• State TI Thinks Resolved
  • Locked Locked
  • Replies 18 replies
  • Answers 8 answers
  • Subscribers 65 subscribers
  • Views 1251 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: Problems completing a

Kristoffer Stenberg
Intellectual 1855 points
Part Number: BQ40Z50-R2
Other Parts Discussed in Thread: BQ40Z80, GPCCHEM, BQSTUDIO

Hi,

Me and a customer are looking into the Battery Charger: BQ40Z50-R2.

We follow the guide: "How to Complete a Successful Learning Cycle for the bq40z80".
Which states:

1. Discharge battery to empty
2. Relax for 5 hours --> Everything looks good this far
3. Charge battery to full
4. Relax for 2 hours --> Problems occur here

We cannot get clear the "VOK", change the "Lstatus" to 0x05 nor toggle the "QMax flag".
Are perhaps OC:s prohibiting the BQ to fully charge? Or what could be the reason behind the Lstatus, QMax & VOK.
I have attached 2 pictures illustrating the registers before and after, which shows that REST has been set.

I would like to share the logs, though they happen to be 39 Mb meaning that could probably be hard...
Any advice would be highly appreciated.

Many thanks,
Kristoffer

  • 0
    TI__Mastermind 42535 points

    Hi Kristoffer,

    Please use update status in dataflash as as reference instead of LSTATUS. That is a more reliable indicator. You can send the zip file here for analysis.

  • 0 in reply to Batt
    Prodigy 120 points

    Hi Batt,
    Could you elaborate where to look when you state "update status in dataflash" instead of looking on the Lstatus?

    BR
    Lars J

  • 0 in reply to Lars Johansson
    TI__Mastermind 42535 points

    Hi Lars,

    The update status is in the gas gauging: state section of the df.

  • 0 in reply to Batt
    TI__Intellectual 1855 points

    Hi Batt,

    You earlier mentioned: "You can send the zip file here for analysis.".
    Here are the zip file containing the three earlier pages attached to this thread, a log file "discharge.log" & a data flash excel file 904discharge.csv.

    You also mentioned that you can skip Lstatus in the learning cycle, how do we proceed from here then?

    Please do tell if you need/ want something else from us!

    Many thanks,
    Kristoffer


    20190920.zip

  • 0 in reply to Batt
    Prodigy 120 points

    Thanks Batt,
    Found it as you described it.

    As Kristoffer pin-pointed, if the Lstatus is skipped, is there any guide to follow to see if one has been able to complete a full learning cycle.   

    The learning cycle process described in this document is used for guidance. http://www.ti.com/lit/an/slua848/slua848.pdf 
    From 4.2.4 you might look on the update status in the “Gas Gauging” in the dataflash for Qmax as you mention earlier.


    Perhaps something like this you can interpret for 4.2.4 and 4.2.5 for the update status register in df, but for the last 4.2.6 remains how to see that.





    Thanks in advance!

    \Lars Johansson C

  • 0 in reply to Lars Johansson
    TI__Mastermind 42535 points

    HI Lars,

    In section 4.2.4 it would be reading update status in df where it will read 0x05.

    In section 4.2.5 update status should change to 0x06 at the end of dsg after Qmax and Ra are updated. It should be noted that RDIS should not be set during this dsg.

    If a relax-chg-relax-dsg-relax cycle is repeated again, you will have update status change from 0x06 to 0x0E.

  • 0 in reply to Batt
    TI__Intellectual 1855 points

    Hi Batt,

    Ok, thanks!
    Is there anything that looks abnormal in the log?

    With Regards,
    Kristoffer

  • 0 in reply to Kristoffer Stenberg
    TI__Mastermind 42535 points

    Hi Kristoffer,

    Please chg to full. You're simply not doing that here. Please perform the chem ID match using GPCCHEM if you have not done so prior to charging. I see the battery charge taper off but the voltage isn't anywhere near full chg. Your passed chg is also not close to design capacity.

  • 0 in reply to Batt
    Prodigy 120 points

    Hi Batt,
    (I work together with Kristoffer)

    In order to a chem ID match, the manual state that we shall do what I interpret as a full learning cycle and we can not do that.

    The batteries in use are LiFePO4 (3.2V) in series, 1500mAhx2=3000mAh. We charge the batteries with (3.6V*2)= 7.2V. I can see in the that we reach in one example 1477mA as passed charge.
    Design capacity shall be 3000mAh, but I have assume that the passed charge is the value times 2. Because when we charge from empty, we reach the voltage of 7.2V and the current goes towards 0A and the Q is around 1477mAh. Is there anything that we are missing? 


    Can you elaborate the statement  "voltage isn't anywhere near full chg."

    BR

    Lars

     

  • 0 in reply to Lars Johansson
    TI__Mastermind 42535 points

    Hi Lars,

    An LFP cell has a very large flat region in which the voltage doesn't change a lot during dsg. We depend on a combination of dV/dt, the IR drop and load rate to get the capacity. Without a correct chem ID a large error is possible and will result in bad estimates.

    I said the voltage wasn't near full chg as I didn't know the chemistry and thought it was Li ion.

  • 0 in reply to Batt
    Prodigy 120 points

    Hi Batt,
    Thanks for the comments.
    With the knowledge that it is a LFP batteries, could you suggest a way forward what you have seen in the .log file?
    I tried to upload the zip file to the GPCCHEM, but for an incomplete learning cycle, as we don't get further. The respond was
    "The following errors were reported by the calculation engine:
    Error: No discharge portion found Make sure discharge current is negative N_points= 2 0 0".

    BR
    Lars

  • 0 in reply to Lars Johansson
    TI__Mastermind 42535 points

    Hi Lars,

    Please upload the log for the GPCCHEM here. You need to follow the instructions on the GPCCHEM page precisely. The log simply needs to be a relax, dsg, relax log with zero current at start, a negative dsg current in the middle, and zero current at the end.

  • 0 in reply to Batt
    Prodigy 120 points

    Hi Batt,
    I have attached the file generated from the software BQstudio "Gauge Parameter Calculator Packer".
    Any good ways how to proceed with the learning cycle?

    7725.GPCPackaged.zip

    BR
    Lars

  • 0 in reply to Lars Johansson
    TI__Mastermind 42535 points

    Hi Lars,

    I uploaded your log to GPCCHEM but it returned an error.

    0 0 0 0 0
    Warning: Discharge rate is too high. High rate reduces ID detection accuracy 10 hr rate is recomeended. Total discharge time is less than 5 hrs t_hrs= -0.00111111111111111 0 0
    Error: No discharge portion found Make sure discaharge current is negative N_points= 2 0 0

    Can you please upload your battery DS here?

  • 0 in reply to Batt
    TI__Guru* 76230 points

    Hi Lars

    You will need to start the chem id identification with a charge step so charge to full, rest 2 hours, discharge using c/10 and rest 5 hours. Then submit the log file to the gpchem tool

    thanks

    Onyx

  • 0 in reply to Batt
    Prodigy 120 points

    Hi Batt,
    Unfortunately there isn't any good datasheet for this battery.
    https://batterigiganten.se/lifepo4-batteri-18650-3-3v-3c-1500mah-18x65mm.html

    I discharge the batteries to empty in the beginning with 500mA, so yes that is less than 5h. There where no recommendation in the beginning of the learning cycle.
    But I didn´t get any error in the learning cycle except after I fully-charged the batteries and I couldn't get any L-status/ right register updates.
    But perhaps is a consequential error ion the end.

    BR

    Lars

        

     

  • 0 in reply to Onyx Ahiakwo
    Prodigy 120 points

    Hi Onyx,
    I guess you refer to the normal learning cycle (wait to get certain register set and so on), or do you mean just doing the procedure as you propose regardless of the registry is correct or not.

    The problem is that I get stuck after fully charged the batteris, and letting the batteries relax, but I don't get any registry update as described in the learning cycle procedure.


    BR

    Lars

      

  • 0 in reply to Lars Johansson
    TI__Guru* 76230 points

    hi Lars

    No i am not refering to the learning cycle. I am refering to the chem id identification process as described in gpcchem our online tool.

    You need to start the test with charge to full first, then relax, then discharge to empty, then relax

    You then submit the log file to the tool and it will return a close id match which you will program on the gauge then perform a learning cycle. App note to guide you on the process below:

    http://www.ti.com/lit/an/slua848/slua848.pdf

    thanks

    Onyx