Another Booleen Problem

 From: Brian (BWTR) 21 Jul 2008  (1 of 14)
 Michael Is there any reason why this array will not booleen join please? Later, refer the other thread, I want to booleen remove this from a sphere to make grooves. (I forsee a problem with that also having tried!) Brian EDITED: 31 Dec 2008 by BWTR

 From: Frenchy Pilou (PILOU) 21 Jul 2008  (2 of 14)
 Boolean union works but before the array you must rotate the object for have the "profil" not on the top!!! EDITED: 21 Jul 2008 by PILOU Attachments:

 From: Brian (BWTR) 21 Jul 2008  (3 of 14)
 1801.3 In reply to 1801.2 Tried that Pilou. Still does not appear to work though? Brian

 From: Frenchy Pilou (PILOU) 21 Jul 2008  (4 of 14)
 1801.4 In reply to 1801.3 I believe that the number of rotation array you ask surrimpose themselves and blocks the boolean process! With 3 or 5 that works fine! For 6 or 8 there are overlaps!

 From: Brian (BWTR) 22 Jul 2008  (5 of 14)
 1801.5 In reply to 1801.4 Thanks.

 From: Michael Gibson 22 Jul 2008  (6 of 14)
 1801.6 In reply to 1801.1 Hi Brian, > Michael > Is there any reason why this array will not booleen join please? There's a few different problems in here... Like Pilou mentions, you actually have some duplicate objects that are stacked up right on top of each other, that definitely won't help. Actually the top one seems to have 3 copies stacked on top of each other. Also like Pilou mentions, having the "end seams" of the surfaces all in one area will just magnify the difficulty for the booleans to figure out which piece intersects with what... Like this end: That's quite a number of edges all sticking kind of partially through each other... Things like that will cause difficulty for the booleans, as it progresses piece by piece, each piece will tend to create little tiny slivers where these things kind of barely graze by each other. Also some of the pieces are not quite aligned in a fully symmetrical way, like here: One piece seems to be sticking out a bit more, if you have more precise regularity to the pieces it would probably help. But you mention that you want to boolean it from a sphere? Please take a look at Danny's illustration here: http://moi3d.com/forum/index.php?webtag=MOI&msg=1792.12 Building the sphere in one segment first and then replicating those grooved segments will be a far more robust way to create that kind of thing because it will avoid this kind of "barely grazing" type of curved overlaps that cause problems. - Michael

 From: Brian (BWTR) 22 Jul 2008  (7 of 14)
 1801.7 In reply to 1801.6 Thanks Michael. However, you may like to look at the two 3dm files and the screen grab? Something odd to me-----why there were those duplicates and out of allignments may be a bug? Brian PS---I think I see the problem, I should have set the angle to 180 not left it at 360! AaaaGhhhhh! EDITED: 31 Dec 2008 by BWTR

 From: Michael Gibson 22 Jul 2008  (8 of 14)
 1801.8 In reply to 1801.7 Hi Brian, yeah if you have something that is closed if you replicate around a whole 360 degree circle you will get duplicates. For alignment I'd kind of need to see your original curve framework to see what is going on there. - Michael

 From: Brian (BWTR) 22 Jul 2008  (9 of 14)
 1801.9 In reply to 1801.8 AND, it should have been 135 deg NOT 180! Not a Brian day is it! Brian And here is the setup for the original sweep. EDITED: 31 Dec 2008 by BWTR

 From: Brian (BWTR) 22 Jul 2008  (10 of 14)
 1801.10 In reply to 1801.9 It was a long, slow learning curve but eventually, with segments etc I think this is close to what I had in mind. The second image shows the first failure! Brian EDITED: 31 Dec 2008 by BWTR

 From: Michael Gibson 22 Jul 2008  (11 of 14)
 1801.11 In reply to 1801.9 Hi Brian, > And here is the setup for the original sweep. That one seems to be looking good. What I noticed in the file that you sent was that other parts were not centered around the origin, like here: Notice how that ring is all to one side of the axis line instead of being centered around it? - Michael Attachments: