Why Your UPS Runtime Is Shorter Than Expected (and How to Fix It)

Start here: the two usual suspects

Almost every "my UPS should last longer" case comes down to one of these, and you can tell them apart in under an hour.

1. The battery has aged

Sealed lead-acid (VRLA) batteries lose capacity steadily from the day they are made. In a typical home or office UPS they last 3 to 5 years, less in a warm closet. Capacity loss is gradual until the last year or so, when it often drops off quickly. If your UPS is more than three years old and runtime has shrunk noticeably compared with when it was new, the battery is the prime suspect. See UPS battery lifespan.

Fix: run a self-test, then a controlled runtime test (below). If measured runtime is well under the manufacturer's chart for your load, replace the battery. See how to replace a UPS battery.

2. The expectation was based on the wrong arithmetic

Many people divide the battery's watt-hours by the load in watts and expect that many hours. For lead-acid batteries discharged in minutes, that overestimates runtime badly, often by a factor of two or more. If the UPS is new and runtime is short from day one, the math is the more likely issue, followed closely by a load higher than you thought.

Fix: measure the real load, look up the runtime chart for your model, and use the UPS runtime calculator, which accounts for discharge rate.

Worked example: where the "missing" minutes go

Take a common consumer configuration: a 1500 VA / 900 W line-interactive UPS with two 12 V 9 Ah VRLA batteries in series (a 24 V string). The connected load measures 450 W, half the watt rating.

The naive estimate

That 9 Ah figure is the battery's 20-hour rating: the capacity delivered when it is discharged gently over 20 hours, at about 0.45 A. A UPS does nothing gently.

What the battery actually sees

The inverter must pull 450 W / 0.85 = about 530 W from the battery string. At roughly 24 V under load, that is about 22 A. Datasheets for typical 12 V 9 Ah VRLA batteries show approximately the following constant-current capability to a cutoff near 1.6 to 1.75 V per cell (figures vary by maker and model, so read your battery's own sheet):

At 22 A, the battery sits right at its roughly 10-minute rate. So the realistic runtime with a healthy, fully charged battery is about 10 minutes, not 24. That matches the published runtime charts of many units in this class at half load, which typically land in the 8 to 15 minute range.

Now halve the load

At 225 W the string supplies about 265 W, or about 11 A. That is close to the 30-minute rate in the table, so runtime is roughly 25 to 30 minutes. Halving the load nearly tripled runtime. The naive formula would have predicted exactly double.

Now age the battery

Suppose the battery has faded to about 65% of its original capacity after four years in a warm room. You might expect 65% of 10 minutes, about 6.5 minutes. In practice it is often worse: an aged battery has higher internal resistance, so its voltage sags further under a 22 A draw and hits the UPS low-voltage cutoff sooner. Runtimes of 3 to 5 minutes at this load would not be unusual, and the UPS may shut off almost immediately on a heavy load spike.

Why lead-acid behaves this way: Peukert, briefly

At high currents, the chemical reaction at the plates cannot keep up: the electrolyte near the plate surfaces is depleted faster than it diffuses in, and internal resistance turns more energy into heat. The voltage sags and reaches the cutoff while usable material remains. Let the battery rest and some of that capacity "comes back," which is why a UPS that shut off can sometimes run briefly again after a few minutes.

Peukert's law is the classic approximation:

For VRLA batteries, k is typically in the range of about 1.1 to 1.3. The law fits moderate rates reasonably but is less accurate in the 5 to 15 minute range typical of UPS use, which is why manufacturers publish runtime charts and constant-power tables. Use those when you can. For a deeper treatment, see UPS runtime explained.

Diagnostic flow

Do these in order. The first two cost nothing and catch a surprising number of cases.

