Fading between class A and AB

[kkrafs]

Is there any situation that would benefit of being able to fade
between class A and AB in a preamp or compressor/limiter?
Im including applications built on tubes, trannies and OP amps.

And i dont find any meta about wich OP amps to use,
or a substitute list, etc? Are this list all tube or have
you folks been discussing this previously? :grin:

 

[toby]

i remeber reading somthing about amplifiers working in pie class, which worked in class A with a low input and moved to AB with higher inputs.

 

[SSLtech]

"Current dumping".

Quad amplifiers did that, operating in class A for low-level stuff, transitioning to class AB for high-power excursions.

I suspect that's not what the original poster's question is all about though.

The difference is actually sort of subtle at times. If it bothers you, build something to do it... I doubt there's a great call for it though.

If you want class A, you use class A: not because it's class A, but because it sounds 'right' for that application. If class AB works well enough or even better, then class AB it is. A blend of the two will have the disadvantages of both, and in a glass-half-empty world that's a bad thing :wink:

Keith

 

[SonsOfThunder]

An observation here:

Usually circuits that are referred to a "classic" or typical class A are also single-ended. Class AB has to be push-pull. So I don't know of a way you could move back and forth?

Class AB is just a class B with biasing and you can add a lot of bias so that the circuit runs in class A mode, but this is different as I said.

Hope that helps clarify. Don't be discouraged by the negative response to your question though. Challanging traditional thought is a part of innovation.

Peace!
Charlie

 

[bcarso]

For certain voltage-controlled amplifier topologies, there are noise advantages to running class (A)B and distortion advantages to running in class A. These VCA blocks would be found embedded in various compressor-limiter products. I don't know of anyone who offers a choice of operating mode for internal details like this, but it wouldn't surprise me if they are out there.

The Quad current-dumping amp is a little different---there the class A low power amp triggers the high current stuff alongside it when the load current gets a little above a threshold, so that the class A amp never has to work very hard. Each section operates in its mode, with the current-dumping heavy lifting portion being very class B. There are some other subtleties that make the thing work better and are part of the Quad patent, although it remains a controversial amp among golden ears.

Brad

 

[PRR]

> amplifiers working in pie class, which worked in class A with a low input and moved to AB with higher inputs.

ALL audio Class AB amplifiers work in Class A for small signals. That's what Class AB means.

A transistor loudspeaker power amp may be biased 50mA idle and will run Class A up to 100mA peak, Class B for peaks of 100mA to 3 Amps. That's lean AB (my terminology).

A tube AB power amp will often be biased to stay Class A up to half power or more. Take about any of the classic hi-fi amps: biased around 50mA per tube, Class A to 100mA peak, going to B for peaks 100mA to 200mA. That's rich AB (my terminology).

Some oddballs like Fisher 7189, several Mcintoshes, and to some degree the bigger Fenders, run fairly lean for tubes (but rich by transistor AB amp standards). And there are some amps working so rich they barely cut-off a tube at peak power, and barely qualify as AB. In fact this is a matter of judgement, because tubes are so hard to push to zero current at high plate voltage that even the leanest AB tube amps don't really cut-off.

There are also wild/wacky combinations of small Class A pushed around by big Class B, as Keith mentions.

> Usually circuits that are referred to a "classic" or typical class A are also single-ended.

True, but there are a few crazy folks out there. Push-pull 2A3 10-watt cathode-bias, or the Williamson, are Class A (at least up to nearly full power). The push-pull output stages in bigger tube console modules invariably made rated power within the Class A zone. H-P 200AB oscillator output stage is resolutely Class A unless you severely overload it. And on the other side of the street, Nelson Pass and others build sand-state true Class A push-pull. While most transistor console modules lean to AB, we had that Flickenger that was for-sure Class A push-pull.

Note that ALL "Class AB" amps have Class A drivers somewhere. You can't make an audio amp that never goes into Class A. (Recently there have been some chips with Class A input stages that shift to Class AB/B under severe provocation, but not in normal linear duty.)

Generally you can make either type perform very well. It is dumb to pre-judge an amp by its "Class", even if you believe the blurb (there are many things sold as Class A that obviously are not). If either class has "a sound", it was not done right.

> For certain voltage-controlled amplifier topologies

OK, that's the exception. VCAs work hyper-close to the silicon and there the A/AB decision makes a difference. I remember when this was controversial.

 

[bcarso]

PRR: "Generally you can make either type perform very well. It is dumb to pre-judge an amp by its "Class", even if you believe the blurb (there are many things sold as Class A that obviously are not). If either class has "a sound", it was not done right. "

I think that prejudgment was what did Quad in among the golden ears. And they could have found a prettier name for the scheme than "current dumping"