Benchmark AHB2 / Class D Purifi Eigentakt / Mark Levinson 333 - Listening impression & Conclusion

[tvrgeek]

Pretty much confirms what I thought. The Benchmark is, well a benchmark. Think twice before spending more. Think hard about spending less.
I would like to see it compared to a March and Vidar. How good can cheap get? As slim as the difference between a JDS DAC and a Gustard?

When folks here talk about dynamics of the 1812, I can only assume they have suitable speakers and hopefully some subs. I know it took both my 12's in the last house. With one here, my system does sound weaker. Not the amps fault.

 

[tvrgeek]

This would be a good opportunity for some in-room high quality recordings at the listening position in order to show some objective data. The relative differences between the amps should be resolvable by a high quality recorder/mic that has more sensitivity than human hearing.

And on what system would you be reproducing those recordings on? Old rule of thumb, your measurement equipment needs to be at least 10 times better than your measurement. Speakers 10 times better? I wish. 802's, well in my personal highly biased opinion, that elevated harsh top end should amplify any amplifier fault. Roll them back down to a realistic EQ and the differences in the amps may be much smaller.

 

[pogo]

How good can cheap get?

You might also be interested in this with a different output. The speakers used here are in a different league.
Link

 

[tvrgeek]

No speakers are good enough.

 

[pogo]

No speakers are good enough.

That is true, but probably a smaller bottleneck in one place or another.

 

[Gorgonzola]

...
Looking at the power x THD+N graphs in 4 ohms of each amp, on the knee before shooting up, we have:
Mark Levinson No.333: 650 W
Benchmark AHB2 (mono): 500 W
Purifi EVAL-1: 250 W

Also Purifi has the lowest output impedance and lowest distortion (IMD) on the treble. This probably correlates with the perceived rolled off treble.

I feel that your remark about Purifi's low IMD in the treble could be right on the money. It wouldn't be the first time that listeners have confused a slightly raunchier top treble with less roll off and/or more detail.

I haven't heard the EVAL-1 buffer. I have a VTV stereo Purifi that I've heard with both the Hypex buffer and VTV's own buffer with multiple op amps. The Hypex I found unlistenable on account of the high treble which sounded harsh and unnatural. With the VTV buffer I found the Sparkos SS3602 op amps to be nicest vs. OPA627, OPA1612, Burson V6 Vivid, and a couple of others. What I'm saying is that the choice of buffer and op amp with your Purifi will make a difference.

Disclaimer: the above are just my impressions -- take 'em or leave 'em.

 

[audio2design]

I have a PhD in Neurosciences from MIT/Harvard, I think I know better what science is than you. You are a cultist.

I will take things that didn't happen for $200 Alex.

 

[Jim Matthews]

I will take things that didn't happen for $200 Alex.

Oh no, it's *all true* there's even pictures of Gueco with a young Ian Malcomb to prove it...(the bowtie is an MIT thang)

 

[audio2design]

Oh no, it's *all true* there's even pictures of Gueco with a young Ian Malcomb to prove it...(the bowtie is an MIT thang)View attachment 176375

Is it wrong that I think this is one of the best movies ever made?

 

[tvrgeek]

I feel that your remark about Purifi's low IMD in the treble could be right on the money. It wouldn't be the first time that listeners have confused a slightly raunchier top treble with less roll off and/or more detail.

I haven't heard the EVAL-1 buffer. I have a VTV stereo Purifi that I've heard with both the Hypex buffer and VTV's own buffer with multiple op amps. The Hypex I found unlistenable on account of the high treble which sounded harsh and unnatural. With the VTV buffer I found the Sparkos SS3602 op amps to be nicest vs. OPA627, OPA1612, Burson V6 Vivid, and a couple of others. What I'm saying is that the choice of buffer and op amp with your Purifi will make a difference.

Disclaimer: the above are just my impressions -- take 'em or leave 'em.

I agree the buffers matter, one also must consider EQ as speakers react differently. Where an amp may be a bit bright on one speaker, it may not on another. To be fair, you need to correct in-room eq for each. Many, if not most, modern speakers are voiced way too bright. I have been told this is for the Asian market which is larger and more profitable than the Western market. Just one more thing to understand different in how the amp behaves with the load, differently from AB. Differently from a transformer. Not better, not worse, different.

