Last spring—March 2024, to be exact—I spent a Tuesday afternoon on my knees in the breakroom of our own office, staring up at a row of Legrand under counter lighting fixtures and a dimmer that would not respond. That is not normally part of my job. I'm the quality and compliance manager for a regional lighting and electrical integration company, and my job is to catch mistakes before something gets installed. I review fixture schedules, driver specs, and wiring diagrams—roughly 150 projects a year. Maybe 170; I'd have to check the log. In 2024 I rejected about 9% of first-delivery submittals because specs didn't match the as-built or the drivers were under-specified. But as I knelt on that vinyl floor, it wasn't a vendor's mistake I was chasing. It was our own.
The Legrand under counter lighting decision that looked simple
For context: we moved our office into an older industrial building in January 2024. The breakroom has high ceilings, a long row of upper cabinets, and one overhead light in the center of the room. Stand at the counter after 4 p.m. and your own shadow sits on the work surface. It was a textbook reason to add task lighting under the cabinets, which is why my facilities manager asked me to spec it.
I went back and forth for the better part of two weeks between a simple line-voltage switch and a connected lighting scheme. A plain wall switch has fewer parts and fewer ways to fail. But the breakroom opens onto two corridors and a conference room, and I knew the lights would be left on overnight more often than not. That pushed me toward controls: an occupancy/vacancy sensor, a dimmer, and a schedule. Not because every breakroom should be smart, but because this one had a predictable human problem.
I ended up ordering dimmable Legrand under counter lighting fixtures, a Zigbee-enabled dimmer, a matching accessory switch, an occupancy/vacancy sensor, and one of Legrand's IoT gateways. The fixture selection was easy. Legrand's under counter fixtures have integrated LED drivers, the controls ecosystem uses Zigbee, and one vendor could take responsibility for compatibility. The controls selection was where I almost outsmarted myself.
The Legrand 3-way light switch wiring diagram was in the packet
The breakroom had two entrances, so we needed control from two switch locations: one by the corridor and one by the conference room door. In the electrical world, that usually means a 3-way switch circuit. We specified the Legrand dimmer at the primary location and an accessory switch at the second location, all according to the manufacturer's layout.
The senior electrician on the job was a 25-year veteran. When I handed him the job packet, I saw the Legrand 3-way light switch wiring diagram sitting on top. He barely glanced at it.
“I've wired three-ways since the 90s,” he said. “What are the odds I'd need the diagram on a kitchenette?”
The odds caught up with us. The symptom was strange: the dimmer at the primary location seemed to work, but the accessory switch only worked in certain combinations. Sometimes the under-counter lights would not turn off at all from the conference room door. After an hour of phone troubleshooting, my technician and I opened both boxes and compared everything to that wiring diagram.
I won't reproduce the full schematic here—if you're doing the work, pull the current PDF from Legrand's website rather than trusting my memory. The important lesson is that a smart 3-way dimmer is not a conventional 3-way toggle. The second location is an accessory switch with a different terminal function. If you wire it the way you would wire an old mechanical switch, you end up with a traveler pair reversed and an accessory switch that appears to be dead. That's exactly what happened. He had treated it like a traditional 3-way, and the digital logic in the dimmer never saw the right signal.
There was a second catch. The old switch box was originally wired as a switch loop, which meant there was no neutral in the box. A conventional mechanical switch can operate on a switch loop, but a Zigbee dimmer needs a neutral for its wireless electronics. We had to pull a new neutral from the junction box above the cabinets. That sounds simple on paper; it meant cutting into a finished wall and running a new cable through an old building's stud cavities. (Should mention: the drywall was already closed, so what should have been an afternoon job turned into a two-day job.)
Total cost of the 3-way mistake: about six hours of labor, a patch of drywall, and one very quiet electrician. I was quiet too, because I had approved the work order without checking that the installer had actually reviewed the current diagram.
Zigbee RF and the breakroom dead zone
Once the wiring was corrected, the lights worked from both entrances. But new problem appeared: the occupancy sensor would drop off the Zigbee network, usually around mid-afternoon, and the scheduled 'off' command would randomly fail. My first instinct was to blame the sensor. My second instinct was to blame the gateway.
The numbers said the gateway should have been fine. It sat in the IT closet, roughly 75 feet from the breakroom, which was inside the published coverage range. The commissioning tool showed a solid connection. But my gut said something wasn't adding up. The breakroom sits between a steel stud partition, a metal prep table, and a microwave. And Zigbee RF lives in the 2.4 GHz band, the same congested band used by Wi-Fi and Bluetooth, where metal and microwave ovens cause real attenuation. Datasheet range is measured in open air, not behind a stainless steel counter.
So we ran a simple test: we unplugged the gateway from the IT closet and set it on a shelf in the breakroom. The sensor re-joined the network in about twenty seconds. Zigbee RF is low-power by design, and in a space like a commercial kitchenette, one obstructed path can be the difference between a reliable network and a flaky one.
People often ask me what practical Zigbee uses are in a commercial lighting job. I usually point to three: wireless accessory switches, occupancy/vacancy sensors, and gateway-based scheduling. But the 'wireless' part is not magic. Zigbee uses a mesh, which means a mains-powered device can relay data to the next device. In our breakroom, the sensor had only one realistic path to the gateway, and that path passed through too much RF interference. We solved it by moving the gateway behind the barista station, which gave the sensor a cleaner path, and by adding a second Zigbee-enabled switch that acts as a routing node. The issue has never come back.
What a flickering fixture taught me about LED drivers
The most humbling part happened after the controls were stable. I did what I always tell my clients to do: I watched the fixtures dim from 100% down to 10%. At low dim levels, three of the fixtures had a slight flicker. Not the kind you'd see on a phone camera, but the kind your eyes notice if you look directly at the light. My first assumption was that the Legrand dimmer was not compatible with the fixture driver. But the flicker followed one specific fixture when we swapped it to another location, which told me the problem was inside the fixture.
We took it back to our workbench and opened the driver compartment. Inside was a small LED driver—a metal module that converts line voltage into constant current for the LED board. LEDs are current-controlled devices, so the driver's job is to hold current steady even as the LED heats up and its forward voltage shifts. That is the part a lot of DIY articles miss.
Around that same time, a younger technician asked whether we should just fix the lights by building a simple LED driver circuit. He had been reading tutorials, and the idea sounded reasonable: a rectifier, a capacitor, a resistor, a regulator. A simple LED driver circuit can power an LED on a bench under fixed conditions. But a commercial dimmable fixture has to handle phase-cut dimming, line voltage variation, and temperature changes. If the current is not regulated properly, the LED doesn't necessarily burn out instantly; it just ages faster and flickers at low dim levels. So no, we did not build a driver from spare parts. We replaced the bad driver and sent the failed unit to the manufacturer for analysis.
That failed fixture changed my inspection process. Now we dim every sample fixture to single digits before we approve a product line, not just run it at full brightness. It found a problem that a standard burn-in test would have missed.
The lesson I keep coming back to
I could tell you that the whole retrofit ended neatly, and mostly it did. The Legrand under counter lighting fixtures look great, the Zigbee controls are reliable, and the occupancy sensor has cut down on lights left on overnight. But the extra labor and rework cost us around $1,800. Maybe $1,600; I do not have the work order in front of me. Either way, it was a preventable expense.
What bothers me more than the money is that I fell into the same trap I warn contractors about every week: I assumed that a small job did not need the full verification process. The fixture schedule was correct. The wiring diagram was correct. The product was correct. The gap was between the drawing and the person who installed it. That gap is where quality issues live.
If you're planning a project with Legrand under counter lighting, a Zigbee gateway, or any 3-way smart switch, here is my honest advice: do not skip the current wiring diagram, and do not let anyone tell you they've done it enough times to not need it. Also check the switch boxes for a neutral before you buy a smart dimmer, think about where the Zigbee RF signals actually need to travel, and test LED fixtures at low dim levels before you install them. These sound like small details. They are small details. But in my experience, they are exactly the details that end up costing you a Tuesday afternoon on your knees.