I brought the infrared theremin design to the point where I had a pitch-only instrument. It worked well, and I put it aside to attend other matters, intending to design the volume section later. After a year went by, some of my Internet-using friends mentioned that theremin-related information was appearing on the World Wide Web. In particular, an article and schematic authored by Alexander Zeyliger, describing his efforts in making a heterodyne-based theremin, came to my attention. I read it with interest, and decided to construct a similar circuit to see how it worked.

The volume portion of the circuit was incomplete in Zeyliger's design, and I didn't attempt to troubleshoot it. The pitch-generating section worked okay provided that the Colpitts oscillators were critically adjusted. Unfortunately, small temperature variations had enough effect on the oscillators' frequencies to cause the unit to drift out of audible range soon after calibration. This, preventing practical use, prompted me to try various schemes to increase the oscillators' stability, such as changing the ferrite core inductors to air core types, voltage regulation, cascode transistor arrangements, and added isolation between the oscillators and mixer. Although  I was able to make some improvements, the temperature stability was still not entirely satisfactory.

Since my goal was to make a unit which would not need periodic manual calibration, I furthered my efforts to that end. I replaced the local oscillator with a crystal-controlled circuit and incorporated a servo control loop which would null the heterodyne difference frequency to zero when the unit was turned on. The calibration value, which I stored in an analog sample-hold circuit, was then left applied to a frequency offset input on the variable oscillator while the unit was played. This approach would assure proper response after power-up, but temperature-induced drift  in the variable oscillator (and sample-hold droop) would necessitate reinitialization after ten or twenty minutes. Accordingly, this scheme was still unacceptable because many performances would require more than ten minutes of operating time. Instead of Colpitts oscillators, I tried Wien bridge oscillators using low-temperature-coefficient resistors and capacitors, and high-bandwidth operational amplifiers. This, coupled with careful physical construction, yielded good radio frequency oscillators which had potential for solving the stability problem. The Wien approach also eliminated the inductors which were hard to specify and obtain.

It is useful to note that the infrared based theremin had none of the frequency drift problems experienced with the heterodyne circuits, and that I could settle for the infrared approach as a final design. All that I needed to add was an infra-red based volume control, yet to be designed, but certainly not hard to implement. By now, however, I was intrigued with the goal of making a capacitance-type instrument, so I proceeded with a different design that used capacitance detection, but not the heterodyne system. That approach eventually led to alternative circuit techniques.

Arthur Harrison
24 February 1996

©1996 by Arthur Harrison

Back to Index Page