Benchmark...first watt....ABX...facepalm!

[Blumlein 88]

https://benchmarkmedia.com/blogs/ap...il&utm_term=0_7c8c792ee5-88d1ec426b-108067001

I rather hate to see this. Benchmark stooping to audiophile press tricks to promote their product. Maybe the cost of doing business. Still this paper is rather over the top to promote their undoubtedly fine amplifier. It simply is rather deceptive in presentation.

Show an AB amp with 1 watt distortion as 70 db, then show it has 73 db referenced to 1 watt when used at .01 watt. Well that means it has -53 db distortion which isn't great performance. I would think it not typical. Plenty of amps for a long time have -80 or -90 distortion at 1 watt, and also at lower levels. So they cherry-picked a poor amp for comparison. Benchmark's own amp is apparently very low distortion. I guess saying a good AB amp has inaudible distortion and Benchmark's is lower still doesn't have the same selling ring to it.

Several other parts are deceptively phrased I won't bother to go over each one.

Anyway, thumbs down on Benchmark for this one.

 

[amirm]

Pretty strange stuff: " Using a double-blind ABX test, we verified that there was a clearly audible difference when the amplifiers drove speakers at an output level of 0.01 watt."

0.01 watt??? That is not first watt, but first 0.01 watt.

 

[Cosmik]

They're just doing what this, and other forums, demand: demonstrate that any claim of superior performance is "audible". Personally, I'd take their word for it if they argued that their amp was superior in terms of design, confirmed with objective measurements, but this would cut no ice for someone like AJ, for example. If it hasn't been demonstrated with an A/B blind test then it doesn't exist as far as he is concerned - and he is not alone.

 

[amirm]

Oh, it goes from bad to worse. It has this picture:

"The class-AB amplifier used in our tests includes a traditional negative feedback system which is designed to correct the distortion caused by the output stage."

That is not a class AB amplifier schematic. It is class B which nobody uses for the reason they show on the graph. No engineer would ever write such stuff.

 

[Blumlein 88]

Oh, it goes from bad to worse. It has this picture:

"The class-AB amplifier used in our tests includes a traditional negative feedback system which is designed to correct the distortion caused by the output stage."

That is not a class AB amplifier schematic. It is class B which nobody uses for the reason they show on the graph. No engineer would ever write such stuff.

Yes you are right. I saw that, but didn't mention it. There are other things too. Like implying that because distortion is above 0 db SPL it will by definition be audible. While in fact with masking it might well not be audible. The whole thing was written to sound deceptively convincing to someone who doesn't really know what is going on.

Maybe they should do a comparison with the more well known feedforward amp which is the old Quad.

 

[NorthSky]

"Dick Olsher once said that "the first Watt is the most important Watt". In my opinion, "the first 10 milliwatts are the most important"!"
_______

You guys know much more than I do. But for many years I always looked @ the very lowest range of an amplifier power for the lowest distortion possible.
From my readings it's an important attribute, one of.

 

[Speedskater]

Who is 'AJ' ?

 

[DonH56]

You guys know much more than I do. But for many years I always looked @ the very lowest range of an amplifier power for the lowest distortion possible. From my readings it's an important attribute, one of.

The problem with this is that most audio (and other) test equipment actually measures signal-to-noise-and-distortion (SINAD) and does not really separate noise from distortion in the power sweeps. As a a result you typically see a rising THD level at very low powers when in fact what is happening is the signal is becoming so close to the noise floor that the SINAD ratio is dropping. Distortion is not really rising, rather the signal is becoming buried in the noise, and the graph makes it look like distortion is rising at low power levels.

IME/IMO - Don

 

[Sal1950]

Who is 'AJ' ?

