Neutron HiFi DAC V1 Review

[dmitrykos]

@CedarX I do not know yet. My intention is to test and to check whether it is capable of sufficiently long Convolver. Taking into account that V1's performance allows 40-band 2-channel PEQ the hope becomes more realistic. There is similar task in Neutron Music Player

 

[pioterd]

engineers at Cirrus Logic about this issue?

The result was similar to RAA test, CS DAC chips always show the distortion.

It’s not the issue of CS chips, it’s the issue of „engineers” who cannot implement it well with all features. Maybe 130 pages PDF chip-dataSheet is too long for some of them.

RAA showed as well that some well done implementations do not have any issues, like Onix Alpha XI1 2XCS43198 + 2xSGM8262

 

[jkim]

It’s not the issue of CS chips, it’s the issue of „engineers” who cannot implement it well with all features. Maybe 130 pages PDF chip-dataSheet is too long for some of them.

RAA showed as well that some well done implementations do not have any issues, like Onix Alpha XI1 2XCS43198 + 2xSGM8262

Exactly.

Note that other than the Onix Alpha Xl1, there must be still some other CS431xx-based devices that do not exhibit the clipping in RAA's DAC Line-Out Multitone Test. I found the JCALLY JM30Pro tested by RAA is yet another case. The JCALLY JM20 MAX I recently tested also does not produce clipping of that kind.

By the way, the known clipping behavior from most of the DACs employing CS431xx cannot be handled by circuit design around the chip since it IS from the chip. Out of the three potential culprits---dynamic range / SNR enhancement, Class H mode, and noise shaping (all adaptive to signal levels)---ONLY the H mode can be defeated as an option. That is why I (and Roman at RAA) believe the distortion is due to the unintended, undesirable functioning of the Class H mode, AND can be resolved by proper firmware design (that disables the H mode).

EDIT. By the way, any further discussion on this topic should better be had in another, more appropriate thread. I am not going to make another post on this topic here.

 

[mc.god]

EDIT. By the way, any further discussion on this topic should better be had in another, more appropriate thread. I am not going to make another post on this topic here.

I think that the most suited place is the dedicated CS thread

A Thread Dedicated to Cirrus Logic CS43131

The USB-C Dongle Market Apparently CS43130 was a Wolfson chip. It was Hidizs / Tempotec / Meizu that first drew my attention to CS43131. Later in 2020 products with dual CS43131 such as iBasso DC03 began appearing. We’ve had these pages...

www.audiosciencereview.com

 

[dmitrykos]

RAA showed as well that some well done implementations do not have any issues, like Onix Alpha XI1 2XCS43198 + 2xSGM8262

To confirm that. Could you play Beats test file from my #304 post, record it and show Spectrogram with Audacity with magnification of timeline like on my screenshot, Just L channel is sufficient. You can magnify to have like 7 beats 3 from area with additional signal and 4 from are with just beats (it is in the middle of Beats file) - make it like my screenshots for easier comparison. My test DAC with 2xCS43198 + Amp (not SGM but it does not matter) shows gradient stripped pattern of the noise floor.

The whole story with CS chip is not about inability of developer reading specs it is about DAC chip which shows nasty noise pattern which to my opinion is not about Class H behavior but a workaround for the low frequency of CS DAC chip (ESS and AKM chips operate at 2x higher frequency) and too close noise from the noise shaper which is lowered by the chip's RMS method (which is causing high-frequency distortion if gain is raised rapidly) - thus we see gradient-like noise image on Spectrogram with beats only and light lines (high frequency distortion, I showed it on waveform) for every beat.

We can continue with this in dedicated CS thread as proposed by @mc.god, please poke me when you got measurement done.

 

[dmitrykos]

Firmware 57 brings new sonic performance for DAC V1.

ESS DAC chip supports compensation of non-linearity of the DAC's output resistors, namely 2-nd and 3-rd harmonics can be compensated. THD Compensation can't fully eliminate harmonics but suppresses them considerably if generally hardware design is correct and harmonics appeared as a result of non-linearity of DAC's internal output resistors. During experiments it was revealed that non-linearity is different based on the analog volume, it is also different for Current and Voltage output modes, 2Vrms or 1Vrms. Left and Right channel. Therefore THD Compensation in DAC V1 is now a vector of C2, C3 coefficients for every supported mode covering analog volume from 0 to -24 dB (below -24 non-linearity is not observed).

The obtained result is quite decent showing improved SNR and DNR in comparison with the initial implementation which was reviewed on ASR.

Results (note: captured with 16-bit ADC, we look at the power of 2-nd and 3-rd harmonics):

2Vrms, Current Mode



2Vrms, Voltage Mode (new starting from Firmware 56)



1Vrms, Current Mode



1Vrms, Voltage Mode (new starting from Firmware 56)

 

[d1m0n]

Firmware 57 brings new sonic performance for DAC V1.

For the optimized THD Compensation the release notes mentions:

….: make sure NConfigurator → Advanced tab → THD Compensation is not NONE or CUSTOM

So to use this I need to select the 32 or 600ohm preset?
So if for example my IEMs and HP were all between 10-150 Ohm. I just set and safe 32, right?

 

[dmitrykos]

So to use this I need to select the 32 or 600ohm preset?
So if for example my IEMs and HP were all between 10-150 Ohm. I just set and safe 32, right?

Yes, 32 preset applies output gain correction, so you won't get clipping with 32 Ohm headset but only if you listen at max volume. If volume is not max then 32 or 600 will work similarly and you can just use 600 (default).

 

[d1m0n]

got it thx.

