[sdiy] SSM2164 state-variable filter

[Tom Wiltshire]

On 16 Aug 2008, at 12:13, Neil Johnson wrote:

> Hi,
>
> On 15 Aug 2008, at 16:18, Tom Wiltshire wrote:
>> Perhaps it is. Roman Sowa says on his page about his filter that  
>> it'll oscillate right down to 0.3Hz, which mine doesn't. It won't  
>> oscillate without a signal input below about 1.5KHz.
>
> Interesting.  I've got my filter to oscillate over the range 4Hz to  
> 16kHz.  The lower frequency may be due to component tolerances  
> (I've used junk box 220p ceramics for the integrator caps in the  
> prototype!).

Perhaps. I started off with junk box ceramics too, but on Harry's  
suggestion changed them for polystyrene. No change.

I've also disconnected the resonance VCA to see if there is any  
residual damping from VCA not being totally off. There isn't - the  
VCA works fine and grounding its input makes no difference.

Then I started experimenting with the resistor values in the initial  
mixer. How close does the matching of the 2:1 ratio between the input  
and the feedback need to be? I've tried replacing R5 (op-amp gain  
resistor) and R6 (feedback input resistor) with a VR to see what  
difference that makes, and whilst it makes *some* difference,  
lowering the lowest oscillation frequency a bit, it doesn't make much  
difference.

So what could affect the LF performance so drastically? I'm getting  
pretty stuck here.

Thanks,
Tom