SMSL D300 Review (Balanced DAC)

[bogi]

Arrived yesterday, happy but scared to check it.

Please go to RESET menu and let us know what for firmware version you obtained with the new unit. Pictures on previous pages show how it may look.

 

[bogi]

I previously wrote that D300 is able of 48k based DSD rates (standard are 41k based). I tested them on 1.0 MCU fw and I am using them with HQPlayer
I am interested to know if 48k based DSD rates work with firmware 1.21 too.
The below dropbox link points to free CC BY 4.0 license content originally of 128k MP3, upsampled by mansr's SoX mod to 48k based DSD64, DSD128, DSD256.
https://www.dropbox.com/s/62i1kp3vj69yvfb/48xDSD-Test.zip?dl=0
Could somebody try if that content is playable with D300 MCU fw 1.21?
Please don't concentrate on sound quality (source is 128k MP3), but about ability to play it.

 

[meowski]

I've had my D300 since Nov 2021 from ShenzenAudio without issues (running Vol Variable using BT & OPT on XLR). Still on original firmware (v1.0).
This week it has started making horrible crackling noises/robot and the audio has been cutting out. Sometimes it works for 5 minutes without issue. No problem with speakers, tried them on another system, so must be the D300

Should I try the firmware/usb update?
From reading here it seems like fw 1.0 is the most stable, but some ppl had success with 1.11?

 

[bogi]

@meowski No other input cable (USB, coax) connected? No change in powering of your system was done, for example on amplifier side?
I would disconnect all input cables including optical and try with bluetooth if the issue persists.
Try to reproduce your issue with battery powered headamp, if you have one available. That would make sure that the issue has no relation to your amplifier.

I would not do firmware update in such a case. It may fail if your unit is defective.

​

 

[bogi]

I am interested to know if 48k based DSD rates work with firmware 1.21 too.

I already know the answer. 48k based DSD rates works correctly on fw 1.0 but don't work correctly on 1.21.
(this information is interesting only for HQPlayer users)

 

[meowski]

@meowski No other input cable (USB, coax) connected? No change in powering of your system was done, for example on amplifier side?
I would disconnect all input cables including optical and try with bluetooth if the issue persists.
Try to reproduce your issue with battery powered headamp, if you have one available. That would make sure that the issue has no relation to your amplifier.

I tested the unit on a different set of speakers and with a variety of playback devices.
It seems my issue only occurs when using Bluetooth or optical. I tried all setting combinations without success.

However, USB works flawlessly. I guess my unit is partially broken :/

 

[Jimster480]

I tested the unit on a different set of speakers and with a variety of playback devices.
It seems my issue only occurs when using Bluetooth or optical. I tried all setting combinations without success.

However, USB works flawlessly. I guess my unit is partially broken :/

What are you trying to play via Bluetooth?

 

[meowski]

What are you trying to play via Bluetooth?

Tried locally stored FLAC and MP3 from Android (with and without LDAC).
Works fine with my portable bluetooth speaker, but I still get horrible crackling on the D300.

Anyone know of any fixes for such an issue? Or am I out of luck

 

[Orlandino]

I want to share some observations and have several questions. I'm using my D300 via ASIO in Foobar2000 and had to install XMOS drivers for xu208 from SMSL site. I have the unit likely from early parties, and everything works smoothly.

Do you have any suggestions on the best way to configure the XMOS Control Panel for D300 in Windows 10?

1. "Format" menu gives variants 24 bits and 32 bits. I set 24 bits as a more common option and because of concerns that the stream will not be bit-perfect if the files are in 24 bits. Am I right, or 32 bit option here is better for this DAC performance, and means something different than DAC bit upsampling?
Interesting that there is no option for 16 bits.

2. "Buffer size" has default "Safe Mode" - is it good, or should we choose some manual buffer size here? The list of variants differs, and sometimes it has options up to 8192 samples. But I didn't notice that it was automatically selected higher than 4096, even for 384/32 or DSD files.

