Benchmark Media Systems AHB2 power amplifier

I first saw Benchmark's AHB2 stereo power amplifier at the 2013 Convention of the Audio Engineering Society, in New York City. On silent display in Benchmark's booth on the convention floor, its compact size and comprehensive features made the amp immediately attractive, and the design was described as a departure from traditional analog and digital amplifiers (footnote 1, more on that later). It was also explained to me that the AHB2 was based on designs by Benchmark's founder, Allen H. Burdick (whose initials it bears). By the time of Burdick's retirement, in 2006, Benchmark didn't yet offer a power amplifier, but the company used a prototype based on his work to evaluate their new digital products, and that amp was soon developed as a commercial product; Burdick died just weeks before the AHB2, now named in his honor, was shown at the 2013 AES convention.

Out of technical curiosity and my own interest as a consumer, I wanted an AHB2 right away, but the model had yet to be launched. Fast-forward to 2015, when I found myself investigating lighter power amps—I can no longer lift my 155-lb McIntosh MC303, or easily maneuver my 65-lb Parasound A31. I didn't recall how well the Benchmark AHB2 might suit me, until the amp began making a buzz (only figuratively!); happily, when I asked John Atkinson if I might review the Benchmark, he told me that he already had one that he could drop off.

Description
When I unpacked the AHB2, I was impressed with its small size and weight: the non-rack-mount version is 11.04" wide by 3.84" high by 9.34" deep and weighs only 12.5 lbs. I was also impressed with a seriousness of design apparent in its features: The fully balanced AHB2 has, on its rear panel, only XLR jacks for signal input, plus a toggle switch that allows the user to select among three input-sensitivity levels. For output, the Benchmark amp offers a choice of traditional multiway connectors or a stereo pair of Neutrik NL4 sockets. A third Neutrik NL4 is provided for use when the amp is in bridged mode, which is selected with a second toggle switch. On the front panel is a pushbutton on/off switch plus dual-mono indicator lights that indicate clipping, excess temperature, and the muting state the amp enters during power-up (or, presumably more rarely, when in fault-detection mode). These features, along with the detailed manual and specs included with the amp, suggest that the AHB2's target market is the audio professional or the recording studio.

Also packed in the box was a slick, gatefold brochure from THX Ltd., which developed the amplifier's patented Achromatic Audio Amplifier (AAA) technology. The THX AAA system combines an intentionally low-bias, class-AB output section with an auxiliary low-power, feed-forward amplifier; the latter drives the former with a correction signal that's claimed to eliminate distortion before it reaches the loudspeaker outputs. Tied to the AAA technology is a system of class-H power-supply rails that track power demands in order to increase amplifier efficiency without the penalty of added distortion; the power supply itself is a regulated switch-mode type that uses resonant switching, claimed by Benchmark to reduce noise.

Apparently, the brief history I'd been told at the 2013 AES Convention was only half the story. To paraphrase Benchmark's John Siau, from a recent e-mail exchange:

"The AHB2 was a team effort between Benchmark and THX. They were very excited about the performance that they were getting from a prototype that used the new THX-AAA technology. They sent the prototype, and we sent them one of our 35W DA101 power-amplifier cards. Our goal was to build a large amplifier that could exceed the performance of the DA101.

"Benchmark set the performance goals, completed the thermal modeling, and designed the heatsinks and the casework. Benchmark also designed the analog input stage, the user interface, and the FPGA-based protection system, and laid out the amplifier printed-circuit board using some of our specialized layout techniques. THX engineers designed the state-of-the-art, low-noise resonant-switching power supply and the core section of the amplifier board.

"The layout of the amplifier board is highly immune to magnetic and electrostatic fields, to achieve the A-weighted signal/noise ratio of 132–135dB. The two magnetic shield plates and all of the transformers are custom-built inside ferrite cores. The radiated magnetic fields produced by the AHB2 are much lower than those produced by other amplifiers of similar power, which means it can be placed directly above or below other, sensitive electronic components."

