STM32G474 strange I2C behavior

[kf_SEA]

Hi,

I hope somebody can explain the following behavior:

I use a NUCLEO-G474RE board for I2C testing. I have an external I2C slave which provides the required pull-ups for SDA/SCL when it's powered.

Now my STM32 started with connected but not powered I2C slave and tried to communicate, which was not possible due to the missing pull-ups and because of missing slave supply. Then I powered the slave but SDA/SCL still show LOW level, measured with a scope. When disconnecting SDA/SCL I see HIGH levels at the slave side, what tells me that the pull-ups do what they should do and that the STM32 pulls the signals low permanently.

When checking the I2C state it shows READY and the last error was a TIMEOUT.

I use the HAL for the I2C interface.

Can anybody tell me what happens with the pin states of my I2C interface?

Thanks and best regards

[Imen.D]

Hello @kf_SEA ,

Make sure that I2C clock speed is correctly configured and check the timing settings.

Check also the I2C pull-up resistors and the speed mode to use within the right specifications (Standard, Fast, or Fast Plus).

This FAQ describes few tips related to I2C peripheral on STM32 devices: STM32 I2C does not work - STMicroelectronics Community

[kf_SEA]

Hi @Imen.D ,

thanks for your respronse.

I don't have problems with I2C in general, but only if I run my application without I2C pull-ups from the beginning.

It seems to be the a problem when trying to communicate without them.

When I enable the pull-ups later (yes, they are switchable in my external hardware), the STM pulls SDA and SCL low continously until a reset.

When running my application with pull-ups from the beginning there is not a problem at all.

Can you tell me, why there is a state after i2c timeout that blocks the bus 'forever'?

Thanks and regards