SizingKitElectronics
UPS runtime calculator
A 1500 VA UPS is not a 1500 W UPS: divide the two figures on the case and you have the unit's own power factor, 0.6 on the familiar 1500 VA / 900 W model, and a load whose power factor is better than that runs out of watts long before it runs out of volt-amperes. This page says which of the two ratings is binding, how full the unit is against it, and how long it holds — 11.2 minutes at 500 W on a 24 V 9 Ah string. It also shows why halving that load buys 1.93 times the runtime rather than twice: the unit spends 20 W on itself no matter what is plugged in.
- 100% free
- No signup
- Both ratings compared
- Unit's own draw included
- 6-point runtime curve
The unit, the load, and the batteries inside
Both ratings from the front of the case, because the pair of them is the whole trick: the ratio between them is the unit's own power factor, and it decides which one your load runs into first.
The number in the model name, and the flattering one.
Those two make a unit power factor of 0.6.
What the equipment actually draws, which for a server is well under its supply's rating.
From the equipment's own label. Power-factor-corrected supplies sit near unity; older switching supplies sit far lower.
Off the label on the battery, not off the model number.
A 24 V string.
A 20-hour figure, on a battery this unit will empty in minutes.
Default 50% — an estimate for a high-rate discharge, not a published figure — the battery's own constant-power discharge table is the authority and this is the weakest number on the page.
Default 90% — the DC-to-AC efficiency of the unit on battery, from its datasheet where the manufacturer publishes one.
Default 20 W — the unit's own consumption on battery, measured with an inline meter and nothing plugged into the outputs — datasheets rarely publish it. It scales with the size of the unit and with what is switched on inside it: a lit display, a network management card and a charging circuit all add to it, and a large rack unit can be several times this.
Runtime and the binding rating
RUNTIME AT THIS LOAD
11.2 min
HOW FULL THE UNIT IS
55.6%
- Load as volt-amperes
- 526 VA
- Against the VA rating
- 35.1%
- Against the watt rating
- 55.6%
The watt rating binds. Your load's power factor of 0.95 is above the unit's own 0.6, so the load reaches the watt rating first and the volt-ampere figure on the front of the case is never the constraint. Buying a larger unit on its VA number would not move this.
Halving the load does not double the time. At half this load the runtime is 21.6 min, which is 1.93 times longer rather than 2 — because the 20 W the unit spends on itself does not get smaller when the load does. The battery bends the curve further still: capacity falls as the discharge rate rises, and that effect is quantified on the amp hour calculator rather than folded in here.
The runtime curve for this unit
Six points against the watt rating, which is the one binding here. This is the shape the chart in the manual has, and comparing the two is the fastest way to find out whether the batteries are still any good.
| Of rating | Load | As VA | Runtime |
|---|---|---|---|
| 10% | 90 W | 94.7 VA | 53 min |
| 25% | 225 W | 237 VA | 23.8 min |
| 50% | 450 W | 474 VA | 12.4 min |
| 75% | 675 W | 711 VA | 8.4 min |
| 100% | 900 W | 947 VA | 6.3 min |
| 110% | 990 W | 1,042 VA | 5.8 min |
The last row is past the rating and is there to show the shape, not as an operating point: a unit above 100% alarms and then transfers or shuts down. The battery figures rest on a sealed lead-acid discharge table drawn to an end voltage — published tables use 1.60, 1.67, 1.70 and 1.75 V per cell — and a UPS runs to the deep end of that range, because in this application five more minutes is worth more than another year of cycle life.
How to work out what a UPS will really hold up
Two numbers off the case, one off the equipment label, three off the battery. The last of those is where the honest uncertainty lives.
Enter both ratings from the front of the unit
The volt-ampere figure is in the model name and the watt figure is usually smaller and printed beneath it. The page divides one by the other to get the unit's power factor, and that ratio — 0.6 on older consumer models, 0.9 on recent ones — is what decides whether your load will be limited by current or by real power.
