• State TI Thinks Resolved
  • Locked Locked
  • Replies 43 replies
  • Answers 22 answers
  • Subscribers 65 subscribers
  • Views 1825 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:

BQ34Z100: About scaling current and energy.

BQ34Z100: About configuring parameters for high capacity application

Hai Nguyen5
Intellectual 450 points
Part Number: BQ34Z100
Other Parts Discussed in Thread: GPCCHEM, BQSTUDIO, EV2400,

Hello,

I'm using 14S6P battery pack, the voltage of the pack when fully charged is 57.75V, the capacity of one cell is 4800mAh.

- I scaled the current of 6: used 6A actual current then calibrate it to 1A.

I set:

- Voltage divider: 57750 mV.

- I calibrated the voltage to (actual voltage/14). For example, applied voltage to the pack is 57750 mV, then the calibrated one will be 57750/14 = 4125 mV. The actual number of series cell is 14 but because I scaled voltage of 14 so I set the Number of series cell  is 1.

- Design capacity: 4800 mAh.

- Design energy: 4.125V x 4800mAh =  19800 mWh.

Did I set the parameters correctly?

Thank you.

  • 0
    TI__Mastermind 24870 points

    Hello Hai,

    The only way to test your setup is to run some gauging tests. Please refer to our blog on how to analyze gauge for accuracy RSOC reporting.

    Thanks!

  • 0 in reply to Kang Kang
    Intellectual 450 points

    Sorry I don't get your answer. I am new so I do not have much knowledge about bq34z100, lithium-ion battery...  

    For more detail, the chemistry ID of the used cell cannot be found in chemistry ID list. So I used the GPCCHEM tool to identify the proper chemistry ID.

    Before that, I need to configure the parameters correctly to get accurate measurement of current and voltage. 

     

  • 0 in reply to Hai Nguyen5
    TI__Genius 10370 points

    Hai,

    The suggestion is that you run cycles to see how the accuracy behaves. So far you have described the setup however you should see your true voltage reported by the gauge. You should set your number of series cells to the correct value, then perform calibration.

    At the end of the day your voltage reported by bqStudio should read correct and the current should be the scaled parameter. 

    Thank,

    Eric Vos 

  • 0 in reply to Vos
    Intellectual 450 points

    Hello,

    The reported current and voltage look accurate.

    I used the GPCCHEM tool to identify the Chemistry ID of used cell but the GPC report returned the result not correct. I used the NCA 4800mAh cell but it returned LMO 3350mAh type. 

    Thank,

  • 0 in reply to Hai Nguyen5
    TI__Genius 10370 points

    Hai,

    The GPC chem tool has no knowledge of the exact cell chemistry. It looks at the entire chemistry database here at TI and uses your log file to find the closest match. You should chose the ID with the smallest error that has the same charging voltage as your cell. The string NCA vs LMO might very possibly be a typo since often times we are not provided the exact chemistry type when cells are submitted.

    Please do use the cell ID that came up in a match. 

    Thanks,

    Eric Vos

  • 0 in reply to Vos
    Intellectual 450 points

    Hello Eric

    Thank you for your answer.

    After learning cycle and generating Golden File, I used the generated file to program for a new device. But in the new device, the reported current and voltage are incorrect.

    Do we need to calibrate voltage and current in the new device? 

    Thank you. 

  • 0 in reply to Hai Nguyen5
    TI__Genius 10370 points

    Hai,

    It is recommended every device have calibration done. The procedure would be Flash golden file to get configuration then calibrate. Calibaration is especially important on the bq34z100 due to the external components involvement.  I would expect he voltage and current to be "close" using a golden file. 

    Thanks,

    Eric Vos

  • 0 in reply to Vos
    Intellectual 450 points

    Thank you Eric for your support,

    How about the SOC, after flashing golden file, the SOC is also incorrect. Do we need charge to full and discharge to empty the battery pack to get correct SOC?

    Thanks,

  • 0 in reply to Hai Nguyen5
    TI__Genius 10370 points

    Hai,

    On initialization the gauge only has voltage, current, and temperature to base its initial SOC on. So if the battery is initialized under a load or charge current or an unrested voltage i would expect SOC error. Once the device rest this should correct the issue. 

    If you are not referring to initial SOC error i would say there is an error with your golden file for the cells that are being used. 

    Thanks,

    Eric Vos

  • 0 in reply to Vos
    Intellectual 450 points

    Thank Eric,

    I will re-identify Chemistry ID and make a learning cycle again. After that I will inform you the results.

  • 0 in reply to Hai Nguyen5
    TI__Mastermind 24870 points

    Thanks Hai,

    Let us know if you have more questions.

  • 0 in reply to Kang Kang
    Intellectual 450 points

    Hello team,

    I did identify Chemistry ID and learning cycle again. This time, the results look better. 

    But when I flash the Golden image into another pack, I cannot calibrate the voltage, the report was "Calibration - a read of data written failed comparison". The different between reported voltage and entered voltage is about 30mV. The reported voltage is above the voltage set in Flash Update Ok Cell Volt.

    CC offset, board offset and current still be calibrated successfully.

    I used EV2400, firmware version 0.28, the bqStudio version is 1.3.86

    Could you help me?

    Thanks

  • 0 in reply to Hai Nguyen5
    TI__Mastermind 24870 points

    Hello Hai,

    Make sure you disable dashboard in bqStudio when you calibrate.

    Can you try the bqStudio test version and see if the issue is resolved.

    Also, please attach your srec and we can try it on our side. Can you tell me what voltage you are calibrating at?

  • 0 in reply to Kang Kang
    Intellectual 450 points

    Hello,

    I did the calibration when dashboard was disable. The voltage was 3851mV when perform the calibration, and I want to calibrate to 3820mV.

    I will check it again with bqStudioTest. 

    Here is my srec file, could you please have check?

    Thank you.

    0100_0_16-bq34z100G1_htt_final.zip

  • 0 in reply to Kang Kang
    Intellectual 450 points

    Hello Kang,

    I have few more question:

    - If we buy high quantity bq34z100g1, does TI support program srec file before deliver the ICs? We will send the srec file.

    - Is there anyway to perform the calibration without bqStudio? It is better if it can be calibrated from programming. 

    It reduces the production performance if we use bqStudio in production process.

    Thank you.

  • 0 in reply to Hai Nguyen5
    TI__Mastermind 24870 points

    Hi Hai,

    1. I am not sure, I need to check with marketing. Do you want to take the discussion offline? Please add me as a friend on e2e.

    2. Yes, http://www.ti.com/lit/an/slua640b/slua640b.pdf

  • 0 in reply to Kang Kang
    Intellectual 450 points

    Thank Kang,

    I sent to you friend request, pls accept the request.

  • 0 in reply to Hai Nguyen5
    TI__Mastermind 24870 points

    Hi Hai,

    I did not get your friend request.

    However, the bq34z100 has FW loaded. You just need to change DF parameters via the host side driver. Are you okay with this approach?

  • 0 in reply to Kang Kang
    Intellectual 450 points

    Hello,

    But if I use that way, I still need program Chemistry ID by using bqStudio. Is it right? 

  • 0 in reply to Hai Nguyen5
    TI__Mastermind 24870 points

    Yes, that's correct.

  • 0 in reply to Kang Kang
    Intellectual 450 points

    Hello,

    Is it possible to program srec file into BQ34Z100 from a host microcontroller(ARM,PIC,...)? The host microcontroller connect to BQ34Z100 via I2C. 

    The program for microcontroller is like a bootloader but will flash to BQ34Z100. 

    Could you provide the protocol to read srec file?

    Thank you.

     

  • 0 in reply to Hai Nguyen5
    TI__Mastermind 24870 points

    Hello Hai,

    There is no protocol to read srec file. You can refer to flashstream app note on ti.com. That converts srec to .fs files using bqStudio.

  • 0 in reply to Kang Kang
    Intellectual 450 points

    Hello,

    I have made some experiment with my battery pack and found out some issue with the SOC reported from bq34z100g1:

    - During discharging with about 1C load, when the reported SOC reach 3%, it instantly decreased to 0%. The bit [SMOOTH] in Pack Configuration C register was enabled.

    - After that, I charged the pack until it reach 19% SOC (actually the true SOC is about 30%: TrueRC = 1299mAh, TrueFCC = 4295mAh). Then let the pack relaxed for about two days, the reported SOC is 32%. It will cause the bad experience for end-user. As I understand when enable bit [SMOOTH], the RC and FCC will gradually tend to TrueRC and TrueFCC, is there anyway prevent SOC increase when relaxing? I known in Pack Configuration C register have [RSOC-HOLD] bit, in [SLUSBZ5B] , it is recommended disable that bit when enable bit [SMOOTH].

    I have attached the log file and gg.csv file. Could you pls help me?

    Thank you in advance.

    Test.csv

    dataflash.gg.csv

  • 0 in reply to Hai Nguyen5
    TI__Genius 10370 points

    Hai,

    If you don't want to SOC in change when in relax mode you just need to clear ([RELAX_SMOOTH][RELAX_JUMP_OK]). This will force SOC to remain the same during relaxation. 

    The gauge will always drop SOC to 0% when termination voltage it reached. At a 1C discharge this is very aggressive, please make sure your [Load Select] is correct to represent this load. 

    Thanks,

    Eric Vos 

  • 0 in reply to Vos
    Intellectual 450 points

    Thank Eric,

    - [RELAX_SMOOTH_OK] and [RELAX_JUMP_OK] bit was cleared during the test. You can check in the attached .gg.csv file. 

    - I thought that the Smooth function will prevent SOC changing instantly. Could you pls tell me when [SMOOTH] is set, at which conditions the Remaining Capacity and Full Charge Capacity will be set equal to TrueRC and TrueFCC?

    - For Electronics Vehicle applications, which Load mode should I select? The Load Mode = 0 and Load Select = 1 are selected now.

  • 0 in reply to Hai Nguyen5
    TI__Genius 10370 points

    Hai,

    The gauge will not jump to any value other than 100% (At valid charge termination) and 0% (when termV is reached). 

    I am not sure on the application load, but Load Mode 1 is for a "Power Dsg" meaning as your voltage gets lower your Dsg current goes up. Load mode 0 is for constant current. 

    Thanks,

    Eric Vos

  • 0 in reply to Vos
    Intellectual 450 points

    Hello team,

    Recently, we test the battery pack with the e-scooter and we still got the the issue: 

    - SOC drop instantly from 7% to 0% when running e-scooter. I known it is because the voltage reach the EDV, but is it true that the smooth function doesn't do anything in this situation? If smooth function cannot handle this situation, is there any way that prevent SOC drop instantly even when voltage reach EDV?  

    - After running, the battery was let relax and the SOC still increase to 15% even the bits [RELAX_SMOOTH] and [RELAX_JUMP_OK] was clear.

    Do you have any document that explain clearly about the smooth function?

    Thank you so much.

  • 0 in reply to Hai Nguyen5
    TI__Genius 10370 points

    Hai,

    The smoothing function is easy. It takes the delta between what the gauge thinks is "True" and "Filtered" and artificially injects/removes accelerated passed charge to smooth the result. Jumps to 100% and 0% are different situations. Since 100% and 0% use measurements of voltage and current their values are absolute and not simulated. If voltage < termination voltage we are at 0% by definition. 

    This means we will always jump to 0% when termV is reached. 

    Smoothing in relax allows SOC to change when the battery has close to 0 current in it. If [Smooth_JUMP_OK] is set then instead of smoothing the adjustment over a time interval it will jump straight to the new value. 

    Thanks,

    Eric Vos

  • 0 in reply to Vos
    Intellectual 450 points

    Hello Eric,

    I don't know why, but in my case, that bits, [RELAX_SMOOTH] and [RELAX_JUMP_OK] was clear already before experiment.

    And could you pls have look at the attached log file? In the log file I noted that at the point SOC drop to 0%, the measured voltage is 3.243V, it is still above the Term Voltage (3V). I want to make it clear that the voltage using to terminate discharging is simulated voltage, not measured voltage, is it right?

    Thank you 

    2376.Test.csv

  • 0 in reply to Hai Nguyen5
    TI__Mastermind 24870 points

    Hello Hai,

    Can you disable fast scaling and re-run the test.

    Also, can you attach your gg.csv file. I think the jump down is due to a grid point.

  • 0 in reply to Kang Kang
    Intellectual 450 points

    Hello,

    Sorry for the late reply. I have attached the gg.csv file. Could you please have check.

    Thank you so much

    P1SampleA.gg.csv

  • 0 in reply to Hai Nguyen5
    TI__Mastermind 24870 points

    Thanks Hai,

    Your gg file update status is still showing no learning cycle, can you follow the learning cycle guide and provide a bqStudio log file prior to doing the learning cycle?

  • 0 in reply to Kang Kang
    Intellectual 450 points

    Hello,

    I already did learning cycle many time. I followed the learning cycle guide and change update status to 02 when learning cycle completed (learn status 06).

    I have attached log files during learning cycle. The log is separated into 4 files. Could you please have check.

    CellLearningLogFile.zip

    Thank you

  • 0 in reply to Hai Nguyen5
    TI__Mastermind 24870 points

    Hello Hai,

    I reviewed the files.

    Do you have the gg file after the learning cycle. The one you attached shows no impedance updates.

  • 0 in reply to Kang Kang
    Intellectual 450 points

    Hello,

    The gg file I sent to you was exported after learning cycle. I will make the test again and export gg file after that.

  • 0 in reply to Hai Nguyen5
    Intellectual 450 points

    Hello,

    In mass production, which tool did you usually use to load .srec file to chip? 

    Thank you,

  • 0 in reply to Hai Nguyen5
    TI__Guru 52115 points

    For mass production,  you can generate a bq.fs or dt.fs file in the bqStudio and then make use of something like the FlashStream Parser mentioned in the following app note to program the bq.fs or dt.fs file to a bq34z100-g1 device.

    http://www.ti.com/lit/an/slua801/slua801.pdf

  • 0 in reply to Andy Liu (BMS)
    Intellectual 450 points

    Hello,

    Do you have any project that implemented that function: program .srec file from a host micro controller (any micro controller)? 

    Could you share with us? We want to use it as a reference.

    Thank you, 

  • 0 in reply to Hai Nguyen5
    TI__Mastermind 24870 points

    Hello Hai,

    Srec files no, you can convert to an fs file and use the flash stream applications note.

  • 0 in reply to Kang Kang
    Intellectual 450 points

    Hello,

    If you have any project that can program bqfs file from a host microcontroller. Could you share with us?

  • 0 in reply to Hai Nguyen5
    TI__Mastermind 24870 points

    Hello Hai,

    No, you will generate this using bqStudio's golden image tab.

  • 0 in reply to Kang Kang
    Intellectual 450 points

    Hello,

    I known that we need to generate golden image file (.srec, bqfs and/or dffs file) after learning cycle from bqStudio. But what I want to ask is the tool to download the golden image file into the chip? 

    bqStudio can be used, but we saw a situation that bqStudio still inform "Download successfully" even downloading process was disrupted. 

    I also already read document SLUA801 about updating bq34z100g1 firmware from host microcontroller. But if you have a project that implemented these function in a specific MCU. Could you please send it to us? We want to use it as a reference source code.

    Thank you.

  • 0 in reply to Hai Nguyen5
    TI__Mastermind 24870 points

    Hello Hai,

    We don't have reference source code for you to use on this part.

    I would refer to the following:

    Smart Flash application note or going to production with the bq275xx as an example on how to parse the srec file. The bq34 uses the same engine as the bq275xx for this.