How do I distort thee, let me count the ways ...

Perhaps (please excuse the fact that I'm not an English native speaker) I should rather have said "differential pair" instead of "push-pull".

Q1 and Q2 form a simple differential pair amplifier such as the input stage of every op-amp. The base of Q1 is the inverting and the base of Q2 the non-inverting input. The advantage of such a differential amplifier is higher input impedance, lower input capacitance and higher open-loop gain than a single transistor stage.

Since the inverting input is grounded via R1, we have a simple non-inverting amplifier with a frequency selective network at the output (L5 + L12 parallel C1) which has its highest impedance and no phase shift at the resonant frequency, so that the amplifier stage has its highest voltage gain and lowest phase shift. C3 does DC decoupling and R3/R4 apply 1/48 V_out as positive feedback. As long as the gain of this two transistor stage exceeds 48, the Barkhausen criteria for oscillation are satisfied: Loop gain >= 1 and phase shift = 0.

Voilà!

*sing* It's soo eaaaasy... */sing*

"*sing* It's soo eaaaasy... */sing*"

Thierry - You have the knack of explaining things clearly and simply, and manage to do this without language being a problem! Congratulations!

Thank you, Fred!

I feel flattered... :-)

I stated before:
[i]Let a1...an be the harmonic coeficients of the fixed osc signal and b1...bn those of the variable osc. After the mixer you'll find the scalar product r1...rn = a1xb1...anxbn.[/i]

As I do never trust myself, I verified this. It's correct, [b]but only under specific conditions[/b]:

The two heterodyning frequencies at the mixer input must be higher than the audible frequency range. Their difference must be in the audible range. The above simplified "formula" represents only the mixer output products which are in the audible range. So it applies perfectly to theremin design.

It does [b]not[/b] apply in other cases, i.e. when ring-modulating two audio signals. In such cases the results will be much more complex.

Hi guys!
Thanks thierry for the explanation of the oscillator. It has helped me a lot! Now, I´m fighting with the multiplier, and I will tell what is happening...
One question... does anyone know something about the digital theremin of Glasgow? because one friend is doing it and he has some doubts about the output of the mixer which varies between 0 to 15kHz, and the question is why he has to use the function generator and the tacometer after the mixer if the signal is already good?
The problem is the output signal after the filter which is after the tacometer, it is a continuous signal as it´s supposed, but without the expected amplitude.
If somebody knows something....
I will tell you my improvements with my Theremin

@esther:

If you have a specific question concerning a specific theremin, please post always a link towards the schematics and/or the description, so that people who spend their rare spare time helping others here don't waste time by "googling".

This is the 26th posting, so this thread will automatically be closed and we have to open this one (http://www.thereminworld.com/forum.asp?cmd=p&T=3925&F=1).

Schematic of 'Glasgow Digital Theremin' (http://www.physics.gla.ac.uk/~kskeldon/PubSci/exhibits/E9/cir2.gif)

The 'Glasgow' 'digital' Theremin consists of a crude pitch oscillator (Schmidt gate with resistor from output to input, which charges/discharges antenna capacitance) and an identical reference oscillator charging/discharging a fixed capacitance.. These 2 oscillators drive a 4070 Exclusive OR gate, which acts as a digital mixer. The digital output pulses from the 4070, when integrated (filtered) produce a difference frequency result (ie audio) which could / can be taken directly as an audio output.

In the Glasgow Digital Theremin, the 'audio' is instead taken to a pulse generator circuit (4098 monostable) which generates one fixed width pulse for every input cycle - these pulses are integrated to produce a voltage proportional to input frequency, and this voltage (after going via an opamp circuit to adjust biasing + gain)drives the 8038 VCO.

>>"the question is why he has to use the function generator and the tacometer after the mixer if the signal is already good?"

The main reason would be to have a choice of output waveforms, and also be able to select tuning and frequency range independently from the HF oscillator difference frequencies.

>>"The problem is the output signal after the filter which is after the tacometer, it is a continuous signal as it´s supposed, but without the expected amplitude."

Two main things to look at here - Check the output pulses from the 4098.. The width of these pulses should probably be about 66us which would give nearly continuous 'high' state output with an input of 15kHz.. at 1kHz input, this would give a 66us pulse every 1ms .. integrating these pulses should give an output swing from near 0 to near maximum supply.. (measure this at point A).. Then check Pin 8 of the 8038 and adjust the trimmers to get the required control range.

thread continues here (http://www.thereminworld.com/forum.asp?cmd=p&T=3925&F=1)

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