277 V, 347 V and 480 V: What North American Lighting Voltages Do to Your Driver Spec

|Power Supply Knowledges|11/08/2026|10.8 min|
Table of contents
Share Post

A lighting contractor in Texas orders a container of 150 W drivers from a European catalogue, installs them on a 480Y/277 V lighting panel, and everything works. Six months later, units start failing in ones and twos — not a batch, not a date code, just a slow trickle from the circuits that run the longest. The supplier’s lab tests the returns, finds the input electrolytic capacitors bulging, and concludes the customer has a “power quality problem”.

The customer does not have a power quality problem. They have a driver that was designed for 230 V nominal, sitting 2% below its capacitor’s voltage rating on a 277 V circuit, in an enclosure at 55 °C, cycling twice a day. That is not a defect either. It is a specification mismatch that takes six to eighteen months to show up, which is exactly long enough for the shipment to be paid for.

This is the most common and most expensive mistake in transatlantic lighting sourcing, and it is entirely avoidable if you understand what the four North American voltages actually are.

Four voltages, one product family

North American lighting runs on four distributions, and they are not interchangeable:

System Line-to-line Line-to-neutral Where you meet it
Single-phase 120/240 240 V 120 V Residential, small retail
Three-phase 208Y/120 208 V 120 V Older commercial, office fit-outs
Three-phase 480Y/277 480 V 277 V Warehouses, big-box retail, parking garages, exterior
Three-phase 600Y/347 (Canada) 600 V 347 V Canadian commercial and industrial

The key point is that 277 V is not an odd variant of 230 V. It is a line-to-neutral voltage on a 480 V three-phase system, and it exists because it lets a building run lighting on a single pole of a 480 V panel without a transformer. In a US warehouse or a Walmart, essentially all the high-bay and exterior lighting is 277 V. In Canada, the same logic produces 347 V.

European and Asian driver catalogues are built around 220–240 V nominal. Ours states 100–240 V on every model up to 150 W, and 175–240 V on the 400 W and 600 W units, because a wide-range front end gets more expensive as the power goes up. Either way, the top of the window is 240 V nominal, and at the +10% tolerance the installation may legitimately present, 264 V. That is not 277 V, and it is nowhere near 347 V.

What 277 V actually does, number by number

Here is the arithmetic that decides whether a design survives. All figures are peak values at the stated nominal plus 10%, which is the normal upper tolerance a US installation is expected to tolerate.

Parameter 230 V design 277 V requirement 347 V requirement
Nominal RMS 230 V 277 V 347 V
+10% RMS 253 V 305 V 382 V
Peak voltage 358 V 431 V 540 V
Bulk capacitor rating typically used 400 V 450 V 600 V
X2 capacitor rating 275 VAC 305–310 VAC 440 VAC
Varistor 275 VAC 300–320 VAC 440 VAC
Creepage (pollution degree 2) 3.2 mm 4.0 mm 5.0 mm

Look at the third row. On a 230 V supply the peak is 358 V and a 400 V capacitor has 12% margin. On a 277 V supply at +10%, the peak is 431 V and the same 400 V capacitor is over its rating by 31 V. It does not explode on the first day. It runs hot, its electrolyte evaporates faster, its ESR climbs, and the driver fails somewhere between month six and month eighteen — sooner if the enclosure is warm, sooner again if the site has voltage sag and recovery events.

This is why the Texas failures clustered on the longest-running circuits and not on a production batch. It was a thermal-aging failure driven by an overvoltage that was present in every unit but only lethal at high duty.

The capacitor that gets missed: X2

The bulk electrolytic gets the attention because it fails visibly. The X2 film capacitor across the input is the one that fails the safety review.

X2 capacitors are rated by the nominal system voltage they are intended for, not by a withstand voltage. A 275 VAC X2 is the standard choice for a 250 V system, which covers 230 V nominal. On a 277 V system, the continuous voltage at +10% is 305 V, and the correct part is a 305 VAC or 310 VAC X2 — which exists, is widely available, and costs about $0.04 more.

