Review and Measurements of Hypex NC400 DIY Amp

[amirm]

Here is the corrected results:

 

[Deleted member 2944]

Amir,

Your AP analyzer applies the AES17 filter to both the DUT excitation signal and the DUT measurement?? You can adjust for DUT latency (if it exists) as well?
I think this is where the confusion comes from regarding the phase response measurements.

Dave.

 

[amirm]

Your AP analyzer applies the AES17 filter to both the DUT excitation signal and the DUT measurement??

No. Speaker output goes through the AES-17 filter. The feed to the amplifier is unmodified (flat).

 

[BYRTT]

Here is the corrected results: .....

Thanks, omiting the hump at about 175kHz below is dirty quick result using Rephase to show belonging minimum phase had we lightspeed bandwidth in measurement gear.

 

[March Audio]

Here is the corrected results:

View attachment 19572

I know we are not testing the same device NC252 V NC400, but I am curious as to why a simple scope measurement would lead to such a different result to this. Any thoughts?

I would surprised if the AES17 output filter caused that much difference, but can you perform a scope measurement without the filter to compare?

I will have the NC500 modules next week so it will be interesting to re-visit.

 

[Blumlein 88]

@amirm has already done a lot of work benchmarking my amp (thank you!), and before he receives something with even lower distortion, I think it is important to determine a reasonable set of standardized power measurements and test conditions that can be repeated with future amplifiers. Comparing apple to oranges creates too much room for interpretation and speculation.

If you want some context to your measurements, consider profiling power at three levels based on use case

1. Normal usage of high efficiency speakers
2. Normal usage of average efficiency speakers
3. Full power (knee in distortion curve which defines usable power bandwidth as Amir did with my amp, not predetermined point like 1%).

To get an idea of what that could mean, calculate power requirements for speakers based on efficiency. One could design a system for a target SPL of say 85dB at the listening position, and this table is just one way to go about it given an 8Ohm load. Given high and average efficiency speakers, we would only care about 1W and 8W respectively (ignore headroom, or add it to your target).

View attachment 19543
For power specific tests, adjust amp Vin to get the predetermined Vout and measure power. Standardizing on a 4Ohm load is probably a good idea, although some amps degrade quickly as the impedance drops away from 8Ohms, but that compounds Amir’s work.
View attachment 19544
Specific tests: This is a starting point based on testing Amir has already completed. Please add, change, correct, argue, specify filters or certain conditions that will affect the outcome. Or, establish a minimum consensus of what matters so Amir remains willing to keep performing these tests for us.

1. Frequency Response & Phase (90kHz BW)
2. THD+N (%) vs Power @1kHz (AES-17 filter, 22kHZ BW)
3. THD+N (%) vs Frequency @ 1W, 10W, full power (90kHz BW)
4. IMD (dB) vs Power @1kHz
5. SINAD @ 1W, full power
6. Square wave 1kHz @ 1W
7. Sine wave 1kHz @ <<1W gauge high order harmonic content
8. 256k FFT 1kHz @ 1W (broadband noise)
9. 256k FFT 1kHz @ 1W, 10W, full power (90kHz BW)
10. 256k FFT zoom 1kHz @ 1W (90kHz BW) gauge mains contribution

Sorry about the giant images. They seems to display in this editor at 2x what they do on their own in the browser.

What you are really looking for I think is the Power Cube measurement. Discussed a time or two here. The Power Cube add on is rather expensive.

 

[March Audio]

As a sanity check I generated a 20 kHz test file in Adobe audition with the right channel with a 20 deg phase shift.

Played it straight from dac into scope. Measured 20 deg, so Im happy the scope measurement is accurate.

@amirm Unless the NC400 is fundamentally different to the NC252, I am very suspicious that the AES filter is adding additional phase shift to your measurement there.

 

[firedog]

If somone is interested in buying an assembled NCore amp from what seems like a reputable builder who does a good job:
https://www.apollonaudio.com/
The also use the NC500 model (OEM instead of DIY model) that allows for builder made input section, so they offer an input section that allows you to "roll" the opamps and color the sound to your taste.
They also use the Hypex 1200 (instead of 600) PS, which eliminates the PS issues mentioned here.

