What Is a UPS? How an Uninterruptible Power Supply Works

The name oversells it a little. A UPS does not make power uninterruptible forever; it makes the interruption invisible to your equipment for a limited time. That time is long enough to ride through the flickers and short outages that cause most data loss, and long enough to save work and shut down cleanly when an outage drags on.

The four parts inside every UPS

Open up any UPS, from a small desk unit to a rack-mount model, and you find the same building blocks. How they are arranged is what separates the different UPS types, but the parts are always there.

1. Rectifier and charger. Converts incoming AC into DC to keep the battery charged. In small units this is a modest charger that keeps the battery at a float voltage (around 13.5 to 13.8 V per 12 V block for sealed lead-acid). In online units the rectifier is sized to carry the full load as well.
2. Battery. Stores the energy. Most consumer and small business units use one or more 12 V valve-regulated lead-acid (VRLA) blocks, wired in series for 24 V, 36 V, 48 V or higher. A growing number use lithium iron phosphate. See UPS battery types.
3. Inverter. Turns battery DC back into 120 V (or 230 V) AC using fast-switching transistors and a transformer or high-voltage stage. The inverter's design decides the output waveform: pure sine or a stepped approximation.
4. Transfer and control. A relay (or static switch) and a microcontroller that watches the input voltage and frequency many times per cycle. When input goes out of tolerance, it disconnects the utility and connects the inverter.

Around those four parts you usually find surge suppression on the input, an EMI/RFI noise filter, often an automatic voltage regulation (AVR) transformer, a USB or serial port for shutdown software, and on business models a slot for a network management card.

What happens when the power fails

Walking through an outage second by second explains most of what a UPS can and cannot do. The timeline below is for a typical line-interactive desktop unit; an online unit skips the transfer step entirely.

The important row is the last battery row. If nothing tells the computer to shut down, the UPS only delays the crash. Configuring shutdown software is what turns a UPS from a delay into protection.

What a UPS protects against (and what it does not)

The table below assumes a healthy line-interactive unit, the most common type for homes and offices. For definitions of each disturbance see power problems explained.

Two gaps catch people out. Surges arriving on a cable modem coax or an Ethernet run from an outdoor camera bypass the UPS entirely, so see lightning and surges for the full picture. And a UPS's built-in suppression is modest; for serious surge exposure consider whole-house surge protection as an upstream layer.

Realistic runtime expectations

Runtime is the number most people misjudge. A desktop UPS battery is small: a common 1500 VA consumer unit carries two 12 V, 9 Ah blocks. That is about 216 Wh of nominal energy, comparable to a large laptop battery or two.

Worked example: why the label energy is not what you get

The naive calculation for a 450 W load on that unit looks generous:

Real units deliver less, often around half that. Lead-acid batteries are rated at a slow 20-hour discharge, and at a 10-minute discharge rate they typically deliver only around half of their rated capacity. The UPS's own electronics also draw power on battery. Rerunning the math with roughly 45% usable capacity:

That is in line with what manufacturer runtime charts typically show for this class of unit at half load. The runtime explainer covers this nonlinearity in detail, and the runtime calculator models it.

Two factors shrink these numbers over time: battery aging (capacity fades steadily over the service life) and heat (batteries kept above roughly 77 °F / 25 °C age faster). Plan on having noticeably less runtime in year three than in month one. If you need hours, look at extended runtime battery packs, a power station or a generator.

Who actually needs one

Not every device benefits equally. Use this as a quick screen.

Misconceptions worth clearing up

"A bigger VA number means longer runtime"

Not necessarily. VA describes how much power the inverter can deliver at once; runtime depends on battery energy in watt-hours. Two units with the same VA rating can have very different batteries. Read VA vs watts for how the ratings relate.

"A UPS is just a surge protector with a battery"

The surge suppression is the least important part. The value is in the switching, regulation and shutdown signaling. The differences are covered in UPS vs surge protector.

"Once it's installed, I'm covered"

Batteries degrade silently. A unit that has never been tested may fail the first time it is needed. Run the self-test periodically and plan on battery replacement every few years; see when to replace a UPS battery and the maintenance checklist.

"Any UPS can back up a sump pump or fridge"

Motor loads draw a startup surge several times their running power and often need far more energy than a desktop unit stores. These are specialist jobs; see UPS for sump pump and UPS for refrigerator before buying.

Where to go next

If you are shopping, the logical order is: decide what must stay on and for how long (how much runtime do you need), measure the load, choose the topology, then size it with the sizing calculator. The buying guide pulls those steps into one checklist.