with this issue 132 -page greenweld catalogue class a vilifieramplifier remains relatively cool...

THE INTERNATIONAL ELECTRONICS MAGAZINENOVEMBER 1991 UK £1.90

FREE with this issue132 -page Greenweld

catalogue

Class A VilifierQuadriform ferrite antenna

Dissipation limiter

24 -bit full -colourvideo digitizer

- Relay card foruniversal bus

Productmodel

9 7 261 111036

11

Please mention ELEKTOR ELECTRONICS when contactin,4 advertisers

ULIF71\7

Five successful yearsTwenty five man years development

Over three thousand five hundred users

To celebrate the continued success of ULTIboard we have a very special:

JUBILEE OFFERILILTU'OrLD), tiL13677COMPUTER AIDED PCB DESIGN SCHEMATIC CAPTURE excluding V.A.TV.A.T

u This is not a stripped down or old version of our software. Updates with new features added,are released every six months.

Ej This is not a low cost package without a growth path. Upgrades are available to our highercapacity, 32 bit, Advanced and Professional systems.U This version includes our high -quality interactive autorouter (not an external batch router) whichallows you to selectively route components/nets/windows as well as the whole board.I ULTImate's Jubilee offer is the only true low cost system available today which gives you leadingedge software with a future.Ask for your demo disk and see for yourself how easy it is to take the ULTImate route.

'

1)1

'

a I 1.

Computer Aided PCB Design

Real Time (Direct) Reconnect, Forcevectors &Histograms guarantee optimal placement.Real Time Design Rule Check prevents illegalconnections and/or clearances during traceeditingTrace Shove & Reroute -while- Move reduceediting time.

Automatic power & ground plane generation.

ULTI©APSchematic Capture

Real Time Electrical Check prevents errors during

Aeduitotingwiring easily makes connections Wire Move and Shove reduce editing time

Perfect Intergration with ULTIboard, even tracewidths may be entered during schematic captureNetlist-toolkit allows interfacing to any CAD/CAEsystemUnlimited attributes on parts, pins & nets

17V7IInternational headquarters I MIEnergiestraat 36 1411 AT Naarden

Tel. 02159-.44424 Fax. 02159-43345 TECHNOLOGY

UK Sales & Support Office 2 Bacchus HouseCalleva Park, Aldermaston Berks. RG7 4QWTel.: 0734-812030 Fax: 0734-815323

November 1991TOR C "O)NTITTFAIT7S Volume 17

Elf R-1 C Number 194

Front coverFor many years now, hi-fiequipment has almostexclusively used class ABoperation. This methodgives acceptabledistortion and goodefficiency, and the outputamplifier remainsrelatively cool duringnormal operation.However, manydesigners and listenersprefer class A operationwith its much lowerdistortion and consequenthigher quality of soundreproduction.To meet their demands.we have revamped theLFA 150 power amplifierwe published about threeyears ago. In the newdesign, a goodcompromise has beenachieved between class Aquality and the high heatdissipation thatcharacterizes this type ofoperation.

In next month's issue

More than 60 constructionprojects and informative.

educative articlesincluding

Connect 4FSK/RTTY decoderfor PCsFunction generator [3]Programmablefilter ICsClass A amplifier [2]Temperature measure-ment techniquesEconomy powersupply

Copyright 1991 Elektuur BV

ii' A i -r.v ig.11iiimfiiir

3

11 Michael Faraday: the father of electricity (179i -i867)

AUDIO & HI -Fl

37 PROJECT: Class A power amplifier - Part 1 (of 2)by T. Giffard

COMPUTERS & MICROPROCESSORS

45 PROJECT: A simple watchdog circuitby Akbar Afsoos

50 Upgrade for MCS BASIC -52 V1.1 - Part 2 (of 2)by Dusan Mudric and Zoran Stojsavljevic

54 PROJECT: Relay card for universal I/O interfaceby A. Rigby

GENERAL INTEREST---

31 Product modellingby R.G. Evans

42 Four -terminal networks - Part 2 (of 2)by Steve Knight

46 PROJECT: The cipher machineby Owen and Audrey Bishop

61 Computer -aided electronics designby our technical staff

! LNTERMEDIATE PROJECT

24 -bit full -colour videodigitizer - p. 16

- -

--.,

,.) -,

Dissipation limiter - p. 19

52 Timer for central -heating systemsby J. Ruffell

POWERPOWER SUPPLIES & BATTERY CHARG ' .

-

\

,

19 PROJECT: Dissipation limiterby C. Zschocke

! RADIO. TELEVISION & COMAIUNICATION8

14 PROJECT: J17 equalizing network for satelliteTV receiversby J. M. Whisor

16 PROJECT: 24 -bit full colour video digitizerby J. Kortink

57 PROJECT: Experimental quadriform ferritetransmit/receive antennaby Richard Q. Marris

60 Z -match II -a reviewby Mike Wooding

The cipher machine - p. 46

. i.- ' v

.1. -r -

-----

1

1

TEST & MEASUREMENT

22 PROJECT: Digital function generator -Part 2 (of 3)by T. Giffard

MISCELLANEOUS INFORNLATION

[ ABC Electronics scene 12; Events 13; Readers services 67;Switchboard 68; Terms of business 68; Index of advertisers 74. Relay card for universal

I/O interface - p. 54I 7 ^

., ,, -=

ELEKTOR ELECTRONICS NOVEMBER 1991

iii Please mention ELEKTOR ELECTRONICS when contacting advertisers

Whether your requirement for surveillance equipmentis amateur. profess:onal or you are just fascinated bythis unique area of electronics SUMA DESIGNS hasa kit to fit the bill. We have been designing electronicsurveillance equipment for over 12 years and you canbe sure that all of our kits are very well tried. testedand proven and come complete v.ith full instructions.circuit diagrams. assembly details and all high qualitycomponents including fibreglass PCB Unless other-r.,se stated all transmitters are tuneable and cart bereceived on an ordinary VHF FM radio

UTX Ultra miniature room transmitter. Smallest room transmitter it inthe world! Incredible lOrnm x 20mm including mic. 3-12V operation, 500mrange

MTX Micro -miniature room transmitter. Best selling micro -miniatureroom transmitter. Just 17mm x 17mm including ink. 3.12V operation, 1000mrangeSTX High-performance room transmitter. High performance transmit-ter with a buffered output stage for greater stability and range. Measures22mm a 22mm including mic. 6-12V operation. 1500m rangeVT500 High -power room transmitter. Powerful 250rnW output providingexcellent range and performance. Size 20mm x 40rnm. 9-12V operation_ Range3000m £16.45VXT Voice activated room transmitter. Triggers only when sounds aredetected. Very low standby current variable sensitivity and delay with I.e dindicator. Size 20mm x 67mm. 9V operation. 1000m range £18.95QTX1130 Crystal controlled room transmitter. Narrow band FM transmit-ter for the ultimate in privacy. Operates on 180MHz and requires the use of ascanner receiver or our QRX180 kit (see catalogue). Size 20mm x 67mm. 9Voperation, 1000m range £39.95SCRX Subcarrier scrambled room transmitter. Scrambled output fromthis transmitter cannot be monitored without the SCDM decoder connected toreceiver. Size 20mm a 67mm. 9V operation. 1000m range £ 22.95SCDM Subcarrier decoder unit for SCRX. Connects to receiver earphonesocket and provides decoded audio output to headphones. Size 32mm x 70rnm.9-12V operation £22.95HVX400 Mains powered room transmitter. Connects directly to 240"., ASsupply for long term monitoring Size 30mm x 35mm, 500m range £1:95ATR2 Micro size telephone recording interface. Connects between tele-phone line (anywhere) and cassette recorder_ Switches tape automatically asphone is used. All conversations recorded. Size 16mm x 32mm. Powered fromline £13.45UTLX Ultra -miniature telephone transmitter. Smallest telephone trans -miner kit available. Incredible size of 10mm x 20mm. Connects to line (any-where) and switches on and off with phone use. All conversations transmittedPowered from line. 500m range £15.45TLX700 Micro -miniature telephone transmitter. Best selling telephonetransmitter. Being 20mm a 20mm it is easier to assertible than UTLX. Connectsto line (anywhere) and switches on and off with phone use. All conversationstransmitted. Powered from line, 1000m range £13.45STLX High-performance telephone transmitter. High power telephonetransmitter with buffered output stage providing excellent stability and perfor-mance. Connects to line (anywhere) and switches automatically with phoneuse. All conversations transmitted. Powered from line. Size 22mm a 22mm.1500m range £16.45TKX900 Signalling/tracking transmitter. Transmits a continuous stream ofaudio pulses with variable tone and rate. Ideal for signalling or tracking pur-poses. High power output gives range up to 3000m. Size 25mm a 63mm. 9Voperation £22.95

CD600 Professional bug detector/locator. Multicolour bargraph readoutof signal strength with variable rate bleeper and variable sensitivity used todetect and locate hidden transmitters Switch to AUDIO CONFIRM mode todistinguish between localised bug transmission and normal legitimate signalssuch as pagers, cellular, taxis etc. Size 70mm a 100mm. 9V operation £50.95

* * * SP ECIAL***DLTX/DLRX Radio control switch. Remote control anything around yourhome or garden, outside lights. alarms, paging system etc. System consists ata small VHF transmitter with digital encoder and receiver unit with decoderand relay output. momentary or alternate. 8 -way switches on both boardsset your unique security code. TX size 45mm a 45mm. RX size 35mm a 90mmBoth 9V operation. Range up to 200m Complete system (2 kits) £50.95Individual transmitter DLTX £19.45Individual receiver DLRX £37.95

A build-up service is available on all of our kits if required.UK customers please send cheques. PO's or registered cash Please add £1.50per order for P&P. Goods despatched ASAP allowing for cheque clearanceOverseas customers send sterling bank draft and add E5.00 per order for ship-ment Credit card orders welcome on 0827 714476.

OUR LATEST CATALOGUE CONTAINING MANY MORE NEWSURVEILLANCE KITS NOW AVAILABLE. SEND TWO FIRST CLASS

STAMPS OR OVERSEAS SENorNO 'RC's.

SUMA DIESIGR1STHE WORKSHOPS95 MAIN ROAD 0827BAXTERLEY, Nr ATHERSTONE 71 4476WARWICKSHIRE CV9 2LE

MICRO BASSThe latest sub bass kit from Wilmslow Audio. Micro Bassmeasures only 305 x 284 x 382 mm and can be positionedalmost anywhere in the confines of the listening room withoutaffecting the stereo image. It can be connected directly intoexisting systems and requires no extra amplifier. Designed asthe perfect partner for most small loudspeakers with asensitivity of 86 to 89 db.

The Wilmslow AudioMicro Bass kitcontains 2 x 8" Bassunits, flat packcabinet kit machinedfrom smooth M.D.Ffor easy assembly,xover kits and reflexport etc.

Other subwoofers availableto match most types of speakers

Impedance:8 ohmsAMP Suitability:20-100w

WiimslowAuer° A

£109.00 inc VATcarriage ins. £10.00

Wellington Close,Parkgate Trading EstateKnutsford, CheshireWA16 8DXTel: (0565) 650605Fax: (0565) 650080

DIY Speaker catalogue £1.50 Telephone credit card =post free (export £3.50) orders welcome

Open Tuesday to Saturday, 4 demonstration rooms available.

IT PAYS TO SUBSCRIBE TOELEKTOR ELECTRONICS

If you take out an annual subscription to ElektorElectronics, you not only save money compared with buy-ing the magazine from your local newsagents, but you havethe convenience of having it delivered to your home and thepeace of mind that you will not miss any issue. The totalcover price for the 11 issues appearing in 1992 will amountto £23.00 in the United Kingdom: more overseas, becauseimporters and their retailers have to add their (perfectlylegitimate) expenses. The (post paid!) subscription rates for1992 are:United Kingdom £22.00Rest of the world - surface mail £26.50AIRMAIL:Europe & Eire £27.50Middle East & North Africa £40.00South East Asia, Central & southern Africa,Central & South America, USA & Canada £40.00Australia, New Zealand, Far East andPacific regions £42.00The differences in these prices are caused merely by thepostage: the basic subscription price is well over 20% belowthe cover price and is the same all over the world.Write now for your subscription to:Elektor Electronics World Wide Subscription Service

Ltd Unit 4 Gibbs Reed Farm Pashley Road TICEHURST East Sussex TN5 7HE England

Phone (0580) 200657 Fax (0580) 200616.


AlICHAF,1 FARADAY: ATHER 0 ELECTRICITY(17914867)IA RARE accolade for a scientist, a por-

trait of Michael Faraday, universally re-garded as the greatest of all experimentalphysicists, has replaced that of Shakespeareon the reverse side of the new British £20note. He is also shown on one of a new issueof postage stamps featuring importantBritish scientists and inventors.

When asked what was the practical valueof the knowledge he was gaining in experi-ments that helped unravel the secrets of elec-trical currents, Faraday is said to havereplied: "Of what use is a newborn child?".The bicentenary of his own birth in 1791, thebeginning of a life whose potential was spec-tacularly realized, is being marked by a se-ries of tributes in the United Kingdom.

London's Science Museum has opened amajor exhibition centred on some of theoriginal apparatus used by Faraday to dis-cover the links between electricity and mag-netism that are at the heart of all present-dayelectrical technology. There are also docu-ments and reconstructions of his most im-portant experiments.

Commenting on the significance of thescientist's work to the modern way of life.ProfessorJohn Thomas, director of London'sRoyal Institution (founded in 1799 and oneof the first scientific research centres whereFaraday conducted his experiments and lec-tured), said: "Michael Faraday changed thevery basis of scientific thought. Newton'spicture of the Universe explains celestial me-chanics, the ebb and flow of the tides and thebasis of flight. It has nothing to say about themarvels of modem communications. Thesestem from a brilliant experiment carried outby Faraday in the quiet of his basement lab-oratory in Albemarle Street in 1845, whenhe demonstrated that light, electricity andmagnetism were intimately related. It is thebasis of all modern communication".

Litttle schoolingFaraday's ideas about the way in which elec-trical and magnetic action passes throughspace inspired the British mathematicianJames Clerk Maxwell to develop his electro-magnetic wave theory, which led to radio, aswell as to the microwave cooker.

Although destined to become one of theworld's greatest scientists, Michael Faradaydid not set out to be an inventor. His discov-eries stemmed from an intense interest in ex-ploring the workings of the natural world.

The new £20 note, bearing a portrait ofMichael Faraday.

He was born the son of a blacksmith on22 September 1791. At the age of 14, withvery little schooling, he was apprenticed toa London bookseller and bookbinder, andthrough reading became fascinated by sci-ence, especially chemistry and electricity.He became a laboratory assistant to SirHumphrey Davy and soon achieved his ownreputation as a leading chemist.

Later, he turned his attention particularlyto electricity and magnetism. All today'selectrical equipment employs discoveriesmade by Faraday. His work led to the electric

motor, the transformer and the generator. Healso found laws linking electrical and chem-ical action.

Wires and magnets ...On 3 September 1821, he made a device todemonstrate that an electric current and amagnet could be arranged to produce contin-uous motion, involving a freely -suspendedwire that swung around a magnet when a cur-rent flowed through it. This simple apparatusis now generally acknowledged as the fore-runner of the electric motor.

Ten years later, Faraday discovered thephenomenon known as electromagnetic in-duction. With two coils of wire wound on aniron ring, he found that a current in one circuitcould produce current in another. This is thebasis of the electrical transformer.

In September 1831, he found he couldproduce electricity from a magnet by movinga wire or the magnet so that the wire passedthrough the 'lines of magnetic force'. Theseexperiments were the foundation of all elec-tric power generated by mechanical means.

In other experiments, he established lawsof electrochemistry, or the chemical effectsof electric currents. These can be used eitherto remove substances from a compound (byelectrolysis) or to transfer a substance fromone electrode to another (electroplating).

... and chemistryAlthough noted mainly for his work on

electricity and magnetism, Faraday beganhis science with chemical analysis. He suc-ceeded in identifying some new compounds,discovering benzene in 1825, and was thefirst to liquefy chlorine.

Electrochemical processes have wide-spread industrial applications. For instance,all the world's aluminium and most copperand chlorine are produced by electrolysis.M

Produced and published by ELEKTOR or +44 580 200 657 (International ) GERMANY NETHERLANDSELECTRONICS (Publishing) Telefax: (0580) 200616 (National) Elektor Verlag GmbH Elektuur BV

or +44 580 200 616 (International) Stisterfeld Stride 25 Peter Treckpoelstraat 2-4Editor/publisher: Len Seymour 5100 AACHEN 6191 VK BEEKTechnical Editor: J. Baiting European Offices: Editor. EJ.A. Krempelsauer Editor: P.E.L KersemakersEditorial Offices: Postbus 75 GREECE PORTUGALDonis House 6190 AB BEEK Elektor EPE Ferreira & Bento Lda.Broomhill Road The Netherlands Kariskaki 14 R.D. Estefini.LONDON SW18 41Q Telephone: +31 46 38 94 44 16673 Voula - ATHENA 1000 LISBOA

England Telex: 56617 (elekr n1) Editor. E. Xanthoulis Editor: Jeremias SequeiraTelephone: 081-877 1688 (National Fax: +31 46 37 01 61 HUNGARY SPAINor +44 81877 1688 (International ) Managing Director: M.M.J. Landman Elektor Elektronikai folyorat Resistor Electronics AplicadaTelex: 917003 (LPC G) 1015 Budapest Calle Maudes 15 Entlo C.

Fax: 081-874 9153 (National ) Distribution: Batthyany u. 13 28003 MADRID

or +44 81874 9153 (Intemational SEYMOUR Editor. Lakatos Andr5s Editor. Agustin Gonzales Buelta1270 London Road INDIA SWEDEN

Advertising: PRB Limited LONDON SWI6 -1DH Elektor Electronics PVT Ltd Electronic Press AB

3 Wolseley Terrace Chhotani Building Box 5505

CHELTENHAM GL 50 1TH Printed in the Netherlands by NDB, 52C. Proctor Road. Grant Road (Et 14105 HUDD1NGE

Telephone: (0242) 510760 Lxterwoude BOMBAY 400 007 Editor. Bill CedrumFax: (0242) 226626 Editor. C.R. Chandarana USA & CANADA

ISRAEL Elektor Electronics USASubscriptions: Overseas editions: Elektorcal P 0 Box 8761Vorld Wide Subscription Service Ltd. FRANCE PO Box 41096 PETERBOROUGH NH 03458-0876

Unit 4. Gibbs Reed Farm Elektor sarl TEL AVIV 61410 Publisher. Edward T. DellPashley Road Les Trois Tilleuls Publisher. M. AvrahamTICEHURST TN5 7HE B.P. 59: 59850 NIEPPETelephone: (0580) 200657 (National ) Editors: D.R.S. MeyerG.C.P. R.,xlerstiorf


DYNAUDIO FROM WILMSLOWWilmslow Audio (see page 4) have becomethe official importers of Dynaudio drive unitsin the United Kingdom. They have a largerange of drive units in stock and full techni-cal information is available upon request.Wilmslow Audio Ltd, Wellington Close,Parkgate Trading Estate, Knutsford, Cheshire,WA16 8DX; telephone (0565) 50605.

NATIONAL CODE CHANGE SETFOR EASTER 1994

BT has announced that, to meet Britain'srapidly growing demand for new codes andextra telephone numbering capacity, as fromEaster Sunday, 3 April 1994, an extra digit1' will be added after the '0' at the begin-

ning of area dialling codes throughout thecountry. For example, Bristol's code willchange from 0272 to 01272; Edinburgh's 031code will become 0131; and the 071 and 081London codes will change to 0171 and 0181.

THINKING COMPUTERS MIMICHUMAN BRAIN

Computer -based programs, such as buildingmanagement systems (BMS), are widely usedin domestic and industrial locations to con-trol temperature and warn of fire and intrud-ers. The systems have been evaluated by theUnited Kingdom's Building ResearchEstablishment (BRE) at Watford, and their im-proved performance over ordinary programshas been shown to result in energy savingsof up to 30%.

Researchers at the world-famous labora-tory, which is devoted exclusively to devel-oping new building techniques and fire pre-vention methods, are now evaluating the useof artificial intelligence techniques to fur-ther enhance the performance of BMS.One investigation at BRE is into the use ofneural networks-computer software that

1111111111111111

A BRE researcher working on the devel-opment of neural computer programsthat mimic the human brain and 'think'their way through problems.

ELECTRONICS SCENE

attempts to mimic the operation of the humanbrain. Neural networks can learn from expe-rience in the manner of the human brain.Building Research Establishment, Garston,Watford, England, WD2 7JR

POCKET -SIZED TELEPHONEThe past ten years have witnessed an explo-sion in the number of telecommunicationsproducts and services on offer, especially inBritain. Everything from in -car telephonesto cordless versions that operate anywhereindoors or out are now common.

Prototype of the new Mercury personalcommunications pocket -sized telephone.

The next few years could, however, seethe biggest revolution of all. By the year2000, one in five of the British populationcould be carrying their own calculator -sizedpocket telephone that will make and receivecalls from any location.

The idea of a personal communications net-work (PCN) is appealing. The potential mar-ket includes everyone from company direc-tors to people in the street, provided the pricescompare favourably with a fixed link tele-phone service. Moreover, it can replace theneed to install new telephone lines in homesand businesses.

PCN forms part of a European standardcalled either GSM (Groupe Speciale Mobile)or the pan-European digital cellular network.This will eventually allow users in ten Europeancountries to use the same portable telephonyequipment to make and receive calls anywherein the GSM area.Mercury PCN, 1 Harbour Exchange Square,London E14.Vodafone, The Courtyard, 2-4 London Road,Newbury RG13 1JL.

Cellnet, 1 Brunel Way, Slough SLI 1XL.Unitel, Elstree Tower, E1stree Way, BorehamwoodWD6 1DT.Microtel, Aztec Centre, Aztec West BusinessPark, Almondsbury, Bristol BS12 4TD.

ELECTRONIC WARNING OFSUSPICIOUS AIRLINE BAGGAGE

The airdisaster involving an American JumboJet at Lockerbie in Scotland on 21 December1988 brought home to the world the vulner-ability of passenger aircraft to attack by ter-rorists and the fallibility of existing baggagecontrol systems.

Using a portable radio terminal to key -inbaggage tag numbers.

In the aftermath of the tragedy, the reportof the United States President's Commissionon Aviation Security and Terrorism stated thatthe total destruction of PanAm flight 103might have been avoided if stricter procedureshad prevented unaccompanied bags frombeing loaded.

Separately, a joint baggage security work-ing group set up by the International AirTransport Association (IATA) and the AirTransport Association (ATA) recommendedthat the positive matching of passenger bag-gage and the mandatory introduction of se-quential bar-code numbering for baggageshould be carried out by all airlines.

To help airlines comply with this recom-mendation and with the United Nations'International Civil Aviation Organization di-rective Annex 17 Standard 4.3.1., BRaLS(Baggage Reconciliation and Location System)Ltd and the British Technology Group (BTG)have introduced the first fully integrated pas-senger and baggage reconciliation system tobe available commerically.

It enables airlines to comply with inter-national regulations at a relatively low cost,estimated to be about 10-30 pence (8-20 0)per passenger, depending on throughput, thesystem configuration and pay -back period.

The equipment can be linked to the main-frame host computer of a departure controlsystem (DCS) or installed as stand-aloneequipment. The primary platforms are allUNIX -based.

At the check -in point, baggage tags bear-


Ing)

ing a unique identification in numerals andbar code is fixed to each item of baggage.The number would be read into the com-puter and identified with the passenger con-cerned. The bar-code reading devices canbe hard -wire or radio linked to the masterradio base station connected to the rampprocessor.BRaLS Ltd, Tring Business Park, UpperIcknield Way, Tring HP23 5AG, England.British Technology Group, 101 NewingtonCauseway, London SE1-6BU.International Aeradio PLC, Aeradio House,Hayes Road, Southall UB2 5NJ, England.

BATAK: ELECTRONICS AND SPORTBatak (the new sport for the 1990s?) is theworld's largest interactive computer -con-trolled game, played against a wall with elec-tronically actuated targets and scored auto-matically by striking the illuminated targetswith a foam -tipped racquet.

Batak is simple to play, with the mini-mum of equipment, by one or more players,for fun, competitively or for calculable exer-cise/keep fit programmes that don't get bor-ing. It requires only a fraction of the space nor-mally required for court and racquet games,yet still creates the same excitement, skill, funand speed that make them so popular.

A computer controls the sequence andspeed of target illumination, gives audio re -

1al ULM= if 44 4 41

NIIIIIIMIN 4111111MMOPI

TEE AND IEEIE PROGRAMME

4 Nov-Reduced instruction set computer(RISC) (colloquium).

4 Nov-Mobile radio communications: thetechnology options.

5 Nov-International manufacturing lecture.6 Nov-Managing software users.6Nov-Electromagnetic compatibility forpro-

ject managers.6-7 Nov-Effective CADCAM '91.*7 Nov -34th HunterMemorial Lecture: Electric

railways-the systems engineering revolu-tion.

7 Nov-Inspection and testing of electrical in-stallations.

8 Nov-Multi-octave microwave circuits.11 Nov-VLSI vision technology.12-15 Nov-Autotech '91.*13-15 Nov-Mobile radio (Nice, France).15 Nov-Interfaces for computer peripheral

equipment.19 Nov-Direct digital frequency synthesis.

ELECTRONICS SCENE

spouses for 'strikes' and 'misses' and timesand scores the game on two large LED read-outs mounted on the playing interface.Quotronics System Engineering, 35 Lee Street,Horley RH6 8ER, telephone (0293) 785 132.

BUFFER PROTECTION UP TO 1500 V

Instrument buffer Type 611, from LaplaceInstruments isolates and protects instrumen-tation from AC/DC voltages up to 1500 V.Its output circuit cannot source more than 15 Veven under fault conditions and its precisionattenuator can select output/input ranges of1:1; 1:10; and 1:100.Laplace Instruments Ltd, Tudor House, GrammarSchool Road, North Walsham NR28 9JH

IEEIE CREATES RAILWAY SIGNALDIVISION

An agreement between the Institution ofElectronics and Electrical Incorporated Engineers

EVENTS

20Nov-The impact of technological advanceson disabled people.

21 Nov-INMARSAT 3 (colloquium).21 Nov-Electricity at work regulations 1989.25-27 Nov-Computation in electromag-

netics.28 Nov-Cochlear implants in the UK.

Information on these, and many other, eventsorganized by the ME or the IEEIE may beobtained from the IEE, Savoy Place, LondonWC2R OBL, Telephone 071 240 1871 or fromthe IEEIE, Savoy Hill House, Savoy Hill,London WC2R OBS, Telephone 0718363357,except those marked (*) for which informa-tion should be sought from the Institution ofMechanical Engineers, 1 Birdcage Walk, LondonSW1H 9JJ, telephone 071 222 7899.

(IEEIE) and the Institution of Railway SignalEngineers (IRSE) enables railway signal en-gineers to register through IEEIE with theEngineering Council as incorporated engi-neers. The special IEEIE/IRSE dual mem-bership will enable those incorporated engi-neers who wish to enjoy the specialist learnedsociety activities of the IRSE-with its 2000members engaged in railway signalling-tohave, at the same time, the benefit of beingpart of the dynamic, 28,000 strong IEEIE,the largest institution of its kind in Europe.IFFIE, Savoy Hill House, Savoy Hill, LondonWC2R OBS.

BBC ENGINEERING INFORMATIONON THE MOVE

The Engineering Information Department ofthe BBC is now located at the new BBCHeadquarters building at White City. Thefull address for all correspondence is

Engineering Information DepartmentBritish Broadcasting Corporation

White City201 Wood Lane,

London W12 7TS

General enquiries 081 752 5040Radio enquiries 0345 010 313

Facsimile 081 752 5503.

IEE TO GUIDE LONDON THROUGHNEW WIRING REGULATIONS

The 16th Edition of the ME Wiring Regulationspublished last June will mean that everyoneworking in the field of electrical installa-tions, whether as designers, specifiers or in-stallers, will need to familiarize themselveswith the changes contained in the new edition.

To meet this need in the London area, theIEE will hold another one -day conversioncourse on 5 December at their headquartersat Savoy Place, London. Both members andnon-members of the IEE can attend.

During November, Frost & Sullivan will con-duct seminars on a variety of topics in thefield of electronics, computers and com-munications. Further information from Frost& Sullivan, Sullivan House, 4 GrosvenorGardens, London SW1W ODH, Telephone071 730 3438.

This year's annual Financial Times conferenceon World Mobile Communications will be heldin London on 31 October and 1 November.Full detailsfromTheFinancialTuuesConferenceOrganization, 126 Jermyn Street, LondonSW1Y 4UJ, telephone 071 925 2323.

Componic '91, the 30th International Exhibitionof Electronic Components will be held inParis from 18 to 22 November. Details fromPromosalons, The Colonnades, 82 BishopsBridge Road, London W2 6BB, telephone071 221 3660, fax 071 792 3525.


