• State TI Thinks Resolved
  • Locked Locked
  • Replies 18 replies
  • Answers 7 answers
  • Subscribers 63 subscribers
  • Views 3139 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.

BQ34110: BQ34110 learning cycle 3S1P NiMh FDK 3/4

ziad jabbour
Prodigy 120 points
Part Number: BQ34110
Other Parts Discussed in Thread: EV2400, BQSTUDIO, GPCCEDV

Hello

We are looking to use BQ34110 with a 3S1P  NiMH FDK HR 4/3 FA

my question is how to get the profile to download into the BQ34110

do TI has the profile?

the application is a low power digital system

i have 3 dis-charge typical currents profile

1 - Sleep arround 150µAH

2 - Standard run mode : 10mAH

3- higth current discharge mode : 500mAH

the temperature range is -40 and 85°C

it will be recharged with a 220mA solar panel

if not, what are the tests that i should do, in case there are several tests at the same temperature, can i do tests in parallel, while using several packs ( what is the methodologie to do, using the ba34110evm and EV2400 ??)

Thanks in advanced

Best regards

  • 0
    TI__Guru* 81034 points

    Hi Ziad,

    I am attaching an example settings file for your configuration to help you get started. There are learning cycles that need to be run in order to tune the gauging accuracy. The bqStudio has a 'GPC Cycle' plugin (at the top of the screen) to help with this data collection and you can read about the online CEDV gauge parameter calculator here:  http://www.ti.com/lit/ug/sluub45b/sluub45b.pdf The tool is located here once you have collected the logs: http://www.ti.com/tool/GPCCEDV

    I think there are a couple of challenges for your application. Charge completion is very hard to detect for NiMH batteries. The best method to detect charge termination is to monitor temperature as it will begin to increase faster when charging has completed (delta T method). Another way to detect charge termination is to watch for a small dip in the voltage as it will peak and decrease slightly. Since the charge current for your application will be very low, neither of these methods will work well. 

    There is a 3rd option that might be best for your application. Charge termination can be achieved by setting Cell Negative Delta Voltage to 0. In this case, the gauge monitors when the pack voltage exceeds Cell Negative Delta Qual Voltage for Cell Negative Delta Time, and if Current() > 0 during this time, then charge termination will be detected.

    Here is the file below you can load to start with. Opening and saving with Excel can corrupt the file, so please use a text browser for viewing (or just load it using bqStudio).

     /cfs-file/__key/communityserver-discussions-components-files/196/1682.bq34110_5F00_3Cell_5F00_NiMH_5F00_SampleSettings.gg.csv

    Best regards,

    Matt

  • 0 in reply to Matt Sunna
    Prodigy 120 points

    Hi Matt

    Glad to have your feedback,

    In facts we have two subjects to deal with : 1 - Charge termination & 2 - learning cycle,

    1 - About the charge termination, with the battery supplyer, and since we are charging with a current less than C/10 ( in our case it will be around 230mA for 3300mAh capacity) we will not manage the issue. In other words the charge termination voltage is set to 1,6VDC per cell, so with 220mA charge current the battery will never reach this voltage regardless what it will be the temperature. since it is a solar charge ( in best cases we will have few hours of charge per day) there is no risk to damage the battery :

    ==> My question : is that will perturb the BQ34110 operation and capacity estimation?

    2 - about the learning cycle, in fact if I understood well, and for my use cas I need to dis charge my battery at 3 constants temperatures ( -15 / 25 / 50) and with two current values for each temperature ( -40mA and -300mA)

    ==> my questions :

    is that a good configuration for my use case?

    if yes, since there are two tests per temperature, can i use two battery packs for the tests?

    is there a simple way to log data? using the EV2400 ?

    test should be done using the battery pack only, and after that i should set the variable "electronics load" in order to integrate my sleep current 150µA ( 135µA for the BQ34110 and 15µA for my electronic board)..., am I rigth?

    looking forward to have your help

    best regards

  • 0 in reply to ziad jabbour
    TI__Guru* 81034 points
    Hi Ziad,

    Regarding question #1, I think this is not the most ideal scenario for the best accuracy as the battery ages. This is because the CEDV gauge algorithm tunes it parameters as it goes through learning cycles in the application. So if the battery rarely goes through a full charge and discharge cycle, this may reduce the accuracy over time.

