Hello Greg,
the ADC will measure the output of linear hall sensors (connected to INP) and substract their common mode voltage (resistor divider 5V / 2 = 2.5V, connected to INN). The output of the hall sensors can swing from 0V to 5V (nearly), depending on magnetic polarity and strongness of the field. At 0 Gauss, the output of the sensors is Vcc/2=2.5V (my common mode voltage above) and therefore INP-INN=0 (theoretical).
The linear hall sensors do not have an absolute output, but their output is ratiometric to the supply voltage. That's why I thought to use also the supply voltage as the Vref, to maintain the conversion independent ov the Vcc-variations.
On the other hand, since Vddana=5V and Vssana=GND, I wanted to have a reference voltage that is slightly larger than the swing from CM.
I planned for all this resistors to use 0.1% tolerance type, with low temperature coefs.
What do you think about my approach? Am I maybe wrong...?
Best regards
Gerardo