How big a backup does an outage actually need

The phrase "what size generator for a power outage" is asked by two different households. One wants to know whether the fridge survives the evening. The other wants to know whether the house is habitable on day four. Those are not the same purchase, and the difference is roughly a factor of three.

There is also a units problem. Machines are sold by output in watts, while an outage consumes energy in watt hours over days. A 2,000 W machine and a 2,000 Wh battery carry the same two numbers and answer different questions: one is a rate, the other is a quantity. This page works in energy and converts to a bank size at the end.

So the calculation has three inputs and one output. The inputs are how much the kit uses per day, how many days the outage lasts, and what the panels contribute. The output is the battery size that covers what the panels do not.

The order matters. Households that start with the machine usually start with a wattage they saw in a review, then discover the kit it can actually run is smaller than expected. Households that start with the kit usually discover the machine they need is cheaper than they feared, because the loads that matter are few.

There is also a useful sanity check at the end. If the answer is several times the capacity of the largest single unit sold, the purchase has quietly become a fixed system, and that is a different budget and a different set of rules. Knowing when the arithmetic crosses that line saves a lot of shopping.

Start with the loads you would genuinely keep running, not with the whole house. Three kits cover most households, and the wattage figures are category typical values rather than measurements of your appliances.

The third row is why heating deserves its own warning. A space heater is a resistance load with no efficiency trick, so a few hours of it exceeds a whole day of everything else. If heat is in the plan, the purchase stops being portable at all.

The load calculator produces this figure from your own list of appliances rather than a kit template, and the appliance wattage library has the per-device numbers behind it.

One caution about the refrigerator row. A fridge is listed with a running wattage and a starting wattage, and the starting figure is several times larger. The daily energy uses the running figure multiplied by duty cycle; the machine specification that matters is the starting figure, because a unit that cannot surge cannot start the compressor at all regardless of capacity.

Duty cycle is the number people guess wrong. A refrigerator compressor runs roughly half the time in a warm kitchen and much less in a cold one, which is why two households with identical fridges report daily figures that differ by a third.

This is the part that surprises people. Doubling the outage does not double the cost, because the panels keep working on every one of those days. It doubles the deficit, which is a smaller number.

Read the last two columns together. Going from one day to three multiplies the bank by three, which is expected. Going from three days to seven multiplies it by about 2.3, which is the panels paying for part of the extra time. The larger the array relative to the kit, the flatter that growth becomes.

This is also why a household with a small array should size for the outage it will actually stay for. A bank that covers seven days is usually money spent on a scenario where the family has already left.

Write the daily balance as a subtraction: production during the day, use over twenty-four hours. If the result is negative, the battery drains a little every day and the outage length decides whether it survives. If the result is positive, the battery is only being used to move energy from noon to midnight.

For a 2,400 Wh kit, a 400 W array at 4.5 sun hours makes about 1,440 Wh a day, so the balance is minus 960 Wh. A 2,048 Wh battery holds about 1,741 Wh usable at 85 % depth of discharge, which covers 1.81 days. On day two of an outage, that configuration is already relying on sunshine it may not get.

The break point is worth calculating for your own numbers, because it is where the whole argument about array size lives. Solve for the array that makes production equal use, then compare it with what your roof or balcony can actually hold.

The battery sizing calculator runs the same balance from the array side, and the storage sizing calculator starts from the load instead.

For the running example that break point is 2,400 Wh divided by 4.5 sun hours and 80 % system efficiency, which is an array of about 667 W. Below that the battery drains every day of the outage; above it the battery only moves energy from noon to evening and the outage length stops being a constraint on capacity.

That single calculation explains most of the disagreement in this category. A household with a balcony array is usually below the break point and genuinely needs a large bank. A household with a full roof array is above it and needs a much smaller one than the load figure suggests.

The deficit is the energy the battery has to deliver across the outage. Dividing it by the depth of discharge gives the nameplate figure, which is the number on the box.

The spreadsheet version of this is one line, and it is worth doing by hand once so the box label stops being a mystery. A maker who permits 85 % is selling you more usable energy from the same cells than one who permits 70 %, which is a real difference in cost per usable watt hour even when the nameplate is identical.

