• State TI Thinks Resolved
  • Locked Locked
  • Replies 17 replies
  • Answers 8 answers
  • Subscribers 65 subscribers
  • Views 555 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

BQ27441-G1: Random jumps in battery % after unplugging and replugging USB charging

Mike Credelle
Intellectual 300 points
Part Number: BQ27441-G1
Other Parts Discussed in Thread: GPCCHEM, BQSTUDIO, EV2400

I have a 1400mah battery that I am using the BQ27441, our system will report normal charging percentages, then when I unplug my usb cable and plug it back in, the percentage will jump 10-15%.  It doesn't drop back down, it's like it wasn't calculating the correct percentage the first time, how can I troubleshoot this?

  • 0
    TI__Guru* 76230 points

    hi Mike,

    You will need to provide a log file extracted using bqstudio so that we understand what is going on.Did you use our online tool gpcchem to determine if this gauge is a good match for your cellss?

    thanks

    Onyx

  • 0 in reply to Onyx Ahiakwo
    TI__Intellectual 2200 points

    Hello Onyx, 

    This part that was recommended to the Mike by my predecessor. I am hoping to clarify the case so that we can help assist. 

    GPCCHEM as not been used. The battery pack being used has a protection IC in it but no gauge so BQ27441-G1B is being used as a side gauge. Should the chemistry of the cells still be considered in this use case? 

    BQ27421 was used in the same way in a previous design and had accuracy issues. BQ27441-G1B was selected as a replacement but when BQStudio is launched with the EVM a compatibility warning is given- can you suggest a better method for this use case or a different solution you would recommend? 

    I appreciate the assistance. 

    Thanks, 

    Lindsey 

  • 0 in reply to Lindsey Fike
    TI__Guru* 76230 points

    Hi Lindsey

    We only recommend using a ROM gauge after running our online tool gpcchem and determining how close a match your cells is to the hardcoded chem ids in the device. 

    To avoid compatibility issues, a flash based gauge is recommended as you can use any chem id that comes up as a close match for your cells with these types of gauges. A good one would be the bq27520

    thanks

    Onyx

  • 0 in reply to Onyx Ahiakwo
    Intellectual 300 points

    I don't really have an option to make changes at this point, unless there is something seriously wrong.  However, even with a slightly improper calibration it doesn't make sense to me how our battery measurement can randomly jump 10-15% by only plugging in or unplugging USB power.  Our system isn't that sensitive to battery life as most use cases it will be plugged in, so if possible I'd like to try and make what we are currently building work.

  • 0 in reply to Mike Credelle
    TI__Guru* 76230 points

    Hi Mike,

    You have to run the gpccchem tool first. That is the starting point for us to understand what is going on with your system

    thanks

    Onyx

  • 0 in reply to Onyx Ahiakwo
    Intellectual 300 points

    I have the eval kit for the bq27441.  To be clear I need to get a ev2300 or ev2400 (A TI usb to i2c adapter won't work?) and then have some sort of circuit that can discharge and charge the battery while hooked up to it.  Does TI make any kit that can help manage the charging/discharging?  Do any 3rd party companies provide services for this?  I have not done much battery testing in my career and I just want to make sure I do it right with our limited time frame.

  • 0 in reply to Mike Credelle
    TI__Guru* 76230 points

    Hi Mike,

    We have the GDK that helps automate charging and discharging. It also comes with an onboard ev2400 so you do not need to purchase that separately.

    thanks

    Onyx

  • 0 in reply to Onyx Ahiakwo
    Intellectual 300 points

    I finally got the GDK, I ran a learning cycle last night but I don't see any data, is that what I was supposed to run?  Is there something else I should run?  Also I am a little unclear if terimnation/discharge voltage is supposed to be the voltage cutoff of the battery back deep discharge protection, or the minimum voltage I want my system to run at.

    Mike

  • 0 in reply to Mike Credelle
    Intellectual 300 points

    I can also run the GDK auto charge and discharge but I don't see any options to output any data, only a data graph, I am using the GDK and a BQ27441G1B

  • 0 in reply to Onyx Ahiakwo
    Intellectual 300 points

    Ok I think I ran it, i used GDK to charge and discharge and did a log of all the bq27441-g1b registers.  The instructions on how to do this are awful, by the way.  Here is my report.  It doesn't recommend using the bq27441-g1b, however I don't want to use a part with internal sense resistor, and we are trying to do everything we can to prevent any further layout changes.  Absolute accuracy is not super important to us, but I do need to figure out my initial question on how are system seems to randomly jump huge percentage just by unplugging and plugging in our usb power supply.cue battery-report (1).zip

  • 0 in reply to Mike Credelle
    TI__Guru* 76230 points

    Hi Mike,

    The error reported by the tool for your chem id is greater than 7%. This is the main source of error you are seeing as the gauge performs remcap simulation at the start of charge which occurs when you insert a charger. Do you have the smoothen flag in opconfig b register set? This should help prevent the jumps you are seeing upon inserting the USB.

    thanks

    Onyx

  • 0 in reply to Onyx Ahiakwo
    Intellectual 300 points

    It is enabled by default and we don't change it.  We just got several hundred units in and I spent a couple days charging up 100 of them, they would all report something like 40-60% battery after the 2 hours or so it should normally take to fully charge, but then unplugging usb and plugging it back in it would immediately jump to thinking it's full.  Is that what you would expect because of the error from the incorrect chemistry?

    Mike

  • 0 in reply to Mike Credelle
    TI__Guru* 76230 points

    Yes such errors has been seen from having incorrect chemistry. Is it possible to hook up your device to bqstudio and log the data so we see what is actually occuring on insertion. can you extract a gg file. Also,   Can you provide a screen shot of bqstudio register screen? Without providing critical log and gg files, all we will be doing is speculating.

    thanks

    Onyx

  • 0 in reply to Onyx Ahiakwo
    Intellectual 300 points

    I don't have the means to connect to our in system device, because our MCU is the i2c master, I would have to get some special firmware setup to use bqstudio.  I am also putting together a test system that will use one of the bq27461evm and see if that makes a difference.  With that setup I might be able to more accurately simulate what is going on and use bqstudio at the same time.  I will provide the info once I have it.

    Mike

  • 0 in reply to Mike Credelle
    TI__Guru* 76230 points

    oh good that you have an EVM. That will be the best way to try to reproduce the issue. Once you have the needed files ps provide it. I still fear that you might put in all this effort to get a solution and it might still boil down to the fact that the ID isn't a good match.  What happens in that case?

    thanks

    Onyx

  • 0 in reply to Onyx Ahiakwo
    Intellectual 300 points

    We have an opportunity to make a change on the fuel gauge circuit, but that would be a last resort.  Our battery life isn't really all that important, for the most part we just don't want people to run a test when the battery is very low, most of our use cases the battery is only a backup so we may just have to add some FW filtering, smoothing, checks to improve the user experience.

  • 0 in reply to Mike Credelle
    TI__Guru* 76230 points

    Hi Mike

    if there is an opportunity to change the gauge, pls go with  a flash based gauge like the bq27542 or the bq27520. This will allow you program a custom id and eliminate errors that using an incorrect chem id can cause.

    thanks

    Onyx