Give the load its own power factor
A modern power-factor-corrected server or workstation supply sits close to unity, while an older switching supply can be at 0.65, and the difference decides which rating you hit first. Where the label does not state it, the ratio of the equipment's rated VA to its rated watts gives it away in exactly the same manner as for the UPS.
Read the battery labels, then replace the high-rate estimate
Open the front panel or the battery tray and take the block voltage, the number of blocks in series and the amp-hour figure straight off the labels. Then find the manufacturer's constant-power discharge table for that block — it is a grid of watts per cell against minutes — and replace the high-rate percentage on this page with what it actually says, because that field is a placeholder and everything else here is not.
Technical specifications
| Two ratings, one ratio | Volt-amperes and watts entered separately, with the unit's power factor computed from them. 1500 VA against 900 W is 0.6; 1500 VA against 1350 W is 0.9, which is why two units carrying the same number in their model name are not interchangeable. |
|---|---|
| Which rating binds | Decided by comparing the load's power factor with the unit's. Above it the watt rating binds; below it the volt-ampere rating does; and the crossover sits exactly where the two power factors are equal. |
| Runtime model | Battery watt-hours multiplied by the high-rate factor and the inverter efficiency, divided by the load plus the unit's own draw. A 24 V 9 Ah string is 216 Wh of nameplate, 97 Wh delivered, and 11.2 minutes at 500 W. |
| The unit's own draw | Default 20 W, measured with an inline meter and nothing in the outlets, because very few datasheets publish it. Editable, and a large rack unit with a display and a network card runs several times higher. |
| High-rate factor | 50% of the block's amp-hour rating, and openly the weakest figure on this page: it is an estimate rather than a published value, and the battery's own constant-power discharge table should replace it. |
| Battery string arithmetic | String voltage is block voltage times the number in series and the amp-hours are one block's. Two 12 V 9 Ah blocks in series are 24 V and 9 Ah — 216 Wh — not 18 Ah. |
| End-of-discharge convention | Sealed lead-acid discharge tables are published to end voltages of 1.60, 1.67, 1.70 and 1.75 V per cell, and a UPS is run to the deep end of that range, trading cycle life for the five extra minutes it is bought for. |
| What leaves the page | Nothing. Ratings, load figures and the labels off your batteries are read where you typed them and go no further. |
Frequently asked questions
My UPS says 1500 VA. How many watts is that?
Whatever the second number on the case says, and there is no way to derive it from the first. The ratio between them is the unit's own power factor, which manufacturers chose as 0.6 for years — hence 1500 VA and 900 W — and now often set at 0.9, giving 1350 W from the same headline figure. A calculator that converts VA to watts with an assumed 0.6 is guessing at the design of a specific product, and it will be wrong by half on a modern one.
Why does the runtime chart in the manual not match what this page says?
Most likely because the high-rate factor here is a placeholder and the manufacturer measured the real thing. Their chart is drawn from a constant-power discharge test on the exact battery in the exact enclosure, which is better evidence than any general figure; where the two disagree, theirs wins and this page is the tool for asking what-if around it. A large disagreement in the other direction — the unit falling well short of its own chart — is the useful signal, because it usually means the batteries have aged out.
Why does an extended battery pack take so long to recharge?
Because the charger inside a UPS is sized for the battery that came with it, not for the ones you added. Doubling the battery roughly doubles the runtime and doubles the recharge, and since the internal charger was already deliberately small, a unit that refilled in four hours can take most of a day with two external packs on it — which matters because a second outage during that window finds the batteries half full.
What load percentage should a UPS actually be run at?
Somewhere around half of the binding rating, for three separate reasons. Runtime at full load is short enough to be nearly useless for a graceful shutdown, efficiency on most units falls away at both ends of their range, and a unit with no headroom cannot absorb the surge when a device restarts. Sizing to 100% also leaves nowhere to add the next server, which is how a UPS ends up overloaded a year after it was correctly specified.
