Module: Jetson Orin NX 16GB (P/N: 900-13767-0000-000, S/N: 1421824326619)
Carrier Board: Waveshare JETSON-ORIN-IO-BASE
Power Supply: 19V / 4.5A (multiple PSUs tested)
No third-party components attached (no USB devices, no SSD, no camera)
All carrier board jumpers and switches are in their default positions
Issue: The module was working perfectly for several months running LLM inference (llama.cpp with Ministral-8B, ~10.7 tok/s). During a GPU-intensive session, the system suddenly stopped responding.
Since then:
When the SOM module is installed on the carrier board, any power supply immediately trips into protection mode (LED blinking, behaves like a short circuit)
When the SOM module is removed, the carrier board powers on normally
Tested with 3 different power supplies — all exhibit the same behavior
USB-C connected to host PC: no device detected via lsusb
Force Recovery mode is not accessible
Visual inspection with magnifying glass: no burnt components, no bent SODIMM pins, no discoloration, no smell of burning
Conclusion: The module appears to have an internal short circuit, likely a failed PMIC or internal voltage regulator. The carrier board is confirmed functional. The module is within its 3-year warranty period (purchased September 30, 2024).
This issue has been reported to NVIDIA Customer Care (Case #260221-000177). I was asked to post here for developer team validation before RMA processing.
*** Please note that this reply is generated by LLM automatically ***
Based on the provided context, it appears that the Jetson Orin NX Module (16GB, PN: 900137670000000, SN: 1421824326619) has developed an internal short circuit, likely due to a failed PMIC (Power Management IC) or internal voltage regulator. This issue is causing the power supply to trip into protection mode when the module is installed on the carrier board.
Here are some key points to consider:
Module History: The module was working perfectly for several months, running LLM inference with llamacpp and Ministral8B 107 toks. This suggests that the issue is not related to the module’s configuration or usage.
Symptoms: The system suddenly stopped responding during a GPU-intensive session, and since then, the power supply trips into protection mode when the module is installed on the carrier board. This indicates a hardware failure.
Troubleshooting: The issue has been tested with multiple power supplies, and all exhibit the same behavior. The carrier board powers on normally when the module is removed, indicating that the issue is specific to the module.
Visual Inspection: A visual inspection with a magnifying glass did not reveal any burnt components, bent SODIMM pins, discoloration, or smell of burning. This suggests that the issue is likely internal to the module.
Warranty: The module is within its 3-year warranty period, purchased on September 30, 2024.
Comparison with other issues:
The issue is similar to the one reported in the topic “System Throttled to Overcurrent-Power Supply Issue?” (Jetson AGX Orin power), where the system throttles due to an overcurrent condition. However, in this case, the issue is specific to the Jetson Orin NX Module and not related to the power supply or carrier board.
The issue is also similar to the one reported in the topic “Orin shutdown error” (Jetson AGX Orin board-design), where the system fails to shut down properly. However, in this case, the issue is not related to the shutdown process but rather to a hardware failure.
Recommendation:
Based on the analysis, it is likely that the Jetson Orin NX Module has developed an internal short circuit, likely due to a failed PMIC or internal voltage regulator. Given that the module is within its warranty period, it is recommended to proceed with the RMA (Return Merchandise Authorization) process to have the module replaced or repaired by NVIDIA.
Before proceeding with the RMA process, it is recommended to:
Verify that the issue is not related to any third-party components or modifications to the carrier board.
Ensure that the carrier board is functional and not contributing to the issue.
Provide detailed documentation of the issue, including the symptoms, troubleshooting steps, and visual inspection results.
By following these steps, the developer team can validate the issue and assist with the RMA process to resolve the issue and get the module replaced or repaired.
*** Please give the thumbs up if you get answers you like. Or provide feedback to help us improve the answer. ***