SMSL D-6s Balanced DAC Review

[HendersonD]

I'm ready to play a pound on it... it's definitely better because it's not the purchase price that counts but the technologies implemented and in this the D-6s is cutting edge.
Before I had a Teac UD-H01 that with the same streamer did not detect 24/192 even though it was compatible.
This is to say that technology is important...

Thanks, certainly the price is right to give this DAC a go, I do think it will best my 11 year old Meridian Explorer

 

[nanook]

Meridian certainly has made the most out of this chip. But of course the PCM5102 is not comparable to the ES9039q2m and the D-6s is a well executed design.

Yes, there are pops from time to time between different songs, but they are not loud at all in my system.
WiiM Pro - > Coax input of D-6s ; XLR to active speakers, D-6s volume 93 (3dB att., this allows to drive the speakers to full volume) ; volume control via WiiM. So most of the attenuation is done ahead of the D-6s.

 

[nanook]

And whether the difference between the Meridian and the D-6s is audible?
Maybe, maybe not...

Anyways, the Bluetooth LDAC capability of the D-6s is a really nice thing.

 

[MI_FI]

Meridian certainly has made the most out of this chip. But of course the PCM5102 is not comparable to the ES9039q2m and the D-6s is a well executed design.

Yes, there are pops from time to time between different songs, but they are not loud at all in my system.
WiiM Pro - > Coax input of D-6s ; XLR to active speakers, D-6s volume 93 (3dB att., this allows to drive the speakers to full volume) ; volume control via WiiM. So most of the attenuation is done ahead of the D-6s.

the light "pops" are related to the recordings of the songs, not perceptible with less sensitive DACs

 

[MI_FI]

Has anyone updated the firmware?
Improvements?
Problems?
Is it worth it?

there is always only the firmware "SMSL D-6S_MCU v1.2_241008" on the site which partially corrects the numbering of the filters. In any case, I updated it immediately as soon as I received it.

 

[MI_FI]

I wanted to give some other comparison ideas on the d-6s. My configuration is: MAC MINI M2 - D-6S - REGA IO - JBL STAGE A130 -SUBWOOFER VELODYNE MINIVIEE X. Before the MINIVIEE X I used a JL AUDIO D110 (currently retired). I bought the amplifier before the ASR review, if I had seen it I probably wouldn't have bought it. In any case, the idea is the following: in my modest listening space (a studio of around 16 square meters with a sloping ceiling and an average height of 2 metres) since I got the D-6S I no longer feel the need to use the subwoofer (and it is something I never thought would happen). The REGA IO's knob is at half way, while the DAC's is at 99db. I check the volume on the Mac and it is at minimum. As far as I'm concerned, this DAC pairs very well with the REGA IO, the bass is present, controlled and dry at this listening volume. I pushed the REGA IO to the maximum, after reading the review here on ASR, and I found only one song from my collection (Lady Gaga - Dance in the dark - flac at 44.1 kHz 16 bit) which causes the amplifier causing (at maximum volume) significant distortion on the speaker driver (subwoofer off). I make both subwoofers work at 45 Hz, the miniviee to get these bass. I wanted to compare myself with those who use class D amplifiers. I wanted to try some other amplifiers for comparison, but I saw that all the amplifiers sold on Amazon and reviewed here have overheating and reliability problems, while the REGA IO, classified as an oven , for me it doesn't get so hot that I can't put a hand on the case, I listen in the evening for about 3 hours on average. I have seen that the speakers manage a maximum of 125W of power, so I would only take into consideration comparisons with amplifiers not exceeding 125W into 6 ohms, which have no problems and above all which do not have noises such as the famous POPs. Let me start by saying that I have already used class D amplifiers (Coktail Audio X40 which left me this year after 9 wonderful years on my main system, replaced by the Eversolo DMP A6).

Given the numerous responses received, I'll answer myself. I bought an Aiyima A70 so as to make a direct evaluation between a class D amplifier that measures well, and a class AB one that measures poorly (Rega IO). As soon as I finish my evaluations I will post my considerations on a dedicated post. Both connected to the D-6S, with the A70 used with both RCA and XRL output (which is the other point I want to evaluate).

 

[HalSF]

I just hooked up my new D-6s — it sounds great so far — and I wonder if I could ask for some help understanding how the DAC's display of the sample rate works.

My source is an M4 Mac Mini and I set the Mac's sound system settings for the DAC at 2 ch 24-bit integer 96.0 kHz because I play a lot of 24/96 files and they're the highest-resolution files I play.

The D-6s display reads 96.0 and stays there even when I'm playing, say, a 16/44.1 file.

Is it safe to assume that the DAC is designed to display the Mac's sound settings, not the audio sample rate of the music I'm playing? And is there any reason for me *not* to leave the Mac Audio MIDI set at 24/96 for this DAC?

Apologies for this pretty basic question, but the D-6s manual and documentation is quite minimal and this is the first DAC I've owned with a display.

 

[MI_FI]

I just hooked up my new D-6s — it sounds great so far — and I wonder if I could ask for some help understanding how the DAC's display of the sample rate works.

