Any thoughts on this as a failure mechanism.?
I've had to replace the +ve rail 10,000uF as it had obviously been leaking ,
probably for some time. I assume in those circumstances, in use, hum ac
increases and the +DC rail decreases. I know from powering at +/- 15V the
amp oscillates. Perhaps -44V/+20V or so, it also oscillates, hence
catastrophic failure, but would one expect one channel oscillating to set
the other one off too. ?

I did not think the loose metal plate was the cause, as not the slightest
"spot weld" mark or smoke staining from a nearby immolation.

I will add 47uF over each of the 4x 12V zeners.

Another noticeable thing about this amp being an early model. The pa power
rails are delivered by mains cable and I thought the green must be 0V. Then
I started hunting for the output lines to the relay and it is the green
wires.
The wires to the speaker terminals is also mains cable


--
Diverse Devices, Southampton, England
electronic hints and repair briefs , schematics/manuals list on
http://home.graffiti.net/diverse:graffiti.net/