BQ27421-G1: Fuel Gauge resetting to default settings

Hello Steve,

Check possibility of power loss/glitch in supply to the gauge.

Monitor the ITPOR bit in Flags (section 4.4 in TRM ) to know if the gauge thinks there was a power reset.

Regards

After checking for ITPOR, it took longer, but eventually the return from 0x3E (designCapacity) reports the default of 1000 mAh and it never sets ITPOR.

Additionally, the SoC is very erratic - it will jump from 100 to 0 to 100 and then suddently set itself to some value in between and count up or down from there (depending on whether we are charging or discharging at the time).

Hello Steve,

I would recommend checking if the chemistry of this cell is a good match to the gauge application and check all settings since SOC is erratic.

For example check Load mode (Reference section 6.4.2.3.6) setting

If you are using bq27421-G1B, your battery curves should match CHEM_ID 0x0312

Regards

I have checked the chemid. This IC part was chosen by our battery supplier based on the battery cell's chem_id. 

The biggest problem is that the IC seems to forget it's design capacity and then the SOC declines at a must higher rate that I would expect based on the average current measurement.  I monitor the design capacity reported via register 0x3C, and I can see it change from 3000 (which is what I program it to) to 1000 (the default capacity).  I am also monitoring for ITPOR and I never see it set.  After 5 hours of running the device with between 100 and 120 mA of discharge, I am seeing 64% SOC and I would expect to see something closer to 85-90%.

I am afraid that if I reprogram the design capacity, it will clear all of the learning that has been done.  Is there a way to update it "on-the-fly" without reseting the SOC calculator?

I am wondering if you are seeing Watchdog Reset

Here is the thread with similar issue https://e2e.ti.com/support/power-management/f/196/t/310543

Thanks.  Let me examine the timing of my communication with the IC.  I will let you know.

I checked the timing and it still isn't working.  I added 1 msec delays between reading different registers.  I am currently reading voltage, current, temperature, state of charge, state of health, design capacity, remaning capacity, full charge capacity and flags every second.  As I said, I am also inserting 1 msec delays between each reading.  In addition to this, when I read the flags, I am checking ITPOR to see if a reset occurred and I have never seen that bit get set.  Despite this, after some period of time - usually when charging - the reading of design capacity goes to 1000 (the default).  If I cannot figure this out soon, I am going to need to change the IC to something more predictable.

As my code is running, I am collecting data about the charger and fuel gauge.  The data below contains the following parameters:

time_in_msec,

BQ25895M_VBUS,

BQ25895M_VBAT,

BQ25895M_VSYS,

BQ25895M_CHG_CURR, 

BQ25895M_STATE, 

BQ25895M_FAULT,

BQ27421_des_cap, 

BQ27421_voltage, 

BQ27421_current, 

BQ27421_SOC, 

BQ27421_flags,

CHRG/DISC

You can see in the data below that at time 3765001, the BQ27421's design capacity goes from 3000 to the default of 1000:

