[sdiy] Moog-type highpass filter

[cheater cheater]

On Thu, Apr 22, 2010 at 18:37, Antti Huovilainen <ajhuovil at cc.hut.fi> wrote:
> On Thu, 22 Apr 2010, cheater cheater wrote:
>
>> yes, but both pre-gain and post-gain are in a single control for ease of
>> use (I know it won't be exactly unity gain, it was just a shorthand).
>
> Should be fairly trivial to do with half of SSM2164 for example, or some
> other linear VCA.
>
>>> Which I have on one of my diy synths... there are separate pots for each
>>> VCO and a master volume after the filter. It was intended for a theremin
>>> so there is ~no~ VCA at all.
>>
>> Correct me if I'm wrong, but wouldn't that just overdrive the input
>> stage? That means for example IC3b on this schematic:
>> http://yusynth.net/Modular/Commun/MOOGVCF/Moogfilter-sch.jpg
>
> No. The input stage _and every following stage_ are driven equally as hard.
> It's the same old equation for each stage (incl. input):
>
> dV/dt*C = Ictrl * (tanh(Vin/2Vt) - tanh(Vc/2Vt))

I meant the input stage as in IC3b on the schematic (input buffer, I
guess), not input stage as in 'first differential pair'. Adjusting
oscillator gain like Harry mentioned would just overdrive that,
wouldn't it? Adjusting the gain of IC3b would make it what we want,
perhaps.

>> I wasn't really even talking about overdrive - just about having
>> enough AC peak-to-peak current so that it can modulate the
>> transistors' resistance in a significant way.
>
> They are the same thing.

I'm not sure how to understand that.

> I find it better to ignore the whole "variable
> emitter resistance" as the differential pair analysis explains it much
> better. Just pick the Vin that gives you whatever amount of distortion you
> like.

if you mean the signal voltage, then yes - that's the idea - change
the amplitude of the input. The general idea stays the same.

And Harry's answer clarifies everything:

> I don't have an IC3b, the audio goes direct into the base of the differential
> pair. In theory you could change the gain of IC3b (lower) and the input resistors
> so that the differential pair is overdriven.
>
> Its a nice sound. Remember this is one of the 'classic' Minimoog ~features~...

:-) Thanks for indulging my curiosity, Harry!

David,
> You seem very interested in this as you've brought it up before, why don't
> you model the ladder part of the filter in LTSpice, and let the group know
> how much signal input voltage you'd need to significantly affect the corner
> frequency set by the control voltage.
>
> It's an interesting question.

I'm not yet good enough with ltspice and electronics to make a model
that I know will reliably mirror real-life performance. I think to get
proper answers I'll have to build a real-life moog filter and measure
the voltages/currents on that, but the possibility of doing that is
currently far away. I keep talking about it to rouse interest and more
importantly not to forget about the idea myself :) On the research
side I'm currently keeping myself busy with additive resynthesis.

D.