[sdiy] SSM2164 state-variable filter

[Ian Fritz]

At 07:45 AM 8/16/2008, Tom Wiltshire wrote:
>So what could affect the LF performance so drastically? I'm getting
>pretty stuck here.

I had a quick look at some *approximate* math, with the limiting stage in 
two positions:  Roman's original (LP clipping) and Harry's suggestion 
(resonance clipping).  My assumption was that the clipping stage puts out a 
constant amplitude sine wave and that there is no input excitation.  The 
results are remarkably different.

For resonance clipping the output voltage has a resonant denominator, 
diverging at the expected frequency.  The LP clipping, OTOH, has a 
non-resonant denominator.  The formulas are not the same as what is 
observed, but the calculation is obviously oversimplied.

At any rate, the answer looks to be what Harry said -- the clipping is in 
the wrong place.  I suggest trying back-to-back zeners across the the (+) 
input of the summer.

   Ian