00> 03757001,12300,4204,4204,1100,2,0,3000,4347,0300,100,0x0000,CHRG
00> 03758001,12300,4204,4204,1100,2,0,3000,4348,0300,100,0x0000,CHRG
00> 03759001,12300,4204,4204,1100,2,0,3000,4348,0300,100,0x0000,CHRG
00> 03760001,12300,4204,4204,1100,2,0,3000,4348,0300,100,0x0000,CHRG
00> 03761001,12300,4204,4204,1100,2,0,3000,4347,0299,100,0x0000,CHRG
00> 03762001,12300,4204,4204,1100,2,0,3000,4347,0300,100,0x0000,CHRG
00> 03763001,12300,4204,4204,1100,2,0,3000,4348,0300,100,0x0000,CHRG
00> 03764001,12300,4204,4204,1100,2,0,3000,4347,0300,100,0x0000,CHRG
00> 03765001,12300,4204,4204,1100,2,0,1000,3600,0000,000,0x0000,DISC
00> 03766001,12300,4204,4204,1100,2,0,1000,4348,0299,000,0x0000,CHRG
00> 03767001,12300,4204,4204,1100,2,0,1000,4348,0299,100,0x0000,CHRG
00> 03768001,12300,4204,4204,1100,2,0,1000,4347,0299,100,0x0000,CHRG
00> 03769001,12300,4204,4204,1100,2,0,1000,4348,0298,100,0x0000,CHRG
00> 03770001,12300,4204,4204,1100,2,0,1000,4348,0299,100,0x0000,CHRG
00> 03771001,12300,4204,4204,1100,2,0,1000,4348,0298,100,0x0000,CHRG
00> 03772001,12300,4204,4204,1100,2,0,1000,4348,0299,100,0x0000,CHRG
00> 03773001,12300,4204,4204,1100,2,0,1000,4347,0297,100,0x0000,CHRG
00> 03774001,12300,4204,4204,1100,2,0,1000,4348,0298,100,0x0000,CHRG
00> 03775001,12300,4204,4204,1100,2,0,1000,4347,0298,100,0x0000,CHRG
00> 03776001,12300,4204,4204,1100,2,0,1000,4348,0297,100,0x0000,CHRG
00> 03777001,12300,4204,4204,1100,2,0,1000,4348,0298,100,0x0000,CHRG
00> 03778001,12300,4204,4204,1100,2,0,1000,4348,0297,100,0x0000,CHRG
00> 03779001,12300,4204,4204,1100,2,0,1000,4348,0297,100,0x0000,CHRG
00> 03780001,12300,4204,4204,1100,2,0,1000,4348,0297,100,0x0000,CHRG
00> 03781001,12300,4204,4204,1100,2,0,1000,4347,0296,100,0x0000,CHRG
00> 03782001,12300,4204,4204,1100,2,0,1000,4348,0296,100,0x0000,CHRG
00> 03783001,12300,4204,4204,1100,2,0,1000,4347,0298,100,0x0000,CHRG
00> 03784001,12300,4204,4204,1100,2,0,1000,4347,0295,100,0x0000,CHRG

Hello Steve,

00> 03765001,12300,4204,4204,1100,2,0,1000,3600,0000,000,0x0000,DISC

The voltage seems to drop suddenly from 4347 to 3600, it is discharging instead of charging, and SOC goes to 0. There is more happening than just design capacity being reset to original.

You may want to log Control status and check if the WDRESET bit is being set.

On the bq27421-G1B datasheet, page 13:

8.5.4.3 I2C CommandWaitingTime

"For read-write standard command, a minimum of 2 seconds is required to get the result updated. For read-only standard commands, there is no waiting time required, but the host must not issue any standard command more than two times per second. Otherwise, the gauge could result in a reset issue due to the expiration of the watchdog timer."

Regards,

I think that the reading of 3600 mV is an artifact of the fuel gauge restarting, but I will check the value against that measured by the battery charger IC (BQ25895M) which also monitors the battery voltage.  

With respect to the timing, I am reading several registers in a row with 1 millisecond delay between requests.  This update is done once every second.  This compiles with statement you mention, right?

"the host must not issue any standard command more than two times per second" seems to imply that a max of 2 commands can be sent per second.

I would recommend that you add 500+ms between subsequent commands to bq27421 and see if the problem goes away.

I assumed that the 500 msec delay was just between a request for the SAME register.  I find it hard to believe that a sophisticated IC like this can't communication a full set of information within one second of time.

I will try that though and see if it makes a difference.

Can you help me with hardware related issues as well if it turns out there is a dip in the battery voltage?  If that is happening, it is probably coming from the battery charger IC as it is switching from being connected to the changer and being connected to the battery instead.  We are using the BQ25895M.  What size capacitors should there be between the fuel guage/battery and the battery charger IC?  It could be that the sudden demand of current from the battery as the charger is disconnected is causing a large enough dip in the battery voltage to reset the fuel gauge.  Right now, we have two 4.7 uF capacitors between the two ICs and a 1 uF between the fuel gauge and the battery.