hrithik18k/metal-touch-panel

VALIDATION.md

# Validation — Metal Touch Rev A

Final layout: 72 x 64 mm, two layers, 14 schematic components including four printed coils, 37 copper trace objects including the four preserved windings.

| Check | Result | Evidence |
| --- | --- | --- |
| Source netlist | 0 errors, 0 warnings | evidence/final-netlist.log |
| Schematic placement | No reported placement issues | evidence/final-schematic-placement.log |
| PCB placement | No placement issues, 0 DRC errors/warnings | evidence/final-placement.log |
| Routed build | Pass, no recorded circuit errors | evidence/final-build.log |
| TypeScript | Pass | evidence/final-typecheck.log |
| Printed winding audit | Four coils, no external top-wire collisions, no recorded errors | evidence/final-coil-clearance.json |
| Gerber-derived shorts | No shorts detected | evidence/final-shorts.log |
| Gerber export | Completed | evidence/export-gerbers.log |

The winding ownership normalization changes only source_trace_id on four objects after routing. It does not suppress errors or remove copper. Native anonymous-footprint copper reporting cannot infer intentional printed-inductor contacts reliably; therefore both geometry and metadata-aware shorts checks are retained.

Nonblocking tool warnings remain about unnamed traces, missing schematic sheet, lack of explicit coil courtyards, and IC pin capability annotations. The reference-design pinout and logical netlist were reviewed independently. Export reports the through-hole header pick-and-place rotation as unverified. The printed coils are marked doNotPlace and must not be purchased as discrete components.

No physical board, impedance measurement, panel press test, temperature test or firmware test was performed. Fabrication outputs are for prototype review; the mechanical target and sensor settings need bench calibration.