We have an application that we would like to search during the place of the product. The stack search feature seams to only work for picking.
Anyway to activate the stack search during the place?
Hi John,
I do not have a solution for you how to search during place,
but maybe you can find a proper work around ?
I know the stack search feature does not support search on place.
Why do you need to search during the place ? E.g. is it the height of the empty work area you need to calibrate or do you want to
compensate for varying (or unknown) heights of the objects that are being placed ?
Best regards,
Anders
Anders,
Yes. Our product we pick up shrinks as we place it. After many layers, we can be off by quite a bit, so we need to compensate for this. Is there anyway to turn on the support for stack search for placing?
Another thought was we could have a signal go to the PLC and the PLC open the tool at the correct position, but the problem is the robot will still try to go to the calculated position. We would need to give a signal to the robot to ‘trick’ it that we have reached the position.
Thanks,
John
Hi John,
No, I do not think there is a way to turn it on. And I think it would
be a real challenge to make it work (if you could turn it on). E.g. stack search
does not work for master work areas. Also, stack search will typically “remove” layers when the measured height becomes too low compared to the expected height. And there could
be other unknown problems. But the idea/request is really interesting.
You could try to modify all coordinate for the actions (before going into PMDoAction), e.g. subtracting an offset representing the difference of the actual height and the expected height.
BR Anders
Anders,
How are you suggesting that we know what the difference between the expected height and actual height is?
Where would we put that number?
John
Is there a spot in the Rapid file that we could modify to allow the robot to go to the target, or till a signal is high (which ever is first)? Then we could cheat the product height down some and have the target be a last resort if the sensor does not get flagged.
Thanks!
Mats/Anders-
There has to be somewhere in the code that I can break into and put some code to stop the robot from continuing to its target if a signal is set high. There are a few case statements that the robot moves through during the place. I would think if we found the correct spot, we could put an “If” statement to trick the robot into thinking it has reached it’s target. We are very comfortable modifying the code, we just need some help on exactly where this code would go.
Thanks!
John
I have an idea that may work.
Use the methods described in the attached document to filter out the place target point (PM_TARGET_POS) for the Outfeeder.
Use the Target and Action data and perform an ordinary SearchL. When the robot stops due to the search, save the current position.
Use the saved position when you resume the pick. The result will be that the robot won’t move to any point, but all the required I/O events will be performed at the point where the robot is located.
When the pick is performed, the robot will continue with the configured depart path.
Good luck!
/Mats
Mats,
I have no problem filtering for the PM_TARGET_POS.
What if I do not care about the I/O events at this point? Does that make this task easier? Can I just set a bit high when the target is found, and jump to a part of the rapid where the robot is trying to reach it’s target?
If I have to do a SearchL, wouldn’t I have start at the pm approach instead of target?
Thanks!
I assume this will need to go into Procedure Operate? Do you have any example of what you are describing to do here? We are willing to give this a shot before scratching PM5 and going back to an all Rapid program that will take a long time to recreate.
Thanks,
John
I think you should put the custom search code where the red code is located on the second slide of the attached doc.