• 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!

miniDSP U-DAC8 8-Channel USB DAC Review

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,394
Location
Seattle Area
This is a review and detailed measurements of the miniDSP U-DAC8 multichannel USB audio DAC. It was kindly purchased and drop shipped to me by a member. The U-DAC8 came out a few years ago but still seems like a current device costing US $255 from the company direct.

The U-DAC8 embodies simplicity and fitment for function:

minidsp U-DAC8  8 channel USB DAC Audio Review.jpg

A USB port in for data and then in the rear, 8 channels of output:
minidsp U-DAC8  8 channel USB DAC Back panel connectors input output Audio Review.jpg

And an external adapter double the size of a typical phone charger.

Wiring will be messy though with that USB connection in the front. Would have been nicer to have a slightly larger box and put the USB in the back.

The U-DAC8 was plug and play with Windows 10, exposing 8 concurrently channels. You could build a PC home theater or multi-channel source with it easily.

Most PCs come with 5.1 output so getting up to 8 channels out of U-DAC8 is a plus. As would be expected performance improvement.

DAC Audio Measurements
As usual, we start with a 1 kHz tone at full amplitude and see what we get out of the DAC:
minidsp U-DAC8  8 channel USB DAC Audio Measurements.png


OK, this is not good. Distortion products are way up there as is enough noise to give us a very low SINAD of 78. This puts the U-DAC together with the rest of the students that have gotten a failing grade from us:
Best multichannel DACs Review.png


The U-DAC8 uses AKM's AK4440 DAC chip. It does everything internally which the U-DAC provides from creating the required supplies to output buffers. Its specifications state a "THD" of 93 dB. We are 15 dB short of that at 78. miniDSP's own spec is 0.006% which translates into SINAD of 84. So whether we take my measurements or theirs, we are way short of the performance of the chip.

I tested the other channels and they produced identical output.

Dynamic range was decent:
minidsp U-DAC8  8 channel USB DAC Dynamic Range Audio Measurements.png


Intermodulation distortion versus level was well below my reference DAC:
minidsp U-DAC8  8 channel USB DAC IMD Audio Measurements.png


Linearity showing was not as embarrassing as I expected it to be:

minidsp U-DAC8  8 channel USB DAC Linearity Audio Measurements.png


Multitone test shows that as frequencies go up, non-linearities increase substantially resulting in more distortion products:

minidsp U-DAC8  8 channel USB DAC Multitone Audio Measurements.png


We will revisit this in a bit later. For now, let's look at jitter:

Untitled-1.png


Noise level is very high as it is but then it rises around our main tone at 12 kHz. I re-ran this test (not shown) at lower frequency and it did that there too. We call this noise modulation although usually the entire noise floor lifts, not just around our main tone. Audibility of this is not as bad but still, something we don't want to see in a high quality implementation.

Real awful response was awaiting us in the form of THD+N versus frequency:

minidsp U-DAC8  8 channel USB DAC THD+N vs Frequency Distortion Audio Measurements.png


What on earth is this? Not only do we have fairly high levels of distortion/noise, it is also frequency dependent. To make sure this was not a measurement anomaly, I tested that trough around 700 Hz versus 1.5 kHz next to it:

minidsp U-DAC8  8 channel USB DAC FFT Audio Measurements.png


Distortion is definitely higher at 1.5 kHz, demonstrating that the previous graph was correct.

Even DAC reconstruction filter response was disappointing with low attenuation out of band:

minidsp U-DAC8  8 channel USB DAC Filter Response Audio Measurements.png


The right side should be down at -90 dB and lower.

Conclusions
What a shame. I walked into this review excited that we would have a clean and simple 8-channel DAC that could let us build our own Audio/Video Receiver in a PC. Performance though is quite disappointing. As a minimum we should have seen performance similar to the AKM DAC chip specs. Where the shortfall is, I can't say given the simple design here. As much as I hate the crapware that comes with PC/gaming sound cards, they have much better performance for much less money than the U-DAC8. I hope miniDSP goes back to the drawing board and builds a proper version of this DAC.

Needless to say, I cannot recommend the miniDSP U-DAC8.

--------
As always, questions, comments, corrections, etc. are welcome.

I threw a bone at one of the panthers and it hit him in the head. No, it was not the headless one. Seems like he needs glasses as he used to be able to grab it mid-air. I am still searching to find an optometrist that specializes in pink panthers. When I do, I need lots of money to pay for the visit and eventual contact lenses (can't imagine the panther walking around with glasses -- the rest of the panthers would tease the heck out of him). So I appreciate generous donations to cover the cost using : https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/
 

JohnYang1997

Master Contributor
Technical Expert
Audio Company
Joined
Dec 28, 2018
Messages
7,175
Likes
18,292
Location
China
??? The tight match between the two channels actually surprised me a bit. Normally when we get this level of bad performance there is difference between two channels. How come??
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,394
Location
Seattle Area
Because everything is on-chip and hence higher precision:

1574927389811.png


1574927446900.png


A few caps and you have the whole solution!

How they managed to degrade the performance so much, is a puzzle.
 

Trdat

Addicted to Fun and Learning
Forum Donor
Joined
Sep 6, 2019
Messages
967
Likes
396
Location
Yerevan "Sydney Born"
So I am presuming you will get better performance with the UDIO8 and purchasing a well measured DAC from your list for each channel?

