Some of what I said yesterday was a combination of “Beginner’s Mind” and the proverbial five blind men describing an elephant by touch. This morning, I woke with a clear vision of the whole 5-axis geometry and a plan to start fresh. It is very simple and gave reasonable numbers right away (X, Y and Z are trivial). The A and B angles looked good but didn’t match the numbers from your mysterious unnamed cluster… This is why:
The mesh for the “orange block” is cockeyed, not aligned with ‘World XY’!!!??? 
So I created this repair utility to align it as best I can before applying my computed angles to it. What a silly way to spend time. How did that happen?
5_AXIS_repair_2019Nov30a.gh (43.8 KB)
So I’ve now integrated that reoriented part with my new code, purged the mysterious unnamed cluster and it looks great at first, until I noticed that sometimes the A and B angles are positive instead of negative. I know why and am working on that.
A “big picture” question though: handling one point on the curve at a time simplifies the conceptual aspects, keeping in mind that the code should also work when fed all points (planes) at once. In the real world, isn’t there a timing issue moving from point to point?
Hope to post my new version soon.