[Amps] Strange behavior (at least to me)

[Steve Bookout]

Hello all,

Since it's gotten quiet here, i thought I would get the collective 'gray 
matter' working again.

I am having a strange problem with a homebrew 8877 amp.  It has worked 
almost flawlessly for 30 years with the same tube and still easily puts 
out legal power.  4000 volts at a load of about 550 ma., driving with 
about 22 watts.

During the night shift in CQWW (I believe) the amp started acting up and 
was replace with a spare.  It is a 40 meter mono band amp.

I was told that there was no output but 'the meters still were 'jumping 
around''.  I opened it up expecting to find an obvious issue with the 
vacuum relay on the output.  Power has to go SOMEWHERE, right?   All 
looked as good as it ever did.

Just before Dayton, myself and two budddies, fired up the amp, after 
once again looking inside.

This is what I am seeing:

We tuned it up with 10 watts drive into a oil filled dummy load and got 
700-750 watts out with a few hundred mils of plate current and about 10 
mils of grid current.  Unkeyed it and then rekeyed it. Upon rekeying, 
the 100 mil grid meter slammed and the grid trip circuit did it's job.  
I have it set for 75 mils of grid current. Wound up reducing the grid 
capacitance so that the max grid current would be more reasonable.    
After trying it on/off again and again, when it goes 'nuts', the grid 
current stays on the scale at about 60 mils and about 'just a little 
bit' when it seems to conduct normally.  When it acts up, if I turn the 
TUNE vacuum variable just a wee bit, I get my plate current as I should 
and the grid current goes down to normal.  My multi turn counter on the 
TUNE cap is marked '0 to 100' on the knob skirt and a 'wee bit' is from 
88 to 96, just 8 marks, or 8% of one turn.  This is on a 375 pf vac 
variable cap.
Taping on the chassis would make the problem come and go, but not every 
time and not easily.    Try it again and I would have to turn the cap 
back the other way to resonate it (seemingly.)

In my mind, it seems like something in the RF side of things, since 
changing the cap brings it back from the 'dark side'.   Keying the amp 
with no drive and swinging the caps back and forth showed no 
instability.  I also took all the connections in the tank circuit apart 
looking for loose joints and corrosion.  None found. Reseated the tube.
Works fine every time till it doesn't.

It's doing the same thing as I think I would expect to see if I manually 
de-tuned the tank circuit with the input cap.  It also showed plate 
current during the 'event', although lower than normal, so where is the 
power going?   Quiescent plate current is about 120 mils.

Any ideas out there?  I can supply pics if needed.  Maybe I should 
record the event and put it on you tube with a link.

Look for some of you this weekend in the WPX CW contest.

73 de Steve, NR4M