> Is it something like two frequencies together produce frequencies which > are both the sum and the difference between the two inputs? And they go > through a low pass filter to get the final tone? Like tuning a guitar the > closer you get to tuning two strings the faster the beat gets? Or is that > how it might be done now? For getting the pitch itself, you use a pair of oscillators, like 700kHz-ish. They sit maybe 200 Hz apart at idle more or less. So the mix products would be 700200, and 200 hz. What I was talking about there though, is what separates a real playable instrument, from a noisemaker. Original RCA design: http://home.att.net/~theremin1/RCA/rca_theremin.html I have a short writeup on it here for a more modern instrument: http://www.dvanhorn.org/Theremin/AntennaTuning.php Some more data here: http://www.dogstar.dantimax.dk/theremin/thersens.htm With the Theremax, the antenna coils were pretty much ignored, and it wasn't clear to me how they should be designed from the information that I've found on the web. I've been able to measure out what the Etherwave pitch antenna is, using my Tek 7L5 spectrum analyzer, and again the Theremin proves to have surprisingly complex behavior. I'm sure this applies to most Theremin designs, as they all seem to have the basic structure of a parallel resonant tank circuit in the pitch osc, and a series LC network in the antenna. I'm sure that some don't take advantage of the behavior that I've noticed here, to their apparent disadvantage. Conventionally measuring the antenna circuit proves to be difficult, because even the input impedance of a Tek scope probe is way beyond the effects that are produced by operating the instrument normally. Attaching the probe drastically detunes the circuit. Still, with special techniques, it is possible to examine the circuit, relatively non-intrusively. As it turns out, the spectrum analyzer is only needed when you don't already know how to look for this effect, or what causes it. Background: The antenna and it's loading coils form a series resonant circuit at some frequency. I had thought, and now confirmed using the Etherwave, that the behavior of the pitch control would be affected significantly by how the pitch osc is tuned relative to the series resonant point (SRP) of the antenna system. Most Theremin circuits that I have seen have the same basic structure. They have a parallel resonant tank circuit in the pitch oscillator, and the antenna and antenna coil form a series resonant circuit. The parallel resonant pair presents maximum impedance, nearly purely resistive, at it's resonant point. If the antenna circuit was dead on resonant at the same frequency, then it would also present some level of purely resistive impedance. The interesting thing, is that the antenna circuit presents it's lowest impedance at resonance, and the tank circuit presents it's highest impedance at resonance. So, the oscillator cannot function properly (if at all), if both circuits are tuned to, or nearly to, the same frequency. What happens: The behavior of the pitch oscillator is very interesting when tuned near the antenna resonant point. Starting at the lowest frequency, it moves smoothly up through it's range, then takes a sudden jump of about 10-20kHz when we cross the antenna resonant point. Due to the interactions, the resonant point is somewhat indistinct, but we can locate it closely enough to improve the system performance. Since the pitch oscillator becomes unstable at the point where it hits the antenna resonant frequency, this obviously would be a catastrophic place to tune your pitch oscillator to. :) So, you get to choose whether to place your pitch oscillator above or below the antenna resonant point. So, I investigated where the range is, relative to this point. I found that which side you select, and exactly where you select, makes a distinct difference. I tried a number of points, starting about 30kHz below the resonant point, and ending about 30kHz above the resonant point. Beyond these points, the difference diminishes rapidly, until you have very similar control behavior above, or below the resonant point of the antenna system. With the oscillator tuned just below the antenna resonance, the control range is somewhat compressed. The closer you get to the antenna resonance, the more compressed it becomes.(!) With the oscillator just above the resonance, the control range is expanded. The closer you get to the antenna resonant point, the more expanded the control range becomes. This makes some sense to me from an antenna tuning/tesla coil perspective. On the low side (pitch osc below antenna resonance), the antenna is slightly capacitive. A given increase in your hand capacitance lowers the self resonant point of the antenna, which makes it LESS capacitive, so the tuning range becomes compressed. On the high side, the antenna is slightly inductive, so you make it less inductive and more capacitive as you approach the antenna. These two changes add to each other, and this explains the expanded range. Different antenna and coil designs will react to different degrees, but the basics should hold. It seems that the shift in the antenna self resonant frequency works in conjunction with the hand capacitance, but only if you are on the high side of the antenna's self resonant point, and fairly close to that point The coils in the Etherwave are still a compromise, being physically very small, and having a lot of parasitic capacitance. I'll be looking at that soon. This behavior near the antenna resonance may also explain some squirrely behavior that I saw when first setting up the Theremax. There is also an area of aluminum tape near the pitch antenna. It is interesting, in that this would not present any sort of "ground plane", even at far higher frequencies. I assume it is being used as a capacitor against the antenna, which could be intended to further increase the control range, by swamping out some of the hand capacitance. I'm not sure on this point yet. -- http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist