing the above, the EPROM functions as a circuit that decodes binary time informa- tion into.a multiplexed 7-segment control signal The circuit ‘The circuit diagram of Fig. 2 shows that the clock generator is formed by ICs, a ‘Type 4060 14-bit counter with on-board oscillator. The oscillator signal is divided by 2", so that output Q13 supplies a 2 Hz signal. Outputs Q6 and Q7 supply the multiplex signals that cause each display to be turned on and off at a rate of 128 times per second. ‘The? Hz signal is fed to the clock input ‘of binary counter IC), This 12-bit counter/divider counts the seconds lapsed since the circuit was started. The use of binary counting in combination with an EPROM allows such a simple sec- fonds counter to be used — the perhaps more conventional alternative would have required at least 8 standard ICs, or two or three specialized ones, The EPROM used in this circuit is available ready-programmed. The reset circuit of IC) has 3 inputs with an OR function provided by Di-D:- Da, The counter is cleared (reset to zero) elther when RESET key $2 is actuated (Di), for when the circuit is switched to the stand-by mode by $1 (Di), or when 1800 seconds have lapsed since the last reset (D3). The last function is realized with the aid of Di-Dr that reset the counter within a few nanoseconds after it reaches state 1800 (111000010003) ‘The lapsed time is passed to EPROM ICs, whose function has been detailed ear- lier. The decoded and multiplexed 7-seg- ment signals on datalines D0-Dé are buffered by Darlington array ICs before being applied to the cathodes of the LED displays ‘The signal on EPROM dataline D7 con- trols sounder Bz: via driver Ts. The sound of the piezoelectric buzzer is made a little less disagreeable by interrupting it with the aid of Ts. A wire link, JP, is fitted to provide 3 sounder options: * position ‘C'for two pips ofa quarter ofa second each; ‘+ position ‘D’ for a single pip of half a second; ‘+ no wire link for a single pip of one sec- ond. ELEKTOR FL CTRONICS FEBRUARY 1990(GENERAL INTEREST r 1 ann e e908 ) Jey “40 le Fig. 3. Printed-circuit board for the dark- room clock. The board is double-sided, but not through-plated. For reasons of safety, and to keep its over- all cost down, the dark-room clock is powered by a mains adapter with a direct, unregulated output voltage of 8-12 V. Regulator ICs reduces this voltage to 5 V. A taal on/off switch is not provided. Rather, the cireuit is switched to stand-by with S), The counter is reset to zero, the displays and LEDs are turned off, and the EPROM is switched to its low-power standby mode. The clock is not re-actu- ated until Sis switched on again. Construction The dark-room clock is built on the double-sided, not through-plated, printed-circuit board shown in Fig. 3. This board is available ready-made. Figure 4 shows the locations of the holes that must be through-contacted by FLEKTOR ELECTRONICS FEBRUARY 1990 Parts tist ‘Roclstors: ‘Ry = 100k Rex 10k a= 10M Ry —Rir Fre? = 10 Fe—Ris= 470. Rusie= 4700 (Capseitore: GirGa= 220p Ga=Ca= 1000 Gr = 100u:25V Semiconductors: DinDr=tNdt48 Da:Oe » LED: ree: 9mm LD: = LDéx 7761 (common anode) Ti Tea 805578 ‘TaTe= 805478 Icy = 4040 IGe = 4060 Is» TaHoT199 IGs= EPROM 27128, Avaliable reacy-pro- ‘graerned undor order numbor ESS 588 {600 Readers Services page) Gs = ULN2004 iGo = 7805 Miscetianeous: : ‘$1 =locking SPOT switch wi black key ‘cap. ITW Type 61-2010400". $2 =SPDT switch wih red kay cap. TW ‘Type 61-1010000", $3:S4 = locking SPDT suitch with integral LED and black Key cap, TW Type 61- -osadot*. 