Marshall JMP 2203 Pair of valves redplatting when cranked.

[henter2017]

sorry for resurrecting the thread, didnt have time to get back into this 2203 until now.
I have removed the series resistor that I had installed on the bias resistor back to stock, set the bias pot to a minimum and installed another new set of matched el34s on it.
With bias probes on all sockets Im getting:

1 - 80mA
2 - 50mA
3 - 80mA
4 - 50mA

extremely hot on valves 1 and 3, where 1 is the closest to the PI ecc83 when looking from the back of the amp...
I have checked the resistors going to all sockets and all tested fine...
Any ideas what could be it?

Thanks in advance!

[Chuck H]

Those current figures are very hot. When you say you set the bias to a minimum do you mean to say that you have less negative bias voltage? You actually want MORE negative bias voltage to reduce that excessive current.

It MUST be telling us something that one tube from each side of the PP circuit is hotter, but evenly so. Since these are matched tubes there simply MUST be something hinky about the wiring or components. Can you please do two things for me? First, swap the tubes around and see if those current figures follow the tubes or stay with the sockets. Then, remove the tubes and post the bias voltage reading from each socket.

[henter2017]

I have changed the bias resistor up from 60k to 75k in order to get the bias lower on all tubes and can test safelly while on.
swaped the tubes and the bias followed the tubes, which made me think are the tubes not matched? but im pretty sure they are, never had problems with matched shuhang quads before from this seller.
then I installed a old a trustful set of winged C, which are for sure matched, (they were installed in several amps before) and had again high discrepancies.
then I installed the original set and managed to get 3 of them at around 40mA and the 4th one at 10mA.
tested it, and the amp still sound as is choking out, the notes are dying soon and sounds horrible.


could an output transformer be causing this at all? I changed an old fried OT to a new 1959HW version, and double checked the wiring and seems fine:
(hardwired to 16OHms)

Output transf. Marshal 100Wl Handwired Series, Model1959HW - Output

GREY 16ohm to speaker output tip
ORANGE 0Ohm to speaker output sleve along with a ground cable.
YELLOW 4Ohm to 100K resistor on board.
GREEN 8 Ohm on air.

then: A1 Red
B+ Brown
A2 White
are verified, (no motorboating) and in the same places as the old transformer.

Also swaped all electrolytic caps, all verified to correct polarity etc...

[henter2017]

forgot to mention that unloaded i still got -39.55V on all sockets.

[Chuck H]

If the current consumption on the aforementioned tubes followed the tubes then that's that. The tubes are NOT matched (for idle current at least). This actually isn't much of a problem as long as you keep one of each higher current and one of each lower current tubes on each side of the PP circuit. That will balance the current through the OT. The downside is that the higher current tubes WILL wear faster carrying most of the load. I'm not the least bit surprised that a trusted distributor would give you something like this. I've experienced all manor of collusion from tube distributors, trusted or otherwise.

On the UP side, the current following the tubes is indicative that the amp probably isn't the problem. It's possible that, as many sets of tubes as you've tested in this amp, it's still been the tubes all along. This has actually happened to me before but I won't go into the story.

Now that you're set up to achieve a lower bias current, just make sure you have balance on either side of the PP circuit WRT current. That is, if you choose to use the "new" tubes see to it that one higher current and one lower current tube are on either side of the OT primary. Set the bias so you're not over dissipating the higher current tube on either side and THEN evaluate the tone.

[henter2017]

Seems that something else must be wrong within the amp, I installed the closest match set that I have, all withing a 25-40mA range on all sockets.
Plugged in and the 3rd tube jumped to 250mA-350mA range while playing, the rest remain more or less within the normal range.
Swaped that tube for another one, and now the 2nd tube jumped to the 250mA range on Idle...

[henter2017]

also the amp sounds choked... very strange....
this is a small clip of the sound:

https://www.dropbox.com/s/hxdvtzd3pt...g%204.m4a?dl=0

[pdf64]

So you are trying to use the amp in that state, ie a tube redplating?
It seems a bad idea to attempt to operate an amp that is known to be malfunctioning so badly.
No surprise really that it doesn't sound good; bizarrely, from the clip to me it sounds like a stage is biased past cut off, ie crossover distortion, though ultrasonic oscillation can have a similar effect.

I suggest to verify that your meter is working correctly and to scope for oscillation.
Tubes that have overdissipated heavily, on a longer than momentary basis, should be regarded as damaged and no longer fit for service.

[Enzo]

Of course the tube current increase while playing. Bias readings are only done at idle.

[henter2017]

Thanks, none of the tubes have actually redplated while testing it, its all momentary, when I saw the bias jumping after taking the standby off I put it back to standby immediatly, also the tubes are testing fine on the tube tester I have.

Enzo just one of the tubes was increasing to that 250mA-300mA range while playing the others went up and down normally without normal ranges.

Could this all be related to a bad OT somehow or something else that Im missing?

[henter2017]

could somebody give me hand installing individual bias potenciometer per single valve (like Oranges/Matamps do).
Or verifying the actual bias circuit in place, I had change the 65k resistor in series with the 22k pot for a 75K, and stil not good enough for some of the sets, something tells me that it must be something else happening.
It will be awesome to be able to bias the valves by singles instead of the actual quad.

[pdf64]

Individual bias can be beneficial but I don't think that attempting to mod out a undiagnosed fault is good practice.

[Enzo]

Pick one tube socket - say the one on the end. Now gather your power tubes. Put one tube in that socket. Bias it up. Is it stable, and does the bias have enough range? Play a few notes to see if it wants to jump. Now remove that tube and put another in its place. Same tests. Now continue one tube at a time. ANy tubes act different? We want to cull out any bad tubes by testing them all in the same socket under same conditions.

Now that we have identified any bad tubes, pick one good tube and stick it in the second socket - the one tube only. Does that socket behave like the first? Now move it to the third socket, and then the fourth doing the same tests. That way we have tested each of the four sockets with the same tube under the same conditions. Does any of the sockets act different?

I agree with pdf64, find your problem and fix it before modifying the amp.

Just my opinion, but having four bias controls is way overkill. It elevates bias adjusting to a level of importance beyond reality. When MArshall builds these amps, they don't used matched tubes. Keep that in mind.

[Chuck H]

also the amp sounds choked... very strange....
this is a small clip of the sound:

https://www.dropbox.com/s/hxdvtzd3pt...g%204.m4a?dl=0

Well something is certainly wrong there! But I noticed that with the guitar turned down the bad affect is the same and when you turn the guitar up the amp DOES get louder with little change to the bad affect or the amount of it. I'm inclined to think that the bad tone is a separate issue located in the preamp or PI.

That's very odd about all but one tube remaining stable and then the one tube acting up, but not the same tube each time. I mentioned that just because you have matched tubes doesn't mean they are IMHE, but the likelihood of none of your matched sets being closer in current than your reports doesn't seem right either. Combined with this newly reported symptom has me scratching my head.

I still think it's possible something is miswired at the sockets. Possibly the PI socket and that is messing with the grid circuits of the power tubes causing the odd current problems. When you read this thread from the start it seems like the problem is changing somewhat and I'm having trouble pinpointing a fixed set of symptoms to solve for.

[Chuck H]

Regarding the individual bias.

Without going long I'll just say that you should fix the amp first. Many 2203's are working fine with the normal, stock bias arrangement so there's no reason changing the bias circuit is necessary to "fix" the amp.