UPS runtime is the number of minutes the system can carry a stated load from its battery alone before the inverter shuts down. It is always quoted against a specific load, which is why a unit advertised at forty minutes may deliver eight when fully loaded: the figure is meaningless without the load it refers to. Runtime is set by the usable energy stored in the battery, the power the load actually draws, the efficiency of the inverter, and two variables most datasheets stay quiet about, namely the temperature of the battery and its age.
Take the energy available in the battery, subtract what cannot be withdrawn without damaging it, divide by the load power, and correct for conversion losses. A string of 16 blocks of 12 volts and 9 ampere-hours nominally holds about 1.7 kilowatt-hours. Planning to no lower than eighty percent depth of discharge leaves roughly 1.38 kilowatt-hours usable, and after inverter losses of about six percent, near 1.30 kilowatt-hours reaches the load. Against a 2 kW load that is 39 minutes on paper.
On paper is the operative phrase. Real strings deliver less, and the gap grows as the load rises.
Doubling the load does not halve the runtime; it cuts it by rather more. Lead-acid chemistry loses effective capacity as discharge current increases, an effect described by Peukert's relationship, because the chemical reaction cannot keep pace with the demand for current at the plate surface.
The practical consequence is easy to state. A string that supports a 25 percent load for three hours will typically support a 50 percent load for around 70 minutes rather than 90, and a 100 percent load for perhaps 22 minutes rather than 45. This is exactly why comparing two products on a single headline runtime figure is unsafe, and why the runtime table at several load points is the part of the datasheet worth reading. Lithium iron phosphate is far less affected, holding close to linear behaviour across normal discharge rates, which is one of its practical advantages beyond service life.
Start by reducing the load. Every watt removed from the protected bus buys runtime for nothing. Non-essential monitors, desktop machines, printers and lighting frequently occupy a quarter of the protected load on a bus that grew by accretion, and moving them to unprotected circuits is a wiring change rather than a purchase.
Second, verify the actual load. Units are routinely bought oversized and then loaded to fifteen percent, in which case the runtime is already generous and the real question is whether the next refresh should be a smaller machine.
Third, add an external battery cabinet. Most units above 3 kVA accept extra strings in parallel, and this is the standard route to longer autonomy. Confirm the charger can still recharge the enlarged bank in an acceptable time, as this is the constraint people hit.
Fourth, change chemistry. Swapping to lithium iron phosphate at equal footprint typically yields more usable energy, because it tolerates deeper discharge and holds its capacity better under load and heat.
Fifth, stop chasing runtime and add a generator. Beyond about thirty minutes of autonomy, batteries become an expensive way to buy time, and the sensible design is a battery sized to bridge the generator start with margin rather than to outlast the outage.
Battery capacity is quoted at 25 degrees Celsius. At 0 degrees a lead-acid string delivers roughly eighty percent of that. Cold reduces available runtime temporarily and reversibly. Heat does something worse: it does not reduce today's runtime much, and may slightly increase it, while permanently shortening the life of the string. The widely used rule is that every 10 degrees above 25 halves the design life. A five-year battery held at 35 degrees is a two-and-a-half-year battery, and the owner usually discovers this during an outage rather than during an inspection.
A battery is considered end-of-life when it delivers eighty percent of rated capacity, and it reaches that point long before it fails outright. The decline is slow for most of the service life and then accelerates sharply in the final year.
Two consequences follow. First, size the string for the runtime you need at end of life, not when new, which in practice means adding about twenty-five percent to the calculated requirement. Second, test rather than assume. A quarterly self-test confirms the string can support a load for a few seconds; it says nothing about whether it can hold on for twenty minutes. Only a timed discharge under a controlled load answers that.
Work from the purpose rather than from a round number. If a generator is present, the requirement is generator start plus stabilisation plus a safety margin, which is commonly five to ten minutes. If the goal is an orderly shutdown of servers and storage, measure how long a full shutdown genuinely takes and double it, which usually lands between ten and twenty minutes. If the site must ride through typical local outages without a generator, examine the utility's own interruption statistics: most short interruptions clear inside two minutes, and covering the ninetieth percentile is far cheaper than covering the worst case.
If you are unsure what runtime your present configuration truly delivers, our engineers can review the load, the string age and the room temperature and tell you where you actually stand. Common questions on autonomy, battery care and testing are collected at https://www.upsboss.com/faq/, and runtime tables at several load points for each model are published at https://www.upsboss.com/products/.
Key takeaway: runtime is a function of load, not a fixed property of the box, and it falls faster than the load rises. Size for the string's end-of-life capacity, keep the room cool, and once the requirement climbs past half an hour, buy a generator instead of more batteries.
Why does my unit shut down sooner than the runtime chart promised?
Three reasons cover almost every case: the load is higher than assumed, the chart was read at the wrong load point, or the string has aged past eighty percent capacity. Measuring the actual load on the output and noting the battery installation date will identify which one applies.
Does keeping the unit permanently at full charge harm the battery?
No, provided float voltage is correctly regulated and temperature compensated, which is what a properly designed charger does. What harms a battery is a float voltage set too high for the ambient temperature, and that quietly dries out valve-regulated cells over a period of months.
Can extra battery cabinets be added later, after installation?
Usually yes on units above 3 kVA, but confirm three things first: that the model supports additional strings, that the charger is rated to recharge the larger bank within your recovery window, and that the floor can take the weight. Retrofitting is straightforward when planned and awkward when discovered late.
Contact: Frank Zhang
Phone: +86-135 5688 8641
Email: frank@upsboss.com
Add: Jufeng Road, Guangming Street, Guangming District, Shenzhen City, Guangdong Province, China
