How do I size a UPS for my homelab?

You size a UPS for a homelab by measuring how many watts your gear actually draws and choosing a UPS whose watt rating comfortably exceeds that load. Then you check that its battery gives you enough minutes to shut everything down cleanly. For most home setups the goal is not to ride out a long outage. It is to survive a short blip and shut down safely before the battery runs out.

Watts versus VA

UPS units are labelled with two ratings: VA (volt-amperes) and watts. VA is “apparent power” and watts is “real power.” Watts divided by VA is the power factor. The watt rating is the one that matters for sizing, because your equipment’s draw is measured in watts.

Do not size from the VA number alone. The watt rating is often noticeably lower than the VA rating, and the exact ratio varies by model. Read the spec sheet for the watt rating and keep your load well under it. A common rule of thumb is to stay at or below roughly 50 to 80 percent of the rated watts, which leaves headroom for startup surges and future additions. The lower end of that range gives more margin.

Measure your real load

Nameplate ratings on power supplies show the maximum a device is rated to draw, not what it does draw. Your real load is usually much lower.

The best way to find it is a plug-in power meter placed between the wall and a power strip holding your gear. Let it run for a day or more and note both the typical and peak readings. Many UPS units also show the current load on their own display, which works too.

If you cannot measure, look up typical idle draw for each device and add them up, but treat that as a rough guess. Include everything you plan to keep on battery: servers, switch, router, modem, and any PoE devices powered through the switch. PoE (Power over Ethernet) draw is part of the switch’s load, so it counts too. Consider leaving non-essentials, like a monitor or a laser printer, off the battery outlets.

Battery energy and estimated runtime

Runtime depends on stored energy and how fast you drain it. A rough estimate:

1. Battery energy (watt-hours) is battery voltage multiplied by amp-hour capacity.
2. Usable energy is that figure times the fraction of the battery you are willing to use. Batteries are generally not drained to zero, as deep discharge shortens their life.
3. Multiply by the inverter efficiency, since some energy is lost converting battery power to AC.
4. Divide by your load in watts to get hours. Multiply by 60 for minutes.

Many manufacturers publish runtime charts for specific loads. Use those in preference to your own arithmetic when available.

Worked example

These numbers are round and for illustration only.

• Measured load: 150 W (a small server, a network switch, a router, and a modem)
• Battery: two 12 V, 9 Ah batteries wired in parallel, giving 12 V and 18 Ah, so 12 V × 18 Ah = 216 Wh
• Usable fraction: 80 percent, so 216 × 0.80 = about 173 Wh
• Inverter efficiency: 85 percent, so 173 × 0.85 = about 147 Wh delivered
• Runtime: 147 Wh ÷ 150 W = about 0.98 hours, or roughly 59 minutes

That looks generous, but real runtime is usually shorter. Batteries lose capacity as they age, and capacity is lower at higher loads and in cold rooms. Efficiency also varies with load. A conservative reading of this example is perhaps half the figure, or around 30 minutes. Check the watt rating too: at 150 W, a UPS rated for 600 W is at 25 percent load, which is comfortable.

How much runtime you need

Decide this from your shutdown time, not from the biggest number you can buy. Add up how long your systems take to shut down cleanly, then add a safety margin. If a hypervisor and its virtual machines take five minutes to stop, planning for 10 to 15 minutes of battery is a reasonable starting point.

Use the UPS’s monitoring connection, typically USB or network, with software that triggers an automatic shutdown when the battery reaches a set level or time remaining. Without that, nobody may be home to shut things down. Once it is configured, test it with a planned test. Do this at a quiet time, with backups current, by unplugging the UPS from the wall and watching what happens. If the UPS offers a built-in self-test or runtime calibration, check its documentation first.

Common mistakes

• Sizing from VA and ignoring the watt rating.
• Using power supply nameplate wattage instead of measured load.
• Trusting the calculated runtime as exact. Treat it as an estimate and test it.
• Forgetting the switch, router, and modem, which may need to stay up if shutdown commands or remote storage travel over the network.
• Plugging a laser printer or other high-draw device into battery outlets.
• Skipping automatic shutdown setup.
• Ignoring battery age. Batteries wear out over the years, and replacement intervals vary by model, so check the manufacturer’s guidance.

Runtime figures here are estimates. Confirm ratings and limits against the manufacturer’s documentation, and consult a qualified electrician if you are unsure about circuit capacity.