Class A or AB ?????

G

granville

Joined
Feb 11, 2006
Messages
138
Reaction score
0
Location
The edge of reality
I would be interested in a discussion from those who know what they are talking about, re the relative merits of class A over AB. Also the best route to achieve a top
Quality 4 channel amp.
The reason for my enquiry is I wish to try a class A/AB in place of my 4 channel class D amp. I have used a Parasound HCA1200 and thought that an improvement over
A Sugden A41 (I think) And I have recently acquired, for free, an Alchemist Forsetti
ADP20A Mk11 which I yet to try with my speakers.
You can find on Ebay circuit boards for class A amps; Pass Labs Aleph 5, KSA250
and others. This route appears to offer the best value but will you achieve a stable
high quality product at the end ? Suggestions and ideas appreciated.
 
Take a look at 4 channel current dumping amps also. The small size of the boards and their simplicity make them very suitable for multi channel use. There are a couple of examples on the web, one a prototype apparently built by a Quad engineer, the other built by Dada.
 
Class-A amps, when well designed, offer the ultimate potential for low distortion because one of the major problems (crossover distortion in the output stage) is totally avoided. But achieving this end requires the whole damn amp is designed with this in mind - just bunging loads of bias current through the output stage is not enough; a claim of 'class A' alone is no guarantee of even good performance.

The penalty to be paid for Class A is massive electrical inefficiency, because the output stage has to run at least 50% of the maximum possible output current all the time, even at idle.

If you can get away with 10-25w/channel or so, then go for it. But if your current power amp is say 100-200w/ch and you 'need ' such power (big room, inefficient speakers / or are deaf) and you want class A to compete on such power outputs, then you'll be sorely dissapointed - it could require an amp that idles at up 600 to 800w per channel. And it will weight a huge amount commensurate with the necessary PSU and heatsinks.

Also be aware that very few 'Class A' amps actually are Class A all the way through due to this thermal problem. Take a look at the classic Krell KSA50 - a true 50w/ch in class A, and unavoidably enormous.


Hence most amps are class 'AB' or just B - designs running just enough output stage bias to deal with corssover distortion, and no more. As with most good things, listen to a few and pick the one *you* enjoy most.
 
Class A is certainly the ultimate both in measured and subjective performance.
As the previous poster said though, there are VERY few genuinely class A amps about. There are plenty which claim to be class A and several that many enthusiasts assume to be class A 'cos they get quite hot.
They become fairly impractical above 25 Watts or so. The Krell KSA50, mentioned previously, is indeed one of the most powerful genuinely class A amps around and that needed fan cooling to be as small and light as it isn't :D
My own homemade class A amp has two fans blowing continuously and is over a meter deep and gives 25WPC.
 
A good example of a 'Class A' amplifier that isn't would be the MF A1.
20wpc rating but only about 5w per side is available in Class A IIRC.
 
Even the large Pass labs class A amps only run In class A to about 70% Of rated output. Surely the way tong is an amp that you like the sound of rather than chasing a particular technology.
 
Did I read somewhere that output transistors are so much better than they used to be: there is no particular advantage in Class A over Class AB.

I ran a Sugden A48 for years - I think is stayed in Class A up to 10 Watts - so that meant most of the time.
 
sq225917 said:
Even the large Pass labs class A amps only run In class A to about 70% Of rated output. Surely the way tong is an amp that you like the sound of rather than chasing a particular technology.
Click to expand...

Yes, and even something like the old NAIM NAP circuit running on very low bias produces very little THD - <0.01% IIRC.

Peter Walker made the point decades ago that even an old 303 amplifier produces around 100 times less distrotion than a vinyl source. That really puts things into perspective.
 
Labarum said:
Did I read somewhere that output transistors are so much better than they used to be: there is no particular advantage in Class A over Class AB.

I ran a Sugden A48 for years - I think is stayed in Class A up to 10 Watts - so that meant most of the time.
Click to expand...

You can use sliding or tracking bias to eliminate crossover distortion, or something like Doud Self's crossover displacement (XD) circuit to supress distortion. These will reduce THD (which includes crossover) to better than 0.001%.
Once you reach those levels - and I'd argue considerably higher - it becomes a non issue.
The distortion distribution is just as important as the raw headline THD figure, but when you get down to these incredibly low levels you just wont hear it, no matter what the order.
We got there in about 1978!
 
MF do now make a 100watt entirely (they claim) class-A amp.

http://www.cinenow.co.uk/articles/22755-musical-fidelity-ams100-power-amp

834.jpg


Ypsilon also make a 50watt class-A monoblock.

http://www.ypsilonelectronics.com/default.htm

ces2007-photo.jpg
 
finally sorted out the music rooms at home, one for the wife and lady friends,
front end, sony dvns900v dvd/sacd, b200, kef ls 3/5a's
my room , avid volvere/sme309/ at 0c9ml, tuner m.f. a5 dab, m.f. 5.5 modified cdp, m.f. a1 pre, m.f. m6i power amp, annd my beloved tannoy's , bliss, cheers to you all.
 
Rob, Actually the MF A1 is a bad example as it is class A!
The A370 is more of a typical example.... Most considered that to be class A but it only made about 20 of its 185WPC in class A.
There is rather more to the advantages of class A than just absence of crossover distortion BTW ;)
 
Arkless Repairs said:
Rob, Actually the MF A1 is a bad example as it is class A!
The A370 is more of a typical example.... Most considered that to be class A but it only made about 20 of its 185WPC in class A.
There is rather more to the advantages of class A than just absence of crossover distortion BTW ;)
Click to expand...

On other benefits of Class A, are you thinking of mutual conductance issues around devices operating in A/B?
That was one of the justifications Quad gave for developing current dumping, with a Class A control amp and Class C dumpers.

I'm surprised to learn the A1 runs in Class A - looks too small for a total 40w capability. Then again it can get extremely hot - like Ouch! hot.
Decent amp though, but I prefer the B200.
 
Much more constant and usually higher transconductance yes, leading to lower and more constant output output impedance before feedback, no Gm doubling distortion, no switching distortion (as opposed to crossover... there is a differance), constant drain on PSU, lack of signal frequency half wave rectification products on the power rails and much easier setting and maintaining of quiescent current.... could probably think of sme more but that's a few to be getting on with :D
I seem to recall having measured an A1 at only 14WPC with both channels driven to point of clipping on a sinewave drive and into 8 Ohms.... and being surprised to find that yes it was pure class A to that point! No wonder they self destruct due to the heat.
I find the A1 a bit "soft" personally and also prefer the B200.
 
constant drain on PSU, lack of signal frequency half wave rectification products on the power rails and much easier setting and maintaining of quiescent current.... could probably think of sme more but that's a few to be getting on with
Click to expand...
I'll give you most of them, but those two..not necessarily. The total power drawn by a class A amp can be notionally constant, but the load on each rail is not..it can vary between 0 and 2x (Imax,out) each cycle.

- And regardless of how the 'class a' amp is approached, it means the amp 'sees' maximal voltage ripple on the PSU at all times, so it's own PSRR had better be damn good. Not necessarily an easy task either! Anyway, I'll continue bashing at the DIY efforts here until I"m happy with it. Might be a while ;)

Best wishes for an 'A' class christmas all.
 
I was wondering if I'd get pulled up on those for exactly the reasons you state! yes neither are notionally perfect.... and the advantages of constant drain on the Rails are inded offset by the need for a higher PSRR. However, if one is going for full on class A then you know you are facing some heavy penelties in terms of size and expense of many of the components needed. Therefore, you will probably be building a PSU that can give low enough ripple in spite of the heavy current drain...
The lack of half wave rectification products is the much more likely to be a moot point because the impedence of the rals is likely to be low anyway.... It still results in much lower radiated fields from the wiring though, which I have found to be capable of giving 0.3 % THD from an amp capable of 0.003 % THD when not correctly dealt with.
 
I measured the output in an A1 many years ago feed from a sine wave sig gen and to me at least displayed clearly third harmonic distortion (cross over ) text book style, for low too moderate power levels valve amps can do not only class A but no feedback
 
Third harmonic distortion and crossover distortion are not the same thing at all.
To do these measurements properly you need a specialised very low distortion sine wave source. An average bench signal generator will have 0.3 - 1% distortion so you just end up measuring the generators distortion!
Solid state can also do class A and no feedback......
 

Latest posts

Back
Top