Theremin Circuits Scratchpad

Heh, I have an old(late 70's ) Heathkit "Dual trace Oscilloscope model IO-4550"

...so unfortunately no USB!   I can take pictures or a short videos of each test

Hi Chobbs -

I have some rough tube models, and am doing some simulations using these.. I will "test" the effects of probe loadings as best I can, and advise.

Hi folks,

Just a quick update - I got some lousy triode models, and simulated with these.. Results were dramatically different to my fet simulations, no matter what I did.

I am now playing with a different option - Again, severe problems simulating this, but with my solderless breadboard and a few components I am working on making a solid-state "valve" using a fast opamp and H11F1 optoisolator.

The H11F1 is slow, but fast enough to manage the RCA frequencies - with the opamps driving it, I can create a high-z "grid" connection and set the gain and biasing for the opto to behave in a region where the cathode-plate charactaristics are closer to what (as far as I can deduce from tube data) is required.

So, in a way, a step backwards - I am "cloning" the tube in order to "clone" the oscillator.. This has distracted me a bit from my focus of further examination of the original RCA.

Simulation is starting to become a bit of a problem, due to unavailability of accurate component models - I could create these models (it was something I specialised in during a past life ;) but it could take a long time - and (as always) I have had plenty of time for the last couple of years - but now that I really need that time, I am getting a lot of work..

Fred.

Here is a quick glimps at the kind of "triode clone" circuits I am playing with:

Chobbs -

This whole issue of "RCA tone" is frought with contradictions -

I have dealt with this in boring depths in past posts - but it comes down to the different way that modern instruments are interfaced - The sound we recognise as authentic RCA, is mostly sound taken using some ancient microphone into an ancient recorder.. Even modern recordings of surviving machines are usually miked..

If we design a clone which has line-level output and drives a good Hi-Fi system, perhaps we shouldnt be looking at the electrical signals - because if we emulate these, the results when fed to a modern Hi-Fi will sound nothing like what we expect..

Perhaps the best compromise for now is to simply take recording with a good quality microphone, from your speaker - record in semitone-ish steps from the lowest to the highest note, hold each note as steady as possible for perhaps 10 seconds, and vary the volume of each note from mute to max, for about 10 seconds (20 seconds total per semitone) - or whatever youi can manage.

Having a simultaneous video showing your hand positions might also be useful - but audio quality is by far the most important.

Yes - having recordings from the innards would be a great help - but perhaps that can come later.

It frustrates me greatly that a database of sounds from original theremins has not been created - We are focussing on the RCA - But Lev created far better theremins.. We should really be looking at these rather than the RCA.

I believe that in the not-too-distant future, analogue theremins will be a historic oddety.. We are, I believe, at the 'cusp' - digital technology still, IMO, has not 'matured' to a level where it is the overwhelming obvious choice for implementing theremins - And, IMO, the simplicity and sonic advantages of analogue make it superior for theremins... But only just.

Sound samples taken today could well be used in the future in a new generation of theremin -

Fred.

Chobbs (and others who have expierience with tube circuits)

Regarding probes and viewing waveforms..

I have no expierience of probing these kinds of circuits with a 'scope..

Photographing waveforms on the test points would be a good step in the right direction - so if this can be done easily and SAFELY, without causing loading problems, then lets do it..

The high z of a 'scope input makes isolation capacitors (if they are needed) requirements far smaller - I was basing my capacitor values on a 10k loading from the audio recorder (PC soundcard or whatever) - but one should be able to drop to about 100n when going into a 1M scope input - so 2*220n in series, with each being class Y2 rated at 250V, will give more than adequate protection.

I would make a clip probe like this:

Probe<---10k-----220n Y2 250VAC --- 220n Y2 250VAC --- Scope probe--- Scope input >= 1M

But I have not used the above with tube circuits - I know the above works safely to block DC at mains voltages.

But not sure if it is "overkill" in terms of viewing valve circuits - any advice on simpler probing would be welcome.

Fred

Update:

I dont think I can use thev H11F1 for a triode 'clone' - too slow, and too non-linear when V4-6 exceeds about 80mV - cannot get it to work.. I had hoped that when biased into its active (linear) region, small modulation of the LED current would be much faster than the 25us on/off time - it is - but it still doesnt seem fast enough and introduces large phase lag.

I am now stripping a spice model of an ECC83 (dual triode) and attempting to construct a model for the UY-227 using the basic triode spice structure as a template.

My feeling is that if I can see (simulate) the oscillator using as close to the 'real' parts, I will be better able to design some solid state triode replacement circuit (or even use fets as in my original - for all I know, the fet output waveform may be closer to reality than my wrong triode simulations).

