• State TI Thinks Resolved
  • Locked Locked
  • Replies 16 replies
  • Answers 6 answers
  • Subscribers 63 subscribers
  • Views 659 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: Sudden large changes of TRUE FCC when current changes suddenly

Lucas RICHARD
Prodigy 30 points
Part Number: BQ34Z100-G1

Hi all, 

I have problem with my BQ34Z100-G1. After performing a successfull learning cycle, and running some satisfying tests, I have encountered a problem. 

I have tried a particular charging/discharging profile which is close to my application. In it, there are sudden changes between charge and discharge so the current can quickly go from 6 A to -6A for instance (with 12 V lead acid agm 90 Ah battery), and it seems that when that happens, the TRUE FCC suddenly changes by a large value, completely false. 

Then after some time, FCC gets to TRUE FCC and my SoC gauge becomes totally wrong (not the good SoC calculated). 

Enclosed is a log file of 5 days of testing. 

At some points in time, it is possible to see those sudden changes, for instance:

22-10-2021 at 17:03:30
26-10-2021 at 01:49:13
25-10-2021 at 23:29:16

and many other times. 

This is TRUE FCC which suddenly changes, sometimes by a very large value (18 965 mAh to 10 335 mAh for instance for 25-10-2021 at 23:29:16). This is obviously completely wrong as my capacity can not decrease that much. Also after a while (too late), or after a complete charge, the TRUE FCC comes back to around its normal value (around 18 500 mAh).

I would like to know how we can prevent this problem from happening as this makes my SoC gauge completely useless and wrong. 

Thanks a lot in advance for your help. 

Regards 

Lucas registre.log

  • 0
    TI__Guru** 105770 points

    Hello Lucas,

    With quick changes between charge, relax, and discharge modes you may need to decrease the Chg Relax Time and Dsg Relax Time so the gauge switches mode faster. You can also try modifying the load select to either 0 or 2 for a more conservative SOC estimation.

    Can you share the .gg file from the test? Ideally we like to see the .gg file from before and after the test to help with debug.

    Sincerely,

    Wyatt Keller

  • 0 in reply to Wyatt Keller
    Prodigy 30 points

    Hello Wyatt, 

    Thanks for your answer. I have launched a new test with Chg Relax Time et Dsg Relax Time decreased from 60 to 1 second. I will update you in 8 hours to see if it works. 

    I can already give you the .gg file linked to the log file given previsouly but they are many. You can find it enclosed. 

    Regards, 

    Lucas Test_22_10.zip

  • 0 in reply to Lucas RICHARD
    TI__Guru** 105770 points

    Hello Lucas,

    That is okay, once the next test is complete if you could share the log and .gg files we can take a look if the issue is not resolved.

    Sincerely,

    Wyatt Keller

  • 0 in reply to Wyatt Keller
    Prodigy 30 points

    Hello Wyatt, 

    I have completed another test of two discharge profils. At the beginning, for 6 to 7 hours, no issues and  TRUE FCC and TRUE RC did not change. 

    But at some point, it changed, again after a sudden change of current from relax to discharge, at 2021-10-27 16:37:44 (TRUEE FCC from 19328 to 8325 mAh).

    Enclosed are the .log file and the .gg file. 

    FCC is still good but normally after some points, it is set equal to TRUE FCC which has become wrong. I had to stop the test unfortunately so I did not happen. 

    Thanks for your help. 

    Regards, 

    Lucas 5684.registre.logTest_27_10.zip

  • 0 in reply to Lucas RICHARD
    TI__Guru** 105770 points

    Hello Lucas,

    The results look much improved. The smoothing function should remove the jumps you see in the true values.

    Sincerely,

    Wyatt Keller

  • 0 in reply to Wyatt Keller
    Prodigy 30 points

    Hello Wyatt, 

    Thanks for your answer. 

    So, what do I need to do to remove the jump in the true values ? The smoothing function is already enabled, right, as I don't have same values for FCC and TRUE FCC and RC and TRUE RC ? 

    I am quite sure that if I have let the tests run for a longer time, FCC would have been set to TRUE FCC which has become 8325 mAh, which is obviously wrong. 

    I can't run any tests before Tuesday but from Tuesday, I will give it another try according to your opinion. 

    Thanks for your help, 

    Regards, 

    Lucas

  • 0 in reply to Lucas RICHARD
    TI__Guru** 105770 points

    Hello Lucas,

    Over time it should correct with an OCV measurement. If the SOC still jumps with the filter on that would indicate a possible chem ID mismatch.

    Sincerely,

    Wyatt Keller

  • 0 in reply to Wyatt Keller
    Prodigy 30 points

    Hello Wyatt, 

    It is not necessarily the SoC which jumps, but more the FCC which then make the coulomb couting of the gauge wrong, and then the SoC gets wrong because of that.  

    For me, a chem ID mismatch would make the SOC jump when an OCV measurement is taken. Now the problem is more than during change between relax to charge or relax to discharge, the TRUE FCC or the FCC is badly updated. Am I right to say that the chem ID does not impact that ? 

    Also, for the chem ID, I got just 6% of deviation error, so I think it's quite good no ? 

    Thanks for your help. 

    Regards, 

    Lucas 

  • 0 in reply to Lucas RICHARD
    TI__Guru** 105770 points

     Hello Lucas,

    The FCC calculation is fairly complex, it depends on temperature, current, and voltage. The learned Ra and Qmax values are using to account for capacity drop due to higher currents and temperature changes.

    Usually we like to have it under 5%, but 6% should not cause the jumps you are seeing, it should only be 6% at certain locations. During any relax period when an OCV is taken the chem ID will have a large impact.

    It looks like the load select is set to 1 still, this also will cause FCC and RemCap calculation changes.

    Sincerely,

    Wyatt Keller

  • 0 in reply to Wyatt Keller
    Prodigy 30 points

    Hello Wyatt, 

    Yes I will run new tests next week with Load Select to 0 or 2. What do you think is best ? 

    Just to be sure to have understood correctly, I need to put the smoothing function to enable and the load select to 0 or 2, right ? 

    Thanks for your help. 

    Regards 

    Lucas 

  • 0 in reply to Lucas RICHARD
    TI__Guru** 105770 points

    Hello Lucas,

    It's hard to say which would be ideal for your system, the load select settings that rely on the present current values are usually a littler more aggressive with gauging, the load select options that use previous discharges are a little more conservative with the SoC estimations. If you can I would test a cycle with each setting to see which tracks better for your application.

    Yes those are the 2 suggestions I would give you to try.

    Sincerely,

    Wyatt Keller

  • 0 in reply to Wyatt Keller
    Prodigy 30 points

    Hello Wyatt, 

    I have tried doing a discharge profile with the load select set to 0 or 2 and this did not change the issue, unfortunately. The TRUE FCC still changes abruptly sometime and sets at an incoherent value. This happens most of the time after a sudden change of current (from relax/charge/discharge to an other mode). 

    Do you think that with smaller change of current, this will not happen ? I have no idea why this is happening so it's hard to know if this will happen when installed on the field. 

    Enclosed are the .log file and the .gg file of the cycles I did on the battery. At the beginning it's with load select at 0, then it's with load select at 2. Both times, the problems happened. TRUE FCC changes incoherently and then after some time FCC gets to the value of TRUE FCC. 

    Any advice on how to avoid this problem ? 

    Best regards, 

    Lucas 4162.registre.logTest_02_11_bis.zip

  • 0 in reply to Lucas RICHARD
    TI__Guru** 105770 points

    Hello Lucas,

    I would do a couple cycles for each load mode to confirm. I would not expect FCC to always be the design capacity, it will adjust according to temperature and the Ra table stored. I would let the gauge relax between load select tests so it can get an OCV and adjust FCC before the next load select test.

    You can also try editing the ResRelax time to 15, this may help reduce the RSOC jumps (and FCC)

    Sincerely,

    Wyatt Keller

  • 0 in reply to Wyatt Keller
    Prodigy 30 points

    Hello Wyatt, 

    Unfortunately, I am going abroad for my work for a couple of weeks (and that's where I am supposed to install the BQ...), so I will not be able to run tests. 

    I am also not expecting FCC to always be the design capacity, I am just surprised that it is changing so much (from 18000 to 8000, it's a lot). 

    Between different load select tests, I have let the gauge relax and the FCC went back to its "normal" value.

    I will try editing the ResRelax time to 15 and see what happens. I hope the problem will not happen when installed on the field, maybe because the current variation will be smaller. 

    Thanks for your help. 

    Regards, 

    Lucas  

  • 0 in reply to Lucas RICHARD
    TI__Guru** 105770 points

    Hello Lucas,

    I couldn't tell from the tests easily where the load select changed, but the different selections can change the estimations significantly.

    The ResRelax time should help with the fast current switches.

    Sincerely,

    Wyatt Keller 

  • 0 in reply to Wyatt Keller
    Prodigy 30 points

    Hello Wyatt, 

    Thanks for the answer. I will try the ResRelax time modifications when I am able to do so. I hope it works. 

    Regards, 

    Lucas