Friday, June 4, 2010

Just another tip from your Uncle Lar...

Pro/E users; do you use external copy geometry features when transferring data from one model to another in an assembly? (Please say “YES”)

When you copy geometry do you copy all the surfaces of the entire model? (Please, for the love of God say “NO!!!”)

Nothing con-volutes a model more than a bunch of unneeded live geometry. It may be the easy thing to do (and it may look cool), but you should only copy the geometry you “absolutely need”.

There are a few reasons for this:

1. It’s simple, the more copied geometry, the greater your chances of failed features due to the original being modified.

2. Regeneration time. Don’t believe me? Try it both ways with a stop watch. Then consider that time difference if everyone who works on your product modeled the same way.

3. If all of those models have full surface geometries over-lapping each other you most likely are selecting surfaces that you do not intend to select. This could very well be creating circular references (talk about a convoluted mess).

These are only a few reasons you want to be efficient when using shared geometry. Don’t get me wrong, shared geometry is a wonderful tool. But like most wonderful things, it can turn quickly if you are not careful.

Just another tip from your Uncle Lar

4 comments:

  1. Excellent post, I'm with you on everything except the first paragraph. I teach folks that they should absolutely use a standard copy geometry when doing top down design, unless there's a reason that you need an external (model shared with many assemblies, need to apply mechanism constraints, etc.)

    The reason is that you can easily have your copied geometry misaligned with your model if you use an external. Because you decouple the model's location and the geometry transfer, there's no guarantee that the external copy geometry is constrained the same way as the component unless you actually assemble your component using the same CS that you used for the external copy geom.

    ReplyDelete
  2. Doug,
    Welcome. I appreciate your input.
    I am with you on the Ext. Copy Geom's need for a common csys. It can somtimes be a pain, but in most cases (at least for me) the common csys fits in with the design intent of large assemblies. I agree that standard copy geom is good for component to component exchange, but for items such as piping and routed systems you run the risk of missing parametric change.
    Also, another thing not many are aware of is that basic copy geom "tags" all of the adjacent components in the assembly. This causes intralink/PDM-Link dependencies (a nighmare to anyone who has ever experienced them).

    ReplyDelete
  3. "Also, another thing not many are aware of is that basic copy geom "tags" all of the adjacent components in the assembly. This causes intralink/PDM-Link dependencies (a nighmare to anyone who has ever experienced them). "

    My understanding was that if I create a copy geom in part ABC from part XYZ in the context of assy 123, then part ABC becomes a child of both part XYZ and assy 123. If I use an external copy geom, then it is only a child of part XYZ. Are you saying that it creates parent child relationships with other parts in the assy not directly related to the copy geometry feature?

    ReplyDelete
  4. Yes. That has been my experiance in the past. It caused something called a "circular dependency" in Intralink which prevented users from being able to check in. By "tagging" I mean that it had a very linear course of dependancies compared to the bridge type dependencies created with ext. copy geom. I hope that makes sense (it's hard to describe).

    ReplyDelete