Hypex Nilai500DIY Amplifier Review

[Scott Lanterman]

350w per channel is even more powerful than the stereo Nilai lol
is the price the only advantage of this one? I'm from Canada so I get eaten on the shipping, but I can order NC400 or Nilai kits from a local distributor
The price difference isn't too big between them, so I figure might as well get the newest one

Hey Jes - I got my Nilai kit from VTV, amp is on back order at many places. No problems and they were great to work with. I think Buckeye is also a great deal. If you are price sensitive you probably cannot get better value than Buckeye or VTV Hypex amps. You have sensitive speakers so I think you might want to care about the sound (super low distortion) at low wattage, and Klipsch can sound good at low listening levels and have amazing dynamics. If price is not a big issue I would consider the Purifi or Nilai. Maybe even go all out with the Topping LA 90 for mid range horns, and then bi amp with some beastly Hypex power for your 15 inch bass drivers and blow the roof of your house, (lucky you are in the basement!?!?). Have some fun with your favorite music man!

 

[jesburger]

Hey Jes - I got my Nilai kit from VTV, amp is on back order at many places. No problems and they were great to work with. I think Buckeye is also a great deal. If you are price sensitive you probably cannot get better value than Buckeye or VTV Hypex amps. You have sensitive speakers so I think you might want to care about the sound (super low distortion) at low wattage, and Klipsch can sound good at low listening levels and have amazing dynamics. If price is not a big issue I would consider the Purifi or Nilai. Maybe even go all out with the Topping LA 90 for mid range horns, and then bi amp with some beastly Hypex power for your 15 inch bass drivers and blow the roof of your house, (lucky you are in the basement!?!?). Have some fun with your favorite music man!

Thanks for the tip, I already have a DIY 18 inch Full Marty subwoofer so I was thinking of doing around a 50hz crossover. I don't have the speakers yet I'm doing a Crites Audio B model cornscala build. I'm not in a rush so I can order the stereo nilai now and get it whenever.

Thanks for the tips

Edit - would the topping amp be a worthy alternative to the class d amps for my scenario?

 

[Scott Lanterman]

Hey Jes - It sounds like you are pursuing a "big" system. This thread is about the Nilai... and I think they are great. I am using them now... they are very clean.

If you just go with this website and care about measurements and value, the Topping is at the top of low distortion. You can bi-amp with Topping for MT (mid-treble) and then use economical Hypex (UCD 400) for the bass. This could be cheaper than Nilai and give you even more performance. If your planned speaker allows bi-amping... go for it. It sounds like a cool project. I like Horn dynamics and I had Klipsch Heresy as a kid and JBL cast offs from my dad.

 

[Mnyb]

Even dense music has quite a high crest factor so I would not hesitate and think it had to much power
The sine wave used for testing draws more continuous power than any music.

 

[jesburger]

Hey Jes - It sounds like you are pursuing a "big" system. This thread is about the Nilai... and I think they are great. I am using them now... they are very clean.

If you just go with this website and care about measurements and value, the Topping is at the top of low distortion. You can bi-amp with Topping for MT (mid-treble) and then use economical Hypex (UCD 400) for the bass. This could be cheaper than Nilai and give you even more performance. If your planned speaker allows bi-amping... go for it. It sounds like a cool project. I like Horn dynamics and I had Klipsch Heresy as a kid and JBL cast offs from my dad.

Thanks I'll look into it!

 

[boXem]

...
If you just go with this website and care about measurements and value, the Topping is at the top of low distortion noise.

Corrected

 

[Whatsthephuss]

Let's look at available power at 4 ohm:

….

I was a bit surprised at the ballooning of the distortion at higher frequencies:

Hi amirm, thanks for all the work put into making this review and your other reviews.

I wanted to point out a couple of things that stood out to me after trying to make some comparisons to the Benchmark AHB2:

1) The two (lower) charts for displaying power vs distortion over bandwidth are not easy to compare at all. It would be much more helpful to stick with one type or the other.

2) The Nilai chart displaying power vs distortion over bandwidth is labeled as the “Low Gain” setting, but when compared to the top chart of power vs distortion w/ 4ohm load it looks to me like you measured the Nilai in “Medium Gain.” This further complicates a fair comparison…

Although I got a lot of useful information from this review, I’m left questioning your data portraying the Nilai’s “ballooning of distortion.” It’s also unclear if this data is with the lid on or off.

 

[Shadders]

I’m left questioning your data portraying the Nilai’s “ballooning of distortion.”

