- Joined
- Aug 14, 2018
- Messages
- 2,769
- Likes
- 8,149
Yes - but he doesn't want to test the HF linearity that way, because it doesn't produce the results he's obsessed with talking about.
-
WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
Hypex Nilai500DIY Amplifier Review
- Thread starter amirm
- Start date
Yes - but he doesn't want to test the HF linearity that way, because it doesn't produce the results he's obsessed with talking about.
Skeeter
Member
Given the HF distortion profile, I suspect the two tone IMD results won’t be good. Getting amplifier mA to show good distortion at 1 kHz is relatively easy, but it gets harder at HF. -110 dBr full 2 tone output power would be a good result.
restorer-john
Grand Contributor
Sadly, there's a cohort who feel its their duty to religiously prosecute the narrative that 20-20kHz is all that matters. Because a deeply vested interest wrote a white paper or they read something a guru wrote and took it to heart as gospel.
But, if you or I had created Class D amplifiers like Hypex or Purifi, we'd attempt to paint them in the best light possible- I've said that before. But I wouldn't stoop to blatant misrepresentation and neither would you. Pure marketing wins and it seems, some people are falling for it. Not me.
I have power amplifiers where the HF distortion is precisely trimmed to a minimum level during setup. A 20kHz signal is injected and the harmonics (up to 200kHz) are monitored and the HF THD trim is tweaked to a minimum. That is so they produce their rated (0.003%) THD across the entire specified bandwidth and meet their advertised specifications in accordance with regulations. Try doing that with a 20kHz measurement bandwidth.
Another, a 1982 Perreaux PMF-2150B 200wpc@8R/400wpc@4R power amplifier has a rated THD at 20kHz and 200 watts RMS continuous of 0.009% or lower. The HF THD adjustment must be made at full power and at 20kHz, monitoring harmonics. Try that with a Class D. The last one I adjusted, was 0.004% at 20kHz at 200w. A difficult trim, but they took their responsibilities to adhere to specifications seriously.
Hypex/Purifi et al, just throw all that out the window and draw a line at 20kHz and say "nothing to see here- move along". Their THD, power and efficiency numbers have been, continue to be and likely will remain, cherry picked. That is until some authority pulls them up and requires honest disclosure.
But we have more in common than not. Everyone here on ASR is interested in fidelity, music and HiFi gear. Some people come at it a different way than others and place more emphasis on standards and disclosure- others just want to press a button and listen to music. Everyone is different.
Sadly there is also a cohort who seem to feel it is their duty to try to come up with reasons why technology that is different from what they are used to is bad. They don't even provide any evidence, they just appeal to authority, supposed expertise or just handed down audiophile folklore.
Such explanation (“like that of zero global feedback amp”) would be a big misunderstanding. The reason is related to feedback loop complex transfer function that corrects the switching amplifier nonlinearities. Multiple poles and is optimized <=20kHz. It shapes and pushes distortion and noise energy >20kHz, similarly as DSD for SACD. This leads to rising noise AND distortion components above 20kHz. This behaviour is inevitable and will remain unless more ideal power switching devices would be available. GaNs are not the case.
This is what it is. Just circuit behaviour. No need for emotions, but also no need for hiding and neglecting facts.
Skeeter
Member
I was commenting on the shape of the distortion curve not how the feedback works. That the Hypex amp use multi-pole feedback compensation is well documented. In the distortion graph shown it bulges between c 7 kHz and HF and espec on the higher order harmonics. One of them is at -80 dB which is high in my view for an amp showing single tone ppm level distortion at 1 kHz. This is not the distortion profile of an ultra low distortion amplifier which is how it’s marketed, so I’m simply asking why this is the case.
So measurements below are made properly, right, taking into account the fact that @BW90kHz is the maximum I can get with my instrumentation. The result is not SOTA, but definitely not bad, good old class AB.
Rob Fens
Member
I may be mistaken and if so please help me to get a better understanding of the matter.
If what you are explaining about bandwith and distortion is true for amplifiers, why is it not true for a cd player, which is also limited to 22kHz?
And if so, why should an amp have a bandwith 5 times bigger than a cd player, i never heard anybody complain about cd players.
Regards,
Rob.
Hypex Nilai 500 DIY
Purifi 1ET400A
Took me 2 minutes to find out, copy/paste included...
And we are repeating ourselves over and over and over...
Edit: I just added Q7 in the Class D FAQ, along with numbering them. Now it is easier to answer: check Q&A 7.
www.audiosciencereview.com
Edit: I just added Q7 in the Class D FAQ, along with numbering them. Now it is easier to answer: check Q&A 7.
Class D FAQ
This topic is dedicated to Frequently Asked Questions about class D amplifiers. If you have been directed here the reason is that it is easier to concentrate the frequent answers here instead of repeating them in several topics. Q1: Class D power ratings cannot be true. How can these small and...
www.audiosciencereview.com
Last edited:
Hi John, this is how the NC252MP measures with 90kHz measuring bandwidth. Spectrum of 1kHz/25W, 20kHz/25W and finally THD+N vs. power at 1kHz, 10kHz, 15kHz and 20kHz, 4ohm load always. This is certainly nothing that one could put into the datasheet, so they had to investigate soft test conditions to make it look better. I will continue posting facts and will prepare more comparisons. Nilai and Purifi probably would not differ much with measuring BW 90kHz.
