I've been trying in vain to get some questions answered about amplifiers and such. Then, I realized that my questions technically fall under sound science, much more so than they fall under DIY.

So lets get this started. Also, if you refer to some external research or source, please post the link to it as others reading may find other information from there useful.

1) Class A, B, C, D, and AB. I understand that a class A amplifier will "look at" the entire waveform over 360 degrees and amplify every single voltage put through it with some gain. I understand that class AB will look at 180 (?) degrees of the waveform and amplify that. And B, C, and D are all other amounts of waveform amplification. My question regarding this is: How in the heck does that make any sense? How does a class AB amp look at only half of the waveform and amplify it, yet we still get a listenable sound out of it? It looks to me like class B does an almost half-wave rectifier amplification but if that is effectively destroying half of the waveform, it doesnt make sense to me how we can even stand to listen to it. I also understand that class AB is basically class B with a voltage bias applied to it (or something).

Any layman's terms definitions/descriptions of amplifier class and their function is welcome for this answer.

2) Lets say I design an amplifier with the idea that I want 2 watts per channel into a 600 ohm load. If I plug in a 32 ohm load instead, does the power of the amplifier change? Or does it simply change the current/voltage relationship to accomodate the lesser impedence? Does sensitivity have anything to do with it? I thought that a 2 watt amplifier will always be a 2 watt amplifier assuming it can handle switching from high voltave/low current to a high current/low voltage power output. I also thought that depending on the amount of power going into the load, the headphone sensitivity is only a standardized measurement of how many decibals you can expect given a standard voltage/current input.

This second questions was kind of 2 parts.

To clarify, I am much more interested in class A and class AB than the other amp classes. To give an idea of my technical knowledge, I have some experience with RLC circuits and the math behind them but I have no idea as to the applications of them. In my current (lol) electronics class we have covered opamps and diodes and we are working on transistors now, mostly MOSFETs. Later on we will be covering BJT's. This is the section of the class that I am extremely interested in, along with vacuum tubes (which are not covered in class). Unfortunately, my class is 99% theory and the professor is new and more of a digital signals guy and is terrible at providing real life applications examples. Great professor though, dont get me wrong.

http://en.wikipedia.org/wiki/Electronic_amplifier#Class_B_and_AB
 
 
For 2) you'd need over 34 Vrms output. Yes, with a 32 ohm load (and 0 output impedance) this would result in over 36 watts and of course a lot more current.
 
P = V * V / R

Ok, so in this case, I'd design the amplifier to output 2 watts per channel into 600 ohms. Using P=V*V/R and P=I*I*R, we get 35 volts and 60 mA. now, if I plug in a 32 ohm load, will the voltage and current remain the same but only the resistance changes? so in this case i'd have 35*35/32 and/or .06*.06*32 watts of power? I thought that headphones would draw on whatever current and voltage combination they were designed for....

Also, it seems that I was confusing around class A and class AB. I need to figure out how a transistor works and what it actually amplifies. That diagram from the wiki makes sense though.

EDIT: regarding the first scenario, rather than tailoring the amplifier to accomodate a specific ohm load, is it better to decide on a voltage and current output goal while focusing on making the output impedence as low as possible?

On an ideal amp the voltage is fixed regardless of the load. Ohm's law is I = V/R which means that the amp would need to deliver over 1 amp into 32 ohms. But I can guarantee you that this will destroy most 32 ohm headphones or your hearing or both.



edit: If you want to use 600 ohm and 32 ohm headphones interchangeably you should think about the voltage both headphones need, your source and the voltage gain you'll need. Probably it'd be a good idea to implement a gain switch.

Just to clarify, the voltage (max) will be fixed, and the current will change?

1:
Before anything else, Class-AB does not exist. It is marketing gibberish that was introduced to help consumers forget how truly terrible class-B amps used to be. Now class-B amps just kind of suck a little I might use the term by accident, I'm so sorry. 
 
Class-A means that a given transistor (includes SS and "tube" parts of all descriptions) does not switch from "on" to "off" at any time. In other words, the transistor is ALWAYS passing current.
 
Class-B means that a given transistor DOES switch from "on" to "off" at some point in time, and for less than 1/2 of any given cycle. Class-B does do a half-wave rectifier thingy to the signal BUT in all "class-AB" or class-B amps you will see complementary devices. One transistor handles the top half of the wave, the other the bottom.
 
The big problem is SWITCHING. This switch from on to off comes with a little "bump" in the output signal that is totally unrelated to the input - can you say mega-whoa-suck-ugly-distortion? yea, good, cause that's what you get without enough feedback to cover this switching noise. 
 
