Fuzz pedal (blocking distortion) makes one EL34 shows early signs of red plating in 50 watt

[Gtr0]

Grid stoppers only make a difference if oscillation is involved. Higher values (in the 100k range) can help to smooth out blocking distortion.

If redplating is not excessive I consider your amp fixed.
If only one of the tubes redplates you might try a different PI tube and/or experiment with different PI plate resistors to achieve better grid drive symmetry.

certainly I have no idea when it is considered excessive... here is a photo of the red plating with the lights off... lights on it is not as apparent of course but I noticed it enough to investigate it..


this is when I am playing with certain guitars or with a fuzz pedal. If it was just the fuzz interaction I wouldn’t worry about it... but it also happens while playing a strat with humbuckers and is made of mahogany (if wood means anything at all).


Finally I replaced the tube socket. The photo above is after the replacement... so it made no difference.

[Helmholtz]

A pedal doesn't directly cause redplating. But it may increase signal level.

I'd say that amount of replating is not unusual with clipping Marshalls.
Better PI symmetry might help, but that would probably mean some re-design rather than repair (see above).

[pdf64]

What’s the resistance of the choke? From the HT voltages, I guess it’s modelled after a higher resistance type, but it’s worth checking.

[Helmholtz]

Choke DCR affects screen voltage but not HT. As he already added 1k screen resistors, I don't think choke DCR is critical.

[pdf64]

My take is that the choke resistance ( eg 690 or 100 ohms) will affect the HT supply Vdc at the g2 node, especially at high power outputs. Which will affect plate current and hence dissipation.

[Helmholtz]

The choke resistance will affect the HT Vdc at the g2 node?

With HT I mean power tube plate supply (otherwise an amp would have several HTs, like HT1, HT2 etc ?).

But my point is that the 1k screen resistors will drop more screen voltage than a typical choke. And preamp or PI supply voltages will have little effect on redplating.

IIRC, a typical Marshall choke has around 100R.

Also individual screen resistors drop voltage immediately with screen current and thus provide more effective screen protection/dissipation limiting than increased power supply resistance, because the filter cap prevents instantaneous voltage drop.


If you look at the scope pictures you see that one of the clipping tubes is "on" for a longer time interval than the other. That tube will have higher plate dissipation and eventually redplate.

[Gtr0]

What’s the resistance of the choke? From the HT voltages, I guess it’s modelled after a higher resistance type, but it’s worth checking.

I can't remember the exact reading, but it a 100Ω.

With HT I mean power tube plate supply (otherwise an amp would have several HTs, like HT1, HT2 etc ?).

But my point is that the 1k screen resistors will drop more screen voltage than a typical choke. And preamp or PI supply voltages will have little effect on redplating.

IIRC, a typical Marshall choke has around 100R.

Also individual screen resistors drop voltage immediately with screen current and thus provide more effective screen protection/dissipation limiting than increased power supply resistance, because the filter cap prevents instantaneous voltage drop.


If you look at the scope pictures you see that one of the clipping tubes is "on" for a longer time interval than the other. That tube will have higher plate dissipation and eventually redplate.

Yes, one side is on for a longer time than the other - could you please explain to me why that is? I assumed it had something to do with this asymmetrical clipping from the PI... but then again as much as I want to learn and know the theory, I am a bit out of my depth.

I meant to reply yesterday with my progress...
So yesterday, I replaced the 82kΩ plate resistor with a 68kΩ, but there was no change in the red plating. The amp at this point started to develop a bit of a hum, so I started with replacing preamp tubes and settled on a set that sounded both pleasant and were noise/hum free. Then I moved to trying different output tubes. ATM, the amp is biased at +/-50% so I just swapped in and out "tube sets" to see if the red plating would change.

I tried some:

• Harma EL34 Retros (I think they are Tung Sol)
• JJ E34L
• Marshall stock (JJ)
• UOS circa 1990 Svetlana EL34
• TAD EL34B-STR.

All sets have been used at some point. Mostly sets I keep when testing new builds for the first time etc.

Well, all tubes sounded good in slightly different ways. I settled with the TADs.

Red Plating: It is at a point that the only time that it red plates is when I use the fuzz pedal in front. It adds a shit-ton of harmonic content to the original signal. Though I would love to sort it out, it could very well be happening to my other amps and I am not aware because they are all tucked away in their head cases.

Before all of the troubleshooting, with the original tube set I had installed (the Marshall tubes) it would red plate if I used my Fender Start (parallel universe Uptown Strat) or if I used my Univibe clone (Fulltone Deja Vibe MDV-3) - the vibe adds quite a bit of low mid harmonic content and the Strat is incredibly huge sounding in the low end / low mids as well. The Strat has "DoubleTap" humbuckers. The Strat can not be used with all amps, certain amps it turns to mud.

At the last point I was at as of last night, all red plating has stopped - regardless of the guitar used - except when using the fuzz pedal.

My only seemingly lingering issue is in the basing. I bought a set of Groove Tubes a few days ago - but really I have had a trust issue with power tubes in general for the last several years. I think I have most faith in TAD tubes, the last two amps I used TAD tubes in were within 1 mA at bias. But I get the sinking feeling that no matter what I use in this amp, the bias will not be matched. And I am totally fine with it other than the possibility that some else is... wrong with it.

I thank you all for your time, thoughts, and efforts. Very much appreciated!

[pdf64]

just checking with my dmm...

plates @ idle = 458VDC
plates @ vol 3 = 399VDC
plates @ vol 5 = 395VDC

G2 @ idle = 453VDC
G2 @ vol 3 = 364VDC
G2 @ vol 5 = 356VDC

Were these measurements taken at the HT supply nodes (ie the fully decoupled nodes at the OT CT, and at the g2 supply, after the choke), or at the actual plate and g2 terminals?

...before the .022 at the PI input...
...

As is typical of this Marshall preamp with the DCCF, with larger signal levels and into clipping, its waveform becomes increasingly asymmetrical. Hence tinkering with the LTP symmetry won't help much.
Maybe asymmetry from the fuzz is exacerbating the asymmetry of the preamp? If so, flipping the polarity of the fuzz signal may counteract it

With HT I mean power tube plate supply (otherwise an amp would have several HTs, like HT1, HT2 etc ?).

But my point is that the 1k screen resistors will drop more screen voltage than a typical choke. And preamp or PI supply voltages will have little effect on redplating.

IIRC, a typical Marshall choke has around 100R...

Apologies for the terminology confusion, my understanding and use of 'HT' is that it's a generic reference to any of the (decoupled) +ve high voltage supplies to the various parts of the circuit. If specifying a particular point in the HT supply, I'll qualify it descriptively, eg HT g2 node. Hence the use of a older type eg RS choke (20H 690 ohms) will result in a lower, saggier HT Vdc at the g2 node, compared to a later JCM800 style 100 ohm choke. Unfortunately I'm unsure of when the transition was made or how long the crossover period was, ie what spec choke was typical of which production era.

[Gtr0]

Were these measurements taken at the HT supply nodes (ie the fully decoupled nodes at the OT CT, and at the g2 supply, after the choke), or at the actual plate and g2 terminals?


As is typical of this Marshall preamp with the DCCF, with larger signal levels and into clipping, its waveform becomes increasingly asymmetrical. Hence tinkering with the LTP symmetry won't help much.
Maybe asymmetry from the fuzz is exacerbating the asymmetry of the preamp? If so, flipping the polarity of the fuzz signal may counteract it


Apologies for the terminology confusion, my understanding and use of 'HT' is that it's a generic reference to any of the (decoupled) +ve high voltage supplies to the various parts of the circuit. If specifying a particular point in the HT supply, I'll qualify it descriptively, eg HT g2 node. Hence the use of a older type eg RS choke (20H 690 ohms) will result in a lower, saggier HT Vdc at the g2 node, compared to a later JCM800 style 100 ohm choke. Unfortunately I'm unsure of when the transition was made or how long the crossover period was, ie what spec choke was typical of which production era.

voltages are at the pins... 3 and 4 (for g2)

And I know the second part is for Helmholtz, but I can say that I have just finished building a '65 era JTM45 with "clones" of the original drake trannies and choke (heyboer) and the choke is 100Ω 3H. But I see pics from amp archives that they were mixed in those early years. I assume by '67 or '68 it was all Drake 3H moving toward C1999 which I think is 10H. ?

as far as referencing HT nodes, I tend to do the same as PDF64. Though when someone asks what my B+ is, I always refer to the output tube plates or CT supply.

[pdf64]

Where do you buy your transformers?

I strongly advise against probing a power tube plate that not at idle; back emf can put >2kV there, especially when overdriven into am inductive load. Putting the test equipment at grave risk of damage.

I only ever probe an operational power tube plate when absolutely necessary; unless an OT primary issue is suspected, the idle Vdc at the HT supply to it is plenty close enough for our purposes.

And with pentodes especially, probing the plate can put the amp into oscillation, screwing up the reading; if nothing else, the amp would no longer be idling.

[Gtr0]

Where do you buy your transformers?

I strongly advise against probing a power tube plate that not at idle; back emf can put >2kV there, especially when overdriven into am inductive load. Putting the test equipment at grave risk of damage.

I only ever probe an operational power tube plate when absolutely necessary; unless an OT primary issue is suspected, the idle Vdc at the HT supply to it is plenty close enough for our purposes.

And with pentodes especially, probing the plate can put the amp into oscillation, screwing up the reading; if nothing else, the amp would no longer be idling.

I bought them at Metropoulos Amps in the US. Importing them to Belgium was a hefty expense... but back in the day I used their trannies (Heyboyer via Metro) - I am originally from the US so it was cheaper then for shipping ;-)


EDIT on Fri, Nov 13: The transformers listed above where what I put in my '65 era JTM45 "clone" (as in the previous post). A different amp from what this topic was about. The amp I am referring to generally in the thread are made in Italy by Inmadout. I have several of his transformers and find them to work quite well. AND he is a super easy guy to work with - if you give him the specs you want, he will build it no problem. However, I am always looking for new companies to check out. Been thinking about Sowter in the UK... they are a pretty penny though. Last time I spent a fortune on transformers were Mercury.... which left me with a tad of buyers remorse.

[Gtr0]

BTW, I typically do not probe the plates either - just too afraid ;-)... even with my 100x 1.5kV probes... I made an acceptation in this case.

[Helmholtz]

But I get the sinking feeling that no matter what I use in this amp, the bias will not be matched

In post #38 you reported an idle current difference of only 2mA. That's a fair match.

[Helmholtz]

Unfortunately I'm unsure of when the transition was made or how long the crossover period was, ie what spec choke was typical of which production era.

The choke in my all original '66 JTM50 measures 113R.

[Helmholtz]

So yesterday, I replaced the 82kΩ plate resistor with a 68kΩ

What effect did this have on the grid signals? You might have lowered the wrong side PI resistor.


Dumble sometimes used a PI symmetry pot: