• State
  • Locked Locked
  • Replies 3 replies
  • Subscribers 62 subscribers
  • Views 984 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.

BQ77910A Balancing LiFePO4 High Current Voltage Tolerance

jonathan digiacomandrea
Prodigy 190 points
Other Parts Discussed in Thread: BQ77910A

We have been doing a lot of testing on the cell balancing function of the BQ77910A. I find that during most of the charge, since we are using LiFePO4 cell the variation from the highest cell to the lowest cell is 40mV or less. The difference between the two highest cells is even less. I have included a graph below showing how the data stacks up.

 

I could not find a tolerance specification in the data sheet for the part. I understand the part compares all the cells and balances the highest cell, for dwell time then checks again. Could you find out if there is a tolerance? I worry we are under the tolerance and balancing is going to be spotty at best.

With all of the testing we are doing we are finding that the Balancing is activating, but not following what is specified in the data sheet. It is supposed to bleed the highest cell, sometimes it is, sometimes it jumps to a different cell. This is causing a large decrease in capacity since the balancing is bringing the pack out of balance.

 

Thanks,

JD

 

  • 0
    Prodigy 190 points

    Bump, Any information, or thoughts?

  • 0 in reply to jonathan digiacomandrea
    TI__Guru 54715 points

    The part uses the same measurement for protection as for balancing decisions.  The measurement vs comparison accuracies are not separated in the datasheet.  Since balancing is typically set at a high threshold, the OV threshold value is likely the best to use from the datasheet. The part will be calibrated at the factory with some resolution, so the differences may be very good near the 4.2V calibraiton point.  As the operating point moves from the calibration point or the temperature shifts, it is more likely to experience tolerances within the possible range. The lower operating voltage of LiFePO4 cells does move the part out of its most accurate region and the flat characteristic of the cell voltage profile makes the decision in that region more important.

    The balance mechanism in the part is very simple.  When enabled it will always balance to maintain a relative balance of the cells, that is it will balance a perfectly matched set of cells to match its calibration values.  It is best to balance only when there is a clear imbalance.

  • 0 in reply to WM5295
    Prodigy 190 points

    WM5295,

    Thank you for your response. I have shifted through the data a bit more considering your advice. If I am reading the data sheet correctly I see the following:

    Delta Vov @ 4.2V (closest to our setting of 3.75V) 0C-50C (range we are testing at) +/-25mV

    Additional shift of 25mV/1V

    Computing the additional tolerance (4.2-3.75=0.45V) so 0.45*0.025V= 11.2mV additional tolerance

    Total tolerance  (+/-25mV) + (+/-11.2mV) = +/-36.mV

    I then plotted the delta from highest to lowest cell, and the delta from highest to second highest cell. I also used the description of balancing in the data sheet to create a decision if balancing should be operating,a nd if so on what cell. I then compared that to what the balancing circuit was actually doing. I called that is balancing working.

    I also over-layed the computed tolerance specification.

    Looking at this plot I can clearly see that when the cell delta is higher, the balancing circuit has a much better chance of balancing the correct cell. But during most of the charge cycle the cell voltage is so close to each other that balancing makes poor decision. This seems to be the cause of the balancing issue.

    Looking at the Cell-Balance Enable Control I can see there are 4 options. Currently we are using 01, which is timer based. Since we can't control how long the user will take to charge this particular battery we don't know what the charge time is. We have been using the 8 hour threshold. At this time I am going to try to change this to 10, which balances as long as the charger is detected, and since I control this signal internal to the battery with a micro I will keep this signal high even during the rest period. If I give the pack long enough to rest I might be able to make up for bad balance decisions during charge. I don't really like this idea because I will be bleeding off an unknown amount of charge during rest, but I see this as the only possible solution.

    Please let me know your thoughts. I have included the plot as well.