Theremin Circuits Scratchpad

"So, bottom line, is there any magic in the Lev oscillator?  I just see it as a split coil design, done more for active component (tube) biasing reasons than anything else, but maybe I'm missing something critical." - Dewster.

I havent done the maths, and I haven yet been able to modify your spreadsheet, and I havent built a Lev oscillator / antenna resonator and run tests on its linearity - so this is "guessing" or "hunch" more than anything else - probably...

The operation of the antenna linearising, as I understand it, is that effectively the antenna resomator behaves like an "variable" inductor in series with the tank inductor.

In parallel tanks, this "virtual inductance" sits across the entire tank inductance (in fact, across the entire tank resonator) - With the Lev oscillator, it is only imposed across one half of the tank inductance ..

I see several possible implications of this - but these are probably unfounded hunches.. 1.) The effect of the antenna resonator will be reduced in "global" terms - as in, loading will be reduced on the primary feedback mechanism, as there is no "virtual inductance" across this. 2.) Although "reduced" (because it is only across 1/2 the total inductance), this "virtual inductance" is likely to have a modified behaviour because it is across an inductance which is coupled in a complex way to its "other half" (the split inductances actually combine magnetically to form a "single" inductance, but this is split with the series tank capacitor forming the series resonant tank circuit).

I am out of my depth a bit in understanding the complex behaviour of this network, but we do know a few things about the Lev theremins.. And amongst the "facts" are that:

A) they are reputed to have exceptional linearity - linearity which SHOULD be common place in all theremins if there is nothing "magic" going on, but which isnt commonplace..

B) they have a smaller range (fewer octaves - less "sensitivity") than most more recent theremins.. But this alone is not the reason for the improved linearity - One can build a parallel tank based theremin with reduced sensitivity, but gain no linearity advantage - you generally just get fewer octaves with little or no improvement in linearity.

To me, there was no basis for the improved linearity over modern designs until the "mysteries" of the Lev oscillator were "revealed" - Now there is a possible reason - not saying it is the reason - But it certainly looks more likely than the other "reasons" we have heard like tube behaviour or large coils or "non capacitive" sensing mechanisms or mysterious antennas!

C) Almost every theremin designed / built other than Lev's originals, use parallel tank oscillators, and it has been believed (since the 30's !?) that this was Lev's topology - even to the level that schematics drawn and analysis of the Lev theremins have had wrong inductance values because these were based on calculating with the assumption of a parallel tank circuit.. All theremins (other than direct physical copies of Lev's) since the '30s have probably been built "wrongly" on "wrong" assumptions!.. Or if not "wrongly" at least "simplistically" - because I think we will find that we completely missed the brilliance of Levs "Front End" and only picked up on the crudest fragments of his ideas.

Is there magic ?  Well, I would bet my last remaining pennies that a correctly implemented Lev oscillator with a correctly implemented antenna resonance circuit will be a lot more linear than any "standard" theremin with paralell tank resonator and optimally implemented antenna resonator, and that the Lev Oscillator will be more stable, particularly when coming under synchronisation of another oscillator.

Is there any other magic? wave-shapes / blocking actions etc? - I dont think so.. Except perhaps for its possible synchronous behaviour / stability in a theremin design when VFO and Ref oscillators are loosly coupled.. But even here I could be wrong - Lev was, undoubtably (IMO) a greater genius than has been realized - I think his "Great Seal" probably demonstrated how subtle and elegant his thinking was - an entirely passive bug built with effectively a capacitive microphone and external excitation provided by the remote reciever - and think about the date that was implemented! - I think we should be on the lookout for some other hidden tricks! ;-)

Fred.

The spreadsheet, particularly the hand capacitance part, is a first order simulation, and I haven't nearly exhausted all the possible component value combinations in it, but I can usually get around three octaves of linearity from the designs I put into it (mine, EWS, etc.).  Even the RC shows decent linearity, though of course there are other factors such as playing field width, absolute pitch, etc.  The LEV trick I believe is to keep sensitivity low and push the linear section of the playing field out so that the pitches are widely spaced and thus easier to play.

My bet with the LEV is that the other side of the split coil also "sees" the linearizing coil load due to coupling.  There is unfortunately no way to easily test this in the spreadsheet because it is essentially just a complex impedance voltage divider, and inductive coupling would require one or more active sources within the network.

In general, for parallel LC networks the drive and pickoff point are the same, and this point is at its highest impedance at resonance (hence the open collector current source drive).  Series LC networks have drive on one end and pickoff on the other, and have their lowest impedance at resonance, and this is what the linearizing leg (linearizing coil & antenna / hand capacitance) looks like.  Does this seem like a good situation?  A high impedance network driving a low impedance network?

