UPS Batteries Fail on a Schedule — Budget for It

The reason battery replacement gets missed is that the failure mode is gradual and the reporting is binary. A valve-regulated lead-acid string does not go from working to broken; it slowly loses the capacity to hold a load, so a unit that once carried a rack for fifteen minutes carries it for eight, then four, then ninety seconds. Nothing about that progression is visible on the front panel, because the unit is still charging, still online and still able to transfer. The number that changed is one nobody was watching.

The industry benchmark for when that fade matters is well established: IEEE 1188, the maintenance recommended practice for valve-regulated lead-acid batteries in stationary applications, treats a string that has fallen below 80% of its rated capacity as due for replacement. That threshold exists precisely because capacity loss accelerates once it starts — a string at 80% is not four-fifths of the way through a long, gentle decline, it is approaching the knee of the curve. Waiting for a fault indicator means waiting for something well past that point.

Design life and service life are also two different numbers, and product records are usually honest about the gap. The replacement cartridges in this catalog publish their own minimum and maximum battery-life figures on the manufacturer's record — commonly two or three years at the low end and five at the high end for standard sealed lead-acid cartridges. That is a range, not a promise, and where in the range a given unit lands is decided by the three things covered below.

Most units run a periodic self-test, and most people treat a passing result as evidence that the battery is fine. It is not that, and it was never designed to be. A self-test is a brief transfer to battery — a matter of seconds — under whatever load happens to be connected at that moment. A string with a third of its original capacity will pass that comfortably, because holding a partial load for a few seconds is a much easier task than holding the full load for the runtime the design assumed.

The number that would actually tell you something is the runtime the unit reports at the load it is carrying, tracked over time. Recorded quarterly against the same load, that figure traces the fade directly, and the trend is far more useful than any single reading. A unit whose reported runtime has halved over two years is telling you exactly what it is going to do in the next outage, and it is telling you now rather than then.

The other reliable signal is simply the install date. Batteries are the one part of an estate where the calendar is a legitimate primary trigger, because the wear is chemical and time-based rather than usage-based. A label on the unit with the cartridge install date, readable without opening anything, is the cheapest asset-management improvement available in this category — and on a fleet, a spreadsheet of those dates is the whole replacement plan.

The financial argument for planning battery replacement is not that the cartridges are cheaper when you plan — they cost the same. It is that an unplanned replacement is never only the cartridge. It is an emergency purchase order, a same-day shipping charge, an out-of-hours visit if the unit is not hot-swappable, and, on a bad day, the cost of whatever the load was doing when the power went. Planned, the same work is a scheduled visit in a maintenance window with the parts already on site.

The mechanics of building the line item are straightforward once the install dates exist. Group the estate by install year, assume the published minimum life as the planning figure for units in warm or heavily-cycled locations and the middle of the range elsewhere, and spread the resulting replacements across the years rather than letting a whole procurement generation come due in one budget cycle. An estate bought all at once will otherwise present a single enormous bill three or four years later, which is exactly the shape of expenditure that gets deferred — and deferring it is how a planned cost becomes an outage.

It is worth deciding the disposal path in the same conversation, because it is part of the cost and it has a lead time. Sealed lead-acid cartridges are recyclable and are handled through established channels; lithium-based packs carry their own transport and handling requirements. Neither is difficult, but both are easier to arrange as part of a scheduled refresh than as an afterthought with a pallet of old cartridges already in a corridor.

The last piece is the units themselves. A battery refresh is a natural decision point for whether the UPS is still the right one — whether the load has grown past the runtime the design assumed, whether the topology still matches what the equipment needs, and whether a unit that has been replaced twice is worth replacing a third time. Uniqcli quotes cartridges, replacement units, the installation and the disposal on one document rather than leaving the estate to assemble it from three suppliers.