Monday, 7:42 AM
I heard it before I saw it. That high-pitched, repeating beep coming from the server closet—the one that makes everyone on the floor look up from their coffee.
The LCD on our rackmount UPS was flashing OVERLOAD. And there was nothing plugged into it except a single 24-port network switch.
I've managed facilities and office purchasing for a 200-person company since 2021—roughly $85,000 a year across a dozen vendors, reporting to both operations and finance. I've dealt with plenty of equipment failures. But an overload alarm with an almost-empty load? That was a new one.
"You're sure nothing else is on it?" I asked our IT contractor.
He traced the cables twice. "Nothing. Just the switch."
"It's probably nothing"
Here's where I made my first mistake.
I'd read enough forum threads to know that a false overload alarm happens. Bad sensor. Firmware hiccup. I told myself it was a glitch.
I unplugged the unit, plugged it back in, and the alarm stopped.
"See?" I said. "Reset fixed it."
For four days, it stayed quiet. And I let myself believe it.
I still kick myself for that. If I'd called the vendor that Monday—just made the call, spent the 15 minutes—we'd have caught what was actually happening. Instead, I filed it under "weird but harmless" and moved on.
On Friday, during a quarterly review meeting, the network switch died. So did the conference room display, the print server, and the badge reader at the front desk. Everything on that circuit. No warning, no graceful shutdown.
The meeting stopped. Our VP of operations—who'd flown in for this—spent 40 minutes watching a blank screen while our IT guy scrambled. One employee had been saving a budget file to a network drive that locked up. We lost it.
That's when I learned what a "false" alarm actually is. It's not false. It's early.
What the tech actually told us
The technician who came out was patient with me, which I appreciated. He explained that when a UPS throws an overload warning with nothing plugged in, it usually means one of three things:
- A battery nearing failure. As internal impedance rises, the unit misreads its own capacity as a load.
- A faulty output stage. Something in the inverter chain isn't switching cleanly.
- Heat. A fan that's starting to fail, or a closet running 10 degrees warmer than it did last year.
He put a meter on our unit. Battery voltage was fine. The output waveform, though, was dirty—visible distortion on the scope. That's not the kind of thing I could have diagnosed by unplugging and replugging.
"You can't just replace the fuel filter and hope the check engine light goes away," he said, half-joking. "And sure, a dirty air filter can cause that light—but it's not always the filter. Sometimes you've got a cracked spark plug ceramic, or a sensor issue, and you're throwing money at the wrong part."
That stuck with me. The idea that you can throw parts at a problem without fixing it—I'd done that with office printers for years, and it never really worked either.
The lithium-ion question
By then, the old unit was six years old. Battery replacement alone was going to run around $180, and the inverter issue pushed the repair well past the cost of a new unit. So we started shopping—and that's where the second lesson came in.
Our IT contractor recommended looking at a cyberpower lithium ion ups as a replacement. Lithium-ion has obvious appeal for a rackmount setup: no annual battery swaps, longer runtime per pound, and—according to the U.S. Department of Energy's Federal Energy Management Program—a typical service life of 8-10 years versus 3-5 for sealed lead-acid.
According to UL 1778 (the U.S. safety standard for UPS equipment), these units are tested for overload tolerance and electrical safety—but the standard doesn't prescribe how long a battery should last. That's on the operator.
But the quotes came back all over the place. One vendor quoted $520 for a 1500VA lithium-ion rackmount. Another quoted $410. A third quoted $375. (Based on quotes I collected in January 2025; verify current pricing.)
I almost went with the cheapest one. (I know. I know.)
What saved me was a habit I'd built after an earlier mess: I started asking "what's not in this quote?" before I asked "what's the price?"
The $375 quote looked clean until I asked what happened at installation. Turned out it didn't include rack rails ($65), the monitoring card ($120), or disposal of the old unit ($75). The $520 quote included all three. The $410 quote included rails but not disposal.
From the outside, the lowest quote looks like the efficient choice. The reality is you're often buying a different scope with a smaller number attached.
The vendor with the highest sticker price was actually the least expensive once the line items were laid out. That's not a coincidence—it's a pattern I've seen repeat across four or five categories now. The company that tells you the full number upfront usually gives you the real number.
What I do differently now
We went with the $520 quote. The unit's been running since late January, and the monitoring card emails me if battery health drops below 80%. (Which, honestly, I check less often than I should.)
The new UPS also taught me to think about failure differently. A UPS isn't a magic box that prevents outages—it's a bridge. It buys you time to shut down cleanly. That's all. Nobody can promise more than that, and the vendors who try to are the ones I trust least now.
Three things I now do before any UPS purchase:
- Ask for a scope sheet, not a price. Every item—rails, cards, install, disposal—should have a line item. If a vendor can't produce one, that's a red flag.
- Test the alarm before trusting the silence. When a UPS throws any warning, I log it, call the vendor, and don't call it "fixed" until someone with a meter has looked at it.
- Budget for the end, not just the start. Disposal, replacement batteries, monitoring add-ons—I put all of it in the initial request now. Finance has stopped calling me about overages, which is its own small victory.
There's something satisfying about a clean rack setup. After the Friday disaster and the quote runaround, seeing the new unit hum quietly with its monitoring card reporting "healthy"—that felt like a real fix. Not a reset. A fix.
The overload alarm with nothing plugged in? Looking back, it was the best thing that could have happened. It just didn't feel that way at 7:42 on a Monday.
Ingrid Bauer
Ingrid Bauer is an industrial automation and electromechanical systems analyst specializing in variable-frequency drives, motors, pumps, and complete power-drive systems. She applies IEC 61800-9-2 drive-efficiency methods, IEC 60034-30-1 motor efficiency classes, and ISO 9906 pump tests while examining torque-speed duty, harmonics, head, flow, and operating-point efficiency. She helps engineers size drive and pump combinations for real load profiles, cable lengths, ambient conditions, process control, energy use, and maintenance access.