Profile with Best Fit Alignment

[[To...]]

I am using a curve to report profile with no DRF on a very small groove. The radial location of the the groove is dimensioned from a diameter with a fairly loose tolerance. Because I need to use a .020" probe in the groove, I'd like to establish the rough location of the groove from the actual diameter before actually measuring the curve. I'm leaning towards taking a point on the diameter and using it in a Geometry Best Fit and allowing translation in X and then assigning this GBF alignment to the curve. Sketch below is view from -Y

Am I on the right track?

Profile groove.jpg

[[Jo...]]

Are you reporting this as sections, as Profile of a line, or profile of a surface ? To solve navigation, just scan thru that geometry and limit the evaluation points. I believe you can select No Datum reference in the Profile of line dialog. If you measured that as a cylinder , torus, cone, then call into freeform maybe best fit Profile of a surface ? I don't understand how a variable diameter could accomodate that though.

[[Ch...]]

self-center any size probe that gets more than half the ball past the edge into this gap ? use midpoint eval and calculate from there?

[[Ch...]]

Here is a strategy I've tried for curve features that are allowed to float a lot. I think you might be able to use it here.

You first do a rough scan of that feature with curve. use a bigger probe and keep your navigation and speeds safe. The idea is to use this rough scan as an alignment for your precise scan. Then you scan it again with curve but this time using your small probe; Use your first curve as navigation alignment for second curve.

[[To...]]

Please sign in to view this quote.

I am measuring line profile. The reason I'm trying to create an alignment prior to scanning the curve to help locate the curve so that my probe has an easier time following the path.

[[Ri...]]

Please sign in to view this quote.

I had to do this on a part recently where I also had to use Unknown Contour to help locate the Curve.

Worked well.

[[Je...]]

Its 6pm on Friday so I will unfortunately not be doing a full write-up at the moment but I'll touch back on Monday morning. This is the fun stuff...

[[Ch...]]

Please sign in to view this quote.

I thought I was the only one to do this. Here is a pro-tip for unknown 3d profiles with no model or nominal information. Scan with Unknown Contour > use results from Unknown Contour to create a new curve and scan again > take actuals from that curve and use as nominals for a 3rd curve > Repeat until deviations from nominal are low. You now have an accurate data for a completely unknown 3d profile.

Theory behind strategy: Unknown Contour in great, but probe radius correction won't be good enough when you want highly accurate results. The iterative process above is intended to generate more accurate nominal and vector information with each iteration.

[[Ma...]]

Please sign in to view this quote.

I am doing that too. But sometimes after 3rd try it doesn't give me better results and still wrong vectors and that's the main problem of UC.

Could you describe for me how you are working with radius correction? I am closing curve after scanning, then i am going to modify nominal points by touch radius, but i was told there is another way to do it.

[[Je...]]

Please sign in to view this quote.

This definitely works but is an extra step with an unnecessary stylus. There are many many ways to get this done. Some necessary data is missing here but the easiest way to properly locate this feature this is to probe the "vertical" surface of this groove with the 0.020" stylus to create a local secondary alignment.

I am assuming the width of the groove is greater than the tolerance band for its location. If so, create a space-point on the vertical surface adjacent to point_A. Give the Retract (and before/after nominal distance) adequate zone to probe properly in the case of feature shift. Also, its a wise move to lower your Probing Dynamic value as to lower the inertia of the stylus. Probing Dynamice (Stylus config inertia) breaks small styli more often than Probing Force. 🙁

Use this Space Point as your locator (X or Y or Z datum feature) for a secondary alignment in the Curve feature which to scan your 2D or 3D Curve. If you wish to provide more specific data in terms of feature size/location, I can assist in a more defined way. I understand the need to be ambiguous however 🙂

To further refine; Recall the Curve into a subsequent Curve (make sure no measured data is present) to eliminate the irrelevant points or use the Point Elimination filtering option (I prefer recall using a defined point box in most cases)

[[Ch...]]

Please sign in to view this quote.

I left out an optional, but crucial step. Sometimes, I will take the additional step of creating a new solid model in Solidworks between iterations. Then I create new curves from that model. This way I can easily measure from more locations and see what is going on.

This isn't without risk. You may have to make some inferences based on your data to make the model. If you are wrong, then it could lead to erroneous data. The fun and games of experimental programming.