OpenMV RT1062, Unable to Connect

When I connect my RT1062 to my laptop, an external drive pops up. Then, when I try to connect to the camera on OpenMV IDE, it says, “Connecting… (Hit cancel if this takes more than 5 seconds)”, but nothing happens next. After I press cancel and try again to connect, a new window pops up and says, “Error: A device which does not exist was specified.!”
Is my camera bricked? Is it possible to fix it?

Hi, the device is not bricked. The drive appearing means it’s working correctly.

This is something else. Which operating system are you on?

Connecting… (Hit cancel if this takes more than 5 seconds)”, but nothing happens next. After I press cancel and try again to connect, a new window pops up and says, “Error: A device which does not exist was specified.!”

Yeah, that will happen when you click cancel as the giving up error.

Could you also try clicking connect and just waiting… does it eventually connect or never? I’ve seen machines which are very slow at doing operations… like 10x what I’d expect.

Thank you for your reply!

I am on Windows 11.
I tried clicking Connect and waiting for about 5 minutes, but it remained stuck on the “Connecting…” window.

Before the problem occurred, I deleted everything on the camera and then pasted a folder generated by EdgeImpulse. Could this have caused the problem?

The folder contains:
.openmv_disk
labels.txt
main.py
trained.tflite

Potentially, could you delete everything in the folder that appears on the camera, then remove it and re-attach it?

The camera was able to connect yesterday when I emptied (a SOUND folder couldn’t be deleted) the drive and pasted the backup of the camera’s old storage.

However, the same problem occurred today. I followed your reply and deleted everything. It finally worked! When I connected it to my computer, a .openmv_disk file was generated. I couldn’t do this yesterday because a SOUND folder couldn’t be deleted for some reason.

Thank you very much!

The camera failed to connect again.

I used “Save open script to OpenMV Cam (as main.py)” under “Tools” to save my code. Then, I pasted in the model and label.txt. The camera connection was fine until this point.

However, after I unplugged the camera and then plugged it back in again, the connection issue came back.

Hi, how big is your model? I.e., how long does it take to run inference? The only thing I can think of is that the system is processing your model and not servicing USB.

Sorry, I don’t know how to run inference without an OpenMV camera. I noticed that if I save the script not as main.py, the camera connects to my computer successfully. But only main.py will automatically run when powered with a battery. Is there a way to work around it?

Can you post your model and code? The system shouldn’t lockup running a model.

tmp.zip (453.7 KB)
Here is my model and code.

@AvidSprinter—Thanks for the script. It appears that you hit a race condition that the IDE wasn’t handling.

When the IDE opens the serial port, the print data you are sending isn’t flushed and sent to the IDE, which doesn’t expect to receive the data on the open the serial port. This causes it to retry… and somehow it retries and runs into the exact issue again. I’m not sure what about your script causes this to align so perfectly… I haven’t seen this issue before.

Anyway, I’m building a new version of the IDE that resolves this. This seems to have something to do with how TinyUSB works and the RT1062 in particular. Other cameras don’t have this issue.

I’ll post a link to the dev release of the IDE once it’s ready. Please note that I’ve been in the middle of refactoring a lot of stuff in the IDE to support the new ROMFS feature. The dev release of the IDE will resolve your issue… but, you will need to update it to the final release to be finished by the end of the month.

https://github.com/openmv/openmv-ide/releases/download/development/openmv-ide-windows-4.5.0.exe

This is a dev release of the IDE. It should be stable. But, could have issues.