Theremin Circuits Scratchpad

Wild a$$ed theory.

It looks kind of like L1 A, B and C1, 2 form a series resonant circuit.

Latest schematic..

UPDATE - At the time the above was posted, I had not understood the implications of what I was seeing - I believe that (thanks to a lot of help from w0ttm) I now do!

This is a LEV OSCILLATOR - note the series configuration of the tank resonator - L1:B to C5 to L1:A

The above has errors:  L1:A and L1:B should have individual inductances of about 158uH, the combined inductance is about 598uH, C5 is 1000pF - with these values, which are the same as the values in the RCA (values given in many descriptions of the RCA are WRONG, and are based on calculations which are wrongly based on the idea that the oscillator is an Armstrong oscillator, with a parallel tank resonator) this oscillator behaves like, and "is" a Lev oscillator - There is some question about whether other "things" happen, but as far as I can see, the above explains everything.

L1:A and L1:B are wound on a common former and act both as a "transformer" (to couple Anode/Drain currents back to the Grid/ Gate and provide 180 degree phase shift) and as a composite inductance behaving as if it was a single winding and therefore having a larger inductance than the inductance which would be obtained if these were simply series connected independent inductors.

w0ttm - you are a genius!

L1:B and C1+C2 must be series, because otherwise C4 would be across L1:B !

But how does this work with the antenna EQ circuit? .. Holy $$$$ ! could it be that the strange large capacitor C5 in combination with  L2+L3 and the player capacitance is effectively "shunting" L1:A ..

Got to think this through before I talk complete nonsense! LOL..

I spent most of my life deliberately avoiding RF.. Only got into theremins because I didnt see a few hundred kHz as "RF" - LOL ;-) .. If I had really seen what was coming, no-one in the theremin world would ever have heard of me (and I would have a lot more money to spend on some other daft persuit! ;-)

What I now need is actual confirmation of the REAL (not calculated) inductance (L1:A) tickler (L1:B) values, and the tank capacitances (C1+C2)..

Fred.

I just calculated the inductance based on this and came up with 165uh each for L1A and B.

Diameter 1.5 inch

Length 1.3 inch

Turn space .017 inch (#26 wire)

Turns 76

That works out to 264 khz.

Now I'm confused.

I suspect we don't have accurate parts values, or the coils have a LOT of capacitance.

Now that is much more likely to be closer to the truth!

Will try it before I post the waveforms on the above! Thanks!

>> Update !

Damn close! 184kHz with 1100pF on the simulation... Not going to edit above schematic, nothing else other than the tank capacitor C5  (300pF -> 1n1) and L1:A and L1:B  (both now 165uH)

I think we have a winner.

The frequency difference is probably from the fet capacitance vs. an antique triode.

This may be the first time anyone has tried to understand the RCA circuit since it was first designed.

I don't think it can truly be called an Armstrong.

I'm gonna' call it a Lev oscillator.

"I think we have a winner" - w0ttm

Damn right! ;-)

Some other notables:

Reducing the inductance did also reduce the harmonics :-( and made them less likely to be musical - big 3rd harmonic peak.

In playing about trying to alter the drive, I discovered something intersting.. Loading L1:A has an effect on the harmonics! - changing its inductance or placing small capacitance across it has a surprising effect...

And this MIGHT mean that there are subtle pitch related harmonic effects on the audio output after mixing.

I am really surprised - I thought the original circuit would have been explored deeper than it appears to have been - Perhaps (like me) most who looked at it never really understood antique valves, and just assumed that these were responsible for any anomolies they saw..

I would have done the same - its only when I see strange things happen on components I fully understand that I bother to ask questions.

Fred.

Its not a Lev oscillator.. By all accounts it seems it was invented simultaneously by Alexander Meissner and Edwin Armstrong

Both published their design somewhere 'round 1913 / 1914 ..

Its not a Lev invention.

- However - The use of combing series resonant and parrallell resonant tank circuits with a series resonant antenna circuit, if that is whats going on - may well be a Lev invention!

I say "If" - Because my brain for some reason is shouting at me that im missing something big!

Fred.

ps - all the best playing with the real stuff - I will be back much later.

Rob (w0ttm).. I think you were right!

I don’t think this is an Armstrong oscillator - It may have originated from a Meissner / Armstrong oscillator..

But it has been changed so hugely now that I think it could fairly be called a Lev (Or Termen) Oscillator.

Major difference I can see is that the "tickler" winding has a much greater role in this oscillator, forming 1/2 of the inductance of a Series tank coil.

Fred.

Just for info, here the capacitance values of the UY-227 RCA Radiotron:

Grid-Plate 3.3pF (The so-called "Miller capacitance" will be this value multiplied by the amplification factor of the tube stage)

Grid-Cathode 3.5pF (The "classical" input capacitance)

Plate-Cathode 3.0pF (The "classical" output capacitance)

Source: The RCA Radiotron manual R-10 from 1932

Another info:

It happened to me too to find "out of range" frequency values (but with very nice and rich harmonics!) when simulating oscillators, especially in Colpitts configuration. These oscillations didn't even not stop when I virtually disconnected the tank inductance in the simulator and I finally found that it was due to some badly chosen "peripheral" R-C time constants and initial simulation parameters.