Skip to main content

How Long Will a Battery Run My House? The Two Tests It Has to Pass

A 100 Ah battery is 1,200 Wh, 600 Wh of it usable — and none of that says whether your fridge will start.

By Mohamed Zakrya

Updated · 9 min read

Share
The two ratings that fail independently Two tests, two units Watts decide whether it starts. Watt-hours decide how long it runs. THE POWER TEST an instant, measured in watts running load 3,000 W largest starting surge 2,200 W peak demand 5,200 W A bigger bank does not change this number. THE ENERGY TEST a duration, measured in watt-hours 100 Ah × 12 V 1,200 Wh usable at 50% depth 600 Wh ÷ 510 W of essentials 1.18 h A bigger inverter does not change this number. Neither number can be worked out from the other A bank can hold days of runtime and still refuse to start a motor.
A 100 Ah battery is 1,200 Wh, 600 Wh of it usable — and none of that says whether your fridge will start.

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.

Watts decide whether; watt-hours decide how long One battery, two separate tests Different units, different inputs, and neither result predicts the other. THE POWER TEST — WATTS Can the load start at all? running load 3,000 W largest starting surge 2,200 W peak demand 5,200 W with 20% headroom 6,240 W Lasts a fraction of a second. Bank size is irrelevant to it. THE ENERGY TEST — WATT-HOURS How long does it keep running? 100 Ah × 12 V nameplate 1,200 Wh usable at 50% depth 600 Wh ÷ 510 W of essentials 1.18 h at 80% depth instead 1.88 h Inverter size is irrelevant to it. A bank can pass either test and fail the other.
Two tests, two units: 5,200 W of peak demand decides whether the load starts, and 600 Wh of usable energy decides how long it runs. Passing one says nothing about the other.

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.

BankUsable energyHours of essentials
1.2 kWh960 Wh1.9 h
2.4 kWh1,920 Wh3.8 h
5 kWh4,000 Wh7.8 h
10 kWh8,000 Wh15.7 h
15.3 kWh12,240 Wh24.0 h
20 kWh16,000 Wh31.4 h
Stored energy buys hours, nothing else Hours of 510 W essentials, by bank size Depth of discharge held at 80 percent. Only the bank changes. BANK USABLE 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 15.3 kWh is one full day — and still has to pass the 5,200 W power test.
Bank size against hours of the 510 W essentials at 80 percent usable depth. Runtime scales exactly with stored energy — and every row still has to pass the 5,200 W power test separately.

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.

Same power, one quarter of the amps One 5,200 W peak, three bank voltages Identical power and identical stored energy. Only the current changes. BANK VOLTAGE CURRENT AT 5,200 W 12 V 433.33 A 24 V 216.67 A 48 V 108.33 A 12 V draws 4.00× the current of 48 V for the same peak Current is what heats conductors and sizes breakers — not the watt-hours in the bank.
One 5,200 W peak, three bank voltages: 433.33 A at 12 V, 216.67 A at 24 V, 108.33 A at 48 V. Raising the voltage cuts the current in exact proportion.

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.

Questions people ask

How long will a battery run my house?

That depends on two things the word "battery" does not tell you: how much usable energy it holds and how much power the house is drawing. A 100 Ah battery at 12 V holds 1,200 Wh, and at the 50 percent depth of discharge a lead-acid battery is designed around, 600 Wh of that is usable. Against a 510 W essentials list — a refrigerator, an LED TV, a desktop PC and a ceiling fan — that is 1.18 hours. Against a full 3,000 W household load the identical battery lasts 12 minutes. The battery did not change; the rate of withdrawal did.

Can a bigger battery start a load that would not start before?

No, and this is the most expensive misunderstanding in backup power. Starting a load is a power question measured in watts, settled in a fraction of a second. A 3,000 W running load with a 2,200 W largest starting surge peaks at 5,200 W, and that peak has to come out through the inverter and the battery-side conductors. Stored energy plays no part in it. A bank holding days of runtime still fails to start that load if the path out of it cannot deliver 5,200 W. More batteries buy hours, not amps at the instant of starting.

What does a 100 Ah rating actually mean in hours?

On its own, nothing. An amp-hour measures electric charge, and the word "hour" inside it is misleading. It becomes energy only after multiplying by voltage — 100 Ah at 12 V is 1,200 Wh — and it becomes time only after cutting for depth of discharge and dividing by a load. Where everything stays at one voltage the shortcut does work: 100 Ah at 50 percent depth is 50 usable Ah, and against a steady 10 A draw that is 5.0 hours. For mixed household loads quoted in watts, dividing usable watt-hours by watts is the version that survives contact with an appliance label.

Why do large backup banks run at 48 volts?

Because current, not energy, is what makes a low-voltage bank difficult to build. Delivering a 5,200 W peak from a 12 V bank means 433.33 A. The same peak from a 24 V bank is 216.67 A, and from a 48 V bank it is 108.33 A — exactly 4.00 times less than the 12 V case. The power is identical in all three, and so is the stored energy. Current is what heats conductors, sizes breakers and punishes an undersized connection, so raising the bank voltage lowers the cost and the risk of every cable between the battery and the inverter.

Does depth of discharge really change runtime that much?

Yes, and it is the largest free gain in the whole calculation. The same 100 Ah at 12 V nameplate — 1,200 Wh either way — yields 600 Wh usable at the 50 percent a lead-acid battery is built around and 960 Wh at the 80 percent LiFePO4 is built for. That ratio is 0.8 divided by 0.5, which is 1.60, so it is 60 percent more runtime from batteries with identical amp-hours, identical voltage and identical stored energy. Against the 510 W essentials it is the difference between 1.18 hours and 1.88 hours, and it is decided at purchase rather than at the outage.

How big a bank do I need to run everything for a full day?

Far bigger than most people expect, because the requirement scales with the load and nothing rescues it. A 510 W essentials list over 24 hours is 12.24 kWh, which at 48 V and 80 percent usable depth needs a 15.3 kWh bank, or 318.8 Ah. Carrying a full 3,000 W load for the same 24 hours is 72 kWh, and the bank 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. Deciding what genuinely has to run is the single largest lever available.

Is a generator better than a battery in an outage?

They answer different halves of the problem, and both still have to pass the same power test. A 5,200 W peak is 5,200 W whether an engine or an inverter supplies it. The difference is on the energy side: a generator makes energy for as long as it is fuelled, so its runtime is set by the fuel supply, while a battery can only give back what was stored beforehand. That is why a full day of a 3,000 W load needs a 90 kWh bank but only a tank of fuel. Neither is universally preferable, and neither escapes the starting surge.

Does the starting surge drain the battery?

Barely, in energy terms, which is exactly why it is so easy to overlook. A surge lasts a fraction of a second, so the watt-hours it consumes are negligible against a bank measured in thousands of watt-hours. Its significance is entirely on the power side: for that instant the bank has to deliver the full 5,200 W peak, which at 12 V means 433.33 A. A runtime calculation built on continuous watts never sees that instant at all, and will happily report hours of endurance for a system that cannot start the load in the first place.