Nano shutting down under load (overheating)

HI all. Sorry if this is a FAQ. Just joined. Some colleagues and I got some Jetson Nano kits and are starting to use them, leveraging azure iot edge

… they have a fan, mounted on the heat sink, and the fan is set to run wide-open. However, consistently, when we put the Nano under load running even just one stream (a YOLO model in a docker container in azure iot edge), the device shuts itself down after about a minute or so. We have them powered through the barrel connector. I’ve tried removing the cover as well to see if there is an airflow restriction. Nothing helps. Also, I was able to run mine successfully for a while a few weeks ago, but now it consistently shuts down running the exact same workload. I can’t claim 100% that nothing changed, but I certainly don’t remember anything changing…

Any ideas of why it would overheat under such low load? Any best practice/tips?

thx,
–Steve

Hi,
We could check whether it is causing by the thermal condition or not by running the tegrastats on serial console.
It would show the temperature before it goes down. Could you paste the result?

Hi Steve,
what power source do you use?
5V⎓4A?

@WayneWWW below were the last few lines of tegrastats right before it shutdown. It ran for about 30 seconds after boot before dying…

Windows Terminal

RAM 1856/3956MB (lfb 153x4MB) SWAP 0/1978MB (cached 0MB) CPU [15%@1479,30%@1479,21%@1479,44%@1479] EMC_FREQ 0% GR3D_FREQ 0% PLL@22.5C CPU@24C iwlwifi@33C PMIC@100C GPU@22.5C AO@31C thermal@23C POM_5V_IN 3364/3
939 POM_5V_GPU 124/123 POM_5V_CPU 1368/1830
RAM 1858/3956MB (lfb 153x4MB) SWAP 0/1978MB (cached 0MB) CPU [33%@1479,17%@1479,21%@1479,45%@1479] EMC_FREQ 0% GR3D_FREQ 0% PLL@23C CPU@24C iwlwifi@34C PMIC@100C GPU@23C AO@31.5C thermal@23C POM_5V_IN 3679/391
3 POM_5V_GPU 124/123 POM_5V_CPU 1612/1808
RAM 1907/3956MB (lfb 140x4MB) SWAP 0/1978MB (cached 0MB) CPU [76%@1479,43%@1479,25%@1479,31%@1479] EMC_FREQ 0% GR3D_FREQ 0% PLL@23.5C CPU@24.5C iwlwifi@34C PMIC@100C GPU@23C AO@32C thermal@23.75C POM_5V_IN 419
7/3939 POM_5V_GPU 123/123 POM_5V_CPU 1890/1815

Andrey1984 - the forum won’t let me @ you, but below is the power supply we are using…

I’m using this 5A barrel connected power supply that came with the kit

Hi Stevebus,

Did you try any other power adapter/supply?

It does not look like a overheat because all the tegrastats shows only 20~35C here.

Such problem may come from the adapter.

Thanks guys - I’ll try another adapter. Odd that this would happen to multiple people on the team, all using the same adapter. I’ll see if I can dig up another 5A barrel adapter from my stash and try it.

One thing to think about is that the power jack is meant for a 5.5mm OD, 2.1mm ID positive tip connector. Looking at your power supply, it appears to terminate in a 5.5mm OD, 2.5 mm connector. Maybe some (very small) electrical gremlins sneak in the gap and gleefully mess things up.