How do amplifiers work?

Nice posts, guys. Even simpler, think of an amp as a signal-modulated power supply for your speakers, properly hinting the importance of its power supply design and execution.

Thanks for the great responses!
One thing that still confuses me is the wattage output during peaks in the music. I'm not sure the best way to phrase this, but for instance, during a heavy bass note, combined with any number of other instruments, when a large amount of power is being drawn, does the amp typically send more wattage than it is rated for(continuous rating)? I read one thread where an owner of a Plinius amp rated at 125watts, said that the amp could output, at peak moments, up to 1000watts before clipping. Where as another Plinius amp, which was rated 150w or 200w, clipped at a much lower level. Obviously the 125w amp is of better design, but how does all this work(stored power?, dynamic headroom?), and what is the typical output of most amps, aside from continuous power. Can small (under 100watt) amps, actually output clean, but brief bursts of wattage that far exceed 100watts? and by how much? and how does one know this? I've never see it in the specs I've read, or don't know what to look for.

Yes, an amp does have peak power normally quite a bit in excess of its nominal continuous rating. The class in which the amp works, the size of its power supply and whether it is stiffly regulated or not are all factors in by how much an amp will exceed its average rated power on peaks. Interestingly enough, in the days when objective measurements were not met with some kind of exorcism rite there was a spec often provided as to the “dynamic headroom” of the amp given in dBs. High Fidelity magazine also touted rating power amps in dB watts so that the single number given would correlate better with the ability of the amp to actually produce an increase in SPLs. The single most significant number that would address your concerns though is the dynamic headroom figure. Dynamic headroom probably accounts in great measure for this subjective audio notion that, based on listening to different power amps (ostensibly in the same system and the same room, but subjective audiophiles are not prone to quibbling over mere details such as limiting the number of variables), you get your "bigger watts" and you get your "smaller watts", then you can also get your "tube watts" and your “solid state watts". All this constitutes, in all likelihood, more bogus observations from the subjective front. It can all be explained "scientifically" and far better than by a duffer like me. Good day.

Wet, for some insight into how it all works, we'll have to talk about another rating, the efficiency of the speakers. The standard efficiency rating is expressed as a certain sound pressure level measured one meter away that is produced using one watt. Loudness is measured in decibels (dB). Say your speaker is rated at 87 dB. For every 3 dB increase in loudness the watts have to double. So at one meter, in this example a pressure of 90 dB takes 2 watts, 93 takes 4, 96 takes 8, 99 takes 16, 102 takes 32, 105 takes 64 and so on. However the efficiency ratings assigned to speakers are often overly optimistic, so as with everything audio take them with a grain of salt. The other interesting thing about the dB scale of loudness is that it takes an increase of 10 dB to make something twice as loud. To get a handle on how loud is loud search the net for a chart of typical sounds. Don't quote me but an idling car might be 50 dB, going 70 mph might produce 80 dB inside the car, a jet airplane at 5000 feet 95 dB and standing beside a jackhammer 120 dB.

Some speaker makers provide the maximum sound pressure they believe their units can reliably reproduce. 115 dB would be a pretty good rating as that's terrifyingly loud. In our example, figure out for yourself how much power that's going to take.

One of the wild cards in the loudness scenario is how the speakers use power at different frequencies. Heavy bass notes indeed take more juice than brushing cymbals. Some speakers are able to extract more potency from the amp at lower frequencies. Which ones? Read reviews or listen to them. Amps and speakers loafing along under heavy load can get louder and do it sounding cleanly. Straining combos have nothing left to give but more distortion.

One thing to look for is physical weight. If one amp has a rating of 100 watts per channel and weighs 50 lbs. and another a rating of 150 watts but weighs 25 lbs. there's a high liklihood the heavier one will supply more peak punch than the lighter one. Power supplies are heavy! Put another way, one thing to ask yourself while shopping is: "How much of a boat anchor am I willing to live with?"

Sorry but there's no simple way to answer the question: "How potent is my amp?" Numbers only tell a part of the story. Don't expect them to replace your own judgement and sensibility. The specs help but when it's all said and done YOU are the ultimate measuring device. Find your own comfort level. Once you can say this amp works for me and I'm happy with it, you're there. If it hurts when Johnny down the block says his amp does 1000 watt peaks, you're not there.

The weight of the amp is some kind of indication of its prowess if you are talking about class A or AB amps. Those with switching power supplies can be made to be very potent and still weigh very little (ask Bob Carver). How they might sound is another issue almost entirely (don't ask Bob Carver). Again, toroidal transformers, I think, usually weigh less than standard ones for the same rating, so that the weight thing is at best a rule of thumb of sorts. Subjective evaluations are all very well and good (in fact they are neither, but that is too contentious an issue to deal with here), but don't expect a proverbial quart out of a pint pot. BTW when you are driving along in your car when you want to gauge how fast you are going don't look at the speedometer, judge it by the seat of your pants. As in subjective audio, if it feels fast, then it is fast. Think about it as subjective autophilia...