[sdiy] SSM2164 VCA

[David G. Dixon]

> You love this IC don't you? :)

I actually love it a lot less since I fried three of them with a stupid
power plug misalignment accident... :(

> I've built a few of the VCA with Irwin lin technique and, sure, they work
well. But how much better (if at all) is this circuit to the LM13700 Mike
Sims

I haven't really looked at the Sims circuit, but I'm trying to get away from
LM13700-based designs...

> Some thoughts I had back when I built the Irwin circuit:
> -would be nice to have a pot to vary from Lin to Log (if possible, would
eliminate full-gain-switch-glitch)

Actually, I'm going to implement Marc Bareille's technique of separating the
functions to different opamps and using an SPDT to switch between their
outputs to the VC pin of the amplifying VCA, rather than what I've done,
which is to combine both onto a single opamp and use a DPDT to switch the
different feedback loops (which leads to the open-gain integrator when the
switch is in transit).

> - cap at the CV input pin is pretty large (1uF on data sheet example).
While this minimizes CV feedthrough 'click' it does so at the expense of
response time.

Those CV filtering caps are unnecessary, and totally undesirable for
sub-audio CV modulation.  I leave them out.

> - related to above; class A or AB operation. AB is for lower THD and/or CV
feedthrough. Data sheet says to leave the pin open for AB.  A 1M resistor
(or there abouts) may be better. Data sheet graph seems to confirm this.

Actually, A is for low THD, and AB is for low CV feedthrough.  A 5k resistor
on the "mode" pin is optimal for low THD, and 560k is optimal for low CV
feedthrough.  I will probably opt for the latter, as the difference in
distortion is probably inaudible, and utterly unimportant for analog synth
signals, which have been thoroughly mangled upstream anyway.

> Just some things to think about...as if you needed any more. :)

Yes.  Too many little hidden trapdoors for the unwary...