BQ27Z561-R1: Unexpected SOC Reading Drop

Hi Team,

My customer is seeing some devices in the field using the bq27z561-R1 fuel gauge reporting large unexpected drops in SOC. Can you give us any advice on this issue? Below is a more detailed description of the problem.

Their devices read the voltage, average current, remaining/full capacity, RSOC, and remaining time to empty (tte) registers and log them every 10 minutes. One example of this issue is a device that reported a drop from 78% SOC to 0% on the next reading. The voltage reading and time frame indicate there should be a much higher SOC. 

Here is an excerpt of the log showing the drop in SOC:

4059mV, -1mA avg, 225/280mAh, 81% SOC, 13500 tte
3905mV, -75mA avg, 219/280mAh, 79% SOC, 175 tte
4016mV, -7mA avg, 217/280mAh, 78% SOC, 1860 tte
4017mV, -3mA avg, 0/280mAh, 0% SOC, 0 tte   <---------- drop occurs here
4012mV, -3mA avg, 0/280mAh, 0% SOC, 0 tte
4011mV, -2mA avg, 0/280mAh, 0% SOC, 0 tte

Their fuel gauges are all programmed with a golden image generated with settings obtained from the calibration and learning cycle procedures and chemID provided by TI.

Couple settings that might be helpful are:

Term Voltage: set to 3000 mV during the learning cycle, set to 3400 mV in the golden image
Design Voltage: 3800 mV
Chem ID: 2817

They have also seen devices report an SOC drop from 100% to 0%, along with other large drops that shouldn't be happening. Do you have any insight into why this may be happening?

Thanks,
Mitchell