Grumble's Theremin Build

First post in this construction forum to check if I'm on the right path in building a Theremin (partly) my way...
Since I'm in the possession of a 3D printer, I might as well put it into use by building a Theremin, and the first thing I did was printing a coil body, coil length = 9cm, diameter  = 3cm, wire diameter = 0.2mm.

As Dewster suggested (somewhere, can't reproduce the thread), I connected one side to a function generator and the other side to an antenna (which in this case is a 50cm long wire). Have a scope channel connected to the output of the signal generator and an other channel of the scope to the junction coil/antenna. (probes are 10x 10Mohm, 14pF)
That's what I did and this is the result:

The resonance frequency is 1.47MHz, the phase is 90° (a Lissajous circle) and the Q-factor (correct me if I'm wrong) is about 25  (50 volt per division / 2 volt per division) 
If I move my hand to and away from the antenna, I can see amplitude of both the signals and the phase of relationship change.
Obviously the impedance of the coil plus antenna is maximal at the resonance frequency, away from the frequency the impedance is lower, so it will affect the amplitude of the function generator.

I have a questions where I didn't find any answers for:

Is there a reason why the volume and pitch oscillators are adjustable while the reference oscillators are fixed? I am thinking about building the pitch and volume oscillators "fixed" so they can be set for the best possible behavior, while the reference oscillators will be a programmable waveform generator that is crystal stable but microprocessor controllable.
When using the clock generator of say an Arduino (16MHz) the resolution of this generator is 0.064Hz and crystal stable!

My knowledge of analog circuit design is limited so I guess I have to "borrow" oscillators schematics and antenna optimization technics from others 

"Since I'm in the possession of a 3D printer, I might as well put it into use by building a Theremin, and the first thing I did was printing a coil body, coil length = 9cm, diameter  = 3cm, wire diameter = 0.2mm."  - Grumble

I would go this way too, but I fear the application of heatshrink tubing to the outside would warp the printed tube (untested).

It seems you are using AWG32 wire?  If so, from your dimensions I calculate ~1.5mH for a single layer solenoid.

"Have a scope channel connected to the output of the signal generator and an other channel of the scope to the junction coil/antenna. (probes are 10x 10Mohm, 14pF)"

You're better off I think connecting that scope probe to a short length of wire placed somewhat near the antenna, rather than directly connected.  Even a simple C divider will load things down and give you false readings.  If you disconnect the antenna you can also use this method to find the self-resonance of the coil.  For a more realistic Q reading you might look at the difference of the 3dB down points and divide by the resonance frequency, rather than trying to measure absolute voltage swing.  I would bet the Q is higher than 50.  Also, you can just look at the drive and probe waves on the scope and easily see the jump in phase at resonance, no need for XY channel math (though that's certainly an interesting way to view what's going on).  Also the 50 ohm impedance of the generator output (I presume?) will lower the Q some because it adds with the coil DCR.

"Is there a reason why the volume and pitch oscillators are adjustable while the reference oscillators are fixed?"

You mean like tuning with ferrite slugs?  I'm not sure I follow you.  The EW has slugs everywhere, and this is mainly due to the coils chosen / available, but they also make it possible to tune the fixed and variable LC tanks to the series LC formed by the EQ coil and antenna C.

"I am thinking about building the pitch and volume oscillators "fixed" so they can be set for the best possible behavior, while the reference oscillators will be a programmable waveform generator that is crystal stable but microprocessor controllable."

If you have sufficient resolution at the generator I don't see why that approach wouldn't work. 

[EDIT] Though I haven't really examined the datasheet you point to.

[EDIT2] I can't tell for sure from your coil photo, but it looks like the leads are possibly pressed against the coil windings, which would increase self C.

I made 2 small holes in the cylinder at the end/beginning of the coil.

Is there a reason why the volume and pitch oscillators are adjustable while the reference oscillators are fixed?
I am thinking about building the pitch and volume oscillators "fixed" so they can be set for the best possible behavior,
while the reference oscillators will be a programmable waveform generator that is crystal stable but microprocessor controllable.

No difference which oscillator to adjust.
BTW, for variable pitch oscillator you can use AD9833 as well (in PLL mode) - sense phase shift between drive signal and antenna voltage or LC tank current and periodically correct generator frequency to keep phase shift locked at 90 or 0 degrees.
PLL (phase locked loop) allows to get really stable oscillator which would filter out noise.

Are you planning to implement (mostly) analog theremin? With exception of the way how fixed pitch generation is implemented?
What do you plan as a mixer (heterodyne)? Analog multiplier IC could allow to achieve ideal sine, but require higher power supply voltage.

BTW, thanks for pointing at AD9833 waveform generator.
While single chip costs $12, there are a lot of small boards available at lower price ($2.5..$3.5), with AD9833 and onboard xtal 25MHz.
With such board you even do not need provide clock signal, 3-wire SPI connection is enough.
AD9833 output is only 0.038V..0.65V and should be scaled up (e.g. using opamp) to get higher voltage swing.

On my simulation, serial LC tank (for current sensing, one side of inductor is drive, another side of inductor is antenna) driven with 5V powered opamp AD8651 with 1K+6.2K resistors giving 0.4..4.6V sine drive via 10 Ohm, with 2.7mH inductor (120 Ohm serial resistance, 1.2pF self capacitance) and 8pF antenna gives ~500Vpp swing on antenna.

I haven’t got a fixed idea how to build my theremin, but in my mind it starts to roughly gets shape.
At the moment I do a lot of thinking what and how to use for the several parts of this build.
First let me tell you that the analog part is way out of my comfort zone, so I will use diagrams from others when building the pitch and volume oscillators.
Let me sketch what I came up with so far.
For the pitch and volume oscillators I am looking at oscillators that use a cmos inverter (like the OpenTheremin), that will give me a square wave, but as I understand it, when mixing with a sine wave, the demodulated signal will be sine wave (?)
The pitch and volume oscillators will be optimised with the used antenna while the local oscillators will be made with the AD9833’s.
For the mixer I’m looking towards a four quadrant analog multiplexer like the AD633
After the mixer in the pitch chain I’m looking at a programmable switched capacitance filter while the demultiplexed signal of the volume oscillator it will probably be a more simple filter, a rectifier and a low pass filter.
Since I need a microprocessor for the tuning of both the AD9833’s maybe I can implement midi as well, and/or control voltages and a gate signal for my diy modular synthesizer.
But for now the design is mainly virtual, and I am reading/learning a lot about theremin design….
It’s easy to change a virtual design, much easier as a pcb 😉

Btw, I recently retired from work where I could sample components from Analog Devices, so I still have some AD9833’s and AD633’s.

The beautiful theremin “voice” that few achieve develops from the proper distortion of a lazy analog sine wave. A perfect sine wave gets you more of a whistle or flute sound. I am jealous of your theremin journey of discovery.

Good luck

Christopher

Edit: I use my programing skills today to help people. . . and reveal the profitable side of the Stock Market.