The Lightning Never Touched Your Generator. It Still Killed It.

The storm rolled through at 2:40 in the morning. The utility feeder dropped at 2:43. The generator started at 2:44 and carried the building through the night on diesel without a hiccup. Then 7:15 a.m. — grid power restored, transfer switch back on utility, and the controller screen went dark and never came back. The lightning had hit a pole transformer two streets over. It never touched our generator. It didn’t need to.

I’ve watched this exact morning play out about a dozen times in twelve years. The generator did its job perfectly — engine fine, alternator fine, fuel fine. What died was the electronics: the controller board, the monitoring module, sometimes the battery charger. And here’s the part that stings — every one of those dead boards died for the same four reasons. Four mistakes I keep finding, site after site, in rooms where the owner is sure he’s protected.

Surges don’t need a direct strike. A lightning hit anywhere on the local distribution network — a pole, a transformer, the warehouse next door — throws a voltage spike down every conductor connected to that network. Your generator is connected to that network. So are its control wires, its charger, its remote monitoring modem. The copper doesn’t care that the thunder was two streets away.

Mistake One: “We Grounded It, So We’re Protected”

This is the most common one, and the most stubborn. The owner points at the ground rods, the fat green conductors, the test report showing under 25 ohms to earth. I wrote a whole guide on getting generator grounding right, and I stand by every word of it — grounding matters. But grounding is not surge protection.

A ground rod gives fault current somewhere to go. It does nothing about the 6,000-volt spike riding in on the utility feeders, or the induced voltage on the control wiring when the sky lights up. Think of it this way: grounding is the drain in the floor. A surge is the flood coming through the window. The drain helps afterward. It doesn’t stop the water.

I’ve argued this one with facility managers holding beautiful grounding reports next to dead controllers. The ground test passed. The surge didn’t read the report.

Mistake Two: One SPD at the Main Panel, Nothing at the Controller

Plenty of sites do install a surge protective device — one big unit at the main service panel — and stop there. Meanwhile the controller sits twenty meters of wire away, full of semiconductors that die at voltages a light bulb would shrug off.

Here’s the physics the single-SPD crowd misses. An SPD clamps voltage, but the wire between the SPD and the controller has inductance, and inductance plus a fast-rising spike means voltage still shows up at the controller end. The fix the industry actually uses is cascaded protection — a heavy Type 1 or Type 2 unit where power enters the building, and a smaller, faster Type 3 unit right inside the controller cabinet. The big one takes the punch. The small one cleans up what gets through.

Your controller is the most expensive fragile thing on the entire set — I break down what’s actually inside one in my controller guide. A DIN-rail Type 3 SPD for that cabinet costs a small fraction of one replacement board and installs in the time it takes to drink a coffee. I’ve never once seen that math go the other way.

Mistake Three: Power Lines Protected, Signal Lines Naked

This one kills more monitoring modules than anything else I see. The power conductors get their SPD. The thin little signal wires get nothing — the RS485 pair to the remote monitoring module, the two-wire start signal from the ATS, the antenna cable on the cell modem. And those thin wires are better antennas. A nearby strike induces voltage on them even when nothing is directly connected.

Two summers ago I kept getting called back to the same site — a logistics warehouse — because their remote monitoring module “died” every thunderstorm season. The module itself was fine each time. Its RS485 port was fried, three years running. A small signal-line protector on that pair would have cost less than one of my three service visits. The third time, I put one in and billed them for the lesson. No fourth visit.

While you’re at it: those same unprotected lines are the ones a fault — or an attacker — can ride in on, which is why my piece on monitoring cybersecurity starts with physical-layer protection. A surge and a hacker both prefer the wire nobody thought about.

Mistake Four: Installed Once, Never Looked At Again

Surge protectors wear out. The metal-oxide varistors inside degrade a little with every surge they absorb — that’s their job, dying a little so your controller doesn’t die a lot. Most units have a status window or a flag: green means alive, red means the cartridge is spent.

And nobody looks at the window. I open control cabinets where the SPD has been showing red for what the dust suggests is years. An SPD with a dead cartridge protects exactly as well as no SPD at all — which is to say, not at all. It’s the same indicator-light syndrome as a charger glowing green over dying batteries: the little signal says fine, so everyone assumes fine, and the assumption is the whole failure.

Checking takes five seconds. Glance at the windows when you do your other rounds. When a cartridge flags red, swap it — they’re made to be replaced, not mourned.

What a Set That’s Actually Protected Looks Like

Three levels, bonded together, checked on a schedule. A heavy Type 1 or Type 2 SPD where power enters the building and at the generator’s output breaker. A smaller Type 3 unit inside the controller cabinet, leads kept short and straight — every extra inch of lead wire adds inductance and steals clamping performance. Signal-line protectors on every copper pair that leaves the cabinet: the RS485, the start-signal wires, the modem line.

Bond all the SPD grounds into the same grounding system — that’s where the grounding work I linked above earns its keep. And write the quarterly window check into your maintenance checklist, right next to the oil and the filters. Five seconds, four times a year.

The engine on that 2:43 a.m. set ran flawlessly all night. Twelve years in this business and the machines still surprise me — but they surprise me in the same four ways, over and over. A modest amount of protection, in the right three places, checked four times a year. That’s the whole difference between a controller that survives the storm and a very quiet Monday morning.

If you’d like someone to look at what’s actually inside your control cabinet before the next storm season, our team does exactly that kind of survey — you can reach us through our contact page.