Most mobile power systems are assembled rather than designed. A battery is bought, then an inverter is bought to suit the battery, then solar is added when the battery keeps going flat, and each purchase is made without reference to the last. The result works until it is loaded, at which point it fails in an order that is entirely predictable.
This note sets out the sequence we design in, and the one decision that constrains every other.
The first decision is system voltage
System voltage is chosen before anything else, because it determines the conductor, the protection and the cost of every subsequent change.
Take a 3 kW continuous inverter load. Ignoring conversion losses, current draw is the load divided by the bus voltage:
- At 12 V — 3000 W ÷ 12 V = 250 A continuous
- At 24 V — 3000 W ÷ 24 V = 125 A continuous
- At 48 V — 3000 W ÷ 48 V = 62.5 A continuous
Those are not three versions of the same installation. A sustained 250 A run requires parallel 4/0 cable and 400 A Class T protection. The same duty at 48 V is comfortably served by 2 AWG. The cable, the lugs, the busbars, the fuse holders and the labour to terminate all of it scale with the current, not with the power.
The second-order effect is worse than the first. Voltage drop is proportional to current for a given conductor, so a 12 V system loses a meaningful fraction of its bus voltage in the cable itself, and it loses it exactly when the load is highest. A lithium bank whose cells are fine will read as sagging, the inverter will cut out on low-voltage disconnect, and the fault will be diagnosed as a battery problem for as long as anyone is willing to keep replacing batteries.
We specify 48 V on anything built to order. The retrofit market is largely 12 V because that is what the vehicle already is, and converting an existing coach is a different calculation — but for a new platform the decision is not close.
Sizing is a load study, not a preference
The usable questions are, in order:
- What has to run at once? Not the appliance list — the simultaneous worst case. An air conditioner and a compressor starting together is a different number from either running.
- For how long, without generation? This is the storage question, and it is the one most often answered by guessing a battery size first.
- What replaces the energy, and how fast? Solar, shore and alternator charging are three different rates with three different availabilities.
- What happens when all of that is unavailable? Which is where a generator earns its place, and where auto-start stops being a luxury.
Only after those four does a component get chosen. Working the other way round — picking a battery and then discovering what it will run — is how a system ends up oversized in storage and undersized in charging, which is the most expensive combination available.
Surge is a separate specification from continuous
Continuous rating tells you what a system will carry. Surge tells you what it will start. Motor loads — condensers, air handlers, pumps, well systems — draw several times their running current for a short period at startup, and a system sized only on continuous rating will trip on the first compressor start and keep tripping.
This is the reason the 30-foot platform is specified at 20 kVA continuous with roughly 35 kVA of surge headroom. The headroom is not margin for its own sake; it is the difference between a platform that powers a work bench and one that can start the equipment being worked on.
Why a single-vendor ecosystem, and where that reasoning stops
We specify Victron throughout — MultiPlus-II and Quattro inverter/chargers, MPPT solar controllers, Cerbo GX, and VRM for remote monitoring.
The argument is not brand loyalty. It is that the components are designed to talk to each other, the firmware is updatable by whoever owns the unit, the parts are stocked by ordinary electrical distributors on several continents, and the documentation is public. A crew chief in a yard at six in the morning can source a replacement without calling us.
The argument stops at the interconnect between units. Units share AC power over standard shore cords, not DC. Each unit runs its own inverter/charger against its own bank. That costs some conversion efficiency, and it buys three things worth more than the efficiency: one unit failing does not affect the others, there are no proprietary connectors anywhere in the deployment, and any electrician can work on it without a training course.
A DC interconnect between units would be more elegant and would make the fleet dependent on a design only we can service. That trade is the wrong way round for equipment a crew depends on.
Batteries: the constraint that is not negotiable
LiFePO₄ only, in commercially manufactured packs with an integrated BMS. Never NMC, and never salvaged EV modules.
This is an insurance and liability position before it is an engineering one. Salvage cells have no traceable provenance, no warranty, no manufacturer standing behind the pack, and no BMS that anyone will certify. The cost saving is real and it is not worth what it does to the insurability of the unit or the exposure of whoever builds it.
Documentation is part of the system
Every circuit is labelled at both ends, a printed legend lives in the panel, and an as-built schedule is delivered with the unit.
This is listed last and it is not the least important. An undocumented system is one that only its builder can service, which means it is one failure and one unavailable phone number away from being scrap. The documentation is what makes the system maintainable by someone other than us, and that is a property the owner is buying, not a courtesy we are extending.