5E3 clone with pos. voltage on the grids

[melvin]

I am hoping someone can help me with a 5E3 Tweed Deluxe clone that is proving difficult.

After about ten minutes of playing this thing, when it's good and warmed up, it gets all mushy and distorted. It has to be closed-up (chassis in amp, back board covering chassis) for it to happen, mind you, but when I open it up, one 6V6 will be drawing a lot more current than the other and there will be positive voltage on the corresponding grid (anywhere btwn 3 and 10 volts). It usually calms down with the back off.

Switching the tubes doesn't make it change sides and it eventually happens with any set of tubes.

Also, I've already swapped out the two blocking caps coming from the phase inverter, the 220k resistors and the grid stoppers. I re-soldered the grounds too.

Can the socket or the output transformer cause this to happen? A resistance check on the output transformer gives equal readings. What gives?

[tubeswell]

So what are all the idle voltages (plates, screens, cathodes, grids)?

[melvin]

from what I recall: 360, 325, 22, and less than .1 until the problem starts. That's when the plate gets yanked down on one side and the grid goes upwards of 3 volts. I can get the exact voltages tomorrow, if you think it will give us a clue as to what might be going on.

[txstrat]

It usually calms down with the back off.

I find it odd that this only happens when the back is on.
How've you fixed the back cover? Maybe the screws or whatever go through the back cover and through the chassis and (if they are long enough) move some wire or stab some solder point inside, this might cause the weird behavior.
All in all it seems to be a mechanical issue to me.

Just my 2C

[MWJB]

Do you have some pics?

[tubeswell]

You better get all the voltages, not just the output tubes. Make it easier to figure out what's going on.

[ThermionicScott]

Try cleaning and tightening the tube socket pins. What do you measure for resistance between the offending tube's grid and ground? (Try leaving off the back panel until it acts up, power off the amp and measure it again immediately.)

- Scott