Hi,
My interpretation is that the ballooning is based on the amplifier characteristics and that the measurement bandwidth is set to 45kHz, as opposed to the usual 22kHz, so more of the harmonics are captured.

I am assuming that the other measurements were implemented using the 22kHz bandwidth filter.

Regards,
Shadders.

 

[Whatsthephuss]

My interpretation is that the ballooning is based on the amplifier characteristics and that the measurement bandwidth is set to 45kHz, as opposed to the usual 22kHz, so more of the harmonics are captured.

I am assuming that the other measurements were implemented using the 22kHz bandwidth filter.

Hi Shadders, thanks for the helpful clarification!

I’m still left wondering a few things… was the AHB2 measured with a 45khz bandwidth? It’s unlabeled…
Edit: just saw AHB2 has both channels depicted in the graph

Just my personal opinion here, but using a 45khz bandwidth for class d amps known to have noise above 20khz seems like looking for trouble, doesn’t it? I personally use two different devices that hi cut and low pass the signal at 20khz, so this data depicting a ballooning of distortion does not seem relevant for my own practical application.

 

[Shadders]

Hi Shadders, thanks for the helpful clarification!

I’m still left wondering a few things… was the AHB2 measured with a 45khz bandwidth? It’s unlabeled, and also unclear if the AHB2 test was done with both channels driven? (While the monoblock Nilai clearly has both channels being driven

Just my personal opinion here, but using a 45khz bandwidth for class d amps known to have noise above 20khz seems like looking for trouble, doesn’t it? I personally use two different devices that hi cut and low pass the signal at 20khz, so this data depicting a ballooning of distortion does not seem relevant for a practical application.

Hi,
It depends on who is doing the testing. In general, the class D amplifier testing uses a 22kHz bandwidth due to the switching frequency causing issues with the Audio Precision measuring equipment :

AUX-0025 / 0040 / 0100 Switching Amplifier Measurement Filters

www.ap.com

For the AHB2 amplifier which is class AB, then the usual measurement bandwidth is 80kHz. I have seen DIY measurement in Elektor magazine using both 22kHz and 80kHz for their class D amplifier design.

As an aside, Hifi News i believe use 80kHz for all amplifiers including class D.

Class D does get "preferential" treatment, but then, if you cannot hear above 20kHz, then it is not an issue. The ballooning seems to be for 15kHz and 10kHz frequencies predominantly, so these 2nd order and greater harmonics will not be heard. Music content at these frequencies is minimal too, so no real impact.

Regards,
Shadders.

 

[Whatsthephuss]

Hi,
It depends on who is doing the testing. In general, the class D amplifier testing uses a 22kHz bandwidth due to the switching frequency causing issues with the Audio Precision measuring equipment :

AUX-0025 / 0040 / 0100 Switching Amplifier Measurement Filters

www.ap.com

For the AHB2 amplifier which is class AB, then the usual measurement bandwidth is 80kHz. I have seen DIY measurement in Elektor magazine using both 22kHz and 80kHz for their class D amplifier design.

As an aside, Hifi News i believe use 80kHz for all amplifiers including class D.

Class D does get "preferential" treatment, but then, if you cannot hear above 20kHz, then it is not an issue. The ballooning seems to be for 15kHz and 10kHz frequencies predominantly, so these 2nd order and greater harmonics will not be heard. Music content at these frequencies is minimal too, so no real impact.

Thanks for the info!

What stood out to me while comparing the Nilai’s 22khz bandwidth graph with the 45khz bandwidth graph, and maybe you can help explain this, but why does THD+N in “Low Gain” mode with 45khz bandwidth get about 5db worse over the entire measured level? Ex. At 1W 22khz bandwidth THD+N is below -105db, while at 1W 45khz bandwidth THD+N is above -100db…..

Also, the total power handling went down. Is the larger bandwidth really the cause for all of these discrepancies?

 

[Whatsthephuss]

Hi,
It depends on who is doing the testing. In general, the class D amplifier testing uses a 22kHz bandwidth due to the switching frequency causing issues with the Audio Precision measuring equipment :

This seems to make the most practical sense to me, since I see no point in extending the bandwidth into areas known to cause issues.

For the AHB2 amplifier which is class AB, then the usual measurement bandwidth is 80kHz.

Why not 200khz?

 

[Shadders]

Thanks for the info!

What stood out to me while comparing the Nilai’s 22khz bandwidth graph with the 45khz bandwidth graph, and maybe you can help explain this, but why does THD+N in “Low Gain” mode with 45khz bandwidth get about 5db worse over the entire measured level? Ex. At 1W 22khz bandwidth THD+N is below -105db, while at 1W 45khz bandwidth THD+N is above -100db…..
View attachment 265452View attachment 265453
Also, the total power handling went down. Is the larger bandwidth really the cause for all of these discrepancies?

