1982 JCM800 2205 is driving me nuts

[TassieViking]

The transistors in the Normal and Boost channels are like switches that connects the signal to ground, they are meant to kill the signal by shorting it out.
When the transistor switch on and conduct its like throwing a switch in those circuits.
Replace the BC184 transistor in the the channel that bleeds through and it should fix that problem.
From memory the IC used in later amps in the switching circuit is just full of transistors, nothing else.

The 4 - 1N4007 diodes at the end of the boost channel might be there for clipping, but I have seen diodes in other amps like that and they are used purely to make sure the signal is not to high.
They might be there for a max signal limit and not for clipping, hard to tell unless checking with a CRO.

 

[arthur.lowery]

Yes the schematic is strange! For example, if the lower HT fuse blows, it will remove the negative bias to the output tubes, but the HT will be supplied through the top fuse! Red-plating, for a while anyhow! The transistors as switches are strange too, because they may work as intended for the top half of the cycle, but rely on the CB junction being forward biased back to the switching circuit, which must absorb the negative part of the cycle. The indicator LED also has no current limiting resistor!

As for your problem, it really sounds like oscillation; however, the feedback does not look strong enough to cause oscillation (100 k to 4 ohm tap), but maybe the bottom resistor of the tails of the phase splitter is a higher value than usual? (Giving more feedback, which can cause oscillation).

My friend swapped his JTM45 for a JCM 800 in the 80's and it blew up. Marshall offered me a job when I told them what to fix! I have the JTM 45 now, and it oscillated with a 47K feedback resistor.

 

[Jethro Rocker]

The 4 - 1N4007 diodes at the end of the boost channel might be there for clipping, but I have seen diodes in other amps like that and they are used purely to make sure the signal is not to high.
They might be there for a max signal limit and not for clipping, hard to tell unless checking with a CRO

The split channel 800s have a lot more gain than the single channel series. Without addition of an extra tube, where is this gain coming from if not diode clipping?
I don't know this stiff, so I ask.