[geofy]

If the transistors are still getting warm then the original fault could still be there, I would, without having a circuit diagram to verify, suspect the red and yellow electrolytic capacitor nearest the two transistors could be leaky, if this is the emitter bypass of those two. Or the driver transistor also an OC81 could be faulty.

I personally would not at this stage start changing all the red/yellow caps on mass, they are not that unreliable. And this set doesn't have waxies in the same sense as older valve sets and the smaller value caps will not suffer leakage in the same way.

For experimental purposes a PP3 battery is satisfactory, and an alkaline PP3 will outlive a zinc carbon PP9, which of course can be fitted for authenticity if this is one of the criteria, at the risk of battery leakage if forgotten about for to long whereas an alkaline will be less likely to leak. I use a rechargeable 200mA/hour PP3 which gives several weeks normal listening.

Just my opinion of course

Geof