• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Answers 2 answers
  • Subscribers 65 subscribers
  • Views 541 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: Problem with learning cycle and battery chemistry for BQ34Z100

Lucas RICHARD
Prodigy 30 points
Part Number: BQ34Z100-G1
Other Parts Discussed in Thread: BQ34Z100, BQ34110

Hi, I'm using the BQ34Z100 to develop a SoC estimator for 12 V 90 Ah Lead-acid AGM battery. I have developed my own electronic card which works well and after calibration and all the necesarry steps, I am trying to do the learning cycle to update the impedance table and the Qmax of the battery. But I have many problems.
Firstly, 2 months ago, I tried to identify the chemistry of my battery within your database and after doing the cycle you indicated, I got the ID 806 with 18% of error deviation for result. After a long talk with TI support, I just got the information that it was unfortunate but that I should continue with that and see if it works.
Now, I am trying to do the learning cycle, but the problem I have is that it is needed to pass 90% of DoD in the battery for a valid OCV measument to be taken. But for a 90 Ah lead-acid battery, it is very complicated or even impossible to pass 90% of the DoD. So, the impedance table update fails because the gauge does not see 90% of DoD passed.
So I was wondering:
- if the impedace table does not update because my equivalent chemistry is too far from the true battery I have ?
- if the impedane table does not update because 90% of DoD is not passed, and if so how to do without that constraint ?
- is the BQ34Z100 appropriate for a large AGM battery ?
So far, when I am using the gauge, the coulomb counter works well but every time a OCV measument is taken, the SoC is readjusted at a value which does not make sense, certainly because the impedance table or the chemistry of the battery is wrong.
Thanks in advance for your help, the BQ34Z100 is a critical component for my PhD and my company so I hope you can solve that issue.
Lucas

  • 0
    TI__Genius 10370 points

    Lucas,

    For this type of application I might lean towards using a CEDV based algorithm. If you are seeing a massive shift in your SOC when you take an OCV reading that means you do have a chemID mismatch. CEDV is a more coulomb count based algorithm. 18% error you are seeing means there is something wrong wtih the ChemID match. Thus we would need to possibly make you a new ChemID. For Lead-acid cells this is a very long process (2Months+)

    A good device to look at would be the bq34110. 

    Thanks,

    Eric Vos

  • 0 in reply to Vos
    Prodigy 30 points

    Thanks Eric for your answer. 

    I will certainly have a look at BQ34110, but I still want to try to make the BQ34Z100 work for my application as I have already worked quite a lot on it. 

    Do you think the massive shift in the SoC comes only from a chemID mismatch or it can also come from the impedance table not properly updated during the learning cycle ? 

    Is there a way to stop the BQ34Z100 to update the SoC when an OCV reading is taken ? 

    Also, to make a new ChemID, how would we do ? I am keen to know the details of the procedure. 

    Thanks for your help, 

    Lucas 

  • 0 in reply to Lucas RICHARD
    TI__Guru** 105590 points

    Hello Lucas,

    As Eric mentioned, having an error of 18% with impedance track will lead to large jumps in SOC when the cells relax because the DOD does not correlate well to the voltage. If there are jumps it would probably be best to switch to CEDV because the characterization doesn't depend directly on the DOD-OCV relationship.

    The correlation between SOC and OCV is pat of the impedance track algorithm, disabling it would make the gauge closer to CEDV type algorithm anyway.

    Sincerely,

    Wyatt Keller

  • 0 in reply to Wyatt Keller
    Prodigy 30 points

    Thanks for your answer. 

    I guess the correlation between SoC and OCV also depends on the impedance table. So I will still give it a try to succeed to update the impedance table and see if it's better, otherwise I would switch to BQ34110 or other CEDV fuel gauge. 

    Regarding the BQ34110, I have looked at it and it does seem very promising. However, it is mentioned that it is for rarely discharged application. But looking at the datasheet and the technical reference manual, I do not get why it is for rarely discharged application. What could go wrong if the battery is often deeply discharged ? 

    For my application (microgrid with a lot of small batteries here and there), the batteries will be discharged daily to around 50% SoC. 

    Regards, 

    Lucas

  • 0 in reply to Lucas RICHARD
    TI__Genius 10370 points

    Lucas,

    There is a specific feature included on the bq34110 that allows for rarely-discharged applications. This feature doesn't need to be used and the device will work as normal. 

    Thanks,

    Eric Vos

  • 0 in reply to Vos
    Prodigy 30 points

    Thanks for the info. 

    I will give it a try then. 

    Lucas