At 11:17 AM -0500 03/12/01, Tkacs, Ken wrote:
>
>You might want to consider making use of the "falling-edge detector."
I want to detect both falling and rising edges. I looked at a few CMOS/TTL
approaches, but most of the flip-flop circuits I found were looking for
much larger edges.
>I don't think you want to full-wave rectify the pulses you get; rather
>use two half-wave rectifiers, one to "peel off" the positive triggers and
>one the negative. Then you can have a separate jack with the falling edge
>>triggers. You can combine the positive and negative triggers to a third
>jack for the 'absolute value' triggers as well. I can imagine some neat
>>effects possible by having a choice of up/down/both triggers.
While that's a very creative idea, I want to keep this circuit as simple as
possible as I'll be building it on perfboard. Extra jacks also cost money
and take up space (which costs even more money). Also, a diode drop (.6V)
is likely to be more than the output of the differentiator, and I have no
idea how to make an "active half-wave rectifier".
I'm hoping I can find a comparator that accepts both positive and negative
going voltages so that I can eliminate the absolute value circuit entirely.
>I'm no EE either, so I can't answer your technical questions. I've dabbled
>(on paper) with things like this myself, based on examples in Forest Mims
>books, but haven't actually tinkered anything together.
>
>This may be the wrong approach, but... if ∗I∗ were going to make such a
>differentiator, I would probably start with a simple lag-integrator circuit
>(cheaper synths have a one-op-amp integrator, low parts count) and use the
>same parts rearranged into a differentiator configuration. I ∗think∗ you
>>get essentially the same kind of specs, just a reversed filter response.
>That at least puts you in the ball park of the range you want to filter,
>>since you're creating a kind of "anti-lag" circuit.
Right, a differentiator is the opposite of an integrator -- the type of
circuit typically used as a lag processor.
>I'm sure a real engineer would do some serious eye-rolling over the above,
>but when you're tinkering... you do strange things.
Perhaps a real engineer like Paul or JH might chime in on this thread :)
>Here are some more thoughts.
>
>Since you're using a differentiator to detect rapid voltage changes,
>you are only going to extract _triggers_ from your input. With a handful
>of extra parts, you can create a gate output as well.
>
>There are ways of using a 555 timer chip as a 'pulse width' circuit. You
>could have the triggers 'trip' the 555, which will then stay high for a
>>time and then fall a specified period later, which can be set with a
>potentiometer. This would give you gates with the same 'on-time.'
>The 555 is used as the basis of a lot of gate delay circuits.
Yes, I have already built a 555 circuit in a project box that converts a
short trigger into a gate so I can use a drum machine trigger to play
various monosynths. It has a knob to control the length of the gate, and
both Roland and S-trigger outputs to control the SH-101 and Mini Moog
respectively.
I have also added a passive "pulse stretcher" (a cap which discharges
through a resistor) to an S-trigger converter so I can use the trigger outs
from my TR-808 to flip the A/B switch on the Lexicon Vortex. The regular
(non-stretched) gate to S-trigger converter I use to sequence the Mini Moog
did not work with the Vortex because the 808 trigger was too short for the
Vortex to recognize it (I presume this is because the Vortex pedal inputs
are debounced).
I'm hoping I will find a comparator circuit whose output is held high long
enough. Horowitz & Hill discuss the use of a Schmitt trigger by adding
positive feedback to a comparator. I don't know how this would apply to an
input that is essentially a spike whose trigger point is at ground.
Anyway, if it is long enough to trigger the motm EG and reset the motm VC
LFO, I should be fine.
Thank you for your comments.