I41

E

SA'EL t TV ir-acrtiv[EigFaq

While the majority of satellite television channels have soundsubcarriers with 50-11s or 75 -us pre -emphasis, certain channels,particularly French and Italian, use an alternative standard called

J17. Since most satellite TV receivers sold in Western Europe lack aJ17 de -emphasis, and suppliers and retailers do not want to know,

the author has come up with a simple add-on filter to solve theproblem.

IE profusion of standards has always1 been a problem in the radio and televi-

sion field. To add to the general confusionabout picture standards and encryption sys-tems, some French and Italian users of satel-lite TV transponders have seen fit to use asound pre -emphasis that differs signifi-cantly from the well -established and farmore generally used 50-ps or 75-ps stand-ards. The J17 pre -emphasis is used with allchannels on the Telecom 1C satellite (orbitalposition 5° West) and with the RAI Uno andRAI Due channels on the ECS1-F5 satellite(orbital position 10° East). The designationJ17 derives from CC11'1 recommendationJ17, published in Geneva in 1972.

Although a small proportion of domesticsatellite TV receivers are compatible with

by J.M. Winsor

TELEVf FRANCOPHONEPAR SATELLITE

Fig. 1. Using a standard 50-.s de -emphasis characteristic (curve b) to receive a soundchannel with J17 pre -emphasis (curve a) yields an undesirable amount of high -boost (curvec).

J17, the majority are not. In particular, thepopular 'Astra' type of receiver is equippedwith 50-ps de -emphasis only.

Different curves, differentsounds

While listening to a broadcast with J17 pre -emphasis and using a receiver with 50-ps or75its de -emphasis (62-ps is also applied insome cases), the perceived effect is thatspeech seems excessively sibilant, and musicharsh and unnatural. The cause of this effectis illustrated in Fig. 1, where curve 'a' repre-sents the J17 pre -emphasis characteristic,normalized at 0 db at 100 Hz, and curve 'b'the 50-ps de -emphasis characteristic. The net

Fig. 2. Depending on the attenuation thatcan be tolerated. only R1 in this networktakes a different value. For 6 dB attenuation:

k(2: for 10 dB attenuation: 171=22Note that the network is designed for Z-600 LI, and Zout 100 kci.


ElJ17 EQUALIZING NETWORK FOR SATELLITE TV RECEIVERS

result obtained by adding curve 'a' to curve13', i.e., receiving a J17 transmission with a50-ps de -emphasis, is represented by curve'c'. Clearly, high frequencies are boosted ex-cessively at no less than about 12 dB at4,000 Hz.

The J17 pre -emphasis characteristic isderived from the equation

18.75- 10logio( 75 + a)/ 3,°°°)2 ) dB1 + ( to/3,000)2

normalized to 0 dB at 100 Hz.By contrast, the 50-tis de -emphasis char-

acteristic is derived from the equation

120 logiO(J

-v1 + (co T/106) -

where T = 50 tts.

A proposed solutionRather than designing a dedicated J17 de -emphasis circuit, the author set out to re-shape the output of the 50 --us de -emphasissuch that acceptable sound reproduction isobtained with the French and Italian chan-nels utilizing the 117 standard.

The circuit shown in Fig. 2 is a simpleequalizing network developed to improvethe poor frequency response that resultsfrom applying a 50-ps de -emphasis to a sig-nal with J17 pre -emphasis.

The network is intended for connectionbetween the audio output of the satellite TVreceiver and the audio input to the TV. It isdesigned for a source impedance of 600 SI orless, and a terminating impedance of 100 Ici2

or greater.The curves in Fig. 3 illustrate the effect of

the network. Curve 'a' represents the net re-sult of applying 50-µs de -emphasis to a sig-nal with J17 pre -emphasis (as curve 'c' inFig. 1), and curve 'b' the frequency responseof the equalizing network. The attenuationof the network is selected to be 6 dB. Theoverall result of adding curves 'a' and 'b' isrepresented by curve 'c'. A total excursion ofamplitude of 6 dB is achieved, which is agreat improvement over that in curve 'c' inFig. 1.

The design of the equalizing network is a

matter of compromise. It inevitably intro-duces attenuation which may be overcomeonly by increasing the volume setting at thetelevision set, or by the provision of addi-tional gain elsewhere. The greater the atte-nuation that can be accepted, the flatter willbe the overall frequency response.

Figure 4 is similar to Fig. 3, except thatcurve 'c' shows the overall frequency re-sponse that results from an equalizing net-work having an attenuation of 10 dB. Hence,the output level will be about one third thatat the input. It is evident that the overall fre-quency response as shown by curve 'c' inFig. 4 is much flatter than that in Fig. 3,showing a total excursion of amplitude of4 dB (only 2.5 dB up to 7,000 Hz).

x+( 4,167/(02

Fig. 3. Output characteristic of a J17 signal after 50-ps de -emphasis (curve a); characteristic of equalizing network (curve b); net result of adding characteristics 'a' and 'b' (curvec). Compare curve 'c' to that in Fig. 1 to see the beneficial effect of the network.

1111111111111Pwra'-'101

Fig. 4. As Fig. 3, but with network component values to give an attenuation of 10 dBDespite the higher attenuation, the output characteristic is better than that of the -6 dBversion of the network.

The frequency responses of the proposed (1/4,167 s) for the 6 -dB version, and 384 usequalizing network are derived as follows. (1/2,604 s) for the 10 -dB version. TheseFor the 6 -dB attenuation network: values are nominal only.

10logio1+(4167/03)2

While for the version with 10 dB attenuation

10 logiox + ( 2,604/03)21 +( 2,604/(13)2

wherex =1 / anti -login ( dB attenuation / 10 ).The time constant C1(121+R.2) equals 240

dB


2 3

G

CJ

L

The circuit presented hereenables the Archimedesvideo digitizer described afew months ago to handlefull -colour images at animpressive resolution of24 bits.

by J. Kortink

idea to process colour pictures withTtlfri aid of the video digitizer was bornshortly after the publication of Ref. 1. Al-though computer -based colour image pro-cessing is far from simple (even asuper -VGA card can only show 256 coloursat a time), there is a growing demand for di-gitized colour photographs. This is mainlyon account of the widespread use of profes-sional DTP (desk -top publishing) packagescapable of supporting 24 -bit colour illustra-tions with the aid of colour separation tech-niques.

The hardware and software describedhere was developed with DTP applications(on the Acorn Archimedes) in mind. The per-formance of the digitizer is quite impressive:it needs only 10 seconds or so to 'read' anddisplay a colour photograph - much fasterthan many commercial products currentlyoffered.

Colour imagesFor black and white screen images, only one'ink' colour is used: usually black. A videoimage may be reproduced with digitalmeans by scanning the image at a suffi-ciently high resolution (both as regards thenumber of pixels and the number of grey le-vels). The video digitizer has a resolution of640x512 pixels at 256 grey levels. The repro-duction on paper of the image requires thediscrete grey levels to be transformed backinto ink spots. On a computer screen, intens-ity modulation is used, while on paper greylevels are reproduced by controlling the sizeof the ink spots. Hence, the resolution of theprinter attached to the computer determinesthe number of grey levels that can be repro-duced. Most 300 -dpi (dots per inch) laserprinters are capable of reproducing 16 simu-lated grey levels. Professional image settingmachines typically achieve 200 levels when

set to a resolution of 2400 dpi. Photographicpaper has no such limitations, offering a vir-tually infinite number of grey levels.

In practice, 64 intensity levels are suffi-cient to faithfully reproduce a photograph.The 256 grey levels offered by the video di-gitizer are, therefore, ample for most appli-cations.

Colour images on a computer screen areinvariably made by 'building' individualpicture elements from three primary colours.In the case of the video digitizer, use is madeof the three additive primary colours red,green and blue (RGB model). By contrast,magazine and book printers traditionallyuse the subtractive primary colours cyan,magenta and yellow, to which is added black(CMYk model).

During the digitizing operation, the satu-ration of the three primary colours red, greenand blue of each picture element ismeasured. This means that the image is ac-tually scanned three times: once for the redcomponents, once for the blue components,and once for the green components, the satu-ration of which is stored as a digital value inthe computer memory. A digitized colourimage is then obtained by combining thethree scanned images containing the pri-mary colours. This technique is not unlikethat used in the graphics and printing indus-try, where colour separations are used toproduce three or four plates (screens) thateach represent a primary colour. The fourthplate represents black, and is used for thebrightness information only. During the off-set printing process, the images on the platesare printed over one another to give animage with composite colours. Note, how-ever, that the individual picture elements arenot printed on top of each other, but side byside with an accurately controlled distance(off -set). In this system, the black plate is the-oretically not required, since black can be re-

produced with the three primary coloursalso. In practice, however, black so printedhas a brownish hue. Grey levels are, there-fore, printed with black ink, which avoidsundesirable sub -colours.

Returning to computers, the number ofcolours is usually determined by the type ofvideo card used. Fortunately, we may availourselves of mixing techniques to take thenumber of colours that can be shown on thescreen over the number of colours actuallyavailable. This is achieved by dithering, orpixel combination. Unfortunately, ditheringhas the inherent disadvantage of reducingthe overall picture resolution.

Dithering techniques are based on errordistribution protocols developed by, amongothers, RW. Floyd and L. Steinberg. Theseprotocols allow the difference between thedesired colour and that produced on thebasis of the actually available colours to bemade as small as possible. Unfortunately, wehave to leave the subject of dithering at thisbrief description, since a discussion of the re-lated techniques is beyond the scope of thisarticle. The soft -ware developed for the pres-ent colour extension of the video digitizerdoes, however, provide pixel dithering rou-tines.

The hardwareAs already mentioned, the black and whitevideo digitizer described in Ref. 1 is taken asthe starting point. To be able to process col-our images, the video information has tosupplied as RGB signals. Lacking an RGBoutput on your video equipment, existingcomposite video or S -VHS video signalshave to be converted into RGB signals first.This may be done with a special converter-see Ref. 2. Cameras, video recorders, tunersand the like capable of producing RGB sig-nals do not require an additional converter,


and can be connected direct to the input ofthe present circuit, which is basically an in-terface between the video source and theinput of the video digitizer.

The circuit diagram of the RGB interfaceis given in Fig. 1. The circuit is really quitesimple. Two control signals supplied by thedigitizer select one of the four signal inputsvia ICia and four relays. As indicated, threeinputs are used for the primary colours, anda fourth for a composite video signal, allow-ing black and white images to be digitized ina simple manner. The fifth input serves topass the synchronization signals to the videodigitizer. This input is not switched becausethe sync signals must always be present.Four LEDs are used to indicate which of thefour video inputs is connected to the di-gitizer.

The printed circuit board used to buildthe interface is shown in Fig. 2. The construc-tion is straightforward, and should not pres-ent problems to those of you who havealready successfully built the video digitizer.

A small modificationThe circuit of the video digitizer requires asmall modification to be made before it canbe used for the colour applications we havein mind. First, a 9 -pin sub -D connector isfitted on the aluminium support bracket ofthe podule. This connector takes the place ofthe existing cinch or BNC socket. For theoriginal black and white application of thedigitizer, the video input is fed to the A -Dconverter (ICio) and the sync separator (ICit)via a buffer stage. Since for colour applica-tions the synchronization and colour signalshave to be applied separately, capacitor C18is removed from the board, and a wire is sol-

24 -BIT FULL COLOUR VIDEO DIGITIZER

Fig. 1. Circuit diagram of the colour interface for the video digitizer.

dered to pin 2 of ICII. Next, a 100-nF capaci-tor (C18) is soldered between the free end ofthe wire and pin 1 of the D9 connector. Thevideo signal is connected to pin 2, the powersupply to pin 3, and ground to pin 4.

Finally, we need two control signals.These are supplied by 1C15, a PCF8574 - oneof the three optional I/O ICs on the podule.Fit this IC and the associated pull-up resistorarray, R9, on the podule board. Output linesPO and P1 (IBO and 1131) are available onpins 1 and 2 of connector K2. These lines areconnected to the D9 connector, for instance,to pins 5 and 6.

The colour interface is connected to thevideo digitizer via a 6 -way cable terminatedwith D9 connectors at both ends.

The above modification still allows blackand white images to be digitized without theinterface when a video connector is usedwith a 9 -pin D plug, of which pin 1 is con-nected to pin 2, i.e., the video signal is con-nected to these two pins, while ground isconnected to pin 4.

Software: IGreyEditThe control program for the colour versionof the video digitizer is called !GreyEdit, andcomes on a diskette (Archimedes format)supplied through the Readers Servicesunder order number ESS 1631. All otherfunctions in the program !VidiDigi (seeRef. 1) are also included on this disk.

COMPONENTS LIST

Resistors:2 10k..(1

4 4700.

Capacitors:1 100nF

Semiconductors:1 74HCT1394 1N4148

LED

R1,R2R3 -R6

Cl

IC1

D1 -D4

D5-08

Miscellaneous:V23100 -V4005 -A000 Ref-Re4(Siemens)

1 9 -way sub -D plug, with hood1 9 -way sub -D socket5 RCA (phono-) socket for PCB mounting1 enclosure, aluminium, size approx.

109x58x25 mm (e.g.. Hammond 1590)1 control software on disk ESS1631

Fig. 2. Printed circuit board for the colour interface.


COMPUTERS AND MICROPROCESSORS

sv

e o

e l

Fig. 3. The modifications to the originalvideo digitizer are simple to make, while allfunctions are retained.

The digitizer functions are available viathe 'digitizer'sub-menu. The buttons on the

select one of the threeprimary colour signals, or the black andwhite signal. The selection must be indicatedby the LEDs on the interface board. The pro-gram also has a utility that allows a colourimage to be digitized automatically. On se-lecting this function, you are prompted to in-form the computer where the digitizedimages are to be stored. Simply drag the iconto a directory on the diskette or hard disk.The digitizing operation starts immediatelyafter this has been done. The files containingthe red, green and blue information arestored successively in the indicated direc-tory. The digitizing process is complete after

Fig. 4. Digitized photographs taken with a Canon ION. Use was made of the S -VHS outputAlso, the picture was improved with the 'sharpen' function provided by the control softwaredeveloped for the Acorn Archimedes.

three scans, and the R, G and B informationis ready for combining into a new, larger,file. Drag the directory with the three pri-mary colour files to the window marked!GreyEdit. Next, the program asks youwhere the photograph is to be stored (saved),and under what name. If everything workscorrectly, the colour photograph will beready after a few seconds of number crunch-ing activity. The photograph is written in the'Clear' format and may be converted into aGIF, 111-.P or PBM file with the aid of the !Cre-ator utility. Mind you, the current versions of

Fig. 5. Interface fitted in a compact enclosure. The unit is connected to the computer viaa short flexible cable.

the GIF protocol do not support 24 -bit col-ours. It is, therefore, recommended to usethe at+ format to export files.

The picture may be edited with !Transla-tor, which is supplied on the disk togetherwith !GreyEdit and !Creator. Finally, notethat the size of the image files requires a harddisk and a minimum memory of 2 MByte.

Black and white imageenhancementA couple of experiments with a Canon IONRC -260 camera showed that the quality of di-gitized black and white pictures can be im-proved considerably when a video sourcewith an S -VHS (super -VHS) output is used.Where such a source is available, connect itsluminance signal (pins 1 and 3) only to thedigitizer. Results dose to perfection may beobtained in this manner, particularly if the'sharpen' function of !Greyedit is used. Thetwo photographs in Fig. 4 were produced inthis way. After digitizing, they were con-verted into Postscript, and printed on a AgfaCompugraphic Type 9800 photo typesetter.

References:1. Black -and -white video digitizer (for AcornArchimedes). Elektor Electronics July/Au-gust 1991.2. S-VHS/CVBS-to-RGB converter. ElektorElectronics September and October 1990.


DISSIPATION LIMITER19

rrHE secondary alternating voltage in power1 supplies stabilized with the aid of seriesregulators is normally calculated to ensurethat with maximum output voltage and cur-rent the potential across the reservoir capac-itor does not drop below the minimum volt-age, U required for satisfactory operationof the regulator. When the supply operatesbelow maximum output voltage and cur-rent, thepower thatis then not required is con-verted into heat. Moreover, the design al-lows for maximum output voltage and cur-rent even when the mains voltage is at i ts low-est specified level.

The secondary voltage may well be spec-ified too high, with the result that when themains voltage is at its highest specified leveland the power supply is open -circuited, themaximum permissible potential across thereservoir capacitor, and thus that at the inputof the regulator, may easily be exceeded.

It is dear that in most power supplies, thespecified voltage levels for the regulator andreservoir capacitor are needlessly high.

Possible solutionsThe uneconomical design outlined abovemaybe improved in several ways. It is, forinstance, possible, with the aid of relays, to se-lect secondary alternating voltages accordingto the load requirement. This is, however, pos-sible only with a very special, and thereforeexpensive, mains transformer. Another pos-sibility is to use phase gating control. This,however, requires fairly high switching cur-rents and causes noise on the mains supply.Moreover, capacitors that can withstand highcurrent pulses must be used and these areexpensive.

A third method is used in the circuit pre-sented here. In this, the charging of the reser-voir capacitor is interrupted when a certainpotential across the capacitor is reached. Thecircuit, which can be added to most existing

by C. Zschocke

Dissipation in electronically regulated powersupplies, manifested by heat, is a problem that isnormally tackled by providing external heat sinks.

The limiter presented here offers a moreintelligent solution. It can be used as an add-on

unit with virtually any power supply.

regulated mains power supplies,minimizes the power dissipation;offers an additional reduction in dissipa-tion during a short-circuit or during cur-rent limiting;is inexpensive to build;uses only standard, readily available com-ponents;needs only little space;is suitable for high -current supplies;

can be used with positive and negativeseries regulators.

Principle of operationIn Figure 1, an electronic switch, for instance,a SIPMOS transistor, is inserted between therectifier and the reservoir capacitor. At thestart of each half cycle, the switch is on (LI r>3 V).A charging current flows into the capacitor

Fig. 1. Principle of operation of the dissipation limiter.


20 POWER SUPPLIES & BATTERY CHARGERS

and the potential, Ur, across the capacitorfollows the instantaneous level of the half-wave voltage. When Lk has reached the min-imum level,Uso, required for the wanted out-put voltage, Uo, the charging current is switchedoff until the end of the half cycle. The levelof Uso must be chosen to ensure that 14 doesnot drop below Ur until charging current be-gins to flow again at the start of the next halfcycle. This voltage is the sum of the wantedoutput voltage, the drop across the regula-tor, Ud, and the voltage resulting from the dis-charging of the capacitor. The difference AUthat is,the series switch -off voltage, Uso(m), be-tween Ur and Us° is easily determined, sinceit is directly proportional to the chargingcurrent; in fact, AU=It/C, where [is the charg-ing current, t is the time during which charg-ing current flows, and C is the capacitanceof the reservoir capacitor. Thus,

LI,,,=U0-1-Ud+AU

=Uo+Ud+It/C.

The time t is, in full -wave rectification,equal to a half cycle of the mains frequency,that is, T/2. Then,

Uso=11,-1-Ud+IT/2C

If now the maximum permissible outputcurrent, /r(m) is substituted for I, the highestpotential drop across the regulator (at the startof each half cycle) is

14000=Uso-Uo=Ud+4,00T/2C.

From this, the instantaneous dissipation, Pcan be computed:

Pr=I01.1,0(m)=IJUd+100J/2C-/oT/4C

As an example: a power supply with areservoir capacitor of 3 300 µF is required toprovide an output voltage of 0---15 Vat a max-imum current of 1 A. Since the potential dropacross the regulator is 3 V, the voltage acrossthe capacitor, Li, must be at least 18 V. Withoutlimiting, the maximum power loss, Pro,* inthe regulator is 18x1=18 W (which is, for in-stance, the case when maximum current flowsat very low output voltage as in a short-cir-cuit condition).

With limiting, the voltage across the ca-pacitor, Uso, needs to be only

1.1,,=Uo+Ud+/0t/C=:14 -3+1x10 -2/3300x10-6=3.03 V.

The maximum dissipation is reduced to

Pqmrldlid+1,00T/2C-10T/4C1.

Since 10=1.04,

Pqm)=10(Lid+10(m)T/4C=

=1(3+1x2x10-2/4x3300x10-6=4.5 W,

which is an appreciable improvement.

Losses uirs kritrg dealt

vtinovt trritaeg Goal

U,. recitict output

U.,. switch -0g voltage

U, = minimum, fogy:slot ingot rouge

U. output voltage

tie= drop across regulator

Fig. 2.Dissipation with and without limiter.

G

72

C 0

BC5568CT G 013

41ATOT

IWO1 ICt

411t7

08

t4 i

2

C

0

BUZIO

C

IC$

vv. EVI

0

Fig. 3. Circuit diagram for both versions of the dissipation limiter.

ELEKTOR ELECTRONIC'S \O\ EMBER 1991

DISSIPA TION LIMITER LAt

00 001-2Fj ap-

of

8 Of o or n, )0°0-01-0 A 010-t1-0OL Rs icoej 0{ R2 f0of R7 }Os 0)R13 f0Of R9 10C2 L'ilat R3 10

01 R5 )0[01 R4 10 001 RB 10 *--14--0 0 00-01102 r'i % ru

--I

01111 1 f0[003;10_,0{14 1 2 )0 '0--1`00{R10 )0 0.21:1 r(74

CLD -() -T-T- E1 C;)

0

N

!.

0

g

0

I

00

01 R1 )00-1-40-014:7), D1 0-N4-0Of R6 IOC Of R2 10

of R7 10 it 0)1413 toof R9 )Om 01 F13 to

0-14-0O( R4 )0

oolfor2001 1 10

0114 12 )0 00 140-42001R10 cr

000-n -m ->-

Fig. 4. Printed circuit boards are available for both versions of the dissipation limiter.

Resistors:R1, R2 = 4.7 kil

R3 = 100 162

R4 = 1 ki-1

R5-9 = 22 Xi/

R10, R11 = 220 I62

R12 = 2.2162

R13 = 100161

Capacitors:Cl = 1 nFC2 = 47 uF, 100 V

Semiconductors:D1 = zener diode, 20 V, 400 mWD2 = zener diode, 39 V, 400 mWD3 = 1N4148D4-6 = 1N4004D7 = see textT1, T3 = BC556B (pos); BC546B (neg)T2 = BUZ10 or BUZ11 (see text)T4, T5 = BC546B (pos); BC556B (neg)

Miscellaneous:Heat sink for T2PCB 910071

The practical circuitThe circuit in Fig. 3 maybe split into the drivefor the current switch, the switch itself ( T2)and a 'thyristor'. The thyristor is formed bytransistors T3 and T5; its cathode is con-nected to the output terminal of the seriesregulator.

The thyristor is switched on via zenerdiode 137 as soon as the voltage across theregulator exceeds the level fixed by the zenerdiode plus the b -e voltage of T5. It switchesitself off just before the next zero crossing.To make certain that it does, it is decoupledfrom the rectifier and the inverse diode of T2via D4 and D5, which are loaded by R5.

When the thyristor is on, T4 switches Tion and this in turn switches T2 off.

Transistor T4 also serves, in conjunctionwith R3, to decouple the thyristor and thecharging -current switch.

Diode 1-2.04 and C2 provide T2 with the nec-essary switching voltage.

Zener diode DI protects the gate of T2against overvoltage.

The inductance of the mains transformercauses a high back e.m.f. when the chargingcurrent is switched off, and this may upset thecorrect operation of T2. This is prevented byD2 and D3, which switch off T2 as soon asthe back e.m.f. exceeds the permissible reversebias. A useful effect of the back e.m.f. is thatit increases the potential across C2, which guar-antees satisfactory switching of T2 even whenthe output voltage is a maximum.

Resistor R12 ensures correct switching ofT2 when the output is not loaded (open cir-cuit). It must be rated at U00,02/R32 Watt. Itmay be omitted if the power supply is per-manently loaded.

Transistor T2 must be fitted, insulated, onits own heat sink or on that of the regulator.

Depending on the component values, thelimiter may be used with secondary alter-nating voltages of up to 50 Vpeak. It is notpossible to say what the maximum load cur-rent can be without knowing the value ofthe reservoir capacitor and the specificationof the mains transformer. What can be saidis that the maximum charging current pulsesmust not exceed the maximum permissiblecurrent through T2, which is 19 A for theBUZ10 and 30 A for the BUZ11.

The limiter can be used without any prob-lem in 5 A power supplies. The prototypeoperated satisfactorily with an output currentof 10 A (reservoir capacitor was 10 000 uF).It is always possible in doubtful cases to con-ned two or three SIPMOS transistors (T2) inparallel.

Here are some tips for calculating the val-

ues or ratings of a number of components.The zener voltage of 137 must be U0 -0.7V(U of T5).

The maximum reverse bias of T2 must begreater than the peak value of the poten-tial across the reservoir capacitor.

The zener voltage of D2 must be equal tothe maximum reverse bias of T2 minus5V.Capacitor C2 must be rated at 11,00,0 plus20 V (Di).Transistor T4 must be able to withstandvoltages of 1.4,,(m) plus 20 V (DaTransistors T3 and T5 must have a reversebias rating equal to the peak value of thesecondary alternating voltage.Resistor 1232 must be rated at 1.10(m)2/R12(watts).All connecting wires should be as shortas possible and of appropriate thickness.

The dimensions of the heat sink dependon the transformer, the reservoir capaci-tor and the maximum permissible short-circuit current.In most cases, it is possible to fit T2 on theheat sink of the regulator.

ConstructionThe limiter described above is, of course, in-tended for positive voltages. One for negativevoltages operates in a similar manner, butthe polarity of some components must be re-versed.

Whichever version is needed, it is bestconstructed on the appropriate printed -cir-cuit board shown in Fig. 4. The differencesbetween the two versions are indicated byan asterisk (*).

It is recommended to 'strengthen' the broadtracks by soldering thick, bare wire on them.

The completed unit is, as shown in Fig. 3,interposed between the rectifier and the reser-voir capacitor. Short, heavy-duty wiring shouldbe used.


DIGITAL FUNCTION GENERATOR PAWby T. Giffard

rriONLY connection between the PLLboard and the discretely designed ana-

logue parts of the generator is via plugs andsockets K2, K5, and Kg. Pin 1 of these connectorscarries the32 Hz-3.2MHz clock, which is con-verted in the various sections to a 1 Hz-1001cHzsinusoidal, rectangular or triangular signal.The o ther pins carry information about the se-lected decadic range.

Sine -wave converterThe conversion of the rectangular outputsignal of the PLL loop into a sinusoidal wave-form takes place in three stages-see Fig. 17.The first of these is a digital -to -analogue(D -A) converter with serial input, IC26-IC28;the second, a filter section that removes anyunwanted harmonic frequencies; and thethird, an output stage based on IC40.

The D -A converter consists of two cas-caded 8 -bit shift registers,IC27and IC28, whose16 outputs are fed via a single opamp, IC 4,

to the filter section.Before the first pulse arrives at the clock

inputs of IC27-IC28, all 16 outputs of the reg-isters are low, whereas inputs B are high.Input A of IC28 is low, but input A of IC27 ishigh, since the (low) level at pin 13 (QH) ofIC28 is inverted by IC26d. Therefore, whenthe first dock pulse is received, output QA

of IC27 goes high; on the second dock pulse,QB goes high (while QA remains high) andso on. When, on the eighth pulse, QH goeshigh, input A of IC28 becomes high, so that theoutputs of this register also go high consec-utively (QA-QH). After 16 dock pulses, alloutputs of both registers are high, whereupon

input A of ICY goes low, as explained ear-lier. The outputs then go low in successionuntil after a further 16 clock pulses the pro-cess repeats itself.

The high level at the outputs is 6 V andthe low level is 0 V (earth). The values of theseries resistors at the outputs have been cho-

Iry

VT

111.

me®m®m

if\/c) 0 e`e/0>ep..V6A___)))

ee°0 © 0 0-0 0

0000 0000*0000p0

1-ccoore

e=.9

*5\ d\l! \ eqi, tre-a

ia1\100 Origg 00t oo otpan

Mifin i-AMM dOnn=9e.8=8= 8 els 4 0 8 e 0$-8

Fig. 13. Third overlay of the PCB for the analogue section of the function generator(belongs to Fig. 10)

Fig. 14. Third overlay of the PCB for the digital section of the function generator (belongs to Fig. 7)

E I .i..K-roR ELECTRONICS NOVEMBER 1991

DIGITAL FUNCTION GENERATOR - PART 2

Fig. 15. Illustrating the two methods discussed to build up a sine wave from stepped voltages.

Fig. 16. Illustrating the fundamental operation of the digital -to -analogue converter.

Table 1

Data for computing the resistance networks in the digital -to -analogue converter

ii U sine sing-sina_i R

-5.625 -0.09800.1960 48 621.23

9 5.625 0.09800.1923 49 573.78

I (I 16.875 0.29030.1811 52 627.25

H 28.125 0.47140.1630 58 476.26

12 39.375 0.63440.1386 68 760.81

1 3 50.625 0.77300.1089 87 515.91

1.1 61.875 0.881190.0750 127 053.40

1 5 73.125 0.95690.0382 249 224.20

16 84.375 0.99520

17 95.625 0.9952

sen to ensure that the total current throughthe resistors is translated into a stepped out

voltage as shown in Fig. 15. The compu-tation of the resistance values is not simple,since more than one solution is possible.

