SMSL DL200 DAC & Headphone Amp Review

[jkim]

ASR has a dedicated thread for this combination DAC & headphone amp from SMSL. Interestingly the OP of the thread titled it "SMSL DL200 the best ever value DAC/AMP?" On paper---in regard to pure, technical performance---I couldn't agree more. And couldn't resist buying and testing it. As of April 2025, it retails at $190.

Although it was measured by some members (e.g., by @nick_l44.1) and presented in multiple posts of the thread, I thought a collection of measurements in familiar forms presented in one place would be informative. In fact, Amir reviewed similar devices from SMSL based on the same DAC chipset, ESS ES9039Q2M: SMSL D-6s (DAC only / no headphone amp) and SMSL RAW-MDA 1 (dual DAC chips plus headphone amp). So, we may easily conjecture the measurements of the DL200 should be very similar to these two devices'. Let's find out.

Measurement Setup​

• AD converter: E1DA Cosmos ADCiso Grade A in Mono Mode, and Cosmos Scaler as a buffer.
• Analog preamp: E1DA Cosmos APU for 1 kHz SINAD and low-level tests.
• Software: Room EQ Wizard (V5.4 beta 73) for most of the tests, and Paul Kane's Multitone Loopback Analyzer (v1.2.9).
• Only the USB input was tested.

Results​

Starting with the 1 kHz sinusoid SINAD tests of the DAC pre outs, both balanced and unbalanced:

Balanced TRS pre outs produce slightly better SINAD than unbalanced RCA outs, but the difference is negligible. This is top-notch performance expected from proper implementation of the ES9039Q2M.

SMPTE IMD vs. output level:

Note. Actual IMD levels are 13 dB higher than shown above. REW (v5.4 beta 73) somehow plots these IMD test results incorrectly. Relative differences between devices are still valid to compare.

I own some other DACs based on the same DAC chip to compare: E1DA 9039S and Topping D50 III. Their IMD levels are essentially the same as each other except for slightly lower IMD for the E1DA 9039S above 1 Vrms.

THD+N versus test tone frequency:

The Topping has a slight edge thanks to its dual chip design. Regardless, this level of performance represents the state of the art in the current DAC technology.

Jitter test:

Digital LP filter responses:

Note that SMSL corrected the filter labels in the latest firmware release. FL 3 (Linear Phase Fast Roll-off) looks great.

Measured the frequency response of both its line-level and headphone outputs using multiple bandwidth settings:

In fact, the headphone output's frequency response is essentially identical to the unbalanced RCA line-level output's response:


Its headphone (and unbalanced line-level) output rolls off a little faster than its balanced line-level output: -0.2 dB versus -0.06 dB @ 20 kHz. Not a problem for any of the outputs.

Let's see how transparent its headphone output is. Although it provides 4.4mm and 6.35mm jacks, both outputs are from the same unbalanced amp circuit.

A THD+N vs. output voltage chart is convenient when comparing all conditions:

To have an idea on how clean the DL200's headphone output is in its Low Gain mode, I compared it to the E1DA 9039S, which is presumably one of the most transparent DAC/headphone amp combination devices:

At a given output level, the DL200 headphone out's THD+N in low gain mode is only 1 dB worse than the 9039S. Yes, 1 dB is a negligible difference.

Noise levels measured above should be higher than the actual levels due to the ADC's noise floor limit. To remedy this (only partially, though) DL200's headphone output performance at a low, 50 milli Vrms was measured with the help of the Cosmos APU preamp:

Conclusion​

Yes, I believe these measurements confirm the question asked by the OP of the DL200 thread at ASR, "SMSL DL200 the best ever value DAC/AMP?" Its measured performance is comparable to that of SMSL RAW-MDA 1. At $190 one can have a state-of-the-art DAC plus a transparent headphone amp that can drive almost all the headphones out there (except a few very demanding ones). Of course, other factors like product support and some advanced functionalities may be lacking. Still, $190 for top performance in regard to pure measurable aspects does not seem to be a lot.

By the way, one benefit of the DL200 is its form factor. It can serve as a sturdy, stable stand for another device or anything you want to put on there

 

[JIW]

View attachment 444143View attachment 444142
ASR has a dedicated thread for this combination DAC & headphone amp from SMSL. Interestingly the OP of the thread titled it "SMSL DL200 the best ever value DAC/AMP?" On paper---in regard to pure, technical performance---I couldn't agree more. And couldn't resist buying and testing it. As of April 2025, it retails at $190.

Although it was measured by some members (e.g., by @nick_l44.1) and presented in multiple posts of the thread, I thought a collection of measurements in familiar forms presented in one place would be informative. In fact, Amir reviewed similar devices from SMSL based on the same DAC chipset, ESS ES9039Q2M: SMSL D-6s (DAC only / no headphone amp) and SMSL RAW-MDA 1 (dual DAC chips plus headphone amp). So, we may easily conjecture the measurements of the DL200 should be very similar to these two devices'. Let's find out.

Measurement Setup​

• AD converter: E1DA Cosmos ADCiso Grade A in Mono Mode, and Cosmos Scaler as a buffer.
• Analog preamp: E1DA Cosmos APU for 1 kHz SINAD and low-level tests.
• Software: Room EQ Wizard (v5.4 beta) for most of the tests, and Paul Kane's Multitone Loopback Analyzer (v1.2.9).
• Only the USB input was tested.

Results​

Starting with the 1 kHz sinusoid SINAD tests of the DAC pre outs, both balanced and unbalanced:
View attachment 444163
View attachment 444335

Balanced TRS pre outs produce slightly better SINAD than unbalanced RCA outs, but the difference is negligible. This is top-notch performance expected from proper implementation of the ES9039Q2M.

SMPTE IMD vs. output level:
View attachment 444166

I own some other DACs based on the same DAC chip to compare: E1DA 9039S and Topping D50 III. Their IMD levels are essentially the same as each other except for slightly lower IMD for the E1DA 9039S above 1 Vrms.

THD+N versus test tone frequency:
View attachment 444168

The Topping has a slight edge thanks to its dual chip design. Regardless, this level of performance represents the state of the art in the current DAC technology.

Jitter test:
View attachment 444167

Digital LP filter responses:
View attachment 444172

Note that SMSL corrected the filter labels in the latest firmware release. FL 3 (Linear Phase Fast Roll-off) looks great.

Wideband frequency response:
View attachment 444174

High-frequency response rolls off a little. But -0.1 dB @ 20 kHz and -1.9 dB @ 80 kHz can't be a problem.

Let's see how transparent its headphone output is. Although it provides 4.4mm and 6.35mm jacks, both outputs are from the same unbalanced amp circuit.

View attachment 444170
View attachment 444169

A THD+N vs. output voltage chart is convenient when comparing all conditions:
View attachment 444171

To have an idea on how clean the DL200's headphone output is in its Low Gain mode, I compared it to the E1DA 9039S, which is presumably one of the most transparent DAC/headphone amp combination devices:
View attachment 444340
At a given output level, the DL200 headphone out's THD+N in low gain mode is only 1 dB worse than the 9039S. Yes, 1 dB is a negligible difference.

Noise levels measured above should be higher than the actual levels due to the ADC's noise floor limit. To remedy this (only partially, though) DL200's headphone output performance at a low, 50 milli Vrms was measured with the help of the Cosmos APU preamp:
View attachment 444164

Conclusion​

Yes, I believe these measurements confirm the question asked by the OP of the DL200 thread at ASR, "SMSL DL200 the best ever value DAC/AMP?" Its measured performance is comparable to that of SMSL RAW-MDA 1. At $190 one can have a state-of-the-art DAC plus a transparent headphone amp that can drive almost all the headphones out there (except a few very demanding ones). Of course, other factors like excellent product support and some advanced functionalities may be lacking. Still, $190 for top performance in regard to pure measurable aspects does not seem to be a lot.

