What's the most desirable keyboard instrument in the world? A Bosendorfer or Steinway grand? Maybe a Synclavier? It seems that age and rareity value have combined to make the Moog Series III the ultimate acquisition for the devotee of vintage synthesizers. But don't bother scouring the Readers' ads...

Had you been a member of the audience at New York's Museum of Modern Art on August 28th, 1969, you would have witnessed a unique musical performance by two men - Herb Deutsch and Chris Swanson. Depending on your perspective, you would have considered the performance a product of either hi-tech electronic gadgetry or a bona fide musical instrument. Either way, what you would have been listening to was the world's first live performance of an electronic sound synthesiser - a Moog synthesiser...

Robert Moog had been working through most of the sixties on making sound synthesis more manageable by improving the giant RCA MkII punched tape machine. By the middle of the decade he had published a paper on the secret to his eventual success - voltage control - and had demonstrated a voltage controlled oscillator and other modules to the AES convention in America.

It didn't take long for his instrument, the Moog Synthesiser, to be seized upon by musicians and not long after that before it spawned a massive hit in the unlikely form of Walter Carlos' Switched-on Bach. This became the biggest-selling classical album of all time with sales of 50,000 in the first six weeks, rising to virtually a million. Suddenly everyone knew what synthesisers sounded like and when Keith Emerson used a Moog live with the Royal Philharmonic playing the Five Bridges Suite - everyone knew what they looked like, too. Moog became a generic term - like Hoover or Tannoy - and new models of greater sophistication and diminishing size were developed.

The most impressive-looking (and sounding) of all the Moog modular systems was the IIIC - three consoles each in a handsome walnut casing stacked on top of each other. If you wanted to gig with a Series III, you could opt for the IIIp version on which the walnut was replaced with leatherette-covered ply. But Keith Emerson seemed to manage OK with his standard IIIC - although this had all three consoles joined together in a single large walnut case.

The largest of the three consoles was situated at the bottom of the stack (raked at an angle of about seventy-five degrees), and the other two were positioned vertically above it, on rubber feet. The lower-most console contained the oscillators, mixers and noise generators; the centre unit held the filters, envelope generators, multiples (see later) and amplifiers; and in the top cabinet were the sequencers and interfaces.

At the heart of the Series III were the VCOs - switchable from 32' to 2', with sine, triangle, sawtooth and pulse waves all available simultaneously. As you might imagine, these produced a very fat sound - though this was as much to do with their tendency to drift as anything else! On the MiniMoog there were three VCOs; on the IIIC there were nine - and a tenth which was usually used as an LFO, but could just as easily be pressed into service as an ordinary VCO if things were sounding a bit thin.

Though adjustable manually, there was no input for pulse-width modulation, but really, this development only came about through the need to fatten up weak or static sounds. With the huge complement of oscillators on the Series III there was little likelihood of that. Later, a more stable oscillator (module number 921) was developed for the 35 and 55 models, and was actually made available to Series III owners as an update. But despite having sync and pulse width modulation, and being less liable to drift, these are not considered as desirable nowadays as the 901B's. As former Moog developer (and owner of two of the very few IIIC's in England) Zak Matelon puts it, "In the hands of a sensitive operator, the earlier VCOs are superior". In other words you have to keep your ears open and your wits about you - and be able to tune accurately. Looking to technology for a solution, Keith Emerson opted for an early forerunner of a digital readout tuner now commonly used by guitarists.

Below each bank of oscillators was a four-input mixer used to combine several signals into a single output, switchable to provide positive or negative voltage. On the same panel there were four jacks which formed a 'multiple' - a patching system which routed input signals to jacks at the rear of the synthesiser - and four switching panels which could duplicate the job of patching control voltages into the oscillators. There was also a white or pink noise generator, a 4 x 4 matrix mixer with bass and treble cut on each of the four outputs, and a patchbay which could link keyboard outputs to any other part of the synthesiser.

Incidentally, each module carried the quietly understated 'RA Moog' trademark along with a symbol that looked like a stylised version of the top of a semi-quaver. Later Moogs had an altogether more 'brash' logo with this symbol incorporated into the first 'o' of Moog, and the initials 'RA' discarded (presumably in the interests of modernity and mass appeal). I know which I prefer.

