UPS Runtime Calculator

Enter your load in watts, then either the batteries inside the UPS or two published runtime points from its spec sheet. You get an honest runtime with the whole derating chain shown: the high-rate amp-hour penalty, battery age, inverter losses, and how hard the answer moves if the load changes.

From battery specs From the spec sheet

Load on the UPS (watts) (a router and modem together draw about 15 to 30 W, a NAS 30 to 60, a desktop PC 100 to 300, a loaded 1U server 200 to 400; a watt meter or the UPS's own display beats guessing)

Voltage per battery (V)

Amp-hours per battery (Ah) (printed on the battery; 7 to 9 Ah is the consumer standard)

Number of batteries (count every battery in the case; a 1500 VA unit usually holds two. Series or parallel does not matter: the energy is the same)

Inverter efficiency (%) (consumer and SMB units run about 85 to 95 on battery; leave 90 if unsure)

Battery age

UPS watt rating (optional) (the W number, not the VA; a 1500 VA consumer unit typically carries about 900 W)

Was this calculator helpful?

How the UPS runtime calculator works

Whether the UPS lives under your desk holding up a router and a PC, or in a server closet holding up something your job depends on, you are asking the same question: when the power goes out, how long do I get? This page answers it two ways. If you know what batteries are inside (the label says: usually 12 V, 7 to 9 Ah, one or two of them), the battery specs mode runs the real engineering, with every derating shown so you can check the work. If you have the manufacturer's runtime figures, the spec sheet mode fits the published curve to your exact load. For consumer units the spec sheet mode is usually the better answer, and we say so plainly: the maker measured their own box, overhead and all, and two of their measurements beat spec-blind physics.

Before either mode, the confusion this corner of the hardware aisle runs on. A UPS wears two ratings, say 1500 VA and 900 W, and they are not the same number. Think of a glass of beer: volt-amperes are the glass, watts are the beer, and the power factor is how much of the glass is beer rather than foam. Old computer power supplies poured a lot of foam (power factor around 0.6 to 0.7), which is why consumer UPS watt ratings settled near 0.6 times the VA number. A modern PC supply with active PFC pours nearly all beer (power factor 0.95 to 0.99; 80 Plus certification requires at least 0.9), so its volt-amperes and watts are nearly equal, and it is the smaller W rating your load slams into first. Size to watts. The VA figure is the bigger, friendlier number on the faceplate, and it is not the one that protects you. If kW versus kWh has ever tripped you up, this is the same family of confusion wearing a different jacket; our EV charging time calculator untangles that cousin.

The formula

battery watts = load ÷ inverter efficiency
aged Wh = volts × amp-hours × batteries × age factor
usable Wh = aged Wh × min(1, (aged Wh ÷ (battery watts × 20))^{k − 1})
runtime = usable Wh ÷ battery watts
spec sheet mode: runtime = a × load^−b, fitted exactly through both published points

Inverter efficiency is the share of battery energy that comes out as usable AC power: consumer and SMB units run about 85 to 95 percent on battery, and we default to 90. The age factor is our editorial derating for sealed lead-acid's decline in service (100, 80, or 60 percent). And k is the Peukert exponent, the heart of the page: we implement Peukert's equation itself, not a flat derating, with k = 1.2. Published exponents for sealed lead-acid run about 1.1 to 1.3, and 1.2 is the value that reproduces the classic landmarks: the full rating at the gentle 20-hour rate, and about 60 percent of it when the pack is emptied in an hour. That 20 in the formula is the rating convention itself: the amp-hours on the label were measured over 20 unhurried hours, a luxury a UPS on battery never gets.

Worked example

A common 1500 VA / 900 W consumer UPS carries two 12 V, 9 Ah batteries: 12 × 9 × 2 = 216 Wh of nameplate energy.

Carrying 60 W of router and NAS: through a 90 percent efficient inverter the load pulls 60 ÷ 0.9 = 66.7 W from the battery. That drains the pack in a couple of hours rather than 20, and at that rate Peukert says only 69.5 percent of the rating is usable: 150.1 Wh. Runtime is 150.1 ÷ 66.7, about 2 h 15 min. Your network shrugs off any outage shorter than a movie.

The same UPS under a 400 W desktop PC: the battery now supplies 444.4 W, the pack would be flat in minutes, and the usable share falls to 47.5 percent: 102.7 Wh. Runtime is about 14 min. Same box, same batteries, and under seven times the load costs nearly ten times the runtime. That asymmetry is Peukert's law doing exactly what it does.

Cross-check with the spec sheet mode: suppose that unit's sheet publishes 12 minutes at half load (450 W) and 4 minutes at full load (900 W). Fitting the curve gives b = 1.58, and evaluating at 400 W lands on about 14 min, within a minute of the physics. At 60 W, though, the fitted curve extrapolates far below the published range and returns 4 h 52 min against the physics mode's 2 h 15 min. Below the range, both modes are guessing; trust the shorter answer and the calculator's honesty note.

