Calculating gain of class A and class B region of class AB output stage

Hi Max

With all these, I was just thinking about SE last night. 3 6L6GC in parallel, option of using parafeed to get rid of the DC current constrain. 300V to 350V +B, 200V screen, Local NFB to lower the output impedance of pentode mode ( no UL), might consider GNFB to raise damping factor more..........I might get like 20W of class A. Too bad I don't see any reasonable priced OPT that has primary resistance below 1K, or else, I can think of 4 tubes in parallel.

Maybe I should build two, one PP and one SE. I just hope with LNFB and maybe GNFB, damping factor can be low enough to tame my speakers. I really think 25W is enough. The crappy kid amp just barely no enough. I don't listen that loud.

Hi Alan,

I think it's a good idea to stay focused on one thing at a time. High powered single ended amplifiers are notoriously difficult to get right, and I think probably it's easier and more cost effective to begin with a push-pull design. Incidentally, class A push pull can also perform well, you could change your existing design to full Class A without even losing much power, and then sleep better at night. I've certainly never looked into the theoretical operation of tubes in class AB with this level of scrutiny, but the questions you're asking do make sense. I do look forward to seeing how they're eventually answered.

Over the holidays I decided to change the capacitors on one of my amplifiers which happens to be class A push-pull, with 807s (similar to 6L6) and I put them into my system a few days ago. It's nice to have the high power reserve, and I can't really say for certain it sounds worse than my single ended amplifier.

-Max
 
Max, the AB thing is interesting...there is some interesting discussion on it here:
http://www.diyaudio.com/forums/tubes-valves/315706-definition-class.html

It is of course the details that make this all interesting to solve. A lot of the GE data sheets have the information needed to take advantage of what is going on. I do like the 'stay Class A' suggestion, it is for many reasons the easy way out...unless you need lots of power. At that point you just need to get high enough plate dissipation w/o exceeding the cathode current limits.
cheers,
Douglas
 
Max, the AB thing is interesting...there is some interesting discussion on it here:
http://www.diyaudio.com/forums/tubes-valves/315706-definition-class.html

It is of course the details that make this all interesting to solve. A lot of the GE data sheets have the information needed to take advantage of what is going on. I do like the 'stay Class A' suggestion, it is for many reasons the easy way out...unless you need lots of power. At that point you just need to get high enough plate dissipation w/o exceeding the cathode current limits.
cheers,
Douglas

Hi Douglas,

Thanks for the above link. Forum member Smoking-Amp explores this issue very well, and draws some interesting conclusions. I would be interested to see one of his completed designs, he seems like a brilliant designer. The conclusion I draw here is that AB1 is a balancing act, which one needs a whole lot of test equipment and time to get the best from, due to gm-doubling. I really like smoking-amp's graphs, and intuitive explanation of this phenomenon.

I have to say the real conclusion I am drawing on a person level, is that going forward I'm going to stick with class A. I'm making one off designs for my own use, so the extra power supply and tube cost is minimal next to the effort required to make AB1 perfect. I don't own anything but high efficiency speakers, and I have no space constraints or power consumption constraints, so an argument for AB is kind of hard to make. It also explains why emperically I prefer how class A amplifiers sound, without knowing why.

Best Regards

MAX
 
Last edited:
Yah...that sums it up pretty well. That of course is why that plot made considering gm as a constant is worthless...LOL
cheers,
Douglas
 
Yah...that sums it up pretty well. That of course is why that plot made considering gm as a constant is worthless...LOL
cheers,
Douglas

Then add to this the fact that real world speakers are a complex impedance, jeez, too many variables, I think it qualifies as an art form... Or at the very least something requiring iterative testing. I'm starting to get your continuous protest against dividing class AB into Class A and Class B regions. Basically what I am seeing is in addition to degrees of conduction, class A means you're biased out of the region near cutoff where gm is most variable. In AB you're right in the curved area.

It's amazing how good a badly designed tube amp can sound though.
 
Hi Alan,

I think it's a good idea to stay focused on one thing at a time. High powered single ended amplifiers are notoriously difficult to get right, and I think probably it's easier and more cost effective to begin with a push-pull design. Incidentally, class A push pull can also perform well, you could change your existing design to full Class A without even losing much power, and then sleep better at night. I've certainly never looked into the theoretical operation of tubes in class AB with this level of scrutiny, but the questions you're asking do make sense. I do look forward to seeing how they're eventually answered.

