Why Your MeanWell 5V 20A Power Supply Keeps Failing (And Why It's Probably Not the Power Supply)
In early 2024, a crate of 13 power supplies showed up in our shop. All had been pulled from customer installations as "failed." Every single one was a Mean Well LRS-100-5 — the 5V, 20A unit that runs control circuits, sensors, and small network gear. Our service manager was ready to strike Mean Well from the approved vendor list. I asked him to hold that thought, and I did something a procurement guy doesn't usually get to do: I bench-tested all 13 units myself.
Eleven of them powered up and ran perfectly at full load. One had a swollen capacitor. One was within spec but noisy. That is when the warranty claims stopped and the actual investigation started.
The Problem Everyone Blames Is the Power Supply
Look, I get why people blame the supply. It has a brand on it, it sits between the wall and the machine, and it's the part that "stops outputting 5V" when something goes wrong. If you search for "meanwell 5v 20a power supply" and read the complaints, you'll recognize the pattern: the load starts browning out, the unit gets boxed up, and someone writes "failed" on the return label.
But here's the thing: a power supply regulates the voltage at its own output terminals. It cannot regulate the voltage at the far end of a bad connection. And on a 5V rail, the difference between "operating fine" and "resetting constantly" is a fraction of a volt.
The Deep Cause: A 5V Rail Has No Tolerance for Bad Crimps
After the bench test, we went back to the actual installations. Not to the power supplies — to the wiring.
Almost every failing installation had the same signature: at least one connector on the output side that had been crimped with the wrong tool, or not crimped at all, just squeezed with pliers and wrapped in tape.
Why does this get blamed so rarely? Because the voltage drop happens after the supply's output terminals. The supply holds 5.03V, the load sees 4.2V, and the tester never checks the load side. Here's the math that should be on every installer's bench: at 5V and 20A, a connection with 0.05 ohms of resistance drops a full volt. The same crimp on a 24V supply wastes a volt that nobody ever notices. On a 5V rail, that's 20% of the available voltage, gone.
In our files, we had a field measurement of a 5V rail sagging to 4.2V at the load while the Mean Well output terminal held 5.03V. The datasheet tolerance for the LRS-100-5 is roughly ±1% — about 50mV. An 800mV gap is not a power supply defect. It is a connector defect.
What most people don't realize — and I include our entire purchasing department in this — is that the comparison sheet stops at the enclosure. We spent weeks comparing price, lead time, and warranty terms across brands, and none of those metrics predicted the field failures. The wire in front of the supply doesn't come with a datasheet, a part number, or a warranty. It just sits there, silently dropping voltage, until someone blames the branded component.
The 2780 terminal incident
A specific story that made this real. One of those 13 units came from a packaging line. The original installer had used generic barrel terminals and squeezed them with combination pliers. The wire was 18 AWG; the terminal barrel was stamped 22–18 AWG. Technically the right range. In practice, the crimp was a squashed oval with air inside, and it ran warm under load and dropped voltage whenever the line sped up.
The site technician re-did that feed with a 2780-series locking spade terminal and a ratcheting crimp tool matched to the terminal die. I had to ask him what "2780" even meant — it's the terminal family code in our supplier's catalog. The terminal cost a few cents more. The tool made the actual difference. That connection has run for over a year with zero return visits.
Gray-market units: the other hidden failure
I also need to be honest about two of the 13 units. They were not genuine Mean Well products.
They had been purchased through an online marketplace listing that looked official, and one of them failed hard enough to take out the small controller it powered. The label looked fine. The behavior ranged from "okay, I guess" to "eventually scary." If you're in North America, this is my strong advice: buy through authorized MeanWell USA channels — the distributors listed on the company's website, plus established electronics distributors. The premium is warranty, safety, and a phone number that answers. I classify that as cost of doing business, not a markup.
