• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Subscribers 63 subscribers
  • Views 742 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.

BQ78350-R1: Gauging algorithm resets SoC too early (EDV) and OCV question

Paul Klarhoefer1
Intellectual 365 points
Part Number: BQ78350-R1
Other Parts Discussed in Thread: BQSTUDIO

Hi,

I think I did not understand completely the SoC algorithm in the bq78350-r1. 

First thing:

1) SoC is completely wrong after a reset of the device. I use a Samsung 50E 21700 cell. In your chemistry I found it and I did program this with (Program selected chemistry) in bqstudio. But for example, if all my cells are nearly full (above 4.1 V) and I do a reset, it resets the SoC to 53 %. Why and how can we change that ? Do I have to program the OCV curve somehow by myself ? Or is there some kind of command to set the SoC? Because it would be nice to set the correct SoC in production.

2) In my application there can be current peaks, 30 seconds long up to 13A (on a 14s1p configuration). So the voltage drops significantly for example on a fully charged battery from 4.2V to 3.7 V with 13A. But as far as I understood this should not be a problem, because the fuel algorithm calculates an Compensated End of Dischrage voltage, which should take the applied current into account. I attached a log. You can see in this log, that the SoC is reset from 41 to 7 percent. But not during high load (13A). It is reset after the high load, during low load. This is because the cell voltage need a couple of seconds to recover again. But in the meantime, the pending EDV switched back from (something 2.7V what I saw) back to 3.5V. And so the condition was fulfilled. So the question is: Is there a setting, which will increase the time evaluation for the EDV pending value ?

Thanks in advance.setting_1505_lab.gg.csv

cycletest_3.log

Paul

  • 0
    TI__Guru* 81034 points

    Hi Paul,

    The Gas Gauging section of this app note describes the role of ChemID and CEDV coefficients in determining SOC: http://www.ti.com/lit/an/slua924/slua924.pdf I think you may already have a good understanding of this material though.

    1) If you have a matching ChemID, it should help to determine the SOC accurately after a Reset. When you program the Chemistry in BQStudio, it writes to a large number of hidden data flash registers. After Reset or booting the device, the highest cell voltages will be used to estimate the SOC. There is another parameter that affects this SOC estimate: Dod at EDV2. That value is learned when the device learns FCC after a full discharge. I think your second issue with the current peaks might be causing the device to learn a wrong value for Learned Full Charge Capacity and Dod at EDV2. Also, let me know the ChemID number you are using in case the solution to issue #2 does not completely solve this initial SOC accuracy issue.

    2) Current peaks can definitely cause issues, but there are some parameters you can work with to meet the characteristics of your application. The parameters EDV 2 Hold Time, EDV 1 Hold Time, and EDV 0 Hold Time should all be increased as needed from their default value of 1 second. This will help to prevent false EDV detection during current peaks. There is also a [VFLT_EN] bit in CEDV Gauging Configuration that may be helpful to turn on. Using these parameters, you should be able to prevent false EDV detection. This is assuming your CEDV coefficients are good for your range of expected current loads. Solving this issue may greatly improve issue #1.

    Let me know how things look after making these adjustments. Look forward to hearing back.

    Matt

  • 0 in reply to Matt Sunna
    Intellectual 365 points

    Hi Matt,

    ok, thanks the SoC jump during current peaks is gone. So this issue is solved.

    What is still a bit weird, is the SoC estimation after a reset has been done. It is still not suiting. It is still setting to something 60 percent on >4.1 V on each cell.

    Is there somehow a parameter I can set by myself or is it always done via automatic calibration of the DoD at EDV2? Or can I set this value by myself or does it not make sense, since it is changed anyway by the chip itself?

    Thanks in advance.

  • 0 in reply to Paul Klarhoefer1
    TI__Guru* 81034 points

    Hi Paul,

    Did you select ChemID 2591 in BQStudio and program this to the device?

    Regards,

    Matt

  • 0 in reply to Matt Sunna
    Intellectual 365 points

    Hi,

    I did program 2520. But I just saw now, that there is also 2591. But it says it is also 50E? What is the difference between those two ?

    IMG_6198.pdf

  • 0 in reply to Paul Klarhoefer1
    TI__Guru* 81034 points

    Hi Paul,

    These two ChemIDs are almost identical in terms of voltage vs. SOC, so either one should work well. I think you meant to say ChemID 2250 which is the one shown in your image. ChemID 2520 will give a very different result since it is a different battery chemistry. 

    ChemID 2591 is the newer one (generated late last year), so it might be best, but like I said they are almost identical.

    Also keep in mind that the battery cell voltages will be lower after they are under a load and can take up to 2 hours to fully settle back to their OCV votlage. So if you are going into and out of Shutdown often then you may see lower SOC values reported. There is a data flash parameter called RemCap Init Percent you can use to scale the SOC value.

    Best regards,

    Matt

  • 0 in reply to Matt Sunna
    Intellectual 365 points

    So I did program the 2591. Now it works pretty accurate, even on a Reset event. 

    Thanks very much. This issue is resolved.