Reverb oscillation fixed

[Chuck H]

Ok. If the wife's camera is in order. My phone camera is awful (about as good as my phone). After I get it back together.

EDIT: If I deliver it myself I'll snap some pics of the "Bulldog" and "Satellite" that this guy also has. It's a 4Xel84 head (designed for power tube clipping), a 1x10 reverb combo amp (dedicated reverb that filches drive from a "host" amp) and a 4x10 cab. All stackable and done in antiqued knotty pine. Perhaps my favorite build. This is his dirty amp (though it also does cleans of course) and the Double Agent is his clean amp.

On the road he just uses rented Peavey Classics

[The Dude]

I have to wonder how there can be enough 15khz getting passed that 1M/220pf combination at the grid of V3B. Unless I'm not seeing things correctly, that should filter well below 15K. You might want to verify that the cap is good or try a different one.

[trobbins]

1M is grid leak only, and not part of any low pass filter.

[The Dude]

Edited: OOPS, scratch that. I was comparing some filtering/reverb schematics and was looking at the wrong one. Apologies. You are correct.

Saying that, a LPF at the grid might work to stop the oscillation. Most guitar amps don't need to respond to 15K anyway.

[Chuck H]

Edited: OOPS, scratch that. I was comparing some filtering/reverb schematics and was looking at the wrong one. Apologies. You are correct.

Saying that, a LPF at the grid might work to stop the oscillation. Most guitar amps don't need to respond to 15K anyway.

Indeed it might! But with only a few millivolts from the pan I (and it seems MOST reverb circuit designers) figured any attenuation was too costly in noise. Reverb recovery stages need to do a lot of amplifying of a very small signal. A little extra impedance and you lose signal. A little extra resistance and you increase hiss.

Just thinking out loud now... Wouldn't a "grid leak" stage (ala grounded cathode, 2M or greater grid load) be just about ideal for a reverb recovery? They don't handle much signal and they don't overdrive gracefully, but these conditions would never occur as a reverb recovery stage. The advantage would be higher gain and lower noise. This seems perfect for reverb recovery. Doesn't it? I wonder why we don't see it being done.
EDIT: Ha! It looks like I started a thread on this about a year and a half ago! They say the memory is the first thing to go... I think that's what they say anyhow.

[trobbins]

Diyaudio have had a few threads in input stage noise - the contribution to noise of a grid-stopper up to about 22k is pretty negligible - so yes I would think that would be a useful addition, or even R plus a few slip over ferrite suppression beads. I'd think some roll-off above 10-12kHz would be advantageous. The pickup coil DCR is pretty low I recall.

[Chuck H]

Running some tests I can see that there's no reason to fear filtering the input. You can actually get pretty aggressive with negligible losses. I didn't think this applied to such a tiny input signal, but it does.

[defaced]

Alternatively, would a plate bypass cap be a good alternate method for making a low pass filter without adding additional series resistance?
Plate Bypass Capacitor Calculator

[Chuck H]

Ok... Some things look obvious, but aren't once you actually analyze them (thank goodness for SPICE type programs). At the location where I have the 220p cap on the recovery input it's only rolling off .001dB at 15k or something. In order to get even a 3dB roll off there the cap would need to be bumped to .047uf!! Really?!? That's what my simulation shows anyway.

I used LTspice with a 4mV voltage at 280ohms parasitic series resistance (the coil resistance) and then just the 1M load and parallel cap. Those were my results. Surprised.