Temperature accuracy and piece-wise linear terms

[Jake.Boyd]

The temp accuracy is stated at +/-5C.  Does that include using piece-wise linear terms for accuracy?  Is there any data on accuracy over a narrow temp range such as 0-40C?  

Thanks

Jake

[Kenneth]

If you are mainly interested in accuracy from 0-40C, then by using a single point calibration in your production, in other words the device do a temperature measurement in production, and receive the actual temperature from the tester or production programmer, and then use the offset between the two measurement for all future measurements should give a much better accuracy than +/-5C. Likely in the <+/-1.5C range. Typically it is at the higher temperatures (>40C) that you will see the most deviation, though it seems to consistency measure too high at the higher level, so if you subtract with 1-2C when passing 40C you should improve accuracy.