Camera error and segmentation fault persist on Raspberry Pi (Unicam timeout) - can I return product?

Raspberry Pi Cameras
1

I’ve been fighting with a camera error on my Raspberry Pi for a while now. It keeps throwing a Unicam timeout and then a segmentation fault, even after doing what support suggested. I’m running the standard Pi camera stack (the error comes from raspberrypi.cpp). Log looks like this:

[0:03:31.684134823] [1763] ERROR RPI raspberrypi.cpp:1875 Unicam has timed out!
[0:03:31.684198324] [1763] ERROR RPI raspberrypi.cpp:1876 Please check that your camera sensor connector is attached securely.
[0:03:31.684260676] [1763] ERROR RPI raspberrypi.cpp:1877 Alternatively, try another cable and/or sensor.
segmentation Fault

Not really sure where to go from here.

2

Those two errors together are a pretty classic sign that the CSI receiver (Unicam) never got any frame data back from the sensor. The segfault right after the timeout is almost always just the app crashing because the capture failed, so I’d treat the timeout as the real issue. The likely culprits are:

  • a faulty ribbon cable or a connector that isn’t seated properly
  • a faulty camera sensor module / bad connection on the CSI data lines
  • a problem with the Pi’s CSI port itself

Let’s rule those out one at a time.

3

Start with the easy hardware checks:

  1. If you have another ribbon cable, swap it in and retry.
  2. If you have another known-working camera (a standard Pi camera module is fine), test that on the same Pi. That tells us if the Pi itself is okay.
  3. Run these commands and paste the terminal output here:

ls /dev/v4l-subdev0 & sudo dmesg | grep ov5647

That will show whether the ov5647 sensor is actually being detected and whether the driver is probing it.

4

So I got another camera module cable and swapped it in - same error. I don’t have another Raspberry Pi camera to test with, but I did set everything up the same way on a different Raspberry Pi and I still get the exact same error. I also reformatted and set it up fresh, same thing.

Here’s the command output you asked for:

pi@raspberrypi:~ $ ls /dev/v4l-subdev0 & sudo dmesg | grep ov5647
[1] 1564
/dev/v4l-subdev0
[ 9.058444] ov5647 10-0036: Consider updating driver ov5647 to match on endpoints
[1]+ 완료 ls --color=auto /dev/v4l-subdev0

5

Thanks, that’s actually helpful. The ov5647 is being detected on I2C at 10-0036 and /dev/v4l-subdev0 is present, so the sensor’s control side is alive. The “Consider updating driver ov5647 to match on endpoints” line suggests the device tree endpoint definition is a bit old for the driver, but that alone usually isn’t enough to kill capture.

What’s notable is that you’ve now hit the same timeout on two different Pis, two different cables, and a fresh reformat. That makes a bad cable or a dead Pi port a lot less likely, and points more at either the sensor module’s CSI data path or the driver/DT streaming configuration.

At this point I’d like to look at the live state of the camera pipeline directly. I’ve sent you a link over private message to arrange remote access to the Pi - check your messages and share the details there when you can.

6

I’ve sent the remote access details through the link you gave me - time, TeamViewer ID and password are all in there. Let me know when you want to connect.

7

Got them, thanks. Quick summary of where we are: the ov5647 is detected on I2C and /dev/v4l-subdev0 is present, but Unicam still times out after testing on two different Pis, two cables, and a fresh reformat. That points toward the sensor module’s CSI data path or the driver/DT streaming setup rather than a simple connector/cable fault. I’ll connect and check the capture state next, and I’ll update this thread once I’ve had a look.