A G310 5G gateway made it personal
The project that finally made all of this click was a G310 5G gateway at a remote monitoring site. The gateway rebooted at nearly the same time every day — right when a nearby pump started and dragged down the shared 5V feed.
The gateway drew under 3A. The Mean Well supply was rated 20A. On paper, the installation had massive headroom. The real problem was one butt splice in the feed line, badly crimped, acting like a variable resistor. Every pump start sagged the rail just enough to reset the gateway. One re-crimp fixed it for good.
What 13 'Failed' Power Supplies Actually Cost
When I closed out that audit in our cost tracking system, the breakdown looked like this:
- Replacement units: about $460 for all 13.
- Labor: diagnosis, truck rolls, removal, reinstall — roughly $3,700.
- Lost production: one six-hour line stoppage, valued at $2,100 by plant management.
The power supplies were the cheapest line item on the whole report. The money went to everything that happened around them. That is a hard story to explain to a CFO when the service ticket says "Mean Well failed," which is exactly why I'm writing it down here. The metric that matters is cost per year of reliable runtime, and chasing the lowest unit price never once improved that metric for us.
There was also a softer cost that never made it onto the invoice: trust. Every "failed" unit eroded confidence in a brand that wasn't actually failing. The service team started doing extra rework to "protect" customers from issues that their own terminations were causing. You can't put that on a purchase order, but you can feel it in every project review.
When I looked back at five years of invoices in our tracking system, the pattern was consistent: the expensive failures were never the units themselves. They were the decisions made around the units — the gray-market buys, the rushed installs, the missing crimp checks. That's the cost nobody budgets for.
One caveat before the fix: my sample is roughly 200 orders and 40-something field installations over five years, mostly in and around Texas. If you operate in clean-room electronics, high-vibration mobile gear, or international logistics, your failure modes are probably different. I can only speak to what I've tracked.
The Fix: Smarter Purchasing and Basic Crimp Discipline
The fix is deliberately short, because once you accept that the supply is usually not the problem, the changes are simple.
Purchasing rules we now follow
- Field units are bought only through authorized MeanWell USA distributor channels. If a marketplace seller isn't a listed distributor, we don't buy.
- When engineering specifies a 5V rail, we ask for the expected current at the load, not at the output terminal. The gap between those numbers is where bad crimps live.
- Connectors get visually inspected and pull-tested before power-on. It's on the sign-off sheet now.
How to crimp connectors (the short version)
I'm a procurement guy, not an electrician. What follows is not a certification course — it's the baseline I now demand from anyone terminating wire for our projects:
- Use a ratcheting crimp tool with a die matched to the terminal. Pliers are not a crimp tool. That's not my opinion; it's physics.
- Match the terminal barrel to the wire gauge. The stamp on the barrel exists for a reason.
- Strip to the specified length, insert the conductor fully, and keep insulation out of the barrel.
- Pull test. A good crimp holds metal-to-metal. If the wire releases, it's wrong. No exceptions.
- For real harnesses, outsource them. A shop that terminates hundreds of connections a week will beat a generalist every time. I know my boundary, and I'm okay with it.
The harness shops we use build to IPC/WHMA-A-620, the industry standard for wire harness fabrication. If your contract shop doesn't know that reference, that's a useful red flag.
An honest closing note
Mean Well makes a solid power supply, and I still buy them. I also don't believe any power supply is failure-free, and any vendor who promises "guaranteed zero failures" is overpromising. That is the kind of claim I distrust more than the occasional dead unit.
Since we made these changes, our field failure rate for 5V control supplies has dropped meaningfully. I won't tell you it dropped to zero — anyone who sells you zero is lying — but it dropped enough that the service manager stopped talking about removing Mean Well from the approved list. That's progress I can track in the same cost system I used to track the failures.
The next time a Mean Well 5V 20A power supply "fails" in the field, test the unit first. Then walk back along the output wiring and check every crimp. I'd bet on the connector before I'd bet against the supply.
Leave a Reply