Hello everyone,
I’m working with a custom carrier board setup that includes three Jetson AGX Xavier modules, all managed by an Enclustra SOM. I was following the power-up sequence according to the NVIDIA Jetson AGX Xavier Series System-on-Module documentation. Initially, the power-up process worked without issues, following this sequence:
- DC power on
- Enable HV and MV power rails
- Set
VDDIN_PWR_BAD_N
signal to high (1) - Press power button (transition: 1 → 0 → 1)
- Set
VDDIN_PWR_BAD_N
signal back to low (0) (this might have been an unintentional error) - Complete remaining steps in the sequence
With this approach, the Jetson modules booted normally. However, after a sudden power loss for whatever developing reasons, the modules fail to boot, often stalling around the fourth second of the kernel boot process. They would only boot correctly if the power supply was fully turned off for a period before attempting to start again.
To address this, I modified the power-up sequence as follows:
- DC power on
- Enable HV and MV power rails
- Set
VDDIN_PWR_BAD_N
signal (but mistakenly left it low instead of high, as the “_N” suffix led me to interpret it as active-low. Additionally, our circuit has an inverted level shifter MOSFET, which added to the confusion.) - Complete the remaining steps
After implementing this revised sequence, one of the Jetson modules overheated, drawing a substantial 20 amps from the power supply, which caused the shutdown of other boards. This module now appears permanently shorted and had to be removed.
Could anyone provide insights into why this overheating and high current draw occurred? Could setting VDDIN_PWR_BAD_N
low at this step lead to such drastic behavior, and what is the recommended approach to avoid this in the future?
Any guidance or similar experiences would be greatly appreciated. Thank you!