So i thought I start trying to model stuff in Blender and now I’m stuck already at a very early point. I created a pretty basic mesh and I’m trying to bevel some of the edges but both CTRL-B in edit mode and the bevel modifier in object mode don’t create nice rounded edges but a huge mess which reminds more of a Star Trek transporter accident than the shape I’m going for.
I created the mesh by adding a cube, resizing it, then adding a plane, resizing it, then joining the 2. Then I went to edit mode and deleted the top face of the cube. After that, I created edges connecting the corners of the plane to the upper corners of the cube and finally created faces accordingly. The blend file is attached.
I’d be most grateful for any comments what might cause the bevel mess…
Unapplied and non-uniform object scale. Tools and modifiers work in object data level so you’re looking at different mesh than what the tools see. Ctrl+A -> scale to apply.
Edit mode, select something from your main object, ctrl+L to select linked, ctrl+i to select inverse and then delete (X), separate § to its own object, or whatever.
And there is also internal geometry in your main object, which will totally screw up your bevel…
See how the face count for that object is 11 (after deleting the separated, free floating face)?
This should read 10, as you have 2 face rings with 4 faces each plus one on the top and one on the bottom.
The additional face is inside the object, making the whole thing non-manifold:
In addition to that, the normals on that object are inconsistent: Some are pointing inwards, some outwards. Tab into Edit mode, hit A to select all and then Ctrl-N to recalculate the normals.
In other words: Your bevel is a mess because your mesh is a mess with almost every possible modeling flaw coming together…
We are dealing with polygonal modeling, which is a surface type modeling paradigm. In Blender faces (triangles, quadrilaterals, n-gons) are polygons but the naming convention may change between applications and context. The important part is that when modeling, we use those to describe a surface.
The polygons need a direction (normal) so that we are able to describe a solid. An uncooked potato for example. Polygons themselves have 0 thickness, so when you model a potato by describing the surface you see, it’s the face directions that define which one is inside (solid potato) and which is outside. If you then flip the face directions, you’re telling that the whole universe is a potato with a potato-size air pocket in it.
It’s the mesh elements, vertices, edges and faces that we use for a surface, or could also have multiple surfaces. But each of the mesh elements need to be part of just one completely closed surface, even if there are several surfaces, that makes them manifold. Any deviation from that causes the mesh to be what is called non-manifold, meaning the mesh has errors.
There are many ways you can make a mesh non-manifold, but some of those errors are worse than others. A face connecting a surface through itself is one of the worst. An interior face breaks many tools. On the other hand if you leave open edges, meaning the potato surface is not completely closed, that’s not too bad. Sometimes even desired when models can consist of intersecting geometry where the hidden polygons get removed.
This will display the face normals as greenish vectors protruding from the faces’ front side in Edit mode. Other than that: If you look very closely at your mesh, you will see that faces look slightly darker when viewed from their “backside”.