Are DOAs class A or AB? (2520, 1731, 990, etc...)

I've been wondering about this since some DOA schematics I've seen seem to have a push - pull transistor pair as the output, for instance, the 2520.  Others (990, OA10) have a transistor pair as an output, but don't seem to have a phase inverter before the output transistor pair, making me think they might operate as single ended class A.
What kind of amplifier are these DOAs? If the 2520 is push pull, is it biased high enough to be  class A (both transistors always on), or class AB (one transistor or the other will shut off)? Are 990s and OA10s SE outputs?
Thanks!

Push pull is by definition class B, while DOA designers have the flexibility to increase the class A part of class AB output stage bias to push the class B transition outside of normal voltage swing for modest load impedances.

JR

> Are DOAs class A or AB?

Can be either. "Discrete" doesn't say....

And WHY do you care?? If the designer did his job good, it isn't a big deal to the sound.

> Push pull is by definition class B

That's worded confusingly.

Push-Pull can be A or B (or C).

Push-Pull can be A for small outputs and B for large output.

Class B Audio *must* be push-pull.

But finding "push-pull" does not say how it is biased or loaded.

We may find a clue in the specs.

If you have a small idle supply current demand and a much larger max-output supply current demand, something appears to be going into class B. Probably the output stage.

But often the specs don't say enough.

990 has a modest idle current demand but can put HUGE output power. We'd guess it runs class B; in good audio "B" often edges to an "AB" mode for teeny output.

One of the ancient discrete modules was very Class A. To make reasonable output with inexpensive devices it ran so close to overheating that a 20% over-voltage would soon smoke it.

> 990... don't seem to have a phase inverter before the output transistor pair

I think you need to understand totem-pole output stages better.

However even the "lack of phase splitter" does not say A or B.

A or B can be the SAME topology. You need to analyze the output stage idle current and compare to the maximum output current.

Here's a link to an article I wrote on the subject so long ago that every actual amp model I pimped with pictures in the article have long since gone out of production and been replaced by newer models.

http://www.peavey.com/support/technotes/poweramps/classact.cfm

FWIW the distinction of class A or B or AB, means less to sound quality than how well the given design has been executed.

JR

I think "push-pull" is maybe the wrong term in cases where the OP stage bias is A or at least A for all practical loads.  I prefer "symetrical" or "complimentary-symetrical" or balanced or _ _ _ , whatever describes the circuit in question's output.  Maybe it's me but "push-pull, class A" doesn't sit right.  Why not "Push-Push"?

I think you need to understand totem-pole output stages better.

Click to expand...

Thanks you! Now I see what is going on in these designs. The emitter and cathode of the two transistors are out of phase, one 'pushes', one 'pulls'.

And WHY do you care?? If the designer did his job good, it isn't a big deal to the sound.

Click to expand...

I'm trying to understand how elements of design affect the sound. It seems crossover distortion would be easy to measure but... how have you proven to yourself that this isn't a big deal to the sound? I haven't found this an easy task. I'm hesitant about drawing any conclusions in this regard, even though I have been recording with DIY preamps for a year. I can say the Neve single ended design sounds fantastic. The UA1108 single ended design also sounds fantastic. The NYD onebottle design sounds fantastic, and is also single ended.  All the discrete preamps I've built and listened to using push-pull DOAs do not sound quite as fantastic to me. I just built a 990, however, with Jensen's best input transformer, that I haven't listened to enough to judge. But since my observations and samples have one thing in common right now (SE vs. PP), I would like to keep testing my theory.

dmp said:

I'm trying to understand how elements of design affect the sound. It seems crossover distortion would be easy to measure but... how have you proven to yourself that this isn't a big deal to the sound? I haven't found this an easy task.

Click to expand...

I will have to choose my words carefully to not offend, but this sounds a little like mythology when people try to explain things they don't understand with things they see and believe are causal. 

Circuit topology and design decisions certainly influence outcomes, but it's not as simple as "all these, sound like this, and all those sound like that".  There are many ways to speed up switching and reduce the impact of class B commutation, before we add the class A bias to chase it away completely.

Properly applied simple linearity measurements (distortion)  will identify how clean or corrupted a path is.

If the differences you hear and are attracted to are in fact euphonious errors, that is another pursuit entirely from clean, linear design.

JR

PS: I recall over the decades trying to listen for crossover distortion while making trims or refining designs. Severe crossover distortion is clearly audible, small amounts not so much, in my experience. Perhaps another reason I don't design by ear (I can measure distortion I can't hear). 

