Alternator to house battery: why direct 12V wiring fails

What a house battery actually wants

Charging any battery well is a sequence, not a voltage. A 12V system's charge profile — the number this site's engine returns for every calculation involving a charge source — is 14.4V, held during the bulk and absorption stages, then relaxed once the battery is full. That number comes from OffgridEngine.vCharge(12), the same constant every controller and charger calculation on this site uses, so it isn't specific to solar; it's specific to what a 12V lithium or lead-acid bank needs to see to charge correctly and not sit chronically undercharged or overcharged.

An alternator was never designed to produce that. It was designed to hold a starter battery near full and cover the vehicle's own electrical load, and its regulation target reflects that job, not a house bank's charge curve.

Why the voltage moves around in the first place

Three things move an alternator's output away from any single number:

• Engine RPM. An alternator is a rotating machine — its raw output before regulation scales with shaft speed, and while the internal voltage regulator tries to hold a flat target across the RPM range, cheap or worn regulators let output sag at idle and creep up at highway RPM.
• Electrical load. Headlights, blower motors, ECU draw and everything else on the vehicle's own circuits pull from the same alternator output before any of it reaches a house battery. Add load, and the regulator's target voltage stays the same on paper, but the available surplus for anything downstream shrinks.
• Smart/ECU-controlled alternators. Many vehicles built in the last ~15 years vary alternator field current under ECU command specifically to reduce parasitic engine drag and improve fuel economy — deliberately running the alternator at a lower voltage, sometimes closer to 12.6–13.0V, whenever the starter battery reads full and load is light. That's the opposite direction of what a house bank needs for a real absorption charge, and it happens by design, not by fault.

None of this makes the alternator broken. It's doing its actual job — keeping the starter battery serviceable and the vehicle's own systems powered — which was never the same job as fully and correctly charging a second, different battery bank with its own chemistry and its own charge curve.

The second problem: the wire itself

Even if an alternator held a perfect, constant 14.4V at its terminals, the cable run from the engine bay to a house battery in a van or trailer isn't free. Every meter of copper has resistance, and OffgridEngine.vdrop() — the same voltage-drop formula used throughout this site for wire sizing — applies here exactly as it does to a solar circuit: drop scales with current and round-trip length, and falls with cross-sectional area.

Take a realistic case: 30A flowing through 6 AWG copper (13.3 mm² — the wire size a 30A circuit actually calls for) over a 5 meter one-way run, so 10 meters of round-trip conductor:

That's a well-sized wire on a short run, and it still eats real voltage before it reaches the battery. Stretch the same 30A circuit to an 8 meter one-way run on undersized 10 AWG (5.26 mm² — a size that's fine for a 30A general-purpose DC circuit at that length by ABYC's 10% limit, but not by much) and the drop grows to 1.57V, 10.9% of the 14.4V target — run through OffgridEngine.vdrop(8, 30, 5.26, 14.4), the exact function this site's own wire-gauge tables use. A house battery sitting at the end of that wire during "absorption" charging is really seeing something well under 13V, nowhere near the voltage the charge stage requires, for the entire time the engine runs.

A direct wire connection has no way to know this is happening and no way to correct for it. Whatever voltage arrives at the battery terminals is whatever the alternator produced minus whatever the wire consumed, and that number changes every time RPM, vehicle load, or driving duration changes.

What a DC-DC charger actually does

A DC-DC charger is a buck-boost converter: it takes whatever voltage arrives at its input — swinging alternator output, minus whatever the input wire dropped — and regulates its output to the house battery's actual charge profile, holding bulk, absorption and float stages correctly regardless of what the alternator is doing upstream. It solves both problems in the same box: it doesn't care that the alternator's voltage is unstable, because it's actively regulating rather than passing voltage through, and several models compensate for drop on the alternator-side wire by sensing and adjusting for it, rather than delivering whatever happens to survive the run.

The Victron Energy catalog in this site's database includes several verified 12V-input, 12V-output DC-DC chargers, differing mainly in rated output current and isolation:

Isolated models galvanically separate the vehicle's starter/chassis electrical system from the house bank — the input and output share no common return path — which matters most on vehicles with sensitive ECUs or where the two battery systems shouldn't share a ground fault. Non-isolated units are lighter and cheaper and work fine when that separation isn't a requirement. Either way, the regulation behavior at the output is the same: a fixed, correct charge profile, delivered independent of what's happening on the input side.

