BQ27426: FuelGaugeIC's Filtered FCC not matched during battery aging

#1: If you want to keep the learned info for a specific cell, you have to save the Ra (0x59), State (0x52) and Chemistry Info (0x6D) classes before you power down the gauge and remove the cell and then load the GoldenImage, followed by restoring these three classes after you powered up the gauge again with this specific cell. These classes are the only ones that the gauge will update during regular use.

#2: FCC (filtered and true) will change with load, temperature, Qmax and Ra. It's not possible to pinpoint the exact reason without a full discharge log for 1. a new cell and 2. the cell under test with the low FCC. This log must include time (not more than 5s interval, ideally 1s interval), current, voltage, temperature, RM true, RM filtered, FCC true, FCC filtered. Each log must start with a fully charged, relaxed cell (gauge must take OCV and update DOD0). The log must be continuous from full to cell voltage < TerminateVoltage.

Hello Dominik,

Thanks for you reply.

#2,For the convenience of analyzing the problem, in addition to the registers mentioned above, which other registers are required to be provided?

For analysis, I recommend logging time (not as a time stamp but in seconds from start), current, voltage, temperature, true and filtered RM, true and filtered FCC, DOD0, passed charge since DOD0, DODatEOC. Once every second, ideally. Not longer than once every 5 seconds.

Hello Dominik，

Thanks for you reply。

I described our validation program in another Case.

Our test conditions do not include full charge and discharge.

Based on our validation scenario, isn't the same register information needed as described above?

Yes, same information is needed.

hello Dominik,

Based on our current battery aging verification, FCC values for FuelGaugeIC have not changed, So we want to know:

1, What are the possible influencing factors that cause the FCC value to stay flat?

2, What validation can be used to rule out possible influencing factors?

3,What should FuelGauge's FCC curve for correctly tracking battery capacity look like?

4,What are the conditions for FCC to track battery capacity correctly?

5,What should be the ideal action and value respectively for the register value to be fetched?

Dominik Hartl11 said:

For analysis, I recommend logging time (not as a time stamp but in seconds from start), current, voltage, temperature, true and filtered RM, true and filtered FCC, DOD0, passed charge since DOD0, DODatEOC. Once every second, ideally. Not longer than once every 5 seconds.

1. The gauge hasn't learned Qmax and/or Ra and/or load and/or charge termination. You will have to let the gauge learn these values before results will change.

2. I am not sure what you mean by this question.

3. FCC is a function of Qmax, Ra, temperature, load, charge termination, end of discharge voltage, load spikes. If these parameters are constant, FCC will be constant. FCC will fall with falling temperatures, higher loads, higher Ra, lower Qmax, lower charge termination voltage, higher taper current, high voltage spikes, higher end of discharge voltage. Because some of theses parameters can change a lot during a discharge (e.g. temperature, load), FCC can be really dynamic (not constant).

4. There are too many conditions to list in an E2E thread.

5. I am not sure I understand the question. The parameters that I listed don't have ideal values, they are just measurements and gauging results that depend on these measurements.

Hello Dominik，

Based on your tips, we added the necessary logs to restart the battery aging test.

Battery design capacity: 1230mAH

Reserve capacity: 350mAh

Tested battery FCC is 1239mAh (after measurement with special instrument)

Running time：26hours

Accelerated aging，FuelgaugeIC FCC is 718mAh.

Natural aging，FuelgaugeIC FCC is 744mAh.

Please help to analyze and provide the basis according to the logs.

1, Accelerated aging. why the result predicted by FuelGaugeIC is not consistent with the real battery FCC?

2, Natural aging. why the result predicted by FuelGaugeIC is not consistent with the real battery FCC?