In the middle of the three consoles came the modules that shape, filter and modify the VCO and noise outputs. First off were a bank of fixed filters: every electronic music studio in the '60s used filters of this kind, and the fact that Moog had pioneered the use of voltage control didn't stop him including a fixed filter bank on the Series III. Fourteen separate controls were used to set up the low-pass, high-pass, and twelve band-pass filters.

Although the latter were only 12dB per octave, the fact that they were band-pass meant they could produce resonant peaks and troughs suitable for the recreation of double reed sounds, for example. With the widespread availability of cheap (and occasionally musical) graphic EQs, it is perhaps difficult to imagine a time when a simple set-up like this was considered state of the art. But the fact remains that even today, only a system like the Neve EQ module could better it for warmth and sound quality.

Perhaps even more crucial to the sound of the Moog were the famous voltage controlled filters. Unusually, the Series III included a high-pass version of the design, which could, for instance, eliminate the fundamental from a timbre, making it instantly thinner, more cutting and less inclined to clutter up a mix. Where this differed from using the EQ (say on a mixing desk or outboard processor) to take out bass or lower mid frequencies, was that through the use of voltage control, every note up or down the keyboard could be given the same degree of timbral alteration.

When the action of the low pass filter was taken into consideration, the possibilities were further enhanced: used in series or parallel, or forming a band-pass/band-reject system (by using the Filter Coupler), the two major voltage-controlled filters could produce almost limitless variations on a basic sound.

For many people, the low-pass filter represented the heart of the Moog synthesiser. Even on the subsequent (and comparatively insubstantial) Minimoogs, the VCF was reckoned to be the key to its great sound. Indeed, early Minimoogs are judged by aficionados to sound better than those with serial numbers over 6000 - in other words, after the take-over by Norlin. One explanation of the difference in sound may be that the later models used 3080 op-amps, as opposed to the discrete transistors of earlier models - and of the Series III modular Moogs, of course.

You might be forgiven for thinking that a VCF is a VCF, but in the case of early Moog filters, common-sense just doesn't apply. They were, and possibly always will be, the best. Their 24dB per octave cut and boost gave them a power which, right up until the patent ran out, always ranked them well above the competition. And even after 24dB became the norm, there was still an indefinable something - perhaps the signal path to and from them, maybe the pleasing distortion they introduced - which made then so distinctive.

It may be a little hard to credit that the Series III with a retail price of approximately $8995 in its day (...something like five times that in today's money) should have had a rather crude spring reverb included in its complement of modules. On the other hand, it is, perhaps, an indication of the machine's quality that the basic sounds were so good even without complex digital signal processing. That said, I love using digital effects with old analogue synths (and equally, warming up modern sounds with a bit of Space Echo or AC30). With vintage Moogs, even a fairly thin and ordinary reverb like that on an SPX90 seems to work them really well. But I digress...

Next in the middle console come the three VCAs - voltage controlled amplifiers - which, interestingly, included linear as well as exponential responses. What this meant in practice was that you had more control over how the envelopes affected the overall sound - exponential was more useful for struck/percussive sounds, and linear for more 'bowed' type sounds.

"Remember that in those days, sequencers were used as much for changes in tone colour as for playing tunes or bass lines. What might appear to be an astonishingly primitive sequencer was actually a useful, and even quite versatile piece of equipment"

Though VCAs and VCFs are still common in analogue synths, the other principle Moog module certainly isn't. The Dual Trigger Delay's job was to delay the trigger to the envelope generators by variable amounts. This being the Series IIIC, you were provided with considerable flexibility in how you choose to use the delays. You could have them both independently triggered, or direct a trigger voltage to the top unit, starting the timing circuit of both simultaneously. Alternatively, the second envelope could be triggered only after the first has run its course.

Complicated stuff, but useful if you needed to maintain full control over complex envelopes. In the 1980s, many synthesiser manufacturers seemed to reach the conclusion that ADSRs have inherent limitations, and as a result added extra stages to their envelopes (Korg's Poly 800 was a good example). Very few of these, however, even approached the complexity of the Series III system.

The final module in the centre cabinet was the Envelope Follower - an early example of a manufacturer trying hard to interface 'real' instruments with synthesisers. Here, various functions could be controlled by the voltages provided by an external instrument. For instance, you could produce automatic wah, or use the external instrument to trigger the envelope and raise or lower the VCA controls - useful as a noise gate or as a more subtle volume control.

