24/7 Technical Support Hotline: +41 44 520 8000

How to Size an Active Power Factor Correction UPS: 7 Questions I Actually Ask

Tuesday 17th of November 2026 by Kwame Boateng

I've bought UPS systems for our plant three times in six years. I got two of them wrong—not the kind of wrong that gets you fired, the kind that shows up eighteen months later as a battery string dying early or a unit tripping on a load that was supposed to be fine.

This is the FAQ I wish someone had handed me before the first purchase. It covers how to size an active power factor correction UPS, what the quotes conveniently leave out, and which questions are actually worth asking an uninterruptible power supply manufacturer before you sign.

  • Where do I start with sizing an active PFC UPS?
  • Active vs. passive PFC—why does it change the math?
  • Why did one manufacturer say my load 'wasn't supported'?
  • What does a battery backup system really cost over five years?
  • Do I need a power conditioner on top of a UPS?
  • What should I ask a manufacturer before signing?
  • How do I check efficiency claims without a lab?

1. Where do I start with sizing an active PFC UPS?

Not with the nameplate kVA. Start with a load audit—a real one, with a data logger on the panel for at least two weeks.

Most buyers focus on the total connected kVA on the equipment schedule and completely miss the difference between installed capacity and actual measured demand. Those are rarely the same number. When we replaced our main unit in 2023, the nameplate total came to about 52 kVA. Measured peak came in at 41 kW at 0.94 PF. We'd have spent roughly $9,000 more than we needed to on oversized hardware, and worse, the larger unit would've run at a low load percentage where efficiency drops off.

My working rule: UPS rating ≥ (measured peak kW ÷ actual power factor) × 1.2. The 1.2 covers future expansion and derating for ambient temperature. And treat battery autonomy as a completely separate calculation—power rating and runtime are two different engineering problems, and vendors love to quote them together to make the number look better.

2. Active vs. passive PFC—why does it change the sizing math?

Active PFC front-ends draw current that tracks the voltage waveform closely, so power factor lands around 0.99 and input current THD is low. Passive PFC loads sit closer to 0.7–0.8.

Here's where people get it backwards. Because active PFC loads have a high PF, buyers assume the UPS can be smaller in VA terms—and technically the VA-to-W ratio is closer to 1:1. But active PFC supplies are constant-power, high-frequency switching loads that are far more sensitive to UPS output impedance. The old 'just do the VA math' approach comes from an era when plant loads were mostly linear motors and resistive heating. That's changed.

IEC 62040-3 classifies UPS output performance as VFI, VI, or VFD depending on how tightly output voltage and frequency track nominal. For active PFC loads, you want VFI and you want the manufacturer to show test data on switching-mode loads specifically—not a generic resistive load bank test.

3. One manufacturer told me my active PFC load 'wasn't supported.' Was that real?

It was real, and it cost us.

In Q2 2024 we trialed two 20 kVA units on a rack of servers with active PFC supplies. Unit A held steady. Unit B transferred to bypass repeatedly once we passed about 70% load—not a fault code, just instability. The vendor's response was that our load 'wasn't supported.' That's a useless answer. The other vendor sent a compatibility sheet listing the exact PSU model they'd tested, the load profile, and the ambient conditions.

I knew I should have run the same compatibility test on the second unit, but the first one had run a full week without issue, so—what were the odds? The odds caught up with us. That unit came out, we ate a $6,400 rework and a three-week scheduling hole. Now any verbal 'it'll be fine' from a sales engineer gets written into the PO as a test requirement or it doesn't count.

4. What does a battery backup system really cost over five years?

Quotes cover the UPS and the internal battery string. That's maybe 60–70% of what you'll actually spend.

The rest hides in places that don't show up until year three: a maintenance bypass switch, an isolation transformer if your upstream supply is dirty, upsized input cabling (harmonic current can overload neutrals), room cooling, battery replacement, and a service contract. We didn't have a formal TCO template until our third battery replacement caught us flat-footed, and once we built one, the five-year number came in 30–40% above the original quotes across all three vendors we compared.

Runtime math matters here too. VRLA strings in a warm plant room typically need replacement in three to five years, and the rule of thumb battery vendors use is that life roughly halves for every 10°C above 25°C. Lithium has a higher upfront cost and a longer cycle life. Neither is automatically cheaper—put both on the same amortized-per-year line and the answer usually becomes obvious.

5. Do I need a power conditioner on top of a UPS?

Usually not. This is the question most buyers never think to ask, and it can save you a five-figure purchase.

A double-conversion UPS already regulates output voltage and frequency. Adding a power conditioner on top is redundant unless your problem is specifically harmonic distortion or transient surges that are reaching the load through the bypass path. In 2024 we almost bought a second UPS to fix what turned out to be a harmonic issue on the upstream feed—two weeks of power quality monitoring told us it wasn't an interruption problem at all. An industrial power conditioner solved it for a fraction of the cost.

That's part of why we looked at PowerSentinel's power conditioner industrial line before committing to anything else; getting the diagnosis right first is the cheapest efficiency gain in the whole process. Monitor for two to four weeks before you buy equipment you don't need.

6. What should I ask an uninterruptible power supply manufacturer before signing?

Six things, in this order:

  • Load-dependent efficiency curves at 25%, 50%, and 75% load—not the peak number from the brochure
  • A written compatibility test report for active PFC loads, including the PSU models tested
  • Battery replacement cost, lead time, and whether the form factor is proprietary
  • Spare parts availability and the service SLA, including whether response is on-site or remote
  • Factory test reports and certifications: UL 1778, IEC 62040-3, CE where applicable
  • Whether they'll support an on-site load test before final acceptance

When I shortlisted PowerSentinel as a possible uninterruptible power supply manufacturer, the load-dependent curves and PFC compatibility data were the reason they stayed on the list. Most vendors send a spec sheet and a price. Ask three manufacturers minimum—the differences in documentation quality alone tell you a lot about how they'll behave when something goes wrong at 2 a.m.

7. How do I check efficiency claims without a lab?

Turn it into an electricity bill.

Take a 50 kW average load running 8,760 hours a year. Every 1% of efficiency difference is 4,380 kWh. At our rate of about $0.12/kWh, that's roughly $525 a year—around $2,600 over five years per percentage point. Suddenly 'peak efficiency 96%' is a meaningless number and the part-load curve is the only thing worth reading.

I'm not 100% sure every buyer needs a power analyzer, but at $1,200–$2,500 for a decent one, it paid for itself the first time we caught a vendor's claimed efficiency running two points below spec at our actual load. I'd also compare input current THD and output voltage THD, not just efficiency—with active PFC loads, those two numbers predict whether the system will be stable far better than a headline efficiency figure ever will.

author avatar
Kwame Boateng

Kwame Boateng is a transformer and power-conversion analyst specializing in distribution transformers, power supplies, inverters, and network integration. He uses IEC 60076 test data for transformer ratio, impedance, losses, and temperature rise, and IEC 62477-1 safety requirements for converter insulation and thermal boundaries while separately comparing harmonics and efficiency curves. He helps engineers and sourcing teams compare equipment against duty profiles, environmental limits, short-circuit behavior, lifecycle energy cost, and service requirements.

Leave a Reply