AVS Forum banner

Class A/B vs Class H

26K views 53 replies 20 participants last post by  droidlike 
#1 ·
hi, im deciding to either pull the trigger on the new emotiva xpa5 gen3 (class h) or save money the money and wait until I find a good class a, class a/b amplifier. Is there really a difference in sound? I play the theater loud!
 
#2 ·
Is there really a difference in sound?
Only if you're peeking! ;)

Mate, to be brutally frank you really should save money on amps FULL STOP . From the look of the gear list your signature, you've already spent way, way to much on electronics/amps in comparison to your speakers.

What do you hope to achieve with this new amp? I count seven speakers and seven amp channels in your sig. Given you play loud (and pending knowing what your room volume is), the obvious limitation in your current system is the single 12" sub. You'll get a transformational improvement in your HT audio if you concentrate your spending and effort on upgrading to at least two high performance subs and getting them properly integrated. A few examples are subs from JTR Speakers, Seaton Sound or Funk Audio. Needlessly(?) churning through amps in search of improved audio is throwing good money away and will not be yielding any actual improvement.

Sorry if I sound harsh, but I'm just challenging you have a rethink about your priorities. :)
 
#3 ·
Oh this wasnt for my current set up... its for a new one at another place...! I have a tv in the terrace and ceiling speakers and planned to set up a 5.1 there but I do see your point except it isnt for my current setup in my signature! Cheers and thanks for your input...
 
#4 ·
Class-A means that there is only one set of transistors, and that they are turned on 100% all the time.
This is of course, highly-inefficient. ~25% when playing music and 0% efficient when no music is being played.
But since it doesn't have to transition to another set of transistors, it is claimed to be the best sounding. (or at least this used to be true, back in the 80's... ;))

Class-AB is like A but with two sets, one doing the positive and the other the negative parts of the wave.
This gives double the efficiency and double the cooling. ~50% when playing music and 50% when not.

Class-H is like AB but with a variable power-supply/bias voltage, so that the PS and transistors are only working as hard as necessary to produce a clean signal.
This produces even higher power ratings for a given weight while remaining cool, reducing the need to engage fans, and ever so slightly more efficient. ~50-60% when playing and 75% when not.

Class-D is entirely different. Instead of leaving the transistors on fully, it only turns them on for as long as necessary and as hard as necessary to produce the signal.
While this is really efficient 60-98% when playing music and 90-99% when not... it also adds lots of HF distortion that needs to be filtered out, the HF extension is limited by the switching frequency and the HF filters.
For this reason they were only used in subwoofers for years and years...

The Gen-3 is a hybrid.
It uses a switching power supply to produce DC (which doesn't have the HF problem) and then a conventional Class-H design to power the speakers.
This avoids needing a 100lb torodial transformer and Class-AB DC power supply, which saves on shipping and production costs, decreases weight and increases efficiency, but rarely does a company pass on those savings to the customer. That said, Emo does a better job at keeping prices low than almost any other company.
The only downside is making the SMPS strong enough to handle the needed RMS load, I haven't seen the benchtests but I would imagine that Emo has done their homework in this regard.

There are many PC's with 1-2kW SMPS's these days for only ~$200-400 so it's not unheard of... CPU's and GPU's have a high RMS load these days.
Quad SLI is pushing the limits of a 15A 1800watt RMS breaker in most cases...
 
#12 ·
All of the classes are 0% efficient when not playing music. Nothing out, some amount of electrical power in. 0 watts output / x>0 watts input = 0%

What Emotiva calls Class H switches between two different power supply voltages.

Unfortunately the audio industry is inconsistent in what it calls Class G and Class H. One class switches between fixed power supply voltages and one class has a tracking power supply, but you will see Class G and Class H used for each of these scenarios.
 
#5 ·
Very good summary from @BassThatHz. I have a few minor changes based upon my experience and understanding.

Class A can use two "sets" of devices in push-pull (for the positive and negative halves of the signal) and that is the way most audio amplifiers are designed. That gets the theoretical efficiency up from ~26% to 50% or so for an ideal output stage.

Class H uses a variable supply rail as @BassThatHz stated so it tracks the input signal. That way less power is wasted in the output stage and efficiency is improved. The actual amplifying circuit could be most anything, but is typically class A or AB as they benefit the most. Up to full power, a class-H design rated for X Watts will require less energy than a pure class AB design even though the intrinsic amplifier is no more efficient; you get the improvement by reducing the power rails. A 50% efficient 100 W amp operating at 10 W with reduced rails is better than a 100 W amp operating at 10 W with the same voltage rails needed to produce 100 W.

Class D is actually most efficient at full output, at least for the designs I have seen. The output devices are always fully switched; how often and for how long depends upon the signal. Class D loses efficiency for small signals because low-level bias, driving, and control circuits don't really scale so you have a base minimum power that is always used no matter the signal level. As the signal decreases, the standing quiescent power becomes a larger fraction of the total power, and efficiency drops.

