Class A vs AB -- Do They Really Sound Different?

[Doodski]

JBL supertweeters which usually don't get used.

Hehehe @ sooper tweeters. I like the sound of that. I like a good top end.

 

[MakeMineVinyl]

Carbon comp is a loony-tunes choice. Used in the original because that was basically the only low-cost choice in 1960. Metal films are in every way superior, but that means carbon comp will now command a premium.

The MK IIIs are original which I built in the 70s, so they have their original resistors. The only modification I made to these amps is placing NOS military oil/paper capacitors in the signal path, because I had them and I could.

 

[Ken Tajalli]

The MK IIIs are original which I built in the 70s, so they have their original resistors. The only modification I made to these amps is placing NOS military oil/paper capacitors in the signal path, because I had them and I could.

What military caps?
I am fan/user of K40Y s.

 

[MakeMineVinyl]

What military caps?
I am fan/user of K40Y s.

They're made by Guideman and are glass bead sealed with metal outer cans, and are individually serial numbered. I picked them up at a surplus store decades ago.

 

[MakeMineVinyl]

So the Dynakit MKIII has some history. Here's a dated bit of literature with wonderful schematics.

Yes, its one of the classic amps.

 

[Doodski]

They're made by Guideman and are glass bead sealed with metal outer cans, and are individually serial numbered. I picked them up at a surplus store decades ago.

Is there any risk of the 3 Ps, Cs and Bs? As is PCBs in the oil.

 

[Doodski]

Yes, its one of the classic amps.

That schematic would look wonderful blown up and mounted in a frame. If it was old school ammonia exposed blue draft copies it would be even better. Give it a enthusiastic bent.

 

[Ken Tajalli]

They're made by Guideman and are glass bead sealed with metal outer cans, and are individually serial numbered. I picked them up at a surplus store decades ago.

Nice.
Russian vintage K40Y's look similar, and are very good too.
But as anything vintage Russian , tolerances are all over the place.
I bought a box full, and sorted them out, matched pairs.

 

[MakeMineVinyl]

Is there any risk of the 3 Ps, Cs and Bs? As is PCBs in the oil.

As long as I don't open the caps up and use the oil on my salad, I'm golden.

 

[Doodski]

As long as I don't open the caps up and use the oil on my salad, I'm golden.

hehe... It's the price one pays.

 

[MakeMineVinyl]

Nice.
Russian vintage K40Y's look similar, and are very good too.
But as anything vintage Russian , tolerances are all over the place.
I bought a box full, and sorted them out, matched pairs.

These are awaiting a home in some future project.

 

[Doodski]

Some nice lathe canned metal caps in there.

 

[levimax]

I picked up a Dynaco ST-70 cheap on Craigs list not working right.... it would sometime turn on but then would blow a fuse... or blow on right away. Turned out it was the PIO caps which were shorting out against the PCB, it was not the first thing I checked . The cap values were wrong and I replaced them with regular caps and it has worked fine ever since.

 

[DonH56]

All correct and agree with, BUT:
Class A is not only superior in sound quality, but also in measurements! it is not a hype, it is based on engineering realities and science.
Any output device, be it solid state or Tube, has a linear range of operation and some non-linear ranges. Biasing the device bang in the centre of its linear range, is called class A.
This makes sure you get max. linearity of operation before resorting to damage limitation, such as feedback - which incidentally works very well on a class A configuration with a simple local-feedback.

Most of the class A amplifiers I have measured over the years perform worse than their AB (or D) counterparts. Perhaps our realities and science differ.

To my knowledge:
Class D amplifiers operate in a manner of speaking, akin to switch-mode power supplies.
In that operational mode, they produce huge amounts of high frequency noise. They ofcourse filter it out, better designed amps do a really good job, but some of that may "leak" into all sorts of places, The ground plane, the PSU, the mains! and even input terminals.
This leakage may find its way back to whatever is connected to the amp, such as a DAC.
DACs, in particular don't like RFI!

AC leakage from 50/60 Hz lines and corresponding 100/120 Hz power supply rectifiers is much worse on many "linear" products -- look at the power supply spurs. It is a problem for many designs irrespective of class of operation. Class D has the advantage that the noise is well above the audio band, ditto for SMPS.

DACs, in particular delta-sigma DACs that dominate the audio world today, have their own high-frequency switching noise.

No point in arguing religion on the 'net...

 

[Ken Tajalli]

Most of the class A amplifiers I have measured over the years perform worse than their AB (or D) counterparts. Perhaps our realities and science differ.

