Issue in the I2C signals in Stm32H745XIH6 microcontroller

[Shivani]

Hello,

1. We are using STM32H745XIH6 Microcontroller for our designed custom boards. We have designed two boards named controller card (i.e. with the mentioned microcontroller mounted) and the other Power supply card.
2. The I2C signals are generating from the controller card to the power supply card through the connectors.
3. There is no problem with the I2C signals in the Controller card and also in the connector side of the power supply card, when two cards are stacked together.
4. the I2C signals are unable to generate from the connector side to the other side in the power supply card i.e a level translator is mounted taking the I2C Signals from the connector in input side and generating the same I2C output signals so that the clock and data signals will go to the respective IC's in the Power supply Card(attached a Png image of our circuit for reference). The speed that I have configured for checking I2C is 100KHz.
5. The provisions are made as per the point 3 & 4, but we are unable to get the signals in the other side of the level translator. 
6. What might be the issue?
7. If I am configuring the same pins as GPIO Outputs and making them toggle every 100msec, the signals are passing from one side to other side in the level translator.

 

[TDK]

What leads you to believe you are "unable to get the signals in the other side of the level translator" ? Specifically since setting the pins as GPIO works. Certainly that shows the level shifter is working. And if I2C signals are correct right out the STM32, it's certainly not an STM32 issue.

Show some more evidence. Probably it's a test setup issue, or due to bad assumptions. An analog trace of I2C before/after the shifter would be informative. Perhaps it's now powered.

[MM..1]

Primary why level shifter with same Vcc values 3,3V ? Not required.

Secondary is GNDs board 1 connected to board 2 GND ?

[Shivani]

Hi,

Yes, the Level Shifter is not required. Instead of Level shifter we have mounted the R101 and R102 resistors for the data and clock pins. But when testing the signals after mounting the above resistors we are not getting the signals both input and output side (2,3 pins as input & 6,7 as output in U6 of the attached file) as expected. It is continuously high when we are checking the clock and data pins in the oscilloscope. If the resistors were not mounted, then the input side we can monitor the clock and data signals correctly. What might be the Issue?

Please refer to the file that I have attached.

And the GNDs board 1 connected to board 2 GND.

Thank you...

[MM..1]

What? Many, but start point is when level shifter is removed, remove too R21 22 25 around it. Check soldering and orientation U4 on pcb. U4 is mirrored on your image is realy right wired? How value resistors above?