Getting accurate results from a small section of large radii

small arc method is the only way to attempt...


1) put yourself in correct workplane
2) create an alignment that gets yourself as local as possible to probe the radius you want to measure. ensure that alignment correctly controls all six degrees of freedom
3) measure the arc with autocircle with 4 hits. use FIXEDRAD as the bestfitmathtype.
4) create alignment to the fixedrad
5) while recalling the alignment to your fixedrad arc, measure the arc in the same spot again (now at (0,0). this time use LEASTSQU as the bestfitmathtype and a normal amount of hits

I have one part where we measure a 0.2200RAD (simple blend for a step on a predominately flat gearface). Only have 12° of arc on the model. Customer refuses to let us use radius gages nor will they accept a profile scan. They demand a darn number lol. The above method works well with the ±0.0075 tolerance range we're given

Congratulations for having the greatest pfp on this forum...

idk what you mean by that comment my friend

Profile picture. 10/10

Finally someone who appreciate high art when they see it =P thank you ! lol

When you put your answer like that it seems so simple lol. I'm using 3d best fit alignment, the part doesn't have one straight edge. Do you think recalling the 3D B.F after measuring the radii and aligning to that will cause issues? Or should I just do a new 3D BF.

When you put your answer like that it seems so simple lol. I'm using 3d best fit alignment, the part doesn't have one straight edge. Do you think recalling the 3D B.F after measuring the radii and aligning to that will cause issues? Or should I just do a new 3D BF.

Sounds like an iterative alignment might be another option.

However, when creating the radius, measure it as a Fixed Radius, then align XY to that radius, then take a bunch of polar edge points and report the deviation.