3. Should we always hold this Control Panel active in the Windows tray to get its options to work? Everything works smoothly even without it, both Foobar ASIO and system sound. So, I prefer when there are no unnecessary applications in the tray.

4. Windows Sound options for sample rate and bits have different lists depending on p. 1 settings - when we choose 32 bits in the XMOS Control Panel, there are 16, 24, 32 bit variants, and the custom one is 96/32. When I set XMOS to 24 bits, there are no 32 in the list (even if the panel is closed, so it also proves p. 3), and I set Windows sound option to 192/24 as more usual one.
Or is it better to select 32 bits in XMOS Control Panel (p. 1), and after it choose something different in Windows options than 192/24?

 

[bogi]

@Orlandino

Windows audio device settings are not used when ASIO exclusive access to device is used. With ASIO in function, you can have anything set in Windows and it will be ignored. It would be used with applications, which don't provide ASIO or WASAPI exclusive access.

Format 32 or 24 ... it does not really matter, since we don't have 32bit audio content available. But you could for example perform some DSP and output it at 32 bit fixed point resolution. So set it at 32 bit, some player DSP or VST plugins could utilize it. XMOS is 32 bit interface and ROHM BD34301EKV is 32-bit DAC chip.

Buffer size is automatically adapted for different input sample rates, but your setting position (for example 4th from top) is preserved. Set it as you want. On this forum it is not popular to tell that different settings could sound a bit differently, but you can try. Computer generated noise, whose possible influences are generally not enough recognized on this forum, can sometimes make a difference.

With very high rate input with some buffer settings you could get ocassional dropouts. I'm using only DSD256 and DSD512 with D300, my personal setting is 3rd from bottom (512 shown for 44.1k input) without the Safe Mode selected.

You don't need to have Asio Control Panel app started. ASIO driver as every driver si working without need to run any GUI - it runs in kernel space and not in process space. The Control Panel app is used only to show or change settings and does not need to run.

 

[Orlandino]

@bogi thank you for the detailed comment. So ASIO completely overrides Format options from the XMOS Control Panel according to the playing file's properties, and even if 32 bits set in XMOS CC but the file has 16 or 24 bits, it would stream bit-perfect to the DAC chip?

 

[bogi]

So ASIO completely overrides Format options from the XMOS Control Panel according to the playing file's properties

ASIO completely overrides Format options from the Windows audio device settings according to settings of your player. With bitperfect setting it is according to the playing file's properties. But you may want for example to upsample in player. So generally it depends on player settings.

even if 32 bits set in XMOS CC but the file has 16 or 24 bits, it would stream bit-perfect to the DAC chip?

Best DACs have about 20 bit resolution so you really don't need to care much about these bits. Everything at such low level appears below DAC analog noise floor.

Adding zeros at the remaining lowest bits does not change the original binary value and the bitperfect condition is preserved. These additional bits are helpful when DAC performs DSP operations like volume control. They ensure that rounding error will not affect resolution of audio content played (for example 24bit recording).

Simply set 32 bit in ASIO Control Panel and don't care more about that. It cannot make anything worse but it allows to be compatible with possible 32 bit output of some DSP or VST plugins.

 

[despoiler]

If the manufacturer erases the marking of the op-amp, it means that he has installed the cheapest op-amp, such as NE5532 or JRC2068.
At best it will be OPA2132...
Therefore, feel free to install all three OPA1612 op-amps and you will be happy!...
...And if you like the imitation of "warm tube sound", then you can put OPA1656.

I'm pretty sure my D300 is already using at least one OPA1612A in the stage just before the output. I cracked my unit open to check the resistors, they look like they match when looking at with my eyes, but on camera they don't. All opa have had their tops scrubbed. On one them is the characteristic yellow/tan dot on the 1612A and in the same position. What's crazy is the pictures previously posted have completely different op-amps. SMSL either has different revisions that they don't indicate to consumers or use whatever they can get their hands on. Either way, it's not good. I wonder if they do this in their other models? Though the newer models advertise using opa1612. I think the other two opa in my unit might be TI NE5532D based on the square silkscreen still showing.