viability of repairing a polymoog keyboard · the omni (and omni ii6) was the best-selling...
TRANSCRIPT
ViabilityofRepairingaPolyMoogKeyboard:acasestudyRichardHallum
SeniorTutor,MAINZSouthernInstituteofTechnology,NewZealand
Abstract
This paper examines the viability of repair of a vintage synthesizer, the MoogPolyMoogKeyboard.Thisinstrumentprovidessomeclassicsoundsandwhenfullyfunctionalispleasingtoplay.Itsmainrenown,however,isitsunreliability.Forthismodelof synthesizer repair cannotbeassumed.These instrumentsweremade insmallnumbersandfewnowexistthatareincompleteworkingorder.TheprojectinvolvedthesuccessfulrepairofadefunctPolyMoogKeyboardwithmultiplefaults,and concludes that, with some provisos, repair is viable. An overview of thePolyMoogdesignrationaleisincluded.
Section1 Background
MyfirstexperiencewithaPolyMoogwasinalocalmusicshopinthelate1970s.Irememberbeingquitetakenwiththeribboncontroller. Idon’trecall theprice,buttheexchangerateoftheNZdollarin1978wasUSD$0.50,sogivenitsoldfor$5000intheUSAitmusthavebeenaroundNZD$10,000.Thiswaswellabovemybudget.
I purchased a PolyMoogKeyboard in 1998. Itwas languishing at the rear of aMusicShopthatwasclosingdown.HowlongithadbeenthereIdon’tknow,butitmusthavebeentradedortakeninforrepair.UpongettingithomeIfounditdidn’t functionatall, sosetaboutrepairing it.Back then, the InternetwasstillfairlynewsoIdidn’thaveaccesstoanyserviceinformation.Ididmanagetofixthepowersupplyandforashorttimegetsomesoundoutofit.However,itdiedagain soon after. At this point in time other priorities took over so I was notprompted to tackle it againuntil 2018. I have found it is actually considerablyeasier to repair now due to being able to source parts online. Apart from theservice information provided by Moog Music, there is still little technicalinformationavailable.
2
FromElectronicMonophonytoPolyphony1
Theinventionofthethermionictriodevalveintheearlypartofthe20thCenturypaved the way for many new electronic inventions, including the musicalsynthesizer.Most instrumentswere focused on the new timbres that could begenerated,andexpressionwasalsoaconsideration.
The engineering restraints of producing single notes (or even a single sound)matchedwellwiththenormsofmanyacousticinstruments,wheretonalinterestand expression are paramount (eg the trumpet or saxophone). Much of earlyelectronics developmentwas dedicated to radio, and heterodyne circuitsweretypically used to form the basis of synthesizers too. Early examples were theTheremin (1920),where hand gestures control the pitch and volume, and theTrautonium(1929)whichusedaresistivewirepressedontoaplate tocontrolthenoteplayed.
Oneobviouswayofmakingtheinstrumentmoreuser-friendlywastoprovideapiano type keyboard for control of the sounds. In the late 1920s the OndesMartenot was invented. It had a keyboard but also a sliding mechanism toprovideportamento.
The Hammond Novachord (1939) had formant filters for modifying theharmonic content of the generated waveform, and envelop shaping to mimicwind and string instruments. Itwas controlled from a keyboard but neither apianoororgantechniquewassuitableforplayingit.Thisprovidedabarriertopotentialcustomersandconsequentlyitfailedtolastlonginthemarket.
It can be seen then, that from early times the design philosophy of thesynthesizerwas split between having a revolutionary instrument, or onewithnewsoundsbutafamiliarinterface.
Themodernhistoryofthesynthesizerbeginsinthe1960swhentwoAmericanengineers worked on parallel (but independent) paths to invent the voltagecontrolledsynthesizer.NeitherBobMoogorDonBuchlaenvisagedakeyboardinstrument.Thedesignidealwastohavetotalsignalcontrol,suchthatanysound(knownorunknown)couldbecreated.Ultimately,R.A.(Bob)Moogcollaboratedwith Herbert Deutsch who presented the idea of controlling the synthesizermodules form a piano-type keyboard. It was this that lead to Wendy Carlos’Switched On Bach (1968) bringing electronic sounds to the mass market,popularising theMoog Synthesizer, and at the same time introducing the ideathatamusicsynthesizerisakeyboardinstrument.
The design dualism between creating an experimental instrument or acommercial one continued into the1970s.By the early1980s thebalancehadswungfirmlyinfavourofsynthesizerswithakeyboard.Withtheintroductionof1Sincethewritingthispaper,aseriesof12videosonthissubjecthasbeenpublishedbyMarcDoty:https://www.youtube.com/watch?v=Yty8Rt7bzOk&list=PLHGZsq10nFV_hYzC5KOXPVfrLfUYNkBed
3
MIDI in 1983 there would be limited further interest in synthesis from anintellectual approach2. Indeed, the original ideal of being able to create anysound, imaginable (or otherwise) was lost. This author remembers manyconversationsofthetime,tryingtoexplainthedifferencebetweenasynthesiserand an electronic organ. Because both had keyboards andwere polyphonic itwasdifficultforthecasualobservertonoticeanydistinction.
In the early 1970s commercial synthesizer designers were preoccupied withproducing a) a portable preconfigured architecture, and b) reliable oscillatortuning(earlysynthesizershadobviousproblemswiththisaspect).
Fig.1.Categorisationofsomenotableearlyelectronicmusicalinstruments.Severaldesignswereinitiallykeyboardless,evolvingtoakeyboardoptionorwithintegratedkeyboard.
2someuniversitieshavecontinuedthisapproach,whilemanymusicdepartmentsnowusecommercialsynthesizers.
Synthesizer
keyboard? YN
polyphonic?YN
Theremin(1922)
OndesMartenot(earlyversions)(1928)
RCAsynthesizers(1950s)
OramicsMachine(1959)
BuchlaBox(1963)
EMSVCS3(1969)
Trautonium(1930)
OndesMartenot(laterversions)
Clavivox(1952)
EMSVCS3(1969)
Telharmonium(1897)
HammondOrgan(1935)
TheWarboFormantOrgel(1937)
Novachord(1939)
Synclavier(1977)
FairlightCMI(1979)
4
FromthisevolutionofsynthesizerdesignitiscleartoseewhattheintentofthePolyMoogwas.Due todemand fromcommercialmusicmakers thesynthesizerbecame a keyboard instrument, and, by 1975, customers were asking forpolyphony. As the keyboard became the acceptedway to play a synthesizer itwasonlynaturalthatplayerswouldrequestmorethanonenoteatatime.
Polyphonic electronic instrumentswere not actually a new idea in the 1970s.ThaddeusCahill’sTeleharmoniumof1906isaninterestingearlyexample.This200 ton gargantuan was the predecessor to the Hammond organ. Both usedelectromechanicalmeans(tonewheels)togenerateafrequencyforeachkey.Theideaofprovidinga fairlysimple tonecircuit foreachkeycontinuedwhen fullyelectronicorgansweredeveloped3.
Thiswayof tonegenerationdesignwasexpensive somanufacturersmoved tousinga ‘dividedown’method. Initiallytwelveseparateoscillators,oneforeachnoteofthescale,wererequired.Thesewouldgeneratethefrequenciesrequiredtosuitthetopoctaveofthekeyboard.Thefrequencieswerethendividedbytwotoproducethenotesoneoctavelower.Thiswasdoneasmanytimesasneededtoprovidetonetoalloctavesofthekeyboard.Inthe1970samorecost-effectivemethodbecameavailableduetotheavailabilityofIntegratedCircuits(ICs).ThismeantaspecialICcouldbedevelopedtoreplacethetwelveoscillators.ThisTopOctave Synthesizer (TOS) could generate the 12 semitones in accurate EqualTemperament for the top octave when driven from a high frequency masteroscillator.
Thehumanear is particularly sensitive to relativepitch so the fact that aTOSarrangement would lock all notes in tune would seem to be advantageous.Although it is highly desirable for instruments to stay in tune overall, slightdetuning between notes is considered a pleasing effect. All electronicinstruments have suffered from the fact that oscillators are essentially blandsounding;acousticinstrumentshavemorerichnessofsound.Tocompensateforthisdeficiency synthesisersmustaddmodulationand time-variant effects.Thegreaterthecontroloftheseparameters,thegreatertheexpressionofthesoundsproduced.ToreturntotheTOS:itprovidedatidytechnicalsolution,butitalsointroducedaproblem.Withnosubtledetuning from independentoscillators itproducesanuncomplexsound.Thetypicalsolutionwastoprovidesomeformofmodulatedeffect(egchorusing)ontheaudiooutput.
Another important way of providing a suitably complex sound is to vary theoscillatorsinsomeway.Techniquesavailableonmonophonicsynthesisincludevibrato,pulsewidthmodulation,andphasingbetweentwo(ormore)oscillatorspervoice.Thisphasingcanalsobelockedorfree-runningtoprovideadditionalvariationstothewaveform.
ThePolyMoogKeyboard4isahybridmusical synthesizer in twosenses. Itusesdigitally controlled analogue circuitry, which was first introduced in Buchla’s
3Unusually,theKorgPE-1000stillusedthismethodinthe1970s
5
System 500 in 1971.5Interestingly, that synthesizer had a 61-note keyboardwhich was highly atypical for a Buchla product. The PolyMoog uses acombinationof7400seriesTTLand4000seriesCMOStocontrol theanaloguecircuits.By the1970s,due to these logic ICsbeing readily available, it becamesensiblefordesignerstousethemforalltheswitchingfunctions.Thisprovidedmore flexibility of design and functionality as well as reduced size, greaterreliabilityandlowercost.
The second hybrid technology used in the PolyMoog is of some interest. Acombination of Master Oscillator (MO) / TOS is used with individual voicecontrolforeachnote.TherawsoundsourceisgeneratedbyfrequencydivisionfromtheMO.Voicecircuits(oneperkey)modifythetoneanddynamicsofthiswaveform, according to signals from the keyboard. The output of the voicecircuits is then fed to a global filter circuit. This resulted in a fully polyphonicinstrument(71notes)withsynthesiscapabilities.
ThiswasaveryambitiousundertakingforMoogandthreeyearselapsedbeforethefirstmodelreachedthemarket.TheR&Dphasewasproblematicduetothecomplexityofdesign.
ARPwerehotontheheelsofMooginthe1970’ssynthesizermarket.Buildingonthesuccessofthe2600andOdysseymono-synths,thecompanyalsobuiltafullypolyphonicsynthesizer.TheOmniwasdevelopedasadirectrivalproducttothePolyMoogsoitisofinteresttocomparethetwo.
