Review and Measurements of Benchmark AHB2 Amp

[pma]

In that case, I would argue that whatever the problems - or I would say limitations - of the SINAD measurement by itself, folks should stop using "load dependence" as a reason that SINAD is a bad measurement, because for loads of 2 ohms or more, amps with high SINAD are also load-invariant, and conversely the load-dependent amps are in the lower SINAD range.

There is absolutely no technical correlation other than accidental between low SINAD and load independent FR. You may design amplifier with quite high output impedance like 0.5ohm - 1ohm which would be SINAD record breaking into resistor load and at the same time with highly modulated FR into complex load. I designed a class A amplifier with 0.5 ohm output impedance and distortion below -120dB. You may also design an amp with purely current output (very high output impedance, ideally infinite) and vanishingly low SINAD into the resistor load.

If it is a class D, then it depends on circuit and feedback topology and loopgain achieved by the design. You may design TPA32XX-like circuit with vanishingly low distortion and again at the same time horrible interaction of the output LC with the complex load.
Currently we may see at least 5 different principles and topologies of class D amplifiers. Still, only BP's designs are "universal" with limited compromises.

The things are not as simple as they look and it is my problem here, the oversimplifying without necessary background.

 

[pma]

Another example is the UcD180HG module that I have recently tested.

Hypex UcD180HG HxR amplifier module analysis and review

Hypex UcD180HG HxR module analysis and review After buying and testing AIYIMA A07 amplifier (based on TPA3255 class D chip) in January 2021 I was quite disappointed with the frequency response modulated by speaker complex impedance and its audible consequences. This is, unfortunately...

www.audiosciencereview.com

It has had moderate to poor SINAD 1kHz/5W/4ohm of 76dB (poor by ASR measures), however excellent load invariance of FR which reflects in very good sound experience with various speakers.

 

[misterdog]

Agreed! I'm surprised nobody here has attempted double blind testing of SINAD differences between the highest performing stack (all Benchmark from DAC to Preamp to Amp), a well regarded modern AVR (Denon 3700H) and an expensive but audiophile approved stack (Pass Labs solid state or McIntosh tubes). Would be interesting to see what the "listener preference" ends up being. Just like there is definitely a listener preference for the "Harman curve", is there a listener preference that's completely divorced of SINAD measurements?

We just need loudspeakers with a similar level of SINAD to hear the full effect of the differences.

 

[misterdog]

I would like to hear Tom's new Modulus 86 against my AHB2, the IMD is spectacular.
I have some of his original design Mod 86 Parallel/bridged monoblocs which bettered N-Cores.

Modulus-86

DIY 65W composite amplifier based on the LM3886 providing better than -130 dBc THD through the use of error correction. The Modulus-86 has no sonic signature allowing you to hear the music and not the amplifier.

neurochrome.com

All below the threshold of hearing obviously, but hey, who is to say it doesn't make a difference.

 

[Spocko]

Agree... but to the point he and I were making is that we likely won't hear any difference between a SINAD of 98dB vs 99dB. Someone might hear a difference between the two preamps, but I highly doubt it would be because of the SINAD variance being 1-2dB... it will likely be from something else.

Or dare we conjecture that SINAD variance of up to 5dB may be indistinguishable?

 

[peng]

Agreed! I'm surprised nobody here has attempted double blind testing of SINAD differences between the highest performing stack (all Benchmark from DAC to Preamp to Amp), a well regarded modern AVR (Denon 3700H) and an expensive but audiophile approved stack (Pass Labs solid state or McIntosh tubes). Would be interesting to see what the "listener preference" ends up being. Just like there is definitely a listener preference for the "Harman curve", is there a listener preference that's completely divorced of SINAD measurements?

This is a completely reasonable comparison I believe as it truly separates the emphasis paid to SINAD alone versus other considerations heretofore unknown that may take priority. We don't know what we don't know, so this comparison is definitely something that needs to be examined. I understand the necessity for high SINAD when processing content as you always lose something after every layer of processing, but for playback, I'm curious to see what listener preferences really are when determining the importance of SINAD.

