Moog Polymoog analog polyphonic synthesizer


Last Update 04-19-2010

The Polymoog was a pioneering design when it was released in 1975 - and by 1977 was an orphan because of its stellar reputation for unreliability and because progress had surpassed it. The other well-deserved common complaint was that the Polymoog didn't sound like a Minimoog. In the early 1970s, polyphony was a tough nut to crack. The prevailing technology before today's accepted voice assignment polyphonics were perfected was the top octave divider systems (TOS) which dates back to organs in the 1950s. String machines such as the ARP String Ensemble or Crumar String Performer were at heart a TOS system.  Bob Moog designed a polyphonic rank module for Wendy (ne' Walter) Carlos in 1969 that was a TOS system. The Polymoog represents the peak of TOS system design (although it is a Dave Luce design, Bob had little to do with it). TOS gives you full polyphony - every key has its own note. Unfortunately what you can do with this note is a very limiting factor in the Polymoog.

Dave Luce (pictured here on the right with an early production unit) designed the Polymoog, which began prototype life as the Apollo which was featured on ELP's "Brain Salad Surgery". Bob Moog had little if any design input as his duties under Norlin removed him from synthesizer development. Luce was something of a perfectionist which was evident when the Polymoog needed a staggering 200 engineering changes before it was ready for production. Internal memos show a company trying to pin Luce down and freeze the design, to force it to closure and get it in production. I have seen a drawing of an alternate panel design, no less confusing than the one we know today. Luce also was fond of complicated convoluted designs where a simple solution would have done the same job. If you don't believe me, I dare you to make sense of the Polymoog block diagram. Any EE who has studied both volumes of the service manual written by Luce will hardly dispute this.

There were six Polymoog prototypes built with a 72 note keyboard with a C to B range. Chick Corea and Larry Fast each had one of these prototypes. The production Polymoog changed the range to F to E, then feedback from musicians competing with guitarists changed it to a E to D keyboard and reduced it to a 71 note system. Changing to an E/D range required a tooling change that was expensive. It had the longest Pratt-Read production keyboard and each keyshell was filled underneath with epoxy and leadshot to give it a weighted feel. While it doesn't feel anything like a piano, it's actually pleasant to play. Pratt-Read never made an 88 note keybed to my knowledge, their target industry was organs not pianos.

Moog developed a promotional movie about the Polymoog. It features interviews and performances with Dave Luce, Michael Boddicker, Chick Corea, and Herbie Hancock as well as a goofy stop-action animation involving a Polymoog in a wooded knoll with two mimes.  Roger Luther of Moog Archives offers this movie converted to disc.

When the Polymoog was released, the buzz was afoot and everybody wanted one.  It had a touch sensitive keyboard, was fully polyphonic, and you could change sounds without re-configuring up a panel full of knobs and switches.  Remember, this was 1975 and nobody else had a synthesizer that offered these features.  It was the first synthesizer with a piano sound that was velocity sensitive, but as you read below the velocity sensing system had a fault.

There is confusion regarding the model names of Polymoogs.  When released in 1975, the eight preset version with a panel full of slidepots (model 203A) was badged the Polymoog Keyboard.  When Moog released the economical fourteen preset version (model 280A) it became known as Polymoog Keyboard while the eight preset fully variable version was re-badged Polymoog Synthesizer.

Presets named "strings", "piano", "organ", "harpsichord", "funk", "clav", "vibes", and "brass" comprise the 203a while the 280a offers "vox humana" (made famous by Gary Numan in his hit single "Cars"), "string 1", "string 2", "electric piano", "piano", "honky tonk piano", "clav", "harpsichord", "brass solo", "brass chorus", "pipe organ", "rock organ", "vibes", and "funk". I never liked the "Funk" preset on my 203a and I replaced the wimpy thin "piano" and "organ" with sounds of my own, which necessitated calculating preset resistors and soldering them in.  "Brass" isn't very authentic or fat either, especially with the limitation of a master filter.

Polymoogs had velocity sensitive keys, which was big BIG news in 1975.  However the velocity sensitivity had a serious flaw.  If you held down the sustain pedal and repeatedly pressed the same key at light velocity, each repeated key press would result in louder volume.  The velocity voltage had an accumulative effect when you repeatedly pressed a key with the sustain pedal held down.  This isn't very practical for percussive sounds like piano.  In practice I rarely use the velocity feature.

The description below applies to the 203A Polymoog (I never owned the 280A)

Despite the "fully variable" tag, you cannot store custom patches on a Polymoog.  There is no microprocessor under the hood.  When you push a preset button, CMOS switching merely selects fixed resistors which configured the sound. If you want to tweak the sound, changing the slidepot doesn't do anything until you press a VAR button that makes the slidepots live.  The VAR button merely replaces the fixed resistors with the resistance element in the front panel slidepot.  Pressing the accomodating PRE button reverts back to the fixed resistors and restores the preset configuration.  If you ever played with a Prophet 5 Rev 1 or Rev 2 polyphonic synthesizer that has PRESET buttons, this convention originated in the Polymoog.

