• State TI Thinks Resolved
  • Locked Locked
  • Replies 15 replies
  • Answers 3 answers
  • Subscribers 62 subscribers
  • Views 399 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: Battery capacity drops significantly when load reduced.

Peter Sherk
Prodigy 30 points
Part Number: BQ34Z100-G1


We're running a BQ34 on a number of 54Ah 12V lead-acid battery packs.  We've run into an issue that when we reduce the load on the batteries from the active load of between 1.0 and 2.4A to a standby load of 50 mA, the capacity indicated by the batteries will almost immediately drop from well over 50% remaining to 12% or less remaining, and it will then often drop to 0% remaining.  The pack voltage and other tests have shown that there is still the previous >50% capacity still in the batteries.  Any idea what could be causing this?

Thanks,

Peter

  • 0
    TI__Guru** 105870 points

    Hello Peter,

    Can you share your log files and .gg configuration file so we can take a look? There are many things that can lead to SOC jumps so reviewing the logs is the easiest way to see any issues.

    Sincerely,

    Wyatt Keller

  • 0 in reply to Wyatt Keller
    Prodigy 30 points

    Hi Wyatt,

    I'll have to try to simulate it on the bench, the drops we're seeing are from field test units that I can't get a real log out of.  It'll take me a few days at least, they're big packs.

    Thanks,

    Peter

  • 0 in reply to Peter Sherk
    TI__Guru** 105870 points

    Hello Peter,

    It may not be possible to debug without the logs, so I think that would be the most helpful. I can take a look at the configuration file while you collect the logs. It sounds like the Ra table may be incorrect, or one of the load select/load mode settings are a little off. Usually when SOC decreases from a change in load it's related to these 2 parameters.

    Sincerely,

    Wyatt Keller

  • 0 in reply to Wyatt Keller
    Prodigy 30 points

    Hi Wyatt,

    I've attached two logs that both show the drop, as well as a .gg file.  Hopefully this shows what you need.  The battery packs are made up of 3x Interstate DCM0018 deep-cycle SLA batteries.

    Thanks,

    PeterBattery Log - 10_15_2021.logBattery Log - 10_12_2021.log1Battery Data Log - 10_12_2021.gg.csv

  • 0 in reply to Peter Sherk
    TI__Guru** 105870 points

    Hello Peter,

    Thanks for the data, it helps a lot.

    It looks like you have RSOC_JUMP_OK enabled, I would change that to 0 so the gauge smooths the RSOC value while in relax. You can see right when an OCV is taken the RSOC jumps, that's the transition into RELAX mode. Also changing the load select to 0 which is more conservative may help reduce the SOC jump. What was the chem ID chosen for your battery? Do you have the GPCCEHM report?

    You can also disable the LFP relax feature.

    Sincerely,

    Wyatt Keller

  • 0 in reply to Wyatt Keller
    Prodigy 30 points

    Hi Wyatt,

    Thank you very much for the info.  I'll get this loaded into my test board and give it a shot.  We're running profile 0808, I've attached what I'm fairly sure is the report, it's been a while.

    Thanks,

    Peter3343.config-report.zip

  • 0 in reply to Peter Sherk
    TI__Guru** 105870 points

    Hello Peter,

    It looks like the chem ID is a good match (less than 3% max DOD error) so I think my above suggestions should help fix your problem.

    Sincerely,

    Wyatt Keller

  • 0 in reply to Wyatt Keller
    Prodigy 30 points

    Hi Wyatt,

    I'm going through the changes you suggested, but I don't actually see an RSOC_JUMP_OK.  Did you mean RELAX_JUMP_OK?  If so, it's already 0, as is RELAX_SMOOTH_OK and SMOOTH.  My Pack Configuration C is currently 0x37.  LFP relax is also 0...  I did notice that the LIFE flag in Flags B is set, indicating LiFePO4 RELAX, should that be disabled?

    Thanks,

    Peter

  • 0 in reply to Peter Sherk
    TI__Guru** 105870 points

    Hello Peter,

    Yes sorry it is RELAX_JUMP_OK. RELAX_SMOOTH_OK and SMOOTH should both be enabled. If they are disabled tat may lead to the RSOC jumping while in relax state. You can disable the LFPRelax since you are using lead acid.

    Sincerely,

    Wyatt Keller

  • 0 in reply to Wyatt Keller
    Prodigy 30 points

    Hi Wyatt,

    Unfortunately, we're still seeing drops.  Does TI by any chance offer a battery characterization service?  As in, could we send you our battery pack and get you to set up the profile for it?  This chip has basically been about eight months of headaches at this point, and we don't yet have the issues totally resolved...

    Thanks,

    Peter

  • 0 in reply to Peter Sherk
    TI__Guru** 105870 points

    Hello Peter,

    I will review your data again to give further suggestions. Can you share the new tests you performed with the .gg from before and after?

    We do not characterize complete BMS units, we only characterize the cells to be used in the final system (or close matches)

    Sincerely,

    Wyatt Keller

  • 0 in reply to Wyatt Keller
    Prodigy 30 points

    Hi Wyatt,

    I don't actually have the logs for the new tests, sorry.  We had to use our test bench to verify capacities on some field trial packs, so I couldn't use it for the testing.  I should be able to get the test pack back on the bench next week.

    Are you aware of any companies that offer full pack characterization services?  We need to get this darn thing fully trustable...

    Thanks,

    Peter

  • 0 in reply to Peter Sherk
    TI__Guru** 105870 points

    Hello Peter,

    There are pack houses that work with our gauges often that setup the parameters, but in the end application testing still needs to be done on the system load. They will perform the learning cycle and make the golden image according to your specifications.

    Sincerely,

    Wyatt Keller

  • 0 in reply to Wyatt Keller
    Prodigy 30 points

    Hi Wyatt,

    You mention system load there.  Is it possibly an issue that our actual final load is nowhere near the learning-cycle test loads?  The minimum I think the manual said is pack capacity/10, which comes out to a little over 5A for us.  Our actual system, however, doesn't draw over 1.5A.  Can this screw up the capacities in use?

    Thanks,

    Peter

  • 0 in reply to Peter Sherk
    TI__Guru** 105870 points

    Hello Peter,

    No this should be fine. as long as you set your charge and discharge current thresholds for your system correctly. We need to have that current to capacity ratio in order to characterize the Ra table accurately.

    Sincerely,

    Wyatt Keller