Distorting reverb weirdness

jpfamps

knuckleberryfinn said:

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.

Good work.

Seems that the 12AU7 was biased quite assymetrically, hence the distortion.

jpfamps

By co-incidence, I've been working on a Gibson GA-30RV which has the same reverb circuit and the same problem with the distorted signal present when the reverb is turned up.

On inspection, the problem is that the signal to the reverb circuit is actually coupling via the power supply.

Essentially the power supply should be a AC ground, however filter capacitors don't have zero impedance, so a residual of the audio signal will appear on the power rails. This can then couple into other gain stages.

The recovery side of the reverb circuit has a lot of gain, so you don't need much signal to couple through for it to be audible in the power stage.

The problem here is that too many gain stages are being hooked up to the same power supply node (both sides of V2, and the return stage V3) with only a 10 uF cap on it. 

The amp I'm working on has a new 10 uF cap, so there it looks to be a "design feature" rather than a problem with the filter cap deteriorating.

You could try rewiring the power supply so that there is an extra stage of decoupling to avoid this problem.

I suspect that Gibson got away with this because this distorted signal was hidden under the guitar signal.

knuckleberryfinn

This is brilliant @jpfamps - thank you.  I'll try putting an extra filter stage in.  The amp sounds a lot better than it did, but still not as good as I'd have hoped - I've lost quite a bit of depth to the reverb by reducing the gain to try to avoid the distortion, and there is a touch of graininess to the sound.  I'll report back after doing the mod... 

knuckleberryfinn

Well @jpfamps, adding another filter cap for the pre-amp and reverb circuits has cured pretty much all the problems with the reverb distortion.  It has also made the amp, when turned down to zero, almost silent (before. the volume was at 2 even when turned down to zero, presumably because of this unwanted coupling problem).  However, it has also unmasked the true sound of the amp, which is not great.  I was planning to sell it, but it simply sounds bad so I'm withdrawing it.  I think I'll strip it for parts and make something decent.  Thanks for all your help with it.  Although I've put a lot of work into rebuilding this thing, it hasn't been for nothing - at least I learnt something.

jpfamps

knuckleberryfinn said:

Well @jpfamps, adding another filter cap for the pre-amp and reverb circuits has cured pretty much all the problems with the reverb distortion.  It has also made the amp, when turned down to zero, almost silent (before. the volume was at 2 even when turned down to zero, presumably because of this unwanted coupling problem).  However, it has also unmasked the true sound of the amp, which is not great.  I was planning to sell it, but it simply sounds bad so I'm withdrawing it.  I think I'll strip it for parts and make something decent.  Thanks for all your help with it.  Although I've put a lot of work into rebuilding this thing, it hasn't been for nothing - at least I learnt something.

You're always learning, and a result of sorts.

I have to admit, I'm often disappointed with with sound of many Gibson amps.

The GA-30RV in the workshop wasn't stellar either.