• State Resolved
  • Locked Locked
  • Replies 30 replies
  • Answers 6 answers
  • Subscribers 65 subscribers
  • Views 1041 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.

BQ35100: SOH result is not correct on EVM

Katsu Matsunaga
Expert 7730 points
Part Number: BQ35100
Other Parts Discussed in Thread: BQSTUDIO, , GPCCHEM, GPCRB

Hello,

My customer is evaluating the SOH mode with LiMnO2 battery on EVM. But SOH seems incorrect as below.

- SOH decreases always 2% every GE toggling even though supply voltage was not changed.

-  Voltage and temperature looks correct value.

Test result and EVM setting are summarized here:/cfs-file/__key/communityserver-discussions-components-files/196/NEC-BQ35100-evaluation_5F00_20200806.xlsx

this is including capture image of register, data memory of BQStudio. 

Please check the result and give your advise to the customer?

Best regards,

Katsu

  • 0
    TI__Guru** 105760 points

    Hello Katsu,

    Could you export the .gg file and share it here?

    Is the exact chem ID chosen and uploaded to the gauge with the battery they're using? Matching the chemID is critical with the bq35100 in SOH mode because it directly compares the OCV data to the chem ID data.

    Also in the excel sheet, you have changed the supply voltage, are you connecting the gauge to the DC supply?

    The gauge is most likely going down in 2% increments because this is the default max change in SOH. To test the SOH functionality they will need to test with a real battery and make sure everything is configured for their application by following the EVM users guide: https://www.ti.com/lit/ug/sluubh7/sluubh7.pdf?ts=1596729586101&ref_url=https%253A%252F%252Fwww.ti.com%252Fproduct%252FBQ35100

    Sincerely,

    Wyatt Keller

  • 0 in reply to Wyatt Keller
    TI__Expert 7730 points

    Hi,

    Thank you for your reply.

    Yes, regulated DC power supply has been used for the evaluation, not battery.

    OK, I ask customer to use battery as actual system.

    I've received additional information from the customer.

    DC supply set 5.889V, SOH shows 92% after "NEW Battery". After that, continuing toggling GE makes reducing SOH 2% each until 38%. This seems not correct SOH at 5.889V. Is this caused by not battery using?

    one question,

    If supply voltage goes up, SOH value stay the previous value?

    what is .gg file? Is it dump data of memory?

    Best regards,

    Katsu

  • 0 in reply to Katsu Matsunaga
    TI__Guru** 105760 points

    Hello Katsu,

    The gauge should work with the DC supply in SOH mode because it is looking at OCV data, but to know why it is acting this way I will need to look at the .gg file. Yes you're correct, it should be under data memory > export on the top bar.

    The gauge is not meant to see an increase in voltage because it is a primary cell gauge, I believe it would stay at the same SOH.

    Sincerely,

    Wyatt Keller

  • 0 in reply to Wyatt Keller
    TI__Expert 7730 points

    Hi Wyatt-san,

    Thank you for your response.

    I'm asking to get the .gg file to my customer.

    Is it possible you to provide the .gg file which is correct functionality?

    Because the customer is out of office in this week, it will delay to send it to you.

    Best regards,

    Katsu

  • 0 in reply to Katsu Matsunaga
    TI__Guru** 105760 points

    Hello Katsu-san,

    Currently I don't have access to an EVM to pull the .gg data to share.

    Each time the GE pin is toggled, SOH will move the max delta until it reaches the OCV reading saved in flash from the chemistry ID that was programmed. So when the gauge settles at 38% SOH that value corresponds to the voltage saved in the chem ID. If you are setting voltage with the supply values similar to your fully charged battery then this could be a mismatched chem ID.

    It would be best if the test it with the batteries they plan to use in their application and see if it operates correctly instead of using the DC supply. If SOH still jumps with the fully charged batteries inserted then we can continue the debug.

    Sincerely,

    Wyatt Keller

  • 0 in reply to Wyatt Keller
    TI__Expert 7730 points

    Hi Wyatt-san,

    My customer has shared .gg file with us. attached here/cfs-file/__key/communityserver-discussions-components-files/196/BQ35100_5F00_export_5F00_202000817.gg.csv

    Please check this and give your comment.