Speaker designer from Central FL and prior member of this forum.
http://www.soundfieldaudio.net/

 

[amirm]

FYI I post a video on differences between class A and class AB amplifiers here: http://www.audiosciencereview.com/f...standing-class-b-and-class-ab-amplifiers.768/

 

[tomelex]

Sorry guys, I do not see anything wrong with benchmarks paper. Yes, they picked an amp with comparable power, a competitors amp, and they showed that it has crossover distortion. Some. Not huge, but some.
They showed that when they dropped the input level by 20 db the resulting distortion waveform remained the same and the distortion level only dropped by 3 db vs the large 20db drop in input. Then they say the feedback mechanism in that amp is not effectively reducing the crossover distortion (it wouldn't do much at that low level any way since it comes from not enough standing bias and noise as Don mentioned too).

They showed that by using feedforward they could reduce their crossover distortion to very low levels. They do have a superbly engineered amplifier based on what they are showing us.

I know Amir judged their picture, but they said it was a simplified picture, which it is. If you bias classs ab enough into class A at low signal levels you can get rid of nearly all that low level crossover distortion, the goal is to reduce charge storage effects in transistors when you get to that low a level.

Northstar is correct that any non class A amp is most stressed at ultra small signals and Benchmark showed how they solved that problem, they needed to show the problem with a competitors amp and they showed how they solved it. They better engineered their amp, simply put. From what I have seen presented by them, they have a superb amp and should be button poppin proud of it and their white papers are remarkably good and informative.

Benchmark rocks dudes! Can we say VALUE for MONEY!

I would sort of consider them the "new" Hafler.

 

[Blumlein 88]

Sorry guys, I do not see anything wrong with benchmarks paper. Yes, they picked an amp with comparable power, a competitors amp, and they showed that it has crossover distortion. Some. Not huge, but some.
They showed that when they dropped the input level by 20 db the resulting distortion waveform remained the same and the distortion level only dropped by 3 db vs the large 20db drop in input. Then they say the feedback mechanism in that amp is not effectively reducing the crossover distortion (it wouldn't do much at that low level any way since it comes from not enough standing bias and noise as Don mentioned too).

They showed that by using feedforward they could reduce their crossover distortion to very low levels. They do have a superbly engineered amplifier based on what they are showing us.

I know Amir judged their picture, but they said it was a simplified picture, which it is. If you bias classs ab enough into class A at low signal levels you can get rid of nearly all that low level crossover distortion, the goal is to reduce charge storage effects in transistors when you get to that low a level.

Northstar is correct that any non class A amp is most stressed at ultra small signals and Benchmark showed how they solved that problem, they needed to show the problem with a competitors amp and they showed how they solved it. They better engineered their amp, simply put. From what I have seen presented by them, they have a superb amp and should be button poppin proud of it and their white papers are remarkably good and informative.

Benchmark rocks dudes! Can we say VALUE for MONEY!

I would sort of consider them the "new" Hafler.

We don't know the amp. We do know they had to search around and find one because few modern good amps would have the low level distortion as the one in their paper. You can go back and find Adcom amps from 25 years ago that had low enough distortion at low levels it was no issue. Benchmarks amps are excellent. Their distortion and noise is better than other very good amps. But if theirs and others both have inaudible low level distortion then it seems deceptive to claim this will make their amp sound better.

 

[tomelex]

I appreciate what you say, but to be pedantic for their benefit, they did prove the crossover distortion was audible, by itself, at midband frequency. Also, that minute crossover distortion will introduce lots of minute IM distortions which taken as a whole could be audible in low sound level parts of music. I am not sure a good adcom amp or hafler was ever measured at 0,01 watts back then, and I suspect it would be worse or comparable to the amp they showed. In fact, I suspect any class AB amp today comparable to their price point would fail to compete at 0.01 watts compared to theirs, simply stunning measurement IMO.

 

[RayDunzl]

I've feel like I've seen a distortion plot like that shown in a past Stereophile review, however, when I looked I couldn't find one.

So, by improving my search query having just written that, here's a couple of notchy examples. They're out there, to be cherry-picked, I suppose.

http://www.stereophile.com/content/herron-audio-m150-monoblock-power-amplifier-measurements-part-2

http://www.stereophile.com/content/class233-cap-100-integrated-amplifier-measurements

 

[Cosmik]

This is interesting from 1978:
http://www.keith-snook.info/wireless-world-magazine/Wireless-World-1978/Current dumping - does it really work DCD.pdf

It is worth remarking that for outputs less than about 200mVrms the notches disappear completely, as the dumpers remain off and the amplifier operates in pure class A. Hence at very low power, crossover distortion is totally absent.

