.1! tortimies *WIN -a - americanradiohistory.com · Printed by Livingstone Printing Ltd. News Stand...

*WINittrautitsp. .1! tortimies

-a Faro ovitt origtaiit I " ON OW fit

VOL. 2 NO. 3MARCH 1978

EditorSTEVE BRAIDWOOD BSc

Assistant EditorGRAHAM WIDEMAN RASc

Marketing Manager and AdvertisingPETER E. PRIEST

Advertising ServicesSHARON WILSON

Advertising RepresentativesJIM O'BRIEN

Eastern Canada JEAN SEGUIN &ASSOCIATES INC., 601 Cote Vertu,

St. Laurent, Quebec H4L 1X8.Telephone (514) 748-6561.

Subscriptions DepartmentBEBE LALL

Accounts DepartmentSENGA HARRISON

Layout and AssemblyGAIL ARMBRUST

Contributing Audio EditorWALLACE J. PARSONS

Editorial Director TOM GRAHAM.Published by Electronics Today

International (Canada) Ltd.

EDITORIAL AND ADVERTISINGOFFICES

Unit 6, 25 Overlea Boulevard,Toronto, Ontario, M4H 1B ITelephone (416) 423-3262

Printed by Livingstone Printing Ltd.

News Stand Distribution Gordon &Gotch, Toronto.

Subscription Rates $12.00 per year,$20.00 for two years. Send to Sub-

scription Dept., ETI Magazine. Unit 6.25 Overlea Blvd.. Toronto,

Ontario, M4H 1 B1.

INTERNATIONAL EDITIONSElectronics Today International

25-27 Oxford:St., London W1 R 1 RF, UK.Editor Halvor Moorshead.

Electronics Today International,Ryrie House. 15 Boundary St.,

Rushcutters Bay, Sydney, AuStraliaEditor Collyn Rivers.

Electronica Top lnternationaal,Postbus 260, Emmen, Holland.

Editor Denis Loos.

Elrad,Kommanditgesellschaft, Bissendorfer

Strasse 8, 3000 Hannover 61, Germany.Editor lido Wittig.

Next Month's ETIETI Panel TransfersETI SubscriptionsETI Binders

electronics treggyCANADA'S OWN ELECTRONICS MAGAZINE

PROJECTSHAMMER THROW GAME 14An LED & Logic version of the olympic sport

TRUE RMS VOLTMETER 22Using an IC gives improved performance

HOME BURGLAR ALARM 28A sophisticated CMOS design

FEATURESEQUALISATION 8Wally Parsons explains one of the fashionable concepts in audio

DATA SHEET 37Super sound effects chip from TI.

THE C2Oth JUNGLE TELEGRAPH 40The amateur radio traffic network

BITS, BYTES, AND BAUDS 43Bill Johnson shows how easy computers can be

MICROBIOGRAPHY 47This month a look at the Signetics 2650

ETI SOFTSPOT 55Software for your programmable calculator

TECH -TIPS 58Ideas for experimenters

FEEDBACK 63Letters from our readers

THE FUN OF ELECTRONICS 65Just for a laugh

NEWS & INFORMATIONNews Digest 4Microfile 6

Canadian Projects Book 727353639

Second Class Mail registration number 3955.Return postage guaranteed. Post Office returnsto Unit 6, 25 Overlea Boulevard, Toronto,Ontario, M4H 1B1.

ETI Circuits Book 39ETI Marketplace 51

Club Call 51

ETI Publications 64Information 66Classified Advertising 66

Copyright: All material is subject to world-wideCopyright protection. All reasonable care istaken to ensure the accuracy of the information.

3

What Is Equalization?A dictionary would tell you that to equalize is to make equal or uniform.Wally Parsons, ETI's Contributing Audio Editor, discusses.

ALL MEN (and women) are said to becreated equal, but they don't all staythat way. Some become fat, othersthin, some tall, some short, some rich,others poor, in other words, some aremore equal than others. And the sameis true with audio electrical andacoustical signals. Sure, they start offokay, but as soon as a musician pushesa sound out of his horn and sends ithurtling to a microphone little gremlinsstart chewing at it as it makes itstortuous way through the air, into themicrophone, the bewildering maze ofwires and transistors, cutters, pickups,loudspeakers, listening rooms. Indeed,it often seems miraculous that itemerges from all this as somethingeven vaguely resembling the original.And then, of course, there is man, everready to show mother nature the errorsof her ways, tinkering with this signalto make it conform to his own conceptof perfection.

Then, too, sound was never meant tobe recorded; therefore we have nochoice but to make our equipmentconform to the nature of sound,because the laws of physics aredefinitely not going to change to suitour convenience, except possibly to dous mischief, as outlined in Murphy'sLaW.

Okay then, what do we equalize, howdo we equalize, and, for that matter,why do we equalize?

TYPES OF EQUALIZATIONEqualization can be divided into two

basic types as regards to function:Correctional, in which the purpose isto correct for faults in some part of thechain which produces deviations fromflatness in frequency response, andAdaptive in which the response isdeliberately caused to deviate from flatin order to improve the operatingcharacteristic of some component in

the chain, or to allow optimizing someother parameter, such as noise ordistortion. In general, correctionalequalization is introduced at anoperator's discretion, while adaptiveequalization is either imposed byphysical conditions over which onehas no control, or is established byagreed upon standards, or both.

ADAPTIVE EQUALIZATIONSeveral problems occur if we

attempt to record or broadcast anaudio signal and reproduce it with a flatcharacteristic from microphone toloudspeaker. As an example let usconsider the case of a cutter engravinga disc recording. If a constant voltageis applied to the cutter, this willtranslate as a constant velocity ofcutting stylus motion. Now, supposewe attempt to record a signal of 1000Hz and at an amplitude such as toproduce a stylus velocity of 10 cm/sec.One cycle will occur in 1 msec and willresult in a stylus swing in eachdirection of 0.025 mm, that is, it willreach maximum displacement in one -quarter .cycle which takes 0.25 msec.Now, if we record a signal of 100 Hz atthe same velocity, the amount ofdisplacement will be TEN TIMES the1000 Hz value, or 0.25 mm. 20 Hz wouldcause a swing of 1.25 mm. Now try tovisualize this on a microgroove disc.To allow adequate spacing betweengrooves, even with no safety factor,would require spacing the grooves atleast 2.5 mm apart. If we recorded overa total of 7 cm of the record surface at33.3 rpm our maximum possiblerecording time would be just over 8minutes. Remember, this assumes noguard space between grooves, whichwould easily cut this time in half, andassumes we can use the full 7 cm,which would be most unlikely forreasons beyond the scope of an article

on equalization. It also assumes wecan cut such an amplitude withoutrunning into formidable problems withthe cutter, and that we can find apickup which would trace such anamplitude. Remember, too, that 10cm/sec is not that high a Velocity.Clearly, some compromise must bemade.

This adaptation consists of "equal-izing" the recording system so that allfrequencies below an agreed uponstandard, namely 500 Hz will berecorded at a constant amplituderather than velocity. This results in anattenuation at the rate of 6 db/octave,and in actual practice this curve ismodified at frequencies below about80 Hz.

Fig. 1. Characteristics of phonograph recordingand playback processes.

CONSTANT'AMPLITUDE

CONSTAN IVELOCITY

FREOUENCY

CONSTANTAMPLITUDE

DISC RECORDINGCHARACTERISTICS

DISC PLAYBACKCHARACTERISTICS

8 ETI CANADA - MARCH 1g78

EqualizationNow, it just so happens that

magnetic pickups are velocity respon-sive devices, that is they give equaloutput voltage for equal stylusvelocity. Since our recording wasmade with constant amplitude below500 Hz, the velocity falls as frequencygoes down, and if played back with amagnetic pickup the response will fallat the rate of 6 db/octave. Clearly, wemust now equalize this response byintroducing a response characteristicwhich increases at this rate asfrequency is reduced.

The reader who has been followingthis closely and who has someknowledge of noise (not the kindsometimes called "music", but theother kind) will realize by now that if wecontinue to record at a constantvelocity as frequency rises above our500 Hz turnover, eventually the pointwill be reached at which noisegenerated by surface irregularities inthe recording will be equal to orgreater than our signal. In addition,noise generated in the pre -amplifierwill assume a high level in comparisonto the signal. The reader will alsorealize that this does not have to be,since if we continue to record at aconstant amplitude we can overcomethis noise in much the same way as wedid at lower frequencies, i.e.: increaseresponse at the rate of 6 db/octave asfrequency rises (this is just anotherway of describing a 6 db/octave roll -offas frequency drops). It will beappreciated that this could result instylus velocities beyond the capabil-ities of the pickup cartridge, andindeed this is one reason for the use ofa modified constant amplitude charac-

OdB

,90

,45.,

MN

AMPLITUDE

PHASE SHIFT

teristic; statistical distribution ofenergy also alleviates some of thepotential problems, which makes a fairamount of boost possible.

Another example of this type ofequalization is encountered withmagnetic tape. A completely loss -freesystem would show a playbackresponse characteristic which rises atthe rate of 6 db/octave when the tape isrecorded with constant flux in the gap,which, in turn, is the result of constantcurrent through the coils of the recordhead. However because of the tape andhead characteristics the response willbegin to level off and ultimately drop asfrequency rises. The 6 db/octave slopecan be readily corrected (and is) by anappropriate low frequency boostcircuit in the playback system, but thehigh end loss is primarily the result oftape self -demagnetization and play-back head losses, with the tape losscharacteristic playing a prominent partat lower speeds. This means thatinherent tape noise will eventuallyswamp the signal so that we cannotrestore flat response in the playbackequalizer, but must do so duringrecording.

Unlike disc recording the amateurtape recordist is in a position to imposeoperating conditions during therecording process which conflict withthe realities which we have beendiscussing. Increasing the signal levelduring the recording process bringsthe risk of overloading the tape orrequires a reduction of overall levelwhich causes deterioration in signal/noise ratio. Boosting response duringplayback also increases noise in theactive region of the equalizer. As is so

ofe

IN

FREQUENCY

1 1020Tr RC 21T RC 2TtRC

Fig. 2. Transfer characteristics of simple single pole RC high pass filter.

often the case in audio work, the endresult is a compromise or, with luck, afine balance between various conflic-ting requirements.

At the present time it is not myintention to get too involved withequalization circuits; volumes havebeen written on individual aspects ofequalization and doubtless more willbe written in the future. However, abrief examination of methods is usefulin order to understand the proper useof equipment.

Fig. 2 shows a simple high-passpassive filter, having a first order, or 6db/octave slope, along with anormalized frequency and phaseresponse curve. Fig. 3 shows its low-pass counterpart. In figs. 4a and 4b wesee these circuits modified to providelow boost and high boost respectively.Notice that the low -boost circuit andits characteristic are derived from thelow-pass circuit while the high -boostcircuit and its characteristic arederived from the high-pass circuit. It isalso clear that a low-pass and a high -cut characteristics are, strictlyspeaking, the same thing, and a high-pass and low-cut are also equivalent.This has led to the mistaken beliefwhich is often encountered even todaythat bass boost and treble cut are thesame thing and vice -versa. Anexamination of the modified circuitused to provide actual boost im-mediately shows the fallacy o thisnotion. ALL equalizers attenuate ALLfrequencies equally, except in therelatively narrow region in which boostor cut is required.

LOCATION OF CIRCUITThe location of an equalizer in a

circuit usually involves the reconcilia-tion of several conflicting elements,and is generally determined by thetype of circuit (bass boost, trebleboost, etc.) nominal signal level, and.the operating band.'

Since a bass boost circuit involvesconsiderable reduction in mid andhigh frequency level it is generallydesirable to put most of the systemgain before the equalizer in order thatnoise components may be attenuatedalong with signal. This is also true forhigh frequency attenuation. However,hum , components would then beboosted along with signal; therefore,too much gain will require specialattention to design aspects aimed atminimizing hum. Conversely, a highfrequency boost circuit should beinserted early enough in the system asto raise signal above the noise ofsucceeding stages.

ETI CANADA - MARCH 1978 9

0dB

40

90

AMPLITUDE

PHASE SHIFT

'Vs

z

IN

FREQUENCY

201TRC 2TIR;

Where signal level is fairly high, aswith some high output magneticpickups such as Decca and Empire,our 'greater concern is amplifieroverload particularly at high fre-quencies, in stages prior to theequalizer.

If you're starting to get the idea thatperhaps equalization, when combinedwith pre -amplification, can best beaccomplished when the equipment isdesigned for, and associated with, thecomponents with which it is to be used,then you're right on target. Indeed, thisis generally considered to be goodpractice in professional circles. Notonly are tape equalizers incorporatedinto the tape machines with which theyare to be used, but turntables mayincorporate the required pre -amplifier/equalizer circuits. In over twenty yearsof audio work I have yet tocomprehend the logic behind thecommon practice of building magneticinputs into a control unit. But it helps toexplain the differences often en-countered between results publishedin equipment reviews and the users'own experience.

PHASEBefore moving on to the subject of

corrective, or elective equalizationsome attention should be paid to thematter of phase. Hundreds of dollarsare often spent on the construction orpurchase of, for example, a phonopreamp, and great attention paid tonoise, channel balance, overload,accuracy of equalization (to within 2db) and yet it is seldom realized that, inproducing a stereo image, one of thethree major factors in directionalperception is relative phase. Inaddition, all matrixed 4 -channel

10

Pr RC

Fig. 3. Transfer characteristicsof simple single pole RClow pass filter.

OUT

systems currently in use utilize a

specified phase relationship betweenchannels to encode and decode theadditional channels. One commoncharacteristic of all of the impressivedemonstrations of quadraphony hasbeen the use of very high qualitycomponents. Since equalizers areamong the first functions to suffer inmaking economy cuts in domesticequipment, it's small wonder thatquadraphony and even stereo repro-duction in the home are oftendisappointing.

How does this happen? Takeanother look at the phase andamplitude characteristics in fig's 2 and3. At the turnover frequency, that is the3 db down (or up) point, the phaseangle has shifted 45° and reaches itsultimate 900 shift a decade away,which also corresponds to the 20 dbpoint. Now, if common 10% and 20%tolerance components are used in twodifferent equalizers (example, each ofa pair of stereo channels), the finalcurve may indeed be well within 1.5 to 2db tolerance in each channel, but ifeach yields a difference in phase over abroad frequency band of as little as 30°the difference between channels mayvary anywhere from 0° to 60°. This isquite considerable in comparison withthe 90° shift called for in theparameters of any quadraphonicsystem, even the Dynaco-Hafflerpassive ambience network. As forstereo perception, although there ismuch disagreement among authoritiesas to the ear's sensitivity to phase shift,much of this disagreement involvessteady tone conditions and singlechannel reproduction. When it comesto the perception of a synthesizedstereo image as in present day 2 -

channel systems, as little as 15° hasbeen observed to have a profoundeffect on imaging, particularly withregard to depth and elevation. In myown experiments I've been able toproduce as much as 10 db channellevel difference with no serious effecton the stereo image other than a shift inlocalization, and yet switching in a

simple tone control circuit constructedof standard tolerance parts and withboth channels nominally in the flatposition will produce a subtle yet realchange in image stability and solidity.The indication is quite clear: theprecise matching of characteristicsbetween channels is probably of evengreater importance than the absoluteaccuracy of characteristics. Thismeans precision parts and theassociated costs.

CORRECTIVEEQUALIZATION

This might also be described as"discretionary equalization", since itrefers to alteration of response in amanner and to a degree completely atthe discretion of the operator. In a veryreal sense the concept of "correct" isirrelevent here. Adjustment is inaccordance with the ear's own conceptof right and wrong. Accordingly, nohard and fast rules can be laid downnor can definite "how to" instructionsbe given. However, most of theconsiderations already outlined applyhere, so we may now proceed with anexamination of the equipment andtechniques available, secure in theknowledge that there is no such thingas magic, and that, rather than usetechnology to make things better, wecan only use it to make things not asbad as they might be.

USES OF EQUALIZERSDiscretionary equalization is used

because either we don't think thesound we're getting is correct, orbecause, correct or not, we don't like itand want to make improvements orproduce some special effects. It'ssomething like the photographer whouses a skylight filter to obtain a morerealistic colour balance, an1d the onewho uses a red filter to simulate aMartian landscape. Since recordingsare made by human beings much of thetime, who monitor through loud-speakers with their own characteristicsin rooms with their own acoustics, andwho apply their own judgement as towhat the final sound should be like, it isnot surprising that the music lover, oraudiophile may be in disagreementwith the producer from time to time.Perhaps you don't agree that the

10 ETI CANADA - MARCH 1978

Equalizationbrasses needed a little extra bite bymeans of a 7 Khz boost, or you feel thatthe strings could have been broughtmore forward and made less disem-bodied by adding a little mid -rangeboost. Indeed, perhaps you don't minda little structural noise of the concerthall, and feel that the producersacrificed too much bass in order tosuppress it. For this kind of correctionyou will want to use an "equalizer" ortone control of some kind to introducethe appropriate compensation. It maybe sufficient to use a simple controlcircuit such as that outlined in figs. 5and 6 to get some bass boost, or treblecut. But then there's no way for you toknow the exact equalization used inthe original recording, and even lesschance that a simple circuit couldduplicate its mirror image anyway, soat best you can still only adjust it until itsounds better.

If this is not satisfactory, you mighttry using what is called a "GraphicEqualizer", so called because such adevice normally uses slider controlsside by side and their settings provide agraphic representation of the responsecharacteristic achieved. With thisdevice you have several controls eachof which controls the response over anarrow range of frequencies. It maydivide the spectrum into as few as fivebroad bands or as many as thirty oddvery narrow bands. These provide veryprecise response control indeed, butyou still can only adjust response untilit sounds right. Just like a simple tonecontrol.

