[i]"It occurs to me that one could pass any simply pitched audio signal through a comparator and use it as data for the "second theremin", along with an envelope follower to track the volume of the signal. Kind of like the opposite of a vocoder." - Gordon[/i]
In theory, yes.. In practice, perhaps..
I am saying (have said) way too much already.. I was working on a heterodyning VCO.. The basic scheme I started with was taking control of the VFO and simply changing its frequency from the CV..
My problem was stability and linearity - the exponentiation required for 1V/8ve combined with the non-linearities of the HF oscillator combined with the tiny % required change in VFO frequency made the design an absolute monster, and the prototype even worse!
Then I got the PLL idea - I thought Eureka!! Just make a standard exponential VCO, feed its output to a phase comparator, use the output from the phase comparator to control the frequency of the VFO in the heterodyning Theremin circuit, Square the output from the analogue heterodyning mixer (actually, I had a seperate D-Latch based digital difference frq detector) and feed this back to the phase comparator, and use the analogue output from the mixer as the audio output..
With this system, one has an entirely analogue heterodyning Theremin - with oscillator interactions changing the waveform, and with the degree of oscillator coupling being fully adjustable - it will never lock up, because the PLL wont allow it to! - The ultimate analogue Theremin!
I also thought about its use with other audio signals as control inputs..
So damn simple - So bl***y obvious..
Alas, low (below about 200Hz) signals screw the system up.. In order to cater for these the phase comparator output needs to be too long a time constant on its filter / integrator.. A long TC slugs the response at higher frequencies - one is limited to about 3 octaves, and performance is better at the HF end..
I have found a fix.. It is not simple, but it is simple when applied to Theremins and specially tailored VCO's.
I have not yet found a way to get the PLL to stay stable for any 'normal' audio input below 200Hz.. and I am not spending time on this at present - I have what I need for the task in hand - I believe that the physics, when related to normal audio frequencies, prevents reliable operation at low frequencies for the same reason that standard pitch to voltage converters fail at low frequencies - It takes 50ms for a single cycle of a 20Hz waveform to be seen, and 50ms is an extremely long time! .. It takes 5ms for a single cycle of 200Hz to be seen - even 5ms is a long time - but it is approaching usable.. In fact, more than one cycle is really required - so 200Hz can be locked in about 10ms.
But, as you said earlier - PLLs occupy volumes of heavy maths etc - When one really starts getting into PLL's, the maths and theory becomes extremely weighty - too weighty for me!
And its not just as simple as multiplying the input frequency by 10 using a PLL! ;-) .. Because the source data for such a PLL would still only update (at maximum) in one cycle..
One can make a fun tracker simply using a PLL (4046) with its VCO tied back to its input, or via a couple of dividers.. I say "fun" because its behaviour at low frequency becomes almost a random tone generator, loosely focussed on the input pitch.. I made quite a few of these back in the early '70s and they sold well - But back then, even Xtal radios using an OA91 sold well ;-helped to pay for all the wonderful electronic junk you could buy - everything having tubes!)
And that is all I am going to say on this subject.. I have already said more than I should, and I know that if the right person read this and spent some time thinking it over and prototyping / experimenting, they will find out the 'fixes' I have worked out.. or come up with different solutions.
Fred.
In theory, yes.. In practice, perhaps..
I am saying (have said) way too much already.. I was working on a heterodyning VCO.. The basic scheme I started with was taking control of the VFO and simply changing its frequency from the CV..
My problem was stability and linearity - the exponentiation required for 1V/8ve combined with the non-linearities of the HF oscillator combined with the tiny % required change in VFO frequency made the design an absolute monster, and the prototype even worse!
Then I got the PLL idea - I thought Eureka!! Just make a standard exponential VCO, feed its output to a phase comparator, use the output from the phase comparator to control the frequency of the VFO in the heterodyning Theremin circuit, Square the output from the analogue heterodyning mixer (actually, I had a seperate D-Latch based digital difference frq detector) and feed this back to the phase comparator, and use the analogue output from the mixer as the audio output..
With this system, one has an entirely analogue heterodyning Theremin - with oscillator interactions changing the waveform, and with the degree of oscillator coupling being fully adjustable - it will never lock up, because the PLL wont allow it to! - The ultimate analogue Theremin!
I also thought about its use with other audio signals as control inputs..
So damn simple - So bl***y obvious..
Alas, low (below about 200Hz) signals screw the system up.. In order to cater for these the phase comparator output needs to be too long a time constant on its filter / integrator.. A long TC slugs the response at higher frequencies - one is limited to about 3 octaves, and performance is better at the HF end..
I have found a fix.. It is not simple, but it is simple when applied to Theremins and specially tailored VCO's.
I have not yet found a way to get the PLL to stay stable for any 'normal' audio input below 200Hz.. and I am not spending time on this at present - I have what I need for the task in hand - I believe that the physics, when related to normal audio frequencies, prevents reliable operation at low frequencies for the same reason that standard pitch to voltage converters fail at low frequencies - It takes 50ms for a single cycle of a 20Hz waveform to be seen, and 50ms is an extremely long time! .. It takes 5ms for a single cycle of 200Hz to be seen - even 5ms is a long time - but it is approaching usable.. In fact, more than one cycle is really required - so 200Hz can be locked in about 10ms.
But, as you said earlier - PLLs occupy volumes of heavy maths etc - When one really starts getting into PLL's, the maths and theory becomes extremely weighty - too weighty for me!
And its not just as simple as multiplying the input frequency by 10 using a PLL! ;-) .. Because the source data for such a PLL would still only update (at maximum) in one cycle..
One can make a fun tracker simply using a PLL (4046) with its VCO tied back to its input, or via a couple of dividers.. I say "fun" because its behaviour at low frequency becomes almost a random tone generator, loosely focussed on the input pitch.. I made quite a few of these back in the early '70s and they sold well - But back then, even Xtal radios using an OA91 sold well ;-helped to pay for all the wonderful electronic junk you could buy - everything having tubes!)
And that is all I am going to say on this subject.. I have already said more than I should, and I know that if the right person read this and spent some time thinking it over and prototyping / experimenting, they will find out the 'fixes' I have worked out.. or come up with different solutions.
Fred.
