[sdiy] Minimoog envelope

[jhaible at t-online.de]

After fixing my moog-modular-style adsr, I was curious about that
minimoog envelope effect. Looked at the schemos, but made no
simulation, I think it's all very clear as soon as you know what
to look for:

The Mini has a simplified 2-transistor flipflop (Q1 and Q4). 
This is SET dynamically (AC coupling) from a rising edge of the 
(V-) trigger (Q1), and RESET with hard DC coupling from the envelope
output (Q4). Nevertheless, it's the SET input that is dominating,
not the RESET input. All the charge / discharge switching 
transistors are connected to the collector of Q1, not Q4 !

In normal operation, the SET signal is limited to a short trigger
pulse (C1 and R26 give a time constant of 100us), and when this 
time (corrected by a certain small factor, to be precise - for the
moment let's keep it with the 100us) - when this time is over,
Q4 can easily RESET the flipflop over R3.

BUT - and this is unique to this design - every new trigger pulse
from the keyboard will *shortly* make Q1 conduct again: for another
100us or so. Only *after* this short span of time, the "peak" 
condition of the envelope can be re-evaluated again (and the
flipflop can be RESET if necessary.) But *during* these 100us,
nothing will stop the integrator from ramping up.

If your initial integrator voltage (envelope output) is near
zero when the new trigger arrives, this time is not long enough 
to push the voltage over its nominal peak: you get the normal
ADS operation, Decay starts when peak is reached (and when 
the 100us are long gone).
But if the integrator voltage is closely below the peak voltage
when the trigger arrives, *and* if the "Attack" time constant is
short enough, the integrator can be pushed beyond the peak level
before the flipflop can possibly be reset. (Q4 will conduct
as soon as the nominal peak is reached, but R3 is not "strong"
enough to pull the base of Q1 to GND as long as current comes over
C1.)

I think that's what it's all about. It should be easy to retrofit
this to other envelope circuits as an add-on feature.

I hope this is of some interest.
(I've just read we're not "sharing" ideas on synth-diy anymore ?!?)

JH.


PS.: Anybody with a Moog Modular VCF ready for some tests ?