This month Andy Honeybone supplies the inside story of the potentiometer. What's that? The fast disappearing "knob". The volume control doomed to extinction by digitisation.

It seems like only ten years ago that the more knobs your synthesiser had, the better it was considered. These days you're hard pressed to find three of these controls in addition to the performance wheels. What has brought about this change? The demand for programmable polyphony at a price, that's what.

A potentiometer (pot for short) is a circuit which allows a voltage source to be measured by reference to another source of known value. OK?

A length of resistance wire is fed from the known source and a sliding contact is used to tap off various voltages. By commoning one side of the two voltage sources and connecting an ammeter (to measure current) between the unknown voltage and the slider, the unknown can be deduced from the position of the slider when a zero reading on the meter is given.

A right performance, I'm sure you'll agree. To us musicians, of course, "pots" are rotary controls. Linear pots always seem to be called "sliders". So a potentiometer, strictly speaking, is an application of a variable resistor. Car enthusiasts apparently talk about pots meaning cylinders.

Early monophonic non-programmable synthesisers were constructed from circuits whose functions were modified by variable resistors — the generic name for the "volume control" mechanism. These variable resistors poked through the front of the case. With educated tweaking, the desired sound could be created. Programmability at the time required banks of multi-pole switches to disable the manual inputs and route in preset voltages.

When polyphony first raised its head the plan was to have a complete synthesiser, knobs and all, for each voice. Changing timbre meant adjusting each voice independently. At the time this was considered a pain. Now, the multi-timbral facility is all the rage. Voice programmability in polyphonic systems came about out of necessity — there was no other option. Eight-gang pots were out of the question, so voltage-control was called to the front.

We could be excused for thinking that synthesisers were all voltage-controlled by this time. Not so. Although the oscillators, filter and amplifier were controlled by voltages, the parameters like ADSR, modulation amplitudes, and signal routing were still set by front-panel hardware. Before true programmability could be achieved, voltage-controlled transient generators, additional control VCAs, and digitally-controlled analogue switches has to be brought in. Only the development of specialist integrated circuits for synthesiser functions made such progress cost-and space-effective.

The corollary to this stage of development is the recent appearance of Oberheim's Matrix series synths which, to some extent, turn these hardware necessities into a sales feature by adding highly flexible control routing.

But for the foetal, wholly-programmable voice module, control was through a set of analogue and digital inputs. Each of these could be commoned with those from other voices such that they could be fed to a single controlling signal.

As often happens with technology, generation-skipping occurs. So instead of that controlling signal being a front panel pot or switch it became the output of a digital-to-analogue converter hung on to a small computer system. For a time it was felt that the familiar control knobs must stay — provision had to be made for the computer to scan each one.

An analogue-to-digital converter was then needed to turn the pot position into a number. This in turn was then passed to the digital-to-analogue converter to be transformed back to a voltage, to drive the sound-producing circuitry. Although this method seems unnecessarily complex, it does allow the timbre to be set by means other than the front-panel controls — namely, settings remembered in computer memory.

As budgets tightened, economies were forced and the first things to go were the massed ranks of variable resistors. Not that they were in themselves expensive, but the knobs which adorned their shafts were. Also, the analogue-to-digital converter could be simplified by substituting a slider data-entry, or done away with completely for numeric keypad systems.

In return for affordable poly toys, the musician had to learn to equate numbers with the pot settings learnt at mother's knee. Possibly for reasons of public relations, the numbers went from 0 to 99 (1 to 100 would have been better but would have required another indicator). As anyone with the slightest smattering of binary knows, 99 is not a round number in computer terms. Seven data bits give 128 values. So it seems that resolution was sacrificed for the sake of the apple cart.

Roland, a Japanese maker of electronica, were keen to hedge their bets by making add-on rotary pot programming boxes for their JX-3P and JX-8P synths. Market research people did their stuff and found that the majority of users were content to use the factory preset voices. Machines like Sequential's Max appeared with no on-board facilities for voice editing.

MIDI was the real death knell for the pot. The five-pin DIN communicator meant that synthesisers could be reduced to keyboardless, knobless boxes that were stirred into action from a remote controller.

The mixing desk, however, has traditionally been something of a refuge for the pot. Controls for gain, attenuation, equalisation, panning, group mixing and fading are generally replicated for each channel giving the engineer far too many things to have to juggle during a heavy mixdown.

The mixing desk's pot is slowly losing ground to computer techniques which use one set of "dummy" controls to input information to however many channels are present in the system. Once the mixer circuitry has been freed from manually operated front panel hardware it is a relatively simple matter to arrive at the luxuries of complete setting-up from disc-stored instructions and computer-controlled mixdown.

To continue in this iconoclastic vein, let's consider effects units. As the five-pin DIN sockets continue to appear on the back of delays and such, so the front panel loses knobs. Digital units already have up/down buttons rather than pots to control delay times. Switches can, of course, be multi-function, and as the long, thin, rackmounted gadgets gain popularity, so space becomes even more tight.

Although this is against doctor's orders, I'd like to make a prediction. The disappearance of the pot is a symptom of going digital.

In the not too distant future, manufacturers may realise that there are far too many conversions from analogue to digital and back in a given chain of instrument, effects and recorder/amplifier. The solution would be to standardise on the digital format of audio signals and to pass this information from source to destination, leaving conversion to analogue as the final operation.

Instead of hooking instruments up with jack leads we could well be using optical fibres to handle the ridiculously high transmission rates necessary. The technology of digital signal processing already exists, as Neve have shown with their installation at CTS studios. All we need is a common data standard. I won't hold my breath.