I have an application requiring exceedingly stable temperatures ~0.001C would be great but I would settle for an order of magnitude would be acceptable. To measure temps, an ADT7310 had been chosen. I am a little unclear about the exact meaning of some of the terms in the datasheet. It seems as if there are three main numbers that characterize the "quality" of this device:
Resolution: 0.0625C or 0.0078C
My question is about the real physical meaning of each of these terms. Accuracy for instance, I suspect it means that any particular sensor will be within 0.5C +/- (of the real temperature of the room) at 24C. But (say if the true offset is +0.25C), will this offset be constant at other temperatures? So if my sensor is measuring 30.25C then I know that it is actually 30C. The actual number is not really important to my application, only that the offset be a constant one.
Resolution I know is the smallest temperature increment that the sensor can resolve. But if the sensor is only good at predicting the temperature within 0.5C then having 0.0625 or 0.0078C resolution would seem superfluous...
Repeatability. This likely means that exposed to the exact same temperature a few times, successive readings would only differ by 0.01C. But how does this affect my problem of small temperature fluctuations occurring constantly? Will the sensor be able to accurately be able to detect a rise in temp of 0.01C reliably but any smaller the result wouldn't be valid?
Any clarification on this matter would be greatly appreciated