In Bryan Berson’s October column on GD&T, Berson says it’s “time to address best practices for selecting Datum Features and for specifying the order in which they are listed in Feature Control Frames.

“Because Datum Features serve to constrain degrees of freedom and establish coordinate systems relative to which the remaining features of a part are to be oriented and located, they ought in general (but with some exceptions) to be those features which actually constrain it relative to the mating part during an assembly process.” Berson continues. “Namely at the end of that process, the objective is for the coordinate systems of two mating parts to coincide in order to guarantee functional relationships between their remaining features.”

LinkedIn Groupmembers had their own commentary to share about the article.

Norm Crawford, dimensional management specialist at Miller Consulting Group:“The only spin I like to put on selecting datum features is to design with GD&T and therefore design datum features. This will cut R&D design interations down considerably and keep the entire team on board with the design.

Roy Jackson, quality engineer at Vermont Castings: “Not so much designing datum features, but datums should be in mind when creating any design...especially when there is interface with other components.”

Norm Crawford: “If you are keeping datums in mind while designing, then it only makes sense that you design the appropriate datum features to simulate those datums. All too often, the theoretical datums are what are kept in mind and then thoeretical datum features. The problem is that through prototyping and physically putting parts together, a design goes through itereations. With a little homework, it is pretty common to find that many design interations are iterations of the actual physical features that assemble the parts within an allowable tolerance. Those features are the datum features.
The bigger issue is that even then those features are ignored in favor of dimensioning a part using theoretical datums rather than the real datum features. That in turn makes the inspection data completely useless. And thus, deviations.”

Roy Jackson: “'You design the appropriate datum features to simulate those datums.'

"As long as the features are important to the function of the part.

"I've also seen the theoretical datums...pretty common in the automotive industry (all components are "in car position" and the primary datums are frequently the front axle for fore-aft and up-down, the centerline of the vehicle is the cross car datum. These datums may be several feet or meters from the component being measured. This actually makes sense from the vehicle design point of view, by the way.

"One thing I've always emphasized when determining features is the datum may be perfect (in theory), but the feature representing the datum may not be...any machined feature has some variation, for example. Even a relatively minor variation can result in significant error over the length of 3' (one meter)."

Norm Crawford: “All datum features are imperfect. That is the hole point of differentiating from theoretical datums and physical datum features.
“I understand vehicle coordinate systems. At a top level assembly, it makes sense to have such coordinates systems that are theoretically defined and if a company chooses to inspect at a top or sub-assembly level to define the datum features to establish the simulated datums for measurement. But the individual parts should have their own datum features that represent the functional assembly of the part with some exceptions in certain circumstances.

“Datum features should always be important to the function of a part. If they are not, then why are they datum features? Again, there can be exception with regard to manufacturing datums and that is an entirely new subject with regard to the purpose of the datum feature.”

To read the full discussion on Quality’s LinkedIn group page, click here.