Yes, that was what I was trying to point out, but failed to specify for clarity. Thanks for the clarification!
Fixed (but adjustable) bias keeps the output valves' idling point the same at all levels of signal. Cathode resistor bias with bypass cap causes the valves' operating bias to change somewhat with signal level. In purely Class A operation it doesn't change much (none at all for a magical distortion-free stage) but for Class AB, bias voltage increases with signal, and significantly with signal above the level where one or the other valve approaches cutoff. As pointed out in post #12 this means that valve operating point is moving around with signal level at higher signal levels, and restoring to small signal operating point over an RC time constant often not too much greater than audio.
Cathode resistor bias without a bypass cap (like a true DTN Williamson) doesn't have this effect, but requires operation pretty much in Class A, and a common cathode resistor. DTNW had this all worked out before I was born.
Our obsession with peak power output, always! expressed on a linear scale, drives us to worry about a 1 or 2dB difference between peak power outputs often to the detriment of more important issues.
Opinionated rant mode off/
All good fortune,
Chris
Cathode resistor bias without a bypass cap (like a true DTN Williamson) doesn't have this effect, but requires operation pretty much in Class A, and a common cathode resistor. DTNW had this all worked out before I was born.
Our obsession with peak power output, always! expressed on a linear scale, drives us to worry about a 1 or 2dB difference between peak power outputs often to the detriment of more important issues.
Opinionated rant mode off/
All good fortune,
Chris
Last edited:
I found cathode bias sound mellow, it by far prefer to know my amplifier is in fixed bias.
A cathode resistor can go to smoke when the tube fails, this is not a problem with fixed bias (or should less be one)
With a fix bias you will have warning signs before the tube fails because it will be hard to bias or run away, a cathode bias will run without letting you know anything and take out the resistor in smoke.
A cathode resistor can go to smoke when the tube fails, this is not a problem with fixed bias (or should less be one)
With a fix bias you will have warning signs before the tube fails because it will be hard to bias or run away, a cathode bias will run without letting you know anything and take out the resistor in smoke.
Yes, if you,ve sized it right. Many online schematics for 5w resistors here, which are great if you want to watch a tube melt down...With fixed bias, the burnt resistor is able to be planned for, as another safety mode. Common values used for bias current metering, like 10, 1 ohm, in fixed bias PP designs are not seeing very much power. 125mW for example, with a 10ohm to ground and a tube running -50v/50ma bias. Therefore the cathode resistor can be sized accordingly to where if the tube starts to run away, the resistor will fail open and the runaway will cease.
Both fixed bias and cathode self bias have tradeoffs.
Both of them can be subject to bias shift problems, depending on the circuit specifics, especially if they have RC coupling from the drivers to the output stages, and depending on how hard the driver tubes drives the output tubes.
I prefer individual self bias when possible. And my latest balanced amplifier with shared self bias only has one bias shift cause. It is exactly the same cause as fixed bias that uses RC coupling.
Believe and trust everything you read on this web site, and then verify if it is true or not.
Trust but Verify.
That goes for all my postings also.
All Generalizations Have Exceptions.
The statement: "fixed bias is better than self bias" is a generalization. Generalizations are often true. But sometimes they are not true.
Trust but verify.
Generalizations allow us to make easy decisions, but not always the best decision for a particular topology and circuit.
Both fixed bias and self bias have other tradeoffs that have nothing to do with bias shift. You have to know what those tradeoffs are, before making an intelligent decision about which is better; regardless of whether there is bias shift or not.
Just my $0.03
Adjusted for inflation.
Both of them can be subject to bias shift problems, depending on the circuit specifics, especially if they have RC coupling from the drivers to the output stages, and depending on how hard the driver tubes drives the output tubes.
I prefer individual self bias when possible. And my latest balanced amplifier with shared self bias only has one bias shift cause. It is exactly the same cause as fixed bias that uses RC coupling.
Believe and trust everything you read on this web site, and then verify if it is true or not.
Trust but Verify.
That goes for all my postings also.
All Generalizations Have Exceptions.
The statement: "fixed bias is better than self bias" is a generalization. Generalizations are often true. But sometimes they are not true.
Trust but verify.
Generalizations allow us to make easy decisions, but not always the best decision for a particular topology and circuit.
Both fixed bias and self bias have other tradeoffs that have nothing to do with bias shift. You have to know what those tradeoffs are, before making an intelligent decision about which is better; regardless of whether there is bias shift or not.
Just my $0.03
Adjusted for inflation.
Using fixed bias, a tube failure has the possibility to burn out your output transformer.
Choose, burned output transformer, or burned cathode resistor.
The exact answer is not simple to determine. It requires very good circuit analysis, and very good understanding of the various types of tube failures.
Here come the generalizations to the rescue; or here come the generalizations to steer you to the wrong decision.
I worked with engineers that were part of one of the world's best failure analysis groups. Just my opinion, but they were Very Good!
The military loved our reliability of our products.
Choose, burned output transformer, or burned cathode resistor.
The exact answer is not simple to determine. It requires very good circuit analysis, and very good understanding of the various types of tube failures.
Here come the generalizations to the rescue; or here come the generalizations to steer you to the wrong decision.
I worked with engineers that were part of one of the world's best failure analysis groups. Just my opinion, but they were Very Good!
The military loved our reliability of our products.
Do we worry about the output transformer's primary impedance? Yes, we need to.
Let's use a 6000 Ohm Plate to Plate primary as an example. Using a non-inductive 8 Ohm load resistor on the 8 Ohm tap, the amplifier is in heaven.
Now, lets use a typical 8 ohm 2-way loudspeaker, either closed box or ported. There is one peak impedance at bass frequencies for closed box, and 2 peak impedances at bass frequencies for the ported box. Those bass peak(s) might be 24 Ohms. on either side of those bass peaks, the impedance might be 6 Ohms.
And at the woofer to tweeter crossover frequency, perhaps at 1500Hz or 2000Hz, the peak impedance at the crossover might be 16 Ohms.
Speaker impedance; Output transformer's Effective primary impedance
6 Ohms; 4500 Ohms; woofer minimum impedance, same as woofer DCR
8 Ohms; 6000 Ohms; at lots of woofer and tweeter frequencies
16 Ohms; 12000 Ohms; at the crossover frequency
24 Ohms; 18000 Ohms; at the woofer impedance peak(s)
Life is not simple here
Let's use a 6000 Ohm Plate to Plate primary as an example. Using a non-inductive 8 Ohm load resistor on the 8 Ohm tap, the amplifier is in heaven.
Now, lets use a typical 8 ohm 2-way loudspeaker, either closed box or ported. There is one peak impedance at bass frequencies for closed box, and 2 peak impedances at bass frequencies for the ported box. Those bass peak(s) might be 24 Ohms. on either side of those bass peaks, the impedance might be 6 Ohms.
And at the woofer to tweeter crossover frequency, perhaps at 1500Hz or 2000Hz, the peak impedance at the crossover might be 16 Ohms.
Speaker impedance; Output transformer's Effective primary impedance
6 Ohms; 4500 Ohms; woofer minimum impedance, same as woofer DCR
8 Ohms; 6000 Ohms; at lots of woofer and tweeter frequencies
16 Ohms; 12000 Ohms; at the crossover frequency
24 Ohms; 18000 Ohms; at the woofer impedance peak(s)
Life is not simple here
Last edited: