Output Current Mystery

[Firestorm]

I've never seen this before (maybe because I never tested for it before), so I wonder if anyone can hazard a guess as to the cause of the following:

This is a 4 x KT88 ultralinear power amp. Each tube is individually biasable and independently measureable via precision 1-ohm resistors on the cathodes. At idle, all of the tubes draw a nice stable 45mA each. But in measuring current while the amp was being played at a moderate volume, one side of the output ranges from 75mA down to 25mA (ain't Fluke meters wonderful?) -- still obviously in the Class A portion of AB -- BUT the other side goes up to 240mA and down to 0mA with the same signal.

Tubes are new and matched, the PI is very well balanced, but the current draw on one side is dramatically huge. Yes, the symptom stays with the sockets.

My gut tells me it's a transformer issue, but I haven't checked it yet. Any other possibilities?

[Andy Le Blanc]

What type of PI? Is the PI balanced for both DC and AC?

[Firestorm]

PI is a long-tailed pair (12AT7) with a somewhat larger than normal (by Fender or Marshall standards) tail resistor to better balance the two sides. Plate resistors are both 82K so it seems the DC is balanced and with a 40K (if I remember right) tail resistor, the AC should be better balanced than a standard Fender PI. There is negative feedback on the second input to the PI (and a presence control across the feedbacl load resistor).

[dobbhill]

I'd double check the AC balance first.
Were the power tubes matched at idle and under load?
D

[Jana]

Use a scope and see what the two sides of the phase inverter are doing, check their levels, etc.

Maybe the power tubes are just reproducing what they are being fed.

Have you swapped out the PI tube? Maybe one half of it is going south.

[Firestorm]

I'd double check the AC balance first.
Were the power tubes matched at idle and under load?
D

I'll check, but -- the AC balance would have to be incredibly, hugely, enormously out of whack to drive the out-of-phase side 700% hotter than the in-phase side. Maybe parasitic coupling.

It ain't the tubes -- I've swapped them around in every possible permutation and the issue stays with the same sockets.

[Andy Le Blanc]

With a long tailed pair you can't balance both DC and AC at the same time at the inverter.
Install a pot and balance the AC at the inverters plate, double check the
inverter tube you can get large differences between halves of twin triodes.
And double check the inverter circuit for error's.
If it's not the tubes and not the AC balance, it's probably the OPT as you suspect.
I still wouldn't expect an ideal balance but what you've described is out there.
Have you double checked the OPT hook up, its so far out that its either blown
or a boo boo in the dress.