The first method is also the most obvi-ous: divide the sine wave into 32 equal parts,so that maximum and minimum values forthe harmonic function are obtained. Thatmeans that 17 discrete values are required asshown in Fig. 15a. Fourier analysis shows thatthis method results in relatively high-levelsecond and third harmonics.

An alternative method, used in the presentdesign, is shown diagrammatically in Fig. 15b.It is based on the symmetry of a sine wavearound the rr/2 (90°) line. To the left and theright of this line, the conevonding steps areidentical. Although the minimum and max-imum values of the sine wave are not reachedexactly, only a 16 -step D -A converter is needed.Moreover, although the 31st and 33rd har-monic have roughly the same amplitude asthose arising with the first method, the 2ndand 3rd harmonic are virtually absent.

The fundamental circuit for the secondmethod is shown in Fig. 16. The maximumvoltage is present when all resistors are con-nected to the 6 V source (high) and the min-imum when all are at earth potential (low).In this method, resistors R86 and R102 inFig. 17 are not used.

The computation of the resistance val-ues is begun at a zero crossing (0°) of thesine wave. Then, one half of the resistors isat 6 V and the other half at 0 V. As an example,resistors R78 and R94 will be computed. First,the difference between the actual and the nexthigher value must be ascertained. After thatit is only necessary to calculate the addi-tional current that flows through the nexthigher resistance.

First, the period is divided into 32 equalintervals: 360°/32=11.25°. From Fig. 15b, itis evident that the level of the output volt-age between 0° and 11.25° must be the same


24 TEST & MEASUREMENT

as the instantaneous value at 5.625°. Next,the sine of the angles an=5.625° and an4=-5.625°is computed and the difference sina-sinaAestablished. That difference is in direct pro-portion to the value of the resistance, that is,R7ri-R94. In the above, n is the harmonic -number.

The resistance is calculated in the man-ner used for an inverting (vamp.

Rn= RGUref/ (sina-sina_4).

In the denominator, the value 0.196 (as givenin Table 1) is substituted and multiplied by1.4 (to obtain the r.m.s. value). Thus,

RTh+R94=2224x6/0.196x1.4=48 621.23 f2.

This value is roughly equal to preferred (E96)resistors of 47.5 kfl and 1.1 kS1 in series. Thesmall difference ofjust over 20 f2 is useful sincethe output resistance of the shift registers is20-30 Q. The higher resistance value (0.1%!)must be just below the nearest preferred valuein the series. Table 1 gives the resistance val-ues from the zero crossing up to the peak ofthe sine wave.

Since the sine wave is not only symmet-rical around the s / 21ine,but also with respectto the x-axis, the steps (1-8) in the negativehalf of the sine wave are identical to the cor-responding ones in the positive half. It is, there-fore, only necessary to compute the values fora quarter wave (n/2)-see the values of theappropriate resistors in Fig. 17. Note thatthe arrangement ensures that the minimumand maximum values of the output voltageremain constant for the duration of two dockpulses.

Operational amplifier IC34 not only con-verts the current into a corresponding volt-age, but also ensures that the 'zero' line is1.2-1.8 V (set with P1) above earth potential,so that a dean, symmetrical sine wave withan r.m.s. value of about 1 V is obtained.The amplifier is a BiFET Type AD711, whichis characterized by a slew rate of 16 Vflts,low off -set voltage, low drift, good bandwidth -gain product (3 MHz at unity gain) and aboveall by an excellent transient response of ±0.001%in 1µs. Capacitors C64 and C62 remove anyresidual r.f. components from the signal andso ensure phase stability of the amplifier.

Locked filterThe filter consists of ten fourth -order low-passsections, of which two are used for each decadicrange. The discrete design offers real advan-tages as far as the harmonic spectrum is con-cerned over a simple switched filter with di-rect input of the clock pulses. The RC elementsare switched on and off by analogue multi-plexers Type 7411C4053, which have a verylow transfer resistance. Each of the filter sec-tions is buffered by a Type AD712 opamp.

The filter sections are fed with signalsD1 -D4 and the output of level comparatorIC17. It can be seen from Fig. 18 that the rel-evant section is active only when the corre-sponding decadic range D1 -D4 is selected.The other sections are then short-circuited

O EN

(!) ill, t6 jut In () it 27

rIej:r®Ien:nr°1°"41®®17261°Cj:

1e26 IC30 1C31 icn IC33

irCOn TOOP TrOo T006

6V

6V

Ci)

9

IC27

74HC

cu, 164

OA

CID

GC

OD

OE

OF

aaOH

R71

6

12

13

C60

OA

A IC28 c'sB OC

OD74

OEHC

C0( 164 OFCG

Q A OH

3

A

MI MIRnCM MIR1MI MI07s

CM I=R76

ITEMnMI MI71

1273370

MI OM

0

11

K5

2 0 00.63

10

12

IC26d

00-1'0 0 7 =1177

0 0-T 10166

0.1.

as

15V

15V

2

12

13

10 96

EMI EMIR111

1=1 MIMI MICM ram

8100

UM MI111:1

Rt01

8102

*se. PA

mom pitimC16I C165 C1114167

700.f 760T00T00o

1.

6V

P1

205

RIOS

15V

*0711

C63 C62Elm

ICOA 10p

IC49

8173 7815 156

8160

1050

C117 a) du.1C27 IC29 1C26

mon 0

15V

Dl

02Dr

015V

Fig. 17. Circuit diagram of the sine wave converter.


DIGITAL FUNCTION GENERATOR - PART 2 El

R105 11127

8110

CH

0111

Cm0112 5H390a

1504

12

13

2

5

C69

620n

C7I

XO

xiIC29

II I

21

74HC

4053

INN

A B C

I

V

z

14

4

9

3

S

IC35a

IC35b

R114 Ril5

C22 C73

C7474

R116 ..R112

R119

12

1C20a

111

0120 141121

154

C75

543

C77

tme

C79

33n

DI.

12 20133t 2

ryo IC30

r5 zo 74

21 41!.1f3

6 201

B C

3IC36a

IC36b

..01/ //1/1., /WM/I ///7/ Z.472i1

4

2

15

IC37a

IC37b

9137

5132

5134

CM C29

1C041

90

5133

RI

115

C91 12

13

1

5

C92 93

112a

RI36 7

02

3

C95

X0

XI X

IC32

Y1

20 74HC

21 4053z

001

A B C

2

14

is

1149

6

IC38a

IC3Bb

5138 5139

C97

sop

5140

C911 C99

RI5012

5142 11143

C1110 C101

100P

122 C103

5I44 45290 330p

2

S

3

50

"0 IC332

74HC

Z1 4053

Doi

A B C

IS

J11 f0 9

31C39a

IC39b

2

3

1C26 - 74HC132IC35 ...IC39 - AD712IC40 = HE5532

15V

15V

15V

0

15V

IC26a IC26b IC264

(I-) IDA (ID ICIOS i102IC35 IC36 IC37

1002 1000 00e,

C109

10044

J) J 116 j) 1J112 'b 114IC38 IC39 1C40

1000 COn 1004

15V

15V

5147

148

=NI

910777 11


TEST & MEASUREMENT

Fig. 18. Illustrating the fundamental operation of one of the five two -stage filter sections.

by switch A. The section for the lowest rangeis, however, always in circuit and, therefore,doesnotneed signal DO. The other two switchedinputs of the analogue circuit are strappedtogether.

When the ratio is lower than 3 200 (thatis, when the frequency of the output signalis at the lower side of the decadic range), RCnetworks with different cut-off frequenciesare switched in. In that way, the filter has anown frequency limit for the ranges 1-410and 410-10 (I10=3.2).

The main function of the filter is attenu-ating the harmonics. Theoretically, only the31st and 33rd harmonics are significant. The31st is 30 dB below the fundamental and isattenuated by a further 70 dB by the filter, sothat ultimately it is -100 dB with respect to theoutput sine wave.

Unfortunately, the cut-off frequencies ofthe filter have to be calculated for the lowestsignal frequency of the decadic range. Thismeans that at higher frequencies the funda-mental frequency may be attenuated by about0.7 dB, owing to an inherent property of theButterworth design used. This attenuationmay be accentuated by the tolerances of thecomponents. The top of the filter character-istic, therefore, has a ripple 0.7 dB). Thismay be remedied by using higher cut-off fre-quencies, which increases the distortion, orusing a different filter design, for example,Chebishev with 0.1 dB ripple, which requireshigh -tolerance (that is, more expensive) com-ponents. Table 2 gives component values fora Butterworth filter with higher cut-off fre-quencies, which give an attenuation of the 31stharmonic of only 50 dB instead of 70 dB, butwhich guarantee a ripple of ..0.05 dB.

The output stage, IC40, uses the well-knownType NE5532 dual opamp. One part, IC40a,functions as a simple impedance converterto decouple the filter. The other part, IC40b,operates as an inverting amplifier. The off-

set voltage can be set between -4.75 V and-1-4.75 V with P2.

The amplitude of the sine wave output canbe set to 0-1 V with P3. If a higher outputlevel is required, P3 can be replaced by a2.2 kfl type. The value of R146 should not bereduced. Also, at higher amplification, it maybe necessary to limit the off -set range by in-creasing the values of R149 and R190.

Rectangular and triangularwaveform converterThe rectangular/ triangular converter, whosecircuit diagram is given in Fig. 19, is con-trolled by the clock from the sine wave con-verter that enters the circuit via pin 4 of Kg.That connecter also feeds signals D0 -D4 (con -

taming information on the selected decadicrange) to the converter.

The clock signal is applied to the non -in-verting input of opamp IC42b, which trans-forms the asymmetrical 0-6 V to a symmet-rical rectangular signal (±15 V).

Circuit IC42b is one section of a dual op-erational amplifier Type OP260. This is arather special device that operates with neg-ative current feedback -see Fig. 20. The in-verting 'input' of this section is, in reality,the low -resistance output of a buffer ampli-fier, whose input is connected to the non -in-verting input of the opamp. If the voltage atthis input rises, the current through R1 in-creases. Since the current drives an impedanceconverter, the output voltage also rises (in pro-portion to the increase in current). Consequently,

Table 2

Component values for higher cut-off frequencies

R106 = 1001d2 R107 = 113 kfl C64 = 680 nF C65 = 100 nFR108=31.6 kf2 R109 = 35.7 kil C66 = 680 nF C67 = 100 nI7R110 = 37.4 kn RIII = 38.3 la2 C68 = 820 nF C69 = 680 nFR112 = 14.3 kil R113 = 14.71S2 C70 = 680 rIF C71 = 560 nFR114 =1001d2 R115=113 kf2 C72=68 nF C73 = 10 nFR116=31.6 kil R117=35.71:0 C74 = 68 nF C75 = 10 nFR118 =37.4 IQ R119 =38.31d2 C76 = 82 nF C77 = 68 tiFR120 = 14.3 kfl R121 = 14.7 IS/ C78 = 68 nF C79 = 56 nFR122 = 1001;12 R123=113 kS2 C80 = 6.8 nF C8I = 1 nFR124=31.6 W. R125 = 35.7 kil C82 = 6.8 nF C83 = 1 nFR126 = 37.4 ki2 R127 = 38.3 Ic.(2 C84 = 8.2 nF C85 = 6.8 riFR128 = 14.3 1:12 R129 = 14.7 kil C86 = 6.8 nF C87 = 5.6 nFR130 = 10.0112 R131 = 11.3 Ica C88 = 6.8 nF C89 = 1 nFR132=31.6 kil R133 = 35.7 ki2 C90 = 680 nF C91 = 100 pFR134 = 3.74 ISI R135 = 3.83 ISI C92 = 8.2 nF C93 = 6.8 nFR136 = 14.3 1:12 R137 = 14.7 kil C94 = 680 pF C95 = 560 pFR138 = 10.0 ISI R139 =11.3 Lc/ C96 = 680 pF C97 = 100 pFR140 = 3.09 kil R141 = 3.40 Ica C98 = 680 pF C99 = 100 pFR142 = 3.74 Id2 R143 = 3.83 ki2 C100=820 pF C101 = 680 pFR144 = 1.43 ki2 R145 = 1.50 k52 C102 = 680 pF C103 = 560 pF


DIGITAL FUNCTION GENERATOR - PART 2 Elhe current through feedback resistor R2 also

increases and this will result in an equilibriumin the current through the buffer amplifier.The OP260 thus operates with a very smallbuffer current. The design has the importantadvantage over conventional opamps of avery fast slew rate: >1000 Wits!

However, the operation is not entirelysymmetrical as far as slope and amplitude areconcerned. This is remedied by P4 -R152, whichmake the signal symmetrical, and R167 inconjunction with Dis-D2i which eqi al i7es thenegative and positive halves of the signal. This

arrangement ensures that the high slew rateis retained at very low signal levels.

The rectangular signal is available at relaycontact ref.

The rectangular signal at the output of ICSis applied to operational transconductanceamplifier (OTA) IC41a via RIM. This ampli-fier has a slew rate of 125 Vitts; its operatingpoint is set with P6. Because- of the currentsource at its output (which sets an OTA apartfrom a 'normal' opamp), the OTA functionsas a variable resistance. This means that inconjunction with a capacitor, C146 -C149, it is

readily made into an integrator. The four ca-pacitors each belong to a different decadicrange; they are selected by two analogue in-tegrated switches, IC45 and IC46. CapacitorC146is connected between the OTA and bufferamplifier IC42a when the DO or Dl line is ac-tive; C147 when D2 is active; C148 when D3 isactive; and C149 when D4 is active.

The use of a switch at either end of thecapacitors is necessary, since the transfer re-sistance of the dosed switch is not linear, butdependent on the supply voltage and thecurrent through the switch_ Switch IC45 is in

15V

1135 j)

t

(D IcIC42 IC43

15V

8154

[la8155

15V

D23 022

350u

21BATES

KI1 1H414602414175

o

.12_0 o-

ti

ti

2i 15Y(12YA

-41--lw

457 405,414

12 13 T1112

141147 Ci144 C149

1 744347n T. 7.

IA 29 NIIC2C

3C 74HC

4 4316

15V

RIS2

8181

K10

SAWS'

14155

IC42

D17

:21

BA%5 X

.7517;112 6V

LI

lc

8170

015

31,3

4051aW

C140 CI41

ISCek 23V

P7 DC -OFFSET

LEMAll V23101-40405- BIM

IC41 CA3280IC42 - 0P260IC43 = 11082

Rue 1411

I 5V

R1S5

8171

15V

51D027 - - 12

rbia/N

MIM

Fig. 19. Circuit diagram of the rectangular/triangular waveform converter.

E I.E K TOR ELECTRONICS NOVEMBER 1991


series with the relevant capacitor and thecurrent cource in the OTA. The current intothe capacitor is independent of the transferresistance of the switch. As far as the follow-ing opamp is concerned, only the potentialacross the capacitor is important. The trans-fer resistance of IC46 is in series with the veryhigh input resistance of IC42, and, therefore,has no relevance.

The signal at the output of IC42a has a vir-tually perfect triangular waveform.

The two zener diodes, D16 and D17, limitthe output level of the OTA to 5.3 V (4.7 Vzener voltage plus 0.6 V forward bias), sothat the signal level at the switches cannotexceed the supply voltage to the ICs. Thezener voltage can, if required, be chosen lowerso as to reduce the regulating time of theOTA even further. An additional effect ofthe diodes is that, in combination with para-

CURRENT FEEDBACK OP AMP

RlAvAI

CURRENT CONTROLLEDVOLTAGE SOURCE

910077 -II 1.-

sitic capacitances, they make necessary a re-duction in the value of 0149 from the calcu-lated 100 pF to 47 pF.

A frequent problem with integrators is theirinability to work with very small d.c. com-ponent This is obviated by the d.c. feedbackloop, R166 -C145, between IC, andand the OTA.This has the disadvantage, however, that thereis no triangular signal available in the low-est decadic range of 1-10 Hz. (Our proto-type delivered a triangular signal down to3 Hz, but it was neither stable nor dean).

The level at the inverting input of IC.tia

may be increased by a d.c. component derivedfrom Ri64-P5-9165 to minimize an off -setvoltage (of the order of a few millivolts),which, owing to the input current's depen-dence on the control current, may arise, par-ticularly at the inverting input

Amplitude control is provided by dualopamp IC43, which is connected as a simpleinverter. Its input and output signals (in anti -phase!) are combined by D13 and D14 andthe rectified output is smoothed by C143. Thiscapacitor is charged via R158 and R159 to thepeak value of the triangular signal. The higher

Fig. 20. Basic circuit of an opamp with Fig. 21. Block schematic of the triangular -wave shaper. integrator and amplitudecurrent feedback. regulation.

0

++

0 r)

20 000000141000000013

S K5 ciie0{R18510 0011001R184}0air R183/0001821001R18010

z-,00110

Of R86 10 Of R 102)0 R 12210

01 R63 10 Of R99 Oc800il-0o QO01 R84 )0 OFR100/0 C82 01100 0100

- c 1 03i114) CH01

1

(*)0

01 R79 }00i 89510 01R 12410PIR 12510Of R85)0 0 01R129)0 r,-0{ R96 cesoffoe ohvcru C109\Of RE31 pot 997 10 C84 011042 00110 160{ 992 100) R98 }A 0 OLF4 27}Ocui c 124

C106Of 978 }00i R94 10 01R1141001R115 00110 00110

1CFY:cr)Cu

O "

IMO

0000

011-0.0

019130)00C8E101100 001

cg°C1100 01100th CO{R133}00

C125

89°°1

C940110- 0{1009 2 01100 0'0110

11

"2+1 r201R134)001R1351.010% (C127 u C112

caw

01913000C35011-00 0109e01140 so&01914 01001R1 4000

0107C1020110C1030110

® ® o"1C132 C130 0 1000{9179}0 .13

000

000

of R77 001 R93 1Oc72011-041 0-01Of R76 }001 R92 }0074011.0 ono* Nu,

P75 O0{ R91 }0 0Of R71 )00i R87 )0 0 011121100{ 972 )001 R88 JO 078 0110* 01,0001 R73 JO Of R89 )007601/00 G0-11001 R74 100[ R90 )0 O{R118}00fR119)0°114

C460.583

CC c13

g.cr,

. 063 Pi0162

R70

FeHDC104

C1 22 .\C1000110-0 01100119142)0 o ung C128

10 CH** 011000115

0120

aa

0Na0

011-00c o5

0"ErT

I

O 0 07 0000 030O 0(-3 0 O h. CO

CT CZ

[-C

P200110 C114

KS P30 0

Fig. 22a. Printed -circuit board (component side) for the rectangular/triangular converter.


DIGITAL FUNCTION GENERATOR - PART 2 29

the potential across C143, the smaller theoutput voltage (=control voltage for the OTA).Since IC43b tends to keep the voltage at itsinverting input, and thus that across C143,equal to that at its non -inverting input, thesetting of P9 will ultimately determine thepeak value of the triangular signal.

A disadvantage of this design is that theregulating time, that is, the time span beforethe amplitude control operates correctly again

after a frequency change, is relatively long,because the time constant of the rectifier cir-cuit has to be sufficiently long even at the low-est frequency. Since the regulating time alsodepends, to some degree, on the setting ofPg, calibrating the two controls properly takesa little practice. This will be reverted to inthe next instalment of the article.

The output of buffer IC42a, like that of am-plifier IC421,, is applied to relay Rel. Which

0 0000 p©© 000

of the two signals, rectangular or triangular,is fed to output stage IC40, depends on thesetting of switch S8.

The use of a relay obviates any problemswith feedback of the rectangular/triangularsignal to the sine wave signal (since bothmay be connected via one cable to the change-over switch on the front panel). One cable can,therefore, be used to take the signal to andfrom preset P8, which sets the peak value of

0000000000000000000000000000000

00co 0 0r=

Fig. 22b. Printed -circuit board (track side and third overlay) for the rectangular/triangular converter.



20000000141 0000000 13

0 uu(Joc135 100001C43

1000001P115210

P9

le

0_0

1:23

cnb.

OiR155i0:3 00 01R15610 9

C144 0{R163[00130-14-0

0 0C138 0-14-0

C143 D14

tp

fi

IR1581

+91

e V

O

I

C1550

IO0O

1C4700

K9

U

0 1:SiR169100 .0

o 0{1-0C137

CC

C140 C142

C141

C15goi

C15200:1171}0m011-0-00001R1 7410 01

0K70J.

B2 0

K10

Fig. 23a. Printed -circuit board (component layout) for the sine wave converter.

Resistors:R70. R160. R161 = 1 McR71. R85 = 2491(0, 0.1%R72. R84 = 127 kit, 0.1%R73. R83 = 86.6 kit, 0.1%R74. R82 = 68.1 kf2, 0.1%R75. R81 = 57.6 kit. 0.1%R76. R80 = 52.3 kf2, 0.1%R77. R79 = 48.7 k52, 0.1%

PARTS LIST

R168 = 82 kt),R169 = 56 kitR170. R176. R177. R181 = 8.2 ki2R173, R174 = 33 kflR178, R179 = 4.7 f2R180 = 10 f2

C132 = 47 uF. 10 V, radialC137 = see textC139, 0141, C156, C157 = 10 pF. 25 VC143 = 1.5 pFC145 = 220 pF. 16 VC150-153 = 47 nF. ceramicC154. C155 = 2200 pF. 40 VC160, C161 = 100 uF. 40 VC162, C163 = 10 uF. 25 V. radial

R78 = 47.5 kit. 0.1%R86, R102 = see textR87, R101 = 196 f2, 1%R88. R100 = 25.5 12, 1%

P1. P4. P9 = preset, 2.7 kitP2. P7 = potentiometer, 5 k.0. linearP3 = potentiometer, 1 kit, log.P5 = preset. 10 k52

Semiconductors:012. D25, D26 = zener, 6.2 V. 400 mW013, D14. D22: D23 = BAT85

R89. R99 = 887 52, 1% P6 = preset, 250 kf2 D15 = zener, 3.3 V, 400 mVIR90. R98 = 634 52, P8 = potentiometer. 1 kit, linear D16. D17 = zener. 4.7 V. 400 mWR91, R93, R95. R97 = 845 a 1% 018-21. D24 = 1N4148R92, R96 = 301 12, 1% Capacitors: B2 = B8001500R94 = 1.1 la 1% C60 = 220 nF 1026 = 74HC132R103, R156, R157 = 2.2 kS1 C61. C149 = 47 pF, polyester IC27, IC28 = 74HC164R104 = 33 kr.1 C62 = 10 pF. polyester IC29-33 = 74HC4053R105 = 5.6 kcl C63. C72. C74, C76, C104-115, IC34 = AD711R106. R107. R114, R115, R122, C119-129. C146 = 100 nF IC35-39 = AD712

R123 = 100 kf2, 1% C64. C66, C68. C116, C142, C144 = 1 uF IC40 = NE5532R108, R109. R116, R117, R124, 125, C65, C67 = 150 nF IC41 = CA3280

R132, R133 = 31.6 kC2. 1% C69 = 820 nF IC42 = 0P260R110, R11, R118, R119, R126, C70 = 390 nF 1043 = TL082 or TL072

R127 = 43.2 ki.1, 1% C71 = 330 nF IC44 = OP64R112, R113, R120, R121. R128. R129. C73. C75 = 15 nF IC45. IC46 = 74H04316

R136. R137 = 34 kit, C77 = 82 nF IC47, 1049 = 7815R130. R131. R138, R139 =10 kfl 1% C78 = 39 nF IC48, IC50 = 7915R134, R135, R142, R143 = 4.32 k i, 1% C79 = 33 nF IC51 = 7806R140. R141 = 3.16 l<12, 1% C80. C82. C84. C88, C92, C147 = 10 nFR144, R145 = 3.4 ki2. 1% 081. C83. C89 = 1.5 nF Miscellaneous:R146, R158, R159. R182-185 = 1 k0 C85. 093 = 8.2 nF K4. K9. K10 = 3 -way PCB connectorsR147. R148 = 1.2 kit C86 = 3.9 nF K5. K8 = 14 -way D -type plugR149. R150 = 15 kfl C87 = 3.3 nF K6. K7 = BNC socket. insulatedR151, R155 = 2.49 kit, 1% C90, C96. C98. C100. C148 = 1 nF Re1 = 12 V relay. 1 change -over contactR152, R171, R172 = 100 S2 C91. C97. C99 = 150 pF. polyster S8 = miniature onioff switchR153 = 330 fl 094. C102 = 390 pF, polyester Tr2 = mains transformer. 2x15 V. 12 VAR154. R175 = 47 kS2 C95. C103 = 330 pF, polyester Heat sink for IC44 (e.g., Fischer ICK14,16B)R162 = 101(0 C101 = 820 pF, polyester [Dau UK Ltd, phone (0243) 553 031]R163. R166= 100 Id/ C117. C118, 0133-136, C138, C140, 0158. Enclosure, Telet LC970 (C -I Electronics. p.51R164, R165 = 39162 C159. C164-167 = 100 nF, ceramic PCB 910077-3R167 = 2.7 kfl C130, C131 = 47 uF. 10 V. tantalum PCB 910077-4


DIGITAL FUNCTION GENERATOR - PART 2

0 0 00

0

0

0 00 no 0

o00000010

ooN

e 0

0 0 0 Li)

Lo

0 0 0 0

0 g

0 s

68BAM0000000

6 O e 8 840(0 0 0 0000 t I) 0

0900 BA e0 0

00

0

800 0 0 au

0

0 00

O 00

0 0 0 0 0

B)80 0000

O 0

0 0

0000

$BB88138800 0000

BBB 90e 0 0 0

O 0 0 G-68 0 0 0 0 0001

0 0

:N=70 0 00

0 0 0 6

0

(D/0

Fig. 23b. Printed -circuit board (track side and third overlay) for the sine wave converter.

the output.The signal is amplified once more in IC44.

This opamp can provide a relatively high out-put currentof±80mA at a slew rate of 130 Vips,so that no difficulties arise with loads downto about 150 LI Since Rinand R1n fix theoutput resistance at 50 LI, the minimum loadresistance may be as low as 100 LI. Note that

the OP64 is not protected against short circuitsand must be fitted on to a special heat sink(see parts list).

The OP64 tends to be overdriven by arectangular signal. Network R169 --C137 in thefeedback loop, therefore, reduces the ampli-fication at high frequencies (that is, the 'cor-ners' of the rectangular signal) slightly.

Capacitor C137 consists of two pieces ofenamelled copper wire twisted together andcut to size as required.

The d.c. component of the output volt-age, ±12 V, is set with P7.

The third and final instalment of this articlewill appear in our Special Large Christmas issue.


PRODUCEMODLLILby R.G. Evans

NG

'EVER since I made my fret circuit (a crystalIL radio) some twenty years ago, I have facedthe problem of packaging the finished circuit.The methods I will describe can sensibly replacethe aluminium or diecast boxes, the 'projeCtboxes' sold by many of the mail order outlets, andat the same time provide very realistic casestailor-made to your projects requirements, look-ing just like (or sometimes better) than the 'realthing'. It is perhaps worth mentioning that I teachDesign & Technology in a secondary school(ages 11-18), and that my colleagues and I usethis technique with our pupils aged 14 and above.The outstanding results obtained by these begin-ners should spur many of the readers of thismagazine on to have a go themselves.

Basic tools and materialsI mentioned that I work in a school which hasworkshops and power tools, but these facilitiesare by no means necessary. Most electronics en-thusiasts will have the tools required for makingtheir own project box: A `stanley' type knife with a scoring blade (or

grind your own from a blunt ordinary knife).If this is not available, a suitable tool can bea made from an old hacksaw blade suitablyshaped and given a simple handle.

A try square of any shape or size. A sanding board - any material that is flat

will do, e.g., a piece of kitchen worktop,blockboard etc. It must be flat. To this glue awhole sheet of 120 grit 'wet or dry' siliconcarbide paper, and leave it weighted down todry.

A workbench (kitchen table??) and, itpossible, some form of vice,. e.g., a Work-mate or a wood or metal vice.

A selection of files, or at least one (fairlycoarse and fairly fine will do).

A tenon -saw, although with care an ordinaryhacksaw will suffice.

Useful, but by no means essential: A wood plane of any type, a small bandsaw

and a sanding wheel (photo 1 shows a usefulassemblage of the above).

Hopefully, the reader will have at least some ofthe above in his or her possession. The next thingto do is to acquire the essentials from which tomake the boxes. Polystyrene is the type of plasticto get hold of (not the expanded ceiling tile stuff)Any colour will do, and two thicknesses areneeded: 1.0 mm and 3.0 mm. Combinations ofthese thicknesses will cover most of our needs.The other main requirement is a suitable solventwith the name dichloromethane, also known asmethylene chloride. This is available from la-boratory chemical suppliers, some model shops Fig. 9. Using a try -square to check the box(under the name of Plastic Weld), and from EMA shape.

cleanly over the edge of a bench.

Fig. 1. Some of the basic tools you will Fig. 2. The pump dispenser. a scoringneed. knife and some plastic sheet.

Fig. 3. Showing sensible planning of your Fig. 4. Showing correct use of a safetywork before you start. ruler, scoring knife and the plastic sheet.

Fig. 5. Showing how to 'snap' the plastic Fig. 6. Showing the removal of the protective plastic film before gluing.

