"But wouldn’t the two coils reverse-orientation make the coupling constructive?"  - BBrinkman

I'm not getting why you would want to stick L12 and L13 together?  If it's for temperature tracking I'd be much more worried about coupling.  Constructive or destructive coupling would most likely be bad here.

I learned a lot about coil coupling via Inca, a small coil simulator.  Find it here along with a couple of useful papers:

http://www.coe.ufrj.br/~acmq/programs/

It doesn't take a lot of distance for coupling to fall to almost nothing, though even a little might be a problem.  High Q resonances at the same (or close to the same, the meaning of which depends on the Q) frequency can couple quite easily over larger distances (hence radio).

Thank you for the link.

I am going to use the Epcos inductors as mentioned earlier. The bifilar arrangement was more a thought experiment prompted by Thierry’s comment regarding experimentation with L12 & L13. I have magnet wire and forms leftover from other projects and if I do any experimentation I’ll post it here.

Here’s another interesting thread:

Alternatives to YAEWSBM and EPSE01 module

I fully agree with Dewster.

Placing the two pitch oscillators apart from each other as long as you can.

In my last build, i used all the space i had on a single sided 300mm PCB.

Too close is not good at all.

Good point BBrinkman, decoupling and mixing can be (and probably should be) combined.

I've got a (simulation only) totem-pole mixer over on my analog Theremin thread for low voltage work.  (Call me crazy but I like to make things work with a 3.3V supply.)

There is another thing i would do in a future build concerning the power supply.

Everybody is looking at inductors, capacitor tempco etc... BUT..

In my last EW build i first used some (too) cheap tank capacitors. Tuning was a pain in the ass, every 30 seconds i had to adjust the coils.

When i searched the reason, i found that the voltage regulators became  quite hot.

They haven´t became hot because of oscillations, but because the voltage of the mains transformer was too high, the regulators were dissipating too much power.

First i used a 15Vac transformer which gave me over 25 volts at the filter caps, even with the single diode rectifier.

Then i changed to a 12Vac transformer which gave me 20 volts, and changed the Regulators from TO92 to TO220 parts.

Those little ac transformers all give a higher voltage output to compensate the power line fluctuations.

The bigger surface of the TO220 regulators helps keeping temperature of the part lower.

The overall dissipation inside of the case stays the same, the air inside will heat up equally.

I suggest for diy builders to build a seperate stabilized power supply with a 3 pole plug.

So the heat generated by the regulators is outside the theremin and will not heat up the stuff inside.

Also one can add heatsinks without bringing metal parts near the oscillators, and allowing a higher current for extra electronics.

The theremin is like an old girlfriend, during dry spells it's nice to revisit once in awhile.

Oilspill you bring up many good points.  Using the T0220 I have mounted them on the bottom side of the pcb to extend through a hole mostly outside of the enclosure. Also I think a T0220 likes a little more current demand so I used a 330 ohm across the its output for better regulation, this could use a little more study.

Christopher

An inexpensive way to facilitate part swapping is to take a pair of needle-nose to one of these bad boys:

The socket pins can be individually popped out and reused in various eyelets on the Etherwave PCB. Most component leads fit in these pins snugly for a secure placement.

I imagine these would be quite useful at the following locations on the Etherwave PCB:

L12, L13, D4 and for use with various decoupling/mixing modules: C2, C6, GND, +12V, -12V & AUDIO.

Oilspill, yes, if the regulators get hot they should be beefier, heatsinked, outside the enclosure (or given separate breathing room).

I wish Theremin designers would get away from the +/-12 V supply as it generally means high current, and most analog ICs don't need it anymore.  My buffered Colpitts NPN oscillator draws 0.3mA @ 3.3V, so basically nothing heats up.  Go for high Q and a lot of problems go away.

On top of the thermal issues, voltage regulation is usually extremely important for oscillator stability.  If the regulators are getting hot and drifting, chances are quite high that your oscillators are drifting too.

BBrinkman, they make those in single rows too.  That's what I buy for generic IC sockets, just clip off however many you need per side.  I also use them for the antenna sampling capacitive divider, where component changes are sometimes necessary to get the swing right.

What accessories do I need for the temperature regulator? I haven't found it yet.

I am still wondering which inductors are used for the 2.5, 5 and 10 mH.?  All the ones I have found have SRF way too low, some at or below the oscillator frequencies.  That prevents tuning the unit.
I did an experiment on the volume oscillator using some radial Bourns inductors.  I injected a signal through a 10pf cap into the antenna with power off.
A scope connected to the detector output.  The resonance was about 325 KHz.  Even shorting out some of the inductors did not help much.  Could not get anywhere near 450 KHz.