NAD C 389 residual noise and first experience

[Randolf]

Hi,
I purchased a NAD C 389, MDC2 BluOS-D Modul and DiracLive full bandwidth license a few weeks ago. It should replace my current setup WiiM Pro->MinDSP DDR-24 (DiracLive)->Yamaha R-N803D. On the one hand I like der C 389 a lot, but I found it's residual noise rather disappointing, especially since I used high sensitivity Speakers (Lowther Fidelio with DX-3, rated ~ 96dB 2.83V/1m). It has a clearly audible hiss even at 3.3m listening distance. The Yamaha R-N803D (70uV residual noise) and also my vintage Yamaha A-590 (35uV (CD Direct) / 90uV (Pure Direct) residual noise) from 1996 have a practically inaudible residual noise/hiss in my setup.
Since I cannot find any residual noise numbers for the NAD C 389 I decided to measure it. I do not have access to uV precision measurement gear so I just performed a simple acoustic near field measurement (12.5cm away from the Lowther Fidelio) of all amplifiers and without any amplifier at all (to get a baseline of the general environmental noise to compare with). The result is:

We can clearly see the C 389 has a significantly higher noise floor above 1 khz than any of the Yamaha amplifiers in any mode.

I found some residual noise measurement on similar NAD amplifiers:
https://www.soundstagenetwork.com/i...ts&catid=97:amplifier-measurements&Itemid=154 (2 NCORE Hypex)
Noise level (A-weighted) <55uVrms <55uVrms
Noise level (unweighted) <82uVrms <79uVrms

https://www.soundstagenetwork.com/i...ts&catid=97:amplifier-measurements&Itemid=154 (1 Ucd Hypex)
Noise level (with signal, A-weighted) <130uVrms <132uVrms
Noise level (with signal, 20Hz to 20kHz) <166uVrms <167uVrms
Noise level (no signal, A-weighted) <102uVrms <97uVrms
Noise level (no signal, 20Hz to 20kHz) <130uVrms <124uVrms

https://www.soundstagenetwork.com/i...ac&catid=97:amplifier-measurements&Itemid=154 (1 NCORE Hypex)
Noise level (A-weighted) <190uVrms <180uVrms
Noise level (unweighted) <310uVrms <280uVrms

So only the C 399 seems to perform well here, I am wondering what the C 379 and C 389 would measure ( I guess similar to C 3050).

Most reviewers describe the NAD amplifiers as very silent:

only

and
https://hometheaterhifi.com/reviews/amplifier/integrated-amplifiers/nad-c-399-hybrid-digital-dac-amplifier-review/ "The first thing I noticed is the noise floor – as in, there isn’t one. This is the quietest integrated amp I’ve yet heard. Even the superb M28 multi-channel amp I reviewed a while back had an audible noise floor. Previous Hypex-based NAD products I’ve reviewed also had audible noise when no signal was present. The C 399 has somehow eliminated this."

notice a potential residual noise issue.

Maybe only the C 399 has a low residual noise suitable for high sensitivity speakers in the NAD product line. Any good or bad experience with C 379, C 389, C 399, C 3050, M10 V2/3 and high sensitivity speakers? I would really like to see residual noise data in any amplifier measurement.

Beside the residual noise issue I mostly like the C 389, minor issues I found

1. Pseudo 2 speaker pair terminals, you cannot use it to compare or just even switch between 2 loudspeaker pairs. In the last 45 year I have never seen such a strange feature, basically just a Bi-wiring helper. The Yamaha R-N803D in contrast allows seamless switching between 2 speaker pairs and is able to activate the corresponding YPAO calibration and level adjustment while switching, which is great to compare speakers.
2. The remote control operates a little bit like controlling 2 different devices (Amp and MDC modul) in one box by having to switch between both via “AMP” and “BLS” buttons, so e.g. switching from TV to Internet radio input is somewhat complicated.
3. After power up BluOS is not able to play where is had previously stopped. So if you want to continue listening to the same internet radio station as you did before shutting down the device you have to select the station again.

Currently I am not sure if I will keep the C 389 or resell it.

Best regards
Randolf

 

[Old_School_Brad]

You seemed to have the perfect setup before -what made you decide to change it?

 

[Randolf]

You seemed to have the perfect setup before -what made you decide to change it?

