BUG in mirroring features?

[fritter63]

See video for explanation, I'm not getting all features completely mirrored.

[fritter63]

Oops, can somebody move this to the modeling forum?

[lichtzeichenanlage]

Is it possible that your plane (mirroring line) is not in the center of your object?

Edit: Would be nice if we could get an export of your design or a link to your project.

[SaeedHamza · Accepted solution]

@fritter63

Try redoing it setting the compute option in the mirror dialog box to optimized

Regards

[fritter63]

@SaeedHamza that did it!

Now the question is.... WHY???? Is it a bug? or just really bad design? Why would one option fail to copy only parts of just one feature?

[TrippyLighting]

Here's the documentation that describes these options.

[fritter63]

thanks @TrippyLighting

I still don't understand the point of it though. 😉

[TrippyLighting]

@jeff_strater @James.Youmatz @innovatenate could you help explaining this behavior ?

[jeff_strater]

I can kind of explain:  There are several modes available for pattern/mirror, which vary among their performance and success profiles:

• Pattern Faces - this is the fastest.  You can select any combination of faces and pattern them, not necessarily those defined by a group of features. This method works by essentially copy/pasting the selected faces to the location specified by the pattern/mirror transform.  But, it is probably the least reliable in terms of success of the pattern - there are some cases where this will fail
• Pattern Features - in this method, Fusion is patterning features, not faces.  But, there are 3 levels of optimization here as well:
  • Optimized.  This is essentially identical to Pattern Faces.  Fusion uses the same capability as above, but the selections are in the timeline, and faces are pulled from those faces
  • Identical.  This option is a bit slower, but is more reliable.  It is not applicable for all selections.  This option copies the selected features "tool bodies" to the new location, then joins/cuts those to produce the final result
  • Adjust.  This option is the slowest, and supposedly most reliable.  In this method, the feature definitions are re-executed at their new positions.  This allows you to make patterns/mirrors that "adjust" to the geometry around them (for example if you have a "To Face" extrude on a slanted plane).  However, this can be quite slow.  In your case, I see that you do have this option selected, so I would expect it to work, and I'm surprised that it does not.  I would say that is a bug.  I will try to reproduce this and log a bug

My personal approach to mirror/pattern is:  I start with the fastest, and work my way up until it works.  Sometimes it takes a few iterations, but I'm always able to find an option that works.

Jeff

[laughingcreek]

For sometime now I've been operating under the assumption that pattern faces was the slowest.  That was conclusion me an someone from AD (forget who) came up with, when he was helping me on a problematic model.  That seems to have been a unique situation, as in other test cases patterning faces does seem to be faster after all.  Thanks Jeff, always nice to have some insight on how things work in fusion.