Hypex nCore vs Class A amps

[Krunok]

You’re right, it doesn’t of course. And being ranked no.1 on ASR will only last as long as it lasts. But that is not crucial to me, and also not what I mean. What I do mean is that probably even you would be amazed to hear how well this “Little Big” amp sounds, and is able to control my rather difficult to drive 3-way ported floorstanders. I haven’t been able to reach this level of bass control at home with several other (owned and loaned), substantially more powerful and well renowned class AB amplifiers, and at the same time the highs and mids leave nothing to be desired. And this same excellent quality is delivered by the Hypex nCores, for me they are in the same league. It underscores the importance of high quality amplification in the chain, and it is great that guys like John Siau, Laurie Fincham and Bruno Putzeys have made this possible for us.

I agree, Benchmark and nCores are state-of-the-art products and I'm certain they provide not only premium SQ but joy of ownership.

 

[maty]

I agree, Benchmark and nCores are state-of-the-art products and I'm certain they provide not only premium SQ but joy of ownership.

I disagree with nCores, they would be at the time in the class D world but not in ALL amplification. And that is evident for a few years, that is why I did not want any for me. I was only interested in them by the SMPS, as an alternative to the large amount of DC that I suffer at mains, but as I learned things about the Hypex SMPS I lost interest, in the PSU and its amplification modules.

 

[boXem]

I disagree with nCores, they would be at the time in the class D world but not in ALL amplification. And that is evident for a few years, that is why I did not want any for me. I was only interested in them by the SMPS, as an alternative to the large amount of DC that I suffer at mains, but as I learned things about the Hypex SMPS I lost interest, in the PSU and its amplification modules.

What's wrong with nCore?

 

[g29]

...
I also compared the ncore to the Benchmark. I can’t tell you how glad I was to return the Benchmark and get my money back.
....

What did you find different between the Benchmark and the NC400 ?

TIA

 

[maty]

What's wrong with nCore?

* Like ALL class D of which we have measurements: limited bandwidth. Same restriction with PURIFI and BOSC? I want at least 100 kHz at -3 dB in home audio.

* NCore modules with integrated SMPS have low PSRR and CMRR values

* The phase shift. In class A and AB the phase shift is less than 6º at 20 kHz. The new PURIFI only 1º at 20 kHz. BOSC even better, starts at 30 kHz.

One thing is that the phase shift is a fixed value and another that it increases with frequency, there is the big problem, the tweeter. If the above were not important, they would not have improved it!

That is why Benchmark AHB2 uses it in laboratories to measure / design speakers. I guess he does not have that problem, which is exclusive to class D. But idea with the "old" class D amp like Ncore or worse still with others.


PS: Always thinking of music, excellent recordings with great dynamic range and acoustic and / or electrified instrumentation, without voices adulterated with Autotune. Without synthesized instruments or that almost all of them are. With modern recordings the problem of the phase will not matter, as with the movies. With such adulterated recordings and poor quality (not always in the digital master but almost always in the audio for sale or now in streaming, modern commercial recordings) it will matter little. Why spend a lot of money to listen to those recordings of low sound quality?

 

[boXem]

* Like ALL class D of which we have measurements: limited bandwidth. Same restriction with PURIFI and BOSC? I want at least 100 kHz at -3 dB in home audio.

Why? The more power bandwidth, the more the out of band noise is amplified. Tweeters may not like.

* NCore modules with integrated SMPS have low PSRR and CMRR values

Not related to nCore but to one implementation. Anyhow, as an example, NC122MP has 71 dB of CMRR, which is not bad at all, PSRR is not specified.

* The phase shift. In class A and AB the phase shift is less than 6º at 20 kHz. The new PURIFI only 1º at 20 kHz. BOSC even better, starts at 30 kHz.

SInce we like data here, I made the calculation for you: considering nCore has a first order low pass behavior with -3dB @ 50 kHz, this gives a phase of -22º @ 20 kHz.

One thing is that the phase shift is a fixed value and another that it increases with frequency, there is the big problem, the tweeter.

I can't think about a physical system with constant phase shift over frequency.
- Any system has a reaction time.
- Phase is the expression of the delay between the input and the output.
So the higher the frequency the more phase shift you have with a basic reaction time. As an illustration, -22º @ 20 kHz is 3us.

Anyhow, we humans don't hear phase shift when it's the same between channels. And I doubt we would be able to perceive 3us.
What we hear is group delay. Group delay is a consequence of a discontinuity in the phase shift. If I remember well, we are able to perceive group delays above 2ms, i.e. we can perceive them only in the bass region.

If the above were not important, they would not have improved it!

What they have done is to give a low pass second order behavior to their amplifier. One of the consequences is lower phase shift at audible frequencies, but I suspect that the real reasons behind this are completely different.

That is why Benchmark AHB2 uses it in laboratories to measure / design speakers. I guess he does not have that problem, which is exclusive to class D. But idea with the "old" class D amp like Ncore or worse still with others.

Sorry, not following you.

 

[Krunok]

Anyhow, we humans don't hear phase shift when it's the same between channels. And I doubt we would be able to perceive 3us.
What we hear is group delay. Group delay is a consequence of a discontinuity in the phase shift. If I remember well, we are able to perceive group delays above 2ms, i.e. we can perceive them only in the bass region.

Related to GD things are not really that simple - you may find this measurement interesting. Reflections screw up everyhing in time domain, GD included, and yet it somehow doesn't seem to matter.

 

[boXem]

Indeed!

 

[maty]

That is why Class D limits bandwidth so much. Presumably, when technology allows it, the bandwidth will also rise.
At that time they will also have to deal with the usual problems of fast amplifiers, limit it to 350 kHz so as not to have problems with RF.

You are right about PSRR. But it probably is lower than the NC400 (Typ: 80 dB) but I am not sure. Better does not credible.

NC400 has 16º at 20 kHz, Amirm measurements. A wonder compared to the most modern IcePower 1200AS2.

So why do many manufacturers strive for the correct integration of the tweeter with the woofer in time? In the loudspeakers design phase oscillates progressively between +/- 45º. If the phase of the amplifier varies we will have a new variable unlike if it were constant. How much phase shift is acceptable? I do not know, but at least we must demand the same as in other technologies.

In a laboratory you want an amplifier that is as neutral as possible -amp as a wire with gain- so that it does not affect the measurements. In the phase, so far, class D does not meet that requirement.

 

[Krunok]

That is why Class D limits bandwidth so much. Presumably, when technology allows it, the bandwidth will also rise.

LOL Maty strikes this topic as well!

Sure M8, as it is a known fact that our hearing goes up to 80 kHz and tweeters usually play linear up to 90dB I say we need amps that can at least reach up to 100kHz (-3dB)!

 

[maty]

Class AB

https://www.diyaudio.com/forums/aksa/294346-brand-super-aksa-saksa-85-a-post4878481.html

by Hugh Dean (AKSA)

[ 6. Many musical qualities are not easily measured. I try to aim at phase shift of no higher than 3 degrees at 20KHz, and ensure that all devices are fast and working with optimal compensation. This forces you to rely on subjective listening, both with the designer and a sample of selected audiophiles who are willing to do the long time listening tests. You do need a lot of tests and measurements regardless; amps must be engineered so that at clip they behave, deliver a higher S/N ratio, that at low levels they sound clear, and with complex music they deliver clean sound with no 'smearing'. ]

 

[maty]

https://orchardaudio.com/bosc

Frequency Response: DC - 60kHz+

PURIFI 1ET400A: 60 kHz+ at -3dB

Hypex NC500: 60 kHz+ at -3dB

 

[boXem]

That is why Class D limits bandwidth so much. Presumably, when technology allows it, the bandwidth will also rise.
At that time they will also have to deal with the usual problems of fast amplifiers, limit it to 350 kHz so as not to have problems with RF.

