So a bit of diagnostic progress:
Removing the earth from the input IEC connector and putting an AVO8 Mk7 on AC volts between the IEC Earth tag and the removed machine Earth, and the machine powered up I was reading 88v AC - roughly what I'd read before and enough to give you a tingle on a grazed knuckle - (ask me how I know!).
Now the internal resistance of the AVO is high enough to limit the earth fault current sufficiently to not trip my 30mA RCD in the fuse box, but is low enough to swamp stray capacitive effects that you would get using a modern electronic meter like a Fluke.
So by a process of elimination working though all the mains using devices in the machine I found that the main spindle drive inverter to be the culprit. Now as inverters switch at high frequencies they do tend to cause quite a bit of earth leakage, but in this case I'm sure it is excessive. I strongly suspect the input filter components in the enlarged view, but getting at them to replace wouldn't be easy.
As I have a 13 amp socket for lighting on my three phase Bridgeport, that is derived from phase 1 and neutral, and the three phase distribution has a 100mA RCD I moved the extension lead from a 30mA protected socket to this 100mA protected socket and sure enough it didn't trip (I'd left the AVO in circuit but rotated the knobs to measure AC 10 amp so effectively a short circuit restoring earth continuity.
Out of interest I turned the knobs back to AC volts and I was measuring very low volts - nothing like before where I had 88 volts AC. All very odd - so I put the machine back onto the 30mA protected circuit with a solid earth by removing the AVO and pushing the faston connection back together, and darn me the bally thing works and doesn't trip the 30mA RCD any more

The only conclusion that I can draw at the moment is that whatever was causing the leakage has been blasted away by the (presumably) less than 100 mA current that initially flowed, and is now in oblivion

So although I don't like disappearing faults, at least it's working and no longer dangerous as it was.
To prove all was working I tried to make another M6 brass bolt, but the darn program wouldn't run - couldn't fathom it, until I noticed the little bit of text saying 'Low Lubrication Oil Level' - of course I hadn't filled the reservoir when I glued the crack to give it time to set. Filled her up, off the program went . . . .BUT the crack still leaks

...ah well . . .
