Adaptive Capacity Estimation
Original Publication Date: 2002-Nov-25
Included in the Prior Art Database: 2002-Nov-25
Related PeopleOther Related People:
In devices with battery supply (e.g. mobile phones) it is necessary to read out the battery capacity in the full lifetime of the battery, adapting to the actual battery capacity. So far a fuel gauge is used. A fuel gauge is a hardware component to be mounted to the battery in the battery-pack itself. It is equipped with a single wire interface in order to exchange data with the micro controller (in the mobile phone). Using a fuel gauge the tolerances in read out of remaining capacity could be less than 1%, but the use of fuel gauges in battery management is not a common solution because of the costs. Normally battery capacity estimation systems are based on measurement on the battery voltage and the voltage drop in the known states. This can be enhanced using assumptions on the current con-sumption when the device is in different power states. The total consumption is determined by adding the values for the enabled states by consulting look up tables. Subtracting the consumption over time from the assumed max. capacity of the battery gives a snapshot of the remaining battery capacity. This method is not reliable, the tolerances are typically in 20% range. The decrease in battery capacity due to ageing is not taken into consideration with this traditional method of capacity estimation. Using a fuel gauge, the battery capacity will be presented in counter steps, where each step equals a certain amount of charge. This counter step can be determined from the sensibility of the fuel gauge and the value of the sense resistor. Considering a battery with a nominal capacity of 650mA/h the counts for full charge will be 3250 (assuming a counter step of 0,2mA/h). The read out of the remain-ing capacity will then be scaled due to the factors 100% equals 3250 counts. But a constant factor will lead to errors in read out due to the fluctuations in battery capacity. The tolerances for new batteries could lead to an empty battery display before the battery is fully discharged (fig. 1). The ageing of the battery could turn off the mobile before the read out shows a zero capacity. The suggested adaptive capacity estimation is a method where the system adapts to the physical (real) capacity of the battery and scales a full charge to 100%. When the battery is fully discharged a flag is set in the mobile phone. Performing a full charge cycle will give a reliable read out of the physi-cal charge capability of the battery in counts. If the counter e.g. reaches 3500 counts the battery max. capacity can be determined as 3500 multiplied with 0,2mA/h equals 700mA/h. In order not to give fluctuating read out, a limit is made on the size of the change in max. battery capacity. Each full charge cycle should not be able to add or subtract more than a certain number of counts to the maxi-mum battery capacity value. In the example (shown in fig. 2) this value is 20mA/h. After a number of full charges – discharge cycles with the battery the system will be calibrated to the exact characteris-tics of the battery. From the system it will be possible to read out both the accurate value of the maxi-mum battery capacity and the remaining battery capacity. And the adaptive capacity estimation feature will secure that the zero and max. values are calibrated with the actual characteristics of the battery. The only disadvantage is that the system demands fully charge cycles.