There are two types of tests: accelerated aging test and natural aging test.
Accelerated aging test conditions:
Test temperature: 70 ℃
Equipment quantity: 3 units
Test steps:
Step 1. Charge to SOC=85%
Step 2, discharge to SOC=51%
Step 3, loop sequentially Step 1->Step 2

Name of attachment:：Accelerated aging Test log_0126.zipAccelerated aging Test log_0126.zip

Natural aging test conditions:
Test temperature: 70 ℃
Equipment quantity: 3 units
Step 1. Charge to SOC=95%
Step 2, Battery Relax, wait for the battery to self lose to SOC=91% (battery self loss current below 1mA)
Step 3, loop sequentially Step 1->Step 2

Name of attachment:Natural aging Test log_0126.zipNatural aging Test log_0126.zip

You'll have to make sure that the gauge learns Qmax and Ra, otherwise it will not track the expected FCC changes.

hello Dominik,

We use new batteries, loaded FS files made based on new batteries, Qmax and RA are already updated.

However, the FCC values predicted by FuelgaugeIC are quite different from the real FCC of the new battery.

And our previous accelerated test, which was run for a month at 70 Celsius, the FCC value was also maintained at 718mAh.

we desperately need to know the reason for this, please help to analyze the logs provided and advise the reason？

You need to make sure that Qmax updates. The main rules are:

1. Relax until voltage is stable at a high or low DOD (>0.9 or <0.1)

2. Pass enough charge between relaxations (at least 37% of Qmax)

3. Make sure that temperature is between 10deg.C and 40deg.C

4. This has to be within a reasonable time (e.g. not a whole day of a single discharge with a C/24 current).

Once the gauge updates Qmax, then you have to make sure that Ra updates. The main rules are:

1. Qmax is correct (the gauge will disable resistance updates if Qmax is grossly incorrect)

2. Current is >C/10

3. Discharge is long enough for DOD to change by approx. 0.1

You will see FCC changes if these conditions are met.

Hello Dominik，

Thanks for your reply, but you didn't reply to my question.

You replied with the Qmax and Ra update rules, but we are using a new battery and the FS file is also made based on the new battery with reserve capacity: 350mAh.

So my understanding is: the new battery doesn't need Qmax and Ra table update when performing charging and discharging, and True Fcc won't change significantly.

1, The new battery loads the FS file, and the Ture Fcc will not change drastically during the execution of charging\discharging. Is the understanding correct?

2, If NO, what is the reason? Please tell me the parameter(ex:Present\Current\passedcharge etc) that causes the Ture Fcc to change drastically?

According to the chart belowNatural aging, True FCC changed from 817mAh to 654mAh at the moment of starting charging of the new battery and remained the same.

According to the following figure of accelerated aging, the True FCC changes from 841mAh to 674mAh when the new battery starts charging, and remains the same all the time.

3, through the log <Natural aging Test log.zip>, can you analyse what factors cause True Fcc to change from 817mAh to 654mAh?

4, through the log <accelerated aging Test log.zip>, can you analyse what factors caused True Fcc to change from 841mAh to 674mAh?

Accelerated aging Test log.zipNatural aging Test log.zip

#1: The gauge uses the data from the configuration to calculate gauging results. So no mattery if the cell is new or old or a different chemistry or a power supply or anything that can power the gauge, it will simply calculate capacity based on the configuration.

Which means if you attach a new cell or an old cell, it will start out with the results for a new cell and then, depending on how old the cell actually is, it will take N update cycles for Qmax and Ra for the gauging results to become accurate. N can't be easily calculated. There is no simple formula.

#2: Charge termination voltage, temperature, end of discharge (Terminate Voltage), Delta Voltage in class State, Ra, Qmax are the main parameters that affect FCC.

#3: The Accelerated Aging Test shows that FCC changes with DODatEOC, which means that the gauge detected charge termination (taper current and voltage at charge termination), which adjusts DODatEOC downward and FCC upward. The Natural Aging Test shows this also but not as drastic. So the main problem here is that the Golden Image doesn't match your test when it comes to charge termination. Make sure that V at Chg Term is set to the actual value when the gauge detects charge termination and FCC will be more stable.

hello dominik,

1. The following descriptions are our two verification methods to verify that the FCC can be matched correctly during battery aging. Our two methods, can we make the FCC update? What is the reason?

