Bizarre biasing problem !!

[hamfist]

I am trying to sort an old Selmer 100W amp, that I have been converting into a two channel "Marshall beast"-type amp.

THe power section is 4 x EL34.

I just can't get the right hand pair to bias the same as the left - the right hand pair always has a much higher plate current.

Whatever tubes I put in, and swap around, it's always the right hand pair, so the problem is associated with those sockets, not the tubes at all.

I've checked all the bias supply, and all the resistors are the correct values.

Resistors coming in to pin 5 on each tube are correct value.

All my 1K Screen grid resistors are new, and good.

Voltages are identical between all tubes.

I even swapped the OT connections on pin 3 over (right pair to left, and vica-versa).... but still the right hand socket pair has a much higher current.

I can't think of anything else that could be different between the two pairs of tubes, except perhaps a dodgy socket

Am I missing something obvious here ?

[stokes]

I would remove the tubes and check the negative voltage at pin 5 of each socket and see if they all match.If the problem is in the tube sockets and not the tubes,they shouldnt.If they do replace the tubes and check again,if they dont match,check the voltages at the tube side of the 220k resistors,if they dont match there,one of those resistors is the problem.Theres not much else to that bias supply.

[hamfist]

I would remove the tubes and check the negative voltage at pin 5 of each socket and see if they all match.If the problem is in the tube sockets and not the tubes,they shouldnt.If they do replace the tubes and check again,if they dont match,check the voltages at the tube side of the 220k resistors,if they dont match there,one of those resistors is the problem.Theres not much else to that bias supply.

Without the tubes, all voltages on all sockets match at Pin 5 .....31V

It should be simple, but I'm flummoxed by this one !

[Gtr_tech]

Check the coupling caps coming from the phase inverter for leakage. Of one of those is even a bit leaky it'll screw with output tube bias.

[R.G.]

Check the coupling caps coming from the phase inverter for leakage. Of one of those is even a bit leaky it'll screw with output tube bias.

+1.

Unsolder the coupling caps from the PI at the output tube end. Now check bias. All the same? If yes, it's the coupling caps.

[hamfist]

+1.

Unsolder the coupling caps from the PI at the output tube end. Now check bias. All the same? If yes, it's the coupling caps.

Another thing I have done since my first post is to swap over the PI coupling caps .... you guessed it .... No difference !

[Gtr_tech]

There's not that much to it....either something is miswired or there's component(s) out of spec. Which model is this anyway?

[Gtr_tech]

This one? http://www.schematicheaven.com/barga...diacsv100w.pdf

[hamfist]

Sometimes ( well quite a lot of the time when I working on my amps, actually), I feel like a complete doofus !!!

Igot by bias probes out to check the bias that way......and the upshot of it all is that the bias is fine between the 4 tubes. I have learned that my DMM measures very inaccurately when down at 1 ohm. THe cathode resistors I put in, must have been wildly different between the two pairs of sockets.

What threw me is that sockets 1 & 2 measured the same, using the CR method, and sockets 3 & 4 measured the same using the CR method......truly as if the problem was between the two pairs.

Ah well, another lesson learned. I hope none of you is as dumb as me !!

[Gtr_tech]

ANother thing....what condition are the sockets in? Thats an old amp...they may be due for replacement. MAybe clean and retension to see if that affects the behavior, but fresh ones would be a good idea.

[Gtr_tech]

I don't do the K resistor thing anymore. They get warm, drift,etc. I built a gadget for biasing that has an octal plug wired to a small box with the socket in it. The plate connection is interrupted and brought out to banana jacks on the box. I put a jack on there for plate V monitor as well. So I can read both current through the tube and plate V to calculate idle dissapation. You need 2 meters, but it sure speeds up the process. On old Fenders I don't even need to pull the chassis to check/adj. bias.