Trevor Wilson wrote:

Got an 'interesting' fault on an ancient Linsey Hood amp.
Switch on a
light etc, and sometimes *one* channel bursts into loud
mains hum - but
carries on working. And the output transistors get very
hot. It's not
normally susceptible to mains born interference - so any
guesses before I
rip it apart?

Both channels seem to match ok before this happens.

**First off: The "Mains hum" is not mains hum. It is 100Hz
hum. It is the sound made by an amplifier when so much
current is consumed from the power supply, that ripple
appears on the outputs.

Is it usual...or was it...to have two separate power
supplies in an SS amp? Note the hum is only on one channel.

Second off: You need to ascertain if the output devices
get hot with no load connected. If they still get hot,
then look for a fault in the bias section, or further down
the chain. Something is causing the output devices to
conduct too much current.

Oscillation, perhaps?

If the output devices run cool when no load is connected,
then it is likely you have a large DC Voltage at the
output. If so, the fault could lie anywhere in the amp.

Since the fault is 'triggered' by some external issue,
it's gonna be tricky.

Except the triggering might be a good clue?

If you have a link to the schematic, it might be helpful.

If the problem could be oscillation, what kind of fault
could narrow the stability margin such that it can be set
off by a spike on the mains?

The schematic of an early JLH amp should be easy to find if
the model is known. Were they commercially produced, or is
it home made? "Improvements" are probably common.

Ian