CVS camera disconnects and causes reboot

[mvonstein]

We recently bought a 1454 and are having an issue where the cvs will (seemingly randomly, but after a very short time) lose the camera settings.   The only way we've found to recover it is to plug the camera into a different port and cold boot the CVS.  Obviously this isn't good for a production environment....

We get various error msgs depending on where you look:

In VBAI if it happens after the inspection is open and run once, it says:  "Resource not allocated" under the step icon.

In VBAI if you try to open the inspection after the camera loses connection, it says:

There is a possibility of noise on the cable

[mvonstein]

Sorry, please ignore the first post.  I accidently posted an incomplete msg somehow...

We recently bought a CVS 1454 and are using an Unibrain (IIDC v 1.3) camera.  Having an issue where the CVS will (seemingly randomly, but after a very short time) lose the camera settings.   The only way we've found to recover it is to plug the camera into a different port and cold boot the CVS.  It acts very similar, to when you unplug the cable while the inspection is running in VBAI.  Obviously this isn't good for a production environment...

We get various error msgs depending on where you look:
 
- In VBAI if it happens after the inspection is open and run once, it says:  "Resource not allocated" under the step icon.

- In VBAI if you try to open the inspection after the camera loses connection, it says:
 "The validation of the inspection steps returns errors.  State: Get Image - Step 2 "Acquire Image (IEEE 1394)1" could not be initialized.

- In MAX:  "The specified image aquisition device could not be initialized properly. Error -1073774583 occurred at an unidentified location"

- In MAX:  "0xBFF68004 Internal error"

I've tried to set the bandwidth down to 100 and the frame rate to 3.75 fps.   It runs longer but still ends up erroring out.

There is possibility of electrical noise or machine vibration on the cable.   We tried holding the cable in place manually to make sure it doesn't vibrate and it didn't seem to help.

I'm wondering: 
- Is there any way noise could cause this?
- Is there anything driver related that could cause this?
- Most importantly-- is there any graceful way (not cold boot and switch port) of automatically reconnecting the camera if there is an error or even if the cable is disconnected intentionally.

Any insights would be greatly appreciated.

Thanks,
Mike
 
 

[MartyP]

Hi Mike, Did youever figure out what the problem was? I am having the same problem with a 1454 and an ISG Lightwise camera. I have tried changing the camera and the 1454 with no success. I have always suspected noise or a problem with the driver. I have three systems running on three different machines and only have trouble with the system on one machine. It is only on this one machine that I have trouble. If you have any info I would appreciate the help.

Thanks,

MartyP

[mvonstein]

Yes it was very strange.  The way we fixed it was to mount the camera on a piece of delrin which isolated it electrically from the electric conveyor it was mounted on.   We found it by accident.  I unhooked the camera from the mount and it was fine.  When I touched it to the mount it would immediately lose connection.

BTW, there is still no "graceful" way that I know of as of version 3.6 to reconnect a lost connection...

HTH,

Mike 

[mvonstein]

I can send you a sketch of the delrin mount we used.  It's pretty simple.  If you need it, just let me know.

Thanks,
Mike

[MartyP]

Thanks, making an isolated mount won't be a problem. I wondered about noise as there are 5 vibratory feeder bowls with inline tracks, 4 servo motors, and one VFD on this machine. The odd thing is that I don't have this problem on any similar equipment. I have two identical systems that run without any problems. I wish I had read this post sooner, I have been chasing this ghost for over a year.  I will try what you suggest and post the result.

Thanks,

MartyP

[MartyP]

The insulated mount didn't seem to have any effect on the problem. Although the noise is measurable whenever the VFD cycles, I don't think it is the problem either. The acquisition step definitely times out at 5 seconds because I lose communication with the camera.  Like you said, there is no way to re-establish communication without rebooting the system.  As a last resort I tried changing the firewire cable. So far it seems to be working properly. I should know better, communication cables are the root of all evil.

Thank you forn your help,

MartyP

[Hony]

I have similar problems with cameras connected to CVS thrue 2 FirWire ports thrue two hubs completely 8 cameras.

Time to time happend that two or four cameras are disconnected from the bus. Then is enough to find them disconnect them from the hub and one after one again connect back. Then it is looking o.k.. I feel that there can be some kind of network overload during inicialization of cameras, then camera itselve is not able to comunicate with CVS (probably due to some error state inside camera) and CVS logically can not see camera due to noresponse. With disconnection of camera from firewire, controler inside camera is reseted and when the comunication after again connection starts properly, everithing is O.K..

That is my theory.

Jan

[mvonstein]

Seems to me it's just a matter of poor error handling within the VBAI Acquire step.  If it errors it should try to reinitialize.  Of course there may be other things that are time-sensitive that are going on in the user program that you don't want to slow down, so may want to make it a checkbox defaulted off on the step.  Something like "Automatically connection reset".

HTH,

Mike 

[mvonstein]

Sorry, I just reread my post.  I should have been more specific.  Meant that as a suggestion to NI for future development, not something Hony can personally do for the problem unless writing a "Run Labview" step.

Mike