[sdiy] MS 20 FILTER

[Tim Stinchcombe]

Hi everyone,
 
> > A note about the circuit you are referring to-I have found that it
> > clips very ugly, the MS20 filter I think is supposed to sound nice  
> > clipping... I believe that's because Tim E's implementation 
> is  designed 
> > for 9v (pedal use) and makes some sacrifices.  
> 
> The diode clipper is missing, that explains the clipping. 
> Also there is 
> no buffering at all for the resonance-feedback signal.

In the Tim E. circuit, since there is only attenuation in the feedback path,
to get the filter to oscillate requires the 2-to-1 ratio of the 220k and
100k resistors at the OTA inputs. An effect of this is to give the overall
filter gain as 2.2^2=4.84. With only 9V to play with, minus about 1.5V at
each rail, this leaves about 6V of swing for the signal - thus if the input
signal is more than about 6/4.84=1.2V peak-to-peak, I suspect one or other
of the OTAs will distort, and this will be made even worse if Vr is not
exactly mid-rail (and it's probably likely the OTAs are not in their linear
zone either, I've not really checked...).

A simple way to check this would be to reduce the 220k resistors (and the
amount of available resonance will be reduced too). If the 220k's are made
100k, to put the resonance back up would then require placing approx 2x gain
in the feedback loop, which is precisely what Rene's circuit does (there the
6x or so is more than enough).

Coincidentally I am in the process of analysing/'deconstructing' the
MS10/20/Korg35 filter - there are a few bits I hope to clear up over the
next few weeks, and then I shall write up my findings in a small paper. I
also intend to build three filters: a (proper) transistor version of the
MS10/20; one of these OTA-based 'clones' (which, to my mind, are not
Sallen-Key filters); and (if I can do it!) an OTA-based version which *is*
Sallen-Key, but which also has the Korg quirks, like the asymmetric
resonance (which appears to be caused by the _signal_ impacting on the
cut-off frequency). I also intend to make the diode limiting switchable, as
I strongly suspect that much of the quality of the sound depends on this,
rather than the filter topology per se.

Tim
__________________________________________________________
Tim Stinchcombe 

Cheltenham, Glos, UK
email: tim102 at tstinchcombe.freeserve.co.uk