Profile - Bets-Fit

[[Cl...]]

I am measuring a six sided post for surface profile evaluation. I'm measuring 6 planes and recalling the points into a FFS. The profile is called back to -A-, which is a cylinder concentric to the six sided post. I'm running a type-1 and my numbers don't look great. Repeatability is 23.56%. I need <20%. I'm thinking about doing a best-fit in the FFS evaluation. The 6 sided post is in the YZ work plane. So, my question is; would I lock translation in YZ an rotation around X, correct?

[[To...]]

Please sign in to view this quote.

If you’re only using Datum A in the profile, then it is already doing a best fit rotation around A. No translation is allowed. Open profile and close it. View Default printout. Should show you what I’m saying.

Historically, I have trouble with FFS on repeatability.

[[Cl...]]

I don't know how I can improve/change the program to get me below 20%?

[[Cl...]]

I can't post any pic's or drawings. The part is a 3.338" long cylinder, laying horizontally along
the X axis. The .984" long x .4716 Ø portion is Datum -A- (more towards the X- end of the part).
On the far X+ side is a .551" long with six .197" flats, (a hex). Drawing calls out surface profile of
this feature back to -A-.

[[To...]]

I wonder if a bunch of touch points on each plane would help.

[[Ja...]]

Whenever I'm faced with a profile callout referencing only one datum, I usually create an alternate alignment using that datum. For example, if A is a plane, I'd recall all points of A into a single point or centroid. Then, I'd create an alignment leveling to A (the plane) and originating from A (the point). Then reference that alignment in your profile characteristic. Give it a shot and see if your repeatability improves.

[[Cl...]]

The primary datum here is a cylinder. I'll give it a try.