• The move to the new server is done. There are some software and database maintenance updates in process. This has us passing the hat around to help out. We appreciate any donations. Seriously, even a dollar helps. The payment page may be found here - https://www.audiokarma.org/support.html

Suggestion for push pull output transformer.

I hardly ever see full curves in transistor data sheets, though I admit I haven't looked at them that much. I think a company designing tube amps for profit should really buy a curve tracer, us hobby guys are at the mercy of data sheets though.
We don't use those curves when designing transistor circuits. It's really not like designing tube amps. Transistor behaves more like a very perfect penthode with collector curves very straight and almost horizontal. Never use load line to design. I think the really old books might have talk about load lines in the 70s that I saw following from the tube days, later on, never seen those load lines.

Transistor has so much gain we never truly ground the emitter and let the transistor running high gain. eg if you ground the emitter and run 1mA through the transistor, the emitter resistor is equal to 26ohm, if you have a 10K load resistor at the collector, you have gain of 10,000/25=400!!! If you run 5mA, you have gain of 2000!! Just from 1 stage. You don't want to run transistor like this. You put an emitter resistor to lower the gain.

Say you run 1mA, you put a 300ohm resistor in series with the emitter and use a 10K resistor as collector load, you get gain of -10000/326=-30.7. Almost like -10000/300=33.3. By putting a 300ohm resistor, the 26ohm internal resistor of the transistor almost got drown out and the little change of the internal resistor does not affect the final gain much. In another words, gm of transistor circuit really does not depend on the transistor, it would be a bad design if so.

The philosophy of design transistor circuit is to take the transistor OUT of the picture and use resistors to determine the gain. So we don't rely on the characteristic of the transistor for the gain. The result is transistor circuit can easily be designed with much much lower distortion, much more stable in characteristics.

This sense, transistors are much more simpler to design.....BUT transistors have both NPN and PNP( N and P FETs). So one open to a whole new world of possibility like complementary design, active current mirrors and all sort of goodies.
 
Last edited:
I contacted Sowter, a pair of OPT that I want including shipping cost about $900USD, that is too high. The shipping is about $180, that's the killer. It's one thing to want to buy good OPT, but for that king of money, I start questioning how much is snake oil effect.

Actually the choice of OPT is very very limited. I cannot find any decent 4K, 3.5K, 2K primary impedance. Mostly are 4.5K, 3.3K, 5K, 2.5K primary impedance.

The better rated Dynakit and Triode only have the Dyna Audio clone of 4.3K primary. Only other brands within reasonable price are Lundahl, Edcor and Hammond. Not much choice.
 
I noticed the highest end brands Monolith Magnetics, Hashimoto, Sowter , all start at 5K primary for a PP trans and as I understood it that's where your better to start off with on bigger tubes like the KT-88. Sowter has one at 4.5K.

You would think if 3.5 or 4K were warranted they would be offering those!

$180 shipping isn't really that bad for those, as I bet they would cost me $150 from the US!

If you want quality tube sound (low distortion etc) you got to pay a little. No big heat sink costs either! From what you've said I don't think you will be happy if you don't spend for quality.

Believe me it isn't all snake oil. Go listen to some tube amps. A lot of people if they can afford it end up with big powerful SE amps in the end as they prefer that sound as they've been through different PP amps and you can use your speakers. Your talking about big & still expensive Output Trans. though and even 1 or 2K voltages to work with!
 
Last edited:
The DIY market is very small. Let's face it, there is not a lot of demand for new production hifi output transformers for tube amps. Sticking to the most common choices reduces costs so the small transformer winder can stay in business.

You can try to contact Heyboer in Michigan. See if they can duplicate something from the heyday era. That may be your best bet at this point if you are moving away from common off the shelf stuff. Or you can learn how to wind your own and then sell them to the rest of us. :)
 
Those old datasheets has holes in them. Like a lot of them show Vg2=250V. What is that good for? Everyone use the +B voltage in penthode mode, nobody create a +250V just to power the screen!!!

