• State TI Thinks Resolved
  • Locked Locked
  • Replies 7 replies
  • Answers 2 answers
  • Subscribers 64 subscribers
  • Views 574 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.

BQ27520EVM: battery SOC suddenly drops to 0

user6300248
Prodigy 90 points
Other Parts Discussed in Thread: BQ27520EVM, GPCCHEM, BQSTUDIO

Hello technical support：

At present, we have successfully learned the battery charge curve on the  bq27520evm.  Were also strictly according to the learning cycle document and get the dffs with STATUS = 2. When we use the EVM to test the accuracy of the battery charge curve , we found that if we use a larger load (2A) , the  SOC will directly  jump value to 0  from a  large value .  The watch log is shown in the figure below. The log and ggfile of the whole process are also provided in the attachment, please help to check this problem. Thanks !

200423.rar

  • 0
    TI__Guru 52115 points

    We will take a look and get back to you.

    Andy

  • 0 in reply to Andy Liu (BMS)
    TI__Guru 52115 points

    I see in your gg file, the chemistry id 3359 is used. It is for a battery that operates from 4.35V to 3.0V.  Did you use the GPCCHEM tool to find out this chemistry id? Can I see your GPC report?

    Also, did you see the same issue when you did the learning cycle? Thanks.

    Andy

  • 0 in reply to Andy Liu (BMS)
    Prodigy 90 points

    Hi andy,

    Thank you for your review and reply！

    I have used the GPC tool to conduct two chemical ID evaluations. One of the original data was generated by the battery manufacturer's charge and discharge test , and the other one was generated  from  bqstudio . Both two tests are strictly followed the  GPC data collection document.

    Could you please help me to see if these two chemical IDs can be used , will it cause the above problems? Do I need to learn chemistry ID again ?

    Reports are in the attachment.

    data-report1.zipdata-report0.zip

    In addition, we use 1 / 7C discharge rate training (900ma) in the learning cycle, but the discharge current of the typical application is about 2A. Is this the main reason cause  above problems? Do we need to use  2A discharge current during the constant current discharge step during learning? But I worry that  may cause a fail learning due to the discharge rate recommended  in official document is 0.1C ~ 0.2C.

  • 0 in reply to user6300248
    TI__Mastermind 24870 points

    Yes, you can increase the rate so that we need to be above 0.1 C for learning cycle.

  • 0 in reply to user6300248
    Prodigy 90 points

    Hi,andy：

    I have tried to train the chemical ID again, but I found that based on our battery, the best matching chemical ID has a big error no matter how I train. In this case, the guage is inaccurately even if the learning cycle is successfully performed. We have  studied on the basis of the chemical ID of 7.3% error previously, and tested the problem of inaccurate battery measurement under the condition of large load discharge. The attachment is a chem id  log. Please help us to see how to match a chemical ID with a deviation of less than 2% . Thanks!

    Fullscreen 2625.config.txt Download 
    ProcessingType=2
NumCellSeries=1
ElapsedTimeColumn=0
VoltageColumn=1
CurrentColumn=2
TemperatureColumn=3

  • 0 in reply to user6300248
    TI__Guru 52115 points

    I have checked the two GPC reports.  One chemistry id #1785 has a 6.38% dedication.  It is not ideal but it should not cause the big SOC jump.

    I would suggest you follow the learning cycle procedure to check the SOC accuracy. The blog below shows how to calculate the battery gauge accuracy step by step.

    https://e2e.ti.com/blogs_/archives/b/fullycharged/archive/2016/11/04/how-accurate-is-your-battery-fuel-gauge-part-2-2

    See