2026-11-06 · Claude Opus 5.5 (high) · data: review-manufacturing.json · related: final geometry review
Disposition. All known blockers (BLK-1 to 3), defects D-1 to D-5 and the stale-export item are resolved in build 9. Stage A coupon and fit files can be prepared for bureau quotation and printing. The full shell remains conditional on measured donor and display fit, coupon results and the acoustic prototype. Production is not validated. No sit, load, tip, thermal, electrical, acoustic or child-use tests and no FEA have been done.
mechanical/out/qa/qa.json generated 2026-11-06T05:24:51.505517Z (build 9).out/docs/part-manifest.json generated 05:24:53.202847Z.manufacturing/release-status.json generated 05:25:46Z.derive-regression.json: ok, 68/68.review/native-manufacturing-audit.json, generated 10:25:02Z: 746 checks, 0 failures, against the same QA and manifest timestamps.mechanical/assembly-manifest.json and mechanical/native-export-audit.json predate build 9 and are awaiting root regeneration. They are not current evidence, and refreshing them is a metadata task, not a CAD rebuild.The QA parameters equal the frozen 355 / rear 92 / 245 / 18° / 180 / deck 62.5 / flat / drop 23 / underlap 39 / pad 7, 1, 10. Key top is 73.3 mm and the envelope is 355 × 338.0 × 415.4.
| Item | Status | Evidence |
|---|---|---|
| BLK-1 shell mesh mismatch (relative tessellation) | Resolved | Absolute tessellation fixed the shells. The only remaining difference is the flexible volume-roller tyre: 2.473 vs 2.490 cm³, about 17 mm³ (0.70%). Spread over the tread surface that is of order 0.01 mm, negligible next to the 0.4 mm stretch fit and TPU tolerance, so I accept it as a minor tread approximation. My earlier 0.5% threshold was my own heuristic, not a standard, and I have withdrawn it. |
| BLK-2 black-key 28.55 ramp value | Resolved | QA and the guide now give the flat top as 25.05 mm, ahead of the CAD lip, separately from the 28.55 outer envelope and the 26.10 viewer value. The optimizer black-key probe shows CAD finger clearance 8.68 mm, against a 3 mm allowance. The white-target 40 mm clearance is not whole-keyboard proof. A hand test is required. |
| BLK-3 dock exclusion and status text | Resolved | The guide states the rear cover-storage dock is a viewer concept with no CAD or BOM. It reports service paths 9/9 as solid sweeps, the mesh difference (0.70%, tyre), the shell order gates (G0–G5, G9) and “no load rating”. |
| D-1 bearing docstring | Resolved | parts.py:207 now says MR84ZZ (4x8x3). |
| D-2 derive regression | Resolved | ok true, 68 rows, 0 failing. |
| D-3 screw and shim engagement | Resolved (documented) | The guide's F14 table gives a 7.5 mm trimmed screw for a 0.5 mm shim, 8 mm as supplied for 1.0 mm, and a 10 mm screw trimmed to 9 for 2.0 mm. Each gives 2.5 mm engagement and 0.5 mm recess. The measured stack overrides the table, and no screw may protrude. |
| D-4 right gasket solids | Resolved (fixed) | gasket_chamber_R is now 1 solid. The guide says “one closed piece per side… not a leak test”. |
| D-5 shim quantity and 3MF semantics | Resolved (documented) | STLs are single geometries. Stage 3MFs are inspection collections that mix materials, not print jobs. Quantities come from print-schedule.csv (6 of each shim). |
| Stale instrument_deck exports | Resolved | Only instrument_deck_rear and wing_L/R exports remain. The native audit checks for extras. |
manufacturing/model-e-sections.png. Its title matches the final build: deck 62.5, drop 23, key top 73.3, depth 415.4, height 338.0.
Crossing outlines in a 2D section are projections of separate parts, not collisions. QA BRep interference is 0. I infer no strength from the outlines.
Packaging, refreshed hashes and manufacturer-reference redistribution belong to root and are not reviewed here.
Native release check, 2026-11-06T10:28:35.496803+00:00: the supervisor assembly manifest and export audit have now been refreshed against final build 9. All 83 published assembly mesh hashes match; both wrappers reference the final QA hash. This closes the metadata-only pending item described above. Detailed manufacturer reference shapes are omitted from the public assembly; fabricated parts are unchanged. No physical validation is implied.