BTW I do not believe their datasheets. Soundstage measurements of NAD C298 (Purifi) show very different CCIF distortion level. Hypex also do not say a word about measuring bandwidth and filters used in their measurements. It is close to marketing policy, hiding facts.
Agree, this is part of why others have accused ASR members of being overly focused on measurements. For that matter, claiming Hypex/Purifi are spinning their specs when most other vendors are doing exactly the same. Not to mention the lack of discussion of whether the measurement differences are audible or not.
This thread's discussion needs to move on or should be closed.
Rob Fens
Member
Hi,
CD is restricted to 22kHz based on the limitations of CD using a sampling frequency of 44.1kHz. The measuring equipment such as Audio Precision will still use 80kHz measurement bandwidth.
Regards,
Shadders.
Skeeter
Member
That’s excellent. But the review graphs here don’t seem to show that. Might be Amir needs to add a 19+20 kHz IMD test. I would recommend that for any amplifier - whether class D or linear AB etc
Please see post #429
Since the Nilai have built in input buffers there is not going to be nearly as much (any?) variation between prebuilt and kits for the Nilai amps. I suppose someone could replace the power supply, but I've not heard anything bad about Hypex power supplies.
I built a pair of the Nilai mono blocks and have no regrets in going this route.
It seems like value only comes up when the price to performance ratio is out of whack. Amir seems to only bring up cost when it is far out of line with what you get for it.
- Joined
- Aug 14, 2018
- Messages
- 2,769
- Likes
- 8,149
No, no reason for you to care at all. But that's because of your particular perspective, approach to amps, and use-case. Many folks, including many members here, view it conversely: why should they pursue a DIY kit if they can get a pre-built unit with a professionally manufactured case, and often a warranty, for the same price?
Neither approach or perspective is wrong; they're just different, and both totally valid.
Welcome to the forum!
You raise a very interesting point, and your point plays into a lot of the discussion here.
The short response to your question is, of course you are correct: the cheapest amplifier that reproduces signals transparently - that is, with frequency response nonlinearities, noise levels, and distortion levels that are below the ability of humans to hear - is by definition the best value.
Of course, in reality it gets more complicated, because there are other questions that are variable based on individual situations and preferences and are sometimes difficult to quantify precisely, but are in my view still within the overall realm of objective qualities of one sort or another:
1. Build quality/likely longevity of the product;
2. Location company and of available vendors - the more inexpensive the product itself, the more the price comparison can change if one item is available domestically with free shipping while another is only available internationally with expensive shipping; and location of manufacturer can also have a major impact on how practical it is to be able to actually take advantage of the warranty if something goes wrong with the amp; and
3. Watts/power/current capacity - do you get an amp with 4x the power for 2x the price of another, otherwise comparable amp? For folks listening in mid-field with highly efficient speakers, the answer is probably No. For folks listening far-field with less-efficient speakers, the answer might be Yes.
Finally, there's the more subjective issue of how a product makes us feel. The science/measurements/engineering ethos of this forum is still an ethos - it still speaks of values that folks bring to their evaluation of products. So having extra power reserves, first-rate build quality, "extra padding" in the performance specs beyond the threshold of audible transparency, and so on - these are things that many folks here like to see in products they consider buying.
Periodically some folks will comment here in threads accusing all of ASR of being a sham or hypocritical because "everyone here chases specs" even when better specs don't make an audible difference, and therefore we are allegedly no different from people who seek out cable lifters, high-end interconnects, and other audiophile "tweaks" even when they don't make an audible difference. The claim is that all this stuff is irrelevant and just makes us feel better, and so therefore measurements are the same thing as snake oil. It's not true of course, but like many bad arguments that refuse to die, it's just close enough to the truth to be misleading.
Last edited:
Skeeter
Member
Seems to me this might be what’s going on here:-
1.The loop gain is VERY high (rough guess 100 dB or more) and the multi-pole compensation suppresses distortion in the audio band up to 20 kHz to very low levels
2. This will also provide very good in-band IMD performance which is clearly evident from boXem’s post above.
3. Above 20 kHz, the loop gain drops off very rapidly as it has to in order to get to close the loop at some reasonable unity loop gain frequency, say 1-3 MHz’ at 20 dB/decade enduring adequate phase margin.
4. So the loop gains above 20 kHz drop off dramatically and that’s why the wide band (90 kHz) HF distortion seem to be higher than one would anticipate looking at the in band distortion performance.
1.The loop gain is VERY high (rough guess 100 dB or more) and the multi-pole compensation suppresses distortion in the audio band up to 20 kHz to very low levels
2. This will also provide very good in-band IMD performance which is clearly evident from boXem’s post above.
3. Above 20 kHz, the loop gain drops off very rapidly as it has to in order to get to close the loop at some reasonable unity loop gain frequency, say 1-3 MHz’ at 20 dB/decade enduring adequate phase margin.
4. So the loop gains above 20 kHz drop off dramatically and that’s why the wide band (90 kHz) HF distortion seem to be higher than one would anticipate looking at the in band distortion performance.
Similar threads
- Poll
- Poll
