We have been testing the ADuCM360 for a new data logger and getting very good results. Our application involves sampling thermocouples during a thermal process during which the logger and ADC can be subject to ambient heating. We have therefore been testing the gain drift while sampling a constant µVolt input with the ADC in an oven starting at 25°C before heating to 85°C and then cooling back to 25°C. The attached spreadsheet shows the results of one such run.
We are sampling at a 50Hz data rate with an effective gain of 16 and taking a single sample every 5 seconds. An internal zero offset calibration is performed before each sample but no gain calibration. The voltage reference is a high stability external 1.25V reference and the constant input is from a calibrating standard microvolt source outside the oven. The trace drifts up during the heating period and then down again while cooling, the cooling phase being much longer as it is in still air. This drift is expected but overall seems to be much less than the data sheet values which is good. Our question concerns the discontinuities in the trace where the reading seems to suddenly jump by something like 700 counts and then carries on from the new level with the rate of drift matching the drift before the jump. Later it seems to jump back down again. It looks a lot like some kind of register overflow wrap around problem.
If we do the same test but using the internal reference then we see exactly the same behaviour but the jumps are much more frequent.
Any suggestions as to what may be causing these jumps in the ADC output?