This is a review and detailed measurements of the Benchmark AHB2 Amplifier using THX technology to reduce distortion. I was going to request one for testing due to membership demand but to my pleasant surprise, the company volunteered to contact me and send not one, but two units! I requested the second one because of something I was seeing in the measurements that turned out to be my issue, not the amp's. The AHB2 retails for US $2,999 from the company direct.
The AHB2 is a very compact form factor, enabled by use of switching power supply to increase efficiency/density:
There is just a power switch in the front. It might be but I constantly reached for the screw next to it to power it on! Having the power switch be slightly different color or shape may help with this.
There are a set of LEDs that show the status of the amplifier. They are driven by an FPGA (field programmable digital logic) that monitors all aspects of the amplifier and shuts down the unit if stressed. It did so very quickly and efficiently in my testing. Usually this is done with purely analog or maybe a small microprocessor. As such, the coverage of scenarios that could damage the amplifier is much less complete than using the approach Benchmark is using in AHB2.
Here is the back panel:
This is a very minimalistic approach given the small amount of real estate available. Only input is XLR balanced which is just fine in my book. A gain switch allows the input level to be controlled so that you can just use an RCA to XLR adapter and still get full power (2 volt max in high gain). At the other extreme, the low gain setting is designed for professional pre-amps with lots of output, getting its max power at 9.8 volts. Using this scheme, the signal to noise ratio can be improved as you will see in measurements later.
For speaker terminals there are two sets: classic heavy duty banana jacks and SpeakOn. The SpeakOn are locking and provide the best performance. I lost about 3 dB of performance using the banana jacks due to less secure connect there. So the measurements you see are with SpeakOn jacks.
Another minor nit on that note: I had a rather cheap SpeakOn cable and its plug was just large enough to hit the silver screws that the sockets are mounted with. Typical round shaped SpeakOn ones are not an issue. Flush mount screws would solve this problem.
There is a switch for selecting stereo or mono bridged output. The latter quadruples the amount of power available and unlike typical bridging, according to Benchmark comes at no penalty in distortion! I plan to test this later as it has a minimum impedance of 6 ohm so I could not use the 4 ohm setup I used for this testing.
There is a very beefy AC mains cable with really nice locking tabs. Insert it and it stays put unless you push the two red tabs on each side. Nothing is more aggravating than the AC cable coming half-way lose from the IEC socket.
Being a proper company of course the AHB2 comes with full set of regulatory/safety/emissions certifications which is super important with power amplifiers given the high voltages and currents running around in them.
In use the AHB2 stays very cool for a power amplifier when it is idling, outputting little power which was a nice surprise.
As to THX technology, it merges a low-power but very low distortion amplifier with a high power but higher distortion amplifier. The distortion of the latter is not seen because it is producing so much power (so the ratio of distortion is lower). The general scheme is not new, dating back to 1980s but new implementation is. We have seen this in stellar performance of
Massdrop THX AAA 789 amplifier which broke new ground in level of distortion and noise. Will the Benchmark AHB2 manage the same? Let's see.
Power Amplifier Audio Measurements
As usual, I start with my 5 watt output test using 4 ohm load using SpeakOn terminals as noted:
View attachment 26575
I hope your jaw is on the floor just like mine was when I saw this picture emerge!
113 dB THD+N in a power amplifier? Are you kidding me?
Look at the harmonic distortion. The worst case spike is below -130 dB! This is insanely good. Of course this type of SINAD (signal and noise ratio) crowns the Benchmark AHB2 as the best I have ever tested:
View attachment 26576
The performance was so good I literally had to rebuild my dummy load to get there. Even the quality of the metal used in the connectors matters to get to this level of distortion. I replace all my dummy loads with higher precision ones that have much less VCR (voltage coefficient of resistance). Resistor values can become voltage dependent creating distortions of their own. Up to about 105 dB of THD+N, it doesn't matter but beyond that, the VCR was the dominant distortion, not the Benchmark AHB2 amplifier!
It was requested that I show the breakdown and distortion so here we are:
View attachment 26577
Our best case hearing threshold is -116 dB SPL so no question that this level of distortion is totally inaudible. Separating the noise from THD, we see that it is noise that we are measuring as THD+N, rather than distortion:
View attachment 26578
And this is with exceptional noise performance of Benchmark AHB2 as we see in the graph of THD+N versus power:
View attachment 26579
Even in high gain, the AHB2 easily outperforms the DIY Hypex NC400 I had tested before which used to be the best amp I had tested.