The Omni (and Omni II6) was the best-selling synthesizer for ARP, and wasperhapsmost notably used by The Cars7where the string-synth sound can beclearlyheard.Analogsynthesizersoundswereoftenratedonensemblesounds(especially synth-brass), and the quality or ‘fatness’ is essentially governed byhowmanyoscillatorscanbeassignedtoeachvoice.
TheOmniusesasingleColpittsoscillatorforitsMO.Thecircuitcanbeseeninfig.2. In the 1960s, ARP Industries founder (Alan R. Pearlman) invented severalmethods for compensating for transistor thermal variations [1], but thesepatents were assigned to Nexus Industries before ARP was established. Thiscouldpossiblyexplainwhythereisnocontrollooporthermalstabilitycircuitryintheoscillatordesign.
To create some movement in the waveform three parallel BBD circuits areemployed before the synthesizer’s audio output to give a chorus effect. Thesimplicity of design compared to the PolyMoog still gives good results,with amuchmorecost-effective implementation.Anaddedbenefitof thisapproach isthatthe ‘rawer’soundoftheOmnihelpedit ‘cutthrough’ inthemixofthepopsongsofthetimes(inthesamewaythatthinorganelectronicsoundsworkedon60s pop songs, rather than a Hammond). The PolyMoog, thanks to its twodetunedMOs (alongwithFMandPWM),doesprovide sufficient complexityoftonetostandasanunaccompaniedinstrument.5http://120years.net/buchla-synthesisersdonald-buchlausa1963/6asimilarversionwithseparatebassoutput7egin‘MovinginStereo’,‘AllMixedUp’
6
fig.2.TheOmniMO,bufferICandTOS.Source:ARPOmni-2ServiceManual.
Ratherthanusingintegratedsquarewavestoprovidethetonesource,TheOmnidifferentiates the square wave (with a suitably long time-constant) then clipsthissignalwithdiodestoprovideasuitablewaveform.
ThePolyMoogiscapableofproducinglushanaloguesoundsandisbestknownforthelegendary'VoxHumana'popularisedbyGaryNuman8.ThissoundisonlyavailableonthePolyMoogKeyboardwhichhasaseparateboardexclusively togenerate this preset sound. To get the exact soundusedbyNuman, theDirectoutput is used (Incidentally, the lower twooctavesdonot function if theBassFilterswitchisoperatedwhileusingthisoutput).
AcomparisonbetweentheOmniandthePolyMoogisshowninappendix3b.
The essence of a synthesizer is that there is user control over timbre anddynamics of the generated waveform. Early attempts at synthesis weremonophonicasitwasnotpracticabletoimplementthistypeofcontrolformorethanasinglenoteatatime.
Japanese and USA manufacturers jostled throughout the 1970s to bringpolyphonytothemarket.MultiplenotescouldbesoundedonanEmumodularsystem due to a microprocessor controlled keyboard. This involved digitalscanning of the keyboard, and then allocating any played notes to free voicechannels. The system solved the problem of how to offer the bestpolyphony/features/costratio.
Overall, the PolyMoog was a significant development in synthesizer design.Moog’s US Patent 4099439 clearly states that one intention of the instrumentwastoprovidethefeelandsoundofapiano.Thiswasbyandlargeachieved.ThePolyMooghas71notes,fullpolyphony,touchsensitivity,andasustainpedal.Bytoday’sstandardsthepianopresetsoundisunconvincing,butmeasuredagainstelectricandelectronicpianosofthetimeitwouldhavehelditsplace9.BobMoogconsideredthatthedesignwasright: ‘Polymoog,ourmostrecentdevelopment,isanexampleof “musicalengineering”at its finest…I’mconvincedthatMoogoffers
8egin‘Cars”,aNo.1hitforNumanin19799thiswasbeforesampledpianokeyboardsexisted
7
equipment that gives you themost quality, playability andmusical control oversound.’[2]
Thestateoftheelectronicsartinthe1970s
Aquestion comesupwhenevaluatingwhether to repair vintage equipment: ifseveralICshavefailed,whatistopreventmoreICsgoingbadinthenearfuture?Ifasignificantnumberoffailuresislikely,thenrepairmaybeuneconomical.
ThepotentialforICstohaveahighfailureratewasidentifiedearlyon,andthefirst symposiumon the topicwasheld in1962.By the time thePolyMoogwasbeingmanufacturedmost of the teething troubles10had been remedied. Othersmall-scale reliability effects were yet to present themselves (due to theintroductionofnanometerscaling).Itisreasonabletoassume,therefore,thatICsmadeduringthistime(thelate1970s)arequitereliable.However,thisperiodofmanufacture was affected by failure due to environmental causes. In the late1970s the main cause of IC failure was ionic contamination (from the plasticcase),and this issue was related to the way integrated circuits were packaged and mounted on a PCB [3]. This raises the point of difference betweensemiconductorsinuse,andinstorage.
A quote fromDubsounds states“AsiswelldocumentedtheolderICshaveaverythin substrate (insulation between internal transistors) and the substratedeteriorateswithtimesoit'sfullypossibleforanICtofailevenwhennotinuse.”[4]ThiswouldseematvariancewiththefactthatmodernICsaremadewithsmallerdimensions on the substrate. The internal transistors in an IC use layers ofvarious semiconductor materials (Si, Ga, In) along with metals (Al, Au), anddielectrics(TiO3,Pb,Sr)toformquantumheterostructures.InmodernICstheselayersmaybeonlya fewnanometers thick.Older IC technologies(TTL,CMOS)were large geometry by today's standards, so there should be less risk ofinherentfailure.
SemiconductorReliabilityEngineeru/afcagroo [5] commentsonReddit that inthe1970splastic IC cases replaced the very robust ceramic cases. Thesewerepronetofailureduetotheingressofmoisture.Inthelate1970sandearly1980sthis problem was corrected by changing the elements used in earlier designs(such as As and Br) that lead to corrosion. This is a likely explanation whyelectronic equipment from the 1970s that has been in storage fails11. In NewZealand, the average relativehumidity is 81% (range: 73% to92%)12so therereally isnowhere in thewhole country toavoidexposurewithout storage inaclimate-controlledroom.
Other factors of aging (at the atomic level) occur so slowly as to not be ofpracticalconcern.Theauthorhasputmany ICs thatareover40yearsold intoservice without any problem. Notwithstanding moisture issues, it is expected
10suchasthePurplePlague11notwithstandingthatdriedoutelectrolyticcapacitorsarealsoamajorcontributor12datasourcedfromNIWA
8
that the shelf life of an IC would be > 100 years, as without voltage andtemperaturestressesthereislittlelikelihoodoffailure.
Wyrwas [6] observes that one particular failure mechanism (hot carrierinjection) happens at room temperatures, and failure analysis data fromtelecommunicationsequipmentshowsprematurefailureofICswithapredictedlifeof15to20years.
Some ICs are nowknown to have a high failure rate. For example, theRolandJunohassixvoiceICs(80017a)thatarenotoriouslyunreliable.This isalsothecasewiththedividerICsusedinthePolyMoog(MM5823).Thereasonisunclear.Unusually, this IC uses ±15VDC and this may make it more sensitive to anyfluctuationsor faultsonthepowersupplyrails.Thepurposeof the-15Visnotspecifiedinthedatasheet,andtheinternalcircuitryisnotdisclosed.PresumablyitprovidessomeformofbiasingtotheFETs.
CommentsonthePolyMoogreliabilityareinAppendix2.
ThereliabilityofthePolyMoogPSUiscoveredinSection3.
9
Section2 SystemArchitectureandCircuitDescription
For any troubleshooting and repair, it is necessary to understand the normaloperationofanydevice.TheintentionofthissectionistoclarifyandsupplementtheMoogservicedocumentsavailable.[7][8]
DescriptionofthePowerSupplyisdealtwithinSection3.
Overview
The PolyMoog uses a combination of monophonic and polyphonic circuittechniquestoproduceitssound.Severaloftheseareuniquetothis instrumentandMoogheldpatentson these:usingadualMO/TOSanddividersystemtoobtain a chorus effect (USP4228717 and USP4145943), use of a special‘modulator’ circuit block to respond to a velocity sensitive keyboard(USP4099439), and a combined monophonic/polyphonic keyboard (USP4282787).ItishelpfultostudythesedocumentsalongsidethePolyMoogServiceManualtogainanunderstandingofthesystemarchitecture.
fig.3.Versionindicatoriconsusedintheservicemanual.Source:MoogMusic.
An overview of the PolyMoog is given in section 1 of the PolyMoog ServiceManual. Several block diagrams are shown, but a sense of self-evidence ispresumed. Only by studying the circuits can one expect to understand theinstrumentsarchitecture13.Somesectionsofthemanualarecommontoboththesynthesizer and the keyboard versions but not every detail will apply in bothcases. Figure 1-1 (Page 1-10) shows the PolyMoog Flow chart. Althoughdepictingbothversions,thefrontpanelfiltergaincontrolsarenotpresentonthePolyMoogKeyboard. It also fails to indicate that the CONTROL INFORMATIONlines also come from the keyboard. Figure 1-2 (page 1-11) shows the overallblockdiagram.Anamendedversionofthisdrawingisshowninfigure4.Therearetwooscillatorranksprovidingsquarewaveswhicharesuppliedtothevoicecards. The outputwaveforms from there are square and sawtoothwaves. Thekeyboardsendssignalstotherespectivevoicecards,aswellastotheTLboard,andtheTRboard.
13TheServiceManualstates“Duetothecomplexityofthisinstrument,nooverallinterconnectingwiringdiagramexists.”(section1.3).
10
fig.4.BlockDiagram(afterMoogMusic).
ModeSelection
TheMode(Preset)switchescontrolthefilter,envelopeandmodulationsettingsfor each of the 14 sounds. The circuit is described in Section 7 of the ServiceManual.
ControlsignalsareencodedintobinaryandsenttotheTRboard.Here,theyaredecodedusingCD4051multiplexers.Thesedecodethebinarycodeandsenditon14linesto54resistornetworks.Thus,dependingonthepresetselected,aspecificbinarycodeisgeneratedandacorrespondingvalueofcurrentissenttothecontrolcircuitsforADSRandmodulationfunctions.Detailsareshowninappendix14.
ModulationCard
TheModulationCard(VoiceCard)containsthemostcomplexICinthePolyMoog.
Eachof the71notes isprovidedwithaModulationCard.Thesearehoused insocketsonthreemotherboards.EachmotherboardalsocontainsaBalanceCard,the purpose ofwhich is to autocorrect for any level changes of the PWM andsawtoothsignalswhenswitchingbetweenpresets.
11
EachModulationCardtakesadividedownsignalfromMO1(T1)andMO2(T2).ThesearesummedtogetherandsenttoaVEMviaagatingcircuitcontrolledbythe keyboard. T1 and T2 are respectively modulated with a separate LFO, asrequired. This produces a squarewave signal fromT1 that can be pulse-widthmodulated,andasawtoothwavefromT2thatcanbeamplitudemodulated.Thearrangementisshowninfigure5.
fig.5.ModulationCardmodulationrouting.
EachmodulationcardalsohasaVCAandVCF.AcompositesignalofDCplusa20kHzcarrierfromthekeyboardcontrolsthedynamics(attackanddecayrate,sustainlevel)andbrightness(viaa2ndorderLPF).ThepurposeoftheVCFistosetwhetherthereisatimbrevariancewithdynamicsforeachpreset.
Oscillators
ThePolyMoogisunusualforaparaphonicdesigninthatithastwoMOcircuits14.These supply two TOS circuits, but with a one semitone frequency offset. Byusing an additional ÷2 stage for the second TOS output a slightly differentfrequencycanbederived.ATOShasafrequencyerrormarginofgreaterthan±1cent, sufficient to provides a slight detune when the two divider signals arecombined.Thisproducesachoruseffect.
14KorglaterusedasimplifiedversionofthisideainthePE-2000,andLambda
12
Note CentDifferenceC8 0B7 1.13A#7 1.786A7 2.035G#7 1.943G7 1.519F#7 0.944F7 0.13E7 3.347D#7 1.786D7 0.172C#7 2.048C7 0
Table1.TypicalCentsdifference.Source:USPatent4228717.
The generator system is shown in figure 6. Two high frequency relaxationoscillators supply a square wave to two top octave synthesisers, which sendsthirteenfrequenciestotherespectivedividerranks.APLLforeachMOisusedtocontrol the frequency. Two reference oscillators of 1.2kHz control theirrespective MO’s, and a similar frequency from the dividers is applied to thephase comparators. The tune control (Fine Tune) adjusts both referenceoscillatorsby±2semitones.Thedetunecontrol(Beat)allowsforadjustmentofthe second reference oscillator by ± ½ semitone. Two LFO oscillators can beapplied respectively to the two reference oscillators, providing FM. Phasemodulationofoneof thereferenceoscillatorscanalsobeappliedtoproduceachorus effect. Both phase comparators can also be fed from a single referenceoscillator, thus eliminating any detuning effects. The primary reason forMoogusingthisgeneratorsystemwastoprovidevariousmodulationmethodstogiveachorusingeffect.Inthepatent,itisalsoidentifiedthatanadditionaladvantageis the long-term frequency stability of the PLL. Selection of a particularmodulationcombinationisactivatedbytheModecircuits.Detailsareshowninappendix14.
13
fig.6.PLL/modulationsystem.Source:USPatent4228717.
Keyboard
TheKeyboardcontrolcircuitconsistsofapolyphonicsectionandamonophonicsection.
Themonophonicsectionisatypicalresistor ladder.Thekeyboardvoltagegoesto a sample and hold circuit via a bilateral switch which is activated by thekeyboard trigger signal. This provides a CV output suitable for an externalsynthesiser.Thearrangementisshowninfigures7and8.
fig.7.Themonophonickeyboardcircuit.
14
fig.8.SimplifiedcircuitofthemonophonickeyboardCVcircuit.
The Polyphonic section is comprised of individual gate signals sent to eachmodulator card via wires to the motherboards. It also contains a sharedcomposite signal that isused to trigger themodcard filterandVCAaswell astriggerthemainVCF.Whenakeyisdepresseda20kHzpulsewavefromKBLLBis sent to the respectiveModulator card. The attack time is controlled by thepulsewidth,andsustainlevelbythepulseheight.DCfromKBLLBisalsoappliedto anRC circuit associatedwith eachkey.This isused to givevelocity sensingthus providing keyboard dynamics. In this way attack, sustain, and dynamicinformationissuperimposedontheonebuss.Adiagrammaticexplanationoftheinteractionsoftheseparametersisshownonpage4-5oftheServiceManual.
15
fig.9.thepolyphonickeyboardcircuit.
16
Section3 FaultLog
The PolyMoog Keyboard that is the subject of this research is serial number164215. For this section, this number will be used to identify this specificinstrument.
ThePowerSupply
There are two variants of PSU used in PolyMoogs: Moog and Faratron. Thetopologyofboth is identical,butsomeactualcomponentvaluesdiffer. It isnotunusual for a designer to sub-contract the PSU design and manufacuture,probablybecauseitissuchacriticalpartofanyelectronicdevice.Itseemsinthiscase this was a mistake as the Faratron version was unreliable. Moog thendesigned their own version, and later model PolyMoog Synthesizers have itinstalled16.AllPolyMoogKeyboardsusetheMoogPSU.
ThePolyMoogpowersupplyreputationforfailureisnotlimitedtotheFaratrondesign,and1642hadadeadPSUwhenpurchased.ThePSUregulatorisaµA723baseddesign.The723wasreleasedin1968andwasthefirstvoltageregulatorIC.Before3terminalvoltageregulatorswereavailableitwascommonlyusedtoprovidevoltageregulationinlinearpowersuppliesforlowvoltageapplications(egintheARP2600).Itisstillalistedcomponentduetoitsversatilearchitecture.This versatilityhas alsobeen itsArchillesHeel as it requiresmore careful usethana‘plugandplay’3terminalregulator.
The723isanadjustablevoltageregulatorwithoutputvoltagesrangingfrom2to37v. By suitable connection the voltage can be of either polarity. The outputcurrent is amere 150mA, but this can be improved to several amps by usingexternal transistors. The 723 contains four separate regulator building blockswhich canbe configuredasdesiredby thedesigner (fig. 10).A shortcoming isthatithasnointernalcircuitrytopreventexcessiveloadcurrent.
fig10.µA723simplifiedschematic.Source:MoogMusic.
15andonMar12,1642,AbelTasmanwasthefirstEuropeantosightNewZealand16aboveS/N3200.Thismeansabout¾ofPolyMoogSynthesizersusedtheFaratronPSU
17
ForthoseinterestedinlearningmoreabouttheµA723,TheArtofElectronics,3rdEd.[9]isrecommended.
Moogused thesamePSUdesign in theMemoryMoog,witha fewminordesignchanges. In the PolyMoog, the total current on the ±15V supplies is limited to0.75A(ontheMemoryMoogitis1.5A).Bothhavethe+5Vsupplylimitedto1.5A.Thereare256ICs,aswellasseveralLEDs.TheLEDs,PolycomICsandTTLICswill be drawing several milliAmperes each. This requires external passtransistors (MPSU05 and TIP41) in the PSU and these are in a Darlingtonconfiguration. By today’s standards the PSU is fairly basic, but does providecurrentlimiting(initiatedbya0.6Vvoltagedropforthe±15Vsupply,anda1.2Vdropforthe+5Vsupply).
The most probable cause of failure is transistors overheating due to beingpushed outside their SOA when the temperature gets too hot. The powertransistors aremountedon an angled aluminiumslab.This then is bolted to ametalsubframethatrunsthelengthofthePolyMoogattherear.Thermalpasteisappliedat this junction.There is limited thermalconductionavailableas thesubframeislightweightandwithoutheatsinkfins,sooverheatingisapossibility.An improvement can bemade by cutting a hole in the case below the PSU toallow for ventilation. It is also recommended to add a line filter to themainsinputasithasnosuppressionwhatsoever.Presumablytherewaslesselectricalinterference in the 1970s. Owners would be advised to have a mains filterinstalled.
Troubleshooting the PSU is reasonably straight forward. The 723 ICs aresocketedtoprovideeasyreplacement,ifrequired.Allthreesuppliesusethe723inthesamestandardconfiguration.However,thereareafewpointstonote:thePSUsensingrailsleavethePSUboardandareterminatedontheCLboard,wherethePSUvoltagesaredistributed.ThismeansthatremovalofthePSUconnectorwillcausenooutputvoltages,evenonahealthyPSU.Theotherthingtonoteisthatwhereasthe+5Vsupplyisindependent,the723for+5Vsupplyderivesitspower from the+15V supply (as it needs at least9.5V tooperate).To test thePSU voltages, connections between these terminals of the PSU connector willneedtobemade:
supply link+15V 1 2
3 4 -15V 6 7
8 9 +5V 12 13 8 9
14 15
Table2.PSUsupplyandsensinglinespinout.
AfurtherconsiderationistheTranzorbs(transientvoltagesuppressors).ThesearelocatedontheCLboard.Duringtroubleshootingitisoftenusefultoremovepowerfromsomeoftheboards.OnthePolyMoogthiscanbeproblematicasthe±15V supply Transzorbs are 16V types. With less than normal load the PSUvoltagescanexceedthis limitandtheTranszorbswillshuntthePSUrails.This
18
was the case on 1642. As a precaution, the three Transzorbs that protect thethreemain supply railswere replaced. TheseP6KE type devices have a 600Wpeakpower rating.Maximumpeak current therefore diminisheswith a higherbreakdownrating,but isstill26.7Aforthe16voltversion.[10]ThewordingoftheirpurposeinthePolyMoogTechnicalServiceManual(…shouldapowersupplyovervoltageoccur.)issomewhatmisleading.Theyareintendedtostoptransientsappearingonsignalandpower linesandwillnotaccepttheirratedcurrent forprolongedperiods17.TheCLboardislocatedinadifficultpositionandthefrontpanelneeds tobe removed to allowaccess.On1642a slightmodificationwasnecessary for reassembly; theboardmountingholesneeded enlarged to allowfor correct positioning, so that the PRE andVAR switches did not bind on thefrontpanel.
The -15Vsupplywasnotworkingwhen I firstobtained1642.Thiswasdue tothe failure of theMPSU05 transistor. This first repair occurred some20 yearsagowhenIwaslivinginasmallremotetown,soIusedatransistoronhandasasubstitute. Thiswas a BD139which is an audio driver/small amplifier outputdevicebutcanbeusedinthisPSUasithassimilarspecifications,astheyapplyhere.Today,adirectreplacement isavailable fromCentralSemiconductor(theCEN-U05).[11] I have not replaced the BD139 as it is suitable for thisundemandingapplication.The lowermaximumcurrent isofnoconcernas theTIP41 conducts the main current (0.75A). Once the PSU is functional it isadvisable toreplaceallelectrolyticcapacitors.Outputvoltagescanbe trimmedwiththepresetsontheboardtowithinatoleranceof±10mV.Thepre-regulatorDCvoltageswerecheckedforrippleandfoundtobewithintolerance(readingswere1.2Vppfor±15V,and1.0Vppfor+5V).
