UPS Runtime Calculator for a NAS or Home Server

The single most common UPS mistake is reading the big number on the box as watts. It is not. VA is apparent power, watts is real power, and consumer units publish a power factor around 0.6, so a 1500 VA unit typically carries 900 W. The second most common mistake is assuming runtime scales linearly with load. It does not, and the gap widens the harder you push the battery. This calculator handles both.

How long will a UPS run a NAS?

Start from energy, not from the VA number. A battery holds watt-hours, and runtime is watt-hours divided by watts, adjusted downward twice. The first adjustment is inverter efficiency and usable depth of discharge: a sealed lead acid pack cannot be taken to zero, and the inverter loses energy converting 12 or 24 V DC into mains AC. Together those take roughly a third off the nameplate figure.

The second adjustment is the discharge rate itself. Lead acid chemistry delivers less total energy when it is drained fast, an effect described by Peukert's law. Pulling half a unit's rated watts does not halve the runtime you would get at a quarter load, it cuts it by considerably more. This is why manufacturer runtime charts curve steeply rather than falling in a straight line, and why the table above prints a linear estimate next to the real one.

The practical consequence for a home server is cheerful: home NAS loads are tiny relative to any UPS you would buy. A four-bay NAS at 46 W sits at a few percent of a 1350 VA unit, which is exactly the region where batteries behave well and runtime is generous. Get your load figure right first using the power draw calculator, because everything on this page depends on it.

What is the difference between VA and watts?

Volt-amps measure apparent power: volts times amps. Watts measure real power, the part that actually does work. On an AC circuit with a reactive load the current and voltage waveforms do not peak together, so some of the current sloshes back and forth without delivering energy. The ratio of real to apparent power is the power factor, and watts equals VA times power factor.

UPS makers advertise the VA number because it is bigger. Almost every consumer line-interactive unit publishes a power factor of 0.6, which is a design limit of the inverter rather than a property of your equipment. Check the watt figure, always printed somewhere on the specification page, and size against that.

Look at the two 1500 VA rows carrying the same 900 W. Identical headline figure, identical real capacity, and one produces a pure sine wave while the other produces a stepped approximation. The VA number tells you nothing about which is which. Then compare them with the 1500 VA unit rated 1000 W, where a higher power factor buys 100 W more real load from the same nominal size. Watt rating and waveform are two separate questions, and the box front answers neither.

Do I need a pure sine wave UPS?

For a NAS, a mini PC or any modern ATX build, yes. Power supplies with active power factor correction expect a real sine wave, and cheaper line-interactive units produce a stepped approximation while on battery. The abrupt voltage steps look like a fault to an active PFC input stage, and the result ranges from loud buzzing through the supply shutting itself off to the UPS overloading and cutting out. It happens at the exact moment the equipment needed protecting.

Active PFC is not exotic. It is required on most equipment sold in Europe, and it is standard on 80 Plus rated supplies, on the internal supplies of Synology and QNAP units, and on the external bricks that come with mini PCs. The safe rule is to assume you have it.

A pure sine wave model costs perhaps 60 to 100 dollars more than the stepped equivalent, and it also tends to publish a higher power factor, so some of that money comes back as real watt capacity. Our roundup of the best UPS units for a NAS separates the sine wave models from the rest and explains which loads can safely stay on a cheaper unit.

How much runtime do I actually need?

Enough to shut down cleanly, not enough to keep working. This reframing saves most people a hundred dollars. The UPS exists so that the array is never mid-write when the power stops, so that a parity operation is not interrupted, and so that the filesystem never has to be checked after a hard stop. None of that needs hours of battery.

A NAS typically needs 2 to 5 minutes from receiving the shutdown signal to being safely powered off: stopping services, flushing write caches, unmounting volumes and spinning the drives down. Add margin for the delay you configure before the shutdown triggers, since you do not want a five second blink to power the house down, and a target of 10 to 15 minutes at your real load is comfortable. Anything beyond that is buying convenience, not safety.

