Simple Boolean Diff Issue

 From: Oliver (FRITZ57) 19 Nov 2016  (1 of 13)
 Hi, simple issue : substracting solid from solid and keep the result solid (!) Any idea ? TIA, Oliver. Oliver o5m6.de Beware of the hun coming out of the sun ;-)

 From: Karsten (KMRQUS) 19 Nov 2016  (2 of 13)
 8186.2 In reply to 8186.1 Hello Oliver, >> substracting solid from solid... Dosen't look like two solids;-) Please explain: What is base object, what is substractor and what should the result looks like. And if the file is not confidential - please post it - so everyone can have a look and maybe find a solution. Have a nice day Karsten p.s.: Und ich dachte, nur der rote Baron kommt aus der Sonne ...

 From: Oliver (FRITZ57) 19 Nov 2016  (3 of 13)
 8186.3 In reply to 8186.2 Hallo Karsten, >> Dosen't look like two solids;-) Please explain: What is base object, what is substractor and what should the result looks like. >> And if the file is not confidential - please post it - so everyone can have a look and maybe find a solution. Base object is the cab, cut object is the cube, both solids. Purpose : Cutting the window(s) out of the cab side(s). 3dm attached. mfg Oliver. >> p.s.: Und ich dachte, nur der rote Baron kommt aus der Sonne ... nein, auch unsere Jäger in der BoB Oliver o5m6.de Beware of the hun coming out of the sun ;-) Attachments:

 From: bemfarmer 19 Nov 2016  (4 of 13)
 8186.4 In reply to 8186.2 The cab appears to be very, very thin, with two tested overlapping side pieces being the same to 7 decimal points, (or maybe more). So maybe scaling up everything by 10 to the power 7 or so would help? - Brian Edit, scaling by 10000000 did not reveal any difference in the dimensions of two tested overlapping side pieces of the cab, so there seems to be no thickness, so the "Solid" designation seems to be faulty, or else ??? I did not try 15 decimal places :-) How is such a thin cab to be manufactured? Even gold leaf only thins down so far:-) EDITED: 19 Nov 2016 by BEMFARMER

 From: Karsten (KMRQUS) 19 Nov 2016  (5 of 13)
 8186.5 In reply to 8186.3 Hello Oliver, I think Brian is right! The wall thickness seems to be lower than the internal tolerance of Moi for identical points (I remeber to read 0.005). Is it created with Moi or an import? -Karsten

 From: Michael Gibson 19 Nov 2016  (6 of 13)
 8186.6 In reply to 8186.5 Hi Oliver, yes it won't work well to have such an extremely thin wall thickness, if your thickness comes too close to around 0.001 units it will confuse the intersection process because it will consider both sides of the solid to be close enough to one another that they are considered overlapping each other. - Michael

 From: Oliver (FRITZ57) 19 Nov 2016  (7 of 13)
 Thank you, guys ! What would be the best way to increase the cab wall strength ? Tried Shell command, but to no avail. Oliver o5m6.de Beware of the hun coming out of the sun ;-)

 From: bemfarmer 19 Nov 2016  (8 of 13)
 8186.8 In reply to 8186.7 The solid is difficult to work with, but after about 8 repeated tries: There appears to be actual bevel thickness on the delta side of the cab solid. There are two copies of the cab. One has a skinny delta shape, which will not quite match up with the other cab. There would be a narrow sliver of a gap, between the "wrong" surfaces, or overlap, near or at the delta shape. First of all, color the box magenta and hide it. Secondly, Separate the solid cab. Pick the "inside" surfaces, one by one, and color them say red. Join the red surfaces. If more than one red joined surface occurs, some of the surfaces are the wrong ones, with a narrow gap or overlap, so swap a surface at or by the delta, until the red surfaces are "continuous", without gap or overlap, and will join to one joined surface. Hide the red surfaces. Color the remaining "outside" surface say cyan. It should Join to one surface, if the red one joined to a different one joined surface. Now the magenta box will trim the red cab. The red cab will Offset Shell just fine, either outwards, or Flip inwards. - Brian EDITED: 19 Nov 2016 by BEMFARMER Attachments:

 From: Frenchy Pilou (PILOU) 19 Nov 2016  (9 of 13)
 It's not more simple to make a simple scale *100 for all objects ? Make the Boolean operation Then Scale * 0.01 EDITED: 19 Nov 2016 by PILOU

 From: Michael Gibson 19 Nov 2016  (10 of 13)
 8186.10 In reply to 8186.9 Hi Pilou, re: > It's not more simple to make a simple scale *100 for all objects ? > Make the Boolean operation > Then Scale * 0.01 That's probably not going to help for this case - on one side of the object there doesn't seem to be any thickness at all, the inside and outside surfaces on one side are directly overlapping on top of each other. That kind of self-intersecting object will confuse the booleans. @ Oliver - does Brian's file above solve it for you? - Michael