Why this site models DC-DC charging as scheduled driving hours, not a constant source

Unlike solar, which this site treats as continuous generation across daylight hours, a DC-DC charger only produces anything while the engine is actually running. The simulation engine (assets/js/simulate.js and its PHP port, Calc::simulate()) models this explicitly: each DC-DC charger entry carries a drive_hours array — one value per weekday — and a drive_start_hour, and during the simulated year, that charger only contributes current during the hours a given weekday's driving window says the vehicle is moving. Outside that window, its contribution is zero, exactly like it would be sitting in a driveway.

Inside the simulation loop, one line does the actual math for every hour where driving is happening:

extDcdc += ch.output_a * v_charge * ch.efficiency

That's rated output current, times the system's charge voltage, times the charger's own efficiency — added to the hour's total charge input only for the hours that fall inside that day's drive window. A charger rated for more amps than it's ever run at, or driven for zero hours on a given day, contributes exactly zero to that day's energy balance, which is the same discipline every other component on this site follows: nothing is credited that the schedule doesn't actually produce.

A worked example: what a day of driving is actually worth

Take the Orion-Tr Smart 12/12-30 Isolated (30A output, 87% efficiency) charging a 12V house bank. The charge voltage target is OffgridEngine.vCharge(12) = 14.4V, so the charger's output power is 30 A × 14.4 V × 0.87 = 375.8 W — this is the charger's own regulated output, not the raw alternator terminal voltage, which is exactly the point: whatever the alternator was actually doing upstream, the DC-DC unit delivers this fixed number.

A day with a short errand run contributes a few hundred Wh — useful, but not a substitute for a real charge source on a day the vehicle stays parked. A longer travel day covers a meaningful fraction of a typical daily load outright. Either way, the number is only ever nonzero during the scheduled driving window; there is no "trickle charge while parked" credit anywhere in this model, because a direct-wired setup wouldn't produce one reliably either — and a properly regulated DC-DC charger, correctly, produces none at all when the engine is off.

Try it: your charger, your drive schedule

Enter a DC-DC charger's rated output current, the house system's voltage, and hours driven per day to see the energy it adds.

When a direct wire is defensible, and when it isn't

There is one narrow case where a simpler connection works: a battery isolator or basic relay-based combiner that ties the starter and house battery together only while the engine runs and both batteries use the same lead-acid-family chemistry with similar charge acceptance. That still inherits every voltage swing and wire-drop problem above — it just accepts the imprecision as an acceptable tradeoff for a cheaper, simpler setup on a bank that isn't picky.

That tradeoff stops being acceptable the moment the house bank is LiFePO4. Lithium cells are far more sensitive to being charged at the wrong voltage — chronic undercharging leaves capacity on the table and accelerates imbalance between cells, while any regulator fault that lets alternator voltage run high risks pushing a lithium bank's per-cell voltage past its safe charge limit, something a lead-acid bank tolerates far better. A DC-DC charger's fixed, chemistry-correct output profile isn't a luxury for a lithium house bank charged from a vehicle; it's the component that makes the pairing safe to leave unattended in the first place.

What this means for a build

Three things follow directly from the mechanism above, not from a rule of thumb:

1. Size the DC-DC charger to the vehicle's realistic driving pattern, not just peak output. A 30A charger run for one hour a day adds roughly a third of what four hours of highway driving adds — match the unit's rated current to how much the house bank actually needs replaced on a typical day, not the longest drive of the year.
2. Isolated models are the safer default unless there's a specific reason not to. Galvanic isolation between the vehicle's chassis electrical system and the house bank removes an entire category of ground-fault and noise-coupling problems that a non-isolated unit doesn't address.
3. A direct wire is never the right answer for a lithium house bank. The regulation and voltage-drop compensation a DC-DC charger provides is exactly the protection a lithium bank's tighter charge tolerances require — skipping it to save the cost of the charger risks the far larger cost of a damaged or prematurely degraded battery bank.

The charger catalog lists every verified DC-DC unit in this database with its rated output, input voltage, and efficiency, so the numbers above can be checked against any specific model under consideration.