My (mostly) digital Theremin AFE is something of a hybrid: a series resonant tank, but the pickoff point at the LC intersection, and in simulation I see 1/2 the maximum impedance at resonance, as you might expect.  With large L and small C it's easy to get hundreds of volts at the LC intersection, and a linearizing coil boosts this even more at the antenna. 

Does anyone know the voltage swing across the capacitor of the LEV tank?

"Does anyone know the voltage swing across the capacitor of the LEV tank?" - Dewster

I dont know what occurs across the "real" capacitor - All I have is the simulation of the fet "clone" which shows the waveforms acress this capacitor:

http://www.thereminworld.com/Forums/T/28796/rca-theremin?Page=5

I think im going to have to do the simulation the hard way - run the circuit with a set of capacitances derived from distances, log the frequencies, and import these into a spreadsheet..

But it will probably be a week - I have a paying job to complete first (this job should be a doddle though - a capacitive sensor which doesnt need to compensate for our weird concept of "linearity" LOL ;-)

Fred.

Concerning pitch range.....

Apparently it can go quite a bit more than a stock RCA. Clara's has a five octave range.

I just confirmed this, from Clara herself.

After dinner with Clara Rockmore.

"Clara's has a five octave range"

Sadly, we really dont have detailed information about Claras theremin (or at least, not much that I have found) - We dont, for example, have details about the tank inductor / transformer windings.. The one clone article I have seen seem to make the assumption that this is the same as thr RCA's, but can we be sure that Lev didnt tinker with it and wind a few more turns on the grid coil than say on the anode coil?

The schematic is, well - sketchy! ;-) .. we can glean some of the differences from the RCA, and these make sense.. But I see nothing in the schematic which would cause more linear playable octaves - I think it likely that this "secret" is in the details of the windings of the tank and/or equalizer / concentrator inductances.

Oh, if only I could get to these theremins with my 'scope and my LCR tester!  - Its just criminal that no one has fully explored and documented these treasures!

It looks like uncle Bob only traced out enough of the circuit to allow him to fix it. That's not unusual. I've done that myself when the goal was just to get it working, and the customer was paying by the hour. It's obvious from looking at photo's of it that not all of it's innards or parts details made it to the schematic.

I think, from looking at photo's, that Clara's concentrator coil was large-ish. About the same size as a toilet paper roll.

I also think that Lev did tinker with it. A lot. There are probably significant variations in almost everything.

I'm considering posting requests on other places like levnet, facebook, etc.

I believe Lydia Kavina has a "New York Skyline" theremin that might have been built by Lev, and Rob Schwimmer has had access to Lucie Rosen's September theremin.

If we can convince the ones that own or have access to the icon instruments how important documenting them is, life will be very good indeed.

If they can't get a tech or engineer to fully analyze them, we might at least get some decent audio. Even a lot of high resolution, close up photo's would help. With a ruler next to the coils and close enough to count the turns and read the values or color codes on other parts.

@ Charlie.

I just thought of something. Lift the tubes up a bit while still making good contact with the socket. Then you can carefully take readings from the pins without inverting the chassis.

@ Fred.

I finally gave in and joined facebook. Friend request sent.....

"I finally gave in and joined facebook."  - w0ttm

They say the best two days of boat ownership are the day you buy it and the day you finally get rid of it.  With facebook only the latter applies.

w0ttm: "I finally gave in and joined facebook. Friend request sent..... "

Alas, you may need to wait a while for me to reply - I was dragged kicking and screaming to join facebook by my daughters - and I only did because it was the only way to access their photos.. But I visited it rarely, hate every moment on it, and get overwhelmed by its interface, and dont trust anything about it..

To me, "facebook" and "skynet" are nearly synonyms! ;-)

"I believe Lydia Kavina has a "New York Skyline" theremin that might have been built by Lev"

Time to get in contact with Lydia again.. She's an hours drive from me.. The only problem is that I told Lydia I knew nothing about tubes - so she may not be happy about letting me poke about in her tube theremins ;-) .. Perhaps I could persuade her if George was supervising..

Fred.

"To me, "facebook" and "skynet" are nearly synonyms! ;-)" Fred.

LOL. The only reason I signed up is because Lydia, Dorit, and a ton of other threministos hang out there.

"I'll be back".

I just found this .pdf on simulating tubes - not tried any of it yet, but thought it may be of interest:

http://users.ece.gatech.edu/mleach/papers/tubeamp/tubeamp.pdf

And this:

http://www.normankoren.com/Audio/Tubemodspice_article.html

Funny - I wasnt looking for tube stuff - Have looked for exactly this sort of thing before and not found these!