• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

Soekris DAM1021 R2R DAC Measurements

March Audio

Major Contributor
Manufacturer
Patreon Donor
Joined
Mar 1, 2016
Messages
3,467
Likes
2,921
Location
Perth Western Australia
#1
I have finally managed to borrow a Soekris R2R sign magnitude dac to measure. Essentially I bought this dac for a speaker project. I was/am agnostic about the DAC technology, however there is no denying R2R measures more poorly than typical delta sigma. I was particularly interested in its FPGA filtering capabilities for crossover duties. Ultimately it wasnt the solution I wanted so the dac now resides with a friend who has put it in a box with very pleasant subjective results.

Note that I had a few issues with 50Hz and harmonics pick up. I will investigate later, so crap below 1kHz please ignore. I have measured this DAC under other circumstances previously and I know its not an issue with it.

http://www.soekris.dk/dam1021.html

This board is the 0.05% resistor version which will exhibit higher harmonic distortion (maybe 9dB or so) than the more expensive 0.01%/0.02% resistor version

The unit is driven I2S over HDMI cable from a Gustard U12.

1.jpg

2.jpg

3.jpg


FS output voltage is approx 4V rms over the balanced outputs, 2V single ended.

0db - As noted typically for r2r, high harmonic distortion. I probably wouldnt be worried if the levels were below -100dB but I think we are possibly in the territory of colouring the sound. I would say that the 0.01% resistor version probably gets below the level where I would be concerned.

upload_2018-2-7_20-19-26.png



-6dB
upload_2018-2-7_20-26-36.png



-60dB ignoring the 50 hz and harmonics which are a measurement issue, the noise floor is very low.

upload_2018-2-7_20-32-48.png



-90.31dB 16 bit

upload_2018-2-7_20-33-25.png

upload_2018-2-7_20-34-7.png



-90.31dB 24 bit - good waveform, seems to confirm low noise floor

upload_2018-2-7_20-36-3.png

upload_2018-2-7_20-36-37.png



12kHz Jitter - minor sidebands too low to be audible

upload_2018-2-7_20-40-43.png



IMD SMPTE - yikes messy! Imd is related to hd so this is no surprise.

upload_2018-2-7_20-42-20.png



White Noise + 19kHz FR and imaging. Now the stock firmware comes with 5 selectable filter types, however IIRC I loaded a set that were developed by a 3rd party. Thus board allows you to design your own filters. This one is fast roll off.

upload_2018-2-7_20-44-33.png



Im planning to produce a linearity chart a la Amirs latest reports

EDIT: Yer tis

I need to add some more data points but the DAC linearity is good to at least 17bits. Ill give the benefit of the doubt until I re-measure, but Ill say its good to the intermediate 18 bit point.

upload_2018-2-8_20-6-49.png


Reversed to be more consistent with Amirs graphs - I wonder how much this would improve with the 0.01% resistor version???

upload_2018-2-8_21-12-17.png
 
Last edited:

Blumlein 88

Major Contributor
Joined
Feb 23, 2016
Messages
5,692
Likes
4,693
#2
When doing the 19khz plus noise test are you using a single signal with combining both parts?
 

March Audio

Major Contributor
Manufacturer
Patreon Donor
Joined
Mar 1, 2016
Messages
3,467
Likes
2,921
Location
Perth Western Australia
#4

March Audio

Major Contributor
Manufacturer
Patreon Donor
Joined
Mar 1, 2016
Messages
3,467
Likes
2,921
Location
Perth Western Australia
#6
Okay, it looked odd. And I intended to download your test files, but somehow never did.

Good job on these measurements btw.
Same track that was used on the Meridian explorer and Chord Mojo tests. Yes please download the files and check them, I have a certain ability to balls things up so it would be good to validate them. :)
 
Joined
Jan 12, 2017
Messages
12
Likes
3
#7
Great work.
Have this dac. Sounds wonderful. I also have a Gustard U12 with hdmi in to the dam1021. However I ditched that in favor for a crappy raspberry pi and i2s (short cables) direct to the dam1021..
I am also using the unbuffered outputs. Notice this test is using the buffered out...
 

svart-hvitt

Major Contributor
Joined
Aug 31, 2017
Messages
2,168
Likes
1,006
#8
Great work.
Have this dac. Sounds wonderful. I also have a Gustard U12 with hdmi in to the dam1021. However I ditched that in favor for a crappy raspberry pi and i2s (short cables) direct to the dam1021..
I am also using the unbuffered outputs. Notice this test is using the buffered out...
Welcome!
 

March Audio

Major Contributor
Manufacturer
Patreon Donor
Joined
Mar 1, 2016
Messages
3,467
Likes
2,921
Location
Perth Western Australia
#9
Great work.
Have this dac. Sounds wonderful. I also have a Gustard U12 with hdmi in to the dam1021. However I ditched that in favor for a crappy raspberry pi and i2s (short cables) direct to the dam1021..
I am also using the unbuffered outputs. Notice this test is using the buffered out...
Even though the PI apparently has a dodgy MCLK, it should work fine with the 1021 as it reclocks the data. Also the i2s is isolated, so I dont actually see the PI as a "crappy" solution :)

I havent tried the un-buffered outputs straight from the DAC. I will see if I can get time to hook that up and test.
 
Joined
Feb 9, 2018
Messages
40
Likes
79
#10
I would just like to point out to you that the dam1021 is level linear down to last bit, thanks to the sign magnitude principle.... So your "DAC linearity" measurement is probably more the measurement of the ADC in your equipment, you're probably using the same QA401 like the other measurements, right ? Try redo it with professional equipment, like the AP used on most other tests....
 

March Audio

Major Contributor
Manufacturer
Patreon Donor
Joined
Mar 1, 2016
Messages
3,467
Likes
2,921
Location
Perth Western Australia
#11
I would just like to point out to you that the dam1021 is level linear down to last bit, thanks to the sign magnitude principle.... So your "DAC linearity" measurement is probably more the measurement of the ADC in your equipment, you're probably using the same QA401 like the other measurements, right ? Try redo it with professional equipment, like the AP used on most other tests....
Hi, great to have your feedback, warm welcome.

We are actually comparing the QA401 with Amirs AP results to validate its efficacy. I have measured a DAC and sent it to Amir for comparison. I fully understand your concern, which I share, which is why we are performing this validation exercise.

Also I have measured several other dacs which prove to be more linear to a much lower level. I will post those results up soon.

Can I ask if you have any test data of your own which you can share, and why you say the dac is linear down to last bit? Also could you comment generally on what reasons there could be for a R2R design not to be linear down to the lowest levels?
 
Last edited:
Joined
Feb 9, 2018
Messages
40
Likes
79
#12
I don't have that good test equipment yet, but over at SBAF they measured the dac1541 to be linear down to -140 dB, using a Dscope III. The dam1021 use the same R-2R Sign Magnitude DAC core. The key is Sign Magnitude, I'm trying to explain it over at the soekris.dk website, but not everybody get it....

R-2R DACs that are not Sign Magnitude, like those using Analog Device R-2R chips (won't mention a manufacture, but you probably know them...), especially those using the cheaper 16 bit or 18 bit version, they have serious problems at low levels.... Low cost Delta Sigma chips can also have limitations, while high end Delta Sigma ones measure good, at least while not playing music....
 

March Audio

Major Contributor
Manufacturer
Patreon Donor
Joined
Mar 1, 2016
Messages
3,467
Likes
2,921
Location
Perth Western Australia
#13
I don't have that good test equipment yet, but over at SBAF they measured the dac1541 to be linear down to -140 dB, using a Dscope III. The dam1021 use the same R-2R Sign Magnitude DAC core. The key is Sign Magnitude, I'm trying to explain it over at the soekris.dk website, but not everybody get it....

R-2R DACs that are not Sign Magnitude, like those using Analog Device R-2R chips (won't mention a manufacture, but you probably know them...), especially those using the cheaper 16 bit or 18 bit version, they have serious problems at low levels.... Low cost Delta Sigma chips can also have limitations, while high end Delta Sigma ones measure good, at least while not playing music....
Well, if there is error in the data we present here we want to correct it. Please trust that we are going through the exercise of validating the QA401 results. Also Im sure if you send a unit to Amir he would be very pleased to measure it on his AP as there has been a lot of interest on the forum regarding R2R dacs.

In the meantime here are the results from a cheap Chinese ES9038Q2M board. Note Flat top window used for amplitude accuracy hence poor spectral resolution. It seems to do quite well WRT linearity, that or it miraculously mirrors the non linearity of the QA401 :)

upload_2018-2-9_11-25-13.png

upload_2018-2-9_11-25-52.png

upload_2018-2-9_11-28-1.png

upload_2018-2-9_11-28-56.png

upload_2018-2-9_11-30-32.png

upload_2018-2-9_11-33-59.png

upload_2018-2-9_11-36-40.png

upload_2018-2-9_11-38-29.png

upload_2018-2-9_11-49-38.png

upload_2018-2-9_11-47-28.png


upload_2018-2-9_12-0-58.png
 

Attachments

Last edited:
Joined
Feb 9, 2018
Messages
40
Likes
79
#14
Could be your measuring setup or configuration, you did mention in your first post that you had problems with power line noise/harmonics, if that don't get filtered then it will probably be included and generate those errors....

I could send Amir one of my DACs, there seems to be a focus on cheaper DACs here, and the cheapest Schiit R-2R one didn't do that good, so a soekris dac1321 would be a good candidate to show how good a low cost R-2R DAC can be....
 

March Audio

Major Contributor
Manufacturer
Patreon Donor
Joined
Mar 1, 2016
Messages
3,467
Likes
2,921
Location
Perth Western Australia
#15
Could be your measuring setup or configuration, you did mention in your first post that you had problems with power line noise/harmonics, if that don't get filtered then it will probably be included and generate those errors....

I could send Amir one of my DACs, there seems to be a focus on cheaper DACs here, and the cheapest Schiit R-2R one didn't do that good, so a soekris dac1321 would be a good candidate to show how good a low cost R-2R DAC can be....
Actually I have looked back at the previous measurements I performed with the board using a completely different ADC and they too have more 50Hz than expected. The connection is simply going from the board to the diff in on the QA401. I am measuring a 1kHz tone and there are no significant mains spuria there. No obvious reason for the mains spuria and its not a problem I see with other dacs, as you can see on the es9038 board above. I will repeat the measurements at a non 50Hz multiple where there are no spuria present.

Lets not dwell on this though, I would rather see one of your finished dac products measured than a diy effort and the potential issues that could create. PM Amir and Im sure he will jump at the chance to measure any of your dacs.

There isnt necessarily a focus on cheaper dacs, but it is very clear that more expensive dacs dont necessarily offer any benefit. Or perhaps I should rephrase that; more expensive dacs with boutique components (you know like femto clocks :) ) need to demonstrate they offer genuine observable technical improvement.
 
Last edited:
Joined
Feb 9, 2018
Messages
40
Likes
79
#16
I agree that we should not spend time on this. The problem is probably the 50 hz power line noise, which don't have anything to do with the dam1021 module itself. You just need to be careful about posting measurements which really are a result of something else.... Try measure with a 400 hz HP filter, if the QA401 software have one.
 

March Audio

Major Contributor
Manufacturer
Patreon Donor
Joined
Mar 1, 2016
Messages
3,467
Likes
2,921
Location
Perth Western Australia
#17
I agree that we should not spend time on this. The problem is probably the 50 hz power line noise, which don't have anything to do with the dam1021 module itself. You just need to be careful about posting measurements which really are a result of something else.... Try measure with a 400 hz HP filter, if the QA401 software have one.
I dont share your view that it is due to the 50Hz, its presence doesnt prevent measurement of other signals. It concerns me its there and that I have seen it in other measurement configurations with the board. I will also repeat with a different ADC.

As I asked earlier, could you explain why sign magnitude makes the dac immune to linearity errors. Entirely linear to -140dB as you mentioned earlier is quite impressive. Also if that linear, why do we still see significant harmonic distortion?
 
Last edited:
Joined
Feb 9, 2018
Messages
40
Likes
79
#18
I dont share your view that it is due to the 50Hz, its presence doesnt prevent measurement of other signals. It concerns me its there and that I have seen it in other measurement configurations with the board. I will also repeat with a different ADC.

As I asked earlier, could you explain why sign magnitude makes the dac immune to linearity errors. Entirely linear to -140dB as you mentioned earlier is quite impressive. Also if that linear, why do we still see significant harmonic distortion?
The 50 Hz is just a possibility, I don't know your exact setup or the QA401, just know that in principle my DACs are all level linear down to nothing, and can't imagine anything that would change that for the specific dam1021 build.

You need to distinguish between absolute linearity (which create the distortion) and level linearity, basically in a Sign Magnitude DAC the distortion is relative to the actual signal level, while in a non Sign Magnitude DAC the distortion is relative to the full signal level, increasing as the signal level goes down.

I have a little tech info at http://www.soekris.dk/products.html, or maybe this snippet from the PCM1704 datasheet help:

The sign-magnitude architecture, which steps away from
zero with small steps in both directions, avoids any glitching
or large linearity errors, and provides an absolute current
output. The low-level performance of the PCM1704 is such
that true 24-bit resolution can be realized around the critical
bipolar zero point.

As I said, Sign Magnitude can be a little tricky to comprehend, don't know how to explain if further.
 
Last edited:

March Audio

Major Contributor
Manufacturer
Patreon Donor
Joined
Mar 1, 2016
Messages
3,467
Likes
2,921
Location
Perth Western Australia
#19
So why arent all dacs sign magnitude if the methodology eliminates linearity errors? Thats a holy grail.

Ok, Im going to plead ignorance here. Lets discuss how the dac creates the output voltage. This is for the purpose of understanding what potential sources of error there could be in the process, because at the moment you are essentially saying there are no potential sources of error. Please excuse me for being a little sceptical about that. :)

Just simply describe what the dac does - a voltage or current is switched across a resistor for example.
 

DonH56

Major Contributor
Technical Expert
Patreon Donor
Joined
Mar 15, 2016
Messages
2,694
Likes
2,790
Location
Monument, CO
#20
This may be helpful in explaining how some DACs work: https://www.audiosciencereview.com/...ital-audio-converters-dacs-fundamentals.1927/ There are other threads on sampling and delta-sigma DACs.

Sign-magnitude implementation has some other issues but does reduce the zero-crossing glitch most R2R DACs exhibit. A purely binary design will generate a half-full-scale glitch at the zero crossing, so is a big deal and especially at low levels. The glitch may be filtered and if unrelated to the signal (that never happens in the real world) does not add nonlinearity but signed-magnitude resolves that problem nicely. Most R2R DACs actually implement the top few MSBs as unary, one cell per step, to reduce the glitch and improve linearity (at the cost of complexity and SWaP, natch).
 

Similar threads

Top Bottom