The Electronic Keyboard (Part 3)
Early Keyboard Instruments
Part 3 The Ondes Martenot
In the last two articles, I have skimmed the surface of some pioneer developments in electronic keyboard instruments. But during this period there were other kinds of keyboard, some designed for research into properties of conventional musical instruments, and some for exploiting tone-colours which were then (40 to 50 years ago) quite new and novel. Three of the latter type paved the way for today's solo keyboards, and since they contain interesting features, I will describe them in turn. But firstly, let me say that the substitution of valves by transistors does not in any way improve the music; it merely permits far more circuitry to be packed into far less space. Weight is reduced, handling is easier and there are no high voltages. Many effects are possible with the microchip, but only because of miniaturisation; all could have been accomplished with valves - cf. the RCA synthesiser or the original Wurlitzer side man.
The natural tendency was to start with solo voices; there were adequate instruments for accompaniment, contrasting sounds were required and chords could not easily be produced together with compactness. Opinion was divided between continuous pitch coverage with no intervals and pitches conforming to the intervals between conventional playing keys. Leon Theremin (French patent 612433) and Maurice Martenot subscribed to the former; Hammond (Solovox), Jennings (Univox) and Constant-Martin (Clavioline) preferred the usual keyboards. The splendid articles in Electronics & Music Maker by Derek Pierce mentioned some of these in passing, but a lot of interest has been shown in how it was actually done and, if transistorised, impecunious experimenters might find some useful ideas here.
Maurice Martenot was a young French engineer who had experimented with sounds produced by reacting radio receivers. Not many readers are old enough to remember the use of inductive coupling to tune and produce reaction in early radio sets. (Can any reader be venerable enough to recall my first book, the construction of amateur valve stations, published by Wireless Press in 1921?) Howls and whistles were only too easily set up, and Martenot set to work to tame them. So did Leo Theremin, but novel though his circuit is, it is difficult to play and limited to the sound of a square wave. He never appeared to expand on his idea, and so lost ground.
Martenot used two approaches, both sometimes combined in the one instrument. Figure 1 shows this compound arrangement. There are two generators; the gliding tones are produced by drawing a logarithmic capacitor ribbon over a fixed plate, a cord with a finger ring being attached to the ribbon. The tuning of one side of a beat frequency oscillator is thus altered, resulting in a continuous change in pitch. Like the Theremin, the sound cannot be cut off, so a carbon strip resistor operated by the left hand provided this facility. The ribbon was so shaped that a linear movement of the tape corresponded very closely with the position of the playing keys above, thus one knew what note was being played.
Above this device was a set of playing keys, 4 or 4½ octaves compass. These controlled iron-cored inductances which were mounted on plates fitted with rollers. Every 4th or 5th had part of the core projecting, so that by lightly pressing keys from left to right, the frequency was modulated to form a vibrato. A very expressive idea, rather similar to playing a cello. A light spring returned the cores to correct pitch. The device was monophonic, the inductances being cut out in turn as in the Hammond Solovox. Further, there were simple tone forming circuits, a transposing key, and means for tuning. The instrument was undoubtedly years ahead of its day.
In another Martenot device, sideways movement of the keys operated a small differential condensor, to de-tune a beat frequency oscillator so giving a touch-sensitive vibrato. A further extension of the idea allowed variable downward pressure on the keys to vary the conductivity of a jelly-like compound (as in the Trautonium), thus varying the loudness; not too easy to play, I imagine.
However, Martenot's most remarkable invention was his loudspeaker, based upon a hollow wooden sound chamber, like a large mandolin, strung on both sides with 12 steel wires. These pass over bridges to tuneable anchorages at the top. At the bottom, all are connected to a rigid bar actuated by a special type of loudspeaker movement. The action of this is lateral and at resonance the strings vibrate forcibly and a loud sound is produced. Since harmonics also energise the strings, a continuous and sonorous sound spectrum is produced. The starting and stopping characteristics are like those of the harp, there being virtually no damping making it admirably suited to the rather legato style of playing of the generator. This ingenious French engineer also discovered that by driving the exact centre of a very hard brass cymbal with a moving coil element, a very effective tweeter was produced! For monophonic music, the Martenot devices open up endless possibilities, limited only by the skill of the player.
Next we come to the solo instruments which, for the first time, had the circuits named with known sounds, eg. flute, trumpet, clarinet etc; and means for playing, from the same keys, in different octaves simultaneously. Also included were different depths and speeds of vibrato. The first of these was the Clavioline, again from France; strange that this country, which pioneered so many musical electronics, should have lost its place in keyboard design.
I have just mentioned octaves; this implies frequency division, and dividers (at that time) required a pulse to drive them. Such pulses are easy to come by if one generates a square wave, and as the Eccles-Jordan circuits do this automatically, the problem is solved. The Clavioline (British patent 653340) is not very interesting circuit-wise, except that Constant-Martin and his wife Marie, who designed the tone circuits (British patent 643846) maintained that all the tone-colours he wanted could be obtained from a square wave which, you recall, is devoid of even harmonics. In the absence of the original, and allowing a little artistic licence, this was a reasonable statement.
However, the designers of the Univox (British patent 722430) took the view that the sawtooth was better. Since this contains even as well as all the odd harmonics, it follows that tones synthesised from such a waveform will be smoother and more mellow. Thus there is room for both, as every synthesiser owner knows.
Perhaps the most interesting instrument, though the least well-known, is the Hammond Solovox (British patent 541911; US patent 209920). Unlike the Clavioline and Univox, Hammond used a string of tuned inductances, progressively removed by the playing keys - fewer coils, higher pitch. At the same time, the playing key contacts controlled attack or expression circuits, which also removed the frequency of the oscillator when a playing key was 'up'. Figure 2 shows a portion of the circuit, which provides four octaves of pitches simultaneously; the apparently primitive tone forming circuits give quite creditable imitations and there is a good vibrato. Once again, all the effects are derived from square waves.