PDA

View Full Version : Please cross check strange boolean artifacts



madno
07-24-2013, 01:38 AM
Hi,
I have some kind of roof geometry which I united by LW boolean command (LW 11.5.1, Win, 64 bit).
After done so, there is some strange geometry which I needed to fix by hand.
Did I made it wrong or is my geometry not valid for a clean boolean?
It would be nice if somebody can cross check.
Thanks in advance.

115789115790115791115792115793

Lewis
07-24-2013, 02:29 AM
Bell booleans are always messy and problematic in most of Polygonal software(s) so I'm afraid you'll have to fix manually lot of stuff :(.

Booleans are making much more sense when working with Nurbs but sadly we don't have that in LW :(.

JoePoe
07-24-2013, 03:36 PM
That Boolean is going to be messy.

But a fixed merge at 2mm seemed to clean it up .... for the most part. :)

madno
07-24-2013, 11:33 PM
Thanks guys,
at least I know now, it is not me doing something wrong ;-)

Amurrell
08-03-2013, 08:53 PM
Booleans also work better on closed geomety with thickness

spherical
08-03-2013, 10:48 PM
I'm a bit surprised that the Boolean worked at all; not being closed volumes. Granted, a Union or Addition isn't cutting anything but the process does go better when the function is given what it expects. Also, it helps to consider that, at least when cutting full objects, you cannot split a line or a point. They are infinitely small and cannot be split. So, overlapping volumes, such that cuts do not intersect a line in a manner that would try to cut the line thickness in half (length no problem, obviously), or points at all, are recommended for clean Boolean operations. A slight shift, such that no lines intersect, often improves results greatly; with less, if any cleanup.

I've also found that Speed Boolean ops sometimes are more successful than the original tool. No idea why but it has saved me a number of times from doing more work to obtain the same result after Boolean failed.