06-06-2025, 07:01 PM
Sorry, bit of a long post. I am trying to do several curve follow projects that all follow the same straight line but have different orientations. The line has normals that are 30degrees off of vertical learning towards the robot. The Kuka ABC orientation values for a program that follows the line with no additional modifications are (-150, 0, 90) which make sense to me. The only angle not parallel or perpendicular to the coordinate frame is A, because I am following normals that are 30degress off of vertical.
If I add one more adjustment to the orientation, it also behaves how I would expect. I applied a 30degree rotation about X so that the end effector leans into the path. The ABC values are now (-150, -30, 90), and it looks how I want it to. However, if I add another rotation, I start to see odd compounding angles. In addition to the normals and the lean, I want to twist 15degrees about the last axis. I would expect the ABC values to be (-150, -30, 75). If I manually drive the robot to this orientation, it looks exactly like what I want to happen.
However, using the "Path to tool offset" settings in the curve follow project, I cannot get the orientation to behave how I want. When I apply the 30degree lean and the 15degree twist and generate a program, the ABC becomes (-158.499, -28.879, 72.808), which is close, but incorrect enough that it is causing us problems. I've experimented with the orientation algorithms ("Minimum tool orientation change" and "Tool orientation follows path"), I've used the "Teach" command with "Minimum tool orientation change", and I've messed with the "Preferred tool X axis" options under "Minimum tool orientation change", but I cannot get the program to run with an ABC of (-150, -30, 75). I also have "Allow a tool Z rotation of +/-" set to 0, but I'm not sure that's related.
Are there any other features I can use to force the robot to maintain the (-150, -30, 75) orientation that we need? Or am I setting something up incorrectly?
If I add one more adjustment to the orientation, it also behaves how I would expect. I applied a 30degree rotation about X so that the end effector leans into the path. The ABC values are now (-150, -30, 90), and it looks how I want it to. However, if I add another rotation, I start to see odd compounding angles. In addition to the normals and the lean, I want to twist 15degrees about the last axis. I would expect the ABC values to be (-150, -30, 75). If I manually drive the robot to this orientation, it looks exactly like what I want to happen.
However, using the "Path to tool offset" settings in the curve follow project, I cannot get the orientation to behave how I want. When I apply the 30degree lean and the 15degree twist and generate a program, the ABC becomes (-158.499, -28.879, 72.808), which is close, but incorrect enough that it is causing us problems. I've experimented with the orientation algorithms ("Minimum tool orientation change" and "Tool orientation follows path"), I've used the "Teach" command with "Minimum tool orientation change", and I've messed with the "Preferred tool X axis" options under "Minimum tool orientation change", but I cannot get the program to run with an ABC of (-150, -30, 75). I also have "Allow a tool Z rotation of +/-" set to 0, but I'm not sure that's related.
Are there any other features I can use to force the robot to maintain the (-150, -30, 75) orientation that we need? Or am I setting something up incorrectly?
