looking for very good SE output transformer

My problem is not when one tube is off, is when both are on. How do you get half the primary impedance, how do you derive this out.

The formula in question is in page 572 of RDH4

RL"=1/2 RL = 2R2(N1/N2)^2

How do you come up with this using transformer formulas? My drawing is just a train of thoughts trying to make sense of this.

Why.

Because you're using formulas for multiple loads on a transformer, not sources.

Think of it like a lamp connected to two batteries which are connected in parallel, and then remove one battery. The load on the remaining battery halves In value.
 
My problem is not when one tube is off, is when both are on. How do you get half the primary impedance, how do you derive this out.

The formula in question is in page 572 of RDH4

RL"=1/2 RL = 2R2(N1/N2)^2

How do you come up with this using transformer formulas? My drawing is just a train of thoughts trying to make sense of this.

Why.

Read my post again! I've already given you an answer!
 
Why making things unnecessarily complicated? I understand the books but I don't understand what your explanation or difficulties is about. The above equation and the picture does not makes sense. I have the book so I know where you find that equation. But the diagram you drawn is quite different and does not mean anything to me.

When the tube is at cutoff (no grid signal or lost of bias), it will have almost infinite (very high to be correct) output impedance according to the triode characteristic. Only one tube conducting at a time at either + or - side of Vin at the grid. Either side of the output tube actually don't know that the other half exists at all in the circuit and don't care at all about Rp of their counterparts. In PP Tube amp, you can remove one output tube and still get half side of the output (not good for the transformer since DC is now unbalanced in the primary.

If you want to know where 4 in the equation comes from, read the book and pays attention to the book including footnotes. See chapter 5, Section 1, sub section ii (Page 201 in paperback RDH4, Figure 5.4). Transformer impedance calculations.

You can't leave out a transfer ratio of the output transformers in the Tube Amp AB topology. It is a critical part of everything!
I read p201 many times, I know where the 4 comes from. This is NOT what I am asking. I gone through to p203 in detail already. It's how they arrive to the formula in p572 I am questioning.
 
Last edited:
Because you're using formulas for multiple loads on a transformer, not sources.

Think of it like a lamp connected to two batteries which are connected in parallel, and then remove one battery. The load on the remaining battery halves In value.

Thanks Max

I need to think about it more, I am stuck today.

Alan
 
I hope this helps!
Look at p202 for origin of equations
Thanks Max

You got it, this explains in terms of the transformer equation. That's exactly what I am looking for. I know the equation in RDH4 is correct, but I just need to have an analytical explanation.

Alan
 
Thanks Max

You got it, this explains in terms of the transformer equation. That's exactly what I am looking for. I know the equation in RDH4 is correct, but I just need to have an analytical explanation.

Alan

I know, and no problem! And thank you for making me think about it. The change in loading upon transition from A to B makes class AB look less attractive or at least less ideal than I always saw it. A Pair of 6550 in Class A1 push pull could make 40W, and avoid this while issue entirely. Now I'm not saying my conclusion is to avoid Class AB, only that by analyzing it I think you found a defect which isn't commonly discussed, and merits further analysis. Good work!
 
I read p201 many times, I know where the 4 comes from. This is NOT what I am asking. I gone through to p203 in detail already. It's how they arrive to the formula in p572 I am questioning.

If you understand it, what was that you don't understand? You don't understand how multi turns transformers works or why the transformer is part of the equation of the Class AB load line?
 
What do you think if I start with two pairs of output tubes to get at least 60 to 80W of power.
With the right supply and OPTs 60 is easy with kt88s. (Dynaco Mk3, Mcintosh MC60)
100+ is doable
Read the mighty bogen (200?) Thread.
 
I know, and no problem! And thank you for making me think about it. The change in loading upon transition from A to B makes class AB look less attractive or at least less ideal than I always saw it. A Pair of 6550 in Class A1 push pull could make 40W, and avoid this while issue entirely. Now I'm not saying my conclusion is to avoid Class AB, only that by analyzing it I think you found a defect which isn't commonly discussed, and merits further analysis. Good work!
Ha ha, it's good or not so good. I just have a habit of not letting it go, got to get to the bottom even if it means I'll get stuck for a while. I think it's important to dig deep as you really learn.

I still have a way to go, I am starting to read p573 onward in RDH4, those feel very important. So it's going to be a little while before I start talking to you on the design.

thanks for you help.