Class-B or class-AB amps ONLY work because they have lots of feedback. Without LOTS of feedback they would be nearly unlistenable. 
 
Said another way, enough feedback can reduce the audibility of the switching to acceptable levels, but think about that for a second. When we run class-B we are knowingly using an inherently VERY non-linear system with lots of distortion. Why not start with a system with lower distortion? AKA Class-A? In the case of headphones NOTHING needs enough power that its ever acceptable to leave class-A in a stationary/wall powered device. With any measure of care an amp can be made that will draw less power from the wall than a decent CD player (about 100W) or computer and NEVER drop out of class-A with a real-world headphone. 
 
Now that I'm done bashing class-B (for the moment, I dint think I'l ever really be done making fun of compromised junk) almost all modern amplifier designs are class-AB for a variety of reasons. The "power war" is probably the largest - you simply cant sell a 30w class-A amp to the masses when they have the option of a 45w class-AB. 45>30. The average consumer wont stop to notice that the 45w amp only runs class-A up to about 20w and rides its feedback loop the rest of the way and simple math says that 45>30. So designers got sucked into finding new ways to get more power out of the same old tired tubes. This was actually a great benefit to the field - without a real challenge many circuits we currently take for granted would never have been invented. Sadly there was as much innovation (if not more) in the field of corner-cutting as there was in actually improving things. Consumers voted with their wallets, as always, and quite a few truly better circuits that cost just a bit more got forgotten. 
 
2a:
yes no maybe. 
You are asking a VERY broad question here. The broad answer is no. Most amplifiers put different amounts of power into different loads but keep reading for some fun exceptions. Since the question is so broad there is no real way to tell what the amp will do with those headphones mentioned. 
 
On that note, there are several ways to make an amplifier that puts equal amounts of POWER into 600 &32ohm headphones. An OTL (SS or tube) amplifier that is voltage limited into 600ohms and current limited into 32 could do just fine. Or a transformer coupled amplifier with separate taps on the transformer for various headphones or speakers would also work very nicely. 
 
There are also several ways to make an amplifier that puts more power into 600ohms than 32, or more into 32ohms than 600. fun fun fun. 
 
Now that that's out there, what do you even want 2w into 600ohms for? Did you find some old 600ohm speakers or do you maybe want to run a 600-ohm cutter head? Some older McIntosh amps have 600ohm taps that will put FULL power into 600ohms... They are very popular for both of those applications. If you are looking at headphones you don't need much more than 50mW for anything but orthos. 
 
2b:
Different MFRs state headphone efficiency differently. Some state SPL@power, others state SPL@voltage. Someone has probably posted how to convert the 2 back and forth.

WOW nikongod, thanks for such an overwhelmingly helpful answer! This has given me a lot to think about. I'm going to research on the finer points of transistor math and function, especially how they amplify and how to calculate their gain and/or use them to achieve a nice gain. I suppose the next question is "where do I go from here" but I think i'm more on the right track.

If you would like to provide a layman's explanation of how a transistor works (mathematically) and how it can be used to achieve gain, I am all ears...er...eyes.

In the meantime I will be scouring youtube for tutorials and explanation videos.

Quote:

Originally Posted by nikongod /img/forum/go_quote.gif
 
The big problem is SWITCHING. This switch from on to off comes with a little "bump" in the output signal that is totally unrelated to the input - can you say mega-whoa-suck-ugly-distortion? yea, good, cause that's what you get without enough feedback to cover this switching noise. 

Click to expand...

 
Since this is the science forum, I guess you have actual measurable evidence of this "mega-whoa-suck-ugly-distortion" in well designed and implemented amplifiers that are less than 20 years old ?
 

nikongod said:

Now that I'm done bashing class-B (for the moment, I dint think I'l ever really be done making fun of compromised junk) almost all modern amplifier designs are class-AB for a variety of reasons.

Click to expand...

Typical case of audiophile FUD. :rolleyes:

Quote:

stv014 said:

 
Since this is the science forum, I guess you have actual measurable evidence of this "mega-whoa-suck-ugly-distortion" in well designed and implemented amplifiers that are less than 20 years old ?
 

Click to expand...

