Defect of the Week: Battery and Shore Power

On 11 and 13 December 2023, a surveyor went aboard the 82-foot yacht Flagship, then lying at Fort Lauderdale in Florida, USA. The survey report recorded that the battery management systems for the vessel's three 48-volt lithium-ion battery banks were not fully functional, and that the associated batteries were 100% discharged.

The vessel was towed to a shipyard on the Miami River that January, where portable chargers were bought to keep the banks up. Four and a half months after the survey, at 10:31 on 28 April 2024, the uncrewed vessel exploded and burned. Firefighters moved her to a sea wall, where she sank. Total loss, put at US$5 million. The investigators found the remains of a portable battery charger beside the battery bank, and the remains of a bypass wire on the 24-volt BMS in the same compartment. The probable cause was thermal runaway of that 24-volt lithium-ion bank, caused by the inoperable battery management systems and the workaround they produced: bypass the BMS with a wire, charge manually with a portable charger, and let a person at the shipyard watch the level instead.¹

The surveyor's report is the earliest independent, dated record of that condition. It named the defect and it fixed the date, months before anyone had reason to care. That is what an electrical section is actually for.

Which is worth remembering when reading the alternative, still common in condition surveys: "Electrical installation found in serviceable condition. Shore power tested and functional." Nine words, covering the system that sits at the top of the fire table. Five years of BoatUS Marine Insurance claim files attribute more than a third of all onboard fires to the DC system alone, with the AC shore power circuit adding a further 9%. That is US claims data, used here because no equivalent UK leisure dataset is published, and the UK picture points the same way: Boat Safety Scheme incident records place electrical problems among the top three causes of boat fires across the past five years.²

Very few surveyors are short of inspection skill here. Almost everyone lifts the battery lid. What separates a defensible electrical section from a worthless one is what gets written down, and specifically whether it describes protection or equipment.

Which Standard, and Where the Scope Ends

Two pieces of housekeeping first, because both appear wrong in otherwise good reports.

The standard for small craft electrical installations is BS EN ISO 13297:2021+A11:2023 (ISO 13297:2020), which absorbed the old extra-low-voltage DC standard ISO 10133 and the previous AC-only edition into one document covering both. A report citing ISO 10133 today is citing a withdrawn document. Alongside it, the RYA/BMEA Code of Practice for Electrical and Electronic Installations in Boats is the practical UK reference.

BS 7671 applies to the marina, not to the boat. Section 709 governs the pontoon installation up to and including the socket-outlet. Regulation 709.531.2 requires each socket-outlet to be protected individually by an RCD, disconnecting all poles including the neutral; one socket-outlet should supply one craft; and a PME earthing facility must not be connected to any metalwork of a boat.³ The vessel's own installation sits outside its scope, so citing BS 7671 at the boat's consumer unit is an easy error for an opposing expert to pick up.

Then the scope statement comes before the findings, not after them. A marine surveyor is not, by virtue of being a surveyor, a competent person under BS 7671. A condition survey does not include dismantling, insulation-resistance testing, earth-loop-impedance measurement, or instrument verification of RCD trip current and time. Say so, say the consumer unit was not opened and concealed runs not traced, and recommend periodic inspection and test by a competent marine electrician plus production of any existing certification. The scope statement is the surveyor's protection, and nowhere more than here.

The Fuse, and How Far From the Battery

Start where the energy is. Every ungrounded conductor leaving the battery needs overcurrent protection close to its source, for a reason worth stating plainly: the cable between the battery and its fuse is unprotected by definition. A dead short in that length has the whole bank behind it and nothing in the way.

ISO 13297 covers DC overcurrent protection in Clause 12 but places the location options in Annex E, which is informative. The European standard does not supply a normative dimension a boat can be held to. ABYC E-11 does, which is why most surveyors reach for the American figures in the field: protection within 175 mm (7 in) of the point of connection to the source of power, relaxing to 1.83 m (72 in) where the conductor runs directly from the battery terminal and is sheathed or enclosed throughout, and to 1.02 m (40 in) where it is similarly enclosed but connected to a source other than a battery terminal. Cranking motor conductors are exempt under E-11, because no practical fuse survives cranking current.⁴

Be honest about what that means in a UK report. ABYC has no legal or regulatory standing here, and none in a surveyor's terms of engagement unless it is written in. Cite ISO 13297 and the RYA/BMEA Code as primary, ABYC as supplementary good-practice guidance where it supplies a criterion the primary references leave open. Then measure the cable and record the number.

The run most often found naked is the inverter/charger DC feed. A 3 kVA unit on a 12 V bank pulls several hundred amps at full output, the cable is heavy, it usually passes through a bulkhead, and on many owner installations there is no fuse at the battery end at all. Record make, model, rated output, and whether the DC input is fused at source. The windlass and bow thruster feeds deserve the same question, being the other two high-current runs typically added long after build. The same visit should locate the battery isolator switch: rating, whether it breaks the positive, whether it is reachable without opening the engine space. Note which circuits legitimately bypass it (the automatic bilge pump feed is the usual one) and confirm each is independently fused.

