ZEDx strange behavior on connection and data oscillating

Hi**,**
We have a setup with 2 zedx connected to a card duo on a jetson nano super.
Let’s start saying that yesterday everything worked correctly in our lab.
The 2 zeds had a lot of difficulties connecting today, i don’t know if it matters, but today we were outside at it was “cold” like 5°C, in the lab is much more. The diagnostic was showing no zed. Nothing suggested that we had a zed connected. After a bit of rebooting a lot of times and cleaning all the cables (even if that doesn’t seem to be the problem), we got a connection on the zed.
The ZED_Diagnostic -c command is now detecting the 2 zed. But there are still a lot of problems. All of the tools availble in the SDK fail to open the camera. I will attach a screen.
Our custom software can connect to the camera, but the reading in position estimation are very strange and wobbly. With the protype still we have a reading that goes from +10m to -10m, sometimes crashing with variance error (see screenshot).

How can we troubleshoot this?

Hi @Ste2038
Can you please send a picture of your setup?

Normally, cameras and capture cards can work at lower temperatures.

The camera is mounted to the prototype and its impossible for me to take a picture.
To describe it to you, we have a jetson nano super, connected to the link duo card via the 2 CSI connectors. Both the Jetson and the card are powered by a 12V line. The 2 ZEDs are connected via the single 4-via connector of which we took off the 2 extra cables. (We also tried with a second version of the cable that has all 4, nothing changed). The 2 camera do have a modified cable having an SMA connector in the middle for faster mounting.

I will point out that the hardware was not changed during the tests, and it was working with no problems.

We also runned some test after the holidays and…unfortunatly I have to say that I see a pattern with cold. (5° outside)

Yesterday we saw the zed losing the ability to grab succesfully after 20 min outside, and if the software was restarted, no zed was detected, not even from ZED_Diagnostic. When we tried to make the link duo board a bit hotter, both zed came back alive.

This happened four time, consistently losing the zed until a bit of hot air was applied.