1 = quart erystal 32.766 Kitz, Bite) oo Heme Gree page). : Enclosure: Heddle Type 222° = TTW Switches + Division of FTW Limited “Emiok Electronic Products Limited « Unit 18a «Industrial Estate « Stanton BUAY ST. | :| EDMUNDS IP31 24R, Telectione: (0389) 511155, : inserting a short piece of wire, and solde- ring itat both the component side and the track side of the board. This arrangement obviates having to solder the ICs direct on to the board, although this is perfectly permissible in this case. Note that the ‘white overlay print must be removed at the component side before the spots can be soldered properly The 3 exclamation marks shown in Fig. 4 indicate PCB design errors that are, fortunately, simple to correct. The holes near pins 22 and 28, towards the ‘ICs’ markon the overlay, must be connected to the associated holes that receive the IC pins (or IC socket pins) at both sides of the PCB. In the case of pin 22, the missing connection is at the track (EPS) side of the board: a litte extra solder tin applied dur- ing the mounting of ICs will bridge the gap between the spots. The extra connec- ° 0°. OO ° ooo9 ek a 00 g'poad 8 90° e 6 © 9 0000 a) PoP 000 ° °° ° ° 0 2| 0 3 ° ° ° Fig. 4. Showing the locations of the holes that must be through-plated before any com- ponents can be fitted. tion for pin 28 is made at the component side of the board with the aid of a piece of thin wire that connects the holes, Before fitting this wire, remove the white overlay printed on the hole for IC pin28. Insert the tip of a sharp pencil in the hole to ensure that it remains open when the wire is sol- dered to the spot Make the connection between Rss and La with a short length of insulated wire fitted at the track (EPS) side of the board. A trackis provided at the componentside, but this makes fitting LDa rather difficult if the centre pin in the right-hand row of, the display is to function as a through contact. The fitting of the remainder of the com- ponents should not present problems. Ifthe HeddicType22? enclosurestated In the Parts List is used, the switches must be raised by mounting them on IC sockets(cut off the non-used pins of 14-way low- profile types). Also note that the LEDs in the key caps are best removed and re- placed by types with long terminals that jeer fan be soldered direc tothe PCB, Use a - drillora smal fils tomodiy the IC socket ae grt if part of itis in the way of the LED terminals SER wenvoicaren wo The switches are ‘Digtast like SPDT oo, (eingle-pole, double-throw) models wit Spe srtey ape srestagane deere oe teehee, an integral LED. With the exception of Se, HEY oes ae ganar they remain locked in both positions. I wa tat aE ste a Se, difficult to obtain locally, the switches may, of course, be replaced by suitable equivalents with the same electrical func- pean ee eee tion, mounted on to the front panel of the FEVTNSo 22" ao atoptayencoouratiydsepiaytit+48o; |aetude oeund bie) Tal oguse, anil ossnaeted bs ecitett (ateuiate a readout in minutos and, seconde) short wires, ee catece "the aizpteyceone) a es Se rrtte sscunai")7 NSEE [g)Sisplayercscuntt)1 nauanalguiiaarsero Stankiog SESS glaeasteandanbeahy Sale lcntenate she ous atepletyeen) hve losttonanttn SHS eegeiggp eae seas on ie auapiayisioatepiere) shee wsee Gialteteytfh Hn Abe cee (esa the sania pase wien 00) ‘(Maximum count 12 1868 seconde) (Gnec £¢ count ig a multipte of 90 and arailer than $1) alse weite (g,displayandsound (ni); s . 4€ count < 100° thon rite Toyaerodl {Lendang™tero blanking cree c gomtt aegis, Wucuaplarenascendisiies =e cise write leyalapiayandsoung{ijts oh leancatas ‘ite. (ovateptay 11) AE count < 100"then write ‘tgines0) | Leading-sara blanking! Side Seite Lrdlepiavt jit # count < 1000 then wait i) Hinadtog-zero blanking! ale Grits qualepayiiit Sean Fig. 5. Listing of the Turbo-Pascal program used to compile the content of the EPROM. The ‘count statements allow you to change the timing intervals to Individual requirement. Like the EPROM content, the program Is divided in two parts: the fist provides the data for the ‘minutes and seconds read-out, and the second the data for the seconds read-out (0-180 8). ‘These blocks may be interchanged. EPROM data Is written to fle cLock par. ELEKTOR ELECTRONICS