Tim Culbertson

EPDM and Skeleton-driven assemblies

Discussion created by Tim Culbertson on Jun 9, 2014
Latest reply on Jun 9, 2014 by Jerry Steiger

I’ve run into a puzzler regarding EPDM and “skeleton” parts used for top-down design.

 

At our company, we’ve proposed a top-down design assembly structure incorporating a “skeleton” part which contains sketches and critical dimensions. This part is used to drive several “real” parts in the assembly in terms of geometry and part location. The coordinate systems for all parts are coincident in the assembly, then geometry is created in each part based off the reference geometry in the skeleton part. For purposes of discussion, suppose our assembly consists of Component A which sits on top of Component B.

 

Suppose we have Assembly1 containing skeleton1 which drives components A1 and B1. We design, validate, and then move these files to a read-only/locked state in EPDM upon their release to production.

 

Suppose now that we need to design a derivative Assembly2, slightly different from but existing in parallel to Assembly1. Assembly2 has a new skeleton (skeleton2) since it’s driving slightly different geometry for component B2 (let’s say it got an inch taller). Assemby2, however, re-uses component A1 in its original geometry.

 

Since A1 is locked, it cannot be updated to reference skeleton2 without an unnecessary/artificial revision cycle.

 

The geometry of A1 itself remains unchanged. However, in Assembly2 component A would have to sit 1 inch higher, so its skeleton-driven location has changed. As a result, Assembly2 has B2 driven by skeleton2 and A1 driven by skeleton1. The geometry for A1 is fine, but it appears 1 inch too low in the assembly.

 

I’m not even entirely sure we want to update A1 to reference skeleton2, because then the component will be positioned too high in Assembly1.

 

I’ve thought about trying to divorce the location of the part from the skeleton (i.e., have the skeleton only drive part shape, not its location, and then position the part in the assembly using normal assembly constraints rather than just having all components share coincident coordinate systems). This feels like it’s breaking from a top-down design approach however?

 

Surely other companies out there have solved this issue. Other people’s experiences?

Outcomes