As primitive as it would appear, this design is not ancient.  The ARP ProSoloist of 1972 used a similar fixed resistor design although it did not offer user tweaking.

In normal operation, pressing a preset button toggled the VAR buttons to PRE so that the preset configuration was called up.  There is a button on the selector panel marked with a dot labeled PART/FULL.  When you pressed this button, a period would appear in the LED.  If you manipulated the sound using VAR buttons and changed to a different preset, the VAR buttons would retain their state.  This is useful if you have an envelope setting or filter setting you wanted share across one or more presets. One other trick I discovered - if I press and hold the PIANO preset then press BRASS, the BRASS sound would use the PIANO preset mode filter (see below), which can offer some variety to the sound.

The best approach to tweaking sounds is to use a preset as a starting point, then tweak the sound to your liking.

The technology for storing patches had not yet progressed to the microprocessor patch storage system of today because microprocessors, RAM memory, and ADC/DACs were very expensive back then.  Oberheim devised their patch storage system using RAMs and ADC/DACs by 1975 in the SEM polyphonics, and microprocessors weren't implemented until 1977 in the Prophet-5.

The early Polymoogs were temperamental beasts and were extremely unreliable.  Thanks to Luce's penchant for fancy circuits, the same principle of mechanical systems applied - the more moving parts, the more likely it will fail. The Polymoog was chuck full of fancy circuits which had a habit of malfunctioning.  Moog had a lot of Polymoog returns for warranty repairs, and that cuts into their bottom line. Today the mere mention of "Polymoog" strikes fear and dread in musicians who had to endure them breaking all the time. The same could be said for repair shops, and few techs will touch a Polymoog today. It is the rare Polymoog today that is 100% functioning.

Polymoogs of serial number 3900 and higher did implement a set of reliability improvements at the factory.  My unit is 4XXX and I can confirm that it has survived years of club gigging with few problems. My unit is dated October 1979 under the hood and is badged "Polymoog Synthesizer", so if you want a later reliable unit then avoid the ones that are badged "Polymoog Keyboard".  Judging by the history of ebay auctions, these later units seldom appeared.

How does a TOS system work? If you divide a periodic waveform by two, the result is an octave lower. A TOS system is a "rank" consisting of a master high frequency oscillator that is divided into the 12 semitones of the highest octave of desire, then the descending octaves are achieved by dividing by two each of those 12 semitones in succession. The complete rank is a "voice" in organ terminology, as opposed to the accepted convention of the synthesizer single note "voice". Every single note in a rank is firing continuously. However the waveform of each divider output is a pure square wave. The Polymoog implements waveshapers per key (to be discussed below) to create ramp and variable pulse waveforms.

The Polymoog is a two rank TOS system - one rank of ramp waveforms and one rank of variable pulse waveforms. 

Because the master oscillators in each Polymoog rank are independent, they can be detuned.  And therein lies a major limitation of synthesis on the Polymoog - only the master oscillator can be modulated, not the individual notes.  You cannot pitch modulate individual notes in a TOS system.

The other major limitation in the Polymoog TOS system is filtering.  The VCF is a single master filter for all the notes.  If you apply a spiked EG to the filter and then stagger the notes in a chord, the VCF will spike for note 0 then note 0+1 then note 0+1+2 then... pretty useless for brass voicings.  While keyboard tracking can modulate the VCF, it isn't very practical in a master VCF system because when you hold a chord then release succeeding high notes, the keyboard tracking for the high notes isn't reset to low notes unless you retrigger the low note.

The ad copy does boast of each key having its own filter - while this is correct, it is not practical. The Polymoog has a custom made "Polycom" IC for each key. This Polycom IC has two waveshapers (converts the square waves from the dividers into ramp and variable pulse waveforms), two VCAs, a mixer, an envelope generator, and a two pole ladder filter. There, each key does have a filter right? But the front panel provides no control over the filter, it is configured by preset resistors under the hood.  The Polycom filter has no resonance capability at all, and there is no facility for modulation - not by an EG, an LFO, nothing. In fact the schematics refer to it as the "brightness" filter, and that is the best description for it.

The lack of filter resonance and modulation in the Polycom doomed the Polymoog in the sound palette department. Why didn't Moog implement it? Remember that back in the early 1970s, customized ICs was unheard of and expensive. A lot of R&D went into the Polycom, and compromises had to be made for the technology of the day. They had to keep the pin count on the IC package down, and filter resonance and modulation increased the pin count to the point the cost was unrealistic. Considering that the Polymoog originally listed for $5295 in 1975 dollars, this is not hard to fathom.

