Moto gps rev a1

MOTO GPS Rev A PCB

Status: ENGINEERING / NOT FOR FABRICATION

The canonical board is generated from the native schematic and its exact pad-to-net data; it is not a hand-written zero-item placeholder.

hardware/rev_a/board/generate_board.sh
hardware/rev_a/board/export_preview.sh
/opt/homebrew/Caskroom/kicad/10.0.6/KiCad/KiCad.app/Contents/MacOS/kicad-cli \
  pcb drc --format json --all-track-errors \
  -o hardware/rev_a/board/build/drc.json \
  hardware/rev_a/board/moto-gps-rev-a.kicad_pcb

generate_board.py creates the four-layer, 1.0 mm, nominal Ø52 mm board with the four V3 R2.70 screw scallops centred on R27.70 / PCD55.40. It places WROOM-1U left, LC76G/RF right, IIS2MDC at 12 o'clock, QMI8658A near centre, and USB/power/display FPC at 6 o'clock. Seventeen test pads are on the back service ring. The placement report is written to build/placement-report.txt.

The H0175 EVT A1 population uses QMI8658A C3021082, IIS2MDCTR C2655002 and Panasonic EVQP7C01P C388883. These are controlled EVT choices, not silent substitutions: firmware identity/interrupt/axis tests, final-assembly compass calibration, and the P-suffix 1.1 ± 0.1 mm actuator/enclosure travel check are release gates. J1's official MPN is U262-161N-4BVC11; the 16XN spelling in the existing footprint name is retained only for library compatibility.

H0175 EVT A1 follows the controlled display PDF page 7 and directly supplies TP3.3/IOVCC/VCI/VBAT/VCI_EN from 3V3_DISPLAY. The assembled board must pass repeated cold-start, low-temperature and maximum-brightness tests before this power choice is production-frozen.

The generator prefers the audited project footprints in MOTO_GPS.pretty. When one is absent it embeds a conspicuously marked engineering envelope so the board can still be density-checked, but the repository fabrication gate rejects the result. Never infer a production land pattern from a generated envelope.

Routing workflow

The routed Rev A0 R4 review source is frozen at build/release_check_r4/moto-gps-rev-a.kicad_pcb. It contains 1555 track segments and 189 vias. KiCad 10.0.6 was run with --all-track-errors, --schematic-parity, all severities and zone refill; the result is DRC 0, unconnected 0 and schematic-parity 0. ERC is also 0. Reports live in ../validation/*release-check-r4.json and the immutable review export lives in ../manufacturing/engineering-candidates/rev-a0-20260903-r4/.

generate_board.py deliberately remains the deterministic unrouted placement generator. The final routed candidate is reconstructed from the saved R15 placement/SES flow plus repair_r3_route.py; do not mistake a fresh generator output for R4 or overwrite the frozen R4 directory.

Blocking physical inputs

• Original AMOLED FPC contact side, pin 1, thickness, insertion direction, connector height and exact MPN.
• Exact USB-C shell/board-edge/Z datum, mating-plug and waterproof-port envelope.
• Second-person pin and land-pattern review for every project footprint.
• GNSS and Wi-Fi antenna/coax/RF-window measurements and controlled-impedance stack-up.
• Protected battery, lead, polarity, connector and swelling envelope.
• Charger/buck-boost power coupon, thermal, switchover and 1.5 A load-step tests.
• Display, battery, antenna, buttons and Garmin mount physical interference checks in the assembled V3 enclosure.
• QMI8658A firmware identity/interrupt/axis regression, IIS2MDC final-assembly calibration, EVQP7C01P button travel, and H0175 direct-3.3 V startup/highlight tests.

A clean DRC has now been achieved, but it does not remove these blockers. Fabrication export must go through scripts/hardware/fab_release.py, which currently fails closed by design.