 

[audio2design]

Statements like this, right at home on Audiogon, do not represent someone strong in the sciences, though we have seen some shocking stuff written by PhD's in the last few years w.r.t. audio, so maybe we should not be surprised. My favourite was the "dude" who did not understand the most fundamental aspects of sampling theory going on and on about microtiming and other made up terms w.r.t. audio .... hmmmm... related? Even 3rd year engineering student understood the fundamental science far better than this "post-grad" did.

https://www.fairobserver.com/more/s...ital-audio-science-news-william-softky-39078/

 

[pogo]

Just one more thing to understand different in how the amp behaves with the load

A good example of this:
Link

 

[rdenney]

Read and understand the first two posts of this thread:

Amplifier Output Impedance (Damping Factor) and Speakers

This is a reprint of an article written ca. 2011-2012. I've been meaning to rewrite and update it, but Life and Work keeps getting in the way, so I am reposting the original as-is. Take it for what it's worth, an introductory take on one of the potential causes for differences in sound among...

www.audiosciencereview.com

This will provide an idea of the real relationship between speakers and amps for real-world examples of both.

Given that buffers are just preamplifiers, I don't understand how they would affect the output impedance of the final amplifier that is actually driving the speaker load. It seems to me that if they don't affect the output impedance of the main amp, they can't affect the interaction of the main amp with different speakers.

They could, of course, be colored or suffer from possible faults on their own. Amir had tested Purifi amps with the EVAL1 buffer and with at least a couple of different aftermarket input buffers (the search will find these). The EVAL1 buffer performed best, based on what I see. But "best" means "least distorted and quietest", not "sounds best to me." So YMMV.

Rick "the Hypex and Purifi amps have extremely low output impedance across the audible band" Denney

 

[BrokenEnglishGuy]

So what's the damping factor number from Purifi eval1?

 

[tvrgeek]

Well, I have been building speakers and amps for 40 years, so I rest on my experience.
I am looking at objective measures. Always first. I then listen, but invariably, my ears usually follow the specs until you get to a "good enough" threshold. Loads have HUGE differences on ampler performance. If you have ever done a Spice model of an amp with various realistic loads, you would see it in the FFT big time. Now, before the "good enough" everyone has their choice of what faults are least objectionable.

One of the things I do not understand yet in ClassD is how distortion varies with current. AB is usually far worse with a "4 Ohm nominal" load than an 8. Sometimes by an order of magnitude.

Super low output Z is no holy grail. If you have ever built a speaker you would know it is darn near BS.

 

[Jim Matthews]

Is it wrong that I think this is one of the best movies ever made?

Not at all.

Totally original screenplay. John Lithgow as a possessed Italian quantum physicist, the single greatest dialog ever written:

Perfect Tommy: Emilio Lizardo. Wasn’t he on TV once?
Buckaroo Banzai: You’re thinking of Mr. Wizard.
Reno: Emilio Lizardo is a top scientist, dummkopf.
Perfect Tommy: So was Mr. Wizard.

 

[Jim Matthews]

Well, I have been building speakers and amps for 40 years, so I rest on my experience.
I am looking at objective measures. Always first. I then listen, but invariably, my ears usually follow the specs until you get to a "good enough" threshold.

This is a genuine difference between generations of home music lovers. As a younger man, the only decent stuff I could afford was in kit form.

Sometime around 1990 the quality of mid priced gear dramatically improved. Now that I can barely see to solder, switching power supply amps with onboard DSP sell for less than my parts cost.

Young audiofiends are still building things, but mostly concentrate on code.

 

[MaxBuck]

... switching power supply amps with onboard DSP sell for less than my parts cost.

You must wear pretty nice pants.

 

[tvrgeek]

This is a genuine difference between generations of home music lovers. As a younger man, the only decent stuff I could afford was in kit form.

Sometime around 1990 the quality of mid priced gear dramatically improved. Now that I can barely see to solder, switching power supply amps with onboard DSP sell for less than my parts cost.

Young audiofiends are still building things, but mostly concentrate on code.

Started with hand-me-down Knight kit.
Code has to run on something. The advantage of powered speakers is the engineer knows the load and can adjust the DSP accordingly.

 

[rdenney]

So what's the damping factor number from Purifi eval1?

Purifi provides this in their data sheet: 0.07 milliohms at 1KHz, and <0.65 milliohms across the band, at 1 amp of output.

I'm seeing a minimum damping value of 8/0.00066=>12K. That's only at 16 KHz--most of the damping values will be much higher. Essentially, the output impedance is effectively zero--that's about as load-invariant as it gets, as long as a low-impedance load does not draw more current than the amp's power supply can provide.

The older Hypex amps have a peak output impedance of 3 milliohms (for the NC502MP), for a damping factor of 2.7K or greater.

These are measured at the module, of course, and the wiring to the speaker terminals along with the terminals themselves will dominate the value measured there.

Rick "far higher than needed to make these amps load-invariant" Denney