Subbing C4 in Yves' ADSR.

[reve]

Hello yu experts!

So for my fourth Yves module, I thought I'd try the ADSR. I've purchased boards from Scott in the past, but I went ahead and etched three
ADSRs myself -- two of which turned out very well. I looked at the parts list, made sure I had everything, and went ahead and put everything together. All went well until I got to C4 and said "oh... _tantalum_ capacitor."

Since the nearest tantalum capacitor is about twenty miles away and I had electrolytics on hand, I figured I'd just throw an electro in and see what happened.

The answer is that the EG _kind of_ works. It seems to work at the outset, but instead of sustaining it keeps (veeerryy slloooowlly) decaying until you release the gate. And even then it never quite reaches 0v -- it hovers around 150mV. Plus something seems a bit wonky with the decay.

Anyway, I'm not very bright but I know:

1. Yves specified tant.
2. Electrolytics leak voltage, and my voltage is leaky.
3. My problems occur around the things C4 is involved with.

But for my own sake as well as the sake of everyone else who might in the future wonder if they can sub an electrolytic, I figured I'd ask here and see if my symptoms are consistent with what I should be experiencing after subbing C4 for an electrolytic. I.e. should I get in the car tomorrow and drive to go by a tant cap, or should I be spending my time looking for other problems with the circuit.

Thanks and cheers!

- reve

[yusynth]

You listed the three points right

My choice of a tantalum was guided by this leakage problem.
However the 150mV resisdual is so to speak "normal" and is also occuring with a tantalum. The problem comes from the fact that C4 has to discharge thru a diode and a transistor, and because of this it only reaches the full discharge after a long while. This is a common problem for all the ADSR based on a similar design.

How to avoid this ?
A possible solution is to connect the emiter of Q2 to a slightly negative reference voltage (-0.2V for example) or not using a FET AOP as an output buffer (thus the cap will also discharge thru the AOP input). In general, this 150mV bias is not considered as a real problem because when driving a VCF this is not relevant and when used with a VCA it is simply a matter of trimming the opening level of the VCA to a higher threshold.

I did try using schottky diodes instead of mere silicon diodes, it is very efficient to remove this bias but unfortunately it creates other more annoying misbehaviour.

[reve]

Just to report back, using a tantalum cap (as Yves specified) worked perfectly. Thanks to Yves for giving us such a great circuit.

[yusynth]

[Jaba]

[yusynth]

I noted that when using shottky at long time duration the curve would somehow reverse, that is it acts normally during ATTACK and DECAY but by the END of DECAY is goes up instead of going down to the SUSTAIN LEVEL.

It was weird but acted the same with all the BAT41 I tested.

[Jaba]

ok, thank you for sharing your experiences with Schottky diodes,
it will save me a lot of time

I guess the reason for those BAT41s was to keep low the voltage drop on diodes; has anyone tried with Germanium diodes ? is it worth doing ?

another idea, probably better: put another diode (the usual 1N4148, silicon) between the positive side of the tantalum capacitor and pin5 of U2 (pos. input); this should compensate the unwanted effects of voltage drop in diodes D2 and D4; you may even match the voltage drops of D2, D4 and the new diode to the desired accuracy for best performance, if required.
I have not tested it yet, but I will soon, and then tell you the results

... hmmmm, but we would still have a small voltage across Q2
... replace Q2 with a FET ?!?
I don't know, just thinking out loud

[yusynth]