Hypex Ncore

Status
Not open for further replies.
Relatively new to the forums, but just read this whole thread with interest (well, some parts more carefully than others).

Simply amazing numbers on this product--can't wait to own one in any form!

I think you've arrived at the perfect DIY module and am planning to build a stereo version to start, possibly expanding to 5 channels later (or maybe cheat and use UcD modules for the other 3 channels).

As for the OEM modules, I would recommend Emotiva as a possible VAR. They have a great product line--their current crop of class AB amps are among the best bang-for-the-buck amps around. I'm sure they would be interested in a product that would allow them to compete at the highest levels. And they're a bunch of nice folks.

Just my 2 cents.
 
Ah, thanks for that! I have been saying exactly the same thing. It is the same in companies, if you let the problem happen and then fix it you are a genius, but if you prevent the problem from happening then it is business as usual. I always explain to would-be programmers that you don't "discover" race conditions, you actively plan against them, otherwise you will spend the rest of your life doing trial and error trying to get the software to work.

@tiki I see your point: you see a problem coming, you solve it and then it fails to come, falsifying your predictive ability :)
This is the usual problem with prophesies of doom which go away when action is taken. One shining example is the Y2K bug. This is now held up just about everywhere as an example of a spurious prediction of doom. Excuse me? It is the prime example of a prediction duly heeded. Thousands of programmers have sifted through billions of lines of code to change 2-digit year fields to 4-digit and to look for time difference calculations with a wraparound problem. The whole effort started years before 2000 and has worked splendidly. Sorry for the OT, but it's an old hobby horse of mine.
 
Last edited:
It is the same in companies, if you let the problem happen and then fix it you are a genius, but if you prevent the problem from happening then it is business as usual.

T'was ever thus - consider this anecdote which forms part of the introduction to my copy of 'The Art of War' by Sun Tzu :

According to an old story, a lord of ancient China once asked his physician, a member of a family of healers, which of them was the most skilled in the art. The physician, whose reputation was such that his name became synonymous with medical science in China, replied, "My eldest brother sees the spirit of sickness and removes it before it takes shape, so his name does not get out of the house. My elder brother cures sickness when it is still extremely minute, so his name does not get out of the neighborhood. As for me, I puncture veins, prescribe potions, and massage skin, so from time to time my name gets out and is heard among the lords.
 
Hello Bruno,

Nice specs! Did you by any change measure "THD vs Frequency" for lower powers (1W, 100mW, 10mW) The plots for 10,100 and 250W are less handy for a guy like me with a high effeciency speaker (Geddes Abbey clone)

I can see from the "THD vs Power" plots that it looks excellent for the frequencies shown, but I would very much like to see the complete picture...

Many thanks,

Wim
 
...Did you by any change measure "THD vs Frequency" for lower powers (1W, 100mW, 10mW) The plots for 10,100 and 250W are less handy for a guy like me with a high effeciency speaker (Geddes Abbey clone)...
?? The graphs at Hypex give THD specs down to 10mW output. Clearly the ncore has excellent behavior and a sensible bandwidth. At power output below 1 watt (perhaps much higher in this case), noise will swamp distortion (I expect noise is the cause of the steady slope to the left of 5 watts output in the graph on page 4, labeled "THD into 4 ohms at 100Hz (blue), 1kHz (green) and 6kHz (red)").

If you need only 5-10 watts max output, you should minimize noise with something like a tiny class A circuit with ultra-clean power supplies: maybe something in Nelson Pass's quiver (Aleph?). Really good amps like that are pretty rare, excluding tube amps which are generally sweet with distortion. IMHO, ultra-high-efficiency speakers sacrifice accuracy for color, but I shouldn't start a war here :) .
 
Last edited:
?? The graphs at Hypex give THD specs down to 10mW output. :) .

This only shows 3 distinct frequencies, I would like to see the complete frequence responce.

At power output below 1 watt (perhaps much higher in this case), noise will swamp distortion (I expect noise is the cause of the steady slope to the left of 5 watts output in the graph on page 4, labeled "THD into 4 ohms at 100Hz (blue), 1kHz (green) and 6kHz (red)").

:) .

I agree it will be higher, but I would like to see if this is caused by noise only, or there is something else also.

If you need only 5-10 watts max output, you should minimize noise with something like a tiny class A circuit with ultra-clean power supplies: maybe something in Nelson Pass's quiver (Aleph?). Really good amps like that are pretty rare, excluding tube amps which are generally sweet with distortion. IMHO, ultra-high-efficiency speakers sacrifice accuracy for color, but I shouldn't start a war here :) .

Why would I need something else? Total unweigthed noise of the ncore is 24uV, and all other specs look fantastic also. I just would like to see if anything changes in the low power region...
 
