How to Run a UPS Battery Runtime Test
The runtime printed on a UPS spec sheet was measured on a new battery at a specific load, and neither of those describes your panel today. A runtime test answers the question that matters: with the real load, right now, how long does this UPS actually hold? Run after a transfer test has proven the mechanics, it is the measurement that decides whether the backup still covers the need.
UPS Battery Runtime Test in one line: To run a UPS battery runtime test, define the required holdover time and a safety stop margin first, then put the UPS on battery with its real load during a planned window and time the discharge while logging battery voltage or the runtime estimate. Stop at your planned margin, restore power, and compare measured runtime against the requirement with room to spare.
What You Need
You need a maintenance window, a recently passed transfer test so the mechanics are not in question, the site's required holdover figure from the design documents, a way to log time and battery voltage or the UPS runtime readout, and an agreed stop point. The required figure is whatever the site's design says must survive an outage: the ride-through until the generator takes over, or the time to execute a clean shutdown, the arithmetic laid out in battery runtime calculation.
Define the Pass Line Before You Start
Write down two numbers before touching anything: the runtime the site requires, and the margin at which you will stop the test. Stopping short of full depletion is deliberate: running a UPS to hard cutoff with the real load connected converts a test into an outage, and deep depletion is wear the battery does not need. If the measured time to your stop point already exceeds the requirement with margin, the test has passed without ever endangering the load.
If your policy requires a full-depletion measurement, do it with the protected load transferred elsewhere or during a window where a hard stop is acceptable, and treat it as a capacity test on the pattern of a rate-aware discharge rather than a routine check.
Put the UPS on Battery With the Real Load
Open the upstream feed as in a controlled transfer test and start the clock. Log the battery voltage or the front-panel runtime estimate at regular intervals rather than only at the end, because the shape of the decline is diagnostic: a healthy battery falls gradually and predictably, while an aged one holds briefly and then drops away fast. Keep the load realistic; testing at a lighter-than-normal load flatters the result exactly when honesty matters.
Watch the estimate against the wall clock. Front-panel runtime estimates recalculate as the discharge proceeds, and an estimate that shrinks much faster than real time passes is itself evidence of a battery whose capacity no longer matches the model inside the UPS.
Stop at Your Planned Margin, Then Restore
End the test at the stop point you defined, restore the feed, and confirm the UPS retransfers and begins charging. Note the total time on battery and the voltage at the stop. The battery is now partially depleted, which means the site is temporarily under-protected: let the recharge complete before considering the system back to normal readiness, and record how long that recharge took, because it defines your exposure window after any real outage.
Verifying the Result
Compare measured runtime to the requirement: the margin between them is the number this test exists to produce. Then trend it: the same test, same load, run periodically, draws the fade curve of the battery, and the point where the curve approaches the requirement is your replacement date chosen on evidence rather than on failure. Feed the on-battery event and its measurements into the monitoring record, alongside the continuous telemetry described in UPS battery monitoring, so the trend lives somewhere the next engineer will find it.
Common Mistakes
Trusting the front-panel estimate without ever timing a real discharge. Testing with a lighter load than the site actually carries. Running to hard cutoff with critical loads connected to make the number rounder. Skipping the post-test recharge window and leaving the site thinly protected. And measuring once, filing it, and never repeating: a single runtime number is a fact, but the trend across tests is the tool.
Frequently Asked Questions
Why is my measured runtime shorter than the spec sheet says?
The spec figure was measured on a new, fully charged battery at a defined load and temperature. Your battery has aged, your load is probably different from the test load, and your ambient temperature shapes the result too. A gap between spec and measurement is expected; what matters is whether the measured time still exceeds your site's requirement with margin, and which way the trend is moving.
How often should a runtime test be run?
Per site policy, with the interval tightening as the battery ages or after deep-discharge events. Between tests, the UPS self-test and continuous battery telemetry carry the watch. The point of the schedule is the trend: consistent tests at consistent load are what make battery fade visible while there is still time to order a replacement calmly.
Automation services
Need help turning this into a working system?
Merobix integrates SCADA, programs Allen-Bradley and Siemens PLCs, and designs and fabricates industrial control panels.
Meeting requests are reviewed before confirmation.