This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

BQ27520-G4:BQ27520-G4: Learning cycle failed

Part Number: BQ27520-G4

Dear TI-expert

I'm evaluating BQ27520-G4 device.
Our battery is UR18650ZY.

There are 3 Chemistry ID values (1146、2026、2678) for Panasonic UR18650ZY cell. 
I have programmed ChemID=2678 into the gauge as shown below.

e2e.ti.com/.../bq27520-g4-chemistry-id-value-for-ur18650zy-cell

Attached is the result of running the learning cycle.
BQ27520-G4_Learning cycle failed_20230308_001.xlsx

During the first charge, VOK changed from 1 to 0 after FC = 1, but Qmax was not updated.
Qmax was not updated during the second charge as well.


I've checked the Initial Qmax Update Criteria for slua903, but I'm not sure what's causing it.

• Temperature — If Temperature is outside of the range 10°C to 40°C.
→no problem. Temp:20.9~26.5℃

• Delta Capacity — If the capacity change between suitable battery rest periods is less than 90% of the
full scale range of the OCV table during the initial QMax Update and 37% during field updates (2nd
and subsequent updates) of QMax.
→no problem. capacity change:2313mAh > 2160mAh(Deseign Capacity2400mAh × 90%)

• Voltage — If Voltage() is inside the flat voltage region. This flat region is different for each ChemID.
The GaugingStatus[OCVFR] flag indicates if the cell voltage is inside this flat region.
→no problem. Voltage:4090mV   > Vqdis_min=3753mV、Vqdis_max=3789mV(ChemID2678)

• Offset Error — If offset error accumulated during time passed from previous OCV reading exceeds
1% of Design Capacity, update is disqualified. Offset error current is calculated as Coulomb Counter
Deadband/sense resistor value.
→unconfirmed(Please tell me how to confirm.)


What is the problem that the learning cycle fails?
Could you give me advice for this?

Best Regards,
Kazunori Morita

BQ27520-G4_Learning cycle failed_20230308_001.xlsx

  • Hello Morita-san,

    The learning process requires a discharge current between C/5 and C/10. It seems that is not the case in the log. Can you check and confirm?

  • Hi Shirish,
    Thank you for your reply.

    From the battery discharge characteristics results, True FCC was 2306mAh.
    C/5=461.2mA

    Considering the margin so as not to exceed C/5, the discharge current is set to 450mA.

    Also, the result of the learning cycle is 5.1h rate as shown below, which is below C/5.
    discharge start[h]:24.69
    discharge end[h]:29.79
    discharge period[h]:5.1

    I think it's C/5.1, is my interpretation wrong?

    Best Regards,
    Kazunori Morita

  • Hello Morita-san,

    The initial FCC in the log is 2067mAh. That puts the discharge current slightly above C/5 and may cause it to be disqualified. I think discharging at around 350mA will help.

  • Hi Shirish,
    Thank you for your reply.

    So, Does that mean that the current value is determined using the FCC estimated by the gauge before learning, instead of determining the current value from the battery capacity?

    Also, C/5 and C/6 are as follows, but is 400mA not good?
    The recommendation for slua903 is listed as C/5.
    Is this a mistake and C/6 is recommended?
    2067mAh / 5 = 413.4mA
    2067mAh / 6 = 344.5mA

    The learning cycle is a very time-consuming task, so we want to keep it as short as possible.


    Best Regards,
    Kazunori Morita

  • The normal recommendation is to keep it between C/5 and C/10.

    I think that line in slua903 is a little misleading because in that same document in "conclusion" it says "Discharge the cell to empty using the typical discharge rate of your application. It must be between C/5 to C/10 rate, otherwise, the learning cycle fails. Resistance tables are updated during this discharge cycle."

  • Hi Shirish,

    I'll try it on C/6.
    Thanks for your support.

    Best Regards,
    Kazunori Morita