how do different wave-forms used as the fixed pitch osc affect the tone?

Has anybody tried using different, more overtone rich wave-forms for the fixed pitch oscillator? or would all of those extra overtones just create a non-harmonic mess?

They have, but maybe not so knowingly.

If you look at the waveforms in the lab pictures in that video about Bob Moog, you'll notice "that ain't no sinewave!". Tube theremins tend to have more nonlinear response than the average transistor ones. But that's not to say that you can't get the same result with transistors.

One of Bob's designs used multiple tuned circuits to select different timbres. Basically, these would select and reject different harmonics on the RF side (before the mixer).

The closer you get to a sinewave, the more whistle-like the sound. The "dirtier" waveforms (typically they look like a rounded-corner sawtooth waveform) sound more like a cello.

Some of the TW experimenters have set about "coloring" the sound on the RF side. Others use effects pedals to affect the post-mixer, audio side.

Hope that answers your question.

Don

yes, thanks! i hope moog releases the plus boards soon so i have a spare board to mess around with! what freq. range does the fixed pitch osc. have to operate within? i am hoping i could just experiment by having a computer spit out the various wave-forms to see if it is worth the time prior to trying to wire in a 'real' oscillator to determine whether the tonal changes of working on the RF side are basically the same as what is already built into the etherwave or whether it is easier to get the same tonal changes by working on the volume side...

i suppose i should also just learn how to play accurately first, hahaha. sigh,

http://www.oldtemecula.com/theremin/ult_antenna/pitchtuner.htm

Go to this site address and download the freeware tuner. This works very well for a "freebie".

You'll just need to connect the audio output of your theremin to your computer's mic input.

If you want to look at the RF side, though, you'll need an oscilloscope. Theremins usually run in the range of the middle to upper range of the AM broadcast band, so you don't need a particularly fast 'scope. Those $200 Velman ones will work fine, or you might be able to find an old analog 'scope at a surplus electronics store.

Once you are "spoiled" with being able to look at every part of the circuit, you become even more impressed by the fact that Lev had no test equipment to use ... or at least none that we ever saw in any of those documentary features.

Don

ack! i didn't know that the pitch oscillators worked at that high of a frequency. i guess having the computer spitting out various waveforms for testing purposes isn't going to work after all... [envisioning dollars walking out of pocket]...

You do not need 'real' oscillators to get an idea of waveforms etc.. This is an area where simulation is quite adequate.

LTSPICE (http://www.linear.com/designtools/software/ltspice.jsp) is an extremely good free simulation package..

One can either use generators to produce the desired HF waveforms for pitch and reference oscillators, or one can 'construct' oscillators - then 'construct' the desired mixer / filters etc.. one can place waveshaping circuits/filters on the HF side, and end up observing the waveform (difference frequency) on the mixer output.. You can directly execute a FFT on this output waveform, and even generate a .wav file from it!

As you know, the output from the mixer is the sum and difference of the inputs - I have found that having a reference oscillator with harmonically rich waveform (switchable ramp,triangle or square) and having a tunable filter / filters on the pitch oscillator, gives huge possibilities for tone control - It is even possible to put a HF 'graphic equaliser' on the pitch oscillator, and have true additive synthesis capability! ... This is one of the ventures which has distracted me from getting my Theremins to market.. When I discovered Bob Moogs idea of filtering before / at the heterodyning mixer, I realised that this idea had awesome untapped potential.. Due to the nature of heterodyning, the filters at the HF end do not need to be 'tight'..

If the pitch oscillator varies from 200kHz to 210kHz, then the 2nd harmonic sits in the range 400kHz to 420kHz, 3rd harmonic sits in range 600kHz to 640kHz... Having filters tuned to these bands allows individual harmonics to be selected, then recombining the bands at ones desired levels, and feeding this with a signal from the reference oscillator, to the mixer, moves all the 'sum' components into a frequency range which can easily be filtered - and one is left with the combined difference components.
I am interested in this technique not only for Theremins.. I believe this principle can be applied to make additive harmonic synthesis feasable for other musical instruments as well.

Fred Mundell
Fundamental Designs Ltd.
Electronics Consultant.
<- See Profile Image for Email.
Designer of Theremins and other alternative electronic music controllers and instruments.

Some more data:
Moog 201 Schematic (http://moogarchives.com/therem03.htm)
The Moog 201 combines both pitch and reference signals in a switch selectable filter (T1,T2 and T3 and related components) - search Levnet for "Moog 201 Harmonics" (07-08/2008) where there is some in-depth discussion on this instrument.

Also, remember that it is the harmonics produced by the harmonic difference frequencies which will appear on the output of the mixer (after filtering everything above say 20kHz out).. So if one has a rich waveform on one input to the mixer, and a pure sine wave on the other input, you will only get a pure sine wave on the output! - You will only get out harmonics which are present on both inputs.. Having one rich waveform on one input, and a waveform with adjustable harmonics on the other input, will allow you to control the harmonic content of the output signal.

interesting! - i always sort of envisioned the heterodyne as a ring modulator but operating in RF, but it seems like the result on a ring modulator is a bit different: on a ring modulator, if one input is a pure sine wave and the other is harmonically rich, the result will be harmonically rich as well... (albeit shifted around a bit in frequency). so does this mean that for, say, the lazy hobbyist such as myself, messing with only the fixed oscillator (the easy one) and not the variable oscillator (the hard one) will not have much of an effect in the end?

[i]i always sort of envisioned the heterodyne as a ring modulator but operating in RF[/i]

[b]It is![/b]

[i] if one input is a pure sine wave and the other is harmonically rich, the result will be harmonically rich as well... [/i]

[b]It is! - But[/b] here is what you are not taking into account.. Heterodyning waveforms results in a SUM + DIFFERENCE output.. If (as is the case when heterodyning audio frequencies in a ring modulator) one uses (outputs) the whole composite waveform, the output waveform will ALWAYS have greater harmonic content than the simple sum of the input harmonics... The SUM component providing the higher harmonics, and the Difference component providing the 'sub' harmonics. Mixing a (say) ramp and pure sine will result in the sum and difference of the sine with EVERY harmonic contained in the ramp.

In Theremin Heterodyning, High frequency (anything > 20kHz) is deliberately filtered out.. So only Difference components are output, and not even every difference frequency is output..
Take a sine @ 100kHz, and a ramp @ 105kHz.. All 'sum' components are > 200kHz.. you will get difference frequencies at 5kHz, 110kHz, 215kHz, 320kHz.... Only the 5kHz will get past the filter!

[i]so does this mean that for, say, the lazy hobbyist such as myself, messing with only the fixed oscillator (the easy one) and not the variable oscillator (the hard one) will not have much of an effect in the end?[/i]

[b] Yes..[/b] - In order to get harmonic variation by adjusting the harmonic content of the reference oscillator (in the above case) you will still need to have all the wanted harmonics available in the variable oscillator waveform - I have found that a selection of ramp, square, triangle and 'original' gives a complete tone 'pallet'.. But this is not too difficult to achieve..

- when waveshaping at high frequency, you do not have the problems one has at audio frequencies with regard to amplitude variation as a function of frequency variation.. at 250kHz, a frequency shift of 10kHz hardly affects the amplitude at all - you can use linear integrators to produce linear ramps/triangle waves.. or any other form of fixed filtering you wish.. The one thing to remember though is that you need damn good op-amps to achieve the required GBW and stability.. The LM6142 is probably the cheapest opamp one can get away with.

[i]"when waveshaping at high frequency, you do not have the problems one has at audio frequencies with regard to amplitude variation as a function of frequency variation"[/i]

This probably needs more explanation..

If one has a (say) 1kHz square wave, and feeds this into a linear integrator which produces a triangle wave going from -2V to +2v, increasing the frequency of the square wave by 1 octave (to 2kHz) will produce a triangle wave output from the integrator going from -1V to +1V..

It can be seen that the amplitude of waveform out of any fixed integrator / filter is directly proportional to the frequency of the waveform input into the integrator / filter..

This problem is overcome if there is a control signal proportional to the input frequency, which changes the integrator / filter time constant / rolloff frequency.. this is the principle used in VCF's - but implementing it in a Theremin involves adding pitch to voltage circuitry etc.

The easier way to overcome the problem is to modify / integrate / filter the waveform at the HF side before the heterodyning mixer..

For example, take a 210kHz reference oscillator, and the variable oscillator which spans 200kHz to 210kHz..

Take the variable oscillator output and using a comparator produce a square wave from it.. Set this oscillator to give an output of say 209.9kHz, and feed this into a linear integrator to produce a triangle waveform of the required amplitude for the following circuit (the mixer).. The output from the mixer will be 100Hz.. Now a decrease of the variable oscillator frequency to 209.8kHz will produce an output from the mixer of 200Hz (a 1 octave increase) but the amplitude of the triangle wave from the integrator will only have dropped a tiny amount.. Even with a full 8 octaves span, the amplitude change on the triangle wave will be less than 5%.

I have been collating notes like the above, and hope to soon publish an E-book ("The Theremin designers cookbook" perhaps) with schematics etc to enable others to experiment and build / modify Theremins 'to taste'.