MPSU05 BD139Vceo 60V 80VHFE(max) 125 250ICBO 0.1µA 0.1µAVCE(SAT) 0.4V 0.5VIC 2.0A 1.5AP@TA=25ºC 1.0W 1.25W
Table3.Q2parametercomparison.[12][13]
The other problemwith the power supplywas failure of two local smoothingcapacitors.OneofthesewasontheTRboard,theotherononeoftheMOboards.Onbothoccasions,theywentshortcircuitwhentheyfailed,causingV+orV-toshut down. This resulted in making all audio circuits inoperable. Thesecapacitorsare tantalumtypeandareprone to failurewhenexposed tovoltagespikes.Allthelocalboardcapacitorswerereplacedasaprecaution(TL,TC,TR,MO). Fault-finding this condition is straightforward with symptoms being noaudio output, but the LED and preset selection still functional (as Vcc isunaffected).Boardscanbedepoweredonebyoneuntilthefaultislocated.
17peakcurrenthandlingreducesto<0.67Aby4.0mS(withtemperaturederating)
19
Toconcludethissectionabriefmentionofotheraspectsofthepowersupplyisgivenhere.The±15Vand+5VrailsaredeliveredtotheCLboard.Reticulationtoallotherboardsisfromthispoint.OntheCLboarda-5.5Vsupplyisderived.Itisused in the trigger circuiton theTRboard.ThePolyMoogcircuitdiagramalsoshows distribution of VCHR. In the PolyMoog Synthesizer this is a regulatedvoltageof4.85V.ThePolyMoogdoesnothavethisandVCHRis5V.VCHRisusedasacommonrailforaudio(egonfront-panelpotentiometers)insteadofanearth.
ModeSelection
ThissectionofthePolyMoogusesTTLICstosendpresetselectionsignalstothedisplay and TR board. A 74148 binary encoder outputs a three bit codecorrespondingtobuttons1to14.
presetno. binarycode tensbit MSBstate1 001 H L2 010 H L3 011 H L4 100 H L5 101 H L6 110 H L7 111 H L8 1000 H H9 1001 H H10 000 L L11 001 L L12 010 L L13 011 L L14 100 L L
Table4.Binarytodecimalencoding.
Thissectionof1642wasfunctionalwhenIpurchasedit20yearsago,butfailedsoonafterthePSUwasrunningagainin2018.Theproblemwastracedtoadead7413 IC.Thepurposeof this IC is to take theEOoutput from theencoderandprovide a one-shot (a delayed 10 µS pulse) to the clock input of a 7475 quadbistablelatch.
The7413 isadual4 inputSchmitt-triggerNANDandon theModeboardbothgatesareusedasinverterswith3inputstiedtoVcc. It is interestingthatactualNANDgateswerenotused.Perhaps itwasdue towhich componentswereonhand (from bulk purchasing).Whilewaiting on a 7413 to arrive I did get thecircuitworkingwitha7400.Thefinalsolutionwasa74LS13asthe7413ishardtosource.TheLSrangeisfullycompatiblewiththe74xxrangeanditisagoodideatoreplacewiththemtolowerpowerconsumption.
20
Fig.11.Timingofthe7475datalatch.
DividerCircuits
Asignificantamountofrepairtimewasspentonthissectionofcircuitry.Thereare24MM5832ICsusedtodividetheMOfrequenciesdowntoaudiofrequenciescorresponding to each note. There are three possible solutions for repair:sourcingNOS,designinganewequivalentcircuitforeachfaultyIC,orpurchasinga replacement from Flatkeys.[14] In hindsight, this would have been thepreferredoption.TosourceanumberofNOSMM5832isdifficult,andthepriceis formidable. An eBay search showed a price of $US50 for a single IC,with asecondsiteofferingsomefor$US120perIC.A difficulty arises in troubleshooting the PolyMoog. There are two probablecausesofadeadnote:a faultyvoicecard,ora faultydividerIC. Ifanote isnotworking at all this is not soproblematic as voice cards are interchangeable soswapping will help prove where the fault lies. Before embarking on this testmethoditisrecommendedtocleanallcardcontactswithisopropylalcohol.Itwasfoundthatasubstantialnumberofnoteswerefaulty,with23keyseitherwithoutsound,orsounding incorrectly.Thereasonswereseveral: intermittentkeycontactfaults,voicecardfaults(deadvoicecards),anddividercircuitfaults.ThedividercircuitICfaultscanonlybelocatedoneatatime.ThisisbecausetheMOfrequencycascadesthroughbanksofthreeICswhicharesolderedontothedividerboard.Theproblemisillustratedinfig.12.
Datatolatcht
A
BC
CLKS
one-shot
tC1
21
Fig.12.TheMM5382ICsaredaisychained.IfIConeisdead,thereisnowayofpredictingifIC2or
IC3arealsodead.UnabletoascertainthesizeofthedividerfaultproblemIoptimisticallythoughtit would be one or two IC dead. On this basis making replacement circuitsseemed feasible. As it turned out, eight ICs needed replaced,which took sometime.TheMM5823isapurpose-builtfrequencydividerforusewithakeyboardTOS.Itprovidesdual÷2and÷4,aswellasanadditionaltwo÷2circuits.[15]Thismakes it an ideal solution for a TOS system. Twenty four ICs are used in thePolyMoog (12 perMO/TOS). It uses PMOS logic running at +15V so from thataspectsubstitutedividersarenotaproblem.ThecloseststandardCMOSICistheCD4520 dual binary up-counter, which contains two four-stage D typeflipflops.[16]
Fig.13.MM5823configuration.To implement the MM5823 functionality two CD4520 are required. This issomewhatwasteful as they are 4 stage counters. They are readily available atlowcost.ThetwoICsmustbehousedonaseparateboard. Inthecaseof1642thissitsperpendiculartothedividerboard.
÷2
÷4
÷2
÷2
÷4
÷2
2 13
14
3 12 4 11
6 9
10
IC1
IC2
IC3
22
Fig.14.Headerpindimensions.Source:https://www.pololu.com.
Thetwo4520ICsareonsmallVeroboardswhichattachverticallyontotheMoogboard using 90º header pins. Ribbon cable is used to make the additionalconnectionsbetweentheboards.ThefirstversionhadtwoCD4520ICsmountedhorizontally on the board. Next, I tried a longer board with the ICs mountedvertically.Thepurposewastotrytofindalayoutthatreducedwiringtime.Thelongerdesignwaseasiertoassemblebutusedmorestraps,sotherewasnogainintime.IalsodiscoveredthattheICsocketsmadethesub-assemblyslightlytoothick toallowthedividerboardwiring loomplugs to fit.Since theseareCMOSICs,ideally,theywouldbesocketedbutitmaybenecessarytosolderthemtotheboard,aswasthecasefor1642.
With multiple divider IC failures, there will likely be more than one faultindication.Somekeyswillhavenooutput,somewillplayinthewrongoctave18,otherswill have seeming tonal, volume, ormodulation issues. Some problemsshowed up on some presets only. This is logical once it is realised that themodulationandenvelopfunctionsareconfiguredforeachpreset.ItisthereforerecommendedtotackleanydeaddividerICsbeforetroubleshootingfurther.
18thisisduetotheleftMOoutputbeing1octavelowerthantherightMO.
23
MM5823 CD4520 wiringpin designation function pin MM5823 4520a 4520b1 Vgg ?,-15v CP1 1 1NC 2 Input1 Not
CP1enable 2 2 1
3 Input2 Q1-0 ÷2 3 3 9 4 Input3 Q1-1 ÷4 4 4 15 NC Q1-2 ÷8 5 5NC 6 Input4 Q1-3 ÷16 6 6 97 Vdd earth MR1 reset 7 7 8,7,15 8,7,158 Vss +15v Vss earth 8 8 16,2,10 16,2,109 Output4a ÷2 CP2 9 9 1110 Output4b ÷4 Not
CP2enable 10 10 12
11 Output3 ÷2 Q2-0 ÷2 11 11 312 Output2 ÷2 Q2-1 ÷4 12 12 11 13 Output1a ÷2 Q2-2 ÷8 13 13 3 14 Output1b ÷4 Q2-3 ÷16 14 14 4 15 MR2 reset 15 16 Vdd +15v 16
Table5.ICpinoutsforthedividerICreplacement.
Otherfaults
Severalotherminorfaultswerediscoveredandrepaired:
o FourofthePRE/VARLEDshadfailedandwerereplaced.o AllVoiceCardconnectorswerecleanedwithIPA.o Oneofthekeyshadabrokentine(whichholdsthekeyinpositionwhen
atrest).Therepairwastofabricateanaluminiumbracketandboltthistothekeyboardsub-frame.ThisisaknownproblemandismentionedinthePolyMoogServiceNotes(1207-33).Thefactorysolutionwastoaddastopbaronlaterinstruments.
o Some keys were operating intermittently. Affected key contacts werecleanedwith CO cleaner. Ideally all key contactswould be adjusted butwithoutthespecialMoogkeyboardadjustingtoolthiswouldbeadifficultprocedure.Togettheresponseofallkeysbacktoperfectoperationwouldrequireconsiderabletimeandcare.Itmaybeworthwhile,asthe‘feel’ofthe PolyMoog keyboard is very good. The Moog Service Notes (993-042314-004) state that ‘Mechanical tolerances in PolyMoog keyboardsareverycriticalfortheydirectlyaffectkeyboarddynamics’.
o Severalrubberkeybushingswerereplaced.o Amissingsliderknobwasreplaced.o An intermittent lossofpower to thePolyPedalwasdue toadirtyCinch
Jonesconnector.Thiswascleaned.
24
o Most of the slider potentiometers were stiff to operate. One wasdisassembledforexamination.Thisestablishedthatthenylonwiperbushwasbindingduetohardenedlubricant.COContactcleanerwasappliedtofreethesliders.
o ThetopC#noteeventuallygavesignalleak-throughwiththeModcardinplace.Thetoptwocardswereremoved.
Calibration
Afullcalibrationwasundertakenoncetheinstrumenthadbeenpoweredforonehour.Mosttestswerestillwithinspecification,withtheexceptionofsomeofthetuningsettings:standardpitch,andfinetunezeroposition.Onecalibrationtestwasnotabletobeperformed: thePhaseModulationamountdidnotshowanychangetothepulsewidthwhenR100wasadjusted.It isunclearwhythiswas.MostofthecalibrationisdoneontheTLboard.OntheTCboardtherearethreeparameterstosetup.Thefiltercutoff,theSwellrange,andtheHarpsichordtone.Thelatterisnotmentionedintheservicemanual,butisadjustedwithR42.Someparameters require setting to a toleranceof±5mVDCsoaDVMwith sufficientresolutionisrequired.