Tuning the g2 voltage to the load line delivered is The Way to get good performance from pentodes. The last amp I built ran only ~280V of B+ and 105 for its g2. If the operating envelope is not properly set up for pentode operation, you are leaving stuff on the table.
cheers,
Douglas
 
Tuning the g2 voltage to the load line delivered is The Way to get good performance from pentodes. The last amp I built ran only ~280V of B+ and 105 for its g2. If the operating envelope is not properly set up for pentode operation, you are leaving stuff on the table.
cheers,
Douglas
That's was old response, I since drawing load lines using lower Vg2 and got much better result. In this thread post #50, I switch to Vg2=350V for +B=400V and I posted the graph: http://audiokarma.org/forums/index....or-a-pair-of-pp-kt88-in-ul-mode.795522/page-3
I tried lower Vg2, I lost too much power. Yes, with lower Vg2, it's easier to make the load line intersect the Vg1=0 plate curve on the right side of the knee to reduce 3rd harmonic. But really low Vg2 loss too much power. I settle with 350V where the load line intersect right in the middle of the knee and I get bigger class A region and reasonable output power.

Please join in the discussion in the other thread.
 
There is no 'bigger class A region'. Either both are conducting at full power, or it is an AB amp, operating in AB( that seems more obvious once it is written...LOL )

Douglas
 
No, every class AB amp has a class A region. Class A region is defined as the region where both tubes are on. If you have 50mA idle current, the second tube will not turn off until the first tube reach two times the idle current which is 100mA. That's the class A region.

This is very well studied in SS amp. That's the key difference between the cheap SS amp and the high end ones that have high bias. I am surprised the old tube books never talk about this. A lot of research has been done in the modern days and things are much better understood.

Like none of the tube books or articles talked about crossover distortion. This is so well studied in the modern electronics. that's why I see a lot of holes in the old stuffs.

These are all very real and important.
 
No thanks, you are not ready.
cheers,
Douglas
I can't help stubborn people. It's hard to accept time has move on and left the people from the 50s behind.

Let me give you some advice as an EE for 30 years owning patents and published papers in America Institute of Physics , electronics gone through a few quantum leaps.......IF any stubborn people care to be humble and learn.

As I see it, anyone cannot comment is because they don't even understand it. AND that is typical. This is show and tell time, show what one really has or not have. If I am not ready, please show me.


Sorry and cheers
 
Last edited:
But tubes sound better and it’s just PFM!

From a recent post:
I rolled my eyes for years, formed opinions about audio snobs that spent too much money on glowing light bulbs and claimed they sounded better. Like photography snobs who claimed "film is better!"

Well......I'm three days into my NOS Valves rebuilt Scott Lk-48. I just played Abbey Road on vinyl and I have to say....I have never heard it sound that good. I can't identify what exactly. My Luxman SS gear sounds damn good, and still does but this little Scott just does something to music. Perhaps that's the "tubeyness" that everyone speaks of? It's just as detailed as the Luxman but like with butter poured over it? That's a silly analogy but all I got. Mingus and Horace Silver last night also just sounded.... right. Does that make sense?

Ahhh, the irony...now I'm a tube guy.....Only took 47 years.

Just saying
:beerchug:
 
Ha ha, nobody put down tube amp here. I am building one!!!

For SOME expert that cannot comprehend the meaning of Class AB and LOL, Class AB is a push pull amp that first run in class A when the signal is small, both top and bottom device are on. When the signal get larger, at some point, one device turns off and only one device take over, this turn into class B. So for ANY class AB, there is ALWAYS a class A region where both devices are on. The class A region is the region where both devices is on.

If there is no class A region, the term is Class B not class AB. It's in the books.
 
Last edited:
Here is an article by Valve Wizard: http://www.valvewizard.co.uk/pp.html
This is the diagram in the article I marked up to explain more clearly about the class A region.

Class A region.JPG

The loadline for each half divided into two section, the section when the amp is in class A where the load resistance each tube sees is 1/2 Raa. As seen in the region I marked class A. You can clearly see the slope is not as steep as the later part when the current is beyond 60mA. The turning point is at 60mA on both top and the inverted graph. Beyond the 60mA point, one of the tube is totally off and the stage goes into class B. You can see the slope of the loadline on both top and bottom are steeper indicating that the load becomes 1/4 Raa as in class B where when one tube turns off, the half disappeared and from turn ratio calculation, the impedance of the half primary is 1/4 Raa.

