• Welcome to ASR. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Class A vs AB vs D amplifiers

There was a tube AVR at the very beginning, I don't remember the brand.
 
rank said:
There was a tube AVR at the very beginning, I don't remember the brand.
Click to expand...
The very beginning of... what? ;)
I mean, harmon/kardon made the first widely-marketed & purpose-built integrated power/pre-amplifier and tuner. Monaural, of course (or "1.0" in AVR-speak). Most preamps in those days had a "TV" input -- so AVRs they were! :cool:

1728047049532.jpeg

1728047088642.jpeg

(random internet photos)

h/k also claim to have marketed the first stereo receiver (the MA-230) for all of your "2.0 receiver" needs. ;)

1728047267189.png

(also a rando internet image)

... or are you talkin' quadraphonic? ;)
 
A tubey Audio Video Receiver. Beginning of Home Cinema.
 
boomtheroom said:
Hi all
Iv always been under the impression that class A AB amplifiers sounded better than class D , however less powerful at same cost etc..
Now days with modules like the purifi , some say the class D exceeds the latter classes... in what respect are we talking, and how does this relate to
sound , detail , low background noise , channel separation, and imaging...
Ive heard the terminology Grainy withe older class d amps.. is this now changed?
thankss!
Click to expand...
Try class G and H
 
kemmler3D said:
Long story short, yes. So-so class-D designs have load dependence and don't always have flat frequency response. So that could easily lead someone to call it "grainy" if the treble was going all over the place in an old, cheap design.


Modern high-end Class D amps have as good or better noise, distortion, and power than most of the Class A or A/B amps out there, and especially when you compare by price.


Simply put, today's good Class D amps are just plain good, in all of those aspects.

To put it another way - many of the Class Ds you can get today (a fun example is the Fosi V3 monoblock) have ideal frequency response in the audible band, inaudible noise, and inaudible distortion. Once you check those boxes, there is no other way for the amp to mess up your sound.

So, yes they've done it, they've made Class D as good as anything else.
Click to expand...
The Fosi V3 uses the TPA3255 class d chip from TI, it has 85 uV of output noise voltage tested at their lab with their specific conditions. 85 uV of noise is very audible if you use a high sensitivity tweeter and listen very close to the tweeter. I've tested it with a ~98dB 4 ohm tweeter and it's very evident and i can hear the hissing noise floor 10 cm away from the tweeter even wiith no volume and no input signal..

Then I swapped the amp with a cheap TDA2003 class AB amp and it's virtually no noise, it's literally dead silent. Therefore we conclude that class AB is by far lower noise than (or even with) high spec'd class d amplifier. Harmonic distortion wise, the TPA3255 amp has lower distortion than the TDA2003 though, but don't mind, it's just noise comparison..

I don't know how purifi's and hypex' class d amps perform in noise since i don't have any of those, it's too expensive for me. The measured noise of all their products averaged at 15 uV of noise voltage and I think it's much better than the TPA based amps. But again, class AB is lower noise, the Topping LA90 class AB has noise voltage of <2 uV.... I don't have them too but, if we make calculations, It's dead silent even with tweeters 100dB of sensitivity and listening close-to-the touch..
 
Last edited:
proplayer69 said:
The Fosi V3 uses the TPA3255 class d chip from TI, it has 85 uV of output noise voltage tested at their lab with their specific conditions. 85 uV of noise is very audible if you use a high sensitivity tweeter and listen very close to the tweeter. I've tested it with a ~98dB 4 ohm tweeter and it's very evident and i can hear the hissing noise floor 10 cm away from the tweeter even wiith no volume and no input signal..

Then I swapped the amp with a cheap TDA2003 class AB amp and it's virtually no noise, it's literally dead silent. Therefore we conclude that class AB is by far lower noise than (or even with) high spec'd class d amplifier. Harmonic distortion wise, the TPA3255 amp has lower distortion than the TDA2003 though, but don't mind, it's just noise comparison..

