[Rich05]

I am modeling up a guitar for a client. Currently my model consists of a point cloud and some lofted surfaces, and some rail surfaces. The model is highly organic.

Now my challenge is to get the surfaces to all meet up properly. I am thinking filets, meeting, joining ect. Is this the right way to go??

Is there any application for using meshes in a project like this? Why use meshes unless editing an stl?

Confused about how to get my model to have nice flowing surfaces and if meshes are usefull and why / how they might be used.

Thanks R.

[JDcam]

For something like a guitar, I think the process you are using is fine. A collection of points to create lofting ribs (or better yet, sweep rails AND lofting ribs), then editing the surface edges together should be just right.

It all comes down to the resolution of the scan vs. desired machining tolerances. Objects with greater detail require a higher scan resolution (lots of data points), where as objects with smooth gradual changes are better off with a low resolution (less points). There are times you may need to combine them for the same part, parts that are overall relatively simple but have isolated areas of fine detail, and blend the surfaces later.

When working with Rhino, avoid meshes as much as possible. In my experiences, even the mesh editing and mesh-to-nurbs plugins can't handle them all, and produce results no better than the method used above. If you have the money to throw around, there are obviously better solutions. If you are budget-limited though, you just need the time and patience of turning point data into a surf 'manually'.


Off topic but related: Has anyone experimented with using BLENDER for ceating meshes from point clouds, mesh editing and possibly conversion into nurbs surfaces? Before I launch into a book-buying spree and setting aside a few weekends to set up Blender and experiment, I was wondering if anyone has already done so.

[Rich05]

great advise thanks

So are meshes mainly done to be able to reduce the model for gaming and such? what tool is going to be my best friend to get all the lofted and rail, and network surfaces to match up?

do i just pull at the control point untill it looks right?

surface match, merge, join seem to only it the surfaces so far?

also do i join the whole surface when I am done and have the model all together?

this model is highly organic.

R.

[michael_e]

For highly organic shapes I'd suggest getting the demo copy of T-Splines for Rhino and using it. It's a bit spendy to buy, but the demo works for 30 days and it's really slick. Here's a link to their site: http://www.tsplines.com/products/tsplines-for-rhino
It's a bit difficult initially since they don't have a manual to speak of, but the tutorials and hacking at it for bit gets you going pretty well. They have a "manipulator" tool which allows you to move mesh nodes or nurbs control points with fantastic precision...very handy.

Oh, as far as using meshes. I find that they are very useful when doing things like closing gaps in nurbs surfaces...things like closing up naked edges and so forth.

Michael E.

[JDcam]

Joining surface edges can be a nightmare sometimes. It really depends on how the surfaces were made. The simplest approach is to extend the surface edges so they overlap, then trim them to each other. If you're building the model yourself using lofts/sweeps, try to extend your contours/ribs so that they overlap. Sometimes this option isn't available, or it simply doesn't work too exact tolerances. Depending on your machining needs, sometimes 'close' is close enough. A .0005 gap is inconsequential when you're using a .5" ball mill on it.

You should also experiment with surface blend, and surface fillet if you can. Depending on the transition, they may work wonders for you.

I use JOINEDGE when I need something for visual purposes, but it's really only a visual 'hack' for rendering purposes, not a physical join. Obviously, the goal is a 'watertight' polysurf, but sometimes it's simply mathematically unachievable with complex organic models in Rhino.

I'll have to play with using meshes to fill such voids next time, Michael, thanks for the tip.

And I demo'd tsplines briefly, it looked great, but I simply haven't had the time to sit down and really hammer on it yet. Based on the hour or two I played with it, it's definitely something that requires some experience to use effectively. Based on some projects coming down the pipeline, I'm hoping to set aside some time to play with it more for them.