Audiopax Model Eighty Eight monoblock power amplifier Manufacturer's Comment
Editor: Finally! Six long years after Eduardo de Lima's first published papers in 1997, John Atkinson and Robert Deutsch have wonderfully proved the point that Eduardo's been making.
Let's face it, there's no real art or science involved in building an amp that'll drive a test-bench resistive load. It's easy to build an amp with "low distortion." The Crown DC-300 broke that ground over 30 years ago. But the test bench doesn't know what to do with the second-order distortion of single-ended amplifiers. Certainly, if it existed as audible distortion at the speaker, we'd consider it flawed performance, too.
Even in this age of (relative) enlightenment, we still hear the tired old myth that music-lovers want single-ended amplifiers because they somehow prefer the pleasant colorations of second-order distortion! Ironically, these thousands of sensitive listeners agree on only one thing: the purity of the sound. That unmistakable purity that they all hear is utterly at odds with the notion of greater audible distortion. Guess what? The two devices that produce the most second-order distortion are: 1) single-ended amplifiers and 2) loudspeakers!
Today we know that, when a single-ended amp is correctly interfaced with a loudspeaker's optimum polarity, the result is sometimes vastly less overall system distortion than is possible using any "low-distortion" amp. This phenomenon is not new. It's been discussed at length by Eduardo de Lima in numerous papers available at our website.
Regarding second-order distortion that the test bench doesn't like (because it doesn't have a speaker to cancel it), we agree—that would be really bad. Fortunately, we do have speakers to attach to our amplifier. And when the optimum polarity is selected, the results are just what RD heard after following Eduardo's suggestion.
JA's measurements and RD's listening tests (along with those of countless other single-ended fans) illustrate why measurable second-order distortion can be a good thing. The total amp-speaker combination is capable of producing significantly less total system distortion when compared to a typical "low-distortion amp" attached to the same speakers.
So we have progressed from certain static test measurements that may not always be relevant, to real-world dynamic testing that proves Eduardo's theory. Still, there are several relevant measurements that do seem to have a direct correlation with sound quality: 1) intermodulation distortion (IMD), 2) bandwidth, and 3) signal/noise ratio.
IMD is audible, and it's objectionable, even in small amounts (in much the way that odd harmonic distortion is offensive). Limited bandwidth results in an overly softened sound, often with poorer imaging capability. And low S/N ratios are bad news for the medium- to high-efficiency speakers likely to be selected. Even so, there's always room for interpretation. For example, the amplifier that JA measured was the same model he'd heard playing earlier on the 110dB efficient Trios in the New York City showroom. Back from measurements, it's once again in the Trio system.
As before, when there's no music playing, most folks standing right next to the horns can't tell the amp is even on. That's how quiet it is on these 110dB "test benches." Go figure!
From our studies of past Stereophile equipment reports, the Model Eighty Eight's squarewave, IMD, and bandwidth are all benchmark achievements for vacuum-tube single-ended amplifiers. And that's without using global feedback as a prop to "help" the test bench.
As the distributors of Avantgarde Hornspeakers, you'd think we'd agree with the notion that the Audiopax is only for Hornspeakers. Au contraire! Naturally, if you took arbitrarily acceptable second-order distortion limits as imposed by the test bench, so-called "usable power" is restricted. However, when you attach actual loudspeakers of reasonable efficiency (regardless of type), the results can be quite extraordinary. They take advantage of the Audiopax's unique system distortion-canceling properties, in exactly the way the Model Eighty Eight was intentionally engineered.
The Model Eighty Eight and its new, lower-priced, lower-powered sibling, the Stereo Eighty Eight, have won many a shoot-out with more powerful "low-distortion" push-pull amps, using loudspeakers in the 89-97dB efficiency range (alas, not our horns, as much as I wish everyone could afford a pair).
Do your intrepid readers still want more info? See our own (very similar) measurements, as performed on each Model Eighty Eight for its final QC. Learn how they correlate with what experienced listeners like Robert Deutsch are hearing. Check it out.—Jim Smith, Avantgarde-USA