Mean Well 48V Power Supplies: The $3,200 Mistake in Capacitors and Connectors (And How to Test Capacitors With a Multimeter)

Most Mean Well 48V power supplies and LED drivers don't fail on their own—they fail because of bad capacitors and under-spec connectors on the load side. That's not a guess; it's a $3,200 mistake I made in 2022, on an order that included a mix of Mean Well LRS-350-48 and Mean Well HLG-480H-48-A units. The fix was a two-minute capacitor test and a proper look at connector ratings.

I'm a procurement lead handling power-supply orders for a telecom equipment integrator. Nine years in, I've personally made and documented six significant mistakes, totaling roughly $18,000 in wasted budget. Now I maintain our team's pre-build checklist. The $3,200 lesson below is the reason that checklist exists.

The $3,200 lesson started with an order that looked fine

In March 2022, we ordered 80 Mean Well LRS-350-48 supplies (48V, 7.3A) and 40 Mean Well HLG-480H-48-A LED drivers (48V, 10A) for a customer's industrial lighting and control rack. Incoming inspection passed: output voltages measured within spec, no damaged units, all connectors seated. We installed the rack, ran a full test, and shipped it.

Two weeks later, six units came back with no output. A week after that, another eleven failed. My first assumption was a bad batch from Mean Well USA Inc. Their support team checked the serial numbers and asked for board-level photos. The Mean Well units themselves were fine. The failures were on the custom interface boards we had built between the supply and the load.

Connectors: the most expensive mistake

The first issue was simple: we used a screw-terminal connector rated for 5A on a DC bus carrying 7.3A from the LRS-350-48. On the HLG-480H-48-A, the same style connector was carrying 10A. We matched the wire gauge and the connector physically fit, so we didn't think about current rating. That was the mistake.

Under load, the connector dropped voltage and generated heat. After repeated thermal cycling, the solder joints cracked. A multimeter at the power supply terminals read 48.1V—which looked perfect. At the load end, the reading was closer to 44V and falling. The Mean Well power supply was doing its job; the connector couldn't pass the current.

Here's the counterintuitive part: choose connectors for worst-case inrush, not just steady-state current. Mean Well LED power supplies can draw several times their continuous rated current for a few milliseconds during startup. A connector rated for 5A won't fail on the first switch-on. It degrades slowly, which is worse because you won't see it until units come back from the field.

What I mean is that the connector's current rating isn't just a number—it includes crimp quality, wire gauge, ambient temperature, and airflow. A 5A connector with a poor crimp may only be good for 3A in practice. Use a generous margin and measure voltage at the load under full load, not at the supply terminals.

Capacitors: the hidden load-side problem

The second issue was on our DC bus capacitor bank. A batch of 100µF electrolytic capacitors from our board supplier measured around 70µF on incoming inspection. We didn't test them because they were from a trusted supplier and looked identical to the good parts. The lower capacitance caused excessive ripple current, which dried out the capacitors over time. That's a classic failure chain in LED power applications.

Once we understood what happened, I added a simple test to our incoming inspection. The name of the test is exactly what you're looking for: how to test a capacitor with a multimeter.

How to test a capacitor with a multimeter

This test takes about two minutes and catches the category of failures that cost us $3,200.

1. Discharge the capacitor. Put a resistor across the leads for a few seconds. A charged capacitor can damage your multimeter.
2. Set the multimeter to capacitance mode. Look for the symbol that looks like –|(– or the letter F.
3. Connect the probes. For polarized electrolytic capacitors, red goes to positive and black goes to negative. Polarity matters for an accurate reading.
4. Read the value. Compare it to the marked rating. A 100µF cap that reads below 80µF is suspicious. A reading near zero means it's shorted.

This quick test won't catch every possible capacitor failure—ESR is a separate issue—but it catches grossly under-spec parts, and that's exactly what caused our failure. Per FTC advertising guidelines (ftc.gov), manufacturers are supposed to substantiate product performance claims. But a datasheet claim is not a physical measurement. You have to be the last line of defense.

What our checklist looks like now

Since we added these checks, our post-commissioning failure rate on the Mean Well LED power supply side of our product mix dropped by about 30%. In the past 18 months, we've caught 47 potential errors that would have gone into the field. Not all would have failed immediately, but several would have come back—and that was the expensive path.

Now, on every order involving Mean Well power supplies, we:

• Check the connector current rating against the supply's maximum continuous current plus inrush.
• Test at least one capacitor from each lot or reel with a multimeter before assembly.
• Measure voltage at the load end, not at the supply terminals, under load.
• Record the results with date and lot numbers.

This adds maybe ten minutes per board. It has saved us weeks of rework and field returns. In efficiency terms, that's a trade I'll make every time.

Where this doesn't apply

To be honest, this checklist isn't universal. It works because we're a small-batch integrator and our load boards come from multiple suppliers. If you're an OEM using pre-qualified boards with properly selected components, you probably don't need to test every capacitor. If you're just replacing a Mean Well 48V power supply in an existing system, you can skip most of this.

I can only speak to our context. If you're doing high-volume production with in-house testing and controlled assembly, the calculus is different. And if you're a reseller that only ships Mean Well boxes, this is likely beyond your scope.

One more honest note: after creating the test procedure, I kept second-guessing. What if we were damaging capacitors during testing? What if the real cause was something else? The answer came in the data: field returns from those failure modes went to zero. A little front-end testing paid for itself many times over. It's not about being paranoid. It's about not letting a $3,200 mistake happen twice.