Hey everyone, I’m having trouble getting an Arducam IMX477 camera module working on my NVIDIA Orin Nano running JetPack 5.1.1. The system just doesn’t seem to recognize the camera at all. I’ve checked the ribbon cable connection a few times and it seems seated properly, but no luck.
Here’s what I’m seeing in dmesg when I boot up:
[ 10.245631] imx477 9-001a: invalid sensor model id: 77
That model ID (77) doesn’t look right for an IMX477. I’m guessing this might be a different sensor variant that needs a specific driver? Does anyone know if there’s a JP5-compatible driver and device tree overlay available for this particular Arducam variant? The stock IMX477 driver in JetPack doesn’t seem to be handling it correctly.
Before we go down the custom driver route, can you clarify what exactly is happening with the native IMX477 camera driver that ships with JetPack 5.1.1? The stock imx477 driver should work with standard IMX477-based modules. The fact that it’s probing but failing on the sensor model ID check is interesting — it means the driver is actually being loaded and trying to communicate over I2C, but the sensor is reporting back an ID of 0x77 that the driver doesn’t expect.
Are you saying the native driver is not being used at all, or is it loading but failing at that model ID check? The dmesg line you shared suggests the latter — the driver is loading, it’s finding something at the expected I2C address (0x1a), but the chip ID doesn’t match what the driver expects for an IMX477.
Have you tried anything else besides checking the physical connection?
Yeah the driver definitely loads — that dmesg line is from the stock driver trying to probe the sensor. So it’s being used, it just rejects the sensor because of the model ID mismatch.
But here’s the bigger issue I’ve run into: when I try to actually test the camera or access it through any capture pipeline, the entire system locks up. Completely frozen — keyboard stops responding, mouse cursor freezes, everything. I have to do a hard power cycle to recover. This happens even just trying to run a basic v4l2-ctl or nvargus-daemon. It’s not just the camera not working, the whole Jetson becomes unresponsive the moment anything tries to talk to that I2C device or the CSI lanes after the driver fails to bind.
Anyone else run into this kind of hard lockup with the Arducam IMX477 on an Orin Nano?
That’s a pretty severe failure mode — a full system hang when the CSI interface is accessed, especially if it’s happening even with basic V4L2 queries. This could be related to how the module is physically connected, power delivery, or a pin/signal mismatch that’s causing the MIPI controller to lock up.
Could you post a couple of pictures of your setup? One from the top showing the camera module and how it’s connected to the carrier board, and one from the back of the Jetson Orin Nano? I want to check which CSI connector you’re using and verify the overall physical configuration before we dig deeper into driver or device tree changes. Sometimes these hard freezes come from signal issues on the MIPI lanes that wouldn’t show up in a dmesg log.