BQ27Z561: Battery training procedure using BQ27Z561 fuel gauge IC and BQ Studio

Hello,

1 - So can i consider that i have successfully achieved learning cycle?

2 - But when i do testing, gauge will take new updated value as a full charge capacity instead of 3088 mAH, then how i can i test properly if there is reduced full charge capacity? How to use 3088 mAH as a starting point for testing?

Note: When I say testing the gauge, I mean to run the scenarios the gauge will be exposed to in your application

What scenarrios are you refering to, please explain me in detail.

Q - Are you using 2 S configuration? 0E means cell balancing has been enabled.

I am using 2 batteries connected in parallel (2 x 1500 = 3000 mAH) that is why i was thinking for cell balancing. Is that correct?

Q - Remember to run some tests on the gauge before mass production. At the test stage, some parameter might be tuned for your application.

What parameters will i have to tune over here?

Q - When I say test, I say putting the gauge under the required application temperature and doing charge-relax-discharge cycles and monitoring RSOC and checking that there are no drops or jumps. 

What will i have to do if there is any drop or jump?

Thanks,

Parth

Hello Parth,

1- Qmax and Ra table updated. The gauge already learned the cell.

If you want to do a sanity check, you can do a quick charge-relax-discharge-relax cycle. Five hour wait is ideal.

2- Manually insert 3088 mAh for Qmax Cell 1.

These scenarios are application dependent. For instance, say you want to test the bq27z561 for a cell phone application where the load changes very dynamically, then you'd want to test the gauge under dynamic loads and check if it responds well.

Cell balancing are required for series applications 2S and above. Your configuration is 1S 2P, no need to balance for 1S, parallel cells will equalize the voltages.

Depending on what errors you run into when testing, then you can ask us to review your log file and configuration file. Make sure you log while you are testing the gauge.

Parth Mehta said:

What will i have to do if there is any drop or jump?

We will let you know in case of a RSOC jump. Make sure you log everything.

Regards,
Jose Couso

Hello, 

1 - But if i do sanity test then update status will change to 0E from 06 as per reference manual. 

2  -If i update qmax manually to 3088 mAH then how full charge capacity gets updated, i could not understand it?

3 - Updating status to 0E is to ensure accuracy, right? So it is good to run last charge-relax-discharge cycle, right? 

Thanks, 

Parth

Hello Parth,

1- If the charge-relax-discharge cycle is done with constant current, it should be updated.

2- Full Charge Capacity is the total amount of charge that the battery can deliver to a system for a specific set of conditions (average load, temperature). The gauge does this automatically, you shouldn't worried about FCC.

3- 0E is for cell balancing, and in your case this is not needed. The accuracy for 1S will be the same if update status is 06 or 0E.

I suggest that you move forward with testing the gauge after you do the sanity test.

Regards,
Jose Couso

Hello,

So after sanity testing i will have to change update status to 02 from 06 to disable IT gauging and also set the Qmax Cycle Count to 0, and set the Cycle Count to 0 too before generating golden image file, right?

Thanks, 

Parth

Hello,

As you said : If you want to do a sanity check, you can do a quick charge-relax-discharge-relax cycle. Five hour wait is ideal.

1 - Which test come under sanity test?

2 - Can i discharge the battery by C/2 for sanity test for fast operation?

3 - 5 hours of relax time is compulsory after charging the battery for sanity test or 2 hours are OK?

4 - Do i have to do calibration for Discharge as well as charge both to get proper current data before generating golden image file?

5 - So after sanity testing i will have to change update status to 02 from 06 to disable IT gauging and also set the Qmax Cycle Count to 0, and set the Cycle Count to 0 too before generating golden image file, right?

Thanks, 

Parth

Hello Parth,

Sorry if I wasn't clear enough when I said "sanity test". The reason I suggested a this is because the last charge cycle you performed did not have a constant current. It is up to you if you want to do a fast learning cycle by only waiting 5 hours after each RELAX. I say fast compared to the 24 hours wait you did before.

CycleCount() is a read-only function. Setting IT or disabling IT won't change CycleCount(). CycleCount() increments each cycle and resets after it reaches Cycle Count Threshold - CCT. 

5- The reason why Update Status goes from 06 to 02 when disabling IT is because the Update Status register contains a bit called ITEN. Please see below, ITEN must be set for the Impedance Track algorithm to function.



1- There's no such test that comes under sanity test, this a word I used for performing a fast learning cycle.

2- Keep in mind you are still doing a learning cycle but fast. I suggest C/5

3- 5 hours is ideal

4- before generating golden image, it is recommended to do calibrations.

Regards,
Jose Couso

Hello, 

1 - As we discussed that Qmax and Ra has been updated already then is it required to do fast learning cycle now? 

4 - Sorry i was not clear enough.What i mean to ask is that do i need to calibrate for charging as well discharging both, because one has to calibrate with positive current input value whereas other has to calibrate with minus current value? And with how much current do i need to calibrate both? 

5 - As per reference manual i need to change update status from 06 to 02 as well as reset all counts which occured during learning cycle before generating golden image for mass production, is that correct? 

Thanks, 

Parth

Hello Parth,

1- Not required, optional. Your gauge already learned Qmax and Ra table.

Jose Couso said:

Sorry if I wasn't clear enough when I said "sanity test". The reason I suggested a this is because the last charge cycle you performed did not have a constant current. It is up to you if you want to do a fast learning cycle by only waiting 5 hours after each RELAX. I say fast compared to the 24 hours wait you did before.

