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.

BQ34Z100-G1: Last product SoC Problem

Part Number: BQ34Z100-G1
Other Parts Discussed in Thread: GPCRA0

Hello;

My cell is Kokam SLPB080085270 (6S1P). Battery features;

- Max Voltage = 4.2V

- Min Voltage = 2.7V

- Capacity = 27000 mAh

The learning Cycle is completed. But there are problems in the last product. After taking the golden image, IT_Enable is being done. 

Then the BMS is starting to be used. The SoC is resetting when the battery is about 17.5-18V.  The log that is received about the subject is below.  What is the problem in the last prouct? I think the problem is that the average current is not updated quickly during sudden current changes. 

BMS_LOG.xlsx

  • Hi Faruk,

    Reviewing the log, the voltage can be seen to be sagging down to 16 V or 2.6 V per cell. The dip may be even more pronounced with a highly pulsed load. Please ensure the device is supplied with at least 2.8 V to the regin pin and can measure at least your Flash Update OK cell Voltage voltage on the BAT pin. If voltage on the BAT pin drops below the Flash Update OK cell Voltage, writes to dataflash will be disabled to prevent the possibility of flash corruption.

    If this does not resolve the issue, please attach your gg.csv file and SREC for analysis.

    Sincerely,
    Bryan Kahler
  • Hello Bryan. Thanks for reply.

    There is no such problem in constant current. I have this problem when discharging in variable currents. For example; Everything is normal when discharging at constant 40A. But when I shoot the current from 40A to 30A to 50A then there are problems. The logs are in both below. We also measured the voltage on the BAT pin. Flash Update OK is not lower than Cell voltage. Associated Data memory.

     NOTE: Current calibration 1/4 ratio.

    Logs and DataMemory.rar

  • hi Bryan,
    we are still waiting for your valuable response. We have to find a solution as soon as possible since we are in a hurry.

    Kind regards
  • Hi Faruk,

    Thank you for the gg file. After further review, please make sure to scale all values in the GG.CSV file that include A/Ah or W/Wh by the current calibration scale factor of 4.

    Sincerely,
    Bryan Kahler
  • Hi Bryan;
    All the data is scale factor correctly. Our biggest problem is SoC jump in variable currents. Our system is operation at high current. There is no problem in constant currents but there is a problem in variable current. We have such a problem last product. We also tired the Load Select modes. Everything is great but we can not prevent the jump of SoC value. (Suddenly resetting in variable currents).
    If the remaining Capacity value can not be read, it will be a jump (SoC 0%). We think that the current value of the mind has been updated in about 1 second. In addition, the Avarage Current value is updated at approximately 85 seconds. Times are too high.
  • Hi Faruk,

    As the SOC drops, the internal temperature increases. SOC1 and SOCF are setting simultaneously on the variable load. Please use the GPCRA0 tool to help improve high rate SOC accuracy. Please make sure to save your current SREC from the gauge prior to following the steps in the user guide for this tool. Also, with a LFP cell, allow the cell to rest for 5 hours after charge, instead of the 2 recommended in the guide (for Li-Ion).

    A fresh, unlearned SREC will need to be programmed to the gauge, followed by your chemID with the chemistry tab. Then export the gg.csv file.

    The steps for the using the tool are discussed in detail in the user manual, link given below.

    GPCRA0 tool page:
    http://www.ti.com/tool/gpcra0

    GPCRA0 tool user manual:
    www.ti.com/.../sluubc9

    Sincerely,
    Bryan Kahler
  • Hi Bryan;
    I am trying to say that. I cary out the Charge Current at 1C, which is allowed by the battery manufacterer. I use the C/3 Current that is mentioned in user manual for the discharge Current. What is the difference between this test and the siple guide to chemical ID selection tool(GPC) test? In this case;

    Charge Current = 27A
    Disharge Current = 10A

    Charge Time = 1 hours
    Relax Time = 2 hours
    Discharge Time = 3.2hours
    Relax Time = 5 hours

    Total test time = 11.2 hours

  • Hi Faruk,

    A lithium polymer cell has a large disqualified voltage region. To ensure you have OCV points that will qualify for a qmax update, please fully charge the battery until FC sets. Allow the device to fully rest (up to 5 hours after FC sets). Then, perform a discharge at your typical system high rate until terminate followed by 5 hours of rest or until relaxation.

    In this test, you perform a log at your high rate discharge instead of C/10 as recommended for the chemID.

    Sincerely,
    Bryan Kahler

  • Hi Bryan;
    I fix the mistake at some point. FCC and Soc are very stable during discharge. When the discharge is over the FCC is updated as needed. But about 1 hour after entering relax mode the FCC falling. True RC goes down negative numbers.

    For Example;
    Discharge is completed --> FCC = 6750mAh ( 6750*Scale Factor = 27000mAh)
    True RC = 0
    True FCC = 6750mAh

    1 hour after discharge --> FCC = 6315mAh
    True RC = -435mAh
    True FCC = 6315mAh

    When I recharge it, the FCC is update as needed. What should I do to prevent this fall?

  • Hi Faruk,

    Please try the steps offered in this thread with a similar issue: e2e.ti.com/.../630767

    Sincerely,
    Bryan Kahler
  • Hello Bryan,

    My problem continues. Pack Configuration I changed the C register values. FCC is being updated while discharging. But in the relax mode the FCC is running out after 1 hour. I share the logs with you.  I'm waiting for your help.

    Sincerely.

    Faruk Sözüer

    033_HexCode.xlsx

  • Hello Bryan,

    My problem continues.  FCC and True Fcc are being updated during discharge. (Increasing). But in relax mode the FCC is diminishing after 1 hour. Package Configuration I changed the C register values. The Discharge log is below.

    0x33 is Pack Configuration C register value; 

    8863.033_HexCode.xlsx

  • Hello Bryan,

    The log I get when the battery is resting is below. I have pointed out the decline of the FCC and True FCC.

    033_RelaxMode.xlsx

  • Hi Faruk,

    Thank you for sharing the logs. Please send the gg.csv file as well so I may provide an update on Tuesday.

    Sincerely,
    Bryan Kahler
  • Hello Bryan,

    Thank you for reply. I tried many things. But I could not solve the problem. I read all the topics on the forum one by one. But I could not find the solution. a very strange problem. The file gg.csv is below. a very important project for me. I'm waiting for your help. 

    Sincerely

    Faruk Sözüer

    033HexCode.gg.zip

    The following file contains all of the Charge-Discharge-Relax logs. 

    deneme.zip

  • Hi Faruk,

    Thank you for also including the log. There are quite a few settings in the gauge that still require scaling. The design capacity and design energy are scaled by 4, however, many other mA and mW parameters when scaling by the factor of 4 have been multiplied by 4 instead of divided by 4.

    For example:

    "Gas Gauging","Current Thresholds","Dsg Current Threshold","187","mAmp" (default: 60)
    "Gas Gauging","Current Thresholds","Chg Current Threshold","175","mAmp" (default: 75)
    "Gas Gauging","Current Thresholds","Quit Current","150","mAmp" (default: 40)

    Please divide the default value by 4 and round up to the nearest whole number.

    These values should instead be:

    "Gas Gauging","Current Thresholds","Dsg Current Threshold","15","mAmp"
    "Gas Gauging","Current Thresholds","Chg Current Threshold","19","mAmp"
    "Gas Gauging","Current Thresholds","Quit Current","10","mAmp"

    Please apply this to all mA and mW/cW values in the device.

    Sincerely,
    Bryan Kahler
  • Hello Bryan

    Thank you for reply.

    I will make corrections and share the results.

    Another thing that thinks of me is chemId. I say it as an assumption. If ChemID is wrong, does this error come into play or was the learning algorithm not complete?

  • Hi Faruk,

    Yes, if the chemID is wrong, gauging will not function properly.

    Sincerely,
    Bryan Kahler