and did a continuity check on the earth between the wall isolator and the machine?
With high levels of 27 Mhz energy in the area an earth lead that shows near zero ohms on continuity with DC test current could have a significant impedance to energy at 27MHz if there are any bends in the cable. This can result in high voltage differences between the ends of the cable. Even running the cable metal to a ferrous object ( girder, iron pipe ) could create an inductive impedance.
Has the cabling to the machine been "tidied up" recently. Such as previously straight runs of cable being re-routed to then have a few bends in them.
Has the machine been moved recently ?
Was it ever intalled properly ?
There may also need to be an earth rod under the machine to disipate stray RF energy. Has that failed ?