One camera black in Arducam 16MP Quad-Camera Kit on Jetson Nano

Hi, I need help with a quad-camera setup on a Jetson Nano. I’m running the Arducam 16MP Autofocus Quad-Camera Kit (IMX519 autofocus, the synchronized version, advertised as compatible with Jetson Nano/Xavier NX).

Setup:

  • Jetson Nano (the one with two camera connectors), HAT connected to the connector closest to the power plug.
  • uname -a: Linux a 4.9.201-tegra #2 SMP PREEMPT Tue Jan 26 16:55:25 CST 2021 aarch64 aarch64 aarch64 GNU/Linux
  • cat /etc/nv_tegra_release: # R32 (release), REVISION: 5.0, GCID: 25531747, BOARD: t210ref, EABI: aarch64, DATE: Fri Jan 15 22:55:35 UTC 2021

Problem: I run the app and get an image, but one camera is black. I moved that camera to another port and there’s still a black area with the other 3 cameras fine. It really looks like one camera is broken.

What I ran:
./install_full.sh -i imx519
python Multi-Camera-Adapter-Demo.py -i 0 (3 cameras’ image correct, one black)
python Multi-Camera-Adapter-Demo.py -i 7 / -i 8 (same log)
v4l2-ctl --set-fmt-video=width=1920,height=1080,pixelformat=BG10 --stream-mmap --set-ctrl=sensor_mode=2 --stream-count=1 -d /dev/video0 --stream-to=frame.bin

Log from -i 0:

Error: Write failed
Encoder null, cannot set bitrate!
Encoder Profile = High
bitrate = 4000000
Encoder Profile = High
Encoder control-rate = 1
Encoder EnableTwopassCBR = 0
Opening in BLOCKING MODE
** Message: 23:07:47.676: <main:4674> iterating capture loop ....
Starting automation...
Error generated. .../gstnvarguscamerasrc.cpp, execute:740 No cameras available
NvMMLiteOpen : Block : BlockType = 4
===== NVMEDIA: NVENC =====

Log from -i 8:

Error: Write failed
GST_ARGUS: Available Sensor modes :
GST_ARGUS: 4656 x 3496 FR = 10.000000 fps ...
GST_ARGUS: 3840 x 2160 FR = 21.000000 fps ...
GST_ARGUS: 1920 x 1080 FR = 59.999999 fps ...
GST_ARGUS: 1280 x 720 FR = 120.000005 fps ...
   Camera index = 0
   Camera mode  = 3
   Output Stream W = 1280 H = 720
   Frame Rate = 120.000005
GST_ARGUS: Setup Complete, Starting captures for 0 seconds
CONSUMER: Producer has connected; continuing.
Image Captured
GST_ARGUS: Cleaning up
CONSUMER: Done Success
GST_ARGUS: Done Success

i2cdetect -l:

i2c-3   i2c   7000c700.i2c              I2C adapter
i2c-1   i2c   7000c400.i2c              I2C adapter
i2c-8   i2c   i2c-6-mux (chan_id 1)     I2C adapter
i2c-6   i2c   Tegra I2C adapter          I2C adapter
i2c-4   i2c   7000d000.i2c              I2C adapter
i2c-2   i2c   7000c500.i2c              I2C adapter
i2c-0   i2c   7000c000.i2c              I2C adapter
i2c-7   i2c   i2c-6-mux (chan_id 0)     I2C adapter
i2c-5   i2c   7000d100.i2c              I2C adapter

i2cdetect -y -r 1 shows nothing. i2cdetect -y -r 7 and -y -r 8 both show 0x24 and 0x6a.

dmesg | grep arducam:

[ 1.329309] arducam-csi2 7-000c: probe failed
[ 1.329342] arducam-csi2 7-000c: Failed to setup board.
[ 1.353763] arducam-csi2: arducam_read: Reading register 0x103 failed
[ 1.353769] arducam-csi2 8-000c: probe failed
[ 1.353805] arducam-csi2 8-000c: Failed to setup board.

