Two questions get asked as if they were one. Will the battery run my house, and how long will it run my house. They are separate questions, they are answered in different units, and a battery can answer one of them well while failing the other outright.
Whether a load runs at all is decided in watts — everything drawing at once, plus the single largest starting surge, in one instant. How long it keeps running is decided in watt-hours, the energy sitting in the bank.
Neither number can be worked out from the other. Doubling the bank doubles the hours and does nothing whatsoever for the surge. Doubling the inverter clears the surge and adds not one watt-hour of runtime.
That is the whole subject. Everything below is the arithmetic behind those two tests, run on the defaults the calculators on this site already use.
Two numbers, two units
The load list for both tests is the same. Taking four presets from the appliance energy calculator — a refrigerator at 150 W, an LED TV at 100 W, a desktop PC at 200 W and a ceiling fan at 60 W — gives a running load of 510 W.
That 510 W figure feeds the energy test. The power test needs a different number from the same house: the worst instant rather than the steady draw, which is where a starting surge enters and where the arithmetic stops being intuitive.
The power test: can it start?
The generator size calculator defaults describe a house drawing 3,000 W with a largest starting surge of 2,200 W. Peak demand is the sum of those, so the instant that matters is 5,200 W.
Applying the 20 percent headroom that tool adds takes the sizing figure to 6,240 W. That surge lasts a fraction of a second, and a motor that cannot get it simply does not start — it hums, draws locked-rotor current, and trips something.
Here is the part that catches people. The surge does not care how large the bank is. Stored energy is irrelevant to an instantaneous power demand, so a bank holding days of runtime still fails this test if the path out of it cannot deliver 5,200 W.
What size generator do I need works this test through in detail — why only the single largest surge counts, and why the headroom is not padding. The same peak applies unchanged to an inverter.
The energy test: how long does it last?
Energy begins with amp-hours multiplied by voltage. The battery watt-hours calculator puts 100 Ah at 12 V at 1,200 Wh, or 1.2 kWh. That is the nameplate figure, and it is not the figure you get to spend.
Depth of discharge takes the next cut. At the 50 percent a lead-acid battery is designed around, 1,200 Wh leaves 600 Wh usable. At the 80 percent LiFePO4 is built for, the same nameplate leaves 960 Wh.
Divide by the load and the hours appear. Against the 510 W essentials, 600 Wh runs 1.18 hours and 960 Wh runs 1.88 hours — from batteries with identical amp-hours, identical voltage, and identical stored energy.
The ratio is 0.8 ÷ 0.5 = 1.60×, so chemistry alone buys 60 percent more runtime off the same nameplate. That is the largest free gain available anywhere in this calculation, and it is decided at purchase.
Load size moves the answer just as hard in the other direction. Put that same 600 Wh against the full 3,000 W running load instead of the 510 W essentials and it lasts 12 minutes. The battery did not change; the rate of withdrawal did.
Amp-hours are not hours
An amp-hour is a measure of charge, and the word "hour" inside it is doing real damage. It becomes energy only after multiplying by voltage, and it becomes time only after dividing by a load. Skip either step and the number means nothing.
The battery life calculator defaults show the short version: 100 Ah at 50 percent depth of discharge is 50 usable Ah, and against a steady 10 A draw that is 5.0 hours. Amp-hours divided by amps genuinely does give hours.
That shortcut holds only while everything stays at one voltage and both sides are expressed in amps. Household loads arrive in watts, so watt-hours ÷ watts is the version that survives contact with an appliance label.
The ladder
Holding the essentials at 510 W and depth of discharge at 80 percent, only bank size changes down this table. It isolates the energy test completely and says nothing about whether any row can start a motor.
| Bank | Usable energy | Hours of essentials |
|---|---|---|
| 1.2 kWh | 960 Wh | 1.9 h |
| 2.4 kWh | 1,920 Wh | 3.8 h |
| 5 kWh | 4,000 Wh | 7.8 h |
| 10 kWh | 8,000 Wh | 15.7 h |
| 15.3 kWh | 12,240 Wh | 24.0 h |
| 20 kWh | 16,000 Wh | 31.4 h |
The 15.3 kWh row is the one worth remembering: 12,240 Wh usable, 24.0 hours of essentials, one full day. Note what the ladder cannot tell you. A 20 kWh bank offering 31.4 hours still fails to start the fridge if its inverter tops out below the peak.
Why big banks are 48 volts
Delivering 5,200 W out of a low-voltage bank takes a great deal of current. At 12 V it is 433.33 A. At 24 V, 216.67 A. At 48 V, 108.33 A — the same power, the same stored energy, one quarter of the amps.
That 4.00× ratio between 12 V and 48 V is why serious backup banks are not 12 V. Current is what heats conductors, sizes breakers and melts undersized lugs, and 433 A is a difficult number to build around.
Current is also the input what size wire do I need takes to work out voltage drop, so bank voltage quietly sets the cost of every cable between the battery and the inverter. The solar battery bank calculator sizes banks at 12, 24 or 48 V for exactly this reason.
Raising the voltage adds no energy at all. A bank holds the watt-hours it holds; voltage decides how those watt-hours are delivered, not how many there are.
What a generator does that a battery cannot
Sizing a bank for a full day makes the difference plain. The 510 W essentials over 24 hours is 12.24 kWh, which at 48 V and 80 percent usable depth needs a 15.3 kWh bank — 318.8 Ah.
Carrying the whole 3,000 W load for that same day is 72 kWh, and the bank required climbs to 90 kWh, or 1,875 Ah at 48 V. Six times the load is six times the bank, with no economy of scale anywhere in the arithmetic.
A generator refuses that trade. It makes energy for as long as it is fuelled, so its runtime is set by the fuel supply rather than by anything decided at purchase. A battery only ever gives back what was put into it beforehand.
The generator has to pass the identical power test, though — 5,200 W of peak demand is 5,200 W whether an engine or an inverter is supplying it. Neither option escapes the first test; they differ only on the second.
Common mistakes to avoid
-
Sizing by watt-hours alone. A bank rated for days of runtime still will not start a motor if the inverter cannot reach the 5,200 W peak. Energy capacity and power capacity are bought separately.
-
Treating amp-hours as runtime. 100 Ah is charge. Multiply by voltage for energy, cut it by depth of discharge for usable energy, then divide by the load. Three steps, and the label gives you one.
-
Spending nameplate energy. The 1,200 Wh battery hands over 600 Wh at 50 percent depth of discharge. Planning against the full nameplate overstates every runtime on this page by double.
-
Comparing amp-hours across voltages. 100 Ah at 12 V and 100 Ah at 48 V are not the same battery, the same energy, or the same price. Convert to watt-hours before comparing anything.
-
Sizing for every load at once. The 3,000 W list needs a 90 kWh bank for a day; the 510 W list needs 15.3 kWh. Deciding what genuinely has to run is the largest single lever in the whole exercise.
-
Forgetting the surge exists. A 3,000 W running load with a 2,200 W surge peaks at 5,200 W. A runtime calculation built on continuous watts never sees that instant and cannot warn you about it.
Sizing a battery system that will be wired into a building is work for a licensed electrician or installer. The arithmetic here tells you which questions to ask and roughly what scale of answer to expect; it is not a design.