BQ27426: BQ27426 I2C has no ACK

Hello,

Please see below



Regards,
Jose Couso

Is it necessary for normal operation?we used with the pull up always on.Is there any other design request?

Hello,

For normal operation, it is necessary to use the recommend 10k-Ω pull up resistor. 

The I2C COMM lines are open drain pins which means, when the device is off, these pins are floating (hi-z). 

Regards,
Jose Couso

our device power on process is: BUB on-fuel gauge on-system power on-pull up 3.3 on-MCU on-I2C on-I2C communication to fuel gauge.

1.Dose the 1M ohm pull down resistor is necessary for the process?

2.when the device is off,I2C is floating,does it can cause I2C work abnormal?why?I think the I2C is not work and is low level.it should not cause fuel gauge abnormal.

Hello,

When the device is OFF, the I2C lines are floating; hence, it will not be stable. For this, we recommend to use a 1 M-Ω pulldown resistor when the pull-ups are disconnected. 

 As long as the I2C lines have a pull up to an external 3.3V at all times during normal operation, the 1 M-Ω pulldown resistor is not necessary. Hi-Z means the voltage is not stable, a pull up during normal operation fixes this.

Regards,
Jose Couso

It is BQ27426 address read I2C logic,why it has abnormal pulse and pull low SCL action?is it normal?

Jose is out of office and will reply when he is back on Monday

Hello,

Can you outline the abnormal pulse? It looks like a normal sequence: write address - send cmd - repeated start - read address - get data.

Sincerely,

Wyatt Keller

the abnormal pulse markd as red,the SCL is pulled low for while is also abnormal.it seems BQ27426 pull it low but we do not know why.

Hello,

It looks to me like the I2C is clock stretching. Our gauges will hold down the bus for a determined specified time in datasheet when it is busy. The MCU must support clock stretching. See below.



Section 6.2 Clock Stretching reference.

Regards,
Jose Couso

Thanks for reply,can you explain the first pulse marked as red.it seems abnormal.

Hello,

It seems like the I2C stop condition. When SCL goes high before SDA. The device is ACK-ing normally. I would not be concerned about this. 

Regards,
Jose Couso