SMSL DL200 DAC & Headphone Amp Review

[forart.it]

How does it compare to the C200?

Which one would you buy for the same price?

 

[olds1959special]

Keeping my DO100 Pro, selling my DL200, I don't need the headphone output, and the DO100 Pro is a dual DAC design that sounds better.

 

[olds1959special]

Nevermind, keeping my DL200 as a source instead of the Duet 3. It seems to sound smoother.

 

[Bleib]

How does it compare to the C200?

Which one would you buy for the same price?

I used to recommend C200 as the budget king but it's for instance not as greatly shielded against cellphones close by like DL200. Since DL200 is also bigger it runs cooler.
Don't know about C200 Pro tho, might improve on DL200?

 

[TheTalbotHound]

Would you be able to make a crosstalk vs frequency measurement for this DAC?

 

[audioje]

Didn't know about this review thread. My dl200 is still working flawlessly after two years of heavy use. Had some hiccups with tidal dac exclusive mode but a driver update solved the issue.

 

[tccalvin]

I've owned this for a while after reading the review. Very happy so far.

 

[tccalvin]

I wanted to add a couple of measurements that I was interested in.

I used a MacBook, the DL200's RCA out (44.1kHz) and the ADC of the UA Volt 2 interface (192kHz). I tried matching the levels as well as I could without making the signal clip. Frequencies shown are from 0dBFS to -120dBFS (this range includes a few dBs of ADC THD+N but makes things easier to see).

One of the reasons that I bought a DL200 was the presence of a digital volume control.

Digital volumes offer some advantages like precise channel matching, fine-grain level control and better linearity compared to analog potentiometers. But the advantage that I was most interested in was the ability to fix inter-sample clipping at the DAC level, since the ES9039Q2M chip offers the ability to lower the sample levels before they hit the DAC's oversampling filter. However...

Test 1: Inter-Sample Clipping

Track 1 (intersample.44k): The test file created by @danadam for testing DAC headroom.

Track 2 (98): DL200 RCA output with volume set to 98/99 (99 made the ADC go into overload protection).

Track 3 (75): DL200 RCA output with volume set to 75/99 (around 12dB quieter than 99).

Track 4 (intersample.44k.noclip): The same test file but normalized to -1dBTP via software (around 3.8dB quieter than the original test file).

Track 5 (98noclip): DL200 RCA output with volume set to 98/99 but this time with the normalized test file.

...it seems SMSL went out of their way to put another digital volume control after the SRC even though they clearly didn't need to...

As you can see, Track 3 (75) exhibits all the same distortion characteristics as Track 2 (98), even though it's several dB quieter than Track 5 (98noclip). This means that it is not possible to fix inter-sample peaks once the signal has hit the DAC. For this specific use-case, the digital volume in the DL200 is not any different from an analog one.

This result is consistent with @NTTY 's findings in their D200 review:

"Well, we get some headroom but not much. If I go to "preamp mode" and lower the volume by a lot more (eg -6db) I don't get more headroom, and that is because the volume control is not done in digital domain, prior to conversion, but post conversion by an IC that attenuates the output voltage (digitally calculated attenuation)."

I'm disappointed, given the price of the device.

In general, it's not that big of an issue if you only use sources that can either digitally lower their volume or normalize music. Some people don't care at all about inter-sample clipping so this test won't mean anything to them.

Test 2: Cirrus Hump

Just because I could.

Track 1 (CMaj): Another @danadam test file that can detect clicks in multitone signals that vary in amplitude (DRE artifacts).

Track 2 (hump?): DL200 output with volume set at 98/99.

...as expected.

If you're wondering about the "chunky" sine waves, they're attributable to distortion and jitter introduced by the MacBook + the ADC.

 

[nick_l44.1]

...it seems SMSL went out of their way to put another digital volume control after the SRC even though they clearly didn't need to...

Setting the volume level in the corresponding DAC register does not change the PCM stream in any way. Signal level adjustment is performed after oversampling, so the intersample clipping embedded in the stream will not disappear. -These are hasty conclusions, sorry.
In the DL200, there are only buffers and passive filters between the es9039q2m and XLR/RCA outputs.

 

[tccalvin]

Setting the volume level in the corresponding DAC register does not change the PCM stream in any way. Signal level adjustment is performed after oversampling, so the intersample clipping embedded in the stream will not disappear.
In the DL200, there are only buffers and passive filters between the es9039q2m and XLR/RCA outputs.

I'm not sure I get what you mean. Have I made a mistake in my measurements or in my conclusions? I am open to feedback.

Going by this schematic, Track 3 (75) should look as clean as it does in Track 5 (98noclip). If the volume control were indeed the ESS one placed before the oversampling filter, there shouldn't be any crazy side bands, as lowering the sample peaks by 3.8dBFS is enough to get rid of any inter-sample distortion. Am I wrong?

 

[nick_l44.1]

Have I made a mistake in my measurements or in my conclusions? I am open to feedback.

I repeated your test. It seems we were both wrong. I get clipping, but it disappears with a slight decrease in volume (94/99 = -2.5 dB). Is this roughly the result you expected? Usually, you need to check that the volume control actually changes the voltage at the XLR/RCA output and that the source of the clipping is not the ADC.

 

[tccalvin]

I repeated your test. It seems we were both wrong. I get clipping, but it disappears with a slight decrease in volume (94/99 = -2.5 dB). Is this roughly the result you expected? Usually, you need to check that the volume control actually changes the voltage at the XLR/RCA output and that the source of the clipping is not the ADC.

Huh. Now that's weird. I would like to understand this better.

If it's my ADC that's clipping, Then wouldn't it also have clipped Track 5 (98noclip)? The DL200's volume was on 98 there. The only difference was that the test file was 3.8dB quieter going into the DAC.

What were your test conditions?

EDIT: Thinking about it, the problem might lie with macOS handling two different sample rates (Out: 44.1k, In: 192k) at the same time, but I'm not sure.

 

[nick_l44.1]

If it's my ADC that's clipping

It's hard to say for sure. It's easier to just check the level. Some ADCs have clipping indicators.

What were your test conditions?

The file was played back by foobar via wasapi/dl200. The Cosmos ADC was connected to XLR. The ADC sensitivity was set to 10 V to avoid clipping.

problem might lie with macOS

Of course, any OS resampling will cause clipping.

 

[tccalvin]

Unfortunately I can no longer edit my above post but I can make a new one.

I take it all back. The DL200 does indeed use the ESS pre-SRC volume.

The clipping came from the CoreAudio mixer. Audacity cannot directly communicate with the DL200 without going through the macOS mixer.

I used a program called "Colibri DAC DSD Test" to play the test file without being limited by CoreAudio.

Now we're talking (volume at 94/99):

My apologies for the mistake and thank you @nick_l44.1 for the tips and for replicating the test.