• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Subscribers 46 subscribers
  • Views 550 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

DAC80502: How to calculate TUE?

Takumi Suzuki1
Mastermind 8610 points
Part Number: DAC80502

Tool/software:

Hi,

Let me ask you a basic question as below about TUE.

Q1) TUE of page 5 datasheet

Will you tell me what does it mean?

The "TUE" includes any other factors (Offset, Gain, etc) ?

Q2)  Datasheet page 35

I suppose that 0.112%FSR(before calibration) is from "equation 2" and 0.05%FSR(after calibration) is from "equation 5".

Will you teach me how to calculate the value?

Which data should I use to calculate it?

Will you tell me how to calculate with concreat value?

Thank you for your support.

Best Regards,

Takumi

  • 0
    TI__Mastermind 41863 points

    Hi Takumi, 

    TUE stands for total unadjusted error. Typically this will include all DAC errors. The value in the datasheet is a measured value (including offset error, gain error, and INL) which is better than a calculated value because the calculated value is a very conservative value. We usually don't see devices with most extreme gain error, offset error, and INL at the same time, but the calculated TUE uses these worst case values.

    The equations on page 35 are showing how to include the drift values and reference error to the measured TUE. You are correct  0.112%FSR(before calibration) is from "equation 2" and 0.05%FSR(after calibration) is from "equation 5".

    For the calculation, follow equation 2. Use the TUE spec from the electrical characteristics table. This spec was measured at one temperature, so equation 2 is adding the effects of the drift and reference. The drift specs in the datasheet are given in unit/C, so you would multiply by your temperature range to know the effect of the drift. In the datasheet example they used 0 to 100C. Before the calculation, all of the errors should be converted to the same units. Either LSB, or %FSR. This is what is shown in equations 3 & 4

    Best,

    Katlynne Jones

  • 0 in reply to Katlynne Jones
    Mastermind 8610 points

    Katlynne-san,

    Thank you so much for your support !

    In accordance with your advice, I have tried to calculate the "equation 2".

    The each factor using calculation is as following.

       Internal Ref 2.5V use, temp = 100C

       TUE : 0.1%FSR  (from page 5)

       TCoe : 2uV/C = (2uV/2.5V)*100C*100 = 0.008%FSR  (from page 5)

       TCge : 1ppmFSR = 0.01%FSR  (from page 5)

       Eref : 0.1%FSR  (from page 6, VOLTAGE REFERENCE OUTPUT)

       GAIN : 1

       TCref : 5ppm/C = (5*100C)/10^4 = 0.05%FSR  (from page 6, VOLTAGE REFERENCE OUTPUT)

    The calculation result becomes 0.1505%FSR, not 0.112%FSR...

    Will you please check the above values I used to calculate?

    About "equation 5", I could succeed to get the same value, 0.05%FSR, in case of INL=1.

    And let me ask you one more question.

    There is a term of "Full-scale error" on the page 5.

    Will you teach me what is this and what do I use for?

    Thank you for your kind support.

    Best Regards,

    Takumi  

  • +1 in reply to Takumi Suzuki1
    TI__Mastermind 41863 points

    Hi Takumi-san,

    I'm still not able to get the exact 0.112% value, but I think the issue is that this section of the datasheet was written while the device was still in preview, and it looks like they updated some specs for the final released version of this part without going back to revise the example in this section. For example the paragraph at the beginning of section 9.2.2 says the TUE is 0.02% FSR typical, but the electrical characteristic table shows 0.04%FSR typical. When I use the 0.02% value I get 0.114%FSR which is closer to the value calculated in the example. 

    Your calculation is correct assuming a 2.5V full scale range. My only comment would be to use the typical value for TUE. The typical value is given at 25°C, and the min/max value is given across the entire temperature range this device is characterized for (–40°C to +125°C). So in this example, since we are only concerned with the 0 to 100C range, you should use the typical spec and then the drift specs will add only the drift for the 0°C to 100°C range. 

    The full-scale error and zero-scale errors are end point errors due to the internal output buffer not being able to output values at the power rails. 

     

    Best,

    Katlynne Jones

  • 0 in reply to Katlynne Jones
    Mastermind 8610 points

    Hi Katlynne-san,

    Thank you very much  for your kind reply !!

    I understood well.

    Regards,

    Takumi