Marshall 2203 JCM800 Lead series (100W) reissue

[Helmholtz]

Assuming a somewhat high heater winding voltage of 6.7V, the total power dissipated in the balancing circuit is P= V²/R = (6.7)²/300 = 0.15W.
This means 0.05W or 50mW per 100R resistor.

[the3secondrule]

R11 open circuit can be just a failed resistor unless burnt. If burnt, I don't think even an anode cathode short can cause that as there is a 100k fanode feed resistor limiting the current but an incorrect valve in V2 position could do that. Pin 1 to pin 8 short ... highly unlikely.
The heater trim resistor chain can fail when an EL34 fails. I have known the chinesium version go heater cathode short circuit on a regular basis. It may be the resutlt of a V2 failure but then again, quite unlikely.
C4 failing will not cause R11 to fail for similar reasons already described.

Just coming back to this after rebuilding the heater / hum balance circuit.
Should an EL34 that's failed like this show a short between heater/cathode when powered off?
Because I metered all the tubes and there was no short - powered up the amp and after a short while the hum balance went up on flames again.
Checked the tubes again after powering down - no short. Is it likely that this is only showing a short while powered up?

I'll be replacing the EL34s before powering up again, but just wanted to check if I'd missed something

[g1]

Should an EL34 that's failed like this show a short between heater/cathode when powered off?

I don't think that is necessarily the case. Some tube testers have a selection to test for h-k short, but I still think it may not always show up until actual working amplifier voltages are in place. Tube testers do not always reach those voltages.

[The Dude]

I can also verify that some shorts won't show until the tube is warm. On my Sencore tester, if you let an HK shorted tube cook a bit it will show shorted after warmup. More often, on good tubes, it's the opposite. The tube will show HK leakage until it warms up and then the leakage disappears.

[the3secondrule]

I can also verify that some shorts won't show until the tube is warm. On my Sencore tester, if you let an HK shorted tube cook a bit it will show shorted after warmup. More often, on good tubes, it's the opposite. The tube will show HK leakage until it warms up and then the leakage disappears.

Ok that makes sense then. I replaced the heater resistors/trimmer, and powered up the amp.
Was getting the same oscillation as previous when plugged in the front end, but was getting clean audio when sending signal through the FX return jack, although it does appear the amp was down on power.
Didn't have a chance to measure the output though, as the amp had been powered on for around 5 minutes and I was setting up to check it on the scope when the heater resistors went up in literal flames.

[the3secondrule]

Did you fix the blown resistors on the Hum Balance circuit ? Floating heaters can cause oscillation because of the feedback path from output tubes back to the input triodes.

have now replaced the hum balance resistors and powered up again - the oscillation is still present when plugging in through the front of the amp, although its gone when plugging through the FX loop return.
However after a few minutes testing the hum balance circuit has gone up in smoke again.

At this stage will be replacing the power tubes as well as hum balance circuit before testing again.

[Greg Robinson]

Yes, it's almost certainly the power tubes, replacing them is highly recommended.
I'd recommend not repairing the stock hum balance circuit, as it is prone to such failure due to under-rated trim pot wattage. I'd instead just use 2x1/2W flame proof resistors for an artificial center tap mounted on one of the power tube sockets where they are less likely to do collateral damage. I'd also consider taking that artificial center tap to node W1 (30V) for some heater elevation rather than to ground.

[Greg_L]

Yes, again, get rid of that dumb PCB mounted hum balance pot circuit and wire it like a standard classic amp. Elevate the heaters, sure, if you want, but get rid of all that other filament circuit nonsense.

[pdf64]

I think DC elevation of the heater circuit is beneficial, eg makes the amp able to use valves that would either cause buzz / hum or become damaged in a DCCF circuit.
And making the heater circuit less 'attractive' to high voltage arcing has got to be good.
But in this case, I think a dedicated elevation potential divider would be far preferable to using the 30V supply. Because the resulting collateral damage if an arc did occur and zapped all the 30V supply and fx loop circuit would be worse.

[Greg Robinson]

But in this case, I think a dedicated elevation potential divider would be far preferable to using the 30V supply. Because the resulting collateral damage if an arc did occur and zapped all the 30V supply and fx loop circuit would be worse.

100Ω resistors should blow before anything else.

[g1]

I think hum balance pots are great.
Maybe it's not the exact right one for the job (wattage wise), but they are the best solution for heater hum in my opinion.

[Helmholtz]

50mW will neither overload the trim pot nor the 100R resistors (see post #16).

[Greg Robinson]

50mW will neither overload the trim pot nor the 100R resistors (see post #16).

No, it should not, however I consistently see the hum balance circuit in this model having failed (typically the trimpot), far more regularly than other amps. Often with damage to the circuit board, and frequently with no signs of a heater short in the power tubes. I shied away from repairing the stock circuit after it failed again in short order on a client's amp with a new set of power tubes that showed no signs of heater short circuit, while 2x100Ω's mounted on a tube socket held up just fine. I can only attribute it to gremlins.

[catalin gramada]

50mW will neither overload the trim pot nor the 100R resistors (see post #16).

Supposed a 100ohm trimpot balanced at the middle of the track ...grossly 3.5V per 50 ohm side meant 0.07A x 3.5V = 0.245 W . A quarter of watt trimpot is supposed to hold those power along its whole track , so there is a certain fact it will burn. ....I don't really know what trim they used....just thinking around...

[Helmholtz]

Supposed a 100ohm trimpot balanced at the middle of the track ...grossly 3.5V per 50 ohm side meant 0.07A x 3.5V = 0.245 W . A quarter of watt trimpot is supposed to hold those power along its whole track , so there is a certain fact it will burn. ....I don't really know what trim they used....just thinking around...

Because of the two 100R series resistors the trimpot only "gets" 1/3 of the heater voltage or 2.3V across, so 53mW dissipation.
The position of the wiper doesn't matter. No heater current flowing through the wiper.