I have a design using LTC4380 for protection and LTC2944 for battery gas gauging. Schematic snipped attached. The +3.3V rail that I2C bus is pulled-up to (not shown in snippet) is powered up well after the LTC2944 VSENSE supply.
On some boards, maybe 5-10%, I have a strange issue where LTC2944 holds up I2C bus and processor can't query the device. If the board is restarted immediately, then it works again. But if the board is left to sit unpowered for more than ~15minutes, the LTC2944 again hangs the I2C bus.
Is there some mechanism by which it's known that LTC2944 can hold an I2C bus busy?
Is there an issue with the I2C pins not being pulled up to +3.3V for quite some time ? The board sits in a sleep state for much of the time, with +3.3V not provided.
I have doubled up the LTC4380 surgestopper and LTC2944 on the same sense resistor - I believe I had asked about this in the past and ADI advised it would be no issue.. but wondering if in fact this is a configuration that could lead to part damage?
If +SENSE / -SENSE voltage exceeds limit of the part (i.e. at inrush current start), would the damage caused prevent I2C bus from working sporadically?