Hypex NCx500 Class D Amplifier Review

[BR52]

If the SINAD leaderboard is indeed intended for just finished products then yes I agree, please do remove all modules on there. including our NC400, the Purifi module. Just all. We don't make the SINAD leaderboard. that is up to Amir. We don't care. We certainly do not manipulate or try to distort your discusions. We do not appreciate such comments and will prevent us participating such forums. So be carefull were you point your finger to. thanks.

To be clear: as soon as we release a new module, OEM or DIY, you readers and contributors on this forum immediately request test results from Amir. As we have nothing to hide we are more the happy to issue samples to Amir for this purpose. But it seems completely irrelevant to most of you that it is or isn't available for DIY customers. but then complaining the use of the module is irrelevant to normal use case. very strange.

Please, we should not forget with buffer the module is really good specially with the high power it delivers.
And we can not stay at class A, AB, ... , in these times anymore. To be energy efficient in all fields of life is necessary. A couple of years ago, a most experts told us it's not possible to have good amplification in class D for example. Hypex, Purify and other drove the progress. Now we see incredible well build amplifier modules.
For sure, where is always headroom for the better.
The little unlucky timing (unbuffered before bufferd) of the presentation is masking the good result with this module.
And discussions how we measure and what is important for interpreting the results, is part of our "science" approach here.
Hypex many thanks for sending your product. Hopefully you do it in future as well.

 

[jsm]

• The product is not commercially available and cannot be purchased as is, right now. It is NOT "production gear"*
• The assembler has no plans to sell the item as tested.
• It is not a complete product. It is a few PCBs in a case with no markings, approvals, warranty, specifications or price.
• It is a test mule for a new Hypex OEM module. A module that can only be purchased as part of a product that Amir did not test.

Remember it has been made clear many times in the past...
View attachment 257428

* Not my term. (see above)

Amir is doing with what he is receiving, will all manufacturers constructing this kind of refence design send final product to Amir? Have we to be deprive of these datas ?
I don't think so, we all known that there will be 2 kinds of use of theses OEM parts: "simple as it", 99% similar to these prototype-like (but just because it is not assembled by Hypex), this review is for them. And other bigger manufacturers using maybe different PSU and buffer stage, these products cannot be assimilated by default to the performances measured here.
It is more or less the same situation with GPU, for the same GPU core, there is a reference card made by the constructor himself (hypex is not an assembler, so a kind ASR member act as it), some almost simple "rebranded" models with light modifications (Apollon/Rouge audio/Audiophonics etc) and heavily modified version with fancy cooling stuff, higher memory frequency...(NAD etc.)
If we have to cut an hair in 4, these module ref is commercially available, but we are not the customers, it's on a B to B market (so all the other point are checked from these point of view). I just hope that if a manufacturer is using ASR as review source for the amp board choice of a futur product or just for a watch, he is also a very generous donator!

 

[Rick Sykora]

Look it is what it is.

I appreciate the sneak-peek review, the discussion and the robust discourse. Perhaps I am a purist when it comes to things like this. But it isn't my website and Amir does a pretty decent job all in all balancing the flak from each direction.

Anyway, it'd be boring if everyone agreed all the time.

Hi John,

Not sure it will satisfy the purist in you but wanted to clarify one aspect here. Just because this specific configuration does not fit my business model, I would still build one if a customer wanted. As some on the forum already know, I have told them when I think there are better values available than I can offer. I do warranty my work and qualification test it before I ship. So, other than volume, my business is very comparable to other small companies using the same tech.

If a customer wanted me to build something other than a Class D amp, my approach would be the same. As you stated, I think Amir is doing an admirable job of balancing what he is willing to test. Even with DIY speakers, he was clear from the beginning that he was not willing to test just any speaker. I think there has been some fun and useful value in testing DIY speakers, seems only fair to screen electronics in a comparable way. So, if I can support it, I will.

Respectfully,

Rick

 

[restorer-john]

I think there has been some fun and useful value in testing DIY speakers, seems only fair to screen electronics in a comparable way. So, if I can support it, I will.

I concede you have a good point there.

Cheers.

John

 

[Stoutblock]

Perhaps we need a seperate SINAD scorecard to compare unbuffered class D modules?

 

[HarmonicTHD]

Perhaps we need a seperate SINAD scorecard to compare unbuffered class D modules?

No we don’t. A clear label is all that is needed IMHO. Amir already labeled the buffered amp accordingly. For the unbuffered amp maybe the label could be changed to “NCx500 unbuffered”. And the Benchmark label added with “low gain” or something similar. (@amirm)

 

[sarumbear]

Still good numbers but dropped from "superb" to "excellent" That is the result of 10dB increase in noise.

I think I will keep my AHB2 bridged pair. 185W is enough to drive my Salon2 pair to the levels I listen music at.

 

[JohnnyNG]

I think I will keep my AHB2 bridged pair. 185W is enough to drive my Salon2 pair to the levels I listen music at.

I think my pair of Audiophonics 1ET400A monos is already enough to fry my M106.

 

[Sokel]

I think I will keep my AHB2 bridged pair. 185W is enough to drive my Salon2 pair to the levels I listen music at.

For the same drivers (give or take) Genelec uses a total of 900 watt per speaker.
And that without passive xover.
Just saying

 

[sarumbear]

For the same drivers (give or take) Genelec uses a total of 900 watt per speaker.
And that without passive xover.
Just saying

You cannot add amplifiers that work on separate drivers as each are working on part of the spectrum. Most Genelecs, like 8351B has 250W max. The only one I know that has more power than this amplifier is 8361A at 700W. Even most of their main studio monitors only have 500W.

 

[Sokel]

You cannot add amplifiers that work on separate drivers as each are working on part of the spectrum. Most Genelecs, like 8351B has 250W max. The only one I know that has more power than this amplifier is 8361A at 700W. Even most of their main studio monitors only have 500W.

I 'm talking about it's main monitors series,I consider them closer to a big-ish 3way.

 

[Pulkass]

See the first post which is now updated.

Where can l actually see the full test ? Not present here!

 

[Rick Sykora]

Just for reference, an amp like this one would retail for $500-600. A stereo, single power supply for about $800.

The Benchmark ABH2 is a great amplifier, but it is $3000. Bridged it does a bit more power, but is 6 times more expensive. Not to mention the Hypex is smaller and costs less to operate. When it comes to value, there is a clear difference here. Audibly, am not as sure.

 

[boXem]

Just for reference, an amp like this one would retail for $500-600. A stereo, single power supply for about $800.

Looks like we have different info on the module pricing .

 

[Rick Sykora]

Looks like we have different info on the module pricing .

may be should have added YMMV.

p.s. along inflation, exchange rates, taxes, etc.

 

[HarmonicTHD]

Where can l actually see the full test ? Not present here!

First page of this thread. Since today there is the data for 12dB and 27dB gain (added today).

 

[pma]

No. A wider bandwidth would actually capture all the harmonics and measure them, just as amplifier measuring has done for the last 70 years. 10th harmonic or 200kHz has been the standard forever.

Now, with so much bandwidth-challenged class D gear, which is also full of HF noise, the rules needed to be changed to accommodate their deficiencies. Don't pretend it is anything else.

But it is what it is and I don't lose any sleep over it. But I'm not so naïve as to read any class d 'datasheet' as anything other than painting their product in the best light possible for the ignorant, and glossing over the deficiencies. Seems to be the standard MO for Hypex and others since day one.

Hi John,

I have some new measurements on NC252MP made with measuring BW 90kHz and I think it might be of your interest. It is not as bad as expected. I will stick with image thumbnails, as this is not my thread. I suppose the results would be quite transferrable to other Hypex NC modules.

1. Frequency response - yes, I do use additional 2. order LC filter which starts at 25kHz with -40dB/decade slope.


2. 15kHz/130W/4ohm spectrum


3. THD vs. frequency at 130W and THD/THD+N vs. power at 15kHz

 

[ZgSTar]

It's Rouge (red in French), not Rogue. They are based in Roma, Italy, but like many Italians it seems that they aslo speak French.

French used to be the upperclass language all over Europe for centuries untill +/- 100 years ago (WOI) and in certain Italian (and not only Italian) upper class envirroments it's still like that. A lot of higher class European brands still prefer French names because it's more "chique" (showing of high class) like we call it here... I don't know that is the reason why they carry a French name, but it could be the reason...

Maybe the owner is just read headed.

 

[Vacceo]

Still good numbers but dropped from "superb" to "excellent" That is the result of 10dB increase in noise.

I think I will keep my AHB2 bridged pair. 185W is enough to drive my Salon2 pair to the levels I listen music at.

Amir aims at 115 db as "no extras added, guaranteed" figure, but personally, 100 db for me is fine enough as it is "CD quality guaranteed, no added extras". In the near future, that will be the threshold to renew my electronics and these modules, just like Purifi's, archive it.

 

[pma]

Not sure if anyone picked up on this, but the IMD of the NCx500 is about 20dB less than the benchmark. And as far as audibility is concerned, IMD is the most important factor.

Not the exact same test, but the results captured by the multi tone show -120dB worse case vs -100dB

View attachment 257479
View attachment 257480

Sorry, these 2 tests cannot be compared in any way. Though twin tone CCIF has amplitudes of individual components only -6dB below composite signal amplitude, amplitudes of individual sines in the 32-tone multitone are far below this. The HF nonlinearities, level dependent, are much better tested with the twin tone. Please do not place company agenda into this thread, try to stay objective.
On the other hand, I agree that AHB should be better in the CCIF test.