Over the holidays I decided to change the capacitors on one of my amplifiers which happens to be class A push-pull, with 807s (similar to 6L6) and I put them into my system a few days ago. It's nice to have the high power reserve, and I can't really say for certain it sounds worse than my single ended amplifier.

-Max
At this point, I think I am right about the gain match ONLY if the internal plate resistance is 0ohm. Take a look at post #19, the higher the plate resistance, the more it is mismatch.

I gave it a lot of thoughts on post #19 to use equivalent circuits to simulate the 3 conditions.

I know people think this is a waste of time to dig so deep into the theory, much easier just build and listen. I just disagree with this idea. I believe digging down as deep as possible gives a lot of insight to the advantage and pitfall of a design. Really understanding the circuit is the key to a good design. One needs to start with a good sounded foundation. This, I am surprised after like two months of studying tubes, that people in tubes just don't go deep into theory. At this point, I really read ( at least scanned through a lot of books and articles), there are HOLES in the the theories, big holes. It's just too bad there are so few people that are interested in tubes these days that nobody actually stop and spend the time to look deeper and cover the holes from the theories of the 40s and 50s.

I know this might be offensive to some people, but I can only say one can never get away in the modern SS world. Everything is analyzed and understood in SS power amp design. Every circuit has explanation WHY it is what it is. I really believe if people would stop and give the same scrutiny to the tube theory, there are room for improvement. It is not about "tubes has so much distortion, that the change are very graduate that you won't notice...". It's about knowing what's the short coming and make the best out of it.

Yes, I thought about that I can get more class A power using PP rather than SE also, it's true. It's just an alternative way of thinking. I am not worry about how difficult it is to implement the parallel tubes SE amp. This is where I really feel differently on this whole thing, and I feel quite strongly about this. Getting circuit to work, tame the circuit from oscillation is just a process. It'll get done one way or the other. It's just take the time to fix it to get it working according to the design. BUT if the foundation is wrong, there is nothing you can do to fix it. As for the implementation part, you just handle it. I am not going to say I can make the 3 tubes SE amp work nicely as I have not try it. But I have never failed in implementation in my years of career. I just handle it as it comes. There's a will, there's a way.

Case in point. I have not seen anyone design like my SS amp. Believe me, I read tons of schematics to make this statement, some in diyaudio came close but not close enough. I went for ultra low distortion, I maximize the forward gain to the extreme by maximizing the gain of each stage. You are talking about over 60dB feedback gain in audio band and bandwidth of over 330KHz. Nobody do this because it's so so hard to tame it from oscillation. I did it in 3 weeks. I tame it even with 20,000pF on the output connector of the amp. To put it in perspective, the famous Nelson Pass Stasis Nakamichi PA-7 BURSTED into wild oscillation with only 1500pF on the output connector. Getting it to work is just a process. Getting it right in theory is the harder and the most important part of the design.

Alan
 
Last edited:
Alan,
I applaud your pursuit. In all honesty i don't think you give many of the people here or the tube community as a whole enough credit. These people (i'm not speaking for myself because i am clearly a novice and likely won't put the necessary time and effort into understanding the theory completely) have been doing, that is, working on, designing, building, modifying existing designs for 30 or more years. They clearly aren't as naive or misinformed as you may think and many have generously given their time and effort to guide you. They have hard earned knowledge. If you don't feel so, that's your right. To dismiss it is your loss. Good luck in your endeavour. Just try not to be so smug about what you do or have done. I don't discount your knowledge and experience, i do doubt you fully grasp everything about audio design electronics that you think you have. Just my 2c I hope you're not offended. This is meant as kindly advice. Not an insult. I hope its taken that way. Cheers
 
Alan,
I applaud your pursuit. In all honesty i don't think you give many of the people here or the tube community as a whole enough credit. These people (i'm not speaking for myself because i am clearly a novice and likely won't put the necessary time and effort into understanding the theory completely) have been doing, that is, working on, designing, building, modifying existing designs for 30 or more years. They clearly aren't as naive or misinformed as you may think and many have generously given their time and effort to guide you. They have hard earned knowledge. If you don't feel so, that's your right. To dismiss it is your loss. Good luck in your endeavour. Just try not to be so smug about what you do or have done. I don't discount your knowledge and experience, i do doubt you fully grasp everything about audio design electronics that you think you have. Just my 2c I hope you're not offended. This is meant as kindly advice. Not an insult. I hope its taken that way. Cheers
I can only go by the stuffs I read. I scan through a lot on Google, I just don't see the question I asked. The only article I've seen that even address this class A and B region is the Valve Wizard article that drew the load line with two distinct region. I think Valve Wizard stuffs are newer and the person did dig deeper. Too bad they did not go into more detail. I would love to see more if anyone can suggest some article people write about this.

