Jetson Nano shutoff During High Computational Load

Hello, I’m curious if the only (main) reason the Nano would shut itself off is a power constraint issue? I’ve been playing around with the dusty-nv/jetson-inference and sometimes when I’m running the object detection model with large camera images (1920x1080) the jetson will shut itself off after successfully detecting pedestrians for a few seconds. The setup I’m running is powering the nano through 3 GPIO pins (GND, 5V, 5V) and a bench top power supply (set to 5V, 6A). I can see the output from the benchtop supply does not go above 2.6/2.7A. I’m controlling the jetson through jupyterlab plugged into my computer through microUSB so there are no peripherals attached except the USB camera. Am I limiting myself by only using one GND pin (maxes out at 3A?)? Could there be another reason the Nano shuts itself down?

You can test if it is a power-related issue by putting your board in 5W mode before running the program:

$ sudo nvpmodel -m 1

If you don’t see the behavior then, it’s likely related to power supply or connection. You might want to try setting your bench supply to 5.1V, using thicker gauge wiring, or connecting another GND pin.

I connected the 2nd GND pin and didn’t have any shutoffs so must’ve been a power throttling issue