What is the standard referenced for Y14.5 Gaussian best fit OUTER plane in 2022 version? (and related questions)

[[Ap...]]

For ASME Y14.5-2009 there is no clear statement of using Gaussian outer plane to serve as the datum. For 2018 it says part has to be adjusted to a single solution that minimizes the separation between the feature and the true geometric counterpart per Y14.5.1 (candidate datum set, chebyshev, LSQ, translation LSQ, etc.).

For Y14.5.1-2019 it says the recommendation is "constrained L2 (constrained LSQ)" from appendix B.

So is the Gaussian best fit outer equal to the "constrained L2" based on Y14.5.1?

 

related questions, 

#1, Refer to the figure, why can't I assign points to fit a Gaussian outer plane?

#2, I use GOM and Calypso frequently and would like to know is this "Gaussian Outer" equal to the "outer tangential" plane in Calypso?

 

Thanks.

 

[[Ch...]]

Hi,

yes the gaussian best-fit (outer) is an gaussian bestfit (i.e. it minimize the squared distance for all points) with the constrained that all points lies on one side of the plane. To our understanding this is what the Y14.5.1-2019 provides at one possibility to compute datums on planar surface. In support of ASME datum system is the main reason for introducing this fitting method.

In general, you can assign points to the outer gaussian plane, but these points needs to have a normal, i.e, if you are using surface points all should be fine. This is needed to decide on which side of the points the outside gaussian plane lies. Of course, the error message is not helpful and therefore we will improve this error message in the future.

 

Best Regards

Christoph Schult

[[Ap...]]

Hi Christoph,

 

Thank you for confirming it is from Y14.5.1-2019.

From what I see now, the constructed gaussian and gaussian outer planes are not parallel. Based on this I assume it is not a “translated gaussian method”, instead, it is a “constrained L2” which refers to B-2.3.2 and figure B-10? Is this correct?

 

Best,

Apris

[[Ch...]]

Hello,

 

yes this is correct. It is a real Gaussian fit with a constraint not a simple translation of the standard fitting result as required from the ASME.

 

Best regards

Christoph Schult

[[Ap...]]

Thank you! (and now I have to think about how to maintain the consistency between GOM(Y14.5) and Calypso)

[[Ap...]]

A further issue I'd like to confirm. While Y14.5.1 states the candidate datum and the 1/3 default rules for a valid datum, is this something being programmed in GOM? 

Or it uses points and assigned algorithm for the "minimum separation" between actual and simulators only?

[[Ch...]]

Nope, we are not using candidate set but computes a single datum according to Y14.5.1 Appendix B, i.e. no candidates and no 1/3 rule.

[[Ap...]]

Good to know, thanks!