The Bank Itself

Restraint. Pushed firmly by hand, the bank should not move. A battery that shifts in a seaway chafes its own cables or shorts across a metal fitting, and on a yacht that can be knocked flat, restraint that resists only sideways movement is not restraint. Record the method: a strap over a purpose-built box is a different finding from two lengths of timber wedged in.

Terminal insulation. Positive terminals need covers. A spanner dropped across an unprotected bank is a spot weld and a hydrogen ignition source in the same second.

Ventilation. Flooded cells gas hydrogen on charge and that compartment must vent. Record whether the bank is flooded, AGM, gel, or lithium, because the ventilation finding depends entirely on the answer. A surveyor who writes "batteries" without the type has removed their own basis for the finding.

Numbers. This is where most electrical sections quietly give up. Record chemistry, nominal capacity in Ah, and the date code where the casing carries one. Then take readings: resting open-circuit voltage after the bank has settled, and terminal voltage with the engine running and the charger live. On a flooded bank where the caps come off, specific gravity per cell. Where no conductance or load test was carried out, say so rather than leave the gap. Age alone is not a finding: a five-year-old flooded bank in liveaboard cycling service and a five-year-old start battery are different propositions, and the report should say which duty the bank is doing.

Terminations. Look under the insulation at the crimps. Green in the strands is capillary corrosion working up the conductor. It raises resistance, and resistance under load makes heat, which is the mechanism behind a great many engine-bay fires. Mixed-age or mixed-chemistry cells paralleled into one bank deserve a line of their own, because the weakest cell sets the behaviour of the whole.

The Wiring That Is Visible Without Instruments

Open a locker on a twenty-year-old cruiser and the findings are usually in plain sight, no instrument required.

Solid-core domestic twin-and-earth on AC circuits is endemic in DIY fit-outs and narrowboat conversions. It is not made for vibration: it work-hardens and fractures at the termination. Untinned copper on a marine DC system is the same problem one voltage down. Then choc-block connectors, twisted-and-taped joints, non-marine crimps, and terminations with no strain relief, all inside a hull that vibrates and sweats. Look for unsupported runs, chafe where a loom crosses a bulkhead penetration, ties across a hot exhaust, and conductors lying in the bilge. Note conductor sizing and volt drop on long runs to critical circuits, and note where successive owners have left old UK, harmonised, and US colour conventions meeting at one junction. That last is a genuine hazard, not a tidiness point.

The Lithium Retrofit

This is the finding most likely to be missed on a 2026 survey, because the boat looks unchanged.

LiFePO₄ drop-in batteries are sold as like-for-like replacements for a lead-acid block of the same footprint. Physically they are. Electrically they are not, and two consequences matter.

Before either, one point the Flagship case makes plainly: a BMS reported as faulty is a live finding, not a note for the file. An inoperable BMS invites a workaround, and the workaround is what burns the boat. Where a BMS is inoperative, bypassed, or shows fault codes, record the condition, record how the bank is being charged in the meantime, and grade it accordingly. A jumper wire across a BMS terminal is visible to anyone who opens the compartment and looks.

The alternator. The failure everyone quotes is the load dump: a lithium BMS protects its cells by opening the circuit, and if it opens while the alternator is charging at full field, the alternator loses its load instantly and the spike goes into the diodes, the regulator, and whatever else shares the bus. ISO 13297 requires DC equipment to function across 75% to 133% of nominal, meaning 9 V to 16 V on a 12 V system; measured excursions on lithium disconnect events run far above that.⁵ But the more frequent killer is thermal. A lithium bank accepts near-full current until it is nearly full, so a standard alternator on a lead-acid regulator sits at high output far longer than it was designed to and cooks itself. The question for the record is the same either way: is there external regulation with an alternator temperature sensor, or is the alternator left on the start battery with the lithium fed through a current-limited DC-DC converter? Many drop-in installations have neither.

Cold charging. LiFePO₄ cells should not be charged below 0 °C without risking lithium plating. Two qualifications matter in the field. It is cell temperature, not air temperature, and a bank of that mass inside a hull lags the air considerably. And several manufacturers permit reduced-rate charging below zero to their own stated limits. So the finding is not "assume it is absent", it is to verify whether the BMS has a low-temperature charge cut-off and record the manufacturer's stated limit.

Compartment siting, cell venting, and thermal-runaway mitigation are addressed by ISO 23625, the small craft lithium-ion standard whose subject matter reached leisure boats by way of the commercial fleet. A non-compliant lithium install is a Category A or B finding on its facts, not a Category C note. Record too whether the retrofit was notified to the insurer and whether installation documentation exists, since added energy storage raises post-construction conformity questions under the Recreational Craft Regulations 2017 and underwriters are actively asking.

