Some 6L6 observations

[GainFreak]

Hi,

The following observations concern only the power output of the tubes mentioned not how they sounded.

I had to retube and amp recently and I got a TAD 6L6 red base quad. After installing it I noticed (on my scope) that power output was ~60W at 5% THD and there was no way to push the amp anywhere near to 90 or 100W even at 10% THD.
I scratched my head for some time then pulled a quad of JJ 6L6 from my amp. No problem getting 90W, hard pushed even 100W.
I put the TAD quad in my amp and got the same lower output (~60W).
I have a new unopened 6L6EH quad I keep just in case so it was next one to be tested. Although a bit better than the TAD it had the same lower than expected output.
In all cases the tubes were biased the same way, anode voltage at idle in both amps was around 450V.

So what gives?

[The Dude]

I would expect slight variations- even with 2 sets of tubes from the same manufacturer, but that's pretty radical. I suspect maybe some of the tubes are bad. It's not uncommon to see defective new tubes these days. When you say, "In all cases the tubes were biased the same way", did you check to see that each tube was conducting? Or, did you mean that you didn't readjust bias?

[GainFreak]

I would also expect slight variations but not 20 or 30W difference. The two quads were brand new. Measured the current while at max power - all were conducting.
By biased the same way I meant they were at 35mA at idle.

[g1]

Do you know what the value of the screen resistors in the amp was?

[Helmholtz]

Power output of a given amp is influenced by the (little known and even less controlled) plate saturation voltage of the power tubes.
This is the residual plate voltage at the peak signal current or the plate voltage where the loadline hits the Vgk = 0 plate curve.
The higher the value the lower the available signal voltage swing and thus the lower power output.
Zollner found that some newer production tubes have excessive plate saturation voltage.
If e.g. the TAD tubes have a saturation voltage being 60V higher than typical, I would expect power to be lower by 30%.

I would try to return the TAD tubes as they obviously don't correspond to 6L6 specs.

[The Dude]

Agree. Closer to 6V6 territory than 6L6.

[GainFreak]

Do you know what the value of the screen resistors in the amp was?

They were the typical 470 Ohm resistors. In my amp they are 1k because sometimes I put there EL34s but still the same result.

Power output of a given amp...

Thanks for explaining that. I already returned the TADs.

[GainFreak]

Two more experiments:

1/ Ran an old beat up quad of EL34EHs in this amp and got 80W at 5% and 90W at 8%.
2/ Ran the 6L6EH quad in my amp where I have a 360V tap on my PT and was able to get 100W at close to 500V at idle.
So obviously these tubes can do 100W but not at 450V. However from what I know running new production tubes at 500V is kind of... not recommended.

[Helmholtz]

Output power is determined by B+ minus Vsat.
So a 50V higher B+ compensates for a 50V higher Vsat.
But of course that's cheating as any tube will deliver more power with increased B+.

[uneumann]

Vsat makes sense, but there is also the question of whether the tubes are actually driven into saturation (or grid current). I didn't see mention of what amp this is and whether you've checked to see if the PI drive is sufficient. Just a guess, but perhaps PI or something upstream is limiting drive. You'd still see distortions, but they could be coming from earlier stages.

[pdf64]

Yes, higher gm valves will need a higher magnitude bias voltage to keep current reigned back.
So the preceding LTP stage would need a higher signal output voltage to reach Vg1-k=0, ie AB1 clipping with a through the knee loadline.

[Helmholtz]

Yes, higher gm valves will need a higher magnitude bias voltage to keep current reigned back.
So the preceding LTP stage would need a higher signal output voltage to reach Vg1-k=0, ie AB1 clipping with a through the knee loadline.

Yes, it would be interesting to know the grid bias with the 35mA idle current as well as the max. clean output of the PI.
But regarding your statement that higher gm requires more negative grid bias for same current, I'm not sure.
An EL34 has almost twice the gm of a 6L6 but needs less negative grid bias.
So Ip/Vgk doesn't behave like dIp/dVgk = gm.

[GainFreak]

Just a guess, but perhaps PI or something upstream is limiting drive.

In my first post I wrote that all tubes mentioned were tested in the same amp, PI included.

[g1]

It's not a question of whether the PI tube itself is bad, but what it's circuit drive capability is, and whether it is limiting some output tubes but not others.
You said you re-biased each set of power tubes so that the idle current was the same. But we need to also know how much variation in the bias voltage (to power tube grids) was needed to achieve that.