I don't know how purifi's and hypex' class d amps perform in noise since i don't have any of those, it's too expensive for me. The measured noise of all their products averaged at 15 uV of noise voltage and I think it's much better than the TPA based amps. But again, class AB is lower noise, the Topping LA90 class AB has noise voltage of <2 uV.... I don't have them too but, if we make calculations, It's dead silent even with tweeters 100dB of sensitivity and listening close-to-the touch..
Click to expand...
It's a good point here, if you have a need for unusually low noise you need to pay closer attention to those values.
 
proplayer69 said:
The Fosi V3 uses the TPA3255 class d chip from TI, it has 85 uV of output noise voltage tested at their lab with their specific conditions. 85 uV of noise is very audible if you use a high sensitivity tweeter and listen very close to the tweeter. I've tested it with a ~98dB 4 ohm tweeter and it's very evident and i can hear the hissing noise floor 10 cm away from the tweeter even wiith no volume and no input signal..

Then I swapped the amp with a cheap TDA2003 class AB amp and it's virtually no noise, it's literally dead silent. Therefore we conclude that class AB is by far lower noise than (or even with) high spec'd class d amplifier. Harmonic distortion wise, the TPA3255 amp has lower distortion than the TDA2003 though, but don't mind, it's just noise comparison..

I don't know how purifi's and hypex' class d amps perform in noise since i don't have any of those, it's too expensive for me. The measured noise of all their products averaged at 15 uV of noise voltage and I think it's much better than the TPA based amps. But again, class AB is lower noise, the Topping LA90 class AB has noise voltage of <2 uV.... I don't have them too but, if we make calculations, It's dead silent even with tweeters 100dB of sensitivity and listening close-to-the touch..
Click to expand...
The new Genelec 8380 announced today, which has a super efficient horn loaded mid/ tweeter, and one of its selling points is "<0 dB self noise" uses a Class AB amp on the HF driver, probably because it is quieter.
 
proplayer69 said:
The Fosi V3 uses the TPA3255 class d chip from TI, it has 85 uV of output noise voltage tested at their lab with their specific conditions. 85 uV of noise is very audible if you use a high sensitivity tweeter and listen very close to the tweeter. I've tested it with a ~98dB 4 ohm tweeter and it's very evident and i can hear the hissing noise floor 10 cm away from the tweeter even wiith no volume and no input signal..

Then I swapped the amp with a cheap TDA2003 class AB amp and it's virtually no noise, it's literally dead silent. Therefore we conclude that class AB is by far lower noise than (or even with) high spec'd class d amplifier. Harmonic distortion wise, the TPA3255 amp has lower distortion than the TDA2003 though, but don't mind, it's just noise comparison..

I don't know how purifi's and hypex' class d amps perform in noise since i don't have any of those, it's too expensive for me. The measured noise of all their products averaged at 15 uV of noise voltage and I think it's much better than the TPA based amps. But again, class AB is lower noise, the Topping LA90 class AB has noise voltage of <2 uV.... I don't have them too but, if we make calculations, It's dead silent even with tweeters 100dB of sensitivity and listening close-to-the touch..
Click to expand...
This wideband noise is probably before the switching stage and hence does not depend on A or D etc. Topping LA90 is SOTA hence <2uV

1759273124851.png

1759273139967.png


Well yea ncx500 is lower SNR by quite a bit as predictable by their respective SINAD performance (noting that ncx500 has slightly higher gain than LA90 here, in fact it's around 2dB higher, which when you add to the 7dB difference in SINAD results in 9dB difference in noise performance, surprise surprise *not really*), but lets look at something with similar SINAD and is not a class D:
1759273358857.png


ncx500 actually has better noise than AHB2.

All numbers are at 5W output

Just to round things up, below are two amplifiers with similar SINAD performance:
1759273578846.png


1759273671945.png


So no, noise level is not a function of class AB or D. But it has good correlation with SINAD. Good amps have less noise, cheaper amps have more noise, that's all. Of course just like opamps you can have designs more tuned towards low noise versus some other priority.

One last thing to note, the datasheet numbers are only best-case scenarios. I'm sure some of us have encountered this, when a speaker's power adapter spoilt and we replaced it with another, it has a buzz that is not there with the original power adapter. Power supply (and the related common-mode a.k.a. ground noise) limit the final product. Which is why 3rd party measurements of end-user products are more useful than first-party datasheet of the individual IC. Although such measurements still do not cover common-mode noise simply because we have not measurement standard or criteria for it.
 
