• State TI Thinks Resolved
  • Locked Locked
  • Replies 20 replies
  • Answers 6 answers
  • Subscribers 66 subscribers
  • Views 672 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.

BQ27Z561: Learning cycle related query

Pratik Panchal1
Prodigy 225 points
Part Number: BQ27Z561
Other Parts Discussed in Thread: BQSTUDIO, BQ40Z50, BQ34Z100

Hello Team,

We have used the BQ27Z561 in our project.

Here are the brief specs of the battery.

  • Battery Voltage - 3.7V
  • Battery Capacity - 30AH
  • Battery Chem - Li-ION
  • Battery Config - 1S10P

We have the Chem-ID for the battery.

We have started the Learning Cycle for the same.

Currently, I have few queries the related to learning cycle.

  • What is the critical parameter which needs to be taken care of during the learning cycle?
  • What happens if charging-discharging relax time will be stretched?
  • In the learning cycle Data parameter for the FC bit is 4150 V set, but still, FC is not set after the charging complete. I found that the battery voltage is 4185 at that time.
  • Do we have IC limitations to read the battery capacity of 30AH?

Here I have attached the gg, CSV, log file, and log error files for your reference.

Let us know the next action items to achieve the successfully golden file generation as we are chasing this activity from 1 year without any fruitful result.

Download.zip30042021.gg.csv

  • 0
    Prodigy 225 points

    Hello Team,

    Awaiting the response on this thread

  • 0 in reply to Pratik Panchal1
    TI__Guru** 105760 points

    Hello Pratik,

    You can follow the guide SLUA903 is very similar process. The current during discharge is very important to be C/5 and C1/10. The relax time can be longer than the stated values, it does not matter.

    Sincerely,

    Wyatt Keller

  • 0 in reply to Wyatt Keller
    Prodigy 225 points

    Hello Wyatt.

    What do you mean by C/5 and C1/10?

  • 0 in reply to Pratik Panchal1
    TI__Expert 3030 points

    Hi Pratik,

    It's current, if your cell capacity is 3000mAh, C/10 is 300 mA,

  • 0 in reply to Abraham Rashid
    Prodigy 225 points

    Hello Abraham & Wyatt,

    We have re-started the learning cycle as per the requirement.

    Battery Charging Current - 3.5A

    Battery Discharging Current - 3.2A

    After the completion charging cycle we found that its FC bit is not set. So we are able to get the only 16A capacity while our battery is 30A.

    Here i have attached the gg.csv file, log file and goldenfile.

    We are following the SLUA903 document and still we are not reaching nearer to the actual capacity. This is our 5th attempt for the learning cycle.

    Let us know what wrong here.

    Kindly provide the data why FC bit is not set after the full charging. 

    What is the basic requirement for the FC bit to set.

    0602.Data.zip

    Goldenfile.zip

  • 0 in reply to Pratik Panchal1
    TI__Genius 10370 points

    Pratik,

    In reviewing your log file i am seeing a major issue. Your DOD0 parameter is being updated to greater than 16K. In addition you are accumulating a passed charge near the 32k limit. 

    At this point i would reprogram a fresh bq27z561 image into the device and start the learn cycle over. Something is corrupted within the gauge. 

    When dealing with capacities this high you need to "Trick" the gauge to thinking is has a smaller capacity then it really does. For this case I would set my design capacity and Qmax to be 15000mAh. Then when you calbrate the gauge apply 2A, but tell the gauge it is 1A. Then everything internally the gauge will work at this scaled values. Your system will need to know your FCC is 1/2 what it really is. 

    In terms of VCT. The gauge looks for Voltage to be within Taper voltage of you "Charging Voltage" established by you advanced charge algorithm settings. Once voltage is in range it looks for current to be below taper current. You are not getting FC becuase the gauge is wanting the voltage to get to 4400, but you are only going to 4200. 

    Thanks,

    Eric Vos

  • 0 in reply to Vos
    Prodigy 205 points

    Eric,

    I am Pratik's colleague and working on the same issue.

    Thank you for sharing your observations, looks like it'll help us to resolve the FC bit issue! I have few more queries:

    1. What impact FC bit has on overall learning cycle? Is FC bit responsible for wrong learning of the capacity (it learnt ~16000mAh only) or any other parameters?
    2. What is the maximum limitation of BQ27Z561 IC? As per datasheet it is 30AH.
    3. We have tried generating golden file for 25000mAh battery previously and it was successfully generated as per the steps mentioned. But struggling with 30000mAh. Do you thing capacity can be one of the reason?
    4. I see you are suggesting to modify some of the parameters in settings. Can you advice with respect to below table available in BQStudio? This will help to make sure we remain on same page as suggested.

    Lastly, our overall production and customer orders are running in delay just because of this failure. Is TI supporting Golden file generation by the expert at TI lab? I'll be more then happy to ship our system or battery to the desired location in USA.

    --

    Thanks,

    Dhvanish Parekh

  • 0 in reply to Dhvanish Parekh67
    TI__Genius 10370 points

    Dhvanish,

    1. What impact FC bit has on overall learning cycle? Is FC bit responsible for wrong learning of the capacity (it learnt ~16000mAh only) or any other parameters?
      1. FC does not play a role in the learning. The important parameter is QMax_DOD and the resistance learning. Triggering a VCT event is critical to learn the "full charge point" 
    2. What is the maximum limitation of BQ27Z561 IC? As per datasheet it is 30AH.
      1. 30000Ah is the limit, but coming close to it is not a good idea. Any minor issue could cause you to roll over. I suggest keeping the range much smaller. 
    3. We have tried generating golden file for 25000mAh battery previously and it was successfully generated as per the steps mentioned. But struggling with 30000mAh. Do you thing capacity can be one of the reason?
      1. It is possible, I suggest doing the scaling to keep the range smaller
    4. I see you are suggesting to modify some of the parameters in settings. Can you advice with respect to below table available in BQStudio? This will help to make sure we remain on same page as suggested.
      1. There are lots of parameters that need to be updated. The ones in regards to VCT are the charging voltages inside the different temperature ranges on the "Advanced Charge Tab"

    Lastly, our overall production and customer orders are running in delay just because of this failure. Is TI supporting Golden file generation by the expert at TI lab? I'll be more then happy to ship our system or battery to the desired location in USA.

    Unfortunately this is not something we have the bandwidth to support at the moment. In addition, it would require a very deep system level knowledge on how you want the gauge to behave in the system.

    Please also refer to the "going to production" application notes on the bq40z50 product. the bq27z561 will behave very similar. 

    Thanks,

    Eric Vos

  • 0 in reply to Vos
    Prodigy 205 points

    Thank you for your instant reply Eric!

    I see what you are saying! Is it possible to arrange a call for exchange of our understandings before we start another cycle with some assumptions?

    Also, if in future if we are able to change the fuel gauge (as of now which is not open option) - which part would you recommend to use instead of bq27z561 for this kind of high capacity (30AH) application? 

    --

    Thanks,

    Dhvanish 

  • 0 in reply to Dhvanish Parekh67
    TI__Expert 3030 points

    Hi Dhvanish,

    You can refer to this app note for a better understanding www.ti.com/.../slua760.pdf

    There are a lot of other gauges which support higher capacity, such as BQ34Z100 or BQ27546, you can choose according to application requirements. But for this situation easiest way is to calibrate current to the half of actual current like Eric said.

  • 0 in reply to Abraham Rashid
    Prodigy 205 points

    Abraham and Eric,

    We'll be starting the new cycle as per your suggestion by Friday.

    Before that let us share the parameters we set based on all your suggestions and get your quick feedback. 

    Also, Eric said that, it requires deep system knowledge on how you want the fuel gauge to behave. Can you elaborate more on it and is there any thing more to it that we need to take care during golden file generation process?

    Like;

    Our in our system we have set charging current to 2.7A and the discharge rate is randomly varies between 10mA to 5A depending on the use case at particular instant.

    For golden file generation we will be going for CC discharge at 2A and charge at 2A.

    --

    Thanks,

    Dhvanish 

  • 0 in reply to Dhvanish Parekh67
    Prodigy 225 points

    Hello Eric,

    Kindly have a look into modified gg.csv file as per your suggestion.

    Let me know if any other parameter needs to be set or not.

    gg_csv.xlsx

  • 0 in reply to Pratik Panchal1
    Prodigy 225 points

    Hello Eric & Abraham,

    Awaiting you response on below thread.

  • 0 in reply to Pratik Panchal1
    TI__Genius 10370 points

    Pratik,

    I have reviewed your GG file. I did notice a few things that did not line up.

    You have design capacity as 15000mA, I assume you have chosen a C/2 scaling, but your QMax value is only ~5000 these values should be around the same for a new battery starting out. 

    Can you also confirm your senseR size you are using? Your CC gain and Capacity Gain are readout as 10 meaning you have roughly a 10mOhm. attached. This value should also be scaled by the same amount. So when you apply 2A the gauge registers 1A. 

    Thanks,

    Eric Vos

  • 0 in reply to Vos
    Prodigy 225 points

    Hello Eric,

    Yes, you are right. We have chosen the C/2 scaling. As far I understand that Qmax value is updated automatically after the completion of the learning cycle. Let me know should I write the value of 15000 in the Qmax or not. ?

    We have used a 10mohm sense resistor in our design. Let us know how much-scaled value should we need to use for the CC gain and Capacity Gain as per our condition.?

    Do we need to modify the sense resistor value?

    Regards,

    Pratik Panchal

  • 0 in reply to Pratik Panchal1
    TI__Genius 10370 points

    Pratik,

    Yes QMax is learned, but it is always good to start with a value close to the expected. 

    If you senseR is 10mOhm you need to re-do the calibration to scale. Please apply 2A, but in the calibration screen tell the gauge it is 1A. CC Gain and Capacity gain should be updated to this "scaled" value

    Thanks,

    Eric Vos

  • 0 in reply to Vos
    Prodigy 225 points

    Hello Eric,

    We have started the learning cycle without any scaled value & failed in the learning cycle.

    Here i have found one thing in the log file.

    "QMaxDOD0_1" became around half value of "QMaxPassedQ" after the OCV results has been taken.

    Currently, We are discharging the battery at C/10 value.

    Here i am attaching the gg.CSV file and log file for your reference.

    let me know what went wrong here. We also didn't crossed the max mAH value for the fuel gauge.

    PS: We have dismantle the whole set-up as its not getting proper value reflected, so you might not get whole learning cycle data.

    0753.Documents.zip

  • 0 in reply to Pratik Panchal1
    Prodigy 205 points

    To add more, we have already completed the learning cycle as per Eric's suggestion - Kept everything to half of the capacity. 

    After successful learning period, when we started the test with actual setup - the behavior was random and we are still seeing the % jump in the battery very often then it should be. 

    This is really frustrating that even after setting all parameters to what it should be every time it ends up learning to exact 50% of the QMaxPassedQ on QMaxDOD0_1. What exactly is going wrong here?

  • 0 in reply to Dhvanish Parekh67
    TI__Guru** 105760 points

    Hello Pratik and Dhvanish,

    Sorry for the late response.

    I'm not sure I understand what you have a question about, the scaling will decrease all parameters with A, mA, mAh, Ah to 1/2 the actual. From the log it looks like it correctly updated to 1/2 of your 30Ah battery.

    Sincerely,

    Wyatt Keller

  • 0 in reply to Wyatt Keller
    Prodigy 205 points

    Wyatt,

    The log file was correct but somewhere during the extensive test we found that the results were not linear. Also, at lower voltage (~3.3V) when we operate the battery for high load (7A) the battery voltage drops down to 2.8V, fuel gauge reads 0%. After that when the load current comes in normal range (~1-2A) followed by battery voltage (~3.2V), fuel gauge fails to read the normal %.