I'm using 2 development cards for initial testing of the system, bq20z45evm and Linear Technologies LTC4100 level 2 charger. The SMB is connected and works, as shown by the fact that the gas gauge and charger negotiate a charge current at the start of the charge cycle. The LTC 4100 card is a proven design, and it works well with the off-the-shelf batteries I have tried.
When I try to run the IT learning algorithm, it always fails in the same way - the first charge cycle ends with the FC bit set, but the update of Qmax is significantly lower than the 3100 mAh I entered at the start. Then, on the second charge cycle, the z45 allows an overcharge alarm to occur before the cells are really fully charged. I suspect that the voltage on the cells is fine, so the z45 does not shut down the charger, but the lowered Qmax causes what *looks like* an overcharge, and the algorithm blows up. Reasonable?
During the first part of the cycle Status Update reads 0x04. After the first Qmax update it goes to 0x0D, which is not allowed, according to the manual. What does it mean?
I am following the instructions in SLUA334b, "When no close Qmax value is known:"
I am setting the dataflash values for 2 cells as laid out in SLVA347. The cells I am using (NCR18650a) have a nominal 3100mAh capacity. I have tried several different sets of cells.
Thanks in advance