We have 185 watts of power at incredibly low distortion of 0.00016%, besting the company specifications.
Note that the FPGA protection mode kicks in and essentially shots the amplifier down past the limit. You get absolutely distortion-less and noiseless performance until there is no more.
EDIT: here is performance in bridged mode using same 4 ohm load:
View attachment 26592
You get 500 watts of stunningly clean power. Protection circuit shut the unit down after that so it is quite safe to try. With a THD essentially matching the non-bridged mode, there is no reason to be afraid of trying this.
Intermodulation distortion versus power level shows the same story of clean power:
View attachment 26580
If you want to know what you get for extra money over bargain amplifiers, you can see the difference in the above graph. We are not talking about 5 to 10 dB but whopping 40 dB better!
I have had requests for intermodulation distortion using dual 19 and 20 kHz tones. Here is that:
View attachment 26582
THD+N versus output level looks far cleaner than anything I have tested before:
View attachment 26583
Even at the limit of our hearing (20 kHz), we have vanishingly low amount of distortion. The graph is exaggerated so shows a rise there but in absolute levels, despite 90 kHz bandwidth of the test, we have incredibly low THD+N of just .004%.
The sharp spike at 45 Hz in green shows the amplifier going into protection mode. So don't pump that sine wave continuously into it at 133 watts.
Frequency response is exceptional too as expected:
View attachment 26581
Since this is not a switching amplifier, there is no filter there allowing the bandwidth to go to 200 kHz and beyond. Lowest band of AM radio is 450 kHz so likely you could use the AHB2 for an AM radio transmitter!
Usually when we test switching amplifiers we see all kind of "interesting" things in their outputs above hearing range. The AHB2 is a classic configuration albeit, with a switching power supply so all is well and clean here:
View attachment 26584
My reference graph for the
Hypex NC400 used an AES filter (by accident) so I ran the AHB2 both ways, with or without that filter. Using the filter (in green) so the two are equal, we see much cleaner spectrum below 200 kHz and of course, no massive switching spikes. Worst case spike is below -115 dB. In other words, the Benchmark AHB2 is clean even in the areas you are not looking! It is like a restaurant scrubbing their parking lot with soap and pad as well as their dishes.
My loose wires on dummy loads is not the best setup for measuring crosstalk but here it is anyway:
View attachment 26585
Where our hearing is most sensitive (2 to 5 kHz), separation is around 100 dB which is way, way more than we need. Despite the small enclosure, the AHB2 manages exceptional numbers here.
EDIT: forgot to run the classic SNR test in the original review:
View attachment 26636
Wow, assuming you play at peak of 120 dBSPL, your noise floor will be at -10 dBSPL! That is absolute silence.
Conclusions
It goes without saying that the Benchmark AHB2 breaks new ground with respect to performance of power amplifiers. Using it, you can be assured that any distortion that you hear is from other sources (speaker, source, content, etc.). This is what I look for in high-end audio: absolutely the best performance so no second guessing is involved. You buy once, and you are happy forever!
All of this comes from a company that is a model of transparency with proper and accurate measurements of their products on their site. And importantly, volunteering to have that data shown to be correct by independent sources such as us. Support form the company for me at least has also been exceptional.
Yes, $3,000 is fair bit of money but is pocket change in high-end audio. Sales tax is higher than $3,000 for most of those products! I hear a lot of talk from audiophiles to have the equipment get out of the way of enjoying the music. Well the Benchmark AHB2 at levels that assuredly passes full transparency. So if that is your moto, you better stop buying boutique products with no specs and independent measurements and get an AHB2. Everything you hear then will be what is in your source, not the dirty dishes that your amp my serve your food on.
Needless to say,
the Benchmark AHB2 gets my strongest recommendation for power amplifier. The pink panther agrees, having hit the ball out of the park after I took that shot!
------------
As always, questions, comments, recommendations, etc. are welcome.
The comedian Chris Rock said it best: there is a difference between being rich and wealthy. If a wealthy person had as much money as I have, he would hang himself! So please donate enough money so that I can become wealthy using either:
Patreon:
https://www.patreon.com/audiosciencereview), or
upgrading your membership here though Paypal (
https://audiosciencereview.com/foru...eview-and-measurements.2164/page-3#post-59054).
