Scan without Curve

[[Pe...]]

Hello all.

I have a round part that gets coated on the OD. During the coating process, the applicator will leave a "hump where it makes contact, then another where it comes off the surface.

Without using freeform, curve etc., is there a way to scan the surface and capture the height of the "humps"?

The size and position of the "humps" vary slightly. Varies just enough where it isn't reliable enough to move to a specific location to capture the point.

The OD (without the humps) is approx. a radius of 57 mm.

[[Ky...]]

Maybe I am missing something, but this is just an OD cylinder nominally, correct? Couldn't you just measure it as a circle and then put a roundness on it? Again, I might be over simplifying, but right now that's just from lack of information at this point.

 

If we are talking about an arbitrary geometry (i.e. not a cylinder or cone, etc), there isn't really a way to scan that without curve or freeform.

[[Da...]]

You could scan it as OD like Kyle suggested and use the evaluation "Radius Points" (e.g. every 1° or so) to see where and how big the bumps are.

[[No...]]

If you scan it as a circle, you can also use "maxiumum coordinates" to find the point with the highest radius. I did something similar to find the position of a knob on an OD and measure its radius on an O-Inspect and it worked very well. Make sure to use a high enough point density.

Edited October 27

[[Pe...]]

Please sign in to view this quote.

I may be over complicating things. Ill try this. I thought about it but I never tried it. I did not think the roundness would be able to get the height.

[[Pe...]]

Please sign in to view this quote.

This is the mind set I had. I was hoping by using the scan I could just "capture" and report the highest point.

I need to figure out how to do this. I tinkered with this, and I was unable to figure out how to find the highest point.

Edited October 27

[[Pe...]]

Please sign in to view this quote.

I already use the radius points on the part, I didn't think about spacing them like that. I will try this also. 

Again, I guess this is what I was aiming for with using the scan.

[[Pe...]]

Thank you all for the responses. I will try these out

[[Lu...]]

You could pick up the non-bumpy part of the circle, then re-pickup the same circle this time including the bumpy part. use formulas to have the nominal of the bumpy circle be the same as the actuals for the non-bumpy circle. then do a diameter characteristic for both. In the diameter of the bumpy circle, constrain the location, then do minimum circumscribed. This should give you a diameter for both the non-bumpy part of the circle, and the distance of the center of that circle to the tops of the bumps (when dividing the bumpy circle's diameter by 2). To then find the height of the bumps you could just use a result element to subtract the radii from each other. Probably better ways to do this but this is what I thought of. Let me know if you have any questions

Edited October 27

[[No...]]

Please sign in to view this quote.

Maximum coordninate is not that hard to use. The maximum point is determined against a reference geometry. You can choose which one you want: The nominal or the actual geometry, which is the fitting element according to the selected fitting method. Usually you would select actual geometry here. The other selection is whether you want a point with or without a vector as the result. 

The maximum point is always calculated "perpendicular" to the feature selected. So for a circle it is the point with the maximum radial distance to the circle edge. For a line it would be the maximum distance perpendicular to the line. Note that you get only one maximum point in return. So if you have two bumps to measure, you have to do this twice with one circle segment for each bump. To get the actual height of the bump you can create a line from the circle center to the max point, intersect it with the circle and calculate the distance between max point and intersection point.

[[Ma...]]

You can use "Form" to show graph with selected deviation. Just restrict origin and radius - nominal can be taken from actuals of real circle.