Some insight on organic modeling

I’m working on a stove burner that was cast back in the 30’s. I’ve made some progress as you can see, but I’m now fairly well stopped when in comes to modeling the port connection. Can someone suggest a resource to pick up some tips on dealing with this sort of problem?

I shall quote my experience with this type of objects: 1) use of a photogrammetry program to get an object reference mesh. 2) Extrapolation of the main curves NURBS object mesh via sections or using the snap summit 3) creating final object NURBS working mainly with surfaces or solids, appropriately cut or modified with the control points.
With objects that have many connections with variable radius is almost never a good idea to work with the classical system (create curves> surfaces and polysurfaces> connect).
To simplify, in these cases it is better to use appropriate tools Rhino namely the T-splice plugin, unfortunately only for Windows.

Yeah, I can see that path as a difficult process. I was thinking perhaps an approximation of the form and then a cage edit but I need to learn how to add and subtract control points as needed for developing the final shape.

Maybe this can help.

I personally would model it with Subdivision Surfaces in a polygon modeller where such forms are relatively trivial and then convert to NURBS with something like T-Splines.



It’s a little indistinct. In the third picture, I don’t even know what I am looking at : ) The pattern maker went artistic with some files.

Looking at the first picture…

I would make the the flange and that shape that affects the side first, that’s blended it onto the cylindrical side of the burner. The solution might even be a surface, node twisted and mangled–boolean subtracted from the first part. Having the holes already in it may to activate Rhino’s WTF are you trying to do system. You might have to experiment with various angles to get the fillets result to match.

The outside of the port can be added later, and it can be hollowed afterwards using a shape to be subtracted.

It may sound strange, but spending a few moments feeling the part might help.

The enclosed file shows how the pattern for a casting like this might be made.
Burner.3dm (401.5 KB)

No pattern maker would make this with T-splines or subdividion surfaces. This is pretty straightforward for any standard CAD program

Thanks, that was helpful.

I like Kyle’s vimeo tutorials.

If you search there’s one for dragonfly that might help as well.
I’m also using mac and too bad we don’t have tspline plugin yet… it seems great for organic stuff…

What is that program, and is it affordable? Details?

My advice is Photomodeler scanner (for automatic operation and manual) and PhotoScan Pro (for automatic use). alternatively you can use Windows Rhino

I use PhotoScan, price $179, rather than PhotoScan Pro, $3499, for boats and similar objects. I also have Photomodeler Scanner but PhotoScan is the software I usually use.

Photogrammetry is a good tool to obtain points on a surface; either a set of discrete points or a point cloud provided the surface has sufficient texture. However some skill and knowledge are needed to convert the points to a Rhino model. It is not automatic.

Perhaps I misuse My English. For automatic, I mean that after entering the photos, it is possible with PhotoScan, automatically create (with 3 steps) a mesh that can be imported directly into Rhino and used. To underline that there is a Mac version of PhotoScan (as opposed to Photomodeler that is Windows only).

I agree that it is frequently possible using PhotoScan to automatically create in a few steps a mesh of an object with sufficient visual texture which can be exported in a format which Rhino can import. However sometimes difficulties arise which require some user input and additional steps. The reality of photogrammetry is not always as simple as many potential users believe.

[quote=“davidcockey, post:12, topic:32141”]
However some skill and knowledge are needed to convert the points to a Rhino model. It is not automatic.
[/quote]This comment referrs to what happens after the mesh is imported into Rhino, not creating the mesh. The same challenges apply to importing any mesh, not just a mesh from photogrammetry. The user needs to know how to create a Rhino model with the general shape and features of the object being modeled.

If you sign up for Kyle’s level 2 class you will get this in no time. Draw the curves with as few points as possible, loft the curves, and then connect the surfaces using the blendsrf command. I think he does this in the sunglass demo but if you can find his tutorial on building a desktop mouse you will see what great workflow this is.

I’m kind of new to Mac Rhino myself so I don’t yet have the other tools or experience yet, but this workflow has been very fruitful for me at least for now. There’s also Brian James’s short tutorial on complex transitions which may be helpful as well:

Thanks for the tip.