Not an opamp but nevertheless (it's current-feedback amp), it's a neat example of a totem-pole output ...

http://sound.westhost.com/jll_hood.htm

(it can be scaled down, minimised and used as a headphone amp or a line driver etc.) ...

> "push-pull, class A" doesn't sit right

Why not?

There is a deep history of push-pull class A.

In the beginning, it was not clear that idle current could be reduced in a push-pull amp. The 2+X power, ripple and DC cancellation, seemed ample reward.

When it became clear that idle current could be reduced, the problem was that class-A biasing was often cheaper.

Domestic tube amplifiers were generally A until solid-state rectifiers became cheap enough to supply bias-voltage.

Class A simplifies power supply design.

Side-note: amps big enough to want class B usually drive loudspeakers. Until after WWII, loudspeakers needed big DC field-current. Taking this in series with a class A stage is wonderfully simple and effective.

The J. L. Linsley Hood amp referenced by tv is forced to work push-pull class A. As is the White Cathode Follower. Nelson Pass and many others get bi money for BIG heatsink push-pull class A loudspeaker amps.

Push-pull makes class B possible and attractive but not mandatory.

> It seems crossover distortion would be easy to measure but...

But it isn't easy to measure, AND simple "measures" do not correlate with "sound".

> I'm trying to understand how elements of design affect the sound.

Self On Audio. He does not say much about "sound", but works to minimize flaws.

> Not an opamp but nevertheless (it's current-feedback amp), it's a neat example of a totem-pole output ...

That output stage can be adapted to a differential voltage input and be a classic op-amp. I think later Hood articles do this.

PRR said:

> "push-pull, class A" doesn't sit right
Why not?

There is a deep history of push-pull class A.

Click to expand...

Yes I know.  I think you believe I was saying that class AB or B is the way to go.  Not so. I was referring to only the name, not the actual class of operation.  And a bit more specifically, I suppose I was talking about discrete transistor op-amps which was what the topic starter had asked about.  Personally, I don't go any other way than 'A' for good audio unless I had to enter in to a little "AB" on big transients. 

Again, I just prefer to name these class A operation amps "Balanced" or "Symetrical", "Antiphase", "Totem-Pole", "Differential" or"whatever" myself. 
But I realize this won't catch on though . 

I just think it makes it less confusing when talking about them.  As John Roberts stated "Push pull is by definition class B".  Having to append "class A" to it because we pushed the B" bit further on out seems a little bit oxymoronic.  In "push-pull" "class A" (oops), we are not really pushing and pulling in the classic sense.  We are pushing and pushing at the same time. 

PRR said:

Class A simplifies power supply design.

Click to expand...

Yes it does.  And having antiphase betters the PSRR, and being perfectly balanced cancels 2nd, 4th, etc. but not 3rd, 5th etc. and etc.
Thanks for the lesson.

PRR said:

The J. L. Linsley Hood amp referenced by tv is forced to work push-pull class A.
As is the White Cathode Follower.

Click to expand...

I know.  I tend to think of the J.L.H. as being a class A common emitter stage having an active load though.
And I think of the W.C.F. as a Totem-Pole.  I have no idea what Eric White called it though?

Edits: Actually, this still isn't clear as they both have those qualities yes?  And also of being a common anode/collector stage with active load.  So?

PRR said:

Nelson Pass and many others get bi money for BIG heatsink push-pull class A loudspeaker amps.

Click to expand...

Yes he and they do.  Most of them are great sounding too. 

PRR said:

Push-pull makes class B possible and attractive but not mandatory.

Click to expand...

Don't worry, I'm not even attracted to the possibility so I won't be making it my own mandate.

While I'm here, I also have a problem with calling indiscriminately all professional lines "Balanced". 
Technically, they are only "Balanced" lines if they are balanced to ground.  But we can send from a transformer and receive with a transformer without being balanced to ground.  Then it would be differential-mode. 
I realize this won't catch on either

Thanks

p.s.  Having to say all this:

PRR said:

> Push pull is by definition class B

That's worded confusingly.

Push-Pull can be A or B (or C).

Push-Pull can be A for small outputs and B for large output.

Class B Audio *must* be push-pull.

But finding "push-pull" does not say how it is biased or loaded.

We may find a clue in the specs.

If you have a small idle supply current demand and a much larger max-output supply current demand, something appears to be going into class B. Probably the output stage.

But often the specs don't say enough.

990 has a modest idle current demand but can put HUGE output power. We'd guess it runs class B; in good audio "B" often edges to an "AB" mode for teeny output.

