OAK-1 disconnection issue with only one device in three

RyanLee

Hi,

We are working on is to capture images by streaming three cameras and manually or automatically focusing on the object. We do not use the AI model supported by the camera for it now.

When we run the Windows program provided buy you (Depthai Viewer), three cameras are captured, and when we select one of the three, it runs, but one camera randomly disappears and is connected again, and it works unstably.

When we tested it with the provided Java source, three cameras were connected and then some time later one disappeared, and it seems the connection was unstable.

Is there a way to set the IP address value or is there a separate setting program? Is there a way to solve it clear?

When coding in Java,three generations are detected

Only 1 unit disconnected

Best regards,

Ryan.

jakaskerl

Hi RyanLee
Is is always the same device? What is the bootloader version of that device? I see it is assigned an IP of 0.2 - could it be that there is another device on the network interfering with this one? I generally would not expect the first available subnet address to be used by the OAK.

Thanks,
Jaka

RyanLee

Hi,

We prepared some answer for you. Please let me know if you need anything more from us.

1. Camera recognition is random, sometimes one is caught and sometimes two are caught. So it's not like only one is having problems. Is it possible for multiple recognitions to cause conflicts?

2. I'll give you the relevant information.

As you can see, 192.168.9.101 : 0.0.22, 192.168.0.14 : 0.0.28, 192.168.0.2 : 0.0.28,

3. Currently, only one development PC and three cameras are connected to the router.

Best regards,

Ryan.

jakaskerl

Hi RyanLee

please update the 0.22 device bootloader to 2.28. The earlier version had an issue where two of the same devices would interfere with eachother
What do you use to power the devices (possibility for a power issue)?
Does the same thing happen in python as well (outside of viewer)?

Thanks,
Jaka

RyanLee

jakaskerl

1. Upgrade to bootloader 2.28

==> It seems to be 0.028. right??

Anyway, I updated the bootloader to the latest version, but the IP conflict phenomenon is the same.

2. Please check if the PoE power is being input safely.

==> I am using the adapter sold by your company for PoE power.

Ubiquiti 802.3AF PoE injector 15W, 48VDC, 0.32A.

3. I know that the same symptom occurs not only in the Luxonis program but also in the code you created. Please check this part again.

==> The symptoms are the same in the DepthAI viewer for Windows, the viewer launched in Python,

and the Java program we created.

Best regards,

Ryan.

HaseenaS

I am also facing the same issues, how to overcome this @RyanLee

RyanLee

HaseenaS

no not yet solve this issue.

jakaskerl

Hi RyanLee
Thanks, for testing.

Can you check with a POE switch perhaps? I find that I have issues with the POE injectors - instability likely caused by insufficient power.
Would you mind checking with luxonis/depthai-experimentsblob/master/gen2-multiple-devices/main.py just so we know for sure it's not a heavy pipeline/bandwidth issue. I figure you should experience the same thing even with this script.

Thanks,
Jaka