• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 46 subscribers
  • Views 385 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.

Original question:

ADS1261-Q1: Combining DC errors and noise

ADS1261-Q1: DC Measurement Error Review

Drew Nakamura
Prodigy 130 points
Part Number: ADS1261-Q1

Hi,

Would it be possible for someone to review my ADC error calculations? If so, what would be the best way to share that? I'm seeing that errors on the order of 0.5mV are causing errors in the corresponding temperature reading of 0.35C at 125C (using an 0.01%, 5ppm/C bias resistor). It seems like a rather large error so I just would like to see if I'm handling the sensitivity analysis correctly.

Thanks,

Drew

  • 0
    TI__Guru 52226 points

    Hi Drew Nakamura,

    Assuming your calculations are clear, we can review them. You should be able to post an Excel document to the thread using the Insert dropdown

    -Bryan

  • 0 in reply to Bryan Lizon86
    Prodigy 130 points

    TTV ADC Analysis TI.xlsx

    Hi Bryan,

    Please see the "ADC Bias" tab in the attached spreadsheet. I have two tables in here because I'm trying to compare the RefOut vs IDAC excitations to convince myself that I'm not missing anything by not going with the more standard IDAC output. I have included circuit diagrams next to each table to represent the approach.

    The main thing I'm skeptical on is the dRTTC/dV_e calculation. I used derivatives in the spreadsheet, but I'm also thinking of taking the RTTC formulas and replacing RTTC with (RTTC+dRTTC) and VTTC with (VTTC+dVTTC) and solving for dRTTC could work. This would give me a much smaller error.

    Please let me know if you have any questions about my work. Hopefully the comments are clear enough.

    Thanks,
    Drew

  • +1 in reply to Drew Nakamura
    TI__Guru 52226 points

    Hi Drew Nakamura,

    Will your product use the entire ADC temp range (-40C to 125C)? If not, that would help reduce the overall error budget

    Also, at first glance it seems you can use the PGA at a gain of 1, instead of bypassing the PGA. This will significantly reduce the leakage current you are calculating

    You might also consider separating out the initial errors and over-temperature errors so you can understand where the largest error source actually is (and then if you need to calibrate)

    I will also admit that I did not fully understand what are you trying to show with the dRTTC/dV_e, etc., calculations. I understand how you got to the "% measurement error", but I am not sure how you determined the other errors, or why they are important? I would have calculated the sensitivity of your sensor (ohms per degree C) and then determined the measurement error from there.

    -Bryan

  • 0 in reply to Bryan Lizon86
    Prodigy 130 points

    Hi Bryan,

    Thanks for the review. As always, I appreciate the quick turnaround. The plan is to use the ADC from -40C to 105C, but I opted for 125C on the top end to account for potential thermal rise in the board. The actual thermal rise likely won't be this much, but I don't have any thermal data yet so I opted for a worst case assumption.

    I had gone back and forth on using PGA = 1 vs bypass, but since it may be difficult for us to apply a full scale input voltage for a gain calibration I opted for the PGA bypass instead. Even though the ADC input leakage will be reduced, I was calculating a pretty significant gain error. I will see if I can design in a way to easily apply that full scale voltage during bring up so that the gain calibration can be performed. 

    The main thing I was trying to show with the final 4 calculations is the temperature error (dT_e) that will result from the DC error (dV_e) at each end of the temperature range. This is the part I was getting a bit confused on, since it was hard to grasp the idea that an 0.5mV error could impact the temperature reading by 0.34C. I may have overcomplicated it, but I couldn't think of a better way to go about it. Do you have any suggestions for this?

    Thanks,
    Drew

  • 0 in reply to Drew Nakamura
    Prodigy 130 points

    Following up on this, I believe there is validity in the derivative approach I took. Calculating the temperature using the errorless voltage and then again with the errorless voltage + error, I get a difference in temperature equal to my dT_e values. At this point I will have to see if that's acceptable for our use case or not.