[sdiy] multimode matrix filter question

[David G. Dixon]

> I'd suspect that heat would not be more of an issue than it would be
> with any other filter. If the expo conversion process (I think David is
> using the inherent expo response of the 2164 here) is temperature stable,
> the whole thing should
> track as a single unit. That leaves resistors (maybe 50ppm easy) and
> capacitors...
> you'd want good quality tight tolerance caps anyway.

The filter core is a 2164-VCAs-and-integrators design.  All of the timing
VCAs are on a single quad chip, and they all take exactly the same control
voltage.  There is no temperature compensation, although I have designed the
control circuit such that one could replace a 2.00k metal-film resistor with
a 3300ppm/C tempco if desired.  In any case, all of the VCAs will be at the
same temperature, so any temperature drift will affect all filter stages
equally.  The filter capacitors are 1% 205pF silvered mica.  The resistors
are all 30.1k 1% metal-film, hand-selected to conform to four significant
figures.

> or would you like some Y5U ???   :^)

No thanks!  Even the 100nF secondary power-rail decoupling caps are X7R.
The little 560pF compensation caps on the VCA inputs are C0G, as are any
opamp compensation caps (of which there are very few in the circuit).