RCA THEREMIN

This may open the door for some experiments with recent production Etherwaves that have two series tank inductors.

Anyone want to take a shot at it?

"This may open the door for some experiments with recent production Etherwaves that have two series tank inductors." - w0ttm

Sorry - but I dont think this will work..

The magic seems to be in the split inductance on a common former, so that the total inductance (without virtual EQ inductance) is 4* the individual inductances.

There appears to be an extremely interesting relationship developed by the EQ circuit across one winding in such a coupled inductor - and as far as I can see, balancing the inductances has an huge linearizing effect -

<Please understand that I am just logging what I find as I go along - Rather than keeping my own on-going documentation, I am doing it directly to TW.. It may all be a wasted read.. From now on, these ramblings will be dimmed..>

I am also now thinking that there is probably some other method used on Clara's theremin to extend its coverage - It seems that the 50:50 winding of the tank inductor gives best linearity, and any deviation from this reduces linearity... I suspect, however, that changes to the total tank inductance and/or capacitance will change the coverage with less loss of linearity than changing the tank inductance ratio.

Tests are ongoing..

 I actually suspect that there is some simple mathematical reason for the improved linearity with the 50/50 winding.. A hunch - something to do with the conversion of square law to exponential.. that the 50/50 "bends" the curve the right amount to get exactly what is wanted.. In my head I can see the operation (but it may be illusory) - But I simply do not have the mental "tools" to even start to formulate an expression or start playing with the maths..

Update:

Just put some of my sim results into a spreadsheet - they dont, when correctly log scaled, look quite as good or convincing as what I was interpreting from the sim results directly.. The 50/50 still looks better than any other configuration - but not as huge an improvement as I had believed.. In fact, having the antenna across 1/2 to 3/4 of the tank inductance seem to have about the same linearity, with 1/2 giving about 3 really linear octaves (I would say 3.5 to 4 usable octaves) and 3/4 giving about 3 really linear octaves and perhaps 5 playable octaves. Bass response is more stretched as one increases the tank inductance across which the antenna acts, but top end compression is almost the same for all inductances >= 50% of the total winding.

Most of the linearity improvement with Lev's system seems to be at the bass (furthest from the antenna) end.

But NONE of the above may be even close to reality!

"From now on, these ramblings will be dimmed.."

Please don't dim :)

I've been inspired by ramblings from a great many people.

The initial ramble may or may not be off base, but it can start a train of thought that leads to something interesting, like when Edison noticed that if he put a metal plate in a light bulb, electrons would flow to it. This inspired De Forest, and now we have tubes.

Ramblings led me to associate the RCA waveform with what I've seen from guitar stomp boxes. I have a prototype built, and the real waveforms look even better than on the sim. I just need to make a couple of tweaks and document them. I'll keep one to experiment on and send a second to Amey for her to try in her Subscope effects loop.

woo!!! :) Can't wait!!

"I have a prototype built, and the real waveforms look even better than on the sim." - w0ttm

I loved the waveform you posted - compared nicely to the signals I have seen for the RCA.

Is there any chance of a peek at the schematic ?   ..... ;-)

Fred.

Back to the maths -

just rambling.. but the square-law relationship of turns to inductance makes me wonder if this could be related to the inverse-square law of distance to capacitance, and if by placing the antenna resonator (inductance) across 1/2 of the tank windings does some mathematical magic ..

I've had to put construction on hold for a few days while I handle a minor crisis with a friends CNC machine. I started to take a photo of the wave and discovered my battery was dead. I'll post a wave ASAP.

There is more than one schematic. The one as built gives a sine wave asymmetrical distortion. It looked great until I plugged in my theremin, which already has an asymmetrical output, and a bit less voltage than I expected.

The new version has a wider input gain adjustment range and selectable symmetrical / asymmetrical clipping.

This is the version as built. It does not have enough gain to work with my theremin, and if Amey's effects loop is low level, it would not work with hers either.

The next one will have the input gain controlled by op amp feedback and will not be quite as busy.

I've discovered I don't need clipping or filtering on the input stage, and symmetrical soft clipping on the output will work fine.

The pot between op amp stages is a one knob tone control that is very effective on over driven signals. The values can be easily tweaked to favor a preferred range. Since a theremin lives in roughly the same place as guitar, I pulled the values from my pedal.

Edit: Diagram moved to E-14.

Bob Moog did it the other way round in the EW Pro: While the pitch antenna circuitry is connected in parallel with the whole parallel resonant circuit of the VPO, there is a split inductance in the FPO. The tuning circuit (amplified virtual and current controlled capacitance through what was called a "reactance circuit" in the 50ies) is connected to the split point.

Hi w0ttm..

Lol - I think you accidentally copied your entire BIOS into that image link! ;-) .. Needless to say its unintellegable.. Unless its meant for skynet... 8-(

Why not upload the image to the (one participant, LOL) "discussion" thread here at Element-14 ..  I upload all my technical images to this site, and then link them at TW.

Hi Thierry ..

Facinating! But you dont particularly need linearity in the FPO - so perhaps theres some other reason.. stability?

You will probably laugh at this - but looking for a mathematical reason for improved linearity from the split inductance scheme, I notice that if one plots turns against inductance ( L = N² *AL ) of  the curve from the mid turns to the full count is linear by comparison to the whole curve .. I was using an AL of 30 (the mid value for the adjustable core I am using) and plotted from 0 to 140 turns (140 turns gives 600uH) from 70 turns to 140 turns is quite a linear zone for obvious reasons..

Is it possible that the "tapping" at 50% was deliberately chosen as optimal - that this linear section (at least, linear compared with the whole inductance) acts as a sort of "attractor" causing the square law to "pull" to a linear tangent and therebye provide a mid-way exponential relationship? .. I am seeing stuff and trying to find any testable hypothesis for what I see - But right now I only have "hunches" or ideas without substance - most frustrating!

I don't know what I did wrong with the diagram. It was just a bitmap like the waves I posted before.

It's up on E-14, but kind of small.

Rob- the schematic looks fine- at least readable when enlarged.-nice little circuit circuit,  pretty simple too! Is this something you cooked up from scratch?