    For the learning cycle, I think it might be better to run these tests with higher currents to get accurate full charge detection. It should be fine to run two battery packs for the test. The easiest way to log the data is to use the EV2400 and bqStudio. There is a GPC Cycle Plugin that will help you collect and package the data. You are correct in your understanding of the "electronics load" feature.

    Best regards,
    Matt
  • 0 in reply to Matt Sunna
    Prodigy 120 points

    Thanks for your feedback,

    about my first question : i did not found a good solar energy harvester with nimh techno managment ... so i will be forced to manage like this, and i get the approval of the battery manifacture ... when you talk abot reducing accuracy over time, can you put numbers ... by how much please ???

    for the second issu,  will be doing the 6 cycles test and i will update you with results, once done whant should i do please?

    best regards

  • 0 in reply to ziad jabbour
    TI__Guru* 81034 points
    Hi Ziad,

    Unfortunately, I believe the accuracy error can be as high as 30% as the battery ages without proper learning cycles during operation. There are two big challenges here: the very low charge rate makes it impossible to detect when full charge is reached, and the absence of learning cycles during operation means the CEDV parameters will not be updated during operation. As a battery ages the capacity decrease - the learning cycles allow the gauge to learn the new capacity as it changes.

    Once you have collected the data from the 6 cycles, you can submit the results to the GPCCEDV tool (follow the instructions at www.ti.com/.../GPCCEDV). The results will be emailed to you and then you need to update the CEDV data flash parameters from the results.

    Regards,
    Matt
  • 0 in reply to Matt Sunna
    Prodigy 120 points

    Hi

    I did the tests but i have error when passing by TI website to get the characteristics

    Your Gauging Parameter Calculator report is here.

    Thank you for your interest in the Gauging Parameter Calculator from TI.
    There was an error in processing your report. The specific error encountered is:

    Error: Computation was not successful, but the specific reason was not identified.
           Please check that all files are prepared according to the documentation.
           The support personnel have been informed


    Please see the Gauging Parameter Calculator for information about the file format and the procedure for creating the input zip file. Should you want to resubmit or need assistance with the Gauging Parameter Calculator, please reference the sites below:

    • Go to Gauging Parameter Calculator to re-submit.
    • Visit TI E2E™ Community on Battery Fuel Gauges or the Gauging Parameter Calculator.
    • TI Battery Management Homepage


    Thank you for using TI's Gauging Parameter Calculator for Battery Fuel Gauges.

    i attached the zip file,

    can you help to get bq init file please

    best regards2262.GPCPackaged.zip

  • 0 in reply to ziad jabbour
    TI__Guru* 81034 points

    Hi Ziad,

    I was able to get a result from the GPCCEDV tool after making some changes to your files. The CellTermV in the config.txt file was set too high, but I think you may have been setting the total voltage when it should be the voltage per cell. I changed this value to 950 and it helped resolve the error. 

    I also deleted the end portions of some of the logs where the voltage is no longer decreasing but the current is changing significantly. Here is the resulting output file.

     /cfs-file/__key/communityserver-discussions-components-files/196/0652.1157.GPC_5F00_CEDV.txt

    If you need to improve the accuracy further, you might see better results running the tests with constant current. 

    Best regards,

    Matt

  • 0 in reply to Matt Sunna
    Prodigy 120 points

    Hello

    thanks for your support,

    please can you tell me how to download this into the bq?

    best regards

  • 0 in reply to ziad jabbour
    TI__Guru* 81034 points

    Hi Ziad,

    You will need to manually enter these parameters from the output file into the data flash. There are seven values you will need to update. You can ignore the VOC75, VOC50, VOC25, and EDVV parameters from the output file since these do not apply for this device. These parameters are under Gas Gauging -> CEDV Profile. 

    Best regards,

    Matt

  • 0 in reply to Matt Sunna
    Prodigy 120 points

    Hello thanls for your feedback

    can you confirm that ui dont need to update the cell number neither the chemestry type or the capacity ...

    i just only need to specify those informations ?

    best regards

  • 0 in reply to ziad jabbour
    TI__Guru* 81034 points
    Hi Ziad,

    I'm not sure I understand the question. The number of cells and design capacity both need to be set, but I think that is already correct in your .gg file. Do you mean for the GPCCEDV tool? The number of cells and chemistry type are in the config.txt file, and I think these were both set correctly in your file.

    Regards,
    Matt
  • 0 in reply to Matt Sunna
    Prodigy 120 points
    Hi
    but should i put those information into my BQ?
    best regards
  • 0 in reply to ziad jabbour
    TI__Guru* 81034 points
    Yes, the .gg file should be loaded into the device with bqStudio. I think you've already done this, right?
    The outputs from the GPCCEDV tool should also be updated. You can either enter these into the .gg file, or you can enter them through bqStudio and then export a new .gg file.

    Matt
  • 0 in reply to Matt Sunna
    Prodigy 120 points
    no i just update the 7 values that you showed earlier
    please can you list all the steps that i should do, i am lost ...
    sorry
  • 0 in reply to ziad jabbour
    TI__Guru* 81034 points
    Hi Ziad,

    No problem. So at the beginning of our conversation I sent a .gg file as a starting point (on August 8). The BQ34110 has a large number (276) of data flash settings because it supports multiple different battery chemistries and multiple extra features. The default settings for the BQ34110 are for Li Ion batteries, so these would not work for your application (3S NiMH).

    1) So the first thing to do is to load the .gg file to the BQ34110 using bqStudio. I think you must have done this because you were able to run the charge/discharge cycles. (In bqStudio this step is done by running Import->'Load from File' on the Data Memory screen.)
    2) Modify any of the flash settings as needed for your application. I created the sample .gg file using what I knew from your application, so you might not need to make changes. (If you make changes, you can run Export from the Data Memory screen to create an updated .gg file)
    3) Run the discharge cycles required by the GPCCEDV tool. You completed this.
    4) Update the 7 CEDV parameters from the GPCCEDV tool output file. This should improve the SOC (state of charge) accuracy reporting for the device. You can update these by writing directly to these parameters in bqStudio (and then save the updated .gg file) or you can modify the .gg file directly and Import it into bqStudio.

    Now the bq34110 should be configured with the optimal settings.
    As I mentioned before, your application is very challenging since NiMH full charge detection is not possible with low charge current. I think your application is using C/10 and is not going to achieve full charge/discharge cycles because it is solar, right? You may need to do a learning cycle (full charge and discharge) using a high current (C/2) so that the gauge can initially learn the capacity. This will give you a starting point with accurate SOC reporting, but the accuracy will degrade over time without full charge/discharge cycles to update the learned capacity.

    I'm sorry, I know this is alot of information. I just want to make sure you understand the challenges for this use case.
    Regards,
    Matt
  • 0 in reply to Matt Sunna
    Prodigy 120 points

    Hi Matt,

    Thanks for your feedback, In fact i did the discharge cyles with our power supply that include data log ... ==> is that impacts the result ?

    for the configuration, Honestly I am lost, can we have a call in order to discus about ? or do it the first time in parallel ... I have lots of errors

    Looking forward to have your feedbacks

    best regards

  • 0 in reply to ziad jabbour
    Prodigy 120 points

    Matt, when i try to download the file that you sjhared earlier i have this message

  • 0 in reply to ziad jabbour
    TI__Guru* 81034 points
    Hi Ziad,

    I think it is okay that the discharge logs were collected without the gauge settings. The important data in the discharge logs is the voltage, current, and temperature - this will be what the GPCCEDV tool uses to determine the CEDV parameter settings.

    The error you are seeing when you load the file into bqStudio is not critical. I think I understand what is causing it - bqStudio is using an older configuration file for the bq34110. To resolve this error, you can follow the steps in section 2.4 of the EVM User Guide: www.ti.com/.../sluubi1a.pdf

    I think Sebastian is going to arrange a meeting next week, so hopefully we can discuss your settings and application in more detail.

    Best regards,
    Matt