Room for the seal.

Research record

The detail behind
the study.

Source-linked work for this revision. Reported checks describe the stated model or test; physical qualification and applicability to the current retrofit remain separate.

R5 widens the sealing groove from2.20 to2.50 mm. This retains the forward continuous mating land, relieved dielectric support and nominal seal squeeze while providing additional space for dimensional variation and seal volume growth.

Defined screening result

All64 corners pass that specified screen. The corners retain the previous illustrative radial±0.02 mm, corner±0.02 mm, seal cross-section±0.05 mm and free-ID±0.20 mm ranges. These are proposed limits, not qualified factory or seal-supplier tolerances. The10% combined volume-growth allowance is an engineering budget, not measured chemical swelling, thermal expansion or a material rating.

The wider groove is a custom departure from the tabulated baseline, requiring seal-supplier and service-motion review. It may affect rolling and positioning during mating. The Trelleborg O-Ring catalog , June2024 page54, supplies the85% reference ceiling. It does not approve this assembly.

The seal remains a nominal constant-volume elliptic occupancy model. The volume-growth screen is a calculation, not a physically deformed second seal model. Actual compound, inside-diameter stretch, manufacturing capability, service chemistry/temperature, extrusion gap, pressure and insertion damage must be resolved before release. The seal still travels past keyway openings on its way to the uninterrupted final land.

Files and checks

complete-termination.step contains457 separate cable and connector solids.

inspection.blend and inspection.png show the CAD-derived exploded interface.

source/ contains editable generation source and locked dependencies.

record.json includes every dimensional corner, the explicit growth budget and remaining limitations.

1019 geometry and screening checks pass. Two focused tests pass. The groove-volume equation independently agrees with the actual CAD volume. A copied-source rebuild matches all457 imported STEP solids by individual Boolean symmetric difference. The render was reopened, its four visible source meshes verified and the image visually inspected. Full regression remains running as session49399.

Retaining-ring installation and actual supplier geometry, bayonet detent/preload, stud anchorage, contact termination, jacket anchorage, electrical performance and environmental testing remain unfinished. This revision removes the previous overfill result within its stated screening assumptions; it is not manufacturing qualification.