A more defined interface.

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.

R3 replaces the R2 custom ring interface with a catalog-derived VSM-32 installed envelope and matching groove. It preserves the complete cable, connector class keys and bayonet arrangement. The ring is represented by a conservative full annulus; this is not exact supplier ring CAD, its free shape, or its installation deformation.

Supplied geometry

complete-termination.step: 457 separate cable/connector solids.

interface-envelope.step: eight mechanical interface solids.

inspection.blend / inspection.png: native CAD-derived inspection render.

The ring envelope is ID30.88, OD35.36 and thickness0.64 mm, placed at Z16.075–16.715. The body groove begins at Z16.00 and is 0.79 mm wide with a 30.88 mm root diameter. These dimensions use the Smalley VSM-32 published interface . The supplier's free ID and end shape differ from the installed annular envelope.

Proposed tolerance basis

Published ring/groove tolerances are included. Custom body/collar radial and axial face targets are ±0.05 mm; groove-start location is ±0.02 mm. These are engineering targets requiring process-capability validation, not existing supplier or factory acceptance limits. The key corner uses radial17.50±0.05 and tangential half-width1.50±0.05 mm. The internal collar radius uses20.30±0.05 mm. Body bore uses12.00±0.05 mm radius. Nominal collar flange faces remain Z14.15 and15.85; rear body shoulder Z14.00; ring relief radius18.40.

The 0.192 mm installation result includes a 0.10 mm study passage allowance. It is not tool clearance or proof that the ring tolerates the required expansion. The axial-float upper bound is an output to evaluate against allowable connector movement; no maximum permitted endplay has yet been selected.

Verification and limits

170 generation checks pass, covering applicable nominal motion/interference, groove restraint and both STEP round trips. Two focused tests pass. A separate copied-source rebuild matches all465 STEP solids by individual Boolean symmetric difference. The native Blender file was reopened and its eight visible meshes checked against CAD vertices within0.001 mm; the image was visually reviewed. The complete R3 regression passed348 tests in2006.54 seconds, terminal exit0. This run predates R4 and later tests.

Actual supplier geometry, selected material, installation stress, installation/removal access, allowable endplay, axial load and fatigue remain unresolved. The supplier's stress guidance matters because our keyed passage exceeds a plain32 mm shaft. The actual spiral installation method needs an accessible assembly sequence. Seal compression, stud attachment, detent/preload and electrical contact termination also remain unfinished.

This revision closes the oversized-ring interface mismatch at the nominal/tolerance-envelope level. It is a development candidate, not a released manufacturing design or an approved VSM-32 application.