Class D principle, audible "higher frequency noise"

[tmtomh]

I forgot to answer the second part of your question: It is mor a "hiss"... somehow like you would take "pink noise" and remove all ferquencies lower than x, means: sounds to me like a mixture of several higher frequencies.

Regarding/Summarizing the other posts:

- Up to now it seems that other members here have same or at least similar experiences (I never had anything similar... and I am quite sensitive to noise, maybe I just was lucky up to now...). So it seems to be a comon thing... not nice maybe, but at least common...

- The noise is most probably not related to the implementation type (class D in this case)

- Regarding the suggestion testing XLR-Input: The noise is also there if NOTHING but the speakers are connected to the ST-10 or STA-9 (no Pre-amp, no RCA connection, no XLR connection, I even tested with no other electrical device in the room, and also in an additional room). I do not have the possibility to test XLR, but I guess the test would not change anything, because even with no connection at all the noise is there. Within my limited possibilities I can connect whatever wherever I want: The noise is there and does not change.

But... does somebody know where this noise is coming from (I mean the technical root cause)?
Or is it just hard or impossible to say because in every system/amp there may be different root causes?

Your description of the sound as pink noise-type hiss but without low frequencies is exactly what the generic electrical noise floor of most equipment sounds like. It's nothing to be concerned about per se, but IMHO if it's audible from any distance it's louder than it should be - reasonable people can debate whether equipment should be 100% totally silent, but I would think we can all agree that any noise should be quite minimal and at a very low volume level.

 

[tmtomh]

Ditto IME, until a few months ago. I have nCore and Purifi amps on hand where there is absolutely no audible noise from the tweeter (or any of the other drivers). I thought I heard some, then kicked myself after figuring out by putting my ear next to the baffle that it was the seashell effect.

One thing I just realized I forgot to mention about my setup: the noise I described above from my Audiophonics amp is when it’s set for max gain. At the factory default medium gain setting (which is 5.3dB lower than max gain), I couldn’t hear anything, even with my ear on the tweeter.

 

[March Audio]

Hi and thanks for the answer.
I already tried quite a lot, including: Connect only the ST-10 or or the STA-9 to speakers. No other connection (means: also no pre-amp), all other devices in the room completely disconnected from any power-socket. I also tried this test with different power and speaker cables, also in a different room. There was no influence to this noise (means: This "higher frequency" noise was not influenced by any other change).
Of course I first thought it would be defect, but the same defect in two devices... hm... and according to some opinions or Nuprime users this noise seems to be "normal" or at least not unusual, which was also surprising to me...

OK then it's not what I thought it could be and is simply a "feature" of the Nuprime. From eetu post above the design appears to have a poor SNR. It's simply noisy. Other class d designs are much better in this respect.

 

[CDMC]

The only known cure to idle background noise (that isn't related to ground loop) is inefficient speakers.

Works like a charm. Never had this issue with any of my Magnepans with countless different front ends.

 

[restorer-john]

I have nCore and Purifi amps on hand where there is absolutely no audible noise from the tweeter (or any of the other drivers).

That is just phenomenal. What did you measure noise wise?

My best AB is 16uV (A) residual into [email protected]/M speakers and I can just hear the tweeter hiss with a shorted input. A 200wpc with <35uV residual is audible from a foot away at night (when all is quiet). Not from a listening position or with a preamp active up front (the preamps make more noise)

 

[donsna]

OK then it's not what I thought it could be and is simply a "feature" of the Nuprime. From eetu post above the design appears to have a poor SNR. It's simply noisy. Other class d designs are much better in this respect.

Yes... from all that I have heard and read so far in the last days, in fact it seems to be "normal", at least within this application/equipment.

 

[donsna]

I was curious and downloaded a "Sound Analyzing" App:
"My higher frequency noise" is (according to the App) mainly within a range from about 6 - 12 kHz, the maximum is in the range about 8-10kHz with a range of 35 - 40 dB(A), measured at a distance of about 1-2 cm from the tweeter.
Quite sure this measurement was not exactly (as I do not know how exact the App is, if the calibration worked really fine etc.), but I guess this is the range we are talking about. Interesting!

Thanks to all contributing here so many useful comments! I might still not know what the root cause of my "personal Nuprime higher frequency noise" is, but now I know that this I could only find out by a brief technical investigation (which I cannot do) or by detailed information from Nuprime (which they will most probably not be very happy to give...).

And a positive side-effect: Now I have again a reason to justify investments to myself in the nearer future, I think I really have to start searching for another class D amp to compare and check!
So next ones to try will be most probably Hypex or similar... or maybe give another Nuprime another possibility... well, I have time within the next weeks and months to check the opportunities!

 

[SIY]

That is just phenomenal. What did you measure noise wise?

My best AB is 16uV (A) residual into [email protected]/M speakers and I can just hear the tweeter hiss with a shorted input. A 200wpc with <35uV residual is audible from a foot away at night (when all is quiet). Not from a listening position or with a preamp active up front (the preamps make more noise)

Here's the noise spectrum of the Purifi. I didn't measure the total voltage in the audio band to compare with your numbers, but I'd be happy to do so later today.