This article explain so much better than all the books I've seen, clearly shows the class A region. AND the transition point is twice the bias current where in this case, the bias current is 30mA, the transition point therefore is 60mA.

This is a good article, better than all the books I read including RDH4 on this part.
 
Last edited:
Back to OPT, I need a 2.5K primary with 4ohm tap. The only ones I found is

Edcor http://www.edcorusa.com/cxpp100-25k

Lundahl LL1659 with 2.6K primary.

I cannot find anything else, I looked at the thread in DIYaudio that listed all the transformer manufacturer. No luck. This is so hard. Any suggestion?
 
A 5K primary will look electrically like a 2.5K primary if the 8 ohm secondary is loaded at 4 ohms. Just an idea. You'd have to verify that the core can handle the magnetic load, etc.
 
There is no 'bigger class A region'. Either both are conducting at full power, or it is an AB amp, operating in AB( that seems more obvious once it is written...LOL )

Douglas
This discussion got kind of heated, I want to be clear I'm joining in to learn/understand, not argue...

It makes sense to me that any amp where the tubes cut off before full power is a class AB design.

However it seems the disadvantage of class B operation (crossover distortion) shouldn't come into play until one tube actually reaches cutoff. In other words the class AB amp does operate in class A, but only for small signals/lowe power.

So you still can't claim a design has "become a class A amp" by moving it closer to class A operation, but you might delay the onset of class B operation. As a result it could produce more power without entering Class B, and you'd be less likely to hear crossover distortion given the same volume level (assuming that level is significantly below full power)?
 
I find that one thing to think about when you're designing, is to pick a transformer first, and then design the circuit for it. It would be nice to do the opposite, but one has to work within the realm of what's actually out there.

2.5k or 2.6k will have no real world difference, in operation. You could go plus minus 10% and still be inside of how much tube parameters are going to vary.
 
This discussion got kind of heated, I want to be clear I'm joining in to learn/understand, not argue...

It makes sense to me that any amp where the tubes cut off before full power is a class AB design.

However it seems the disadvantage of class B operation (crossover distortion) shouldn't come into play until one tube actually reaches cutoff. In other words the class AB amp does operate in class A, but only for small signals/lowe power.

So you still can't claim a design has "become a class A amp" by moving it closer to class A operation, but you might delay the onset of class B operation. As a result it could produce more power without entering Class B, and you'd be less likely to hear crossover distortion given the same volume level (assuming that level is significantly below full power)?

If you look at the area directly around the operating point though, and look only at small signals, of course it looks like class A. If you look at the maximum signal, it looks like class B, hence class AB. Since tube tube cuts off gradually it's hard to say exactly where is the transition from A to B. I think the transition from Class A to Class B could be empirically determined by finding the point where the output stage begins to consume more current. .
 
This discussion got kind of heated, I want to be clear I'm joining in to learn/understand, not argue...

It makes sense to me that any amp where the tubes cut off before full power is a class AB design.

However it seems the disadvantage of class B operation (crossover distortion) shouldn't come into play until one tube actually reaches cutoff. In other words the class AB amp does operate in class A, but only for small signals/lowe power.

So you still can't claim a design has "become a class A amp" by moving it closer to class A operation, but you might delay the onset of class B operation. As a result it could produce more power without entering Class B, and you'd be less likely to hear crossover distortion given the same volume level (assuming that level is significantly below full power)?


That's another way to look at it. This is not "become a class A amp", like you said, there is in the region where both tubes are on when the signal is small That's by definition is Class A. There is no electrical crossover distortion.( not magnetic crossover distortion in OPT, that's another subject.)

There is straightly no "class AB", class AB by definition, that there is a % of over lap where both devices are on and doing the job( notice I use "devices", transistor push pull circuit works exactly the same way). In that region, the amp is operating in CLASS A. But where the signal gets bigger, at some point of time, one device is totally turn OFF, only the other device doing all the work. This is class B. So Class AB is an amp that has overlap where both devices are on ( class A) and region where only one device is on ( Class B).

If you read the definition of class B, that is when the overlap region is reduce to zero, which means each device is on 180deg, or exactly half the cycle, there is no overlap. In another words, there is no class A region in class B amplifier.
 
Back
Top Bottom