Many of the old tube amplifiers designed and built during the old golden age, and some later designs too . . .
Push pull output tubes that either had fixed bias or fixed adjustable bias; but which were not separately adjusted for each tube;
Or, push pull output tubes with a single combined cathode self bias resistor (cathodes directly tied together);
. . . The one common manufacturer requirement of all those amplifiers was: you have to use very well matched output tubes.
The lesson for all who agree, and for those who disagree:
1. Use very well matched output tubes, Or . . .
2. Use separately adjustable fixed bias for each tube, Or . . .
3. Use separate individual self bias resistors for each tube, Or
4. Expect less than optimum performance.
Caveat Emptor
The above are just my personal opinions, and my personal observations.
Push pull output tubes that either had fixed bias or fixed adjustable bias; but which were not separately adjusted for each tube;
Or, push pull output tubes with a single combined cathode self bias resistor (cathodes directly tied together);
. . . The one common manufacturer requirement of all those amplifiers was: you have to use very well matched output tubes.
The lesson for all who agree, and for those who disagree:
1. Use very well matched output tubes, Or . . .
2. Use separately adjustable fixed bias for each tube, Or . . .
3. Use separate individual self bias resistors for each tube, Or
4. Expect less than optimum performance.
Caveat Emptor
The above are just my personal opinions, and my personal observations.
I might add to your comments and say that self/cathode bias appears absolutely useless if you want any sort of consistency in your amplifier performance from tube set to tube set. Even if I had bought a matched pair of the higher biasing tube, I would still be stuck with two channels of 4% THD @ 1W. This is truly abysmal performance after getting used to ~1% THD with all the tubes I had tried previously.
Unfortunately when buying matched pairs we don't have the option to say "I want the ones that bias at about 45V and 70mA".
Thank you all for the education, manual bias only from now on.
GregH2,
Are we talking about single ended outputs?
Or is your amplifier push pull?
Do you have a schematic handy that you can post here?
Thanks!
Are we talking about single ended outputs?
Or is your amplifier push pull?
Do you have a schematic handy that you can post here?
Thanks!
6A3sUMMER,
It's single ended and can bes used in triode/UL and with and without cathode feedback in the output stage. You can see an approxmiate schematic in post 23 of the thread below:
https://www.diyaudio.com/community/...-of-first-tube-amp-design.379321/post-6916707
The current circuit is slightly different - I upped the source follower load resistor to 40k, so its only running at 5mA or so. I also increased the screen resistor to 1k3 for U/L mode (still 100R in triode). The cathode resistance for the output stage is now 650R.
As I said, I have now run several pairs of EL34's, KT77's and KT88's in this amp from different manufacturers. The EL34's/KT77s have all biased within 1V of 37V or so (~60mA) and produce about 1.4% THD @ 1 Watt. All of the KT88's have biased within 1V of 45V (~70mA) and produce about 1.1% THD @ 1 Watt. The exception is the one tube I have been asking about, which biased at 50V (~77mA) and produced >4% THD at 1 Watt.
Even the cheapest and nastiest chinese EL34's biased within 1.5V (unmatched).
Note that I realise the KT88's could be run hotter but my xformer is at it's current limit already - I'm already planning the next build with a larger transformer.
Look forward to your reply,
Greg.
It's single ended and can bes used in triode/UL and with and without cathode feedback in the output stage. You can see an approxmiate schematic in post 23 of the thread below:
https://www.diyaudio.com/community/...-of-first-tube-amp-design.379321/post-6916707
The current circuit is slightly different - I upped the source follower load resistor to 40k, so its only running at 5mA or so. I also increased the screen resistor to 1k3 for U/L mode (still 100R in triode). The cathode resistance for the output stage is now 650R.
As I said, I have now run several pairs of EL34's, KT77's and KT88's in this amp from different manufacturers. The EL34's/KT77s have all biased within 1V of 37V or so (~60mA) and produce about 1.4% THD @ 1 Watt. All of the KT88's have biased within 1V of 45V (~70mA) and produce about 1.1% THD @ 1 Watt. The exception is the one tube I have been asking about, which biased at 50V (~77mA) and produced >4% THD at 1 Watt.
Even the cheapest and nastiest chinese EL34's biased within 1.5V (unmatched).
Note that I realise the KT88's could be run hotter but my xformer is at it's current limit already - I'm already planning the next build with a larger transformer.
Look forward to your reply,
Greg.
KT88 in SE and not running it at recommended voltage and current means that in your design it is not running in pure class A but somewhere in AB. No wonder you measure a difference when the tube bias turns out a bit different. No problem with the tube.
The Quad II would still have acceptable performance when there was 25% tube mismatch. https://www.keith-snook.info/quad-ii-valve-power-amplifier.html
The Quad II would still have acceptable performance when there was 25% tube mismatch. https://www.keith-snook.info/quad-ii-valve-power-amplifier.html
How does a single ended output stage run in class AB?
This amp has a low current, high voltage output stage mated to a 5k transformer. One possibility is that the the higher current of the tube I mention is saturating the output transformer....but it is supposedly rated for 150mA so should be ok in theory.
I dunno…. Is it still considered “class A” if it’s severely underbiased, and undergoes a significant shift up in DC consumption when driven hard? Along with 20-30% 2HD? I know that in the RF world with a tuned load it wouldn’t be. We’d call it AB.
Been thinking about this some more and I think my suggestion above that maybe the increased current of the "bad" tube is saturating the OPT might be on to something. I will try one of the "good" tubes at a higher current and see if the results are the same.
OK. I am now convinced this is a dud tube. My previous comments about the output transformer possibly saturating were incorrect.
I took the 5 other KT88's I have in my posession and installed them one by one, altering the cathode resistance to provide a standing current equal to what the "dud" tube had in my previous configuration. This required me to drop the cathode resistance from 650R to 550R. In all cases, the THD dropped with the increased bias current, as I would have expected. All 5 KT88's displayed under 1% THD, much better than the "dud" tubes 4-5%. So it definitely has nothing to do with the dud tube saturating the output transformer with it's increased current.
I took the 5 other KT88's I have in my posession and installed them one by one, altering the cathode resistance to provide a standing current equal to what the "dud" tube had in my previous configuration. This required me to drop the cathode resistance from 650R to 550R. In all cases, the THD dropped with the increased bias current, as I would have expected. All 5 KT88's displayed under 1% THD, much better than the "dud" tubes 4-5%. So it definitely has nothing to do with the dud tube saturating the output transformer with it's increased current.
GregH2,
Thanks for giving us conclusion as to the cause of the original problem!
I think we had some good discussion, and that we all learned something along the way through this thread.
Thanks for giving us conclusion as to the cause of the original problem!
I think we had some good discussion, and that we all learned something along the way through this thread.
Thanks 6A3sUMMER,
No problem and thanks for your comments which are generally pragmatic and a genuine effort to help or explain the problem.
It certainly has been interesting and I think a good lesson for everyone that they can't always expect consistent performance from their amplifiers/tubes.
Whether this "dud" tube falls within normal variability or not, the tube would not have performed to expectations in any amplifier. This is irrespective of manual or cathode bias, feedback or class A or B. It simply does not perform well under any conditions.....by a long shot.
I am fortunate to have an audio analyser that allows me test this. Others would have just thought it sounded a bit"off".