• State Resolved
  • Locked Locked
  • Replies 2 replies
  • Subscribers 46 subscribers
  • Views 77 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

DAC8760: Regarding total unadjusted error

Daisuke Higa
Genius 3515 points
Part Number: DAC8760

Hi support team.

I have two questions regarding TUE (total unajusted error).

1. Is the formula for calculating TUE as follows? Also, are the calculations given in the example correct?

TUE=sqrt(GE^2+OE^2+INL^2)
GE:gain error
OE:offset error

Example: When using at 4mA-20mA
FSR=16mA -> (4mA+16mA/2^n)
1LSB=16mA/2^16=0.244uA

Case 1　Internal Rset @Ta=25℃
GE_max=0.08*16mA/100=12.8uA
OE_max=0.07*16mA/100=11.2uA
INL_max=0.024*16mA/100=3.84uA

TUE=sqrt(12.8uA^2+11.2uA^2+3.84uA^2)=17.44uA=0.11%FSR

2. The specification in the data sheet is 0.08%FSR for Case 1, which is the most fluctuating case, so it is smaller than the calculated value.
I can't understand what this is.
Does this mean that an error of 17.44uA will occur in the worst case regardless of the code?

Best regads,

Higa

  • +1
    TI__Expert 3370 points

    Hi Higa-san,

    I believe your calculation to get a potential 0.11% FSR is correct.

    This question is related to how TI measures their typical and min/max datasheet values.

    Typical values is the range seen in ±σ (68%) of devices and min/max values is the range seen in ±5σ (99.9999%) of devices.
    So for each of your max cases for GE, OE, and INL, those are the highest values for ±5σ of devices, but extremely unlikely all on the same device.

    The TUE for this device appears to be directly measured and not calculated. The chances of these maximum error values for GE, OE, and INL appearing on the same device is very low, leading to the measured TUE range to be a lower value of ±0.08%.

    Daisuke Higa said:
    Does this mean that an error of 17.44uA will occur in the worst case regardless of the code?

    As the datasheet error values are based on a large sample size of testing, it would be very, very unlikely you would have a device with that large of an error for Case 1.

    Thanks,
    Lucas

  • 0 in reply to Lucas Burnette
    Genius 3515 points

    Hi Lucas-san

    Thank you for your answering.

    I understood well.

    Thanks!

    Regards,

    Higa