The board was out of service while I copied the rootfs from micro SD card to SSD drive using the script provided here : https://github.com/JetsonHacksNano/rootOnUSB/blob/master/copyRootToUSB.sh. So, I did not have time to modify anything. Also, I later tested with a freshly flashed SD card as well.
This morning (after 2 days), I tried my luck by plugging the power supply and surprisingly the board booted up, I could then copy the rootfs to SSD drive and made Ubuntu run from that SSD.
But when I tried to run my application which is a DNN face detection (which is a resource-consuming application) using USB webcam (which shares the bandwidth with the SSD drive), the board easily went out of service for some times and finally ended up with the same state as 2 days before: the 3V3 disappeared, LED not lit. I used external power supply 5V-4A while running my face detection application for info.
I’m curious to know if it would return to operational state again in some days.
I already noticed about using high current power supply (actually I used the 5V-4A one, theoretically, i.e. I bought the adapter which says that its output is 4A max. but I did not really measure). What seems to be strange to me is that after the board went out of service a while, suppose that was due to not-enough-current, it should be back to operational state but it was not my case.
I noticed from other threads that the Jetson Nano is sensible to voltage drop that could happen with my power adapters. So I attached my new power setup in attachment.
One strange thing I also noticed that after the board went out of service, if I re-applied the power supply immediately, it did NOT work. But if I left it for about 1 night then re-applied the power supply, it was back to operational state again.
There might be some capacitors not discharged yet. You can try shorting pin 7 & 8 of J40 to disable auto power on first, then shorting pin 1 & 2 to implement a manually pressing power button.
Thanks for your kind support, I will try this to see if I could make it become operational again. But I assumed an ESD damage since after 1 night it the 3v3 came back + green LED lit but not stable, it went rapidly to out-of-service state.
By shorting pin 8 (LATCH_SET) of J40 to GND, I can have the board in operational state every time I power it up. I need to read more about the J40 pins’ functionalities.
I have similar problem on my Jetson Nano. The board seems to start correctly (except i2c errors) but I have no voltage on 3V3 (J41). I tried and let it be for few days if it recovers but it didn’t help. Also tried to disable the Auto On and start manually with no difference.
If I measure 3V3 against GND with ohmmeter (when the board is not powered) there’s near 0 ohm. I was trying to track the problem to the 3V3 DC/DC U9 using the schematics from nVidia webpages but there seems to be a different IC used on the board. There’s supposed to be MP2384 but on the board there’s EM5841C which I didn’t manage to find a datasheet for.
I’m persuaded that the Jetson Nano is quite sensitive to the quality of the power supply. I used to think that mine unstable or even died but I’m still using it today. You can try my “solution” in previous comment related to pin 8 of J40 to wake up your board again knowing that is not a documented way but at least it worked for me. One more thing is that I sometime have problem when using the 5V-4A power supply (thru the 5V jack) while plugging the micro USB port to my laptop.
I tried and started the board the way you had (using a jumper on pin 7 & 8 of J40) with no difference.
Could you do me a favour and measure resistance (ohms) between 3V3 and GND when your board is NOT powered. I suspect there might be a short connection on my board but it could also be the DCDC IC shorting the output when no powered. A datasheet would be helpful.