Most LED bulb tests are useless. The bulb goes into the socket, it lights up, and everyone calls it a pass. But that check tells you nothing about whether the fixture will match its neighbors in three years, whether the driver can survive a lightning storm, or whether the lumens on the box are even real. For a single lamp in a broom closet, that's fine. For a contractor or specifier buying 100 downlights or 50 path lights for a client, it's false confidence.
So here's my advice in one sentence: test the documentation before you test the bulb. If a vendor can produce an LM-79 photometric report, LM-80 lumen maintenance data, a TM-21 lifetime projection, and a stated MacAdam ellipse tolerance, you already know more about that product than any ten-minute hands-on test will ever tell you. If they can't, no amount of “it looks good to me” should win the argument.
I'm the quality and brand compliance manager at bega lighting, a manufacturer of architectural outdoor fixtures. I review every product and component that ships under our name—roughly 200+ unique items a year. Actually, closer to 250 in 2024. I've spent the past four years building the verification process we use now, and I built it mostly because we got burned early on. What follows is what I actually check, and how you can run a meaningful version of it without a photometry lab.
What Fails First in an LED Fixture
If you ask most people what dies first in an LED fixture, they'll say the LED. They're almost always wrong. The LED package itself is good for 50,000 to 100,000 hours of operation. The components that actually fail are the parts nobody thinks about:
- The driver. This is the power supply, and it is the weakest link. Electrolytic capacitors dry out, surge protection lets through a spike, thermal shutdown trips too early. I've seen drivers fail in under 6,000 hours inside fixtures claiming a 50,000-hour system life. The LED was fine. The driver killed it.
- The thermal design. LEDs hate heat. If the heat sink is undersized, or the emitter is overdriven, color shifts and output drops far faster than the datasheet suggests. You can't spot this in a one-minute test, but you can often feel it after thirty minutes at full power.
- Color consistency. This isn't a total failure, but it's the most visible problem on real projects. Fixtures from the same batch can vary by 300–500K in color temperature. Side by side along a hotel pathway, that difference is the line between “architectural” and “discount.”
Here's the frustrating part of my job: none of these problems appear in a quick bench test. The fixture runs. The light looks warm and pleasant. Then the full batch arrives on site and 12% of the units are visibly a different color than the other 88%. That's the failure mode that costs everyone money, and it's invisible in a single sample.
We learned this the expensive way. Back in 2022, we didn't have a formal incoming inspection process for LED drivers. Cost us a villa project in southern Europe. A batch of drivers with marginal surge protection made it into 30 fixtures on one system. One nearby lightning event, and 26 of the 30 drivers were dead. The redo cost the customer around $22,000—I want to say the exact invoice was a bit higher, but that number stuck with me. Every specification we write now requires surge immunity tested to IEC 61000-4-5 at 2 kV for outdoor products. That rule was born from a bill we never want to see again.
Then there was the vendor I almost switched to. The numbers said go: 15% cheaper than our incumbent, comparable paper specs, every box ticked. My gut said no. They were slow to answer technical questions, and something about their application notes felt recycled. I stayed. Six months later, we learned their driver's surge protection was 1 kV—half of what our spec requires. They never even read the spec. Slow to reply was a preview of slow to comply.
How to Test LED Bulbs and Fixtures: Three Checks That Matter
If you've been searching for how to test LED bulbs properly, you've probably seen the usual advice: check the resistance, look for a burnt mark, try another socket. That's how to check whether a bulb is dead. It's not how to test whether a fixture is worth installing. Here's the practical process I recommend to contractors and specifiers. You don't need an integrating sphere. You need the right questions and, ideally, two tools that cost less than a hundred dollars total: a lux meter and an IR thermometer.
1. Test color consistency
If you have a spectrometer, measure the fixture and compare it to its rated CCT. Most people don't, and that's okay. There's a cheaper screening method: light two fixtures of the same model side by side on a white wall in a dark room. If you can see a color difference in the reflected light, the mismatch is probably 300K or more. That's meaningful.
But the more important test is a question, not a measurement. Ask the manufacturer: “What is your color tolerance in MacAdam steps?” A spec-grade product holds a 2-to-3-step ellipse. Budget products often don't state one at all, or they sit at 5-to-7 steps. You will not see that difference looking at one sample. You will see it in a 40-unit installation.
2. Test the driver, not just the LEDs
First, ask whether the driver is replaceable. If it's integrated—potted into the fixture as a throwaway component—the entire unit becomes e-waste when the driver dies. That's typically year three to five. This is a real cost difference between a $45 fixture and a $120 fixture, and it never appears on the invoice.
Then run two checks that take minutes:
Dimming test. If the product claims to be dimmable, dim it to its lowest level and watch it through your phone camera. If you see moving bands on the screen, the fixture is flickering at a frequency many people can perceive. Ignore this and you'll be dealing with complaints—or migraines—before the punch list is done.