You didn't get me. I am not writing about the out of band noise created by the amplifier but about the out of band noise at it's input. Think NOS DACs without reconstruction filter as an example. Imagine the mess is you start amplifying their crap until 350 kHz.

So why do many manufacturers strive for the correct integration of the tweeter with the woofer in time? In the loudspeakers design phase oscillates progressively between +/- 45º. If the phase of the amplifier varies we will have a new variable unlike if it were constant. How much phase shift is acceptable? I do not know, but at least we must demand the same as in other technologies.

These are different things. Transducers being out of phase create irregularities in the overall response that need to be addressed.
For passive speakers, you have one amplifier at the input, whatever it's phase is will not change anything to the phase of each transducer, the speaker response will not be affected.
For active speakers, if you start playing with different types of amplifiers for each transducer, the phase shift may add problems. If all amplifiers have the same response, you don't care anymore. And anyhow, you certainly have a DSP, just use it.[/QUOTE]

In a laboratory you want an amplifier that is as neutral as possible -amp as a wire with gain- so that it does not affect the measurements. In the phase, so far, class D does not meet that requirement.

First I am living in a house not a laboratory.
Second several class D amplifiers are now reaching the marketing gimmick -amp as a wire with gain-. Not sure this is the case with a lot of class A.
Last, but not least, class D amplifiers are the only ones able to convert more energy into music than into heat.

 

[maty]

I have not write about class A. I only say that class D must meet the minimum requirements of other technologies. I do not write about consumption, heat or ...

Yes, one key is the DSP... well implemented.

By the way, why does KEF LS50W use class D on the woofer and yet class AB on the tweeter? It has a DSP if I am not wrong.

PDF very interesting about this amplifier, with a lot of pictures: https://fccid.io/UXD16001/Internal-Photos/Internal-Photos-3130815.html

 

[Deleted member 2944]

I only say that class D must meet the minimum requirements of other technologies.

Why is that??

It seems to me that "other technologies" seldom meet their own minimum requirements. (Numerous examples tested here on ASR.)

Dave.

 

[boXem]

I have not ruled on class A. I only say that class D must meet the minimum requirements of other technologies. I do not write about consumption, heat or ...

In my eyes, consumption is a minimum requirement since 1974 (45 years ago...)

By the way, why does KEF LS50W use class D on the woofer and yet class AB on the tweeter? It has a DSP if I am not wrong.

Because they made an engineering choice leading to this decision. I guess that BOM cost and engineering budget were major factors.
Their class D amplifier has certainly the filter outside of the loop, leading to poor distortion figures in the highs. So they use it for bass, and take a class AB for the tweeter which doesn't need a lot of energy. And they use DSP to deal with potential phase problems.
I have my doubts that the KEF LS50W amplification stage meets a lot of your exigences (which doesn't mean it's not good at what is does). Why are you taking it as an example?

 

[scott wurcer]

Class AB

https://www.diyaudio.com/forums/aksa/294346-brand-super-aksa-saksa-85-a-post4878481.html

by Hugh Dean (AKSA)

IIRC Mr. Dean also believes in scalar waves.

 

[maty]

Because in the end it is always the same, the best or worst implementation. In the cheapest KEF LSX, 4", amplification is only class D. With DSP too I think (I am not sure).

I do not recommend neither, it is simply a good observation.

PS: Other technologies = class A and class AB

 

[maty]

@scott wurcer

Easy question: how much phase shift?


TI - HiFi Audio Circuit Design by Wayne Liu, August 2017

[PDF] http://www.ti.com/lit/an/sboa237/sboa237.pdf

The capacitors must be large enough to maintain stability but not introduce too much phase shift and amplitude attenuation at audio frequency. With a 620-pF capacitor, the fp = 276 kHz, the phase shift at 20 kHz is 8°, and the amplitude attenuation is 0.02 dB; this is a very good balanced solution.

 

[Krunok]

Are you guys aware that having such discussions with Maty is as futile as resisting the Borg?