Surface normal from a curved surface

I'm new to GH and I'm sure this is a very simple solution, however, I am unable to resolve the issue.

Essentially I am attempting to find the surface normal for all UV corner points in order to offset a second set of points (each with specific offset distances) in the correct direction.

The result, however, seems to be random in terms of angle and direction.

Any suggestions?

Thank you in advance,

David

Attachments:

Replies to This Discussion

Permalink Reply by Mårten Nettelbladt on March 4, 2010 at 2:22pm

Hi, can you please post your .ghx definition as well?

Permalink Reply by David Bourque on March 4, 2010 at 2:47pm

sure - here is the definition

Attachments:

surface normal.ghx, 136 KB

Permalink Reply by trevor.patt on March 4, 2010 at 3:27pm

The problem is that "V"-vertices coming from Explode BRep component are in XYZ-space, but after going through the List-Item component you are using them to evaluate the surface as though they were UV-coordinates.

Instead of using Divide>SubSrf>Explode>Item just plug your sliders directly into Surface Frames. You'll get UV-coordinates as an output, put them into Eval Surface to get XYZ-coordinates for your vector Display anchor. Grasshopper will even let you cheat and use the Frames (F) from SFrames immediately as the Normal Vector.

Alternatively create the divisions by using a range (0 to 1) and when you use Eval Surface, right click 'S' and select reparameterize.

Attachments:

surface_normal_fix.ghx, 143 KB

Permalink Reply by David Bourque on March 4, 2010 at 4:05pm

Thank you for the advice!

The vectors seem to be much more controlled with the Surface Frames, however a few still seem to be missing the mark, and all points are showing multiple vectors.

Is it perhaps something to do with the multiple points?

Attachments:

surface normal_vector_03.jpg, 354 KB
gh_def_srfnrml_02.gif, 52 KB

Permalink Reply by Mårten Nettelbladt on March 4, 2010 at 4:21pm

Hi David, try disconnecting your SubSrf/Explode/Item/Pt chain and only use either the "SFrames" or the "Eval" technique described by Trevor. Good luck /Mårten

Permalink Reply by trevor.patt on March 4, 2010 at 4:23pm

Right. The number of points you are using as anchors doesn't match the number of vectors you have, so certain vectors are being duplicated (remember vectors don't have any defined position in space). Most points are showing multiples because a lot of your points are coincident (the top-left corner of one subsrf is the top-right corner of the next, etc..) Simply use the points (P) from the Eval component and everything will match up.

Note also that the SFrames gives you a 4x4 grid of points in this case (0,.33,.66,1) whereas the SubSrf gives you 3x3.

Permalink Reply by David Bourque on March 4, 2010 at 4:40pm

Great - it is working perfectly now. Thank you so much Trevor and Marten for the great info!

Best

RSS