Xpander VCF

[jh.]

> I tried a spice simulation of this, using several models for the
> 4-pole filter as a basis (including the JH 2040 clone). I did not
> note where self oscillation occurred.

Finding a self oscillation point is something that's extremely
difficult in Spice. When you're running an AC simulation, you
can see a resonant peak rising with increasing feedback gain, but
you will never see an "infinite" peak which would be self oscillation.
To make things worse, if you're further increasing feedback gain,
the simulated resonance peak will get smaller again !
You can only know if you're in the valid range if you're making a lot of
AC runs with parameter sweep and see where the peak will stop
increasing and apparently start decreasing again.

It's good to have some theorethical calculations to estimate
the self oscillation point, and then just verify with Spice.

JH.