Another type of device is known as a"Parametric Equalizer" because itvaries the parameters which define a

0--WV--411 O

FREQUENCY

IN OUT

response characteristic, that is centrefrequency, bandwidth, and degree ofboost or attenuation. In general, suchdevices offer fewer choices than agraphic equalizer with regard to thenumber of centre frequencies whichmay be operated upon simultaneously.However, in actual use it is generallymore flexible largely because of theability to vary the bandwidth and tochoose centre frequencies. In sometypes it is possible to operate on thesame frequency band twice or tooperate on two closely spacedfrequencies and to combine charac-teristics to obtain a final responsewhich is completely unobtainable withany other type of component. Theparametric equalizer has been widelyused in professional work, and is thetype of equalizer normally found oneach channel of a recording orbroadcast production console. Therearen't too many in commercialproduction for consumer use yet, butthere is every reason to expect thatincreasing numbers will be offered tothe audiophile. For my money it is thepreferred unit of choice for operatingon the programme characteristics,provided it is not required to serveother functions.

The graphic equalizer is probablymore familiar to the amateur, since it'sbeen around much longer and severalsuch units have been published asconstruction articles (See ETI Oct.1977). The great virtue of this unit.especially the 1/3 octave type lies in itsability to notch out or boost one ormore very narrow bands of fre-quencies, making it especially useful incompensating for the irregularities of

IN

FREQUENCY

OUT

such components as pickups (RIAAEqualized), loudspeakers, and roomacoustics. Indeed one of its earliestand still common applications amongprofessionals is in equalizing controlroom/speaker systems. For thispurpose, the speakers are each fedwith pink noise and the responsemeasured via a calibrated microphoneand a real time analyzer or other meansof measuring response characteristics.The graphic equalizer is inserted in thespeaker channel before the poweramplifier and adjusted until the desiredresponse (usually flat) is achieved.This procedure is then repeated foreach speaker in the system. It shouldbe noted that the response is valid onlyfor the location of the measuringmicrophone and for the exactacoustical conditions which exist atthe measurement.

Two problems arise here, especiallyfor the amateur audiophile. To beginwith the previously referred to phaseproblem can impair the systemimaging characteristics. On the otherhand it may improve a characteristicwhich was already deficient because ofthe phase shifts inherent in thepreviously uncorrected irregularities.

DI-

-J

IN

TREBLE

CONTROLACTION

BASS

CONTROLACTION

FREQUENCYFig. 6. Range of boost and cut action available

from "treble" and "bass" controls.

BASS I FIEBLE

OUT

Fig. 4a. Low boost circuit and response curve.

ETI CANADA - MARCH 1978

Fig. 4b. High boost circuit and response curve. Fig. 5. Typical tone control circuit.

11

OW A

.46 .06;18

Yi H

noWn,W} a24 18 4 , tv 14 41 nd8 018

Fig. 7. Example of a parametric equalizer, in this case a two band, two channel model, the SAE 1800.

In other words, the assets may exceedthe liabilities, both in qualitative andquantitative terms. That's what I meanby an inability to lay down hard andfast rules.

The second problem is more serious.After purchasing or building such anequalizer, especially the graphic type,there is the temptation to use it tocorrect faults in what is actually apoorly designed system. I can recallone enthusiast who built such a unitfrom a kit and installed it in a systemwhich was little more than junk. Asidefrom sounding terrible, it also burnedout speakers and destroyed an outputtransistor.

Why? Well, take a look at fig. 8 whichis the representative response andimpedance curve which might beexpected of a small bookshelf speakerof the sort which promises tooutperform speakers three times itssize and selling for ten times the price.Although it may boast a responsedown to 30 Hz, its response at thatfrequency is down a good 20 db, and itsimpedance is equal to the voice coilresistance, around 6 ohms. Now, toflatten the response of such a speakerrequires a power increase of 100 times.If, in order to operate at high soundlevels, it requires 10 Watts of power inits mid -range, 1000 Watts would berequired. That's quite a lot to demandof a 60 Watt amplifier whose ratings arealready optimistic, to say nothing ofwhat such power would do to the poorlittle speaker. Actually, most suchequalizers only offer about 12 db ofboost, but if this is combined with a so-called loudness control, it's easy to seethe kind of abuse possible.

Another point worth considering isthat if you use the control to correct forthe equipment faults, you can't use itfor programme correction at the sametime. If you have 10 db boost available

and you use it all for boosting speakerresponse, you have nothing left overfor programme correction. And theamount of boost available is a functionof the equalizer and power amplifierreserve plus speaker handling capa-city.

Another consideration is energydistribution. A common assumption isthat power levels tend to be about thesame at all frequencies. This is nottrue, as is demonstrated in fig. 9 anenergy distribution curve averagedfrom the results of a variety of studies.It shows that the largest amount ofenergy in orchestral music is concen-trated in the range between about 100Hz and 500 Hz dropping off rapidlyabove_ and more gradually below thisband. It should also be rememberedthat 500 Hz is a common cross -overfrequency in 3 -way speakers and thatwith a current trend away fromconstant resistance networks, manysuch speakers exhibit high impe-dances and considerable reactivecomponents in this region, whichimposes severe limitations on thepower capabilities of many amplifiers.This, in turn, limits the usefulness ofmany equalizers in this region,especially in the boost mode, wherethe result is often high distortion anddamaged equipment. With a great dealof hard rock, electronic and synthe-sizer music high frequency energytends to be considerably greater thanwith orchestral music. Now, thetweeter of a 60 Watt speaker systemmay be capable of handling typicallyfrom 5 to 10 Watts of actual power. Thisis reasonable enough in relationship tothe orchestral distribution curve, butan excessive boost in the regionhandled by the tweeter carries with itthe distinct possibility of requiring it tohandle anything up to the full output ofthe amplifier, which may be 60 Watts or

, p.a et 1,2 94 .40

VVIraff),41148Y

more. Further, if the tweeter level ispadded down (with an L -Pad, I hope)to match the other drivers, this mightsave them, but much of this power is

then dissipated in the pads, and mayexceed their ratings.

Then there's the tape recorder.Remember the high frequency boost inthe record mode? Well that uses uppart of the headroom , available. Younow have the same problem as withspeakers.

And one final word about highfrequency boost. Every phonographpickup has limitations to its trackabil-ity. And when it does mistrack it cangenerate very large high frequencycomponents, which is one reason whyit sounds so bad. Moreover groovedamage also results, and even if thedamaged groove is later tracked with abetter pickup there are still quite a lotof extraneous high frequency "gar-bage" signals generated. If you boostthese component signals along withthe desired signal, the rest of yourequipment doesn't know the differenceand will react in the various waysalready outlined.

CIRCUIT LOCATIONThe same considerations apply as

were outlined with regard to adaptiveequalizers; locate at a high enoughpoint in the system to avoid overloadproblems without boosting noise, andstill be functional. Most commercialcontrol units provide a recordingoutput and a monitor return just beforethe volume control which effectivelybypasses all controls including volumeand tone, which are used only formonitoring. Installing a graphic orparametric equalizer at this pointusually is the most satisfactory as itcan then be used for recording and forlistening. In general, the most useful

12ETI CANADA - MARCH 1978

Equalization

OdB-

10d8-

AMAITUDE

IMPEDANCE

10 100 11C1 -1z 10kHz

FREQUENCY

Fig. 8. Typical performance characteristics oflow price small size (and high hype)loudspeaker example.

point is at the same level as is used forswitching. However, if the device isused to correct for speaker/roomacoustics the more logical location isimmediately before the power ampli-fier. Unfortunately, this cannot be.done with most receivers or integratedamps without opening and modifying

the unit. And this is a good argumentfor the use of separates. But that'sanother story.

CONCLUSIONAt this point the reader might well

wonder what useful purpose is servedby these equalizers. In many instancesthey do more harm than good, but thisis largely the result of using them as asubstitute for good design. If theperformance level sought requires theuse of a large Klipschhorn, get aKlipschorn, not a $99.95 supercompact speaker special and a magicbox. It won't do the job. DO payattention to the acoustics of the roomand the proper placement of a welldesigned speaker driven by a suitableamplifier. And DO use pickups andother equipment of appropriatequality. Then select the appropriateequalizer if you have use for one, anduse it to deal with lesser acousticproblems or to make small alterationsto programme quality.

In light of this the reader may beinterested to know that my own systemuses no equalizers of the discretionarytype. A set of large transmission linespeakers in a properly treated room

provides high level performancethrough the full audible range with animaging matched by only a very fewprofessional systems. Phono preampsare matched to their own Stanton andShure pickups each on its - ownturntable, and are not interchanged.The only need for additional equaliza-tion occurs occasionally when tapingradio and TV broadcasts and 78 rpmdiscs. Under those conditions a fixedequalizer is designed and inserted inthe line. So far the, only real problemencountered is with a particularrecording in which a phasor is used. Inderived quadraphonics it sweeps thesignal around the room and drives thecats crazy.

Fig. 9. Energy distribution graphfor various types of music.

ROCK,ELECTRONIC

ORCHESTRAL

10 100 Ikiiz 10kHz 20kHzFREQUENCY

cisVV1 AG Y1 IA M. electronics.inc.72 - 74 STAFFORD STREET 364-6754TORONTO, ONTARIO, CANADA M6J 2R8

RESET - BREAKERS

TRIP ATTRIP AT

TRIP AT

ALL ONE

PRICE

51.00 EACH

1,6 Amp2 Amp3 Amp.

ETI PCBs for recent and current projectsAll boards are superior quality predrilledand guaranteed no C.O.D.s

MINIATURE SINGLE ENDED

ELECTROLYTIC CAPACITORS

-

ForFor printed Circuits

1000 MFD 35 Volt1000 MFD 25 Volt1000 MFD 16 Volt100 MFD 35 Volt

4.7 MFD 50 Volt

.65 Each50 Each40 Each20 Each10 Each

CUSTOM DESIGNERSMANUFACTURERS AND DISTRIBUTORS

OF ELECTRONIC EQUIPMENT

NOW...the Mobile Power SleuthA TRANSISTORIZED AUTO ANTENNA TOIMPROVE AM, FM, RECEPTION.

MAGNUM ELECTRONICS, DEVELOPERS OF THEWIDELY ACCLAIMED 85 FM POWER SLEUTH FOR THEHOME, IS NOW OFFERING THE MOBILE SLEUTH FORYOUR CAR.

Two separate transistorized amplilers In, AM and FM Increases signal strength approz 70 db to

improve reception quality and range Built in cross modulation suppression and

shielding from interference Fits all 6 to 12 volt application,

1 year warianty. Antenna rod locks at 30 60 9Q

Easy mounting on fender. door post.rear deck, etc

Solves wind shield antenna problems. Electronic circuitry in base 100'. sealed

twin moisture Antenna rod folds completely into base to prevent car wash darn

age Ideal for vans, boats, iecreational vehicles. Developed in Germany, sold throughout Europe. after extensive

field testing, the North American model is now available

$49.95

SPECIFICATIONSWith built-in cross modulation suppression and shielding for interterence.Turing ranges marine band 150-340 Khz. AM 550-1600 Khzshort wave - all bands FM 88-108 Mhz.

FACTORY TO YOUPRICES ON

8 -TRACK PLAYERS

Formerly sold at a much higherprice by a national audio chain.

-=1,..z2,La. '1.11

High quality 2-4 channel playersin stylish Walnut grained wood

cabinet. Automatic switchfrom 2 to 4 channels.

Output 350mV

Kit T1 complete $35.00Kit T2 mechanism& chassis $19.50

Add $1.00 for shipping & handling if total less than $20.00 - Minimum order $5.00

Ontario Res. add 7% P.S.T. -


ETI Project

HammerThrow

AMMER THROW'`. 2 4 8 8

O THROW o m

00

0 0 0ON PL AY

RESET

110111111111111111

An exciting game of skill and luck that will help pass those long and lonelywinter evenings.

IF, LIKE MOST of the ETI staff, youhave more brains than brawn, andwould not boast about the quality ofeither, it is likely that the mere thoughtof swinging a massive weight aroundyour cranium is enough to strain yourbodily systems. This probably means -and we are sorry if this comes as adisappointment - that your chancesof selection for the Olympic hammerthrowing team are, shall we say, nil.

Some may say that this is a pity asthe sheer thrill of an event such as thehammer throw is probably very stimu-lating to those chunky brutes that arelucky enough to be able to take part.This is where we come to the rescuewith our armchair version of the game.We think it has a number of distinctadvantages over the real thing. One ofthese is that anyone, from an anemicsparrow upwards, can play the game. Asecond being that it is nowhere near asmessy if, when playing in your livingroom, you get things wrong.

The game, as can be seen from ourphotographs, has a front panel with acircle of sixteen LEDs together with aline of eight LEDs at a tangent to thecircle.

To play, after pressing reset, firmlypress the play button. The LEDs in thecircle will light one at a time simulatinga spot of light moving in a circle. Atthe same time a distinctive, not to say

loud, sound will be generated.The spot will at first travel slowly

round the circle, but will soon beginincreasing in speed until it is travellingquite fast.

The object of the game is to releasethe play button at the instant that thelop' LED of the circle is lit. If success-ful the line of LEDs will light to indicateyour score, the faster the spot wasmoving when you scored the more willbe your score. If you miss, the circle ofLEDs will continue to rotate at thesame speed as they were when youplayed.

BIG ONES AND LITTLEONESA game will consist Of, say, eightrounds - the score from each beingadded to the last. At the end of a gamethe person who scored the most is thewinner. The skill comes in decidingwhether to go for a number of lowscores that are relatively easy to get, orfor a few big ones.

As befits the design of a project ofthis nature we were in convivial moodand pleasant surroundings when wefirst discussed the game. We producedthe first design sketch (well a few lineson a napkin - yes in the tavern again)which used digital devices. Upon seeingthis some likely person said that hethought most games featuring LEDsdesigned over the past few years should

generically be called "spot the 4017".Our initial reaction was to defend

our design but a moment's thoughtshowed that he had a point - the4017 CMOS counter is over -used whenit comes to games. At this stage wedecided to rise to the occasion andproduce the game using an all analogueapproach.

The result can be seen in the circuitdiagram. We are pleased with thiscircuit: It uses some unusual ICs andfeatures a number of interesting circuitblocks - and of course there is not a4017 in sight.

CONSTRUCTIONConstruction of the game is greatlysimplified if the PCBs are used. Threeboards are required, one for the powersupply, one for the display, and finallythe main control board. Begin bybuilding and testing the power supply.Take care to ensure that all componentsare mounted as shown in our overlay.

Next assemble the control anddisplay boards. These carry a largenumber of components and mistakesmade during assembly can be difficultto trace later - so take care at thisstage. Do not insert the link betweenIC3/4 and IC9 at this stage.

It is best to test the boards beforemounting them in the case, as it isdifficult to get to some of the deviceswhen the boards are in their final


Hammer Throw

000U

In -XIn toCX LI/

_GEt

r- to to

O

U,

C

+ - I

T034

-,A7s.Ar-t

CV ,to

-V

toeuix

uCV -

+ I JN

U

01

CV

CX

CC

In

O

O

In

CL

CL

01Cs(

cn-X

In

U'

In 0

0

of

CL

0 Xt" tog

No

o

0

O3

04:1

0C0

0_

Eco

O

rn

0

.6)

positions. We used a sloping front Verobox to house our game and the generallayout adopted can be seen from ourphotographs.

SETTING UPThere are five preset potentiometers onthe board and all must be correctly setup before the game can be played.

The first adjustment to be made is toRV4. To calibrate this control first pressthe reset button and then the playbutton for a few seconds. At this stage asound should be heard from the speakerand the game display LEDs should beseen flashing. Adjust RV4 until theLEDs produce a continuously rotatingspot of light. The speed at which thecircle of light rotates can be adjusted byRV 1.

The next operation is to set up thescore display. To accomplish this, pressreset and then operate the play buttonuntil the spot of light is rotating atmaximum speed. Release the playbutton and enable the score display byapplying a positive pulse (from supply)to the junction of R29 and IC6. RV5should now be adjusted so that theseventh score LED is just extinguishedand the eighth lit.

The final adjustments concern the'window' discriminator. To make thisadjustment R38 (the end remote fromIC9) should be connected to the sliderof RV,. Adjustment of RV2 shouldilluminate successive LEDs of the gamedisplay. RV2 should be set to the pointat which the top LED just extinguishesand the LED to the left just lights.

Now connect the input of IC9 to theslider of RV3. Adjust this pot so thatthe top LED just extinguishes and theLED to the right is just on.

This completes the adjustments andthe link omitted during construction,should now be fitted.,

Now is the time to get in trainingand, if you're good enough, you mayyet make it to Moscow.

BUY LINESThe "line-o'-leds" ICs are made by

Seimens. Their distributors are:Amphion in Halifax, and Moncton;Prelco in Montreal, Ottawa andMississauga, Carsten in Toronto,Electrical Supplies Ltd and WES Ltd.in Winnipeg, Radio Supply andService in Regina, Cardinal Industrialin Edmonton, Paar Industrial inCalgary, with RAE Industrial andWestern Telecom in Vancouver.


ETI Project

C1I100n

R404k7

R41 10k

GI

R1.2100k

R43 10k

R41.33k

R1.510k R47

R46 10k1k

01

9

ICIO

6

R49 10k

SI

0ENABLE

R48100 k

R50 33k

R51 10kR52 R5310k 1k

IC11

16

'VVV`R54 10k

O0y

R55

o 0ETI HAMMERTHROW PSU

B1

A

Ct

01

C1

B1

LEDS1-16

Al

AlB1

C1TNNDI

NN C1

B1NuAl

A2,82

C214 D2

14 02LEDS17-24A2

B2 012C2

1.4 B2C2

A2

A02

A2'B2C2

D204-11

02

C2

®Q2Fig. 2.