My source is an M4 Mac Mini and I set the Mac's sound system settings for the DAC at 2 ch 24-bit integer 96.0 kHz because I play a lot of 24/96 files and they're the highest-resolution files I play.

The D-6s display reads 96.0 and stays there even when I'm playing, say, a 16/44.1 file.

Is it safe to assume that the DAC is designed to display the Mac's sound settings, not the audio sample rate of the music I'm playing? And is there any reason for me *not* to leave the Mac Audio MIDI set at 24/96 for this DAC?

Apologies for this pretty basic question, but the D-6s manual and documentation is quite minimal and this is the first DAC I've owned with a display.

you must set exclusive management of the DAC on the player you use (if it lets you do so), while on the midi configuration set the maximum sampling value.

 

[staticV3]

Is it safe to assume that the DAC is designed to display the Mac's sound settings, not the audio sample rate of the music I'm playing?

For a DAC, only the sample rate of the incoming audio data exists.

It cannot differentiate between the sample rate of the underlying audio track and the sample rate of the final audio stream sent by the operating system.

By default, macOS will resample all audio to one uniform sample rate, that you specify in the Midi control panel.

There are ways to assure that audio is played at its native sample rate though, like for example using the LosslessSwitcher application.

 

[Esprit]

Try Audirvāna

 

[HalSF]

you must set exclusive management of the DAC on the player you use (if it lets you do so), while on the midi configuration set the maximum sampling value.

Thanks for the reply. It sounds like I'm set with the MIDI configuration.

But I'm not sure about the first part of your answer regarding setting exclusive management of the DAC on the player I use.

I use the MacOs Sequoia Music app for the vast majority of my listening (and a minimal player app called Tiny when I play FLAC files). I'm completely in the dark about whether or how I can set exclusive management of the DAC in the Music app — I'm not seeing anything like that in the app settings.

 

[HalSF]

For a DAC, only the sample rate of the incoming audio data exists.

It cannot differentiate between the sample rate of the underlying audio track and the sample rate of the final audio stream sent by the operating system.

By default, macOS will resample all audio to one uniform sample rate, that you specify in the Midi control panel.

There are ways to assure that audio is played at its native sample rate though, like for example using the LosslessSwitcher application.

Many thanks for this. Looks like I can easily set the appropriate native sample rate on the fly and have it displayed on the D-6s DAC using Rogue Amoeba's excellent SoundSource menu app.

Out of curiosity, is there a downside to having a 44.1 file resampled to 96.0 by macOS in terms of sound quality?

 

[Esprit]

I'm completely in the dark about whether or how I can set exclusive management of the DAC in the Music app — I'm not seeing anything like that in the app settings.

You can't, without third party software there is no way to make the frequency change automatically in MacOS.

 

[nanook]

I have played with pkane's Multitone tool this afternoon and the current version runs extremely stable with my setup.
--> Thank you so much @pkane !

Setup was DAC(USB input, 48kS, Dither 24bit) - APU(notch 1kHz, Gain 26dB) - ADCiso(Range 3v5, 192kS, Mono).
The APU notch was in place because the ADC otherwise makes a distince hump itself.

Multitone:
I took the default THD-vs-Level test and just modified the FFT length to 1M.
I did load Ivan's calibration for the APU notch 1kHz, 192kS (APU_1kHz_26dB_192kS.cal)
The graphs are plotted along the measured level. (next time I would try to plot them along drive level)

Measured DUTs:
SMSL D-6s (1x ES9039q2m)
SMSL DO400 (1x ES9039Pro)
SMSL PO100AK (1x AK4493, USB-bus-powered)

1) D-6s, L, Vol: 99 (Dithering at 20bit did not make a difference)
2) D-6s, L, Vol: 95 (just slightly shifted)
3) D-6s, L, Vol: 58 (just shifted)
4) D-6s, R, Vol: 95 (hump is smaller)
5) DO400, L, Vol: 0dB
6) DO400, R, Vol: 0dB
7) PO100AK, L
8) PO100AK, R
9) D-6s, L, Spectrum @ -50dBFS
10) D-6s, L, Spectrum @ -40dBFS
11) D-6s, L, Spectrum @ -20dBFS

Can someone falsify or confirm this strange behavior I observed on the SMSL D-6s (1x ES9039q2m)?
(It should suffice to measure the spectrum of a 1kHz sine wave at -40dBFS and -50dBFS)
--> Is this just my unit or is this the typical behavior of the ES9039q2m?
--> Is my measurement flawed?

The DO400 and even the tiny bus powered PO100 (1x AK4493) were very well behaved.

EDIT: Just to avoid any misunderstanding - yes it surely is inaudible.

 

[trungdtmc]

I have played with pkane's Multitone tool this afternoon and the current version runs extremely stable with my setup.
--> Thank you so much @pkane !

Setup was DAC(USB input, 48kS, Dither 24bit) - APU(notch 1kHz, Gain 26dB) - ADCiso(Range 3v5, 192kS, Mono).
The APU notch was in place because the ADC otherwise makes a distince hump itself.