Zi11111."'hor.yea

Fig. 7. Edges may be squared up using a Fig. 8. Edges may be squared up on aconventional wood plane. sanding board with equal ease.

Fig. 10. Adding solvent to each sidepossible) of the joint.

(if


PRODUCT MODELLING

Fig. 11. Box with two sides, also showingslightly oversized end pieces.

Fig. 12. Using a sanding board to squareup the box ends.

Fig. 13. A sanding wheel provides a quick Fig. 14. Gluing the ends on to the box.alternative, but check its angle first!

Fig. 15. Sanding the top of the box ready Fig. 16. After gluing the lid on, use somefor the lid (shown upside down).

Fig. 17. Reducing the edges with a file -always work into the joint.

Fig. 19. Cutting the box on a band saw.Always use a fence!

weights to keep it there for half an hour.

Fig. 18.all sides.

A finished box, sanded smooth on

Fig. 20. With care, a tenon -saw can beused for the same task.

Model Supplies. The latter also supply a very neatpump dispenser which makes the use of the sol-vent much easier (see photo 2).

PlanningHaving assembled the tools and materials, it istime to plan your box a little. I teach my students'design based problem solving' which simplymeans that as much research and investigationshould go into the generation of initial ideas aspossible. There are several ways of packagingyour circuit, and your first idea is unlikely to bethe best. There are several things to consider,such as serviceability, replacement of battery, po-sition and accessibility of the controls such ason/off switches, sensitivity adjustment and rangeswitches. It is worth jotting down a few vitalmeasurements at this stage (remember to leaveplenty of room inside -I once had to squeeze abattery in a vice to get the lid on! Never again!),and to sketch a few ideas of the product (photo 3).

Box assemblyOnce a few dimensions have been decided, con-struction can start. Mark out the base of the box(biro or pencil will suffice) on the 3 -mm sheet ofpolystyrene. Cut it by using a ruler and a scoringknife (score to about Vz the plastic's thickness -photo 4). Place the scored line over the edge of abench or table, and apply sharp downward press-ure to both sides. The plastic will snap cleanlyalong the scored line (photo 5).

The secret of making these boxes lies in mak-ing good butt joints, for which a square edge isessential, and a little time spent here will makeyour job easier later on. Note that some poly-styrene sheets have a protective layer of poly-thene on one or both sides. For our application,this layer should be peeled off (photo 6). Next,cut a strip of the required height, and twice thelength of the box. Leave a few millimetres extrafor squaring up. At this stage, only one long sideof each piece will need squaring. Use a woodplane if one is available, or sand on the sandingboard, taking care to keep it at 90° to the board asyou do so (photos 7 and 8). Keep sanding until auniform square edge is obtained.

Next, place this edge on top of the base in therequired position, and hold it up to a set square(photo 9). Using a small paint brush (maximumsize 5 mm), dip the bristles into the solvent(photo 10), and then run the brush along the in-side of the joint. If it is a long joint, replenish thebrush as nececnry. Once the solvent has been ap-plied, capillary action will put it into the jointarea, and there is little to be gained by continuedapplication of the solvent. If both sides of thejoint are accessible, the solvent may be applied toboth sides. Hold the two pieces together while en-suring that the joint is at 90° to the base.

After some 15 to 30 seconds, the joint will becapable of supporting itself, and the other sidecan be glued in a similar manner. Although thereare several proprietary polystyrene cements onthe market, none have the ease of use and speedof bonding offered by dichloromethane. Unlikeother cements, however, dichloromethane ispurely a solvent, and as such it does not have anygap filling properties. It is therefore well worthmaking sure that your butt joints are good fits for


34 GENERAL INTEREST

best results. Gaps can be filled at a later stage, butit is a lot quicker and easier to get it right firsttime.

Leave the two sides and base to dry while youprepare the ends and lid of the box. This is wheresome planning will have paid off. For ease andspeed of construction, it is a lot easier to overlapjoints, and sand down after construction, than toget it all to fit first. The two ends need to overlapboth sides and base (photo 11), so cut out somesuitable pieces. As they will be glued onto theend of the sides and base, there is no need to givethem square edges. Cut a lid out on the samebasis, that is, overlap on both sides and ends.

By now the base and sides will be workable,and now is the time to square up the ends. Thoseof you with a disc sander will find this easy, al-though the plastic does tend to melt more thansand. Those with sanding boards only will haveto carefully sand the ends square, taking care toavoid 'rocking', and consequent rounding of thebox ends. This process can be made easier if (a)you did not leave that much spare in the first place(careful marking out!), and (b) if you drag the boxacross the board towards you, rather than at-tempting to work both ways (photos 12 and 13).

Once each end is square, the end pieces can beglued on as per the sides (photo 14). Here, no setsquare will be needed, and pressure can be ap-plied by hand until the solvent 'bites'. Give bothends a chance to dry as the next stage will putthem under strain!

Gripping the box firmly upside down, sand itbackwards and forwards (and circular) on yoursanding board, until an even top surface is ob-tained (full width of 3 mm all the way round -photo 15). At this point, the lid can be glued onwith liberal application of solvent, as the insidewill not be accessible.

Leave the box weighted down (photo 16), andforget about it for 12 hours or so. Whilst the sol-vent is drying completely, it is worth mentioningwhy the box needs leaving. Although the solventsticks the plastic very quickly, and allows workto proceed, it does nevertheless cause softeningof the plastic in joint areas, and if fine finishingof the box is attempted, 'plucking out' of thissoftened material will result, and there will be adepression left to fill later. Be patient!

Once the box is dry, it can be externally tidiedup. Large amounts of plastic hanging over edgescan be removed by using a sanding wheel, ifavailable, or simply filed away (photo 17).People used to working with harder materials willfind this plastic quick to file and shape. With onlysmall lips remaining (less than I mm), the wholebox can be sanded on the sanding board untileach and every surface has been 'sanded' (includ-ing the middle of plain sheets, photo 18). Thismay seem a long process, but better results areobtained this way.

Back to the planning stage, and how the boxwill be divided. Thought needs to be given to fit-ting the controls, switches, displays, etc. It ispreferable, but not essential, that they are allmounted in one half of the box, but for mainspowered projects it may be preferable to fit thetransformer and circuitry in the base, and haveflying leads to the front panel controls. Planahead now to avoid disasters later!

Most project boxes can be cut around theirheight anywhere between I, and 2,5 the total

Fig. 21. You should now have two similarpieces of box. The heights can differ.

Fig. 22. Carefully sand the cut edges untilthey are free of saw marks.

1

Fig. 23. The thin (1 -mm) 'joining sheets' Fig. 24. All four thin strips fitted andare shown resting in a box.

Fig. 25. Plan and mark out where your con-trols are going to be.

Fig. 27.with a file. then with wet and dry paper.

glued.

Fig. 26. Mounting pillars for PCBs aremade by stacking up small squares.

Radius the box edges. starting Fig. 28. Surface preparation is vital - usesoap. water and wet or dry paper.

Fig. 29. There are two main types of letter- Fig. 30. A variety of effects can being for surface detailing. achieved with relative ease.


PRODUCT MODELLING

Fig. 31. All manner of 'things' will provideextra surface detail.

Fig. 33. A wide range of paper stickers areavailable.

Fig. 35. Small patches showing throughare not a problem - smooth finish is vital.

height of the box. This is easily done with a smallhandsaw, but take care and use a substantial fenceto hold the box against (photo 19). Many boxeshave been spoilt by the saw wandering off, so becareful! Those of you without a bandsaw will findthat a tenon -saw (fairly fine toothed - seephoto 20) will suffice, or failing that, a conven-tional hacksaw will work but may prove difficultto control directionally.

Assuming your box has now been cut into two(photo 21), the cut surfaces are sanded smooth onthe sanding board once again, until any marksfrom the sawing have gone (photo 22). Take careto sand as evenly as possible, otherwise your boxmay develop a list. Using the 1 -mm polystyrenesheet, cut a strip, or strips, of sufficient length andwidth to fit around the inside of the box, from thebase up to nearly the full interior height. Usingthe solvent again, secure the two long sides firstin the deeper part of the box, and then the twoends (photo 23). Squareness and close fits are notso important inside, and gaps for things like swit-ches can be left in the thin strip. Leave all theseto dry for 10 minutes or so. Then reassemble yourbox, and hopefully you should still have a reason -

Fig. 32. Plasticard letters are easily pickedup using the tip of a knife.

Fig. 34. A range of suitable products tohelp finish your product.

Fig. 36. Removing errant Letraset with apiece of masking tape.

ably square 'shape' (photo 24). If there are smalldiscrepancies between the top and the bottom ofthe box, these can be sanded out on the sandingboard before the next stage.

Fitting upWith the basic box finished, the fitting up canbegin. Holes for switches, dials, controls, cablesand the like should be added (photo 25). Insidethe box, battery compartments can be fabricatedquickly by using suitable pieces of plastic andsolvent, as can PCB pillars made from small`piles' of 3 -mm plastic glued on top of each other(photo 26). The PCB and the pillars can then bedrilled to suit suitable self -tappers which happilycut their own thread in the core diameter pilotholes. Please note that unlike acrylic (perspex),polystyrene does not take fine threads particu-larly well, so it is best to stick with self -tappers.

These pillars of material can also be used togood effect if you wish to 'sink' securing screwsbelow the surface. In this way, results very simi-lar to commercial injection moulding can beachieved. Simply build up a suitable thickness of

material in the area that you wish the screw to besunk (probably around 12 mm). Drill a hole ofsufficient depth and diameter to clear the screwhead (allowing for paint and perhaps screwdriversize). Drill a clearance hole right through to allowthe self -tapper to sit in the recess so created.

Next, build up a suitable pilar on the oppositeside of the box, to the required height for the self -tapper to screw into. Take care to get the pillarheight correct, or box distortion can occur. Whilstthis provides a secure method of holding yourbox together, you will often fmd that the thinstrips inside provide more than enough force tohold things together, especially non -critical itemslike small battery -powered test equipment, tor-ches, etc. If all this sounds too much, countersunkself -tappers may be used.

Appearance!With all the fitting and bracketry finished, the ex-terior of the box can now be finally prepared.Radiusing the edges and corners of the box willhelp its appearance, and, if it is hand-held, its'touchability'. Use a fine file with caution, andaim for an even radius all over the box, just totake the sharp edges off (photo 27).

The next stage requires warm soapy water andseveral grades of wet -and -dry silicon carbidepaper. Starting with 220 grit, rub the box all over,using plenty of water to avoid clogging the paper.This is the time to finally blend in the radiusingwork that you may have done earlier. When youare sure that the whole of the box has beencovered by this grade, move on to 320 grit and re-peat the process. At this stage, you will probablybe able to 'feel' the surface through your fingertips better than by visual inspection.

When you are sure the whole box has beencovered by the 320 grit, move on to 400 grit andrepeat the process. Finally, go over the whole boxwith 600 grit, feeling and looking for any scrat-ches left over from coarser papers. If any arefound, remove them gently with the appropriategrade, and then move on to 600 grit (photo 28).

With care you will have got this far, and havea very smooth box. If, however, there are somelarge holes, these can be filled with ordinary P38car body filler or cellulose putty. It is much easier,however, to avoid them in the first place, whichis not as difficult as it may sound. Give the box afmal rinse inside and out in warm to hot water, patit dry with a lint free cloth, and leave it some-where warm and dry for a few hours.

LetteringWhile the box is drying out (especially in the jointareas), some thought needs to be given to any sur-face detailing that you may wish to add. Simplethings will help make your product look moreconvincing, and can make all the difference.Raised lettering can be added to give the im-pression of embossed or moulded -on letters.There are two main types, Plasticard polystyreneletters (usually sold to model railway enthusiasts)in several sizes, and Edding self-adhesive vinylletters. Of the two, the latter are fairly low -profile,but are useful to give a 'makers name' to the pro-duct, as they will be painted over later, and so willform part of the product. The Plasticard letters arehigh -profile, and can be painted over also. The


36 GENERAL INTEREST

tops of the letters can be carefully sanded clear ofpaint to create an additional effect should you sodesire (photos 29 and 30), but their higher profilemeans that they are ideal for identifying import-ant controls such as on/off switches, withoutbeing as obvious as, say, the devices' names'which can be added later in Letraset.

Other features that you can add include'grips' made by gluing (with liberal amounts ofsolvent) strips or pieces of 180/220 wet or drypaper, and spraying over them later. This createsvery effectively the 'moulded -on' appearance.Using coarser grades of paper spoils the effect.Embossed self-adhesive metal film is also avail-able from EMA (details further on), and 'pin-stripe' tape as sold in motor accessory shops canprovide a useful way of breaking up large flat sur-faces. This can either be sprayed over, or addedlater as a colour feature. It does not need much tomake the product look real, so take care not to getcarried away! (photo 31).

To get perfect edging on some of the itemsmentioned may necessitate going back to a finefile and the range of wet -and -dry papers, just toblend in an overhanging edge. This usually pro-duces better results than just cutting a piece outand sticking it on.

The Plasticard polystyrene letters can be cutout and trimmed with a suitable craft knife. Thetip of this knife is the best tool to pick the lettersup and put them in place (photo 32). A dab of sol-vent on their rear immediately prior to placingeach letter produces good results. Positioning theletters dry, and then applying solvent tends toleave a mark, so be careful. Getting all the lettersin a neat line is also important, and a pre -drawnfaint pencil line at 90° to an edge will help a lothere. The vinyl letters mentioned are self-ad-hesive and may be positioned similarly. If you aremaking a 'novelty' product for perhaps a child, itmay be worth a visit to your local craft and sta-tionary shop. These often have rolls and rolls ofsmall paper stickers that can also be added to thesurface of your product, and then sprayed over(photo 33). In this way, a variety of themes couldbe imparted on your model.

PaintingAfter the surface detailing is complete, the nextstage is to paint the object. For this, aerosol paintsas sold by car accessory shops are ideal(photo 34). The two halves of the box should beseparated, and can then be given the first coat ofgrey cellulose primer. A reasonable coveringshould be given by starting the spray to, say, theleft of the object, moving across it and clear to theright before returning. If you wish, you can stopspraying at either end, although some brands ofaerosol tend to 'spit' a little when used in thisfashion. Spraying this way helps prevent unevenpaint buildup and runs! Leave the primer to dryovernight, and then very carefully rub down thesurface with 800 grit wet -and -dry paper usedwith lots of soapy water. It is not necessary to re -expose the plastic underneath, just to achieve asilky smooth surface ready for the top coat (useyour finger tips). If small areas of plastic re -ap-pear, don't worry. So long as the preparation wasgood, the primer does not have a lot to do, and thesmall gaps in the coverage are not a problem(photo 35). When the box has been 'flatted' all

Fig. 37. A selection of finished articles.

over, a fmal rinse in water is needed (inside andout), and then the box can be dried.

The top coat of paint is also aerosol cellulosecar paint, and so there is an almost endless varietyof colours. If you want black, choose Satin black,which shows fingerprints less than either gloss ormatt. Metallic colours are also very effective, butchoose carefully after looking at commercial pro-ducts. Probably the best is Ford 'Graphite', Ford'Pearl Grey', and one or two others, notably theblues.

A useful tip to improve the spray power ofaerosols is to place them in hot water (max. 50°)for a few minutes. This raises the internal press-ure, and helps them spray better for longer. Makesure that the whole aerosol is completely dry, asa single drop of water will play havoc with yoursmooth mirror-like surface!.

If you want an ordinary colour, try and stickwith primary colours rather than sickly mixes.Some aerosols (especially the metallics) are whatis known as '2 -pack', which means that they needa coat of lacquer to give them the final gloss. Ifyou intend to use any Letraset, then this will beneeded anyway, and should not be feared.

After the paint has had time to dry, Letraset orsimilar dry transfer lettering can be added. Giventhat these dry transfer letters sometimes transferthemselves at the wrong time to the wrong place,the following may be useful. Take a piece of or-dinary masking tape, and rub the sticky sideagainst a cotton T-shirt or similar. This will helpreduce its stickiness. Use this treated piece ofmasking tape to provide the baseline for your rub-down lettering. If any do come unstuck, they willprobably be on the masking tape, which can becarefully removed after you have finished. Dotake great care not to press it down too hard, orpull it off too briskly - you may find it brings thepaint with it! If, despite this precaution, the oddletter does appear in the wrong place, or youmake a mistake, do not despair, as individual let-ters can usually be removed by the careful appli-cation of a small piece of masking tape(untreated), wrapped around a finger tip, and'dabbed' at the offending letter (photo 36). If thisfails, it should be possible to remove it with the

aid of a scalpel or craft knife, and a scraping mo-tion, although this method is likely to leave amark.

If you have used a dark colour to spray yourbox, then white or red lettering can be used togreat effect. Controls can be labelled, and someform of product name added. Very often codenumbers at the bottom of the Letraset sheets pro-vide convincing model numbers. It all adds to therealism of the finished article, and will makepeople look twice or even thrice in their attemptat determining the origin of the 'device' they see.

Finally, a coat of clear lacquer needs to be ap-plied. Car accessory shops sell this as clear cellu-lose lacquer. Be careful how much you spray onin one coat. Sometimes the solvent attacks the Le-traset, causing it to wrinkle up, and destroy allyour hard work. Apply just enough to 'flow' intoa gloss (allow it to stand a few seconds before ad-ding more!), and spray it somewhere warm tohelp the solvent evaporate quickly.

Add extra coats as needed until you are happywith the appearance of the finished object(photo 37). All this paint will take quite a whileto harden, depending on the number of layerspresent. I have known some objects (admittedlyheavily laden with lacquer) to take up to a weekto dry sufficiently to not take fingerprint im-pressions when handled. If you get the odd fin-gerprint, it can usually be polished out by using'T -cut' colour restorer (also from your car acces-sory shop). That concludes the description of theprocess - it probably takes as long to read it asit does to make a box!

* EMA Model Supplies, 58-60 The Centre, Fel-than], Middlesex TW13 4BH. Telephone: (081-890) 5270. Fax: (081-890) 5321.


CLASS -A POWER AMPby T. Giffard

The power amplifier presented is a revamped version of theLFA-150 we published in late 1988. It provides rather less powerthan its popular predecessor, but retains its class A quiescent -

current setting up to -3 dB below its clipping limit. The aim of thedesign was to find a good compromise between class A quality

and the dimensions of the necessary heat sinks.

Fig. 1. Prototype of the class A amplifier.

TECHNICAL DATA

Poy.'er output tat 1 kHz)

Music power (500 Hz burst,5 periods on, 5 periods off)

Power bandwidth (48 W into 8 f2)

Slew rate

Signal-to-noise ratio (I W into 8 S2)

Harmonic distortion (25 W into 8 fl)

Intermodulation distortion(50 Hz: 7 kHz; 4:1)

Dynamic IM distortion (rectangularwave 3.15 kHz; sine wave 15 kHz)

Input sensitivity

Input impedance

Quiescent current

48 W into 8 f283 W into 4 f2122 W into 2 K2

50 W into 8 SI90 W into 4 f2150 W into 2 fl

1 Hz -270 kHz (+0 dB, -3 dB)

>45 Vigs

>105 dB (A -weighted)

<0.005% (1 kHzO<0.02% (20 Hz -20 kHz)

<0.03% (25 W into 8 f2)5.0.005% (1 W into 8 12)

<0.015% (25 W into 8 f2)

1.1 V r.m.s.

25 kf2

1.25 A

A LTHOUGHmost of our readers will know.L-1.what is meant by class A, AB or B, itmay be useful for others to recapitulate theproperties of these classes of operation.

Most modern a. f. output amplifiers are de-signed as a push-pull configuration. In this,two power transistors (or valves) are con-nected in series between the positive andnegative power supply lines, and the loud-speakers are connected to their junction. Thetransistors are driven by opposing signals,that is,whertone conducts,theother isswitchedoff.

The transfer characteristic (I b/I c curve)of a transistor is, unfortunately, not a straightline: particularly at its lower end, when thecurrents are low, it is very curved. In princi-ple, the characteristics of two transistors inpush-pull complement one another nicely,butthe base -emitter transfer voltage puts a span-ner in the works.

To prevent distortion, therefore, use ismade of a bias that causes a small constantcurrent to flow through the transistors. Theresult of this is that the curved part of the char-acteristic is not used. If virtually no currentflows through the transistors, class B ampli-fication is obtained, which is characterized bycross -over distortion. Nowadays, this methodis used only in inexpensive, portable equip-ment where economy of battery usage is theprime requirement.

For many years now, class AB operationhas been used in hi-fi equipment. In this, asmall quiescent current flows through the out-put transistors, which makes them work inclass A when the drive is small (say, 0.1 W)and in class B at higher drives. This methodis characterized by acceptable low distor-tionand good efficiency (about75%). Moreover,heat dissipation depends on the energy ofthe drive signal, so that the amplifier remainsfairly cool during normal operation.

Many designers and listeners do not findAB operation good enough and prefer class A.In this, such a large quiescent current flowsthrough the transistors, that even at maximumdrive neither of the transistors switches off (ashappens in dass AB). The disadvantages ofclass A operation are its low efficiency (<50%)and large heat dissipation. Even during qui-escent operation, a class A amplifier usesabout twice its nominal power.Naturally, commercial class A amplifiers are


38 AUDIO & HI-FI

available at many hi-fi retailers, but, althoughtheir superior quality cannot be denied, theyare very expensive, have very large heat sinksor fans to get rid of the dissipated heat, andprovide only a modest power output. Thesefactors make it easy to understand why a25-50 W (lags A amplifier is not exactly themost popular item in the hi-fi trade.

In the present design, we have tried to com-bine class A operation with an acceptable heatdissipation. Output power was specified at50 W into 8 a. In true dass A, that would meana dissipation of rather more than 100 W perchannel. To get rid of more than 200 W ofheat in normal enclosures without the use offans is virtually impossible. Fans are, how-ever, anathema, because even the best onescan be heard, particularly during soft musicpassages.

The design is, therefore, near -class A, thatis, the quiescent current is set to a level thatensures that the amplifier can deliver half itsnominal power to the loudspeakers. The classA range is thus 25 W, which means that inactual use the amplifier operates in class Aup to -3 dB below full drive. The dissipationthen amounts to about 140 W and that canbe got rid of with 'normal' heat sinks (basedon an ambient temperature of not more than35 °C and a maximum heat sink tempera-ture of 75 °C).

It should be noted that a class A amplifieroperates in its class only into its specifiedload impedance, normally 80. It is, of course,possible to provide more power into 4 f2 or252, but that will not be in class A. This pointsto a serious disadvantage of class A amplifiersthat ishardly ever highlighted: into a lower thanspecified load, it operates in class AB. Since theinput impedance of virtually all modern loud-speakers varies between 4 a and 10 C, am-plifier manufacturers have the near impossi-ble task of designing an amplifier that re-mains in class A over this range of impedances.The power required to operate a class A out-put stage into a 4 a load is twice that for op-erating it into an 8 52 load. A 25 W class Aamplifier designed to operate into 8 S2 dissi-pates more than 100 W, but to enable it tooperate in class A into 452, the quiescent cur-rent would have to be increased to a level wherethe dissipation would exceed 200 W.

The real requirement of a top-quality a.f.output amplifier is that it can deliver a rea-sonable power, say, not less than 10 W, intoan impedance varying between 452 and 10 52.That will meet the demands of all hi-fi lis-teners in 95% of the time. The present de-sign provides 25 W into 8 52;12.5 W into 452and rather more than 6 W into 2 a all inclass A. At the other 5% of the time, in classAB, it gives 48 W into 8 0,83 W into 4 a and122 W into 2 a.

The revamped circuitThe renovating of the LFA-150 started witha lowering of the supply voltage from ±56 Vto ±29 V to ensure an appreciable loweringof the dissipation.

Next, a number of transistor were replaced,partly because some of them were no longer

Cermetrnattiturn

Cp.= 34114

Cp.= 31V

Cp.= rrvs

Cl

A 1W5

R2 7%

0 Pe47%77,0,*

o Styrene.

c,2

77- 63V

R251%

TS

BC639am

10w 4011

1114148im C7

I.iV

R29 T1S80139 +2.4. I

10P40V

4W

R15 C

BC560C

719

- Vce=3V3

T7

T9

80140 -44VL_ J

Fig. 2. Circuit diagram of the class A amplifier.


CLASS -A POWER AMPLIFIER - PART 1 39

r- -

L-K2

ERROR LOW IMP POWER ONft ft _tt_

29

Protection Circuit

880092-3

+30Vgf

-30V f.'S

X13

22V I

22V I

20 40 60 80

10305 70

J

p

C54

660n 1000p1009 639

I C571 C55

61110n 1000v100v eav

R106

040

'

039

C52 C53

22n 22n035 D36

037 038C50 C51

22n 22n

I -70V

I 035..-039 = 1N4002- -

K1 40

1 305

e6100

7 90

319"-I

2 Fi

I77mA

5C ,A7

F2 SI

2

2A7

C

V

3.4 ilductacceate

C27 RS9

680,

726

1001

3 Irei

012 2x I

1N4148

729

R71

13C5468R69

1500

670

MIM

loop 40v

LS I

+30V

+30V

L

22V".

C31

41croov v

221/.s,

'N.0

4400DTV

0-30V

0---30V

131= 8100C35000

Is 2

(2:22M411)

910124-11


40 AUDIO & HI -F1

available. Motorola's BF762 and BF759 havebeen replaced by the BF470 and BF469 re-spectively, while Sanken's Types 2SC2922and2SA1216havebeensubstituted for Toslulm's2SC2565 and 2SA1095 output transistors.

Third, the amplification of some stages hasbeen altered. The gain of the first differential

amplifier is now 19 dB instead of 11 dB, whilethat of the second differential amplifier hasbeen reduced by 6 dB. This results in theopen -loop amplification rising from 2300 to3 000, which lowers the distortion to evenlower levels than in the LFA-150.

Finally, the compensation and d.c. oper-

Fig. 3. Printed circuit board for the voltage amplifier.

ating points have been changed at a numberof places.

Because of all these alterations, the val-ues of many components in the original LFA-150have had to be changed. The amplifier is con-structed on four PCBs: two for the outputstages, one for the protection circuit, and one

Resistors:R1 = 1001(12, 1%R2 = 681 n, 1%R3 = 33.2 kfl, 1%R4 = 562 S2, 1%R5 = 15 11R6 = 10 kfl, 1%R7 = 1 kfl, 1%R8 = 953 12, 1%R9. R10 = 18.2 12, 1%R11. R65 = 82 SIR12 = 3.3 k12R13 = 5.6 k1-2R14 = 22 kftR15 = 165 a 1%R16. R17 = lon,1%R18 = 2.2 k12, 1.5 WR19. R22 = 100 12, 1%R20. R21 = 4.7 k12, 1.5 WR23 = 2.2 Id.2R24 = wire bridgeR25 = 3.32k12,R26, R34 = 8.2 kt.).R27 = 12 kS.2R28, R32. R35, R39 = 5.6 kit

PARTS LISTR29, R38 = 1 SI, 1.5 WR30, R37 = 100 SIR31. R36 = 3.9 k12R33 = 6.8 k52R45 = 39 12R46 = 680 12R47 = 270 12R48 = 47 IIR49, R56 = 56 S2R50, R57 = 10 S2R51, R53. R58, R60 = 2.2 12R52, R54, R59, R61 = 0.22 1.-2,

3 W. inductance -freeR55. R62 = 470 a 2.5 WR63. R64 = 4.7 12. 1.5 WR65=8252R66 = 150 12R67 = 10 ki-.1R68 = 270 12R69 = 1501-2R70 = 68 12R71 = 47 k12P1 = 100 12 multiturn preset

P2, P3 = 2.5 k12 presetP4 = 1 kf.-2, multiturn preset

Capacitors:Cl = 4.7 uF, MKTC2 = 680 pF. polyesterC3 = 12 nFC4, C25 = 100 nFC5, C7= 47 tiF. 63 VC6, C11. C17 = 10 uF, 40 VC8 = 2.7 nFC9 = 68 pF. 63 V. polyester010, C15, C16. C21. C26.

C27 = 680 nFC12. C18 = 1 ttF. 63 V013. C19 = 47 nF014. C20 = 100 pFC28 = 120 nFC29, C30 = 100 ttF, 40 V

Semiconductors:D1. D2. D11-13 = 1N4148D3. D4 = zener. 18 V, 400 mlAI

D9. D10 = 1N4002T1. T2 = 2SK146VT3. T4 = BC550CT5 = BC639T6. T7 = BC560C (match!)T8. T9 = BF470T10. T11 = BF469 (match!)T12-14 = BC546BT15, T20 = BD139T16-18 = BC556BT19 = BD140T21 = 2S02238T22 = 2SA968T23. T24 = 2SA1216T25. T26 = 2SC2922

Miscellaneous:Ref = 24 V for PCB mounting

with 1 change -over contactL1 = 6 turns. IID 15 mm, of1.5 mm enamelled copper wireK1 = 10 -way straight headerHeat sink. 0.4 KM (see Part 2)PCB 880092-1PCB 880092-2


CLASS -A POWER AMPLIFIER - PART 1

4'

for the auxiliary power supply,Only an outline description of the circuit

in Fig. 2 will be given-for more details, seeRef. 1.

