• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

S.M.S.L PO100 AK - Measurements / Round 2 (DAC)

VintageFlanker

Major Contributor
Forum Donor
Joined
Sep 20, 2018
Messages
4,940
Likes
19,692
Location
Paris
S.M.S.L PO100 AK - Measurements / Round 2
Main.jpg


Hi folks,

This is an update over my previous measurements of the S.M.S.L PO100 family. I will focus precisely on the PO100 AK, which not only is an USB to S/PDIF converter (as are the PO100 and PO100 PRO) but also a budget DAC, that claims to perform excellent for its price point (89$ MSRP, but I saw it discounted at 75€ right now at Amazon FR). The unit I bought back in December (and returned since) showed kind of mixed bag results. Pure SINAD was very good for one channel, while the other appeared to be significantly worse and gave me inconsistent numbers, runs after runs. In addition, it was not particularly great either in some other tests, like Jitter or Frequency Response.

The manufacturer, represented here by @SMSL-Mandy finally reacted to my review and suggested that my unit could faulty: "we suspect that the PO100 AK may be defective - parts soldering caused by this test". @SMSL-Mandy offered to send me another sample for a second evaluation, which I accepted:

I received it a few days ago. As far I can see, it seems to be a regular production sample, with the exact same packaging as the previous PO100 AK I owned... So let's assume that it is not a cherry-picked unit, even if it had been safer to buy another random one for sale on the market.

I will only talk about overall performance of the second VS first sample in this post, for any other information, just check the original review:

Measurements

Disclaimer: Measurements you are about to see are not intended to be as precise or extensive than what you get from a 30k€ AP. There is obviously both hardware and software limitations here, so not quite apples to apples comparison with Amir's testing. Still, this data is enough to have a pretty good idea if the gear is bad or not, stellar, broken, or sub-par...

- Instruments : E1DA Cosmos ADC (Grade B). RME ADI-2/4 PRO SE. E1DA Cosmos APU 60dB preamp is used for DR measurements. Output voltage is measured separately, using a DMM with a 0dBFS 1kHz tone.
- Software : REW V5.20.14, Multitone Loopback Analyzer 1.0.75 and RMAA 6.4.5 PRO,
- Method : 8 runs for each test, then I choose the closest to the average. Bandwidth and sampling rate to be specified for each test.
- S.M.S.L PO100 AK is running at full output (2.1V).

Note: Both my measurements process and graph presentation evolved a bit since the original review a couple of months ago. Anyhow, to give a fair apples to apples comparison, I used the exact same instrumentation and settings for the comparison charts below. I will then provide more extensive measurements, now made possible with my new equipment.​

Summary
Tests
1rst Sample
2nd Sample
Frequency response (from 40 Hz to 15 kHz)
+0.04, -1.28dB​
+0.02, -1.27dB​
Noise level (RMAA)
-117.0dB​
-119.1dB​
Dynamic range, dB (RMAA)
116.9dBA​
119.0dBA​
THD (REW)​
0.00024%​
0.00009%​
THD+N (REW)​
110.7dB​
115.8dB​
IMD + Noise (RMAA)​
0.00078%​
0.00042%​
Stereo crosstalk (RMAA)​
-102.1dB​
-119.4dB​
IMD at 10kHz (RMAA)​
0.00048%​
0.00039%​

Note: For the two samples, these numbers are the average of both channels. As such, Right channel (the faulty one) is obviously degrading the overall score of the first sample.

As a reminder, this is the original THD+N (SINAD) graph I posted, that clearly shows the disparity between both channels:
Total Harmonic Distortion + Noise (SINAD) - 24b/44.1kHz
REW FFT THD+N.jpg

And this is what I now get out of the second sample:

Total Harmonic Distortion + Noise (SINAD) - 24b/44.1kHz
REW FFT SINAD COSMOS.jpg


These are good news! Results are right on par with manufacturer's specs (-116dB THD+N), not to mention with excellent consistency regarding Noise and Distortion between Left and Right channels. Also, it is remarkably good for some 89$, USB powered unbalanced DAC. In fact, actually better than the venerable and pricier Topping D10S. Just for peace of mind, I double-checked it with my ADI-2/4 PRO SE:​

Total Harmonic Distortion + Noise (SINAD) - 24b/44.1kHz
REW FFT SINAD RME.jpg


I use the digital trim of RME, to get closer to 0dBFS in REW RTA. Apart from absolute level, both results are consistent enough.

For comparison, @SMSL-Mandy published this SINAD FFT, done with an AP:

12.jpg


This is about 1.5dB better than my results, but I wondered if this could be A-Weighted. It appears to be the case, when looking at parameters.

