[sdiy] Filter problems

[Neil Johnson]

I wrote:
>> You certainly don't want them anywhere near the integrator
>> capacitors otherwise you'll end up with the poles not aligned
>> due to the diode capacitance slugging one of the integrators.

David G Dixon wrote:
> We put zeners across the first integrator cap in our SVF design, and it
> works great.  Gives very stable oscillation at the critical damping point,
> and doesn't color the filter's sound very much in filtering mode.  The only
> issue is that the self-oscillatory sine waves do not have a constant
> amplitude with changing frequency, but this is a pretty minor problem.
> Cascaded-stage filters oscillate in a much simpler fashion in my experience.

Zener diodes have a measurable capacitance of about 100pF.  Two in
series gives around 50pF.  If we consider the schematic that Tim
posted:

https://www.dropbox.com/s/acjvwb17nmuok1n/PrismFilter.pdf?dl=0

then the first integrator has C = 270pF (220pF capacitor plus 50pF in
parallel from the two series-connected zener diodes) and the second
has C = 220pF.
It will sound slightly different from a filter that keeps the two
integrators more closely aligned (subject to capacitor tolerance).

Neil
-- 
http://www.njohnson.co.uk