The input signal is applied to differentialamplifier T1 -T2 via C1 (the only capacitor inthe signal path) and low-pass filter R2 -C2. Thefilter has a cut-off frequency of 340 kHz to pre-vent transient intermodulation distortion (TIM).The d.c. operating point of the two FETs is setby a constant -current source based on T5.

Transistors T3 and T4 and the FETs form acascode circuit that has an amplification ofabout 9.2 (=gain of about 19 dB). Preset P1serves to eliminate any inequalities (off -set)in the d.c. operating points.

Lag compensation is provided by R5 -C3.The amplified signal is fed to differential

amplifier T6 -T7 that together with T8 and T9form a second cascode circuit to keep thebandwidth as large as possible. The amplifi-

cation of this section is 308 (nearly 50 dB).The output of T8 is fed to the current am-

plifier via current mirror T10 -T11.The power supply for the voltage ampli-

fier is taken from discrete voltage regulatorsT12 -T15 (positive) and T16 -T19 (negative). Thesupply voltage for this section is higher thanthat for the current amplifier, so that the volt-age drop across the output transistors re-mains small, even at maximum drive levels.

The current amplifier consists of a quies-cent -current control based on T28, and driversT21 and T22, followed by power transistorsT23, T24, T25 and T26, which are connected ina compound configuration. In contrast to theusual emitter -follower, this design providessome voltage amplification and combines verylow distortion with a low output impedance.The output signal is fed to the loudspeakersvia an inductor and a protection relay.

When the load current exceeds 15 A (peak),

transistors T27 and T30 switch on and actu-ate the protection circuit via T28 and T29. Thatcircuit then de -energizes the relay.

The power supply uses two mains trans-formers in series, Tr1 and Tr2. Note that Fig. 2shows the power supply for a mono ampli-fier; two (housed in one endosure) need tobe built for a stereo amplifier. The supply forthe current amplifier is around ±30 V, and thatfor the voltage amplifier, ±44 V. The voltageamplifier, however, operates from +38.5 V and-35 V, because its voltage drop on the posi-tive line is about 3.5 V higher than that onthe negative line. This supply arrangementensures that the positive and negative drivelevels are exactly equal.

Reference."LFA-150: a fast power amplifier", ElektorElectronics, November 1988, pp. 20-26;December 1988, pp. 42-46.

Fig. 4. Printed circuit board for the current amplifier.


\IEL1NOIZ(S M LJ

The design of attenuators

by Steve Knight, BSc

In this second and final part of our investigation into the designand uses of four -terminal attenuator systems, we shall first of all

derive some simple relationships between the characteristicresistance of an attenuator T -section and the degree of

attenuation provided by the section in terms of the resistiveelements that make up the section.

rrHE calculations are not difficult, and1 those readers with an arithmetical bent

might like to confirm some of the answers.We will relate our results with those of theit-section, consider insertion loss, and havea look at the practical design of attenuators.

Finding valuesFor the properly terminated symmetrical T -section shown in Fig. 8, we let the attenua-tion be expressed as N=U1/U2, in whichnumbered subscripts are used for conve-

Fig. 8. Deriving the attenuation factor ofa section.

nience. Since the input resistance is R0, theinput current, /1, must be 1/1/Ro and the volt-age, U, across the centre shunt arm will be

U=111-114

Hence,

U= /Ro=Ui [ I -R tiRoi

But the output voltage, U2, is clearly theproportion of U developed across R0, givenby

U2= U[R.I(Ri+Ro)],

whence

U2= U1 [ 1-(R11R0)]IR0/(RI-ER0)].

From this, the ratio

1/11U2=N=(Ro+R1)1(Ro-R1).

This gives the attenuation factor in termsof Ro and R1. What we need now is a rela-tionship involving R1 and R2 in terms of Roand N. Working from this last result, we fmdthat

R1=Ro[(N-1)/(N+1)].

Using the earlier result that

R0=1/(R12-i-2RIR2),

we get

16=R0[2N1(N2-1)].

When Ro and N are known, these simpleexpressions enable an attenuator to be de-signed that gives a desired signal reductionand the proper matching conditions for thecircuit system concerned. Figure 9 illus-trates the meaning of the two formulae.

Fig. 9. Relationships to determine theelements in terms of R0 and N.

Trying things outA couple of examples will show how wecan apply these results to solve some ele-mentary design problems.

First, suppose we want a symmetricalT-section network to match into a 300 nline and have a voltage attenuation of 14 dB;what values of resistors do we want?

Well, the attenuation factorAntilog 14/20=antilog 0.7=5.In other words, the output willbe one fifth of the input. We thus get

R1=300[(5 -1)/(5+1)1=300x4/6=200 SI;

and

R,=300[(2x5)/(25-1)1=300x10b4=125 D.

Figure 10 shows the completed section.

3000

2000

4 anon

910103.2-13

Fig. 10. Example of en elementaryT -section design.

Suppose now that the output stage of asmall transmitter has an internal resistanceof 600 fI and is intended to supply currentto a 600 SI load. We need to design a T -sec-tion which, when inserted into the line con-necting generator and load, will reduce theload current to one third of its initial value.

Fig. 11. A further design example relat-ing to correct matching procedures.

This situation can be illustrated after themanner of Fig. 11. The section matches be-tween equal impedances of 600 S2 and should,therefore, have R0=600 fl also. The attenu-ation factor, N, is 3, so that


FOUR -TERMINAL NETWORKS - PART 2

R3=1:10[(N2-1)/2N1 Rb=13.[(N+1)/(N-1)]

Fig. 12. These attenuator sections have the same total series Fig. 13. Finding the element relationships between the T- andand shunt resistances; then. ZoTZ=Ri R2. 7r-sections.

R1=600[(3 -1)/(3+1)].600x2/4=300 f2

and

R2=600[(2x3)/(9-1)1=600x6/8=450 f2.

In a case like this, it is important to no-tice that current It supplied by the generatoris the same whether the section is in circuitor not. In both cases, the generator sees a re-sistive load of 600 fl. With the section con-nected, the 'unwanted' current (2/311) flowsalong the shunt arm, as an application ofOhm's law will immediately verify.

The ic-sectionIt might appear that the it -section has beenrather neglected, but there is a circuit rela-tionship between the T -section and it -sec-tions that makes an analysis of the it -sectionquite easy. Figure 12 shows a ladder net-work of series and shunt resistances; froman examination of this network, is it a chainof T -sections or it -sections? The answer isthat it depends how you look at it. For, ei-ther a T -section can be taken from the net-work by considering the line division AA,or a it -section taken by considering the linedivision BB. The T -section cuts through thecentre of each of the series elements RI,while the it -section effectually splits theshunt element into two parallel parts, eachof value 2R2. Hence, the laddercan be viewedas a string of T -sections or it -sections.

We might expect, therefore, that therewould be little difference between the rela-tionships derived for the two types of sec-tion, but for the matter of equivalence it isn't

just a job of halving the series arms and dou-bling up the shunts as this last analysis mighthave led us to believe.

We can make an exact comparison betweenthe sections by looking at Fig. 13. Here, weuse letter subscripts for the it -section resis-tors to avoid confusion.

For the two networks to be equivalent intheir characteristics, the resistance seen be-tween terminals 1 and 2 of both networks mustbe equal, as must the resistance between ter-minals 1 and 3 of both networks, with theoutput on open circuit. This means that Ri+R2on the 'F -section must equal Rb(Ra+Rb)/(Ra+2Rb)on the it -section. Also,

or

2R1=2RaRb/(Ra+2Rb)

R1=RaRb/(Ra+24).

Substituting back for R1 in the first expres-sion then gives us

121=Rb2/(Ra+2Rb).

These relationships provide us with the el-ements of the T -network such that this willcorrespond to a given it -network.

It is also not too hard to demonstrate thatfor the basic it -section shown in Fig. 13

Ro=RaRbhi(Ra2+2RaRb)=4Rocke

and that

Ra=R0RN2-1)12A9

and

Rb=RoRN+ I)/(N-1)]

where the bracketed term is simply invertedfrom that of the T -section and R, and Rb arethemselves swapped over.

A useful relationship follows from the firstof these: since -4(R12+2RIR1) is the R0 of aT -section (ROT), we can deduce that for sec-tions having the same total series and shuntresistances (see Fig. 12 again)

R0,R0T=R1121=RaRb.

Don't confuse this with the equivalence re-lationships.

So, fundamentally there is no differencein the functions of these alternative forms ofthe sections and either may be used in a par-ticular situation. The actual choice dependsupon which form, given a particular valueof characteristic resistance R, yields themore readily obtainable values of resistancefor the elements.

In general design work, where stringentconditions are not of vital importance, the useof 5% resistors is quite acceptable; the re-sulting variation in attenuation will normallybe no more than 0.5 dB and the mismatch inthe characteristic resistance itself of the orderof 5%. We will return to these points a littlelater on.

Insertion lossThe important thing in designing attenua-tors for general bench use is not to bothertoo much about getting exact answers to thecalculations. It is no use working out a re-sistance to three or four decimal places and

R,-.R.[(ti-1) 2N] IR,,R,,t(114-1) (N-1)]

Fig. 15. The relationships should be com-pared with those for the T -section givenearlier.Fig. 14. Designing an attenuator knowing R, and the required attenuation.


44 GENERAL INTEREST

then discover that you've got to use a pre-ferred value anyway. The attenuation factorcan often be rounded off in the same way;6 dB, for instance, is a voltage ratio of 1.9953...but we wouldn't put this into a formula: weuse the nice round figure of 2. Of course, thingsdon always work out quite so conveniently,but anything beyond two decimal places isa waste of effort.

Let us now design a 5 -section attenuatorfor a 7042 line giving switched positions of3 dB, 6 dB, 12dB and two 20 dB reductions.For interest, let the first three stages be de-rived from it -sections and the last two fromT -sections. The job will look like Fig. 14,which also shows the appropriate switching.

Look first at the 3 -dB section; the re-quired attenuation is 3 dB for which attenu-ation factor N is found to be antilog 3/20 or1.41. Using the relationships shown for con-venience in Fig. 15, and with R0=70 0, wefind that R3=24.8 fl and Rb=410 LI The lastvalue is a bit of a problem, since it falls be-tween 390 S2 and 430 fl in the E24 range.One way out is to parallel 9.1 kfl with a430 S2 or get hold of a 412 f2 from the E96precision range. Othenvise, use 24 f2 and390 f2 in series; the mismatch is not veryserious: R. then works out at 66 52.

Going through the same procedure forthe 6 dB (N=2) and the 12 dB (N=4) It

we find for the elements the respec-tive values: Ra=52 f2; Rb=210 52; Ra=131Rb=116 f2. Practical values here would be51 Q; 220 Q; 130 f2; and 120 fl. Amuseyourself by calculating the above three rangesin terms of T -sections; do the resistance val-ues fit any better?

For the two T -sections, we require an at-tenuation of 20 dB (N=10); hence, for Ro=70we find

RI=RoRN-1)1(N+1)]=70x9111=57

and

16=R0f2NI(N2-1)1=70x20/99=14

Practical values here would be 56 f2 and1352. The completed attenuator is shown inFig. 16.

Now, try designing a series of five T -sec-tions with switched ranges of 1 dB, 2 dB,4 dB,8 dB and 16 dB; this enables a maximum at -

Fig. 16. The completed design.

tenuation of 31 dB to be achieved in 1 dBsteps. The Ro should suit your own particu-lar fancy. As a help, the respective N valuesare 1.12; 1.26; 1.58; 2.45 and 6.30.

Practical considerationsWhen we talk about purely resistive attenu-ators, we are, of course, in the realms of fan-tasy; it is not possible to make up an attenu-ator system having a number of sections intandem without introducing some inductiveand capacitive elements. The object of anydesign is to keep these to their absolute min-imum, just as in filters, where inductanceand capacitance are the necessary elements,we try to eliminate resistance.

It is necessary, then, to keep all intercon-necting leads as short as possible and toavoid the proximity of these leads, as wellas the resistances themselves, to any sur-rounding metal parts. Further, there must beno capacitive coupling between the sectionsor certain frequency components of the sig-nal will sneak through without the desiredattenuation, but with definite phase shifts.The resistances should not, for obvious rea-sons, be wirewound, though types with a non -inductive construction might be used.

Simple aluminium boxes are availablenowadays that can be used to house bench -type attenuators. Types measuring some120-150 mm in length with perhaps 50 nunwidth and a depth of about the same areideal. Figure 17 shows the general methodof assembling such attenuators.

The box is divided up into the requirednumber of compartments by cross screensthat can be made of thin aluminium or tin-

plate. If aluminium is used, the screen willhave to be flanged and screwed to the sidewalls of the box, but with an all tinplate con-struction, soldering is the best approach.

Small holes are drilled centrally in eachscreen (before fitting!) to permit the series -arm resistors to feed through, and the shunt -arm resistors are returned either directly toeach screen or to a convenient position onthe box floor itself. A relatively heavy com-mon wire running through the screens some-times makes dependence on the box metalunnecessary.

Slide or miniature toggle switches are usedfor the attenuation selection, though it ispossible, with certain forms of construc-tion-see Fig. 18-to use rotary switches.It is often easier to use the base of the box tocarry the terminals and switching, the ac-tual lid being fitted last and becoming theworking base. This avoids having long leadsto connect the switches which would thenbe separated from the rest of the network inthe body of the box.

Attenuators that are to be housed insideequipment, such as a signal generator forexample, can be built into the general inter-nal design method, but a divided -compart-ment assembly is still necessary if the bestresults are to be obtained. For less stringentwork, the resistances can often be simplymounted directly to the tags on a rotaryswitch wafer or wafers as they could for thesystem shown in Fig. 18. Most simple 'di-viders' are assembled this way.

The power ratings of the resistors used inattenuators must, of course, be such as toensure that no appreciable temperature risecan occur in normal use.

Fig. 17. Practical method of assembly. Fig. 18. An alternative switching system-more difficult toscreen adequately.


IN3

A SIMPLI WAT

SINGLE -CHIP microcontrollers havebeen designed and optimized for specific

applications such as industrial control andautomotive applications. The Intel MCS-51family of microcontrollers is a popular andtypical example.

The presence of various types of electricalnoise and interference in the above men-tioned applications is by no means taken forgranted, but generally accepted as hard toavoid. By virtue of some special features,such as powerful Boolean processing capa-bilities, single -chip microcontrollers aremostly used in industrial control applica-tions which are known as electrically noisyenvironments.

In spite of a well -designed power supply,a shielded enclosure, and adequate decoup-ling provisions, noise may still enter a micro -controller system, upsetting the normaloperation, or even causing a system crash.These situations are usually marked by themicroprocessor going into its static state, orperforming unpredictable operations. Insuch a case, normal operation can be re-stored only by a hardware reset (PC usersknow this as BRST - Big Red Switch Time).

A problem occurs when the microproces-sor system operates unattended, which isusually the case in industrial applications. Atotal lockup of the system can then disrupt aproduction line, or cause expensive 'downtime'. In such applications, hardware reset-ting must be performed automatically on de-tection of a software malfunction. The circuitdescribed here (Fig. 1) provides a simplemethod to accomplish this.

Circuit descriptionThe watchdog circuit consists of a power -upreset, a gated astable multivibrator, and afailure detector circuit. In fact, the total cir-cuit simulates a retriggerable monostable

DOG Cby Akbar Afsoos

Re

Fig. 1. Circuit diagram of the watchdog. The input is connected to a CPU port line.

multivibrator, which detects a softwareupset whenever it is not triggered at a properrate.

In the MCS-52 family of microcontrollers,PX.X can be any pin (bit) of the ports 1, 2 or3. Ports 2 and 3 may not be used when per-forming their secondary functions. Compo-nents R3 and C5 form a power -up resetcircuit that applies a reset pulse to the CPUvia gate ICib, whenever power is applied.

During normal operation, and in themain loop of the program, a periodic pulsemust appear on port line PX.X. The fre-quency of this signal must be at least 10 Hz.The software determines which port line isused to keep the watchdog triggered. In 8051accembler code, this is simple to implementby the instruction CPL PX.X, where X.X isthe port line identification, e.g. 1.0 (bit 0 ofport 1).

The software failure detector consists ofCi, Ri, Di and Dz. The pulse train applied tothe PX.X input of the circuit continuouslydischarges C3 via D2 and Ci. As a result, thegated bistable around gate ICia is disabled,and its output goes high. Hence, during nor-mal program execution, the two inputs ofICH, are high, so that the CPU reset line isheld low.

As soon as a software upset occurs, the

CPU very likely leaves the main loop of theprogram. This means that the pulses on PX.Xfail, so that the port line takes on a staticlevel. This enables the astable circuit whichconsists of ICU, R2 and C3. After C3 has beenfully charged by R2, the output of ICla goeslow. Consequently, a logic high reset pulseappears at the output of ICTh.

Because ICia functions as an astablemultivibrator, C3 is discharged again afterabout one second, so that the output of ICiagoes high This causes the CPU reset line torevert to logic low, and the control programis executed starting at the reset address. TheCPU has enough time (about one second) torestart the pulse train on PX.X, and resumeits normal operation.

The reaction time of the watchdog may beshortened by using a smaller capacitor in po-sition Cs. Capacitors Cz and C4 reduce the ef-fect of noise on the operation of the circuit. Ifa reset key is required, you can connect apushbutton in parallel with capacitor C5.

This circuit may also be used with any ofthe devices in the Intel MCS-48 family ofsingle -chip microcontrollers. This, however,requires an inverter gate to be inserted be-tween the output of Km and the CPU resetline, because these controllers have active -low reset inputs.

(PX. X ) Asa. aak.

D.C. ( 0 or 1 )(software upset)

Loh. a= 11=6 CI .1=

I

ICUs output z.ier-up reset41

- watchdog reset(CPU reset)

normal operation again

sintsu

Fig. 2. Ideal waveforms illustrating the operation of the watchdog circuit.


CJ

E CIPHEg

This project is a playful introduction into message scrambling,intended as a mind -teaser or a toy for the junior members of the

family.

SINCE antiquity, the scrambling of theletters in a message has been a popular

though elementary method of encryption.One of the earliest techniques was to writeout the message in the rows of a square, orrectangular array of prearranged shape andsize:

THISISISNOTAHIGHLYSECURESYSTEM

The message is then written out by columns:

TIHSSHSIEYINGCSSOHUTITLRESAYEM

Needless to say, anyone who knows any-thing at all about ciphers will have no diffi-culty in breaking such a simple scrambledmessage. Another well -established methodis to group the letters of the alphabet intopairs, and then exchange one member of apair for the other when writing out the mess-age. The pairs are usually determined by be-ginning with an agreed key -word, followedby the remaining letters in order:

MACHINEBDFGJKLOPQRSTUVWXYZ

To encrypt the message, letters in the toprow are exchanged for those in the bottom

by Owen and Audrey Bishop

.61K

X

3 1) 3 ,7,0 D®0

00 LOLJIIr r ro ro r0L U 4 LO 00 L.0

Fig.1. Representing letters and numerals ona 7 -segment LED display.

row, and the other way about.

Message:ALSO FAIRLY EASY TO BREAKCipher.ONNA WORIMJ TONJ EA UITOZ

In this project we carry the idea of transposi-tion into the letters themselves. We take thestrokes used to write the letter, and scramblethem by swapping them in pairs. The resultis either a different letter or a weird and un-recognizable symbol. Most alphabetic andnumeric characters can be written using the7 -segment matrix commonly employed inpocket calculators and digital docks. Unfor-tunately, there are some problems in repre-senting certain letters such as 'M', 'W' and'X'. For these and a few others, a 16 -segment'starburst' matrix gives a better image, butsuch devices are expensive. This is intendedto be a cheap, easy -to -wire, project for you tobuild for the youngsters. You may even wantto build two, so that they can exchange mess-ages.

So we have settled for a 7 -segment dis-play, even though there are difficulties witha few of the more 'awkward' letters. Thescheme works better if we use a lower-casealphabet (Fig. 1). We make use of the eighthsegment of the display, the decimal point, todistinguish 'm', 'u', 'w' and 'x'. This is usedto represent a central vertical segment in thebottom half of the array.


THE CIPHER MACHINE 47

gmDISPLAY

\ \I;61

ON/0 FP/MODESUPER

SCRAMBLERCLEAR

37:C.57-12

Fig. 2. Suggested layout of the panel of the

Practical useThe Cipher machine is used as follows. Theletters or numerals of the message are en-tered one at a time by touching the appropri-ate segments of a 7 -segment key -panel witha stylus (See Fig. 2.). The scrambled versionof the character appears on a 7 -segment LEDdisplay, and this is copied down onto paper.The receiver of the message has a Cipher Ma-chine with identical wiring (or uses the samemachine), and enters the symbols on to thekey -plate. The machine unscrambles the seg-ments, reproducing the original letter ornumeral on the 7 -segment display.

Choice of displayThe original version of this project simplyscrambled the segments, producing charac-ters such as those in Fig. 3. The weakness ofthis technique lies in the fact that the scram-bled character has the same number of seg-ments as the regular character. This is a due

Cipher machine.

to the identity of the letter. For example, a 3 -segment symbol obviously comes from a 3 -segment letter, either 'c', 'n' or 'v'. It isusually easy to decide, for other reasons,which one of these it must be. Worse still, ifthe letter has only a few segments (forexample, 'c' in Mg. 3), when the symboltends to end up as a number of disconnectedstrokes. Such symbols are difficult to writeconvincingly. They also further advertise thefact that the system depends on interchang-ing segments.

To make unauthorized deciphering moredifficult, the Cipher Machine logically in-verts the scrambled segments when there arefewer than four. For example, the letter' (orfigure '1'), has only 2 segments, but is dis-played as a 5 -segment symbol (Fig. 4). Thisprocedure has the advantage of increasingthe average number of segments making upthe symbols, giving them a more 'connected'appearance. There are a few exceptions tothis inversion routine, for reasons which willbe explained later.

Table 1. Ciphering rules

Case Original character Ciphered symbol b lit?

A 1, 2 or 3 segments, not including h All segments inverted Yes

B 1. 2 or 3 segments, plus h h inverted, rest not inverted No

C 4 to 7 segments, not including h None inverted No

Table 2. Deciphering rules

Case Ciphered symbol Deciphered character h lit?

A 4 to 7 segments, including b All segments inverted No

B 1, 2 or 3 segments. not including b b inverted, rest not Yes

C 4 to 7 segments, not including b None inverted No

0.0 <=k) El

I) r_.>U 0,(1

,,, rli

0 4/ch7.inal letter ciphered cisplay -..irilten symbol

Fig. 3. Simple scrambling.

Fig. 4. Scrambling with inversion.

->u

-->

original letter ciphered display written symbol

Fig. 5. Use of the 'wild card' segment.

Scrambling twistsAnother twist to confuse the snooper is theway we use the decimal point (segment h ofthe display), when this comes up as one ofthe scrambled segments. We think of this asa 'wild card' which can be added to the sym-bol anywhere that is not part of the 7 -seg-ment matrix, as in Fig. 5. You can add the'wild card' in different places each time it oc-curs, producing a confusing set of variationsthat all represent the same letter.

Finally, there is the super -scrambler op-tion. This consists of a 3 -position switch thatvaries the scrambling of four pairs of seg-ments. This switch is set to a given positionbefore ciphering begins. The authorized reci-pient knows the setting, and uses this whendeciphering. All sorts of routines can be usedby prior agreement, such as to alter theswitch at the start of each paragraph, or evenfor each word or letter, according to a regu-lar system.

Circuit detailsThe eight segments of the key -panel are con-nected to the SET input of a set -reset bistable(flip-flop) (Fig. 6). The sEr input is normallyheld logic high by a pull-up resistor, Ri-Rs.

LLEKTOR ELECTRONICS NOVEMBER 1991

48 GENERAL INTEREST

When a segment is touched by groundedstylus, the low input to the corresponding bi-stable causes its Q output to go high. TheRESET inputs of these bistables are all con-nected to a single CLEAR pad on the key -panel, and are held logic high by R9.Touching the CLEAR pad with the stylusresets all the bistables, and their outputs golow. Thus, the bistables store the segmentsthat make up the character.

The output of each bistable goes to oneinput of an XOR (exclusive -0k) gate. Whenthe inputs of the gates are low, the output ofeach gate has the same state as the corre-sponding bistable. If the other inputs arehigh, the output of each gate is the inverse ofthe bistable outputs. For example, if pad a istouched, the first bistable in ICI is set, and its

output, pin 13, goes high. If pin 2 of IC3 islow, pin 3 of ICz goes high, and the LED seg-ment g is lit. If pin 2 is high, the segment isnot lit. The pads and the LED segments arein exchanged pairs so that, for example,touching pad a lights (or turns off) segmentg on the display. Conversely, touching pad glights (or turns off) LED segment a.

The circuit diagram shows just one per-mutation for scrambling. The super scram-bler switch, Si, gives a variable pairingbetween segments c to f This part of the cir-cuit is open to modification. Many differentcombinations of any four of segments a to gcan be wired to the super scrambler switch,and those four segments not so wired can bepaired in three different ways. Th only re-striction is that li and its partner may not be

wired to the super scrambler switch_ If youwant to be able to use one machine for ci-phering, and another for deciphering, bothmachines must be wired exactly alike.

The lower part of the circuit diagramshows the logic switch that decides whensegments are to be inverted. The logic issummarized in Table 1.

In race 'A' we obtain a symbol with fiveto seven lit segments, plus b (assuming thath has been scrambled with b, as in the sche-matic). The letters included in case 'B' are the'm', 'u', 'w' and 'x', all of which require theuse of the h segment. Since we rely on thestate of It (scrambled to b) when deciphering,we have no compromise by not inverting thesegments of these few letters. Cace 'C' in -dudes all other letters with three or fewer

a

ta

$

e° )1D

IC2d

)D-

19

Display

525

SZSoper scrambler

SI /0

SIofliaa mode

1

II66148

RIFTL071

FIS3.-11.101ICS

1144166

6

CS=A011

6V

0

r.3:.71.5

Fig. 6. Schematic diagram of the Cipher machine.


THE CIPHER MACHBCE

from IC111

from ICU

lo ICas. IC70. IC3d.SC4b, ICU

Fig.7. (a) Wiring of Si. as seen from the rear. (b) Wiring of S2. as seen from the rear.

segments. There are no characters that havefour to seven segments and h.

From Table 1, we establish three rules forciphering:1. Invert a and c tog if there are one to threesegments of a tog set, and h is not set (case'A');2. Invert b for the same conditions as in Rule1 (Case 'A');3. Invert 1: if there are one to three segmentsof a and c tog set (case 'B').

The number of segments set is 'counted' byan operational amplifier, IC6, wired as asumming amplifier. We could have used adigital equivalent, but no 7 -input gate isavailable 'off -the -shelf'. This substitute isconsiderably less complex to wire up than adigital version made of discrete gates. Thesumming action is also accompanied by in-version. When three or fewer bistables havea high output, the output of the operationalamplifier swings fully high. When four ormore have a high output, the amplifierswings fully low.

When there are one to three segments lit,the output of IC6 is high; this is fed by way ofSic to IC4d, the XOR gate for segmentwhich is thus inverted, causing segment b toremain off. This complies with Rule 3 above.The output of the h bistable is inverted byIC5a, so its output is high when Ii is not set.The output of IC6 is high when one to threesegments are set. The two high inputs to theNAND gate, ICsb, make its output go low,and this is inverted by IC5c. Through swit-ches Sia and Slb, this causes high levels at theinputs of the XOR gates, inverting segmentsa tog, and conforming to Rules 1 and 2.

The rules for deciphering are slightly dif-ferent, which is why we need Si to switch be-tween ciphering (C) and deciphering (D)modes. The conditions for deciphering aregiven in Table 2. The logic for Rules 4 and 6is effected by Sia and Sic connecting the out-put of the b bistable to the inputs of the XORgates a and c tog. The outputs of IC6 and theb bistable need to be ORed together to per-form deciphering Rule 5. Instead of install-ing an extra IC to provide an OR gate, wehave used 'Mickey -Mouse' logic: the twodiodes Di and Dz, and a 10-kfl pull -down re-sistor, Rzo.

ConstructionSince the Cipher Machine was intended as a

not -too -serious toy for the junior members ofthe family, the prototype was built in abrightly -coloured novelty form (see the in-troductory photograph). The enclosure is afood storage box with a dear snap -on lid.This exposes the 'works' for all to see, so thebattery holder, PCB, switches, ICs and othercomponents are painted in red, blue and yel-low. We also bought plastic knobs for Si andS2.

There is no limit to the variety of the col-our schemes you can employ but, if youprefer a rather different image, the projectcould be realized as a rugged all -black 'mili-tary -look' device. Alternatively, you coulddisguise it as a radio set, or even as a boundvolume to hide away on the bookshelf.

The circuit should not be too difficult tobuild on a piece of Veroboard or shipboard.The battery holder is glued to the bottom ofthe enclosure, and special self-adhesive PCBmounting strip is used for the circuit board.The area around the display is shielded fromdirect light by a cylinder painted matt blackinside, and a bright tint outside. The capfrom a spray can makes a suitable shield.

