Solid State Power Amp Reviews

After the null was as good as I could get it by ear, I played some music signals and listened through the monitor to the nature of the residual sound. It was very quiet, so much so that I had to plug headphones into the appropriate outlet on Hafler's box in order to tell that the residual seemed very, very clean. Even with discs which I know to be rather worn (and thus a rich source of mistracking transients), the sound from the monitor speaker remained sweet and clean.

This, I thought, bade well for the sound of the XL-280.

Sound Quality
But when I finally got around to auditioning the XL-280 on its own, my initial reaction was quite different from how I expected to respond to what might well be the most accurate amplifier made. Although the 280 was superbly clean and uncolored through its entire midrange, I was immediately aware of two areas in which it sounded very different from my reference amplifier (the awesomely expensive—$2900 per channel—but similarly powered Threshold SA-1).

First, the 280 produced quite a bit more low end than the Thresholds, and with this went (as is usually the case where there is simply more low end) an impression of diminished control and detail. The low end from some Telarcs—notably The Stokowski Sound— was positively billowing and sounded very loose, yet when I used that CD later as a test source for nulling capability, there was absolutely no indication, audible or visual, of any loss of woofer control by the 280.

Second, the 280 sounded somewhat rougher and drier at the high end than did the SA-1s. Massed violins sounded subtly dirty (although there was no added steeliness), and the whole sound had less of what I think of as "suavity."

Third, there seemed less of a feeling of space and air around the instruments from the 280. Some of their roundness and solidity seemed lost.

And finally, although the 280 sounded no more forward through the midrange than the Thresholds, it seemed to produce a slightly shallower soundstage.

Actually, the 280 reproduces both depth and spaciousness very well—better, in fact, than many of the higher-priced amplifiers I have auditioned during the past three years, but not quite as convincingly as some others. All of them, interestingly, were much more costly than the XL-280. In fact, the only comparably priced amplifiers I can think of which clearly surpassed the 280 in any respects were the $440 B&K ST-140 and Carver's $600 Model 1.0 amplifier after it was worked over for our "1985 amplifier challenge." The B&K had a much sweeter high end than the 280, but was not quite as good in any other respect. The souped-up Carver prototype sounded amazingly like one of our favorite tubed amplifiers (footnote 2).

It is important to remember that these are comparative assessments, and that I do not pretend to know which of the amplifiers are more "right" in any areas, because I really don't know what the original program material sounded like.

Were my speakers exaggerating bass that the 280 merely revealed, while the Thresholds complemented it? Was the XL-280 mercilessly exposing a roughness in all my program material that the Thresholds somehow prettified? And just how wide and deep were the original soundstages on which the recordings were made? I am not prepared to say. But there was no question in my mind that, with the two loudspeakers used for the comparisons (the Altec Lansing 301s and Sound Lab A-3s), the Thresholds produced what I judged to be a more musically realistic sound, with sweeter highs, better proportioned and controlled bass, and a more persuasive impression of the sound coming from a real performing space.

But look at the price difference! A pair of the Threshold SA-1 amplifiers costs a staggering 12.5 times as much as the 280! For the Hafler to not sound laughable by comparison is reason enough to view it with great respect.

Final Thoughts
As I write this, my null tests on other amplifiers, including the Thresholds, are still in the future. Perhaps, after those are completed, I will be prepared to say that the Thresholds sound better than the XL-280 because they outperform it on Hafler's own SWDT, and are thus more accurate. As of now, however, I can only speculate as to whether, in fact, accuracy is, or should be, the ultimate test of a component's suitability for use in a given system.

I know, I know, this is heresy. Accuracy is supposed to be the audiophile's touchstone, but there is accuracy and there is accuracy. There is intrinsic accuracy, which is the quality of perfection in an individual product, and there is ultimate accuracy, which is the ability of a complete system to produce convincingly musical noises despite the intrinsic inaccuracies of its components.

Some of our readers have written to express the conviction that we at Stereophile are obliged to perpetuate the "myth" of subjective testing because, were anyone to devise a test which could predict with absolute accuracy how a component will sound, we would all be out of a job. Well, the SWDT may prove to be just such a test. But we do not see the function of subjective reviewer as threatened by it, because as long as one component in an audio system is intrinsically inaccurate, the system will sound more ultimately accurate when that component is paired with another whose inaccuracies complement its own.

For example, loudspeakers are still very imperfect components, and vary in sound far more than do today's halfway-decent power amplifiers. Since one cannot listen to a loudspeaker without using an amplifier to do so, it is impossible to determine what the inherent sound of a particular loudspeaker actually is. So, even were a given amplifier demonstrably perfect, there is no guarantee that it will elicit convincingly accurate sound from a given loudspeaker system.

Clearly, the way to design an accurate loudspeaker would be to do so in conjunction with that amplifier which outperforms all others on Hafler's SWDT, but until loudspeaker manufacturers actually start doing that, choosing a power amplifier must continue to be done by ear, on the basis of what it makes our favorite loudspeakers sound like. (Or, conversely, we can start with the most accurate amplifier, as determined by the SWDT, and look for speakers that sound the most musically accurate with it.)

So, let's just forget about the SWDT for the moment and consider the Hafler XL-280 as a product.

Okay, so it's perfect—but how good is it? I am not even going to try to excuse it for not "sounding" as good as the Thresholds. They aren't its competition. In its own price class, I venture to say the 280 might well be a hands-down winner.

But—to quote Darth Vader—"There is another." I refer to the $440 B&K ST-140, an unknown factor at this point. That unit has apparently undergone substantial modification since I tested a sample back in August of 1984, with a resulting increase in output from 70 to 100 watts per channel. If the latest version has lost none of the high-end sweetness and openness of the original, it could well be a better choice than the Hafler. But since higher power often goes hand in hand with high-end roughening, I am not going to climb out on a limb and recommend the B&K over the Hafler without an audition. But it is with only that qualification that I recommend the 280.

(Significantly, the 280's high end was still growing gradually sweeter after about 20 hours of listening. I don't know how good it will become, with time, but I'll have a follow-up on the XL-280 when I do.)

---

Footnote 2: Carver is now mass-producing that modified amplifier as the 1.0t. We recommend that you read our forthcoming review to find if it sounds as good as the custom-tweaked version.