Brief update.
I removed and cleaned the contacts on all the sub PCBs in the signal/RGB stages and touched up a few suspect joints. No effect. I have to say the standard of soldering in this set is first class. Especially when you consider it is very cramped in there, with a few components that run hot by nature, and will have been used in a hot environment for decades.
I discovered the chroma board had a small daughterboard rather crudely mounted to it, obviously modified at some later stage. Maybe for PAL? Since I picked up the set here in Madrid from a French ex-pat, that would make sense. There is one loose wire from it, which I have no idea where it was connected.
I tried adjusting the R,G and B drives and the blue had absolutely no effect, the red could be varied slightly. Disconnecting the green output still gave the green screen, this time with a few retrace lines. So it may be a HK short after all
Next I will try the old trick of isolating the heater winding (been a few years since I last did that!)