Field Notes: Sea Trials for Hull Nine

There's a particular smell to a new hull — resin, paint, and the faint electrical tang of a battery bank that's never been warm. Hull Nine had all three when we craned her off the low-loader at the ramp south of Fortescue Bay. Four days of sea trials ahead of us, a checklist with ninety-one items on it, and the quiet understanding that at least five of those items would fail, because they always do. Sea trials aren't a pass/fail exam. They're a negotiation between the boat you designed and the ocean you're about to hand it to.

For anyone new to the fleet: our Drifter vessels are autonomous, solar-electric survey-and-seeding platforms — six metres of catamaran carrying sonar, cameras, a line-laying rig and enough smarts to run a reef survey pattern alone. Hull Nine is the first of the Block 3 builds: same bones, new power bus, redesigned seed drum, and a sensor mast that folds instead of unscrewing — a change we owe entirely to Drifter 7's complicated history.

Day One: Does It Float Like a Boat?

Day one is deliberately boring. Trim, displacement, bilge behaviour, low-speed handling. We load her to operating weight — dummy payload in the seed drum, full batteries — and drive slow circles in the lee while two people take notes and one person just looks at the waterline.

Boring is the point. Every problem you find at two knots is a problem you don't find at twelve in a seaway. The one catch of the day: she sat three centimetres bow-heavy with the drum loaded forward, which is precisely the kind of thing that doesn't matter until the day it matters — trim affects solar incidence, and solar incidence is range. We added ballast aft, re-marked the waterline in the log, and moved on.

Day Two: The Power Budget

A Drifter is an energy accountant that happens to be a boat. Day two is the accounting audit: measure everything. Solar harvest versus sun angle, hotel load of the compute and sensors, propulsion draw at survey speed and at transit speed, and the heat behaviour of the new power bus under sustained load.

The Block 3 bus was the question mark — it consolidates charging, propulsion and payload onto one managed backbone instead of the old split system. The bench tests said it would run cooler. The sea said the bench tests were right about temperature and wrong about one connector, which developed an intermittent fault every time the mast folded — movement plus a marginal crimp. That's two checklist items for the price of one discovery: the crimp got re-done to shop spec, and "fold mast 50 times under full load" got added to the acceptance test permanently. This is why the gear graveyard exists: every prototype that failed taught us a test we now never skip.

By evening the numbers said what we hoped: at survey speed in autumn sun, Nine is net-positive on energy across a working day. Not by much — autonomy at sea is a game of margins — but positive.

Day Three: The Autonomy Trials

Day three is when the boat stops being a boat and starts being a robot. We ran her through the standard course: waypoint lawnmower over a mapped reef patch, station-keeping against wind and tide, obstacle response with a drifting float, comms dropout behaviour, and return-to-home from the far edge of range.

The lawnmower was flawless — the sonar mosaic it built of the trial ground was the best quality we've pulled off a new hull, cleaner calibration than anything in the shop predicted. The interesting item was station-keeping. Holding position over a mark in a knot of tide, Nine hunted — small corrections back and forth instead of settling. Not a fault, a tuning parameter: the new drivetrain responds faster than the control loops expected, so the loops were over-correcting. Twenty minutes with a laptop on the chase boat and she was sitting on the mark like she'd been glued.

What we never tune away is the dropout rule: lose command link for ninety seconds, the vessel finishes its leg, climbs to safe depth of operation, loiters, then comes home. We test it by simply turning the radios off and watching. It's the least exciting test on the list and the one that matters most, because an autonomous boat in kelp country will lose its link. The only question is what it does about it.

Day Four: Seeding Systems, Wet

Everything until now could have been done by any survey craft. Day four is what makes her a Drifter: laying Holdfast line. We spooled two hundred metres of seed-line through the redesigned drum and ran the full laying profile — tension control, depth control, the laying head's response to speed changes.

The new drum was the best surprise of the week. The old design needed a crew member's judgement on tension through turns; the new one holds it automatically within the band the hatchery's seed viability checks care about. The failure of the day was human, not mechanical: our deployment check sequence had the mast-fold step before the drum-spool step, which doesn't work with the Block 3 mast geometry. Reorder the checklist, update three other documents, tell the shop. Paperwork as safety equipment.

What "Passed Trials" Means

Four days, ninety-one items, eighty-three passed first time, six fixed on the water, two deferred to the shop (the crimp batch inspection, and a latch redesign on the drum hatch that's already drawn). Hull Nine signs on to the fleet next month.

The phrase I keep from the week is the trials skipper's: the slow tests are the safety tests. Anyone can make an autonomous boat that impresses you in a demo. The work is in the ninety-second dropout loiter, the fifty mast folds, the trim measurement at two knots — the slow, damp, unphotogenic diligence that means that when Nine is surveying alone on the far side of the bay, the exciting thing about her is the kelp data, not the drama. Curious about the people who build and run these boats? Our ocean-tech careers guide is the honest version of the job.

Key Takeaways

• Sea trials are a negotiation, not an exam: expect failures, and treat each one as a permanent addition to the acceptance checklist.
• Trim and power margins decide an autonomous solar vessel's real range — measure them loaded, at working weight, in real water.
• The best trial results are boring: energy-positive at survey speed, clean sonar mosaics, uneventful dropouts.
• Day-one fixes (a crimp, a tuning gain) are cheap; the same faults discovered mid-season are not. Slow tests are safety tests.
• Every fleet improvement traces back to a failure somebody logged honestly. The checklist is institutional memory with boxes.

FAQ

How long does sea-trialling a new Drifter take?

Four intensive days on the water for a Block 3 hull, bracketed by bench weeks in the shop before and punch-list work after. A genuinely new hull design — not a block update — takes a full season of progressive trials.

What's the most common failure in trials?

Connectors and anything that moves: folding, spooling, latching. Salt, vibration and marginal crimps are an unbeatable combination, which is why the trials script now hammers moving parts under load, repeatedly.

Why test comms dropout by switching the radios off?

Because simulated dropouts test the software and real dropouts test the whole chain — antennas, power sag, failsafe timers. The vessel's designed behaviour on link loss is its most important safety feature at sea.

Who crews sea trials?

A small mixed team: vessel skipper, controls engineer, and an operations diver with authority to halt anything. Trials are engineering work conducted in an operations environment, and both cultures need a voice on the chase boat.

Do failed items stop a launch?

Depends on severity. Structural, power or failsafe faults stop everything. Ergonomics, calibration drift and workflow faults get fixed in parallel with early operations. The rule is simple: anything affecting the boat's ability to come home safely is a launch blocker; everything else is a task with a date.