How to Evaluate LED Power Supply Manufacturers: A 7-Step Checklist

Who this checklist is for

This is for buyers and engineers evaluating LED driver manufacturers—especially for constant current drivers and dimmable drivers used in commercial or OEM projects. If you’re sitting on three or four quotes and need to narrow down before a first 5,000-unit order, run them through this sequence.

Seven steps. The order matters.

Step 1: Request the model specification table, not the brochure

A brochure tells you what the manufacturer wants to sell. A specification table tells you what they actually make.

Ask for the product matrix—the one that lists every model with output rating, input range, dimming protocol, and temperature derating. If the reply is “we’ll send you our catalog,” that’s a signal. Not a dealbreaker, but a signal that your question hasn’t reached the right desk yet.

Look for two things: whether every SKU gets its own row (not “TBD” ranges), and where the “pending” values sit. If “efficiency” and “THD” both say TBD, you’ll be paying for the testing later.

Step 2: Match the dimming protocol, not just “supports dimming”

Most buyers check the box that says “dimmable” and move on. That’s the gap. The real spec is the protocol plus the driver topology.

Put these into your RFQ response template:

• Dimming method (0-10V, 1-10V, DALI-2, TRIAC, DMX, wireless modules)
• Dimming range—down to what percentage
• Curve type (linear vs logarithmic)
• Number of dimming channels per driver
• Whether dimming side is isolated

In early 2024, we evaluated a supplier who replied “DALI-compatible.” Turned out that meant “compatible,” not native. For commercial installs, that distinction meant either a reprocess or a full return.

This is also where the spec sheets from manufacturers like Inventronics tend to get pulled out as reference—their dimmable driver lines list protocol by model number, not by marketing claim. Whether or not you buy, that’s the level of specificity you should expect from anyone on your shortlist.

Step 3: Verify certifications against the database, not the logo

Safety certification is where buyers get burned most often. A supplier can put a UL mark on the box and have exactly one model covered.

Do this:

1. Ask for the certificate number (not just “UL listed”)
2. Go to the UL Product iQ database (or equivalent) and search
3. Match the model numbers on the certificate against the ones you’re quoting

Same process for CE, ENEC, CCC, and TUV. A single certified model doesn’t cover the line.

One step almost everyone skips here: ask the manufacturer whether any component change (e.g., swapping the capacitor brand) would void the certificate. Some will. Get that in writing.

Step 4: Read the thermal derating curve, not the rated wattage

A 100W driver is 100W at 25°C. In a 60°C fixture housing, it might be 70W.

Ask for the derating curve—the graph that shows maximum output over ambient temperature. Size against that line, not against the nameplate.

I’ve lost count of the spec tables where the headline says “100W” and the footnote says “derates above 45°C.” For outdoor or industrial projects, that footnote is where deployments quietly fail.

Step 5: Ask for aging test data

A certificate that says “100% burn-in passed” is not data. It’s a statement.

Ask for:

• 500-hour burn-in results across lots/date codes
• Sampling ratio used
• Instrumentation for the parameters measured (voltage, current, ripple, efficiency)
• AQL sampling plan for routine production

If they can’t produce it, that isn’t an automatic rejection. It just tells you samples are going to be your substitute for data. But a mature factory should have this ready or at least be able to promise it.

Step 6: Nail down private label and OEM terms

If you’re sourcing LED drivers for private label, labeling isn’t your bottleneck. Certification is.

Ask:

• Who re-certifies when the label changes—them or you?
• Does a private label SKU need new safety certification?
• What’s the minimum order per SKU?
• Who provides the label and packaging artwork?
• If the packaging ships to export markets, what needs to change?

I’ve watched three brand owners do what they thought was a simple “stick our logo on it” private label run, only to find out re-certification was 8-10 weeks. Sort that question before procurement, not after.

Step 7: Run a blind sample test

Last step. Don’t test brands. Test units.

Order ten pieces from two vendors. Run them under the same load for 24 hours. Remove the labels. Have your engineering team score them blind.

We did this in 2023—two vendors, 30% price difference, nearly identical spec sheets. Blind scoring clustered around the cheaper one, calling it “cooler running.” Turned out the premium unit ran marginally cooler, but its rough cable jacket biased the testers before they’d even looked at the data.

Blind tests don’t always change the decision. They always change what you know about the decision.

Common mistakes to watch for

Comparing price before derating. A driver rated at 70°C can look 30% more expensive until you compare it to what the cheaper one actually does at 70°C. Compare specs at deployment conditions, not at 25°C.

Confusing “does everything” with “does anything well.” A vendor that says yes to every scope is running a different business than one that names its limits. When a supplier told me once, “this isn’t our strength—here’s who does it better,” I trusted them more on everything else, not less. Focus is the quietest quality signal there is.

Skipping the certification database step. It takes twenty minutes. It costs you a reprocess later if you skip it.

Signing before the sample is validated. Get a conditional PO with a mass-production validation clause if the supplier passes this round.