• State Resolved
  • Locked Locked
  • Replies 10 replies
  • Answers 5 answers
  • Subscribers 63 subscribers
  • Views 288 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

BQ28Z620: Battery Learning for Relative State of Charge

Sebastien Gravel
Prodigy 40 points
Part Number: BQ28Z620
Other Parts Discussed in Thread: GPCCHEM

Tool/software:

Hi,

We are using a BQ28Z620 BMS with LiPo batteries in 2s configuration and are having some issues having a valid SOC value.

When searching for the procedure to execute a battery learning cycle, we had trouble finding all the information, but we were able to find pieces of procedures in multiple documents and forum.

Here is the latest procedure we executed on a brand new BMS evaluation kit (BQ28Z620EVM-071).

  1. Update the base image of the BQ28Z620 firmware from version 1.00 to version 1.01 using BQ Studio.
  2. Configure the minimal settings for the BMS to be functional with the batteries (2s config, enable FET).
  3. Perform a discharge (until 6V), relax, charge (until 8.3V), relax, discharge (until 6V), relax cycle while logging using BQ Studio.
  4. Rename the log file into `roomtemp_rel_dis_rel.csv`, remove the file header, zip the file with a config.txt.
  5. Upload the zip file to your Battery Chemistry Online Tool and check the chemical ID in the report (default 1210 is good in our case)
  6. Change the chemistry using BQ Studio and configure the following settings indicated in the learning procedure.
    1. Design Capacity: 2100mAh (also set the 1554cWh)
    2. Charge Termination Taper Current: Default 250mA
    3. Discharge Current Threshold: 20mA
    4. Charge Current Threshold: 50mA
    5. Quit Current: 10mA
    6. Termination Voltage: 6V
  7. Activate the GAUGE_EN setting and perform a discharge (until 6V), relax, charge (until 8.3V), relax, discharge (until 6V), relax cycle following the timing from the learning procedure.
  8. Once completed, the golden image was exported using BQ Studio.

After that, the SOC was tested by charging the battery to 100% capacity and discharching at a steady rate of 430mA (~C/5). The SOC, voltage and current was manually logged to observe the discharge.

We would expect the SOC to be mostly linear with a discharge rate that should match with the orange line for 2100mAh. However, the relative SOC procides invalid value and indicates 0% charge about 45 minutes before the real empty state.

  20250203-SOC.xlsx

Our application requires to have a good estimate of when the battery will be depleted, so we need to address this issue.

I also attached the golden image output files in case it can provide you more information.

20250123-BQ28Z620-LiPol Golden Rev2.0.zip

Thank you,

Sebastien

  • 0
    TI__Mastermind 24931 points

    Hello, 

    Firstly, I recommend you refer to this FAQ to ensure you are calculating the SoC Accuracy correctly. Additionally, can you please provide me with the report you receive from the GPCCHEM tool? 

    Regards, 

    Jonny. 

  • 0 in reply to Jonny
    Prodigy 40 points

    Hi Jonny,

    Here is the report from the GPCCHEM tool and the measurement that was done to generate the file.

    Note, I had to re-generate the report using the tool and the chemID changed to 1304 in the report.

    Both 1210 and 1304 are LiCoCO2/carbon 11 type.

    20250123-LiPol-Rev1.zip

    20250123-LiPol-Rev1-report.zip

    I checked the document in the link that you provided (https://www.ti.com/lit/pdf/slua684).

    I'm happy to see that the procedure in section 2.2 is the one I did with a constant current load and that my analysis is valid.

    Let me know if you need other information.

    Regards,

    Sebastien

  • 0 in reply to Sebastien Gravel
    TI__Mastermind 24931 points

    Hello, 

    I programmed your .srec file onto my EVM, and it appears that you have not successfully completed the learning cycle, and update status is still 0x04. Please ensure that you have successfully completed the learning cycle before referring to the SoC accuracy. You can refer to this application note on Successfully Completing the Learning Cycle. 

    Regards, 

    Jonny. 

  • 0 in reply to Jonny
    Prodigy 40 points

    Hello,

    Thanks for the tip. We are re-running the learning cycle as we speak. I should be able to validate if it solved the issue by the end of the week.

    I'll keep you posted.

    Thanks,

    Sebastien

  • 0 in reply to Sebastien Gravel
    TI__Mastermind 24931 points

    Sounds good, 

    Please update this thread if you have any other questions. 

    Regards, 

    Jonny. 

  • 0 in reply to Jonny
    Prodigy 40 points

    Hi Jonny,

    We have completed the learning cycle all over again. Everything went as expected, except for the update status register which is still 0x04 after the last relaxation. According to the procedure, the update status should change to 0x06 when the VOK bit is cleared. However, it did not happen.The log and golden image are attached with this message.

     20250205-BQ28Z620-LiPol Rev2.1.zip

    We have been trying to run the learning cycle (which takes multiple days each time) for a while now and it still isn't working. Do you have any idea why the learning cycle is not completing or if we can force the learning cycle to finish?

    Thanks,

    Sebastien

  • 0 in reply to Sebastien Gravel
    TI__Mastermind 24931 points

    Hello, 

    Please refer to this E2E post. 

    Regards, 

    Jonny. 

  • 0 in reply to Jonny
    Prodigy 40 points

    Hi Jonny,

    I checked the post you provided. Looking at the log file, the DOD0_x registers were close to 90% change, but did not reach it most probably due to the batteries' internal protections when discharging. The undervoltage protection kicks in at 2.75V.

    In the previous cycles, our charging voltage was set to 8.3V. Would increasing the charging voltage to say 8.4V or 8.5V help to increase the DOD0_x change?

    Regards,

    Sebastien

  • 0 in reply to Sebastien Gravel
    TI__Mastermind 24931 points

    Hello, 

    I believe this would help, although I recommend making sure that you do not exceed the maximum spec of the cells. 

    Regards, 

    Jonny. 

  • 0 in reply to Jonny
    Prodigy 40 points

    Hi Jonny,

    Good news!

    We successfully completed the learning cycle by increasing a little bit the charging voltage.

    The Update Status went from 04 to 05 after charge and to 06 after discharge.

    For anyone curious, here are the 3 things we did differently this time.

    1. Just after the first discharge, we manually balanced the batteries too reduce the discharged voltage before the DOD measurement
    2. We reduced the termination current to allow the battery to charge more before disconnecting
    3. We increased charging voltage a little to increase the difference in DOD measurements

    We did run the same kind of test as before to measure the accuracy of the SOC and the results are much better.

    Thanks for your support,

    Sebastien