I spent a good chunk of the day working on the swarf toolpath problem I've been facing on the truck project.
I sent an email off to Autodesk to ask for assistance in getting a better surface finish on the side of the cab of the vehicle. The problem I'm having is that it will leave facets and jagged edges on the side face. It seems to stutter as it moves along in its 5-axis motion, and changing the tolerance and smoothing settings wouldn't do anything to fix it.
Autodesk got back to me this morning with a few changes to my swarf toolpath. They increased the points along which the tool would follow and changed one of the selected settings from 'automatic' to 'entity,' which I'm not too sure what that does. They also checked a box that minimizes axial motion as the tool is contouring the part. However, after running the program, I had a very similar result, with only reduced facets. I did some more playing around with the settings and finally got a somewhat decent finish by adding more points for the tool to follow and using a different selection method when getting the rail guidelines.
My boss thinks the issue is around the computing for the toolpath; since the two straight line segments are different lengths, it's trying to plot the best course of action for the endmill using one point at the short line and more points at the longer segment. His thought is that if you change the points on both lines to be equal, so say a thousand on the first line and a thousand on the segment; the points should match up and have a more fluid motion. He came to this conclusion by how the corner fillets turned out, which was a mirror finish. The corner radius has the same length and line points on the top as the bottom does and so it doesn't need to compute a lot to get the toolpath.
I attempted to draw out a bunch of line segments to try simulating the idea manually; however, after I got to 8,000 points, my computer crashed, which wasn't a viable option. The other thought was that since it's trying to convert a straight line into a bunch of smaller segments and getting confused by straight lines. I could turn the straight line into a small ark. However, after trying this out, it didn't make a difference in the final result.
I also spent a bit of time playing around with my other finishing toolpaths and got some amazing results from slowing the RPM down on the tools I'm using. I really want to build a good tool library of proven toolpaths that I can pick from without having to wrestle with this over and over, and this is the first step toward that.