We use a AD5258 digital pot to adjust the gain of an non-inverting amplifier (see schematic). In a small number of units in the field, which originally passed factory test and calibration, now have the pot at 0 Ohms. Sometimes power cycling the product will correct the problem, but sometimes it will not. Resending the pot setting to the unit does not fix the problem. Unfortunately, we haven not taken any particular care to manage the power up sequence, and it could be that pot leads A, B, W see some voltage before the supply voltages come up. Is the zero Ohms state an expected fault if the power is not sequenced properly? Is there an application note or sample circuits that suggests proper power up of these type of components?
Could you check if the digipot ack the commands when the faulty state is detected?
Could you provide a plot including the inverting, non-inverting and output voltages in the opamp?