Last edited:
boomtheroom said:
Hi all
Iv always been under the impression that class A AB amplifiers sounded better than class D , however less powerful at same cost etc..
Now days with modules like the purifi , some say the class D exceeds the latter classes... in what respect are we talking, and how does this relate to
sound , detail , low background noise , channel separation, and imaging...
Ive heard the terminology Grainy withe older class d amps.. is this now changed?
thankss!
Click to expand...
My experience has been:

1. Class A: My Musical Fidelity A1 Class A integrated amplifier is rated at 25 wpc, but had no problem driving my Klipsch KG4s. Sound wise it was just okay. Needed to be left on for about 30 minutes just to get hot (not warm) enough to sound right. Ultimately, I got tired of all the waiting for it to get hot and now it’s for sale.

2. Class D: My NAD C 3050 is rated at 100 wpc and drives my Klipsch KG4s very well, with solid bass and wide sound stage. It sounds surprisingly good and balanced across the audio spectrum. This is the first Class D amp I’ve owned. I’m mostly satisfied with it (volume controls are quirky).

3. Class A/B: My Kenwood KR 10000 II is rated at 80 wpc and doesn’t drive my KG4’s so much as rolls thunder into them. This old school receiver (50 years old) is an absoulte beast. I can’t say it’s balanced across the audio spectrum, but it has the most satisfying sound of the three amps and currently occupies the primary amplifier position.

I’ve had other Class A/B amps (Vincent SV 500, Black Ice Audio 3502S) that were pretty good too. Between those amps, I’d choose my Class D NAD C 3050 over them.

It really comes down to what works best for you and provides the right synergy with your audio gear.
 
What is so surprising that an amplifier with 1/50 the rated output power (6 W @ 4 Ω, 10% distortion) of the TI TPA3255 has lower idle noise?

If one is using the TPA3255 with a high sensitivity speaker or driver, there should be plenty of power to spare. Just pad it down with a L-pad and you'll get rid of the idle noise issue.
1759275519406.png
 
wwenze said:
This wideband noise is probably before the switching stage and hence does not depend on A or D etc. Topping LA90 is SOTA hence <2uV

View attachment 479665
View attachment 479666

Well yea ncx500 is lower SNR by quite a bit as predictable by their respective SINAD performance (noting that ncx500 has slightly higher gain than LA90 here, in fact it's around 2dB higher, which when you add to the 7dB difference in SINAD results in 9dB difference in noise performance, surprise surprise *not really*), but lets look at something with similar SINAD and is not a class D:
View attachment 479669

ncx500 actually has better noise than AHB2.

All numbers are at 5W output

Just to round things up, below are two amplifiers with similar SINAD performance:
View attachment 479673

View attachment 479675

So no, noise level is not a function of class AB or D. But it has good correlation with SINAD. Good amps have less noise, cheaper amps have more noise, that's all. Of course just like opamps you can have designs more tuned towards low noise versus some other priority.

One last thing to note, the datasheet numbers are only best-case scenarios. I'm sure some of us have encountered this, when a speaker's power adapter spoilt and we replaced it with another, it has a buzz that is not there with the original power adapter. Power supply (and the related common-mode a.k.a. ground noise) limit the final product. Which is why 3rd party measurements of end-user products are more useful than first-party datasheet of the individual IC. Although such measurements still do not cover common-mode noise simply because we have not measurement standard or criteria for it.
Click to expand...
Thanks for your explanation.. So your studies/ explanations showed that Class D's is on par with Class AB's and performs almost equal nowadays.

Now I know that the wideband audio frequency noise (human hearing) is not directly related to class D switching because it's greater than the factor of 10 below class D switching.

But how do you think about the theory of class D to have higher noise than linear class AB because of it's switching nature and carrier based signal amplification? Is this true?
 
Last edited:
wwenze said:
This wideband noise is probably before the switching stage and hence does not depend on A or D etc. Topping LA90 is SOTA hence <2uV

View attachment 479665
View attachment 479666

Well yea ncx500 is lower SNR by quite a bit as predictable by their respective SINAD performance (noting that ncx500 has slightly higher gain than LA90 here, in fact it's around 2dB higher, which when you add to the 7dB difference in SINAD results in 9dB difference in noise performance, surprise surprise *not really*), but lets look at something with similar SINAD and is not a class D:
View attachment 479669

ncx500 actually has better noise than AHB2.

