[sdiy] SSM2164 state-variable filter

[Oakley Sound]

 > http://www.electricdruid.com/ResonanceGraph.gif

This is quite common in an uncompensated SVF. I still think its phase 
lags in the op-amps and variable gain sections that is causing this. It 
doesn't take much to take the SVF from high-Q to oscillation, so just a 
little change in the overall phase delay may be enough to give this sort 
of graph shape. After all, this is a measure of oscillation level not of 
Q itself. I bet if you measured Q, with the SVF set to be just below 
oscillation at 10kHz, it wouldn't look that different across the audio 
band until higher up.

Also watch out for tracking errors in the two OTA/VCA stages, as well as 
timing cap quality. Using 1% caps usually gives a more uniform self 
oscillation level. Although, I haven't quite got my head around the 
reason why...

Ever wondered why most classic synth SVF designs didn't oscillate.

If you desiging a state variable oscillator you'd probably provide a lot 
less feedback around the loop to ensure it would oscillate over the 
working frequency range. You'd then control the amplitude by a simple 
regulator type circuit, eg. zener or a Vbe multiplier.

Tony