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.

TMP117: Clarification of temperature accuracy

Part Number: TMP117


I would like some clarification on the stated accuracy of the TMP117A under section 6.5 Electrical Characteristics. Given the "Typical" specifications are implemented (V+ = 3.3V at 25C), what is the minimum and maximum error for the range -20C to 50C? Or put another way, does the +/- 0.05C accuracy apply only at 25C or the entire -20C to 50C range?

Thank you

  • Hi Greg,

    Thank you for your post!

    The typical error of ±0.05°C applies to the entire -20°C to 50°C range. The typical specifications of 25°C and 3.3 V are unless otherwise noted. In this case, the table specifies -20°C to 50°C.

     

    Regards,

    Nicole

  • Thank you for clarifying that up. I do have a couple of follow-up questions.

    1) Figure 2 shows 3-sigma average of the temperature error close to 0.1C for the -20C to 50C range.

    2) Figure 7 graph indicates errors going out to -60 mC.

    How do I reconcile the graphs with your statement?

    Thank you

  • The 3-sigma average shown in Figure 2 corresponds with the maximum error ratings specified in section 6.5. This is to ensure that no device will fail to meet the maximum specification. The typical error of +/- 0.05C represents typical offset of the devices.

    Figure 7 indicates errors out to -60 mC and this supports the +/- 0.1C maximum error for the -20C to 50C temperature range.

    Regards,

    Nicole

  • Hi Nicole,

    I would agree with your interpretations if the "typical conditions" were not specified for these graphs. You affirmed in a previous response that the TMP117 is accurate to +/- 0.05C in the range of -20C to 50C when operated at 3.3V so Figure 2 seems to contradict this.

    I appreciate your patience. I just want to make sure that I understand the datasheet information.

    Best regards,

    Greg

  • Hi Greg,

    I apologize for any confusion. The use of the word "typical" in the caption of Figure 7 is to show the distributed accuracy of tested devices, and includes those that fall into the +/- 0.05C error range. Additionally, Figure 7 shows that no device will have an error greater than +/- 0.1C.

    The word "typical" in section 6.5 is referring to a 1-sigma value to illustrate that most of the devices will have that accuracy.

    Best regards,

    Nicole

  • Hi Nicole,

    Thank you for clarifying. Now sections 6.5 and 6.8 make sense to me. One (hopefully) last question:

    If a one-point calibration was performed for the offset of an individual TMP117 sensor, would the sensor then be able to meet a 3-sigma error of +/-0.05C across the -20C to 50C range? Has there been any testing of this by TI?

    Thank you,

    Greg

  • Hi Greg,

    It may be possible to improve on factory calibration, but there will still be errors that a one-point calibration cannot eliminate. Any improvements cannot be guaranteed outside of the datasheet specifications.

    Additionally, we have not had application-level testing of this.

    Best regards,

    Nicole