How long will it actually keep things running?

Capacity in kilowatt-hours divided by average draw in kilowatts gives hours. That is the idealized arithmetic behind the runtime this worksheet prints, and it is correct as far as it goes. What it omits is that the same pack has a continuous power rating, the most it is rated to deliver for as long as its energy lasts — 3.84 kW ^[1] for the smaller unit here, and 7.08 kW ^[2] for the larger. The larger pack’s specification line prints that rating in kVA while its product description gives the same number in kW; this worksheet uses it as kilowatts and does not model power factor, so a motor-heavy set of loads close to the rating is one to check with the installer.

Ask the smaller pack to carry a load above its continuous rating for more than a few seconds and the answer is not a shorter runtime: the pack is not rated to deliver it at all. That is why this worksheet checks power beside energy. When the running watts of everything you selected add up to more than the continuous rating, the headline says so instead of giving hours.

The subtler version is the surge. Compressors, well pumps and air-conditioning condensers draw a multiple of their running power while they start; the presets here assume 3 to 5 times. A load that sits comfortably inside the continuous rating while running can be well outside it for the moment it takes to spin up. Whether that moment trips the pack depends on its short-term peak rating and on what else is running, not on the continuous rating alone: the smaller pack’s maker also publishes higher peak output that it can hold for a few seconds ^[1]. The worksheet names the largest surge among your selected circuits and, when it is above the continuous rating, flags it as something to check against the peak rating rather than as a verdict.

The most useful exercise here is not a runtime figure at all. It is deciding, in advance and in daylight, which circuits you actually want alive at three in the morning. Backing up a whole house asks the most of a pack. On the presets below, the freezer, the lights, the router and a medical device together use 7.09 kilowatt-hours a day, about 11 percent of the 66.91 that every preset together would use.

Read the last column against the 10.0 kWh the larger pack holds ^[3]. On these presets a fridge, lights and the internet together come to 4.86 kilowatt-hours a day, so a full pack would carry them for about 49 hours, or 2.1 days, with no recharge. Add the air conditioner and the same selection uses 55.26 kilowatt-hours a day, enough to empty the pack in about 4.3 hours if it can start the compressor at all; the worksheet flags that start-up surge for checking. Both figures are idealized: they divide the pack’s stated usable energy by the presets’ average draw, start from a full charge, and add nothing for standby draw or conversion losses.

These running watts, on-times and start-up multiples are fixed presets we chose ourselves: assumptions, not sourced figures and not measurements of your appliances. The worksheet lets you switch each preset on or off, set the state of charge and enter a daily solar recharge; it does not let you change a preset’s numbers. If your fridge or pump differs, a reading from a plug-in energy meter or a clamp meter on its circuit describes your own appliance, which these assumed presets do not.

The payback tool on this site accepts an outage value and defaults it to zero. That is deliberate. A freezer of food, a night in a hotel, a spoiled work day and an oxygen concentrator are worth wildly different amounts, and no published national figure can price yours.

What we can say is how to make the estimate honestly: count the hours you actually lost last year rather than the hours you remember, multiply by the load you would want carried, and price that energy at what avoiding the disruption is genuinely worth to you. If that produces a number large enough to justify the purchase on its own, the arbitrage arithmetic elsewhere on this site is a bonus rather than the case.