The key -pads are laid out as an array ofstrips of coloured PVC insulating tape. Anarrow hole is drilled at the centre of eachstrip, and through the lid below. A brassthumb -tack is pushed into each hole to formthe contact for the stylus. Wire is wrappedaround the pin of each tack, and soldered toit. The lid of our endosure was made of high -impact polystyrene or similar plastic, so carewas needed not to shatter it by over -enthusi-astic drilling. The two switches are alsomounted on the lid. Their decorative effect isenhanced by looping out the connectingwires in the shape of flower petals, as shownin Fig. 7. A hole is drilled in one side of theendosure for the lead to the stylus.

There are no alignment problems: if cor-rectly assembled, the Cipher Machineshould work first time. The detailed descrip-tion given above should help you iron outany discrepancies.

OPERATINGINSTRUCTIONS

Ciphering

1. Switch on and turn to cipher mode (C).

2. Set the super -scramble switch to anyone of the positions you have agreedwith your correspondent.

3. Touch the stylus to the CLEAR pad.

4. To encipher the first letter or numeralof the message. touch the stylus to thesegments that contain it (Fig. 1).

5. Copy on to paper the symbol that ap-pears on the display. If possible, make itmore 'rounded' like a hand-written char-acter. If the decimal point on the displayis lit, add a stroke anywhere you like. butnot where the seven segments of the dis-play are located.

6. Repeat steps 3 to 5 for each characterof the message. Leave a space betweenwords, or devise a special 'space' sym-bol, making sure it is not one that de-ciphers into a regular letter or numeral.

Deciphering

1. Switch on and turn to decipher mode(D).

2. Set the super -scramble switch to theposition that you have agreed with thesender of the message.

3. Touch the stylus to the CLEAR pad.

4. To decipher the first symbol, touch thestylus to the segments that it contains. Iiit includes a 'wild card' stroke, touch thedecimal point.

5. Copy on to paper the character thatappears on the display. If the decimalpointis lit. it indicates one of the letters

'u'. ;,.1 or 'x'. depending on whichother segments are lit.

6. Repeat steps 3 to 6 for each symbol ofthe message.


in)QQAD FOV. LT

Following last month'sdealings with thefloating-point nucleus andthe hex -to -BCD conversionroutine in the MCS-52BASIC interpreter, theauthors now tackle someproblems withmultiplications.

by Z. Stojsavljevic and D. Mudric

Continued from the October 1991 issue.

INCONSISTENCIES in the multiplicationof two numbers as performed by version

1.1 of Intel's MCS BASIC -52 interpreter canbe demonstrated by running the followingthree small programs, all of which producewrong results.

1020

a=1.E-65b=1.E-65

30 ?a*b (result: 1.0E+126)

10 a=1.E-6520 b=1.E-6430 ?a*b (result: 0)

10 a =1.E-6420 b=1.E-6430 ?a*b (result: 1.0E-0)

In all three cases,produced

ERROR: ARITHLINE 30

the interpreter should have

. UNDERFLOW - IN

The above examples point to inconsistenciesin limit cases when two numbers are multi-plied. On investigating the operation ofIntel's BASIC interpreter, a multiplication al-gorithm of the type shown in Fig. 1 wasfound. Apparently, the inconsistenciesbrought to our attention by the aboveexample programs were caused by the expo-nent adjustment routine, which is listed inFig. 2. Unfortunately, the errors found in thisroutine cannot be corrected in BASIC -52 ma-chine code, because some expansion of themachine code is in order. This means that anumber of lines should be added to the as-sembler source file for compensation.

A problem occurs when the result of theinstruction SUBB A,I181H is OFFH, which

VC 3AS

Fig. 1. Flowchart of the routine thathandles multiplication of two numbers.

equals exponent E+127. If the program partfor mantissa multiplication includes a resultthat begins with a 0 after the decimal point,the exponent is not incremented but remains

-52

1A9AH 9175 ACALL1A9CH 8E0002 CJNE1A9FH 6188 AJHP1AA1H 8112F NOV1AA3H EF NOV1AA4H 60F9 J21AA6H 2E ADD1AA7H 20E705 JB1AAAH 100706 JBC1AADH 6152 AJNP1AAFH 5002 JNC1A13111 6IAI AJHPIABAH 9481 SUBBIAB5H FE NOVIA86H 7188 ACALL1A88H 7804 NOV1A8AH ACOI MOVIABCH 8C01 NOVIABEH E3 MOVXIABFH FA MOV1830H 7834 NOV1832H E6 NOV1833H FE NOV:334H 6003 JZ:336H 717F ACALL1338H 18 DEC:839H 08 INCIBAAH 7408 NOVIB3CH F9 NOV18308 28 ADDIBAEH F8 NOV183F0 860500 CJNE11342H 4013 JC1844H D3 SETB11345H E4 CLR1846H 18 DEC1847H 36 ADDC11348H D4 DA11349H D6 XCHD104AH 30E409 JNB184DH D9F5 DJNZ184FH 18 DEC18508 7601 NOV1852H 717F ACALL11354H 8006 SJIIP

1856H 19 DEC1857H E9 NOV1858H C3 CLR1859H C8 XCH185AH 98 SUBS1135138 F8 NOV185CH 792B HOV1135EH E6 NOV185FH C4 SWAP11360H 08 INC1861H D6 XCHD1862H 4206 ORL11364H F7 MOV1865H 08 INC1866H 09 INC181678 892FF4 CJNE1B6AH EE MOV106BH 7003 JNZ1860H 753000 NOV1.117FH 0530 INC1881H E530 NOV1883H 70F9 JNZ1685H DOE0 POP1887H DOE0 POP11389H 61A1 AJHP

1C75HR6,1100H,IAAIH188882FH,R5A,R71A9FHA,R6ACC.7,1AAFHCY,IAB3H11382HIAB3HIBAIHA,1181HR6,A188BHR3,#04HR4,01H01H,R4AAR'R2,ARO,#34HA,@ROR6,A1839H1B7FHR0ROA,108HR1,AA,R0RO,A9R0,50514,18421-105714C

805,01/0AA4R0ACC.4,11356HRI,11344HRO980,101HIB7FH1135CHRIA,R1CA,R0A,R0RO,A81,12BHAAROAR0A,@R006H,A@Rid,R0RIR1,112FH,1135EHA,R61070H3OH,100H30HA,30H187EHACCACC113A1H 910126.2-.2

Fig. 2. Original program part for exponentadjustment.

ELERTOR ELECTRONICS NOVEMBER 1991

UPGRADE FOR NICS BASIC -52 V.1.1 (Part 2)

ORG 1A9AH

; TOS - Top Of arithmetic Stack; NXTOS - NeXt TOS (position behind)

; TOS_NUL1; routine for multiplication of two FP numbers, one of; is located on TOS and the other on NXTOS (NXTOS'TOS)

TOS_MUL1:ACALLCJNE

TOSN_0: AJMPNXTNI4_0: NOV

NOVJZ -

ADDJBJEWAJMP

CIIPM_EXP:JNC')VER_HD: AMP

1

CORM_EXP:CLRSUBBINCJNCSETS

NMARK_L: NOVACALLMOVNOV

HUL_NOOV:110VMOVXMOV

PREP_MUL; clearing of FP workingR6,80,NXT/411_0 ; is NXTOS equalZERO_HANTISSA ; in case TOS orSIGN,RS ; result mark inA,R7 ; is TOS equal 0?TOSM_OA,R6 ; additionACC.7,CMPM_EXP ;

CY,CORM_EXPUNDERFLOWCORM_EXP; in thisOVERFLOW; in this

which

space and reg. prep.0?NXTOS are 0SIGN

of exponent degree (exp. multipl.)exp. with different mark or overflow?result is bigger than 0.1in this case exp. sun is < -127case result is smaller than 0.1case exp. sum is > 127

MUL_LIMIT_CASE ; flag of multiplication limit caseA,#82H ; exp. multipl. results are within the limitsA ; deduction with 82H because of reducedNMARK_L ; conditionsMUL_LIMIT_CASE ; limit caseR6,A ; exp. adjustment to real valueBCD2_1 ; disassembl. of TOS mantissa in LEN_MANTISSAR3,8LEN_BYTE ; acc. (one number in each byte)R4,AR1 ; RI is pointer of NXTOSAR1,R4 ; @R1 contains few numbers and disassembledAAR1 ; mantissa is successively multiplied by eachR2,A ; of them in argument accumulator (R2 is

; auxilliary register)

; TRANSFER; routine for adjustment of multiplication and division; results of FP numbers in argument stack and mantissa; conversion in BCD packed format in argument accumulator

TRANSFER:MOVNOVNOVJZACALLDEC

MSB_EQ_O:INCMOVNOVADDNOVCJNE

ROUND1: JCC_BCD: SETB

CLRDECADDCDAXCHDJNBDJNZDECNOVACALLSJHP

NO_C_BCD:DECNO_ROUND:MOV

CLRXCHSUBBNOV

BCD1_2: NOVJNBJZINCJZ

N_LCASE: NOVXFER_1: NOV

SWAPINCXCHDORLNOVINCINCCJNENOVJNZNOV

; is no rounding; carry takes one because of rounding; transfer of carry into higher byte; addition of 1 to higher byte; adjustment to BCD format

A,P0 ; result storing in higher byteACC.4,NO_C_BCD ; transfer in higher BCD nibble?R1,C_BCD; is @RO the address of NSB number?RO ; transfer oemori2ing in the place MSB+i@RO,#01INC_EXP ; transfer is outside decimal point frameBCD1_2R1A,R1 ; address return in RB to the first bigger BCDC ; number which is not equal 0A,R0A,R0RO,AA,EXPONENT ; test of exceeding in limit caseHUL_LINIT_CASE,N_LCASEUNDER_ND; message about underflowAUNDER_ND; message about underflowRI,8ARGUMENT_ACC; repacking of BCD number in one byteA,@RO ; in two BCD numbers in one byte

; higher number in BCD packed format is in80 ; upper nibbleA,@RO ; BCD packed format is in ACCAR6,A ; the indication if mantissa is 0 is in R6@R1,A ; packing in argument accumulator mantissaRO ; two BCD packed numbers in one byteRIR1,8SIGN,XFER_IA,R6RESULT_MAT_OP_TOS ; is mantissa equal 0?EXPONENT,80

ROAIASCII_DEC ; pointer to HSB result numberA,@RO ; HSB result number is in acc.R6,AHSB_EQ_0; is MSB result number equal 0?INC_EXP ; exp=exp.1 because from the start it wasRO ; taken that HSB result number is equal 0RO ; pointer to 108-1 numberAALEN_MANTISSA ; positioning to LSB-1 numberR1,A ; RI is counterA,R0RO,A ; RO contains the address of LSB-1 number@110,885,ROUNDI ; test of remainder to roundingNO_ROUND; in case remainder is smaller than 5 thereC

ROA,@ROA

INC_EXP: INCNOV.JNZPOPPOPJBCAJMP

UNDER MD:AJIIP

EXPONENT; exp. adjustment by 1 (consequence ofA,EXPONENT ; transfer after addition)EXP_OK ; 'test of exponent exceedingACC ; removal from return address stackACCNUL_LIHIT_CASE,UNDER MD ; in limit case it isOVERFLOW; underflowUNDERFLOW

910128-2-13

Fig. 3. Source code of the improved multiplication routine. The vertical bars indicate linesthat correct the limit case multiplication errors mentioned in this article.

OFFH. After the result formation (see Fig. 1)and the serial output (display) routine, thisgives an exponent E+126.

To avoid inconsistencies, and eliminateerrors in limit case multiplications, the pro-gram listed in Fig. 3 was developed. Thelines marked by vertical bars in particularcorrect the limit case multiplication errors.

On the basis of the information containedin this and last month's instalment, theauthors developed an improved, error -freefloating-point (FP) arithmetic for the 8051,on the basis of the FP nucleus extracted fromthe 8052AH-BASIC. The new FP nucleususes a modified way of accessing the arith-metic stack, and has faster, shorter code for anumber of algorithms. It allows the length ofthe mantissa of a FP number to range from 2to 16 digits, while the entire memory map ofthe FP arithmetic variables is located in theinternal memory space of the 8051 microcon-troller.

AZ -e7


V IJ IJ

omEcT

A series of projects for the not -so -experienced constructor. Although each articlewill describe in detail the operation, use, construction and, where relevant, the

underlying theory of the project, constructors will, none the less, require anelementary knowledge of electronic engineering. Each project in the series will be

based on inexpensive and commonly available parts.

TIMER FOR CENTRAL HEATING SYSTEMS

To achieve a comfortableroom temperature quicklyit is often required toswitch on the centralheating (CH) boiler for tenminutes or so, forinstance, in the morning,or when you get homefrom work. Unfortunately,most CH systems can beswitched on only byturning up one of the roomthermostats. This works,but if you forget to turn itdown again, the resultanthothouse is not onlyunhealthy but also a wasteof energy.

by J. Ruffell

rrHE circuit described here is a timer thatcan be connected to most types of CH

room thermostats, including the well-knownHoneywell types and look-alikes (see theabove photograph). Its function is to actuatethe CH boiler (and pump) for a certainperiod, overriding the thermostat output.Unlike you, it is not in the habit of forgettingto turn the boiler off again - all you have todo is press a button to start the timer. The cir-cuit is simple to build, and convenientlyfitted next to the room thermostat.

How it worksThe timer is actually a monostable multivi-brator with an adjustable on -time. The cir-cuit diagram in Fig. 1 shows that a 4060ripple counter IC is used as the timing de-vice. The 4060 has an on -board oscillator, Fig. 1. Circuit diagram of the central heating timer.


TIMER FOR CENTRAL HEATING SYSTEMS 111

Fig. 2. Component mounting plan on board UPBS-2.

which is connected here to a resistor -capaci-tor network, R1 -Pi -R2 -C1 -C2. These compo-nenis, connected to the PO (phase out), PO(phase out inverted) and PI (phase in) pins ofthe 4060 determine the oscillator frequency.Preset Pi gives the oscillator a frequencyrange of about 6 Hz to 27 Hz. The Q13 out-put of the 4060 goes high after 2" (8,192) os-cillator dock' pulses, that is, after 5 to20 minutes depending on the setting of Pi.

The ripple counter in the 4060 operatesonly if the reset input of the IC is held logic

low, which is the case when bistable IC2a isset. This happens when push-button S2 ispressed. When the Q13 output of the countergoes high, the logic level at the data (D) inputof the bistable is docked ('latched'). Since theD input is permanently connected toground, a logic low is docked, which has thesame effect as resetting the bistable (note thatresetting is also possible by pressing push-button S2).

The resetting of the bistable halts thecounter. This happens either when the preset

wire to room thermometer

circuit

1---.. mainsadaptor

r -

L _MOM

thermometer

1

push-button

916034X-12

Fig. 3. Connection of the unit to the CH wiring.

COMPONENTS LIST

Resistors:3 1 OkS1

1 1 M

2 18k1).

1 27052

1 50ki-2 preset H

Capacitors:1 1nF

1 111F solid MKT1 100(F 16V axial1 100nF

Semiconductors:1 4060

1 40131 76L052 BC5501 LED1 1N4148

R1;113:R4R2

R5;R7R6

P1

ClC2C3C4

IC1

IC2

IC3T1;T2DID2

Miscellaneous:1 V23127 -A0001 -A101 R el

(Siemens) or similar5-V PCB -mount relay

1 push-button with LED hole1 push-button1 2 -way terminal block1 printed circuit board

S1

S2

K1

UPBS-2

time has lapsed, or when the user presses S2while the timer is still counting down. If theQ output of the bistable is high (Le., when thecounter is active), transistors Ti and T2 areswitched on via R5 and R7 respectively. Ti inturn switches on a relay whose contacts areconnected in parallel with the room thermo-stat contacts. The dosed contact causes theCH boiler (and pump) to be switched on. Atthe same time, T2 turns on a LED that indi-cates the activity of the CH timer.

The timer has an internal 5-V supplybased on a 100-mA 5-V regulator Type78L05. The direct input voltage of between8 V and 12 V may be supplied by a mainsadaptor.

Construction andconnectionThe CH timer is built on a universal proto-typing board size -2 (UPBS-2), which may beobtained through our Readers Services. Asuggested component mounting plan isshown in Fig. 2. The two ICs are best fitted insockets. Do not forget any of the wire links,and observe the polarity of the diodes andthe electrolytic capacitors. Make sure con-nector Ki is soldered securely to the board toprevent it coming loose when the connectingwires cause strain. If you want a front panelcontrol instead of the preset on the board,simply replace Pi by a 47-1d2 linear poten-tiometer. A simple timing scale may be pro-vided around the pot by measuring theon -time of the relay at a few settings. Finally,the timer is connected to the CH system wir-ing as shown in Fig. 3


54

RELAY CARD FOR UNINTERFA E

S A

This relay card connects to the universal I/O interface for IBM PCsdescribed in the May 1991 issue. Simple to build and program, it

offers a safe and easy way of controlling all sorts of equipment bymeans of a PC.

ALTHOUGH the relay card describedhere is designed specifically as an ex-

tension for the PC I/O interface (Ref. 1), itsinput control signals are readily found, ormade, in non -IBM (PC) computers.

The circuit diagram of the relay card,Fig. 1, shows that it is linked to the I/O cardvia connector Kt. It is also seen that the I/Ocard signals are buffered and fed to connec-tor K4, which allows up to four extension cir-cuits to be connected in series. Only twosignals 'change' between Kt and K4: addresslines AO and Al are interchanged on K4(with respect to Ki), and A0 is inverted. Thisallows all extension cards connected to thePC I/O interface to make use of a single,simple, address decoder. All cards con-nected in this way respond internally to ad-dress 00 (binary), but the successiveinterchanging of AO and Al causes their ac-tual address to be determined by the order inwhich they are connected. Table 1 shows theaddress assignments.

More relays than ICsThe data flow between the PC I/O card andthe relay extension is controlled by a bidirec-tional buffer, IC2. Although a unidirectionalbuffer would have been in order for the relaycard (which functions as a write -only exten-

by A. Rigby

Table 1. Address overview

Addressbase address + 0base address + 1base address + 2base address -3

Relay card1

9

3

4

sion), bidirectional buffering is applied be-cause two-way data flow may be required byother cards in the system.

The actual relay interface starts at registerIC4, which is used to latch data when therelay card is addressed. The addressing is ac-complished via the ENABLE and WR lines.When both are low, the output of ICib is lowalso. To ensure that the data are stable at theinput of the register, they are latched whenthe output of ICib reverts to logic high. Thelogic pattern stored in IC4 is fed to driver ICs,which controls the relay coils. The relays areactuated by a logic high data bit written to1C4, so data inverting is not required.

The relay contacts are brought out to pinson connectors K2 and 1(3. Connector K2 car-ries the mother contacts and the normallyopen (NO) contacts of the relays, and 1(3 themother contacts and the normally closed(NC) contacts.

Table 2. Relay specifications

Siemens V23040 -A0001 -B201Contact specifications

Max. switching voltage:

Max. switching current:Max. continuous current:Max. switching power:

Max. si.vitching frequency:Mechanical lifetime:Mechanical lifetimewith contacts loaded:

150 V d.c.125 V a.c.

2A2A

35 W d.c.60 W a.c.

100 Hz108 s.o.

103 - 108 S.0

s.o. = switching operations

Building and testingThe relay extension is built on a double -sided, through -plated printed circuit board,of which the track layouts and the compo-nent overlay are given in Fig. 2. The con-struction is mostly straightforwardsoldering work. The completed card is con-nected to the PC I/O interface via a 20 -wayflatcable with IDC connectors at both ends.


RELAY CARD FOR UNIVERSAL I/O INTERFACE El

Re1...Re8 = Siemens V23040 -A0001- 8201

Fig. 1. Circuit diagram of the relay extension. Up to four of these circuits may be connected in series and controlled by a PC.

The current consumption of the relaycard is determined by the number of actu-ated relays. When all relays are actuated, thecurrent consumption is a little below150 mA.

The switching functions of the relay cardmay be tested with the aid of the program

listed in Fig. 3. When run, this programcauses the relays to be actuated and deactu-ated in succession. The program may beused to test up to four relay cards connectedin series.

Because of the track layout of the printedcircuit board, the maximum voltage that

may be switched by the relays is 42 V a.c. or60 V d.c. This means that the relay card maynot be used to switch mains loads directly.

Reference:1. Universal I/O card for IBM PCs. El ektorElectronics May 1991.


56COMPUTERS AND MICROPROCESSORS

O

C2 Re I

io o cscp

ms -son `Rn 9?°) u13Ts4rioc+cw w//3 w

3

cs8 888888BB--(6

cr(kkk P°k ° 0)

Zax-im

au! ob\o\Iw(1) °

CD DJ CD CD CD CD CD CD CD CD CO CD CD OD

N0

o c oo-I to

o1 IC Re 2

O FO

Re3

O -0'

Re4

O

200000000161 00 0 000 0 01 5

K2

Re5

0 -0

0-8-00-)-0

Rea

0 -0'

0-0-00-1-0

Re?

E°

ore

0 -0

0-8-00-1-0

2000000001610000000015K3

Rea

N0000

00Q00

0

0

Fig. 2. Printed -circuit board for the relay card. Note that this board is double -sided and through -plated.

10 CLS20 ' ibmio interface test30 X=0 ' address definition40 ' X=0: 611300-6H303 X=1: 60304-307 X=2: 6H308-6H3OB X=3: 61130C -30F50 ' X=4: 01310-6H313 X=5: 6H314-317 X=6: 60318-6H31B X=7: 0131C -31F60 X=611300+X.6H470 addresses'

80 A1=X+0: A2=X+1: A3=X+3: A4=X+2- ' I/O addresses90 ' Al100 '110 CLS120 PRINT "Testing I/O"130 FOR I=0 TO 7

- A4 = relay card 1 to relay card 4test of I/O ports

140 OUT A1,2 -I ' close relay number i of card 1150 OUT A2,21 close relay number i of card 2160 OUT A3,2-1 ' close relay numh.r i of card 3170 OUT A4,2 -I '

180 GOSUB 280 '

close relay number i of caficalit

190 NEXT200 FOR 1=0 TO 7210 OUT A1,255 -2-I ' relay rim,hr i of card 1220 OUT A2,255 -2-I ' relay number i of card 1230 OUT A3,255 -2-I ' relay number i of card 1240 OUT A4,255 -2-I ' relay number i of card 1250 GOSUB 280 wait260 NEXT I270 GOTO 130 ' return for next cycle280 ' subroutine to execute a wait period290 FOR J=0 TO 1000:NEXT300 RETURN

openopenopenopen

COMPONEN=TCapacitors:2 100nF

Semiconductors:1 74HCT321 74HCT2451 74HCT041 74HCT5741 ULN2803

C1:C2

IC1

IC2IC3IC4IC5

Miscellaneous:2 20 -way pin header with K1;k4

side latches2 16 -way pin header K2:1<3

8 PCB mount relay Re1-Re8V23040 -A0001-13201 (Siemens)

1 enclosure (e.g.. Heddic 222)1 Printed circuit board 910038

Fig. 3. Run this little BASIC program to test one to four relay cards.


WAAL QUADRIFORM ELIZRTVA\SMiT/RECEIVE ANTENNA

by Richard Q. Marris, G2BZQ

The quadriform is an experimental 3.5 MHz transmit/receiveantenna of mini dimensions, consisting of four high-grade robust

ferrite rods in a 9 in x 9 in (230x230 mm) loop configuration.Chambers 20th Century Dictionary defines quadriform as

"fourfold; having four forms or aspects".

Fig. 1. Assembly of the quadriform ferrite transmit/receive antenna.

MI, a decade or so after the end ofWar II, the transistor and ferrite

rod antenna made their entry on the market,they quite revolutionized the domesticLW/MWradio scene with the arrival of small portablereceivers with built-in aerials. One mighthave expected that ferrite antennas would alsobe used extensively for transmission, butthis wasnot tobe for various reasons. Obviously,however, experiments were taking place, es-pecially in the American defence industry, butlittle or nothing has been published on thesubject.

Nowadays, ferrite rod materials can bedivided broadly into manganese -zinc mix-tures and nickel -zinc mixtures. The former areused invariably for VLF and LF receivingapplications between 1 kHz and 1 MHz, andthere are several different mixtures. Nickel -zinc rods, again in several mixtures, haveappeared with permeabilities between 40and 850. In the latter types, apart from theiruse in domestic LW/MW radios, it has be-come possible over the years to obtain rodsthat can be used as HF band receiving an-tennas. An excellent booklet (1) is availablefrom Amidon Associates (USA), which givesthe full characteristics and sizes of manytypes of manganese -zinc and nickel -zinc rod.It may well enlighten the reader as regardsferrite rods for reception, but the contentsare not geared to use of the rods for trans-mission.

Of particular interest are the Amidon Type61 nickel -zinc rods, which are quoted as beingsuitable for use in receiving antennas for fre-quencies of 0.2-15 MHz. However, I have usedthese rods experimentally for reception inthe VHF bands with acceptable results. I havealso made several experimental single rodtransmitting antennas from Type 61 ferritematerial which gave widely varying results.A brief report on a small part of earlier ex-periments was published in Ref. 2. For con-venience, all ferrite transmitting antennas usedin my 'on the air' experiments have been de-signed for the 3.5 MHz (80 metres) amateurband.

Some of the problems that revealed them-selves as the experiments progressed duringthe design of ferrite rod transmit antennasare listed here.


58 RADIO, TELEVISION & COMMUNICATIONS

1. The difficulty of locating small suppliesof sufficiently large and robust rods suit-able for high frequencies and, most im-portantly, for transmission.

2. How to couple the transmitter to the fer-rite antenna: to feed RF power into theantenna? And then how to persuade theantenna to radiate that power over satis-factory distances to other stations?

3. Solving the serious problem of saturationof the core. As power into such an an-tenna is increased, a point is reached quicklywhere core saturation occurs. This satu-ration manifests itself by the heating upof the core, which appears to produce thefollowing effects:(a) de -resonating the antenna from theselected frequency;(b) harmonics that cause intereferenceand TVI;(c) general instability, similar to (b).

Itmust be stressed that these findings are basedentirely on the results of my own experi-ments.

Assuming that theseproblems canbe solved,a tremendous advantage appears in that wehave a tiny antenna-only a few inches long-compared with a traditional wire antenna:on 3.5 MHz, a wire dipole is about 140 ft (al-most 43 metres) long!

It has been discovered that the main dif-ficulty is that of core saturation, even whena suitable rod for hi transmission has beenfound. Experiments have shown, however,that this problem can be overcome to a largeextent by leaving a large gap between theferrite rod and the wire turns and distribut-ing the turns over the entire length of therod. From these findings, it was decided to tryto produce a realistic unidirectional ferritetransmitting antenna with reasonable gain.The result is the Quadriform Ferrite Antenna.

DescriptionFrom Fig. 2, it is seen that the antenna con-sists of a 9x9 in. (230x230 mm) loop withsmall inductances, L1 -L5, in each side. Thewire sides are formed around a 7.5x0.5 in.dia. (190x12.5 mm) high-grade nickel -zinc(Type 61 material) rod enclosed in a plastictube as shown in Fig. 3. It is fed at the bot-tom, between Li and L5, with leads to plugs

and P12 that mate with appropriate sock-ets, Skt1 and Skt2.

The loop is resonated by a 100 pF ceramic,insulated, variable capacitor, C1. The feed toand from the transmitter/receiver is a TypeRG58 coaxial line via C2 (a 150 pF ceramic,high -voltage disc type). Capacitors C2 andC3 form the actual coupling and impedancematching element to match the loop to a 50impedance. Socket Skt3 is an optional earth-ing connection. The antenna is unidirectional.

The general layout is shown in Fig. 1. Theloop is supported vertically by a 9 in (230 mm)long piece of plastic tube fastened with wallclips to the front vertical edge of the side ofa 6x4x4 in (150x100x100 mm)aluminium boxthat houses the variable capacitor and someother components. The box is large and heavy

A

A

A

B

01000 0 L 0 0000 0 100101=lo0L2

L4.1,010.1011,00

"I" = Hi - current (RF)"E" = Hi - voltage (RF)

A = 9" (230 mm)B = 7.5"(190 mm)

C2

150p1000V

RG58 cable(Z=501))

L5

maximumradiation

II

ferrite rod (.4)7.5" long 0.5" dia.(190.12.5 mm)(see text)

9101.13-11

Fig. 2. Circuit diagram of the quadriform antenna.

enough to act as a base that stops the looptilting. Sockets Skt1-Skt3, together with Cr,are mounted on the front panel; the RG58 cableexits through a grommet at the rear. The boxis a two-piece interlocking double 'U'; the loopis clipped to the removable lid.

