Hi Colin,
Across X3 I get 179mV and between X3 anode and R21 slider I get 178.5mV so not a lot of difference.
I think I will have to put the silicon diode back in and try gently heating the output stage whilst monitoring the bias current carefully. I never seem to have much luck with bias arrangements, I have also been working on a Ferranti set which uses a thermistor to stabilise the bias, this has failed, a normal resistor works but then you end up with the problem of thermal runaway and of course correct thermistors are no longer available
Josh.