Repository navigation
CadQuery Workplane.cut() produces degenerate results for overlapping spheres #2020
Description
Description
I've run into an issue where a boolean cut between two overlapping spheres using Workplane.cut() appears to produce incorrect geometry. The resulting solid retains the volume of the original uncut sphere, and the cut face has zero area. I noticed that performing the same operation with OCP.BRepAlgoAPI.BRepAlgoAPI_Cut directly gives the expected result, so I wanted to flag this in case it's helpful.
It's possible I'm using the API incorrectly happy to be corrected if so!
Reproduction
import cadquery as cq
from OCP.BRepPrimAPI import BRepPrimAPI_MakeSphere
from OCP.BRepAlgoAPI import BRepAlgoAPI_Cut
from OCP.gp import gp_Pnt
d, r = 3, 5 5
# Via CadQuery
sphere1 = cq.Workplane().transformed(offset=(-d, 0, 0)).sphere(r)
sphere2 = cq.Workplane().transformed(offset=(d, 0, 0)).sphere(r)
crescent_cq = sphere2.cut(sphere1)
print(f"CadQuery volume: {crescent_cq.val().Volume():.2f}")
# 523.60 — same as the original sphere, expected ~414.69
for f in crescent_cq.val().Faces():
print(f" Face area: {f.Area():.4f}")
# Face 0: 314.1593 (full sphere surface area)
# Face 1: 0.0000 (cut face appears degenerate)
# Via OCC directly
s1 = BRepPrimAPI_MakeSphere(gp_Pnt(-d, 0, 0), r).Shape()
s2 = BRepPrimAPI_MakeSphere(gp_Pnt(d, 0, 0), r).Shape()
crescent_occ = BRepAlgoAPI_Cut(s2, s1).Shape()
print(f"OCC volume: {cq.Solid(crescent_occ).Volume():.2f}")
# 414.69 — expected crescent volume
for f in cq.Shape(crescent_occ).Faces():
print(f" Face area: {f.Area():.4f}")
# Three faces, all with non-zero area Expected behaviour
sphere2.cut(sphere1) should produce a crescent-shaped solid with volume ~414.69 and faces with non-zero area at the cut boundary.
Actual behaviour
The result has the same volume as the uncut sphere (523.60). The cut boundary face has zero area and receives no mesh triangles from BRepMesh_IncrementalMesh, which leaves the solid effectively unclosed.
Context
@jmwright and I came across this while working on CAD-to-DAGMC neutronics geometry conversion. The degenerate face prevents the mesh from being watertight, which causes particle transport failures. Using BRepAlgoAPI_Cut directly works around the issue for now. I think this is related to the errors we saw on the model-benchmark-zoo PR today
Environment
Tested on both CadQuery 2.7.0 and 2.8.0.dev0 (installed from main today) — same behaviour on both.
Activity
adam-urbanczyk commentedon Mar 28, 2026
This is related to failed clean. As a workaround try:
crescent_cq = sphere2.cut(sphere1, clean=False)amaelFr commentedon Apr 14, 2026
Got the same kind of issue with this code
Is there a way to correct it ?
One error on the cuts and one on the union of the tread and core