Re(2): 2nd generation VCFs (was: Xpander VCF)

[jh.]

I just come from reading, 3:43 AM - you're still awake ?

> > > Right. I would strongly advice you to read the patent behind the SSM
> > > 2044. It is US Pat. 4,404,529. That patent expired in 1995 BTW, so
> > > it's content is out for grabs!
[...]
> The bleedthru may not be covered, but there are many other aspects...

No it isn't. Nor is it in the preceding patent 3.969.682.
Dave R. talks about the need of buffering the final stage, but for reasons
of driving low Z stages, not to prevent bleedthru. For driving low
Z stages, voltage amplification (divider resistors) is not needed. But
as you need extra gain for resonance/oscillation anyway, it's clear that
gain and buffering would be combined in the same stage.

Now the 2044 is rather close to the Moog ladder, where chaining
stages together without intermediate boosting and attenuating is quite
"natural". The same thing built with OTAs is covered in the Rossum /
Oberheim patent 3.969.682, Figure 9. That's what I was refering to
in my initial posting.

I found an interesting sentence in one of these patents about the
possibility of the last stage of several to have a lower cutoff
frequency than the preceding ones, providing the benefit of low
noise by filtering noise from previous stages. But - was this
really used in synth filters ? I'm sceptical, because normally
you try to start with a quadruple pole at no resonance and the
"X"-shaped pole migration with increasing resonance. (And
4-pole filters which don't use that scheme, as EMS or 303,
have a different sound.) And there was some mention about
*that* in the patents as well. So I guess the cap is just different
because of the different gain of the 4th stage, thus resulting
in the *same* pole location as the preceding 3 stages.
I may be wrong here - somebody has to calculate it.

JH.