The reason for the limited palette of the Polymoog's TOS design is because it cannot implement voice modulation - if you wanted to modulate the VCF or the PW using the oscillator of an independent voice, you could not do it. Want to modulate the pitch of a single note in a TOS? Can't be done. The effects that can be done on the Polymod section of Prophet-5s or voice modulation section of Memorymoogs are not possible on the Polymoog.

Contrast to the accepted synthesizer architecture of VCO->VCF->VCA per voice, the Polymoog architecture is VCO->brightness filter->VCA per key.  Then the Polymoog audio can be processed by preset filters (called "mode" filters), by the master VCF, or by a "Resonator" (the model 280A Polymoog Keyboard omits the Resonator). The filtering is done in parallel and there are separate outputs for each filter system on the rear panel (but not the 280A). There is also a DIRECT output that bypasses the filter so you can hear the "brightness filter" in action (and witness how thin and wimpy it sounds). You can also use the direct output for unfiltered processing.

Two VCOs are offered. VCO#1 is ramp only and can be switched to 8' and 4' ranges while VCO#2 is variable pulse only and can be switched to 16' and 8' ranges. The Ramp waveform can be detuned and its level is variable, but not the pulse waveforms. There are three master LFOs available - two for frequency modulation (one for each ramp/pulse rank) and one for pulse width modulation.  This is one of the few saving graces of the Polymoog as it works to create nice string or vocal pads.

There is a LOCK button that syncs the ranks together.  But it is not hard sync, it is phase lock sync.  Phase lock sync locks the VCO pitches exactly.  If you attempt to use the detune control, both ranks will change in pitch and stay in sync.  Hard sync is much more interesting in that detuning the synced VCO creates new timbres.  Due to the TOS architecture, hard sync is not possible (the dividers can only create pure square waves).  In LOCK mode, the frequency modulation is converted to phase modulation.  But the phase shifting is a wimpy sounding single stage phase shift and has little effect on the sound.  You can hear the phase shift in the HARPSI preset.

The "Mode" filters are preset filters which are standard electronic opamp textbook filters, some are simple 12dB low pass and others are a mix of 12/24dB LP/BP/HP.  These are NOT the famed Moog transistor ladder filters and they are not editable.

The Resonator is a three-band multimode filter system - it is NOT the Moog parametric EQ, totally different circuit.  It is not the famed Moog transistor ladder filter either.  Each band covers low/mid/high ranges and has level, frequency, and resonance controls.  There's a switch that changes the filters from lowpass to bandpass to highpass, on all three bands at once.  While this sounds interesting, Luce blew it because the filters are 6dB/oct.  The 6dB/oct Resonator is a standard electronic textbook state-variable filter, and it isn't as useful or colorful as the 12dB/oct Oberheim state-variable filter.  The 6dB/oct passes too many harmonics, resulting in a buzzy sound even at maximum resonance.  It isn't very useful for formant sounds because there aren't any overlapping frequencies between the bands.  The Resonator can't be modulated either.

The master VCF *is* the Moog transistor ladder filter but it sounds pretty lame, in fact of all the Moogs I own this one is the worst sounding Moog filter of the lot.  It also won't self-oscillate.  Why?  It's the last device in the audio chain, there's no post VCA to kill the oscillating VCF when you let up on the key.  Remember, the Polymoog is VCO->brightness filter->VCA *per key*, THEN to master VCF.  That's another reason why the Polymoog has no noise source.

Thanks to its generous rear panel, you could process external audio through the mode filter, the resonator, or the master VCF.  The 280a had a less generous rear panel.

There is a panel section for the loudness contour generator, which controls the envelope generators in the Polycom ICs.  It is the ADS/shared-D/R system similar to the Minimoog (this is "progress"?!?).  You have separate DECAY slidepots for upper and lower splits, which is really limiting when you read the paragraph regarding bi-timbality.  The EGs have a flaw - if the ATTACK time is greater than the DECAY time, you get next to no envelope and near silence.  The LOCK button engages the release stage of the EGs, same as holding down the sustain pedal.  You don't want this engaged all the time because the unit will slowly build up a high frequency hiss in the audio.

Some ad copy claim that the Polymoog is bi-timbral.  Very limited bi-timbrality, actually.  Remember - master VCF limits the bi-timbrality.  You can split the upper and lower section (split point is fixed at two octaves above the lowest key).  Each split has control over waveshape selection, frequency modulation, pulse width, PWM, ramp level, decay time, and which section is processed by which filter system.  If you wanted a spikey bass in the lower split with a mushy pad in the upper - can't be done because you don't have separate ATTACK controls for each split.  Also the keyboard has three sections whose audio can be balanced with three slidepots but again not very practical.  Occasionally I use these balance slidepots to get some overdrive in the audio.