B2

O A44 A2_,

Circuitdiagram of gamedisplay section.

PARTS LIST

Fig. 3. Foil pattern of power supply board shownfull size (120 x 45 mm).

RESISTORS all Y4W 5% unless stated

R1,5,17,26,42,48 100kR2,13,15 10MR3,6,8,14,16,25,

C5 10u 25 V electrolyticC6 10n polyesterC7,8 2u2 polyesterC9,10,12 100n polyesterC11 100u 25 V electrolytic

30,37,38 1MR4,40,55 4k7 SEMICONDUCTORSR7,18 2M2 C1 78L15AR9,31 560k C2,3 LM3900R10 4M7 C4,5,7,9 741R11,12,28,29 820k C6 ' CD 4016820,21,34,36,39,41,43 C8 555

45,47,49,51,52,54 10k C10,11 UAA 170 (SEIMENS)R22 220k Q1 2N3905R23 150k Q2 MPS6515R24 2k2 D1,2,3,4-12 1N914R27 120k LEDs 1-24 .2" typeR32,46,53 1k BR1 4 pin DI L TYPE:R33 56k 0.9 A 400 VR35 100RR44,50 33kR19 47k TRANSFORMER

T1 120V - 15 V 6VA

POTENTIOMETERS LOUDSPEAKER Telephone type insert orLS1 similar

RV1,3,4,5 100k min hor trimRV2 10k min hor trim SWITCHES

PB1,2 Push to makeSW1 SPST toggle

CAPACITORSCASE Vero type 65-2523

C1 1000u 25 V electrolytic MISCELLANEOUSC2 220n polyester PCBs as patterns, LED mounting clips, fuse and holder to suit.C3 470n polyesterC4 1n polystyrene For clarification of component notation see "Reader Service Information".


Hammer Throw

IA

Fig. 4. Component overlay of PSU,Iine groundis connected to TI by a solder tag under themounting bolt.

Fig. 5. Full size (160 x 110 mm) foil pattern of display board.

100mA 0,..-0-cp<o

6VA

SW1

0Ground

FS1TI

120V / 15V

O

I + BR1

ICI

CI1000u

25V

C2

I/P G0/P

220n 470n

+105V

OV

0

Fig 6. Circuit diagram of the game's power supply.

I CANADA - MARCH 1978 17

ETI Project Hammer Throw

C4011

Fig. 7. Full size foil pattern of main control board (160 x 110 mm).

4:rti,`,:.1I t t 1

r

Tfinp -Ean-

1.7] 1.11

_IU..1 COC[ Cl_

18

Fig. 8. The overlay for score board Fig. 9. Overlay for the control board. Continued on page 20.


ETI ProjectHOW IT WORKS

MATRIXOUTPUT

TOLEDS

GND

H

G

E

D

B

- 3

-5

-6

-7

-8

12-

11-

10-

9-

U44 170

1 -

2

3

GNU -4

-87- .Ve

6

-5741

1BRIGHTNESSCONTROL

STABILISEDOUTPUTVOLTAGE

VREFUPPER

VREFLOWER

INPUT

Ve

A MATRIXOUTPUT TOLEDS

CO

Fig. 10. Pinouts for the integrated circuits usedin the hammer throw.

The circuit may be broken down into anumber of major blocks - viz the displaysections for both game and score, a voltagecontrolled oscillator, a ramp and holdcircuit whose output controls the oscillator,a 'window' discriminator, a sound generat-ing circuit and finally a power supply. Aswell as these major blocks there are also anumber of latches, buffers and switchesthat are necessary for circuit operation.

The block diagram shown in Fig i.1 showsmost of the circuit blocks and, togetherwith the circuit diagram, should be read inconjunction with this how it works.

SYSTEM OPERATIONThe game display is based on a UAA 170 IC.This device is for driving LED displays andwhen connected to a line of sixteen LEDswill illuminate any one of these dependingon the magnitude of the analogue voltageapplied to its input. For the game displaywe need to produce the effect of a spot oflight moving in a circle. To achieve this wearranged the sixteen LEDs in a circle andfeed a sawtooth waveform into the UAA170. A moment's thought will show that thiswill produce the desired effect.

In order to make the display rotateslowly at first, but speed up as play pro-ceeds, we made the sawtooth generatorvoltage controlled. The control voltage isproduced by a ramp and hold circuit whichis reset to zero at the start of play, butbegins to ramp up, thus increasing thesawtooth's frequency as play continues.When the play button is released, thevoltage reached is held by the ramp andhold configuratibn until it is reset. Thisvoltage is used for score purposes as de-scribed below.

The game requires that if, at the instantof releasing the play button, the 'Top' LEDof the game display is lit, a score is indi-cated, the magnitude of the score beingproportional to the speed at which thecircle of LEDs was moving at the instant ofrelease. From the description of the gamedisplay it will be seen that in order to light aspecific LED the voltage input to the dis-play driver must lie within a specific vol-tage range, thus in order to detect whetheror not the 'top' LED is on we must look atthe output of the sawtooth generator (thisis input to UAA170) and decide whether itlies within the range that will light thespecific LED at the instant the play buttonis released. The circuit that accomplishesthis is the 'window' discriminator.

This is formed from two voltage compa-rators together with two analogueswitches. Detailed action is described be-low, but briefly the circuit, when fed withthe sawtooth output, will provide an indi-cation whenever this waveform passesthrough an (adjustable) 'window' voltagerange.

At the instant that the play button isreleased a short pulse is produced from amonostable. If this pulse is coincident withan indication from the window circuit thatthe top LED is on we must arrange toindicate a score.

The score must be proportional to thespeed of the LED circle which is in turnproportional to the voltage level reached bythe ramp and hold circuit. Thus, to producea score, we feed the output from ,the rampand hold, via an analogue switch, to asecond UAA 170. This second display con-sists'of eight LEDs in a line.

This completes a brief description ofcircuit action; we shall now deal with eachblock in more detail.

RESET CIRCUITRYThe game is initiated by operation of thereset button (PB1). This zeros the ramp andhold circuit described below, as well assetting latch 1 IC2/3 and resetting latch 2IC3/1. Latch 1 enables the play buttonwhen its output is high (set) - latch 2enables the score display when low (reset),the game display when high (set)

Each latch is based on two of the am-plifiers of an LM 3900 Quad Norton am-plifier package. This device is unusual inthat instead of amplifying the difference involtage applied to its input terminals, itamplifies the difference in input current.

The + and - inputs of these Nortonamplifiers are both clamped to one diode -drop above ground and thus all inputvoltages must be converted to currents (byresistors) before being applied to the inputs.This is the basis for the current -mode(Norton) type of operation.

In operation the current flowing into the+ input must equal that flowing into the-input, the difference between the currentdemanded and the current provided by anexternal source must flow in the feedbackcircuitry.

Operation of both latches is the same andwe shall only describe the action of latch 1.

Assuming that the latch output is low(the latch is reset) the current injected intothe - input of IC2/3 will ensure that theoutput remains low. If now sufficient cur-rent is injected into the + input the outputvoltage will rise as the device attempts toreduce the input current differential tozero. Positive feedback via R9 will enhancethis action and cause the amplifier to latchhigh. This is because the current injectedinto the + input via R9 in this case isgreater than that into the - input due 'toR8. A positive pulse via R11 to the - inputwill however once again bring the outputlow.

C5 and R4 ensure that when power is firstapplied the game is reset.RAMP AND HOLDThe ramp and hold action is provided byIC2/2 and IC2/4. A positive voltage via R5and Dl causes the output to ramp downwhile a similar voltage via R10 causes theoutput to ramp up. The reset button causesthe downward ramp while play causes anupward ramp.

In any sample and hold application a verylow input bias current is required if the holdperiod is to be stable. The existence ofmatched amplifiers within the LM3900allows one amplifier to bias another.

In operation the LM 3900 requires a biascurrent to be applied to its - terminal.IC2/4 has its + terminal grounded andfeedback applied via R15 and R16. Theoutput voltage of this device will attain alevel such that the current fed back viathese resistors is equal to the bias currentdemanded by the input. This same currentwill flow via R13 and R14 into the - inputof IC2/2 reducing the effective bias currentof this amplifier to almost zero. DI isolates'this bias current from the rest of the inputcircuitry.

If now a positive current is injected intothe - terminal, the output voltage will fallas it attempts to feedback a current of thisvalue in order to reduce the input currentdifferential. This constant current acrossC7 results in a linear voltage ramp ap-pearing across C7. Input to the + terminalcauses a positive going ramp, to the -terminal a negative going ramp.

The rate at which the voltage across C7.changes is proportional to the value of the


Hammer Throwconstant current supplied which is in turnproportional to R5 and R10. As R5 is some40 times larger than R10, the ramp down(reset) is far quicker than the ramp up.

The output from the ramp and holdcircuit is fed, via IC6/1 to the score displayand via IC3/2, a non -inverting scaler, to thesawtooth VCO.NON -INVERTING SCALER

The scaler is required because the outputfrom the ramp and hold configuration canvary over nearly the whole supply voltagewhereas the VCO requires only a small vol-tage swing to provide the requiredfrequency change.

The scaler is based on another Nortonamplifier arranged as a non -inverting am-plifier. Feedback is applied via RVI and R19and output is fed to a pcitential dividerformed by R22 and R23 and thence to theVCO.VOLTAGE CONTROLLED SAWTOOTHOSCILLATOR

The VCO is formedby IC3/3 and IC3/4.Action of IC3/4 is much the same as that ofIC2/2 described above. The special inputbias circuitry is not required as there is nohold requirement.

IC3/3 acts as a comparator and circuitaction is as follows: while the output ofIC3/4 is high and ramping down (input to- terminal) the current into the - input ofIC3/3 due to R26 is greater than that to its+ terminal due to R25 - its output is thuslow.

As the output of. IC3/4 ramps low how-ever, there comes a point where this situ-ation is reversed. The output of 103/3 goeshigh (this state being maintained .bypositive feedback via R7 which injects alarge current into the +input of 1C3/3 asR7 is much smaller than R25.) '

The output if IC3/4 thus goes high, res-toring current flow via R26 and starting thecycle again.

By varying the current injected via R22the time taken for the output of IC3/4 toramp down to the point at which the com-parator triggers is lessened. This results inan increase in the frequency of the saw -tooth.

The output from the VCO is fed to thegame display section RV4, to the 'window'discriminator formed by 1Cs 4 and 5 and viaIC7 to the sound generator IC8.WINDOW DISCRIMINATORThe window discriminator is formed by twocomparators IC4 and IC5 and two of theanalogue switches in IC6.

Operation is as follows: If we assume thatthe output of the sawtooth VCO is high andramping down the voltage on the - input ofIC4 will be higher than that on the + input(a reference level established by RV2) andits output will be low. The output of IC3 willbe high as the input to its + terminal ishigher than that to its - input.

As the voltage ramps down, a point'willbe reached where the output of IC4 goeshigh as the voltage at its - input falls belowthat set by RV2 at its + terminal. At thisstage the outputs of both IC4 and IC5 arehigh, as IC5 has not switched. As thevoltage continues to ramp down, however,the voltage on IC5's + input falls to a pointbelow that on its - input and the output ofthis IC goes low.

Thus the outputs of both ICs will be highfor a small range of input voltages (thewindow) defined by the difference in vol-tage between the sliders of RV2 and RV3.

The outputs of these ICs are fed to theinputs of two analogue switches. A positivevoltage applied to these 'switches turnsthem "on".

PLAY

-0 0

ORESET

*15V

0 1101.00A° 0 E

RAMP AND HOLD3900

OUTVu

LATCH3900

ll900

NON INVERTINGSCALER

MIT"LIMI

Fig. 11. Block diagramof the hammerthrow game.

IVIVAT011(II

ql3 IN UT

AND AV1010

IIL

NON INVERTINGD UFFER

1,41,11

Thus during the window a signal pathexists from the input of IC6/4 to the outputof IC6/2.MONOSTABLE

The monostable is formed by IC2/1 this '

produces a short positive going pulse uponreceipt of a negative spike produced by therelease of the play button.

Current injected into the - terminal viaR3 will normally hold the output low,however a negative pulse applied via C4and RI will "rob" this current from theinput and causes the output to go high.

R7 latches the gate in this state after thenegative pulse is removed. At this stage C6begins charging, feeding back an increasingamount of current to the - input as thevoltage at the junction of R6 and R3 rises.

There comes a point when this current isgreater than that fed back via R7 and theoutput returns low. Diode D2 rapidly dis-charges C6 to provide reliable re -triggering.

The leading edge of the output pulse iscoincident with the release of the playbutton. This pulse is used to turn on analo-gue switch IC6/.3 It will be rememberedthat if the voltage of the VCO is within the'window' at this point - switches IC6/4and IC6/2 will also be on. This allows thesupply voltage input to IC6/3 to set latch 2and thus initiate the required actions, ie, .

blank game display, enable score display,etc.

The monostable also resets latch 1 IC2/3to remove supply from the play button, thisprevents cheating.GAME DISPLAY

The output of the sawtooth VCO is fedvia an inverting buffer, IC9, and a potentialdivider, RV4, to the input of ICIO a UAA170.The input circuitry of this device consists ofa series of differential amplifiers with oneinput of each connected to the input ter-minal (pin 11) via an emitter follower. Theother input of each is connected to a pointin a potential divider chain consisting ofequal value resistors. The differential am-plifiers thus operate as analogue voltagecomparators and as the input exceeds thereference voltage of a particular compa-rator, the 'output of that comparator willchange state.

To reduce the package pin -out the LEDsof the display are not driven individuallybut are arranged in a four by four matrixpattern controlled by the row and columnoutputs of the UAA170 (A -D and E -F res-pectively). By enabling the appropriate rowand column output any one of sixteenLEDs may be selected. The matrix outputsare controlled by the internal' logic of theUAA170.

The resistor chain R42, R44 and R45.setsup the reference voltage inputs of thedevice. The voltage on pin 12 establishes

SWITCH-CF\9015

/117

OWL01501 AY

the lowest voltage to which the UAA170will respond. If the input voltage is belowthis point the first LED of the displayremains lit. As the voltage rises above thislevel the first LED is turned off, the secondon - as the input rises the spot moves upthe chain, until the voltage reaches that seton pin 13. This is the maximum voltage towhich the display responds and if the inputis taken above this level the last LEDremains lit.

In addition to defining the indicationrange the voltage between pins 12 and 13determines the abruptness of transitionbetween any two LEDs. With this dif-ference set to 1V4 the light point glidessmoothly along the scale, with increasingvoltage difference the passage becomesmore abrupt until at 4V the light spot jumpsfrom one LED to the next. We have set thisvoltage to a point between the two ex-tremes.

The resistors R46, and R47 controlthe brightness of the display. Q1 suppliespower to the display and is driven fromlatch 1 IC2/3. This, you will recall, is reset,ie, its output is low, at the start of a game. Alow voltage applied to Q1 via R41 turns thistransistor on and enables the display. Thelatch is returned high at the end of a game,this turns Q1 off and blanks the display.

SCORE DISPLAYThe score display is formed by a second

UAA170 (IC10). Much of the circuitry is thesame as that of the game display exceptthat we only wish to display eight LEDs.The diodes from unused outputs to the+ VE supply act as 'dummy' LEDs, res-tricting the display to eight LEDs, youcould use LEDs for extended scoring - buta larger box is needed. This display ispowered by Q2 which is again fed from theoutput of latch 1 (1C2/3. This time, how-ever, the display is blanked, Q2 off, whenthe latch is low and enabled, Q2 on, whenthe latch output is high.

SOUND GENERATORThe sound is generated by IC8 an NE555

operated in its astable mode.The reset pin(4) is normally held low by

R32 and hence circuit action is inhibited. Apositive voltage applied from latch 1 via theplay button enables the sound during thegame.

The output is frequency modulated byapplying the output of the sawtooth VCO,via buffer IC7 to provide the necessary lowimpedance drive, to the voltage controlinput (pin 5) of IC8.


ETI Project

True RMSVoltmeterThe use of a special IC results in performance greatly improved overconventional designs.

MOST METERS which can measure acsignals do so by rectifying the signal andthen measuring the average voltage.With a sinewave the average voltage is0.637 of the peak voltage while the rmsvalue is 0.707 of the peak. Therefore acorrection factor of 1.11 is built intothe meter to give the rms value of thesignal.

Provided you stick with sinewavesignals these meters are adequate. Withany other waveform, however, they arenot accurate. With a square wave theerror is 11% and with pulse wave formsthe error increases.

Before continuing we should explainwhat rms means and its significance.Without getting mathematical, the rmsvalue of any wave form is the same as adc value which would produce the sameheating effect in a resistor. For example:

Power in a load can be varied byusing phase control (i.e., light dimmer)where the time the load is connected tothe line is variable. The rms value isdifficult to calculate except at the pointwhere it is half on-half off. The powerthen is obviously half power.

//

/Half power point.

0 DECIBELS 4 -4

VOLTS

ON

.4f

1mW 600

I I Al

True RMS ac Voltmeter

ETI 134

30 mV-30 dB

10 mV-40 d13

3 my-50 dB1 mV-60 d8

-70 dB

100 mV-20dB

300 mV-10 dB

+10 d810V

+20 dB

30 V+30 dB

If the input voltage is 120 V and theload is 120 ohms the power (maximum)is given by

P= E2 or 120 x 120 = 120WR 120

Half power therefore is 60W. The volt-age corresponding to this is given by

E=VPxR or 85 V (rms)

On a "normal" meter this will read60 V or an error of 30%.

