Circle fit point bug?

circle fit test.3dm (64.5 KB)
Useing the botton “Circle: fit points” on the set of points in the uploadet .3dm file I get:


Which I think is the correct fit but transladed to some plade.

Yeah, that is kind of odd… the size of the point circle is relatively small relative to the file tolerance, but that shouldn’t matter all that much I think… Don’t know how the circle fitting algorithm works in this case with non-planar point sets.

Attached I ran PlaneThroughPt to find the “best fit plane”, flattened the points to the plane, then ran Circle>FitPoints… The result is quite different…

–Mitch

circle fit test-2.3dm (71.7 KB)

1 Like

I scaled the points by a factor of 1000 and moved them close to the origin and the results were the same so it’s not due to the scale of the points set or tolerances.

When I used just the larger group of co-planar points the resulting circle lies in the best fit plane for those points as expected.

Looks like a bug in the algorithm. My guess is the cause is probably a typo in the coding rather than the logic.

Got that, thanks.

http://mcneel.myjetbrains.com/youtrack/issue/RH-32838

-Pascal