feat(assets): local_jar_lid — the operator's own printed screw jar
Third project-authored screw asset (local_bottle_cap / local_juice_shot
family), split from the operator's own Jar__Lid.stl print plate (included
unmodified as Jar__Lid_source.stl for reproducibility).
THE THREAD IS MEASURED FROM THE MESH, not guessed — the juice_shot lesson
applied at authoring time: helix fits through the lid's inner-skirt crest
vertices and the neck's outer crest vertices independently converge on a
3-START thread, lead 13.35 mm/turn (phase concentration 0.73 / 0.65).
Engagement ~7.5 mm => the real lid disengages in ~0.56 turn; the task gate
(10 mm) sits at ~0.75 turn — inside the one-wrist-rotation collectability
regime the juice_shot 22-take failure established.
Mechanism: lid_slide (FIRST joint, overlay contract) + coaxial lid_twist,
equality polycoef[1] = 0.01335/2pi; 0.055 N*m thread friction + 5 N skirt
friction, both solreffriction-stiffened (the hammer-nail sag lesson) =>
straight-pull threshold ~31 N (measured: 20 N moves the lid 0.59 mm), 11x the
0.28 kg assembly weight. lid<->jar contacts excluded. Geometry: 90x90x110 mm
chamfered square body + O77.4x11 mm neck; 90x90x33 mm lid, 15 mm skirt.
ESTIMATED pending measurement: part masses (PLA, 0.20 / 0.08 kg authored).
usd/jar_lid_render.usda from RoboVerse/scripts/mjcf_screwcap_to_usd.py; the
screw ratio rides the prismatic joint as wuji:screwPitchMPerRad customData
(PhysX mimic-drive wiring stays the Isaac-side follow-up).
Consumers: hand_track.open_jar_lid + tests/test_jar_lid_asset.py (8 physics
regressions: lead cross-check, slide-first, no pair contacts, holds closed,
pull resisted, opens at the measured ratio, tighten blocked, print envelope).