Native GH vs Rhino Curves-Why Are Lofts Breps Instead of Surfaces?

Replies to This Discussion

Permalink Reply by Benjamin Anderson on May 28, 2009 at 6:11pm

Here's the Rhino file with curves if you're curious...

Attachments:

• Tower Massing Curves.3dm, 39 KB

Permalink Reply by Dirk Anderson on May 28, 2009 at 8:13pm

Sorry but you may need to upload the rhino file as well. If you're referring to (kinked?) rhino curves this would result in a Brep (poly-surface).

Permalink Reply by Benjamin Anderson on May 28, 2009 at 9:40pm

The Rhino file was uploaded in the first reply...but to pick up on your point-the curves generating the loft have hard corners...is this the cause of the problem? Is there a way to create a kinked shape with a continuous surface?

Permalink Reply by Dirk Anderson on May 28, 2009 at 10:02pm

strange... you're reply wasn't up when I posted?
ok, so theres a few issues - with polylines, you're always going to get polysurfaces with the loft unless you do a rebuild with the surface with a lot of points:

loft options: rebuild surface (100), straight sections

you could also re-draw the polyline as a nurbs curve with midpoints and change the weightings of the corners. the issue with this is the point count of each curve - since you have chamfers on some profiles it changes the point count of each level, resulting in a twisted surface. Not sure if you require it to be one continuous surfaces as you had selected straight sections on the loft options.

Permalink Reply by Benjamin Anderson on May 29, 2009 at 1:13pm

Wow Dirk. This is phenomenal. Thanks so much...ideally I'd like to solve the issue of the twisting, but nonetheless I appreciate the breakdown between polylines and the resultant surfaces they produce-I just didn't know that (still new to Rhino).

While the nurbs curve solution seems to work really well-I'm curious whether the twisting issue can be resolved by giving all the poly/nurbs curves the same amount of initial points? (GH isn't trying to tie a 6 point curve to a 5 point one, and so on...) It looks like GH / Rhino is trying to 'track' a point down the surface, and where there should be a sharp corner (2 points converging), it pulls away and connects somewhere else instead (see comparison to basic geometry from sketchup). Is there an opportunity to edit the start point of those lines within GH like a normal rhino loft (see second image)? Would that help solve the twisting too?

Attachments:

• Base Geometry Skin vs Rhino Skin.jpg, 154 KB
• Loft Curve Control Points.jpg, 224 KB

Permalink Reply by Luis Fraguada on May 29, 2009 at 6:39pm

Benjamin, the relationship between topology is imperative in Surface modelling...points to curves to surfaces, if you are making one from the other, there are always consequences. Its not a Rhino thing...your thinking about how many points a curve has and its resulting geometry is probably correct.

Permalink Reply by Dirk Anderson on May 29, 2009 at 7:08pm

Ben, you could try straight lofting for the external skin and then create a number of floor plate sections that you subdivide for your ordered point set:

Attachments:

• towerSlice.3dm, 75 KB
• towerSlice.ghx, 121 KB

Permalink Reply by Benjamin Anderson on May 30, 2009 at 5:05pm

This is great Dirk...but I'm not sure how to incorporate your cluster into my definition or how to use it further down the command chain. Is your cluster just an example, or can that be plugged into the definition (see my modified one...I've hot-rodded a previous definition to create variable diagrid shells on the surface..._) to use. Right now I can't make use of your definition, although it's pretty much exactly what I was trying to use to get a uniform breakdown of the surface...

For the time being I may just import the section curve for each floorplate (the base and tower vary) and use that to generate the form & point sets, which seems to me to be the same thing, right? Ideally it would be great for the whole thing to be controlled within GH_just using the baked form from Sketchup as a guide...but I'll have to toy around with this a little more to master, I think.

I appreciate the help so far.

Attachments:

• 09-0530_Tower Massing.ghx, 1.2 MB
• 09-0530_Tower Massing.3dm, 75 KB