repl pots with vc

[Haible Juergen]

	>Perhaps my mail wasn't too clear, it is ok for the VCS3 filter,
because
	>it's allready there, ie. a clone can do this, because it's in the
original
	>drawing.
	>
	>OTOH the distortion, even with anti distortion bias network, is
also there,
	>it is so much that you can even see it on the scope. This means,
that adding
	>a FET attenuator to any kind of filter might change the sound. But
the aim was
	>just to add control eg. to the wasp filter clone, and not to change
the
	>sound at all, because this isn't a clone any more then.

Ok, I see. No extra distortion is not - so one should take care not to
introduce
remarkably higher amount or different type of distortion. Now the SEM
filter,
and the WASP filter, too, have resistor dividers to make the OTA distortion
not
that dominant - they have diodes in the feedback path that limit the
amplitude
at high Q or at oscillation, i.e. they're increasing the (damping) feedback
gain
for higher amplitudes. That's contrary to the amplitude limiting behaviour
of an overdriven OTA (and FET, too (?) ), so you still need the diodes in
conjunction with an OTA. OTA should work in linear range, or at least the
OTA's signal compression during overdrive should be weaker that the diodes'
signal exspansion. This should describe the OB-X and WASP method.

Another idea: Make the damping feeddback path a composite path of a
direct (non-inverting) part (simple resistor, without diodes), plus an
overdriven
OTA in inverting configuration. Feedback signal would be resistor signal
*minus* OTA output current then. Resonance would be incresed by
increasing OTA control current (not decreasing like the normal SVF), and the
OTA's overdrive nonlinearity would now indeed limit signal amlitude by
decreasing
Q at high signal levels. Should be more difficult to calculate, but probably
better SNR that the Oberheim solution, as you can drive the OTA harder. And
you can even get self oscillation, when the (negative) OTA feedback
outweights
the (positive) resistor feedback. 
Wild guess: This wasn't how the old EMU SV filters did it, was it ??

(BTW, the elektor Formant SVF has no diodes, but can have the integrator
otas overdriven. Surely a bad choice in the light of the above
considerations,
but it has a nice sound of its own, nevertheless.) 

JH.