I also want to know why there is no such comparison. It doesn't even have to be a Denon AVR, it could even be one that sits at the bottom of Amir's chart such as the NAD T758 V2. If compared in pure direct mode, at output level below 50 W into 4 ohm, I would like to bet... Would it be possible that many might have done something like that but would not post the results for fear of........?

Now someone will say you can't go by the ASR tests because there are things that matter but not measured, and the measurements were done with a load resistor blablabla instead of a complex load, but Stereophile does use a simulated load and from what I could see, the differences in FR don't seem to be too great, say within 1 dB at the extremes for even class AB amps such as an Emotiva amp and given that the AVR-X3700H likely have much higher output impedance than the AHB2, if one can tell a difference if will most likely be due to FR, not SINAD. While it could be audible, but as others mentioned, difference ≠ better, and given not everyone has the best discerning ears/brains it is also possible that in A/B/X if double blind, most people probably would not score much better than 50/50 either.

 

[Willem]

See here for something similar: the old blind comparison between three generations of Quad amplifiers: https://keith-snook.info/wireless-w...-World-1978/Valves versus Transistors DCD.pdf The panel could not tell them apart. Their designer and Quad's owner, the late Peter Walker, gave me a private version of the test, and although I thought I could tell them apart, Peter Walker cheerfully told me afterwards that I had been unable to do so. It was a useful lesson.

 

[rdenney]

I also want to know why there is no such comparison. It doesn't even have to be a Denon AVR, it could even be one that sits at the bottom of Amir's chart such as the NAD T758 V2. If compared in pure direct mode, at output level below 50 W into 4 ohm, I would like to bet... Would it be possible that many might have done something like that but would not post the results for fear of........?

Now someone will say you can't go by the ASR tests because there are things that matter but not measured, and the measurements were done with a load resistor blablabla instead of a complex load, but Stereophile does use a simulated load and from what I could see, the differences in FR don't seem to be too great, say within 1 dB at the extremes for even class AB amps such as an Emotiva amp and given that the AVR-X3700H likely have much higher output impedance than the AHB2, if one can tell a difference if will most likely be due to FR, not SINAD. While it could be audible, but as others mentioned, difference ≠ better, and given not everyone has the best discerning ears/brains it is also possible that in A/B/X if double blind, most people probably scored much better than 50/50 either.

The first sticky in the psychoacoustics subforum summarizes what we know about audibility thresholds. Amir has been trained to listen for specific distortion but most people have not, leading that thread to present strict and lenient standards. If people would read that before reading measurement reviews, it seems to me, they’d be empowered to draw their own conclusions.

Rick “who can’t hear harmonic distortion below about -36 dB” Denney

 

[Spocko]

I also want to know why there is no such comparison. It doesn't even have to be a Denon AVR, it could even be one that sits at the bottom of Amir's chart such as the NAD T758 V2. If compared in pure direct mode, at output level below 50 W into 4 ohm, I would like to bet... Would it be possible that many might have done something like that but would not post the results for fear of........?

Now someone will say you can't go by the ASR tests because there are things that matter but not measured, and the measurements were done with a load resistor blablabla instead of a complex load, but Stereophile does use a simulated load and from what I could see, the differences in FR don't seem to be too great, say within 1 dB at the extremes for even class AB amps such as an Emotiva amp and given that the AVR-X3700H likely have much higher output impedance than the AHB2, if one can tell a difference if will most likely be due to FR, not SINAD. While it could be audible, but as others mentioned, difference ≠ better, and given not everyone has the best discerning ears/brains it is also possible that in A/B/X if double blind, most people probably scored much better than 50/50 either.

Somewhere out there is a grad student looking for a research topic...

 

[peng]

Somewhere out there is a grad student looking for a research topic...

That would be great for sure.

By the way, just noticed a series typo in my post but I hope you noticed that too.

".....most people probably scored much better than 50/50 either......" should have been "....most people probably would not score much better than 50/50 either....."

 

[Donovan]

I have never seen an amp that measures this way. I don't know that anybody has. It isn't the first amplifier to use feedforward -- indeed, John Siau points out that feedforward is actually an older technology than negative feedback -- but it's rare in doing so.

