Thanks. Make sense.
On my side I was on
I find it easier to know what and how if I use familiar basic terms/entities - vertex and edge offset.
For tangential and radial, I would first need to test to see what these do and then I would know.
You are right.
If you use bending on a real material the only thing that is preserved is the volume.
On the bend the material will get thinner and the inner radius is that of the tool.
The outer radius will be such that the volume is preserved
I disagree.
Volume remains nearly constant (material is only redistributed).
Inner radius is formed by the tool, but springback occurs after unloading
-> use a slightly smaller tool radius to compensate.
-> use overbending to correct for springback
The part gets slightly longer overall because the outer material is stretched more than the inner is compressed (neutral axis shifts inward).
This is the more precise answer - the first one was a simplification. I assume that you do the corrections for springback.
To me the main thing is volume preservation. I would ignore the elongation - it makes things too complicated
Actually, neither FredoBend (nor TrueBend nor ShapeBender) preserve the volume when bending. This is due to the segmentation of curves.
This may not be a problem since these tools are in most cases designed to create geometry rather than to mimic a real physical deformation. For instance, if you create a bent wall, you donât really care about preserving the volume of an hypothetic straight wall.
Preserving volume in a context where curves are built out of segments (i.e. not continuous curvature) is doable with orthonormalization. But I am not sure this is worth developing, as it may also alter the simplicity of the shape by creating triangulation.
Thank you @Fredo6 for answering the question that has been on my mind since this topic started: was the intent to create a physics-based bend or just a better geometric one. The latter.
The [originally shown] option A. must be âbetterâ because the start/end parts are 90° [keeping it the same as the original âstraightâ shape] - this is by using a half-segment to the curve path arcâs start/end.
In contrast option B. has the path arcâs start/end parts âskewedâ [when compared to the âstraightâ original] - this is because the curve path has a full arc segment at the start/end.
Because of the segmentation of curves, everything is approximate when it comes to preserve lengths, surfaces and volumes.
At least, this was the occasion to realize that Sketchup itself uses the two options when offsetting curves, depending on whether the curve is declared as a true Arc of Circle (or Circle) or something else. So, the Sketchup developers came to identify the two options A and B and made a reasonable choice to mitigate the approximation, based on the intention of the user (a circle will privilege the preservation of radiuses, otherwise the preservation of offset chord lengths).