[sdiy] 2164 Korgasmatron VCF

[David G. Dixon]

> > In simulation, if I feed the input into both the HP and LP destinations
> on a
> > single filter, I do indeed get a kind of gentle notch.  The minimum
> > amplitude at the cutoff frequency (with zero resonance) is 6db (i.e.,
> 50% of
> > the input amplitude).  It's not a sharp notch like one would expect from
> a
> > proper notch filter, but more of a smooth dip in (log) amplitude at the
> > cutoff frequency.  Is that going to sound like much?
> 
> I think the phase response is much more important than the amplitude
> response here! You have not created a notch, what you have is an
> all-pass.

Yes, it is close to an all-pass.  Mathematically, it's a series combination
of a 12dB bandpass and a 12dB highpass.

> >> I don't know if you posted any details before... but what is your
> >> "dial in a zener voltage" method. Is it just adding series resistance
> >> or something more complex ?
> >
> > Yeah, basically.  In this case, the zeners are connected on one end to
> the
> > negative input of the opamp, and on the other end to the centre tap of a
> > 100k pot.  One end tap of the pot is connected to the opamp output, and
> the
> > other to a 30k resistor which is grounded.  That gives me full zener
> range
> > from 3.3V to the rail.  The pot can be a trimmer as well.
> 
> You mean the wiper?

Yes.  I meant the wiper.  I'm not always "down" with the "lingo".

> > One thing, though: I can't seem to reproduce the "formant" sound with
> the
> > notch setting at high resonance like that soundclip on the MOTM 420
> webpage.
> > All the other clips I think I can emulate.  I'd like to know exactly
> what
> > was done to get that sound, to see if my filter can do it as well.  I
> think
> > they might be using two filters.  When I get the other half wired up to
> the
> > panel, then I should be able to get all kinds of interesting timbres
> with
> > both filters in series.  However, I won't be able to do that until I
> return
> > from Saskatoon.
> 
> From what I understand, to create a notch you need to use separate
> cutoffs for the LP and HP.. dependin on the spread, you can get
> formant style sounds too (about mid resonance, and you should use a
> nice difference of maybe one octave or two. the LP band should also
> not track as much as the higher band does. for best results the high
> band is a BP)

I've got it sorted now.  It requires two filters to get proper formants
(although you can squeeze something like formants out of a single filter,
with some doing).  You can hear some of the results in my soundclips on
Muffwiggler.

> > However, I'll definitely want to normal filter A's CV
> > inputs to filter B, and also normal filter A's output to one of filter
> B's
> > inputs.
> 
> The typical topology with two filters that gives the best results is
> LP->HP with a soft clipper inbetween. This is because the LP output
> still maintains most of the amplitude and can give the clipper a go.
> If you use LP as the second filter, it'll sound weedier. Of course if
> you normal to HP on the second one you can still have BP and HP in the
> first one which is nice too. Some settings of high pass can give nice
> big amplitude to (e.g. slightly self-resonant HP at a low freq
> setting)
> 
> BTW, can you send me the final schems you have used? :) Thanks..

Sure.

> I'll make sure to listen to the demos once I'm at home :)
> 
> Really good going, keep it up!

Thanks.

> > I've finally got some soundclips of the Korgasmatron up on Muffwiggler.
> > Three of them also show off my redesigned CGS analog shift register.  Go
> > here:
> >
> > http://www.muffwiggler.com/forum/viewtopic.php?t=21058
> >
> > If you can't access the soundclip files, let me know and I'll email them
> to
> > you privately.