FactoryModifications
1642waschecked forall relevantmodifications.Mostof thesehadbeendone,eitherinserviceoratthefactory.Missingwasonethatputsa1MΩresistorfromearthtothePLLlineintotheMO,sothiswasadded.TheMOboardsinstalledin1642 are the redesignedversion2.On the circuit diagram the1MΩ resistor isshownas a requisite component for this version, however thiswas only fittedfromserialnumber1693on.ThismodificationpreventschirpsintheMOcausedby coupling between the PLL and the MO. An interesting discovery was thatalthoughthedateinsidethelidwas11/78thefiltercircuitboardisdated3/79.There are two plausible reasons for this: 1642 was in the factory for severalmonths,ortheTCboardwasreplacedatsometime.
25
SummaryofFindings
1. ThechanceoffindingaPolyMooginfullworkingconditionisverysmall(seeAppendix2).
2. These days it is potentially much easier to repair vintage equipment,thanks to the Internet. This hopefully provides company serviceinformation,sourcesofparts,andhelpfulforums.
3. The PolyMoog exhibits typical small production run electronicengineering for the era. After this time manufacturers producedsynthesizersusingspecialisedLSIcircuitdesign.
4. Asaresultofthispre-LSItechnologybeingused,understandingofcircuitfunction and obtaining available replacements is generally possible19.There are some exceptions in the PolyMoog: the proprietary voice ICmakesitdifficult(orveryexpensive)toreplaceastheycannowonlybecannibalisedfromanotherPolyMoog,orsourcedfromascarcityofNOS.Ontheotherhand,asthevoicecardsplugintotheMotherboarditiseasytoputdeadcardsontheextremitiesofthekeyboard.
5. Suppliesofmostoftheactiveelectroniccomponentsaregood.Pricesanddeliverycostsdovaryconsiderably.SomeNOS is still available.Asboththe7400seriesand4000serieslogicICsweremass-producedforalongperiodthepossibilityofobtainingthemisgood.However,certainspecificICs in these ranges are becoming rare. All the opamps used in thePolyMoog are still available. OTA’s are now difficult to obtain. A fewspecialised ICs are very difficult to source, namely the MM5823N,MK50240P, µA726, CA3094, D16P1, MPSU05, 4051AE, Polycom VDM7670,andthePolycomIIDM8670.Thepossibilityofgettingno-nameICsisgoodbutitisdubiouswhethertheyareasreliable.Incircuitswheretolerancesare important it isunlikely thatknock-off ICswould functioncorrectly.
6. The technical documentation supplied by Moog for the PolyMoog iscomprehensiveandreasonablegood.Alargenumberofmodificationsaredocumented in Section 2. There are few errors, and descriptions ofoperation are sufficiently detailed to enable an experienced repairerunderstandthecircuitry.Thepublishedscanneddocumentsdolacksomequality and some are too small to read without enlargement. Althoughtherearesomeblockdiagrams,ittakessometimetounderstandexactlywhat everything does, and how all the sections of the PolyMoog fittogether.Thisispartlyduetothedocumentationlackingsomeconnectiondetails, and also to some functional blocks being spread across severalboards(sometimesinanillogicalmanner).
7. The overall construction of the PolyMoog is good, with a heavy solidkeyboard.Whereas it ispossibletorepairelectronicsthatarehousedincheap plastic, there is something satisfying about repairing somethingthatwasbuilttolast.
8. Timerequiredtorepair/restoreaPolyMoogwillbeconsiderable.Thisisduetoboththecomplexityandthefactthatitwasneverareliabledesign.Thismeansthattherewillbemultiplefaultstotraceandfix.
19manyvintagesynthsused3rdpartyproprietaryICs(egCEM,SSM).Since2016someofthesearebeingremanufactured
26
9. The competence level required to tackle aPolyMoog ishigh.Due to theoverall complexity of the design it requires an experienced technician.Some advanced knowledge of digital and analogue circuit techniques isneededtodiagnose,repairandcalibrateallpartsoftheinstrument.Thisassumesavailabilityofthetestequipmentlistedinappendix10.
10. Access is an important part of any repair process. The PolyMoog isgenerallyverygood in this respect.Thecover iseasily removedmakingthe three top boards easily accessible. Most connections are on Molexconnectors,sothereislittleornosolderinginvolvedtoremoveaboard.These boards hinge to allow access to the motherboard, and the voicecardsandMOcardsaresocketedallowingeasyremoval.TheCLandVoxHumanaboardsarenot readilyaccessible,as theyare tuckedunder thefrontpanel.
11. Viabilityofrepairisusuallyrelatedtothevalueoftheitem.InthecaseofthePolyMoogthisisnowquitehigh.Furthertopoint8,foranenthusiastwho is a PolyMoog repair novice, considerable time will be spent infinding theirwayaround this synthesizer. It isexpected thatat least50hourswillberequired.Ifaprofessionalisemployedthetimemaybenoless, but a full refurbishment couldbe expected.Aswell as repairs, thiswould entail replacement of all electrolytic capacitors, allmodificationsperformed,fullcalibration,allkeyresponsesset,andcosmeticdetailsasneededtobringtheinstrumentupto‘asnew’condition.
12. It should be noted that repair is not the same as restoration orrefurbishment. The PolyMoog front panel is quite durable and shouldgenerally not needmuch attention. The keyboard is a differentmatter.Extratimewillbeneededtoachieveidenticalresponsefromeverykey.
27
References
1. TillmanJD.ARPPatents[Internet].2007.Availablefrom:
http://www.till.com/articles/arp/patents.html
2. MoogR.,Norlin.AbluntandtotallybiasedviewpointonelectronicsynthesizersfromBobMoog[Internet].n.d.Availablefrom:http://www.fdiskc.com/syn/
3. MaricauE.,GielenG.AnalogICReliabilityinNanometerCMOS,NewYork:SpringerMedia;
2013.
4. Dubsounds.PolymoogReliability(onceandforall!!!)[Internet].2010.Availablefrom:http://dubsounds.proboards.com/thread/2623/polymoog-reliability-all
5. Reddit.Whatisthelifeexpectancyofanintegratedcircuit?Ifyouburiedoneintheright
environment,what'stheplausibleupperlimitonitstillbeingfunctional?[Internet].2013.Availablefrom:https://www.reddit.com/r/askscience/comments/1nvu7x/what_is_the_life_expectancy_of_an_integrated/
6. WyrwasE.PerformanceandReliabilityofIntegratedCircuitswithinComputingSystems
[Internet].11.07.2011.Availablefrom:https://www.eetimes.com/document.asp?doc_id=1279208
7. MoogMusic.TechnicalServiceManualforMoogPolyMoogSynthesizerandKeyboard.No.
993-042314-004[Internet].1979.Availablefrom:http://www.fdiskc.com/syn/moog/polymoog/servicemanual/polymoog_service_manual.pdf
8. MoogMusic.TechnicalServiceManualforMoogPolyMoogSynthesizerandKeyboard
[Internet].Availablefrom:http://www.univertron.com/SAS/Moog/Polymoog/
9. HorowitzP.,HillW.TheArtofElectronics,Cambridge:CambridgeUniversityPress;2015.
10. Vishay.TRANSZORBTransientVoltageSuppressors[Internet].18Sep12.Availablefrom:https://docs-apac.rs-online.com/webdocs/12fc/0900766b812fc170.pdf
11. CentralSemiconductorCorp.CEN-U05Through-HoleTransistor-BipolarPower(>1A)
NPNGeneralPurposeAmplifier/Switch[Internet].Availablefrom:https://my.centralsemi.com/product/partpage2.php?part=CEN-U05
12. AllTransistors.MPS-U05Datasheet[Internet].Availablefromhttps://alltransistors.com/transistor.php?transistor=62327
13. STMicroelectronics.BD135–BD136.BD139–BD140.Complimentarylowvoltage
transistor[Internet].2008.Availablefrom:https://www.st.com/resource/en/datasheet/bd140.pdf
14. Flatkeys.MM5823N6-stageDividerChip.2018[Internet].Availablefrom:
http://www.flatkeys.co.uk/Dividers/MM5823N.html
15. NationalSemiconductor.MOS/LSIDatabook,SantaClara:NationalSemiconductor;1977.
16. HarrisSemiconductor.CMOSDualUp-Counters.CD4518B,CD4520BTypes[Internet].TexasInstruments.Mar.2004.Availablefrom:http://www.ti.com/lit/ds/symlink/cd4520b.pdf
28
Appendices
1. AshortbiographyofthecareerofDaveLuce2. PolyMoogReliability
3. a)PolyMoogandPolyMoogKeyboardDifferences
b)PolyMoogKeyboardandARPOmniIIComparison
4. ActiveComponentsList
5. ListofAbbreviations
6. ICManufacturersChronology
7. ComponentSuppliers
8. PartsSuppliers–ComponentPrice&ShippingCostComparison
9. ICManufacturerPrefixes
10. TestEquipmentUsed
11. SynthesisCircuitTopologiesusedinthe1970s
12. RepairAssessmentChart
13. ThePolyPedal
14. ParameterSettingsforthePresets
15. SynthesizerTerminology
16. PolyphonicSynthesizersofthe1970s–early80s
29
Appendix1
DaveLuce(1936-2017)
WhileresearchingthistopicIcameuponthefactthatDaveLucediedlastyear.Ashortrecountofhiscareerispresentedhere.
Dave Luce was the designer of the PolyMoog. His exceptional ability withelectronicsgainedhimapositionwithMoogMusicin1972,wherehisrolewastodevelopthispolyphonicsynthesizer.
“FollowingthePolymoog,heheadedtheEngineeringDepartmentatMoog,becamepresident in 1981, and a co-owner in 1984. In these capacities, he was directlyinvolvedwithandoversawthedevelopmentofmanyotherinstruments, includingthe Micromoog, the Mutlimoog, the Moog Taurus, the Moog Source, the MoogLiberation,andtheMemoryMoog.”20
In1987Moogcloseddown,andheacceptedaseniorR&DpositionatAmericanOptical, now called Reichert Technologies. Here he began research on theNoncontactTonometer(NCT),adevicewhichutilizestheresponseoftheeyetoan air puff to measure the pressure in the eye as a means to detecting thepotentialonsetofglaucoma.
In2000,hediscoveredameanstodisentangletheeffectsoftheviscoelasticityofthe cornea from the pressure inside the eye, thereby vastly improving andextending the diagnostic capability of the instrument,whichwas renamed theOcular Response Analyzer. His new measurement method is called CornealHysteresis.
This last phase of his scientific career is of personal interest, as I have aninheritedriskofglaucoma,andgettestedbi-annually.
