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

[Helmholtz]

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.

The asymmetrical grid signals with grid current also develop with a symmetrical PI input signal. Primary reason is an asymmetrical shift of the DC operating points of the 2 PI triodes, caused by the grid currents (charge pump effect?). Zollner analyses the effect in his book "PoteG".
After re-reading I'm no longer sure if balancing the PI gains cures the problem.

I have to admit that I don't completely understand all the interactions involved. Too many nonlinear effects.

[pdf64]

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

Yes, the Ingo Gorges JTM45 transformer set I've got has that type of choke too. But from other suppliers, I remember seeing an alternative type offered, with higher inductance and resistance.I have a hypothesis that those early JMP50s that didn't have screen grid current limiting resistors may have had such chokes, as that would help the EL34s to cope. It's just a hypothesis though, I've no evidence to support it.

Does your JTM45 have the RS Deluxe OT? How about the 1k safety resistor across the output jacks?

[Helmholtz]

Yes, the Ingo Gorges JTM45 transformer set I've got has that type of choke too. But from other suppliers, I remember seeing an alternative type offered, with higher inductance and resistance.I have a hypothesis that those early JMP50s that didn't have screen grid current limiting resistors may have had such chokes, as that would help the EL34s to cope. It's just a hypothesis though, I've no evidence to support it.

Does your JTM45 have the RS Deluxe OT? How about the 1k safety resistor across the output jacks?

My JTM50 might be one of the first EL34 amps. It has the original JTM45 circuit (470R screen resistors) but has an early EL34 OT.

The 20H choke can be seen in early original JTM45 schematics.

But it's easy to see if wiring a 560R (>3W) in series with the existing choke helps. I don't expect much difference.

[pdf64]

My JTM50 might be one of the first EL34 amps. It has the original JTM45 circuit (470R screen resistors)...

Does it have the 1k shared screen grid resistor, which then feeds the 2 x 470 resistors? As per https://drtube.com/schematics/marshall/jtm45tr.gif
Ah yes, there's the 20H choke too!

...it's easy to see if wiring a 560R in series with the existing choke helps. I don't expect much difference...

My guess it that there's a good likelihood it would help - perhaps not much but maybe enough

[Helmholtz]

Does it have the 1k shared screen grid resistor, which then feeds the 2 x 470 resistors?

Excellent point!

I had forgotten about that. It originally came with two parallel 22k (!) resistors in that place. Considered it an error.
Resistors looked original and they might have been necessary if they used KT66s with the wrong (too low impedance) OT. The amp came with KT66s.

(I much prefer original KT66s to EL34s and will change the OT sometime.)

As per https://drtube.com/schematics/marshall/jtm45tr.gif
Ah yes, there's the 20H choke too!

I guess you noticed the wiring error?

My guess it that there's a good likelihood it would help - perhaps not much but maybe enough

Well, it will drop maybe 35V at full power. Probably more at full clipping.
Might be more effective to increase screen resistors to 1.5k.


But as said before, this all is no longer fixing but rather redesign territory and might change the amp's tone.

[nickb]

Hi Friends.

I understand that blocking distortion can cause red plating - though I have never experienced the possibility until recently while using a Fuzz pedal.

I missed this thread but this statement got my attention. Blocking happens when the amp is driven hard and that means the tubes are close to saturation and that means less power dissipation since the voltage across the tube is less. So it begs some questions.

How does the power dissipation vary with drive and how does using a fuzz pedal affect that?
Does a lower bias current would help (although I find very cool say < 50%) biased amps sound a bit gritty but depends on the feedback).
Lastly, is this a problem specific to this amp?

So here is some data that try to answer these. The model used a couple of EL34 in push pull, B+ 490V (very stiff). The input was varied from 0 to 180Vpp, used sine a square ( like a fuzz box) waves. Bias varied from 30% to 80%.

Here is the power dissipation for a square wave with varying drive and bias.


Here is same using a sine wave.

Observations.
You can see the dissipation slightly is higher when using a square wave.
The maximum dissipation is exceeded. Of course you might not be driving it flat out but you can see the potential to red plate.
Also note that bias has a reducing effect as the drive increases.
The max power dissipation happens at about 2/3 of the way to clipping, so turn it up and the dissipation goes down.

So, my guess is, you are seeing a normal result as a combination of style, effects and volume level.

[Helmholtz]

Does your EL34 model incorporate grid conduction effects?
And does the sim show the duty cycle shift of the grid signals (assuming an LTPI) ?

[nickb]

Does your EL34 model incorporate grid conduction effects?

Yes.

Does the sim show the duty cycle shift of the grid signals (assuming an LTPI) ?

No.

Here is the effect of power supply sag:

[Helmholtz]

Here is the power dissipation for a square wave with varying drive and bias...

