AD5422 accuracy out of specification

Dear community,

I use the 4-20mA current output and cannot achieve the specified accuracy.

Example at 20°C: set value 20mA, measured value 19.9772mA which gives an error of -0.114% FSR.

Such a high deviation is not expected at a constant temperature since the TUE is +/-0.15% FSR over the full temperature range!

Several configurations (internal/external reference, Rset, boost etc.) were tested without success. Schematic and measurement results are attached.

Please consider and give a hint. Thank you very much!

DAC-failure_description.pdf
  • Hi,

    Apologies for the delay. Just some quick points:

    • The data sheet version mentioned in the pdf file is Rev. G. The latest version of the data sheet as of this writing is Rev. L.
    • The calculation for the FSR at 4-20mA range is (note that the full scale range is 20mA-4mA = 16mA):
      • error = (difference mA/ 16 mA ) * 100%
    • In the schematic on page 3, the version of the part being used is the A version. Could you verify what was really used in the actual board?
    • It should be noted that the TUE specifications in the data sheets are measured using the ideal Rset and ideal external reference. You would need to take the initial error and the temperature coefficients into account. You can add them to the TUE value in the data sheet using the root sum of squares equation. The same thing is done if you are using the internal reference. You can see a sample computation using the AD5421 by clicking here.
    • Using the internal Rset will give a worse number for TUE because the internal resistor will have drift, tolerances, and self-heating.

    Regards,

    Mark

  • Thank you, Mark!

    Indeed, our measurements comply to the datasheet, if interpreted correctly. They have been confirmed by measurements on the EVAL board as well.

    Unfortunately, the definition of the TUE and FSR% is not clear. The above link is an eye opener. Why isn´t it explained in that way in the dataseet (there is a comment, but quite hidden)?

    To be honest, it is quite difficult to estimate the real behavior of the device by reading the datasheet only. We have invested high effort in fundamental investigation and analysis to figure out the characteristics. Only after measuring we understood the datasheet..... should be vice versa......

    Just to complete the discussion:

    - We have used the B-type of the AD5422 in all measurements (the schematic say A-type which is misleading).

    Thanks a lot for your advice!

    Regards, Tino