Battery design capacity: 1230mAH

Reserve capacity: 350mAh

Natural aging test conditions:
Test temperature: 70 ℃
Equipment quantity: 3 units
Step 1. Charge to SOC=95%
Step 2, Battery Relax, wait for the battery to self lose to SOC=91% (battery self loss current below 1mA)
Step 3, loop sequentially Step 1->Step 2

Accelerated aging test conditions:
Test temperature: 70 ℃
Equipment quantity: 3 units
Test steps:
Step 1. Charge to SOC=85%
Step 2, discharge to SOC=51%
Step 3, loop sequentially Step 1->Step 2

2、our application scenario：the battery is in Relax state for a long time, and the battery will be recharged only after the battery self-depletion (current <1mA) reaches a certain value.

FuelGaugeIC's Sleep mode has a current of 10mA. i.e. the process of battery self-depletion is not sensed by FuelgaugeIC.

Recharge interval:RM<278mAh to recharge; RM>318mAh to stop recharging and keep Relax.

① Based on our application, what parameters will affect the FCC update?

② How can we manipulate or modify the parameters so that the FCC can track the battery capacity?

Attached is our FS file.

Dominik Hartl11 said:

#2: Charge termination voltage, temperature, end of discharge (Terminate Voltage), Delta Voltage in class State, Ra, Qmax are the main parameters that affect FCC.

3, the V at Chg Term was verified: modify the FS file V at chg Term (4252mV --> 4400mV), FS file is loaded, when charging is performed, the TRUE FCC reduction value is changed from 166mAh to 5mAh, the effect is very obvious. This is equivalent to solving the problem of instantaneous FCC drop during the first charge. Thanks a lot.
We would like to know:

① What is the role of V at Chg Term? Does this parameter belong to battery parameter or application parameter?

② V at chg Term is a Chem Data data, why is it 150mV less than the rated voltage of 4.4V of the battery and what is the basis for 150mV?

③ Can V at chg Term be modified manually? What is the range of modification in our application?

④ What are the parameters associated with V at chg Term? After modifying V at chg Term, which parameters still need to be modified synchronously? Is it necessary to recreate the FS file?

⑤ According to the logs, what parameters other than the V at chg Term parameter caused the FCC to not match the battery capacity?

Dominik Hartl11 said:

#3: The Accelerated Aging Test shows that FCC changes with DODatEOC, which means that the gauge detected charge termination (taper current and voltage at charge termination), which adjusts DODatEOC downward and FCC upward. The Natural Aging Test shows this also but not as drastic. So the main problem here is that the Golden Image doesn't match your test when it comes to charge termination. Make sure that V at Chg Term is set to the actual value when the gauge detects charge termination and FCC will be more stable.

2627.0418_New_Bat_FS_withChemID_0419.gm.zip

#2: The gauge will still sense the self-discharge because the gauge uses OCV (which will drop) to update DOD. The gauge will not learn anything in this use case because the passed charge is too small for a Qmax update and the current too small for Ra update. You'd have to periodically discharge at least 37% of Qmax to get a Qmax update.

#3: V at Chg Term is used by the gauge to calculate the OCV for charge termination (=V at Chg Term + R * Taper Current) which is then used to calculate DOD (DODatEOC), which is then the starting point for FCC predictions (from DODatEOC to estimated cell voltage < Terminate Voltage + Delta Voltage). That's why it is so important. V at Chg Term in the default settings is just a typical number, which may or may not apply to a specific application. It must be set by the customer or better, let the gauge measures it automatically with a learning cycle. You have to make sure that the Taper Rate is correct for your application. See https://www.ti.com/lit/pdf/sluubn3