Thermal test. Run the fixture at full output for 30 minutes, then place your hand on the housing. If you can't hold it for five seconds, the thermal design is marginal. Heat is the enemy of LED life and color stability. Extra heat sinking is one of the things you're actually paying for when you buy a heavier, better-built fixture.
One surprise I keep in mind: not all cheap fixtures deserve the bad rap. We once evaluated a budget downlight LED for a temporary installation and its color consistency was genuinely good—better than a premium competitor tested in the same session. The surprise wasn't in the LEDs. It was in the driver, which flickered visibly below 30% dimming. Good color, bad driver. That fixture would have driven a facilities manager crazy within a month. That's why testing the whole system matters, not just the pretty light.
3. Verify the output claim
If you have a lux meter, there's a rough but useful check. Open the fixture's IES file and find the candela value at beam center. Place the meter one meter from the fixture, aligned with that beam center—below a downlight, perpendicular to a path light lens. At one meter, lux equals candela, so your reading should come close to the IES value. If it's more than 10–15% below, either your setup is imperfect or the photometry is optimistic. Trust the discrepancy, re-check, and if it persists, ask the manufacturer to explain.
The Four Documents That Separate a Manufacturer From a Trader
This is the heart of my advice, and I want to be blunt: the single most reliable test in lighting is asking for paperwork, then noticing how fast it arrives.
- IES LM-79. The approved test method for measuring a complete solid-state lighting product—luminous flux, input power, CCT, CRI, chromaticity. When a vendor claims “12,000 lumens,” the LM-79 report is what actually backs that claim.
- IES LM-80. The test method for measuring lumen depreciation of the LED source over time. It tells you how the specific chip in your fixture ages, based on real test data rather than hope. Good LM-80 data is worth more than any warranty promise.
- TM-21. The projection that converts LM-80 data into a lifetime estimate, usually expressed as L70—the number of hours until the fixture retains 70% of its initial light output. When you see “50,000-hour LED,” TM-21 is where that number comes from.
- MacAdam ellipse tolerance. The manufacturer's answer to “how consistent is your color?” If they can't answer in steps, they don't know. And if they don't know, the consistency is not being controlled.
Test the response time. Ask for an LM-79 report by name and watch what happens. A company that has actually tested its products will email the PDF within the hour. A company that's rebranding imported fixtures will change the subject or say “we'll get that to you.” Then they don't. I've run this exact test with dozens of vendors, and the response time is remarkably diagnostic.
Why “Three Times the Price” Can Still Be the Cheaper Option
I hear the price objection on bega path lights almost every month. At unit level, the gap is real. Let me show you the math I use when a customer raises it.
A hotel installed 24 path lights at their entrance. Budget fixtures, $45 each. Spec-grade, $120 each. The spec-grade option costs $1,800 more upfront.
Year two: a storm sends a surge through the site. Nine of the 24 budget fixtures lose their drivers. Each replacement costs $95 installed—the fixture isn't the expensive part; the truck roll and the labor are. That's $855 in year two.
Year four: the facilities manager notices the entrance path is half “warm white” and half “yellow” in the same photo. The surviving units have drifted. They replace all 24: $1,080 in fixtures plus labor—roughly $2,280 total, installed.
Four-year total for the budget route: $4,215. Four-year total for the spec-grade route: $2,880, zero service calls, and the entrance never looked bad even once. Same path, same purpose, different ledger.
I'm not saying the spec-grade fixture will never fail. Nobody in this industry should promise that, including us. But the failure probability is lower, the color tolerance is controlled, and the documentation tells you exactly what you bought. From my experience reviewing thousands of fixtures under real conditions, the lowest quote has been the most expensive option more often than chance would predict.
When the Full Test Is Overkill
Not every purchase justifies this level of scrutiny. If you're buying a floor lamp for a reception corner or a budget downlight LED for a storage closet, screw it in, see if you like the light, and move on. The cost of being wrong is low, and the aesthetic judgment is yours, not a spec sheet's.
There are also honest limits to what you can verify from outside the factory:
- You can't test lumen depreciation in an afternoon. That's what LM-80 and TM-21 exist for. A vendor can exaggerate in conversation but is less likely to stretch the truth on a formal document.
- A single sample says nothing about batch consistency. I've seen beautiful first articles followed by chaotic production runs. Order two samples from different production dates, not two units from the same box.
- “We use the same LED chip as [premium brand]” is a sentence that should raise your skepticism, not lower it. The chip is maybe a third of the story. The driver, the thermal path, the assembly quality, and the testing are the rest.
So: run the three checks in this article if you have the tools, and by all means take photos, compare samples, and ask uncomfortable questions. But the strongest test is still the one that happens before you touch a fixture—demanding the documents and noticing who can produce them. The lighting that looks good in a box is easy. The lighting that still looks good on a wall in year four is a different product entirely.
That's the difference I look for every day, and it has nothing to do with the price sticker.