A 275 VAC part on a 277 V circuit will not fail immediately either. It will lose capacitance faster than specified, its self-healing events will become more frequent, and in a humid enclosure it can track across the surface and produce a line-to-line arc. That is a fire scenario, not a warranty scenario, and it is why the X2 rating is one of the first things a UL inspector looks at.

For 347 V in Canada, both the X2 and the bulk capacitor change class: 440 VAC X2, 600 V bulk. At that point you are not modifying a 230 V design, you are designing a new input stage. The 600 V bulk capacitor is physically much larger, which usually forces a bigger enclosure — and that is the point at which the project stops being a catalogue order.

60 Hz, not 50 Hz

A smaller difference that catches people out: the mains frequency changes the ripple.

At 50 Hz, the rectified ripple on the bulk capacitor is 100 Hz. At 60 Hz it is 120 Hz. The bulk capacitor’s ripple current rating is specified at 100 Hz or 120 Hz with a frequency correction factor, and the difference is 5–10% — usually not decisive. What is decisive is the output ripple: a single-stage design passes a portion of the input ripple to the output, and the output capacitor has to absorb it at 120 Hz instead of 100 Hz. On a 24 V output with a 240 mV ripple specification, the extra 20 Hz can push a marginal design over the limit.

The second-order effect is EMI. An input filter tuned for 50 Hz conducted emissions (150 kHz to 30 MHz, measured with a quasi-peak detector) has a slightly different corner when the source impedance and the line frequency change. In practice, a design that passes EN 55015 at 230 V/50 Hz will usually pass FCC Part 15 at 277 V/60 Hz, but “usually” is doing a lot of work in that sentence, and it has to be tested.

Certification: the part that determines the timeline

If you are selling into the US, the driver needs to be listed by an NRTL — UL, ETL, TÜV SÜD America, CSA — and the listing has to cover the product as installed. The relevant standards:

  • UL 8750 for LED equipment. This is the one that applies to a driver inside a luminaire.
  • UL 1012 for power supplies, when the driver is sold as a standalone unit.
  • UL 1310 if the output is Class 2 (limited to 100 VA and 60 V, which is its own subject).
  • CSA C22.2 No. 250.13 for Canada, with the 347 V rating called out separately.

Two things about this that European suppliers consistently underestimate. First, a CE declaration and a CB report do not substitute for an NRTL listing. US inspectors and AHJs want a listing mark; a CB report is useful evidence but it is not the mark. Second, the listing is per model and per construction, so adding a 277 V variant to an existing 230 V family is a new listing — with new samples, new thermal testing at the higher input voltage, and a new file. Budget eight to fourteen weeks and $6,000–15,000 depending on the NRTL and how much of the existing test data carries over.

This is the real cost of the 277 V decision. The components are cheap: 450 V bulk capacitors, 310 VAC X2 parts, a 320 VAC varistor and a slightly wider creepage slot add perhaps $1.80 to a 150 W driver. The listing adds more than the entire bill of materials, and it cannot be rushed.

What to send for a 277 V quote

If you are sourcing for a North American project, five pieces of information will get you a real answer instead of a maybe:

  1. The system voltage and the panel type — 480Y/277, 208Y/120, or 600Y/347. Not “US voltage”.
  1. The output requirement — voltage, current, and whether the load is constant voltage or constant current. A 277 V input with a 24 V/6.25 A output is a very different driver from 277 V in and 0–10 V dimmable constant current out.
  1. The certification target and the NRTL you want to use. UL vs ETL changes the cost and sometimes the design (a few UL interpretations on spacing are stricter than the IEC equivalents).
  1. The enclosure and the mounting orientation, because 277 V input components are physically larger and the thermal design at 60 Hz is different.
  1. The volume and the annual split between 230 V and 277 V builds, because a dual-range design (180–305 V) that covers both is possible and may be cheaper than two separate SKUs.