RCD, MCB, RCBO: Not the Same Thing

An MCB is an overcurrent device: it protects the cable against a sustained overload or a short. An RCD compares live and neutral and trips on the imbalance that means current is leaking to earth, which is to say through a person. Neither implies the other. An RCBO combines both functions in one module on one circuit.

So "consumer unit with 30 mA RCD fitted" is not by itself a finding of adequate protection. Split-load boards put only one bank of ways downstream of the RCD, and older installations often run the immersion heater or a ring main on the unprotected side. Identify which ways sit behind the RCD and say so. Press the test button and record that it operated, but claim no more than that: the button proves the mechanism moves, not the trip current, the trip time, or the integrity of the earth path.

RCD type is the modern finding. A Type AC device can be blinded by smooth DC residual current, and a boat with an inverter/charger, a DC-DC converter, solar controllers, LED drivers and switch-mode supplies is exactly the environment that produces it. Current practice has moved firmly away from Type AC. Record the type marking (AC, A, F or B) and flag a Type AC unit on a vessel carrying that kit.

There is a second way an RCD is quietly rendered useless. When the vessel runs on its inverter rather than shore power, a downstream RCD only works if the inverter's output neutral is referenced to earth. Many inverters close an internal neutral-to-earth relay on transfer to invert mode. Many do not, and a floating output produces no imbalance for the RCD to see. Record whether the inverter provides that reference, and confirm the changeover arrangement: ISO 13297 requires a device that simultaneously breaks both active and neutral when changing power sources, with protection against backfeeding to shore power.

Then test the outlets. A plug-in socket tester takes half a minute per outlet and is the right field tool, but it screens rather than verifies. It will not detect a borrowed neutral, high earth-loop impedance, or an earth improperly connected to neutral, and certain combinations of reversal fool it. State that limitation alongside the result. What it catches reliably is an open earth and reversed polarity. Reversal leaves the appliance switch in the neutral, so the wiring stays live at the socket with everything switched off. It may originate at the pontoon, the inlet, or an on-board socket, and each has a different responsible party, so locate it rather than merely reporting it. ISO 13297 expects a reverse-polarity indicating device in shore power systems, excepting unpolarised systems on double-pole branch protection and craft where an isolation transformer establishes polarity on board.⁶

The inlet deserves a line: type and IP rating (BS EN 60309-2 "commando" on any properly fitted installation), mounting above the waterline, condition of the flap and seal, state of the pins. Then the shore lead. Look for a domestic 13 A adaptor arrangement, joined leads, a cable reel left coiled while loaded, and brown heat discolouration around the live pin. Browning is arcing, and arcing is the last visible stage before a fire.

Last, the galvanic isolator: a corrosion-control device fitted into a safety-critical conductor, which is exactly why it earns attention. Back-to-back diode pairs in the shore earth block the millivolt-scale galvanic current that would otherwise eat the most active metal on the pontoon, roughly below 1.2 to 1.4 V, while passing AC fault current so protection is unimpaired. Here the primary standard is the stricter one, and it is worth quoting. ISO 13297 clause 6.9: "When a galvanic isolator is fitted in the protective conductor, failure of the isolator shall not result in an open circuit." That is a normative fail-safe requirement, with no alternative offered. ABYC A-28 is the more permissive of the two, requiring a status monitor that gives audible or visible indication of failure unless the unit is of fail-safe design.⁷ An older unmonitored, non-fail-safe unit that has failed open has silently removed the vessel's earth and nothing on board says so. Record the make, whether a status indicator is fitted, and whether it reads healthy. Check too that nothing bypasses it: a TV aerial coax or shore-side data connection can carry the earth straight past the isolator and defeat it. An isolation transformer does the job more completely where weight and cost allow.

Steel Craft and the Earth Path

On a steel narrowboat or widebeam, the AC protective earth must be connected to the hull. If that connection is missing, broken, or high-resistance, two things follow. A fault in any appliance can raise the hull to a live potential relative to the water. And the fault current has no low-resistance route back, so the RCD may never see enough imbalance to trip. Someone stepping aboard from a wet pontoon closes the circuit themselves. This is the highest-consequence electrical finding on inland craft, and unlike most it is traceable by eye: follow the earth conductor from the consumer unit earth bar to its bonding point on the plating, and look hard at that joint. ISO 13297 supplies two criteria to judge it against. Clause 6.3 requires a final, single connection of the AC protective conductor to the hull, so a second or improvised bonding point is itself a finding. Clause 6.4 requires that connection to sit above any anticipated water accumulation, which rules out the bilge-level joint found on a good many conversions.⁶

That on-board bond is not in tension with the PME prohibition above. The two work together: the craft bonds its protective earth to the hull, and the marina must supply that earth from a TT arrangement with a local electrode, a TN-S supply, or an isolation transformer, never from a PME earthing facility.