It's not about smug, it's about frustration that I read through a lot of the RDH4, scanned through book by Gray, book by Crowhurst Cooper and articles I can find on Google, I cannot find the few questions I asked. And to me, those are important questions. These question were asked and answered in SS world. This class A to class B crossover region is proven to be very important and very well studied in SS world, it's not even mentioned in the tube world. But it's there. Just resort to saying, it's been done, tested for over 30 years just not going to do it for me.
 
That DIY Audio thread Douglas posted a reference to, really gets deeper into this theory. At first it starts with just talking about the definition of different classes, but about half way though a forum member called smoking amp comes in, and starts some very interesting posts. I suggest reading the part about gm doubling, it speaks about exactly the issue you're asking about in this thread, in more detail than has been covered here.

I suspect a lot of people use tube manual suggested operating conditions and just run with it, instead of trying to reinvent the wheel at every turn. They actually work really well most of the time, and are really the conditions the tubes were designed to work under. My logic on that matter has always been ok RCA designed the tube, let's listen to their instructions how to use it. Then design the rest of the circuit around that. I never really questioned it in this much detail, because like I said, as is it works great. I think other designers follow a more technical oriented path, for me it's about the result more than the method. Like anything analog I suppose it's always possible to make an incremental improvement, but once it hits a certain level I sort of lose interest in making it better, and want to try something else.
 
Is your question about the difference between the A and AB regions? Or is it about the crossover region? Those are two different subjects.

The difference in gain (if there is any) between the A and AB region is not a problem to solve. But crossover distortion is. So what are you really interested in? The confusion you're having with this is the reason you don't find anything about designing for gain across A and AB phases. Gain is not a problem.
To me, it's both. It's the transition between A and B and B to A, which is transition between one gain to another.

I have not even think about the transition yet, just from one gain to the other. There might be ringing during the transition that cause more problem, I don't know.

It's like in SS, the transition is very important. I can only take the words that in tubes, the transition is very gradual so it is not a problem. But the gain is a big problem as shown in post #16, that the whole waveform is distorted if the gain in class A and B region is different.


This is an example of transition distortion ( crossover distortion) in SS:

Wing spread.JPG

This is what we called the wing spread diagram where the Y-axis is the gain of the output stage, and X-axis is the output current. There is a big kink in the gain when it's transition from class A to B. I can only assume the tube is a lot more graduate than this.


Believe me, it's there for the tube output stage, just I don't know how bad it is. In SS, not only we identify the problem, we solved the problem by making the emitter resistor value equal to the 1/gm of the transistor to minimize the crossover bump( we called the Oliver's condition). This is what I am looking at and hopefully find a way to minimize it.
 
That diyaudio link provides insight in to how gm varies with grid-voltage swing. The variation in combined gm in a PP output stage, around a given idle point shows there are many aspects that could noticeably influence the level of raw distortion.

However, a lack of gm curves for commonly used output tubes is noted. I could only really identify the STC 6BQ5 report from vintage days. Luckily, modern curve tracers come to the rescue and have the benefit of providing data from actual tubes, rather than averaged curves - UL papers by Rudolf Moers offer even more insight, and data on common tubes:
http://www.oestex.com/tubes/ul.html
https://www.dalmura.com.au/static/Moers UL_1.pdf
 
No


http://www.diyaudio.com/forums/tubes-valves/315706-definition-class.html

Like I said get through to smoking amp's post, this guy is doing good work here.

Thanks Max for taking the time to download the graphs and corresponding to the post # for me. I reading through part of it and scanned through the rest. Yes this is kind of relate to my question, but it's not really.