Wim,

The plot you're asking for would be a set of horizontal lines because the analyser would see only noise. The high power THD vs frequency plots were taken to show that at other frequencies THD remains roughly at the level found at 1kHz. If you look at the FFT spectrum at 1W you see that distortion is already quite far down. The spectral composition remains the same at other frequencies. Distortion at 100mW and 10mW is below the distortion of the AP's generator. I take it your question is: does this amplifier do anything unexpected at power/frequency points not in the data sheet? Answer: no. The three THD vs power plots and the three THD vs frequency plots are made to give a representative cross-section of distortion performance and you can safely interpolate between them.

I should add that also at these low powers Ncore has way lower distortion than esoteric circuits purportedly invented specifically to have good low-power behaviour. When some designers (I saw a name flash by just now) claim their amp is "best below 1W" this simply means their amp gets steadily worse above 1W, not that below 1W it does better than any other amp. Of course they want you to think they're saying the latter. This is called the "bait and switch" fallacy. Don't fall for it.
 
Wim,

The plot you're asking for would be a set of horizontal lines because the analyser would see only noise. The high power THD vs frequency plots were taken to show that at other frequencies THD remains roughly at the level found at 1kHz. If you look at the FFT spectrum at 1W you see that distortion is already quite far down. The spectral composition remains the same at other frequencies. Distortion at 100mW and 10mW is below the distortion of the AP's generator. I take it your question is: does this amplifier do anything unexpected at power/frequency points not in the data sheet? Answer: no. The three THD vs power plots and the three THD vs frequency plots are made to give a representative cross-section of distortion performance and you can safely interpolate between them.

I should add that also at these low powers Ncore has way lower distortion than esoteric circuits purportedly invented specifically to have good low-power behaviour. When some designers (I saw a name flash by just now) claim their amp is "best below 1W" this simply means their amp gets steadily worse above 1W, not that below 1W it does better than any other amp. Of course they want you to think they're saying the latter. This is called the "bait and switch" fallacy. Don't fall for it.

Thanks for your answer. Now I cannot wait until I can buy some of these modules :)

This will be the amp I (and I am sure many others) have been waiting for, quiet, low output impedance, powerful and overall excellent measurable performance, without functioning as a “class A” central heating system…
 
Good Lord, I finally got the measurements of Halcro's DM88 monoblocks for comparison. You have to twist your head a bit, but the ncore BEATS the Halcro in high-power THD *AND* IMD--the latter measurement was at 400W@2Ohms for ncore and 495W@4Ohms for Halcro! Caveat, Stereophile has a curious testing history.

EDIT: OK the Halcro is close to clipping there, no other IMD graph available for it, ncore is just at ~40% of clipping. Halcro seems to be a little better at THD at 8 ohms, hard to discern, but at 4 ohms and below is owned by ncore. Low level THD is outstanding for both.

I haven't studied these graphs for long so would appreciate another interpretation.

In any case ncore actually is rated for and measured at 2 Ohms, is perhaps 10% the finished price, <15% the weight, and many times greener.

See:

Halcro dm88 Reference monoblock power amplifier Measurements | Stereophile.com

http://www.hypex.nl/docs/nc1200 folder web.pdf

I would have found this months ago, I just stopped short. Not that I was about to buy a pair of Halcro monoblocks at $40k...
 
Last edited:
Good Lord, I finally got the measurements of Halcro's DM88 monoblocks for comparison. You have to twist your head a bit, but the ncore BEATS the Halcro in high-power THD *AND* IMD--the latter measurement was at 400W@2Ohms for ncore and 495W@4Ohms for Halcro! Caveat, Stereophile has a curious testing history. Halcro seems to be a little better at THD at 8 ohms, hard to discern, but at 4 ohms and below is owned by ncore. I haven't studied these graphs for long so would appreciate another interpretation.

See:

Halcro dm88 Reference monoblock power amplifier Measurements | Stereophile.com

http://www.hypex.nl/docs/nc1200 folder web.pdf

I would have found this months ago, I just stopped short. Not that I was about to buy a pair of Halcro monoblocks at $40k...
 
Let’s hope some of the selected VARs are no-nonsense engineering companies that make an affordable design, so that this excellent performance becomes available for non-DIYers also…

If you look at the performance data of the Amplifier modules produced by Douglas Self (class-A design by another of the audio guru’s), you’ll will see that the ncore beats even their performance in many ways
 
Good Lord, I finally got the measurements of Halcro's DM88 monoblocks for comparison. You have to twist your head a bit, but the ncore BEATS the Halcro in high-power THD *AND* IMD--the latter measurement was at 400W@2Ohms for ncore and 495W@4Ohms for Halcro! Caveat, Stereophile has a curious testing history. Halcro seems to be a little better at THD at 8 ohms, hard to discern, but at 4 ohms and below is owned by ncore. I haven't studied these graphs for long so would appreciate another interpretation.

