Now trying a test cell mix of NaCl and calcium carbonate, as no convenient
source of calcium chloride to see if a mix of those , via deliquescence,
will go thermally conductive without addition of liquid water. Until anyone
gets more info I will go with chloride contaminated calcium carbonate as the
suspect pcb filler

As nothing is learnt by handing over money to Mr Marshall, modded this main
board. As this 06 variant is almost a google-whack I doubt an exact
replacement board is available ,off the shelf, anyways. Missing opto devices
etc
Now I've worked out how to make a 5mm diameter hollow end face cutter , to
go in a Dremmel. Once the main board is disconnected ( how many
connections?) it is now a simple matter of cutting away a neat hole of pcb
around the grid socket pins. And while at it, did so to all 12 grids ECC83
and EL34 and hard wired with silicone sleeving , so all grids are now
isolated from the pcb.
Now no trace of any hum over normal background even over an hour and also
no sensitivity at all from blowing hot air around V1 or the other valves.
Just leaves the mystery of what the exact hum intrusion mechanism was ,
something to do with DC+AC to V1 and V2 heaters?
Now all that is left is to get more history from the owner, why and what
Marshall had farted about with when returned to them before, other than the
all too obvious cut and fly 1/3 W resistor replacements.