I2C lines short with GND in STM32H745XIH6

[ABora.4]

Hello,

I am using the MCU : STM32H745XIH6 in my currnet product from last couple of the years and it is working fine till.

Recently, we face the issue in I2C line.

I2C3 line connected with,

1. Super capacitor Board through connector(there is I2C IO expander to contol the charging IC)
2. Also sensors are connected through connector

I am not sure but the issue is coming after adding the super capacitor board.

ISSUE : - After some power cycle of the MCU board, I2C3 lines are shorted with the GND so we need to remove the MCU and solder the new and then it is working but after some power cycle of CPU,same issue is observed. So it is not easy to solder the BGA package manually and we lost the MCU also.

As we are planning to purchase the new stock of the MCU as we have used lots of MCU in rework but without solution we are worried to purchase new stock.

Kindly look in to the matter and resolve the issue so we can move to purchase th new stock for the next production batch at earliest.

We can also share you the schematic under NDA.

Regards,

ALpesh

[ONadr.1]

It is likely that the I2C from the additional board remains unde voltage when the MCU is already off and the driver or substrate diode has broken down. A schematic would be needed for a more detailed analysis.

[ABora.4]

Hello,
Thanks for the reply.

In this case, Is the position of pull up resistor matter?

What will be the best position for the pull up resistors when the Mater and slaves are in different two boards and they are connecting with wire harness?

Please find the schematic in one page here,

[ONadr.1]

I would leave the overall connection as it is, but instead of R47, R48  0R I would put resistors of approx. 20-40ohm. These should prevent extreme current spikes while maintaining proper I2C line function.

[ABora.4]

Hello,

Thanks for your suggestion.

I will update the design as per your suggestion and test it.

If you found anything to make change, please update me.

Regards,

Alpesh