Valid Question, only minor issues with the old setup:

• The PCM3168A DAC of the DDRC-24 is not really state of the art.
• The headphone jacket of the Yamaha R-N803A has very high output impedance which is bad for driving my low impedance headphones (although I found a workaround for that: https://homeaudio.jimdofree.com/headphones/)
• I had problems with LG TV optical out via WiiM Pro and DDRC-24. LG TV SPDIF is known to have high jitter and WiiM at least initially had serious issues with optical input (got better after various updates). This resulted in somewhat distorted TV sound.
• I sometimes had the impression the sound is not always ok but couldn't really find the root cause (Yamaha is 7 years old, Minidsp just 2 but already had to be repaired once)
• Esthetics: 2 black plastic boxes + a nice silver vintage amplifier vs. single black brick

The TV sound seems to be ok now with the NAD and I somehow have the subjective impression that the C 389 sounds better in general, but unfortunately I have not yet done a direct A/B comparison via an amplifier switch box (the NAD seems to have some detection circuit which disables output if no speaker are connected, strange, i need to investigate). If the Yamaha R-N1000A would have DiraceLive instead of YPAO this would be my choice, unfortunately it hasn't.

 

[Randolf]

When switching from db to voltage scale and looking add the relative difference to the Yamaha measurements I would estimate a residual noise of about 170 uVrms A-weighted for the C 389.

 

[Old_School_Brad]

If I calculate the residual noise from the speakers, assuming 4 ohms and 170 uV, it comes out to about 15 dB @ 1 meter. However, you mention it’s audible from 3 meters, so something might not be adding up.
The residual noise from the NAD isn’t great, but it’s not terrible either. That said, it’s clearly not ideal for your setup.

Yamaha for the win..

 

[AnalogSteph]

We can clearly see the C 389 has a significantly higher noise floor above 1 khz than any of the Yamaha amplifiers in any mode.

As an aside, what did you use for a microphone? If that is calibrated as suggested by the presence of a mic cal file (and not some random dynamic vocal mic), then the speaker frequency response would seem to be quite concerningly nonlinear. While an on-axis treble tilt is quite expected for a wideband speaker (a large one at that), ~2.1-2.5 kHz seems a bit much and the 3-4 kHz region has a major dip.

 

[Sokel]

As an aside, what did you use for a microphone? If that is calibrated as suggested by the presence of a mic cal file (and not some random dynamic vocal mic), then the speaker frequency response would seem to be quite concerningly nonlinear. While an on-axis treble tilt is quite expected for a wideband speaker (a large one at that), ~2.1-2.5 kHz seems a bit much and the 3-4 kHz region has a major dip.

Judging by the SPL I think that what we see is room noise plus the hiss of each amp,no signal,at all.
One of them is with no amp at all.
I wonder what this 100Hz spike is.

 

[Old_School_Brad]

Judging by the SPL I think that what we see is room noise plus the hiss of each amp,no signal,at all.
One of them is with no amp at all.
I wonder what this 100Hz spike is.

A slightly lower harmonic of the 50 Hz mains frequency?

 

[Sokel]

A slightly lower harmonic of the 50 Hz mains frequency?

Acoustically???That's underlying hum if it's that,grounding problems probably,it must be addressed.

 

[Old_School_Brad]

Acoustically???That's underlying hum if it's that,grounding problems probably,it must be addressed.

That’s what it seems.

 

[Randolf]

Hi,
I was also surprised by the 50Hz and 100Hz (1. upper harmonic) peaks. They even exist in the base line (green, amplifier turned off) measurement. Since the measurement was done using a laptop running on battery power and the UMIK-1 it probably does also not come from the measurement gear itself and seems to be real environmental noise perhaps from the neighbours apartment. Anyway it is not audible, no hum just hiss. So we can concentrate on the result above 1 khz. The low frequency section of my measurement is also not valid since the Lowther Fidelio is a backloaded corner horn design, so you cannot measure it's base response 12.5 cm away from the driver. In general the entire measurement it not really accurate since it also includes the rather uneven frequency response of the loudspeaker. Only the delta between the different amplifiers and the idle environment should be considered.

In general residual noise becomes audible only if it exceeds the human hearing threshold at the listening position, which is defined by the threshold of Fletcher–Munson curve and its refinements:

So in most practical cases only hiss and not hum becomes an issue with residual noise. Therefore it is usually valid to just look at the a-weighted voltage (+ speaker sensitivity + listening distance) to identify if residual noise may become audible or is just of no practical impact at all. For average sensitivity speakers it is usually never an issue.

