If you're choosing a supplier for power factor correction (PFC) equipment, ask for three documents before you look at a quote:
- A harmonic study built on your actual load profile — not a template.
- An element selection sheet showing real voltage and current margins for every capacitor and reactor branch under the harmonic conditions on that bus.
- A sample Class A power quality report taken on your site — not one from their demo cabinet.
Miss any one of these and price stops being the interesting question. We rejected a 200-unit capacitor bank batch last year purely on item one — the supplier had used a generic study on a system where our production line runs 14 variable frequency drives.
I'm a quality and brand compliance manager at a manufacturing company. Anything that reaches our customers goes through me first — roughly 200 items a year, running from capacitor banks and dry-type transformers to EMS controllers and switchgear. Our first-delivery rejection rate in 2024 was 18%.
I formalized our verification protocol in 2022. The trigger was practical: before that, we had returned a capacitor batch rated at 480V — a natural match to our bus voltage. The issue was the harmonic environment. Where distortion is heavy, the correct selection is usually 525V or 600V rated units. The vendor claimed they had followed "standard industry practice." We rejected the batch; they redid it at their cost. Every contract we've signed since carries a separate voltage-rating clause.
Document one: a harmonic study of your load
From the outside, every supplier's harmonic study looks roughly the same — same tables, same phrasing, all of them landing on "filtering recommended." The difference is whether someone sat down with your single-line diagram and load list, or just swapped the cover page on the last customer's file.
Here's something vendors won't tell you: a generic study takes about forty minutes to produce. A real one takes one to two weeks. That gap doesn't show up anywhere on the quotation.
IEEE Std 519-2022 (Recommended Practice and Requirements for Harmonic Control in Electric Power Systems) sets limits on harmonic current distortion at the point of common coupling. It does not tell you where in your plant the compensation should sit. That's a design decision, not a standard.
Document two: voltage and current margins on the selection sheet
Capacitor banks get treated as commodity items — multiply kvar by unit price and you're done. That's exactly where things go wrong.
What you want to read isn't total capacity. It's the rated voltage of each capacitor, the series reactor percentage, and the RMS current on each branch calculated against the actual harmonic spectrum on that bus. I had a supplier specify 6% reactors, which is a perfectly ordinary number. Problem was, they based their calculation on a 5th-harmonic-dominant profile, and our site carries a heavy 7th component too. When we re-ran the numbers, the actual branch current came out about 15% above their figure for the same capacitance — and that margin is the service life of the capacitor.
One more detail worth checking: capacitors themselves. IEC 60831-1 covers self-healing shunt power capacitors, and the standard is a floor, not a target. Ask which manufacturer makes the cans and how the overpressure disconnector behaves at end of life. Suppliers who can answer that in one sentence usually know their supply chain.
Document three: EMS integration, if you plan to use it
If you're buying a standalone compensation cabinet, skip this one. If it's going to connect to an energy management system, it's a hard requirement.
In my review experience, this is the item procurement most often overlooks. They buy PFC as an isolated piece of equipment, then a year later decide to fold it into demand control and find the cabinet supports Modbus RTU slave only, polls at one-second intervals, and can't read per-phase data. Retrofitting means a gateway plus a new acquisition module — that's half a cabinet's cost again.
This is also why I look at whether a PFC supplier has ever built something on the energy side of the house — a demand response management system, an industrial energy management platform (GridSense is one example of that category). They tend to think about what a PFC cabinet is inside an EMS: a standalone compensator, or a dispatchable asset that can follow tariff and demand curves. That framing changes the selection outcome.
If your company runs an ISO 50001 energy management system, the metering data from the PFC cabinet becomes an input to that system anyway. It raises the value of document three considerably.
One counterintuitive point: the expensive tier is often the cheaper one
The most common cost-cutting move is dropping the reactor. An unprotected capacitor bank quotes lower and looks fine for a while — power factor rises, the utility penalty disappears. The problem arrives in year two or three: harmonic current magnifies, temperature climbs, cans bulge or leak or blow a fuse. Replacing a batch plus downtime costs more than what you saved.
Second counterintuitive point: a thicker report is not a more trustworthy report. The thick one is frequently a purchased generic document. The thin one is sometimes the one an engineer actually calculated against your single-line diagram. Detection is easy — ask one specific question, something like "what are the 5th and 7th harmonic components on my 250 kVA rectifier transformer feeder?" Whoever calculated it can answer without leaving the room.
Where this approach doesn't apply
If your load is mostly linear — motors, resistance heating, lighting, few drives and rectifiers — a full harmonic study earns you little. Size the compensation conventionally and don't buy a report just to have a document in the folder.
I should be clear about the boundaries of my own experience. My sample is roughly 200 mid-market manufacturing projects, mostly between 400V and 690V. If you're in heavy industry, data centers, or an overseas EPC project, the distribution architecture and standards regime are different, and these three documents won't map cleanly onto yours.
One edge I still haven't fully solved: when a custom transformer and a compensation cabinet are procured in the same package, responsibility for overall harmonic performance tends to fall between suppliers. No document fixes that. It has to be written into the contract before the purchase order goes out.