For my sensitive IEMs I set the Output to 1V and reduced/limited the max vol via the slider accordingly as a precaution.
Is it then fine if I go till max here as I would assume I never reach the actual internal max volume you mentioning of the DAC?

 

[CedarX]

About the THD compensation… I believe the results are also somewhat dependent on the load—the HP/IEM. I think the further “away” (downstream) from the DAC, the less effect you can get from THD comp.

Have you evaluated this? How much dependent is the optimum THD comp. from the HP/IEM impedance characteristics?

 

[dmitrykos]

For my sensitive IEMs I set the Output to 1V and reduced/limited the max vol via the slider accordingly as a precaution.
Is it then fine if I go till max here as I would assume I never reach the actual internal max volume you mentioning of the DAC?

For 1Vrms mode DAC V1 does not apply output gain correction because it does not clip, clipping is only a matter of 2Vrms mode. So if comfortable volume for you is with 1Vrms mode and you can get closer to max volume then it is safer for ears of course, just in case something happens with software and it suddenly sets max volume (with too powerful output it can have consequences for hearing I believe).

 

[dmitrykos]

About the THD compensation… I believe the results are also somewhat dependent on the load—the HP/IEM. Have you evaluated this?

ES9219 DAC chip has integrated OPAs (amplifier stage) which act as a buffer effectively isolating DAC's internal output circuitry from the load therefore there is no direct effect of impedance on THD Compensation. The reason why 32 and 600 presets exist is because preset for 32 Ohm load applies volume trimming (Master Trim) in order to avoid clipping in 2 Vrms mode and this change in volume changes 2nd and 3rd harmonics because it changes voltage between DAC's output and its internal OPAs. That is why in Firmware 57 were covered all modes which affect output stage (2 Vrms - 1 Vrms, Current or Voltage) and the whole analog gain range [0, -24]. For each volume value and output mode there is a separate THD Compensation coefficient (consisting of 2 values - for 2nd and 3rd harmonics) applied by firmware to the DAC chip.

 

[d1m0n]

Hello @dmitrykos,
I have some improvement / wishes. Maybe would be possible to implement this in the future?

- Minimum phase fast and slow filter in NConfigurator presets like you did with the linear types
- Larger display brightness range for the slider, so that the display can be set even darker than now when it is set to the smallest level.

Thank you very much
BR

 

[dmitrykos]

- Minimum phase fast and slow filter in NConfigurator presets like you did with the linear types

Noted. Meanwhile, for the fast minimum phase filter you can use DAC chip's built-in FAST_ROLLOFF_HYBRID:

- Larger display brightness range for the slider, so that the display can be set even darker than now when it is set to the smallest level.

Currently smallest value in NConfigurator corresponds to the lowest brightness of the OLED display, making it more dimmed is not possible.

 

[OHtaru]

Hello @dmitrykos,
Does the DAC support ASIO? Didn't see any information on the website. I want to use the DAC with DAW for music production.

 

[dmitrykos]

Does the DAC support ASIO? Didn't see any information on the website. I want to use the DAC with DAW for music production.

ASIO driver is not provided but you can use FlexASIO (https://github.com/dechamps/FlexASIO) configured to use Exclusive WASAPI. WASAPI backend is provided by PortAudio (https://github.com/PortAudio/portaudio/) and it is currently very stable and allows low-latency operation.

Could you clarify why you need ASIO for USB DAC - low latency, bit-prefect output, specific formats, software supports only ASIO?

 

[Egoquaero]

I saw that there are no physical buttons on the device? There's no way to change Eq preset on the go?
My setup: iPhone 16, Tidal, Xenns Top Pro iem
What's the fastest way to change EQ preset on iOS?
Cheers

 

[dmitrykos]

What's the fastest way to change EQ preset on iOS?

So far there is no fastest way for iOS, only via connection to PC. On Android it became possible recently with mobile/web version of NConfigurator because Android OS has no limitation in relation to USB device access by the app.

But Apple, I do not know the reason behind it, does not allow apps accessing USB device, at least HID access would do the trick in case of DAC V1 (it uses HID interface for NConfigurator needs). If they change this policy NConfigurator will appear on iOS immediately.

 

[CedarX]

But Apple, I do not know the reason behind it, does not allow apps accessing USB device, at least HID access would do the trick in case of DAC V1 (it uses HID interface for NConfigurator needs). If they change this policy NConfigurator will appear on iOS immediately.

Speculating a little... I suspect Apple reason(s) are the same as the Mozilla Foundation: it is potentially opening a backdoor into the web browser, as the system doesn't really know what kind of USB device is connected and whether it could be malicious...
Although there is a potential security risk, the net effect is negate the use of any WebApp to communicate not only with audio devices, but also a myriad of industrial and medical devices which is, IMO, an unfortunate and excessive restriction...

 

[Egoquaero]

@dmitrykos Thank you so much for your reply. It's so great to see your direct involvement in here. I can feel your passion.
It's such a shame that I can't find my perfect dac/amp right now.
Something with hardware PEQ, bluetooth and enough power to drive HD800 with EQ effortlessly. Am I asking that much?
Basically it was the Qudelix T71, but for some reason they decided not to implement bluetooth capabilities.
Or the Topping G5 (which I own), but it lacks PEQ functionalities.
I went on buying a FiiO BTR17, which was perfect on paper, but boy.. This is quite a frustrating device. Plenty of bugs and the PEQ capabilities are so basic. And the power is just a tiny bit on the short side when EQ is applied to the HD800. I'll have to return it.
I guess I'm waiting for the Neutron Hifi DAC V2/MAX
You guys know any other device that might fit the bill right now?