L515 not working

[DarthBogdo]

Required Info
Camera Model	L515
Firmware Version	1.5.8.1
Operating System & Version	Ubuntu 20.04
Kernel Version (Linux Only)	5.8.0
Platform	PC
SDK Version	v2.50.0
Language	ROS/python
Segment	Robot

Issue Description

The camera doesn't display a depth image in the realsense viewer or does so shortly only to report unreachable streams afterwards and show a black screen with occasional colorful dots. After unplugging the camera, the pointer submits a value of null and the device doesn't get recognized.

I am trying to use the 5.8 kernel and SDK v2.50.0 for compatibility, but I can't get past the patched kernel modules step:

[MartyG-RealSense]

Hi @DarthBogdo Could you first try unplugging the micro-sized end of the USB cable from the base of the camera, reversing the connector's orientation and plugging it back into the L515 camera, please (USB-C cables are two-way insertion at the micro-sized end). Does doing so make any difference?

[DarthBogdo]

Hello @MartyG-RealSense
I tried this previously on other versions of the SDK - didn't bear any results. Mostly doing so would initiate the "null pointer" issue as soon as I reconnected the device.

[MartyG-RealSense]

Have you tried building the 2.50.0 SDK from source code and including the flag -DFORCE_RSUSB_BACKEND=TRUE in the CMake build instruction, please? An RSUSB = true build of librealsense is not dependent on kernel versions or Linux versions and does not require a kernel patch script to be applied.

[DarthBogdo]

I have added RSUSB to the CMake step of the process described here: https://dev.intelrealsense.com/docs/compiling-librealsense-for-linux-ubuntu-guide

However, when I switched back, the issues with the depth stream:

and connecting:

persisted.

[MartyG-RealSense]

The instructions at #9931 (comment) for an RSUSB source code build are a little clearer than the guide that you linked to.

These are the CMake build flags that I recommend to use.

cmake ../ -DFORCE_RSUSB_BACKEND=true -DCMAKE_BUILD_TYPE=release -DBUILD_EXAMPLES=true -DBUILD_GRAPHICAL_EXAMPLES=true

[DarthBogdo]

I've built 2.50.0 SDK according to your instructions. Unfortunately, the depth stream issue didn't go away. The feed is visible when I start the the SDK after booting and then it becomes unreachable as soon as I turn it off or it disappears by itself after a short period of time:

[MartyG-RealSense]

A quickly disappearing depth image can be caused by GLSL settings in the Viewer's settings interface. #8110 (comment) has instructions on how to disable GLSL.

[DarthBogdo]

After looking into it, the GLSL options were already turned off.

[MartyG-RealSense]

The Viewer images that you have provided look like a hardware fault in the camera that is not repairable. If you purchased the camera less than a year ago then I would recommend enquiring to the retailer that you purchased it from about an RMA return.

[DarthBogdo]

I see. Thank you. Is there anything else I can try/do?

[MartyG-RealSense]

You could run the rs-capture example program to confirm whether the depth display problem is limited to the RealSense Viewer tool.

[DarthBogdo]

I can't properly test it because of the irregular behaviour of the camera in terms of device detection (the null pointer case I described before). Could you give me any advice for that?

[MartyG-RealSense]

It is highly likely that your camera has a hardware fault on its USB port on the base of the camera that causes the irregular behaviour, unfortunately. It is not repairable.

[DarthBogdo]

The camera now displays irregular behaviour: the depth stream actually works after launching rs-capture and waiting for a couple of minutes. Albeit, the device heats up considerably.

Thank you for your support and efforts to help me!

[MartyG-RealSense]

You are very welcome!

I would add that if a camera heats up considerably after only a few minutes after starting the camera then it can indicate that there is a problem with the USB port or with the USB cable (such as a cable that has internal damage from having been bent a lot).