[sdiy] SVF question

[David G Dixon]

> You can set it up with enough feedback from the lowpass 
> output that the filter will oscillate, but with enough 
> damping from the bandpass output that the oscillation won't 
> get driven into clipping. Since the diodes set the point at 
> which that extra damping kicks in, they're a key part in 
> this, but the weights on the feedback paths are also crucial.
> 
> The same problem came up some time ago with SSM2164-based SVF 
> designs. It's tricky to balance it just right.

Yes.  I think we came to the conclusion that there were a couple of good
places to put the diodes.  I settled on putting them across the first filter
cap (the one which gives the BP output, and which generates the feedback for
resonance), with a tiny bit of the BP signal shunted around the resonance
amp network through a large (1M or so) and back to the input to ensure
robust self-oscillation.  At self-oscillation, simulation gives sines from
the HP and LP outputs with 0.042% THD, and from the BP output with 0.144%
THD.  I haven't built this filter yet -- haven't had the time -- but I think
it's going to be a good one.