Scoping my 5E3 - is this the problem?!

knuckleberryfinn

Following on from an earlier thread (http://www.thefretboard.co.uk/discussion/14356/can-t-work-out-if-i-have-a-problem-with-my-amp), I've been having some trouble with the 5E3 clone I recently re-built to its stock circuit (after having fitted reverb to it for about a year).  At low volumes there is a rustley distortion riding on top of the notes, though this seems to go away when the volume is turned up.  Anyway, on the advice of @ICBM I thought I'd take the plunge and get an oscilloscope and see if I can find the problem.  I've been playing with it today, and here's what I've found...

First of all, scoping the preamp (between first gain stage and second, where the volume and tone controls are) I got a perfect sine wave at all volumes, though the trace did keep jumping vertically up and down the screen as I adjusted the volume.  Does this indicate a grounding issue, or nothing at all?

Second, the phase inverter shows significant clipping (see first picture below, where volume was set at 5) which I think is normal isn't it?  If I turn the volume down to as low as it will go, one side distorts before the other (see the second picture below).  Could this be my problem?  

http://i1309.photobucket.com/albums/s639/knuckleberryfinn/20140507_153559_zpsd93754ea.jpg

http://i1309.photobucket.com/albums/s639/knuckleberryfinn/20140507_153534_zps5a847ade.jpg

From the way the volume pot is being scratchy, I think it's probably a problem with the earlier part of the circuit, either grounding or maybe some swarf/wire trimmings/solder has got into the pot (I've replaced the bright channels first coupling cap and this made no difference, so I don't think there's DC on the pots).

What do you people think?

(And whatever the problem is, oscilloscopes are really good fun.)

ICBM

It sounds like something might be up with the PI - in theory the two waveforms should be identical and opposite up to the point of clipping since they're each derived from a resistor of the same value with the same current through them. Have you checked for a DC leak on the cap between the two halves of V2?

camf

I love science.

jpfamps

ICBM said:

It sounds like something might be up with the PI - in theory the two waveforms should be identical and opposite up to the point of clipping since they're each derived from a resistor of the same value with the same current through them. Have you checked for a DC leak on the cap between the two halves of V2?

Actually that's only the case as long at the power valves have not gone into grid conduction. The traces shown above looking to me like this is happening.

If the power valve driven from the cathode goes into grid conduction you then effectively have a triode gain stage with a fully bypassed cathode and will experience a gain spike on in the signal. Normally this will not be audible because the power valve will have been driven into clipping at the point anyway (depending somewhat on the loading), although blocking distortion can be exacerbated by the jump in gain. If the anode driven power valve goes into grid conduction there is little effect at the cathode.

To differentiate between clipping in the PI and the power vales entering grid conduction try scoping the PI with the power valves removed.

In PI itself can also go into grid conduction, although this is unlikely at low volume levels.

What brand of valve have you been using by the way? JJ ECC83s used to be known for going into early grid current (I don't know if this is still the case;, I should try investigating it)

The trace jumping up and down on the screen as you adjust the volume controls is most likely a small shift in the DC conditions are you rotate the pot.

It's not uncommon in my experience for amps to have some unpleasant distortion artefacts at lowish volumes which are masked at higher volumes by greater signal distortion.

knuckleberryfinn

Thanks @ICBM and @jpfamps.  I'll do some more probing early next week, which is unfortunately the earliest I can get back to it.  I haven't checked for DC leakage at that cap, so will check that out.  The ECC83 is indeed a JJ - nice detective work.  

57Deluxe

I have a Lab coat in the shed...