Being modular meant that there could always be variations in individual Series IIIs, but because it contained the least essential of the modules, the top panel often tended to reflect the individual requirements of the buyer. Though in the usual run of things you could safely assume the cabinet would house the sequencer and interface equipment, this was by no means always the case.

The Moog on which this retrospective was based included an interface, two sequential switch modules, and a double complement of sequencers - though the standard format comprised only one sequencer, with a blanking panel. The sequencers were driven by an analogue voltage controlled clock oscillator and provided a maximum of twenty-four steps in three rows of eight - your position in the sequence being marked by a series of indicator lights.

But hang on a minute - how did you set up a sequence in the first place? The answer lay in the twenty-four pots, which could be adjusted to provide a different voltage for each step. Further, for each row you could control the voltage range - doubling or halving the value from a nominal centre figure of four volts. Using this facility, it was possible to vary the pitches of different rows of the sequencer for different octaves, or produce different degrees of VCF or VCA change for each row. And of course with two parallel outputs for each row, you could (for example), control VCAs and VCOs simultaneously.

You could also use the output jack under each column to trigger anything you wanted. This was in addition to the jacks for trigger inputs beneath each column (which could trigger voltages independently of the clock) and (one of my favourite bits on any modular synth), the manual trigger buttons.

Remember that in those days, sequencers were used as much for changes in tone colour as for playing tunes or bass lines. And what might at first appear to be an astonishingly primitive sequencer was actually a useful, and even quite versatile piece of equipment. You could skip or stop with dedicated switches; you could treat a sequence as three rows of eight, one row of twenty-four, or eight rows of three - or you could reduce it to sixteen steps, using the third row to determine the length of the steps. Being in the analogue voltage control domain, you could also use an external clock voltage either exclusively or (...and this is the interesting bit) in combination with the Moog's internal clock.

As I said, this was sequencing quite unlike anything heard today. People produced things on the Series III sequencer which I've never heard anyone achieve on Cubase or Notator. It's not that computer software isn't up to the job - it's just that no-one seems to investigate the possibilities of working with a strictly limited palette - not when there are virtually limitless possibilities available to them at the click of a mouse button.

In addition to the sequencer, Series III Moogs all had an interface module in the top cabinet combining various circuits - all available individually or in combination. One of them was for audio-in to voltage-out, another linked S-trigger in to V-trigger out, and a further two converted V-triggers to S-triggers. It was a pretty utilitarian module; its function was simply to enable all the other modules to talk to each other. The conversion of V-triggers to S-triggers, for instance, was often used to change a voltage trigger from the sequencer into an S-trigger needed to trigger an envelope.

The last of the modules on the Series III was the Sequential Switch system. As its name suggests, this was used to select the various sequencer parameters outlined earlier: whether you used two or three rows of eight voltages, for example, or how you chose to program the steps. It was also used to step sequentially through different VCOs when you were playing the keyboard - a facility that Oberheim were later to assimilate into their machines. You may also have come across it on something more recent like the Yamaha TX81Z.

The whole instrument could be controlled by the voltages from two performance sources - a ribbon controller (a beautifully engineered piece of gear ideal for portamento or smooth changes of volume or timbre) and a keyboard. The keyboard was monophonic, of course, but it did have a generous (for the time) five-octave range, and variable portamento. You could also alter the scale and range of the keyboard, for microtonal work.

With talk of the sequencer and keyboard, this tour of the IIIC may seem to have finished on something of a low note for devotees of the Moog sound. Undoubtedly, the real power of the instrument lay in the combination of VCOs and VCFs: there is still nothing to touch it in the eyes of many people. If you're on the look out for a series III... good luck. This is one very scarce machine. Only about three hundred Moog modulars systems were ever made and only a handful of these ended up in the UK. They are, as a result, one of the most valuable late twentieth-century instruments. For any IIIC, you're talking about a price tag of five figures and for one in pristine condition, you can expect to pay in the region of £20,000.

The machine supplied for this article was bought by ace synth collector Bob Williams, and lovingly restored to its original condition by Stephen Dyer of Cornwall, who has made something of a speciality of restoring and renovating Moog products for the growing band of enthusiasts for whom only the best is enough. Bob owned it for less than a year before he had an offer (he couldn't refuse) from an intermediary for a 'well known name', but this was time enough for him to reach the conclusion that no synthesiser yet made has rivalled the awesome power of this, the king of modular synths.