Modeling for Downstream Processes

Engineering CAD models can be built with different purposes in mind. Generally the difference between different types of models is the way in which those models are set up to handle change. Models built for different purposes in mind can be built differently. Most CAD models are intended first to be easy to change as part of the evolution that comes with the product development process. There is a lot of inherent change during development, and the models have to be built to be able to flex in the directions needed by product development. I’ve seen companies where they make a development model, which tends to be sloppy and full of errors. And then they later come through and do a production model, which is cleaned up, corrected, and the “design intent” is purified.

Secondarily, models need to be able to drive 2D drawings, or MBD models with 3D annotations. These are really meant to help in downstream manufacturing processes. Someone will need drawing information to set up machines efficiently. Many non-manufacturing downstream processes do not require tolerances, but where tolerances are required their absence can actually stop progress on the project.

When I develop a new model, the sketches, features, dimensions and relationships I use to make the model editable (to create the “design intent”) are usually different from the dimensions I would use to actually build physical parts or make the model easier to work with given a certain process, such as sheet metal, injection molding or CNC fabrication. The dimensions and tolerances are established to allow parts to be assembled with other parts and interchangeable. These are valid every-day sort of purposes for CAD models with very different requirements that might actually call for a different model to be used for mold building vs quality inspections, for example.

But there are other purposes for CAD models. Often times as part of an analysis you might want to try several versions of a design, and to create those versions through a method meant for the task (say configurations) you might need the dimensions and sketches set up a certain way. And that way is probably different from how you need the model set up for a different downstream task such as animation.

Sometimes a model has to be used to create a mold, and the mold cavity must account for the rate of shrinkage in the plastic or cast material. Parts that are bent are usually over-bent to compensate for spring back. Welded, soldered or brazed parts have to compensate for thermal expansion during the process. Sometimes you make a model that you know will need to be analyzed for mechanical or thermal stress, and you might build the model specifically to accommodate the types of changes that you anticipate you’ll need to make based on the experience of having done this type of work (and made this type of change) several times in the past.

I know I have made a variation of an assembly model when I knew I needed to do a motion analysis of that assembly. You can’t just use one single model for everything. Configurations are not robust enough to handle all of the types of edits that you need to make. Eventually you’re going to hit Delete instead of Suppress, and it becomes obvious you should have done a Save As instead of just make a New Config. Right?

Insert Part

If you are doing some analysis work on a single part, and you want the part to update with changes, but you don’t want to affect the original, use the Insert Part command. This gives you one way associativity. You can use direct edit (or what passes for direct edit in the SW realm) to make edits to the copy that won’t go back to the original.

For example, if you wanted to add a thickness to a part all around, insert the original design into a new part, select all the faces on a model and use Offset Face to add (or remove) a thickness.

Save As

For a motion analysis or an animation, save a copy of the assembly file, and make new mates specifically for the analysis without worrying about the production model used for drawings and other documentation.

You can do the Save As routine with parts as well, but I usually prefer the Insert Part. Save As creates a non-associative part with no link for updates. I generally save the Save As for assemblies, where all of the original parts are still linked, but the top level assembly is a different file altogether.

Configurations

I usually don’t support the use of configurations for this kind of thing unless the changes you make are small or predictable. For example, the use of a Scale feature to show a part scaled up for thermal expansion or other purposes.

Display States

If you are just making versions for display or visualization, you can use display states to turn parts on and off. If on/off is all you’re changing, you don’t need to add the extra burden of configurations. Display States will work just fine for that kind of thing. Configurations are needed to show assemblies in different positions or show different configurations of parts in the assembly.