Care is needed in winding the inductors.First, cut three 7.5 in (190 mm) lengths of the11/16 in (17.5 mm) dia. tubing and two piecesof about 2.5 in (63 mm). The two shorterlengths, when fastened together with a stan-dard T junction, form the bottom side thatshould also be 7.5 in long. The 9x9 in squareis completed with four standard 90° plastic el-bows. All plastic tubing and junctions are read-ily obtainable from most DIY stores (in the UKunder thenamePOLY-YORC). A75in (190 mm)long,0.5 in (12.5 mm) dia., ferrite rod is in-serted into each side before the corner el-bows and T junction are fitted. It is neces-sary to wind a couple of turns of masking taperound the ends of each rod to ensure rigidityin the tube-see Fig. 3. The loop windingsare made from PVC covered, single -strand,1/0.6 mm wire, 1.2 mm o.d., rated at 1 kV,1.8 A. They are wound as shown in Fig. 3.The wire linking inductors L1 and L5 is tapedclose along the outside of each of the tubes;over the 90° elbows it is held in place by a small'V' filed in the outside radius of each elbow.

Black tubing, black PVC wire and black PVCtapes were used throughout.

The whole wire length, including the smallcoils, forms part of the ferrite -cored loop.Inductors L1 -L5 are close wound, clockwisethroughout.

Variable capacitor Ci is mounted at deadcentre of the front panel of the base box andit and the other few components are hardwired. A 9 in (230 mm) length of tube is in-serted into the T junction of the finished loop;the bottom end of this is clipped to the frontedge of the side of the lid of the box withtwo standard plastic wall clips-see Fig. 2.

CAUTION! It is essential that the specifiedferrite rods are used-see parts list. Theyshould be treated with great care. If they aredropped on to a hard surface, they may break,chip or suffer from core destabilization. Theyare expensive! At the time of writing, they cost$15.00 (about £8.50) each plus $8.00 for air-mail shipping direct from Amidon. Deliveryis by return mail and is usually within threeweeks from airmailing the order to Amidon.Master Card and Visa are accepted.

Testing, operation and resultsInitial testing of the antenna is carried out witha receiver tuned to 3600 kHz. Capacitor C1 is


EXPERIMENTAL QUADRIFORM FERRITE TRANSMIT/RECEIVE ANTENNA 59

pad rod ends with masking tape 7.5" x 0.5' Cia ferrite rod

90' plastic elbow(each corner) tape wire en outer edge

4 turns

2 turns "I- piece

Ele small "rto hold wire

reach rile:)

910143.13

Fig. 3. Assembly details.

rotated for resonance, which is indicated byan increase in signal strength. The plates ofthe capacitor should then be enmeshed about50%. The 3500-3800 (or 4000) kHz band isaccommodated without any problem.

The polar diagram of a single horizontalferrite rod antenna is a figure 8 off the twolong sides. The polar diagram of the Quaciriformis rather more complex and not easy to es-tablish indoors, possibly owing to reflectionsoff walls, and so on. However, it is, to all in-tents and purposes, unidirectional with onevery large lobe off one end, which reducesinterference appreciably. This lobe can bemoved through 180° by reversing P11 andP12 in sockets Skti and Skt2, which is a sim-ple way of achieving all-round operation. Ifnecessary, further small adjustments can bemade by a small rotation of the antenna. Thisradiation pattern holds good on transmitand receive.

Before the antenna can be tested for trans-mission, it is necessary to first set up thetransmitter into a 50 SI dummy load at thedesired frequency. The antenna is then res-onated with the receiver, at the same fre-quency, and subsequently substituted forthe dummy load. It should be fully loadedby the transmitter on low power, after possi-

ble minor adjustments of C1. The usual band-width on transmit, without readjusting C1,is about 25 kHz.

The antenna is essentially a low -power de-vice. In my experiments, it has been drivenby over 20 watts RF without detectable coresaturation. However, it is intended for useindoors near the transmitter/receiver, wherea transmitter power of <10 watts has been used9323MCNIECILISIMIIEMLint..

'on the air' with very satisfactory results.There was no detectable harmonic or TVI ra-diated. The use of an RF earth connection madeno apparent difference.

WARNING!IN THE INTERESTS OF DOMESTIC SAFETI,

THE TRANSMITTER OUTPUT SHOULD NOTEXCEED 10 WATTS.

ReferencesIron -powder and Ferrite Coil Forms, publishedby Amidon Associates, P.O. Box 956, Torrance,California 90508, USA.

"Fe -ONE ExperimentalTransmitting Antenna"by Richard Q. Marris, Practical Wireless,January 1989.

PARTS LIST4 nickel -zinc rods. 7.5 in long by 0.5 in. die.

(190\12.5 mm), available from AmidonAssociates Inc. P.O. Box 956, Torrance,Ca. 90508, USA; Tel. +1 213 763 5770:Fax 1-1 213 763 2250.

1 aluminium box 6x4x4 in.(153/102/102 mm) order code Box/J10from Marco Trading.

1 variable capacitor, 100 pF, Jackson BrosType 11 or equivalent style and qualityplus knob to suit.

1 ceramic disc capacitor, 150 pF, ?_1 kV.1 ceramic disc capacitor, 350 pF, kV.2 4 mm banana plugs3 4 mm banana sockets1 m Poly-Yorc (or similar) black tubing,

o.d. 0.7 in (18 mm), i.d. 0.6 in (13 mm);available from many DIY stores.

4 90' elbows to match the black tubing.1 T junction to match the black tubing.2 wall clips for fitting the black tubing.5 ft RG58 coaxial feed line: Z=50 0. plus

coaxial plug to suit transmitter/receiver.PVC covered. single strand 110.6 mmwire (black). 1.2 mm o.d., rated at 1 kVand 1.8 A. Order code CBUEWliblackfrom Marco Trading.Black PVC tape and masking tape as re-quired.


Z-MA"C

17 -MATCH II is a Smith chartLaRF design software package.Quoting from the supplier, Num-ber One Systems: 'In spite of theavailability of modern designaids, such as the hand-held pro-grammable calculator, sophisti-cated circuit simulation andComputer Aided Design (CAD)software. the Smith chart is stillwidely used as a radio frequencycircuit design tool. The Z -MATCH II program enables theSmith chart design process to beperformed easily and accuratelyon a personal computer.'

For those of you who havenever seen a Smith chart, the fol-lowing is going to seem likewhite man's magic! However, Ishall attempt a short explanationof what a Smith chart is, and howone is used. This is not going to be the easiestthing for me to do, as it is a very long time sinceI learnt how to use them. Therefore, with the kindpermission of Number One Systems Limited, Iam going to reproduce, in part, the introductoryexplanations of Smith charts and their uses fromthe Z -MATCH II instruction manual.

Smith chart circlesThe Smith chart is made up of two sets of circles;one set represents the resistive part, R, of a com-plex impedance, the other set represents the reac-tive part, X.

Normalised resistance and reactanceIn order to avoid the need for a different chart foreach characteristic impedance, Zo, the paperSmith chart uses normalised values of resistanceand reactance circles. The normalised imped-ance, 4, is given by:

4 = z/zo = (R/zo) + i(x/zo)

Where Z, R and X are the actual values of imped-ance, resistance and reactance respectively.

Series impedance valuesThe impedance of any point on a transmissionline can be represented by a point on a Smithchart.

The series impedance value of a point on thechart is found by reading the values of the inter-secting resistance and reactance circles. The Z -MATCH II program displays the seriesimpedance value corresponding to the cursor po-sition on the chart.

Parallel admittance valuesAn admittance value (Y-) Smith chart can also bedrawn. The circles on this chart represent values

Circle DisplayLoss Point--.---..

Jelocity Factor=0.66

1n A girAf

by Mike Wooding G6IQM

node cOlour Utility QuitR/G circ SUR circ X,B circ

Z ChartD/C

s vv

0.000-p.393

Impedance -Mims:

0.0000 -j175.47 -

of constant conductance (I/R) and constant sus-ceptance (1/X). By using a Y -chart overlay on animpedance chart, it is possible to convert fromseries impedance to equivalent parallel admit-tance.

Standing wave ratio (SWR)A circle that is concentric with the centre of aSmith chart has a fixed value SWR. The SWR ofsuch a circle is equal to the value of R/4 at thepoint where the circle crosses the horizontal axison the right-hand side of the chart.

Intersections of the SWR circle with the hor-izontal axis on the left of the circle representpoints of voltage minimum, intersections on theright represent voltage maxima. Moving round aconstant SWR circle is equivalent to travellingalong a lossless transmission line; successivevalues of impedance indicated on the chart corre-spond to the impedances seen along a losslessline with the same SWR.

Wavelengths towards generator and loadThe distance moved on a transmission line is di-rectly proportional to the angle of rotation arounda constant SWR circle; one revolution is equal toa half wavelength movement.

Moving around an SWR circle in a clockwisedirection is equivalent to travelling towards thegenerator, whereas moving anti -clockwise is thesame as travelling towards the load. The wave-lengths towards the generator and load (back-wards and forwards respectively) are shown onthe periphery of the standard paper Smith chart.These peripheral scales on the paper chart areused by drawing a straight line from the centre ofthe chart through the point of interest. The Z -MATCH H chart indicates directly the wave-length (or length in metres) corresponding to thecursor position.

By convention, the starting point for the

elength scales is the left-hand minimum position; this isbecause in practice it is easier toaccurately locate a voltage mini-mum than a voltage maximum ona line. The angle of the reflectioncoefficient is zero at the opposite,oltage maximum, point. Since

the Smith chart repeats itselfevery half wavelength around aconstant SWR circle, lines longerthan half a wavelength are dealtwith by subtracting multiples of ahalf wavelength from the actualline length.

Lumped L and C circuitsThe Smith chart can also be usedfor the design and analysis of dis-crete L and C circuits. When asingle component L, C or R isadded to a network then either the

resistance (R), reactance (X), conductance (G) orsusceptance (B) parameter of that network willnot change. The point representing the networkimpedance, or admittance, on the Smith chart willtherefore move on a particular chart circle whena single component is added.

By switching between the Y and Z charts andmoving the cursor on constant reactance or con-stant susceptance circles, it is possible to movefrom any one point on a chart to any other. Mov-ing on a Y chart susceptance circle is equivalentto adding parallel inductance or capacitance to anetwork; moving on a Z chart reactance circle isequivalent to adding series inductance or capacit-ance. Using the Z and Y charts in this way, it ispossible to build up networks to impedancematch from any source impedance to any givenimpedance load.

The reference mode facility provided by theZ -MATCH II (see below) program is particularlyuseful in this type of process; the reference modeenables the value of inductance or capacitance re-quired to move between any two points on acircle to be read directly.

With the conventional paper Smith chart therules for the correct direction of movement on theconstant parameter circle need to be known. Z -MATCH II simplifies the procedures involvedconsiderably, by displaying directly the equival-ent value of R, L and Cat the operating frequencychosen. The change in L. C or R can therefore beseen as the cursor moves round any of the con-stant parameter circles.

Z -MATCH IIZ -MATCH II is a CAD package for designingand calculating Smith charts. The basic require-ments to run the package are an IBMPC/XT/AT/386 or PS2 computer, or compatibleclone, running under MS-DOS 2.0 or later, a col-


Z -MATCH II -A RE \ I F \\ a3o

our graphics adaptor (CGA, EGA or VGA) and acolour monitor, and a minimum of 256k of freeRAM. For hard copy of the output an IBMGraphics printer, or compatible, and adaptor cardare required. Z -MATCH II does not require amaths co -processor to be installed, but it will in-crease the speed of operations substantially. Amouse may be used with the software, but, as willbe seen later, owing to the necessity of precisionlocating of the cursor, this method is not whollysatisfactory.

User manualThe presentation of the package is very good. Theuser manual comes in an A4 ring binder, allowingfor updates to be added easily. The software issupplied on one 5.25 -inch 360k disk or one 3.5 -inch 720k disc, both being supplied with thepackage.

The user manual begins with the usual soft-ware and copyright licence conditions and agree-ment, followed by the introduction, installation,running and basic operating instructions. The ma-nual then continues with a brief description ofwhat a Smith chart is and what it can be used for.This is, as is explained in the manual, by nomeans meant to be able to teach you the whets,whys and wherefores of Smith chart use, but is ageneral guide for those already conversant withthe subject. For those less used to working withSmith charts, a tutorial section of workedexamples is included later in the manual.

Following the initial section of the user ma-nual is a comprehensive detailing of all the fea-tures of the package and explanations of all thefeatures, menus and command key functions.

Cursor controlThere are four methods of manipulating the cur-sor around the chart:- in straight lines using the numeric keypad

keys (2, 4, 6 & 8). The speed of movementcan be changed by holding down the'SHIFT' key simultaneously;

- if your keyboard is an enhanced version withseparate cursor control keys, then these canalso be used. In this case the shift key will notchange the speed of movement with thesekeys, but the speed will always be opposite tothat with the numeric cursor control keys;

- the 1 and 3 keys on the numeric keypad('End' and 'Pg Dn') move the cursor aroundselected circles (see below). On the enhancedkeyboard the separate 'End' and 'Pg Dn'keys also do this, but at the slower speed;

- using the mouse, by pressing the left-handbutton, moving the pointer to the desired po-sition and then releasing the mouse button.However, as previously noted, using themouse to position the cursor is less accuratethan using the cursor keys which is the rec-ommended method.

FeaturesAlthough, owing to the complexity of the soft-ware, I am unable to give a complete descriptionof all the features available, I hope to briefly de-scribe some of the major functions.

CirclesThis function allows the user to draw various

circles on the chart:- a circle centred on the current cursor posi-

tion;- a constant conductance circle that passes

through the current cursor position when inthe impedance mode;

- a constant resistance circle that passesthrough the current cursor position when inthe admittance mode;

- a constant Standing Wave Ratio (SWR)circle that passes through the current cursorposition (this circle passes through all the im-pedance, or admittance, points that would bepresent on a half -wavelength of transmissionline with the same SWR and Zo value);

- a unity conductance circle to be drawn on thechart when in the impedance mode.

DisplaysThis allows the method in which the parametersare displayed on the chart to be changed to suitthe user's requirements in the following ways:

- a rectangular or polar coordinate displayshown at the bottom of the screen. This pro-cess serves the same function as using a Car-ter chart overlay with a paper Smith chart;toggling between a wavelength scale fromsource and load to a distance in metres scalefrom source and load, depending on whichconstants or parameters are already known;

- redraw the entire chart, maintaining thevalues of frequency, Zo, etc., already entered,but clearing the display of any circles drawn,whilst maintaining the current cursor posi-tion;

- the ability to enter the various known par-ameters, i.e., characteristic impedance (Zo),frequency and dielectric constant or velocityfactor.

LocateThis set of functions allows the user to easilymanipulate the graphics cursor around the chart:

- the ability to compensate for transmissionline loss, by simply entering the loss in dB,which updates the cursor to a new position,taking the loss into account and giving thecorresponding SWR, impedance, etc.;the ability to move the cursor to a specificpoint on the chart relative to the promptedinput of values for series impedance, paralleladmittance, polar impedance or scatteringparameter,the facility to permanently mark a cursor po-sition on the chart for future reference;

- selection of cursor movement around con-stant resistance or conductance circles, SWRcircles, or the constant reactance or suscept-ance circles.

Other facilities are available which allow the userto switch the program to reference mode, whichenables further calculations to be made with ref-erence to the current cursor position. Also, thedisplay can be switched between series imped-ance and parallel admittance displays, using asingle function key stroke. The background anddrawing colours used in the display can also beuser selected.

Main features of Z -MATCH II

Z -MATCH II displays a Smith chart whichshows:

Actual Impedance and AdmittanceNormalised Impedance and AdmittancePolar ImpedanceDistance towards Generator and LoadReflection CoefficientStanding Wave RatioEquivalent Inductance or CapacitanceZo, Frequency and Dielectric ConstantNetwork Q

Z -MATCH II provides these features:

Conversion between Impedance and Admit-tanceCircle drawingDetermination of the effect of line lossLine Transformer calculationsLocation of any given Z, Y, S or Polar par-ameterMovement of the screen cursor on chartcirclesAmplifier design using S -parametersDisplay of ANALYSER II program outputfiles

Note: ANALYSER II is an advanced AC LinearAnalysis program that calculates and displays thesteady-state AC frequency response of a circuit interms of gain, phase, group delay and input/out-put impedances. It is also available from NumberOne Systems Limited at a cost of £195.00 exclu-sive of VAT and p&p. This package may also bereviewed at a later date.

ConclusionsI found the package easy to use and the results ob-tained were as accurate as those that could be ob-tained by manual charting, but much quicker andeasier to obtain! The ability to quickly changeparameters and observe the changes on the chartis an absolute boon to a designer.

I agree with the comment in the user manualthat a mouse can be used but is not recommended.It is not really possible to place the cursor accur-ately enough using the mouse. However, movingaround the chart with the mouse and then makingfinal precise adjustments with the cursor keysworked fine.

Obviously aimed at the professional RF cir-cuit designer, this package represents excellentvalue for money, especially when taking into ac-count the time that could be saved using such autility, instead of the 'Bob Crachett' methodusing quill and ink. Highly recommended.

I wish to thank Mr. Espin of Number One Sys-tems Limited for his help and advice, and for thereview software.

Z -MATCH II is priced at £195.00 + £4.75 p&p +VAT and is available from: Number One Sys-tems Limited, Harding Way, St.Ives, Hunting-don, Cambridgeshire PE17 4WR. Telephone:(0480) 61778. International: +44 480 61778.Fax: (0480) 494042.


16NI

CON- PULT .-)-) A DrD E

It is hard, if not impossible, to think of professional andindustrial electronics design without a computer enteringat some stage along the long way via concept, designrule checks, electrical checks, PCB design and productmodelling, etc., to a working model that can be used forproduction. In this article we examine some of thesoftware tools (for IBM PCs and compatibles) availablecommercially for schematics drawing, circuit simulationand PCB design.

ALTHOUGH the Bob Cratchit method(quill, ink, and the back of an envelope)

of recording new electronic circuit designsis still widely applied, an increasing numberof electronics designers avail themselves ofa PC to produce circuit diagrams (sche-matics) electronically.

The operation of these so-called sche-matics editors is basically identical. In mostcases, components selected from a libraryare placed on the screen in the desired con-figuration. On completion of the componentplacement phase, the individual parts are in-terconnected by lines, representing electri-cal connections. The circuit diagram somade may be stored on disk, or on the com-puter's memory, or it may be sent to a printeror plotter to obtain hard copy.

Once you are used to seeing only a sec-tion of your drawing on, say, a 14 -inchmonitor, drawing a circuit diagram elec-tronically is hardly more difficult than usingpen and paper for the same purpose. Vir-tually all schematic editor packages allowthe user to zoom in on a certain section of thediagram, and also to view the complete di-agram at a certain reduction factor. Duringthe first attempts at using a schematics edi-tor, the displayed section is nearly alwaystoo small to have a good overview, or thecomplete drawing is shown at a size reducedso far that it is difficult to place a componentat the right place. These problems areusually caused by lack of practical experi-ence, hardly ever by faults in the program.Obviously, a larger monitor solves all thesedifficulties at a stroke, but the cost of, say, ahi-res 19 -inch colour display may be pro-hibitive to many.

Whether or not a schematics editor is auseful tool depends on the size of the com-ponents library, and the user interface.While the components library can usually be

edited and expanded to your heart's desire,there is very little you can do to improve aless than efficient user interface. Fortu-nately, the choice of schematics editors isfairly large, so that fmding one that meetsyour personal requirements at a certain out-lay should not be too difficult.

Circuit simulationA circuit diagram is but the start of a longway towards the production of a workingmodel. The second phase is almost alwaysthe construction of a prototype on whichelectrical measurements are carried out tosee if the circuit (on paper or on the screen)lives up to its expectation. In most cases, itwill not, and some redesigning is in order.Building a prototype from scratch withoutthe least idea whether or not it will work is aproblem to many designers because it is timeconsuming, and requires a basic set of testequipment. Provided you have acquired theskills necessary to work with them, elec-tronics simulation programs are a good wayof preparing for prototype construction.These programs allow you to run a (purelytheoretical) analysis of the electrical perfor-mance of the circuit taken from your sche-matics drawing package. For instance, aninput signal of 100 mV may be simulated,and a frequency sweep of 20 Hz to 20 kHz.The software calculates the resulting voltageresponse, group delay, phase shift and, insome cases, the distortion, at the output ofthe circuit, or at a user -defined point (node)in the circuit. In digital circuits, 'analogue'parameters such as the distortion and thephase shift are usually irrelevant, and otherparameters, such as signal set-up times, de-lays and system timing, are called for to seewhat happens if, for instance, a certain bitpattern is applied to the circuit. In many

CS

cases, digital circuit simulation optionsallow 8 or even 16 outputs to be monitoredsimultaneously.

Whether analogue or digital, all circuitsimulation is based on component par-ameters available to the program. Like sche-matics editors, simulation programs use acomponents library for this purpose. Here,again, the usefulness of the simulation pro-gram depends entirely on the accuracy of thecomponent data, and the number of compo-nents included. A component store -and -editfunction is indispensable for any seriouswork, and quite some time may have to bespent on entering data on new components,the libraries supplied with most simulationpackages being fairly rudimentary. Fortu-nately, most of the leading semiconductormanufacturers supply data on their compo-nents on diskette, in a format that can be rec-ognized by simulation programs.

Circuit simulation is a pretty complexmatter, as you will soon find out from thetime your PC spends on analysing evenbasic circuits like a one -transistor amplifier.Unfortunately, it will be found that a stand-ard PC -XT is hardly suitable for this appli-cation. Also, you will need quite someexperience in numerical circuit analysis be-fore you can use a simulation program inany successful way.

Prototype constructionEven the best simulation program will nottell you if a circuit works reliably. Prototypeconstruction is, therefore, a must after theusual series of corrections and additions tothe circuit. Prototype construction shouldtake as little time as possible, whence theimmediate need for a printed circuit board.Construction on veroboard or stripboard isnot usually considered at this stage becauseof the high risk of errors, and the time re-quired for assembly and debugging.

There are a number of ways to produce aPCB track layout, and dedicated software isavailable for this purpose. For instance,there are programs that enable a track layoutto be drawn only, while other, more ad-vanced, packages are capable of drawingtheir own layout on the basis of the circuitdiagram. Called autorouters, the latter in-variably insist on the user setting certain in-itial values and parameters related to thePCB layout. Even the most sophisticatedautorouters still require components to be


COMPUTER -AIDED ELECTRONICS DESIGN 63

picked up and placed on to the board ma-nually with the aid of the mouse. That iswhere the problems begin for inexperiencedusers: even the best autorouter will not pro-duce an acceptable track layout if the com-ponent placement is not given sufficientthought.

Schematics entryAutoSketch 3.0 plus electronics library is agood example of a schematics drawingpackage without interfaces to PCB layoutsoftware. The program is ideal as a low-costentry level to schematics edition, and can berun from XT -class computers onwards witha minimum of 512 KByte RAM and a 10-MByte hard disk. A mathematics coproces-sor is supported but not strictly required.

AutoSketch is a spin-off of the muchmore powerful AutoCAD program, andboth have been developed by the Swisscompany AutoDesk AG. Originally a low-cost CAD program for machine constructionand architecture, its library extensions forelectrical and electronics applications en-able it to be used for drawing circuit sche-matics.

The program supports a mouse or agraphics tablet as pointing devices, and aplotter, matrix printer or laser printer as out-put devices. The program installation isstraightforward with the aid of the instruc-tion manual. The operation via menus islearned rapidly, which is an advantage if theprogram is used only occasionally. The fin-ished drawing may be saved in the HPGL(Hewlett Packard Graphics Language) orDXF (Data eXchange Format) format. Al-though neither of these is of any use for fur-ther processing by PCB layout programs,they do allow drawings to be copied to DTP(desk -top publishing) or NC (numericalcontrol) programs.

OrCAD SDT W (Schematic Design Tools)is a schematics editor with added featuressuch as annotation, back -annotation, netlistand parts list output, and design rule check-ing. The construction drawing facilities of-fered by the program are modest, but theelectrical circuit drawing capabilities verypowerful. Version 4 of the program has nofewer than 20,000 library items, as com-pared to 6,000 in version 3. Also new is theshell-like user interface, ESP, throughwhich the user has access to the various sub-programs and utilities. ESP is a welcomeand state-of-the-art improvement to theDOS -based parameter entry implemented inprevious versions.

Orcad supports EMS (extended/ex-panded memory system) of any size, doingaway with the dreaded 640-kByte basememory limit in PCs. As to the hardware en-

vironment, OrCad can be used cheerfullyonly when an AT -class PC is used with ahard disk. The package comes with driversfor virtually all of today's graphics adapters.Supplied on four 31/2 -inch disks and com-plete with a hefty manual, OrCad is pro-tected against illegal copying and use by aso-called dongle.

Producing circuit diagrams with OrCadis straightforward work. By virtue of thenew shell, post -editing, for instance, togenerate a parts list, is much simpler thanbefore.

OrCad SDT has a link to OrCad PCB, aseparate PCB layout program. SDT is alsocapable of supplying output to 20 odd stand-ards, including Spice and Calay.

Schema III is a another schematics editor.This program comes on two 1.2-MByte51/4 -inch diskettes, along with two 360-Kbyte disks that contain the library and theutility programs. Schema III is the onlyschematic editor discussed here to provide areal-time pan (panorama) function, whichmeans that the visible section of the drawingis moved over the screen automaticallywhen the cursor reaches one of the borders.Other programs may offer a similar func-tion, but invariably redraw a section of thescreen. By contrast, Schema III ensures thatthe full section is always visible. As withOrCad, the subprograms that make upSchema III are accessible via a user inter-face called the Manager.

Schema m supports HP, Calcomp andHouston Instruments plotters, and savesdrawings in the 111-1-, DXF or Postscript for-mat. It provides the basic schematic editingfunctions: component replacement, compo-nent loading, parts list and netlist gener-ation, and design rule check. According tothe manufacturer, it can be run on PC -XT,AT and 386 -based computers with a mini-mum of 640 KByte RAM and 4-MByte offree space on the hard disk. A mathematicscoprocessor is not supported.

Schematic 3 is a schematic editor in a seriesof CAD products from the Australian soft-ware house Protel. The program is ideal forfast drawing of circuit diagrams and the

generation of netlists and parts lists (`bills ofmaterial'). Remarkably, the package has anintegrated word processor, which allowstext to be placed freely in the circuit diag-ram.

The design rule check is performed as aseparate utility, i.e., not during the entryphase (on line), which is more usual. Errors,if found, are written to a file. The package isquite extensive, and requires an AT -classcomputer (or faster) to run at acceptablespeed. EMS is fully supported.

The libraries number a total of 15, andcontain over 3,000 components. Among theoutput devices supported are HPGL and Ro-land plotters, Epson (compatible) matrixprinters and HP (compatible) laser printers.Postscript output is not supported. The pro-gram, which is supplied on 5 diskettes, is notprotected by a dongle or any other means.

Combination packagesCombination packages avoid problems thatmay occur in standards exchange and fileconversions, by integrating the functions ofa schematic editor and a PCB design pro-gram. Almost all combination packagesoffer an autorouter.

Eagle 2.05 is a combination package withextensive libraries and an autorouter capableof handling double -sided board of10x16 cm (Eurocard size) either automat-ically or interactively. The PCB layout pro-gram offers 1/moo-inch resolution, andvarious grids on a PCB surface of 1.6 m2maximum (approx. 64x64 in.). Each PCBtrack can have its own width. The operationof the autorouter can be followed on thescreen, and may be interrupted and conti-nued at any time. The router has been op-timized for two-sided boards, and has apreference for taking directions at an angleof 90°. The current version of Eagle can notbe called state-of-the-art with its 50 -mil gridand its inability to handle SMA (surface -mount assembly) components. However,the manufacturer has promised that a smal-ler grid and SMA support will be included ina future release of the program. The newrouter, to be released towards the end of this


64 GENERAL INTEREST

year, is said to offer a resolution of 4 mils atno fewer than 12 layers. SMA should also besupported.

The Eagle package supports all currentprinter and plotter formats, as well as Post-script. A converter supplied with the pro-gram allows netlists and libraries made withOrcad to be loaded.

Eagle is simple to install on an XT or ATcomputer running under DOS 3.30 or later.The program comes on three disks, and isprotected by a dongle plugged into theprinter port of the computer. While testingthe Eagle package, this caused a problem be-cause on our computer the port happened tobe in use for an NEC P2200 printer.

Eagle distributors should be able to sup-ply a demonstration disk at nominal cost.

Boardmaker 1 is an inexpensive programthat combines a PCB design program and aschematic drawing program. PCB layoutscan be edited, and the program comes with acomponent library and a library editor.

Boardmaker supports Excellon NC drill,Gerber photoplotter and Postscript outputformats, but unfortunately lacks a netlistoutput. The upgraded version, Board -maker 2, does include a netlist output, but isconsiderably more expensive than the stand-ard version.

Boardmaker 2 version 2.30 comes ontwo 51/4 -inch diskettes, and is complete witha ring -bound instruction manual. The pro-gram is capable of accepting and processingnetlists produced by Schema II, Schema III,Protel, Tango Schematic and OrCad.

Tailor-made for Boardmaker 2 is TsienLaboratories' Boardrouter, an interactiveautorouter. A number of different par-ameters included in the command string thatcalls up the program enable either sectionsof the board, or the entire board, to berouted. The router can be interrupted and re-started at any time. Boardrouter is protectedby a dongle.

