output tube dissipation unbalance in overdrive

[Dreams]

I've got an amp on the bench, a late 60's (?) AB165 Fender Bassman. It seems to be working fine; rebuilt power and bias supplies, all voltages good, 50 watts at threshold of clipping, good bias voltage and idle current, etc.

Current through the output tubes remains generally balanced from idle up to clipping, but one I start to overdrive the power stage, current in one tube goes way up while current through the other tube stays relatively the same. (It's a later "balance bias circuit, btw. I get around 100mv through the "fixed bias" tube and as much as 200mv through the "bias balance" tube.)

I've noticed that some amps do this and some don't. Could anyone speak about what causes this? Is it just....normal? Is this just gonna wear out one tube prematurely?

[shooter]

anyone speak about what causes this?

There is one check that you can do, swap tubes and make sure it "follows" the tube.  As to cause if that's "normal or not", it's an analog world

[Dreams]

Doesn't follow the tube; happens to whichever tube is in the one socket.

[pompeiisneaks]

Doesn't follow the tube; happens to whichever tube is in the one socket.

Then I'd check all the components in each side of the bias balance chain, maybe one of the resistors is a little flakey and as it heats up, tends to drift a bit?  causing less than favorable behavior under heavier load conditions?

~Phil

[tubeswell]

Also could be caused by the LTP inverter bias being to hot or too cold (which affects the PI duty cycle when the PI's inverting stage is overdriven, which can cause the red plating on one side of the output stage under overdrive conditions)


See Randal Aiken's post about this at TAG a few years back. Its also mentioned in Merlin Blencowe's 'designing tube preamps' book (2nd edition)


http://ampgarage.com/forum/viewtopic.php?f=6&t=10502&start=15

[shooter]

which can cause the red plating on one side

+1
if nothing pans out from these, swap the OT input wires (disconnect any NFB)
I had a bad OT that only failed when heated

[Dreams]

Also could be caused by the LTP inverter bias being to hot or too cold (which affects the PI duty cycle when the PI's inverting stage is overdriven, which can cause the red plating on one side of the output stage under overdrive conditions)

I think you hit the nail here. I ended up converting the bias to the older style blackface circuit, and the "problem" remained, so an issue with the bias supply was ruled out. Red herring; just a "coincidence" that the issue was on the adjustable side. Next I swapped the OT leads, also with no change. When i swapped the grids, the "problem" finally followed.

At this point I'm more questioning my test method than the health of the circuit itself; maybe slamming the hell put of the PI with 400Hz is not always the most ideal way to test all aspects of the output section. At the very least, I'm now familiar with the duty cycle imbalance of the LTP and at least some of it's effects. Ya learn something new every day.

Still not quite sure what to do about this, but it's late and I'm going to bed. Any other thoughts are appreciated. Thanks, everyone.

[sluckey]

Look at the output tube grids and/or the PI plates with a scope.

[tubeswell]

How is the LTP set up? What is the bias resistance (in your LTP)? If you're using 420R, try 820R. If you're using 820R, try 1k2.

[Dreams]

Sluckey: What would I be looking for on the scope? Evidence of uneven duty cycle?

Tubeswell: it's an ab165 Bassman, stock values. 470 PI bias resistor, all voltages seem correct:

Plates sit at 250v
Screens are at 70v
Cathodes are at 109.2v
Across the tail measures 106.8v

Do I even have a problem here? Is my problem with the way I'm testing it?

[sluckey]

You would be looking for differences between the two grid signals. Maybe high frequency oscillation on the grid of the tube with excessive current.

[tubeswell]

Tubeswell: it's an ab165 Bassman, stock values. 470 PI bias resistor, all voltages seem correct:

The 420R may be leaving the inverting triode too hot. Try 820R

[Dreams]

As designed and built, dead stock values, does the AB165 Bassman phase inverter run the inverting side too hot?

Increasing the PI bias resistor from the stock design cools down the inverting side, yes.

[sluckey]

If the tubes are not redplating under playing conditions I'd leave it alone.

[Dreams]

You would be looking for differences between the two grid signals. Maybe high frequency oscillation on the grid of the tube with excessive current.

No oscillation. Only difference is amplitude. Both clip at the same time, although the non-inverting side clips at a slightly lower amplitude.

Inverting clips at 30vac rms
Non clips at 27vac rms

[Dreams]

If the tubes are not redplating under playing conditions I'd leave it alone.

That's what I'm asking/saying. Thank you

[tubeswell]

As designed and built, dead stock values, does the AB165 Bassman phase inverter run the inverting side too hot?

Increasing the PI bias resistor from the stock design cools down the inverting side, yes.

Fender 'designed' many their LTP stages running 12AT7 tubes with 420R bias resistance, but in amps using 12AX7 LTP tubes (many of their 'Brown series amps), they used 820R. Not sure what type of tube you're running in the LTP, but if the output tubes are red-plating under overdrive and it is (evidently) related to the PI set-up, then if you want to cure the red plating, try going to a bigger size, (like 560R, 680R, 820R or even 1k).