QuietWatts
Real hardware. Real benchmarks. No cloud required. Method v1.0 Runs 1 Machines 2

What Size UPS a Home Lab Needs

Researched

Seeded with published specifications and cited sources rather than results measured here. Every seeded figure links to where it came from.

Measured at the wall, everything you will plug in.
0.95 is a safe assumption for modern supplies.
Battery, for the runtime estimate
A UPS cuts off well before the battery is flat.
263 VA Minimum VA rating
250 W Minimum watt rating
47 min Runtime ceiling
211 VA Your load in VA
216 Wh Battery, nominal
156 Wh Battery, usable

Look for at least 263 VA and 250 W, and check both numbers on the box. 47 minutes is the ceiling, so plan the shutdown for well inside it.

This calculator needs JavaScript. The conversion is one division and it is written out below.

Runs in your browser. Nothing you type here is uploaded, logged or sent anywhere. The page is a static file and the arithmetic happens on your machine. Same principle as the rest of the site.

How this is calculated

Two separate calculations. Sizing is exact arithmetic. Runtime is an upper bound, and it is important to know which is which.

load VA = load watts ÷ power factor
minimum VA rating = load VA ÷ (headroom percent ÷ 100)
minimum watt rating = load watts ÷ (headroom percent ÷ 100)
battery Wh = volts × amp hours × number of batteries
usable Wh = battery Wh × depth of discharge × inverter efficiency
runtime minutes = usable Wh ÷ load watts × 60

InputDefaultWhere it comes from
Power factor 0.95 A computer supply with active PFC sits near 1. Lower it for older or simpler kit
Headroom 80% Leaves room for startup surge and for the load growing. Your choice, not a standard
Inverter efficiency 0.9 Losses converting battery DC to mains AC
Depth of discharge 0.8 A UPS shuts off before the battery is flat, to avoid destroying it
Battery volts and amp hours Your unit's spec Printed on the battery itself

Worked example, the one above: 200 W divided by 0.95 is 210.5 VA. Divided by 80% that is 263 VA of rating to look for. The battery is 2 × 12 V × 9 Ah, or 216 Wh nominal, of which 156 Wh is actually available after discharge limits and inverter losses. At 200 W that is 47 minutes at best.

What this can't know

Sources

FAQ

What size UPS does a home lab need?

Size it in volt-amps, not watts, and leave headroom. A 200 W load at a power factor of 0.95 is 211 VA, and loading a UPS to only 80% of its rating means looking for at least 263 VA and 250 W. Check both numbers on the box, because a unit can have enough VA and not enough watts.

What is the difference between VA and watts on a UPS?

Watts is real power, the part that does work and heats things up. Volt-amps is apparent power, voltage times current, which is larger whenever current and voltage are out of step. The ratio between them is the power factor. UPS units are rated in both and you have to satisfy both, which is why a 1500 VA unit is often only 900 W. The watt rating is the one people overlook, and it is usually the binding constraint for a lab full of computers.

How long will a UPS actually run my server?

Less than the arithmetic says. The figure here divides usable battery energy by the load, which assumes the battery delivers its rated capacity at whatever current you draw. It does not: pull harder and you get proportionally less out, an effect that gets worse as the battery ages. Treat 47 minutes as a ceiling, halve it for a battery a few years old, and trust the vendor's published runtime chart over any calculator including this one.

Do I actually want a long runtime?

Usually not, and this is where money gets wasted. For most home labs the job of a UPS is not to keep working through an outage; it is to hold everything up long enough to shut down cleanly, which is a few minutes. Buying an hour of runtime costs a great deal more and buys very little unless you genuinely need to keep serving. Spend the money on a unit that notifies your machines to shut down, and on testing that the shutdown works.

Does my server have a power factor of 1?

Close to it, but not exactly. Any computer power supply with active power factor correction, which is essentially all of them made in the last fifteen years, sits somewhere around 0.95 to 0.99 under load. It drops at very light load. Network gear, older kit and anything with a simple linear supply can be considerably lower. If you do not know, 0.95 is a safe assumption that errs toward a slightly larger UPS.

Where to go next