20https://www.forevermissed.com/david-alan-luce/#lifestory
30
Appendix2
PolyMoogReliability
It is not hard to find anecdotal evidence that the PolyMoog is complex andunreliable.Hereisaselectionofcomments:
“ThePolyMoog'sbiggestproblem,however,wasreliability.Basically,itdidn'thaveany!! It is reputed to be perhapsmusic history'smost unreliable synthesiser andmostofthemspentmoretimeinservicecentresthanstudios!!”21
“TheMoogPolymoog—nowaservicingnightmare.”22
“Polymoogsareoneofthosesynthsthatarealmostinfinitelybetteronpaperthantheyareinreallife.Ohyeah,theysoundHUGE,butactuallygettingthemtomakethatsoundreliably?Nightmareisaveryappropriateword!”23
“ThispolymoogwasoneofthemostchallengingunitsIeverworkedon.Thecircuitdesigniscomplicatedandnotatallintuitiveorfamiliar.”24
“If I could give it a minus 20 I would. The Polymoog is always breaking down.Watchebay,4outof5polyssoldarebroken insomeway.Thepoly isdifficult torepair. In part because of the convoluted design, andmostly the inavailability ofspecialty(OEM)partssuchasthe"polycomchip"andtheMM5328dividerchips.Mostrepairshopswon'ttouchthem.”25
“I'd ratemine at 7maybe 8 on the reliability scale [of 1-10]. If you've seenmypreviouspostsIhadherallworkingaftermyPSUincinerationandthenshemeltedanIContheCLboardfornoapparentreason,otherthanmaybeage?Notagoodsignforthefuture!!!”
“Overallreliabilityratingvaries.Ifyouknowhowtoworkonit:7.Ifyoucanaffordtokeep itwellmaintainedandnevermove it:6. If it isnotwellmaintained,or ityoumoveitaroundalot:4.”26
21http://www.hollowsun.com/nostalgia/vintage/polymoog/index.html22AnalogSynthesizers-MarkJenkins.200723https://www.reddit.com/r/synthesizers24http://fixingelectronics.blogspot.com/2013/05/polymoog-repair.html
25http://www.harmonycentral.com/reviews/product/moog-polymoog/619052
26http://dubsounds.proboards.com/thread/2623/polymoog-reliability-all
31
APolyMoogowner’sdirectoryhasbeensetupontheDubsoundswebsite27.Here82PolyMoogSynthesizersand44PolyMoogKeyboardsarelisted.Arating(outof10)for‘Condition’,anda‘LastRestored’columnareincluded.Fromthese,andcomments, the average state of PolyMoogs can be surmised. The averageconditionratingforthissampleof126instrumentsis7.06/10.
Given that 51 (41%) of these Polymoogs have already been restored this is afairly low rating for reliability. Several of the listings put a ‘not working’commentalongsideaconditionratingof7or8(indicatingthephysicalconditionisgoodbutthekeyboardisnon-functional),sotheoverallreliabilityisthereforesomewhatlessthan71%.
@moogpolymoogonFacebookrecounts thestoryof finallygettingaPolyMoogworkingafterseveralyears(andseveraltechnicians)offaultfinding.28
There is sufficient anecdotal evidence to show that the problems encounteredwith1642arenotanisolatedcase.InordertomakethePolyMoogsoundlikeasynthesizer, rather than an organ, the design became quite complex. ThisresultedinashorterMTBF.29However,notallfaultscanbeputdowntodesigncomplexity.ThedividerICsweresimplypronetolong-termfailure(Moogcouldnothavepredictedthis).TheprimarypointoffailureseemstobethePSU.Thedecision to initially use a third-party supplier may have been driven bymarketingdemands,ratherthangoodengineeringpractice.
27http://www.dubsounds.com/pm_owners.htm28https://www.facebook.com/pg/moogpolymoog/notes/?ref=page_internal
29MeanTimeBetweenFailure
32
Appendix3a
PolyMoogandPolyMoogKeyboardDifferences30
PolyMoog PolyMoogKeyboardYearsofManufacture 1975-80 1978-80Price $5295USD $3000USDVoxHumanapreset No YesAudioReferenceVoltage +4.85V +5VCRboard Yes No
Appendix3b
PolyMoogKeyboardandARPOmniIIComparison
OmniII PolyMoogKeyboardno.ofkeys 49 71cost £1200(1975) £2295(1978)31TOSICtype 1xMK50240N 2xMK50240PdividerICtype CD4520 MM5823Nno.ofICs 81 256touchsensitivity setsreleasetime setsdynamicsenvelopecontrols ADSR attackattacktimeresponse firstnoteonly allnotesfilterADSRcontrols Yes Noenvelopetrigger multiple single/multiplebasssplit Yes Yeslayering Yes(synth+strings) Nofootpedals Yes YesPhaser Yes(3xMM3002BBD) NoPresets No 14PitchController No Ribbon
30TheoriginalsynthesizerwasnamedPolyMoogKeyboard.ThischangedtoPolyMoogSynthesizer,whenthepresetversion(whichtookthenamePolyMoogKeyboard)wasreleased.
31http://www.dubsounds.com/pm_history6.htm
33
Appendix4
ActiveComponentsList
ItemTL TC TR HFOscs Divider Mode CL PSU MBs bal voxhum totals
boardno. 7 8 9 1&2 3 12 10 13 4,5,6 &voice 14OpAmpsLM358N 15 7 6 28MC1458CP-1 3 13 4 3 2 25741 3 1 1 53130 4 4726 2 2LF353 1 1CA3094 1 1TL082 1 1LF356N 1 1LM3080AN 8 8
Transistors2N3904 1 8 2 2 132N3906 2 2 2 62N3392 6 6D16P1 2 2TIP30 1 1TIP41 4 4MPSU05 3 3
TTL7413 1 2 1 47426 1 2 2 2 1 87401 1 17473 2 274148 1 17475 1 17447 1 1
CMOSMM5823 24 244007 2 2 44011 2 24046 2 24051BE 2 24016AE 3 1 44051AE 6 64052B 2 2
MISCCA3046 1 1723 3 3MK50240P 2 2MAN6630 1 1PolycomVDM7670 71 71PolycomIIDM8670 3 3
256
Board
34
Appendix5
ListofAbbreviations
MO MasterOscillator
TOS TopOctaveSynthesizer
PLL Phase-LockedLoop
OTA OperationalTransconductanceAmplifier
Opamp OperationalAmplifier
VEM VirtualEarthMixer
CMOS ComplementaryMetal-OxideSemiconductor
TTL Transistor-TransistorLogic
PSU PowerSupplyUnit
MOD Modulator(Voice)Card
CL LeftControlBoard
DIV DividerBoard
MBL(M)(H) MotherboardLow(Medium)(High)
MODE Presetselector
TC TopCentreBoard
TL TopLeftBoard
TR TopRightBoard
VCC +5VDC
VCH AudioReferenceVoltage
V+ +15VDC
V- -15VDC
SOA SafeOperatingArea
PCB PrintedCircuitBoard
NOS NewOldStock
LSI LargeScaleIntegration
Voice Asignalchannelassignedtoasinglenote
PWM PulseWidthModulation
FM FrequencyModulation
LPF LowPassFilter
LFO LowFrequencyOscillator
CV ControlVoltage
35
Appendix6
SemiconductorManufacturersChronology
ManyICmanufacturingcompanieshavebeensubsumedinto largercompanies.Current companies are shown on the left. The year of acquisition is show inbrackets.ThosewhomadeICsforthePolyMoogareshowninboldtype.
AnalogDevices(2016)
LinearTechnologies-----------------------------------------------------------------
ONSemiconductor(2016)
Fairchild-----------------------------------------------------------------
Renesas(2017)
Intersil(1999)
Harris(1988)
RCA-----------------------------------------------------------------
Maxim-----------------------------------------------------------------
Microchip(2015)
Micrel-----------------------------------------------------------------
TexasInstruments(2011)
NationalSemiconductor-----------------------------------------------------------------
NJR-----------------------------------------------------------------Rohm-----------------------------------------------------------------
NXP(2015)
Motorola
PhilipsSemiconductor(1975)
Signetics
-----------------------------------------------------------------
36
Appendix7
ComponentSuppliers
In general, supplies of 4000 series CMOS and 74xx TTL are slowly drying up.However,manytypesineachseriesarestillreadilyavailable,asaremostoftheopamps.Thehardtogetitemsareshownhere.
