[sdiy] More elegant trigger derivation?

[Gene Stopp]

This is what I have done with good results:

Gate generated by the usual method (whatever that may be).

(I typically use a one-shot that is triggered by the key-down mux signal
from the digital keyboard interface. One key down, one pulse per scan, many
keys down, many pulses per scan. One-shot is retriggerable with a time
constant slightly longer than one keyboard scan.)

CV is differentiated by two op-amp circuit (hi-gain amp followed by
differentiator). Output is positive pulse for CV step up, negative pulse for
CV step down. This goes into comparator that has threshold above ground, so
only positive pulses trigger comparator. Output of comparator is negative
during positive pulse, open collector, tied to gate so the gate gets dragged
to ground momentarily on trigger.

This does not provide a "true" trigger, but rather puts "glitches" in the
gate! Works like a charm, especially on Minimoogs etc. that don't have a
pulse trigger to restart the EG attacks in the first place.

I usually use a dual op-amp (nothing special) and an LM339 quad comparator.
The LM339/393 will tolerate the negative pulses even if powered with V+ and
ground, otherwise you can diode clamp the differentiator output.

Theoretically this is a race-condition no-no since the gate will supress its
own rising edge for the duration of the comparator groung-clamping, and a
teeny-tiny rising edge may sneak out depending on circuit parameters. Never
seen this, always works, nothing that an extra cheater-cap won't fix! :)


- Gene


-----Original Message-----
From: sbernardi at attbi.com [mailto:sbernardi at attbi.com]
Sent: Thursday, September 19, 2002 12:49 PM
To: synth-diy at dropmix.xs4all.nl
Subject: [sdiy] More elegant trigger derivation?


I've been thinking about Harry's idea of deriving a 
trigger by differentiating the control voltage. So, I 
run the CV into a differentiator, which outputs a 
positive or negaitive going pulse depending on whether 
the the CV went up or down. Then this goes into a window 
comparator, because I don't care what the polarity is, I 
just want to indicate that it has changed. Say we set 
the window threshholds at +/- 100mV. 
I also want to generate a trigger when the gate goes 
high regardless of whether the CV changes (i.e., you 
could press the same note multiple times). So I run the 
gate into a positive edge detector and OR it with the 
window comparator output. Then I use this to trigger a 
oneshot such as a 555 to generate the actual trigger.
This can all be done with an op amp, a dual comparator, 
some logic gates, and a 555. Seems like a lot of parts - 
is there a simpler, more elegant way to do this?