temperature drift

" -800ppm/C for the 6310 is pretty horrible for Theremin use! " -- dewster

I’ve been playing around with a simulator, and it turns out that a temperature drift of +/- 800 ppm for the equalizing coil isn’t that fatal:

The black curve shows the original curve for EWS with an equalizing coil of 30 mH. On its sides are two curves corresponding to an increase and decrease of inductance by 800 ppm (i.e. for +/- 1 degree Celsius). As you can see, the zero-beat point shift is only about 5 cm. However, for, say, 10 degrees, it will be much more fatal!

It’s also interesting to know which inductance drift has a stronger effect on the “zero point” shift. Below are two more graphs, where the first one changes the inductance of the variated pitch oscillator (VPO), and the second one changes the fixed pitch oscillator (FPO). Both changes are +/-800 ppm:

And here is the result of their simultaneous (“correlated”) drift:

It’s evident that the total drift is not compensated – this is due to the non-identity of the circuits, i.e., the presence of an equalizing coil in one of the oscillators. But the most interesting thing is that the magnitude of the total drift turns out to be the same as that of the equalizing coil’s drift, but opposite in sign.

Therefore (and this is the main point!), the trick is that if all three coils have the same drift, the “zero point” will stay in place:

Clarifications:
1. For correct simulation, the "Auto calculated mode" for FPO must be turned off.

2.The model does not include the FPO oscillator coil, so its temperature drift is modeled indirectly – by changing the FPO frequency (I had to get out the calculator and figure it out).

BTW, the simulator already has tools to change parameters by a small value (1000 ppm or 0.1%). To do this, just hover your cursor over the parameter name and right-click to bring up the context menu. Just don’t forget point 1!

Thanks for that good illustrations, ILYA. Does that mean, if the three coils L1, L2 and L for the FPO have the same temperature drift in ppm/K, the zero point doesnt change? 

Exactly!

Can you explain this a bit? The FPO with low L/C and the VPO with extremely high L/C result in the same frequency deviation if the drift parameter for the inductors is roughly the same?

Thanks for doing this "basic experiment" ILYA!  Very interesting results.

"Therefore (and this is the main point!), the trick is that if all three coils have the same drift, the “zero point” will stay in place:"  - ILYA

Does anyone know the tempco of any of these coils though?  And tempco often isn't a linear function.

I'm not a radio guy so I have no experience with the coils used in e.g. the EW.  Transformers / coils designed for filtering the IF might have looser specs (i.e. higher tempco) than coils designed to be part of a local oscillator?  Are the EW coils IF transformers?