I manage the purchasing for a mid-sized company. Not IT infrastructure, not facilities—just the person who places the orders. So when my team asked me to spec out UPS units for three new office floors, I thought I had it figured out. Pick the model with the highest VA rating that fits the budget, right?
Wrong. So, so wrong.
The Surface Problem: My UPS Keeps Beeping
A year ago, I ordered ten APC Back-UPS units for our workstations. The specs looked good—sufficient VA, decent surge protection. We installed them, and within a week, the beeping started. Not all of them, but enough that people were complaining.
I checked the logs. 'On battery.' 'Low battery.' This wasn't a power outage problem. This was a capacity problem.
The Deeper Issue: Sizing Isn't What You Think
Here's what I didn't understand. The VA rating on a UPS is not the continuous load capacity. It's a maximum. A peak. If you run it at 80% or above consistently, the battery runtime drops faster than you'd expect—and the unit runs hotter, which shortens its lifespan.
When I dug into our actual load, each workstation had a desktop, a monitor, sometimes a second monitor, maybe a dock. The combined draw was higher than the UPS could sustain for more than a few minutes.
But the real kicker? The model I chose was designed for occasional power fluctuations, not frequent brownouts. Our building has older wiring. We get minor dips in power multiple times a day. The UPS was working exactly as designed—switching to battery for protection—but the battery wasn't sized for that frequency of use.
What Nobody Told Me About Runtime Curves
I should have used a UPS calculator—not just the one from APC, but any reliable sizing tool. The APC UPS calculator (https://www.apc.com/us/en/tools/ups_selector/) would have shown me that at 60% load, I'd get maybe 8 minutes of runtime. At 80%, it dropped to under 4. And the beeping starts well before the battery is actually empty.
Take this with a grain of salt, because every model is different: but running a UPS above 70% continuous load is asking for trouble.
The Cost of Getting It Wrong
Let me quantify the damage.
- Direct cost: $3,200 for the ten units (plus installation).
- Replacement cost: $5,800 for the upgraded models we actually needed.
- Downtime: Each time a unit switched to battery and then shut down, the user lost work. We estimated 12 person-hours of lost productivity per month.
- Goodwill: Hard to measure, but my reputation with the team took a hit. I looked like I'd cheaped out.
But here's the part that stung most: I could have gotten it right from the start if I'd understood the actual problem—not just the listed specs.
The Deeper Problem: Surge Protection Isn't Enough
When I started researching our options, another issue surfaced. Some of my colleagues were using standard power strips with surge protection—like a 220 surge protector—instead of proper UPS units. They thought a surge protector and a UPS did the same job.
They don't.
A surge protector handles spikes (voltage above normal). A UPS handles dips, outages, and noise—and provides battery backup. If you're in an area with unstable power, a surge protector alone is not enough. But if you're in a building with perfect power, a full UPS might be overkill.
We're the former. Our building is old. I needed both: surge protection for the power lines, and battery backup for the brownouts.
When the Solution Isn't Right for You
I want to be honest about something. I recommend UPS units for most offices—especially if you're running servers, critical workstations, or network equipment. But if your power is rock-solid (you're in a new building, no history of outages), you might be fine with a good surge protector.
That's not me trying to sell you a smaller solution. It's me admitting that I almost ordered a massive 3000VA unit for our server room—which would have been overkill. Our actual load was 1200VA. The 3000VA unit would have cost three times as much and taken up twice the space.
Better to match the solution to the problem.
I'm not 100% sure of the exact numbers, but I think the 1500VA model we ended up with was the sweet spot. It handles the load, provides enough runtime for a clean shutdown, and doesn't beep constantly. If you're doing the math, the online UPS calculators are your friend.
What I'd Do Differently Today
If I were starting this process over, here's my checklist:
- Measure actual load. Don't guess. Plug in the devices, measure the draw. The UPS calculator will ask for this anyway.
- Consider the environment. What's the power quality like? If you're in a building with dirty power, get a unit designed for frequent switching.
- Plan for runtime. How much time do you need? A few minutes for a graceful shutdown, or 20 minutes to ride out a brownout?
- Factor in surge protection. Even with a UPS, a dedicated surge protector on the input line adds an extra layer. I've started using a 220 surge protector for the office power drops, and the UPS handles the rest.
- Check the warranty and support. APC's warranty is decent, but I've had mixed experiences with battery replacements. YMMV.
So glad I caught the mistake before we rolled out the big server room UPS. Almost went with the 3000VA unit, which would have been a $2,000 mistake. Sizing down—and using the UPS calculator—saved us money and space.
Did we still spend too much? Maybe. Could I have negotiated a better price? Possibly. But at least the equipment works, the beeping stopped, and my team stopped giving me the side-eye.
Dodged a bullet. Barely.