Now, let's go deeper in the performance analysis...

Frequency Response was kind of an issue with the first sample. Still is. The PO100 AK uses Slow Roll-Off filter from its AK4493 chip, and it doesn't allow choosing another one... Which is very unfortunate when you have such an aggressive highs roll-off, running regular 44.1kHz sample rate. On the other hand, I am please to see almost identical voltage and amplitude for both channel (rare for budget DACs).
Frequency Response - 24b/44.1kHz
REW FR 44.jpg


Beware: Manufacturer's explanations for this could drive some of you mad... :p
The next item to note is that we are using one of the Slow Roll-Off filters in the AK4493S due to MQA certification.

In addition, they shared this FR capture:​

123 (1).jpg

Well, I was effectively able to reproduce the exact same FR, but only with 88.2kHz sampling rates and above (separated FR, but amplitude is the same):​

Frequency Response - 24b/192/96/88.2/48/44.1kHz
FR all SR.jpg


Let's see the rest of benchmarks, with updated process (REW replacing RMAA, which remains for Crosstalk):​

Noise Level - 24b/44.1kHz
REW Noise L 120.2 R 120.3.jpg



Dynamic Range - 24b/44.1kHz
REW DR.jpg



Intermodulation Distortion SMPTE - 24b/44.1kHz
REW IMD SMPTE R 105.8.jpg



Intermodulation Distortion SMPTE Versus Level - 24b/44.1kHz
IMD SMPTE VS Level.png



THD Versus Frequency - 24b/48kHz
THD VS Frequency.png

A new test that was lacking in my previous reviews. The PO100 AK shows near-ideal consistency within audible range.​


Jitter - 24b/48kHz
Jitter 2.png

This one is now a lot better than previous results: While Left was a touch better than Right, it was not "great" either. Now, we got the cohesive results out of both channels.​


Multitone 32 - 24b/192kHz
Multitone 32.png

Multitone gave me the same TD+N of -106.3dB I got out of the first sample's Left channel... Expect that Right is now "fixed".

Stereo Crosstalk - 24b/44.1kHz
cross.png

Again, quite a bit better than it was before.

That is all. I am done...

Conclusions

I will be short. Overall, PO100 AK truly showed an excellent performance for the class. It virtually nailed all tests, apart from Frequency Response. Perhaps, some of you could help me to speculate about present versus previous results, because I still have no clue about what went wrong with the first unit. Symptoms looked presumably like a ground loop issue, but I found quite uncommon to observe it in one channel only (I noticed that @amirm had one case recently, tho). Now that I may compare both my new and previous data in details, it is clear that bad performance of the Right channel indeed bled into Left's one in some tests, such as Crosstalk and Jitter. With this properly working sample, we just have very consistent behaviour for both, and I am surprised to observe such a good implementation of the AK4493 DAC chip... particularly at that price.

I would like to address my sincere thanks to @SMSL-Mandy for reacting to my review, sharing their own data and for sending another product. For now, I assume (say, I hope) that we could exclude the hypothesis of a cherry-picked SMSL PO100 AK...

So... are we good now, S.M.S.L? Not quite. The Frequency Response is still barely acceptable for my standards. It is fine to have this AKM Slow Roll-Off filter, but not when it remains the only one available. Since the PO100 AK does have no physical switch, I am sure that this still could be upgraded by firmware. At least, give your customer the choice of getting what any DAC is supposed to achieve: a flat Frequency Response from 20Hz to 20kHz. Just basics.

Flanker rating: Competent
 
Last edited:
OP
VintageFlanker

VintageFlanker

Major Contributor
Forum Donor
Joined
Sep 20, 2018
Messages
4,940
Likes
19,692
Location
Paris

taner

Active Member
Joined
Jun 13, 2020
Messages
199
Likes
136
This slow roll off issue related to "MQA certification" is a very good example of how manufacturers ditching their products for sake of pleasing the crowd and following nonsense trends. MQA is a scam in purest form.
 
Last edited:
OP
VintageFlanker

VintageFlanker

Major Contributor
Forum Donor
Joined
Sep 20, 2018
Messages
4,940
Likes
19,692
Location
Paris
This slow roll off issue related to "MQA certification" is a very good example of how manufacturers ditching their products for sake of pleasing the crowd and following nonsense trends.
I don't know which crowd would be pleased with such FR (maybe it's supposed to sound more "analog":facepalm:), even more when it is justified for MQA certification. Imagine that PO100 AK could have been 10-15€ cheaper without it, and it would have actually performed better.

I couldn’t hear that 20kHz drop anyway….
Probably most adults won't. But it still is a roll-off that starts from 9kHz.
 

JktHifi

Senior Member
Joined
Mar 9, 2023
Messages
387
Likes
195
Been using this device for a week. When amp volume cranked up to the max, using analog RCA, still can’t here any hissing/humming sound. Compare to other device such as DAP or TV Box, sound humming very bad.

The built in DAC of Digital coaxial input of my amp is Wolfson WM8718. I found it better on louder volume (has to be very loud) than AKM AK4493S DAC of this device. But when the amp volume below 12 o’clock, this device win. Is it because of the amp or the DAC (SMSL PO100 AK)?
 

nerone

Active Member
Joined
Feb 2, 2019
Messages
239
Likes
52
Location
ITALIA
Thanks for the new review, the border color looks different, maybe it's a new version.
Mine works well, in terms of sound quality compared to the Topping E30 it seems to have cut the high frequencies, I can't explain better.
The negative points are:
It doesn't work if you use Tidal from smartphone and iPhone, with smartphone it only works with UAPP while with iPhone I couldn't get it to work.
I hope SMSL knows how to fix these problems otherwise I see wasted money
 
OP
VintageFlanker

VintageFlanker

Major Contributor
Forum Donor
Joined
Sep 20, 2018
Messages
4,940
Likes
19,692
Location
Paris
Question to SMSL: Will the version with the index v2 appear on Aliexpress or Amazon? From which serial number should I buy the device?
I think you misinterpreted my post.

There is no "V1", nor "V2" revision whatsoever. There are just (supposedly) one specific defective/faulty unit, and the other that is not.

Is it because of the amp or the DAC
Definitely your amp. The PO100 AK have some extremely common 2.1V output for single ended. Your amp should ideally have about the same internal sensitivity for both its integrated DAC and analog inputs.

the border color looks different, maybe it's a new version.
It's not. ;)
 
Last edited:

CtheArgie

Addicted to Fun and Learning
Forum Donor
Joined
Jan 11, 2020
Messages
509
Likes
773
Location
Agoura Hills, CA.
I don't know which crowd would be pleased with such FR (maybe it's supposed to sound more "analog":facepalm:), even more when it is justified for MQA certification. Imagine that PO100 AK could have been 10-15€ cheaper without it, and it would have actually performed better.


Probably most adults won't. But it still is a roll-off that starts from 9kHz.
Yes, but it is less than 1 dB until around 14 kHz, isn't it?
 

nerone

Active Member
Joined
Feb 2, 2019
Messages
239
Likes
52
Location
ITALIA
The seller of this P0100AK hasn't answered any questions, I guess there won't be any updates for this gizmo so I'll be returning it
 

DanielT

Major Contributor
Joined
Oct 10, 2020
Messages
4,751
Likes
4,633
Location
Sweden - Слава Україні
Yep.

BTW, I just updated the FR graphs with sampling frequencies from 44.1 to 192kHz. 88.2kHz appears to be also usable:
View attachment 272321
That was a very strange, bad FR for a DAC.

You say:

I will be short. Overall, the PO100 AK truly showed an excellent performance for the class. It virtually nailed all tests, apart from Frequency Response

The problem is that deviations from Frequency Response are the most negative thing for sensible sound reproduction. It is like a basic requirement that the FR is straight. Low distortion, SINAD in all its glory but without a sensible FR it becomes rubbish.

The Frequency Response is still barely acceptable for my standards...The same to everyone else.We are talking about a DAC, not speakers FR deviations.
 
OP
VintageFlanker

VintageFlanker

Major Contributor
Forum Donor
Joined
Sep 20, 2018
Messages
4,940
Likes
19,692
Location
Paris
Low distortion, SINAD in all its glory but without a sensible FR it becomes rubbish.
Agreed, but I cannot give it a failing grade considering how good it is overall for such a cheap and little DAC. That's about it.

I still have some pretty naive expectations that @SMSL-Mandy could add a different filter by firmware...
 

Toku

Major Contributor
Joined
Apr 4, 2018
Messages
2,391
Likes
2,776
Location
Japan
Why they insist pushing that mqa-evil is beyond me.

Found it for $93 shipped. Not bad.
It looks like we won't need to talk about MQA anymore.

MQA filed for bankruptcy on April 3.
It seems that there will be no supply of new MQA sound sources to the market in the future.
If you search the web for "MQA bankruptcy" for more information on this matter, you will find a lot of information.
Therefore, PO100 AK should also shift to a model without MQA.
 

flowjm

Member
Joined
Aug 24, 2019
Messages
95
Likes
197
It looks like we won't need to talk about MQA anymore.

MQA filed for bankruptcy on April 3.
It seems that there will be no supply of new MQA sound sources to the market in the future.
If you search the web for "MQA bankruptcy" for more information on this matter, you will find a lot of information.
Therefore, PO100 AK should also shift to a model without MQA.
Let's hope so. I'd buy one of the models in a heartbeat without MQA (and licensing fees) and fixed FR.
 

Morpheus

Active Member
Forum Donor
Joined
Jun 5, 2019
Messages
132
Likes
144
Location
E.C
S.M.S.L PO100 AK - Measurements / Round 2
View attachment 271967

Hi folks,

This is an update over my previous measurements of the S.M.S.L PO100 family. I will focus precisely on the PO100 AK, which not only is an USB to S/PDIF converter (as are the PO100 and PO100 PRO) but also a budget DAC, that claims to perform excellent for its price point (89$ MSRP, but I saw it discounted at 75€ right now at Amazon FR). The unit I bought back in December (and returned since) showed kind of mixed bag results. Pure SINAD was very good for one channel, while the other appeared to be significantly worse and gave me inconsistent numbers, runs after runs. In addition, it was not particularly great either in some other tests, like Jitter or Frequency Response.

The manufacturer, represented here by @SMSL-Mandy finally reacted to my review and suggested that my unit could faulty: "we suspect that the PO100 AK may be defective - parts soldering caused by this test". @SMSL-Mandy offered to send me another sample for a second evaluation, which I accepted:

I received it a few days ago. As far I can see, it seems to be a regular production sample, with the exact same packaging as the previous PO100 AK I owned... So let's assume that it is not a cherry-picked unit, even if it had been safer to buy another random one for sale on the market.

I will only talk about overall performance of the second VS first sample in this post, for any other information, just check the original review:

Measurements

Disclaimer: Measurements you are about to see are not intended to be as precise or extensive than what you get from a 30k€ AP. There is obviously both hardware and software limitations here, so not quite apples to apples comparison with Amir's testing. Still, this data is enough to have a pretty good idea if the gear is bad or not, stellar, broken, or sub-par...

- Instruments : E1DA Cosmos ADC (Grade B). RME ADI-2/4 PRO SE. E1DA Cosmos APU 60dB preamp is used for DR measurements. Output voltage is measured separately, using a DMM with a 0dBFS 1kHz tone.
- Software : REW V5.20.14, Multitone Loopback Analyzer 1.0.75 and RMAA 6.4.5 PRO,
- Method : 8 runs for each test, then I choose the closest to the average. Bandwidth and sampling rate to be specified for each test.
- S.M.S.L PO100 AK is running at full output (2.1V).

Note: Both my measurements process and graph presentation evolved a bit since the original review a couple of months ago. Anyhow, to give a fair apples to apples comparison, I used the exact same instrumentation and settings for the comparison charts below. I will then provide more extensive measurements, now made possible with my new equipment.​

Summary
Tests
1rst Sample
2nd Sample
Frequency response (from 40 Hz to 15 kHz)
+0.04, -1.28dB​
+0.02, -1.27dB​
Noise level (RMAA)
-117.0dB​
-119.1dB​
Dynamic range, dB (RMAA)
116.9dBA​
119.0dBA​
THD (REW)​
0.00024%​
0.00009%​
THD+N (REW)​
110.7dB​
115.8dB​
IMD + Noise (RMAA)​
0.00078%​
0.00042%​
Stereo crosstalk (RMAA)​
-102.1dB​
-119.4dB​
IMD at 10kHz (RMAA)​
0.00048%​
0.00039%​

Note: For the two samples, these numbers are the average of both channels. As such, Right channel (the faulty one) is obviously degrading the overall score of the first sample.

As a reminder, this is the original THD+N (SINAD) graph I posted, that clearly shows the disparity between both channels:
Total Harmonic Distortion + Noise (SINAD) - 24b/44.1kHz

And this is what I now get out of the second sample:

Total Harmonic Distortion + Noise (SINAD) - 24b/44.1kHz
View attachment 271982

These are good news! Results are right on par with manufacturer's specs (-116dB THD+N), not to mention with excellent consistency regarding Noise and Distortion between Left and Right channels. Also, it is remarkably good for some 89$, USB powered unbalanced DAC. In fact, actually better than the venerable and pricier Topping D10S. Just for peace of mind, I double-checked it with my ADI-2/4 PRO SE:​

Total Harmonic Distortion + Noise (SINAD) - 24b/44.1kHz
View attachment 271988

I use the digital trim of RME, to get closer to 0dBFS in REW RTA. Apart from absolute level, both results are consistent enough.

For comparison, @SMSL-Mandy published this SINAD FFT, done with an AP:

View attachment 271989

This is about 1.5dB better than my results, but I wondered if this could be A-Weighted. It appears to be the case, when looking at parameters.

Now, let's go deeper in the performance analysis...

Frequency Response was kind of an issue with the first sample. Still is. The PO100 AK uses Slow Roll-Off filter from its AK4493 chip, and it doesn't allow choosing another one... Which is very unfortunate when you have such an aggressive highs roll-off, running regular 44.1kHz sample rate. On the other hand, I am please to see almost identical voltage and amplitude for both channel (rare for budget DACs).
Frequency Response - 24b/44.1kHz
View attachment 271992

Beware: Manufacturer's explanations for this could drive some of you mad... :p


In addition, they shared this FR capture:​
Well, I was effectively able to reproduce the exact same FR, but only with 88.2kHz sampling rates and above (separated FR, but amplitude is the same):​

Frequency Response - 24b/192/96/88.2/48/44.1kHz
View attachment 272317

Let's see the rest of benchmarks, with updated process (REW replacing RMAA, which remains for Crosstalk):​

Noise Level - 24b/44.1kHz
View attachment 272003


Dynamic Range - 24b/44.1kHz
View attachment 272004


Intermodulation Distortion SMPTE - 24b/44.1kHz
View attachment 272005


Intermodulation Distortion SMPTE Versus Level - 24b/44.1kHz
View attachment 272006


THD Versus Frequency - 24b/48kHz
View attachment 272020
A new test that was lacking in my previous reviews. The PO100 AK shows near-ideal consistency within audible range.​


Jitter - 24b/48kHz
View attachment 272007
This one is now a lot better than previous results: While Left was a touch better than Right, it was not "great" either. Now, we got the cohesive results out of both channels.​


Multitone 32 - 24b/192kHz
View attachment 272008

Multitone gave me the same TD+N of -106.3dB I got out of the first sample's Left channel... Expect that Right is now "fixed".

Stereo Crosstalk - 24b/44.1kHz
View attachment 272026
Again, quite a bit better than it was before.

That is all. I am done...

Conclusions

I will be short. Overall, PO100 AK truly showed an excellent performance for the class. It virtually nailed all tests, apart from Frequency Response. Perhaps, some of you could help me to speculate about present versus previous results, because I still have no clue about what went wrong with the first unit. Symptoms looked presumably like a ground loop issue, but I found quite uncommon to observe it in one channel only (I noticed that @amirm had one case recently, tho). Now that I may compare both my new and previous data in details, it is clear that bad performance of the Right channel indeed bled into Left's one in some tests, such as Crosstalk and Jitter. With this properly working sample, we just have very consistent behaviour for both, and I am surprised to observe such a good implementation of the AK4493 DAC chip... particularly at that price.

I would like to address my sincere thanks to @SMSL-Mandy for reacting to my review, sharing their own data and for sending another product. For now, I assume (say, I hope) that we could exclude the hypothesis of a cherry-picked SMSL PO100 AK...

So... are we good now, S.M.S.L? Not quite. The Frequency Response is still barely acceptable for my standards. It is fine to have this AKM Slow Roll-Off filter, but not when it remains the only one available. Since the PO100 AK does have no physical switch, I am sure that this still could be upgraded by firmware. At least, give your customer the choice of getting what any DAC is supposed to achieve: a flat Frequency Response from 20Hz to 20kHz. Just basics.

Flanker rating: Competent
The thing that really bothers me in this review/product no one has mentioned..
I can live with the filter: although I absolutely know I should be using a sharp rolloff on my RME ADI 2 DAC for impecable performance, I still end coming back to NOS, even with all the problems that has somehow I prefer it, i'll be damned..And the sadest part is I don't even hear past 13,5 khz, the irony..!!
Anyway, for me the elephant in the room for me is QC. There is a disturbing number of items on this site that were measured and proved to be defective,completely off spec, and cheap chinese gear although far from the sole offenders, have been in the forefront. As a consumer, I' d rather spend 200e and have assurance of reliability and that the unit I got is within spec, than tossing out 100 to the trash, risking a dud.
Yes, only one sample gets tested here, but somehow defective items keep coming up, and it is especially concerning as most of us only have their hears to detect flaws ( usually just the blattant ones), not test equipment...I think that wether you pay 50, or 5000, getting exactly what it says in the tin, good or bad, is a minimum.
 
Top Bottom