I have opened the unit, and re-tightened all the bolts that hold the PCB's
Most seemed OK, but some could be turned a little.
Tested it, and until now (2 days) I haven't seen the fault re-appearing, while before it was almost permanent on 6m.
So it seems to do the trick.
I tried to understand what happened. I can't believe that the screws are actually too loose. I noticed they are (probably) zinc plated. I guess that over time some oxyde layer develops,causing high impedance at certain positions. Moving the screws a bit (losing and re-tightening) will probably damage the oxydelayer enough to lower the impedance to within tolerances to avoid oscillation.
Another explanation can be oxydation in the tapped holes in the alumin(i)um frame they are connected to?
The solution may be temporary, will keep an eye on it.
73
Peter/PA3FQH