IRC5P - Programs

EmilyReid

Hi, We have an IRB5500B IRC5P Paint Robot.

What is the process for loading modules created in RS into the actual controller?

We have been creating a module in RS, and loading them into the controller when the pc is connected. However, each time we do this there is always an error, and fails to open on the TPU due to semantic errors. When created in RS, the programs always run smoothly with no errors, it's just when transferring the modules errors appear.

Can someone advise me if this is the incorrect process for a paint robot, or what I am doing wrong please? 

lemster68

Try saving to a usb drive and then load from there.  That will help to find if it is getting corrupted by the file transfer.

jmf

Can you send some pictures of the semantic errors you get when loading? This might help to fault find.

EmilyReid

Thank you for your fast response lemster68, you’re help and knowledge is so important to us newbies! We are now running with programs from RS into the controller with no errors. However, we have realised that the robot in RS and the real life one are still out. E.g we created a simple routine with only 4 linear moves upwards, and the amount the robot is away from its intended target progressively gets bigger, like a scale. We noticed around 4mm for every 300mm travel, so about 1.3%. This therefore offsets all of our positions slightly… Do you have any ideas why this might be? 

EmilyReid

Also thank you jmf, I have somehow stopped the semantic errors?! I deleted all the unnecessary programs / data and the problems seem to have gone.

lemster68

Make sure that you have matching robot types.  Ensure that the calibration offsets match and check the zero position.  Check the TCP's match and are good.

Grahammy

So the robot is a IRB5500.
We have the correct robot in RobotStudio, there is only really 1!
We have jogged the robot to the centre point of our table, and placed the table in RS at the same point. All good.
Checked calibration and updated rev counter

We have done a Move J, axis 1 only, point. By tape measure in the real world it is 1155mm away, RS says it is 1130mm, thats an inch difference! Over say a 3m working envelope, Robot studio is going to be about 70mm to 80mm out. That can't be right.
We have done similar tests with other axis to ensure its not 1 rogue axis. 

Is there a scaling factor in RS somewhere?
We have an error factor of about 1.5%.

Obviously you don't see it on small movements, but once we import our finished product CAD drawings and start a painting program in RS, a 3" difference is massive.

Thanks, we appreciate the help guys.

lemster68

TCP's match and are good?

Grahammy

Yes, TCP is good. That would give the same static position error though.
Ours is on a sliding scale, the further the arm moves, the high the positional error

lemster68

That makes sense.  How fast are the motions?  Is the robot secure to it's base/floor?

xerim

In jogging mode, when you move the TCP to the origin of the work object, does it read 0,0,0? 

graemepaulin

Are the motor offsets in the controller the same as the label not sure where this is on a paint robot?

Does the robot jog in a straight line when jogging in linear mode (jog alongside a flat surface)?
If it is not jogging linear then the fine calibration may need to be redone (have had it before where it was wrong out of the factory).