Should the VPTAT temperature accuracy of the TMP01 be included to determine the temperature tolerance at which over and under outputs switch. Is this accuracy specification included as part of the output Vref specification?
Yes, the set-point error is the total of all the errors in the system:
For example, with a set-point of 25 degrees C and no hysteresis, your divider output would be 1.49V -- with a 2.5 volt reference, the needed divider ratio is 1.49 / 2.5 = 0.6
Your worst-case error using the E-grade part might look like this:
VPTAT tolerance 1.5 deg
V ref tolerance 5mV x 0.6 = 3mV = 0.6 deg
Resistor tolerance 1% = 14.9mV = 3.0 deg
Offset variation 0.2mV = 0.4 deg
Total +/- 5.5 degrees
This is a worst-case absolute error, which means that in practice you can probably build hundreds of these units and find few if any this bad. Although ADI guarantees the worst-case errors as shown, their "typical" error specs are about three times better. Also, the individual errors may subtract (cancel) as easily as add.
The largest contribution to the total error is the resistor divider. You can improve this by using 0.1% resistors or by calibrating out the error in your software.
Retrieving data ...