Distorting reverb weirdness

knuckleberryfinn

Just looking for a bit of general advice before I start tearing components out of my amp.  The Epiphone Futura amp I recently resurrected sounds bloody marvellous - until it gets hot.  After about 20 minutes or so, the reverb starts getting grainy and distorted.  What would be the most likely culprit?

ICBM

Something in the reverb drive valve circuit most likely - bias resistor or cathode bypass cap, hopefully. (Rather than the transformer.)

Assuming this is a transformer-driven low-impedance (Fender-style) reverb and not a high-impedance 'crystal' (Danelectro-style) reverb...

knuckleberryfinn

Yes, it's transformer driven, and the whole circuit is pretty similar to a Fender circuit.  I've already clipped the bypass cap on the reverb return which improved matters.  Maybe I'll fiddle with the bias resistor values.  Odd that it works fine until the valves are up to their highest temp - do they draw more current at higher temperatures or something? 

ICBM

knuckleberryfinn said:

Yes, it's transformer driven, and the whole circuit is pretty similar to a Fender circuit.  I've already clipped the bypass cap on the reverb return which improved matters.  Maybe I'll fiddle with the bias resistor values.  Odd that it works fine until the valves are up to their highest temp - do they draw more current at higher temperatures or something? 

Not really - or not a sudden change unless the valve is faulty, anyway. If you've tried swapping the valve already, my guess would be the cap on the drive valve. It could also be the plate resistor on the return, sometimes those go very high. Sometimes this sort of thing only happens when the component is hot.

knuckleberryfinn

There's no cap on the drive side, so I reckon it's the plate resistors.  They are all original, so 50 years old.  I don't want to open a can of worms, but should I use carbon comp here?

knuckleberryfinn

Judging by this article, it seems that this is the place to use carbon comp.

http://www.geofex.com/article_folders/carbon_comp/carboncomp.htm

ICBM

Don't bother in a reverb circuit.

knuckleberryfinn

Ah yes, good point. 

jpfamps

Sounds heat related.

I'd investigate the cathode resistor on the drive circuit.

The 12AU7 driver draws quite a lot of current and the cathode resistor will get hot.

You can try a resistor with a higher power rating and also one with a higher value to reduce current.

knuckleberryfinn

Thanks JPF, I'll do that too.

knuckleberryfinn

I'm still having problems with this bloody reverb circuit.  I've replaced the single 1K cathode bias resistor on the reverb driver with two 2200R resistors in parallel to bias it slightly cooler and to share the current between the two resistors.  I had already got rid of the bias bypass cap on the reverb return to lower distortion.  It worked fine for a bit - I even successfully gigged with it - but a week or two ago something went pop, and the volume dropped dramatically.  Now, the volume has returned to normal, but the reverb is weak and distorted.  Also, when I turn the reverb right up, it distorts massively, and then there's no reverb at all.  But if I then turn all controls to zero and turn the reverb back up, it usually returns!  What the hell is going on?  Does this sound like a dodgy cap which needs the controls zeroed in order to discharge?  I'm really confused.  Here's the schematic...

http://www.gibson.com/Files/schematics/EA-12RVT.jpg

ICBM

Could be. Do you mean it needs the channel volume zeroing before the reverb will come back? I also notice the reverb level is a 'dwell' not a 'mix' (ie input level not output) and operates in the same part of the circuit as the channel volume and tone.

My first guess would be the .02uF coupling cap from V1B to the top of the reverb level pot. If that's putting DC onto the grids of V2A and V3B it might do what you're describing - V2 would be more affected since it's directly connected when the reverb is up full.

knuckleberryfinn

It was a 500pF running from the reverb pot to the grid of V2A - I'd added it to try to stop lows and mids going through the reverb circuit in attempt to stop it distorting.  The cap must have gone bad when I cranked the amp, though the cap was rated at 500V.  Cheers for the pointer @ICBM.  The whole amp still doesn't sound right.  It's quieter than I would expect and it distorts early and not particularly smoothly.  I think it's V4B: the voltage across the cathode bias resistor is over double what it should be at 3.2V rather than 1.4V.  This would mean a low current, which would account for the low output and early distortion wouldn't it?   I've changed the anode resistor, I've checked all the resistors in that part of the circuit, and I've changed the valve.  Nothing seems to be bringing that voltage down.  Any ideas?