Nevertheless both versions are capable of making some unique and exceedingly powerful sounds.  While they should not be dismissed, you should also be prepared for the price of admission as any used unit will need TLC and $$$ to get it in top shape.

A common complaint (rightly justified) was that the Polymoog didn't sound like a Minimoog. Luce had a different approach to sound than Bob Moog.  Bob exploited distortion while Luce adhorred it.  Luce put more focus on circuit stability and control than he did on sound, IE great pains were taken to ensure minimal distortion in the signal path.  The reason the Minimoog sounds so good is the subtle distortion from overdriving the filter and VCA, and because Luce neglected that the sound of the Polymoog is too clean.  You can crank the levels with the BALANCE sliders next to the mixer section but the distortion isn't at all pleasant, it's not the same effect as overdriving the Minimoog filter.

An optional foot pedal controller called the Polypedal (model 285a) was available. This was a substantial unit with two sweep pedals, two foot pedals, three footswitches, and two thumbwheels on the rear.  The unit connected to a Polymoog with a multicable snake that fanned out into plugs.  I had replaced these fanned cables with a multipin connector and replaced the TRIGGER IN jack with the mating connect.  TRIGGER IN on the Polymoog was rather impractical and not used in my applications.  One sweep pedal controlled volume, the other could control pitch or filter cutoff.  One of the thumbwheels will adjust the amount of pitch bend.  One of the footpedals controlled sustain while the other could control trigger mode (single or multiple triggering).  Frankly I found the latter not very useful.  The footswitches selected sweep pedal destination to filter or pitch, and a third EXT SYN enabled triggering of external syntesizers.  You could use the Polypedal on other Moog synths like the Minimoog, Micromoog, Multimoog, anything that had the six pin cinch-jones "accessories" power connector, although the Minimoog will need to be modified to provide 5DVC for the Polypedal.

What's this "external synthesizer" bit?  You mean the Polymoog keyboard can be used as a CV controller for my modular?  Yes, *BUT* not across the full 71 keys.  The Polymoog keyboard is split at two octaves above the lowest key, and this split point can't be changed.  You can use the keyboard as a CV controller with either the lower 24 keys, or the upper 47 keys.  A shame you can't use the entire 71 keybed at once.  However glide and pitch ribbon does work on the external CV output.

Yes folks, the Polymoog is in effect a fancy organ.  It's a dual TOS system with PWM and VCA per key, and a few master filters thrown in.  Despite its promising technology, it quickly became outclassed and even derided in many circles.

Moog had their share of teething problems with the Polymoog and it became an orphan with the competing synths.  It also gained a reputation of unreliability (it is the rare Polymoog that is 100% functional today and few repair shops will touch them).  Then the assignable voice polyphonics that followed - Oberheim FVS Four Voice SEM in 1975, Yamaha CS-80/60/50 in 1976, Prophet-5 in 1977, Oberheim OB-X in 1979 - showed the world the superior implementation of polyphonic synthesis.  Yes there was the Yamaha GX-1 in 1975, but these were prohibitably expensive and less then 70 of them were made.  Moog finally discontinued the Polymoog in 1980 and released the Memorymoog in 1982 - a pretty late entry in an already crowded market!  Alas, before the Memorymoog could establish a foothold, the Yamaha DX-7 effectively killed the analog synthesizer market in one sweep.

The Memorymoog was a worthy successor to the Polymoog in that the assignable voice system is the superior polyphonic design.  There are a lot of tricks in  that are not possible in a TOS sytem like the Polymoog, chief among them voice modulation.  That is why the Memorymoog has a broader palette of sounds over the Polymoog.

The Polymoog was my first Moog synthesizer and my first endeavor in polyphonic synthesizers, the velocity and full polyphony appealed to me as a piano player.  But it didn't take long to learn of its limitations.  Strings, pads, organs, and harpsichord are the Polymoog's specialty - and that's about it.  I bought my Polymoog synthesizer with Polypedal, legs, and case in 1985.  Later as I acquired my other moogs (Source, Memorymoog, Minimoog, Taurus, Voyager, etc) the Polymoog got used less and less.  Being my first moog synth, I'm too attached to it to sell it and it's been with me for many gigs.  It's still in really good cosmetic shape and it looks great in the studio.  At one point it served my studio as a coffeetable (good looking one too).   Today I am exploring new sounds with external processors, notably my Korg SDD-3300.  I can get some really nice vocal pads with this combination.  The Polymoog is undeniably an important instrument in synthesizer history, but it's not terribly variable and most are not very reliable.

If you're looking for your first analog or your first polysynth, don't pick this one.

contact info