See:

Halcro dm88 Reference monoblock power amplifier Measurements | Stereophile.com

http://www.hypex.nl/docs/nc1200 folder web.pdf

I would have found this months ago, I just stopped short. Not that I was about to buy a pair of Halcro monoblocks at $40k...

Hello Sam_Lord

Dont forget that the Ncore amplifier is measured with a AES17 Low Filter at the input of the Audio Precision which has a stop band attenuation >60 dB, 24.0 kHz–200 kHz (for the 20Khz setting) which of coarse improves the THD performance at the high frequencies.

AES17 Filter

If we look at a high frequency like 12Khz for example the second harmonic 24kHz is highly attenuated by the filter , hence lower THD . Using this filter the Halcro and for that matter any amplifier would also benefit in the high frequency THD performance stakes.

I dont mean to be an apologist for the Halcro.


Regards
Arthur.
 
That's why one should do the IMD test. Not just on class D amplifiers but on any amp. You want to find out if a given distortion mechanism can give rise to mix products inside the audio band or not. Two amps could show a similar rise in THD at >6kHz (as measured without low-pass filter), but they could do so for different reasons. If the cause is merely loop gain dropping off, this means any mix products inside the audio band will still be met with the full force of the in-band loop gain. If the cause is some nonlinear capacitance, in-band IMD will be quite high.

Because outband harmonics are in themselves inaudible (you can hardly hear musical content above 20kHz, let alone distortion) whilst mix products inside the audio band are not, it is OK to cut the THD measurement above 20kHz provided a HF IMD plot is supplied. This is more relevant to sound than a THD plot with a wide analyser bandwidth. Of course, if an amp has low THD with a wide measurement bandwidth, you can dispense with the IMD plot. In the case of the Self amp linked earlier it's hard to say whether the distortion rise indicates in-band IMD problems or not. My guess is not.

I'm a bit confused about the measurements in Stereophile. First there's the description of the test condition. They're saying 453W *peak* (not average or "rms"). That's 42.5V peak. This means each test tone alone has a 21.3V peak value (56.5W/4ohm). The average ("rms") power output during this test was 113W/4ohm.
That is of course unless the reviewer meant something like "peak average" power in which case the figures are 60.2V, 30.1V, 113.25W and 226W. This seems reasonable and entirely possible except that it's hard to believe that this is how the phrase "peak" should be read.

Luckily we can still work out how big the uncertainty is that results from this ambiguity. The THD plot shows onset of clipping at 500W/4ohm or 63V. Depending on how we read the above, the IMD test either excercises the amp to 67% or 95% (in voltage terms!) of clip. I guess we'll never know. On the other hand, we do know that THD rises monotonically with power right until the thing clips hard. In that case we can use the simple polynomial rule to extrapolate. According to the IMD plots, the strongest IMD product is the 3rd, which rises 2dB for every 1dB of signal level (indeed this is what the THD plot does). So if we interpret the test conditions incorrectly, the maximum error we can make is 6dB. So we err on the side of caution and presume that the Halcro was indeed tested at 95% output swing. The 3rd IMD product reads around -87dB

In the case of the Ncore measurement the 4ohm test was done at 200W average, being 100W per tone. Peak voltage of 56.6V. Because of the gradual clipping nature of self-oscillating amps (I mean THD starts rising sharply before actual clip) it's hard to say which is the clipping point. If we take 0.05% THD it's about 600W (69.3V). The IMD test stressed the amp to 81.6% (in voltage terms) of clip, where IMD was around -103dB.

Scaling the Halcro measurement down to the same ratio we get about -90dB. In the high power IMD stakes Ncore wins by a margin of at least 13dB (or 19dB if we take the other interpretation). Note however that Ncore's distortion does not drop off monotonically so it's quite likely that at some lower power the two will meet.

At least that's how I read the Stereophile plots.

And here's my second issue. I just find it difficult to believe that Bruce would claim <120dB THD if his own lab sample measured as in the Stereophile test. I won't speculate on the cause and it's more a human judgment than a technical one, but I'd like to measure a Halcro first before fully believing the Stereophile test data.
 
Last edited:
Just a quick update to say things are still moving on the NC400 front. The whole discrete shebang (regulators, op amps) is working OK, the power stage seems to do its job as well. Modulator (unloaded) linearity took some tweaking. Attached pic is the THD vs power at +/-65V (Apologies for the colour).
 

Attachments

  • nc400thdvp.GIF
    nc400thdvp.GIF
    20 KB · Views: 871
Just a quick update to say things are still moving on the NC400 front. The whole discrete shebang (regulators, op amps) is working OK, the power stage seems to do its job as well. Modulator (unloaded) linearity took some tweaking. Attached pic is the THD vs power at +/-65V (Apologies for the colour).
Is this plot directly comparable to the one of the datasheets of the UCD400HG etc ?
If so, very impressive !
 
Status
Not open for further replies.