All numbers are at 5W output

Just to round things up, below are two amplifiers with similar SINAD performance:
View attachment 479673

View attachment 479675

So no, noise level is not a function of class AB or D. But it has good correlation with SINAD. Good amps have less noise, cheaper amps have more noise, that's all. Of course just like opamps you can have designs more tuned towards low noise versus some other priority.

One last thing to note, the datasheet numbers are only best-case scenarios. I'm sure some of us have encountered this, when a speaker's power adapter spoilt and we replaced it with another, it has a buzz that is not there with the original power adapter. Power supply (and the related common-mode a.k.a. ground noise) limit the final product. Which is why 3rd party measurements of end-user products are more useful than first-party datasheet of the individual IC. Although such measurements still do not cover common-mode noise simply because we have not measurement standard or criteria for it.
Click to expand...
I always see that class AB amps have lower noise voltage than class D design s.

Do class D switching always contribute to the noise of 20-20kHz even if it's well attenuated by LPF? I.e., the high frequency carrier modulates this low frequency audio noise (20-20kHz) and add it to the system noise floor?
 
Do class D switching always contribute to the noise of 20-20kHz even if it's well attenuated by LPF? I.e., the high frequency carrier modulates this low frequency audio noise (20-20kHz) and add it to the system noise floor?
Click to expand...
ALL active analog electronics generate SOME noise.* You'd have to check the measurements and compare, regardless of class... A particular class D amplifier may have less noise than a particular Class-A amplifier or vice-versa. You do need comparable measurements because there's more than one way to measure noise. The manufacturer's specs aren't that helpful.

Of course if the noise isn't audible it's not an issue... You don't necessarily need the quietest amp. Noise audibility depends on a lot of factors, including how close you are to the speakers, the sensitivity of the speakers, and other ambient noise in the room.

Class A isn't the best, it was simply the 1st. I don't know when class A/B was invented but when I was a kid in the 60's almost all radios & TVs had class-A amplifiers because they could be built with one less tube or transistor and it was cheaper than class A/B .

Class B isn't good for audio because there is crossover distortion. But class-B amplifiers might have been used when power-output was more important than distortion... I don't know.

Class A/B (and class-B) requires one more "device" (tube, transistor, or MOSFET) but it's more energy efficient than class-A so it can run cooler and use cheaper devices and smaller heatsinks so the cost of the additional parts is insignificant and the overall cost-per-watt is lower than class-A. Nowadays you can get all of the electronics in one inexpensive chip as long as you don't need more than several Watts. The cost of one more device inside the chip is "nothing" so it makes no sense to make a class-A chip.

Class-D is a lot more complicated and it requires a LOT more electronics. But since all of that complex circuitry is usually all in a chip, Class-D is the most economical per-Watt.




* Resistors (and other passive electronics) generate noise too (thermal noise) but it's inaudible unless amplified.
 
proplayer69 said:
But how do you think about the theory of class D to have higher noise than linear class AB because of it's switching nature and carrier based signal amplification? Is this true?
Click to expand...
No, there's no rule or guarantee that a well designed Class AB amplifier costing $600 is audibly quieter than a well designed Class D amplifier costing $600. Where the gains are identical.

However, there are probably more badly designed, very cheap, chip-based Class D amplifiers with lots of gain which are noisy.

Very high quality discrete Class D (such as Purifi and NCore) are extremely low noise and have switching artefacts many times higher than bats can hear, let alone humans.
 
You must log in or register to reply here.

Similar threads

Hart
Amplifier power: Class A, AB vs D
2 3 4
Replies
62
Views
5K
Sal1950
Sal1950
danishchelsea
Why use Class AB amplifier modules in Active speakers when Class D already delivers?
Replies
9
Views
735
-Matt-
M
J
Class D: Hypex/ICE/Purifi vs. Pro Amps
2
Replies
26
Views
2K
jwr159
J
A
Class D Amplifiers Explained: Complete Beginner’s Guide
Replies
5
Views
1K
audioman91
A
L
Class A/AB amplifier recommendations
2 3 4
Replies
60
Views
7K
Chrispy
Chrispy
Back
Top Bottom