• State TI Thinks Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 1 answer
  • Subscribers 63 subscribers
  • Views 744 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.

Original question:

BQ78350-R1: Unexpected results of Chemistry ID matching tool

BQ40Z50-R1: Unexpected results of Chemistry ID matching tool/failure of learning cycle

Hamish Henderson
Prodigy 80 points
Part Number: BQ40Z50-R1
Other Parts Discussed in Thread: GPCCHEM, BQSTUDIO,

I have completed a full ChemID Cycle (partial charge, relax, discharge, relax), following the instructions in http://www.ti.com/lit/an/slva725a/slva725a.pdf. The only deviations from the instructions were that the relax periods were significantly longer than 2hr post charge and 5hrs post discharge specified (battery left to relax overnight).

The log file from this that I submitted to GPCCHEM was:

/cfs-file/__key/communityserver-discussions-components-files/196/8546.chem_5F00_id_5F00_testing.zip

and the report received:

/cfs-file/__key/communityserver-discussions-components-files/196/0508.chem_5F00_id_5F00_testing_2D00_report.zip

As you can see in the report, the recommended ChemID is 1582. As shown in the screen cap from bqStudio below, this is listed as a 1920mAh model. I do not have the exact chemistry of the cell under test, but I do know it is ~5400mAh, which is significantly different. Does the difference in capacity matter for Chemistry ID? Should I disregard the result and try again, or trust the matching tool?!

I ask because when subsequently running a learning cycle the learning cycled failed during the discharge stage with RDIS and NSFM bits set. From /cfs-file/__key/communityserver-discussions-components-files/196/0020.7418.Achieving-The-Successful-Learning-Cycle.pdf, common failure number 6 states

Which I would not expect as I have just applied the ChemID matched by GPCCHEM. I have confirmed that the ChemID was applied correctly with SMBus command 0x6, and I reset the bq40z50-R1 prior to starting the learning cycle to ensure the update was 'applied'.

My questions are:

  1. Is the matched Chemistry ID implausible given the large difference in capacity between the cell under test and the model listed in the chemistry database?
  2. If so, can you see what in the submitted log file is causing the problem and how I might fix this?
  3. If not, what else can I try to successfully complete a learning cycle?
  4. How could I verify that I have an appropriate Chemistry ID, other than by running a learning cycle and not getting RDIS set?

Thanks,

Hamish

  • 0
    TI__Expert 6575 points
    Hi Hamish,
    The capacity indicated in the chemID selection screen is only for reference.
    The chemID match process does not look at absolute capacity, but rather tries to get the closest voltage curve match as well as tries to match the shape of the impedance curve to select the right chemistry.
    Once the chemID has been matched, it is important in the learning cycle process that you first get a Qmax update before attempting to learn the Ra Table.
    From your procedure, it appears that the Qmax update would not happen before the Ra table is updated, and this is the most likely reason that the [R_DIS] is getting set.
    The best way to rectify this would be as follows:
    1. Disable gauging
    2. Discharge to the cell specification minimum voltage for the pack.
    3. Let it rest for 5 hrs (Nothing connected)
    4. Enable gauging. "Update Status" is 04
    5. Charge at C/2 to the cell specification maximum voltage with a C/10 taper current condition. [R_DIS] will be set.
    6. Let it rest for 2 hrs (Nothing connected)
    7. Check that "Update Status" is now 05 and Qmax is updated. [R_DIS] will be cleared.
    8. Discharge at C/5 rate
    9. "Update Status" should now be 06 and Ra Table should be updated.

    Regards,
    Swami
  • 0 in reply to Swaminathan Ramanathan75031
    Prodigy 80 points

    Hi Swami,

    Thank you for your response. Good to know that the ChemID is not necessarily wrong, thank you for that explanation.

    Could you explain please what it is about the learning cycle process you think is wrong? I did get a Qmax update before the Ra table was updated. That part of the cycle ran fine. On the discharge stage, after relaxation, the Ra table was updated for a while; I could see GaugingStatus[RX] bit toggling, which according to the Tech Ref indicates a resistance update.

    However, at approx 75% SoC, RDIS and NSFM bits were set, meaning no further Ra table updates would be made. As quoted from the SLUA597: Achieving the Successful Learning Cycle, point 6 states that this behaviour "indicates that the Chemistry ID choice is incorrect", which is why I became suspicious of the ChemID.

    Are there any other reasons that could cause a negative scale factor to be detected and prevent Ra table updates?

    Thanks,

    Hamish

  • 0 in reply to Hamish Henderson
    Prodigy 80 points

    Hi Swami,

    Any update on this? I have completed another fresh learning cycle with nearly the same result; RDIS was set only at ~15% SoC, so nearly a fully complete discharge.

    Thanks,

    Hamish

  • 0 in reply to Hamish Henderson
    TI__Expert 6575 points
    Hi Hamish,
    Could you post the bqStudio log of your learning cycle as well as the gg file before and after the cycle?
    I can take a look and see why resistance updates are getting disabled.
    Typically this would only happen if the calculated resistance update come out negative.
    Regards,
    Swami
  • 0 in reply to Swaminathan Ramanathan75031
    Prodigy 80 points

    /cfs-file/__key/communityserver-discussions-components-files/196/learning_5F00_cyle_5F00_rdis.zip

    Attached is log file for the full learning cycle, showing RDIS set during discharge. I do not have a *.gg from before the learning cycle anymore, but have attached a *.gg from afterwards. Note that despite RDIS set (see logs) update status changed to 0x6 on the second run and the Ra table looked to have been updated sufficiently.

    The *.gg attached has the following known changes from the learning cycle run:

    • CAPM was set to 0, which disables the reserved capacity as it is set in cWh
    • TD Clear Voltage Threshold increased from 3000mV to 3050mV
    • Static DF signature changed