Residual noise typically originates from any electronics after the volume control (including the volume control itself, often an electronic device today).

The residual noise of NAD C 389 is not terrible (otherwise I would have returned it immediately) but it is also not good. Many Yamaha amplifiers are usually good here and Yamaha publishes residual noise in any device manual. Usual the Yamaha spec lines up with my experience, but for some low end devices it doesn't. E.g. R-S202D, R-N303D, R-N402D are rated 70uV, just the same as the R-N803D. However I know that these devices have a clearly audible hiss on my Lowther Fidelio in contrast to the R-N803D. Some amplifiers like the R-N402D use an additional mute switch, which gets activated when the volume is turned down completely or a digital 0 source signal is amplified. This can fake a good residual noise level, which gets a lot worse once you only slightly turn up the volume or the music starts. Maybe this explain the low residual noise Yamaha spec for this device. In the power on/off graph of https://www.audiosciencereview.com/forum/index.php?threads/yamaha-r-s202-receiver-review.52587/ we can see ~500uV idle voltage. Not sure how that is exactly measured but it somewhat matches with my experience running a R-N402D on the Lowther Fidelio, which was really bad. You can also guess from the technical details of the R-N402D that it must be significantly worse than a R-N803D. It's volume control chip BD3491FS is rated 5 uV residual noise, which gets amplified by the power amp section and is 5 times higher than the one used by the R-N803D (BD34703KS2).

The good news is, the majority of people run average sensitivity speakers and usually do not need to care about residual noise. For those probably almost any NAD amplifier is just fine.

 

[restorer-john]

OK. Let's start doing this properly or not at all.

Residual noise in amplification is measured (electrically, not acoustically) with a nominal load, a specific bandwidth, a weighting (or not) and controls set at particular positions as well as input conditions considered (shorted/open etc).

If you are to actually measure residual noise, across a specific bandwidth, you also need to consider the contribution of the input device itself, and remove that from the calculations or at least take it into consideration. D/A converters, regardless of their excellence will contribute considerable 'noise' especially if they are implementing digital attenuation.

 

[AnalogSteph]

Since the C389 is an integrated amp like the others, I would assume that the OP would have turned its volume pretty far down for a residual noise test... right? That would address any DAC noise level concerns.

The NAD's paper specs suggest a residual noise level of about 50 µV (that's what -95 dB ref. 2.83 V out translates to), so nothing concerning there. It could still be a settings issue of some kind, or the unit is just a dud. Actual residual noise level seems to be more like 140 µV (going by the AX-590), missing the spec by over 9 dB. In other words, you would normally expect it to be roughly on par with the R-N803D.

 

[Randolf]

OK. Let's start doing this properly or not at all.

Residual noise in amplification is measured (electrically, not acoustically) with a nominal load, a specific bandwidth, a weighting (or not) and controls set at particular positions as well as input conditions considered (shorted/open etc).

If you are to actually measure residual noise, across a specific bandwidth, you also need to consider the contribution of the input device itself, and remove that from the calculations or at least take it into consideration. D/A converters, regardless of their excellence will contribute considerable 'noise' especially if they are implementing digital attenuation.

I fully agree, but measuring it electrically requires some expensive measurement gear like an Audio Precision analyzer. The rest of the requirements are actually fulfilled, I usually use shortened RCA input or not connected digital input and volume turned down completely (which usually means the chosen input is not relevant). The residual noise of the C 389 is effectively the same regardless which input you choose as long as the volume is turned down or even Mute is pressed. Maybe it's accurately measured residual noise is 150uV or 200uV and not my delta estimated 170uV. Maybe someday Amirm or soundsatgenetwork will measure it more accurately. My current assumption is that C379, C 389, C3050, M10V2, C700V2 will all operate in that region and only the C 399 might be an significant really state of the art improvement when it come to residual noise.

 

[Randolf]

Since the C389 is an integrated amp like the others, I would assume that the OP would have turned its volume pretty far down for a residual noise test... right? That would address any DAC noise level concerns.

The NAD's paper specs suggest a residual noise level of about 50 µV (that's what -95 dB ref. 2.83 V out translates to), so nothing concerning there. It could still be a settings issue of some kind, or the unit is just a dud. Actual residual noise level seems to be more like 140 µV (going by the AX-590), missing the spec by over 9 dB. In other words, you would normally expect it to be roughly on par with the R-N803D.