That is the reason to look for a usable figure rather than a nameplate figure, and the reason the chemistry comparison puts depth of discharge above cell chemistry in importance.

There is a practical ceiling to keep in mind. Two units of the same brand can usually be linked, and the combined capacity is the sum, but the output often is not. A pair sized for a three-day outage may still be unable to start a fridge if each unit is small, which is why the surge figure has to be checked on the unit that actually feeds the compressor.

Linking also changes the wiring question. Two units on a bench are still two appliance loads rather than a house supply, so anything beyond plugging devices into them needs the same electrical work as a single larger unit.

Four omissions matter enough to state before the result is trusted.

Heating and cooling. A heat pump or a central air conditioner is a 240 V load and outside the portable category entirely. A resistance heater fits the arithmetic and ruins it, because it dominates the daily total.

Overcast days. Sun hours are an annual planning average. A week of storms can halve production for days at a time, so a system that is exactly balanced on paper fails exactly when the outage happens.

Standby losses. Inverters draw a few watts doing nothing, which matters over seventy-two hours in a way it does not over one evening.

Transfer losses. Any changeover equipment between the battery and the house wiring costs a little energy and, in most homes, requires an electrician.

The honest way to use the result is to add margin and then check it against the worst day rather than the average one.

A useful habit is to run the calculator twice: once with the sun hours you would see in June, and once with half of them. The gap between the two results is the margin the plan actually needs, and it is usually larger than people expect in the northern half of the country.

Standby losses deserve a number rather than a mention. An inverter idling at five watts spends 120 Wh a day doing nothing, which is a tenth of a minimal kit and worth knowing before a three-day outage is planned around a tight bank.

The same outage asks for very different hardware depending on what is on the list.

The refrigerator kit. A fridge is a compressor, so it needs a surge rating several times its running watts. Capacity alone will not start it. This is the kit where the surge figure on the box decides the purchase.

The medical kit. A CPAP or a concentrator runs all night, which shifts the demand to the dark hours. Panels contribute nothing at the time of use, so the battery carries the whole load and the array only refills it. Size the battery for the night and the array for the refill.

The work-from-home kit. Modems, a laptop and a monitor are small and steady, and they are the easiest kit to cover. A 288 Wh station runs this for most of a working day, which is why the smallest units exist at all.

Those three are different purchases, and the calculator above is what separates them before money is spent. Two other situations point away from a portable battery entirely, and it is worth saying so on a page that sizes them.

Two situations point away from a portable battery and it is worth saying so on a page that sizes them.

A multi-day outage with 240 V loads in scope. A well pump, a dryer or a central air conditioner cannot run from a 120 V box, so no amount of capacity helps. That is a generator or a fixed system with a transfer switch, and this calculator is the wrong tool.

A household that will evacuate. If the plan is to leave for anything longer than a night, a large bank is insurance against a scenario that will not happen. A small station for the first evening plus a decision rule to leave is cheaper and just as effective.

Where fuel generators are the realistic alternative, the comparison of a battery station against a gas generator covers the trade honestly, including the outdoor siting rule that decides it for many households.

These are recorded units across the range, ordered by capacity. None of them carries a printed price here, because a printed price is wrong within weeks; the retailer's page is the only current source.

For the ranking behind these picks, the portable power station comparison sorts the same units by capacity and output, and the home backup comparison covers sustained multi-day coverage.

Sizing by watts instead of watt hours. A 2,000 W output says nothing about how many days it lasts, and the days are what an outage costs.

Using the whole house as the kit. The daily total then becomes 30,000 Wh and the answer is a fixed system, which may be right, but it is not the purchase being searched for.

Forgetting the surge. A battery that holds enough energy but cannot surge will not start a fridge, and the compressor is the whole point of the purchase.

Planning around the average day. An outage happens in bad weather, when panels produce least. Add margin for the worst day, not the average one.

Buying for a seven-day event you would evacuate from. The most common overspend in this category is a bank sized for a scenario the household will not stay for.

If a figure here disagrees with a calculator on this site, the calculator is the source of truth and the changelog records each correction.

The common thread across all five is planning in the wrong unit. Outages are measured in days, batteries are sold in watt hours, and machines are advertised in watts. Converting everything to watt hours before comparing anything prevents most of these mistakes at once.