Benchmark appears to have a new power amp. The blurb claims it to be better than all present day power amplifiers.
AHB2 - Power Amplifier | Benchmark Media
So did anyone get to hear it at RMAF ? How was it ?
Cheers.
AHB2 - Power Amplifier | Benchmark Media
So did anyone get to hear it at RMAF ? How was it ?
Cheers.
No, but it is always handy and fun to lookup the company address on Google Street Maps. These guys occupy Suite 2 in this building. They don't even display their company name. Makes you wonder.
Attachments
Here are THX'es patents, https://www.google.com/search?tbo=p&tbm=pts&hl=en&q=inassignee:"thx+ltd"&num=100 - Most likely are the Low Dissipation Amplifier, and Power Supply ...
The efficient power amplifier was described as a summer project in Elector of 1982 and was also used by several Japanese amplifier companies at the time, so nothing new here. The distortion reducing schemes correlate with publications from Sigfried Linkwitz also from the eighties.
Why not simply patent electronic engineering and nobody can do nothing without paying everyone something.
Why not simply patent electronic engineering and nobody can do nothing without paying everyone something.
This is John Siau, VP and chief engineer at Benchmark
The THX patents are:
8,004,355
8,421,531
The patented topology uses feed-forward error correction to virtually eliminate crossover distortion. The technique is so effective that the output stage can be run in class-B operation while achieving very low distortion. This also allows class-H or class-G tracking rails without the usual distortion problems associated with these designs. THX has demonstrated a design that rivals the efficiency of class D amplifiers.
The THX topology was interesting to Benchmark for an entirely different reason:
The same feed forward error correction can be used to remove crossover distortion from a class-B amplifier can be used with class AB biasing to create an amplifier with extraordinarily low distortion.
Our goal was to achieve very low distortion, with a focus on low crossover distortion. We were willing to increase the power dissipation as much as necessary to optimize the distortion performance. The optimum solution uses some bias current, but much less than a traditional class AB. Distortion of the AHB2 is at the measurement limits of our AP 2722 and 2522 test stations.
The THX topology also allowed us to utilize two power supply rails in a class-H (or G) configuration without any measureable distortion penalty.
The AHB2 is significantly more efficient than a traditional class AB, and has much lower distortion. Power consumption is only 20W idle. A traditional class-AB of equivalent power would consume 100 to 120 W idle. For example, the similarly sized Bryston 3B consumes 120W idle.
Benchmark chose to build the amplifier with very low gain. Gain is only 9 dB instead of the more typical 20 to 30 dB gain found in most power amplifiers. This means that the AHB2 clips with a 22 dBu input instead of an 8.2 dBu (2 V RMS) input. The AHB2 is designed to accept studio-level input levels and this can significantly improve noise performance between connected devices. In my opinion, most power amplifiers have far too much gain.
Our design goals included a 130 dB SNR relative to maximum output. To achieve this performance we used very low impedances, and some very unique PCB layout techniques. We also chose to use a switching power supply so that we could eliminate 60 Hz magnetic components. The magnetic fields produced by a high frequency switching supply are much lower and are out-of-band.
We also chose a tightly regulated power supply. Most power amps use unregulated supplies because this is the most efficient use of resources if switching supplies are not used. Given switching supplies, tight regulation comes almost for free. Better yet, the power supply control loop can be made fast enough to react to all audio frequencies. The AHB2 does not rely on secondary capacitive storage.
We will be posting performance plots and measurements on our website soon.
Member
Joined 2009
Paid Member
Very interesting. From the description I suspect that Vanderkoy and Lipshitz would recognize some of this (i.e. it reminds me of their paper on how a low power small amplifier drives current-dumping Class B amplifier). Or is this something more akin to Graham Maynards GEM amplifier.
Well John, your new amp sounds like it will be very nice; it would be interesting to see a DIY version, perhaps if the patents are adequate protection you'd be willing to volunteer a simplified schematic for us to think about ?
Well John, your new amp sounds like it will be very nice; it would be interesting to see a DIY version, perhaps if the patents are adequate protection you'd be willing to volunteer a simplified schematic for us to think about ?
Last edited:
The noise for Benchmark's AHB-2 is 9uV A-W. But I don't think it is a state of art noise level of an audio power amplifier
I my bench, 8uv without weighting for a 50W 8R Class A Power amplifier is just a common noise level. Every one knows that Class A means higher Johnson noise form resistors/transistors......
I have a ultimate solution for a preamp, 30Vrms balance output with only 3uV noise level, 6dB gain. That means you can get 140dB SNR from such a preamp.
From diyaudio, a topic 'yet another amp ... (I don't really remember the accurate title)' claim that his power amp presents 1ppm distortion for over 200W 8R and full audio frequence range. If the measurement by the poster is true, that will be the most fantastic power amp.
Last edited:
I have made several variations of output stages that operate at very low bias yet show no indication of crossover distortion(or Gm doubling), in effect, they operate class B.
About 8mA for the BJT OPS and about 50mA for vertical mosfet OPS. They are based upon the amplified diode concept that is certainly not new news. It is true that with a higher bias the error signal is smaller, but there is little difference in the output. The output stage Gm is greatly linearized. With modern materials and tighter layouts, the concept works wonders as it did 35+ yrs ago when it was invented. And it creates a dynamic output impedance which IMO greatly enhances the ability to provide to the speaker a more ideal 'voltage source'.
About 8mA for the BJT OPS and about 50mA for vertical mosfet OPS. They are based upon the amplified diode concept that is certainly not new news. It is true that with a higher bias the error signal is smaller, but there is little difference in the output. The output stage Gm is greatly linearized. With modern materials and tighter layouts, the concept works wonders as it did 35+ yrs ago when it was invented. And it creates a dynamic output impedance which IMO greatly enhances the ability to provide to the speaker a more ideal 'voltage source'.
low gain is the only way to get low noise from the technology we have 21st century. I agree with you.
I am very interest in the performance measurement, especially 20kHz 100W 8R distortion. For 1kHz seems so simple to get -120dB THD now days.
Best regards
Paul
Last edited:
Every time I hear someone say "did you hear that amp what did you think of it" I want to PUKE.
Correct me if I am wrong here, but you are perhaps hearing A LOUDSPEAKER connected to the amp? AND how many times do we hear that such and such loudspeaker sounded like CRAP at the big ol audio show (CES, RMAF, etc., etc.). held in a hotel room...
So, did anyone hear the amp? NO!
Your wrong.
different amplifiers DO sound different.
try it sometime, you might be suprised....
That's a tell-tale sign. Rather than just avoiding/ignoring people who talk about amp sound, you get annoyed by them?
Well you did invite correction so yes, you are wrong. People do indeed hear amps through speakers. Speakers distort for sure (but that's largely inaudible being low order), but amps add noise. So when a person hears an amp they in general hear what its power supply is adding to the original input. That's assuming the amp's low-level IMD isn't intrusive, which it is on some less-well-engineered devices.