This design uses an rms detector IC,which is basically a small, special-purposeanalogue computer to mathematicallycalculate the true rms value for anywaveform.

DESIGN FEATURESThe design of the voltmeter is basicallysimple, starting with an attenuator inthe front end, then an amplifier with ahigh input impedance and switchablegain which, with the attenuator, givesthe range selection. A filter is thenadded to give the "A" weighting and therms detector IC (LH0091) does the rest.

The output of the input amplifier is60 mV, independent of range selected,for an input corresponding to the fullscale reading. This gives a maximumgain of 46 dB on the 0.3 mV range.There is a loss of about 2.3 dB in thefilter (at 1 kHz) and the spare amplifierin IC2 is used to provide a gain of 20 dBgiving 500 mV (for full scale reading)before the rms detection is done. The


True RMS Voltmeter

SPECIFICATIONS Meter Type

Ranges

Accuracy

Input Impedance

Weighting Networks

Frequency Response

rms reading ac only

0.3, 1, 3, 10, 30, 100, 300 mV1, 3, 10, 30 V

+3% nominal (crest factors up to 3)-8% at crest factory of 10

1 megohm in parallel with 25 pF

Flat or 'A' weight

10 Hz - 20 kHz

Fig. 1. Meter scale shown full size.

rms detector section has unity gain with500 mV rms in giving 500 mV dc out.

However things are never that simple.With a total of 60 -odd dB gain, alongwith the requirement for a 1 M inputimpedence, we have an excellentformula for an oscillator. With the thirdtry (yes, we have failures too) with ade-quate shielding and layout, stability wasobtained and this final design is presentedhere.

The spare IC in the LH0091 is nor-mally used to buffer, filter or amplifythe output of the rms converter (seedata sheets in this issue but we used itbefore so as to buffer the filter networkand save an additional op amp (theinput of the rms converter is only 5 kohms). The output voltage from theconverter is only 500 mV but this isadequate to drive a meter. We couldhave provided more gain in the bufferstage so giving a higher output but thiswould lead to greater errors with highcrest factor waveforms.

We have limited this instrument to acsignals as this eliminates the need forbalance controls to correct for driftwhen measuring low level signals. Thisnormally is of no consequence as mostsignals, i.e., output of a tape recorder,sound level meter, etc., have no dc com-ponent. If dc capability is needed,capacitors C1, 8, 9, 14, 15 and 16 haveto be shorted out, a zero adjustmentpotentiometer added to IC1 along withthe potentiometers needed to offsetadjust IC2 (see data sheet).

CONSTRUCTIONIf the printed circuit board is used alongwith the layout and shields as describedthere should be no problems with con-struction. The wires associated with therotary switch should be no longer thannecessary to minimise any pickup. Thebox should be grounded to the lineground and the front panel groundterminal (left hand one) should also beconnected to ground.

USEWhen measuring low level signals theremay be 60 Hz pickup unless thecommon side of the input signal is con-nected to ground. This may be doneeither in the unit under test or on themeter (hence the ground terminal).Also with the meter terminals opencircuited the meter will give somereading. However, as the outputimpedance of low level signals (0.3 mVand less) is normally relatively lowthis is usually no problem.


m

INP

UT

0---

I CI

100n

C2

R1

15p

1M

C3

R2

150p

100k

C4

R3

1n5

10k

C5 27

n 7

C6

J-27

n

SW

1 a

R4

R5

12k

150k

R6

100k

C7

l2

15p

C

OV

OV

Fig

. 2. C

ircui

t dia

gram

of th

e vo

ltmet

er.

OV

0.3m

V10

*15V

IC1

CA

3140

4

-15V

C9

100n

T1C11F0

00n

R19

RV

147

k10

0k

C11

10n

R21

27k

C13

4n7

R7

R18

R20

C12

10k

27k

27k

T82

0p

C8

100:

.

R8

56

SW

1 b

1mV

0

3mV

0

10m

V 0

30m

V

100m

V 0

300m

V 1V0

3V 10V

30V

R9 lk

SW

2

Fla

t

A t

+15

V

OV

CA

LIB

RA

TIO

N P

RO

CE

DU

RE

a) S

ELE

CT

TH

E 1

VO

LT R

AN

GE

.b)

CO

NN

EC

T A

SIN

E W

AV

E O

SC

ILLA

TO

RD

TO

TH

E IN

PU

T A

ND

AD

JUS

T T

O1

VO

LT A

ND

1kH

z.c)

SE

LEC

T 'A

'wt P

OS

ITIO

N A

ND

AD

JUS

TR

V2

TO

GIV

E C

OR

RE

CT

RE

AD

ING

.dl

WIT

H N

O IN

PU

T, A

DJU

ST

TH

E M

ET

ER

R10

R11

AD

JUS

TM

EN

T T

O O

V. R

EC

HE

CK

120

3k9

ST

OP

C.

el S

ELE

CT

Fla

t PO

SIT

ION

AN

D A

DJU

ST

RV

1 T

O G

IVE

CO

RR

EC

T R

EA

DIN

G.

R12

R13

390

47k

R14

<R

151k

5 <

39k

R16

<R

1712

k <

39k

+15

V

OV

-15V

R22

270k

lOw

R23

270k

R24

27k

C15

13

1611

15

2IC

210

LH00

91

I1

OV

-15

V

120V

- 2

8V C

T6C

ma

C16

-±,

10,T

RV

220

0

M1 1m

A F

SD

R25

330

OV

J

1-0

GR

OU

ND

o.

NCry

HO

W IT

WO

RK

ST

he in

put s

igna

l is

atte

nuat

ed b

y th

e ne

t-w

ork

R1-

R5

and

C2-

C6;

the

appr

opri

ate

atte

nuat

ion

is s

elec

ted

by S

W1a

. Thi

s gi

ves

0 dB

, 20

dB, 4

0 dB

and

60

dB. T

he o

utpu

tof

SW

1a is

buf

fere

d by

1C

1 w

hich

is a

FE

Tin

put o

p -a

mp.

Thi

s am

plif

ier

has

a ga

inw

hich

is s

witc

habl

e gi

ving

5.5

6 dB

, 15.

56dB

, 25.

56 d

B, 3

5.56

dB

and

45.

56 d

B. B

yse

lect

ing

a co

mbi

natio

n of

thes

e tw

ova

riab

les

the

elev

en r

ange

s fr

om 0

.3 m

V to

30 V

are

obt

aine

d. T

he o

utpu

t of

1C1

for

full

scal

e re

adin

g is

60

mV

.T

he o

utpu

t of

IC1

goes

to th

e 'A

' wt

filte

r ne

twor

k an

d al

so d

irec

tly (

via

R19

)an

d R

V1)

to S

W2.

Thi

s se

lect

s ei

ther

'A'

wei

ghtin

g or

fla

t res

pons

e. A

s th

e fi

lter

has

2.3

dB lo

ss a

t1

kHz

the

"fla

t" p

ositi

on is

also

atte

nuat

ed (

henc

e R

19, R

V1)

tom

aint

ain

calib

ratio

n.T

he r

ms

dete

ctor

IC

pro

vide

s a

gain

of

20 d

B b

efor

e th

e de

tect

or; t

he o

utpu

t of

the

dete

ctor

is a

bout

500

mV

for

ful

l sca

lere

adin

g.T

he p

ower

sup

ply

is s

impl

y a

full

wav

ere

ctif

ied

supp

lygi

ving

both

plus

and

min

us v

olta

ges

of a

bout

20

V, w

hich

are

then

reg

ulat

ed to

±15

V b

y 1C

3 an

d 1C

4.

A K

it of

par

ts f

or th

is p

roje

ct is

ava

ilabl

e fr

omD

omin

ion

Rad

io, s

ee p

age

60.

C11

C::)

Ri=

"18

TO

SW

1

> w (114

113

0R

194:

:::C

7a)

"-71

".R

20cc

N 0 Z

4=

0 11

1-41

1R

23--c..)-- +00 ..

IC3

IC4

Fig

. 3. C

ompo

nent

ove

rlay

and

wiri

ng d

iagr

am.

See

fig

5 fo

r sw

itch

wiri

ng.

O 0 C z C

77 mm

m

RE

SIST

OR

S

All

Y2

W 5

%, e

xcep

t whe

re m

arke

d.

R1

1M1%

R2

100k

1%

R3

10k

1%

R4

12k

1%

R5

150k

R6

100k

R7

10k

1%

R8

561%

R9

1kR

1012

01%

R11

3k9

R12

390

1%

R13

47k

R14

1k5

1%

R15

39k

R16

R17

R18

R19

R20

,21

R22

,23

R24

R 2

5

PA

RT

S L

IST

12k

1%

39k

2%27

k47

k27

k

270k

27k

330

POT

EN

TIO

ME

TE

RS

RV

110

0ktr

imR

V2

200

ohm

trim

CA

PAC

ITO

RS

C1

100n

C2'

15p

C3'

150p

C4*

1n5

poly

este

rce

ram

ic

poly

este

rC

5, 6

*27

nD

l -D

41N

4001

or

sim

ilar

For

clar

ific

atio

n of

com

pone

nt n

otat

ion'

see

"Rea

der

Serv

ice

Info

rmat

ion"

.

C7

C8

C9,

10

C11

C12

C1,

3C

14 -

C18

C19

, 20

15p

100µ

100n

10n

820n

4n7

10µ

100µ

cera

mic

25V

ele

ctro

poly

este

r

cera

mic

poly

este

r25

V e

lect

ro25

V e

lect

ro

The

se c

apac

itors

sho

uld

be a

s ac

cura

te a

spo

ssib

leas

they

affe

ct a

ccur

acy

abov

e10

kHz.

SEM

ICO

ND

UC

TO

RS

IC1

CA

3140

op

amp

IC2

LH

0091

RM

S co

nver

ter

IC3

78L

15re

gula

tor

IC4

79L

15re

gula

tor

MIS

CE

LL

AN

EO

US

PC b

oard

ET

I 13

4SW

12

pole

11

posi

tion

OA

Ksw

itch

SW2

SPD

Tm

inat

ure

togg

lesw

itch

SW3

DPD

Tm

inat

ure

togg

lesw

itch

T1

Tra

nsfo

rmer

120

V.2

8Vct

(28V

ct)

M1

Met

er1m

Asc

aled

assh

own

3 te

rmin

als

(red

, bla

ck g

reen

)B

oxM

etal

bra

cket

s an

d sh

ield

s (s

ee F

ig 7

)3

core

fle

x an

d pl

ugA

lum

inum

fro

nt p

anes

16 p

in s

ocke

t for

IC

2K

nob

= CD

CD

CD

ETI Project48

True RMS Voltmeter

Fig. 4. Printed circuit layout.Full size 90 x 60 mm.

TO COMMONINPUT TERMINAL

Fig. 5. Connection of the range switchdrawn in the 30 V position.

TO INPUTTERMINAL

A

R3

+10

10

20

30

40

50

60

"11111,

10 2 3 4 5 6 7 8 9 0' 2 3 4 5 6 7 8 9 03

WEIGHTING CURVE A

Fig. 6. The response in the "A " weight position.'100

FOLD UP 90°

L.40.- 70

MAI JI1IAt. 188 AL

Fig. 7. The details of the shield,transformer bracket and PC board supportbracket ( two required 1.

2 3 4 5 6 8 0' 2 3

FREQUENCY IN Hz

80

28 28

MATERIAL: 189 AL.

10 r

BEND UP 90° 7 I%

Le,

TMATERIAL: 189 AL.


ETI Project

House AlarmIn these days of increasing crime and vandalism an alarm system for thehome can add greatly to ones peace of mind. To be effective however,not only must the alarm circuitry be well designed, it must also becorrectly installed. This article describes a sophisticated alarm systemand how best to commission it.

WE HAVE EXAMINED the most common forms ofbreak-in and came up with an overview of the burglaryproblem, and how to most effectively counter thewould-be thief. First we'll look at general techniques,and then we present a quite sophisticated alarm unitwith a variety of useful features.

We cannot emphasize enough though, that anyalarm system - no matter how sophisticated - canonly be of use if it is installed correctly. Further theinstallation of an alarm should only be considered aspart of a general awareness of the need for greaterattention to be paid to security. For this reason, beforegoing on to describe the alarm in detail, we shall dealwith domestic security in general, the installation ofalarms and how the specification of our alarm evolved.

HOW THEY GET INNearly 30% of all burglaries are committed by thievesentering via unlocked doors or windows. A further24.4% are committed via forced door locks, and aboutthe same percentage via forced windows.

Thus nearly four out of five potential breakins can beavoided by installing adequate door and windowlocking mechanisms.

Use 'deadbolt' locks on all external doors. Thesecan only be opened with a key - even from the inside- so that even if a thief enters via a window he cannotremove any large items as the doors remain locked andfew thieves will risk passing out items through awindow.

Do have locks fitted by an experienced locksmithunless you have experience in this field - and do notfall for the door-to-door lock salesman - it is notunknown for such people to retain a duplicate key.

Consult a security expert about window lockingdevices. Innumerable type's are available for metal,wood framed and sash windows. A burglar might breakglass but few will risk climbing through a windowframe with broken glass in it.

The precautions outlined above will reduce yourchances of being burgled by about 80% - theremaining 20% can be reduced to almost zero byinstalling a good burglar alarm. The emphasis must beon the word good, a poor alarm may go off erratically,or worse, not at all.

SENSORSFor most premises, it is necessary to install sensors toprotect front and rear doors, windows and garageentrances.

A few forced entries are made through the walls orroof or very occasionally via the floor. Although rare,such forced entries may be guarded against by placingsensors in a strategic passage or area through whichany intruder is likely to pass.

The simplest and most reliable switching device foralarm installations is the magnetic reed switch. Thisconsists of a pair of ferromagnetic contacts in a smallhermetically sealed glass enclosure. The switchcontacts are cantilevered from the ends of the glasstube and overlap slightly at the centre, with a small airgap between them.

When a magnet is brought near the reed switch, theattracting forces increase and overcome the stiffness ofthe reeds, bringing them into contact. When the

Fig. 1. Care must betaken if the reed switchconnecting leads needshortening. Hold wiretightly with pliers (asshown) to preventbreaking the glass seal.


House Alarm

Force door byjemmy or pressure

11.6%

Manipulate door locksby jemmy or pressure 24.4%Open doors by duplicate keys 8.3%

Entry through unlocked doors and windows 29.2%

Fig. 2.This sketch shows how criminals generallyenter a house. Note the very high percentageof entries made via doors and windows foundunlocked.

magnet is removed, the contacts open. The relativedistance for pull -in is less than for drop -out, a valuablefeature as small movements of doors and windows willnot cause false triggering.

Reed switches purchased for alarm installationsmust be of a type specifically intended for thepurpose - standard reed switches will not do.

Many professional security companies install reedswitches and magnets encased in plastic mouldings.Whilst these are neat and simple to fit, it is better toconceal both reed and magnet within the framework ofthe door or window to be protected.

In Figs. 3 and 4 we show just two of the variousmethods of fixing the reeds and magnets (note that themagnet is to be fitted to the moving part of any door orwindow).

Fig. 3. Set the reed switch into the window frame andthe magnet into the moving part.

Roof, wall & Floor entry0.3%

Remove louvres 2.1%Force windowsby jemmy orpressure 24.4%

Break glass in window andoperate catch

6.3%

Window glass may be protected by glueing on aloop of aluminium foil tape (or using a self-adhesivetype of foil). The foil is quite thin and breaks if the glassis fractured. Foil will deter all but the most determinedof burglar. After all, why risk being caught when nextdoor does not seem to be protected by an alarm?

Vibration sensors may be used to protect large areasof glass but these are prone to false triggering duringthunderstorms etc.Many other types of intruder sensing devicesThay alsobe included in the system. Pressure mats for examplecan be placed under carpets in strategic passageways- or even under the door mat. The mats contain alarge number of normally open contacts some of whichwill be closed when the mat is trodden on. The systemcan also include more sophisticated intruder detectors

Fig. 4. To protect a door set the reedswitch into the doorframe.


ETI Project

PE RIME It 13ALARM071((E11 .T12 I

U 116

SILL NINIRY

C113E1111

CIRCUIT 2

1:1111 :1111

111110E11 4

114

ORO

C1113:1111

(ERMA I I

lv

(1' ROM BAl EBY)

OV

SICI111Iro 3,111.3.0

Pill ° nINIC3 I

/ 1C3

9 I.I f 11.111.

,PlY

(AUX 181111.1

Fig. 5. CircUit diagram of the A' board.

. such as infra -red type sensors.The intruder alarm itself should be reasonably

accessible to people entering and leaving the premisesvia a 'silent entry' door, but will be hidden from the sightof an intruder. The alarm's output stage should be arelay which latches when an alarm signal is received.

WARNING DEVICESFor household use a good quality 12 Volt bell shouldprove an adequate warning device. Being mechanicallyresonant, bells have a very high conversion efficiency;in fact, the average bell draws less than 500 mA at12 V yet can be heard several hundred metres away.

Good sirens can be heard well over a few kilometresaway, but they draw a lot of current and cost morethan a good bell. Small cheap sirens cannot berecommended.

If at all possible, householders should make mutualarrangements with neighbours to contact the police ifthe alarm is heard. Similar -arrangements should also bemade so that neighbours can switch off the alarm whenthe police arrive.