I also followed the install instructions and got images with -i 7, but the images came out random. Any idea what’s going on? Thanks!

Thanks for the detailed report and logs, that helps a lot.

First, one important version note: the IMX519 driver currently only supports L4T 32.6.1 and 32.7.1. Your board reports R32.5.0, which isn’t a supported release, so keep that in mind as we narrow things down. The driver install, for reference:

wget https://github.com/ArduCAM/MIPI_Camera/releases/download/v0.0.3/install_full.sh
chmod +x install_full.sh
./install_full.sh -m imx519

Reference doc: https://www.arducam.com/docs/camera-for-jetson-nano/16mp-autofocus-camera-on-jetson-nano-and-xavier-nx/

Now, since 3 cameras work and 1 is black, let’s work out whether the fault follows the module or the port/channel.

Step 1: switch the channel and detect the i2c signal. Run:

i2cdetect -y 10

Send me the result.

Step 2: then connect ONLY the camera that outputs the black area (just that one, by itself) and test it with these register commands:

i2cset -y 10 0x24 0x24 0x02
i2cset -y 10 0x24 0x24 0x12

Let me know what i2cdetect shows first, then we’ll go from there.

Ok, i2cdetect -y 10 doesn’t work at all for me:

Error: could not open file '/dev/i2c-10' or '/dev/i2c/10': No such file or directory

There’s no i2c-10 on my board. Here’s what i2cdetect -l lists:

i2c-3   i2c   7000c700.i2c              I2C adapter
i2c-1   i2c   7000c400.i2c              I2C adapter
i2c-101 i2c   tegradc.1                 I2C adapter
i2c-8   i2c   i2c-6-mux (chan_id 1)     I2C adapter
i2c-6   i2c   Tegra I2C adapter          I2C adapter
i2c-4   i2c   7000d000.i2c              I2C adapter
i2c-2   i2c   7000c500.i2c              I2C adapter
i2c-0   i2c   7000c000.i2c              I2C adapter
i2c-7   i2c   i2c-6-mux (chan_id 0)     I2C adapter
i2c-5   i2c   7000d100.i2c              I2C adapter

I checked the buses I actually have:

i2cdetect -y 0:

     0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f
00:
10:
20:
30: -- -- -- -- -- -- -- --
40:
50: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
60:
70:

i2cdetect -y 7:

     0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f
00:
10:
20:
30: -- -- -- -- -- -- -- --
40:
50: 50 -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
60:
70:

i2cdetect -y 8 gives the same as -y 7 (0x50 at 0x50), and i2cdetect -y -r 8 looks the same as before.

So the whole i2cset -y 10 ... test can’t run. What should I use instead?

Thanks. The key thing to figure out here is whether the fault follows the camera or stays with the port. There’s a useful diagnostic rule: if you swap the cameras but do NOT swap the cables, and the black image stays with the same camera, that points at the camera board itself rather than the port.

Some things worth checking:

  • Is the camera’s cable connected properly?
  • Is the connector on the camera module loose?

Please unplug and re-insert the connector on the camera module (the location shown in the docs) and try the demo again. If it’s still black after re-seating, the camera board is likely damaged.

Let me know if re-seating changes anything at all.

I checked both ports and switched the camera module over to the different cables. Same result, still one black camera.

I also tried unplugging and re-inserting the connector on the module, no difference there either. So it really does look like that one camera is dead.

Alright, based on everything so far: 3 cameras work, one is black, and it stays black even when you move it to a different port/cable and after re-seating the connector. That pins the fault to the camera board itself, not the port or the cabling, so it’s a hardware fault with that module rather than anything we can fix in software.

I’ve sent you the arrangements for getting that camera replaced through private message, please check your messages and provide the requested details there so it can be set up.

One more tip for when you test the replacement: bring it up one camera at a time first (single module, check i2cdetect on the correct mux channel) before reconnecting the full quad array. It makes it much easier to confirm each module is detected. Let us know if anything looks off.