Item Usedin Quant. Suppliers Lowestcost(Feb’18)
µA726opamp TLboard 2 UTSourcePortabellabz*VintageSynthParts
unknown10eu
CA3094opamp TCboard 1 UTSource $US2.15LM3080ANopamp TLboard 8 Futerlec $US1.90D16P1transistor HFosc 2 UTSource unknownMPSU05transistor PSU 3 Donberg 13euSN7413TTL TL,HF,mode 4 Futerlec $US0.95MM5823CMOS Divider 24 FlatKeys*
£12.30
CD4051AECMOS TRboard 6 UTSource unknownCA3046tran.array TCboard 1 Futerlec $US1.10MK50240P HFosc 2 VintageSynthParts 31euPolycomVDM7670 Voicecard 71 Noneknown PolycomIIDM8670 Bal.card 3 VintageSynthParts 39.95euPotentiometerknobs Frontpanel Syntaur $US2.95 Keybushing Keyboard 71 Syntaur $US0.70 *equivalent
SuppliersList
General
RSComponents
https://nz.rs-online.com
Element14
http://nz.element14.com
DigiKey
https://www.digikey.co.nz
37
Mouser
https://nz.mouser.com
Jameco
https://www.jameco.com
Futerlec
https://www.futurlec.com
Arrow
https://www.arrow.com
UTSource
https://www.utsource.net
Donberg
https://www.donberg.ie
Vintage
davenjan2000
https://www.ebay.com/usr/davenjan2000
SurplusElectronics
http://www.surplus-electronics-sales.com
SmallBear
http://www.smallbear-electronics.mybigcommerce.com
Portabellabz
http://www.portabellabz.be
VintageSynthParts
http://www.vintagesynthparts.com
Syntaur
https://www.syntaur.com/index.php
FlatKeys
http://www.flatkeys.co.uk/Dividers/MM5823N.html
SynthChaser
http://synthchaser.com
38
Appendix8
PartsSuppliers–ComponentPrice&ShippingCostComparison
Jaycar SIcomp SurplustronicsGlobalPC RS Element14 carzx™ Jameco Futerlec DigiKey WES Arrow UTsource Donberg
$2.15+ $1.90 $1.27+ $0.73 $0.39US $0.22US $0.58 $0.95A $0.40US$2.15 $1.38+ $1.60 $0.45US(10) $0.30US $1.48 $1.77A $0.53US$2.40 $2.17+ $3.10 $0.55US $0.22US $0.80 $0.82A $0.52US
$3.64+ $4.54 $4.09+ $1.95US $1.10US $4.16 $2.40USY
$2.65 $2.48+ $1.48+ $1.11+ $0.49US $0.32US $0.83 $1.14A $0.53US$2.15US
$1.82+ $2.60 $1.20+ $1.15+ $0.69US $0.40US $1.20 $1.50A $0.71US$3.25+ $1.57+ $1.31+ $0.79US $0.85US $1.19 $2.14A $0.77US
$3.95US $1.90US $1.30US(10) 5.9eu
$1.98+ $0.32+(10) $0.33+ $0.08US(10) $0.06US $0.30 $0.36A $0.13US$0.72+ $0.30+(10) $0.39+ $0.12US(10) $0.10US $0.30 $0.25A $0.15US
$0.15US $1.02 $0.51US
Y
$1.00+ $0.35US(10) $0.60US $0.91 $0.37US$5.39+ $1.15+(5) $1.00+ $0.49US $0.45US $1.05 $0.81A $0.54US
13eu
$7.65+ $1.95US $0.95US $1US$1.09+ $0.55US $0.90
$1.85 $0.20US $0.47A 1.09eu$5.30 $1.40+ $1.25 $1.64 $2.03+ $1.79US $0.45US $1.89US
$2.42+ $1.19US $0.75US $2.56US$2.15 $1.22+ $3.25+ $1.89US $0.55US $2.03US$2.40 $3.17+ $1.59US $0.85US $5.00 $1.14A
$1.65 $1.15 $0.87+ $0.45US $0.25US 0.69eu$1.40 $0.81+(10) $0.86+ $0.55 $0.45US $0.25US $0.82 $0.47A $0.45US 0.89eu$2.40 $0.55US $0.45US $1.14A 1.13eu$2.15 $1.75 $0.93+ $0.55 $0.45US $0.50US $0.95A 0.89eu$1.40 $1.20 $1.25 $0.87+ $2.20 $0.45US $0.30US $0.81A 0.38eu
Y$1.40 $1.20 $0.93+ $0.49US $0.40US $1.00A 0.89eu
$2.95US $1.10US $3.59A $0.32US(5)$2.80 $1.95 $0.55US
$0.43 $0.74+ $0.26 $0.40 $0.21+(10) $0.14US $0.04US $0.28+ $0.07A$0.45 $0.28+ $0.33 $0.40 $0.22+(10) $0.14US $0.06US $0.28+ $0.08A$0.49 $0.61+ $0.42 $0.50 $0.07US $0.10A
$1.75 $1.64+ $1.25$1.50
$1.75$2.40 $1.02+ $1.95 $0.33(25) $0.85+ $0.45US $0.45US $1.09A
$6
Jaycar SIcomp SurplustronicsGlobalPC RS Element14 carzx™ Jameco Futerlec DigiKey WES Arrow UTsource Donberg- - $5 - free $16 $3 $17US $5to$7 $26 ? $38US $2to$4US 10eu
free>$66 min$10US free>$66 free>$49US min20eu
39
icbreakout.com www.surplus-electronics-sales.comMouser Syntaurdavenjan2000(ebay) smallbear VintageSynthParts
OpAmps pinsLM358N 8 $0.50USMC1458CP-1 8 $0.60US741 8 $0.65US3130 8 $2.10US726 metalcanLF353 8 $0.65USCA3094 8 $8.35USTL082 8LF356N 8 $0.60USLM3080AN 8 $2.99US $4.95US
xistors sub2N3904 $0.45US(5)2N3906 $0.10US2N3392 $0.92
D16P1
TIP30TIP41MPSU05 9.95eu
TTL7413 147426 14 $8.99US(10) $0.40US7401 147473 1474148 16 $2.587475 167447 16
CMOSMM5823 144007 14 $0.45US4011 14 $0.25US $0.55US4046 16 $0.75US4051BE 164016AE 144051AE 164052B 16
MISCCA3046 14 $2.25US723 14MK50240PMAN6630 LEDdispPolycomVDM7670PolycomIIDM8670
EXTRASICsocket8pin $0.10USICsocket14pin $0.10USICsocket16pin $0.12USbreakout14pin $US4.25headerpin40way90ºheaderpin40way74LS144520DIPveroboard110x140mm
davenjan2000(ebay) smallbear Mousericbreakout.com www.surplus-electronics-sales.com$US12.59 $US11.50 $US14.59 $US16.25 $26
free>$66
40
Appendix9
ICManufacturerPrefixes
Prefix Manufacturer AD Analog Devices Am Advanced Microdevices (AMD) CA, CD RCA DM National Semiconductor H Harris HA Hitachi I Intel ICL, ICM Intersil IDT Integrated Device Technology L, LD Siliconix LF, LH, LM National Semiconductor LT Linear Technology MC, MM Motorola N, NE, SE Signetics PM Precision Monolithics SN, TL Texas Instruments SP Plessey WD Western Digital XR Exar µA Fairchild Semiconductor
Amorecomprehensivelistcanbefoundhere:
http://www.dialelec.com/semiconductorprefixes.html
41
Appendix10
TestEquipmentUsed
Fluke77multimeter
Comment:forcalibrationpurposesaDCvoltmeterwithanaccuracyof±1mVisrequired.
Trio40MHztripletracedelayoscilloscope
Comment:minimumrequirementwouldbeadual trace20Mhzoscilloscope.Amixed signal oscilloscope will remove the need for a Logic Probe (or LogicAnalyser).
PeakLCRmeter
Comment:severalprecisioncomponentsareusedinthePolyMoog.AnaccurateLCRbridgeisrequired.
STCLogicProbe
Comment: a probe with both TTL and CMOS capability is required. A Pulsecapturefunctionisdesirable.
HP5300BFrequencyCounter
Comment:anaccuratefrequencycounterisrequiredtosetthetuningtoA=440Hz
42
Appendix11
SynthesisCircuitTopologiesusedinthe1970s
The electronic circuitry available in the 1970s led to engineers using someinnovative building blocks in analogue synthesiser design. Some of these areparticulartosynthesis,whereasothersweretakenfromotherfields(egradiointhecaseofthePLL’sandtheMC1495four-quadrantmultiplier).
1. TheCMOSAnalogueSwitch
The CD4016 quad bilateral switch was used extensively to replace individualJFET’s for thepurposeof switchingaudiosignalpaths.Here, the termbilateralindicates that each electronic switch (a SPST configuration) is non-polar. Theinternal circuitry to achieve this function is an N-channel MOSFET in parallelwithaP-channelMOSFET.
AuseoftheCD4016isshowninfig.1.IC3AandIC3DareconfiguredasapartialDPDTswitch.IC3BactsasaninvertersothatwhenIC3Aison,IC3Disoff(andviceversa).TheCD4016isalsousedintheTRboardtoselectattacksettings,andPWMandFMamountforvariouspresets.
fig.1.Thebassfilterselectcircuit.Source:MoogMusic.
43
When used to switch op amp inputs, care is needed to ensure that the onresistance(RON)doesnotaffectthegain.Forthisreason,itispreferabletouseanon-invertingopampconfiguration toavoidvariations in the input resistance.Ideally, theswitchshouldbe followedbyaunitygainbuffer. Ifan invertingopampisrequired,largeresistorvaluesshouldbeusedtominimisetheeffect.
A later pin-compatible version (the CD4066) superseded it with betterspecifications (a lower and more consistent on-state resistance). As well asswitching,highfrequencypulsingtechniqueswereused,forexampleintheETI/Maplinsynthesizeroscillator.TheanalogueswitchspecifiedistherareCD4416.All three typesexhibit the samebasic characteristics.However, care shouldbetakentoreplace likewith like.TheCD4016hasa typicalRONof280Ω,whereasmodernanalogueswitchescanhaveRONfiguresof<1Ω.TheRONohmicvalueandlinearity are affected by supply voltage (with RON reducing with increasedvoltage),andinputvoltage.
fig.2.RONversesVIN.Source:MaximIntegratedApplicationNote5299
Subtle design differences will affect performance in analogue circuits. The onresistancewill be of interestwhere transistor biasing is involved, and pulsinguse will require equivalent dynamic characteristics32. A lower RON figure willresult in increased input capacitance. For S/H applications the CD4016 isrecommended.33DistortionoftheCD4016ishighbytoday’sstandards(at0.5%itistwoordersgreaterthanthebestperformingmodernanalogueswitch).
32examplesincludetheETIsynthesizerVCOandVCFcircuits33T.I.Datasheetschs051g–November1998–revisedJune2017
44
fig.3.CombinedN-channelandP-channeltransferfunction.Source:AllenandHolberg-CMOS
AnalogCircuitDesign.
2. ThePhaseLockedLoop
ThePhaseLockedLoop(PLL)isatypeofnegativefeedbackcontrolsystem,withthetransferfunction:
𝑌(#)𝑋(#)
= 𝐺(#)
1 +𝐺(#)𝐻(#)
fig.4.ThePLLmodel.Source:AnalogueDevisesMT-086.
Referringtofigure4,aHFVCOgeneratesanoutputatFo.ThisoutputisalsodividedbyNandsenttotheerrordetector.FREFcanbeaLFreferenceoscillatorthatisalsoaninputtotheerrordetector.Thiselementactsasaphasecomparatorbetweenthesetwosignalsgeneratinganoutputaccordingtothephasedifference,whichisfedtotheVCO.TheVCOwillchangefrequencytoeliminateanyphasedifferencebetweenitandFREF.Aloopfilterisrequiredtoband-limitthecontrolsignaltoensuresystemstability.
45
fig.5.exampleofaPLLcontrollingawaveformgenerator.Source:http://www.seekic.com.
3. TheLadderFilter
Possiblythemostfamousofalltheanaloguesynthesizerbuildingblocks,theMoogladderfiltergivesadistinctivesound.Thisispartlyduetodistortionofthetransistors,butthemostnotablecharacteristicisthatofaLPFthatresonateswithoutlosingmuchlowfrequencycontent.Itwaslargelyresponsibleforwhatdefined‘the’analoguesynthesizersound.
SuchwasitssuccessthatARPcopiedtheladderfilterdesignintheirOdysseysynthesizer.Moogissuedalawsuitforthisinfringement,withthetwocompaniessettlingoutofcourt.34TheRolandSH2000filteralsoshowsaninfringementoftheMoogpatent.
DetailsofthefilterareshowninUSpatent3475623.35Thedesignwasdrivenbytheproblemoffindingasuitablecircuittoproducevoltagecontrolforthefiltercut-offfrequency.DrMoogcameupwithanovelsolution,whichwastousethedynamicresistanceofastringoftransistoremitterjunctionsinconjunctionwithcapacitorstovarythefiltercut-off;
ƒ𝑐 =𝑟𝐶2
Wherer=dynamicresistanceofthebase-emitterjunctionpair.Usingtheexponentialtransferrelationshipoftransistorsprovidedawideoperatingrange(1000:1),sothatthefullaudiospectrumwasundervoltagecontrol.