By the way, one benefit of the DL200 is its form factor. It can serve as a sturdy, stable stand for another device or anything you want to put on there

What output did you use to measure frequency response? According to SMSL's own measurements, there is a greater roll off on the RCA output.

Further, according to measurements of the RAW-MDA1 by SyncerTech on iXBT.com, the headphone output has a more pronounced roll off compared to the line out.

 

[jkim]

The roll-off also depends on the bandwidth of ADC which needs to substantially higher than that of DUT for the result to be accurate to the higher end of the BW. I measured the FRs of both the TRS line out and headphone out in multiple sampling freqs. Will update the post later today.

EDIT. BTW I doubt the RCA out has different FR than the TRS out. They must have been derived from the same buffer. There's no reason to do it differently. I suspect the difference was from the cable capacitance.

 

[JIW]

The roll-off also depends on the bandwidth of ADC which needs to substantially higher than that of DUT for the result to be accurate to the higher end of the BW. I measured the FRs of both the TRS line out and headphone out in multiple sampling freqs. Will update the post later today.

EDIT. BTW I doubt the RCA out has different FR than the TRS out. They must have been derived from the same buffer. There's no reason to do it differently. I suspect the difference was from the cable capacitance.

iXBT also uses E1DA Cosmos ADC so should be the same.

Here is a picture of the board from the other thread.

SMSL DL200 the best ever value DAC/AMP?

Second, back to the question that started off this blather: From what I read on that forum, it appears that some guy (IVX, apparently well versed in the topic) tested the chip and found something wrong with the upper frequencies that appears to deliberately artificially increase SINAD and has...

www.audiosciencereview.com

 

[jkim]

iXBT also uses E1DA Cosmos ADC so should be the same.

Just updated the post with new frequency response measurements. I am confident the results are accurate, which are more or less consistent with iXBT's measurements.

I will measure the RCA output when I get a chance.

 

[Sokel]

That one is the ONLY of the series of cheap DACs we have seen paraded these years that temped me to get.
(must be the size also, can stand these tinny stuff with hard to read, small screens that become unreadable due to burn after some time )

 

[JIW]

Just updated the post with new frequency response measurements. I am confident the results are accurate, which are more or less consistent with iXBT's measurements.

I will measure the RCA output when I get a chance.

Thanks. It should not amount to anything except for the most golden eared maybe that is the way they tamed the sharp sabre...

 

[jkim]

Thanks. It should not amount to anything except for the most golden eared maybe that is the way they tamed the sharp sabre...

Added unbalanced line-level FR measurements to the review.

You're right. The RCA output's FR is indeed different (rolling off faster) than the TRS output's. It turns out that the RCA line-level FR is essentially the same as the headphone output's FR. To see why we need to see its circuit diagram---the PCB layout is difficult to read. This behavior must be due to the two line-level outputs having either separate buffer stages or some specific output coupling structure. One interesting thing is that according to SMSL's specs both outputs seem to have the same output impedance (100 ohms). Usually the impedance of a balanced output is twice that of an unbalanced output.

 

[JIW]

Added unbalanced line-level FR measurements to the review.

You're right. The RCA output's FR is indeed different (rolling off faster) than the TRS output's. It turns out that the RCA line-level FR is essentially the same as the headphone output's FR. To see why we need to see its circuit diagram---the PCB layout is difficult to read. This behavior must be due to the two line-level outputs having either separate buffer stages or some specific output coupling structure. One interesting thing is that according to SMSL's specs both outputs seem to have the same output impedance (100 ohms). Usually the impedance of a balanced output is twice that of an unbalanced output.

It might be in the differential summing stage to get a single ended signal. The output impedance might just be given for one of the outputs. You can check whether they are the same by seeing if the relative voltage drop over a low impedance load is the same for an unclipped signal. The 300 Ohm load may be fine, but the 32 Ohm load will give greater deviations but also require much more current and thus probably lower level.

