OTE for a Cylinder

[[Ri...]]

Hello, everyone,

I'm back with another challenge. I first wanted to start off by thanking everyone who assisted me with the challenge of finding a solution of finding the 3-points used to construct an OTE Plane - and especial big thanks to Phillip Keller <3.

I thought that I would take what I learned with the OTE Plane, and apply it to Cylinders. It appears though that it is more complicated than that though.

There are cases where you will only see 3-points that construct an OTE Circle on the Cylinder. I suspect this is due to the shape of the actual cylinder - conical, or hourglass shaped.

Below is an example of data only showing 3-points.
Below is an example of data showing 5-points.
I ended up finding an article written by Vijay Srinivasan from NIST going over this very topic - I'm attaching the paper if anyone wants to read it. The paper talks about a constrained least-squares definition, but I'm not certain I entirely understand. Are they talking about constructing say an LSQ cylinder, and then having it shifted to the OTE diameter? Or we aren't entirely sure what they are doing because it's something they developed?

Either way, I'm looking for insight on what Zeiss does to calculate the OTE of a Cylinder to ensure stability in the cases where the diameter is conical or hourglass shaped.

Thank you.

ASME2016IMECE2016-67753.pdf

[[Th...]]

I had a copy of this paper as well a couple of years ago, although I don't know where I stashed it. Can't speak to the hard math, but based on the visual figures, I think your understanding is correct, which is that LSQ is being used to define the derived feature's location and axial orientation, and then its size is being constrained to the contact points on the measured surface.

[[Da...]]

A simple outer tangential element for a cylinder requires five points, which is the minimum required number of points to construct a cylinder in the first place.

However, the standards ISO and ASME have established more stable ways of creating OTEs.

In ISO, it’s basically a filtered minimum zone element that is expanded or shrunk to the one highest point, depending on whether it’s an inner or outer cylinder.

In ASME, a method called „constrained L2“ will be used, which is a filtered least squares cylinder that again is expanded or shrunk to the highest point.

These methods will be used for all OTEs, because they will result in more stable OTE elements than the old methods.

[[Je...]]

Please sign in to view this quote.

.

Richard,

Thank you very much for this post. OTE versus constrained LSQ is a highly relevant conversation in CMM programming and manufacturing and has been getting louder between the 2009 and 2018 versions of ASME Y14.5. I actually missed your first post on planar OTE from Aug-16 of this year, so I went back and read that as well.

The theory that propped up OTE as default best practice seems to be losing ground to the stability offered by LSQ.

Great question about wanting insight on what Zeiss does to calculate the OTE of a Cylinder to ensure stability. I have no idea. I hope that an application specialist or other personnel from Zeiss will be able to answer you. I'm definitely following this post.

.

Please sign in to view this quote.

Daniel, thanks for clarifying the difference between the ISO and ASME methods for calculating the OTE of a cylinder. I appreciate that ASME uses an L2 filtered cylinder shrunk to the highest point.

.

[[Ri...]]

Please sign in to view this quote.

I have heard of the gaussian shifted outside algorithm, but I was only aware of this for Planes, not for Cylinders.

I'm looking in Y14.5.1-2019 right now to see if this is covered there.

Thank you.

[[Ri...]]

Please sign in to view this quote.

I'm just going to go back to my old ways of stating that functional testing on a CMM is nonsensical.