I am trying to replicate a model of a rudder I made in Rhino using Curve Network and sweep 2 rails tools.
So far I am struggling to get anywhere near close to what I want. The loft tool doesn't seem to have the right functionality.
Any help would be appreciated.
Here is video in how you can build it in Fusion but it is not really easy to use compared to the ease of use in Rhino.
Fusion Team, if you see this there are certain problems in the sketch mode I mentioned at the end of the video.
Maybe I use it the wrong way, otherwise I think there are few areas where the system can be improved.
https://drive.google.com/file/d/0Byzv_NlyKp_2WnprYlp4LWpiRVk/edit?usp=sharing
Claas Kuhnen
Faculty Industrial Design – Wayne State Universit
Chair Interior Design – Wayne State University
Owner studioKuhnen – product : interface : design
Under the covers Sculpt Loft is creating a BRep surface and fitting a TSpline. The tolerance allows you to determine how you want to balance between accuracy and smoothness. The Extrude and Sweep use another method based on fitting the curve input. To be honest I think the Loft method is better but we need to update the fitting strategy for those commands.
BTW I played a bit with another method I would have used in the past with surface modeling. Those methods of lofting may be acceptable for some applications but the density of the controls makes it seem less attractive. In this alternative I created a base surface and modified the controls to find the intersection at the tip. This creates much smoother geometry. Of course it will still need to be converted and trimmed. I will try to get that finished and share the final results.
I'm very interested in the method you are showing here. I tried making this foil and found using T-splines very cumbersome compared to a surfacing program where you could define a net of curves. Lofting in Fusion seems to have problems if you want to end up with a midline termination as with the tip of this rudder.This is similar to the issue I had with the boat hull at I3D/ AU. Creating accurate surfaces based on cross section data is something I haven't figured out how to easily do in Fusion.
Claas Kuhnen
Faculty Industrial Design – Wayne State Universit
Chair Interior Design – Wayne State University
Owner studioKuhnen – product : interface : design
Here is a video that shows my method for creating the rudder surface. In NURBS surfacing, especially "Class A" design, the emphasis is on keeping the surface definition as simple as possible while still describing the shape. Using methods such as loft generally wouldn't be acceptable. One strategy is to try to match the boundary and shape and then trim the surface. I applied that to strategy in this case with the T-Spline and then converted it. It may not satisfy what you wanted to accomplish here particularly in terms of precision but I thought I would share the strategy because there was some interest.
I think this exersize also points out some of our shortcomings with T-Spline fitting to curve input. There are areas of overlap between the freeform shape definition and the manufacturable model that we need to pay closer attention to. Feel free to share your comments.
Thanks
Interesting approach - you can also an approach that simulates a sweep trick from Alias.
But it looks T-Splines does not yet allow you to extrude vertices into lines and create surface inserts (offsets)
Here is a link to a video outlining the missing tools.
https://drive.google.com/file/d/0Byzv_NlyKp_2Wk1fd2hHOHhyVms/edit?usp=sharing
thats why I modeled the geometry outside but it works great in Fusion when imported.
Claas Kuhnen
Faculty Industrial Design – Wayne State Universit
Chair Interior Design – Wayne State University
Owner studioKuhnen – product : interface : design
Hi Claas a Sweep might be more precise but Fusion only has 2 rail sweep right now. This is a very good approach as well but you have to do the sweep in another program then import into Fusion. It works for you great but most users might not have the resources to own that program plus Fusion360. But it is a very good precise way I agreed! Great work here!
OK bare with me. After much perseverance I have come up with a solution that is not 100% accurate in terms of how it matches the Rhino model but it would be acceptable for manufacture. It's a pretty long winded procedure compares with the curve network surface I used in Rhino but I'll try and explain step by step.
First I created a trailing edge loft using the profile edge as a rail.
The I needed to make sure both edges of the profile were G1 so I added a fillet at the corner. This was to be able to produce a full surface loft.
Then I created a loft using the Forward edge and Trailing edge surface as profiles and the foil section at the top and bottom as rails. The loft wouldn't work the other way round even though it would seem more logical. Using the foil sections as profiles and the edges as rails just didn't work.
Then I created another loft to get the bottom shape using the Foil section and point as the profiles. This loft wouldn't work with rails so I had to give the first profile a direction and judge the distance until it looked close to the bottom edge curvature. The point profile was set to tangent.
Then I merged the 2 T-Spline bodies and converted them into a surface mirrored it patched up the gaps and created a body.
So it was a bit of a long winded process and is probably not 100% accurate. Some of the steps don't seem all that logical either. But I think this is something that could be manufactured depending on what level of accuracy is acceptale.
Cheers
Russell
BTW - I tried the same process in the patch workspace and everything looked good up until the point where I tried to merge the 2 faces. It turned out real ugly.
Could this possibly be a bug??
Claas Kuhnen
Faculty Industrial Design – Wayne State Universit
Chair Interior Design – Wayne State University
Owner studioKuhnen – product : interface : design
Claas,
I think you are raising valid points about the capabilities you like in a polygonal modeler. There are obvious similarities between T-Splines and polygonal modeling but there are also some self imposed restrictions in working directly with T-Splines in manner you describe in Blender. The T-Spline topology won't tolerate extruding points or edges because it will create non-manifold surfaces. There have been discussions about providing SubD modeling with the ability to convert to T-Splines when required thus getting the advantage of the poly workflow until you need the T-Spline ability to convert to NURBS or isolate complexity. Your familiarity with polygons makes it an obvious advantage to have it all in one place.
I would also mention again on this posting that the extrusion or sweeping of T-Splines from curve data is not very effective. This is because the T-Splines are all degree three based surfaces and the curves we are handing it are higher degree requiring a fitting which leads to the trouble. There are solutions we can apply similar to working with T-Splines in Rhino so I appreciate the discussion on this posting and encourage you to make your requirements known.
Thanks for all your comments and discussion on this posting. It definitely helps highlight shortcomings that must be addressed.
Hi Paul,
That makes sense. To really harness the power of subD like modeling with T-Splines I think this should really be considered if possible.
The power of polygons is the way you can sculpt with it and working with poly lines is part of it.
Thats the main reason why I still do the modeling in Blender and not Fusion because I have more tools at my disposal there.
This particular model is already a good example where it makes it very complecated to build the shape if you do not want to use
a box model approach which will lead to typical subD rounded corners issue like in the following images if not desired:
Claas Kuhnen
Faculty Industrial Design – Wayne State Universit
Chair Interior Design – Wayne State University
Owner studioKuhnen – product : interface : design
A little mistake to be corrected:
In my last reply on page three of this discussion the phrase : "a Sweep might be more precise but Fusion only has 2 rail sweep right now."
Should be read as: " a Sweep might be more precise but Fusion only as 2 rail Loft right now."
Sorry about that. Too early in the Hong Kong morning for that reply!