 

[josiah]

Thank you! I found this to be a huge upgrade over my Topping DX3 Pro+ for multiple reasons! I was lucky enough to do some A/B testing as I have both units. My fiancé uses my DX3 Pro+ now. It is nice to see some numbers to back up my listening experience.

 

[mc.god]

Uhm... I was pretty sure this was tested by Amir, bot nope, it wasn't... maybe the other thread is so long that my brain automatically translated it into a review...
Another sneaky trick by my (de)generative (un)artificial intelligence
Thank you @jkim, another great review, you are doing wonders!

 

[Power Pop 23]

Uhm... I was pretty sure this was tested by Amir, bot nope, it wasn't... maybe the other thread is so long that my brain automatically translated it into a review...
Another sneaky trick by my (de)generative (un)artificial intelligence
Thank you @jkim, another great review, you are doing wonders!

Don't feel too bad....

Amir has measured a SMSL 'DO200 (Pro)' and SMSL 'DL100'.

 

[nick_l44.1]

SMPTE IMD vs. output level

SMPTE levels far below -120 dB looks absolutely unreal. Check the manual fundamental override or something else maybe. Can you show the SMPTE FFT?
You were able to measure distortion at high output levels. With my dl200 there was clipping in the APU. And THD start to grow a bit later. I've noticed the same behavior from raw-mda1 in Amir's measurements.
BTW, the 9039s hpa is a way better for heavy loads. Here is the THD+N vs Voltage for 300/32/16 Ohm:

The 16 Ohm line is shorter for a clipping above -.5 dBFS.

 

[obsessive9119]

Is this essentially the same as the C200?

 

[jkim]

SMPTE levels far below -120 dB looks absolutely unreal. Check the manual fundamental override or something else maybe. Can you show the SMPTE FFT?
You were able to measure distortion at high output levels. With my dl200 there was clipping in the APU. And THD start to grow a bit later. I've noticed the same behavior from raw-mda1 in Amir's measurements.
BTW, the 9039s hpa is a way better for heavy loads. Here is the THD+N vs Voltage for 300/32/16 Ohm:
View attachment 445534
The 16 Ohm line is shorter for a clipping above -.5 dBFS.

Thanks for the info.

I lent the DL200 to my son, so won't be able to measure it again for a while. But I will check the IMD measurement setting in REW.

The reason why my distortion vs. output level results (for either the DL200 or 9039S) are different from yours is that I do not use the APU for sweeping tests. This is because the main point of varied output level tests is to see the trend and clipping behavior, not accurate THD numbers.

 

[jkim]

Is this essentially the same as the C200?

Some differences. The main thing is the DAC chip adopted. The C200 is based on the previous generation chipset, ES9038Q2M.

EDIT. I notice that there is the C200 PRO. Weirdly the SMSL website does not show this product. It seems that the PRO version adopts ES9039Q2M.

 

[olieb]

Is this essentially the same as the C200?

I would hope not. I experienced a few annoying glitches with the C200.
The first one died on me after 364 days. I was sent a replacement with no problems, that was great.
With the new one I found that with REW sweeps in one channel after seconds some kind of protection kicks in and I have to either power cycle or (as this does work only sometimes) to unplug USB.
Sometimes the USB looses the connection spontaneously and again only unplugging and reconnecting helps.
Seems to me as if the firmware could use some work.

 

[obsessive9119]

I would hope not. I experienced a few annoying glitches with the C200.

Other than my first C200 being DOA, mine has been really solid. I had some weirdness when using USB (depending on the device) but with optical it has been flawless.

 

[josiah]

I would hope not. I experienced a few annoying glitches with the C200.
The first one died on me after 364 days. I was sent a replacement with no problems, that was great.
With the new one I found that with REW sweeps in one channel after seconds some kind of protection kicks in and I have to either power cycle or (as this does work only sometimes) to unplug USB.
Sometimes the USB looses the connection spontaneously and again only unplugging and reconnecting helps.
Seems to me as if the firmware could use some work.

My unit has been solid! I have not experience any issues. I do not use Bluetooth.