10 016

/k

OV 48

0 'BOARD

I011"4';,rsl),`,Ird)

czn10011

11V

C21

T)0011OV

1231

1140

1410430

SW3/2

CIRCUIT A.IC1-3 ARE 4001IC4 IS A 4068IC5 IS A 4001IC6 IS A CA3130IC7,8 ARE NE555THE POWERRAILS OF IC1-1C5ARE NOT SHOWN.PIN 7 IS OV, PIN 14IS +12V.01 IS A TIP 2955Q2 IS A MPS6515D14711 1N914D12,D13 1N4001

0-8. 51 ,..-ELME RGENC Vf1,A

r,9 &NORMALLYOPEN ITIRCE111.

PB3ALARM

I

()V

5 11,13 3

D1311

1 TV

1

11*

VITV

Fig. 6. Circuit diagram of the "B" board.

30 ETI CANADA MARCH 1978

(House Alarm

An Mein which toilets ohm n period of time,Silencing the bell or siren, IN rt IiMllfttl device, thet will bemuch appretiated by the neighbours, Crtro must hetoken to ensure however, that the nhtrm when IiIj et ridnod reset, still provides some measure of protection tothe property.

Whatever the warning device chosen, it should bemounted unobtrusively high up in an inaccessibleplace. The leads to the device should be of on roleorietogauge to avoid any voltage drop associated with a longrun. The wires should be concealed from view,

We strongly recommend that a sepe rate 12 Vbatterybe used in any burglar alarm. This should be checkedat regular intervals to ensure it is still in good conditionand should be replaced as a matter of course when ithas been in service for a period of one year.

ALARM UNITThe specification of our alarm unit is shown in Table 1.From this one can see that the alarm has seven'normally closed' circuits (A2 -A8) plus a silent entry

UNLIKE SOME ALARMS that use a singlesensing loop with all the switches wired inseries, this design features a number ofdifferent alarm groups. These are brokendown into two groups designed fornormally closed (N/C) switches -Perimeter Group (inputs Al -BO andInternal Group (inputs B1-B,m) - togetherwith one group for normally open (N/O)switches (inputs to A).

The Inputs to each of the circuitsdescribed above have their own inputcircuitry.

PERIMETER CIRCUITThe normally closed sensors associated

with the perimeter circuit (inputs to A FA)are connected to the circuitry around IC,and IC2.

These ICs are Quad NOR gates which, inthis application are configured as inverters.The sensors are connected to the inputs ofthese gates via the resistors R ,,FR ,,. Withthe sensor switch closed the output of theassociated IC will be high. If the switch isopened the output will go low as the inputsto the gates are then tied high via resistors12FR,, R, is included to ensure that theinputs to the CMOS ICs are terminatedunder all conditions. The capacitors C FC,together with the resistors RI,FR 17, providea filter to ensure that transients on theinput lines do not trigger the alarm.

In each output of ICI and IC2 there is aLE) which is connected to the SecurityCheck Button (PB,), Upon operation of thisbutton power is supplied to the LEDs whichwill light if the IC they are connected to hasII low output, is the Input Is triggered. Thediodes in series with the LEDs arenecessary because of the low reversevoltage breakdown of the LEDs. Diode Dsupplies power to the Input circuitry duringthe security check. The Input Al providesthe silent entry feature and is describedbelow,

10,1 / Witillp (31 IlittSWI/C/10.1/1/11/p1/010111/0118

HOW IT WORKS

SW1

13/ Alb

POWER

A17

119

A16

SW2

PERIMETER

as S1

SW3

NTERNAL

Ala

PB1

B6SECURITYCHECK

PB3

A19

A9

fp

12V BATTERY

A20

38

ALARMTEST

PB2

S2

ALARMCANCEL

The other sections A2 -A, have theiroutputs fed via an RC network, whichgenerates a negative pulse upon triggering,to one of the inputs of IC4. Thus if any of theinputs are triggered a positive pulse at theoutput of IC4will result.

SILENT ENTRY CIRCUITWith the silent entry circuit a 30 second

delay due to R 2c, C 1, and IC 3 overrides theoutput of ICI, immediately after the alarmhas been energised. After this time if theinput is triggered the output of IC, will gohigh, having been inhibited from doing sountil now by the high output of IC3,1, andwill toggle the RS flip flop formed by IC3,and IC3,4, taking the output of high.After another 30 second delay due to R38Ci,, the input to IC, will be high and itsoutput low.

TRIGGERING CIRCUITThe same output results if one of the

other inputs is triggered and the output ofIC 4goes high momentarily.

This output is used to toggle, via IC,,, theRS flip flop formed by IC 5,4 which is used tocontrol the alarm and resetting circuitrydescribed below.

IC5,2also has two other inputs. The first,consisting of the network Ra C22 and D13.This circuitry disables the alarm functionwhen the Perimeter Switch is in the offposition and for a short period of time afterthe switch is moved to the on position byholding the input of IC5,2 high. Thisprevents spurious triggering.

The second input to IC5,2 is from thenormally open input (A ), as well as theemergency and alarm test switches. If anyof these switches are taken low a negative

i,..going pulse is coupled to IC5,2to trigger thealarm. These functions operate even if theperimeter sensors are off. This input can be

used for emergency inputs such as firealarms.

OUTPUTThe positive going pulse at the output of

I('',,;, mutil the IrtS flip flop IC5,4 and inthis triggered state IC5,3 output is low andI( ' 414 high,

The delay circuitry, uses a CA3130 (IC)configured as a comparator. C ',is normallycharged to + 10V until the flip flop istriggered allowing it to discharge via R.12.When the voltage on C, has fallen to about20mV (the level set by R44 and R45 on thenon -invert log Input of IC,). The output ofthe IC will go high resetting the flip flopsformed by IC two IC ftri and IC3,,o IC3,4. R47 isincluded in the feedback loop to providesome hysteresis.

The output device can either be a relay orsiren circuit. We have provided for bothoptions. The siren output is formed by two555s, one operating at a high frequency anddriving the speaker via driver transistor Q,and the other at about 2Hz which is used tomodulate the frequency of the first. '

The relay and 555s are energised when Q2is turned on by the high output of IC5,4 asthe flip flop is set.

Addition circuits can be added in blocksof four at a time (as Board 13) and connectedto the Aux. input.

AUXILIARY BOARDThe circuitry of board B is almost

identical to that of Board A. The maindifference is that the negative goingoutputs of each IC are ORed using diodesDTI) , as opposed to a logic gate.

This board can only be energised if theperimeter board is powered up. Thecapacitor C,, together with R , and Dprovide a short positive going pulse uponswitch on to disable the main alarm for abrief period of time.

I I ANADA MANCI 11911)31

House Alarm

under these conditions.We have provided a test button so that a check on

the security of the house can be made before the alarmis set, indicating immediately which window or door isopen.

As well as the external circuits the system hasprovision for connecting a number of internal circuits.These may be actuated by normally closed switches -in which case they should be connected to B1 -B4 - orby normally open sensors connected to A9.

It may well be worth considering installing a seriesof emergency push buttons. Such switches should bemounted on the doorframes of the front and rear doors

BOARD ARESISTORS all 1/2 W 5%

R1-8,5489,37,39,41,46,51,53R10-25,48,52R26,38,42R27-34,36,43,49R35F(40

44R45,47R50

CAPACITORS

C1-8,16,18,22,23C9-15,17C19C20,21C24,26C25,27

SEMICONDUCTORS

IC1-3,5IC4IC6IC7,8LED1-8Q1Q2D1-11D12,13

22k47k1k4M71M10k100R220k680R2k2

10u 16 V tantalum47n polyester22u 16 V tantalum100n polyester15n polyester100L: 16 V

CD 4001CD 4068CA 3130555.2" type LEDTIP 2955MPS6515

1N9141N4001

MISCELLANEOUSPCB as pattern,12 V 185R relay.

BOARD BRESISTORS all 1/2 W 5%

R1-4R5,10R6-9,11-14R10R15-18

22k47k1k47k1M

or in a readily accessible position near the doors. Theyenable the occupant to set off the alarm if a caller forceshis way into the house when the door is opened.Although this is not a common event, emergencyswitches provide elderly or timid people with a feelingof security.

Use good quality bell pushes for these circuits andconnect them to the A9 inputs on the circuit board.

FIRE ALARMSFire sensors may be wired across the A9 input. Theactual fire sensors should be mounted in the ceilings of

PARTS LISTCAPACITORSC1

C2-5C6-9

SEMICONDUCTORS

IC1D1-11LED1-4

MISCELLANEOUSPCB as pattern.

33u 16 V electrolytic10u 16 V tantalum47n polyester

CD 40011N914.2" type LED

GENERAL FOR BOARDS A & B.SWITCHES

SW1SW2SW3PB1-3

SPST toggle switchSPDT toggle switchDPDT toggle switchsingle pole press tomake push type.

MISCELLANEOUSCase to suit, 12 V batteryterminal strip, bell or speaker

B6 B1

1972.1crirj1"Ci

Na4ve.igi

Itt61

TO POINT AON BOARD A

OVB8

B7B9

TO AUXI/P

N BOARD A

Fig. 9. Component overlayof the 'B' board.


ETI Project

Al .01---0SILENT ENTRY

CIRCUIT 2

CIRCUIT 3

CIRCUIT 4

CIRCUIT

CIRCUIT 6

CIRCUIT 7

CIRCUIT 8

CIRCUIT 1

CIRCUIT 2

CIRCUIT 3

CIRCUIT 4

N/0 CIRCUITS

EMERGENCY

N

NA

A10 .14 0 N/0

All . 0 N/C RELAY

A 121 0 COMM.

Al3SPEAKER

A14

Cr-43 04)% (lb

88 a \A;

Fig. 10. Connection of the rearterminal block.

1Fig. 11. PCB foil pattern of A' board shown full size (130 x 1161110

CP'42

811141VOC)\4:

1.7-ttn:81

O ETI 00 0Fig. 12. PCB foil pattern of '8' board shownfull size (75 x '65mm).

34 ETI CANADA - MAIRCH 19111

House Alarm

11111111,1111 which there is a fire hazard - kitchen, livingif Him, moms with electrical or heating appliances orwhew people smoke (don't forget the bedroom ifyi111 vil .1 habit of smoking in bed!). Sensors should alsoIii leittalled in the roof of the garage especially if this isanal Neil ip the house - the laundry, workshop etc.

GUNSTRUCTIONII II, to the number of components, it is recommended

Ihril the unit should be built using the PCBs shownItem

Aamemble the components, watching the connectionill ell the polarised components. Also solderHamm:don of all the polarised components. Also solderIlia CMOS ICs last and then solder pins 7 and 14 first.lila allows the protection diodes inside the IC to be

',fleetly°. The LEDs should be mounted parallel to thelit It 4i shown in the overlay as these have to protrude101/111110 holes in the chassis.

hexing of the alarm unit is largely a matter of choice.MI layout can be seen in the photographs. Note thatwa did not fit a key switch to our alarm, but installed itIli n locked cupboard which could also be used for thealoinge of valuables.

SPECIFICATIONS

I yims Of Inputs

',Omit Entry

Porimeter Circuits

Internal Circuits

I mergency Circuits

current Drain

Emergency onlyAlarm activeAlarm sounding

Alarm Time

Silent entryPerimeter circuitsInternal circuitsEmergency circuits

Single circuit,30 s exit delay,30 s entry delay.

7 circuits, N/C contacts,can be expanded in units of 4.

4 circuits, N/C contacts,can be expanded in units of 4.Any number of N/O circuits.

Any number of N/O circuits.These circuits are active evenif perimeter and internal circuitsare switched off.

2.5 mA9 mA500 mA

12 minutes.

Rapippeet, prtalIVILsFERs,,_

-t-

Rub-down panelmarkings

A really high quality systemfor finishing off your pro-jects. The sheets include amass of lettering and controlscales for both rotary andlinear pots.

The lettering is trans-ferred simply by laying on tothe panel and rubbing down- it's strong and permanent.

The markings are on IWO

New from ETI

nol

sheets (a full-sized one cut inhalf for easy postage) andcontain sufficient letteringfor dozens of projects.Send $3.50 (including post-age) to ETI PANEL TRANS-FERS. Unit Six. 25 OverleaBlvd.. Toronto, Ontario, M4H161. Ontario Residents add 7%PST.

Printed Circuit Board materials forthe Hobbyist and Technician.

PAIN D CIRC

11°""40 flr:41.1,1TIINI:i'41;211-4 YOUR OWII 0010010 CIFICADI 000005

0-40MAKE TM OWN PRINTED

CIRCUIT HARM

MAO

1111111111INJECTORALL - from Fingers to Doughnuts a verycomplete line of quality circuit board materials.

omnItronix ltd.

I I I CANADA - MARCH 1978

2056 SOUTH SERVICE 110 IIIANS I /103110 HWY DORVAL, GUE HOP 2N4 PHONE 1014) 083 0093

35

ETI Data Sheet Sound Effects Genet

Texas I nstruments7647/This flexible device should prove useful in a variety of applications,including video games, alarms and more.

THE SN76477 is a bipolar/ I2L device thatprovides a noise source, VCO, low frequencyoscillator, envelope generator, plus variousmixing and control logic on a single 28 pin DILpackage. By the connection of appropriateexternal components and application of logiclevel control signals a wide variety of complexsounds can be synthesized.

The block diagram in Fig. 1 shows the maincircuit blocks, each of which is described indetail below.

SLF (SUPER LOW FREQUENCYOSCILLATOR)The SLF can be operated in the range0.1-30 Hz, the specific trequency Isdetermined by a control resistor connected topin 20, and a capacitor connected to pin 21.The frequency being given by the followingequation:

0.64Fs,. F= Hz

CSIF

VCO (VOLTAGE CONTROLLEDOSCILLATOR)The VCO provides an output whosefrequency is dependent upon a voltage fed toits input, the higher the voltage the lower thefrequency. The control voltage may be eitherthe SLF output, or an external voltage appliedto pin 16, the SLF output being selected whenthe voltage applied to pin 22 is a logic '1', andthe external source when pin 22 is at logic '0'.

The "range" of the VCO is internally set ata ratio of 10:1. Tlie minimum VCO frequencyis determined by a control resistor connectedto pin 18 and a capacitor to pin 17. Thisminimum frequency is given by the equation:

0.64F MIN VCO-

../C0 ,VCO Hz

The "pitch" of the VCO's output is changedby varying the duty cycle of the output. This isachieved by adjusting the ratio of the voltagesat pins 16 and 19. The duty cycle is given bythe following equation:

VCO Duty Cycle=0.5V pin 16

V pin 19

leaving pin 19 high produces an output with50% duty cycle.

MIXERSELECT

C

MIXERSELECT

B

MIXERSELECT

A

MIXEROUTPUT

PIN 27 PIN 25 PIN 26

0 0 0 VCO0 0 1 SLF0 1 0 NOISE0 1 1 VCO/NOISE1 0 0 SLF/NOISE1 0 1 SLF/VCO/NOISE1 1 0' SLF/VCO1 1 1 INHIBIT

Table I

NOISE OSCILLATORThe "noise oscillator" supplies randomfrequencies for the "noise generator". Thenoise oscillator requires a 43 k resistor toground at pin 4. The "noise oscillator"controls the rate of the "noise generator". Anexternal noise oscillator may be used toprovide this control. The external source isapplied to pin 3 and provides an automaticoverride of pin 4.

NOISE GENERATOR/FILTERThe output of the "noise generator" feeds

an internal noise filter. This ''roundsrounds off" thegenerator's output, reducing the HF content ofthe noise. The upper 3 dB point is given by

F UPPER=1 .28

R Nf C NE

where R NF and C NF are external componentsconnected to pins 5 and 6 respectively.

MIXERThe "mixer" logic selects one, or acombination, of the inputs from the SLF, VCO,and noise generator. Selection is according toTable X.

SYSTEM ENABLE LOGICThe "system enable" input provides anenable/inhibit for the system output. Theoutput is inhibited when the voltage at pin 9 isa logic '1', and enabled when logic '0'.

ABSOLUTE MAXIMUM RATINGSAT TA = 25°C (Unless otheiwimlspecified)

SUPPLY VOLTAGE, Vcc (1),PIN 15 6.0V

SUPPLY VOLTAGE, Vcc (2),PIN 14 12.0V

INPUT VOLTAGE APPLIED TOANY DEVICE TERMINAL 6.0V

STORAGE TEMPERATURE-65°C to +150-1:

OPERATING TEMPERATURERANGE . -55° C to +120' (:

LEAD TEMPERATURE1/16 INCH FROM CASEFOR 10 SECONDS . +260'C

RECOMMENDED OPERATINGCONDITIONS

SUPPLYVOLTAGE, Vccl,

MIN TYP MAX WHYS

PIN 15 4.5 5.0 5.5 VSUPPLYVOLTAGE, Vcc2,

PIN 14 5.7 9.0 VOPERATINGFREE -AIRTEMPERATURE 0 25 70 "C

OPERATING CHARACTERISTICSAT TA=25° C AND Vccl = 5.OV


ETI Data Sheet Sound Effects Gonorator

ONL SI (UT LOGICThe "one shot" logic can be used to

,,ounds of a short duration. Theduintior, if the "one-shot" is given by thelollowing equation:I os- 0.8 Ros Cos

where A1,,. and Cesare external componentsconnected to pins 24 and 23 respectively. Theiiiiiximiim duration of the "one-shot" is about

nods.I 11. one-shot" logic is triggered by the

tirtiliii11 edge of the system enable logic controleltp1111

ADL (A F I ACK/DECAY LOGIC)6, The ADL determines the envelope for the

!neon 's output. The envelope selected isdeleinonod by the ADL control inputs to pins 1NIA 78, thu output selected being shown inTable 2.

ENVLLurE GENERATOR ANDMODULA TOR,. 111 ii A decay characteristics of theoutput mu determined by the componentsnonnectecl to pins 7, 8 and 10.

