R-J3 2000/iB 165 F robot crashing with V-500iA/2D vision system

[inbigtrouble]

Hello everyone. I'm here for the first time, looking for some expert help.
We're currently using 10 robots - connected to the vision system: V-500IA/2DV FANUC and controllors are R-J3iB Mate Controllor.
Robots are R-J3 2000/IB 165F.

The situation is as follows:
Two of the ten robots are used to depalletize. The parts are stacked 5 high, 4 parts on a skid. The camera is mounted on the robot - robot moves to the position above the object; snaps the picture of the part and picks it up. The vision is only looking for a circular area 60mm in diameter in centre of the object and it calculates the offset.
The issue is that the robots is crashing on the part. It will go through a continuous motion for about 300 parts smoothly before it starts crashing. When it crashes the picture is high 90% ,good pass on the vision system, yet the x axis offset coordinate is off for about 50mm compared to the previously taken pictures,for the same part position.
When I reset the system, it will run fine for another couple hundred cycles before it will crash again.

 

[itsmoked]

Sounds like a problem.. Doh!

I mean something not working correctly as opposed to just not working very well.

Possibilities:

1) Software error that is not clearing a counter, or a buffer overflows, and isn't properly being detected.
2) The encoder is messing up.
3) Electrical noise is affecting the encoder signals.
4) ?

Can you tell if this error only starts happening in a certain geometry, like only at the lower parts in a stack?

Is this 50mm error always on the same side?

Does this error creep up or is it a sudden offset?

Keith Cress
kcress -

http://www.flaminsystems.com

 

[inbigtrouble]

Sorry, I didn't get a chance to thank you sooner. I appreciated your quick reply.

I found out that the vision system is taking 2 different pictures, but it is carrying over the same offset. I have talked to the Fanuc guy, and I sent him the program. He said it was "strange". I found my way around and have put the logic to check the vision offset, and if the same as previous, to repeat the snap program. It seems to be working now.

Just writing to say thanks, again. Hope you have a Merry Christmas, and a Happy New Year.