Why UPS batteries fail early in Pune
A VRLA battery rated for five years will often manage two and a half here. The cause is almost never the battery — it is the room it sits in.
Batteries are the single most common cause of UPS failure by a wide margin. They are also the component customers feel most misled about, because a bank quoted with a five-year design life fails in under three and the obvious conclusion is that the batteries were poor quality. Usually they were not.
Design life is quoted at 25 °C
Every VRLA design life figure you have ever been quoted assumes an ambient of 25 °C. The rule of thumb across the industry is that life halves for roughly every 8 to 10 °C above that. A bank rated for five years at 25 °C is a two-and-a-half-year bank at 35 °C.
Now consider where these actually sit in Pune. An unventilated server cupboard through April and May. A terrace machine room above a lift, which regularly passes 40 °C in the afternoon. A panel room with the UPS itself rejecting a kilowatt of heat into the same space. None of these are 25 °C environments, and none of them were when the life figure was quoted.
The cheapest thing you can do
Ventilate the battery room, or move the bank somewhere cooler. It costs less than one replacement cycle and it is the only intervention that reliably changes the outcome.
Cycling is the other half of it
A battery wears by being discharged and recharged. A bank on a stable supply that goes to battery twice a month will outlast an identical bank on a feeder that dips several times a day, even at the same temperature.
This is worth understanding because it is fixable. If the UPS is transferring to battery frequently, the cause is usually that the incoming voltage keeps leaving the machine’s input window — not that anything is faulty. An online UPS with a wide input range, or a stabilizer upstream of a line-interactive one, cuts the cycle count dramatically and the bank simply lasts longer.
Why the display says the bank is healthy right up until it is not
This is the part that catches people out. A VRLA cell that has lost forty per cent of its capacity still reads a perfectly normal 13.5 V on float. Voltage tells you the state of charge, not the capacity. The bank looks fine on the panel and on any spot check, and then the mains fails and it collapses in ninety seconds.
Two measurements tell the truth. Internal impedance, taken cell by cell, rises as a cell degrades and identifies the weak one in a string — quick, non-invasive, and it belongs in every quarterly visit. And a timed discharge test at rated load, which reports delivered capacity as a percentage of rated. That second number is the only one that answers "how long will it actually hold".
Replace the string, not the cell
In a series string the weakest cell limits the whole bank. A new cell installed alongside aged ones is charged and discharged to suit the old ones and degrades to match them within months. Partial replacement feels economical and buys almost nothing. The only real exception is a bank under six months old with a single manufacturing failure.
Industry practice is to replace when measured capacity falls below 80% of rated, because degradation accelerates sharply past that point. That is a measurement, not a calendar date, which is why annual testing costs less than either premature replacement or a failure.
Is lithium the answer?
Increasingly, above roughly 20 to 40 kVA. LiFePO₄ costs two to three times more up front and lasts three to four times longer, takes about a third of the floor space, tolerates higher ambient temperatures — which matters a great deal here — and gives real state-of-charge telemetry rather than a voltage you have to interpret.
Over a ten-year horizon, counting the two VRLA replacements you would otherwise buy, it usually comes out ahead. Below that band the up-front premium often does not pay back within the equipment’s life. It is worth running the ten-year cost both ways for your specific bank rather than taking either answer on faith.
Questions we get asked
Anything not covered here, call us — most of it is faster to answer on the phone.
How often should UPS batteries be tested?
Impedance readings cell by cell at every quarterly visit, and a timed discharge test at rated load once a year. The quarterly readings show the trend and flag a weak cell early; the annual discharge gives you the delivered-capacity figure that tells you how much runtime you genuinely have.
My UPS says the battery is 100%. Why did it only last two minutes?
Because that percentage is derived from voltage, which reflects state of charge rather than capacity. A degraded cell charges to a normal voltage and holds almost nothing. Only a discharge test distinguishes the two, which is precisely why a float voltage check is not a battery test.
Does keeping the room cool really make that much difference?
It is the single largest factor. The industry rule of thumb is that life halves for every 8 to 10 °C above 25 °C, so a bank at 35 °C gives you roughly half the years of the same bank at 25 °C. Ventilation is cheaper than one replacement cycle.