Neutralizing cap blowout

[W9RAN]

What would make a neutralizing cap all-of-a-sudden start to arc over?  Cap in question is in a G-76, not exactly a powerhouse but plenty enough juice to light up the rear quadrant of the rig in a pretty impressive way.  My experience is, once a cap arcs it'll always arc, but this is an APC style variable with 2x the normal plate spacing so a replacement may not be easy to find.  So I bead-blasted the plates clean and tried it again, and at first it would only arc a bit when I'd unkey.  But after a few kerchunks, it started up again.   I can't find anything wrong with the driver or final.  The cap was set with the plates about 20% engaged, so I turned it to the minimum capacitance position and it's stable and no signs of neutralization weirdness, at least on 75.  The higher bands might be a different story. 

Anyone run into this or have a good suggestion for a substitute?

73, Bob W9RAN

[W1AEX]

Any chance that RFC 6 has opened up? I've seen weird changes in tank circuit RF voltages when the choke at the output fails.

[VE3GZB]

Almost sounds like a parasitic has cropped up somehow.

[k4kyv]

An air variable cap shouldn't be any more prone to arc after an initial flash-over, unless carbon tracks were etched into the ceramic insulators.  Usually the arc is directly across the air gap, from plate to plate.  Look  for pitting or beading of molten metal where the arc took place.  Sometimes the heat can cause a plate  to warp and distort the plate spacing.

I have had caps to arc over until a hole started to melt in the plate, but fixed it by bending the plate back straight, and, using a file and emery cloth, smoothed over and re-polished the burnt spots.

[WQ9E]

Bob,

Do you have a high pot tester or a variable source of HV so you can test the cap outside of the rig to see if it is breaking down at normal operating voltages or whether it is a rig problem?

Is it producing normal output on 75 meters?  If the DC blocking cap were open this would create very high voltage at the tube anode but then you wouldn't have output either.

It is odd that it happens when "kerchunked" since the RF output isn't relay switched in the G-76 so it isn't a relay or sequencing problem.  Are you using the "official" Gonset AC supply?  Is the relay that kills the HV in receive working OK?  What is your plate voltage on the final reading?

Rodger WQ9E

[W9RAN]

Thanks for the comments guys.  The G-76 is on the factory p/s and the 75 meter output is clean up thru 100 mhz on the spectrum analyzer.  The arcing is from plate-to-plate just as you described, Don, and makes a very impressive zorch!  Under magnification you can see some real craters and molten spots in the plates, even though it's only happened for a few seconds at a time.  I've had some success in polishing plates too, but since a buddy has a parts rig, I think I'll swap this one out. 

I'd like to identify a root cause however.  Rob,  I hadn't checked that RFC, but in fact it was open, so I think you nailed that one, although I'm not sure it caused the problem.  The old Handbooks used to say that adding that choke  reduced the voltage rating for the plate tuning cap by providing a DC shunt across it, but mostly I remember the choke being installed as a safety precaution in case the plate blocking cap failed.  But my neutralization cap is on the other side of the blocking cap, so it's going to have to withstand both DC and full modulated RF voltages.  In any event, the cap still arcs after the choke has been replaced.

My bet is that something like dust created an initial path for arcing to begin on a voltage transient on the trailing edge of the HV, and the degraded plate surfaces now allows it to continue.

73, Bob W9RAN