Two shapes of outage argue for slightly more than the bare minimum. The first is the stutter, where power drops, returns after four seconds, then drops again a minute later. Configure a delay before the shutdown triggers so a blink does not power the house down, and that delay comes out of your runtime budget. The second is the outage that arrives during a parity scrub or a large restore, when the array is at its busiest and drawing closer to its load figure than its idle one. Size against the load number, not the idle number.

Beyond that, more battery buys convenience rather than safety. A 1500 VA pure sine wave unit holding a 46 W NAS for several hours does nothing the same unit at fifteen minutes would not have done, unless your goal is to keep working through the outage, which needs the router, the modem, the access point and the workstation on battery too and changes the sizing problem entirely.

What should go on the battery outlets?

Every consumer UPS splits its outlets into two banks. The battery-backed bank is small, and everything in it eats your runtime. The surge-only bank protects against spikes but goes dark the instant the power does. Sorting your equipment between the two is free and it is the single biggest lever on runtime after the load figure itself.

Two entries deserve emphasis. A laser printer will trip a home UPS, because the fusing element pulls over a kilowatt in bursts and the unit reads that as an overload. And the network gear belongs on battery even though instinct says otherwise: a small 2.5GbE switch costs perhaps 5 W of your runtime and, without it, a NAS being told to shut down by a monitoring host on another machine never receives the message.

How do you make sure the NAS actually shuts down?

A UPS with no data connection is a delay, not a protection. It keeps the array alive for twenty minutes and then drops it mid-write exactly as the outage would have, just later. The USB or serial cable between the UPS and the server is the part that turns battery into safety, and it is the part most often left in the box.

Every serious NAS platform has this built in. Synology DSM has a UPS page in Hardware and Power, so a Synology DS925+ needs nothing but the cable and a checkbox. TrueNAS and Unraid both ship the network UPS tools daemon, usually called NUT, which does the same job and adds something more useful: one machine acts as the master with the cable attached, and every other machine on the network subscribes as a slave. That is how you protect a NAS, a mini PC and a virtualisation host from a single unit.

Four settings decide whether it works when it matters:

• Trigger on remaining capacity, not elapsed time. Battery capacity fades as the pack ages, so a rule such as "shut down at 40 percent remaining" stays correct for years while "shut down after 10 minutes" quietly stops being safe.
• Set the initial delay deliberately. Something like 60 seconds ignores blinks without wasting much battery.
• Decide what happens when power returns mid-shutdown. The safe answer is to complete the shutdown and power back on afterwards, rather than aborting halfway through stopping services.
• Test it once, on purpose. Pull the wall plug with the array idle and watch the whole sequence run. This is the only way to learn that the cable was in the wrong port before an outage teaches you.

How long does a UPS battery last?

Sealed lead acid packs last roughly three to five years, and heat is the main variable: a unit in a warm enclosed cupboard ages considerably faster than one in open air. They fail by losing capacity, not by dying outright, which is the dangerous part. A unit that genuinely delivered 40 minutes when new can be down to 4 with every light on the front still green.

Run the self test the monitoring software offers a couple of times a year and record what it reports for runtime at your normal load. A sharp drop is the replacement signal. On a unit with a user-replaceable pack that is a fifteen minute job and a fraction of the price of the whole unit. On a sealed model the replacement is the unit, which is worth knowing before you buy rather than after.

One more thing worth checking on the specification page: automatic voltage regulation. Line-interactive units correct mild brownouts and overvoltage by switching transformer taps without touching the battery, which matters if your supply sags when a compressor starts. Every cycle a unit spends on battery is a cycle off the pack's life, so a unit that can fix a sag without discharging is a unit whose battery lasts longer.

Finally, remember what the UPS is not. It protects against a power event reaching the array mid-write. It does nothing about a drive dying, which is what parity is for, and nothing about deletion, ransomware, theft or fire, which is what a real backup is for. Those three protections are independent and you need all of them.