One of the ancient discrete modules was very Class A. To make reasonable output with inexpensive devices it ran so close to overheating that a 20% over-voltage would soon smoke it.

Click to expand...

Almost makes my point! 

Not so fast!

"PURE" Push-pull Class-A (at least in my book) can ONLY exist if two identical SE class-a amp stages are stacked on top of each other, like in this headphone amp (the bottom one):

http://www.tcaas.btinternet.co.uk/jlhphones.htm

or like Hiraga did it (pardon my French, pun intended, haha):

http://www.tcaas.btinternet.co.uk/monster27f.htm



But, if you only search for "SE-footprint sound", try this (see attachement).

It's a "pussyfied" JLH, good enough to drive 600 Ohm, will idle at 250-300 mA, depending on beta of output transistors, will keep you warm on long winter nights (see dissipation graphs, IOW use similar heatsinks you would use for a cca 5W amplifier), AND, most importantly, will generate ample supply of "good" hormonics.




And, if you want to investigate further, see what I did to the Melcor circuit here:
http://www.groupdiy.com/index.php?topic=45107.msg565318

tv said:

Not so fast!

"PURE" Push-pull Class-A (at least in my book) can ONLY exist if two identical SE class-a amp stages are stacked on top of each other, like in this headphone amp (the bottom one):

http://www.tcaas.btinternet.co.uk/jlhphones.htm

Click to expand...

I would tend to agree with you.  So why are we calling other things "push-pull, class A" as well? 

tv said:

or like Hiraga did it (pardon my French, pun intended, haha):
http://www.tcaas.btinternet.co.uk/monster27f.htm

Click to expand...

Le Monstre is a good one

tv said:

But, if you only search for "SE-footprint sound",

Click to expand...

I'm not searching for any particular sound. Only clarity in definitions so I am not left wondering what things are.    Giving some help to cancel harmonics is OK in my book. 

tv said:

try this (see attachement).
It's a "pussyfied" JLH, good enough to drive 600 Ohm, will idle at 250-300 mA, depending on beta of output transistors, will keep you warm on long winter nights (see dissipation graphs, IOW use similar heatsinks you would use for a cca 5W amplifier), AND, most importantly, will generate ample supply of "good" hormonics.

Click to expand...

A "pussyfied" Le Hiraga for line stage will probably do the same.  So maybe "push-pull, class A" should only be used when there is non cancellation of even harmonics generated within?  A legitimate question. 
Are you sure 250-300mA is enough to drive 600 ohms though? 

tv said:

And, if you want to investigate further, see what I did to the Melcor circuit

Click to expand...

I noticed your post but haven't looked closely yet.  Anyway, thanks for posting the designs and I'm sure theyre nice ones.
I have investigated the various circuit choices ability to cancel harmonics self generated. 

My only point here was that a definition "push-pull, class A" doesn't sit well with me only because it still doesn't tell me more of what is happening.  I see this and have to still look at the static bias and place an imaginary load on it to see if it leaves A.  I'm not even saying that leaving A to AB is bad.  Only that I still don't know the story.

Thanks TV.

It would seem easier to start at the beginning, in defining PP class A.  There are a billion pre-WWII PP class A amps.  Class B became wildly popular in the mid 1930's, then fell out of favor.  I have heard a mid 1930's class B speaker amp that sounded fantastic; it beat out several of it's mid 1930's PP class A brethren in a simple listening comparison. 

Push Pull versus Push Push seems mostly semantics to me, in regards to class A.    I can interpret from either direction. 

emrr said:

It would seem easier to start at the beginning, in defining PP class A.  There are a billion pre-WWII PP class A amps. 
Class B became wildly popular in the mid 1930's, then fell out of favor.  I have heard a mid 1930's class B speaker amp that sounded fantastic; it beat out several of it's mid 1930's PP class A brethren in a simple listening comparison. 

Push Pull versus Push Push seems mostly semantics to me, in regards to class A.    I can interpret from either direction.

Click to expand...

I suppose it is semantics.  It certainly has nothing to do with preference or better.  If you can interpret "Push-Pull" and "Push-Push" as being the same, then that is OK for you.  I will still try to put a better label on them (in my own head from now on) so the story is a bit more filled out for me.  And this is OK for me too.
Again, I realize this will not catch on.  I have tried before saying to very good engineers that a particular line is not "Balanced" but "Differential Mode" and that didn't work too.  I must be odd... I'm OK with that
Please, carry on with the more important discussion of the thread.

