Owner, METASONIX, Redwood City, California USA
Classic analog synthesizers commonly achieved equal tempering
with one of two schemes: either a linear-response
VCO (or VCOs) with a keyboard attached to a voltage
divider network with exponential scaling; or, more
commonly, a VCO with an exponential-to-linear voltage scaler circuit,
and a keyboard with a linear divider network. Both schemes have
their weaknesses. The former, when an LFO or other modulation voltage
was added to the VCO input, did not offset in equal tempering.
These problems do not apply directly to our tube synthesizer. Instead, we have a different problem: our very simple VCO is temperature-stable, but has a response that varies semi-randomly with the individual thyratron tube used. So, we not only have to use a non-linear divider on the keyboard; we have to use one that can be ADJUSTED for the individual thyratron. If we want to use two or more VCOs in parallel, we have to match the thyratrons for pitch-to-voltage response curves.
The Trautonium never had this problem--it used a ribbon controller, with no fixed keys. We could always do that with our tube circuits. Refer to the classic ELECTRONOTES newsletter for a number of suitable circuits. However, if you insist on having an equally-tempered keyboard, we must use a divider made of trimpots.
Figure 1 shows how this is done. Normally, the output voltage of the divider drifts up to the supply voltage (you can use +12 or +15), because none of the keys are down and the voltage is not divided. When a key is pressed, the voltage goes to a lower level. By appropriately adjusting the trimpots, the output voltage can be set to driven a tube VCO to follow equal tempering. (It should be obvious that you adjust the pots starting at the LEFTMOST key, i.e., the lowest-pitch key.)
We follow the buffer IC1a (which can be any inexpensive single-supply quad op-amp) with a sample-and-hold, Q1. IC1b is used as a comparator to detect the drop of the divider voltage from the +V, and raises its output. This is the GATE signal. It triggers the S/H, and may be used to operate any conventional ADSR or otherenvelope generator to feed the VCA. Q1 can be any inexpensive n-channel junction FET, such as the commonly-used 2N3819 or 2N5486. (Obviously, +v should come from a well-regulated supply. Less than 20 mA are needed by this design.) The PITCH BEND/ADJ pot shown on the output buffer IC1D is optional--novices would be smart to just ground that point from the 10k resistor, and leave this pot out. It complicates tune-up. Note that it will not give a full pitch-control range on the lowest few keys. If you desire a true up-or-down pitch bend, this would be more easily done with the master-tune pot in the anode of the thyratron.
Our controller circuit will give about 2 1/2 octaves of control range on the tube VCO, from a few hundred millivolts above ground to nearly +V. This should be enough for most experimental music applications. (Getting a greater range out of the thyratron VCO is an advanced subject, which we will not get into at this time.) The AGO keyboard sold by PAIA Electronics can be used with this circuit. It has 37 keys. We can only use about 28, so remove the unused keys after the keyboard is installed in your intended cabinet, and cover the open areas with aluminum or plastic plates. I have built a few such circuits, and can assure you that they work very well.
Alternatively, if you already have an old Moog or ARP (or other) analog controller keyboard and want to convert it to drive the tube VCOs, just substitute the trimpot divider for the existing voltage divider. This may ruin any collectibility value of the keyboard, but one gains a small side benefit; the keyboard can be used to drive ANY VCO, tube or transistor, linear, exponential, what-have-you. And it can produce any desired scale; equal-tempered, just, microtonal. Simply drive the VCO desired and adjust each trimpot (while holding down its key) for the desired pitch.
Finally, MIDI control of your tube circuits. You can NOT use a conventional MIDI-CV converter, due to to difficulty of converting the (usually linear) CV output to the voltage the tube VCO would like to see. So, just replace the keyboard in Figure 1 with a JW Electronic MTP-1 MIDI-to-relay interface board. Its relay outputs are open-collector, and perform the same function as the keyboard by shorting the required divider point to ground. JW is at: email@example.com, web site http://www.warwick.ac.uk/~essfc/mtp1.html. I have used an MTP-1 for this successfully. Its major advantages are that it is less expensive and easier to use than other MIDI-parallel interfaces now available. (And yes, there is nothing to keep you from wiring the MTP-1 AND the keyboard to a single controller. All they do is ground a point on the divider. So, as long as one is not being used, the other can control the synth.)
About the Author
Eric Barbour holds a BSEE degree from Northern Arizona University. He has been a regular contributor with GLASS AUDIO magazine since 1991, staff editor of VACUUM TUBE VALLEY magazine since its founding in 1995, and has written articles for many other music and audiophile publications.
A note to readers: this circuitry is intended for the more advanced builder. Because high voltages are used, a shock hazard exists. We do NOT recommend that the novice DIY musician try to construct this synthesizer. Some experience with tube electronics is highly recommended.