No point in arguing religion on the 'net...

You are obviously qualified for me to ask this question.
from a theoretical point of view, can you explain why a class AB amplifier should be superior to Class A in performance ?
I am talking performance , not practicality or efficiency .
thanx

 

[restorer-john]

My experience aligns with @DonH56. Class A amplifiers are often worse than A/B due to the huge standing currents, which often result in more noise, PSU modulation (hum/buzz etc) and obviously, considerably less available power on tap.

I have a TOTL Marantz PM-95 (and a PM-94 class A/AB) MOSFET integrated from 1990 with switchable Class AB (120wpc) or Class A (30wpc) and apart from the increased physical transformer noise when switching to Class A and the massive increase in heat produced, the subjective difference is zero, that is until you run out of power in Class A and it is very obvious.

A huge amplifier, 27.5kg built on a diecast chassis, all copper plated and several kilogram heatsinks to get rid of the heat in class A.

 

[DonH56]

You are obviously qualified for me to ask this question.
from a theoretical point of view, can you explain why a class AB amplifier should be superior to Class A in performance ?
I am talking performance , not practicality or efficiency .
thanx

Not really qualified, as audio amplifier design is not my day job. But @restorer-john provided excellent examples in his post. Here are my thoughts, though hopefully some of our more learned audio designers will chime in to correct my mistakes:

The class A designs I have seen have much higher (often an order of magnitude or more) noise and distortion than comparable or higher-powered class AB designs. How much of this is design choice (e.g. lower feedback) and how much is a limitation of the operation (higher bias means more noise) I could not say.

Class AB offers much more power and the potential for even larger dynamic headroom for the same size, weight and power (SWaP) as a class A design which has much lower power output, greater heat, and typically (IME) higher noise and distortion. The best a class A output stage can do is about 50% efficiency (for push-pull, ~27% for single-ended) so at least half the power is wasted heat. The heat shortens the lifetime of the components inside and increases thermal (Johnson) noise. The high bias current increase current (shot) noise. Additional devices are needed to support the constant current in class A designs, which means they add their own noise (from internal resistances and such) and reduce bandwidth (from all the extra parasitics, though of course you can beef up the driver stages to compensate). That also reduces slew rate that can be achieved and potentially increases TIM (transient intermodulation distortion). Class AB amplifiers typically have greater loop gain and higher feedback so their distortion is lower, though of course part of that is design choice -- lower feedback seems to be a marketing point these days. The high constant bias of the output stages constrains the gain and feedback you can provide, though this is a more practical than theoretical issue. Given the high peak-to-average power ratio typical of music and movies the extra power can be very helpful compared to clipping a lower-powered class A amplifier.

One thing you really have to watch is what happens when the amp clips. Class A clipping means the output runs out of bias current, perhaps dumping to the load, and overload recovery can be "interesting" as the feedback loop (what there is) opens up and it can take some time to "pull" the devices back into linear operation due to all the charge storage in the output stage devices. Class AB can have similar issues, but as only half the waveform is immediately impacted, tends to recover better. Or there may be other reasons; not something I have thought about much, I have just observed a trend favoring AB for better overload recovery. But this is my experience and I have not tried to do a detailed theoretical analysis, nor is it something I have spent a lot of time researching on paper or in the lab.

So aside from the price, heat, and so forth, performance-wise the trades for class A are higher noise, greater heat leading to shorter lifetime and worse reliability, higher distortion, and lower power. Where class A is typically marketed is at low power levels because they have no crossover distortion, although push-pull designs still exhibit some distortion at low levels because of gm-modulation of the devices as load current switches from one set ("side") of devices to another. However, a class AB design runs in class A at low levels so also has no low-level crossover distortion, so this tends to be a wash.

HTH - Don

 

[Chrise36]

Without using an active crossover there is little or no advantage to adding extra amps. Active crossovers in general have much higher performance than passive crossovers so not sure what you mean by "measure poorly".

There is a big advantage the way i see it because:
a. I can control the level of the tweeter which i find a bit hot without using dsp which most of the time i dont like
b. I avoid inserting an active crossover for the reasons mentioned in my previous post
c. No worry about dac interfering with my class d amp since i will be using a class ab amp
d. My current amp (ncore nc252mp) will have to drive the mid and the bass driver only

 

[Ken Tajalli]

Not really qualified, as audio amplifier design is not my day job. But @restorer-john provided excellent examples in his post. Here are my thoughts, though hopefully some of our more learned audio designers will chime in to correct my mistakes:

Thank you for replying.

