Should Voltage Sag Be Even Across all Power tubes?

[Zer09]

I replaced (upgraded) the output transformer in my 5e8a twin today and while I had the chassis out i tried a few different rectifiers and watched plate and screen voltages as i played. I didn't see what I expected and I'm assuming there is a problem somewhere (OT?). I found one socket dropped significantly more plate voltage than the other. I swapped tubes and it did not follow the tube so it's specific to that socket. I played at volume levels between 1 and 8 and while one socket was dropping under 40 volts on heavy chords the other was dropping up to 100+ with a max of around 130v. Drop in screen voltage was similar (a few volts) on both sockets. I normally use a single 5U4. I tried a GZ34 as well as Weber Coppercap versions of both rectifiers and it made no difference. I'm guessing it's the new OT. Does anyone have any other ideas before I swap it out again? I guess swapping the primaries will tell me all I need to know?

[trobbins]

Trying to measure plate voltage is not a good idea imho. Perhaps use a known matched pair of 6l6's if you're concerned.

If you're happy to go in to testing, then check the bias level is constant. The screens are at the same voltage level, unless you've added screen stoppers. If you're happy to add cathode sense resistors then you can check if your 6L6's are idling at similar level - if you then join the cathodes and sense the cathode voltage across the parallel sense resistors with an oscilloscope then you may be able to check for assymetry with a sine wave, which would indicate poor matching of the 6L6's. But in essence the amp is mainly wanting matched 6L6's.

[Zer09]

I should have included that in the original post. They are matched within 2 ma. I even tried different matched pairs. There are 1k scrren resistors and the amp has 2 watt 1ohm 1% resistors for measuring bias current. Bias was measured using 3 different DMM's all with similar results.

[loudthud]

If you are using a DVM to measure the plate voltage, what you are seeing is a difference in duty cycle.

Guitar amps do not distort evenly. If you feed them with a sine wave, you get a distorted square wave with about a 40/60 duty cycle. This make the tone not sound hollow like a 50/50 square wave would. The DVM takes an average of the AC voltage on the plate. One plate has a 40/60 wave, the other has a 60/40 wave.

[EFK]

Does that socket run with a voltage drop at idle? My simplistic thought is that you either have an oscillation in the wiring to that socket only or more likely the DVM itself hanging off the plate is causing that socket to oscillate.

[diagrammatiks]

you can't get an accurate reading with a digital multimeter.

you need at least a vacuum tube volt meter.

[Zer09]

I've narrowed it down to an oscillation in the PI circuit. It's not audible but I can see it on my scope. I'm guessing I either disturbed something while routing the wires for the new OT or it's the lead dress of the new wires. It starts at the treble pot wiper which connects to a PI grid and I can induce an audible hum by poking that wire. That pot has always been noisy. I just happened to have ordered a pot to add a dwell control to my Deluxe Reverb so I'm going to use it to replace this one and rewire the PI socket first and take it from there if it doesn't solve the problem.

[trobbins]

Maybe a 10k grid stopper on that 12AX7 driver to the PI would help. But lead dress and screening of longer cables from that point would help.

[Zer09]

I spent hours trying all sorts of different thing including replacing coupling caps, rewiring the phase inverter, and on and on. I finally lifted the board to and was able to get slight improvement by re-routing the plate wires which led me to where I should have started in the first place...approximately a 10-15 degree turn of the new OT. Problem solved.