We would like to ask your assistance, clarifying the accuracy measurement of the device TMP100NA/250.
I assume if you repeat a measurement of a steady temperature, those measurement values will form a distribution roughly centered around the actual value. The spread of the distribution is related to the precision or standard error of the instrument. See the “Quantification” section of this webpage:
An example I found of this being quantified is in this engineering standard. See section 6.1:
“It is usual practice for instrument suppliers to express accuracy data as an upper limit based on test data inaccuracy values that none of their instruments will exceed (the confidence level is greater than 99.7%). It is conservative to assume that the vendor accuracy data unless otherwise specified represents a probability distribution of three sigma with a confidence level of 99.7%.”
So, my question is, what is the confidence level of the +/-0.5 and +/-2 accuracy specifications you have in your datasheet? i.e. “I’m 99.7% sure that the true temperature is within 2 deg C of the measurement, and 95.5% sure it’s within 0.5 deg C” would be the kind of statement I’m looking for.