Decay caps and resistors - is it always better to use a larger cap?

[digi2t]

Oh Mark... I guess there's something to be learned here, as we try to age "gracefully";

Develop a sense of humour.... or die!

[Mark Hammer]

Me, I'm going to cackle all the way to the grave. 

[PRR]

> D1>---R1--------[variable resistance element]
>               |  |
>               C  R2
>               |_|
>                G
>
> If R1 = 100K and R2 = 100K and C = 1uF

If so, then the output is *half* the input, and the decay time is, in some sense, twice the attack time.

Attack is (R1||R2)*C or 50K*1u or 0.05 Sec or 50mS.

Decay is R2*C or 100K*1U or 0.1 Sec or 100mS.

In many speech/music cases we want short attack long decay. R2>R1. And minimal interaction between the two knobs. Then a preferred circuit is to put R2 "on the left" of R1. Now attack is 100K*1u or 100mS, decay is 200K*1u or 200mS, and you can use a _smaller_ (cheaper!) cap to get back to the desired timings. Also it eventually rises to FULL voltage out of D1 (for any attack/decay ratio), instead of to half voltage (more or less depending on attack/decay ratio).

[Mark Hammer]

The quirk about plucked strings as a signal source is that envelope ripple tends to be maximal a few hundred msec after the string has been plucked, as the string is "settling down".  That's always presented a bit of a conundrum.  On the one hand, if one imposes a long enough decay, via a bigger averaging/"smoothing" cap, the envelope follower will not feel responsive.  On the other hand, if you aim for greater responsiveness, letting the string ring juusssstttt a little too long will get you that raspberry sound from a filter or OTA being modulated at audio rates, that is often misinterpreted or misclassified as "distortion".

Paul's right that in many cases, we want any sweep that occurs to linger just a little, via a longer decay time.  As I've tried to convey here and elsewhere, since perception functions in terms of amount of change per unit of time, having a wide sweep go from one extreme to another (and back again) in a short period of time can be jarring and unpleasant.  Stretching a sweep out a bit makes it more palatable.  Unfortunately, we have to strike a compromise based on what we'd like to hear, and what strings do.

[midwayfair]

Thanks, everyone, for the help, the graphs, and for the formulas. You guys are the best.

Samhay: Every time I try a circuit simulator, I feel dumb and I'm not always sure what I'm looking at. But I have tried to use them.

Mark: Fortunately, the circuit I'm sticking this into is an overdrive pedal, so I'm not too concerned about ripple.