The class A designs I have seen have much higher (often an order of magnitude or more) noise and distortion than comparable or higher-powered class AB designs. How much of this is design choice (e.g. lower feedback) and how much is a limitation of the operation (higher bias means more noise) I could not say.

There is no technical reason for this assertion, perhaps you have been unlucky or tested crapy amps. In case of "Single Ended" amps you may have a point (actually a lot of points) but a like for like, Differential class A amp vs Differential class AB of similar manufacture, class A wins, it will have lower noise and lower distortion, I claim it is inherently superior, but bad design and or manufacture is another story.
Class A or class AB reffers to the output stage only, so let's keep that in mind.

Class AB offers much more power and the potential for even larger dynamic headroom for the same size, weight and power (SWaP) as a class A design which has much lower power output, greater heat, and typically (IME) higher noise and distortion. The best a class A output stage can do is about 50% efficiency (for push-pull, ~27% for single-ended) so at least half the power is wasted heat. The heat shortens the lifetime of the components inside and increases thermal (Johnson) noise. The high bias current increase current (shot) noise. Additional devices are needed to support the constant current in class A designs, which means they add their own noise (from internal resistances and such) and reduce bandwidth (from all the extra parasitics, though of course you can beef up the driver stages to compensate). That also reduces slew rate that can be achieved and potentially increases TIM (transient intermodulation distortion). Class AB amplifiers typically have greater loop gain and higher feedback so their distortion is lower, though of course part of that is design choice -- lower feedback seems to be a marketing point these days. The high constant bias of the output stages constrains the gain and feedback you can provide, though this is a more practical than theoretical issue. Given the high peak-to-average power ratio typical of music and movies the extra power can be very helpful compared to clipping a lower-powered class A amplifier.

No argument that class AB is more efficient (more power less heat, less electricity). Class A is also heavier on the PSU and components life, agreed But the issue in question is fiedility only.
Constant current in class A designs?? perhaps you can clarify.
A diffrential class A power section has noise rejection same or better than class AB. If designed correctly, TIM of a class A amp should match or better class AB. Class AB requires HIGHER global feedback than class A to control any residue crossover distortions, so gain must be worse (but who cares) output stages are mostly unity gain. One should not compare a 10W class A amp to 50W class AB, let's keep it like-for-like. Not everybody requires tons of power. My office amp is rated at 6W, and it is plenty for my use.

One thing you really have to watch is what happens when the amp clips. Class A clipping means the output runs out of bias current, perhaps dumping to the load, and overload recovery can be "interesting" as the feedback loop (what there is) opens up and it can take some time to "pull" the devices back into linear operation due to all the charge storage in the output stage devices. Class AB can have similar issues, but as only half the waveform is immediately impacted, tends to recover better. Or there may be other reasons; not something I have thought about much, I have just observed a trend favoring AB for better overload recovery. But this is my experience and I have not tried to do a detailed theoretical analysis, nor is it something I have spent a lot of time researching on paper or in the lab.

I don't believe that is the case, perhaps someone can shine a better light on this. Properly designed class A amp should recover no worse than class AB.

So aside from the price, heat, and so forth, performance-wise the trades for class A are higher noise, greater heat leading to shorter lifetime and worse reliability, higher distortion, and lower power. Where class A is typically marketed is at low power levels because they have no crossover distortion, although push-pull designs still exhibit some distortion at low levels because of gm-modulation of the devices as load current switches from one set ("side") of devices to another. However, a class AB design runs in class A at low levels so also has no low-level crossover distortion, so this tends to be a wash.

HTH - Don

Again, class A being less efficient is agreed, but I dispute higher noise and distortion, unless you are talking single ended output stage.
We should keep in mind that class AB is basically class A for small signal, and class B for higher power. It is a technique/trick to get more power out of a device, it was not conjured up to cure any other shortcomings.

If one needs higher powers, class A is not convinient to maintain, but that was not the question.
A member with a horn loudspeaker needs only a few watts, of super clean signal.

 

[manisandher]

I challenge you to find a solid state amplifier (or tube amplifier) which can pull that off with such efficient horns.

My CD horn tweeters are also 107dB/W@1m. I use class A/B Neurochrome 286 amps in a 4-way fully active system - all amps connected directly to drivers. I can barely hear any hiss with my ear up against the horn of the tweeter.

Mani.