Why nameplate amp-hours lie at UPS speeds

The 9 Ah printed on a UPS battery was earned the easy way: drained over 20 hours at a current so small the chemistry never breaks a sweat. Ask for the energy faster and the battery simply has less to give, because the reaction cannot reach the deeper active material before the voltage sags below useful. This is Peukert's law, and it is not a small effect: a sealed lead-acid pack emptied in one hour delivers only about 60 percent of its rating, and a UPS carrying a real load drains its little pack far faster than that. Our worked example's 400 W case gets 47.5 percent of nameplate. Marketing did not lie to you, exactly; the label just answers a question a UPS never asks. It also runs the other way, which is the good news on the page: half the load is more than double the runtime. In our spec-sheet example, dropping from 450 W to 225 W takes the published 12 minutes to 36. Lithium UPS packs, for what it is worth, hold nearly flat at high rates (Peukert exponents around 1.05), which is one honest reason they are taking over the data center.

The other quiet thief is age. Sealed lead-acid fades in service whether you use it or not, which is why VRLA batteries are on a 3-to-5-year replacement schedule everywhere that people check. The failure mode is cruel: the UPS sits there passing self-tests and chirping cheerfully for years, and the one time the grid actually drops, the tired pack sags in seconds and the "protected" computer goes down with everything else. Our age select prices this honestly: the worked example's 2 h 15 min becomes 1 h 43 min on a pack a few years old and 1 h 13 min on one due for replacement, and those tiers are labeled estimates, not measurements. Write the install date on the battery with a marker, run the UPS's real load test once or twice a year, and treat a shortening self-test runtime as the battery telling you the truth early, which is the only time the truth is cheap.

What to actually put on the UPS

Here is the doctrine the runtime table teaches: put the network first. Twenty watts of modem and router runs for around eight hours on the same UPS that carries a gaming PC for a quarter of an hour. Keep the network gear on battery and an outage becomes an inconvenience with wifi: phones stay connected, laptops run on their own batteries, and the smart home keeps its brain. The PC, the monitors, and the laser printer (never the laser printer; its fuser pulls more than the whole rest of the desk) belong on the surge-only outlets, with the PC getting just enough battery time to save and shut down. And mind the flagged rows in the results table: loading a UPS past its watt rating does not buy a shorter runtime, it buys an alarm and a protective shutdown, which during an outage means everything goes dark anyway.

A word for the server closet, where this math has a budget attached: nobody sizes lead-acid for hours, and you should not try. The standard playbook is 5 to 15 minutes of battery, enough to ride through the blips that make up most outages and to cover the gap until either a generator picks up the load or a shutdown agent closes everything gracefully. Runtime beyond that point costs more per minute than a generator does, precisely because of the high-rate penalty this page calculates. If you are sizing the whole room rather than one box, our data center power calculator does the facility-scale math, and the same honesty applies at every scale: the spec sheet knows things the physics does not, and the battery knows things the spec sheet has forgotten by year four.

Sources

Where the numbers on this page come from. We go to the body that publishes the figure, not to another calculator. See how we verify.

What is UPS Battery Efficiency and how it is counted?, Right Power Technology. UPS battery-to-output efficiency runs about 80 to 95 percent depending on chemistry, age, and load; the basis for our 85 to 95 band and the 90 percent default.
How to Calculate Your UPS Inverter Battery Back-Up Time, Lento. Corroborates the inverter efficiency band (80 to 90 percent, with 85 commonly used for safe calculation) in backup-time math of exactly this page's form.
Peukert's law, Wikipedia. The Peukert equation itself and the published exponent range for lead-acid batteries, about 1.1 to 1.3. Our k = 1.2 sits mid-band.
Battery capacity and Peukert exponent (SmartShunt manual), Victron Energy. Corroborating exponent ranges by chemistry: AGM at the premium end runs 1.05 to 1.15 (so our 1.2 is slightly conservative for a fresh AGM pack), and lithium sits near 1.05, the one-sentence basis for the lithium-holds-flat note.
BU-402: What Is C-rate?, Battery University. The classic landmark our exponent is anchored to: a lead-acid battery delivers its full rated capacity over 20 hours but only about 60 percent when discharged in one hour. A test asserts our model reproduces both figures.
Lead Acid battery Downsides and Maintenance, PowerTech Systems. Corroborates the high-rate honesty: Peukert losses mean high-current loads can leave a lead-acid bank delivering as little as 60 percent of its normal capacity.
Understanding the difference between VA versus Watts, Eaton. Watts = VA x power factor, and sizing belongs on the watt rating. The W rating is the binding one for modern high-power-factor loads.
Watts and Volt-amperes: Confusion about Power, Cabling Installation and Maintenance. The classic explainer of the consumer-UPS convention: cheap units are built around a power factor near 0.6, so a 1500 VA faceplate typically means about 900 W of real capacity.
PSU Terminology, PC Power and Cooling. Active PFC raises a PC power supply's power factor to about 0.95 to 0.99, and 80 Plus certification requires at least 0.9: why a modern PC's volt-amperes and watts are nearly equal.
How often should UPS batteries be replaced?, Eaton. The standard VRLA service life of 3 to 5 years behind our age-derating tiers, with heat and discharge cycles shortening it further.
What Happens if Your UPS Overloads?, Eaton. Overload behavior: past its rating a UPS alarms and shuts its output down (or drops to bypass on utility power) to protect itself. Why our over-rating rows warn of a shutdown, not a shorter runtime.
How many Watts do a Modem and Router use?, PC Tech Kits. The network-gear load anchors: a router runs roughly 2 to 20 W and a modem 5 to 15 W, so a typical pair lands in our 15 to 30 W hint.
Server Power Consumption Calculator, The Home Server Blog. Corroborates the larger load anchors: NAS boxes in the tens of watts, desktops from roughly 100 W to the low hundreds, and loaded 1U servers in the 200 to 400 W class.

