AB167 bias balance

[Randall]

I've got a '67 Bassman AB165 head on the bench for a recap and a mighty cleaning. It has been sitting for decades unused. I finished the work, and was checking the tubes and bias. The old set of EH 6L6s seem fine and reasonably matched, but are idling at around 12 watts, too cold for me. It has that blasted bias balance contol wthat only adjusts one tube. I am getting -60v off the diode. Customer wants to keep it as original as he can, so I doubt he is going to want me changing to the black face bias circuit. But, I doubt he would mind if there was a simple change to lower the bias supply.

Would increasing the 470 ohm 1 watt on the bias board be the easy answer?

http://ampwares.com/schematics/bassman_ab165.pdf

[Leo_Gnardo]

There's a bias setting resistor on the back of that balance pot, right? Has it drifted upwards in value? They often do. If it has you can swap in the correct "what's on the blueprint" value. Which might nudge things in the right direction. Or you can always tack a parallel resistor on that puts you in a better bias zone altogether. Be sure to use a period-correct 10% carbon comp half watt resistor. And charge accordingly.

[Helmholtz]

What are actual grid voltages?

[Enzo]

And why is that too cold for you? How does the amp sound? The guy wants it stock, and it has been this way for over 50 years now. I wouldn't hand it back and say "Oh, I changed things..."

[christarak]

The old set of EH 6L6s seem fine and reasonably matched, but are idling at around 12 watts, too cold for me.

Would increasing the 470 ohm 1 watt on the bias board be the easy answer?

http://ampwares.com/schematics/bassman_ab165.pdf

I wouldn't make any changes until you confirm that it is too cold. Too cold, to me, means that you are getting into the crossover distortion point. If you have access to a signal generator and oscilloscope, see if there is a slight notching at the zero point. If so, you're right...it's too cold. If there is no notching, I would definitely not change the bias values. Making the bias less negative will decrease the clean headroom that your customer loves about this amp. You might like a more "chewy" sound that a warm/hot bias gives, but your customer may not. I, for one, prefer a cooler bias.

[Helmholtz]

I agree that it is most important how the amp sounds, but...

meanwhile actual mains voltage may be about 10% higher (check heater voltage) and resistors may have drifted by 10% or more.
This easily results in an increase of bias voltage of 20% which roughly decreases idle currents by 30%-

I wonder if running the tubes at 30% lower current than designed can still be considered "stock".

[Randall]

My grid voltages are -51v. I don't believe this fella has ever played this amp, so he has nothing to compare it to.

By resistor on the back of the bias pot, are you talking about the 15K on the shell? If so, it is clocking in at 17K. There has already been resistor replacement, about 15 of them, I'm not going to worry about period correct, since non of those are.

I will run a sine through it later and report about the crossover dist.

[vintagekiki]

http://ampwares.com/schematics/bassman_ab165.pdf

The "lower" 6L6GC receives a fixed bias voltage via a 10k resistors.
The "upper" 6L6GC receives variable bias voltage via bias balance pot 10k.
Bias for both tubes at Bassman AB167 determines the resistor 15k from the bias balance pot 10k to ground.
Higher resistor from 15k = higher bias voltage = lower quiet current
Smaller resistor from 15k = lower bias voltage = higher quiet current

[Randall]

Good idea, that way it is easily reversed if anyone wanted to. Danke.

[christarak]

That's actually only around 10% from the schematics value of -45V. You could leave the 15k in place and just wire a 100k in parallel - if sound improvement justifies the mod.

An excellent suggestion. I use my old resistor decade box parallel to the 15K and set the the value just warm of no cross-over distortion. When the amp gets booked in, one of my stock questions is what sound are they after (ie Clean or Early Breakup). I then use my ears for the final taste test before soldering in that final parallel resistor value. 90% of the time it comes in at the Fender defined voltage or extremely close to it. Those guys sure knew their stuff.

[Helmholtz]

BTW, is there actually an AB 167 version? Or did you mean AB 165 like the schematic?

[Randall]

Sorry, I meant AB165. I fixed it.

So the resistor I paralleled with the 15K on the pot, that had drifted up to 18K ended up being 47K. 100K didn't have much of an effect. That combination puts me at 15.5 watts, just a smidge over 50% max diss., which is about where Fender sent amps out. Interestingly, my bias voltage is still a little higher than the schematic -45v, but I'm good with it now. I will explain to my customer what I did and why, and if he doesn't like it, it is no big deal to remove it, but I think he will approve.

Thanks guys.

[pdf64]

For a given plate current, bias voltage is proportional to the screen grid voltage; if the magnitude of the latter is X% greater or smaller than the schematic nominal, then the bias will need to follow suit.
So if the mains voltage is X% different to 117Vac then screen grid voltage is unlikely to be 425Vdc and so, for bogey tubes, -45Vdc at the bias supply outputs will be 'incorrect', ie it's wrongheaded to view the bias voltage in isolation, it has be be seen in the context of the the other stuff that constitutes the tube's operating point.
https://el34world.com/charts/Schemat...b165_schem.pdf

For class AB1 amps, setting bias according via solely via crossover kink elimination is somewhat dodgy in my view, as the transfer function will tend to inherent non linearity in the transition area. And high signal levels, ie approaching and beyond clipping will increasingly cause bias shift, which it's kinda topsy turvey to try and bias out, eg https://youtu.be/VZor5pDOy3Q?t=4925

[Chuck H]

So the resistor I paralleled with the 15K on the pot, that had drifted up to 18K ended up being 470K. 100K didn't have much of an effect.

???

I'm thoroughly confused by this. The 100k, being a lower value, should have reduced bias voltage more (more 'effect") than 470k. Did you mean 47k?

[Helmholtz]

For a given plate current, bias voltage is proportional to the screen grid voltage; if the magnitude of the latter is X% greater or smaller than the schematic nominal, then the bias will need to follow suit.

Interesting.
I wouldn't have expected such a simple and general relation. Where did you find this?

Consequently, when mains voltage rises by 10%, bias as well as screen voltage can be expected to rise by 10% and both effects on plate current should compensate.