Thanks for your help, I think we will buy a new board.
I think there is some shorting/static charging things happened just like what mentioned in this post Computer won't boot unless I remove and reinsert the 24 pin connector after shutdown + turning off the power bar | Tom's Hardware Forum
We have exactly the same phenomenon that is: if I keep the cable plugging after it crashed, i won’t never able to turn it back on again. But if I unplug for a min or so and replug the power cable, I will be able to turn it back on.
@linuxdev have you seen this kind of issue by any chance? Thanks.
The tomshardware post is for a PC, but most every computer in existence has capacitors in it which can hold a charge after turning off…and starting after a complete discharge can make something behave differently than starting without a complete discharge (it shouldn’t happen, but on marginal hardware it does). There are a number of sequential steps in booting any system, and if a capacitor were to throw something out of order on a system where something went wrong it is plausible that a complete power off and discharge might temporarily fix something.
On the other hand, you’d need to have serial console attached and basically nothing else to see what is going on. Attached peripherals quite often are the end problem. Testing with just the console and seeing what happens (and getting a log) would be a next step, but you should probably start a new thread since your Jetson issues may be quite different than the oringal poster’s Jetson.
@linuxdev I am the one who started this thread. We (robotdreamers and khuong.hust) are actually the same team and talked about the same board :). As you remembered in earlier discussions, we already tried everything including attaching nothing to the board except the serial connection.
My last message is just the continue of our debugging process.
Sorry, the tomshardware info threw me off there…made it look like a new topic.
Since this message is still for the same board, I still suggest it is time for RMA since there is no way to do component level debugging on the board:
write raw data fail - Success
…this message during a clone is unrecoverable under most circumstances (and perhaps even at the factory). A reset of some component via complete power off and waiting for capacitors to discharge is just more evidence of hardware failure. If you are just trying to figure out what actually went wrong, then capacitors are always high on the list, but there is no way to actually know under the circumstances.
This does bring to mind the idea that if you can completely power off and get a valid start in recovery mode that way, then perhaps you can clone and save the data. Other than that I believe there isn’t much else you can do.