• State TI Thinks Resolved
  • Locked Locked
  • Replies 7 replies
  • Answers 1 answer
  • Subscribers 64 subscribers
  • Views 468 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:

BQ4050: GPC CEDV Accuracy Error

BQ78350-R1: GPC CEDV - Accuracy Error Deviation is so high that it is most likely due to anomaly in the data. Please check that data files have recommended format, units and test schedule

Abhishek Jethva
Prodigy 20 points
Part Number: BQ78350-R1

After uploading the files as described in the GPC CEDV user guide, I get following error 

Your Gauging Parameter Calculator report is here.
Thank you for your interest in the Gauging Parameter Calculator from TI.
There was an error in processing your report. The specific error encountered is:

The following errors were reported by the calculation engine:
Accuracy Error: Deviation is so high that it is most likely due to anomaly in the data. Please check that data files have recomended format, units and test schedule

Not really sure which deviation is the error referring to. Please advise 1348.GPC.zip

  • 0
    TI__Guru* 81034 points

    Hi Abhishek,

    I looked at your data. It looks like the stack voltage is actually rising during discharge for the first few minutes in some of the data files, which does not make sense. Are you sure the voltage measurement is accurate? I also notice that you are discharging to different voltages for the different files. I think you should try to discharge to a lower percent (less than 7% SOC) where the discharge curve is steepest. CEDV is compensating the end-of-discharge point, so if you are not reaching that point it will not be able to calculate accurate coefficients. You are specifying a 2500 mV termination voltage in your Config.txt file.

    Regards,

    Matt

  • 0
    Prodigy 20 points

    Hi Matt, 

    Thanks for your response. The battery pack voltage that you see going up during the discharge is for the cold temperature testing. Typically the battery cells have a higher internal resistance during colder temps (-15degC) thats why you see a huge sag in the voltage and as they get little warmer due to the discharge current the DCIR of the battery gets less and due to RC polarization the voltage rises a little which is pretty normal behavior.

    I was trying to discharge it down to 32,500mV (2.5V/cell since i have 13cells in series) the BMS on the battery protected it to 36,000mV (~2.75V/cell). I did mentioned 2500mV termination since its the hard limit of the battery cell. Since I am not going all the way down to that level what should I mention in the Config.txt file. Should I mention 2750mV or 2750mV x 13 cells = 35,750mV. The config.txt file calls for cell term V so I thought its asking for cell termination voltage and not pack termination voltage. 

    Your response is much appreciated. 

  • 0 in reply to Abhishek Jethva
    TI__Intellectual 1090 points

    Since LFP generally has worse accuracy, in some cases this result is the best you can achieve.
    But you can also experiment with chosing different values for 

    FitMaxSOC%=12
    FitMinSOC%=6

    for example:

    FitMaxSOC%=9
    FitMinSOC%=3

    ...as long as LearnSOC% is between these two values.
    Different fit range can sometimes get you out of a local minima in the fit.

    If it is possible to discharge to lower SOC % for your system, you can also reduce


    LearnSOC%=7

    to 

    5 or other value. Same change should be done in firmware.
    Lower learn SOC will result in higher accuracy, but you will learn at later point which is not always useable in your system.

  • 0 in reply to Yevgen Barsukov
    Prodigy 20 points

    I am not using LFP, I am using NMC

  • 0 in reply to Yevgen Barsukov
    Prodigy 20 points

    If you can explain FitMaxSOC, FitMinSOC and LearnSOC in little detail that will be great, no where in the user guide it explains it in detail

  • 0 in reply to Abhishek Jethva
    TI__Intellectual 1090 points

    Fit Max SOC is the upper % range of the voltage curve, that would be used in the fit, adn Fit Min SOC is the lower % range.

    Using wider range generally makes it harder to fit 7% value more accurately. Since fit of 7% portion is the main goal of this fit, so it makes sense to maker the range smaller.
    But at the other hand, making the range wider sometimes allows to capture the "bend" portion of the voltage curve which gives the most information for the non-linear fit and allows better convergence. So it is best to try to find a range that includes the bend, but is as narrow as possible otherwise.