It's great to see someone breaking all the "rules" of amplifier design, because it's been in a rut for some years now, with the best designers squeezing pretty much everything that could be squeezed out of available devices using conventional topologies.

Anyone who thinks all amps sound the same will be in for a rude but happy awakening when they compare an inexpensive amp to an amp like the Benchmark.

That said, anyone who thinks that the Benchmark is a *perfect* amp because of its steady state measurements will learn that that doesn't yet exist.

I've owned an AHB2 for several years now, and it's the cleanest amplifier I've ever heard. I was listening to it last night and I can only describe the sound as ravishing. Like silk. Or like Class A without the absurd size, weight, and heat. (If you've ever heard crossover notch distortion you know why Class A amps sound better than A/B ones, it's incredibly audible and annoying in small amounts.) And the low weight, tiny form factor, and efficiency of the AHB are really welcome. It's a save-the-planet kind of amplifier, without the sonic compromises of Class D.

But perfect amplifier? No. It excels in steady state measurements, but it isn't as dynamic sounding as other good amplifiers with which I've compared it. Again, this isn't a question of steady state measurements or amplifier power -- I'm not clipping it. But listen to a piano on the Benchmark and another good amp and you'll hear what I mean: it rounds off the attack.

Amplifiers actually differ surprisingly in the degree to which they do that. My Parasound A21, for example, is a less refined amplifier than the Benchmark, but it happens to be one of the best amps I've heard in this particular regard. There are those who say that the Benchmark robs music of excitement, and after several years of listening experience, I tend to agree. It's strength is in beauty of sound. The highs in particular are to die for, like gossamer.

Another issue is, paradoxically, lack of distortion masking. Lower order harmonic distortion is known to render higher order harmonic distortion euphonic. This is because of the mechanism by which we detect the timbre of sounds. And higher order distortion is unfortunately quite common on poorly made recordings and badly-designed equipment. What this means in practice is that the AHB2 will sound great on clean recordings, but will pass the harshness through on more distorted ones, while some more colored amplifiers with lower order harmonics will tend to mask that harshness.

So the Benchmark isn't a perfect amplifier, but anyone who has never heard a really good amplifier or thinks that all amps sound the same is going to be blown away by the sound. It's a bargain, too, by the standards of high end amplifiers, and with its sophisticated protection circuitry, quite bulletproof!

Great insight, I’m actually looking at this Benchmark amp, Parasound, and Rotel for a 2-channel amp for a future Spatial Audio Sapphire M3 speaker purchase. Is there particular music styles you preferred with the Benchmark and Parasound amps? I like 70’s and 80’s rock and pop, electronica, chill out (like Enya, Clannad, B-Tribe, etc..), and some orchestral classical.

 

[peng]

Anyone who thinks all amps sound the same will be in for a rude but happy awakening when they compare an inexpensive amp to an amp like the Benchmark.

Agreed, but I am still wondering who would say such a thing (that all amps sound the same..)? I know there are no shortage of people who claimed that "amps sound the same.. but there seemed to be always a "but" in such claims. That is, with qualifications, caveats etc., not a blanket statement. Clearly someone may think amps like your Parasound and AHB2 would sound the same under some conditions, but I doubt anyone would think a $60 tube amp and the AHB2 would sound the same even when used within their the cheap amp's limit.

 

[josh358]

Agreed, but I am still wondering who would say such a thing (that all amps sound the same..)? I know there are no shortage of people who claimed that "amps sound the same.. but there seemed to be always a "but" in such claims. That is, with qualifications, caveats etc., not a blanket statement. Clearly someone may think amps like your Parasound and AHB2 would sound the same under some conditions, but I doubt anyone would think a $60 tube amp and the AHB2 would sound the same even when used within their the cheap amp's limit.

I think that there usually is a qualification, like that guy who had the amplifier challenge said that the amps had to be within their linear range and any response differences have to be equalized out. Which is a bit like saying "They don't sound the same, but we can tweak them until they do," LOL. Though I'm not convinced that FR is all there is to it.

One guy hypothesized a few years back that subjective differences in sound quality were due to a very minor overall tilt in the response curve. Not sure if anyone ever followed up on it.

