I have a Jetson Nano module that is not booting up anymore on my devkit.
I’m using a 5V 5A power supply and the devkit board is working, but the Nano module doesn’t work. I’ve flashed different SD cards to try it out, but no success.
I’m aware that when the screen doesn’t show up anything I can use the serial debug ports on the devkit board, but nothing is logged into the serial port. (When using a working Jetson Nano devkit, the serial debug console is working fine)
The device was being used in a project and it had to be powered 24/7, but in one day it stopped working and we couldn’t start it again. The area that it was being used stated that they had a good electrical stability, but we’ve noticed that sometimes the device stopped and it would only start again if we removed it from the power supply and energized it again.
I’m pretty sure this problem is related to the module, because I tried switching the module with a working one and the devkit board was ok and the system was booting up fine.
I’ve not personally seen this, but it sounds like hardware failure. One simple test might be to see if the module can be put into recovery mode, and then the host PC examined to see if the Jetson shows up as a flashable device from “lsusb” on the host. If the unit is able to be flashed, then this might fix the issue.
Note that SD card models (dev kits) have QSPI memory on the module itself, and that it is this QSPI memory which contains boot software. JetPack/SDK Manager can flash this, and that the SD card only gets referenced for boot once QSPI has completed its work. See if you can flash the module.
I am surprised that serial console fails, and yet flash gets part way in. Having the serial console work with other units though pretty much says it has to be this particular Jetson. The SDKM has an export log button on it, and I’m not very good with SDKM logs, but the NVIDIA people here should be able to use that to get an idea on whether this might be hardware failure. Can you attach that log file zip to the forum?
I have a similar situation, but it’s not exactly the same. I’ve a Jetson Dev Board that was woking for months as a part of a Kubernetes Cluster, powered by 5v 4A Power Jack (J25). This week it stopped working. After I removed it from cluster and disconected all interfaces (SSD, Ethernet), I realized that I can boot it only by micro USB 2A (J28) interface, but it’s not a viable solution.
Today, after read again some technicals docs, I realized that I can power it over Expansion Header (J41) using the two 5v header pins and GND. It can draw 6A in this configuration. I’d like to ask if someone knows any problem on powering the module in this configuration. It colud even be bether (6A vs 4A), but it’s safe? There is any kind of power regulator behind J25 power jack that I’m bypassing powering it by expansion headers? Anyone has any ideas if some circuit behind Power Jack could be damaged and if I can do anything that I can do to repare this? Thanks in advance for any advices.
Your description is good, but it is different than the problem in this thread (which is still ongoing), so you would be better off starting a new thread. I will say though that it wouldn’t be unusual for the carrier board power delivery to have an issue (the module itself rarely fails). Also, the barrel jack connector is often chosen as a superior solution to powering via micro-OTG USB since the power available by barrel jack is higher (USB itself has limits, and if you can imagine powering USB peripherals via the Nano which is itself running off of USB, then you can reach the limits by adding otherwise valid peripherals).