• State TI Thinks Resolved
  • Locked Locked
  • Replies 16 replies
  • Answers 6 answers
  • Subscribers 63 subscribers
  • Views 771 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.

BQ40Z50-R2: Unable to learn battery RA table

shphilippe
Prodigy 80 points
Part Number: BQ40Z50-R2
Other Parts Discussed in Thread: BQ40Z50, BQ24780S, BQSTUDIO, BQ40Z50EVM-561

Hello,

I've 4 LG HE2 in series and I'm trying to perform the learn cycle. The first Qmax went fine (UpdateStatus = 0x5). But then I'm unable to perform the next learning cycle.

The RA tables in flash is updated, thus I don't fully understand what is happening. I've attached two learn cycle I did to this post.

Thanks for the help,

Philippe

cycle-2.zip

  • 0
    TI__Mastermind 42535 points

    Hi Philippe,

    Your VOK bit never clears for a proper DOD0. My suggestion is to repeat your test with a fresh unlearned pack. Dsg it fully. Then wait for VOK to clear in relaxation. Then perform a learning cycle. Your Qmax should update first, update status 0x05, then wait for VOK to clear again at chg relax. Then dsg, that way VOK will set and your DOD will allow Ra updates and finally set update status to 0x06. Repeat the learning cycle again for update status of 0x0E.

  • 0 in reply to Batt
    Prodigy 80 points

    Hi,

    I redid a cycle from scratch, and the VOK never clear after discharge, even after 48 hours.

    I've attached the cycle to this post, any idea why it fail to learn ?

    Thanks,

    Philippe

    cycle3.zip

  • 0 in reply to shphilippe
    TI__Mastermind 42535 points

    Hi Philippe,

    Looking at your log, your voltage never stabilizes and always continues to trend downwards. Therefore the gauge is not able to take an OCV. This is why VOK never clears.

  • 0 in reply to Batt
    Prodigy 80 points

    Hi,

    Thanks for helping me on that, but how can I solve this ? The system is not drawing any current on the battery and the battery capacity is quite large (2500mAh, 4 cells). How can I perform a learning cycle in a such condition ? 

    Thanks,

    Philippe

  • 0 in reply to shphilippe
    TI__Mastermind 42535 points

    Hi Philippe,

    A crude hack to do it would be to disable the FETs until OCV is taken and then re-enable them. However you'll have to do this at top of chg and bottom of dsg to get an OCV. If you are using your own board, you may want to look at leakage due to board layout. Such a drop is highly unusual and not seen.

  • 0 in reply to Batt
    Prodigy 80 points

    Hi

    I don't think there is strong leakage in the board layout. I suspect more the power consumption for the bq40z50 because of the constant logging.

    And if this is the reason, then disabling the FET will not change anything because the bq40z50 bypass the FET for its supply.

    How can I solve that ? 

    Regards,

    Philippe

  • 0 in reply to shphilippe
    TI__Mastermind 42535 points

    Hi Philippe,

    The bq40z50 doesn't take that much current in normal mode, please see table 7.5 for power draw. Even if it's 1mA which is unlikely, you can't see that high of a drop.

  • 0 in reply to Batt
    Prodigy 80 points

    Hi

    Well, 1mAh during 72 hours, it's 72mAh. It's not negligible, the battery is fully depleted at this point.

    We have a BQ24780S charger which is in learn mode when doing the wait, thus the battery is totally disconnected for the system.

    Do you have any idea on how we could solve it ?

    Kind Regards,

    Philippe

  • 0 in reply to shphilippe
    TI__Mastermind 42535 points

    Hi Philippe,

    When we use the battery and log using bqstudio for the learning cycle the voltage does stabilize. In your case, I'm not seeing that at all. If your board has design artifacts then the only remedy is a redesign of the board. If you use an evm the learning cycle should work successfully.

  • 0 in reply to Batt
    Prodigy 80 points

    Hi,

    I've checked the board impedance, it's 2MOhm. Thus this is not produced by any leakage of the board.

    Logging at the speed requested by TI during 72 hours while the battery is fully depleted *will* drain the battery because of the power consumption of the bq40z50.

    Should I discharge less the batteries ? If yes, at what is the acceptable level ? 

    Thanks for your help,

    Philippe

  • 0 in reply to shphilippe
    TI__Mastermind 42535 points

    Hi Philippe,

    You can dsg the battery up to the manufacturer rated terminate voltage. The gauge in normal mode with comms will consume 336uA typically. If it's in sleep with no comms, the consumption would be even less. However, for learning purposes, you need to log, 76mAh in a 2500mAh is 3%. Relaxation should not take more than 5 hours tops if there's no drain. This gauge has been verified by many customers, all issues point to your board being the problem. If you want,you can buy an evm and check out the learning cycle on it.

  • 0 in reply to Batt
    Prodigy 80 points

    Hi, 

    So, we got a bq40z50EVM-561, and it's even worse. Here is the log.

    I had to stop the log because the voltage was getting really too low on one battery.

    What can we do about it ?

    Regards,

    Philippe

    Cycle-evm.zip

  • 0 in reply to shphilippe
    TI__Mastermind 42535 points

    Hi Philippe,

    Please allow me 2 days to look into this.

  • 0 in reply to Batt
    TI__Mastermind 42535 points

    Hi Philippe,

    I see that you have logged over 2.2 days and seen the voltage drop. Now, even in relaxation the gauge is going to consume current in normal mode since comms are on. If your battery is fully dsg at terminate voltage it can drop below that voltage quickly even under light or no load. I would suggest that you chg the cell since the gauge is already in rest and proceed with the learning cycle.

  • 0 in reply to Batt
    Prodigy 80 points

    Hi

    I'm sorry, I don't understand. The whole point of doing this is to do a QMax. I was told to do:

    "Your VOK bit never clears for a proper DOD0. My suggestion is to repeat your test with a fresh unlearned pack. Dsg it fully. Then wait for VOK to clear in relaxation."

    which I interpret as:

    Discharge fully. Wait until REST bit is on and VOK is cleared.

    Charge fully, wait until REST bit is on and VOK is cleared.

    You are telling me to charge without waiting the VOK bit to clear ? Then I fail to understand your first answer to this thread. Can you clarify ?

    Thanks,

    Philippe

  • 0 in reply to shphilippe
    TI__Mastermind 42535 points

    Hi Philippe,

    Please do the following

    1. send the reset command

    2. dsg the battery fully

    3. wait until rdis clears and the rest flag is set

    4. start charging and then perform the learning cycle.

    This should help complete it successfully.