When does shell create radii to hold uniform wall?

bruce_williams

@armindilo Has a nice elegant solution for intersecting pipes at https://cad.onshape.com/documents/f6fec9930959f5e092add16f/w/cd8884dce5ef97e6734a4977/e/651a8e3cd8465d8cbf73b4ef    see forum discussion at  https://forum.onshape.com/discussion/comment/24264#Comment_24264 

Shell working here was a surprise to me.  I expected radii would be created at internal intersection (see picture).  So my question is - What are the rules for shell?  When will a fillet/round be created to hold even wall thickness?

Jake_Rosenfeld

Hi Bruce,

This isn't my area of expertise, but I tried something else as an experiment.  I copied @Armindilo 's document and added a fillet to the joint between the two pipes before the shell and got the following result:



I suspect that our geometry kernel has a hard decision to make when faced with a non-tangent intersection, and solves it by extending the intersecting faces and taking an intersection of their shells. If you need constant wall thickness for something in the future, I would suggest adding a very small (or larger, appropriately sized) fillet to any non-tangent edges.

bruce_williams

I did some experimenting and answered my question.  A fillet is required to force shell to round the corner holding 'even wall'.  Now I see this is a very good behavior as it gives an option for the corner on inside.  Note the pictures below, I have the smallest fillet allowed (greater than 3.938 e-4; or 1 micron) and then shell creates a radius to hold wall thickness.  By suppressing the fillet, shell creates a sharp corner on inside of part.  NICE!  Have it your way....

Jake_Rosenfeld

Hi Bruce,

This isn't my area of expertise, but I tried something else as an experiment.  I copied @Armindilo 's document and added a fillet to the joint between the two pipes before the shell and got the following result:



I suspect that our geometry kernel has a hard decision to make when faced with a non-tangent intersection, and solves it by extending the intersecting faces and taking an intersection of their shells. If you need constant wall thickness for something in the future, I would suggest adding a very small (or larger, appropriately sized) fillet to any non-tangent edges.

bruce_williams

I did some experimenting and answered my question.  A fillet is required to force shell to round the corner holding 'even wall'.  Now I see this is a very good behavior as it gives an option for the corner on inside.  Note the pictures below, I have the smallest fillet allowed (greater than 3.938 e-4; or 1 micron) and then shell creates a radius to hold wall thickness.  By suppressing the fillet, shell creates a sharp corner on inside of part.  NICE!  Have it your way....

Armindilo

Nice catch and nice solution @bruce_williams.