Tempco measurements on ASM-1

[jh]

>From what I recall, this is because of how the ASM-1 expo
>converter is dimensioned. The scaling drift will be virtually
>zero when the collector currents of the two transistors are
>equal, i.e. 15/150k = 100 uA. This yields an oscillation
>frequency of approx. 9 kHz, which is pretty high, IMO. The
>further you move from this frequency, the greater the error
>without compensation, as your measurements proved.

You're perfectly right (as always). Too stupid that I never thought
about that before ! my own design goal often was to make the tuning
procedure as easy as possible, so my idea was setting the
0V point of the transistor pair to the lowest note on the keyboard
(CV = 0V); then the frequency offset and scale trimmers would
work independently. Thank God that I never actually built it that
way, 'cause with this method I would have chosed the worst possible
point in terms of drift !
So if I get you right, we should put the 0V point somewere in the range
of the highest note we would normally play on the keyboard, as
the beat rate decreases with lower frequencies, but not much higher than
that point. (?)

JH. (still learning)