We're using the AD7714 ADC in a circuit connected to a strain guage load cell. We normally run the ADC at 2.4576MHz with a 20Hz output rate, the buffer off and a gain of 128 (20mV unipolar input). The reference is about 2.6V (derived from the load cell excitation). We're now trying to run this at 50Hz output rate.
We've monitored the coefficients and the ADC reading produced (divided down to give a 16-bit value). When running at 20Hz the ADC value step-changes over 10 self-calibrations to give a maximum variation of about 8 counts in 65535. This is acceptable for our application.
When running at 50Hz, repeating the tests show a variation of about 50 counts and the step change is much more obvious when monitoring the ADC reading. This is not acceptable.
Is this performance to be expected at 50Hz output data rate? Or should we be getting better results than this? Is there something I can check or experiment with to try to improve the results?