 

[tomelex]

Class AB is always a compromise (when compared to class A) whether you are current dumping and or feedforward, etc. Now, virtually all of these plots show these distortions when driving a resistor, they will look much worse when driving a reactive speaker at such super low levels, thats also a reason the Benchmark is so superb, driving a speaker, at least as far as their white paper abx test, showed their unit sounded better, another Kudo to them and their circuit design.

There are other tests, for example cycling of the drive with high power then low power, on a conventional class AB amp, when you do this, at 0.01 you will see more crossover distortion as the transistor parameters change with heat, and the conventional circuits can not correct for this etc blah blah. However, I should note that even class A needs to stabilize for some time to alleviate this effect (tube or transistor) but once stabilized, no matter if high and low cycling the final output is still near perfect. As the class A is first turned on, the tube or transistor general heating is taking place, and thus even though the bias is solid, the heating of the transistor parameters or the tube elements requires a settling in time.

Class AB Semiconductor amps, no matter what kind , need feedback to stabilize them and drive down distortion from their semiconductor inherent double digits, and Benchmark has created the right kinds to do the job right. Again, real value for money there, and USA made.

 

[Cosmik]

Class AB is always a compromise (when compared to class A) whether you are current dumping and or feedforward, etc. Now, virtually all of these plots show these distortions when driving a resistor, they will look much worse when driving a reactive speaker at such super low levels, thats also a reason the Benchmark is so superb, driving a speaker, at least as far as their white paper abx test, showed their unit sounded better, another Kudo to them and their circuit design.

There are other tests, for example cycling of the drive with high power then low power, on a conventional class AB amp, when you do this, at 0.01 you will see more crossover distortion as the transistor parameters change with heat, and the conventional circuits can not correct for this etc blah blah. However, I should note that even class A needs to stabilize for some time to alleviate this effect (tube or transistor) but once stabilized, no matter if high and low cycling the final output is still near perfect. As the class A is first turned on, the tube or transistor general heating is taking place, and thus even though the bias is solid, the heating of the transistor parameters or the tube elements requires a settling in time.

Class AB Semiconductor amps, no matter what kind , need feedback to stabilize them and drive down distortion from their semiconductor inherent double digits, and Benchmark has created the right kinds to do the job right. Again, real value for money there, and USA made.

Performance-wise, how does it compare to a Hypex NCore?

 

[tomelex]

I am a little bit old school and so do not have any practical experience with Hypex, however, I have no doubt that Hypex type amps are the future for high power and medium power. Very low power stuff will remain the domain of class a imo.

Speaking of amp class, there is also AG and AH.. These amps (don't know if anyone makes one) are class A up to several watts or even tens of watts, then if the input signal is higher their power supplies output higher voltage during that time, so if done right, you get "relative" efficiency and also capability to reach huge peak output when needed. So you get advantages of Class D and class A both.

 

[DonH56]

Class A is not a complete panacea. A Class A amplifier can go unstable and misbehave quite badly if ever its current capacity is exceeded by a combination of load and signal. Class AB typically draws until the voltage rails sag and usually handle transient spikes better since the output current is not (or less) controlled the way pure Class A design are biased. I suspect the "greening" of marketing has resulted in much lower biasing (quiescent bias current, "less Class-A") than years ago so AB designs draw less power, and require more feedback, to control crossover distortion. Like many things, however, a nasty-looking distortion on a plot scaled 100x~1000x, has much less impact on the actual sound we are hearing.

All IME/IMO/FWIWFM/my 0.000001 cent (microcent)/YMMV/etc. - Don

 

[Speedskater]

In the first link, down at the bottom of the page is a graph. It's labeled:
"THD+N vs Power - Relative to 1 W"

I wonder what the "Relative to 1 W" part means?
I also don't recall ever seeing a graph where the noise doesn't increase as the power decreases.

***************************************
Here is a very nice DIY chip amp kit graph.