• State TI Thinks Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 2 answers
  • Subscribers 63 subscribers
  • Views 1322 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.

BQ34Z100-G1: Confirmation of Golden Image for production programming

Peter Money
Prodigy 140 points
Part Number: BQ34Z100-G1
Other Parts Discussed in Thread: BQSTUDIO

Hello,

Our system uses the BQ34Z100-G1 on the system board and a 2S 5000mAh LiPo battery which is not user replaceable.

During development I have used the bq34z100-G1 EVM to log the battery charge/discharge profile and this data was uploaded to the TI GPC calculator and several Chem IDs were suggested.

I have successfully programmed the best Chem ID into the EVM using bqstudio, and then ran several learning/optimisation cycles, charging at 2A and discharging at C/10 (500mA) constant current with rest periods between the charges and discharges. The Update Status register has now successfully incremented to 0x06.

So I think I have successfully learnt a battery profile? So I have generated the golden image .srec .bqfs, and .dffs files.

My question is, are these golden image files good enough to be used to configure the bq34Z100-G1 parts on our first production units? The golden image files are based on a sample of 1 battery and 1 EVM. We only need the gas gauge to give fair representation of battery capacity and not to be highly accurate on this first build (accuracy within 10% would be good)

Also, going forward, do I need to perform the learning/optimisation procedure on several batteries to get an average profile? If so then how do I do this and merge the data from each battery into one configuration file?

Finally, could you please take a look at the register settings to check that I have correctly set everything up. I have attached the GG file for review (I hope)

Thank you

Peter

2S_5000mAh_Export_25Jan18.gg.zip

  • 0
    TI__Mastermind 25955 points
    Hi Peter,

    With respect to the single images, please increase your sample size to at least 10-30 before creating the golden image. With the average values between these samples (assuming low variance), optimization may not be required in production.

    To test your image, please operate the gauge under normal application conditions and check to see if the SOC is appropriate.

    When creating your golden image, please set your status update to 0x02 and reset your cycle count, and lifetime values. Please also make sure to set your Auth Key.

    The values of the gg file look good assuming a max 4.2 V cell (similar to default settings).

    Sincerely,
    Bryan Kahler
  • 0 in reply to Bryan Kahler
    Prodigy 140 points

    Hello Bryan,

    Thank you for the information and checking the GG file settings for me.

    I will reset the status update register, cycle count, and lifetime values as suggested and re-generate the golden files. I will test this file on our next sample batch.

    I will also continue to test some more batteries to increase the sample size. Can you please advise on the following:

    1) I will be using the same EVM to perform the learning cycles on the new battery. How do I reset the bq34z100-G1 before testing a new battery to clear the learnt data from the previous battery test?

    2) For each new battery, I will perform the multiple charge, rest, discharge, rest cycles until the update register increments to 0x06. Discharge will be at C/10.

    3) Once the update register is at 0x06, what register values data do I need to record to create average values? Can you list which registers need to be averaged please.

    4) After I have run the learning cycles on a number of batteries and averaged the register values, how do I put these values back into the bq34Z100-G1 so that I can then generate golden files from bqstudio?

    5) Is it essential to set an Auth Key before generating the golden files? The bq34z100-g1 is embedded on our system PCB and not in the battery, and is not user accessible.

    Thank you

    Peter

  • 0 in reply to Peter Money
    Prodigy 140 points

    Hi,

    Could you please tell me:

    • How to reset the registers on my EVM before performing the learning cycles on a new battery pack? Do I need to erase the data from the previous battery?

    • What register values need to be averaged from completed learning cycle data from multiple cells? Also do I average this data in excel, then modify a .gg.csv file and then import it back into bqstudio?

    • Once I have averaged register values, how do I enter these back into bqstudio to create a golden file?

    Thank you

    Peter

  • 0 in reply to Peter Money
    TI__Mastermind 25955 points
    Hi Peter,

    1. You can reset the registers by saving the SREC file and gg.csv from the gauge (as a backup). In data memory, erase all the values you do not want to keep and save the SREC (to program the new gauge with).

    2. The values that need to be averaged will be the calibration data and data updated during the learning cycle such Qmax and RA tables.

    3. You will follow a similar process as in step 1, but replace the values to keep with the calibrated values.

    4. The auth key defaults are known by anyone that downloads the TRM. I would recommend changing them from default when going into production. If you are worried about someone in the factory seeing the key in a gg.csv file, for example, programming the key via commands is possible in a production environment.

    Sincerely,
    Bryan Kahler
  • 0 in reply to Bryan Kahler
    Prodigy 140 points
    Hi Bryan,
    Could you please provide a comprehensive list of registers to reset so that I don't overlook any and then mess up the next learning cycle.

    Thank you for your help
    Peter