Best Fit Alignments/ Spherical Radius

[[Ha...]]

I am currently trying to measure several features but I have a couple different questions. The views on the print are not that great so I will do my best explain. There is a Spherical Radius (datum D), but there are only 4 small portions to measure and there is a Composite Profile callout to the Sphere. First question, how should I call out the first part of the profile? Should I create a sphere with the small amount that we have and constrain the X-Y-Z for the size of the Radius, then constrain to the Radius for the location? Or, should I create Sphere Points from the additional features column? Also, what would the second part of the Profile be, form?

Next question, there is a second Composite Profile callout to 3 separate surfaces with the DRF "DAB". Datum D is the primary and is the Spherical Radius with 4 small portions. What type of alignment should I create to call this out correctly? Any help would be appreciated thank you!

Also, I attached 3 separate views. The first should be the Isometric view, then the second and third refer to the callout questions.

3d best fit 3.PNG3d best fit 2.PNG3d best fit 1.PNG

[[Br...]]

Do you have Curve and Freeform? If you have freeform you can measure the spherical radius as seperate 3D curves, call them into a freeform surface. Use this to do your first Profile callout to ABC, and then you can also do a Profile to no datum references. You can use the 3D curves you mesured to do a best fit of curves alignment to create the other Profile. I don't know what Datum feature A is so I'm not sure how to construct this one.

[[Ha...]]

Unfortunately, we do not have Curve or Freeform. For the most part we've been able to get by without it. Do you use 3d curves and freeform surfaces much or notice that is more accurate/useful? Also, you can call out a profile to no datum references? And Datum A is just the plane on the back side of the part.

[[Br...]]

Oh yeah Curve and Freeform are a must if you deal with profile a lot. So I think you can use general surface instead but you can only take individual points. No scanning. I'm not sure if you can do the Profile to no datum reference with general surface. You'll have to try it out. Also not sure if you can do a best fit alignment with it. You'll have to play around with it. The problem with just calling it a "Sphere" feature is that it will project the missing portion of the sphere to get a size and center-point. So when you use it as a datum feature, the center-point it uses could be way off relative to the actual surface. The alignment for the DAB Datum Reference Frame should constrain Spacial rotation on Datum Feature A, Planar Rotation on Datum B, and XY&Z orgins to Datum D. The lower segment of that composite FCF effectively constrains nothing considering Datum Feature D has no rotational constraints.

[[Me...]]

Can you use a tooling ball as a Datum Feature Simulator? Maybe use space points to verify the 4 surfaces of D and then the ball to get the correct center.

As Brett pointed out, you're going to be hard pressed to get an accurate centroid for the sphere based on the small amount of surface data. Curve is one of the greatest tools in Calypso, IMO. It does everything.

Profile to no datums is form only, Best Fit to all 6 degrees of freedom.

[[Ha...]]

A tooling ball might be difficult because it is pretty large radius and we definitely don't have that size in house. Right now I'm programming offline so I don't even have a physical part yet. However, I know this will give us issues because like you said, we're going to have trouble getting an accurate center. Also, I'm trying to figure out if it is worth the upgrade to Curve, considering the frequency of the more complicated parts we keep engineering.

[[Ha...]]

Would using Sphere Points be that bad of a way to inspect this part? You could preset the dimension from ABC using sphere points and put a +/- .002 tolerance on the size of the radius (datum D). That should cover the second part of the Profile callout (.004 no datum reference) somewhat, correct? Now that you've established that the Sphere is the correct size with preset locations, could you not just preset from ABC and call that alignment DBC? Technically, I know it's not the right way, but a .010 profile to DBC is a fairly wide open tolerance. Then, for the ".002 D" callout to the 3 surfaces, call it out as flatness and put +/-.001 on the associated basic dimensions. Again, I know it's not the best way to inspect the part, but would this sufficient?