Success.... after replacing a FET in a PA stage, I quickly measured the resistance of the device from input/output to ground. Whilst all others on the board (and other boards) were measuring ~2.2K the new device was measuring ~1.2K

After walking away in disgust a few times, I invested a couple of hours yesterday to determine if other components were causing the result. I'd been advised to look at any of the tantalum and electrolytic caps. Removing all of these associated with the replaced device produced no improvement.

A closer inspection under a bright light showed that, although I'd cleaned the board after inserting the new FET with IPA, flux was built up between the input leg and ground. Scratching away the flux, cleaning with IPA and repeating a few times and the flux 'bridge' was gone!

Measured the impedance - ~2.2K

Reassembled on heatsink. Applied 50vDC... no smoke (and all voltages tested fine).

A happy day!