• TTom

    • Post #1
    • Edited

    Hello,

    I am having some issues with camera disconnections after a period of time being connected. I am on the latest DepthAI 2.29.0, and have the latest camera firmware installed.

    Here is my hardware:

    • Jetson Orin Nano
    • 3x OAK-FFC-4P, each has 1 IMX577.
    • 1x OAK-1 W (OV9782)
    • USB 3.2+ cables, less than 1m.

    What happens is after some period of time, sometimes as small as 2-3 minutes, sometimes as long as several days, I will get this XLINK error below, and the camera will disconnect.

    F: [global] [    985675] [Scheduler00Thr] dispatcherResponseServe:928  no request for this response: XLINK_READ_REL_RESP 1

    Sometimes I'll also get an error about writing:

    F: [global] [    800972] [Scheduler00Thr] dispatcherResponseServe:928  no request for this response: XLINK_WRITE_RESP 1

    What could be causing this? Typically it is the IMX577 modules that disconnect, but sometimes the OV9782. For example, this weekend I kept the hardware on and the OV9782 stayed up the entire weekend, but this morning, I restarted the pipeline and the OV9782 failed after just 5 minutes.

    I have read into power being maybe an issue here, but I still get this issue with the FFC boards plugged into the external power adapter. I have also read about the Jetson USB drivers potentially being an issue - tonight I will leave the cameras connected to a regular machine and see.

    It is difficult for me to provide a MRE since the DepthAI code is heavily wrapped for my use case, and reproducing locally takes sometimes days.

      Tom
      Yeah that is still an issue that hasn't been solved. I might get around to try and figure out the cause but it doesn't seem depthai related.. Let me know how it goes on a different machine.

      Thanks,
      Jaka

        • TTom

          • Post #3
          • Edited

          I plugged in two boards to my laptop overnight, and I'm still having problems, although a different message.

          I turned off sleep, etc so the device was active the whole night. This happened around an hour after I set it up.

          What's interesting to me is they failed at the same time. Both devices were connected via one USB hub.

            Tom
            If that is the case, then very likely there is a power issue with the hub. If they both die at the same time that would signify a voltage drop.. Might be worth checking the dmesg logs for more info on exact timing.

            Thanks,
            Jaka

            • Tom replied to this.

                • TTom

                  • Post #5

                  jakaskerl I'll try again with my laptop and powering the boards. I only tried powering them when on the Jetson, so it could be a combination of both the power and Jetson USB drivers potentially.

                    5 days later

                    • TTom

                      • Post #6

                      I set up a spare 4P board with two modules (AR0234 and IMX577) and left it plugged into a Linux machine for around 4 days with no issues at all. I didn't even have power plugged in.

                      So now what I'll do is take this board / module combination, and see if I have any issues connected to the Jetson and to power. If I have issues, that should essentially isolate the Jetson as being the problem.

                      • TTom

                        • Post #7
                        • Edited

                        The device disconnected overnight while connected to power, so that should isolate the Jetson as being the problem.

                        One more thing I'll try is to disable USB suspension as mentioned here.

                          Tom
                          Looks promising, LMK if you find anything - I'd like to fix that just as much as you.

                          Thanks,
                          Jaka

                          • Tom likes this.

                            Seems like this did not help, unfortunately.