Closing opening problem

 From: Brian (BWTR) 9 Feb 2008  (1 of 8)
 A wierd one. For some reason, going back through several approaches, I can not get this opening to be sealed/closed Helppppp!! EDITED: 31 Dec 2008 by BWTR

 From: Michael Gibson 9 Feb 2008  (2 of 8)
 1359.2 In reply to 1359.1 Hi Brian, instead of trying to construct a cap directly in place on that 3D curve, it is a lot easier to create a more simple capped end and use a boolean to cut it and that keeps everything as a solid. So for example, instead of creating a wavy hole at the end, construct your object to come out to a simple flat cap like this: Now to get your curved end, draw a curve in the side view: And then use Boolean difference to cut the solid with the curve. Here I have dragged the piece apart slightly: Just delete the piece you don't need. Often times it can be a lot easier to build a more simple version of your shape that is extended somewhat, then cut it back to get your final result. It can be a lot more difficult to try and work directly off of a more complex boundary hole. There is another approach you could use - you happen to already have 2 edges in that opening, so you could select both of them and then use Construct / Loft to make a surface through them. But that surface will be collapsed down to a single point at the top and bottom, it will tend to cause problems later on with fillets and such. The simple cut that I showed above is far more fillet friendly. Hope this helps! - Michael Attachments:

 From: Brian (BWTR) 9 Feb 2008  (3 of 8)
 1359.3 In reply to 1359.2 Thanks Michael. Even so, I am surprised that I could not find a closing solution as it is? One of those odd ones I guess.

 From: Michael Gibson 9 Feb 2008  (4 of 8)
 1359.4 In reply to 1359.3 Hi Brian, > Even so, I am surprised that I could not find a closing solution as it is? There is actually a solution as it is - you can select the 2 edge curves and then make a Loft between them. But like I mentioned previously, that will create a surface that is collapsed / pinched down to a point at either end of it which will cause problems later on. That's why the other method is better. It may be possible for me to add a tool in the future to fill in a hole like this where an extrusion will fit it. That takes quite a bit of extra analysis in the hole filler code - right now it only knows how to handle a flat planar hole. - Michael

 From: Brian (BWTR) 9 Feb 2008  (5 of 8)
 1359.5 In reply to 1359.4 Michael in those two posts of yours I see maybe where some of the previous filleting problems I have encountered could have arrisen from. I think there is a "better to do this than that" set of "basic" rules item here but I dont know how to put in words myself? Here is the end rsult of the fly spray can which was good fun--except for that puzzle!-- EDITED: 31 Dec 2008 by BWTR

 From: Michael Gibson 9 Feb 2008  (6 of 8)
 1359.6 In reply to 1359.5 Hi Brian, > I think there is a "better to do this than that" set of "basic" > rules item here but I dont know how to put in words myself? I think some images will help! So say you have 2 curves like this (here just one is selected): You can select both of them and use Loft to make a surface there. But such a surface is collapsed at the points where the curves touched each other. To better explain it, imagine that the surface is a kind of gridwork of curves (that is basically what NURBS surfaces are actually). The grid of such a surface will look like this: Notice how there is a collapsing down to a kind of "pole" area at the top and bottom where the curves touched. These pole areas tend to cause problems with more advanced calculations such as offsets and intersections (both of which are used in the process of filleting). With freeform surface generation commands like Loft, it is easy for there to be little tiny ripples and bumps right in that pole area. That causes a lot of confusion about the precise direction of the surface normal in that area, and that's what causes these problems. Some other kinds of surfaces like spheres or surfaces of revolution also have these poles, but they tend to be much more precisely formed at that area and without little wiggles or bumps. So they are somewhat more behaved despite having these poles, although still it can sometimes cause problems even then. Now here is basically what happens with the alternate way that I was describing. In this method things start with a central flat curve like this: Then that gets extruded out to a surface: Then that extruded surface gets trimmed to an outline: That will create a very similar shaped surface. However, if you examine the structure of this surface, it has a lot more simple rectangular kind of layout to its grid, like this: That's because when you trim an object, new trim curve are calculated with mark areas of the surface as being "holes", but the surface data is still there underneath the trim curves. A trimmed surface like that has far more "stable" properties to it compared to the pinched one - even if it has some ripples or bumps in it, it doesn't have any spot where the surface might actually fold back over on top of itself, that's kind of what can happen sometimes with the pole situations unless things are very exact in that area. So the thing to avoid is creating surfaces that squish down on themselves, unless you don't plan on filleting it. If you plan on filleting, it is better to have things like extrusions which generate more rectangular grid type surfaces and then have trims that cut away parts. I hope that might help? - Michael

 From: Brian (BWTR) 9 Feb 2008  (7 of 8)
 1359.7 In reply to 1359.6 Beautifully explained. Now printed out to add to my MoI special help file. Brian

 From: Brian (BWTR) 9 Feb 2008  (8 of 8)
 1359.8 In reply to 1359.7 And here is the end result. With EASY filleting! Aahhh. EDITED: 31 Dec 2008 by BWTR