I'm using an AD7793 for resistance measurements in a simple ratio-metric configuration (see schematic) with one excitation current and a reference resistor. I try to measure a 0.1% tolerance resistor for testing purposes and I use a 0.01% 5k reference resistor.
I noticed that I get slightly different readings (+/-10 Ohm) for the resistor if I connect the circuit to different AD7793s and I figured that calibration might help me solve that problem and lead to more equal readings across devices.
However, when I carry out internal zero scale and full scale calibrations, things get much worse. While without calibration I might just be off by a few Ohm, after every calibration my readings seem to get distorted completely. I get widely differing results for the 2.2k resistance, ranging anywhere from 50-5000 Ohm. This is the case even if I calibrate again right after a calibration.
Can you help me figure out what could be the problem here? It seems like something is interfering with the calibration but I can't figure out what.
Thanks a lot!