ICBM

knuckleberryfinn said:

The whole amp still doesn't sound right.  It's quieter than I would expect and it distorts early and not particularly smoothly.  I think it's V4B: the voltage across the cathode bias resistor is over double what it should be at 3.2V rather than 1.4V.  This would mean a low current, which would account for the low output and early distortion wouldn't it?

No, it indicates a very *high* current - more current develops more voltage. A slight DC leak through one of the two coupling caps feeding it would account for it, because it will be trying to forward-bias the valve. That would definitely cause early/bad distortion too.

knuckleberryfinn

Okay, I'll investigate those two caps when I can bring myself to open up the chassis again.  So bored of fiddling with this amp.  Thanks again.

ICBM

Just to make you feel better, I have an old FAL 100W solid-state PA amp here that I was given which has a noise problem in the preamp. I've already spent far more hourly rate on it than the damn thing is worth and I still can't find it - I've changed just about everything in it apart from a couple of old transistors which have no visible codes left, and which actually meter OK anyway. I should have just binned it in the first place but stupidly I didn't start to investigate the noise until *after* I'd already rebuilt the mains wiring to make it safe, so now I am at the point of throwing good money (or at least time) after bad...

knuckleberryfinn

That does actually make me feel better.  I'm determined to conquer the bastard!

jpfamps

knuckleberryfinn said:

It was a 500pF running from the reverb pot to the grid of V2A - I'd added it to try to stop lows and mids going through the reverb circuit in attempt to stop it distorting.  The cap must have gone bad when I cranked the amp, though the cap was rated at 500V.  Cheers for the pointer @ICBM.  The whole amp still doesn't sound right.  It's quieter than I would expect and it distorts early and not particularly smoothly.  I think it's V4B: the voltage across the cathode bias resistor is over double what it should be at 3.2V rather than 1.4V.  This would mean a low current, which would account for the low output and early distortion wouldn't it?   I've changed the anode resistor, I've checked all the resistors in that part of the circuit, and I've changed the valve.  Nothing seems to be bringing that voltage down.  Any ideas?

Did you use a silvered mica 500pF cap?

These are notorious for leaking DC, and this seems to have got worse post RoHS. Apparently this is caused by the silver migrating through the mica.

I would use a ceramic disc, or film cap (you can get MK1813 470 pF 630 VDC film caps).

It's easy to diagnose leaky caps by measuring the DC on the valve grids (or say on the reverb pot).

The reverb pot is quite a high value (2M), so only a small amount of current leakage will result in a relatively large DC voltage.

knuckleberryfinn

Yes it was a silvered mica one.  I'm sure I've got some film caps somewhere so will use them instead.

A query on DC leakage - I was measuring for DC leakage and got none except when a signal was applied to the input, does this mean DC leakage, or just that the multimeter is registering the AC signal as DC?

jpfamps

knuckleberryfinn said:

Yes it was a silvered mica one.  I'm sure I've got some film caps somewhere so will use them instead.

A query on DC leakage - I was measuring for DC leakage and got none except when a signal was applied to the input, does this mean DC leakage, or just that the multimeter is registering the AC signal as DC?

The multimeter is trying to measure the signal as DC.

knuckleberryfinn

Well I've fixed it.  None of the caps were leaking DC, but the .02 before the grid of V4b looked like it had seen some heat (that would have been my fault).  So I replaced it.  The voltage of the cathode of V4b was unaffected, so I decided that the schematic must be wrong saying the bias resistor should be 2200R.  After all, the schematic has the power valve bias cap the wrong way round, so there might be other errors.  And I haven't seen 2200R on a single 12AU7 triode in any other schematics (2200 is used for triodes in parallel I think?).  So I added another 2200 in parallel (no 1K resistors left) and the voltage on the cathode came down to 1.3V.  So it's now within range.  What a bloody relief.  Thanks to the usual suspects for the help: @ICBM and @jpfamps.