Theremin Circuits Scratchpad

Small signal JFET's, TO92 package, N channel.

Some that should work are:

2N5484, 2N5485, 2N5486, J111, J112, J113, BF256B.

All of these are in stock at Mouser, and are cheaper than dirt.

The J111, J112, and J113 are designed for switching, but might work and have a higher voltage rating. At 19 cents each, worth a shot...

The voltages I'm seeing at the drain with a 12 volt supply are close to 25 volts, the limit for 2N548X and MPF102. The BF256B can handle 30, the J11X can do 35.

I'm thinking a 9 volt supply for the oscillators, and a diode on the gate might be a good thing to keep the transistors safe, and with a 9 volt regulator fed from a 12 volt regulator, improve stability a bit.

Dual FET's are usually either surface mount, or huge, high powered devices, and usually MOSFET's. They are a whole 'nother critter.

I forgot to mention that Im ordering from Jameco and have to settle for their lame supply.

oh they have 2n548x's  @ .69$ each.    Minimum order of 30.  !!

j111's are cheap enough tho'.     Ill have 30 to try.

What about 2n5457s?   I think I already have a bunch of them.

If you havent ordered them yet, I think thev J111 may be a bad choice - they have a wide range for cut-off voltage .. anything from -3V to -10V..

J112 is better at -3 to -5

The J111's might be ok - all depends where their cut-off voltage sits (and whether it matters with the signal levels - but I dont have time to do the calcs or sim right now).. I would put a socket on your board ( 1/2 of a 6 pin DIL turned pin socket does nicely) so you can swap parts if it doesnt work - You should get some which are ok..

This is what can be done with op amps, diodes, resistors and caps.

The input signals are both 1 volt peak sine waves.

This is from the sim, but I have the real thing with only minor variations in a stomp box, and it performs exactly the way the sim says it should.

1.5 khz.

100 hz.

wont get the 111's ....jameco doesnt have  j112's...so Ill try j113

(a haiku?)

J113 cut-off is -0.5V to -3V .. I think this is a better choice, as it allows you to add +Ve biasing if needed, but its cut-off voltage is not so -Ve that the signal wont be able to get there without -Ve biasing if it needs to..

By controlling the amplitude and perhaps adding a little +Ve biasing (which is a lot easier than adding -Ve biasing) you should be able to control the output waveforms quite extensively I think..

"This is what can be done with op amps, diodes, resistors and caps." - w0ttm

Real nice! And these waveforms look like the right direction..

Have a look at these: http://rcatheremin.com/tone.php

Fred.

ps - looking at your waveforms, and the RCA waveforms, I am wondering if all this halabaloo about the RCA sound is really justified -

The mixer is what creates the majority of the waveform - But if you can get the same end-result waveform from processing (distorting) a sine wave, well, why go to all the bother of replicating the mixer waveforms? Producing a clean sine difference from the Lev oscillators can be easily achieved using a MC1496.

I really dont know - Is there "Magic" in Lev's audio? or is the "magic" confined to the "front-end" - His brilliant oscillators and resulting playability?

I am not ready yet to completely write off "magic" ;-)

Yes, I looked at the RCA site and thought "that looks familiar".

The circuit is a guitar overdrive stompbox with a lower gain front end and modified for asymmetrical clipping.

I'm still holding out for magic as well, but this was just too easy....

Amey's Subscope has an effects loop, so I'm going to whip one of these up and send it to her.

"I'm still holding out for magic as well, but this was just too easy...." - w0ttm

Yes - "too easy" always worries me! ;-) - I think "if its that easy, why hasn't it been used / done extensively before?"

Not to say that one should write-off "easy" solutions - but caution is probably advised.. For example, the "RCA Switch" on the Theremax is an "easy" "solution" - just switch a 100k resistor into the circuit, but to my ears, its sound is nothing like the RCA. (on this matter, I completely disagree with Uncle Howie!)

I have not really tried to recreate the RCA sound - my focus has been on pre-mixer processing, using additive techniques to allow adjustment of each harmonic and craft whatever waveform one wanted, with frequency dependent control over the harmonic profiles - this (in theory at least) should allow any sound and behavior of any theremin (including the RCA) to be accurately emulated..

But it is complex - If a simple post-mixer circuit can do a reasonable job of replicating the RCA sound, then this will certainly have a place and be a lot cheaper.

As I see it, the Lev theremins had/have two main desirable features - First was/is playability and linearity.. I believe that the secret of this has now been revealed, and can be replicated with the simple circuitry which has been published here on TW - Simple Fet oscillators which actually have the same topology as Lev's original oscillators.

The other feature was / is the tone.. If this turns out (and its starting to look like it might) to be simple, then the job is done - only the volume circuit / VCA remains to be finalized, and the RCA's volume circuit is bloody awful - no problem in bettering it!

I suspect that an entire theremin design for a reasonable solid state RCA "clone" which is simple to construct with low cost, readily available parts, could easily be completed this year - IF there is no "magic" we have yet to discover.

Fred.

Ah, the volume circuit.

Uncle Howie recommends going with the Melodia VCA. I was ready to dismiss that idea at first, but after some thought, I realized I have a functional version of it under my nose, and it does work pretty good.

This is a "white elephant" theremin and while rare, has generated almost zero collector interest so I have no problem experimenting with it, and I even have 2 spare coil forms for it.

There is plenty of room inside to add all sorts of mods and test circuits, and I can save the original parts in case some collector looses their sanity and wants it.

I wonder how it would perform with a Termen oscillator on the volume side?

Just thinking...