 

[VintageFlanker]

Hello @amirm
Would be curious to see your measurements of this Hypex Ncore assembled power amp:
The Audiophonics PA-S125NC.

https://www.audiophonics.fr/fr/ampl...stereo-class-d-ncore-2x75w-8-ohm-p-12756.html

Perhaps @Audiophonics could send you one for review?

 

[restorer-john]

Unless the NC400 is fundamentally different to the NC252, I am very suspicious that the AES filter is adding additional phase shift to your measurement there.

Curiously, on AP's site they say this:

"Note that only the 2700 Series analyzers accept optional AES-17 filters. The APx Series and ATS-2 don’t need them—their standard low-pass filters are implemented digitally and already have a brick-wall roll-off that meets the AES-17 specifications."

Also this in relation to an earlier AES-17 AP filter (active):

• The AES17 filter affects phase measurements, since one channel is filtered, and the other is not. Do not perform phase measurements when this filter is enabled.
• The AES17 filter also affects Level and Frequency meter readings on the selected channel. Flatness measurements should be performed with the filter disabled.

Why is the crosstalk of the AP AUX-0040 only 90dB? The unit will have separate boards and the connectors are sufficiently separated I'm sure. Perhaps the 90dB is a worst case, at extremely high input voltages? If they couldn't do better than 90dB, why not just make them two separate boxes for each channel?

 

[JustIntonation]

As a sanity check I generated a 20 kHz test file in Adobe audition with the right channel with a 20 deg phase shift.

Played it straight from dac into scope. Measured 20 deg, so Im happy the scope measurement is accurate.

@amirm Unless the NC400 is fundamentally different to the NC252, I am very suspicious that the AES filter is adding additional phase shift to your measurement there.

I think I get now how the phase shift works thanks to posts in this thread.
It is completely due to the natural minimal phase behavior of the frequency response (knew they were related just didn't know it is this much already for such small amounts of -dB). So the NC400 being around -0.5dB at 20kHz means it is also has the related minimal phase behaviow phase shift of about -40 degrees. (and the first phase response post probably automatically corrected for this and only showed phase deviations from minimal phase behavior)
Your measurements of the NC252 probably has less than -0.5dB at 20kHz, more like -0.2dB (perhaps due to DAC or ADC etc) hence the -20 degree phase shift.
So it basically shows these amps have completely natural minimal phase behavior and no phase anomalies.

 

[March Audio]

I think I get now how the phase shift works thanks to posts in this thread.
It is completely due to the natural minimal phase behavior of the frequency response (knew they were related just didn't know it is this much already for such small amounts of -dB). So the NC400 being around -0.5dB at 20kHz means it is also has the related minimal phase behaviow phase shift of about -40 degrees. (and the first phase response post probably automatically corrected for this and only showed phase deviations from minimal phase behavior)
Your measurements of the NC252 probably has less than -0.5dB at 20kHz, more like -0.2dB (perhaps due to DAC or ADC etc) hence the -20 degree phase shift.
So it basically shows these amps have completely natural minimal phase behavior and no phase anomalies.

Good point, I will check the frequency response of the 2 units, but I thought they were very similar.

Edit. Yes they are almost identical, if anything the 400 is slightly lower rolloff.

 

[BYRTT]

...So it basically shows these amps have completely natural minimal phase behavior and no phase anomalies.

Think it looks modern ways of circuit compensention inside various chips or amps can sometimes lead to more or less deg of phase deviations at output verse a expected pure minimum phase device, often its hard to see if its close to expected target curve but a dirty quick way up at HF stopband is look for -3dB points where phase then often shall pass -45º or -90º points. Whether even small phase deviations is audioable is probably different from person to person, and should it be a problem then active systems or players including DSP engines can go for a dialed in textbook minimum phase domain.

 

[March Audio]

Curiously, on AP's site they say this:

"Note that only the 2700 Series analyzers accept optional AES-17 filters. The APx Series and ATS-2 don’t need them—their standard low-pass filters are implemented digitally and already have a brick-wall roll-off that meets the AES-17 specifications."

Also this in relation to an earlier AES-17 AP filter (active):

• The AES17 filter affects phase measurements, since one channel is filtered, and the other is not. Do not perform phase measurements when this filter is enabled.
• The AES17 filter also affects Level and Frequency meter readings on the selected channel. Flatness measurements should be performed with the filter disabled.

Why is the crosstalk of the AP AUX-0040 only 90dB? The unit will have separate boards and the connectors are sufficiently separated I'm sure. Perhaps the 90dB is a worst case, at extremely high input voltages? If they couldn't do better than 90dB, why not just make them two separate boxes for each channel?

Well it's simple to confirm with a (filter less) scope measurement.

 

[March Audio]

Hello @amirm
Would be curious to see your measurements of this Hypex Ncore assembled power amp:
The Audiophonics PA-S125NC.

https://www.audiophonics.fr/fr/ampl...stereo-class-d-ncore-2x75w-8-ohm-p-12756.html

Perhaps @Audiophonics could send you one for review?

Same module as my amps for which I will be publishing independent measurements soon.

 

[Thomas savage]

This thread reads quite well and seems constructive , thank you to those involved in the discussion here.

 

[March Audio]

It's interesting as there isn't a whole lot of independant data about class d amps. They seem to polarise subjective opinions, so its good to technically delve in.

 

[maty]

So far, class D amplification is problematic with tweeters, at least the old class D.

When I was n-reviewing the specifications and measurements of the IcePower 1200AS2 module PDF manual, the only graph I found different was that of the phase. That it varied so much was an alarm signal in my brain.

This would explain why unaltered voices (you know, without Autotune, vade retro satana) and acoustic instruments sound bad with this technology.

Many audiophiles reneged on the NC400, which is why I supposed that he also had a big problem with the phase. They have the problem but it is not that huge. Maybe the NC500, with custom buffers mask that problem.

With the current music, so badly recorded and manipulated should not be problematic. This would explain why many do not dislike sound. Let is face it, what percentage of people listen to (high) quality recordings with acoustic instrumentation and unaltered voices via software? And with DR > 11dB. The percentage must be ridiculous.

 

[Frank Dernie]

I cross once in a while an electronic made by Tag McLaren on eBay or on other second-hand listings, wondering what the relationship with the car racing company was, if any ... or perhaps just a branding name coincidence ...

Somewhat off-topic but
TAG is/was a trading company that originally made most of its money as the agent to the Saudi Arabian minister of defence. They got a percent. age of all arms sales.
The eldest son of the founder was a motor racing fan and wanted to increase his father’s company’s brand image and di a sponsorship deal With Williams Grand Prix Engineering around 1980, which is where I met him. He wanted more involvement by buying in but Frank Williams wouldn’t sell. McLaren heard about this, allegedly by bugging the Williams motor home, and proposed to Mansour (the son) to let him take over the Porsche turbo engine project they had ordered and call the engine the TAG engine - rumour has it McLaren couldn’t afford to pay either for the engine of the rebuilds and Mansour took shares in lieu of payment. Whatever he ended up with a controlling interest in McLaren.
The Tag engine used Bosch’s electronics and the engineer in charge of the project was very ambitious and convinced McLaren to form their own electronics dept, which he ran. He is a big HiFi enthusiast and when the FIA changed the rules to limit electronics TAG-McLaren bought Audiolab for him and renamed it TAG-McLaren audio. Early products were more or less the Audiolab stuff in restyled cases.

The company was not a success and they sold to IAG which later re-launched the Audiolab name.
TAG had also bought a controlling interest in the Heuer watch company and re-named it TAG-Heuer, this has been more successful.
I knew the ex-Bosch engineer really well from F1 but can not think of his name at the moment.

 

[Thomas savage]

I moved some posts out, can we stick to the relatively technical discussion regarding the measurements here and challenges of measuring class D as that’s a intresting evolution of this thread.

If you just want to promote/highlight the many class D options out there you can do this here, https://www.audiosciencereview.com/forum/index.php?threads/class-d-amps-what’s-out-there.5934/ .

Cheers .