1. Check the shutdown settings. If a computer shuts down with battery left, look at the threshold in PowerChute, PowerPanel, NUT, apcupsd or the operating system (see Windows UPS shutdown). A "shut down after 2 minutes on battery" rule looks exactly like a short runtime.
2. Make sure the battery is fully charged. After an outage or a new installation, leave the UPS plugged in for at least 8 hours, ideally 24, before judging runtime. If it never reaches full charge, see UPS not charging.
3. Measure the load. Read the load percentage or watts from the LCD or software, or use a plug-in wattmeter. Check it under real use, not idle. GPUs, monitors at full brightness and laser printers (which should never be on battery outlets) are the usual surprises. See UPS load percentage.
4. Compare with the runtime chart. Find the chart or table for your exact model on the maker's site or in the manual. Our runtime calculator gives a reasonable estimate if you cannot find one.
5. Check the room temperature. A UPS in an unheated garage in winter will deliver less; one in a hot closet will have aged faster.
6. Run a controlled runtime test. With a full charge, connect a known, non-critical load of roughly the size you care about (a lamp with incandescent bulbs or a spare PC), unplug the UPS from the wall, and time it to the low-battery alarm. Do not test with irreplaceable work open.
7. Check battery age and condition. Look for a date code on the battery label. Any swelling, cracking, leaking or heat means stop and replace; see swollen UPS battery.
8. Calibrate if the estimate is the problem. If the measured runtime is fine but the display estimate is wrong, or if you just replaced the battery, run a runtime calibration.

Symptom, cause and fix

Temperature: two different effects

Temperature affects lead-acid batteries in opposite ways over different timescales.

• Cold reduces available capacity now. Battery ratings are referenced to about 77°F (25°C). Near freezing, 32°F (0°C), a VRLA battery typically delivers somewhere around 15 to 25% less, and the loss is larger at high discharge rates. The capacity returns when the battery warms up.
• Heat shortens life permanently. A widely used rule from battery makers is that service life roughly halves for every 15°F (about 8°C) of sustained operation above 77°F (25°C). A UPS in a closet that sits at 95°F (35°C) may need batteries every 2 years rather than every 4.

Placement matters more than most people expect. See where to place a UPS.

The estimate on the display vs reality

A UPS does not measure remaining energy directly. It estimates runtime from battery voltage, load current and a stored model of the battery's capacity. Several things push that estimate away from reality:

• A new battery without calibration. The UPS may still be using a capacity figure learned from the old one, in either direction.
• Fluctuating loads. A PC that idles at 120 W and games at 450 W will show wildly different estimates minute to minute. The estimate is only as good as the load at that moment.
• Battery aging between calibrations. Capacity falls, the model does not know, and the estimate stays optimistic until the next calibration or failed self-test.
• Aftermarket batteries. Batteries of the same size can have different high-rate performance. See OEM vs aftermarket UPS batteries.

What calibration does and does not do

On many APC Smart-UPS and some Back-UPS models, runtime calibration is started from PowerChute or the front panel; CyberPower, Eaton and Tripp Lite units with the feature typically expose it in their software. The unit usually requires a full charge and a minimum load, then runs on battery until it is nearly depleted and records the result. The exact requirements vary, so check your manual.

Calibration fixes the estimate, not the battery. Each deep discharge consumes some of a lead-acid battery's limited cycle life, so calibrate after a battery replacement and perhaps once a year after that, not monthly.

Fixes, from cheapest to most expensive

1. Shed load on battery. Move printers, speakers, secondary monitors and chargers to surge-only outlets. Because runtime is nonlinear, removing even 100 W from a 450 W load can add several minutes.
2. Correct the shutdown settings so the computer uses the runtime you actually have, while still leaving a margin for a clean shutdown.
3. Improve the environment. Move the UPS out of hot closets and cold garages.
4. Replace the battery with one of the correct type, voltage, capacity and terminal style. See UPS battery sizes and terminals.
5. Add an external battery pack if the model supports one. See extended runtime battery packs.
6. Buy a larger UPS or split the load across two. A larger unit at lower percentage load benefits twice: more stored energy and a gentler discharge rate. Start with how much runtime you need.

A note on lithium UPS units

UPS units with LiFePO4 batteries show a much smaller drop in capacity at high discharge rates, so the naive Wh estimate is closer to reality, and they typically tolerate heat and cycling better. They cost more up front. If short runtime is driving you toward a new unit, the trade-offs are covered in lithium UPS pros and cons and lead-acid vs lithium UPS.