This is much ado about semantics. The final product (AC audio), involves pushing and pulling the speaker load, (or opamp output).  The distinction "class A" refers to a circuit whose active power device(s)  do not ramp down to zero current output, avoiding turn on delays and distortion between full off and on. Class A devices are always on.

My previous post was an oversimplification, while class B use active modulated push and pull devices that can turn off, class A will often also involve a second fixed push or pull stage working against the modulated stage. In some advanced class A designs both stages are modulated while still avoiding turning completely off. 

There are variants on class AB where the power devices do not turn completely off but they still call it AB because there is a major shifting of current output between the push and pull devices with near class B efficiency. 

Before we get too carried away about the merit of class A, it is worth note that early cheap transistor radios and consumer gear (before ICs became widespread for such applications) were predominantly class A discrete, and often delivered questionable audio quality. Many of the cheap ICs use class A internal paths, not because it's better, because it's cheaper. 

JR

PS: yes, "balanced" and "differential" are not the same thing with important distinctions we'll save for another thread. 

JohnRoberts said:

This is much ado about semantics.

Click to expand...

Yes it is. I'm sorry I brought it up.  I find it interesting though that it seems to have brought on strong emotional responses. 
And mainly ones to do with "quality".  I never mentioned quality.

JohnRoberts said:

The final product (AC audio)...  (lesson for the day removed for brevity)
...Before we get too carried away about the merit of class A...
JR

Click to expand...

Again, thank you for the lesson.  Who is talking about merits?  Not me, although I did express a preference but it wasn't as a point of contention.

Question: Why do a lot of engineers or moderators launch into such lessons?  If I were more sensitive,  I would be taking offense that they are already assuming I don't know quite a lot of obvious stuff and am a complete idiot.  I am confident enough with myself to admit when I don't understand something or have learned something.  I have done so on other threads.  I didn't learn anything from this one but:
Again, I am sorry for the diversion.  It is semantics and I was wrong to bring it up.  You are right on that part.

JohnRoberts said:

PS: yes, "balanced" and "differential" are not the same thing with important distinctions we'll save for another thread.

Click to expand...

Yes they are different.  I do not need to know the distinctions myself so no need for another thread on my benefit.  Again, was just expressing a tiny issue I have with sometimes incorrect labels.  No big deal. 

Peace and understanding to all

SRPP

http://www.freewebs.com/valvewizard1/srpp.html

Is it a push-pull or push-push? ...


One reason why I put up the JLH "output stage" (...) is that it is in fact a) "single ended" AND a  push-pull at the same time, but uses a "pair" of output devices that aren't driven "one from each other" (like a SRPP).


Some "amps" are "mirrored" or "symmetric" and of course ran in class-A up until the AB output stage.. which is good for "by-products" cancelation if you want clean.


And there is the important question:

"Do I want/need CLEAN gain or do I want HORMONICS?"


The latter IMHO tends to sound better when single-ended designs are employed (just look at the JLH FFT, practically the 2nd, 3rd and 4th harm. are "present", the higher ones go down into the noise floor, you aren't getting that with class AB unless a BRUTAL dose of NFB is applied.)

Euphonics EQUALS the "good" harmonics, IMHO.

The "final" answer was implied with this statement as well (again IMHO).


...err .. on the second thought, WTF I'm writing this? ...

8)

Jean Clochet said:

Question: Why do a lot of engineers or moderators launch into such lessons?  If I were more sensitive,  I would be taking offense that they are already assuming I don't know quite a lot of obvious stuff and am a complete idiot. 

Peace and understanding to all

Click to expand...

Don't take me bloviating as anything personal. If it was personal I would be quoting you, and even then I am usually quoting comments just to highlight a topic to better explain to the full group. Perhaps a little arrogant but i don't consider this a dialog between two people but threaded posts for the benefit of many.

Don't assume that because you understand the nuances surrounding a topic, that everybody reading here does. 

JR

PS: to your specific question, people are often inclined to answer with all they know about a given subject whether it actually answers a direct question or not.

Thanks John.  I didn't take it personally, honestly   Sometimes it's hard when you're reading or writing to convey or understand the real intent.  No face to look at for clues?  I am as (or more) guilty as anyone.

And I too launch into lessons easily, I think it is natural.  It's just interesting to observe I think

TV, I haven't read your post through yet but I will do when I have a bit more time.  The subjects are interesting to observe although maybe it's another distraction from the real thread?
Thanks TV, I'll get back later iif I may...