Class A, AB

I saw some amps that are Class A, Class AB. Do these amps switch to AB at a certain power level? Where do they usually change to AB?
Most amplifiers are called "Class A/B" because the operate in Class A at very low power outputs, then switch to Class B operation at higher outputs. A 100 watt amp may put out 5 watts in Class A, then switch to Class B above that level. Even the most powerful A/B amps can only output about 15% of rated power in Class A. Typically, it's more like 5%.
OOPS- I should have said, "Most Solid State amplifiers".
Neilmc , if the amp is AB it will be class A (does not go into cutoff) to a certain point and above a certain power the amplifier will operate in the 'B' region, IOW part of the amp goes into cutoff as the waveform amplitude increases. There is no such thing as combining Class A with Class AB- such an amplifier is by definition Class AB only.

Class A means that at no time during the amplification of the signal does the amplifier or a portion of the amplifier go into cutoff, such that only a part of the circuit is handling the signal. All single-ended amps are Class A BTW.
"...There is no such thing as combining Class A with Class AB- such an amplifier is by definition Class AB only."

I beg to differ. I believe quite a few Plinius amps do exactly that, i.e., combine in a single amp either class A operation or class A/B. Which modality one employs/enjoys at any particular time is up to the end user via a simple toggle switch. In my mind, that is a simple yet monumentally significant design feature which even taken alone, shoots the Plinius amps way out in front of their competitors. Hard to fathom why it is not a more universally adopted design.
My Crown D-150 is described as a class AB + B. Which means, I guess, that it is a class AB up to power point and then switches to B.

Salut, Bob P.
I too have Class AB amps that run in class A mode until two-thirds power then run in AB mode. I a not sure of the design principles used - perhaps a sliding bias or the final ouput stage only goes to work at the higher operating range....
Maich, if you were describing two independant amplifiers in one chassis that somehow both drove the same speaker without blowing each other up then you would be right. Personally I've not heard of such a thing. The definition of Class AB is: Class A to some given power, it might be 5% or 75% (doesn't matter), and then 'B' operation after that.

Your amp, Inpepinnovations' amp, and Shadorne's amp all fall within that definition. There may be some nuance - Class AB1 or AB2, but they are all Class AB.

In recent years some companies have sold 'enriched AB' amplifiers, claiming a high level of A operation, but what you have to get about that is that that is marketing. 50% Class A operation is unusual, and I think an amp company would have to come up with *something* to point that out, so I am not saying that this is misleading on their part, but it can cause confusion with the market.
The definition of Class AB is: Class A to some given power, it might be 5% or 75% (doesn't matter), and then 'B' operation after that

Surely an amp that is capable of operation to 75% of power in Class A is much more expensive to build than one that runs only Class A to 5% of same rated power.

Why does it not matter or am I missing something?
From a Plinius manual:

"The Plinius [model] power amplifier operates in class A and is therefore capable of generating a moderately high temperature..."

Later in the same manual referring to the toggle switch I referred to:

"...A/B position reduces the bias on the output stage to operate in class AB. This is a bias configuration used by the majority of High End amplifier designers for their products."

Sounds to me like it operates in class A as well as A/B.
Example to explain:
Class A: If amp peak current is 10A then with no signal all 10A flows thru the + rail transistors also the - rail transistors and none thru the speakers. At half power 5A flows thru the speakers and 5A flows thru 1 of the transistor rails while 10A flows thru the other (depending on polarity of signal).
Class A-B: If peak current is 10A, then with no signal maybe only 1A (amount can vary according to how heavily biased it is) flows thru both transistor rails and none thru the speakers. Above 1A of speaker current one transistor rail turns off and the other carries the full speaker current (which rail depends on polarity of signal)
And in class A mode, Plinius amps run hotter than you want to touch for very long, even while outputting no music. In class A/B, they are almost stone cold even while operating at amazing capacity.
Gyphon makes several amps that operate pure class A and consequently are very big, very hot , have huge power supplies, and only put 150-175 watts into 8ohm loads due to inherent class A inefficiency. Their power supplies do allow them to double down into difficult loads ( 300w 4ohms, 600w 2ohms ). There is a switch which allows for AB bias operation for "non critical" listening which saves electricity and heat. I never use it. I always run class A. Maich, the reason pure class A operation is rare in ss amps (even with a switch for AB)is for the above reasons and the expense of the power supplies, required capacitance, and heat sinks. It's nearly impossible to build one thats operates in pure class A with significant output that's less than 200 lbs or so and they get very very hot and use lots and lots of electricity. 95% of the market would rule them out on these grounds alone not to mention the $/watt ratio they represent to less picky consumers. Most people would not justify the sonic benefits they(class A) confer.IMHO they are great for difficult loads (4 ohm and under) or for high efficiency speakers. I personaly am hooked on these big ugly beasts. - Jim
Here's my question about the Plinius, and others, but I'll use the Plinius SA100 for example.