Does a power-factor-corrected supply change which rating binds?
Yes, and it usually flips it. A corrected supply at 0.99 draws almost the same volt-amperes as it does watts, so on an older 0.6 unit it exhausts the watt rating while less than two-thirds of the VA rating is used — the VA figure that sold the unit is unreachable. The same equipment on a 0.9 unit sits much closer to using both ratings together, which is the practical argument for the newer designs.
Why do UPS batteries only last three to five years?
Because they live on continuous float charge inside a warm box, which is the pair of conditions sealed lead-acid tolerates least well. Rate of grid corrosion roughly doubles for every 10 °C above 25 °C, and the inside of a rack-mounted unit with its own losses in it is rarely at 25 °C; on top of that, the deep end-of-discharge voltage a UPS uses costs cycle life every time it runs. The failure is gradual and invisible from the front panel, which is why a periodic self-test against the manual's runtime chart is worth more than the age on the label.
Can I put a laser printer or a fridge on a UPS?
Not on a unit sized by its steady-state ratings, which is what this page and the case both describe. A laser printer's fuser and a compressor's locked-rotor current draw multiples of their running figure for a fraction of a second, and a UPS that is comfortable at the running load will transfer to bypass or shut down on the surge. Manufacturers list these as loads to keep off a UPS for precisely that reason, and the fix is a larger unit chosen against the surge rather than the average.
About the two ratings, and the shape of the curve behind them
Every UPS carries two capacity figures because it is limited in two different ways at once. Its output devices and its wiring care about current, which makes volt-amperes the natural rating; its inverter and its battery care about real power, which makes watts the natural rating; and the ratio the manufacturer chose between them is stamped into the product for life. Consumer units were designed around 0.6 for a long time, which is where the familiar 1500 VA / 900 W pairing comes from, and the modern move to 0.9 is the same box sold with a third more usable watts. Which rating you actually run into is decided by your load rather than by the unit: the crossover is exactly the point where the load's power factor equals the unit's, above it the watts bind, and below it the volt-amperes do. That single sentence explains most of the confusion around UPS sizing, and it is the reason a VA-to-watts converter with a fixed factor in it should not be trusted.
The runtime curve has a shape of its own, and the shape is not the straight line intuition supplies. A conversion stage consumes a fixed number of watts simply by being switched on — the inverter, the control board, the display, the charging circuit — and that draw does not shrink when the load does, so a light load spends a larger share of the battery on the unit itself. Halve the load on the default figures here and the runtime grows by 1.93 times rather than 2, and the gap widens the smaller the load gets. The battery bends the same curve further at the heavy end, because capacity falls as the discharge rate rises; that effect is quantified on the amp hour calculator and is deliberately not folded in here, so that the two pages cannot disagree with each other silently.
The weakest figure on this page is the fraction of the battery's rating you get at a UPS-like discharge rate, and it is worth saying plainly rather than burying. An amp-hour rating on a sealed block is measured over twenty hours; a UPS empties that block in minutes, and no general percentage describes what that returns. Manufacturers publish the real answer as a constant-power discharge table — watts per cell against discharge time, run down to a stated end voltage — and their table beats this default the same way an equipment plate beats any generic target on the superheat calculator. For the same arithmetic applied to a battery bank you own rather than a product with two ratings on it, use the battery runtime calculator; for the conversion between the amp-hours on the block and the watt-hours this page divides, the watt hour calculator; and for the supply circuit the UPS itself needs on the input side, kW to amps at the unit's own input power factor.
What a load list leaves behind
Every number on this page is worked out by JavaScript running in the tab you are reading it in. Nothing you type — loads, lengths, nameplate ratings, the rates your utility charges you — is uploaded, logged or kept, which is also why the calculators carry on working in a mechanical room with no signal.
The ratings and load figures you enter describe somebody's equipment rack, so they are treated as somebody's business: they are read in the page and never sent anywhere.