Hi,
I think you have copied the wrong graph - you have pasted the AHB2 by mistake ?

Yes, the extra 23kHz (from 22kHz to 45kHz) can add the extra noise which is summed by the AP equipment.

Regards,
Shadders.

 

[Shadders]

This seems to make the most practical sense to me, since there is literally no point in extending the bandwidth into areas known to cause issues.

Why not 200khz?

Hi,
Probably it is a remnant from how it has always been done, given the capability of the equipment, but then measuring the extra 180kHz has no real benefit apart from presenting those amplifiers that excel at linearity.
Regards,
Shadders.

 

[Whatsthephuss]

Hi,
I think you have copied the wrong graph - you have pasted the AHB2 by mistake ?

Yes, the extra 23kHz (from 22kHz to 45kHz) can add the extra noise which is summed by the AP equipment.

Regards,
Shadders

Thanks for pointing that out. I’ve corrected the graph but the numbers are the same.

Pretty fascinating that extending the bandwidth beyond the useful hearing range can add so much distortion noise! But as someone who looks for data that is useful in real world applications I’m still left wondering what the THD+N would be with a proper filter.

This is what I found on the Nilai datasheet

 

[restorer-john]

Probably it is a remnant from how it has always been done, given the capability of the equipment, but then measuring the extra 180kHz has no real benefit apart from presenting those amplifiers that excel at linearity.

Total harmonic distortion is measured and calculated for 10 harmonics of a fundamental- hence a 200kHz measurement bandwidth for a 20-20kHz audio amplifier.

With only a 22kHz measurement bandwidth, the THD figures are meaningless above 2kHz and the plots do not represent the actual distortion present. A 45kHz bandwidth means you can get to 4.5kHz with accuracy. The plots are totally misleading as people look at the falling HF distortion and think "this is great" when it is the harmonics falling off the measurement bandwith and disappearing.

You need to basically extrapolate from the loss of the first group of harmonics and draw your own trend line.

 

[Shadders]

Total harmonic distortion is measured and calculated for 10 harmonics of a fundamental- hence a 200kHz measurement bandwidth for a 20-20kHz audio amplifier.

Hi,
Thanks - i was unsure what number of harmonics were used to produce THD. 10 seems OK for 2kHz, but for 1kHz and an amplifier that has high order harmonic distortion, which may be heard from 10kHz to 20kHz, then maybe more is a benefit ?

The blog from AP indicates that THD+N may measure all noise upto the bandwidth of the filter - so for the Nilai the additional harmonics above the 10 is counted as noise and increases the noise - reduced the THD+N.

Regards,
Shadders.

 

[Shadders]

Pretty fascinating that extending the bandwidth beyond the useful hearing range can add so much distortion! But as someone who looks for data that is useful in real world applications I’m still left wondering what the THD+N would be with a proper filter.

This is what I found on the Nilai datasheet

Hi,
The distortion above 20kHz is not heard, so the degradation of THD+N is not an issue. The test here for the 1kHz i believe used the 22kHz filter, and the result probably reflects the datasheet values.
Regards,
Shadders.

 

[Whatsthephuss]

Total harmonic distortion is measured and calculated for 10 harmonics of a fundamental- hence a 200kHz measurement bandwidth for a 20-20kHz audio amplifier.

Hi John, thanks for your input.

With only a 22kHz measurement bandwidth, the THD figures are meaningless above 2kHz and the plots do not represent the actual distortion present. A 45kHz bandwidth means you can get to 4.5kHz with accuracy.

I agree with your point of using wider bandwidth as a more scientific method of measuring THD figures, but someone using class d in commercial sound (with low pass filters) will want data that is more accurate to spec. Doing a 22kHz bandwidth measurement will give THD+N figures that are much more relevant to real world use than >45kHz. Not saying we can’t measure both, in fact that’s what I would prefer.

 

[Julf]

Total harmonic distortion is measured and calculated for 10 harmonics of a fundamental- hence a 200kHz measurement bandwidth for a 20-20kHz audio amplifier.

Any pointers to standards specifying "10 harmonics"? The usual way of specifying THD+N measurements is simply by bandwidth, as nobody is actually adding individual harmonics when they measure. What you normally measure is the difference between a pure sine tone and what the amp produces when given that sine tone as input - the difference is distortion and noise.