Anyone who knows where I can obtain spice models for other tubes, please let me know.

I may go quiet for a while - probably wont have much more to say until I have a running oscillator simulation using a triode - and that could take a while.

One final wild thought.. Does anyone know of anyone in the UK who has a RCA, and who might be willing to have me probe it with a small handheld 'scope ?

The ideal would be to actually look at the oscillator waveforms - but to do this one needs rather fancy 'scope and active probe - stuff I have, but the probe alone cost me £400..

Perhaps I need to go on one of these on-line sponsorship sites, and see if I can get people to donate enough that I can pay Chobbs a visit ;-)

Fred.

Guys, got this working- I stopped by the 'shack and picked up some mpf102's like, w0ttm suggested.  For coils, Im using one of the  'trial runs/ duds' left from the RCA build, floating around the junk box.  This coil is on a slightly larger dia tube, like 1.58" instead of 1.5"   -thus the inductance is a little higher- like 188uH and 198uH.   On my model, w/ only 1000p  for C5,  it's running exactly  at  172.5khz.- havent looked at it on the scope yet. What is R1- 1e6?
 I've left it out.

I have plenty of 'proper-size' phenolic- will be ordering more magnet wire soon.

Any ideas for a mixer?     

R1 is a 1M resistor - not needed - I fit this for protection against ESD primarily, so that I can safely mess at the antenna side without risking "floating" the gate.

Great that this works! Would be great to know how waveforms compare to the tube versions.

I have a mixer circuit published here somewhere.. < first page of this thread > but I think it will need a lot of tweeking.. I strongly suspect that it is the '24 based mixer which gives this instrument most of its character.

I have been thinking along the lines of having a diode circuit of some sort at the input stages of the mixer, to emulate the distortion caused when the cap (CG) is is operating with +ve bias (as in, outside its linear zone) and any other distortions (perhaps at the grid (SG) which may occur.. But was hoping to be able to simulate the the RCA's 224 mixer before designing any required distortion circuit/s.

Fred

been busy -Sorry for not getting back sooner

Fred -So I have everything for the probe, just havent put it together yet.   Send me the marked up schematic w/ the test points ( or even a verbal list- ie: grid of v3 ;or  junction of T6 and r15)      and Ill get started on "catching some waves" for you.

-Mixer,  Ive been looking at your mixer on the first page- Lots of parts.   (if I may)... it looks too thought out.      My .02$, as a builder,  I think the most logical next step is to get a pair of FET termen oscillators up and singing, any way possible- then (re)evaluate .        The original RCA mixer is one tube, right ( ok,....and a resistor.... and a 1:5 tx, perhaps).  Im pretty sure that the 24 was one of the newest tube types used when the rca was designed and built- so there werent a lot of other choices -( tubes w/ different performance characteristics)- I would suspect that the 24 was chosen because it was the simplest way to mix/extract an audible tone-

Fred (and others)  - "let go -trust your instinct - use the Force"  

What is the simplest (or even dumbest) way to "mix" the pitch oscillators?  

 edit-oops   just saw the sim models you and thierry made.

"edit-oops   just saw the sim models you and thierry made.."

These are theoretical only - for one thing, look at the voltage levels! - You would need level shifting and gain / attenuation circuitry - and things get a lot more complicated than my mixer schematic!

The simplest mixer is the diode type as per EW.. I have not examined the effect of loading the oscillators into such a (IMO Horrible ;-) circuit.

You say "The original RCA mixer is one tube, right ( ok,....and a resistor.... and a 1:5 tx, perhaps).  Im pretty sure that the 24 was one of the newest tube types used when the rca was designed and built- so there werent a lot of other choices -( tubes w/ different performance characteristics)- I would suspect that the 24 was chosen because it was the simplest way to mix/extract an audible tone-"

The above is true.. But the closest equivalent to the tube is probably a dual-gate fet - I do not know much about these - AFAIKS my schematic will probably behave similar, and all the dual gate fets I can get easily are SMD.

" Send me the marked up schematic w/ the test points ( or even a verbal list- ie: grid of v3 ;or  junction of T6 and r15)      and Ill get started on "catching some waves" for you. "

Here we go ;-)

I have moved my reply to a new thread RCA Signals Waveforms Etc  as I feel that this is an important "reference" undertaking which should be easy for others to find in future.

I tend to access TW far more often than I read my emails - LOL - So the best way to get a quick reply from me about any issues with 'scope or whatever would be to post your questions there - You are welcome to email me - but having everything public means it adds to a sort of "FAQ" for other / future investigators ;-)

Fred.