So i am continuing this process a bit longer than I thought. So here is my few findings:
I found out that my linear voltage regulator is damaged in some way. LM9036MX the part number U140 on schematics. It was not regulating and 3.3V line was at 1.365V. I removed it and gave it 12V and it outputted 8V. So i am a bit concerned after this result cause all components that connected to 3.3V might got 8V or maybe worse.
After removing regulator i gave 3.3 from power supply so it can initiate the power sequence for type-c.
I did all the things using only type-c power supply to test if it can communicate or not.
It worked like this if i used jetpack 5 cause even with external power supply to directly to 3.3V trace it is not working probably something on the trace is burned or damaged to. If i use jetpack 6 usb ports will not work probably due to jetpack 6 driver checks but it works with jetpack 5. I am currently unavailable to further investigate this cause i did make some mistakes like giving 12V accidentally to 3.3V so i think i will have a much bigger problem than before.
First of all please check schematics before using dev kit with another custom boards you made. We did check it in version 1 board that we designed but forgot some of the diodes and relevant power regulating circuits so we end up with reverse voltage (amount is unknown) directly to 3.3V rail because we used non-isolated can line directly connected to motors. Probably we had multiple spikes during operation of motor. Probably can bus line carried that voltage if our transceiver is fried which it did and end up releasing all of the remaining power from that spike to directly can line. So after that our Orin side transceiver is also fried from same voltage so it end up releasing all the reverse voltage to its input pin. We powered up that chip directly through 3.3V pin of AGX Orin pins. So after that moment it is not 3.3V anymore, depending on the power we might damaged a lot of the components that not protected connected to same rail.
So in conclusion do not use non-isolated can like this setup. Also probably add diodes before connecting anything to 3.3V pin to protect orin in case of something like this. What we did is absolutely not recommended and it is not correct use of it. Wrote this in case something similar happened to anyone. If it is not booting up or that white LED is immediately turning down when releasing power button you have probably similar problem. In that case try to find type-c controller power rails so you can check if it is correctly receiving required power or not.
A custom board is defined as anything that is not a devkit.
So there is nothing called “a custom board that we directly connected to AGX Orin dev kit’s pins.”.
I guess you are talking about connected to Orin AGX “module” pin.
If you are using a custom board and this issue is on Jetpack6, you may need find someone for software to help. This sounds a typical software issue which we’ve seen a lot.
So let me be more specific. We had AGX Orin dev kit 64Gb. We connected a pcb that we designed to Dev Kits 40 pin GPIO headers in order to use internal can controller with rest of our system.
If you are talking about you added something on the devkit 40 pin headers, then it is still a devkit.
You don’t need to say something like “custom board” which is confusing others.
Also, I think the question here is, if you don’t do those hardware change, would you still observe the issue?
I don’t know what do you want to continue from that old post. It seems our guy already checked and thought your board had hardware issue.
We won’t really discuss an issue when the board is already defective.
I’ve seen this problem multiple times, both on my own kit and in other forum threads. At some point, the AGX Orin just shuts down and never comes back. After that, it won’t even take power from the Type-C input anymore. That’s why I wrote my original post: this isn’t just a one-off, it looks like a recurring hardware failure.
It’s really frustrating because it feels like the hardware simply gives up without warning, and there’s no clear documentation about why it happens or how to prevent it. If more than one user is facing the same issue, wouldn’t it make sense to have official documentation so people don’t waste time chasing the same dead ends?
Either your board or the devkit seems to be damaged due to reverse voltage fed back to the dev kit or applying voltage above the recommended operating range.
Please check your design carefully with our reference devkit design files provided and available for download in the Jetson Download Center and follow the AGX Orin design guide guidelines to avoid this failure mode repeating. If any specific questions, we can try and help answer them.
If using for motor control or high power control applications it might be better to use optocouplers between the devkit and your application board to provide electrical isolation between the low power control circuit side and the high power circuit side.