A/D + A/R schem available

[Christopher List]

Hi List - 

I've completed and debugged a new module that I've been working on & I thought 
I'd tell you about it.

Full credit for this goes to Juergen who posted the schem for his A/R generator 
a long time ago. 
I based my circuit on this and added some other stuff. 
When some of my changes didn't work, I went back to Juergen for advice and he 
gave me some tips and tricks to try that got it all going.

The basic circuit is a voltage controlled A/R generator based on the LM13600. I 
wanted to maximize the range of control for this so I took Juergen's original 
schem and reduced the value of the resistor that goes to the base of the tranny 
pair for the CV. My goal was to send an even smaller control current to the 
LM13600 allowing me to get even longer rise/fall times out of the circuit with 
the same control voltage. This had the unfortunate side effect of causing the 
envelope generator to lock up when BOTH the attack and the decay were set too 
high. Turns out this was due to current leaking through the diodes while it was 
trying to charge or discharge the cap. Juergen suggested using a FET as a fancy 
diode to prevent this reverse current - and sure enough - it worked! Gate = 
anode, source tied to drain = cathode.

The time range for my circuit is now .1ms to about 50sec for the attack and 
.9ms to about 2minutes for the decay (for some reason decay has always been 
longer than the attack). This is with a .01uF cap. Changing the cap value will 
(obviously) change these rise/fall times accordingly. I don't know what you 
guys think, but I am *VERY* pleased with these numbers. 

What I did next was build a flip-flop around this A/R generator so that now 
there are two inputs: gate and trigger. The gate in will make it work like an 
A/R, while the trigger in makes it work like an A/D. I also added an "end 
pulse" output - like on the Serge Dual Slope Generator. This sends out a pulse 
when the envelope gets back to 0v. This can be tied to the trigger in so that 
once you start it going (with a gate input) it will work like an LFO - with a 
max frequency of about 1KHz. The only odd thing about the way I did this was 
that the reset for the flip-flop uses it's own comparator which has to be set 
just below the peak voltage of the envelope or else it won't reset and the 
thing will lock up in the "high" state.

I have a schem (in .gif format) available to anyone who's interested....
...I'd welcome any suggestions on the design.

- Chris