interlock (was:Re: ASM-1 oscillators)

[Haible_Juergen#Tel2743]

Joachim:

>I was wondering, with these fast-reset sawtooth VCOs, are there any
>hints of VCO interference (interlocking) at all? How does local supply
>decoupling affect the performance in this context?

Tom:

>Correct me if I'm wrong, but it seems the oscillator reset would cause
>a large current spike in a very small loop which does not include the
>power supply, just the capacitor and whatever is discharging it.  The
>current available from the supply is limited by the current source
>controlling the cap's charging current, right?

Gene:

> The VCO's are next to each other on the circuit board, but as far as I
>     can tell there is *no* cross-influence between them at all. I can tune 

>     them to an identical (zero-beat) frequency and they exhibit the very
>     slow evolving timbre sound that is characteristic of two very close
>     oscillators. You can hear the second harmonic drift in and "take over" 

>     and fade out again which is a good indication of un-synchronized
>     sawtooth mixing.

Good design, Gene !

Even the 3340s, which don't have this fast discharge at all, are known to
be sensible to interlock. At least that's what Doug Curtis says in
Sythesource. He says that the 3340's can be soft synced by spikes
that come in thru the 1.5meg resistors for bias current setting.
I wonder if the rather low impedance resistor divider in the SEM
modules has some advantage here - could be possible at least!
(I think that they used the 15k divider to limit the voltage across the
array, but avoiding interlock might be a good side effect!)
Doug Curtis pointed out that a proper layout and  good decoupling
are important to avoid syncing of 3340's below a beating cycle of
some seconds.
Now this might be different in later OB-synths, and might be another
jigsaw piece in the quest of seeking reasons for sound differences
in OB-Xa and OB-8.
Well, I said "might", not must. I don't want to create myths here (;->).
But I wonder if anybody has checked differnt synths in regards of this!
Did anybody do some mesurements of this effect on commercial
synths? (besides MS-20, where it's obvious, and Paia Fatman, where it's
well reported and where there is a workaround published)

Thanks for any hints,

JH.