Predicting the remaining capacity of lithium-ion batteries can be difficult given the many factors influencing gas-gauge ICs; cold temperatures are one of those factors. There are several gas-gauging ICs on the market that have several features to provide accurate performance for operation at cold temperatures that we will highlight in this blog. In this blog, I’ll discuss some of the parameters to accurate gas gauging are and how to tweak parameters for the best performance. Applications engineers, battery pack engineers and even systems level engineers can often do these tweaks prior to mass production.


A customized battery model

To get the best performance from gas gauges, you will need to use flash-based gas gauges where you can fully customize the battery model.. For cold-temperature performance, the first step is always having a customized battery model. This process takes a month and is only available on gas gauges with configurable data flash.


Thermal model

Almost all gas gauges have a thermal model feature to help predict capacity more accurately. Like a battery model, a thermal model accounts for self-heating of the battery during charge or discharge. In short, the thermal model uses the impedance information and discharge current to get temperature data from a power calculation. This feature is configured using data-flash parameters, which are calculated by performing a discharge at low temperature with the battery in a thermally isolated box. By calculating the self-heating temperature, we can predict the capacity of the cell much more accurately, because temperature increases affect battery impedance and will lead to capacity increases during discharge.

To calculate a thermal model, the Gas-Gauging Applications team will ask for a bqStudio log containing a charge to full-relaxation-discharge-relaxation cycle. When you are collecting this data, the discharge rate needs to be C/5 to C/3 and the ambient temperature will have to be tightly controlled in a thermal chamber. The battery should be connected to the evaluation module (EVM) and placed in a thermally enclosed box to simulate the end-system heating conditions.


Transient model

Lithium-ion batteries are often modeled as a DC source with a constant resistance. This is true if the cell has been discharging for a long-enough time to saturate the impedance. However, it takes some time in order for the resistance to fully saturate.

Many Texas Instruments gas gauges like the bq27542-G1, use transient modeling to more accurately improve cold- and room-temperature discharging accuracy. The easiest example is when a battery is discharged. If you let it rest for a while, it can still provide a few minutes of energy. This is because the cell impedance is not fully activated after some time is passed when we apply the discharge current. At cold temperatures, the cell impedance is large. When you apply a current, it takes time for the cell impedance to saturate to the final steady state impedance; this means you will have some remaining capacity until the IR drop forces the OCV to the system shutdown voltage. In order to estimate remaining capacity more accurately at colder temperatures, this is another parameter we have to tweak.

Thermal models, transient models and customized battery models help gas gauges make much more accurate state-of-charge predictions at different temperatures. Check the data-flash settings of the gas gauge to ensure that these parameters are set up properly when testing its discharge performance. These are the most common and essential parameters to check when you run into low-temperature performance issues such as SOC jumping to zero too soon or erratic SOC behavior.


Additional resources: