Industrial UPS Systems: Honest Answers to What Buyers Actually Ask

100kVA three-phase online UPS — what's it actually for?

A 100kVA three-phase online UPS is the awkward middle child of power protection. It's too big for a small workshop, not big enough for a full plant. The sweet spot is medium manufacturing floors, small regional data centers, hospital imaging suites, and distribution rooms where the load is balanced across all three phases.

Last quarter alone, we shipped 7 units in the 100kVA class. The pattern I keep seeing: clients size the UPS against today's load, then forget the transformer inrush current when the plant adds a production line or a third shift. A unit running at 60% load suddenly sees 85% during startup, and the batteries age twice as fast.

My rule of thumb: if your continuous load sits between 60 and 75kVA, and your worst-case startup current stays under 90kVA, 100kVA is right. Push past that and you're better off with a 125kVA or a modular frame. Nobody likes hearing it, but the cost of replacing batteries every 18 months usually dwarfs the cost of upsizing from the start.

Double conversion UPS and IEC 62040 — what's the connection?

Double conversion means the UPS converts AC to DC (charging the battery in the process), then converts DC back to clean AC. The load never touches raw mains power. There's no transfer switch or delta conversion in the normal path.

According to IEC 62040-3:2011, UPS systems fall into three performance classes: VFD (voltage and frequency dependent), VI (voltage and frequency independent), and VFI (voltage and frequency independent). A true double conversion UPS is classified as VFI — the highest tier, where output stays stable regardless of what the input does.

Here's where it gets practical. When you're comparing quotes, look for the VFI label on the datasheet, not the marketing copy. I've seen vendors label a line-interactive unit as "online UPS" because the brochure says so, but the IEC classification tells the real story. More than once, a client has sent me a spec sheet asking "is this double conversion?" and the answer was no — the fine print said VI.

When should you replace an APFC UPS instead of repairing it?

Active power factor correction (APFC) keeps the input current waveform clean and the power factor near 1.0 while reducing harmonics. The failure mode is rarely dramatic — it's a slow creep. Power factor drops from 0.99 to 0.94, harmonics climb, and the input current gets choppier. Eventually you're dealing with nuisance trips or an upstream breaker that decides to open for no obvious reason.

My cutoff rule: if the APFC module replacement costs less than 30% of a new UPS and the rest of the unit (inverter, battery string, control board) is under 8 years old, repair it. If the module is over 40% of replacement cost, or this is the second APFC failure in two years, replace the whole unit.

I still kick myself for a call I made in 2023 — recommended a third APFC repair on a 10-year-old unit instead of pushing for replacement. Two months later, the inverter failed. The client paid for both repairs, then bought a new UPS anyway. That one stings.

Industrial UPS versus power transformer — same thing?

I get this question a lot because PowerSentinel builds both, and the confusion is understandable.

A transformer shifts voltage up or down. That's it. It doesn't store energy, it doesn't ride through an outage, and it doesn't condition power quality. When the grid drops, a transformer does nothing.

A UPS system stores energy in batteries and takes over when mains fails. It filters voltage sags and swells, holds the load up for minutes to hours depending on battery capacity, and on double conversion units, isolates the load from input disturbances entirely.

Where they overlap: many industrial UPS systems include an isolation transformer at the output stage for voltage matching and galvanic isolation. So the question isn't "transformer or UPS." It's "what voltage does the load actually need, and how long do you need to ride through an outage."

When the UPS has to be on-site in 48 to 72 hours, what's your process?

This is what I do every day. Here's the sequence.

First, confirm the actual load. Don't place an order until someone measures the real kVA and startup current. In March 2024, a client called at 4 PM on a Thursday needing a 100kVA unit by Monday. The plant had measured 65kVA continuous load. We found a vendor with stock, paid $3,200 extra in rush freight and expedite fees on top of the $14,000 base cost, and delivered Sunday night. The client's alternative was shutting down production for three days — roughly $28,000 in lost output.

Second, get the delivery commitment in writing. In a 72-hour window, a verbal promise is worth nothing. If the vendor won't put a date and time in an email, they aren't committed.

Third, check the installation site before the unit shows up. I've seen a 100kVA unit arrive Friday afternoon with no clearance for airflow and no forklift path into the electrical room. That conversation is never fun.

What's the most common mistake procurement teams make when ordering industrial UPS?

Skipping the site survey. I get it — it adds a week to the timeline, and everyone's in a hurry. But I've seen more problems from missed survey details than from any other single step in the process.

The usual misses: transformer inrush current from downstream loads not accounted for, ambient temperature in the electrical room not measured (batteries age noticeably faster above 25°C, and plenty of electrical rooms run hotter than anyone admits), input breaker capacity not verified, and existing cable runs not checked against the UPS input current.

One client ordered a 100kVA unit, then discovered the existing input breaker was 150A — barely enough for the UPS's own input current, let alone the startup surge. The rework cost $1,800 and added five days. Better than a blown breaker, but that week wasn't fun for anyone.

Leave a Reply