Buying Event and Decorative Lighting? Check the Power Supply First

The thing that breaks first is never the LED

After six years of ordering lighting for company events, the pattern is boring and consistent: the light doesn't fail—the power supply does. Of the decorative lighting returns, write-offs, and dead-on-arrival claims I've processed since 2020, I'd put roughly 7 in 10 down to a driver, battery, or seal failure. Not one of them to a dead LED.

That flips what you should look at when you buy. Not color. Not shape. Not the photo on the product page.

Who's asking, and why it matters

I'm the office administrator for a 180-person company. I manage all facilities and event ordering—roughly $60,000 annually across nine vendors. I report to both operations and finance, which means I get judged on cost and on whether things actually work when the guests arrive.

Part of that spend is boring: paper, coffee, janitorial. Part of it is the fun stuff—the lobby display that gets swapped every season, the summer party, the trade show booth, the bar setup for the annual client dinner. That second category is where I learned the most, mostly by getting it wrong.

When I took over purchasing in 2020, I treated decorative lighting like party supplies. Buy what looks right, buy enough of it, move on. Three years of watching fixtures die early taught me that it's really a small-scale electrical purchase, and the electrical part is where all the risk sits.

Floating lights: the IP number tells you less than you think

First real mistake. I bought pool floating lights based on the rating alone. IP68. Done, right?

Half of them were fogged or dead within three weeks. Not because the LED failed, and not exactly because the rating was a lie. Because the seal around the battery compartment was a single O-ring assembled dry, and the charging port was covered by a friction-fit plug that nobody seated properly.

What I check now is boring stuff. Is the battery compartment sealed with a compression gasket or just a rubber plug? Is the charging port covered by a screw-down cap or a push-in plug? Does the manual tell you to lubricate the seal and give you an interval for it?

That third question is the one most buyers skip. A seal that needs servicing is fine. A seal the manufacturer pretends will last forever is a design failure.

One more thing worth knowing: IP codes come from IEC 60529, and IP68 means dust-tight plus protected against continuous immersion—but the standard lets the manufacturer define the depth and duration. "IP68 at 1 meter for 30 minutes" and "IP68 at 3 meters continuous" are not the same product, and they're rarely labeled differently on a marketplace listing. Read the footnotes.

Solar lights: buy the panel and the battery, not the lamp

We put solar lights in three planters at the entrance in 2021. They worked beautifully from April to September. By the second winter, two of six were dead and the rest were dim.

The panel wasn't the problem. The battery chemistry and the charge controller were.

Cheap solar lights use whatever cell is cheapest—often a small NiMH pack or a low-grade LiFePO4 with no low-temperature cutoff. Cold weather drops usable capacity, and repeated deep discharge kills the cell outright. There's no battery health indicator on a garden light, so you just watch it get dimmer until one day it's off. What I ask suppliers now: panel wattage at standard test conditions (not "high efficiency"), battery chemistry and rated cycle life, whether there's low-temperature charge protection, and whether the cell is replaceable or the whole unit is disposable.

That last one is the reverse-intuitive part. A replaceable battery on a $40 solar light beats a sealed battery on a $70 one, because at least you get a second life out of it. I've thrown away more "premium" sealed units than cheap serviceable ones.

Color-changing chairs and chair LED setups: color consistency is a real spec

Nobody warns you about this one.

We rented a lounge setup for a 2023 holiday party—twelve color-changing chairs, LED integrated. Identical in the showroom photos. Identical in the staging photos. Under the event lighting, four of them rendered purple as a slightly pink magenta and the other eight were a bluer violet. Ten minutes into the event, three of us were in a corner with one remote trying to find a color that made all twelve match.

There's a number for this. Color matching tolerance is usually expressed as Delta E, where Delta E under 2 is a tight, brand-critical match, 2 to 4 is noticeable to a trained eye, and above 4 is obvious to most people—thresholds that trace back to the Pantone Color Matching System guidelines and the underlying color science. A supplier who can't tell you their Delta E for a batch, or who's never been asked, probably isn't testing per-unit consistency at all.

Two other things I check on LED chairs. First, the control protocol: an IR remote means one chair, one line of sight, while RF or app control means one remote for all twelve—but only if the receivers genuinely pair as a group. Second, startup and fade behavior. Chairs with matching colors but different firmware can fade at different rates, so the "breathing" effect looks drunk. Ask for a demo of the sequence, not a static color. Ask about the chair LED controller specifically; it's usually a different component from the LEDs and it's the one that decides whether the set stays in sync.

Bar table lights: the case for spending more on the driver

For our annual client dinner we light about twenty high-top tables. We started with cheap battery-powered table lamps at roughly $8 a unit, and they were a nightmare. Different brightness per unit. A touch switch that woke up in the transport case. Half of them dead by 8 p.m. because something had brushed the sensor in the crate.

We moved to a cordless design with a real driver and a physical on/off instead of a touch pad. Higher unit cost. And genuinely cheaper, because I stopped replacing them and stopped getting asked why table seven was dark.

For the permanently wired bar lighting—the fixed counter we installed in 2024—we used Tridonic drivers. The reason wasn't the logo on the box. It was consistency. At a 5% dim level, twenty fixtures on twenty different no-name drivers will not hold the same low-end curve; they flicker, drop out, or visibly differ. Using one driver model across the whole installation plus DALI control meant one dimming curve for the entire room. If you're wiring something permanent and controllable, that's the level of specificity you want in your spec sheet.

Light-up ice cubes: this is a battery safety purchase, not a lighting purchase

Light-up ice cubes are where I've seen the widest gap between expectation and reality. They're molded plastic with an LED, a coin cell or small rechargeable, and a switch. That's the whole product.

A few things I've learned the hard way. Treat them as consumables unless you verify otherwise—the cheap ones have essentially no moisture protection, and a drink is a moisture environment, so water gets in and the cell corrodes. Ask about battery and safety documentation instead of just color options: small lithium rechargeables fall under UN38.3 for transport and IEC 62133 for the cell itself, and many markets have specific warning requirements around coin cell ingestion. A supplier who can't produce those documents is a compliance risk you're carrying, not them.

And rechargeable isn't automatically better. A sealed rechargeable cube has a fixed number of cycles and then it's landfill, while a cube with a replaceable coin cell lasts longer in practice for occasional use. Not glamorous. But when I bought forty of these for a company anniversary and two of them leaked into the linen on the bar top, glamour was not the issue.

When you should ignore all of this

If you need fifty light-up ice cubes for one Friday happy hour and they're going in the trash on Monday, buy the cheap ones. Please. The documentation homework costs more time than the product is worth.

Same logic for a single seasonal display or a one-week pop-up. If the fixture lives in a crate and the budget is the real constraint, buy on appearance and accept the failure rate going in.

The driver-first, battery-first approach earns its keep when three things are true at once: the product gets used more than once, multiple units have to look identical at the same moment, and somebody—a client, a guest, an auditor—will notice if it doesn't work. That covers most recurring corporate event and installed work. It doesn't cover everything.

To be fair, even with all the vetting I've still had a batch arrive and die immediately. No inspection process removes bad luck. The difference is that now, when it happens, I can point to the specification, the raised gasket, the replaceable cell, and the battery documentation—and the vendor usually eats the replacement without an argument.