BQ35100: Force reset won't clear the accumulatedcapacity buffer

Jorg,

Setting GE pin low or issuing the reset command does not clear the accumulated passed charge. That is an intended feature. If you send the Gauge_stop command any charge accumulated will be saved into flash so the next time you boot it will start there. 

If you want to start at 0 every time, do not issue the gauge stop command and just read accumulated charge right before powering GE down. Likewise you can issue the "New Battery" command which will clear out the accumulated charge in Flash and should start you from 0.

Thanks,

Eric Vos

Hey Eric Vos,

Thank you for the fast response. I tried clearing the accumulated charge using the new battery command, but nothing happens. Same as reading the accumulated charge before powering GE down. Before step 1 I set the gauge in ACC mode and reset the battery. Then my code looks as followed:

///Step 1: Set GE high to power up the BQ35100 device and wait for ALERT to go low due to INITCOMP = 1
enableGauge();
controlStatusRet = 0;
do
{
controlStatusRet = readControlStatus();
} while ( ( ( controlStatusRet >> 7 ) & 0x1 ) == 0 );

///Step 2: Clear ALERT (read BatteryStatus()) and send GAUGE_START()
I2cRead( &I2c, c_BQ35100I2CAddr, BQ35100_READ_BATTERY_STATUS, bufferSOH );
DelayMs( 100 );
controlSubCmd( BQ35100_GAUGE_START );
///Step 3: Read AccumulatedCapacity() for the latest passed discharge data since GAUGE_START().
stateOfHealthACC();

Running code here...
Skip step 4 (gauge stop)

///Step 5: Read final AccumulatedCapacity() value
stateOfHealthACC();
///Step 6: Set GE low to power down the BQ35100 device
disableGauge();

What am I doing different?

Yours sincerely,

Jorg 

Jorg,

I don't understand the issue here. Reading above you want to just start each active gauge section with 0 charge in the accumulated charge register. If so you should do the below

GE_High

Send Gauge Start

(Do stuff)

read 0x02 (Accumulated Charge)

GE_Low 

This process will not commit the charge you counted to flash so the next time you boot you will start with 0. If right before GE_Low you send gauge stop command this will update flash and the next cycle will start with the previous value.

The New Battery command will erase anything stored into flash. 

Please add more details on what you are wanting the system to do. 

Hey Eric Vos,

Every time the device (microcontroller) starts I want to begin with a new battery. After that it should be reading the accumulated capacity only for the active states. So right before the device is in sleep, I set the GE_Low for the gauge. When the device wakes I set the GE_High followed by the gauge start command. Before GE is Low I read the accumulated capacity again and so on. 

The problem is that while starting the device (microcontroller) the new battery command seems to have no effect. Because I have been sending a gauge stop command before (thought I had to send the gauge stop command before setting GE_Low), and that data will not disappear.  

Thank you,

Jorg

Dear Eric Vos,

Onyx told me the accumulated charge should increase. But for some reason unknown my accumulated capacity register begins at about 4.29e9 and then slowly discharges (decreases) every time current is used from the battery. I think this makes my problem more clear too. There are no errors or overflows found in the control register, so what is going wrong?

Greetz,

Jorg

Jorg,

Please send me the full srec you are using. I have verified the functionality on the bench and do not fully understand the issue you are seeing.

the accumulation counter will count up and down depending on the current direction. While primary cells should not see charge current in Accumulation mode the gauge does not care. 

thanks,

Eric Vos

Dear Eric Vos,

The clear battery command works now! (timing issue...),

But after the first discharge the accumulated charge value changes from 0 to 4294967287.

Is it right to take that discharge value of 4294967287 off from 42949667295? (2^32-1)? To see how much charge is used.
It makes sense since a uint32_t value is form 0 to 42949667295 and since discharging means decreasing, the accumulated charge wouldn't be able to be negative so then it decreases from 42949667295 to 0? If this assumption is right the problem is solved.

Thanks for the help! 

Jorg

Dear Eric Vos,

Since I have no permissions to PM, I should answer over here again. I have checked the connections and they are correct. I have done no calibration yet, because there is no example code.  I'm using the typical application of the BQ35100.pdf. The battery is connected between BAT+ and the sense resistor. The microcontroller (load) is connected between BAT+ and PACK-. Since the CHEM_ID doesn't matter in ACC mode, I'm only changing the values of the battery to 17Ah to calculate the percentage left. It is a non rechargeable battery. The config A is changed to ACC mode. 

This is the code for reading the AccumulatedCapacity(): 0x02/0x05 register:

For reading a direct read command I use:

uint16_t readRegister( uint8_t readRegister, uint8_t *bufferRead, uint16_t readRegisterSize )
{
    uint16_t pData[2] = { 0 };
    pData[0] = readRegister & 0xFF;
    pData[1] = ( readRegister >> 8 ) & 0xFF;

    if ( I2cReadBuffer( &I2c, c_BQ35100I2CAddr, (uint16_t) ( pData[0] + pData[1] ), bufferRead,
            readRegisterSize )== FAIL )
    {
        printf( llerror, "readfromDataRegister: reading the register failed" );
        return FAIL;
    }
    DelayMs( 100 );
    return *bufferRead;
}

Then I use the read command in the following code:

//Read the accumulated capacity
uint8_t capacity[4] = { 0 };
uint32_t accumulatedCapacity = 0;

readRegister( BQ35100_READ_ACCUMALETED_CAPACITY, capacity, 4 );
DelayMs( 100 );
accumulatedCapacity = (( ( (uint8_t) capacity[3] << 24 ) + ( (uint8_t) capacity[2] << 16 ) + ( (uint8_t) capacity[1] << 8 ) + capacity[0] ));
printf( lldebug, "Accumulated capacity read: %u uAh", (uint32_t) accumulatedCapacity );

The "funny" thing is that the amount it decreases is pretty accurate. This is exactly the charge used by the BQ35100 + microcontroller. With some coding the IC is still useful. But because it should be increase, I don't trust it...

I have been reading that there are more people with this problem:
https://e2e.ti.com/support/power-management/f/196/t/793885?tisearch=e2e-sitesearch&keymatch=read%252525252520acc%252525252520bq35100
But since Onyx only told him about the CHEM-ID is not needed for ACC mode, this forum is not useful for me. 

Thank you for all the help!,

Jorg