Ooops, cannot find it in the NAD specs / manual, can you provide a link to it? If there are NAD specs we could compare it with the various NAD measurements on https://www.soundstagenetwork.com/index.php?option=com_content&view=article&id=17&Itemid=141 and have a look at the differences between C 379, C 389, C399, C3050, M10V2, C700V2 in the NAD specs.

 

[Randolf]

Ooops, cannot find it in the NAD specs / manual, can you provide a link to it? If there are NAD specs we could compare it with the various NAD measurements on https://www.soundstagenetwork.com/index.php?option=com_content&view=article&id=17&Itemid=141 and have a look at the differences between C 379, C 389, C399, C3050, M10V2, C700V2 in the NAD specs.

I think I got your point NAD specs say:

LINE IN, SPEAKER OUT (ANALOG BYPASS ON)
C700: Signal-to-Noise Ratio>84 dB (A-weighted, 500 mV input, ref. 1 W out in 8 ohms)
C3050: Signal-to-Noise Ratio>85 dB (A-weighted, 500 mV input, ref. 1 W out in 8 ohms)
M10V2: Signal-to-Noise Ratio>85.1 dB (A-weighted, 500 mV input, ref. 1 W out in 8 ohms)
C379: Signal-to-Noise Ratio>85 dB (A-weighted, 500 mV input, ref. 1 W out in 8 ohms)
C389: Signal-to-Noise Ratio>95 dB (A-weighted, 500 mV input, ref. 1 W out in 8 ohms)
C399: Signal/Noise Ratio>95 dB (A-weighted, 500 mV input, ref. 1 W out in 8 ohms)

From this one would calculate (although I don't know what 500 mV input means, 1 W out in 8 ohms equals to to 2.83V):
C700: 178uV
C3050: 159uV
M10V2: 159uV
C379: 159uV
C389: 50uV
C399: 50uV

However soundstagenetwork measurements are:
C700: 240uV
C3050: 130uV, 97uV
M10V2: 180uV
C399: 55uV

Some equally rated devices according to NAD specs result in rather different soundstagenetwork measurement, but at least in the same range. But if NAD specs are accurate this would mean my C389 is already the best I can get from NAD, upgrading to C399 makes no sense than. It would also mean Yamaha specs are measured differently and I cannot get anything close to mid price Yamaha from NAD. If this is actually true I am pretty disappointed by NAD.

I did not put the C 389 in analog bypass, maybe this explains the difference, I will give it a try. But I turned the volume completely down pressed even Mute (no difference). And since all my sources are digital and I need DiracLive "analog bypass" is useless for me anyway.

 

[Randolf]

As expected using a line input with analog bypass enabled and short cutting the input doesn't improve the residual noise with the volume turned down completely. I have send a service request to NAD, I guess the C 389 spec is simply wrong and it's real residual noise is somewhere in the 150-200uV range like c3050, C379 or M10V2. This would match with my measurements. I will post NAD's feedback here.

 

[Old_School_Brad]

@Randolf

I’m having trouble understanding your in-room measurements above.

• Why does the noise floor of your room show such a significant rise in low frequencies?
• What is the source of the 50 Hz and 100 Hz tones? You mention they aren’t audible, but they appear to be around 35 dB SPL. If I play a 50 Hz or 100 Hz tone at 35 dB SPL, it would be clearly audible to me. Annoyingly so, even.

 

[pogo]

What is the source of the 50 Hz and 100 Hz tones?

These will be disturbances within the measurement setup and will not be present in reality.

 

[AnalogSteph]

@Randolf

I’m having trouble understanding your in-room measurements above.

• Why does the noise floor of your room show such a significant rise in low frequencies?

The UMIK-1 is kind of a noisy SOB. Condenser mics tend to have predominantly 1/f noise, and this is a small and cheap one with an inexpensive USB audio codec to boot.

• What is the source of the 50 Hz and 100 Hz tones? You mention they aren’t audible, but they appear to be around 35 dB SPL. If I play a 50 Hz or 100 Hz tone at 35 dB SPL, it would be clearly audible to me. Annoyingly so, even.

I have my doubts. 50 Hz at 35 dB SPL isn't even above the hearing threshold, and 100 Hz barely 10 dB. There is a lot more mains hum going around than you might realize, from all kinds of sources.