We used the Xavier NX 8GB module on the baseboard of the Xavier NX devkit. When I used the I2C1 in the picture below to connect to our test chip (chip model is: 24C02), using “i2cdetect-y-r” can normally read the address of the I2C chip. However, when using the I2C0 shown in the following figure, the address of the I2C chip cannot be read, and the process takes about half an hour.
We used Jetpack4.6.1, L4T32.7.1
When I replaced a new module, I found that there was no such problem. Could you please tell me how to operate this module suspected of I2C failure so that I2C can communicate normally?
This is the I2C chip that we tested. And the related wiring mode.
Has anyone experienced the same problem as me? It looks very similar to the Xavier NX module failure. Are there any other verification methods?
please see-also Topic 173325 to check you’re using correct i2c bus.
As I said in another post, we have a batch of NX modules that are used on devkit boards, and only one I2C bus can’t get the device, so we can be sure that we are using the correct bus.
Xavier nx flash error: return -8 - Jetson & Embedded Systems / Jetson Xavier NX - NVIDIA Developer Forums
please capture kernel logs when you trying to communicate with I2C0.
I am very sorry for not replying you in time. Attached is the relevant log.
These are the screenshots and logs that pass the test, and the screenshots and logs that are tested abnormally.
i2c_detect_1_pass.log (68.0 KB)
do you have other modules for testing expansion header with I2C0_SDA/I2C0_SCL?
Yes, I have tested with other modules, and “i2c_detect_1_pass.log” in the above log is the normal log, and “i2c_delete_1_pass.png” is the normal module test and was able to get the address 0x50.
To prevent confusion, I uploaded the test screenshots and test logs of normal modules from new.
it should be module issue… as I2C0 is ID_I2C from module directly.
That is to say, I need to apply for RMA for this module, right?
do you have two of more Xavier NX to have cross validation.
Yes, we have used about 50 NX in this project, and this is the only part that has such a situation at present.
you may contact with the NVIDIA Customer Care team for the RMA process.