# NOPE003: FreeCAD - "Macaroni Cut" - subtractive pipe Non-Obvious Practical Examples 003: "directional fillet", or cutting a softer bending radius for paracord looped through 3D printed holes. ## Starting Materials FreeCAD Version: 0.21.1 ## The Problem How to make nice holes for paracord attachment points in 3D printed objects where refined shapes come at no extra manufacturing cost? Cearly, the minimum bending radius of a chamfered hole is somewhat offensive: ![photo_2024-03-14_18-16-07](https://hackmd.io/_uploads/BykK-2x0p.jpg) Generally, curvature introduces uneven load sharing of the fibers in a rope, or the wires in a steel cable. Here are some climbing knots and the associated reduction in breaking load: ![grafik](https://hackmd.io/_uploads/BJ3KQhlRa.png) (source: https://edelrid.com/eu-en/knowledge/knowledge-base/strength-reduction-of-textile-materials-by-knots) The relation between relative bending radius and breaking strength of steel cable is perhaps a bit cleaner, since the radii are larger and the material does not deform as much under regular use as does climbing rope. ![grafik](https://hackmd.io/_uploads/rkPfznxAp.png) (source: https://practicalmaintenance.net/?p=576) ## The "Macaroni Cut" - sketch with a circle matching the hole diameter (probably not a good idea because OCCT?) - sketch with a path to be followed, defining a moderate bending radius matched to the plate thickness - sketch with circle offset to the start of the beginning of the pipe path ![grafik](https://hackmd.io/_uploads/Sy4k0heCa.png) ### What can go wrong? - Wrong orientation of the circle sketch ![grafik](https://hackmd.io/_uploads/Bk9syTeCT.png) - OCCT can screw up (I don't fully understand why. One candidate explanation is that the subtractive pipe shape asymptotically approaches the cylindrical hole shape and right in the center, they coincide within numerical errors) ![grafik](https://hackmd.io/_uploads/SJcxlpx0p.png) ### Occasional Result As of now, I can't give better guidance than offsetting the pipe path by 0.5 mm outward, or by making the circle diameter smaller to avoid locally coincident surfaces. There certainly is nothing wrong with using 3.5 mm radius instead of 4.0 mm. Suggestions are welcome. ![grafik](https://hackmd.io/_uploads/Hkki-pgC6.png) ### Addendum It's surprising that a 3D printed plate withstands 30 kg load without deforming when printed in PLA. I also printed the part in PETG, but.. eventually this will be made from Al plate, so I bought a quarter circle cutter: ![quarter circle cutter](https://hackmd.io/_uploads/B1merMGAT.jpg)