Your sim results with square wave drive are most interesting and essentially conform with Zollner's measurements.
They confirm that Aiken's dissipation calculator gives wrong results for squarewave drive at least at high output (https://www.aikenamps.com/index.php/...why-70-percent)
It seems impossible that the plate dissipation tends to zero with increasing output.
I assume that he didn't factor in the finite plate "saturation" voltage, which prevents zero dissipation.
I actually wrote him an email about this some months ago - no response to date.

[Gtr0]

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

Right - I had forgotten that once I swapped the tube positions they seemed to balance better. I was so focused on the readings that were there before I swapped their positions.

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:
[ATTACH]temp_52415_1605195164614_217[/ATTACH]

Very interesting indeed. I'm going to take a closer look into this. As far as the readings that resulted, I will have to refer to my notes and get back to you on that... but I have a few images I can share in a few moments....

This morning I was fiddling around with Ampbooks LTPI balance calculator...
https://ampbooks.com/mobile/amplifie...ir/calculator/

Seems MesaBoogie did it better with a 91k/100k - seems to balance out better. Mine was more about trial and error from clues on a few things I have read... but yes, it seems I may have gone the wrong way... (??)

As a general update: I recently purchased a TAD BiasMaster. I didn't have a chance to compare it's readings to doing it the old fashioned way (CT to Pin 3 voltage drop / resistance), but if nothing else I thought it may be a quick way to see how well a given set of tubes are matched.

Having said that, I put in a set of brand new TAD EL34B-STR in the amp. According to the BiasMaster they are within about 1.5mA. Good enough for me.

Overall, the amp has more overdrive than any other "super lead" type amp I have ever had. It's not insane, but it's a noticeable amount. So my volume pot is on 8 where on my 100W super leads... one I have built, and one Marshall RI from the 90's, the volume is on 10 to be "on par" with amount of overdrive. I's not a huge difference, generally, when you get from 8 to 10... but there is definitely more "swirl" (intermodulation distortion?) - which I am typically a huge fan of. Also because of negating the effects of the bright cap, it can get muddy if the bass is not on 0.

So, couple a circuit that seems to have more overdrive/distortion than usual with a fuzz pedal... this was why I originally termed it with "blocking distortion" - certain notes get filled out in a way that it just sounds like a super square wave. I assume this is describing blocking distortion. It's been a while since I have read about it, I will have to go over my books again, but I know this causes all kinds of unusual behavior.

[Gtr0]

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:
[ATTACH]temp_52415_1605195164614_217[/ATTACH]

Here are some images....
Things seemed to have..... gone up. More voltage swing... larger signal. At least that's what I am seeing... mind you I do not have enough experience to know for sure... and so this is a great learning experience for me. :-)

The pics were taken just before I replaced the tubes with a new set. But are with the 68kΩ in the PI plate #1 (v4 side)

pi grid 1 (v4 side)


pi grid 2 (v5 side)


pi output 1 (v4 side)


pi output 2 (v5 side)



What do you think?

[Gtr0]

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:
[ATTACH]temp_52415_1605195164614_217[/ATTACH]

Ahh I see - 10k pot.

What type of pot would be used here? I think the voltage is too high for one of those 15mm Piher trimpots... but the .25 wattage rating would be fine, right? If my in-brain calculations are to be correct (which doesn't surprise me if they are not) :-)

Some pots show they can handle 500V if the pot in linear... huh???... where their log taper counterpart claim 250V. Bournes HRT.

[Gtr0]

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

Jealous. In a friendly way of course.

I just took the measurements last night.. it is 10H, 125mA, 116R (actual, in circuit)

[nickb]

Ahh I see - 10k pot.

What type of pot would be used here? I think the voltage is too high for one of those 15mm Piher trimpots... but the .25 wattage rating would be fine, right? If my in-brain calculations are to be correct (which doesn't surprise me if they are not) :-)

Some pots show they can handle 500V if the pot in linear... huh??? Bournes HRT.

The voltage across the pot is small, something like 20V so just about anything will work.
0.25W is plenty.

[Gtr0]

So I have one, final, question related to this topic.

Some schematics show 2x 10k dropping resistors to the PI B+, some do not. And I saw at least 1 that were 2x 8.2k.

This amp build utilises one single 10k resistor in this area.

Any thoughts on that? Will it drop voltage down stream the entire line or just the PI supply? I suppose I could just try, but I thought I would defer to those of you guys who may have the real deal late 60's JMP 1987.

Here are my vastest voltages... in detail

v3 (PI):
1 - 271.3
2 - 25.42
3 - 42.75
6 - 251.8
7 - 28.28
8 - 42.75

v2 (CF/EQ):
1 - 328.1
2 - xx
3 - 1.335
6 - 194.4
7 - xx
8 - 196.2

v1 (Input):
1 - 179.9
2 - xx
3 - did not measure
6 - 226.3
7 - xx
8 - did not measure

NODES:
OT CT: 464.1
SCREENS: 462.8
PI: 380.1
v2: 328.2
v1: 307.4

Output tubes, 2x TAD EL34B-STR:
3 - 467
4 - 465
5 - -41.90

BIAS out: -43.00

once again, thank you all fr your help!!!!