Alan
 
If you understand it, what was that you don't understand? You don't understand how multi turns transformers works or why the transformer is part of the equation of the Class AB load line?
Read the posts, it's been very clear, Max hit the problem dead on and I understand it now. The question I asked was "why the plate of each tube sees 1/2 or the total primary resistance in class A operation."
 
With the right supply and OPTs 60 is easy with kt88s. (Dynaco Mk3, Mcintosh MC60)
100+ is doable
Read the mighty bogen (200?) Thread.
It's too early for me to determine what I want. I am still reading.

From reading so far, I am thinking about running in triode mode mainly because plate resistance of triode is much lower than panthode or UL. This will have a big advantage of lowering the output impedance which increase the damping factor of the amp. I am even thinking of using two pairs of output tubes to further lowering the output resistance to increase damping factor.

As for power, I don't think it's that important for me to get over 40W, so power is not important.

I think crossover distortion ( I think it's the same for SS and tube, is when one half turns off and the the other half take on the whole load) get less as output impedance goes lower. that's the reason high end SS amp uses many pair of output transistors to make very high damping factor possible.

Also, I think the theory is the same for SS and tube, even running in class AB, run at high bias so one can get a big class A region. the amp actually work in class A in the first few watts. This is how I design my SS amp, the first 5W is in pure class A to avoid the "U" shape distortion performance. Literally put the crossover point in much higher signal level where a little of the crossover distortion is not important anymore.

Well, this is my theory at the moment. I might change my mind when I learn more. I still yet to learn how to calculate output power in tube PP amps. So I have no idea how much I can get.

thanks
 
Last edited:
Read the posts, it's been very clear, Max hit the problem dead on and I understand it now. The question I asked was "why the plate of each tube sees 1/2 or the total primary resistance in class A operation."

Class A operation? If I were you, I would read more and stop digging my own grave! I understand you fine! You missed to understand the importance of the Page 201 with diagram 5.4. The rest is just the explanation of putting them into words. Perhaps read a bit more of the entire books a few times?
 
Ha ha, it's good or not so good. I just have a habit of not letting it go, got to get to the bottom even if it means I'll get stuck for a while. I think it's important to dig deep as you really learn.

I still have a way to go, I am starting to read p573 onward in RDH4, those feel very important. So it's going to be a little while before I start talking to you on the design.

thanks for you help.

Alan

Sure, knowledge is power, and you're just scratching the surface. I want to see what happens once you get deeper, I suspect you will work some magic in Spice.

One thing you cannot under estimate is the value of a good technical library. A good university library beats the Internet hands down, I used to take piles of books about tube theory and design, and read them during my spare time.
 
Class A operation? If I were you, I would read more and stop digging my own grave! I understand you fine! You missed to understand the importance of the Page 201 with diagram 5.4. The rest is just the explanation of putting them into words. Perhaps read a bit more of the entire books a few times?

I've heard that kind of pleas before. I have also seen the helps given are suddenly turned around or overlooked by a Professor who knows all the very next day and even suggested his theories works a wonder. It is important to mention some Alan cooked capacitive magic cables and very inductive wirewound "non linear resistors" are used in the experiment. And finally ,the conclusion is a very low DF with not a very good THD is why this amp sound bad while all the crazy things ever happened to this amp could as well be a result of 100% builder's fault and half cooked measurements. Ironic I say. But this whole thread is becoming a very entertaining read for me and may be for others too. Kudos to that, some threads Alan quoted are even worth leaving a milestone in the Twilight Zone!
With all that I say May the force be with you Max!

Who's the only one that doesn't get the question? Max understand just fine and he gave the answer already.

The question to you is why you even join in? I really don't want to get into fight and keep this pure technical.
 
It's too early for me to determine what I want. I am still reading.

From reading so far, I am thinking about running in triode mode mainly because plate resistance of triode is much lower than panthode or UL. This will have a big advantage of lowering the output impedance which increase the damping factor of the amp. I am even thinking of using two pairs of output tubes to further lowering the output resistance to increase damping factor.

As for power, I don't think it's that important for me to get over 40W, so power is not important.

I think crossover distortion ( I think it's the same for SS and tube, is when one half turns off and the the other half take on the whole load) get less as output impedance goes lower. that's the reason high end SS amp uses many pair of output transistors to make very high damping factor possible.

Also, I think the theory is the same for SS and tube, even running in class AB, run at high bias so one can get a big class A region. the amp actually work in class A in the first few watts. This is how I design my SS amp, the first 5W is in pure class A to avoid the "U" shape distortion performance. Literally put the crossover point in much higher signal level where a little of the crossover distortion is not important anymore.

Well, this is my theory at the moment. I might change my mind when I learn more. I still yet to learn how to calculate output power in tube PP amps. So I have no idea how much I can get.

thanks
The tube amps with the greatest damping factor achieve it through feedback, have a look at the Harman Kardon Citation II (MUCH better... listen to one), triode-connecting the outputs does not make not make up for the loss of power in terms of control.

AB looks imperfect on paper but somehow the Citation, Marantz 2,5,8 and 9, Mac 225 and MC30, Fairchild 255, 260 and 275, Heathkit W6M among others don't seem to bother too many people sonically.
 
The tube amps with the greatest damping factor achieve it through feedback, have a look at the Harman Kardon Citation II (MUCH better... listen to one), triode-connecting the outputs does not make not make up for the loss of power in terms of control.

AB looks imperfect on paper but somehow the Citation, Marantz 2,5,8 and 9, Mac 225 and MC30, Fairchild 255, 260 and 275, Heathkit W6M among others don't seem to bother too many people sonically.
It is important not to rely on GNFB or any FB to raise damping factor. You want to start with good design and FB is the final touch.

I still need to study a lot more, this is only my thinking after a few days of studying. Just saying out loud.

Actually there are two things that bugs me in class AB PP amp.

1) If one follows page 573 of RDH4 how the author use power series to represent the output current of both tubes. In equation (4), when id=ib1-ib2, you left with all the odd harmonics. All the even harmonics is gone. Even harmonics are equated to better sound, odd are not so much. That's not a good thing!!!

2) The hysteresis of the output transformer occurs at 0 crossover regardless of idle current. This will cause crossover distortion that cannot be fixed by running higher idle current. I don't know how much it's a factor. Max said it's not important. I don't know what to think.
 
The tube amps with the greatest damping factor achieve it through feedback, have a look at the Harman Kardon Citation II (MUCH better... listen to one), triode-connecting the outputs does not make not make up for the loss of power in terms of control.

AB looks imperfect on paper but somehow the Citation, Marantz 2,5,8 and 9, Mac 225 and MC30, Fairchild 255, 260 and 275, Heathkit W6M among others don't seem to bother too many people sonically.

What's interesting to me here, is how Alan noticed that formula and decided to look into it more. I've always just drawn load lines using a slope of 1/4 * Rp-p, and been done with it, as though everything is always working in class B. What I ignored, is the step change in impedance which occurs upon transition from Class A to Class B. In fact it was absolutely not on my radar at all. Now practically, this can be drowned out with feedback.. no arguments from me, and practically, there's many many examples of excellent sounding Class AB amplifiers.

But theoretically, it would be interesting to quantify what's going on during that transition, not from a practical perspective but just from a theory/understanding perspective. It's true this phenomenon can be mitigated with brute force by going to Class A. I would be really interested to see what happens to the voltage output of a resistive loaded pure pentode amplifier without any loop feedback, as it transitions between Class A and Class B.
 
It is important not to rely on GNFB or any FB to raise damping factor. You want to start with good design and FB is the final touch.

I still need to study a lot more, this is only my thinking after a few days of studying. Just saying out loud.

Actually there are two things that bugs me in class AB PP amp.

1) If one follows page 573 of RDH4 how the author use power series to represent the output current of both tubes. In equation (4), when id=ib1-ib2, you left with all the odd harmonics. All the even harmonics is gone. Even harmonics are equated to better sound, odd are not so much. That's not a good thing!!!

2) The hysteresis of the output transformer occurs at 0 crossover regardless of idle current. This will cause crossover distortion that cannot be fixed by running higher idle current. I don't know how much it's a factor. Max said it's not important. I don't know what to think.

1. That's in theory that they're all gone. You still get second order distortion from the driver tubes. One huge benefit you haven't go to yet is hum is also cancelled, so power supplies are much less critical.

2. Look at this as a high quality transformer: http://www.sowter.co.uk/specs/u027.htm Note distortion due to core saturation is 0.5% at 20Hz, but distortion due to hysteresis is only 0.05% at 1000Hz. When you compare this to what the tubes will be doing at those frequencies, well, I believe the transformer's contribution is not significant. That is of course, if it's a good transformer, which in your case it very well will be. Note that's also open loop distortion, it's only gonna go down in the final product.
 
Back
Top Bottom