• State Resolved
  • Locked Locked
  • Replies 2 replies
  • Subscribers 30 subscribers
  • Views 314 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.

TMP117: Precision vs Accuracy

Ya Zhang
Prodigy 30 points
Part Number: TMP117

Hi,

I'm a bit confused about what the +-0.1deg accuracy of the TMP117 realistically means. Is it an absolute accuracy? E.g. suppose there is an ideal sensor that measures the "real" temperature without any errors, the TMP117's results lands within +-0.1deg of that of the ideal sensor?

For my application I am only concerned with temperature change, not absolute temperature. Is my measurement noise floor then limited by this 0.1deg accuracy, or the 0.0078deg resolution, or somewhere in between? Within the same sensor, when the reading increases by 1 bit (0.0078deg), does that reflect a real temperature increase, or should it be treated as purely quantisation noise?

The above is all assuming 20-30deg temperature range, 8 averages, continuous cycle, 1Hz cycle time (basically the default configuration).

Thanks in advance,

Ya

  • 0
    TI__Mastermind 28070 points

    Hello Ya,

    Thank you for your post. Your understanding of the TMP117's +/-0.1C accuracy claim is correct. The TMP117 will be within accuracy specification when compared to a "Golden Reference" that reflects the actual temperature without error. 

    Ya Zhang said:
    For my application I am only concerned with temperature change, not absolute temperature. Is my measurement noise floor then limited by this 0.1deg accuracy, or the 0.0078deg resolution, or somewhere in between? Within the same sensor, when the reading increases by 1 bit (0.0078deg), does that reflect a real temperature increase, or should it be treated as purely quantisation noise?

    Your measurement noise floor would really be limited by the thermal noise in your environment and the repeatability of the TMP117(which is typically +/- 1LSB). So in an environment that is completely thermally stable, you could typically expect the TMP117 to return the same temperature within +/- 1 LSB (0.0078125C). If you saw a change of 2 LSB or more, you could then say that there was an actual change in temperature. 

    In practice we do not see many temperature environments which are that stable. You may want to try placing a TMP117EVM or a reference probe of similar accuracy/resolution into your setup to try and determine what level of temperature oscillation represents "normal" in your application. 

    Best Regards,
    Brandon Fisher

  • 0 in reply to Brandon Fisher94
    Prodigy 30 points

    Thank you for the explanation. It was very helpful. The system I'm working with is very thermally stable, so I am actually seeing the +/- 1LSB noise floor above 0.001Hz.  I have wondered if the tiny fluctuations above this 1000 sec period is valid data and reflects real temperature change, and you have confirmed that it is. Thanks very much.