# Jimny V2 — local interface fit coupons

**12 separate samples: one STEP solid and one STL per sample.** XYZ millimetres, Z up, every file already bed-oriented at Z=0. Largest extent is 43 mm. These are small local fit experiments, not physically validated manufacturing releases.

## Print and identify

Print at **100% scale**, one copy of each selected file, using the intended printer, filament, nozzle and slicer settings. Do not let the slicer auto-rotate the supplied orientations. Record actual settings and measured fit; avoid sanding mating surfaces before recording the first result. Review the sliced layers and bed adhesion before printing. No slicing or physical printing has been performed here.

| Matching set | Individual file stems | What it tests |
|---|---|---|
| Engraved **1**, one dimple | `pod-1-male-c03`, `pod-1-female-c03` | 0.3 mm Y-coordinate flank clearance |
| Engraved **2**, two dimples | `pod-2-male-c04`, `pod-2-female-c04` | 0.4 mm Y-coordinate flank clearance |
| Engraved **3**, three dimples | `pod-3-male-c05`, `pod-3-female-c05` | 0.5 mm Y-coordinate flank clearance; current V2 female profile |
| Capsule | `capsule-tab-2mm`, `capsule-ledge-04` | Actual 2 mm tab under a 2.4 mm ledge, 0.4 mm vertical running gap |
| Swivel | `swivel-sleeve-8mm`, `swivel-bore-88mm` | 8 mm sleeve OD, 3.4 mm sleeve through-bore, 8.8 mm leaf bore |

Each stem has `.step` and `.stl`. Pod keys are identification numerals and counted dimples, **not geometry that prevents mismatching**. All three males deliberately share the same profile; match the engraved pair first, then a single male can compare all three females.

### Dovetail

The working profile is the current V2 45° trapezoid: root 8 mm, crown 14 mm, rise 3 mm above a 3 mm base. The `0.3/0.4/0.5` sweep changes only the female Y-coordinate expansion, preserving the actual female template's vertical positions. Thus the perpendicular flank gaps are `c / sqrt(2)` (approximately 0.212 / 0.283 / 0.354 mm), **not** 0.3 / 0.4 / 0.5 mm normal offsets. Crown-to-roof gap stays 0.6 mm; shoe-to-base vertical gap stays 0.4 mm.

Print males base-down, dovetail up. Females are exported **roof-down, channel-up**, avoiding a closed slot roof bridge. After printing, turn the female over, align the rail/channel axes, and slide in from either open end. The external number pad is outside the working flank. Compare start force, full short-stroke sliding, lateral play and gentle lift capture. Do not force a jammed fit.

### Capsule tab/ledge

**Current source is 12 × 6 × 2 mm, not the older brief's 12 × 8 × 2 mm.** The actual 12 × 2 × 5 mm stem placement and 6.4 mm tab-to-lid underside datum are preserved. The ledge begins 0.4 mm above the tab top; the stem has 0.6 mm side clearance. Only the nearby lid skin is simplified to a rectangular slab.

The tab prints **on the end of its constant section**, original 12 mm X becoming print Z; this avoids supporting the mating lip. The ledge prints inverted with its broad face on the bed. For testing, turn both to the functional orientation: stem beside the ledge edge, tab below ledge, lid slab above ledge. Slide along the ledge; a gentle upward motion should contact it. This is a single straight ledge segment, without capsule windows or a full slide-then-lift release. End-on coupon layer orientation may differ from the final lid: use the result for rough local clearance screening, not a universal printer compensation.

### Swivel

Both parts print with the cylinder axis vertical. Place the annular bore sample over the sleeve and compare sliding/rotation. The stem is the actual 15.5 mm sleeve height above a shortened 3 mm test foot; the bore sample is 14.7 mm thick. A full M3 fastener stack, bottom nut trap, leaf shape and stop track are intentionally absent.

**8.8 − 8.0 = 0.8 mm diametral clearance = 0.4 mm radial clearance per side.** A claim of “0.4 mm diametral” would describe a different 8.4 mm bore. The digital test places the bore 0.4 mm above the foot; no spacer is included, so hand seating on the foot does not reproduce that axial datum.

## What is verified

`manifest.json` contains per-file SHA-256, bounds, positive volumes, one-solid STEP round-trip, watertight single-component STL round-trip, planar bed-contact area and explicit **material/void probes on reimported STEP solids**. Five interface reports check clear sliding/insertion at discrete positions plus intentional interference beyond clearance: dovetail lift/lateral motion, capsule premature lift, swivel eccentric displacement. Transform reimported `Solid`s, not the enclosing located STEP `Compound`.

The maximum mesh/BREP bound comparison tolerance is 0.04 mm, consistent with the 0.035 mm curved tessellation chord tolerance; the 0.0001 mm vertex weld is numerical mesh cleanup, **not** a manufacturing clearance. All planar interfaces retain their exact BREP dimensions.

These rough slot/rail and pivot coupons **do not prove complete assembly stops, release sequences, retention, heat resistance, creep, wear, strength, vehicle fit or crash safety**. No physical validation is claimed. Keep small pieces away from children.

## Rebuild and integrate

Extract `coupons.zip` and run from its root, retaining the `public/coupons/` layout. The kit includes root CAD profile source, export helper, dependency lock and cassette generator, plus the two current cassette source STEP inputs in `public/models/`. No main-model rebuild is needed.

```sh
uv venv --python 3.12 .venv
uv pip sync --python .venv/bin/python requirements.lock
.venv/bin/python -B public/coupons/source.py
.venv/bin/python -B build_cassette_coupon.py
```

The first command after setup rebuilds ten samples; the second appends the cassette pair, yielding twelve STEP/STL pairs in `public/coupons/`. This standalone kit is not the full gallery/source release; use `jimny-v2.zip` for that.

The script reads the current parent profile source without importing/executing its build/export code, guards the capsule/swivel numeric profile assumptions, reuses the actual pod profile functions through AST extraction, and checks that the parent file did not change during the build. It writes only inside this directory. Parent profile changes require reviewing/rebuilding these coupons; the manifest pins source SHA-256 and source excerpts.

Manifest shape: `{schema_version, title, units, coordinate_system, status, sample_count, source, samples:[{id, family, role, files:{step,stl}, print_orientation, print_to_test, dimensions_mm, bounds_mm, ...validation, ...interface_parameters}], interface_checks:[...], limitations}`. Artifact paths are **relative to this directory**. `print_to_test` means rotation about the stated global origin axis first, then translation in mm. Missing rotation components are zero. Use `samples`, not a directory glob, as the integration/download allowlist.

## Cassette runner / groove pair

`cassette-runner` and `cassette-groove` crop the actual reimported cradle/drawer STEP profiles over a 24 mm rail length. A coupon-only backing web outside the mating face reconnects the groove roof after cropping removes the real drawer end walls. The active rail/groove profile and 0.4 mm running gap remain unchanged. Keep the exported Z=0 orientation; support review is needed for the overhang/channel. Use `print_to_test.translation_mm` to align canonical print files for the digital mating pose. Physical placement: slide the projecting tongue into the side groove; do not push the roof vertically over it. The report checks five sliding offsets and a deliberately blocked 1 mm lift. This tests local fit, not the full stop key or latch.

Rebuild the first ten samples with `public/coupons/source.py`, then append this pair with root `build_cassette_coupon.py`, in either the standalone kit or full project. The combined manifest reconciles twelve samples.