Hello, I have a protoype case I am trying to design. I have a surface model of the outside of the shape that I want. When I do an offsetsrf to give it thickness, I am having some issues getting a closed polysurface. What can I do to get a better closed polysurface? Is the problem in the surface model or do I just need to edit the “almost” closed model?
Some general comments would be that you’ve got some areas there with ‘messy’ geometry that just isn’t going to shell well, though it looks like most of it is okay on one side and just needs to be mirrored over or duplicate how you did the ‘good’ side. Also your file tolerance settings are far too tight, don’t go any tighter than 0.001 for mm, 0.01 is probably adequate, while the default angle tolerance of 1 is too loose, it should be 0.1.
Also, don’t hide your isocurves while working, you need to see what your surfaces look like.
Jim’s suggestions about absolute tolerance and angle tolerance are good.
Some of the curved surfaces in the initial model could be much simpler. I would rebuild the surfaces, but depending on the requirements it could be possible to fix what you have.
How much experience do you have with Rhino? With 3D design in general?
Thanks David,
I will work on rebuilding the surfaces, it will probably be faster for me. This surface model is originally from another that was twice as big and built by someone else on solidworks. I shortened it up but the messy stuff comes from the original file.
Experience. Been playing with Rhino since V4 was out. AutoCAD since V10 back in the early 90’s. Never learned more than the basics. Usually can make the part fit, just need to use a bigger hammer 
I really appreciate you and Jim taking your time helping people like me out.
Kevin
thank you Jim, I will make those changes
prototype left 002.3dm (6.2 MB)
Rebuilt it, but I am having problems with the complex radii and it not offsetting well. Suggestions on how to rebuild it cleaner?
How are you making these fillets?
I did a bit amateurish approach and failed to make it good enough, but I guess it’s still usable for 3d printing. Edge tolerance was quite nice at less than 0,000001 mm before splitting the two bottom areas at the end, but then it dropped to 0,0002876 mm. I changed the file’s distance tolerances to 0,001 mm and angle tolerance to 0,1 degrees.
prototype left 001 - Bobi.3dm (5.7 MB)
The video tutorial showing the complete modeling process will be ready tomorrow.
Edit: Still needs 8 more hours for processing of the 4K quality.
Hi Bobi
It looks good from the rendering. I was unable to open the file since I have not yet upgraded to Rhino 7. I will watch for the video to be posted. Thank you
Kevin
Because of the tight radius curves where I am having the issues, I have been using mostly sweep2 and sweep1. VariableFilletSrf and FilletEdge have not worked on anything over R10mm
What do you suggest?
Hi Kevin, I just exported the same model as Rhino 6 file. Hopefully it helps.
from the Isocurves, I think I understand how you did it. will try and replicate it tomorrow. my only question now is how you did the 3 sided corner? I used sweep1 but it did not turn out that clean
Thanks again.
Kevin
I used a few of the original surfaces, including the same 3-sided corner, which I left unaltered. I would do it with the “Revolve” tool, because it’s the one way to get clean revolved surfaces with minimum amount of control points.
By the way, my video is processed to 360p by YouTube at the moment, but it says that the 4K quality will take another 8 hours. But in a few seconds I will add the link to the video in my first post just in case.
Thank you. I forgot about the Revolve tool. Doing something like this only a couple times a year I start to forget in my old age.
You won’t believe how many times I forget about the existence of certain tools, for the very reason you mentioned. I may end up creating a new custom toolbar will all the important tools collected together.
The ‘cutouts’ on the side are totally tripping you up here, you should remove them and do them after, before or after the shelling your choice. I rebuilt this from scratch, and where I ran into a limitation of FilletEdge I just used FilletSrf, which is how I do most filleting. There’s very little reason to ever be hacking what are supposed to be ball fillets with sweeps and the like, when you find yourself there it’s time to take a step back and rethink.
prototype left 002.3dm (1.0 MB)
The way to do a manual ball fillet corner with 3 or 4 identical radii is with the “hidden” test command–it’s not considered ready for user use yet, or is just for debugging use–called “testsphericalpatch.” It does the same thing as creating a sphere that fits the edges(Sphere with the “4 point” option) and trimming it off.
It’s illegal to hide such a useful command! No beer for the programmers unless they make it a proper Rhino tool. 
