Cylinders vs. 3d Line

[[To...]]

Many times, I've seen people post to use a 3d Line constructed from 2 circles instead of using a Cylinder. Somebody may have explained why but I honestly don't remember. Can anyone explain why?

Also, there is information in the cookbook that states that if a cylinder depth is less than it's diameter, a circle feature should be used instead of a cylinder. However, nowhere in ASME Y14.5, have I seen anything stating this practice should be done. All I ever see is that the entire axis of a cylinder should be considered in a position or other GD&T call outs. I understand there are applications where measuring a very short cylinder may not be practical but there are plenty of "depth less than diameter" applications where a cylinder could be used with reliable results. Plus, if I measure 2 circles to create a 3d Line, the length of line is not the full length of the cylinder.

[[Ri...]]

Please sign in to view this quote.

I think a lot of people do this from their old DMIS days. The 3d Line method is incorrect because the axis line is drawn from the center-point of the circles. The Cylinder/Cone/Step-Cylinder using all of the points to create the axis line is the correct method. This is why you will see differences in the A1/A2 when comparing a Cylinder that is created from all of the scanned points, and a 3d Line created only from the center-point of the circles.

This topic gets covered in AUKOM GD&T as there is great discussion regarding how ISO & ASME describe datum construction, and I know our trainer talked about doing this in Calypso.

[[Cl...]]

Would it make a difference whether the hole(s) were interpolated, vs drilled?

[[To...]]

I would assume that creating a 3d Line from 2 circles or creating a cylinder using 2 circle paths would be nearly the same. Length of 3d Line and length of cylinder axis would be the same. The difference would come with a cylinder extracted from a model measuring with 2 circle paths due to the extrapolated end points. However, if this was a datum feature, would it really matter as the axis is infinite. Obviously, the distance between the 2 circle paths affects the reliability of this axis.

[[Ri...]]

Please sign in to view this quote.

Do some testing, and see what results you see. The two will be different. As your form goes up, so will the differences.

[[Cl...]]

The Step-Cylinder is a fairly new addition to Calypso. What is it, 3 or 4 years old? Prior to
that, if you had two different size circles to work with, then the 3d line was your only option.

[[Ma...]]

Theoreticaly if cylinder is rather cone, then it will get another results for dimensions, but axis should be the same i think.

I wonder how it is for length and start point of this cylinders for calculations. ie. for 3d line it is fully defined by circles, but if you measure cylinder a few mm from edges, would it be shorter?

[[An...]]

Recall Feature Points into a 3D Line
versus
Cylinder

There is not any difference (in my opinion)

Just showing the travel of center point by using different circle pathes.

See attached.

Contribution_09_09_2020.pdf

[[Da...]]

Please sign in to view this quote.

Thanks Andreas. Looking at sheet #6, is there graphical bug possible? Shouldn't be right in the center?

[[An...]]

Please sign in to view this quote.

Least Squares evaluation:
Axis right in the center.

Feature points in 3D Line:
Axis 0.12mm below the center

Has to do with the point density/distribution on the surface.

All illustrations are the result of CALYPSO-calculations.
The surface points (surface geometry) are designed
by a CAD application.

Contribution_22_09_2020.xlsxContribution_15_09_2020.pdf

[[Jo...]]

I don't think a 3D line between two circle and an axis per ASME Y14.5 is the same. Looks like depending on where the circles are scanned will affect where the center is calculated. Axis of a cylinder is the most accurate evaluation as some have already stated.

3D line_vs_Axis.png

[[Mi...]]

Please sign in to view this quote.

I agree it's not per ASME, but it isn't quite as bad as your picture suggests - the real issue is only scanning at two different heights. That would cause problems even if evaluating as a cylinder.

[[Jo...]]

Please sign in to view this quote.

I drew it like that so that the error was actually visible and the point was clear, that's all. If you evaluate this as a cylinder and only use two circles then yes, it's no different. That's why you would scan at least 3 circles.