1)In post #53, I absolutely understand the gm is not constant and according to Smoking Amp, gm increases with current. But the idea of different gain between the class A region and B region is still different.....be that it's graduate and not as abrupt as I make it to be, but it's still there. Good to know though that in class A region, gm is almost flat.

2) In Post #89, funny he talked about using Ramp, I was drawing triangular wave to show more obvious how change of gain affect the output waveform.

3) In post #101, He actually talked about the crossover distortion, the Wing Spread gain wobbling during the crossover from class A to class B. This is a totally DIFFERENT from what I am talking in this thread. I did not address the crossover distortion at all as I have no idea how it looks like. Smoking Amp actual plotted out the crossover region and you can see the "V" shape gm curve if it is way off adjustment, but becoming close to the Wing Spread diagram I posted in post #34 here.

4) Post #112 talked about adjusting the current etc. Like SS, we adjust to a prescribed current and using the prescribed emitter resistor to minimize the kink of the Wing Spread graph. Still the question is from what theory it is adjusting to. Sounds like he use the scope to adjust. We actually calculate out the resistance value and current and adjust to spec. I would like to here the theory of minimizing the wing spread deviation.

5) Actually in post #91, Smoking Amp made a comment exactly what my other part of the question:
"I would be happy to settle for a well designed class AB amplifier. Using local output stage N Fdbks and moderate global N Fdbk.

If purchasing one, I would demand the gm sum "wing plot" be posted. (SS design books typically post these for different topologies) From that, one can tell if the amplifier is going to be ear grating (noticeable and many ripples in the gm wing plot) or smooth sounding (a flat wing plot, or at least smoothly varied gm for low harmonics). I would value this far above any THD specs."


All in all, this is interesting information. It tell me something that I eventually will raise question, the crossover distortion which is the Wing Spread graph. This also tells me tube has crossover distortion that I have not seen described in books.

Smoking Amp really dig deep into this also. Good to know there are other people that are obsessed in get into this also. I always wish to have a way to display the wing spread on the scope.

Thanks Max
 
Last edited:
That diyaudio link provides insight in to how gm varies with grid-voltage swing. The variation in combined gm in a PP output stage, around a given idle point shows there are many aspects that could noticeably influence the level of raw distortion.

However, a lack of gm curves for commonly used output tubes is noted. I could only really identify the STC 6BQ5 report from vintage days. Luckily, modern curve tracers come to the rescue and have the benefit of providing data from actual tubes, rather than averaged curves - UL papers by Rudolf Moers offer even more insight, and data on common tubes:
http://www.oestex.com/tubes/ul.html
https://www.dalmura.com.au/static/Moers UL_1.pdf
Thanks for the links, I have to take some time to read it. It's new years day, not particularly in the mood to read.
 
Thanks Max for taking the time to download the graphs and corresponding to the post # for me. I reading through part of it and scanned through the rest. Yes this is kind of relate to my question, but it's not really.

1)In post #53, I absolutely understand the gm is not constant and according to Smoking Amp, gm increases with current. But the idea of different gain between the class A region and B region is still different.....be that it's graduate and not as abrupt as I make it to be, but it's still there. Good to know though that in class A region, gm is almost flat.

2) In Post #89, funny he talked about using Ramp, I was drawing triangular wave to show more obvious how change of gain affect the output waveform.

3) In post #101, He actually talked about the crossover distortion, the Wing Spread gain wobbling during the crossover from class A to class B. This is a totally DIFFERENT from what I am talking in this thread. I did not address the crossover distortion at all as I have no idea how it looks like. Smoking Amp actual plotted out the crossover region and you can see the "V" shape gm curve if it is way off adjustment, but becoming close to the Wing Spread diagram I posted in post #34 here.

4) Post #112 talked about adjusting the current etc. Like SS, we adjust to a prescribed current and using the prescribed emitter resistor to minimize the kink of the Wing Spread graph. Still the question is from what theory it is adjusting to. Sounds like he use the scope to adjust. We actually calculate out the resistance value and current and adjust to spec. I would like to here the theory of minimizing the wing spread deviation.

5) Actually in post #91, Smoking Amp made a comment exactly what my other part of the question:
"I would be happy to settle for a well designed class AB amplifier. Using local output stage N Fdbks and moderate global N Fdbk.