Multitone:
I took the default THD-vs-Level test and just modified the FFT length to 1M.
I did load Ivan's calibration for the APU notch 1kHz, 192kS (APU_1kHz_26dB_192kS.cal)
The graphs are plotted along the measured level. (next time I would try to plot them along drive level)

Measured DUTs:
SMSL D-6s (1x ES9039q2m)
SMSL DO400 (1x ES9039Pro)
SMSL PO100AK (1x AK4493, USB-bus-powered)

1) D-6s, L, Vol: 99 (Dithering at 20bit did not make a difference)
2) D-6s, L, Vol: 95 (just slightly shifted)
3) D-6s, L, Vol: 58 (just shifted)
4) D-6s, R, Vol: 95 (hump is smaller)
5) DO400, L, Vol: 0dB
6) DO400, R, Vol: 0dB
7) PO100AK, L
8) PO100AK, R
9) D-6s, L, Spectrum @ -50dBFS
10) D-6s, L, Spectrum @ -40dBFS
11) D-6s, L, Spectrum @ -20dBFS

Can someone falsify or confirm this strange behavior I observed on the SMSL D-6s (1x ES9039q2m)?
(It should suffice to measure the spectrum of a 1kHz sine wave at -40dBFS and -50dBFS)
--> Is this just my unit or is this the typical behavior of the ES9039q2m?
--> Is my measurement flawed?

The DO400 and even the tiny bus powered PO100 (1x AK4493) were very well behaved.

EDIT: Just to avoid any misunderstanding - yes it surely is inaudible.

I only have the ADC and no APU so I'm not sure I can confirm this, but I'll try again when I have some free time.

 

[nanook]

I only have the ADC and no APU so I'm not sure I can confirm this, but I'll try again when I have some free time.

When I measured without APU I got a distinct hump also for the DO400 (and even for the AK4620 based RME FF UC) so I attribute this to the ADC (see attachment).
-> You might not really see this hump as clearly. In the spectrum you should actually see the lower harmonics coming up when going from -50dBFS to -40dBFS.

 

[JIW]

I have played with pkane's Multitone tool this afternoon and the current version runs extremely stable with my setup.
--> Thank you so much @pkane !

Setup was DAC(USB input, 48kS, Dither 24bit) - APU(notch 1kHz, Gain 26dB) - ADCiso(Range 3v5, 192kS, Mono).
The APU notch was in place because the ADC otherwise makes a distince hump itself.

Multitone:
I took the default THD-vs-Level test and just modified the FFT length to 1M.
I did load Ivan's calibration for the APU notch 1kHz, 192kS (APU_1kHz_26dB_192kS.cal)
The graphs are plotted along the measured level. (next time I would try to plot them along drive level)

Measured DUTs:
SMSL D-6s (1x ES9039q2m)
SMSL DO400 (1x ES9039Pro)
SMSL PO100AK (1x AK4493, USB-bus-powered)

1) D-6s, L, Vol: 99 (Dithering at 20bit did not make a difference)
2) D-6s, L, Vol: 95 (just slightly shifted)
3) D-6s, L, Vol: 58 (just shifted)
4) D-6s, R, Vol: 95 (hump is smaller)
5) DO400, L, Vol: 0dB
6) DO400, R, Vol: 0dB
7) PO100AK, L
8) PO100AK, R
9) D-6s, L, Spectrum @ -50dBFS
10) D-6s, L, Spectrum @ -40dBFS
11) D-6s, L, Spectrum @ -20dBFS

Can someone falsify or confirm this strange behavior I observed on the SMSL D-6s (1x ES9039q2m)?
(It should suffice to measure the spectrum of a 1kHz sine wave at -40dBFS and -50dBFS)
--> Is this just my unit or is this the typical behavior of the ES9039q2m?
--> Is my measurement flawed?

The DO400 and even the tiny bus powered PO100 (1x AK4493) were very well behaved.

EDIT: Just to avoid any misunderstanding - yes it surely is inaudible.

@nick_l44.1 did some measurements of the DL200 a while ago (See here, here and here). The below is IMD not THD for the headphone output in low gain.

 

[JIW]

Also, Archimago measured the E1DA #9039S using Cosmos APU, ADC and Scaler (Link). The black line is THD from second to ninth harmonic. There is also a distinct hump starting around -50 dBFS.

 

[nanook]

I measured my 2nd D-6s today. Same picture and - for whatever reason - again the left channel is the worse one again.
Sampling Rate: Sampling rate change from 48kS to 96kS does not make a difference.
Dither: I'm not sure anymore if I really applied dither to 20bit in one of the measurements in my last post - either I misunderstood the GUI or it doesn't work

This time measurements are plotted vs. generated level.
D-6s: Vol = 99
1) L, 48kS
2) L, 48kS, zoom
3) L, 96kS
4) R, 48kS

I'm just wondering, why I do not see something similar on the DO400. I thought the DAC core of the ES9039q2m and the ES9039Pro are the same.

 

[nanook]

I only have the ADC and no APU so I'm not sure I can confirm this, but I'll try again when I have some free time.

It's clearly visible in the spectrum without APU in my measurement.