Dr. Q envelope detector question (solved)

[Danich_ivanov]

Hi everyone. Is there a way to change direction in which envelope detector reacts? I need transistor to be closed when there is a signal and open when there is no signal. The idea is that when i'm playing, intesity control of the tremolo should be where it is set, but when i'm not playing it basically goes to the ground, and this detector works perfectly, just the wrong way around. Thanks.

[PRR]

I do not know if this will work.

Add about 100K to keep the transistor "on".

Flip the rectifier so it goes negative, opposing the 100K.

Some experimentation will be needed.

[Keppy]

There are some nuances to flipping the direction, and it sounds like you're trying to graft the detector onto another circuit. It would help to see what that circuit is.

EDIT: Sorry, I posted some messed up steps earlier, not realizing the correct parts of what I said were already in Paul's drawing. Referring to that drawing, though, it looks to me like opamp (+) input and polar cap should connect to V+, not ground. As shown, the polar cap is reverse biased and the opamp has no downward headroom on the output.

[Keppy]

Looking at it even more, I think this won't work without additional mods. The base of the transistor can't rise more than one diode drop above ground, but there are two diode drops (plus ~22k) from the transistor base to the opamp output, meaning that for those two diodes to conduct the opamp would have to produce voltages negative to ground.

That could be altered by moving Paul's 100k bias resistor to the other side of the 22k resistor. That seems like it should work, but might screw up the action of the Range trimmer.

Another option would be to AC couple the opamp to the smoothing cap by adding a 47uF cap between the diode nearest the opamp and the 100R, (-) towards the opamp. This would remove the DC path from the opamp output to the transistor input, allowing for more isolated biasing.

[Danich_ivanov]

Thanks for you'r input. I managed to make it work by pluging in germanium PNP transistor instead of a suggested npn one, and it works marvelously actually. The other problem that i had as with npn as with pnp is a weird ringmod'ish noise that follows the signal, but i suppose it should be pretty easy to solve with some filtering. And just for the record, i eliminated a few things from detector that didn't seem important for my particular scenario, such as first 100k pot, 47k resistor, .05 cap, 25k range trim, and 22k resistor i replaced with 50k (at least for now) pot wired like normal pot, not like variable resistor, also .05 cap made it work in such a way that it pretty much only worked on a low notes.

[Keppy]

Glad you got it working.

weird ringmod'ish noise that follows the signal

Sounds like rectifier bleed into the signal path.

i eliminated ... first 100k pot, 47k resistor, .05 cap

I'm still not sure how you're hooking this up, but inverting opamp stages like this rectifier need some input resistance to function well. Additionally, in the circuit pictured, the components you removed are the only thing isolating the rectifier opamp from the signal input, meaning you might have caused some rectifier bleed into the signal path by removing them.

Or, the bleed might be coming straight through the transistor, or a result of saturation in one opamp affecting the other, or a problem with the smoothing cap, or whatever else. But since this is a known design, I'd first suspect the parts you changed.

[Danich_ivanov]

Yes, you are probably right, i'm also looking forward to add a buffer befor this thing. It is not in it's final form yet, what was important for me is to make it react in the right fashion, all the other stuff i'll add and fix in a process.

[Danich_ivanov]

I added 100k before opamp, and there is no bleed now. Sensitivity is a little worse, but with a buffer i think it will work a bit better.