If purchasing one, I would demand the gm sum "wing plot" be posted. (SS design books typically post these for different topologies) From that, one can tell if the amplifier is going to be ear grating (noticeable and many ripples in the gm wing plot) or smooth sounding (a flat wing plot, or at least smoothly varied gm for low harmonics). I would value this far above any THD specs."


All in all, this is interesting information. It tell me something that I eventually will raise question, the crossover distortion which is the Wing Spread graph. This also tells me tube has crossover distortion that I have not seen described in books.

Smoking Amp really dig deep into this also. Good to know there are other people that are obsessed in get into this also. I always wish to have a way to display the wing spread on the scope.

Thanks Max

HI Alan,

Keep reading, it's one of the most interesting series of posts on a forum I've seen in a while. Like I said before, I would really like to see one of smoking amp's final designs, I imagine it's something special

Enjoy the holiday

Max
 
HI Alan,

Keep reading, it's one of the most interesting series of posts on a forum I've seen in a while. Like I said before, I would really like to see one of smoking amp's final designs, I imagine it's something special

Enjoy the holiday

Max
I don't follow post #112, how to have grounded cathode with CCS. Does he mean CCS drawing current from the cathode of each power tube and using bypass cap from each cathode to ground to give an AC ground? I don't see any difference from just simple grounding the cathode. Usually people adjust the bias current to adjust the wing spread AFTER optimize the design.

Using tube to do CCS is too complicated. If I were to go with modern design like CCS, I would use BJT as CCS. BJT is perfect as the output impedance ( like rp in tubes) is way higher than any tubes can touch. ONLY reason to use CCS is to provide a very high impedance current source so the current source approach an idea current source. Using bypass cap just does not make sense to me.

Unless one actually hook up the two PP power tubes as differential pair ( long tail pair ). I actually thought about that before also. Then it will make a whole world of sense to use CCS tail to make it ideal. But that only work for class A, not class AB.
 
Last edited:
I don't follow post #112, how to have grounded cathode with CCS. Does he mean CCS drawing current from the cathode of each power tube and using bypass cap from each cathode to ground to give an AC ground? I don't see any difference from just simple grounding the cathode. Usually people adjust the bias current to adjust the wing spread AFTER optimize the design.

Using tube to do CCS is too complicated. If I were to go with modern design like CCS, I would use BJT as CCS. BJT is perfect as the output impedance ( like rp in tubes) is way higher than any tubes can touch. ONLY reason to use CCS is to provide a very high impedance current source so the current source approach an idea current source. Using bypass cap just does not make sense to me.

Unless one actually hook up the two PP power tubes as differential pair ( long tail pair ). I actually thought about that before also. Then it will make a whole world of sense to use CCS tail to make it ideal. But that only work for class A, not class AB.

Mostly what smoking amp is doing, is food for thought. The non-linearity he's showing with the transconductance plot is also shown on the regular tube curves with how the spacing between the grid voltage lines get grouped together toward the bottom of the pentode curves. He's just showing the same thing in another way, and looking for ways to mitigate this effect.

A regular design, simply push-pull, kept in class A, and biased so there's lots of linear region above and below the operating point, will go a long way towards mitigating both of these issues. I'm not seeing a big reason to get too fancy here.. just push pull Class A with the operating point chosen to bias 6L6GC tubes right to the limit of dissipation. The more I study this, the more I think the Quad II looks like a good design... it sure checks all the boxes for quality, and none of the boxes for gadgetry.
 
Mostly what smoking amp is doing, is food for thought. The non-linearity he's showing with the transconductance plot is also shown on the regular tube curves with how the spacing between the grid voltage lines get grouped together toward the bottom of the pentode curves. He's just showing the same thing in another way, and looking for ways to mitigate this effect.

A regular design, simply push-pull, kept in class A, and biased so there's lots of linear region above and below the operating point, will go a long way towards mitigating both of these issues. I'm not seeing a big reason to get too fancy here.. just push pull Class A with the operating point chosen to bias 6L6GC tubes right to the limit of dissipation. The more I study this, the more I think the Quad II looks like a good design... it sure checks all the boxes for quality, and none of the boxes for gadgetry.
Yes, I understand what he showed in #53 the spacing is very even with PP class A region.
What is Quad II? I don't know what that is.

Main think I got out of the thread is tubes do have crossover distortion just like SS output stage, the wing spread is very similar. This is on top of the gain difference between class A and B region.
 
Back
Top Bottom