Three previous shops left a tangle. One day to clean it.

§ 01 · The vessel

F/V Northern Runner is a 42-foot Calvin Beal Jr. lobster boat, built on Beals Island in 2003. She fishes inshore lobster out of the south end of the harbor, with a Cummins QSB 6.7 (425 hp) and a small Phasor K2 (5 kW) for when she's plugged in or for the trap-hauler hydraulics. Bridge electronics: a Garmin 8600 chart-plotter, a Furuno NavNet TZT2 backup, a Garmin GMR Fantom radar, an Airmar B175 transducer, an SI-TEX MDA-5 AIS, and a Maretron magnetic compass.

§ 02 · The problem

Six different brands on the bridge, three different shops over six years. The captain had been to Bridgeport for a one-day repair, to a shop on the Cape for an electronics swap, and to a Maine yard for the autopilot. Each shop had touched the NMEA 2000 backbone and none had left it clean.

What we found on the walk:

• Two terminators missing — the bus had open ends at both directions on the bridge.
• Two daisy-chained drops — the GMR Fantom radar dropped to a tee that branched to the AIS and the magnetic compass.
• A 12-meter drop on the magnetic compass, against the 6-meter spec.
• One unfused 12 V power injector run from the wrong battery bus, dropping the bus voltage at the far tee to 10.9 V.

The captain reported the chart-plotter would freeze at irregular intervals, the AIS would drop sentences without warning, and the autopilot occasionally lost heading reference for two-second windows. None of those is a small thing on a working boat.

§ 03 · What we did, in one day

The fix was a complete backbone rebuild, executed in one working day at the State Pier, slip 14:

1. Pulled the entire existing backbone — 28 meters of mid-cable in three colors.
2. Replaced with a single continuous backbone in Maretron mid-cable — 16 meters — with a 120 Ω terminator at each end.
3. Re-located the magnetic compass tee onto the backbone (4-meter drop, well within spec).
4. Wired one drop per device, in the order: Garmin 8600 → GMR Fantom → Airmar transducer → AIS → autopilot computer → Maretron compass → Furuno NavNet (via gateway).
5. Re-fused the 12 V injection at the helm panel, on a known-good house-bank circuit, with a 5 A ATC fuse.
6. Verified bus integrity with the Maretron N2KMeter: 60 Ω across, 12.4 V at the far tee, every device announcing on its own address.
7. Printed the bus map and sleeved it at the helm.

§ 04 · Outcome

The captain texted the shop on a Tuesday six weeks after the work to report that the chart-plotter had not crashed once. We took that as the test result we wanted.

The work is documented at length in the August 2025 bulletin on backbone discipline — written before this job, but the principles applied to it.

Cross-references: Plate P-06 (NMEA 2000), August 2025 bulletin.