The bq78PL114S12 and bq78PL116 do not need an initial learning cycle for every pack that comes off a production line. An off-line learning cycle is done on one pack at the completion of product development. This Off line Auxiliary Chemistry Test, taking upwards of 10 hours, is described in application note SLUA505b. The designer should identify one pack that is indicative of all the packs that will be produced – a production exemplar. It is important that the cells and electronics used in the exemplar pack are identical to that which will be used in production. Any deviation between the exemplar pack design and the production pack will show up as a reduction in accuracy for the initial cycles in the end product. Again, this exemplar is used during the Off line Auxiliary Chemistry Test. The results of the test are used by the bqWizard to generate an .aux file. The .aux file is loaded to the production exemplar. Then, the off-line learning is “taught” to each production pack via the golden flash file (.dat).
The Off-Line Auxiliar Chemistry file (.aux) provides the following information to the Gas Gauge:
|
Parameter |
Description |
|
Qmax: |
QMax is the chemical capacity of the cell. That is, it is the capacity obtained at a very low discharge rate with no compensation for the cell impedance and corresponding IR drop. The equivalent compensated value is the Full Charge Capacity, FCC. |
|
Qrem: |
Qrem is the value of the remaining chemical capacity of the cell or pack. That is, it is the capacity with no compensation for impedance. The equivalent compensated value is the Remaining Capacity. This is initially set to 50% of Qmax by the .aux file. |
|
Tau10: |
This is the time constant used for current taper during end of charge. It characterizes the shape of the exponential decay in current as the pack nears end of charge. |
|
Ra Values: |
These are the impedance values at each of the SOC grid points in the gas gauge algorithm. |
|
NDI Initial: |
This is the initial normalized dynamic impedance. |
Now if there is a great variation from pack to pack, then each properly qualified cycle will attempt to adjust parameters like Qmax – this is the Qmax learning described in the TRM (SLUU330b). Also, impedance values like Ra and NDI will be adjusted if qualified current step changes exist and they differ from the initial Ra and NDI information stored in the .aux file. If all of the packs are very similar, then only minute changes in gas gauge performance will be observed.