I want to fillet the bottom edge of every half sphere. But I am struggling to only select the correspondng curves. I tried using nodes to check if something was curved or not, but since the spheres intersect with a cylinder, every edge is curved. Is there a good way to get the curves I want?
In the blue group there is an example of how the edges are selected for the fillet.
In the yellow group the criterion for selecting all the edges involved.
Here only the top and bottom circles are deleted from the selection.
There is also the coincidence of the lower edge with edges 905, 906… which gives an error.
A more accurate realization of the initial geometry would solve many problems. Fillet_question a.gh (873.8 KB)
Could you describe more about what you mean of “more accurate realization of the initial geometry”? I am new to grasshopper so any suggestions are welcome.
In this case there are only two small improvements to make.
Leave enough space to allow the fillet.
For example between the base and the hemispheres.
The presence of the seam can cause problems with the fillet.
The best thing is to redo the brep.
I don’t know the criterion you used to position the hemispheres.
In any case, cut the part of the base sphere that contains the seam, so there shouldn’t be any problems.
After I saw the first image I just couldn’t get that idea out of my head. I gave up trying to figure out how to do the fillets, but then I realized that for my purpose fillets were not needed. So I made this:
The hemispheres make a particularly nice decoration for 3D printing because insetting them just a small distance eliminates the need to print overhanging (or nearly so ) loops on the outside.
My thanks to Drew for showing this concept - I’m going to try using it on more complex base shapes.
I too am 3d printing them, and the whole point is to be able to have the tops of the spheres open while still allowing for a vase mode print. You are correct that the fillet isn’t needed, but I am just trying to learn new grasshopper skills and selecting these edges has posed a significant challenge for me. So that’s why I keep trying
If you have more cool shapes you come up with, I would love to see
Well that’s an interesting idea for sure. It seems to me you’d have to have fairly large spheres to make that work. And that would require a base shape that would be too large for my printer.
This is the next level of shape I’m working on (constant convex curvature) before I try one that has both convex and concave curvature.
Placing the spheres on points on horizontal planar curves doesn’t work, so I’m thinking I’ll switch to one of the Morph components so I can use the surface’s normal vector at each location to properly orient the hemispheres.
Well I just don’t have the patience for a 0.1 mm nozzle - so I use the standard 0.4 mm ones. Even with that diameter I found vase mode to be pretty fragile, so I use a standard wall thickness of 3 loops, which typically ends up being at least 6 because I almost always have an inside and outside surface.
Making the base shape is actually quite simple: it is a Loft surface formed from 3 circles at heights of 0, 50, and 100. The top and bottom circles are 50 mm radius, and the middle one is 65mm. Right now I’ve got an unnecessarily complicated method for making the circles, but I’ll be happy to share my GH file after I fix it up a bit more.
I opted to use the Orient component instead of one of the Morph ones because I think Orient processes faster. So with that my result looks like this:
I think that is a bit better than before, but still probably not as good as a Morph would do. However, it’s not clear to me that normal people would notice the difference (whatever it might be), so I’m likely to leave it like this.
Sure - I have no problem sharing my GH file. I post all my STL files online for free anyway - so why not? But I reckon you’ll have quite a few questions about why my GH is what it is. So feel free to ask any question.
Right now the GH is in 2 parts because I added components to make shapes like this:
To avoid overhangs at the bottom I had to tweak a few numbers in the Gene Pool to increase the bottom radius a small amount, and also slightly reduce the XY value for the ScaleNU component. Overall I’m quite happy with the results.