The Ranger 1 package offers a schematiceditor and a PCB layout editor with auto-router. Ranger 1 has no on-line design rule

checker. It requires an AT or 386 -based PCwith a 40 MByte (or larger) hard disk,640 KByte of RAM, and DOS version 3.30or later. The package includes two 51/4 -inchdiskettes, a manual, and a dongle for copyprotection.

The schematic drawing program canhandle up to eight pages. Components takenfrom the library can be placed, rotated andmirrored. It is also possible to design newcomponents and add them to the library.

A small word processor enables netliststo be imported, that is, circuit diagrams can,in principle, be made without any drawingactivity on part of the user.

The program uses the schematic togenerate the parts list and the netlist. Afterdefining the size of the PCB (max.32x32 in.), the netlist can be used by thePCB design software. Next, the parts areplaced on to the board using the well-knownrubber -band technique. The Lee track routercan be used in manual, interactive or auto-matic mode. The track placement is directlyvisible on the screen. The router uses a 1 -milgrid.

The program has an option to check therouted board for short-circuits, missing con-nections or too close by tracks. Errors foundare immediately indicated on the screen.

Challenger is a combination package withan excellent price/performance ratio. Sup-plied by Ultimate Technology, it uses theUltishell user interface to integrate the Ulti-cap schematics editor and the UltiboardPCB design program. Both Ulticap and Ul-tiboard are 'small' versions of much moreexpensive high -end programs. Here, 'small'has nothing to do with the performance -only the maximum PCB size is reduced toabout 700 pins, corresponding roughly totwo times the Eurocard size.

Challenger supports EMS for schematicsentry as well as board design, an on-line de-sign rule check, a library containing about4,000 components, and interfaces to Work -

View, DASH, Orcad and Schema. Apartfrom the very useful 'force vector' and his-togram utilities for component placement,

the program offers a 'trace shoving' func-tion in the layout phase. This function allow-s PCB tracks to be moved to make room fornew ones, without having to delete and repo-sition the ones that had been drawn already.

The integrated autorouter, Ultiroute, canhandle up to 32 PCB layers simultaneously.Two remarkable features of the Challengerpackage are the automatic re -wiring aftercomponent repositioning in the schematicseditor, and the 'hardware zoom' support forcertain graphics adapters. In additions tothese, Challenger offers the usual functionsof auto -panning, parts list and netlist gener-ation, and data output to several plotter andNC standards, as well as to two differentPostscript formats.

The Challenger package is not copy pro-tected, and comes with two extensive, well -organized user manuals.

Circuit simulation softwareElectronic circuit simulation on a shoestringdoes not exist, and is marked by hugeamounts of data and a high calculation loadon the processor. Professional simulationprograms such as Accusim require a 400-MByte hard disk, 16 MBytes of main RAMmemory, and a Sun-SPARC WorkstationType 2 to work comfortably.

Two simulation packages that can beused on PC -AT computers have found theirway to the advanced hobbyist and the labor-atory worker. One of these, Microcap III, is

Min IttilkW 11C.1.3

_tiors Intl } at, Itn_esunt I ts rOminme

supplied in several versions, each with aspecific maximum number of nodes (com-ponent junctions). The student version, forinstance, can handle up to 30 nodes, how-ever does not include a graphics editor or aMonte -Carlo analyser function, and has arelatively small, but extendable, compo-nents library.

Microcap III offers four ways of examin-ing an electronic circuit: a.c., d.c., transientor Fourier analysis. Initially, the program isused to draw the circuit, which is achievedby loading components from the library, andplacing them on the screen. Next, the par-ameters to be examined are entered, and a


COMPUTER -AIDED ELECTRONICS DESIGN

predefined signal source is 'connected' tothe input. The program calculates the re-sponse of the circuit, and shows the resultsin the form of a high -resolution graph. Thelibrary editor is, of course, much more ex-tensive than that of a schematics drawingprogram, since the number of characteristicsrelated to a new component is much higher.Consider, for instance, the following basicparameters of a transistor. current gain, cur-rent gain in an inverting circuit, temperaturecoefficient of the current gain, junction re-sistance, saturation current, etc.

The Monte -Carlo analysis is particularlyuseful to check the reliability of the circuitin regard of component tolerance. With theaid of this option, the program creates com-ponent deviations (within the specifiedtolerance) in a quasi -random manner, afterwhich a new analysis is run. The Monte-

Carlo analysis thus produces a very usefulinsight into the performance distribution ofa particular circuit intended for larger -scaleproduction.

Microcap runs on MS-DOS compatiblecomputers of the AT class and higher, sup-ports expansion memory, and makes use ofa coprocessor, if installed. The programcomes on four 51/4 -inch floppy disks, and iscomplete with a user manual. The installa-tion should be carried out with care, as it isrecorded on one of the disks as a means ofcopy protection.

The other widely used simulation programfor IBM PCs is Pspice from Microsim Corp.Pspice requires an AT, 386 or 486 computerequipped with a maths coprocessor. Pspiceis available in three versions, ranging fromPspice Evaluation to Pspice DeLuxe.

Pspice consists of several parts: a sche-matics editor, an analyser, a signal sourceeditor, a graphics output function (calledProbe), and a utility to establish componentparameters (called Parts). In addition tothese, the package offers Monte -Carlo anddigital simulation utilities. All programfunction are accessed via a user interfacecalled the Control Shell.

Pspice supports direct circuit entry with

the aid of the mouse, as well as entry viaparts lists or netlists supplied by the Orcadschematics editor.

PCB design packagesThese are basically drawing programs tai-lored to meet the often exacting demands ofPCB (printed circuit board) designers.

Rule is a simple and inexpensive programcapable of producing PCB layouts that meetrelatively high standards. Offering the bestprice/performance ratio in its class, Rule hasbut one major disadvantage: practically alldesign work has to be done manually.

Rule can be used on any PC/XT/AT with640 Kbyte of RAM, one floppy disk drive

higher. The program supports Hercules.CGA, EGA and VGA display adapters.which are automatically recognized. The in-stallation is simple and without problems,and may be done on a hard disk. The soft-ware is not copy protected. The 5 V4 -inchfloppy disk contains the main program,drivers for Epson compatible printers (8-, 9 -and 24 -pin types) and HP-LaserJet com-patible printers, a components library and afew sample PCB layouts. A 20 -page usermanual (in English) completes the package.

Rule handles up to 16 layers in a PCBwith a maximum size of 23x23 cm. The pro-gram is capable of generating the artworkfor the component overlay, the drilling tem-plate and the solder resist mask. The PC key-board is used for the basic programfunctions: select, delete, move, rotate, mir-ror and copy. These functions may also beselected with the mouse. Lacking an auto -router, Rule forces you to draw tracks fromsolder pad to solder pad. Fortunately, this issimple to learn, and can be done quitequickly after a while. Output drivers forPostscript, Gerber and Excellon, as well asan extended components library are avail-able separately as upgrades. A HPGL plotterdriver is in preparation.

Easytrax from Protel is a PCB design pro-gram with an autorouter, offering an excel-lent price/performance ratio. The current

version is 2.06, supplied on four 51/4 -inch360-KByte floppy disks. The program'handles PCBs with a maximum size of32x32 inch, and up to 10 layers (6 of whichare for tracks). Easytrax has an integral li-brary which, unfortunately, does not includeSMA components. Simple automatic pan-ning is supported, as well as variable zoomfactor setting.

The copper track width can be set to be-tween 12 and 100 mils, while solder padscan have any diameter between 40 and250 mils. On completion of the layout work,the utility Easyplot is used to export the art-work in the HPGL, Calcomp or Roland plot-ter format, complete with Gerber orExcellon photoplot or drill files. Postscriptoutput is also supported. The autorouter inthe Easytrax package is a simple pad -to -padrouter that draws a track between two solderpads selected with the mouse.

Another Protel product, Autotrax, is an ex-tended, much more powerful, version of Ea-sytrax. Not surprisingly, it is also muchmore expensive. Although the maximumPCB size is 32x32 inch, Autotrax handlesup to 13 layers, of which 6 are used for sig-nals. The use of SMA components is fullysupported. The program allows track widthsbetween 1 and 255 mils, and offers 6 differ-ent solder pad shapes of a diameter between1 and 1,000 mils.

The autorouter can be used with any pre-defined track width and design grid, whichis of great importance when working withSMA components. Unlike Easytrax, Auto-trax has a netlist importing function, whichwidens the application range considerably.The program reads netlists produced bySchematic, Tango Schematic, Orcad SDTand Schema. After having the program readthe netlist, you can run the Auto -Placer. Thisutility places the components on to the PCBin a manner that results in the shortestpossible tracks and the lowest possible num-ber of through contacts. The decisions of theAuto -Placer can be overridden manually, al-though this is slightly cumbersome. Auto-trax is copy protected by a dongle attachedto the printer port of the computer.


r \ ERAL INTEREST

Layo 1 Plus is offered in two versions, aJunior version (Layo Plus Junior) and a full-blown version (Layo Plus). The two ver-sions differ only in regard of the maximumPCB complexity. Version 4.8 of Layo Plusis an extremely powerful, all-round PCB de-sign program offering flexible editing fea-tures, schematic capture, netlist import(from, among others, OrCad SDT III), anextensive components library containingnearly 500 components (shapes), SMA sup-port, forward annotation, design rule check,project management, real-time rubberband-ing, macros and high-speed auto -routing.Layo Plus is capable of workin with16 layers in a PCB of 650x650 mm maxi-mum size. Remarkably, every layer can beshown individually as well as in combina-tion with any other layer(s). The resolutionis Vino inch, or about 0.0019 mm. Theminimum system requirements to run thePlus version are MS-DOS version 3.3x, anEGA card and monitor, a hard disk with5 MBytes of free space, and an 8 -pin matrixprinter. A more comfortable system con-figuration would, however, look somethinglike this: MS-DOS 4.xx, a VGA card andmonitor (resolution up to 1,024x 768), a 40-MByte hard disk, EMS 4.0 with a minimumof 2 Mbytes of extended memory, and alaser printer and/or a plotter.

Input to the program is via a mouseand/or keyboard. A wide variety of outputformats (Postscript, DXF and others) andoutput devices is provided, includingprinters, plotters, photoplotters and drillingmachines. Layo. Finally, Layo is copy -pro-tected by a dongle.

AutoSketchAutodesk, Inc.World Headquarters2320 Marinship WaySausalito, CA 94965U.S.A.Telephone: (415) 332 32 44Fax: (415) 331 80 93

Autodesk Ltd.Cross Lanes

GuildfordSurrey GUI 1 UJUnited Kingdom.Telephone: (0483) 303 322Fax: (0483) 304 556

BoardmakerBoardrouterTsien (UK) LimitedCambridge Research Laboratories181A Huntingdon RoadCambridge CB3 ODJUnited Kingdom.Telephone: (0223) 277777Fax: (0223) 277747

EagleAmerican Small Business Comp. Inc.One American WayPryor, OK 74361U.S.A.Telephone: (918) 825 63 59Fax: (918) 825 4844

P.M.S. Instruments Ltd.Waldeck HouseReform RoadMaidenheadBerkshire SL6 8BRUnited Kingdom.Fax: (0628) 770 562

Easytrax, Autotrax, Schematic 3Protel Technology Pty Ltd.Technopark Dowsing PointHobartTasmania 7010Australia.Telephone: (+61) 02 730100Fax: (+61) 02 730944

Layo 1International HeadquartersBaas ElektronikaRijksstraatweg 423281 LW NumansdorpHolland.Telephone: (+31) 1865 4211Fax: (+31) 1865 3480

distributor for the UK:Pentagram Electronic DesignMr. Andy Coe6 Pasture CloseBradford BD14 6LYUnited KingdomTelephone: (0274) 882609Fax: (0274) 882295

Microcap IIISpectrum Software1021 S. Wolfe RoadSunnyvale, CA 94086U.S.A.Telephone: (408) 738-4387

OrCad SDT IVOrCad Systems Corporation1049 S.W. BaselineSuite 500HillsboroOregon 97123U.S.A.Telephone: (503) 640-9488Fax: (503) 640-6491

PspiceMicroSim Corporation20 FairbanksIrvine, CA 92718U.S.A.Telephone: (714) 770-3022

Ranger 1distributor addresses fromConrad ElectronicKlaus-Conrad-Strasse 1D-8452 HirschauGermany.Telephone: (-1-49) 9622 30111Fax: (+49) 9622 30265

RuleIngenieur-Biiro FriedrichSudetenstrasse 14D-6405 EichenzellGermany.Telephone: (+49) 6659 2249Fax: (+49) 6659 2249

Schema IIIOmation801 Presisdential DriveRichardson, TX 75081-9990U.S.A.

Ulticap, UltiboardUltimate TechnologyInternational HeadquartersEnergiestraat 361411 AT NaardenHolland.Telephone: (+31) 2159 44424Fax: (+31) 2159 43345

UK Sales & Support Office2 Bacchus HouseCalleva ParkAldermastonBerkshire RG7 4QWUnited Kingdom.Telephone: (0734) 812030Fax: (0734) 815323

Note: further CAD software for schematicsdrawing and PCB design is available fromNumber One Systems (0480) 61778, andLabcenter Electronics (0274) 542868. Forbasic product information, please refer tothe relevant advertisements in this and pre-vious issues of Elektor Electronics.


Please mention ELEKTOR ELECTRONICS when contacting advertisers

POWER AMPLIFIER MODULES -TURNTABLES -D MERS-LOUDSPEAKERS-1S INCH STEREO RACK AMPLIFIERS

* ----- dal d- * AI SERVICE * LARGE LAM r STAMPED FOR CATALOGUE *

OMP MOS-FET POWER AMPLIFIERSHIGH POWER. TWO CHANNEL 19 INCH RACK

THOUSANDS PURCHASEDBY PROFESSIONAL USERS

,1r.I.11.,111-43246.11ra at.1!,121112I .1,111111111 SUPPLIED READY BUILT AND TESTED.nese modules now enjoy a etorkt-wide reputation tor crsalay, rehab...Ay and parkehuteire VI IIIVISLC pm. Picamodels are arailabie to stet the needs ot the professional and hobby mutat La. Irekestry. Leince.. Inshurnental andet. When comparing prices. NOTE that all models include toroidal pow*, supply. integral heat sire._ glass fibre PC B. anddn.. cirvas to power a co4.04b414 Yu meter All models ate open ar4 shod circul pool

THOUSANDS OF MODULES PURCHASED BY PROFESSIONAL USERSOMP/MF 100 Mos-Fet Output power 110 wattsR.M.S. into 4 ohms, frequency response 1Hz - 100KHz-3dB. Damping Factor :300, Slew Rate 45V/uS,T.H.D. typical 0.002%. Input Sensitivity 500mV. S.N.R.-110 dB. Size 300 z 123 a 60mm.PRICE £40.85 - £3.50 PAP

THE RENOWNED MXF SERIES OF POWER AMPLIFIERS -FOUR MODELS:- MXF200 (100W - 100W) MXF400 (200W - 200W)

MXF600 (300W - 300W) MXF900 (450W - 450W)ALL POWER RATINGS R.M.S. INTO 4 OHMS, BOTH CHANNELS DRIVEN

FEATURES: *Independent power supplies with two toroidal transformers * Twin L.E.D. Vu meters *Level controls * Illuminated Wort switch * XLR connectors * Standard 775mV inputs * Open arid short circuitproof * Latest Mos-Pets for stress free power delivery into virtually any load * High slew rate Very lowdistortion * Aluminium cases * SAXES.° 8 81XFS00 fan cooled with D.C. loudspeaker and thermal protection.

USED THE WORLD OVER IN CLUBS, PUBS, CINEMAS, DISCOS ETC.SIZES:- MXF200 W1 9"sH3' r" (2U)oD1 1"

MXF400 W113"sH5' (3U):131 2"MXF600W19"1115' (3U)s1:113"MXF900W19":H5' (3U)sD14.."

PRICES:- MXF200 £175.00 MXF400 £233.85MXF600 C329.00 MXF900 0449.15SPECIALIST CARRIER DEL 01250 EACH

111,..1:&71 ki 4.11:1:k Ci xi

* Manual arm * Steel chassis * Electronic speedcontrol 33 & 45 R.P.M. * Van pitch control * Hightorque servo driven DC motor * Transit screws *12' die cast platter * Neon strobe * Calibratedbalance weight * Removable head shell * 'r'cartridge fixings * Cue lever* 220/240V 50/60Hz* 390x305mm * Supplied with mounting cut-outtemplate.

PRICE £61.30 - £3.70 P&Pmaitehrmi,i_tehimiotoui.it.tilici*. STANTON AL500MktI GOLDRING G950

PRICE C115.95 - 50P PRICE £7.15 - 50P[4:14.1 ):4 4:i 0711.44Ix;

STEREO DISCO MIXER 'inn 2 a 7 bandL 8 R graphic equalisers with bar graphLED Vu meters MANY OUTSTANDINGFEATURES:- including Echo with repeat &speed control, DJ Mit with tone control

talk -over switch, 7 Channels withindividual faders plusHeadphone Monitor. Usehill combination ofthe following inputs:- 3 turntables (mag), 3mica, 5 Line for CD, Tape, Video etc.

Price £134.99 - £5.00 P&P

* WITH ECHO*

SIZE 482 a 240 a 120rrim

Join the Piet. revolution! The low dynamic mass (no voice coil) of a Piezo tweeter produces an improvedtransient response with a lower distortion level than ordinary dynamic tweeters. As a crossover is not requiredthese units can be added to ezisting speaker systems of up to 103 watts (more if two an put in series. FREEEXPLANATORY LEAFLETS ARE SUPPLIED WITH EACH TWEETER.

TYPE 'A' (KSN1036A) 3' round with protective wire mesh. Ideal forn- bookshelf and medium sized Hi.Fi apeakers. Price C4.90 - 50p PAP.

%," TYPE 'It' (KSN1005A) 3'/,' super horn for general purpose speakers.disco and P.A. systems etc. Price £5.99 - 50p PAP.TYPE 'C' (KSN1016A) 2'x5' wide dispersion horn for quality Hi-Fi sys-tems and quality discos etc. Price 05.99 - 50p PAP.TYPE '0' (KSN1025A) 2"x6" wide dispersion horn. Upper frequencyresponse retained extending down to mid -range (2KHz). Suitable for highquality Hi-Fi systems and quality discos. Price C9.99 - 50p PAP.TYPE 'E' (KSI11038A) Pr' horn tweeter with attractive silver finish trim.Suitable for Hi-Fi monitor systems etc. Price C5.99 - 50p PIP.LEVEL CONTROL Combines, on a recessed mounting plate. level controland cabinet input jack socket. 85185mm. Price 04.10 - 50p PAP.

TYPE C TYPE.

.0.'.12111:1:14iikellilizi:141:1 :1:4

THE VERY BEST IN QUALITY AND VALUE

Yaae espevally to scat toda, s need for compactness mm high outputsc...-d levels. finished in hard wearing olack vynide with protectiveccrners grille and carrying handle Each unit incorporates a 12' driverplus high frequency horn for a lull frequency range at 45Hz.20Hz. Bothmodels are 8 Ohm impedance. Size: H20" s W15 012'.

CHOICE OF TWO MODELS

POWER RATINGS QUOTED IN WATTS RMS FOR EACH CABINET

OMP 12-100WAITS (100dB) PRICE C153.50 PER PAIROMP 12-200WATM (200.113) PRICE C214.55 PER PAIR

SPECIALIST CARRIER DEL E1250 PER PAIR

OMP/Mr 200 Mos-Fet Output power 200 wattsR.M.S. into 4 ohms, frequency response 1Hz - 100KHz-3dB. Damping Factor - 300, Slew Rate 50V/uS.T.H.D. typical 0.001%, Input Sensitivity 500mV, S.N.R.-110 dB. Size 300 x 155 x 100mm.PRICE £64.35 - C4.00 P&P

OMP/MF 300 Mos-Fet Output power 300 wattsR.M.S. into 4 ohms, frequency response 1Hz - 100KHz

p -3dB, Damping Factor .300, Slew Rale 60V/uS.T.H.D. typical 0.001%. Input Sensitivity 500mV. S.N.R.-110 dB. Size 330 x 175 x 100mm.PRICE £81.75 - C5.00 PAP

OMP/MF 450 Mos-Fet Output power 450 wattsR.M.S. into 4 ohms. frequency response 1Hz - 100KHz-3dB, Damping Factor -300, Slew Rale 75V/uS,T.H.D. typical 0.001%, Input Sensitivity 500mV. S.N.R.-110 dB. Fan Cooled, D.C. Loudspeaker Protection. 2Second Anti -Thump Delay. Size 385 x 210 x 105mm.PRICE 0132.85 - £5.00 P&PNOTE: 1105 -PET NODULES ARE AVAILABLE IN TWO VERSIONS:STANDARD. INPUT SENS SOOrnV, BAND WIDTH I OOK Ka.PEC (PROFESSIONAL EQUIPMENT COMPATIBLE) - INPUT SENS775mV, BAND WIDTH SOKHz. ORDER STANDARD OR PEC.

Vu METER Compatible with our four amplifiers detailed above. A very accuratevisual display employing 11 L.E.D.s (7 green. 4 red) plus an additional onroffindicator. Sophisticated logic control for very fast rise and decay times_ Toughmoulded plastic case, with acrylic tinted front. Size 84 a 27 a 45mm.PRICE £8.70 50p PAP

LARGE SELECTION OF SPECIALIST LOUDSPEAKERSAVAILABLE, INCLUDING CABINET FITTINGS, SPEAKERGRILLES, CROSS-OVERS AND HIGH POWER, HIGHFREQUENCY BULLETS AND HORNS, LARGE (A4) S.A.E.(50p STAMPED) FOR COMPLETE LIST.P -From McKenzie Professional Seriesi`s- From McKenzie Studio Series

1;1'41 121:1-111:11I!,14:11izilL'Aillgiefoli41401ALL McKENZIE UNITS 8 OHMS IMPEDANCE8' 100 WATT P C8-IOOGP GEN. PURPOSE. LEAD GUITAR. EXCELLENT MID. DISCO.RES. FREQ. 80Hz. FREO. RESP. TO 7KHz. SENS 96dB. PRICE C31.45 - C2.00 PAP10' 1 oowArr s Cl 0-1 00GP GUITAR. VOICE. KEYBOARD. DISCO. EXCELLENT MID.RES. FREO. 72Hi, FRED. RESP. TO 6KHz. SENS97dB. PRICE 038.89 - C2.50 PIP10' 200WATTS Cl 0-200GP GUITAR. KEYB'D. DISCO. EXCELLENT HIGH POWER MID.RES. FREQ. 69111. FREQ. RESP. TO 5KHz. SENS 97dB. PRICE 053.21 - C2.50 PIP12' 100WATT P C12-100GP HIGH POWER GEN. PURPOSE. LEAD GUITAR. DISCO.RES.FREO. 49HEFREO. RESP. TO 7KHz. SENS 98dB. PRICE £40.35 - C3.50 P&P12' loowarr P Ct 2.100TC (TWIN CONE) HIGH POWER. WIDE RESPONSE, P.A.. VOICE. DISCO.RES. FREO 45Hz. FREO. RESP. TO 12KHz. SENS 97d8. PRICE 041.39 - £3.50 PIP12" 200 WATTS Cl 2-200B HIGH POWER BASS, KEYBOARDS. DISCO. P.A.RES. FREO. 45H1. FREQ. RESP. TO 5KHz. SENS 99dB. PRICE 071.91 - C3.50 PIP12' 300WATTS C12 -3000P HIGH POWER BASS. LEAD GUITAR. KEYBOARDS. DISCO ETC.RES. FRED. 49Hz. FRED. RESP. TO 7KHz. SENS 100d8. PRICE 095.66 - C3.50 PAP15" IOOWATT PC15-10085 BASS GUITAR. LOW FREQUENCY. PA.. DISCO.RES. FREO. 40111, FRED. RESP. TO 5KHz. SENS 9848. PRICE C 59.05 - C4.00 PIP15" 200WATT P CI 5.200B5 VERY HIGH POWER BASS.RES. FREO. 40Hi. FREO. RESP. TO 3KHz.SENS 984B. PRICE C80.57 C4.00 P&P15' 250WAT1 S C15-250BS VERY HIGH POWER BASS.RES. FREQ. 39Hz. FREO. RESP. TO 4KHz. SENS 99d8. PRICE C90.23 - £4.50 PAP'15' 400WATTS C15-400115 VERY HIGH POWER. LOW FREQUENCY BASS.RES. FRED. 40Hz. FREQ. RESP. TO 4 KHz. SENS 100d8. PRICE £105.46 - C4.50 PIA119" 500WATTS C18-500135 EXTREMELY HIGH POWER. LOW FREOUENCY BASS.RES_ FRED. 27Hi. FRED. RESP. TO 2KHz. SENS. 98dB. PRICE £174.97 - 25.00 PIP7.171715.111511nalraraITTIEWEWirelALL EARBEN DER UNITS 8 OHMS tEacrpt ESS5.1 & £915.50 which are dual impedance tapped 4 A 8 ohm)B ASS, SINGLE CONE, HIGH COMPLIANCE, ROLLED SURROUND8" SOwatt EBB -50 DUAL IMPEDENCE. TAPPED 418 OHM BASS. HI -F1. IN -CAR.RES. FREQ. 40Hz. FRED. RESP. TO 7KHz SENS 97dB. PRICE £8.90 - C2.00 PIP113' 50WATT EB10-50 DUAL IMPEDENCE. TAPPED 418 OHM BASS. HliFt. IN -CAR.RES. FRED. 40Hz, FREQ. RESP. TO 5KHz. SENS. 99d13. PRICE C13.65 C2.50 P&P10" 1 00WATT EB1 0-1 00 BASS. HI -Fl. STUDIO.RES. FREQ. 35Hz. FRED. RESP. TO 3KHz, SENS 96dB. PRICE 230.39 - £3.50 PIP12- 100WATT EB1 2-100 BASS. STUDIO. HI-FI, EXCELLENT DISCO.RES. FREQ. 26Hz. FRED. RESP. TO 3 KHz. SENS 93dB. PRICE 042.12 - C3.50 PAPFULL RANGE TWIN CONE, HIGH COMPLIANCE, ROLLED SURROUND

60WATT E85-150TC (TWIN CONE) HI-FI. MULTI -ARRAY DISCO ETC.RES. FRED. 63Hz. FREQ. RESP. TO 20KHz. SENS 92dB. PRICE C9.99 - £1.50 PIP6' 2- 60WATT EB6-60TC (TWIN CONE) HI-FI, MULTI -ARRAY DISCO ETC.RES. FRED. 38Hz. FRED. RESP. TO 20KHz. SENS 94dB. PRICE 010.99 - 1.50 PIP8' 60WATT E198-60TC (TWIN CONE) HtFl. MILTI-ARRAY DISCO ETC.RES. FRED. 40Hz. FRED. RESP. TO I 8KHz. SENS 89413. PRICE 012.99 - C1.50 PAP10" 60WATT EB10-60TC (TWIN CONE) HI-FI. MULTI ARRAY DISCO ETC.RES. FREQ. 35Hz. FRED. RESP. TO 12KHz. SENS 98dB. PRICE (16.49 - C2.00 PIP

PRICES: 150W 049.99 250W 099.99400W C109.95 PAP C2.00 EACH

THREE SUPERB HIGH POWERCAR STEREO BOOSTER AMPLIFIERS150 WATTS (75 - 75) Stereo. 150WBridged Mono250 WATTS (125 - 125) Stereo. 250WBridged Mono400 WATTS (200 - 200) Stereo. 400WBridged MonoALL POWERS INTO 4 OHMSFeatured* Stereo. bridgable mono * Choice Clhigh & low level inputs * L 8 R levelcontrols * Remote oniort * Speaker 8thermal protection

PROVEN TRANSMITTER DESIGNS INCLUDING GLASS FIBREPRINTED CIRCUIT BOARD AND HIGH QUALITY COMPONENTS

COMPLETE WITH CIRCUIT AND INSTRUCTIONS3W TRANSMITTER 10-1DLVIIL WACO CONTROLLED PROFESSIONALPERFORIUKE RANGE UP TO) If LE& SIZE 11.121,00. SPINS tN .:: 011.111.

PIKE C14.115 - MOO PAPFM MICRO TRANSMITTER 103-13136h., UNCAP IIINED. COMPUTE WMVERT 50365 FET EC. RANGE 1C0-30:wis. Siff 56 a Nam RIPPLY SV BATTERY.

PHOTO: 3W FM TRANSMITTER

B.K. ELECTRONICSinPOSTAL CHARGES PER ORDER E1.00 MINIMUM. OFFICIAL

ORDERS FROM SCHOOLS, COLLEGES, GOVT. BOOMS, PLC. ETC-PRICES INCLUSIVE OF VAT. SALES COUNTER. VISA AND

ACCEEIS ACCEPTED BY POST, PHONE OR FAX.

EI.EKTOR ELECTRONICS NOVEMBER 1991

UNITS "I S. 5 COMET WAY, SOUTHENO-ON-SEA,ESSEX. 552 BTR.

Tel.: 0702-527572 Fax.: 0702-420243

5

0 <IT

.0

VC

bdC

s0'1

k:"

with this issue 132 -page greenweld catalogue class a vilifieramplifier remains relatively cool...

Documents