Custom Nvidia Jetson Orin Nano Board Entering Coldboot Mode

Issue Overview

Users have reported an issue with a custom-designed board for the Nvidia Jetson Orin Nano (model p3767) entering coldboot mode repeatedly. The problem occurs when the module is installed on the custom board, despite functioning correctly on the development kit. The issue is observed during the boot process, with specific errors related to I2C communication and EEPROM reading.

Key points:

• The module works fine on the development kit
• The custom board keeps entering coldboot mode
• Debugging is being performed via UART2
• Errors are observed in I2C communication and EEPROM reading
• The issue appears to be related to the custom board design or configuration

Possible Causes

1. EEPROM Absence or Malfunction: The error logs suggest issues with reading from an EEPROM device, which may not be present or properly configured on the custom board.

2. I2C Communication Failure: The logs show errors in I2C communication, which could be due to incorrect wiring, faulty components, or misconfiguration.

3. Incompatible Jetpack Version: The initial Jetpack version used may not have been compatible with the custom board setup.

4. Incorrect Boot Configuration: The boot sequence or configuration files may not be properly set up for the custom board.

5. Hardware Design Issues: There might be discrepancies between the development kit and the custom carrier board design that are causing boot problems.

6. Power Sequencing Problems: The power-up sequence circuit mentioned in the original post could be contributing to the boot issues if not properly implemented.

Troubleshooting Steps, Solutions & Fixes

1. Verify Jetpack Version:

   • Check the Jetpack version using the command:

     cat /etc/nv_tegra_release
     

   • Ensure you are using Jetpack 6.0 or later, as this version was mentioned in the discussion.

2. EEPROM Configuration:

   • If your custom board doesn’t have an EEPROM, follow the instructions for EEPROM modifications:
     • Locate the file tegra234-mb2-bct-misc-p3767-0000.dts in the BSP package on your host PC.
     • Modify the file according to the Jetson Orin NX and Nano Series — EEPROM Modifications guide.

3. Reinstall Jetpack:

   • Use a fresh microSD card (32GB or larger) to install Jetpack 6.0.
   • Follow the official NVIDIA documentation for a clean installation.

4. Custom Image Preparation:

   • Refer to the Jetson Module Adaptation and Bring-Up guide for Jetson Orin NX and Nano Series.
   • This guide will help you customize the image for your specific board design.

5. Hardware Verification:

   • Compare your custom carrier board design with the development kit.
   • Check for any discrepancies in power supply, I2C connections, and other critical interfaces.
   • Verify that all required connections for boot are properly implemented.

6. I2C Debugging:

   • Use I2C debugging tools to verify communication between the module and any I2C devices on your custom board.
   • Check for proper pull-up resistors and correct voltage levels on I2C lines.

7. Power Sequencing Analysis:

   • Review and potentially revise the power-up sequence circuit mentioned in the original post.
   • Ensure that power rails come up in the correct order as specified in the Jetson Orin Nano documentation.

8. UART Debug Output:

   • Continue monitoring the UART2 output during boot.
   • Look for any additional error messages or unexpected behavior in the boot sequence.

9. Bootloader Configuration:

   • If you’ve modified the tegra234-mb2-bct-misc-p3767-0000.dts file, ensure the changes are properly incorporated into your boot image.
   • You may need to recompile the bootloader with these changes.

10. Consult NVIDIA Developer Resources:

    • For unresolved issues, consider posting detailed information about your custom board design and boot logs on the NVIDIA Developer Forums.
    • Provide information about any deviations from the reference design to get more targeted assistance.

Remember to document all changes and test results thoroughly during the troubleshooting process. If the issue persists after trying these steps, you may need to conduct a more in-depth analysis of your custom board design or seek direct support from NVIDIA.