Xpander VCF

[jh.]

> What I wanted to say is that you can't simulate closed loop behaviour
> in time domain near oscillation.
>
> I think one criterium to be able to integrate via Newton/Raphson
> or other methods is stability of the system.
>
> I know no simulator that is able to do this.  So you
> have to do it open loop.


But on *this* I disagree !

In *time domain* you can simulate an evolving self oscillation.
This is, if you're carefully choosing the initial conditions, and
if your models show enough real life limiting (like opamps
clipping at the rails) to prevent infinite values.
There is this delayed LFO in the Roland keyboard vocoder
for instance. Unlike other such LFOs which use a VCA to fade
it in with an envelope, that Roland device is a sine oscillator
that is switched from damped to under-damped, and which
produces an evolving sine wave. I have sucessfully simulated
that in PSpice. It's not that difficult.

Oscillations due to parastitic effects are a different thing, because
the parasitics are normally not modelled well enough.

JH.