.111.ick.ind decay times are given by the

Pip. I. A voltage fed to the input of the VCO will014117110 the output frequency of this oscillate,.

VOLTAGE FEDTO VCO

T ATTACK= RA C A, p secs

T =RDC. °secswhere CA 1.) is the attack decay capacitorconnected to pin 8, and R and R D are resistor-,connected to pins 7 and 10.

OUTPUT AMPLIFIERThe output amplifier provides a low

impedance output. The peak output voltage isdetermined by the following equation:

3.4 RsV OUT=

AG

where Rs is a summing resistor connected topins 12 and 13 (set equal to 10 k) and RG is again resistor connected to pin 11.

NOTES:1. Supplies greater than 5V0 may be used,

in which case they should be connected to pin14 to allow the internal regulator to supply theinternal circuit requirements.

2. For dedicated sound logic inputs (pins 1,9, 22, 25, 26, 27 and 28) may be hard -wiredto high or low logic levels,

IIFSULTINGVCO OUTPUTWAVE FORM

Int .1 ;en( I 111,1,;.1 i;

SLFCONTROL

O

NOISEOSCILLATOR 0

CONTROL

r

20

121

4

SUPER LOWFREQ. OSC

(SLF)

11.1 11.. It.,1 11, 1.

ONI 111111rNvel 111'1

1.1ATTACK

......,..?" L

ATTACK

ADLSELECT 1

PIN 1

ADLSELECT 2

PIN 280 0

01 0

1

1111 As

I II 1 11'1 I 1

V1 I 1

MIKIII IINI !

VI II WI I II f I II' I 1011'

VCOSELECT

VEXTERNAL

VCO CONTROL0

22

EXTERNAL VCOOR SLFSELECT

16

pan IcoNrihit

n

19

VCO

EXTERNALNOISE OSC. 0(OVER IDE)

SYSTEMENABLE

NOISEOSCILLATOR

9SYSTEMENABLE

LOGIC

ONESHOT

CIRCUIT

NOISEGENERATOR

NOISEFILTER

i

_IL+)

I IN 111111

MIXER

23i 24)3

ONE SHOTCONTROL

ATTACKDECLOGICIt

ENVELOPEGENERATOR

ANDMODULATOR

11 281 26 25 270 a a

81 101

ATTACK LA B Ci ATTACK ATTACK DECAY AMI I. TTUDI

DECAY DECAY CONTROL CONTROL CON I VIOL

t .1 I

It

NI llhlI II II HI. I If) 111111

1

111VI

VIT I 111/1

.4-LI11I 111111111-I V

---- AMP1:I

-'CI

SELECT MIXER TIMINGSELECT CAP

I All 111011r111MMIN11111 FM111111

38ETI CANADA MAR(,' I HUH

The Jungle Telegraph ofthe Twentieth CenturyTrafficing: messages across the continent. Mike Goldstein VE3GFN,discusses this fascinating facet of amateur radio. It's also a service YOU canuse.

WE'VE ALL SEEN THOSE Africanadventure films where, after several "Ithought I saw a man in the bushes"sequences, the castaway hero andheroine clutch each other as the drumsbegin to beat! First, the drums are loud,as the locals "call up the troops". Then,a drummer further away picks up themessage, and repeats it. We know that,within the hour, the message will havebeen relayed to the far reaches of theDark Continent, and every jungleresident not otherwise engaged will behastening to the slaughter. We leaveour hero at this point, counting thebullets in his Webley revolver . . .

The modern version of this jungletelegraph is a message -handlingservice, operated by the radioamateurs of North America.

Briefly, this system consists of aseries of message networks, spanningthe continent, and which operates on adaily basis throughout the year.Operated by the radio "hams" as partof their hobby, the system enablesmessages (radiograms) to be relayedfrom network to network, until the laststation receiving the message deliversit by mail or telephone.

This service is absolutely free, andYOU can use it.

This system of organized, message -relaying networks, known as theNational Traffic System (NTS) issponsored and administered by theAmerican Radio Relay League, (yes,that's where the name comes from!)the international amateur radioorganization in North America. Thesystem was established for emergencycommunications purposes, and that isits prime function. However, emergen-cies being of a rather sporadic nature,the NTS practices daily on ordinarymessage traffic, and any message of anon-commercial nature is cheerfullyaccepted. Messages are the "lifeblood"of a message -handling network, sothey are always "looking for business".

WHO CAN USE IT?Any person who knows about theservice can use it. The user need not bea radio amateur, but need only know aperson who is; not all radio hams areinvolved in the system, but most canstart a message on its way, one way oranother. The service is absolutely freeto the public, and radio amateurs arenot allowed to accept any reward forthe use of their services - beyond agrateful "thanks!".

There is no limit to the number ofmessages one can send, and the lengthof each message can be twenty orthirty words, if necessary. If properaddresses and telephone numbers areprovided, to aid quick delivery, theoriginator often finds himself receivinga reply the same day! Not bad, for afree, volunteer service.

HOW DOES THE SYSTEMWORK?The NTS is a sort of pyramid structure,with each level of the pyramid havingliason with the next higher level, andthe next lower level.

Basically, the NTS divides thecontinent into three Areas - theEastern, Central, and Pacific Area -and each Area has its own major net,the EAN, CAN, and PAN. Represen-tatives, or liasons, carry traffic backand forth between these three Areas.These liason operators form theTranscontinental Corps (TCC), theelite of the NTS.

Each Area has within it a number ofRegions, each Region having its ownRegion Net. The 11th Region, forexample, is made up of EasternCanada the Atlantic Provinces,

AMATEUR RADIO SIMULATED EMERGENCY TESTA massive snowstorm struck the Toronto areaon the afternoon of January 28, paralysingtraffic, creating a breakdown in law -enforcement leading to looting situations, andstranding hundreds of motorists on Highway401, in danger of freezing to death as the stormroared unabated through the ,af ternoon.

As it became obvious that civic authoritieswere going to require assistance in handling awide variety of communications problems, thelocal Amateur Radio Emergency Service swunginto action.

Being a Saturday, a large number of hammobiles were on the roads, stranded along withmany others. These operators collected namesand addresses of motorists wishing to advisefamilies of their predicament, and kept the localpolice advised on traffic conditions.

Amateur radio stations were set up at RedCross headquarters, the local office of theOntario Hospital Association, local Boroughutility offices, and at least one Metro radiostation. Liason was established with the policeradio system.

Soon, traffic reports from on -the -spot mobilesbegan to flow in to the radio station, forbroadcast to the public. The police werereceiving a steady stream of reports ofaccidents, and some looting areas. Boroughsnowplows and trucks were directed by radio tothe Worst -hit areas. Dozens of motorists wereassisted by Red Cross, with food and,

accommodations, after being rescued by

snowmobile teams coordinated by the radiohams.

Ham stations set up at various hospitalscoordinated the arrival of frostbite and heartattack victims, coordinated by the O.H.A. office,using the radio system.

The amateurs, using vhf repeaters for localMetro -wide coverage, and high -frequency radiofor the long -haul messages around the GoldenHorseshoe, demonstrated the ability forassistance in disaster situations, for which theyhave long been famous.

The preceding situations represent the formatfor the annual Simulated Emergency Test, heldon the afternoon of Saturday, January 28, bylocal -area radio amateurs:

Sponsored by the American Radio RelayLeague, this Test is an annual event designed totackle the many problems that exist in any

\ disaster situation, and examine methods ofeliminating them, while training radio amateursin the art of disaster communications. Amateurradio has a long tradition of renderingassistance in time of emergency, and it is onlythrough constant training that the amateurs areable to field a large, experienced group ofcommunicators when the need arises. This Testserves that purpose.

The Test is coordinated in the Toronto area byMichael Goldstein, helped by his assistants PaulEdgley, Lyle Stanway, Saltus Jones, KeithBallinger, and over a hundred radio amateurs.

40 ETI CANADA -- MARCH 1978

Quebec, and Ontario - and this 11thRegion Net is called the EasternCanada Net (ECN).

Each Region net meets once, beforethe Area net convenes, and once afterthe Area net has terminated - somessages are collected for trans-mission out of the Region, on the Areanet, and incoming messages to theRegion are picked up on the Area net,and distributed on the Region net.Therefore, ECN will send both a"Transmit" and a "Receive" liason tothe Area net, EAN.

Each Region net is made up fromliasons from the Section nets which(hopefully) cover all of the geo-graphical area within the Region. TheECN collects liason stations from theAtlantic Provinces Net APN, theQuebec Section Net QSN, the OntarioSouthern Net OSN, the Grey Bruce NetGBN, and the Ontario Phone Net OPN.Messages designated for addresseswithin the Region are passed betweenSections on the ECN, while messagescoming into the Region from the Areanet are distributed ,on ECN to theproper Section net, for local delivery.

Section nets may collect represen-tatives from local nets which are notaffiliated with the NTS, or they may actas the only message -handling systemwithin a Section, or a portion of theSection.

INSIDEThe National Traffic System repre-

sents the only "professional" side ofamateur radio: the traffic schedules arekept religiously, the procedures areadhered to, and the normal "social-izing" that is a tradition in amateurradio is highly discouraged on theformal nets . . . when they open forbusiness, the discipline of theoperators is immediately apparent.The foundations of the NTS were laiddown prior to the Second World War,and many of its present operators havebeen involved in handling traffic formore than thirty years - theirproficiency can only be imagined!

To make life easy, the techniquesand procedures used on all NTS netsacross the country are identical; if aham can operate in one net, he canoperate in any of them, if his codespeed is up to scratch. The formats ofthe radiograms are also identical.

While a certain percentage of NTSbusiness is conducted on radio-telephone, the dyed-in-the-wool trafficoperator is generally a "CW man" -aMorse code operator. The use of Morsecode, far from becoming an obsoletetechnique, has been brought to a highlevel of efficiency on the traffic nets.Operators on traffic nets employ"perfect break-in" - a techniquewhich allows them to receive andtransmit virtually INSTANTANEOUS-LY, and the faster they transmit, thebetter it works. The transmittingstation can immediately detect noiseon the channel, interference, or hischum on the receiving end trying tointerrupt him.

Message -handling in the traffic

networks is only one of manyinteresting aspects of the hobby ofamateur radio. This pastime is pursuedby hundreds of thousands of people,from all walks of life, across thecontinent! Not only does this activityprovide many with excitement of being"in the midst of things", but it preparesand keeps in tune what can be in timesof emergency a vital message service.

"73, old man. I must check into thelocal traffic net. I can hear the drumsbeating . . . "

The keen traffic operator oftenteaches himself to type, as Morse canbe "copied" much faster on atypewriter than with a pencil. Forseveral reasons, formal, writtenmessages can be taken much faster onMorse code than by radiotelephone.The traffic nets use a system ofabbreviations and coded instructions,which allow them to "say a lot" in ahurry. Net operation is often at speedsin excess of twenty-five words -per -minute, at the higher levels.

THE AMERICAN RADIO RELAY LEAGUE

RADIOGRAM7 virrq- 14

To B. CORNWELL,

44 AFFERTON DR.,

TORONTO ONT. 1433 MAR 14TN., 14.010 VSSA01 WAS IMCSIVE0 T

tMATSUN STATION

KKKKKK 00111S11

CITY AND

NANAIMO, BC.

WILL BE IN VANCOUVER NEXT WEEK, HOPE TO MEET YOU MON.

OR TUES. PLEASE TRY TO GET IN TOUCH WITH TONY.

REGARDS,

FRANCES

01 001 4090 NON 001 CCCCCC Kt

REC'D 7F,SENT phoned

ToroM7"ont. 374 1444-'n CCCCCCC

Cs

tNAISION 91000094.0 oat ..... Or 1,49 4 it OC0011009. ..... ....... Alt rmonitA0 iiittiliTIZ9,VISIIII.70107111 11,90:11:4979,

&Olt 0111 11'0: " 01" :Or:: 0: OM CAA OE tiCCOIT11 tie 1001 OWAIRII. INtIVIIIO it AOT euttaim0 On late 0A-i;

A typical radiogram . . someday soon a friend of yours will be receiving a. message on a form likethis.

ETI PROJECT KITS Starting in the APRIL ISSUE - We will be offering

COMPLETE COMPONENT KITS for

ivingstone CURRENT PROJECTS EVERY MONTHI

ElectronicsALSO, we can supply kits for previous ETI PROJECTS

including p.c. boards with component layouts

(SUITE 201) - 801 YORK MILLS RD. and those HARD TO FIND PARTS'DON MILLS, ONT. M3B 1X7 available for immediate delivery


BITS, BYTES and BAUDSby Bill Johnson VE3APZ

Many parts make a whole; Bill Johnson puts it all together.

DURING THE LAST few articles in thisseries, I have attempted to impartenough wisdom to those who haveclosely followed my every word toallow me to go into an explanation of acomplete system in this article.

We have so far learned about variousmeans of communication, the basicarchitecture of the computer, binarynumbering systems, the structure ofthe bus, peripheral and memoryaddressing, I/O transfers, and massstorage devices.

My own introduction to computerstook a very similar approach to thepattern that I have followed in thesearticles. When I had finished learningabout the hardware, I could take aparta computer, shuffle the circuit cards,and put it back together, havingconfidence that I would get it workingagain. (Eventually).

If you are reading these articles toget an insight into how a computerworks, then my task, but for a fewdetails here and there, is finished.However, I suspect that many of youare interested in building a system ofyour own, and once it is built, wouldlike to be able to do something with it.

At this point, you will most likelyhave a picture in your mind of a millionand one things that have to be doneinside a computer to make it workproperly and in a usable fashion. Youprobably have a feeling of utterhelplessness when faced with theenormous task of making a computerdo something. A married man whobuys a computer with no programs willfind that by the time that he hassomething worthwhile to show for hishuge investment in time and money, hedoesn't have a wife to show it to. (Orvice versa). Don't worry, this is normal.

DOING THE HOUSEWORKOne of the 'extras' that should be

bought with any computer is a set ofprograms known as an OPERATINGSYSTEM. An OPERATING SYSTEM,

or 0/S for short, will usually comecomplete with the following:-a) A TEXT EDITOR - for creating andchanging SOURCE (human -readable)programs in plain language,b) An ASSEMBLER - for changingSOURCE programs into OBJECT (ormachine -usable) programs,c) A LINKER - for taking the outputprogram or programs generated by theassembler and joining, or LINKINGthem in such a way that they can runtogether in memory.

Some other handy programs thatmay or may not be included in anoperating system package fall underthe heading of SYSTEM UTILITIES.UTILITY programs allow the user tointerchange files between peripherals,make corrections or PATCHES topreviously -assembled programs with-out the need to re -assemble them,destroy old programs, change pro-gram names, search a program or filefor any particular bit pattern or

instruction, print a listing of a program,or a host of other small, but importantchores.

One type of program that is onlyused while a program is being writtenand debugged is called an ODT, shortfor ONLINE DEBUGGING TECH-NIQUE. This program, when run at thesame time and in conjunction with aprogram that has just been written,allows the programmer to run his newprogram under very tight control, andwith the option of printing a log ofeverything that his program does,starting it under simulation conditions,or stop it if it accidentally gets out ofcontrol. Since this type of program isgenerally used to get rid of bugs in aprogram, it is sometimes called aDynamic Debugging Technique, afterthat other well-known bug -killer, DDT.

An operating system consists of. amain program, called a MONITOR, inaddition to the above programs. Themonitor, as its name implies, monitorsthe entire computer system. One of the

most important functions of themonitor in this regard is to handle allinput/output operations. The reasonfor this is simple: in a large system, ifthere are ten programs runningconcurrently in a computer, thechances are very high that eachprogram will want to send or receivedata to or from an external device atsome time or another. If all theseprograms were allowed to handle theirown I/O by sending instructionsdirectly to discs, cassettes, terminals,etc., the result would be catastrophic- since each program would workwithout reference to all the otherprograms then one program could givea write instruction to a disc whileanother program is trying to' read,giving unpredictable and usuallycatastrophic results.

For this reason, all I/O is handled bythe use of I/O requests made by theuser program to the monitor in theform of codes that can be left in aspecial place for the monitor to findthem. The operating system willreceive requests from all programs inthe system, store the requests in strictorder of priority, and execute the I/O asthe speed of each requested peripheralpermits. The monitor will take care ofall HOUSEKEEPING, such as waitingfor the device to be ready to acceptcommands, formatting the data intothe right -sized blocks for transfer tothe various peripherals, and each userprogram will simply have onecommand to execute to tell the monitorwhat to do and where it can find thedata to be transferred.

SOMEBODY HAS TOLISTEN TO THEOPERATOR

Another of the monitor's functions isto handle all communications with themain console terminal, and itsindispensible accessory, the computeroperator. One may think that this is the


same thing that was just mentionedunder the heading of 'handling the I/O',but there is one major difference -that is that the console terminal canoverride all system activity. Through it,the operator can monitor, suspend, orchange system activity simply byentering pre -defined commands. Itshould be understood at this point thatthese commands are not computerinstructions themselves, but aresequences of letters that meansomething to the monitor. Forexample, assume that a multi -taskoperating system for a large computerhas just been started up. It does notknow the time or date, since itsmemory has just been cleared. (Fromhereon, all computer output will bepreceded by an asterisk (*), and alloperator input will be prefixed with adash (-) ). Here is an example of typicalcomputer -operator dialogue: chart A.

The above example shows how anoperator can talk to the system, andmonitor and control its operation byChart A. Operator - computer dialogue.

