[sdiy] 2164 gain response variability

[David G. Dixon]

Well, I finally completed two identical 2164 Expo VCOs with all the little
improvements, and I can now say that there is indeed a little bit of
variability in the gain response of different V2164 chips.

Also, the gain response is not perfectly linear, but typically dips a little
bit at 0V.  In fact, the tracking response looks a bit like a seagull in
flight.  Hence, setting 1000Hz as the reference frequency at 0V CV and
adjusting to exactly 125Hz at -3V, then 250 and 500Hz are a little bit high
(0.2 to 0.5Hz), and 62.5Hz is a tad low (0.2Hz).  Likewise, pulling up to
exactly 4000Hz at 2V with HF trim, then 2000Hz may be 1 or 2Hz high, and
8000Hz may be a few Hz low.  Tuning tends to be a bit of a compromise
between the low and high ranges.

In any case, if two reasonably similar chips are selected, then the two VCOs
can be tuned almost identically up to the top of the useful range, and they
are pretty stable.  In fact, my two VCOs sound absolutely stunning together!
If unmatched chips are selected, then the frequencies might differ by 2 or 3
Hz in the 1000 to 4000Hz range when the lower frequencies are optimally
tuned.

Also, the occasional 2164 chip even has an "upside-down seagull" response.
I'll try to reserve these for use in VCFs or VCAs.