Amplifier Power: Better by the Pound?
Some fellas look at the size.
Some fellas look at the sheetmetal they wear;
Some fellas look at topology.
First I look at the power supply.
I am, of course, paraphrasing “First I Look at the Purse” by the inimitable Peter Wolf and the J. Geils Band, who were covering early Motown’s The Contours (the song itself was co-written by Miracle Smokey Robinson, by the way). “Some fellas look at the eyes/Some fellas look at the nose/Some fellas look at the size/Some fellas look at the clothes…First I look at the purse.” I have a very funny Peter Wolf story from my misspent youth, but on advice of counsel I ain’t gonna tell it here.
Okay, I’m squeezing a metaphor until it hurts, so let me put it another way. When shopping for amplifier power, most of us calculate, consciously or otherwise, watts-per-dollar. I’m here to suggest another metric: pounds-per-watt, (kilos, if you prefer), divided by dollars. Amplification by the pound.
See, once upon a time, copper and steel were as cheap as the dirt from which they sprang. Silicon, at least in the form of output transistors, was very expensive. And so even relatively low-power amplifiers were big and heavy, thanks to massive power transformers (large steel bobbins wound with yards of heavy copper wires) and banks of high-current storage capacitors (essentially, steel cans full of copper-foil jellyroll). In conventional “linear” amplifiers these parts, always among the most costly components in any amplifier design, work together to create reservoirs of electrical power that, controlled by the output devices, become the output signal routed to your loudspeakers.
What a difference a few decades make: today, transistors are almost as cheap as the beach sand they’re made from (your iPhone contains more than a billion), while copper and to a lesser extent steel, are just the opposite. This matters because, in a linear world at least (we’ll tackle Class D another time!), power-supply beef is one of the two fulcrums upon which amplifier performance balances: the other is output “safe operating area,” a three-axis map of the current, voltage, and time capacities of the transistor that’s gate-keeping the output.
Back when copper was cheap, standard practice was to design a massive power supply able to deliver all the current you might every wish for, up to and beyond the maximum capacity of your output devices. Engineers usually designed heavily regulated, “stiff” power-supply voltages (“rails” in the lingo) that remained at spec regardless of the demand; whether the music was loud or soft, dense or spare, ±60v (or whatever) remained ±60v. This is of course wasteful, since most of the time nearly all of that voltage potential is just sitting there doing nothing other than generating a bit of heat. But what the hey—copper’s cheap.
Or was. So engineers began dubbing around with variations on “smart,” or at least smart-er power supplies: Class G, Class H, and so on, where the power-supply voltages are deliberately allowed to “sag,” or step among lower, preset levels, when demand is low, and can reach higher, peak-demand values only for relatively short periods of time. (Switch-mode supplies are another option.) Consequently, over the past decade or two power supplies have been shrinking, as designers strive to mitigate the impact of ever-rising raw material costs.
There’s no better place to witness this trend than among AV receivers, which have steadily gotten smaller, and especially, lighter, as designers downsize power supplies and engineer multichannel amps more for real-world conditions than spec-sheet competition. This explains why more and more receivers specify power as one- or two-channels-driven rather all-channels: While the receiver’s supply has plenty of current to drive one or two channels to full output simultaneously, it will fall short when three or more are demanded for more than a few hundredths of a second. This is not so clearly the “cheat” that many of us reflexively assume; it actually reflects real-world conditions better than the brute-force, all-channels approach, since analyzing actual multichannel recordings (movies or music) reveals that multiple channels virtually never require full power, simultaneously and in-phase.
In pursuit of this slimming program, designers have turned increasingly to the “smarter” amplifier topologies mentioned above, though marketing materials often decline to name them, turning instead to proprietary trade names like Intelli-Amp or Smart-I-fier. (Hey, those are pretty good! Anybody buying?) I’ve no proof, but suspect this is because many audiophiles, without knowing exactly why, turn up their noses at non-conventional topologies like Class G and H. The village wisdom is “they don’t sound as good,” and the bad smell has seeped into the mass market. (This is largely nonsense in my book: you can find both superb and under-performing examples from every amplifier class, even including the high-end favorite Class A.)
Nonetheless, the majority of separate power amplifiers remain Class A and A/B, and among these, big, heavy, high-current power supplies continue to be a hallmark of quality, and usually performance. Of course, big, heavy amplifiers tend to be expensive, esoteric, fancily finished ones with billet-aluminum front panels and gold-plated connectors. But if you can find a cheap’n’cheerful model that’s still boat-anchor heavy, absent lead weights built into the chassis (don’t laugh, we’ve seen it done) it might just be your best value.