STC3115 wrong SOC estimation

[matteoriga]

Good morning,

I'm currently using the STC3115 in my application for estimating the SOC of a 1000mAh LiPo battery.

The application drain a quite costant current, between 20 and 25mA, so the coulomb counter should be an easy approach for monitoring the SOC. For that reason the STC3115 has been configured with a relax current lower than 10mA.

Unfortunately the SOC estimation apparently is not reliable: during a long period discharge phase, the SOC decreases faster than expected and, moreover, below 50% it increases the descending slope. for comparison, I used the current value provided by the same component for implementing a pure coulomb counter and the behaviour is completely different, despite the original data (i.e. the current) is the same. The figure represents the two SOC estimations (STC3115 and pure coulomb counter).

Because the internal STC3115 SOC estimation algorithm is not described, I kindly ask to some expert from ST if the observed behavior make sense and/or is due to a wrong STC3115 configuration.

By the way, I calculated the battery internal resistance with the suggested approach in AN4324 (result around 1ohm), but I tried values ranging from 100mhom to 2.5ohm, without any impact on the above mentioned behavior.

Thank you in advance and best regards

[Daniel3]

Hi,

I was in vacations and I think that nobody received this mail otherwise you will have a first answer.

The behavior described is a bit strange to me. If the current is low (around 20mA), normally you must still use the Voltage Mode with the internal OCV curve and the calculated resistance of 1ohm.

Just to be sure, you are using a LiPo battery, classical float voltage at 4.2V max. so the STC3115AIJT (for WLCSP) or the STC3115AIQT (for DFN).

I will have a more deep look at the results because It is not normal behavior, the CC has an error but not so high.

BR

Daniel 

[matteoriga]

Hi Daniel, thank you for your reply.

I confirm we are using a normal LiPo battery, so the STC3115AIQT is the version we adopted.

It sounds strange to me to use the Voltage Mode with a current around 20mA because if I use the STC3115 provided current value for implementing the coulomb counter, results are quite perfect (see my previous chart); it means the coulomb counter properly works with such a current range.

Thank you and best regards

Matteo