[sdiy] vdPol oscillator

[Ian Fritz]

At 10:07 AM 11/22/05, M.A. Koot wrote:
>Well it might be a bit out of the context but this seems an interesting
>article anyway so thanks in any case ;)
>I have managed to get it working though by rectifying the signal and to use
>the VCA, some offset some sign-flipping and it's ok.
>I gotta rewrite the schematic first, because it's a mess right now ;)
>But actually the complete vdPol osscilator works now, with a variable
>setting to regulate the feedback (so it's resonance is variable, great for
>bandpassing). It's sounds nice, though the frequency range is somewhat
>limited yet, I think due to some opamp's GBP, but that's a different thing
>to look at.

Good news that you got it to work already!  The way I remember the vdP 
circuit is that it has a slow part and a fast part.  You could see 
slew-rate limiting problems unless you use a decent opamp.

>P.S. Ian, I don't quite understand the zener-thing you are doing in your
>circuit yet, but it certainly looks interesting, is it some way to flip the
>sign during the osscilation?

Yes, exactly.  It's quite easy to understand if you think about the two 
inputs separately.  The signal through the (+) input is just attenuated by 
the opamp circuit, so its part of the total response function is just a 
straight line.  The zeners block the input to the (-) input until they 
break down.  Beyond the breakdown point that signal path gives a negative 
slope.

So draw a picture of a line with a (small) positive slope and one of a 
curve that is zero up to a point then breaks upward into a (large) positive 
slope, and subtract the two.  You will see right away that the result is a 
folder, with a positive slope near the origin and negative slopes beyond 
the zener turn on.

   Ian