+ Post New Thread
Results 1 to 7 of 7
  1. #1
    Member level 5
    Points: 1,100, Level: 7

    Join Date
    Jan 2016
    Posts
    83
    Helped
    0 / 0
    Points
    1,100
    Level
    7

    Gradient problems in current steering DACs

    Hello all,
    I have been hearing about gradient errors in current steering DACs. Could someone help me understand what it is?
    Thank you

    •   AltAdvertisement

        
       

  2. #2
    Advanced Member level 3
    Points: 5,771, Level: 18
    Dominik Przyborowski's Avatar
    Join Date
    Jun 2013
    Location
    Norway
    Posts
    755
    Helped
    318 / 318
    Points
    5,771
    Level
    18

    Re: Gradient problems in current steering DACs




    •   AltAdvertisement

        
       

  3. #3
    Member level 5
    Points: 1,100, Level: 7

    Join Date
    Jan 2016
    Posts
    83
    Helped
    0 / 0
    Points
    1,100
    Level
    7

    Re: Gradient problems in current steering DACs

    Quote Originally Posted by Dominik Przyborowski View Post
    Thank you Dominik. But most of the talk about how to correct it. I wanted to understand what are the causes for it?



    •   AltAdvertisement

        
       

  4. #4
    Advanced Member level 3
    Points: 5,771, Level: 18
    Dominik Przyborowski's Avatar
    Join Date
    Jun 2013
    Location
    Norway
    Posts
    755
    Helped
    318 / 318
    Points
    5,771
    Level
    18

    Re: Gradient problems in current steering DACs

    Have you ever cut bread? Have the slices always been the same, ideally equal with uniform thickness? Have the air bubbles in the bread been always identical, with the same shape and volume?


    1 members found this post helpful.

  5. #5
    Full Member level 1
    Points: 2,388, Level: 11
    Achievements:
    7 years registered

    Join Date
    Jan 2009
    Posts
    105
    Helped
    16 / 16
    Points
    2,388
    Level
    11

    Re: Gradient problems in current steering DACs

    Non-linearity errors (DAC or ADC) are usually caused by IC production tolerances and/or bad IC design (power distribution)



  6. #6
    Advanced Member level 5
    Points: 39,802, Level: 48

    Join Date
    Mar 2008
    Location
    USA
    Posts
    6,429
    Helped
    1878 / 1878
    Points
    39,802
    Level
    48

    Re: Gradient problems in current steering DACs

    Across-wafer uniformity of transistor attributes is never
    ideal and often contains a gradient in addition to random
    mismatch. The gradient is not necessarily linear nor
    monotonic (e.g. wafer bowing may cause depth-of-field
    related photo errors that make center lower than both
    edges for some param, or whatever).

    If you have a wafer-scale gradient then you have a
    chip-scale gradient (of lesser magnitude).

    There can also be thermal gradients especially on parts
    that have an output buffer / driver that dissipates much
    power. In such cases special care with matched devices'
    location, orientation relative to the on-chip thermal
    gradient is required to avoid "bad IC design". Far from,
    and matched-rack orthogonal-to, is about all you can
    do. However even this may be a compromise, such as a
    case where the output buffer high and low sides can't
    be merged neatly, and the thermal gradient in the low,
    max load case will be different vector than the high,
    max load output case. Then you have to do your best
    and eat the rest.

    Distributing MSB, MSB-1, ... higher order bits as segments
    can help major carry DNL and gradient issues.


    1 members found this post helpful.

    •   AltAdvertisement

        
       

  7. #7
    Member level 5
    Points: 1,100, Level: 7

    Join Date
    Jan 2016
    Posts
    83
    Helped
    0 / 0
    Points
    1,100
    Level
    7

    Re: Gradient problems in current steering DACs

    Quote Originally Posted by Dominik Przyborowski View Post
    Have you ever cut bread? Have the slices always been the same, ideally equal with uniform thickness? Have the air bubbles in the bread been always identical, with the same shape and volume?
    No. I understand now. Thank you



--[[ ]]--