Frequently asked questions

How long will a 1500 VA UPS run my router?

A long time, and that is the best argument for owning one. A typical 1500 VA unit carries two 12 V, 9 Ah batteries (216 Wh), and a 20 W modem-and-router load runs about 8 h 25 min on it by this page's math. Add a NAS for 60 W total and you still get about 2 h 15 min. The same box under a 400 W desktop lasts about 14 minutes, which is why we suggest putting the network first.

Why does my UPS not last as long as the rating suggests?

Three honest reasons stack up. Amp-hour ratings are measured over a gentle 20-hour drain, and at UPS speeds a sealed lead-acid pack delivers roughly 45 to 70 percent of that number (Peukert's law; this page models it). The inverter eats another 5 to 15 percent turning battery DC into AC. And the battery itself fades every year toward its 3-to-5-year replacement date. A few-years-old pack under a real load can honestly deliver a third of what the fresh-battery, gentle-drain arithmetic promised.

What is the difference between VA and watts?

Volt-amperes are the glass, watts are the beer, and the power factor is how much of the glass is beer rather than foam. A UPS publishes both a VA and a W rating, and the W rating is the one to size against: modern PC power supplies with active PFC run power factors of 0.95 to 0.99, so their watts nearly equal their volt-amperes and the smaller W limit binds first. On consumer units the W rating is typically about 0.6 times the VA number, so a 1500 VA unit really offers about 900 W.

How many minutes do I get at full load?

Usually single digits. Consumer spec sheets typically publish 3 to 5 minutes at the full watt rating, and our worked example uses 4 minutes at 900 W. That is not a defect: it is enough to ride through the sags and blips that make up most power events, and enough for the UPS software to save your work and shut the machine down cleanly. If you need more minutes, the cheap lever is less load, not more battery: halving the load more than doubles the runtime.

How often should I replace UPS batteries?

Sealed lead-acid UPS batteries have a service life of about 3 to 5 years, less in a hot closet, and they fade whether or not they ever carry a load. The dangerous part is that a dying pack still passes the quiet self-test; it fails under the one real load that mattered. Write the install date on the battery, run a genuine runtime test yearly, and replace on schedule rather than on failure. Replacement packs cost far less than the UPS did.

What should I plug into the battery outlets?

The modem and router first: 20 W of network gear runs for hours and keeps every phone and laptop in the house on wifi through an outage. Then the desktop PC and one monitor, sized for a clean shutdown rather than a work session. Keep laser printers, space heaters, and vacuum cleaners off the battery side entirely; a laser printer's fuser alone can out-pull the whole rest of the desk and trip the overload.

What happens if I overload a UPS?

It does not just run shorter. Past its watt rating a UPS alarms, and depending on the model and the severity it either drops your equipment to raw utility power or shuts its output down entirely to protect itself, which during an outage means everything goes dark at once. That is why the over-100-percent rows in our results table carry a warning instead of just a smaller number. Size to the W rating with headroom to spare.

Are lithium UPS units worth it?

They fix both of lead-acid's honest weaknesses: lithium packs hold nearly their full rated capacity at high discharge rates (Peukert exponents around 1.05, against about 1.2 for sealed lead-acid), and they last roughly 8 to 10 years instead of 3 to 5. You pay more up front and get more of what you paid for when it matters. For a router-and-modem load the cheap lead-acid unit is still excellent value; for a server closet the lifetime math increasingly favors lithium.