Benchmark AHB2 Amplifier

Essentially Perfect

Equipment report
Solid-state power amplifiers
Benchmark Media Systems AHB2
Benchmark AHB2 Amplifier

I don’t have much to say about the sonic qualities of Benchmark’s new AHB2 power amplifier because there isn’t that much to say. Used within its limitations and for its intended purposes it is in any practical sense perfect. I know this is not the sort of thing we’re supposed to say about products, but it has been evident for a very long time now that solid-state electronics, particularly linestage preamplifiers and power amplifiers, have reached a point where they are effectively a solved problem such that it is exceeding difficult to tell one from another even in the most exacting A/B comparisons. Ferretting out differences typically involves zeroing in on a very specific and limited characteristic or set of characteristics with such concentration as to leave one tired, uncertain, or both: in other words, an activity that is the very antithesis of what is involved in listening seriously, even critically—to say nothing of pleasurably—to music. [I must point out that this view is not shared by the TAS editorial staff. —RH]

That said, even by contemporary standards of the most sophisticated, exotic, and expensive electronics, this new amplifier is something of a technological tour de force. The AHB2 boasts a signal-to-noise ratio that exceeds 132dB and total harmonic distortion figures of –131dB (3rd harmonic) and –122dB (2nd harmonic) under no-load conditions; it also exhibits, so Benchmark claims, “no trace” of crossover-notch distortion, though it is Class AB, not Class A. (I should immediately add that these figures have not been achieved by heaping on huge amounts of negative feedback, à la those wretched solid-state designs of the Seventies: assisted by a feed-forward error-correction circuit, the AHB2’s noise and distortion hold up under dynamic, not just static, conditions.) The power output is 100Wpc, with a claimed 18 amperes of current in stereo mode, bridgeable to 380W mono. The amp, which is THX certified, is outfitted with a number of protection circuits, the operation of which is monitored by front-panel LEDs, is unconditionally stable, and, like all Benchmark products, is backed by an impressive five-year warranty.

I front-load my remarks with these numbers to get your attention. While Benchmark products have made serious incursions into the audiophile community, they are made for the professional market (where they are ubiquitous) and designed to the most exacting standards of performance and reliability. They are also for the most part minimalist as regards function and size: You won’t herniate yourself toting this amplifier around because it measures 11" x 4" x 9" and weighs a mere 13 pounds. I auditioned the AHB2 on my Quad 2805 electrostatics, Harbeth Super HL5pluses, and the new Falcon LS3/5as. (Digital sources are a Marantz 8000 SACD player and a Benchmark DAC; vinyl, a Basis 2200 and Vector IV ’arm with an Ortofon Windfeld pickup.) Allowing for differences in the speakers, the amp sounded identical on all, its 100 watts more than sufficient to drive them to levels far higher than I could comfortably sustain for more than a few minutes. The sensitivity of all these speakers is around 85–87dB and they present reasonable impedance loads.

An adjective that came to mind when I first fired up the Benchmark was “crisp.” “Ah, see,” I hear several of you saying, “solid-state.” Well, yes and then again no. For the past year and longer, the amplifiers I’ve been mostly listening to have been the Zesto Audio Bia, an all-tube unit, and the Quad 909 solid-state, but solid-state with limited bandwidth both top and bottom (the AHB2’s bandwidth is extremely wide). Now both of these amps are, generally speaking, neutral within what I call acceptable boundaries of neutrality, but neither of them is dead neutral. As I pointed out in my review, the Bia is perhaps best described as classic tube sound brought up to date: very dynamic, excellent transient response, well-defined and articulate bass, gorgeous midrange. But it is decidedly a tube sound with all the lusciousness and romanticism thereby suggested. The 909 is bandwidth-limited at the very bottom (below 13Hz), which inevitably translates into some phase shift that confers a little extra warmth on the presentation, and also at the very top, which lacks the almost crystalline clarity you get from frequency response that extends out to 100kHz and farther. But its supreme musical authority and naturalness are second to none and never fail to satisfy me when it’s in the circuit (which is often).

So, yes, coming after both of these units, the Benchmark would by comparison sound a bit crisp. Within a remarkably short period of time, however, this adjective dropped from mind as my mental processes shifted from comparison mode to enjoying music, all the attention drawn to the source material and the rare sensation of feeling as if I were listening back to the source. The Tokyo Quartet’s valedictory recording of the late Beethoven quartets is close up and slightly dry but the instruments are nevertheless warm and beautifully rendered. By comparison, the Belcea’s traversal is recorded more distantly, the acoustical space—a lovely one—far more enveloping, and it sounds that way. In neither case does the Benchmark appear to overlay any sound of its own—nor would I expect it to. What is there to overlay? Used within its specifications, distortion is nonexistent, while the noise floor is claimed to be lower than the lowest digital by as much as 10dB.

You will notice, by the way, that I’ve said almost nothing about detail and resolution. There are two reasons for this. First, the Benchmark resolves and thus reproduces all the detail there is on every recording I played that I know well, so there’s no real need to accrete half a dozen examples to buttress the point. But let me cite the toughest resolution test I know: for the a cappella introduction of “Moon River” on her Johnny Mercer album, the soprano Jacintha was placed in an isolation booth while occasional chords from a piano were played through her headphones so that she could stay in tune. Despite heroic measures to ensure isolation, the chords nevertheless bled through her headphones and can be heard during the silences. Now these are extremely far down in level; in a couple of instances they are all but inaudible. If a component can reproduce them, that’s about as much resolution as you are ever likely to need. Suffice it to say the Benchmark did.

But second, and more important, is that I’m getting tired of hearing detail and resolution used as be-alls and end-alls in audio reviewing. Too much detail is unnatural and it’s certainly not realistic, and because you hear it with some components and not with others does not necessarily indicate the former are accurate, the latter not. The reality is that it’s easy to phony up detail with rising top ends. This is the case with speakers, phono pickups, and even some kinds of electronics. Try it sometime: Borrow an equalizer or simply a control unit with a treble control and boost the highs even slightly—voila!—more detail. I’ve gone out and auditioned many components that are raved about for their so-called resolution and in virtually every instance I’ve found they do not sound truly neutral or natural when it comes to reproducing musical instruments and voices. [See my editorial “Too Much of a Good Thing?” in the February 2016 issue for my substantially different perspective on this subject. —RH]