--- title: "Watts to VA Calculator" description: "Convert watts to VA for UPS sizing. Why a 1000 VA UPS will not run a 1000 W load, how to read a nameplate, and how runtime changes the answer." standard: "VA = W / PF" source: "https://energycalchq.com/tools/watts-to-va" --- This is the conversion that decides what size UPS you buy — and the one that gets it wrong most often. Your equipment is rated in watts. The UPS is sold in VA. They are not the same number, and the difference is not small. ## The formula ``` VA = W / PF ``` Where `W` is real power in watts, `PF` is the power factor, and `VA` is apparent power in volt-amps. The power factor here is the **UPS's own**, printed on its nameplate — not your load's. This is the detail that catches people out, and it is worth being precise about why. ## The nameplate you are actually buying Every UPS states two numbers: ``` 1000 VA / 600 W ``` Both describe the same machine. The **VA** figure is what the inverter's switching devices and transformer can carry — a current limit. The **watts** figure is what the unit can actually deliver as real power — a thermal and inverter limit. Whichever you hit first is your ceiling. Connect a 900 W load to that UPS and it overloads instantly, even though 900 is comfortably under 1000, because you have exceeded the 600 W limit. The ratio between the two numbers is the UPS's power factor: ``` 600 W / 1000 VA = 0.6 ``` | UPS generation | Typical PF | 1000 VA delivers | | --- | --- | --- | | Older line-interactive | 0.5 – 0.6 | 500 – 600 W | | Common consumer UPS | 0.6 | 600 W | | Better line-interactive | 0.7 | 700 W | | Modern online / double conversion | 0.8 – 0.9 | 800 – 900 W | | High-end online, unity | 1.0 | 1000 W | So "1000 VA" tells you almost nothing on its own. Two units with the same VA rating can differ by 400 W in what they will actually run. ## A worked example A small server rack drawing 600 W: ``` VA = 600 / 0.6 = 1000 VA ``` On a 0.6 PF UPS, you need a 1000 VA unit — and that leaves nothing spare. At 0.9 PF: ``` VA = 600 / 0.9 = 667 VA ``` A 700 VA modern online unit would carry the same load. Same watts, very different purchase. ## Add headroom — 25 % is not generosity Sizing a UPS exactly to its load is a mistake for three reasons: **Heat.** A UPS at 100 % load runs hot, and heat is the single largest factor in how long its batteries last. A battery bank rated for five years at 25 °C can manage half that at 35 °C. **Runtime collapses at full load.** Battery capacity is not linear with discharge rate. Running at 100 % load does not give you half the runtime of 50 % load — it gives you rather less than half, because of the Peukert effect. A unit quoting 10 minutes at full load will often give 30 at half. **Whatever gets plugged in next.** It always does. So take the calculated VA and multiply by 1.25. Our 600 W load at 0.6 PF: 1000 VA × 1.25 = 1250 VA, so buy a 1.5 kVA unit. The calculator above shows this figure alongside the raw conversion. ## Crest factor, and the load the arithmetic misses Computer power supplies, LED drivers and anything else with a rectifier at the front do not draw a smooth sine wave. They draw current in short pulses at the peak of the voltage waveform. The ratio of that peak current to its RMS value is the **crest factor**. A sine wave has a crest factor of 1.41; a switch-mode supply without power factor correction can reach 2.5 or 3. It matters because the UPS inverter must supply the *peak*, not the average. A UPS rated for a 3:1 crest factor handles this; a cheaper unit rated for 1.4:1 may go into overload on a load that looks fine on paper. If you are supplying a rack of older equipment, check the crest factor specification, not just the VA. Modern equipment with active power factor correction draws a near-sinusoidal current at close to unity power factor, which sidesteps the problem entirely — and is why a unity-PF UPS pairs so well with modern IT loads. ## Working out your load in watts Do not use the figure printed on the equipment's power supply. That is the supply's **maximum output** rating, not what the machine draws — a 750 W PSU in a desktop PC typically draws 150 to 250 W in normal use. Better options, in order of reliability: 1. **Measure it.** A plug-in energy meter costs very little and removes all guesswork. 2. **Read the equipment's specification** for typical consumption, not the PSU label. 3. **Estimate from the class of equipment** — a desktop and monitor around 200 W, a 1U server 300 – 500 W, a network switch 30 – 100 W. Then add them up, and add the 25 % headroom to the total. ## Runtime is a separate question VA sizing tells you whether the UPS can carry the load. It says nothing about how long for — that is the battery, and it is sized independently: ``` Battery Ah ≈ (W × runtime hours) / (V_battery × η_inverter × DoD) ``` A 600 W load for 30 minutes at 85 % inverter efficiency on a 24 V bank at 50 % depth of discharge needs about 29 Ah. The [battery bank and backup time calculator](/tools/battery-backup) does this properly, including the C10 rating derating that makes most first estimates optimistic. ## Common mistakes **Buying VA equal to watts.** A 1000 VA UPS for a 1000 W load overloads on day one. This is the single most common UPS purchasing error. **Using the load's power factor instead of the UPS's.** Your server's PF describes your server. The UPS's rating is set by its own design. **Sizing from the PSU label.** A 750 W power supply is not a 750 W load. **Forgetting the monitor, the switch and the modem.** They are small individually and meaningful together — and a UPS that keeps the server alive while the network switch dies has achieved nothing. **Ignoring runtime until after purchase.** A correctly sized UPS with four minutes of battery will not survive the outage you bought it for.