Keep the two corrosion mechanisms separate in the report, because they carry separate remedies. Galvanic corrosion is the millivolt-scale dissimilar-metals problem the isolator addresses. Stray-current corrosion is a leakage current finding a path through the water, orders of magnitude more aggressive and capable of eating plating in a season. Where it can be measured, record hull potential against a silver/silver-chloride half-cell reference electrode alongside anode condition and bonding continuity. That is the numeric record pairing with the ultrasonic thickness readings, because stray current wastes plating from the outside in and the gauge is where it shows.

Inland boats also concentrate the multi-source charging problem: alternator, solar, and mains charger all feeding one bank, added by different hands over two decades, sometimes with no single point of isolation. The narrowboat and inland craft primer sets out the regulatory frame. Record the Boat Safety Scheme certificate and its expiry alongside the survey findings. The BSS examination covers battery security, terminal insulation, fusing and isolation directly, so a current certificate sitting beside a serious electrical finding is worth a sentence of explanation.

Classification: A, B, or C

The test is the one governing the whole IIMS Category A, B and C system: does the defect present a danger to life or vessel if the vessel is used? Electrical findings split cleanly on it. Anything that can ignite, or that can put a person into a fault path, is Category A. Anything that degrades protection without removing it goes to B.

Two boundaries move, so state them. Category B items escalate to A where an arcing or ignition path is present: corroded terminations are B, but significant corrosion with compromised insulation inside a battery compartment is A, and the same escalation applies to inadequate ventilation on a gassing bank. Conversely, the absence of a galvanic isolator is not a defect. No standard requires one, an isolation transformer is a legitimate alternative, and a boat that never connects to shore power needs neither. It is a recommendation with reasoning attached, and grading it B is the kind of over-classification that costs a surveyor a broker relationship.

The error to avoid mirrors the one made with seacocks: downgrading because the boat is working fine today. An unfused inverter feed worked fine every day of its life right up until the moment it did not. Equally, do not inflate cosmetic terminal oxidation into an ignition hazard. Grade the mechanism, not the appearance.

Documenting It

Inadequate: "Electrical installation in serviceable condition. Shore power tested and functional."

Professional: "Category A — Shore power, 230 V AC nominal. All four outlets tested with plug-in socket tester: reversed polarity recorded at the galley outlet, remaining three correct; reversal not present at the pontoon pedestal, therefore on board. Shore lead plug shows brown heat discolouration around the live pin, consistent with arcing at the contact. Both to be corrected by a qualified marine electrician and the shore lead renewed before the vessel is next connected to a shore supply.

Category B — Consumer unit beneath the chart table. Single 30 mA Type AC RCD fitted; test button operated and device tripped. Two ways (immersion heater, aft ring main) connected to the unprotected side of the busbar and carrying MCB overcurrent protection only. Vessel carries an inverter/charger and solar controller, for which a Type AC device is no longer appropriate. Recommend RCBO protection on all ways to a Type A or F device within six months.

Scope: consumer unit not opened; no insulation resistance, earth loop impedance or instrument RCD testing carried out. Photographs attached."

The second version names the circuits, the protection actually present, the test performed, where the fault lies, the evidence, and the action. It grades two different problems separately instead of bundling them, and it fixes its own boundaries. Months later it gives an adjuster a dated, evidenced picture of the installation, and it gives the surveyor a defensible account of what was and was not done.

Marine Inspect separates the electrical section into the items that actually get argued over: shore power inlet, consumer unit and RCD, earth bonding to hull, galvanic isolator, inverter/charger, and DC bonding. The make, the reading, the test result and the photographs go against each item at the point of inspection, and they carry into the report as separate graded findings rather than one line for the whole system. That is what a buyer's pre-purchase report needs if its electrical section is to be worth reading.

A closing field note. Count the sources of energy aboard and account for every one: start bank, domestic bank, alternator, shore supply, solar array, generator, and the portable power station somebody has left in a cockpit locker. There are more of them than there were ten years ago, they are usually connected to each other, and the one that starts the fire is generally the one added last and inspected least. Alongside that count, note the smoke and heat detection, the CO alarm and its expiry, the extinguisher service dates and any engine-space suppression. A piece that opens with a burnt boat should close by asking what would have found it early.

Grading an electrical installation and need the record to hold up? Marine Inspect provides a structured electrical section with separate items for shore power, RCD protection, earth bonding, the galvanic isolator and the battery bank, so each finding carries its own reading, test result and photographs. See what a defensible pre-purchase report contains to judge the output.

Read next in the series: Defect of the Week: Through-Hull Seacocks, Defect of the Week: Osmotic Blistering, and Defect of the Week: Chainplate Failure.