*MASTER OPERATING SYSTEM VERSION 02B*ENTER DATE AND TIME YY/MM/DD HH:MM:SS

-DATE 78/01/10 23:15:00*DATE 78/01/10 23:15:00*READY-RUN PRINT

*PRINT 3 0100 05 01FF

-RUN LISTING

*LISTING 3 0200 05 05FF

-RUN VERIFY

*VERIFY*PROGRAM TOO BIG*LOAD ABORTED

-FR

*20 kB FREE

very simple means, using simple,English -like commands. However,before the programs listed above canbe run like this, they must be specified,designed, flowcharted, encoded, anddebugged.

ENTER THEPROGRAMMER

Since the human operator does nothave to think like a machine in order tooperate and understand a computer, itseems only fair that the humanprogrammer should not have toremember the horrendous number ofdifferent possible combinations of bitsthat make up the machine's instructionrepertoire.

Every instruction can be representedby a combination of bits, which form abinary number. When a program isloaded and executed, it must be in thisform so that the computer canunderstand and make use of it. Thisdoes not mean that the programmermust code it in machine language

Monitor announces that it has just been loaded.Asks for date and time, and tells the form inwhich it is to be given.Operator enters information.

Computer indicates that it is ready.

Operator tells monitor to run, or execute programcalled 'PRINT'

Computer prints information about the programand starts executing it.Operator starts another program called LISTING

Computer announces the loading of the secondprogram, and runs it.

Operator tries to start another program.

Computer tries to load new program but findsthat the two programs already in memory takeup too much space for the new one to run.

At this point, the operator is uncertain how much memory is free. He thinks that he shouldhave enough, but obviously, something has been put into the memory that he thinks is free.He thinks that there are 40 kBytes of memory free.

He asks the monitor to report on how muchmemory is free.

Computer replies.

-SY' Operator asks for system status to find out whyhe only has 20 kB free.

*SYSTEM STATUSPROGRAMS IN MEMORYPRINT 3 0100 05 01FFLISTING 3 0200 05 05FFDATA 7 0600 00 2EFF

-KILL PRINT

*EOJ PRINT 05

-RUN VERIFY

*VERIFY 3 0600 05 10FF

*EOJ LISTING 05

*EOJ VERIFY 05

This tells him that one of the programs hasloaded a data buffer that is taking up memory.He remembers that this is done by thePRINT job.

He gets rid of the PRINT JOB. Computerannounces that the PRINT job has gone toend -of -job (i.e. has terminated) and givesits status code.

He tries again.

This time, since there is room in memory,the load is successful.

If left alone, the jobs that are running willannounce their end -of -job as soon as thetasks assigned to them are finished.

though, and in practice this is veryrarely done.

There are several 'levels' ofcomputer programming languagestoday. At the machine level, everyinstruction occupies one, two, or threememory bytes in a microcomputer,depending on the instruction. Theinstruction itself is contained in thefirst byte, and if no memory referenceis needed, then this represents acomplete instruction. An example ofthis is the CMP A instruction in a 6800which complements (changes all thezeroes to ones and vice -versa) the 'A'accumulator. Since no memoryreference is needed, this is a single -byte instruction. If an instructionreferences data in another locationthat is very close, only two bytes maybe required, since only the last eightbits of the address need be given, thefirst eight being the same as theinstruction. If the referenced locationdoes not have the same first eight bitsof address, then the instruction mustbe a three -byte one to give the fullsixteen -bit address.

EFFICIENCY VS.SIMPLICITY

Machine -level programming oper-ates most efficiently at the time the.program is run, because the codeexecuted exactly fits the requirementof the program. It is, however, a verytedidus language in which to programsince a great number of instructionshave to be written, coded, anddebugged, to perform a relativelysimple task. It is called, therefore, low-level programming.

In practice, when a programmeruses low-level programming (i.e. whenhe is forced to use it), he will write thecoding in MNEMONIC form. Thismeans that instead of writing down theactual numbers that the computer willuse, he writes down simple combina-tions of letters that represent theinstruction and are easy to remember.For example, the following code addstwo numbers in the 6800: chart B

As you can see, it is much easier toremember the MNEMONIC forms ofthese instructions.

There is one handicap to the systemdescribed above, however. Let usassume that the above three lines formpart of a large program. The programhas been run and a bug has beendetected, so the programmer wants tochange the program by adding anotherinstruction early on in the program. Allinstructions past this point where thechange takes place will have to bemoved down by one, two, or maybe


Bits Bytes and BaudsMACHINE MNEMONIC

CODE CODE

960024 LDA A 0024

960025 ADD A 0025

970026 STA A 0026

OPERATION

Load accumulator A with thenumber in memory location 24Add accumulator A to thecontents of memory location 25Store the result in location 26

Chart B. Adding two numbers in 6800-ese.

more bytes. This may not seem to beimportant, until you realise that anymemory location past that point thatreferences another memory locationas data will have to be changed, sincethe data address wih now bedifferent.See chart C.It can now be seen that the LDA A, ADDA, and STA A instructions no longerreference the correct data, so they willhave to be changed. The jump will alsohave to be changed, or the computerwill start executing the data in location28 as an instruction. As serious as thisproblem may seem, it can all be solvedwith the addition of LABELS into thesource coding. A program with labelswould look like this:

START:

DATA 1:DATA 2:DATA 3:DATA 4:

CONTINUE:

INSTRINSTRINSTRJUMPDATADATADATADATALDA AADD ASTA AHALT

CONTINUE

DATA 1DATA 2DATA 3

The above program, when convertedto machine code by the assemblerprogram, will generate the code asillustrated on the left side of theprevious example, provided that youtell the assembler that the symboliclabel START is equivalent to location0020. All other labels will be assigned avalue at assembly time referenced tothis base address of 0020 andwhenever an instruction refers to oneof these labels, the assembler willsubstitute the calculated value of thelabel's address into the objectprogram. This way, the program can bechanged at will, leaving the tedious jobof calculating all the memory locationreferences to the number -crunchingcomputer.

MACRO TO THE RESCUEThere are some limitations to the use

of machine code. One is the number ofinstructions required to perform evensimple tasks, and so the general rulewhen designing a system is to use ahigher -level programming language if


at all possible, such as BASIC, orFORTRAN. However, for those whoare constrained for various reasons touse an assembler -level language (andthere are many of them), there is atime- and exasperation -saving deviceknown as the MACRO assembler. Amacro assembler will perform all thefunctions of a straight assembler with avery popular addition: if you don't likethe instructions that the computer iscapable of decoding, then you canmake up your own, using theinstructions that are available, ingroups. For instance, let us assumethat you are writing a process -controlsystem and you have a lot of long-winded multiplications to do. Tomultiply two numbers in a micro-computer can take twenty, thirty, oreven more instructions at the machinelevel. Instead of writing down all theseinstructions every time you want to doa multiplication, the MACRO as-sembler allows you to specify yourown MNEMONIC code and equate itwith a long group of machine -levelinstructions. Every time the assemblerrecognises your own mnemonic in theSOURCE program, it generates all theinstructions that you equated with itand puts them into the output OBJECTprogram.

HIGHER LEVELS YETThe higher -level programming

languages such as FORTRAN, BASIC,COBOL, ALGOL, etc. have theadvantage that they require theprogrammer to write only in simple,

predefined, English statements. Theironly disadvantage is that the code thatthey generate is sometimes muchbigger, therefore taking more memorythan the corresponding assembler -level equivalent program. They wouldtherefore run somewhat slower, and beless attractive to the memory -boundsmall system user. However, duringthe last few years there have emergedsome very good scaled -down, orlimited -function versions of the highlevel languages quite usable by thesmall system user.

X=A+Bis an example of a BASIC programstatement to add 2 numbers. BASICand FORTRAN are'very similar in theirsource language except that FOR-TRAN contains far more sophisticatedstatements. Also, BASIC and FOR-TRAN work in two different ways.

FORTRAN is a COMPILER -typelanguage. That means that once thesource program is written, it must beprocessed by the FORTRAN COM-PILER program, which will generateoutput or OBJECT code that can bedirectly run by the computer withoutfurther need for the compiler program.

BASIC, however, is called anINTERPRETER -type language. There

object code cannot be generated,stored away, and used again on itsown. When a program has been writtenin BASIC and the "RUN" button ispushed, the BASIC program is used asdata for the machine languageINTERPRETER program. The se-quence of events that then take place isanalogous to a line by line compile -then -execute process of the BASICpr6gram.

The big advantage to a languagesuch as BASIC is, of course, its inter -activeness, that the programmer caninteract directly with the program, afeature which makes personal com-puting so attractive.

Chart C. Problems if you don't use labels.

Old New, after the addition of two instructionsAddress Contents Address Contents

0020 instruction 00200021 instruction 00210022 instruction 00220023 jump to 28 00230024 data 00240025 data 00250026 data 00260027 data 00270028 LDA A 0024 00280029 ADD A 0025 0029002A STA A 0026 002A002B instruction 002B002C instruction 002C002D instruction 002D

instructioninstructioninstructioninstructioninstructionjump to 28 (wrong, should be jump to 2A)datadatadatadataLDA A 0024 (wrong, should be LDA A 26)ADD A 0025 (wrong, should be ADD A 27)STA A 0026 (wrong, should be STA A 28)instruction

45

rtnrinrinnnnnnnirinimicrobiooraph

ULIUULJUL1111111JUILI111111

Considering how little has been heard about Signetics in personalcomputing circles, the 2650 is a surprisingly powerful chip.

UP UNTIL 1975, Signetics was afairlywell respected IC manufacturer, es-pecially in the area of bipolar circuitry.However, slipping behind in the field ofMOS technology resulted in financialproblems, and in mid '75 the companywas acquired by Philips NV.The resul-tant sharing of marketing power andtechnology interchange has and will nodoubt will result in benefits to bothorganizations.

nnnnnnnnnnririnri

UULJULJUULJULILJULJU2CMD

The 2650, then, may be considered asPhilips' entry into the microprocessorfield, and they're looking for somepretty hot action around this product. Itis certainly no run of the mill mpu, andincorporates a number of very niceunusual features. Heavy emphasis hasbeen placed on interface -ability. In fact,while other manufacturers almostuniversally released their XXX mpuPLUS support chips (without suchsupport using the XXX is difficult),Signetics originally brought out the2650 all by itself, for use with standardmemories, latches, buffers, etc.

In addition, much attention appearsto have been paid to "minimum" andsmall configurations. Indeed, the 2650would look very good in a controllerapplication, unlike some of the mpuswhich show up in big systems, but arenot in their element as simple con-trollers.

HARDWAREThe 2650 fits into the set of speci-

fications generally expected of a "thirdgeneration" mpu. On the hardware sidethis includes single +5V supply at

100mA, and single phase TTL clock offrequency DC to 1.25MHz, ie, the 2650is all static logic. The instructionexecution times range from 4.8 to 9.6 us

The block diagram and pinouts areshown in Fig. 1 and 2 respectively. Theinternal organization may be roughlydivided into the three sections: Control,Addressing, and Data processing.

Control. In accordance with theemphasis on ease of interfacing, inputand output to/from the mpu is asyn-chronous, which is to say external chipsare not driven from the same clock thatruns the mpu. Two control lines handlethis function. OPREQ informs externaldevices that all information from thempu is valid, and ready foraction. Whenthe memory, (or whatever it might be)has done its thing, it signals back viaOPACK that it has finished.

Other control inputs include tri-stating controls for data and addressbuses, and a "pause" input. Controloutputs are a memory/I0 select line,read/write, write pulse. and run/waitindication. There. are also interruptrequest and acknowledge lines.

Addressing. The 2650 address bus is15 lines wide allowing 32k of memoryspace. Referring to Fig. 1 again, oneadvantageous feature is the built inaddress adder which allows fast in-dexed addressing. In addition there isan 8 level subroutine return addressstack.

Data processing. The main itemsrequired for this function are registersfor storing data and an arithmetic logicunit for processing it. In the 2650 thereare 7 data registers, but only four are inactive use at one time. This is bestvisualized by numbering them RO, R1a,R2a, R3a, R1 b, R2b, Rib. Those in useare then RO plus the "a" series or the"b"series. Note that there is no accum-

ulator as such, since in most instruc-tions any register (of the four in use) canbe thesourceordestination register. ROwould however, appear to be used in amore accumulator like fashion, with theother registers used as loop counters,index registers and so forth.

The "program" status word contains14 bits (in 2 bytes) of information withthe following functions. "Register bankselect" determines which set of regis-ters are in use, "Carry", "Logical/Arithmetic Compare", "Overflow","With/Without Carry", "InterdigitCarry", are all used in processing data.A two bit condition code is providedwhich generally gives >/0/< or +/0/ -based on instruction results.

The PSW also includes the stackpointer indicating the level of thesubroutine stack in use, and the"sense" and "flag" I/O bits, to be dealtwith in the next section.

The two PSW bytes may each beread from or written into.

INPUT/OUTPUTMany forms of I/O are available, so

that external circuit sophisticationneed only match requirements.

The very simpleSt I/O is theFlag/Sense output/input line pair. Flagmay be set to 1 or 0 by writing into oneof the PSW registers, while sense canbe read similarly. With only two gatesattached as buffers we have a Teletypeinterface!

The next level is the simplest parallelI/O scheme. Provision is made for asingle 8 bit I/O port to be readfrom/written to in conjunction with anyone of the four active mpu registers.This port would have a control and adata register. The mpu communicateswith the port via the data bus,Memory/I0 select line and upper two


ADDRESSBUS

A

SUBROUTINE RETURNADDRESS STACK

8.15LIFO

A-N--

INSTRUCTION ADDRESS REGISTER

OPERAND ADDRESS REGISTER

It

REGISTERSTACK25358

t7RO

PROGRAMSTATUSWORD

ULTIPL NE

--v

ALU

CONDITION CODEAND

BRANCH LOGIC

ADDRESS ADDER INTERRUPT 15

INTERRUPTI HOLDING REGISTER I I INSTRUCTION

REGISTERREQUEST

414]. 4 INTERRUPTLOGIC

ACKNOWLEDGE

Fig. 1. Block diagram of 2650 internals.

SENSE

AOR 12

AOR11

A01110

APRA

AORS

GDR]

APRS (-

GORE

A13101

0.0113

A0112

ADA 1

AORO

ADREN

RESET

unisus

/501118.17/C

MIST

25

30

33

32

31

30

29

28

ST

26

75

20

23

22

21

LAG

9CC

CLOCK

MITOPACK

111 N 017 -UT

NT ACK

-3 Deus°DOUS1

0.62DODS]

D01154

MUSE

DOUS6

DEWS/

DBUSEN

OPR EO

A/W

-3 IN RP

CND

Fig. 2. Signetics 2650 pinouts.

address bits which select this mode ofI/O and also tell whether theinformation on the data bus is data orcontrol.

The next level of sophistication is the"Extended" I/O mode. In this case thesame instruction is used, plus one byteto select one of 256 I/O port locations.Then the same lines plus 8 of theaddress lines to actually activate theappropriate I/O device.

Finally, one can always use I/Odevices that masquerade as memory,which the mpu and programmer treatlike any other memory location.

SOFTWAREThe instruction set, shown in Fig. 3,

is a fairly standard collection. Some ofthe interesting features are the

110CONTROL LINES

LOGIC DECODING AND CONTROL LOGIC

conditionals, and especially theaddressing modes.

As previously mentioned, the two bitcondition code determined by theresult of some previous instruction willin one of three states. Each of theconditional instructions allows theprogrammer to include in theinstruction a two bit code. Then,depending on the instruction theconditional action is, or is not taken ifthe programmer's code matches t1-.9condition code.

ADDRESSING MODES

Again, a respectable set of ad-dreSsing modes are provided, registeraddressing, Immediate, Relative,

vcc DNO

2650

SENS

FLAG

WRPBAY

OPACK

OPREO

Am

AD -A,

CLOCK

404B

TIMING LOGIC

DATA BUS

CLOCK

Absolute, the usual favorites.In addition there are Indirect and

Indexed modes. Indirect addressingworks as follows: the instruction opcode is followed by two bytes whichform the address at which to find theaddress of the operand. Indexed modemeans that the two bytes, following theop code, added to the contents of oneof the mpu registers (selected by partof the op code) equals the address ofthe operand. But here's the real heart -winner, auto increment and autodecrement, which Signetics thought-fully included. This mode is the sameas indexed, but will automatically add(or subtract) 1 to/from the indexregister.

Suppose you've a table of data to be

.5

TTYUNIT

DATA BUS

7326

7439

2608

255 X 4 RAM

CE

RISE

2686

256 X 4 RAM

2604

1024 XB ROM

CS

11 .5

24122

11-505/,-0 .6

ADDRESS BUS

NOTES:

1. ONE .50 SUPPLY' SEVEN IC PACKAGES

2. 'CMOS RECEIVER USED FOR HIGH NOISE IMMUNITY.

Fig. 4. A minimal system based on the 2650 and currently available parts.


Microbiography

C.)

W2

cc

00

Fig. 3. Instructions available on the 2650.

MNEMONIC DESCRIPTION OF OPERATION

LOD

STR 1

Z

I

R

A

-ZR

A

Load Register ZeroLoad ImmediateLoad RelativeLoad Absolute

Store Register Zero (r 0)Store RelativeStore Absolute

ADD

SUB

DAR

ZI

R

A

ZI

A

Add to Register Zero w/wo CarryAdd Immediate w/wo CarryAdd Relative w/wo CarryAdd Absolute w/wo Carry

Subtract from Register Zero w/wo BorrowSubtract Immediate w/wo BorrowSubtract Relative w/wo BorrowSubtract Absolute w/wo Borrow

Decimal Adjust Register

AND

IOR

FOR

Z

I

R

A

ZI

R

A

Z

I

A

AND to Register Zero (r 0)AND ImmediateAND RelativeAND Absolute

Inclusive OR to Register ZeroInclusive OR ImmediateInclusive OR RelativeInclusive OR Absolute

Exclusive OR to Register ZeroExclusive OR ImmediateExclusive OR RelativeExclusive OR Absolute

COM .R

Z

I

A ,

Compare to Register Zero Arithmetric/LogicalCompare Immediate Arithmetic/LogicalCompare Relative Arithmetic/LogicalCompare Absolute Artihmetic/Logical

RRRRRL

Rotate Register Right w/wo CarryRotate Register Left w/wo Carry

BCT

BCF

BRN

BIR

.

