Machine Design

Better models, better translations

Much of the difficulty with IGES translations today comes from poor model building techniques on the part of designers, rather than from inadequacies in translation software.

So says International TechneGroup Inc., a firm which translates large numbers of CAD models between formats. Older surface modelers such as versions of Catia before V4 use the approach of sewing surfaces together to create solids. Using the same techniques with newer parametrically based packages creates difficulties in IGES or STEP translations, but also in any sort of downstream efforts that employ the solid model as a basis for operations such as FEA or manufacturing operations.

Here are a few examples of model quality problems that ITI personnel have uncovered, and practices that would have avoided the difficulty.

Some early IGES history

Early versions of the IGES spec were based on the Boeing Database Standard Format (DBSF) which was heavily influenced by the CAD systems in use at Boeing during the 1970s: CADDS 3 and Gerber IDS. Both used simple geometric elements and basic text and dimensioning abilities.

Ground rules for adding to the IGES specification were that an entity had to exist in three major CAD systems before it would be considered. Thus anything unique to a given CAD system would be excluded. Items in a CAD system that were special, probably also constituting a good portion of its competitive advantage, had to be converted to a different concept in IGES or be dropped altogether. This practice sheds some light on why useful features found in CAD systems have been slow to find their way into IGES.


Interior Faces
This solid is created from outer surfaces that are thickened toward the center. Because they are not perfectly tangent along the reflection plane, two faces are created inside the model. The angle between their common edge in the back is 0.0°. They overlap each other 0.005 mm at the lower, front corner. The solution is to have the right and left top surfaces defined by a profile revolved about an axis. Repositioning the axis onto the reflection plane makes the surfaces tangent so the thickening operation does not create interior faces.

A thin crack appears between a boss and the upright. The boss is defined as a linear extrusion that does not compensate for the draft angle in the upright. This crack makes complete fillets and rounds impossible to create. The solution is to constrain the boss to compensate for any draft angle in the upright. The boss is also defined before any fillets or rounds are applied.

Knife Edge
This fillet is defined as a profile swept along a curve and then joined to the part. The curve does not precisely match the shape of the part. The fillet primitive extends through the tip of the part at the end. The interaction of these two features creates a razor-sharp edge and a crack. The solution is to define the fillet as a blend surface between the side of the part and a reference surface extracted from the mating part below (not shown). The lower edge of the blend is used to trim the reference surface, forming a V-shaped, open shell that is joined to the part, leaving no spaces between the fillet and the part.

Realism-Pinched Face Microstep
The top section of this upright is defined as a linear extrusion feature joined onto the part. The front faces do not match. There is a 0.08-mm microstep between them formed by a single, pinched face along the bottom of the linear extrusion. The solution to the problem lies in making the profile of the top section part of the profile of the entire upright. The upper and lower front faces match precisely, and variable radius blends are completely defined.


Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.