What I do find interesting is that I can identify differences between the AHB2 and the Benchmark in blind testing that don't seem to have anything to do with power or frequency response. But then, John Siau once told me had ABX'd crossover notch distortion, and the A21 is high bias, but still a class AB amp. John Curl says that you can improve the A21's sound by biasing it higher and adding some fans, LOL.

 

[tmtomh]

Always a good story to recall:

Carver Challenge | BobCarverCorp.com

www.bobcarvercorp.com

(It depends on the presumption that Stereophile's editors could reliably tell amps apart to begin with, but still, I think an instructive story nonetheless.)

 

[Laserjock]

Great insight, I’m actually looking at this Benchmark amp, Parasound, and Rotel for a 2-channel amp for a future Spatial Audio Sapphire M3 speaker purchase. Is there particular music styles you preferred with the Benchmark and Parasound amps? I like 70’s and 80’s rock and pop, electronica, chill out (like Enya, Clannad, B-Tribe, etc..), and some orchestral classical.

Someone in the PA5 thread talking about speakers similar to yours. Might go check it out. Can’t remember the name

 

[josh358]

Great insight, I’m actually looking at this Benchmark amp, Parasound, and Rotel for a 2-channel amp for a future Spatial Audio Sapphire M3 speaker purchase. Is there particular music styles you preferred with the Benchmark and Parasound amps? I like 70’s and 80’s rock and pop, electronica, chill out (like Enya, Clannad, B-Tribe, etc..), and some orchestral classical.

I think the differences I heard were much more apparent on classical than on rock, mostly because classical has a reference (live music) that you can compare it to, while rock is heavily processed. But if you wanted to play rock at high levels and didn't have efficient speakers, the Parasound might be a better choice since it's a bigger amp (you can always bridge the Benchmark).

Essentially, these differences are so subtle that I'd boil them down to price, how much power you need, and whether you want a behemoth or something that's small, easy to move, and doesn't make glaciers melt. If it were up to me, I'd prefer the Benchmark to any old style amp. Wish Benchmark made a more powerful model, I'm already biamping and four Benchmarks would cost almost as much as my dental bills for the week.

 

[pogo]

Though I'm not convinced that FR is all there is to it.

The FR is measured in the quasi steady state. What is dominant in music, however, are rapidly changing transients. So you would have to generate a complex test signal and monitor the FR over a certain time range at a real load (lsp) to see the swing-in/out behavior.

 

[peng]

The FR is measured in the quasi steady state. What is dominant in music, however, are rapidly changing transients. So you would have to generate a complex test signal and monitor the FR over a certain time range at a real load (lsp) to see the swing-in/out behavior.

That's often been said but really overrated. Complex or not, the amplifier would see just one value at any moment, one value at a time.. That is a concept that while obvious, I never thought of it until I read one of Elliot Sound's article. Again I am not saying testing with just a single sine wave is all that is needed, just saying the point about music's complex waveform is often overrated. We are actually lucky that we have ASR, Amir at least would measure just about whatever he could think of that his AP can measure, including the 32 tone test even if he might not believe it was important to do.

 

[KSTR]

Frequency reponse (of magnitude, additionally of phase as well) can be measured in many many ways with pretty much any signal (including music). Only requirement is that all frequencies of interest must contained somehow (that is, some minimum frequency density) at sufficient levels (vs. noise levels) for robustness.

Using different methods is often used to double-check if a system is linear and/or time invariant. Simple example: measure a tweeter's response at higher powers and with long sweeps, then the FR changes with the orientation of the sweep as power compression reduces the level after some time.

 

[pogo]

Complex or not, the amplifier would see just one value at any moment, one value at a time.. That is a concept that while obvious, I never thought of it until I read one of Elliot Sound's article.

I'm fully with you on that. But which discrete values are seen here? And this is exactly the grey area, which is very complex and cannot be answered with a voltage divider assuming a constant voice coil resistance of the loudspeaker and certainly not with transients. And the different transducer concepts will also have a great influence on this.

The short paragraph on '7.4.3 Damping factor' is sufficient to understand how complex the theory can be and I think this one is closer to reality:
Link