Rated at 100W stereo into 8 ohms. More into 4 ohms and all the way up to something like 650W into 4 ohms as a bridged monoblock. It would be silly to think all of that would be class A. How much, I not really concerned about but it does indicate that the load plays a role. It still draws the same 500W from the wall, in class A bias, which was about 40% efficiency for that 2 X 100 watts. I haven't bothered to put an ammeter in there to see if the draw jumps up when I get it loud. That would have to be VERY, VERY loud.

The origional Pass XA series (unique feedback/stabilization) and most tube amps are different. They don't vary output to the load, as much.
I have a Forte 4a which was advertised as a "class A" amp. The amp is biased higher than usual. It is still NOT a true, pure "class A" (100% class A operation)
It uses the most electricity idling! More than when playing, and thus is HOTTER just sitting doing nothing if it is on.
And yeah, class AB IS part class A operation and part class B operation. And nearly all amps that are called class AB are just class A up to a point. Most biased pretty low. The exceptions that are biased fairly high are usually called "class A" operation.
Very few amps are 100% pure class A. (overkill I think)
I am the opposite orientation on the A vs. AB choice -- after some intial experimentation and excitement about the sonic benefits of the class A mode, now I almost never go that route -- not so much concerned about the electrical use as the heat factor -- that is troublesome in a way that is not so easily overcome. Also I will say that in the case of Plinius at least, they are phenomenally close in AB, so I guess that makes me a "non-critical" listener for the most part.

According to the SA100 manual (page 6) "The current from the mains/line supply by the amplifier in BIAS A is now approximately 1,000 watts." So looks like the current draw actually is up by quite some margin to use it this way, also should be constant (see below).

Page 21 same manual:
"A class A circuit topology is one in which the total current the amplifier is capable of delivering, is kept flowing in the circuit regardless of the demand....
In a Class A circuit, current draw should be constant therefore there is an absence of the power supply modulation common in Class AB design amplifiers."

This under the heading of WHY IS CLASS A BETTER? So again, it sounds to me like these amps are switchable from Class A to AB.
>>Also I will say that in the case of Plinius at least, they are phenomenally close in AB<<

That depends to a large extent on the speaker as well.

It's very difficult to evaluate an amplifier based on one speaker.

Your Plinius may sound terrific on speaker A and like a dog on speaker B.

My analogy is also applicable to tonearms and cartridges.
Oh yes, there is little doubt that when it comes to audio review, people are not hearing the same thing. That is why the only real way to determine if a given component is for you, is to buy and listen in your own environment -- enter Audiogon.

Interestingly, I try to stay away from really narrowing down what something sounds like because of this factor, and whenever I break from that, inevitably it has a predictable result. Had a buyer recently work really diligently trying to get me to characterize the sound of a piece I was selling. I explained my position, but gave at least that it was one of the more "dynamic" pieces I had heard. He bought it, and was happy, but one of his first comments: "Does not strike me as particularly dynamic." All I could say is "there you go..."
Maich, I didn't mean to imply anything about your listening. I used quotes to emphasize that Gryphon refers to AB that way in their manual. Sorry if I offended you - jim
Oh, not at tell the truth, I was not even referring to your use of the term "non-critical" -- the Plinius manual also refers to AB listening in those same terms. I was just amusing myself by restating that characterization because it strikes me as funny that my listening habits might EVER be considered "non-critical" -- I ALWAYS, even on a moment by moment basis, find something to criticize in whatever sound I have going (but nothing that ever really prompts me to want to use the Class A switch, I must say...)
Maich, it was not clear to me initially that you were talking about a switch on the amp to change class of operation.

There are a variety of amplifiers made over the years with this feature- we have built a few ourselves.

Generally, when switching to Class AB, the power supply voltage is increased and the bias current is decreased whether the amp is tube or solid state.

However to be accurate, the amp when in the AB mode is that and that only. FWIW most amplifiers have noticeably smoother sound and greater authority in Class A, as distortion cancellation occurs throughout the entire waveform where in AB mode it does not.

Shadorne, I used the example of 75% to make a point- I don't think there are any amps out there that operate with that much 'enrichment' in AB mode.
Shadorne, I used the example of 75% to make a point- I don't think there are any amps out there that operate with that much 'enrichment' in AB mode.


Most of the Chord line of power amps are 'A' weighted to two thirds power or 67% - not far off 75%.


While we are on the discussion of Class A, remember that the reasons for these amps, with massive heat sinks and huge power consumption, are to avoid crossover distortion affecting 3rd, 5th, 7th and 9th and much higher odd-order harmonics which can work its way up from a room pounding mid/bass frequencies to be audible, even in tiny amounts, in the 2 Khz to 8 Khz range(where your hearing is naturally very sensitive and where the harmonics may bear little in relation to anything musical).

A trick that works:

Rather than use only a super massive Class A (very expensive) to drive your entire speaker. The cheap trick is to bi-amp your speakers and use Class A for the tweeter (only very modest power needs) and use AB for your woofer - this way you keep crossover distortion completely away from your tweeter AND you have a massive powerful punchy bass without having so much heat and huge expense. Anyway just a thought...and although it may be a cheap option it will still sound sweet.
Thanks. Please ignore my newer post on the subject.