Page 312 - Mastering SolidWorks
P. 312

|
        284   CHAPTER 9  PAtterning And Mirroring


                       You might hear a lot of conflicting information about which features are better to use in
                    different situations. Users coming from a 2D background often use a function such as sketch
                    patterning because it’s familiar, without questioning whether a better approach is available, and
                    often without having a way of measuring how it performs. When in doubt, you can perform a
                    test to determine which features work best for a given situation.
                       To compare the performances of various types of patterns, I made a series of 20 × 20 patterns
                    using circles, squares, and hexagons. The patterns are both sketch patterns and feature patterns,
                    and I created them with both Verification On Rebuild and Geometry Pattern turned on and off.
                    Verification On Rebuild is an error-checking setting that you can access through Tools ➢
                    Options ➢ Performance. Geometry Pattern is a setting that is applicable only to feature patterns,
                    and it disables the intelligence in patterned features.
                       Table 9.1 shows the rebuild times (in seconds) of solid geometry created from various types of
                    patterns as measured by Performance Evaluation. (It is found at Tools ➢ Evaluate ➢ Performance
                    Evaluation; this was formerly called Feature Statistics.) Sketch patterns are far slower than
                    feature patterns, by a factor of about 10. The biggest speed reduction occurs when you use sketch
                    patterns in conjunction with the Verification On Rebuild setting, especially as the number of
                    sketch entities being patterned increases.



                       Verification on Rebuild
                       When the Verification on rebuild (Vor) setting is turned off, a given face is tested against each of
                       its neighbors (for example, to see if the model is self-intersecting). When it is turned on, Vor tests
                       each face against every other face in the model. on a simple cube, the difference between on and off
                       is four tests versus five tests. on a larger model, it could well be the difference between four tests
                       and a thousand tests. the type of error this would catch would be, for example, if you had a box and
                       shelled it out, removing one face, and then you put a large fillet on an outer edge. You would see the
                       inside of the shell break through the outside of the fillet. if Vor were turned on, this would cause
                       an error. if Vor were turned off, SolidWorks would allow this, and you would have invalid geometry
                       with no error. to examine this in more detail, open the sample model Chapter 9 - VOR.sldprt.
                       of course, having Vor turned on for large models is very costly from a performance point of view.





















                       Generally, the number of faces and sketch relations being patterned has a significant effect on
                    the speed of the pattern. The sketch pattern times are taken for the entire finished model,
   307   308   309   310   311   312   313   314   315   316   317