I've chosen the AD7799 to measure the output of a thermocouple. I'm using bipolar mode so that the thermocouple works either way it is connected. I'm using a 2.5V voltage reference with 2ppm / deg C drift with temperature.
After reading the data sheet several times, I get the impression that 0V on the inputs will correspond to half-full-scale on the internal ADC (with full-scale being 2.5 volts). If so, should I expect the converted value (with 0V on the inputs) to drift 1.25V*2ppm / deg C?
On the other hand, if that is true then the drift errors on negative voltage measurements will be superior to that with positive voltage measurements. Well, at least it seems that way.
I'm not asking about calibrations or anything like that. I'm not worried what gain I'm setting on the inamp - I'm just trying to understand the errors due to drift of the external reference.