FWIWFM - Don
 
  • Like
Reactions: Madmax67
#6 ·
Thanks everyone, I have an EPISODE 12 channel amplifier, 70 watts per channel.. I connected my mx122 to the the front center and left... from what I can read on the specs of the episode its a class d amplifier (i have it set for yard work, rocks speakers, outdoors, etc...)...

When I play music in stereo the music sounds FLAT, like tinny... very different from the way it sounds on the monoblock emotiva amplifier or the mc7205 amplifier (120watts per channel)... my question comes down to if all amplifirs eventually sound the same? why would someone pay the 9k from an mc207 when you can pay for the emotiva 7 channel emotiva amp running on class h? would there be a difference in sound between the emotiva xpa7 gen3 and the mc207?
 
#10 ·
Not all amplifiers sound the same. More of them do than don't, however, IME/IMO.

This does not sound like an amplifier issue but I am not sure exactly what you are describing. What changes in your system when you run in stereo mode? Sounds like an EQ problem or maybe you prefer the surround sound (many do, including me, when done well).
 
#7 ·
It was my understanding that part of what makes a class D amp tick is its ability to change the analog waveform into a PCM signal. This "full-on/full-off" signal is what gets amplified, which is why it's so efficient: there is no transition or intermediate state to worry with. The amplified PCM signal is decoded back to the analog wave form, and Bob's your uncle, as they say.

I'm not completely sure I'm contradicting anything anyone else is said yet. Electronics is hard.
 
#8 ·
any discernible difference in sound at any playing level between class a, class a/b, class h and class d?
 
#9 ·
Depends on who you ask. When driven within their limits, amps aren't supposed to add or subtract anything to (or from) the sound. Some do anyway, of course, but it's debatable whether or not any of it can be heard by human hearing. A/B/X testing seems to confirm that none of them do audible things to the signal.

There are some who will take extreme exception to this opinion. That's fine. Do some research, draw your own conclusions.

Personally I went with class D and always will. It may not do anything to the signal other than amplify it, but it does so orders of magnitude more efficiently than other designs. Small size, big power, doesn't heat the house, doesn't up the electric bill. What's not to love?

Again, others will disagree. Enjoy yours.
 
#13 · (Edited)
Even two orders of magnitude greater efficiency would make Class D 100 times more efficient that other amplifier classes. Even if another amplifier class is only 2% efficient at some level of output then the Class D would be 100 times that or 200%. This would make it a net generator of power thus greatly reducing your power bill, and ultimately solving global warming if enough units are deployed.
 
#11 ·
Well it sounds tinny lol... i cant find a better way to describe it.. its like ultra bright... when playing the same on my mcintosh it sounds deeper, theres depth to it...
 
#17 ·
Well, I need to backtrack a bit... I missed in his post the bit about the Episode 12 amplifier. It does not appear to be made any more, but what I could find online indicates it is (was) a 50 W/ch amp into 8 ohms, 70 W/ch into 4 ohms. And the OP likes it loud.

What speakers are being used and how far away is the listening position? It may be the little amp is simply running out of steam. However, unless the speakers are a tough load, I would not expect problems at low volume. Listening outside, if that is the normal use, I would expect a significant amount of power and the Episode might not be up to it.
 
#19 · (Edited)
It would have to be benchtested, but I would imagine the McIntosh would have a slightly lower noisefloor, cleaner transient response and better power reserves. But you can't say for sure without testing both under lab conditions.

From 20hz to 20khz and from 0watts to just below clipping, I would imagine that almost all amplifiers are within 1% of each other. If we are talking modern-day SS amps.

Some amps have fans, some have more power, some have lower noisefloors, and some have really wide frequency responses at all loads.

When people say "the speaker really opens up" they are probably just hearing an increase in peak-power / dynamics or a lack of previous clipping... and maybe less hiss.

THX is a 20db peak vs the average, which requires a 100x increase in peak-power.
Which could be hundreds or thousands of watts required.

If your average is 5watts to reach 85db at the seats, then you'll need 500watts to reach 105db peaks.

A 90db/w speaker @ 10ft needs 285watts to reach 105db and 3watts to reach 85db.
If you only have a 100watt AVR and you aren't using >95db/w speakers, you might be clipping the amp HARD at THX levels during action scenes, which will sound nasty.

With compressed Top 40 music (of almost all genres) it is even worse, since the average vs peak values are closer to 6-10db, so the amp is working 4-10x harder on average; depending on how loud you crank it of course!

In a room, the SPL will drop about 3db per doubling of distance and you get a free 10db from the 6 walls, but outside in a backyard, it drops off at 6db per doubling of distance and you only get maybe 3-6db of add SPL from ground reflections (if you are lucky).
That's why speakers sound weaker when placed outside and are more likely to be overdriven by about 20x too far...
 
#21 ·
Please note:

THX is 85dB average which is 88dB peak. (Sine waves) Maximum is 105dB peak (102dB average or RMS or whatever) which is +17dB or 50 times more power.

If the average power at 85dB is 5 watts then an amplifier rated at 250 RMS (average) watts (500 peak watts, that is +3dB) will be required for 105dB peak. Amplifiers of this rating are readily available, especially if the power is to be supplied at closer to 4 ohms than 8 ohms.

A speaker with a sensitivity of 90dB output at 1 watt input will take less than 1 watt to produce 85dB, not 3 watts. Perhaps the intended value was 80dB at 1 watt input which then fits better with the other values. Note that the 285 watt value is peak watts or about 142 average or RMS watts, that is, an amplifier rated to supply at least 142 watts RMS is required.

Just for numbers, 100 watts from an AVR driving a 95dB per watt speaker will supply 115dB average, 118dB peak. 100 watts from an AVR from an 90dB per watt speaker will supply 110dB average, 113dB peak. The output at the listening position depend on the factors you mention. Even 100 watts provides a lot of output, especially when many people report listening at average levels of 75dB to 80dB.
 
#23 ·
distortion, noise, headroom all have to play a role in SQ to a certain extent at certain output levels. but I hear all the time it all sounds the same, dont really agree with that, but maybe it does.
 
#25 ·
I remember when that wasn't true, but maybe things have changed a lot in the last 25-30 years.

I'll still stick class A amps for musical instruments though.

Cross over distortion on class B was quite noticeable, and A/B designs often had some odd effects in certain cases.

Classes G and H weren't even used.

Class C amps are useless for audio.

Class D requires a good output filter to get rid if switching of noise. For large spaces I can see a point, bit seems overly complex for home theater power levels.
 
#30 ·
Class D requires a good output filter to get rid if switching of noise. For large spaces I can see a point, bit seems overly complex for home theater power levels.
My Pioneer Elite SC-Series AVR disputes this...
 
#27 ·
Think about it this way. What is the purpose of an amplifier?

In general, it is to make a small signal more powerful. For audio, you want an amp that produces a more powerful signal that is a replica of the input signal, but more powerful.

So for that purpose, there are several designs that attempt to do this like class A, class AB, and class D. In general, if they do a good job of reproducing the input signal, then you really cannot tell amplifiers apart - if the amplifier is good, then the output signal has a high fidelity to the input signal, and they will be indistinguishable.

Now, there are some kinds of amplifiers that do not reproduce the incoming signal with good fidelity - they add distortion. Some are inherent in their design, like class c, other are just because their active components just inherently do it (e.g., tube amps). Now, some people like the sound of the distortion, but that's what it is - distortion, so the output waveform is not a replica of the input. You can reproduce this if you add a effects unit to your high fidelity amplifier to create a distorted waveform.

FYI, class C amplifiers, though having high distortion, are very efficient, and are used where the wave shape is less important. This is true for phase-modulated signals like FM radio. This gives a very high efficiency amplification that has little effect to the signal as FM does not care about the wave shape - just the phase.
 
#40 ·
Snark aside, my SC-61 AVR pumps out 125WPC @ 8 Ohms - ALL CHANNELS DRIVEN. Meaning, that at any time, I get 125W going to all seven of my speakers. Clearly, Pioneer is doing something right with their Elite SC-class AVRs.
 
#49 · (Edited)
about audio perspective on
class d amp versus class a or a/b amp

in the following discussion


https://www.avforums.com/threads/class-d-improving-in-leaps-and-bounds.2034982/


you can read at the very end
that the krell 600 monoblocks were replaced by the new class d amps


=============================
nord One UP 500 DM ST Dual Mono Stereo Amplifier Review
nord One UP 500 DM ST Dual Mono Stereo Amplifier Review
By George Sallit on 26/06/2016

.
.
.
It has now become my summer amplifier as the Audionotes put out too much heat in the summer. Oh and the Krell 600 monoblocks have gone.

If you want to be a hero well just follow me.


no wonder why these amps are used in the NAD D 3020

Hypex Electronics BV - News

NAD D 3020 Integrated Amplifier Review


===============================


and
even Ati-Amp is using the same class D amp module above
in the Theta products
with more than 800 wpc@4 ohm


http://www.thetadigital.com/press_release_new_dreadnaught_d_introduced.shtml
 
#48 ·
G or H can be applied to A or AB fundamental amplifiers and usually are as those are less efficient. Class D is a different beast and switching/tracking the rails is not worth the effort.

So the real question is A vs. AB vs. D for all practical purposes. For any decent SS design A and AB are going to sound the same and may be different from D. Class D is a newer technology and more sensitive to implementation so how much difference you hear is going to depend on the amp and the speakers (and how loudly you like to play them, natch). Class D is generally more sensitive to load impedances but that is not always true, usually matters only for extremely low and/or capacitive loads, and may or may not be audible. Which is what you can say about A or AB amplifiers as well.

So what Noah said.
 
#53 ·
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top