.

R

A

R

R

A

R

Branch On Condition True RelativeBranch On Condition True Absolute

Branch On Condition False Relative'Branch On Condition False Absolute

Branch On Register Non -Zero RelativeBranch On Register Non -Zero Absolute

Branch On Incrementing Register RelativeBranch On Incrementing Register Absolute

U4ccco

U)

2

MNEMONIC DESCRIPTION OF OPERATION

BDR

ZBRR

BXA

R

A

Branch On Decrementing Register RelativeBranch On Decrementing Register Absolute

Zero Branch Relative, Unconditional

Branch Indexed Absolute, Unconditonal(Note 5)

BST

BSF

BSN

ZBSR

BSXA

'

R

A

R

A

R

A

C

E

Branch To Subroutine On Condition True,Relative

Branch To Subroutine On Condition True,Absolute

Branch To Subroutine On Condition False,Relative

Branch To Subroutine On Condition False,Absolute

Branch To Subroutine On Non -Zero Register,Relative

Branch To Subroutine On Non -Zero Register,Absolute

Zero Branch To Subroutine Relative,Unconditional

Branch To Subroutine, Indexed, AbsoluteUnconditional (Note 5)

Return From Subroutine, ConditionalReturn From Subroutine and Enable

Interrupt, Conditional

WRTDREDDWRTCREDCWRTEREDE

Write DataRead DataWrite ControlRead ControlWrite ExtendedRead Extended

HALTNOPTMI

Halt, Enter Wait StateNo OperationTest Under Mask Immediate

LPS

SPS

CPS

PPS

TPS

U

L

U

L

U

L

U

L

U

L

Load Program Status, UpperLoad Program Status, Lower

Store Program Status, UpperStore Program Status, Lower

Clear Program Status, Upper, MaskedClear Program Status, Lower, Masked

' Preset Program Status, Upper, MaskedPreset Program Status, Lower, Masked

Test Program Status, Upper, MaskedTest Program Status, Lower, Masked


Microbiography

processed, starting at location L. Thefollowing miniprogram shows howhelpful auto increment can be:

Index Reg=0Start Loop

Get byte of data fromlocation L+ Index Regand automaticallyincrement Index Reg.

Process

End loop

Auto increment packs a lot of workinto one statement, possibly resultingin large savings in time and memoryspace.

SUPPORT CHIPSAlready available are two serialcommunications ICS.

2651: Programmable CommunicationsInterface (PCI) is a combined USARTand baud rate generator (50 to 19.2kbps). It includes modem control andsupport of IBM's BISYNC protocol,asynchronous echo and self testing.

2652: Multi -Protocol CommunicationsController (MPCC) is able to receive,format and transmit serial data inSynchronous Data Link Control andother serial systems, at speeds up to500k bps.

To be available in mid 1978 are:2655: Peripheral Programmable Inter-face (PPI), a general purpose parallelI/O device having 3 8 -bit ports.2656: System Memory Interface (SMI)is the other half of a 2650 2 -chip systemand contains 2k bytes ROM, 128 bytesRAM plus an 8 bit I/O port and timerfunction.

Signetics integrated circuits areavailable from Hamilton Avnet inToronto, Ottawa and Montreal andCESCO in those same cities plusQuebec. 'Check with a local Philipsoutlet for Signetics distributors inother areas.

Inside view of the 2651 chip.

50 TI CANADA MARCH 1978

ETI Softspot Horse Jumping Game

Horse JumpingFor CalculatorsThe first in our series of calculator programs, this one for the SinclairProgrammable was submitted by Mr. P. Cornes. A flow chart is given to aidowners of different machines in writing their own versions.

Object - To simulate a show jumpingcourse in such a way that: -

1. The player enters a guess as tohow many strides of accelerationhe thinks will be required by ahorse to clear a fence H feet high.

2. The player is given an indicationof right and wrong guesses.

3. The player's total score is madeavailable to him at the end of thegame.

4. The player's score is madedependent on the value of hisguesses and on his successfullyclearing the fences.

The biggest problem with thisprogram was trying to find a realisticrelationship between the number ofaccelerating strides input and theheight that these strides would enablea horse to jump. The following curveshows the sort of relationship that isrequired.

As you can see from the curve theextra height that the horse can jumpdecreases as the number of stridesincreases, such that after a certainpoint no increase in height is gained byincreasing the number of strides. Thisis the sort of curve you would expect inreality. I have simulated this curve byusing the arctan function. The tan of anangle can take any value between zeroand infinity so the arctan of anynumber between zero and infinity has aradian value between 0 and 1.57 andyou will find that taking the arctan ofany number greater than about twentygives approximately 1.57 as an answer.The only thing to be done now is toscale the arctan values up to give areasonable range of heights, to do thiswe multiply by five.

Looking at the plan of the course youwill see the path connecting the fifteenfences together. The number along-

side each fence is its height (H) and thenumbers on the paths between thefences are the distances in strides toeach fence. If you input these numbersas your guesses then you areguaranteed to clear the fences but youwill find that it is possible to clear mostof the fences in less strides thanshown.

Your sco 'e is calculated by totallingall your guesses round the course andby adding a penalty of nine points foreach fence you do not clear. Youshould consider yourself to be

HEIGHT

FEET

disqualified if you knock down morethan four fences.

If you clear every fence in theminimum number of strides you willend with a score of ninety-five but youshould consider a score of onehundred and ten or less as good.

When you master this course it is asimple matter to change the heights ofthe fences and this creates your owncourse but remember that no fenceshould exceed 7.5 feet in height or youwill not clear it.

2 4 6 8 10 12 14 16 18

NUMBER OF STRIDES

FLOW CHART SYMBOLS

Information Output

/ Data to be Input

Branch

V0

Operations

Program Entry Point

End


ETI Softspot

START7.1'

Horse Jumping GameFLOW CHART

A

Above: a suggested course for the horse race game. All the fenceheights are given in feet, guess the number of strides between the fences.

SOFTSPOT is ETI's programmablecalculator software department. Weknow there are many of you who havegone to a lot of effort to write routinesfor your machines - how aboutsharing the fun. Send us a copy of yourpet program, preferably with flowchart. To make things interesting wewill restrict our choices to only thoseprograms making use of loops orconditionals.

All programs we publish will be paidfor.Mail to: ETI Softspot

Unit 6, 25 Overlea Blvd.,TORONTO, OntarioM4H 1B1

Don't forget to mention what kind ofcalculator you use - and we'd also beinterested to know where you boughtit.

EXECUTION0/AV /sto/AV /AV /go to/0/0/

input H fence 1 /RUN

input strides/ RUN/right-wrong

input H fence 2/RUN/


input last H/RUN


V /Rcl/score

VARIABLESSTOTAL SCORE

STRIDES GUESSH HEIGHT OF NEXT LNG!J HEIGHT OF J111,AU

V/R0109 acorn . s

T 0

Enler 9910 ofnom ttnce

ENTER 11

/1,90 00095110,1 0, 10 10. lonro,

ENTER S

11,011,10 nom,

r15

91.11,09 limo. 9000

J 5 AI:TAN ISI

(scull (

T9

Display 0

PROGRAM

INPUT STRIDES

006 01

Stop 02

UPDATE SCOREWITH STRIDESINPUT

CALCULATEHEIGHT FROMSTRIDESINPUT

VALUE -VE FORFENCE CLEAREDEVE FOR FENCEDOWN

UPDATE SCOREWITH'FENCEDOVVN'PENALTY

FENCE CLEARED/DOWN DISPLAY1 = CLEARED0 = DOWN

A 03MEx 5 04

05Rcl 5 06

07

V A 08MEx 5 09V

arctanA9

10

11

12

3 13

5 14

15

6 16

17

V

GINA 18

193 3 202 2 21

3 229 23

E 24Rcl 5 25

26Sto 2 27Mr 3 28

29'F 30E 31

3 32

INPUT HEIGHT

3334

Stop 0 35


tech -tipsELECTRONIC`SPIROGRAPH'A. Sharp.

The circuit will generate 'Spirograph'patterns on a conventional oscilliscope.The circuit consists of two sinewavegenerators followed by allpass filterswhich we use to phase shift the inputsignals by 900. Applying a sinewave tothe y input gives a circular trace. If asecond set of sin and cos signals aremixed in, a 'Spirograph' pattern is

obtained. A block diagram of thesystem is shown in Fig 1.

RV1 is a balance control which variesthe contribution of each oscillator tothe pattern withdlut affecting the size,so that once set up there is no need toreadjust the gain controls on the oscill-iscope. This type of control can only beused if the oscillators have a lowimpedance output.

SW1 is a reversing switch which hasthe effect of turning the pattern insideout.

An exi ting sinewave oscillator can ofcourse be used and the 60 Hz linecould be employed (attenuated to about2 V RMS from a low voltage trans-former secondary) as the fixed oscill-ator. However flickering is a problemwith lower frequencies (complexpatterns requiring four or more cyclesto complete will flicker at about 10 Hzusing the line frequency as an oscill-ator. I found 150 Hz to be a goodcompromise (higher frequencies requiremore critical tuning).

The allpass filter is recommendedfor phase splitting as it has a unity gainfor all frequencies and settings of RV5.

First connect the y input of thescope to the output of an oscillator andadjust RV2 until a two volt RMSsinewave is obtained, repeat for secondoscillator. Then connect up the x and yinputs as shown in Fig 1, turn thebalance control to one end so as to lookat the output of the fixed oscillatorthen adjust the 100 k pot until a circleis obtained (with suitable x and y gains).Now put the balance control in themiddle and adjust the frequencycontrols until a stable pattern is pro-duced. SW1 and RV1 the balancecontrol can be used to alter the natureof the pattern without affecting itsoverall size, stability or symetry. Ad-just RV5, the phase control (followingthe variable oscillator) for symmetry.- Have fun!

SINE WAVEOSCILLATOR FIXED

(150Hz)

Tech -Tips is an ideas forum and is not aimed at the beginner.ETI is prepared to consider circuits or ideas submitted by

readers for this page. All items used will be paid for. Draw-ings should be as clear as possible and the text should prefer-ably be typed. Circuits must not be subject to copyright.Items for consideration should be sent to ETI TECH -TIPS,Electronics Today International, Unit 6, 25 Overlea Blvd.,Toronto, Ontario, M4H 1B1.

Fig. 1. Block diagramof the 'spirograph' swi

SINE WAVEOSCILLATORVARIABLE

150Hz-1.5kHz

RV1a250k

P-1

O

TO y INPUT

.4111-1°

90" PHASESHIFTER

Fig. 2(a) suitable oscillator for the 'spirograph'

INPUT

R6 100k

+Ve

7

RV5100k

OUTPUT

4

-Ve

90" PHASESHIFTER

Flf1

R910R

RV410kGANGED

Fig. 2(b) Arrangement to give fine control of thefrequency of the oscillator shown in Fig. 2(a). For150 Hz fixed frequency use Rf'=RP=10k

Fig. 3. Phase shifter circuit for use in 'spirograph' circuit.

OV

Fig. 4. PSU for 'Spirograph.

+9V

OV

-9V


Tech Tips

SAMPLE AND HOLD FORMUSIC SYNTHESIZERS

L. Robinson

Sample and hold is a useful effect foruse with music synthesizers and consistsof 'sampling' an input voltage functionsuch as a waveform for a very short timeand then 'holding' it at this selectedvoltage level for the duration of theclock period. This voltage is then usedto control the frequency of a voltagecontrolled oscillator, filter etc.

It is therefore possible to producerandom or repeating sound patterns byvarying the input waveform and fre-quency, pink noise can be used as asample source to create authenticrandom voltages.

The circuit shown is much simplerthan previously designed sample andhold circuits, this is possible by theuse of CMOS technology. The clockoscillator is a standard CMOS squarewave oscillator as found in RCAapplication notes, and this is used toprovide a variable frequency rate from0.2 Hz to 45 Hz. The output then goesto the synthesizer envelope shaperwhich should be of the ADSR type for

KITS FOPETII PIQOJECTS

Commencing with this issue

Watch for

ETI Project KitsComplete with Case

available from

DOMINION RADIO &ELECTRONICS

535 Yonge St. TorontoOntario M4Y 1Y5

Now AvailableETI Project No. 134

True RMS Voltmeter

Price only

$89.95

+9V

C2200n

+7V2

R4 680R

R310k

SAMPLEINPUT

OV

ZD17V2

TOENVELOPESHAPER

IC2/1

-7V2

R5 680R

C4 C550u 50u SAMPLE

INPUT

maximum effect. The clock outputalso goes into a monostable whichproduces an output pulse of approx-imately 20 mS which opens the 4016analog gate for this period. Thevoltage input is therefore sampled andthe value of the amplitude at this pointof the waveform is remembered by thehigh input impedance (1012 Ohms)CA3140 voltage follower. This output isthen used to control the VCO etc. The

_oscillator and monostable can beconstructed from either a CMOS 4001or 4069, ensuring that unused pins areconnected to the high or low power

PROGRAMMABLE GATEThe Programmable Gate is a gate whichconverts an AND gate to an OR gate byapplying a logic '1' on the functioninput.

The logic design uses 8 x 2 inputNAND gates. The' number of gates maybe reduced by replacing the 5 NANDgates enclosed by the dotted line, witha 2 input exclusive OR, such as theTT L 7486.A

1u

-9V

IC1 = 4011IC2 = 4016IC3 = CA3140

OUTPUTFROM HOLD

0/P

supply line via a 1k resistor. The inputwaveform to the analogue switch canhave an amplitude of ± 7 V maximum.If a FET was used as the gate, it wouldonly respond to negative voltages, sothe more expensive analogue switch isused for this reason. The total cost ofthe circuit, including the ± 7V supply,is less than $5.00.

P. Mead

FUNCTIONINPUT

INPUTSOUTPUTA B

0 0 0 0

0 0 1 0

0 1 0 0

0 1 1 1

1 0 0 0

1 0 1 1

1 1 0 1

1 1 1 1

ANDFUNCTION

ORFUNCTION


Tech TipsFM SIGNAL CONDITIONERR. N. Soar

As an alternative to an extra IF stagein an FM tuner, a PLL IC can be usedas a signal conditioner. The VCO ofthe PLL tracks the input signal toprovide a less noisy and stronger signalat its output.

The circuit shown is built aroundthe Signetics NE561B PLL. The onlything necessary is adjustment of the3/30 p trimmer which sets the VCO'scentre frequency to 10.7 MHz.

The circuit should be effectivelyshielded to avoid interaction with theFM front end that provides the circuit'sinput.

MINIMIZING CONNECTIONSM. T. Clarke

Anyone who has connected togethermemory ICs may well be appalled at thenumber of connections, especially thosewhich simply parallel the IC pins.

.4.4

2604

GAKTRON,S

New C BSound Saddle

let's you hearWhat you'vebeen missing

V00

+18V

330R

47n

/I I/

3130p

1.100n

///1//

10.7 MHz IN

Realizing that the address pindesignations are purely notational meansthat address lines can be rearrangedbefore they reach an IC, as convenient.This eases considerably PCB design.

An example is shown where connect-ion of 4K dynamic RAMs (2604) wasundertaken on Vero -board. The copper

2604

21 CONNECTIONS

Von

.19

.510

511

CSIN

OUT

2

V

VSS -'B

47

n6

VOO

CE

NC-45

'3

09W

9 CONNECTIONS

Oaktron Industries puts your CB radio inthe SB Sound Saddle to give you reception younever before thought possible. The speciallymade, built-in 3" x 5" voice communicationspeaker virtually eliminates unwanted high andlow frequency interference - then directsthe sound to you, not to the floor.

Oaktron's CB Sound Saddle is fully adjus-table to almost any transmission hump. Even

100pN/C TOOTHER PINS

10.7 MHz

I-°°UT100p

ov

tracks provide all address connectionsfor every alternate IC without anywiring from the surrounding ICs (thissaved almost 100 connections on a

4K x 16 board).Dynamic RAMs require segregating

the row and column addresses, butwithin each they can be freely mixed.

2604

van

21 CONNECTIONS

if it is not permanently attached, it's designedto ride out any kind of trip with ease, yet isfully portable if you want to remove entireunit.

CB Sound Saddle includes all hardwareneeded, takes 4-6 minutes for custom assembly.No tools needed. Your choise of grilles; BlackEnamel, Walnut Woodgrain or Chrome Plated.All American made for dependability's sake.

Meet a wholenew conceptin CB soundand convenience.1. Powerful 3" x 5"voice communicationspeaker aims clear,crisp sound directly atyou.

2. Puts your CB radiocontrols within easyreach.

3. Customfits completely secure on mostany transmission hump.

4. Permanent or porta-ble installation.

2056 SOUTH SERVICE RD. TRANS CANADA HWY. DORVAL, QUE. H9P 2N4 - PHONE: (514) 683-6993


Date post:	13-Jun-2020
Category:	Documents
Upload:	others
View:	0 times
Download:	0 times

.1! tortimies *WIN -a - americanradiohistory.com · Printed by Livingstone Printing Ltd. News Stand...

Documents