    Best regards,

    Katsu

  • 0 in reply to Katsu Matsunaga
    TI__Mastermind 24870 points

    Hello Katsu-san,

    Does the customer also have a log of where the SOH is jumping?

    Thanks!

  • 0 in reply to Kang Kang
    TI__Expert 7730 points

    Hi,

    I have received a result of experiment with actual battery as CR123A.

    There is some gap with battery discharge characteristics, please check attached spreadsheet and give your comment Do they need some modify for setting?

    /cfs-file/__key/communityserver-discussions-components-files/196/BQ35100-experiment-0821.xlsx 

    in addition, they have question about device configuration.

    Question of gas gage configuration

    Capacity of CR123A is 1550 mAh and will use two in series.

     

    About a cell design capacity mAh, is it OK to set 3100 mAh?

    If four batteries in series and 2 para, cell design capacity is 6200 mAh?

     

    About Cell Design Voltage,

    Max voltage in spec is 3V, but actual is 3.2V, which is value input to?

    Or in case of two in series, is this 6V or 6.4V?

     

    About Terminate Voltage, is this stop voltage of two in series?

     

    About internal resistance of battery, is this needed to set anywhere?

    If yes, please let me know the register name and setting value.

     

    Best regards,

    Katsu

  • 0 in reply to Katsu Matsunaga
    TI__Guru** 105760 points

    Hello Katsu-san,

    They are comparing the voltage readings and SOH values from the manufacturer and the gauge? Is the gauge inaccurate with a real battery pack when tested? The load and cell resistance will change the SOH reading from the gauge.

    For the BQ35100 it is important to have an exact match with the chem ID for SOH mode. The design capacity doesn't change for the cells in series, it is only increased by cells in parallel. For a 2S2P pack the design capacity should be the mAh rating for one cell times 2.

    Design voltage should be set to the nominal voltage of the cells.

    The terminate voltage is based on the voltage of a single cell and is calculated from the stack voltage divided by the number of cells in series you have programmed to the gauge.

    The internal resistance of the battery is calculated by the gauge over the life of the battery. During the first pulse the gauge sets a scale factor for the next resistance values calculated during discharge, this is why a chemID match is important.

    Sincerely,

    Wyatt Keller 

  • 0 in reply to Wyatt Keller
    TI__Expert 7730 points

    Wyatt-san,

    Thank you for your support.

    About first question as 0.1V gap between battery characteristic and gauge value, it is being asked to customer.

    ,

    I have received other question.

    Q2. How does the temperature used for SOH correction?

    Q3. Can I have discharge carb of Chemi ID = Pana CR123A?

    Best regards, Katsu

  • 0 in reply to Katsu Matsunaga
    TI__Expert 7730 points

    Hi Wyatt-san,

    Regarding 0.1V gap, I have received customers experiment result.

      In case of no load condition, the gap between actual voltage and gauge's voltage is 0.001~0.002V. This means it is good correlation at no load.  

      What is the reason of the 0.1V difference when it has load?

    Additional question

    Q4. In low temperature test in -20c, gauge measured temperature and initial voltage correctly. After NEW BATTERY command, GE is repeatedly toggled, SOH got 0 even though actual voltage was 6V. In this situation, temperature calibration didn't work with below error

    "Calibration - Calibration temperature value entered is out of range. This entered value is in degrees centigrade. please check value and retry."

    What is problem?

    Could you please respond to Q2 and Q3 too?

    Best regards,

    Katsu

  • 0 in reply to Katsu Matsunaga
    TI__Expert 7730 points

    Hi Wyatt-san,

    Regarding 0.1V gap between actual voltage and gauge voltage, my customer did experiment.

    There is no difference the voltages when no load condition. the 0.1V difference was existed when connected to load. Do you have any idea?

    Additional question

    Q4. In the low temperature test as -20degC, the gauge got correct temperature and initial voltage. But they observed strange phenomenon. After NEW BATTERY command, GE was repeatedly toggled, then SOH got 0 even though actual voltage was 6V. and in this situation, temperature calculation didn't work with below error.

    "  Calibration - Calibration temperature value entered is out of range. This entered value is in degrees centigrade. please check value and retry."

    Why does the SOH get 0 even though actual voltage is 6V? 