That last point is worth expanding. A dual-range input covering 180–305 V is technically achievable with a 450 V bulk capacitor, a 310 VAC X2, and a PFC stage with enough headroom. It lets one SKU serve both the EU and the US, which sounds ideal until you look at the efficiency: the same driver optimised for 230 V loses 0.8–1.2% efficiency when it has to work across the wider range, because the PFC inductor and the switching frequency have to be compromised. On a 150 W unit that is 1.8 W of extra heat for every unit sold into Europe, forever, to serve a market that may be 20% of the volume.

The honest answer is usually two SKUs: a 100–240 V build for the EU/Asia volume and a 249–305 V build for 277 V, sharing the same PCB with different input components. Same tooling, same housing, same secondary side, two bills of materials. That is what we do on the 150W IP20 platform for North American OEM customers, and it keeps the certification files separate without duplicating the mechanical design.

What it costs, in the order you will feel it

Item Cost impact on a 150 W driver Timeline impact
450 V bulk capacitor + 310 VAC X2 + 320 VAC varistor +$1.80 per unit None
Wider creepage, revised PCB +$0.60 per unit at 1,000 pcs 2 weeks
277 V thermal re-validation Engineering time 1 week
NRTL listing for the new variant $6,000–15,000 one-off 8–14 weeks
Dual-SKU inventory (230 V and 277 V) Working capital Ongoing

The last line is the one finance notices. Two SKUs means two safety stocks, two forecasts, and the risk of having the wrong one. For a distributor serving both markets, a single dual-range SKU often wins on total cost even with the 1% efficiency penalty, because carrying cost beats component cost. For an OEM building one product for one market, two SKUs is cleaner.

The mistake to avoid

The mistake is not buying a 230 V driver for a 277 V job. Everybody knows that is wrong on paper.

The mistake is buying a driver whose datasheet says “input: 180–264 V” and assuming that covers 277 V because 277 is close to 264. It does not. 264 V is the top of a 240 V +10% window. 277 V is a different nominal, and at +10% it is 305 V — 15% above the top of that window, and 31 V above the peak rating of the capacitor inside.

If a supplier tells you their 180–264 V product “works fine at 277, we sell a lot to the US”, ask them two questions: what is the bulk capacitor voltage rating, and what is the X2 rating. If the answers are 400 V and 275 VAC, they are selling you a driver that will fail after the warranty period, and the only remaining question is whether it fails before or after the invoice is settled.

FAQ

Can I use a 230 V LED driver on a 277 V circuit? Not safely. At 277 V plus 10%, the peak is 431 V, which exceeds the 400 V bulk capacitor used in most 230 V designs. The failure is a slow thermal aging process, not an immediate breakdown, so it typically appears after six to eighteen months.

What is the difference between 277 V and 347 V drivers? 277 V is the line-to-neutral voltage of a 480Y/277 system in the US. 347 V is the equivalent on a 600Y/347 system in Canada. The 347 V design needs 600 V bulk capacitors and 440 VAC X2 parts, which usually means a larger enclosure — it is not a component swap on a 277 V design.

Do I need UL listing or is a CB report enough? For installation in the US, you need an NRTL listing (UL, ETL, TÜV SÜD America, CSA). A CB report is supporting evidence, not a listing. Inspectors check for the mark.

Does 277 V change the output or only the input? The output stage is largely unchanged. The input stage — bulk capacitor, X2, varistor, creepage, PFC inductor — changes, and the thermal design has to be re-validated because input components run hotter at the higher voltage.

Can one driver cover 120–277 V? Yes, with a wide-range PFC front end, at the cost of roughly 1% efficiency and a taller input section. It is a reasonable choice for a distributor carrying one SKU; it is a poor choice for an OEM whose volume is 90% in one market.

Sourcing for a 480Y/277 or 600Y/347 project? Our OEM/ODM LED driver manufacturing line builds 277 V and 347 V variants on the standard indoor and IP67 platforms, with the input components, creepage and NRTL listing handled as one package. Send the panel voltage, the output requirement and the target NRTL, and we will come back with the bill of materials delta and a realistic certification timeline — including the cases where the honest answer is that the volume does not justify a listing.

SANPU

Stay in the loop

Subscribe to our free newsletter.