 

[pjug]

I found this review of the ST-10 with measurements. So if this is typical, noise level is above 0.1% at 1W into 4 ohms.
https://translate.google.com/transl...901-nuprime-dac-10h-st-10&prev=search&pto=aue

 

[Eetu]

I found this review of the ST-10 with measurements. So if this is typical, noise level is above 0.1% at 1W into 4 ohms.
https://translate.google.com/transl...901-nuprime-dac-10h-st-10&prev=search&pto=aue

View attachment 74117

But the test was with NuPrime DAC-10H + ST-10. This fig. is THD+N so you can't tell what the actual noise level is based on that.. They got a SNR of 80db (A) with the combo.

 

[pjug]

But the test was with NuPrime DAC-10H + ST-10. This fig. is THD+N so you can't tell what the actual noise level is based on that.. They got a SNR of 80db (A) with the combo.

Yes, I did not read it carefully. Maybe looking at the DAC specs you can tease out whether it is contributing. Seems likely that the noise would be mostly from the amp.

Whatever the measurements, the OP said the hiss is not audible a foot away from the speakers. So I wouldn't worry about it in this case.

 

[SIY]

That is just phenomenal. What did you measure noise wise?

My best AB is 16uV (A) residual into [email protected]/M speakers and I can just hear the tweeter hiss with a shorted input. A 200wpc with <35uV residual is audible from a foot away at night (when all is quiet). Not from a listening position or with a preamp active up front (the preamps make more noise)

19uV over 24k bandwidth, which is likely a bit high because my test leads are about 7 m long and nothing special for shielding. That's -103dB from 2.83V.

My speakers are more like 86-87dB rating at 2.83V, so there's exactly zero chance that I'd be able to hear any of that noise, even if I were 50 years younger.

 

[restorer-john]

Here's the noise spectrum of the Purifi.

That is just incredible. The noise in the audio band is dead flat. And absolutely no mains harmonics either. So that's a shorted input on a non-buffered Purifi using a Hypex SMPS?

 

[SIY]

That is just incredible. The noise in the audio band is dead flat. And absolutely no mains harmonics either. So that's a shorted input on a non-buffered Purifi using a Hypex SMPS?

It has the Purifi buffer- basically, it's their EVAL1 board in a chassis. 26dB gain. Yes, shorted input.

 

[dc655321]

Here's the noise spectrum of the Purifi. I didn't measure the total voltage in the audio band to compare with your numbers, but I'd be happy to do so later today.

View attachment 74111

Nice! Anything interesting to see at 20-80Hz? The plot is a bit truncated.

 

[SIY]

Nice! Anything interesting to see at 20-80Hz? The plot is a bit truncated.

Not much, other than tiny amounts of pickup from the test leads.

 

[Panelhead]

I only keep amplifiers that pass the no noise with ear in tweeter test. Have had several, including one tube OTL. Rating was -107 dB S/N. It was quiet.
The Purifi amp nor should not show the main frequency harmonics. Switching supplies and resonant mode supplies do not show the mains frequency.
If the OP and others can live with noise at close distances it is fine. But it is not unavoidable. There are reasonable priced components that are silent.

 

[donsna]

If the OP and others can live with noise at close distances it is fine. But it is not unavoidable. There are reasonable priced components that are silent.

Initially, I wanted to understand the main reason (root cause) of the noise. If it is acceptable or not, is a different and subjective question.
I personally agree with you: I prefer quite equipment (my Lyngdorf e.g. is absolutely silent, means I cannot hear any noise from speaker or device), so as indicated in my last (summarizing) post, one of my conclusions is that in the next weeks/months I will look for other implementations... that's for me part of the hobby, even if this "try and error" sometimes can get a little expensive, I like it somehow...

 

[Panelhead]

Initially, I wanted to understand the main reason (root cause) of the noise. If it is acceptable or not, is a different and subjective question.
I personally agree with you: I prefer quite equipment (my Lyngdorf e.g. is absolutely silent, means I cannot hear any noise from speaker or device), so as indicated in my last (summarizing) post, one of my conclusions is that in the next weeks/months I will look for other implementations... that's for me part of the hobby, even if this "try and error" sometimes can get a little expensive, I like it somehow...

I diy amps too. Isolation between power supply ground and signal ground lowers this grit or hash. Something like 10 ohms is fine.
SIY, TOMC, and March must use these buffering resistors to lower noise floor. If I can build an amp that has no noise on 100+ horns, anyone can.

 

[Panelhead]

Yes, I did not read it carefully. Maybe looking at the DAC specs you can tease out whether it is contributing. Seems likely that the noise would be mostly from the amp.

Whatever the measurements, the OP said the hiss is not audible a foot away from the speakers. So I wouldn't worry about it in this case.

But the test was with NuPrime DAC-10H + ST-10. This fig. is THD+N so you can't tell what the actual noise level is based on that.. They got a SNR of 80db (A) with the combo.

Well, a SNR at 80 dB is terrible. Most dacs are way over a 100 dB, so that points to the amp. As far as resolution less than 14 bit. This is SET territory.
People argue that 16 bit is enough resolution. I want way more, at least 20. But 13.4 bits is not enough.
Not to just throw turds, layout, parts sizing, circuit design all play into noise floor. Too many get it right to be unachievable.