Could the UDIO8 also possibly have issues with it?(I ask cause not sure how that works in conjunction with a well performing DAC). Or might just be better to get a descent multichannel DAC in the first place...
 

restorer-john

Grand Contributor
Joined
Mar 1, 2018
Messages
12,580
Likes
38,281
Location
Gold Coast, Queensland, Australia
Does the Khadas tone board and this minidsp use the same AKM DACs? Is that the reason for your comparison, beacuse there are some striking similarities in THD+n v Freq, albeit it a much higher level?

1574930024377.png


In terms of measurement points, the above graph also suggests there are only three measurement points above the 10KHz point, is that correct? Is that sufficient in your opinion for 50% of the audible spectrum?
 
Last edited:

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,201
Likes
16,983
Location
Riverview FL
Good catch!
 

restorer-john

Grand Contributor
Joined
Mar 1, 2018
Messages
12,580
Likes
38,281
Location
Gold Coast, Queensland, Australia
Has the Khadas tone board simply paralleled 4 D/A converters for each stereo channel if it uses the same chip?
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,201
Likes
16,983
Location
Riverview FL
Last edited:

headwhacker

Active Member
Forum Donor
Joined
Dec 25, 2018
Messages
165
Likes
224
Does the Khadas tone board and this minidsp use the same AKM DACs? Is that the reason for your comparison, beacuse there are some striking similarities in THD+n v Freq, albeit it a much higher level?

View attachment 40595

In terms of measurement points, the above graph also suggests there are only three measurement points above the 10KHz point, is that correct? Is that sufficient in your opinion for 50% of the audible spectrum?
KTB use an ESS DAC chip. I belive the mobile version of 9038
 

digicidal

Major Contributor
Joined
Jul 6, 2019
Messages
1,981
Likes
4,838
Location
Sin City, NV
Suddenly a number of AVRs and AVPs did a little victory dance (or at least a sigh of relief). I understand that the principle application is different, but still if you count the cost of the PC and amplification in the package... this would be bested by a number of them for the same cost - and that's without taking the significant amount of additional features into account. :confused:

I wonder (hope) that the nanoAVR HDA is better. At least it's a more complete package with DSP & HDMI switching - and all the cables connect on the same side. For $130 more I'd think it's the better deal though slightly different application.
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,524
Likes
37,057
This looks similar to results from the ViewHD HDMI switcher when using the SPDIF output to feed a good DAC. Which makes me think this is all in the clocking. The switcher optical output fed to a Tact RCS 2.0 DAC in red and the same Tact DAC being fed via optical from the Musical Fidelity V-link is in green. 48 khz sample rate in both cases. This output was very similar, but worse in some ways if using that switcher's DAC analog output. To me this points to the clock being the problem. So what is doing the clocking in this uDac8?

1574940419826.png

1574940478714.png
 
Last edited:

Vovgan

Active Member
Forum Donor
Joined
Jan 9, 2019
Messages
188
Likes
346
Location
Moscow, Russia
Ah, such dismal performance from the manufacturer of the (probably) most popular measuring mic (UMIK-1) for audio enthusiasts is a shame indeed.

As an aside: Is my understanding correct that even if their UMIK-1 has a hundredfold worse performance (THD+N for ADC), there would still be absolutely no concern for using it for the purpose of room measurement?
 
Last edited:

AudioSceptic

Major Contributor
Joined
Jul 31, 2019
Messages
2,691
Likes
2,534
Location
Northampton, UK
Does the Khadas tone board and this minidsp use the same AKM DACs? Is that the reason for your comparison, beacuse there are some striking similarities in THD+n v Freq, albeit it a much higher level?

View attachment 40595

In terms of measurement points, the above graph also suggests there are only three measurement points above the 10KHz point, is that correct? Is that sufficient in your opinion for 50% of the audible spectrum?
50% linear, but only the top octave. Isn't this why we use log scales for frequency (and dB are log anyway for the Y-axis).
 

pozz

Слава Україні
Forum Donor
Editor
Joined
May 21, 2019
Messages
4,036
Likes
6,827
As an aside: Is my understanding correct that even if their UMIK-1 has a hundredfold worse performance (THD+N for ADC), there would still be absolutely no concern for using it for the purpose of room measurement?
The UMIK-1's self noise is very low, although the overall quality is not exceptional. See this discussion: https://www.audiosciencereview.com/forum/index.php?threads/noobie-question.9799/#post-263381

miniDSP in general hasn't produced great measurements:
 

ElNino

Addicted to Fun and Learning
Joined
Sep 26, 2019
Messages
557
Likes
722
Those jitter measurements probably aren't good news for the MiniDSP MCHStreamer digital transport, which seems to the basis of this design's digital interface.

I agree with the speculation that this could be related to clocking. The thing I've never understood about the MCHStreamer is how it's designed to accept digital inputs (S/PDIF) but can't slave its clock to its inputs, unlike the MiniDSP USBStreamer, while also not having a dedicated ASRC onboard.
 

daftcombo

Major Contributor
Forum Donor
Joined
Feb 5, 2019
Messages
3,687
Likes
4,068
Willing to see more miniDSP measurements now.
Thanks Amir!
 

trl

Major Contributor
King of Mods
Joined
Feb 28, 2018
Messages
1,967
Likes
2,523
Location
Iasi, RO
[...]

How they managed to degrade the performance so much, is a puzzle.


Maybe there's something inside we can spot. I wonder if those non-shielded ribbon cables would be at least partially responsible by the bad measurements.

Also manufacturer's webpage: https://www.minidsp.com/products/minidspkits/minidac8
 

trl

Major Contributor
King of Mods
Joined
Feb 28, 2018
Messages
1,967
Likes
2,523
Location
Iasi, RO
Top Bottom