Jetson Orin NX – LPDDR5 Case Temperature Over 85°C: Expected Behavior or Failure Mode?

Robotics & Edge Computing Jetson Systems Jetson Orin NX
, ,
1

Hi Team,

We’re investigating a shutdown issue on Jetson Orin NX during thermal stress testing and would appreciate insight into potential failure modes related to LPDDR5 memory.

🔬 Test Scenario:

  • External lab test conducted at ambient temperature: 60 °C
  • Full CPU and GPU stress test
  • When CPU/GPU temperatures reached ~85 °C, the NX module abruptly shut down

We suspect the LPDDR5 RAM case temperature may have exceeded 85 °C, which might have caused the system to trip.

❓ Questions:

  1. What failure mode or protective behavior is expected when LPDDR5 temperature crosses its thermal limit? (e.g., shutdown, throttle, crash?)
  2. Are there internal sensors monitoring LPDDR5 temps, or does shutdown happen at the SoM level?
  3. Has anyone in the community experienced shutdowns under high ambient + stress that could be attributed to LPDDR5 overheating?

We’re trying to determine whether this was an expected protective shutdown (e.g., thermal safeguard) or indicative of a design/thermal margin issue on our carrier or enclosure side.

Any insights or similar experiences would be highly appreciated.

Thanks,
CEZWA

2

*** Please note that this reply is generated by LLM automatically ***

Thank you for reaching out to the community for insight into the thermal stress testing issue you’re experiencing with the Jetson Orin NX.

Expected Behavior when LPDDR5 Temperature Crosses its Thermal Limit

According to the Thermal Specifications document, when the internal temperature of the Jetson Orin NX module reaches an unsafe temperature, the hardware is designed to initiate a shutdown. This is a protective measure to prevent damage to the module. The expected behavior in this scenario is a thermal shutdown, which is a hardware shutdown triggered by on-chip and on-board sensors when all other cooling strategies have failed and software shutdown has failed to occur.

Monitoring LPDDR5 Temperatures

The Jetson Orin NX module has built-in temperature monitoring capabilities. The module can report and take action when internal temperature exceeds limits. However, it’s not clear if the LPDDR5 temperature is specifically monitored or if the shutdown happens at the SoM (System-on-Module) level. It’s possible that the LPDDR5 temperature is not directly monitored, but rather the overall module temperature is monitored, which includes the LPDDR5 component.

Community Experiences with LPDDR5 Overheating

There are a few threads in the community forum that mention thermal-related issues with the Jetson Orin NX, but none specifically mention LPDDR5 overheating. However, one thread mentions a shutdown issue with the Jetson Orin NX 16G module, which could be related to thermal stress.

Determining if the Shutdown is an Expected Protective Shutdown or a Design/Thermal Margin Issue

To determine if the shutdown is an expected protective shutdown or a design/thermal margin issue, you may want to investigate the following:

  1. Check the thermal design of your carrier or enclosure to ensure it meets the thermal specifications of the Jetson Orin NX module.
  2. Verify that the module is properly cooled and that the cooling system is functioning correctly.
  3. Review the power and thermal management settings on your system to ensure they are configured correctly.
  4. Check the system logs to see if there are any error messages or warnings related to thermal stress or shutdown.

If you’ve checked all of these and the issue persists, it’s possible that there’s a design or thermal margin issue with your carrier or enclosure.

Comparison with Other Jetson Modules

The Jetson Orin NX module has a similar thermal shutdown mechanism to other Jetson modules, such as the Jetson AGX Xavier. However, the specific thermal specifications and monitoring capabilities may differ between modules.

In summary, the expected behavior when LPDDR5 temperature crosses its thermal limit is a thermal shutdown, which is a protective measure to prevent damage to the module. While there are some community threads that mention thermal-related issues with the Jetson Orin NX, none specifically mention LPDDR5 overheating. To determine if the shutdown is an expected protective shutdown or a design/thermal margin issue, you should investigate the thermal design of your carrier or enclosure, verify proper cooling, review power and thermal management settings, and check system logs.

*** Please give the thumbs up if you get answers you like. Or provide feedback to help us improve the answer. ***

4

Please refer to the Orin NX/Nano Thermal design Guide available in download center. Jetson Download Center | NVIDIA Developer

Max Tcase temperature for LPDDR5 is 85C, so anything above this means the thermal solution/heasink design may not be sufficient. Please refer to Table 3-1 to 3-7 as appropriate for your module and operating mode.

image
image1206×1712 209 KB

5

@sgursal,
Thanks, and appreciate your response.

We obtained the 85°C temperature limit from the thermal design guidelines. However, we’re more interested in understanding the potential failure modes when this temperature is exceeded.

According to the available data, the LPDDR5 is not in direct contact with the heatsink. This raises concerns, especially in high ambient temperature conditions — even if we’re able to maintain the CPU and GPU temperatures below 85°C (with thermal protection triggering at 105°C), the LPDDR5 may still be at risk.

As per forum information, LPDDR5 does not have a dedicated thermal sensor to trigger thermal protection beyond 85°C. In our case, the SoC unexpectedly tripped without any error log, which makes it more critical for us to understand the failure mode involved.

Given the limited thermal management and lack of protection for LPDDR5, we are puzzled. If LPDDR5 cannot reliably operate beyond 85°C, then why is the SoC thermal limit set as high as 105°C?

Any insights or clarification on this would be greatly appreciated.

6

Are you able to get UART logs when the shutdown happens? The shutdown temperature for Orin SOC is based on its specifications. Each component could have a different Tj or case temperature. Have you done thermal characterization using thermocouples attached with your thermal solution to check LPDDR5 temperature when operating at 60C? If it is exceeding 85C, then you may have to check if adding some thermal pad/gap filler helps to bring it in operating range.