Soekris DAC1421 Measurements

Discussion in 'Source Measurements' started by purr1n, Apr 21, 2019.

  1. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    51,047
    Dislikes Received:
    85
    Trophy Points:
    113
    Location:
    Antarctica
    upload_2019-4-21_11-34-40.png

    More or less the same as Amir's.

    All the distortion is -85db down or lower. That's 0.005% THD, this means the measurements are excellent. Yes, R2R DACs measure worse relative to the sigma-delta DACs uses opamps, but that does not mean R2R DACs measure badly as some have interpreted.

    Saying this measurement looks bad or poor is like saying the 2015 Camaro ZL1 which does a 1/4 mile in 12.4 @ 117 mph is "bad" because the Corvette ZL1 will do the 1/4 mile in 10.8 @ 135mph. In real world street driving conditions, it doesn't matter.

    Amir measurement for comparison.
    upload_2019-4-21_10-52-30.png
     
    Last edited: Apr 21, 2019
    ultrabike, Jinxy245, haywood and 4 others like this.
  2. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    51,047
    Dislikes Received:
    85
    Trophy Points:
    113
    Location:
    Antarctica
    THD+N for 1kHz signal.
    SINAD at 0dbFs for 1 kHz signal is indeed about 85db. While not as good as the 100db numbers from modern sigma delta chips with opamps, 85db is still excellent and near the limit of microphones / mic-preamps.
    upload_2019-4-21_11-17-58.png
     
    ultrabike likes this.
  3. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    51,047
    Dislikes Received:
    85
    Trophy Points:
    113
    Location:
    Antarctica
    A few more THD+N:

    20-20kHz at -10dbFS
    upload_2019-4-21_11-27-11.png

    -30dbFS
    upload_2019-4-21_11-29-28.png

    -60dbFS
    upload_2019-4-21_11-32-29.png
     
  4. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    51,047
    Dislikes Received:
    85
    Trophy Points:
    113
    Location:
    Antarctica
    Back to the 1kHz FFTs.

    -30dbFS. Note scale is normalized to the -30dbFS signal = 0db.
    upload_2019-4-21_11-47-37.png

    This time at low output -60dbFS. Note scale is normalized to the -60dbFS signal = 0db.
    upload_2019-4-21_11-39-24.png

    And near the limit of 16-bit audio at -90dbFS. Note scale is normalized to the -90dbFS signal = 0db.
    upload_2019-4-21_11-38-48.png

    Note that the harmonic distortions go away at the lower levels.
     
    Last edited: Apr 21, 2019
  5. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    51,047
    Dislikes Received:
    85
    Trophy Points:
    113
    Location:
    Antarctica
    Conclusion:

    Nothing wrong with the measurements. In my experience, nothing seen here will correlate well or at all to human observation*.

    While R2R or discrete resistor designs do not measure as well as modern sigma-delta DACs with opamp output stages, the distortion and noise are well below that of the threshold of human hearing and on par with microphones and preamps on a good day in the studio.

    If we wanted to be nitpicky, the odd order distortions do seem to be more evident than the even order, and there is a bit of AC mains which isn't being totally filtered out. But again, these are so low in level that humans will not be able to hear them. Note that the harmonic distortions essentially go to zero for lower level signals - this is not the case for some R2R DACs.

    What constitutes a bad measurement? See here: https://www.superbestaudiofriends.o...ically-incorrect-audio-blog.7261/#post-249242

    *I strongly suggest the use of volume matched double or triple blind tests (and actually attempt via practice to make out perceptual differences of DACs - put your expectation bias away): https://www.superbestaudiofriends.org/index.php?threads/dac-blind-test-series.7518/
     
    Last edited: Apr 21, 2019
    Josh83, Ash1412, soekris and 3 others like this.
  6. soekris

    soekris MOT - Soekris Engineering

    Friend
    Joined:
    Mar 25, 2016
    Likes Received:
    609
    Dislikes Received:
    1
    Trophy Points:
    93
    Location:
    Denmark
    A note, it's not that there are no distortion at lower level, but a Sign Magnitude R-2R DAC's distortion is typically relative to the signal level, so at lower signal levels it disappear into the noise....
     
    Vtory likes this.

Share This Page