Open theremin v3 loses calibration

Hello musicians, engineers and everyone! I built an open theremin v3, used an arduino clone and am very happy with the result, I have 3 octaves at a distance of 30 centimeters. But there is one bug: after the artist leaves the antennas, and this is required by the theatrical action, the calibration goes astray, the adjustment changes the value and the range of approach of the hand on the pitch antenna becomes 10 centimeters instead of 30 centimeters. Farther from the antenna there is zero beat and even farther away - a slight increase in pitch. The calibration went well and the instrument is playable again. is this a design feature and does it need to be calibrated before each playback session? Is it possible to fix the calibration settings so they don't change, in code? Maybe it's a hardware problem? 

A friend told me that the problem is the lack of reference voltage in my arduino.

Yes, I've noticed the same problem. I think this is a hardware problem. In my opinion, it is due to 8.000 and 7.3728 crystal tolerances.

One of my Open theremins plays just the little calibration melody then it makes some noise after touching the pitch antenna. The volume side seems to be ok but the pitch side is not possible to calibrate. Probably the MCP4922 chip is not working properly..

It would be very kind and welcome if any expert could help with this issue.

Yes, I've noticed the same problem. I think this is a hardware problem. In my opinion, it is due to 8.000 and 7.3728 crystal tolerances.One of my Open theremins plays just the little calibration melody then it makes some noise after touching the pitch antenna. The volume side seems to be ok but the pitch side is not possible to calibrate. Probably the MCP4922 chip is not working properly..It would be very kind and welcome if any expert could help with this issue.

Okay, I ordered the original arduino uno and will be able to test it in a week.

Yeah, well, you can try. But I don't think it's the Arduino's problem, the fault is somewhere in the components of the Open Theremin V3 Shield itself.

I have the original Arduino and regardless, it still does the above-mentioned misbehavior.

Studying a little the principle of operation of a Theremin, you'll discover that the theremin's antennas are kinda "blind" - they can't distinguish if it's your hand playing, or, at a little bigger distance, your whole body. Thus, in general, moving away from the instrument and approaching again will never let find you exactly the same field geometry as before, at least a little tweaking with the pitch and volume potentiometers is required in such a case, independent of the theremin brand and model.

Now, to the open.theremin: Pitch and Volume tuning are done in two steps. The coarse tuning is done through the calibration process. At its end, the "correct" tuning values are latched into the DACs which control the variable pitch and volume oscillators via varicap diodes. This value should not change abruptly, besides the little and slow thermal drift over several minutes, and which can normally be compensated by tweaking the potentiometers a little. But a full calibration cycle should definitively not be required if the circuit works correctly.

If you encounter problems, the first step of diagnosing would be connecting a computer with the Arduino IDE installed to the open.theremin and opening the Arduino serial monitor during the calibration cycle. It will show a lot of verbose information, like, among others, the initial oscillator frequencies, the tuning limits, and the final oscillator frequencies. I invite you to copy/paste this verbose output here for a deeper analysis.

Thierry, thank you very much for your input!

well, not to speak for Minix but I completely understand the principle of theremin antennas and its "blindness". What is a kind of mystery for me is the open.theremin tuning. I know that the coarse tuning is done through the calibration process and fine-tuning is made via pots.

I have built many of these open theremins myself. Some are working more or less nice but some of them have thermal or whatever possible kind of drift so big that it is not possible to fine-tune this beast via potentiometer. After playing for maybe 20 minutes the playable field around the pitch antenna is so small or big (depends on every piece) that the re-calibration is necessary. That's pretty annoying. So I assume that it may be a problem with the 8.000mhz or 7.3728 mhz crystal. Or I am wrong?

I want to ask you what are the less important components in this circuit? The DAC where calibration values are latched is MCP4922? How important are 74hc74 chips?

I have to admit I can't work with Arduino serial monitor so I cannot copy-paste the message.

Thank you in advance for your help

My first guess for thermal drift would be the ferrite core coils.  My second guess would be the varactor diodes.  The 4069 has a fair amount of input capacitance that might be temperature dependent.  The crystals would be my very last guess.

Thank you very much for your advice, Dewster!

OK, I try to resolder the diodes, coils, and eventually the 4069 and see what happens. I hope that it resolves that problem. Thanks a lot

"OK, I try to resolder the diodes, coils, and eventually the 4069 and see what happens. I hope that it resolves that problem."  - moonlightdog

I'm not sure resoldering them would help?  I mainly meant they are the things I would look at in terms of drift in a design.  If you have a frequency counter you might be able to characterize what is going on.  The debug messages might do that too (using the crystals as a reference).