4- What current calibration means for the gauge is the bqStudio has a calibration feature which tunes the ADC Coulomb Counter of your board for best current accuracy. I do not know what you refer to by "calibrate for charging as well discharging both"

Please refer to calibration section in the bq27z561 TRM

5- You shouldn't be worried about changing the update status 02 from 06 at this moment. You still have to test the gauge under the application the battery will be put under. 

Note: The SLUA777 is a reference guide for learning cycle completion only. You should always reference back to bq27z561 TRM for your technical questions, not SLUA777

Regards,
Jose Couso

Hello,

I referred TRM as you suggested but i could not find some answers which i was looking for. Please help me to find some answers.

1 - Can you help me to identify the test cases which i need to perform after completion of learning cycle.

2 - When should i change update status to 02 from 06?

3 - Full charge capacity is getting changed if i update Qmax cell 1 value manually. My battery actual battery capacity is 3000 mAH (2 x 1500), so in my case what value do i have to set in Qmax cell 1 to get accurate Remaining capacity data?

4 - Also i observed that my RM is 0% although i am getting battery voltage = 3.15V, so when RM = 0 then voltage should not be zero?

Because if RM =0 that means battery is fully drained so at that time battery should not be able to provide any voltage, is that correct or not?

Even though RM = 0, i am getting voltage = 3.15V and fuel gauge is ON and i am able to see fuel gauge data in bqstudio, but is this possible?

5 - I am getting same FCC and RM even though i am lowering actual battery capacity to 1500 mAH (Connecting only one 1500 mAH battery instead of both), so what could be the reason behind it? FCC and RM should reduce with actual capacity, Right?

You asked -  I do not know what you refer to by "calibrate for charging as well discharging both"

If we calibrate gauge using bqstudio then we have to give input current value in applied current. So when we put battery in charge mode and calibrate gauge then we have to give positive input current value in applied current option and similarly when we put battery in discharge mode and calibrate gauge then we have to give negative current value in applied current option. So my questions is, do i need to calibrate fuel gauge in charge as well as discharge mode both?

Thanks,

Parth

Hello Parth,

1- The test cases are for the customer to decide based on customer application. How are you going to be using your battery cell? Under what scenarios is the cell going to work? Is it at room temp? Is it at low temp? how big is the load in your application? This is what I mean by testing the gauge to see how it responds to your application. 

Short answer: The test is to charge and discharge the cell and monitor RSOC. If there are not jumps and everything works fine, then you can move forward with mass production.

2- Before saving the Golden File, set Update Status to 02 to disable IT gauging, set the Qmax Cycle Count to 0, and set the Cycle Count to 0. For a pack side gauge, this ensures that when the pack is assembled and gauging and lifetime data is enabled, the cycle counts start from 0.

You should only procedure to golden file stage, once you have tested the gauge and you know 100% it works well.

3- For Qmax Cell 1, you should use the value your gauge learned "3088 mAh". The gauge continuously learns Qmax and updates this value, but this is accurate enough as I don't expect Qmax to change much from this value, unless something abnormal happens. For instance, a huge jump in temperature,  a big load, etc.

4- This is expected and normal. In your current settings, you have defined Term Voltage as 3V and Term Voltage  Delta at 200mV

EDV-End of Discharge = Term Voltage + Term Voltage Delta = 3.2 V

So, you should be expecting the gauge to report 0% for RM about 3.2 V

Parth Mehta said:

Even though RM = 0, i am getting voltage = 3.15V and fuel gauge is ON and i am able to see fuel gauge data in bqstudio, but is this possible?

RM = 0 only means you have discharged the battery to the point where you want the discharge to stop. Term V is the point you stablish as 0% RM.

Batteries in general cannot be discharged to 0 V as this is unsafe region, if you look at your cell manufacturer datasheet, it should state the lowest voltage you can discharge the cell.

5- FCC value should maintain nearly constant unless there's big load or temperature changes.

RM capacity should only be equal or close to FCC when battery is fully charged. RM capacity goes down as you discharge and goes up as you charge. RM capacity is proportional to RSOC. 

RSOC = RM / FCC

Please send a charge and discharge to further analyze this.

Parth Mehta said:

So my questions is, do i need to calibrate fuel gauge in charge as well as discharge mode both?

No need to do it for both. 

Please see below instructions, and ignore bq40z50EVM. Same principle for the bq27z561.



Regards,
Jose Couso

Hello,

This is much helpful. Now i have understood pretty much everything. I will generate golden image after doing full testing as you suggested and will update you the status.

Just one question : Can i update term voltage and term delta voltage after completing learning cycle?

Thanks,

Parth

Hello,

Can we update term voltage and term delta voltage after completing learning cycle?

Thanks,

Parth

Hello Parth,

Term Voltage and any Data Flash parameter can be updated anytime during testing. To what value are you looking to update Term Voltage. It is recommended not to set Term Voltage in the flat region of the cell OCV.

Regards,
Jose Couso

Hello, 

I want to set remaining battery percentage 0% at 2900mV so in order to do that i am thinking to update term voltage = 2700mV and term delta voltage = 200mV so that at 2700+200 = 2900mV i can get remaining battery percentage = 0%

Thanks, 

Parth

Hello Parth,

I am glad learning cycle worked fine with ti EVM board. 

Can you open a new thread with the new issue you are experiencing? It is better to close this thread to further help customer with learning cycles in the future.

I will close this thread.

Regards,
Jose Couso

Yes, sure

Thank you so much for your help to implement learning cycle successfully.

Thanks,

Parth