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

Hey, IMHO the OP has "asked for it" right here in reply No.5:
http://www.groupdiy.com/index.php?topic=45117.msg565032#msg565032

Asking about single-ended vs. "opamp sound", the latter being "not quite as fantastic".

So, all examples I posted here serve to underline some basic differences in "sound" - or to underline what usually happens to be "different" in single-ended designs, making them sounding perceptually better. (perhaps I should get paid for taking time to explain...)

Also, when reading into all that, one should be aware of our - let's say - brain's "reward systems". That's what it's all about. I.e., what "tickles your ears", in fact, tickles your brain. That's why certain fuzzes (a two-transistor job) cost more than a 0.0001% THD headphone amp with a power supply. Because they "tickle it", while a headphone amp "just reproduces". An important difference, in fact, "a world apart".


see here for some introduction
http://pedagogic-verses.blogspot.com/2011/01/music-and-brain-dopamine.html


these are broad topics... but that's how it all works in the end.

I think a lot of moderators and other well educated types give lessons, not necessarily to address the person posting, but for the silent ignorant masses reading along.  Helps keep the stupid questions down, assuming they can read with comprehension.  I'm usually too impatient to play teacher. 

I gave up about balanced audio a long time ago too.  There's really not much of it out there, but used to be a lot of it in telephony and broadcast.

tv said:

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

Click to expand...

TV, I would try not to use either name here but would define it simply as shunt regulated.  The same as a W.C.F.  - both using two devices that contribute to load current.
The J.L.H. example you used before is the same to me as well, just not self contained as it uses an external phase splitter.  Just as a W.C.F. can use an external phase splitter. 



I don't know that I will have an opinion here as to specific harmonic cancellation/addition as an effect or preference.  This is, to me, deeper than simple explanations allow.  And it doesn't help clarify things other than someone's preferences or approaches.
We know there is more than one way to reach the end goal. Some good, some not so good. But there is no way to define the "Best Way".

Edited for clarity.

emrr said:

I gave up about balanced audio a long time ago too.  There's really not much of it out there, but used to be a lot of it in telephony and broadcast.

Click to expand...

Thanks EMMR.
Are you saying that mostly you use single-sided unbalanced send and receive?  If not, then I hate to be semantic again but most pro stuff IS Balanced because the send and(or) receive ports are electronic rather than galvanic these days.
A lot of old differential-mode lines did end up being balanced somewhere though (with a centre tap grounded) to help with Hi Frequency.
My example with another good engineer was related to me putting a filter in only one side of a short, fixed differential line to save on costs.  I was told it would unbalance the line which, if it is differential-mode, it won't of course.

I think I've probably said this before, but for the benefit of onlookers who may be somewhat confused by all these semantics, it's worth noting that 'single ended' in this context is a very different concept to 'unbalanced'. You can run a 'single ended' circuit (where one active device amplifies the entire waveform, as opposed to 'push pull', where a complementary pair of transistors each handle 180 degrees) in a balanced configuration if you want to. You can also run a PP stage in balanced / differential mode if desired.

In some designs PP circuits measure higher amounts of IMD (they tend to in my tests) when they leave the Class A region due to the fact that it's near-impossible to create devices that match so well they mirror each other in distortion artefact (hence the IMD is created because each phase contains different harmonics, which, when added together create sum + difference harmonics). Basic SE circuits tend to measure highly in terms of even-order harmonics and some designers have gone to great lengths in order to linearise SE circuits, on the basis that their inherent distortion is more benign.

There's a lot of generalisation in what I've said (I'm sure there are circuits which turn all these dogmas on their head). As John says, generalising in these circumstances is short sighted and there are many ways to skin this animal, but it could be helpful background to anyone bemused by the concepts / nomenclature in this thread.

Justin

thermionic said:

In some designs PP circuits measure higher amounts of IMD (they tend to in my tests) when they leave the Class A region due to the fact that it's near-impossible to create devices that match so well...

Click to expand...

There's also the issue of what is happening as you get near to approaching grid current in valves.
And if there's a transformer on the output , in AB, the valves see less inductance/load. 
In A, the valves see the full traffo winding inductance & load in series. 
And also the current imbalance giving possible core saturation.
That's why different spec transformers for A and AB with the same valves.  And a beefier driver.

Probably better to have a pretty clean AB on peaks though than just running out of A juice.  Another thread again...

I'll probably regret stirring this stew............

> I think you believe I was saying that....
> ...Not so. I was referring to only the name

I understood you were saying "push-pull, class A" doesn't sit right. Since class A push-pull IS (in electronic theory) a very clear object of study, I think it should "sit right" in your mind. It makes discussions clearer.

> I would be taking offense that they are already assuming I don't know quite a lot of obvious stuff

None of this stuff is "obvious". It has to be learned. And eventually organized for easier reference and discussion.

Note too that this is a PUBLIC forum with 9000 nominal members, 20,000(!) views on some threads... most of the regulars are very aware that many members are "silently" reading and many need to be brought up-to-speed on some points that the participants already know.

I don't see "idiot" anywhere except your post. That's completely off-base.

> Why do a lot of engineers or moderators launch

Don't be over-awed by a moderator hat. Here, "moderator" is un-paid extra duties. John won't stay away, so Ethan gave him a mop in case he sees a mess (spam, nasty-talk). But he's here to learn/share/teach because he wants to be here, not to crack heads.


> This is much ado about semantics

Indeed.

The semantics of amplifiers are well-organized, in theory.

There's small-signal (invariably SE A).

Next chapter you push the amp to the max for A Power.

Next chapter is push-pull (or push-push... totem or transformer does not change the theory). Some nice improvements.

