Soekris dac1541 technical measurements

Discussion in 'Source Measurements' started by atomicbob, Aug 18, 2017.

  1. johnjen

    johnjen Friend

    Friend
    Joined:
    Sep 29, 2015
    Likes Received:
    543
    Dislikes Received:
    5
    Trophy Points:
    93
    Location:
    Well Grounded
    That there is a TON and 1/2 of work.
    Thanks for all the effort.

    It is curious that there is such a difference between balanced and single ended.
    Which makes me wonder if it is SE internally with Bal -> SE/SE -> Bal 'converters' on the inputs/outputs.

    JJ
     
    Changeling likes this.
  2. Cspirou

    Cspirou They call me Sparky

    Friend
    Joined:
    Sep 27, 2015
    Likes Received:
    3,152
    Dislikes Received:
    4
    Trophy Points:
    113
    Location:
    Irvine,CA
    I recall for the DIY modules that SE was unbuffered while the BAL output went through an opamp. I don't know what the situation is for the dac1541
     
  3. soekris

    soekris MOT - Soekris Engineering

    Friend
    Joined:
    Mar 25, 2016
    Likes Received:
    438
    Dislikes Received:
    1
    Trophy Points:
    53
    Location:
    Denmark
    First, thanks to atomicbob, I know myself how long time it can take to make so many the measurements, even with good tools, although measuring the Windows WDM driver model seems a little waste of time, we all know how the WDM path ruins things....

    I am a little puzzled by the first "Frequency Response" measurements, it look like the start to fall of way too soon, while the "imaging" measurements looks correct.

    But let me explain the output buffer / amplifier, before too many theories comes up....

    The dac1541 is fully balanced, with two complete R-2R DAC strings per channel. They feed into a fully differential discrete high speed amplifier with zero negative feedback, the two outputs are a result of both R-2R stings, so out+ = in+ - in- and out- = in- - in+, the out+ and out- goes to the balanced outputs while only out+ goes to the single ended outputs, the single ended outputs are still a result of both R-2R strings. The same amplifier is used for both line and headphone outputs, when used for the headphone outputs it is set for higher bias current, and additional gain as needed with higher volume setting. There are no opamps in the signal path, just that single amplifier.

    The balanced and single ended outputs should sound and measure the same, but the balanced output can deliver higher levels and could balance out some things.... So it's a puzzle to me with that big differences between balanced and single ended, especially the balanced "inferred jitter" measurement which should look exactly like the single ended.... You might want to redo that, something must be wrong....

    To me the measurements look like what a very very good R-2R DAC should look like, but the measurement equipment seems to run into its limitations at times, not really a big surprise with such a good DAC....
     
  4. atomicbob

    atomicbob dScope Yoda

    Friend
    Joined:
    Sep 27, 2015
    Likes Received:
    6,304
    Dislikes Received:
    2
    Trophy Points:
    113
    Location:
    On planet
    Appreciate the thoughts on how much time this takes. A considerable amount of extra time was spent with diagnostics run on behalf of the very issue you mention. I made 10 to 20 measurements on those that gave me pause such as the inferred jitter. Checking ALL dScope settings, changing a few to see what would happen, changing cables, orientation of physical devices, cycling power, checking AES vs spdif, etc. When every measurement provided a similar answer I finally picked what was representative of the multiple measurements and moved on. The results reported are repeatable. I note the dac1541 is representative of an excellent R-2R implementation.
     
    Last edited: Aug 19, 2017
    aufmerksam, landroni, Grahad2 and 7 others like this.
  5. atomicbob

    atomicbob dScope Yoda

    Friend
    Joined:
    Sep 27, 2015
    Likes Received:
    6,304
    Dislikes Received:
    2
    Trophy Points:
    113
    Location:
    On planet
    Updates to section 14 for data resolution and added some comments. Correction to 20 Hz square wave -3 dBFS Balanced output.

    27 bit R-2R architecture is impressive for resolution. Well done @soekris !
     
    Asgeir101, aufmerksam and Cspirou like this.
  6. Nanekiu

    Nanekiu Rando

    Joined:
    Jan 3, 2016
    Likes Received:
    3
    Dislikes Received:
    0
    Trophy Points:
    8
    Location:
    Trondheim Norway
    Just curious but when is high Common Mode Rejection needed?
     
  7. atomicbob

    atomicbob dScope Yoda

    Friend
    Joined:
    Sep 27, 2015
    Likes Received:
    6,304
    Dislikes Received:
    2
    Trophy Points:
    113
    Location:
    On planet
    A serious example: running 50 meters of cable from stage to control room through the catwalk cable trough with lighting dimmer packs also located in the catwalk cable trough.
     
    Dino likes this.
  8. Armaegis

    Armaegis Friend

    Friend Contributor
    Joined:
    Sep 27, 2015
    Likes Received:
    2,806
    Dislikes Received:
    0
    Trophy Points:
    113
    Location:
    Winnipeg
    Next to a 16 channel snake, possible speaker cable runs, and a microwave in the break room.
     
    atomicbob likes this.
  9. Junki

    Junki Acquaintance

    Joined:
    Aug 13, 2017
    Likes Received:
    95
    Dislikes Received:
    53
    Trophy Points:
    18
    Location:
    United States
    Highlight for me personally: this is the best linearity I've seen in a R-2R DAC. Wonderful resolution.
     
    Last edited: Apr 3, 2018
  10. Clemmaster

    Clemmaster Friend

    Friend
    Joined:
    Sep 28, 2015
    Likes Received:
    693
    Dislikes Received:
    0
    Trophy Points:
    93
    Any plans on measuring the DAC1421 that I saw in your stack recently? :D
     
  11. atomicbob

    atomicbob dScope Yoda

    Friend
    Joined:
    Sep 27, 2015
    Likes Received:
    6,304
    Dislikes Received:
    2
    Trophy Points:
    113
    Location:
    On planet
    It is in the queue ... along with many others. I've had the ADI-2 Pro for many months now and it is in constant service, yet haven't found the time to perform a moderately complete test suite on it.
     
    Clemmaster likes this.
  12. FrivolsListener2

    FrivolsListener2 Rando

    Joined:
    Mar 27, 2018
    Likes Received:
    7
    Dislikes Received:
    0
    Trophy Points:
    3
    Location:
    Loveland, Colorado
    Now I want one.
     
  13. looks significantly better than the denafrips ares results across the board.
     

Share This Page