Follow-Up, July 1988
JGH wrote again about the Hafler XL-280 in July 1988 (Vol.11 No.7):
This budget-priced ($650) power amplifier has been the center of some controversy over the validity of a nulling test that Hafler devised, wherein an amplifier's output is fed, in reversed phase with its input, to yield an overall signal cancellation. Hafler contended that any amplifier that gave a perfect null must, by definition, be a perfect amplifier, and that, since the Hafler XL-280 gave by far the most nearly perfect nulling, it must be the most nearly perfect amplifier in existence.
The argument sounded airtight, but the acid testlisteningburned holes in it. When reviewed in Vol.10 No.1, although the XL-280 was a very good amplifier, it did not sound like what I thought the most nearly perfect amplifier should sound like. Its low end was slightly bloated and deficient in impact and control, its soundstaging was a little cramped, and its highs were dry almost to the point of graininess. We were forced to conclude that, perhaps, real-world loudspeakers don't like nearly perfect amplifiers.
Readers' letters in response to that review, however, suggested that some of the problems we heard may have been direct results of the measures Hafler had to take to eliminate the effects of phase shift on nulling errors. (The very slight time taken for signals to pass through a typical amplifier can cause incomplete nulling, even though they may have no effect on the sound of the amplifier.) Hafler, on the other hand, took the attitude that if "real" loudspeakers don't like perfect amplifiers, it is the loudspeakers that should be changed, not the amplifier. We couldn't disagree with him, but the options here would have been to change every loudspeaker in existence or to change one amplifier. There was little question as to which was more practical.
A few months ago, we received another sample of the XL-280, which was in effect claimed to be even more nearly perfect (ie, it produced a deeper null) than the first. Whether this is because it has lower distortion, less phase shift, or both, was never revealed to us, but I gave it a listen.
I don't have that original XL-280 on hand for direct comparisons, but from what I remember of it, and after rereading my comments about it, there is no question that the latest one is substantially better. The high end is much less dry than it was, with none of the grittiness I heard from the first one, and although it's still not quite as sweet as that of, say, the B&K ST-140, it now has about as nice a high end as any solid-state amplifier in its power class. The low end is a bit more under control, having somewhat more heft and impact than before, and even the soundstage seems to have opened up. Only at the low end does it yield to higher-current designs in terms of control and apparent extension. A very good buy, highly recommended (putative Class C, lower end).J. Gordon Holt