All of that makes sense, although the AHB2 can hardly be called a physically "large" amplifier. Installing the AHB2 was trivially easy, as it should be for a power amp. I simply connected my Audio Research MP1 preamp to the Benchmark's XLR inputs, and my banana-terminated speaker cables to its multiway output terminals, plugged in the power cord, and pushed the Power button. All seven of the AHB2's indicator LEDs briefly lit up, a relay clicked, and out came music.

Listening in the City
Benchmark and THX make much of the AHB2's lack of audible noise: something good systems aren't troubled by in any case. That said, it was probably the AHB2's low noise level that revealed to me much more apparent low-level detail in already-familiar recordings. I qualify that statement with apparent because, after hearing the AHB2 uncover previously unheard subtle details, I found I could now hear them when I returned to my other amps. I suspect that, being newly informed of their existence, my ear/brain could more easily extract those details from the output provided all along by my other amplifiers. This made a much greater impression on me than any subjective awareness of a lower noise floor, per se.



Footnote 1: For a full description of the design and engineering technology packed into the AHB2, click here.
COMPANY INFO
Benchmark Media Systems, Inc.
203 E. Hampton Place, Suite 2
Syracuse, NY 13206-1633
(800) 262-4675
ARTICLE CONTENTS

COMMENTS
eriks's picture

Nice review, but I'm still trying to get my head around a single person liking both Monitor Audio Silver series speakers as well as B&W 800 series speakers. Maybe B&W has changed much in the last 5 years. Also might be nice to add the MA Silvers to the "Associated Equipment" list. Now that I think about it, I have only heard B&W's high end played by Krell equipment when my buddy Dan was still in charge. I may need to reset my expectations of the brand now.

John Atkinson's picture
eriks wrote:
Also might be nice to add the MA Silvers to the "Associated Equipment" list.

It's always been there: look under the "Connecticut System" subhead.

John Atkinson
Editor, Stereophile

eriks's picture

Didn't realize the two were separated! Makes more sense though.

Kal Rubinson's picture

This helps to minimize any corrupting interactions between them.
;-)

HammerSandwich's picture

This is one heck of an amplifier!

"However, I was surprised to find that the S/N decreased slightly when the amplifier was very hot."

JA, I believe you have (re-)discovered Johnson noise.

John Atkinson's picture
HammerSandwich wrote:
JA, I believe you have (re-)discovered Johnson noise.

:-)

Seriously, the rise in the noisefloor, though small, was a little greater than I was expecting.

John Atkinson
Editor, Stereophile

billyb's picture

Great article! Thank you for reviewing this amplifier and testing it so thoroughly. I wish more equipment could spec out so well. As this highlights the issue of measurements vs subjective listening, would you consider dispatching this amplifier for a second opinion? Would one of the other writers at Stereophile care to listen? Perhaps Mr. Dudley or Reichert could give their impressions since they have different equipment preferences? It would be fascinating to juxtapose viewpoints on this amplifier.
Regards,
Bill

dce22's picture

This amp is designed to show good THD+N on the AP test you can't correlate this measurements to the other amps tests in stereophile because the amp has no gain it puts AP generator/analyser in the low distortion field if you loopback AP2 you will see that from 10-500mvRMS has 0.015-0.0004 THD+N if you add 23db gain to that line the amp that has no distorion has to have THDvsPower graph line from 0.003-0.0003 downward slope strait line from 100mW to 30Watts then flatten's out toward 100 Watt on 80kHz bandwidth, if put the analyser in 20kHz bandwidth the line will be lower but still the same shape like this

http://www.hardwareanalysis.com/images/articles/large/12004.gif

http://www.hardwareanalysis.com/images/articles/large/12003.gif

http://www.hardwareanalysis.com/images/articles/large/12005.gif

as is measured on a class a amp with lower distortion then AHB2 and when you crank the volume control on the pre-amp to compensate for the low gain you will amplify the noise of the DAC or Phono-pre up to levels of the other amps so low noise in the AHB2 is just red herring if you lower the designed gain of any amp the noise gets lower as well.

PS. To John Atkinson you cant measure AES48 compliant device's noise by shorting it to the GND you will pickup noise from the enclosure, you need to short out the signal wires or put a resistor between them.

Ishmael's picture

You might have a good point in there somewhere, but the use of punctuation would sure help us decipher it. Did you hammer out your comment on a vintage cell phone or something?

dce22's picture

My apologies im always not very good at this grammar... today :)

AHB2 is a excellent amp you just can't look at the THD+N graph and directly compare to the other amps with "normal" gain structure.

elvisizer's picture

the ahb2 is not "designed to show good THD+N on the AP test". It's designed to work with professional line levels, which allow lower noise and distortion because the amp has lower gain.
If your equipment doesn't support professional levels, you can switch the ahb into higher gain mode.
there's nothing shady going on here, you just don't understand the design.

eriks's picture

Hi Stereophile,

This isn't necessarily directed at Benchmark, but I would really like to see that Stereophile will be publishing THD+N at 1 watt instead of at full power. Of course, it IS included in the charts, but I think reporting this as it's own number would go a long way to improving the quality of amplifiers instead of the power of amplifiers. Even if we set a target for 2017, I think it would be a good improvement for the 21st century.

Best,

Erik

johnf's picture

Will M22 be reviewed soon?

dumbo's picture

As a current owner of a few DIY Hypex NC400 amplifiers I see some of the measurements of this Benchmark amp looking very similar to what has been seen in the Hypex amp/smps modules in terms of low distortion figures. Given that both units use an SMPS power supply I wonder how different the two units would sound from one another. Or more appropriately, how little sound signature each of the two amps impose on the original source material.

There are many folks who claim that Class D amps sound very flat and lifeless (I don’t agree) but is this claim due to the use of an SMPS or more to do with how a Class D amp works compared to their Class A/B brothers?

I dont think Stereophile would ever measure a Hypex NC400 amp given its DIY status but it would be very interesting to see a comparison of this benchmark unit which appears to be well liked by the reviewer up against the Hypex offering. Both units would fit within the criteria of being light weight, quite powerful and fairly inexpensive. Maybe the reviewer could consider one of the Hypex units in his search for a replacement amp which meets his needs.

If measuring a DIY Amp against an OEM would burn too many bridges for Stereophile then I think another alternative would be to put the Merrill Audio Veritas on the test bench instead as it should be just as impressive in terms of measured performance.

pablolie's picture

I had both poweramps in and compared them for over a month. In the end I picked the NAD M22. It just seemed more *fun*, seemed to add some sparkle. Mind you, it was of course extremely close, and sometimes I thought I was just imagining it all. But during extended listening I kept thinking I was having more fun listening to the M22, there seemed to be additional extension and a nano-ounce of extra clarity. BTW the NAD M22 has built-in Hypex Ncores, not a DIY, and supposedly NAD adds some tweaks to the design.

Associated equipment was
Source: Benchmark DAC2HGC (I feed it via Squeezebox Touch, library for more critical listening is FLAC)
Speakers: Totem Element Fire and also KEF LS50.

PS: It's also worth noting that the $12k Theta Digital Prometheus -which received much praise from Stereophile ("measured performance of Theta Digital's Prometheus is superb")- is also based on a Hypex Ncore design.

johnny p.'s picture

If Kal says he can 'hear new details in his old amp', he at least owes us an update. Say in a few months - to see if the details were retained.

TJ's picture

Kal, it's reasonable that the AHB2 could be a better match with one set of speakers than another. However it seems puzzling that once you heard a mismatch with the Monitors, you could then identify a sound problem with this amp on the B+Ws, but only on one recording. How would one know which playback is more faithful to the original recording, the AHB2 or collectively the other amps chosen for comparison? Is it possible that the nice "bell-like" quality of her voice on the other amps could reflect some coloration that's cancelled in the AHB2 circuitry?

RaimondAudio's picture

Measurements were made with what gain?

hifiluver's picture

astonishingly good measurements. good luck benchmark. real cutting edge stuff. you deserve all the success.

JRT's picture

On the measurements page, the caption under figure one mentions red colored curves twice, while there is only one red curve, and does not mention the green colored curve at all. I suspect that the green colored curve is the one with the 2_Ohm load.

Typos happen. Thank you for the measurements.

BTW, I am surprised that I have not noticed that before, because I have probably looked at this webpage more than a dozen times since it was first posted more than 5 years ago.

X