I am currently working on a 4S1P battery pack that uses a BQ20Z65R1 to monitor the cells. What we have noticed is that after we run the packs through an environmental stress test some of the packs think that they have dropped in capacity. This is manifesting itself as a single LED, or a single flashing LED showing up on the fuel gauge, when in reality the pack has not been discharged at all.
We are currently in the process of rerunning the environmental test with batteries that are connected to the EV2300 so that we can log data during the test.
There is no discharge applied to the batteries while they are in the environmental test, and the voltage of the cells before and after the test are nearly the same, with at most a 5mV difference.
The environmental test method is as follows:
1. Drop to -20 degrees C from ambient in 30 minutes
2. Hold at -20 degrees C for 60 minutes
3. Increase to 70 degrees C from -20 degrees C in 60 minutes
4. Hold at 70 degrees C for 60 minutes
5. Drop to -20 degrees C from 70 degrees C in 60 minutes
6. Hold at -20 degrees C for 60 minutes
7. Increase to ~25 degrees C from -20 degrees C in 30 minutes
8. Hold at ~25 degrees C for 60 minutes
9. Repeat 1-8 for a total of 3 cycles
10. Hold at ambient until total test time reaches 24 hours
Not every battery demonstrates this problem, and what we have found is that if we send the reset command 0x41 that the battery pack SMBUS registers fall into a point that we would expect them to be. We also notice that after running a few cycles on one of the packs that it reset its state of charge to the correct indication and the capacity values make sense.