See the cyan curve in space? See the cyan line that was “projected” on the plysurface? How? I am pulling my hair out trying everything. While I can get it using “Projectcurves”, see the red curve on the surface. It needs to be projected in the same way, in the exact same spot as the cyan curve on the surface.
I am trying to learn how to do it myself. Someone else did this CAD. I was just given it to me to learn how to do it myself.
check out the difference between project and pull.
project
will use a unique direction. Therefor all sample points are projected in parallel.
similar to the sun (nearly parallel rays) is projecting silhouettes on everyday objects.
pull
will use the surface normals - or in a reverse: each sample point will mapped to the closest point on the target surface.
looks like the cyan curve got projected to the polysurface.
there should be one parallel view (front, right, …) where both curves match.
by the way - the surface does not look like very proficient modelling.
if you need more feedback post a file.
kind regrads - tom
Help.3dm (465.2 KB)
I appreciate the help, I come from the Solidworks world and Rhino can be frustrating at times. I will play around with pull and project more with what you said in mind.
If you don’t mind, please tell how I could have made the surface better? Or what was done wrong?
EDIT: When I select project, I can select the whole polysurface as one when prompted. But when I use Pull, it separates the polysurface into the individual surfaces. When I continue with the command it shows separate line segments based on the individual surface and the lines don’t connect or line up
the curve was projected.
the green line/errows show the direction
setting a corresponding viewport with parallel projection - the 2 cyan curves will “fall” together:
it depends a bit, on what you re gona do with the surface.
it is possible to model this surface with a single nurbs surfaces, that is trimmed or has a so called degenerated edge.
it is also possible to model this surface with only 3 or 4 nurbs patches.
my critique on the given surface:
maybe a starting point for this kind of surfaces:
hope this helps - kind regards -tom
There is a lot to take in there so forgive me for relying late.
I can’t replicate the view of that curved line at this moment. But I will work on trying to get it.
As far as the patchwork of surfaces. Trying to understand your view, making a nice nurbs patches with nice continuum, what end result would that give me? Or why would I want to strive to make it a single nurb surface? At the moment the only way to get it to fit those lines is to do multipole surfaces like that.
I honestly would like to learn why. This was created and taught to me by a whole team who (in my view) seems to know a lot about Rhino. It would be cool to maybe bring something to the table that they don’t know about. It will make me look good! lol
The lines are all derived from points taken on that boat hull. So the surface would still have to go through all those lines. Once the surface is created, we never move it or manipulate it. Ever. So having better control points won’t help us. Although a smoother surface would. We always “roll” over the “U” and “V” lines with the viewport (works great with a space mouse!) to ensure they are smooth and no weird bumps or humps.
Dear “NewtoRhino”
read above link. check the rhino files posted there, lock at the curves and surfaces with the analysing tools.
search this forum for boat / ship hulls.
learn something about curve / surface analysis, understand _curvatureGraph and how to interpret _zebra.
A lot of stuff is explained in the help.
_curvatureGraph (Curves - also works for surfaces)
these are the curves from your model - it s hard to get a nice surface with them, as curvature is not nicely distributed / stable / or curves even have inflection points.
_zebra (surfaces only)
the reflection stripes should form a smooth overall look. The edge continuity is not tangent / has Kinks.
what the usage for Cad data ? visualisation ? model makeing (scaled ? 1:1 ?), cnc- milling … ?
for the view:
set a custom cplane by normal.
use both end-points of the cyan curves
(now turn on project)
set cplane by x-Axis (“rotate the initial custom cplane”)
_plane command will set a parallel projection view onto the cplane.
Surface fairing for a boat hull or any hydrodynamic shape can be accomplished through many means. Here is an article with perhaps too much information:
I used the method in the article and got this result. It is a very quick run, so I expect there needs to be a little more editing.
To make it even better I would edit the control points just a little to fit the intersection curve with the base just a little closer to the orignal base curve.
Model included.(upload://yaPL0mDQDlAAXCFNzsoupwPRvKN.3dm) (633.4 KB)
Because this is a single surface the project and pull onto the surface will make clean curves.
Tom_P:
I understand zebra and curvatureGraph and other analyzing tools. Used them in the past. But we don’t use them in this team I am on. None of us see the benefit. That’s way i ask, what benefit will it give me making that surface a single NURBs smooth surface?
I am making boat covers, canvas. Those individual surfaces get turned into a polysurface and then flattened, cut and sewn. To make sure all the surfaces and lines are G2 or close to that would require an extreme amount of time added to the project (I assume, although I could be wrong.) As these points were taking with a 3D digitizer, the Proliner. Granted, not the best job taken nice points. Intern. I actually cleaned up the points quite a bit before I posted that file.
In most cases the surface HAS to be split up as textiles won’t stretch that far.
Scott:
Thank you for the Fairing PDF. I am reading it now.
So this answers my question for the usage of the data.
As it is a double curved surface and the unrolled / flattened version always is an approximation there will be tolerances anyway.
One benefit of a single surface will be less issues if you want the seams not identical to the initially captured curves. if you flatten a non tangential edge, it will result in a kinky cutting pattern at the former edge.
as you do not night super high precision:
_networksSrf with a quite rough tolerance for inner curves might do a good job - or even _patch.
you might also use _networksrf for an initial build, and then _rebuild.
_pointDeviation (combined with _divide or _mesh) can help you to measure the final deviation.
kind regards - tom