    Could you please respond to Q2 and Q3 too?

    best regards,

    Katsu

  • 0 in reply to Katsu Matsunaga
    TI__Guru 57697 points

    When the battery is fully charged, the temperature has no important effect on the SOH. But when the battery is discharging, the effects of the temperature become bigger and make it harder to estimate the remaining battery. TI uses a proprietary algorithm to ensure that SOH is reported accurately.

    The discharge curve of a battery should be provided in its datasheet. You can use the GPCCHEM tool  to check if the actual battery curve matches the selected chemistry ID.

  • 0 in reply to Shirish
    TI__Expert 7730 points

    Hi Shirish-san,

    Please let me confirm your comments and add new question of Q5.

    Q1: 0.1V difference between batteries discharge carve and measured voltage

    You have suggested using GPCCHEM. This is used for a problem of 0.1V difference between battery discharge carve and measured voltage?

    Customer has set chemiID of Pana CR123A. Does the ChemiID’s information need to be modified using the GPCCHEM in case of my customer?

     

    Measurement result and .gg file have been shared with you. Do these information shows you some strange things?

     

    Q2: SOH correction by temperature

    We cannot provide the relation between temperature and SOH as equation?

     

    Q4: SOH reached to 0% even though VBAT=6V at -20degC.

    In your comment, TI algorithm can calculate correctly. But customer has observed the strange phenomenon. Could you please advise how to solve the problem?

     

    Q5: When customer uses other battery which doesn’t have ChemiID.

    Could you tell me a procedure to use the NON ChemiID battery?

    Does TI prepare a ChemiID for the NON ChemiID battery?

    Now I'm confuse what your comment was replayed for.

    Best regards, Katsu

  • 0 in reply to Katsu Matsunaga
    TI__Guru 57697 points

    Hello,

    For operation below 0degC, there are additional steps needed for accuracy. Since this is a primary cell battery, there may be variations to the procedure, so i have asked for help from our chemistry expert.

    Here is the top level link to chemistry optimization.

    Please refer the "GPC for Impedance Track (IT) algorithm gas gauges" section for more information.

    If there is a new battery model that is not listed in TI's chemistry database, then there is a procedure to send it to TI for characterization. Please provide us with the details.

    Regards

  • 0 in reply to Shirish
    TI__Expert 7730 points

    Hi,

    Could you please put Q# to your comments?

    I'd like to ask you to provide a answer as soon as possible for Q1 and Q2. Because evaluation result on EVM were not expected.

    For Q1:

    I would like to share my customer concern about Q1 with you again.

    There is 0.1V difference between Gauge's measurement and discharge profile of the battery. For example, When voltage was 2.9V, the gauge says about 30% but discharge profile in datasheet of the battery says about 50%.

    Could you please provide a discharge profile of ChemiID (Pana CR123A) of BQ35100?  I want to compare it to battery's profile (Pana CR123A)

    Does customer need to make procedure of GPCCHEM?

    If yes, how long does it take after customer send data to TI?

    For Q2: correction by temperature

    Please check below figure it shows a relationship between voltage and SOH of the gauge for three temperature as 9c, 25c and 35c.

    It seems storage that 9c line was cross to other higher temperature under SOH 30%.

    Could you please provide how it is calculated with the formula?

    Could you tell us how to solve this?

    I would like to fix these Q1 and Q2 soon, please make response sa soon as possible.

    best regards,

    Katsu

  • 0 in reply to Katsu Matsunaga
    TI__Guru 57697 points

    Hi,

    Thank you for the graphs.

    I have looped in the chemistry and algorithms team for help. 

    For Q1 please confirm that voltage was calibrated on the gauge.

    Regards

  • 0 in reply to Shirish
    TI__Expert 7730 points

    Hi,

    Yes, voltage calibration was done correctly.

    And customer confirmed both input voltage and gauge voltage.

    In this case, voltage calibration does not effect to the difference between gauge SOH value and battery's discharge profile. I think that SOH should be on the battery discharge profile with the gauge voltage, right?

    It seems that the discharge profile of the ChemiID(Pana CR123A) is no matched with actual discharge profile.

    Thanks and best regards,

    Katsu

  • 0 in reply to Katsu Matsunaga
    TI__Guru 57697 points

    Here is the response from our algorithm team

    How did the customer get the chemID for this primary cell? Was it obtained through cell characterization or was it by a chemid match. If it is chemid match, I would not recommend it since we don’t have a lot of primary cell chemids in our database and hence we may not get a good match.

     Q1. Also from this discharge profile that I found on e2e, it seems they are comparing OCV to voltage under load, so obviously it will be off. Even though the load 20mA is a relatively light load, the Ra values are pretty large as you can see below and hence the IR drop will be large.

     

  • 0 in reply to Katsu Matsunaga
    TI__Guru 57697 points

    Hi

    Can you confirm that bq35100 has been programmed with Chem_ID 0616?

    Thanks

  • 0 in reply to Shirish
    TI__Expert 7730 points

    HI,

    Thank you for your support.

    Yes, customer programed with Chemi_ID 0616.

    The figure of Q1 was measured without load. So it is matched to OCV?

    You mean that we need other chemi-ID for the Pana CR123A(0616)?

    Best regards, Katsu

  • 0 in reply to Katsu Matsunaga
    TI__Guru 57697 points

    Hi

    Figure of Q1 has bq35100 at no load but Panasonic datasheet at 20mA load. When load is lighter, the voltage is expected to be higher for same SOC. We expect the curves to match at load of 20mA.

    Regards,

  • 0 in reply to Shirish
    TI__Guru 57697 points

    Hi

    I received additional information

    The OCV correlation operates only under completely relaxed conditions, long after any load has been applied. This gauge cannot be used if continuous current is applied - a wait is required after current has stopped until cell is fully rested (typically more then 5 hrs) before SOH.gets updated.

    Please verify the use case. The curves that were measured at constant load are not applicable to the use case that bq35100 supports.

    Regards

  • 0 in reply to Shirish
    TI__Expert 7730 points

    Hi,

    Let me confirm my understanding.

    Now customer is comparing SOH result of the gauge and battery's discharge curve at 20mA. this is not correct, right?

    How does customer verify the BQ35100 work correct?

    Could you please advise me how to verify the gauge for the battery?

    best regards,

    Katsu

     

  • 0 in reply to Katsu Matsunaga
    TI__Expert 7730 points

    Hi,

    Customer is checking the discharge curve what you mentioned.

    Customer is trying to make gauging parameter of GPCCHEM using primary battery of pana CR123A.

    https://www.ti.com/tool/GPCRB

    The customer generates data log file in room temperature and cold temperature. then submitted them with config and gg files. But they received error email as below, it seems gg file has problem.

    customer has below question

    1. Is it available to use the GPCRB for primary battery with BQ35100? because the system has only discharge, so teat procedure doesn't meet the usage condition.

    2. Why does the error occur? 

       Customer doesn't use BQstudio for getting data log parameter. in this case, .gg file doesn't matter to generate the gauge parameter. What kind of .gg file does customer upload to GPCRB?

    error email is here.

    Best regards,

    Katsu 

  • 0 in reply to Katsu Matsunaga
    TI__Guru 57697 points

    Hi

    GPCR  tools requires GG file. The GG file must be of the correct format. It is not necessary to use bqStudio as long as the format is correct for all files.

    Regards

  • 0 in reply to Shirish
    TI__Expert 7730 points

    Hi Shirish-san,

    customer use gg file here

    /cfs-file/__key/communityserver-discussions-components-files/196/7343.BQ35100_5F00_export_5F00_202000817.gg.csv

    Could you please check the gg file?

    export on BQstudio is only way to generate .gg file, right?

    Thanks and best regards,

    Katsu

  • 0 in reply to Katsu Matsunaga
    TI__Guru 57697 points

    Hi

    The format looks ok. If any parameters values are configured incorrectly, then the GPCRB tool will provide an error message.

    Regards

  • 0 in reply to Shirish
    TI__Expert 7730 points

    Hi,

    Thank you for your support.

    I still don't understand why customer has the error of GPCRB. I would like to know what a reason of the error?

    Do the configuration file or data log file have something problem?

    Does someone use GPCRB with BQ31500?

    Best regards,

    Katsu

  • 0 in reply to Katsu Matsunaga
    TI__Guru 57697 points

    Hi,

    This issue is now a very specific debug activity. It is best to take it offline and connect you directly to the chemistry experts.

    Regards,