Class AB reduces the crossover disortion of class B so it's no longer a problem from the wiki,
 
"Here the two active elements conduct more than half of the time as a means to reduce the cross-over distortions of class-B amplifiers. In the example of the complementary emitter followers a bias network allows for more or less quiescent current thus providing an operating point somewhere between class A and class B. Sometimes a figure is added (e.g., AB_{[size=x-small]1[/size]} or AB_{[size=x-small]2[/size]}) for vacuum-tube stages where the grid voltage is always negative with respect to the cathode (class AB_{[size=x-small]1[/size]}) or may be slightly positive (hence drawing grid current, adding more distortion, but giving slightly higher output power) on signal peaks (class AB_{[size=x-small]2[/size]}). Solid-state class-AB amplifier circuits are one of the most popular amplifier topologies used today."
 
So if he is talking purely about tube amps he might be correct but modern solid state no.
 
 

 
Quote:

stv014 said:

 
Since this is the science forum, I guess you have actual measurable evidence of this "mega-whoa-suck-ugly-distortion" in well designed and implemented amplifiers that are less than 20 years old ?
 

Click to expand...

 
"well designed amplifier" is a very subjective term. 
Do you mean in amps without global feedback? I took care to note that the distortion is largely covered by the feedback loop.
 
Quote:

xnor said:

Typical case of audiophile FUD.

Click to expand...

A little bit, sure, but when it costs so little to do better for headphones why shouldn't someone? 
Considering what people (measurements guys included) spend on an inaudible zero going from 0.001%thd to 0.0001%thd knowing where to find that last zero can be very valuable. 
 
 
 
You know that you didnt actually define class-A or class-B, which was the OP's key question. 
You also didnt really answer question 2. There are tons of situations where anything can happen. 
 
Quote:

jrg1990 said:

Class AB reduces the crossover disortion of class B so it's no longer a problem from the wiki,
 
"Here the two active elements conduct more than half of the time as a means to reduce the cross-over distortions of class-B amplifiers. In the example of the complementary emitter followers a bias network allows for more or less quiescent current thus providing an operating point somewhere between class A and class B. Sometimes a figure is added (e.g., AB_{[size=x-small]1[/size]} or AB_{[size=x-small]2[/size]}) for vacuum-tube stages where the grid voltage is always negative with respect to the cathode (class AB_{[size=x-small]1[/size]}) or may be slightly positive (hence drawing grid current, adding more distortion, but giving slightly higher output power) on signal peaks (class AB_{[size=x-small]2[/size]}). Solid-state class-AB amplifier circuits are one of the most popular amplifier topologies used today."
 
So if he is talking purely about tube amps he might be correct but modern solid state no.
 
 

Click to expand...

Nope, solid state too. 
Its definately there in the few pieces of solid state gear that dont use global feedback. 

I would say that class AB doesnt sound as good as class A, but unless you dont have a portability requirement, class AB is pretty much the only way to go because class A is simply not efficient enough or small enough to be packaged into a portable rig. One question regarding this: Look at the Cmoy amp. That uses an opamp, which is a specific topography of transistors inside. Can the cmoy be classified as AB? If you have an opamp, how do you classify the topology inside of it? I'm assuming other portable audio devices dont bother using individual high power transistors and instead opt for using a simple opamp, or at least an IC that has a ton of transistor connections available so they can hook it up themselves.

Also, can someone clarify how a single transistor can act as an amplifier? I see that it can act as a current amplifier but where does that extra current come from? I also see that you can take that current increase and change it to a voltage increase as a tradeoff for lower current, and in that way you can use it as a voltage amplifier too.

nikongod said:

A little bit, sure, but when it costs so little to do better for headphones why shouldn't someone? 

Click to expand...

Because class A, low amounts of or no feedback is not (always) "better".

Considering what people (measurements guys included) spend on an inaudible zero going from 0.001%thd to 0.0001%thd knowing where to find that last zero can be very valuable.

Click to expand...

Reducing thd can be achieved quite cheaply, except for when you lock yourself into some designs or choice of components which audiophiles consider superior. Examples: it has to be class A, the opamps have to cost at least 15 $, dual mono blabla ....

You know that you didnt actually define class-A or class-B, which was the OP's key question.

Click to expand...

Why should I re-define what is already well defined? Click the link and read. If that doesn't satisfy you maybe take a look at Google.

You also didnt really answer question 2. There are tons of situations where anything can happen. 

Click to expand...

"yes no maybe. anything can happen." ? aha.

shrimants said:

I would say that class AB doesnt sound as good as class A.

Click to expand...

Any factual proof or measurements?.

shrimants said:

Can the cmoy be classified as AB?

Click to expand...

Depends on the op-amp you're using. An op-amp could contain a class A gain stage and a class AB output stage. Therefore a simple class X classification is not possible.
But does it even matter? No, what matters is the achieved performance.

but where does that extra current come from?

Click to expand...

The power supply.