Other Parts Discussed in Thread: BQSTUDIO
We have a design incorporating the BQ27421-G1D Fuel Gauge in production. The design incorporates a ~300 mAhr integrated battery and we have a normal system shut-down voltage of 3.40V. Generally the system & fuel gauge behave as expected, but we have recently received several field returns where even fully charged devices will power down within a few minutes of operation.
For one of these returned units, I have monitored the actual battery voltage and confirmed that even though the fuel gauge is indicating 0% SOC, the battery voltage is well above the 3.40V threshold (as high as 3.8 - 3.9V).
Using bqStudio i have been able to extract the register values and data memory from the in-situ device and found that the register values for Full Charge Capacity, Remaining Capacity and related values are drastically lower than expected, even for a fully charged battery. This state seems to persist across several power-down / charge cycles.
As a test, I disconnected and then re-connected the battery (effectively performing a POR on the Fuel Gauge) and again extracted the fuel gauge registers and data memory. After this the registers look closer to those found in a normally operating unit and system shut-down is occurring closer to the design value of 3.40V.
My question is what would cause the fuel gauge to get into this state where the actual battery capacity is being grossly under-estimated; and apart from removing the battery, how may this condition be detected and corrected by the system?
I have attached an excel spreadsheet with the register values and data memory that were extracted through bqStudio, both for the original state and after the battery was removed and re-attached.
