Bias current vs. voltage

[Lectroid]

I'm just finishing a BF Princeton Reverb-ish clone with 432V plate voltage. For 6V6s, 70% of 12W max dissipation is 8.4W. I've adjusted the bias to where I'm showing cathode currents of 20mA and 18mA respectively.

Taking 19mA as an average, this gives me:  432V x  0.019A = 8.2W
I can certainly live with that.  But when I measure the actual bias voltage at the grid pin, it's not Fender's -34.  It's around -37.5V.

If setting the cathode current is the accepted way to proceed here, why this discrepancy with the bias voltage? 

[Willabe]

1st off, who says Fender biased the tubes for 70%?

2nd, different tubes from different makers and even from the same maker will bias up differently, that's why we go by current, not voltage.

PRR wrote here before that the only difference between a 12AX7/12AT7/12AU7 was the distance between the control grid and the cathode. They all had the same guts and their voltage and current outputs were pretty different. So I'm thinking it's a matter of how closely they can line up the K, grids and plate in each tube.

Fender, at the time, was buying really good RCA tubes. It was the golden era of tubes and the tubes they got were all pretty close to each other current wise. RCA's machining was still very tight probably.

So Fender probably had a vdc -bias that they found to work with all the tubes they were receiving at the time from RCA. They probably had it a little low just in case they ran into some hotter tubes. And Leo liked clean, not distorted sound.       

[tubeswell]

Fender, at the time, was buying really good RCA tubes. It was the golden era of tubes and the tubes they got were all pretty close to each other current wise. RCA's machining was still very tight probably.

And there was a high reject rate of tubes that weren't 'within spec'. Piles of otherwise good-enough tubes were trashed, all in the name of ensuring quality control. The days of plenty.

Testing from 31:07
https://www.youtube.com/watch?v=GDvF89Bh27Y

[pdf64]

Anode current can be derived from the published triode anode characteristic plots. Just use the closest control grid plot, use the screen grid voltages up from the x axis and see what anode current corresponds to that on the y axis.

Using that method and Fender schematics, they seem to idle 6V6 hotter, and 6L6GC colder, than might be expected from the 70% guideline.
The above method seems to align well with online loadline calculators, eg https://www.vtadiy.com/loadline-calculators/loadline-calculator/

[Merlin]

• Tubes are not perfectly identical when new
• Tube bias difts with the age (usage) of the tube
• B+ and OT winding resistance will affect the bias voltage needed
• Meter loading will affect the measured voltage. Fender was using a low-ish impedance meter but you're probably using a much higher impedance modern meter.

You have nothng to worry about.

[Lectroid]

@willabe,

1st off, who says Fender biased the tubes for 70%?
2nd, different tubes from different makers and even from the same maker will bias up differently, that's why we go by current, not voltage.

So most 'discrepancies' are mostly due to normal tube variation and biasing for current is the correct process.  Thanks for validating that. Thanks for the history review too.   I grew up 15 miles from Fullerton and revered Fenders.  I tend to forget they aren't perfect exemplars of the amp-maker's art.  Just commercial products, made as cheaply as possible.

@tubeswell,
I loved that Mullard video--never seen that automated tube manufacture before.  Amazing use of robotics for their time.  I had no idea they were that advanced.

@pdf,
That technique for finding anode current sounds useful. I have to try it.

@merlin,
Thanks.  As always, comprehensive and concise. 

Q: If an early tech had had no way to actively measure the bias current in a BF PR, wouldn't they have been fairly "safe" if they biased to reach -34V grid voltage?

[Willabe]

So most 'discrepancies' are mostly due to normal tube variation and biasing for current is the correct process.

What 'discrepancies'? There's no 'discrepancies'. We already went through this. The normal thing is to bias tubes up by current. Fender picked those #'s for the tubes they were buying. They checked enough of them and then probably gave them a little grace, lowered current draw.

I tend to forget they aren't perfect exemplars of the amp-maker's art.  Just commercial products, made as cheaply as possible.

They were very good examples of tube amps. Very reliable, held up on the road, sounded great, looked good, easily serviceable.

They were not made as cheaply as possible, Leo could have made them cheaper, but then they wouldn't have held up on the road or even in town gigging. Think of Valco and others. He found a balance of making a good looking, good sounding amp that held up for the money. And the sales would seem to prove that.

Q: If an early tech had had no way to actively measure the bias current in a BF PR, wouldn't they have been fairly "safe" if they biased to reach -34V grid voltage?

Already been explained. If back then they got a hold of the same RCA tubes, then probably yes, but you ALWAYS check bias when installing new power tubes, ALWAYS.

Put in a new carburetor in a car, you set the corroborator for idol. Same thing.

[Latole]

I never bias tube ( in class AB ) over 60 % .
70% is a too hot

[Lectroid]

@willabe,

So most 'discrepancies' are mostly due to normal tube variation and biasing for current is the correct process. 

What 'discrepancies'? There's no 'discrepancies'. We already went through this. The normal thing is to bias tubes up by current. Fender picked those #'s for the tubes they were buying. They checked enough of them and then probably gave them a little grace, lowered current draw.

Maybe I expressed myself badly.  Putting 'discrepancies' in quotes was my attempt to acknowledge what you and  everyone had told me: that there is no 'discrepancy' in the way I was asking about it. I guess I need to work on my phrasing.   What I'm walking away from this thread with:

- Always bias for current not voltage. 
- Do not bias to match the "bias voltage" shown on a schematic.   
- Minor variations are due to the items on Merlin's list and probably to be expected.

Thanks for the answers and again, sorry for the confusion.

[Fenders] were very good examples of tube amps... And the sales would seem to prove that.

I wouldn't argue that in any particular. I like Fender's BF amps a lot.  I was trying to say that they were mass-market consumer products, built subject to all the pressures of making and selling a quality product with competitors all around them.  It's kind of a compliment to what Leo and his people were able to accomplish.