Transistorsareusedinacomplimentarypairconfigurationtoprovideabalancedcircuitthatpreventsanyofthecontrolvoltageappearingontheoutput.
34http://www.vintagesynth.com/arp/odyssey.php35https://patents.google.com/patent/US3475623
46
Fourstagesareincorporatedintothedesigntogiveaslopeof24dB/octave.Thetransferfunctionis
𝐻 # = 1
(𝑠 + 1)2
Tomakethefilterresonateanamountoffeedback(𝛽)isneeded:
𝐻 # = 1
(𝑠 + 1)2 + 𝛽
Becausethecut-offfrequencyisproportionaltoanexponentialcontrolvoltage,musicalintervalsareproportionaltothecontrolvoltage.Thiswasanothersignificantadvantageofthisdesign.
AdetailedanalysisoftheMoogladderfilterhasbeencompletedbyTimStinchcombe.36
fig.6.TheLadderFilter,asimplementedinthePolyMoog.NotetheuseofaCA3046transistorarray
anddiscretetransistors.Source:MoogMusic.
36http://www.timstinchcombe.co.uk/synth/Moog_ladder_tf.pdf
47
4. TheOperationalTransconductanceAmplifier
TheOperationalTransconductanceAmplifier(OTA)formsanimportantbuildingblockforanaloguesynthesizercircuitry.ThisduetoitsabilitytoactasthecontrolelementforaVCAorVCF.BeforeDSPtechniquesbecamecommon,theOTAhadnumeroususes(suchasmultiplexingandmodulation)butisnowanobsoletecomponent.Consequently,devicessuchastheCA3080andCA3094arebecomingdifficulttosource.
TheOTAoutputscurrentproportionaltoadifferentialinputvoltage:
𝑔5 = 𝜹𝑖89:𝛿𝑣=>
where𝑔5istransconductance.
Gainiscontrolledbyaseparateinputcurrent,thusmakingtheOTAaready-madeVCA.TheCA3080exhibitsgoodlinearityuptoabiascurrentof200µA,witha1:1ratioofoutputcurrenttobiascurrent.RolandusedtheCA3080inseveralofitssynthesizerfilters,beforedevelopingitsintegratedfour-OTAIC(theIR3109).
5. TheFourQuadrantMultiplier
AnotherwayofachievingcontrolofaVCAorVCFistouseafour-quadrantmultiplier.FourQuadrantmeansthatbothinputscanbeabipolarsignal.Apopularchoiceinthe70swastheMC149537IC.Thisdeviceproducesalinearproductoftwoinputvoltages,sothatwhenaDCvoltageisappliedatVC,theamountofVINpresentedatVOUTwillbedirectlyproportionaltothatvoltage.AlogarithmicconverterisrequiredtousethedeviceasaVCA.
37asimilarIC,theMC1496balancedmodulator,wasalsoused
48
fig.7.TheMC1495configuredasalineargaincontrol.Source:Motorola
6. TheMonolithicTransistorArray
These were used to provide thermal stability to oscillator circuits to controltuningdrift. ICexamples include theµA726and theCA3046.TheµA726hasalong-termvoltagedriftofonly5µVperweek.Thisisduetotheinclusionofovencircuitrytomaintainthetransistorsataconstanttemperature.PriortotheseICsbeing available, temperature compensation techniques included usingthermistors or discrete transistors bonded together in an enclosure. Thesemethods were less effective due to mismatches, and the non-linear nature ofthesedevices.ApartfromthePolyMoog,theµA726wasusedinseveralRolandsynthesizers.38
VCOcircuitswilltypicallytaketheinputvoltageandconvertittoanexponentialcurrent.This function is thenatural characteristicof the transistor junction,asdefinedbytheEbersMollequation:
Ic=Is[exp(qVbe/kT)–1] (1)
WhereIsiscollectorleakagecurrent,andislargelydependentontemperature.
38egSH2,SH5,SH7,Jupiter4,System700
49
fig.8.TheµA726equivalentcircuit.Source:FairchildSemiconductor.
byrearranging(1)weget∂Vbe=(kT/q)ln(IcQ2/IcQ1) (2)
This is the usual case of a differential pair used, where the second transistorprovides temperature compensation. By providing a current mirror with areferencecurrentsourcetheeffectsofIscanbecancelled.
ByusingtheµA726inbothofthereferenceoscillatorcircuits,short-termpitchstabilityof0.02%(±1/3Cent)wasobtainedinthePolyMoog.
50
Appendix12
RepairAssessmentChart
PowerSupply
CheckDivider
CLboard
KeyOutput?
CheckMod.Cards
checkM.Osc
MajorityofKeys
function?
N Y
N Y
51
Appendix13
ThePolyPedal
AdiscussionofthePolyMoogsoundandperformancecapabilitiesisincompletewithout mention of the PolyPedal. This rather large unit provides control offeatures that are otherwise lacking. Moog decided which effects and controlswouldfunctiononthepresetsaccordingtothenormsoftheinstrumentsthatarebeingemulated39.
The PolyPedal changes the PolyMoog Keyboard into a versatile performanceinstrument. When the PolyPedal is employed the range of synthesis tones isgreatly improved. This aspect of design reveals the Moog philosophy: thatcreatingintriguingsoundswasonlypartofthedesign;justasimportantwasthatthe synthesizer should be a playable musical instrument with a good userinterface.
Thecontrolsareasfollows:
Swell:thisisavolumepedal,andisparticularlysuitedforRockOrgan
Sustain:thisactivatestheenvelopesustain.Thesustainlengthisdeterminedbythepresetselected.
Leftpedal: this canbe assigned topitch and/or filter sweepby theuseof twofootswitches.ThefiltercutofffrequencyisonlycontrollableontheBrasspresets.Thepitchshiftisapowerfuleffectasitsweepsfrequencypolyphonically.Arangeadjustmentallowstheascendingsweeptobesetanywhereupto1octave.Thisismuchmoreusefulthantheribboncontrollerthathasarangeof±major6th.
Trigger Mode: when depressed this pedal changes the trigger mode betweensingleandmultiple.AswitchattherearofthePolyPedalchangestheoffstatetoeithermode.Theusefulnessofthiscontrolcannotbeoverstated.DaveLucewasabrilliantelectronicsengineer,butitwasBobMoogwhooversawthecontrolleraspects of the design. Providing a piano style pedal for triggermode indicateshowmuchthoughtwasputintothePolyMoogbeinganexpressiveperformanceinstrument.Bothsingleandmultiple triggeringof the filterareuseful,and thispedalgreatlyenhancesthepolyphonicplayability.
ModulationDepthModification
39seeappendix13
52
ScottJuskiw40haspublishedamodificationtothePolyPedalthatreusesthePitchpedal as aModulationDepth control.With the addition of a suitable switch itshouldalsobepossibletoselecteitheroption.
fig.1.ACinchJonespowerconnectorisusedonthePolyPedal.Source:https://www.vintagevibe.com
Appendix14
ParameterSettingforthePresets
Modulationparametersforthe14presetsarelistedinTable6-3ofthePolyMoogservicemanual.
Themodulation amount effects the PWM only on the Strings and Rock Organpresets.FMisactivefromosc1andosc2onlyontheStringspresets.Otherwiseit is fromoneFMosconly,withtheexceptionbeingthatthere isnoFMontheRockOrganpreset.
Themodulationrateisfullypre-setfortheStringspresets,andforPWMfortheClav preset. PWM rate is only variable on the Rock Organ preset. The otherpresetshaveeitherFMosc1orosc2withvariablerate.
TheBeatcontrolisonlyactiveonthesepresets:VoxHumana,Strings,andHonkyTonkPiano.
Attackispre-setforallthepercussivekeys,RockOrgan,andFunkpresets.40http://tellun.com/polymoog/polymoog.html
53
Appendix15
SynthesizerTerminology
Monophonic
Onlyasinglenotecansoundatatime.
Duophonic
Twonotescansoundsimultaneously.
Polyphonic
Morethantwonotescansoundsimultaneously.
Paraphonic
Allnotescansoundsimultaneously,butthroughasharedvoice.
x-Note(eg8note)
Themaximumnumberofnotesthatcansoundsimultaneously.
x-Voice(eg8voice)
Themaximumnumberofgenerator/articulationchannels.Thisissometimesexpressedasvoicespernote.
NotePriority
Thepriorityofnotesplayed.Foramonophonicsynthesizer,thismeanswhatnoteoverridesthepreviousnoteplayed.TheordercanbeLowNotePriority,HighNotePriority,orLastNotePriority(desirable).FirstNotepriorityiswherethefirstkeymustbereleasedbeforeasecondnotewillsound(undesirable).
54
ControlVoltage(CV)
Thevoltagesenttothevoltagecontrolmodules(egVCO,VCF,VCA).
ForcontrollingaVCOtherearetwostandards:
VoltsperOctave(linear)
• eachvolt=1octave,eg2v=A2,3v=A3,4v=A4
• usedbyMoog,Arp,Oberheim,SequentialCircuits,Roland,andmostmodernsynthesizers
HertzperVolt(exponential)
• oneoctave=double(orhalf)thevoltage,eg2v=A2,4v=A3,8v=A4
• usedbyKorg,Yamaha
Gate
AsignalthatissentfromthekeyboardwhenakeyisplayedtostartanEnvelopeGenerator.Thesignalstatechangewillbeheldaslongasthekeyisdepressed.
Trigger
AtransientsignalthatissentfromthekeyboardwhenakeyisplayedtostartanEnvelopeGenerator.
SingleTriggering
TheEnvelopeGeneratorwilltriggeronthefirstnoteplayed,butnotonsubsequentnotes,untilallthekeysarereleased(onaPolyphonicSynthesizer).
MultipleTriggering
TheEnvelopeGeneratorwilltriggeronthefirstnoteplayed,andonsubsequentnotesplayed(onaPolyphonicSynthesizer).
55
Appendix16
PolyphonicSynthesizersofthe1970s–early80s
Model YearIntro.
Polyphony Price
ARPOdyssey 1972 2voice MoogSonicSix 1972 2voice Buchla237 1970s 3voice Buchla238 1970s 4voice E-muModularSystem 1972 2voice MoogApollo 1973 61voice YamahaGX1 1975 2x8voice $60,000Oberheim2Voice 1975 2voice Oberheim4Voice 1975 4voice MoogPolyMoog203A 1975 71voice ARPOmni 1975 49voice KorgPE-1000 1976 60voice YamahaCS-80 1976 8voice $6900KorgPS-3300 1977 48note SCProphetV 1978 5voice $4595RolandJupiter4 1978 4voice $2895OberheimOB-X 1979 4,6and8voice $4595(4voice)MemoryMoog 1982 6voice YamahaDX7 1983 16voice