10 Inch Sphere Question

[[Sh...]]

10 inch diameter Partial Inside Sphere (Bowl) with a 2 inch diameter hole at bottom center. Need to measure the Sphere radius and location. My approach was to two ways first was a heliacal scan with 5 revolutions and the second was to do a partial circle starting at the base of the sphere just outside of the 2” hole to the top of the bowl then patterned that around to sphere every 15 degrees. Then I took all the circle segments and created a sphere. My issues is that the individual circle radii are about .003” smaller that that of the combined Sphere radius. The Created Sphere from the Segments are with in .0003” of the Helix Sphere method. Any thought on why this would occur.

Sphere.PNG

[[Ke...]]

What is the form error on the sphere(s) vs the circle path(s)?

[[Ri...]]

Please sign in to view this quote.

Small partial arcs are typically less accurate and less repeatable. I wouldn't worry about the individual, use the combined evaluation.

[[Er...]]

There seems to be a theme of recalled points not being as accurate. For the work I do recalling is accurate enough.

My hypothesis is that recalled points do not have their vector recalculated when brought into the new feature.

[[Da...]]

All of the sphere algorithms are based on making your results be "a sphere" it will average the heck out of the data to do that. Try making a circle section like you mentioned axially. Ast for the radius and look at the location. Now do one opposite the first. Usually what you'll find is that either the radius is off or the location is off. If the radius is off then the person making the part will have to correct it. Possibly a tool radius that wasn't compensated for or it's worn down to a different radius. If the location is off plus you'll get a form the looks like a doughnut too far minus and it looks like a football. Again the person making the part will have to compensate.

[[Ma...]]

You can analyze two of your scans that are opposite one another to create a circle that covers a larger cross section of the sphere. You can also constrain the location and/or diameter of this circle to match the output from the sphere and look at the results with a roundness plot to see the differences.

The flexibility of Calypso is great for analyzing this kind of problem.