Rock 5 ITX+ Crashing

#1

Hey just wondering if anyone else has had this or knows how to troubleshoot it further.

I’ve got a heatsink and fan on it, and have tried 3 different power supplies (all 30W+), along with two different OS images - the official one, and also Joshua Riek’s Ubuntu for Rockchip.

The issue I have is that it just hard-crashes. Sometimes a few seconds after boot, sometimes hours. It’s completely unpredictable. It does seem completely tied to using the NPU unfortunately, which I’m doing with ffmpeg (Frigate). That being said, I believe I’ve had it crash without NPU use too, that just seems to make it more frequent.

Having stuck a serial bridge on the debug/UART ports, I can watch it boot and get to a login screen, but then there’s seemingly nothing extra output when it crashes. On HDMI and UART the cursor just suddenly stops blinking, and at the same time it stops responding to ping.

I am absolutely stumped, and I would hate for this thing to go to waste as I REALLY wanted to make use of the NPU since it’s such an awesome unit.

Thanks in advance :pray:

1 Like
OS or user issues
#2

Please send your crash log and we can help to check if it’s a possible hardware issue.

#3

Hey @jack great to hear from you, and thanks for helping out!

Where would I get a crash log from? As mentioned, there is no extra output via UART or on the screen, all IO just stops - no keyboard input, no ping, etc.

If you can advise on the log location, I can reenable rknn so it crashes, then send a fresh copy :pray:

#4

Okay some further testing, and first just to recap:

  • It crashed repeatedly with the official Debian 12 b3
  • It crashed repeatedly with Josh Riek’s Ubuntu
  • It crashed whether using a 12V 3A or 12V 5A power supply, both on UPS
  • I then found it crashed even when not using RKNPU
  • It crashes with no USB or HDMI connected (only LAN)
  • It crashes with HDMI and USB connected too
  • I deployed Armbian on SD card in case it was onboard eMMC; still crashed
  • I moved my NVMe SSDs to USB connection in case of M.2 problems; still crashed

At this point, it cannot be just software? It’s just OS + Docker + Frigate. This should definitely not crash! But I have ruled out all hardware I can. It is cold here too; 5C overnight, 20C max in the day, and the ITX+ board has it’s heatsink & fan on.

When it does crash it appears to completely stop accepting IO; no ping response, console takes no input, and even UART with a serial bridge just stops blinking. I have to unplug and reconnect power.

If there are logs I can check, I would love to, but I do not know where they would be as I would have expected/hoped that the UART output would have some data! But alas, I think this is hardware failure :disappointed_relieved:

#5

Hi @jack just following-up again? I have tried everything I think now, and it still crashes!

#6

Hi, @Platima

This seems quite unusual, I’ve sent you a PM for a replacement and we need to check what’s the issue with this one.