Now back up. Off-bias the amp for class B. Show that SE class B won't make "undistorted audio", but two antiphase B units can combine for "undistorted", AND with wonderful improvement in idle-heat and max-power efficiency.

In here the text should touch class C, but there's no "undistorted audio" possible here (it was very important that young EEs learn their C because radio was major jobs.)

Comp/quasi-comp, totem or transformer (output or driver) are implementation details. Yes, this is critical. But it isn't basic A or B, SE or P-P, as the thread-starter question asked.

> early cheap transistor radios and consumer gear ...were predominantly class A discrete, and often delivered questionable audio quality...

Not so predominantly. Huge numbers of pocket radios (and other gear) used "Class AB/B" outputs with such sloppy biasing that they would not play soft. (The alternative with poor tolerances was possible runaway or bias-trim, both bad for business. Implementation details.)

> place an imaginary load on it to see if it leaves A

If it leaves "A", it better be designed for "B" operation; and surely is.

If it was _sold_ as "A!", and leaves A, that's marketing misrepresentation. The sins of marketing are too numerous to count.

> "PURE" Push-pull Class-A (at least in my book)

"Equal and opposite swings" is pure enough to be "push-pull class A". Your book seems to overlay this with requirement to be push-pull class A in ALL stages to be "PURE". OK, but that's one box of the 69-box chain from artist to CD/MP3 to speaker. Some days we have to take one stage at a time.

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

Broskie's analysis shows it is both depending on load (not level).

Oh.... "push-pull" is catchy but open to confusion. "Equal and opposite signal swings" would be a better phrase, but won't roll off the tongue. "Equal and opposite signal swings" can be done either with transformer or totem-pole... implementation detail.

There "is" also a "push-push" mode. Signal currents are NOT opposite. (And to be useful, they must also be unequal.) This is a "bad" mode, wasteful. FWIW, both the WCF and SRPP enter this mode for large load impedances (larger than Rp). To get 1mA _signal_ current, they shift 2mA in top device and 1mA in bottom device. 2mA-1mA= 1mA output signal. But the pair is working 3X harder than necessary. You can, for hi-Z, get a better amp with both devices parallel resistor-loaded (but this sacrifices the option for useful power in lower-Z loads). I've only seen this as totem-pole. I guess when you have to pay for a transformer you don't consider such wasteful working. (Actually there may be a tube guitar amp which can be mis-tuned this way for less power more flavor.)

> think of the J.L.H. as being a class A common emitter stage having an active load

It is not any sort of emitter follower. There is an input section, then a bootstrapped-resistor (or CCS) loaded driver, then two same-type output devices. The driver acts as a differential OUTput current source. 200mA down through the resistor is split 100mA to each output idle, 0-200mA to the top and 200-0mA to the bottom. The current sources have voltage compliance (are not low-Z). This current is multiplied by hFE of the output devices to be load current. 0A-1A-2A in top device, 2A-1A-0A bottom. The difference, +/-2A, is load current. Output impedance (before NFB) is high.

PRR said:

I'll probably regret stirring this stew............

Click to expand...

I give my word, except for a couple of points, I'll let the stew simmer...

PRR said:

> Why do a lot of engineers or moderators launch

Don't be over-awed by a moderator hat. Here, "moderator" is un-paid extra duties. John won't stay away, so Ethan gave him a mop in case he sees a mess (spam, nasty-talk). But he's here to learn/share/teach because he wants to be here, not to crack heads.

Click to expand...

Yes I understand.  I was asking the question with regard to human nature.  I did say that I am just as guilty of it.  I find it interesting, that's all. 
Someone asks a simple (on the surface) question and an engineer instinct is to write down all he knows on the subject. 
Maybe it is a desire to instruct?  Maybe it is beating one's chest?

PRR said:

> This is much ado about semantics

Indeed.

Click to expand...

And indeed from  me too.  I thought I'd agreed so.

PRR said:

> place an imaginary load on it to see if it leaves A

If it leaves "A", it better be designed for "B" operation; and surely is.

If it was _sold_ as "A!", and leaves A, that's marketing misrepresentation. The sins of marketing are too numerous to count.

Click to expand...

I have no idea if the amps the original poster was talking about were sold as A.  It was a long time ago.  But I could place an imaginary load and a voltage requirement on it to see if it is A for that particular situation. 
Of course, it probably doesn't really matter if and when transition from A happens if the job was done nicely with regard to details.
So the question of whether it is A or AB is, itself, possibly semantic too?

PRR said:

Oh.... "push-pull" is catchy but open to confusion. "Equal and opposite signal swings" would be a better phrase, but won't roll off the tongue.

Click to expand...

Exactly. 

PRR said:

> think of the J.L.H. as being a class A common emitter stage having an active load

It is not any sort of emitter follower.

Click to expand...

Yes I know.  I didn't say emitter follower, I said "common" emitter.  The emitter being common for in and out. 
In other words, a gain stage with an active load
I did liken it to an SRPP and a W.C.F. though.  But only in regard to it, and them, being shunt regulated. 

Thanks. 

If any readers here find themselves confused, just remember that end users often think 'Class A' is just a 'quality' of circuitry, not an engineering term... Like 'Grade A chicken'... Marketing has a lot to answer for. I can think of one client I heard referring to 'class A resistors'...I knew he'd be trouble from the off...

J

class A resistors

Click to expand...

Ha! I always buy resistors that have tube warmth.

There are also "classes" of wiring or insulation that could be a little confusing with amp classes..

JR

> I didn't say emitter follower, I said "common" emitter.

I apologize for reading that part way too fast.

PRR said:

> I didn't say emitter follower, I said "common" emitter.

I apologize for reading that part way too fast.

Click to expand...

Thanks.  Not a problem, I understand.