The battery box lid landed six feet away. Acid on the wall. Acid on the ceiling. Tuesday afternoon, routine site visit, a 500 kVA set in a concrete room the size of a parking space. Nobody was hurt. Pure luck.
The charger had been stuck in boost mode since Friday. It cooked two batteries dry over the weekend, and hydrogen had been collecting at the ceiling the whole time. All it needed was one spark. It got one.
Here is the thing most people never connect: your generator’s batteries are chemical plants that happen to start engines. Every hour they sit on a charger, they make gas. Explosive gas. And most battery rooms are designed as if that gas does not exist.
Charging a lead-acid battery splits water. Hydrogen comes off the negative plates, oxygen off the positive ones. On float charge — the gentle 13.5 to 13.8 volts per 12-volt battery your charger should be holding — the amount is small. A trickle. Nothing to panic about.
Boost and equalize modes are different animals. They push voltage well above float, and gas production climbs with it. A charger stuck in boost over a weekend, like the one I found that Tuesday, can dry out cells and fill a small room’s ceiling with a mixture that needs only about 4% hydrogen by volume to ignite. Four percent. That is the lower explosive limit, and hydrogen reaches it fast in a sealed space.
It also lights off sparks so small you would not feel them — roughly an order of magnitude less energy than natural gas needs. A relay contact. A light switch. Static off your sleeve. That is the whole story, and it ends at four percent.
Every vent cap on a battery top is a tiny chimney. When the charger works the cells hard, that is exactly what they become.
Your generator might run fifty hours a year. Its batteries charge 8,760 of them. The charger is the hardest-working device in the whole installation, and it is the one nobody looks at.
I meter float voltage on every site visit now. Two minutes, once a month. The green light on the charger means nothing — I have said that before and I will keep saying it until people believe me. A drifting charger holding 14.5 volts instead of 13.7 does not just shorten battery life. It turns the battery room into a gas generator. (The unglamorous checks that actually catch this are in our generator battery maintenance guide.)
An old-timer I work with disagrees with me on this one. He has kept batteries in a plywood box for thirty years, he says, and nothing ever blew up. He is right about the thirty years. He is wrong about what it proves. Thirty years of luck is not a ventilation design. I told him that. He laughed. Then he moved his battery box away from the furnace-room wall — which tells you everything about how the argument really ended.
This is the mistake I find most often, and it is almost touching in its logic. Somebody thought about ventilation. They cut a vent into the battery room. Down low, near the floor, where vents usually go.
Hydrogen is the lightest gas there is. It does not drift down to your thoughtful floor vent. It rises, hits the ceiling, and pools there in an invisible layer. A low vent in a battery room is decoration.
What works: openings at the highest point of the room, or better, a small exhaust fan up top that runs continuously or switches on with a hydrogen detector. No dead pockets — that dropped ceiling section above the door, the boxed-in beam, the top of the stairwell. Gas collects in exactly the architectural features nobody thinks about. (Room airflow deserves its own treatment; see our generator room ventilation design piece.)
I once walked into a site where they had done everything right except the detector. They had bought a hydrogen detector — good — and mounted it at knee height next to the door, exactly like the carbon monoxide detector in their office. CO weighs roughly what air weighs. Hydrogen does not. That detector would have reported clean air right up until the ceiling layer reached the light switch. We moved it to six inches below the ceiling. The facilities manager looked embarrassed. He should not have been. The diagram in the box showed the thing mounted at waist height.
A vent grille near the ceiling and a detector mounted high — that is where hydrogen actually goes. Anything lower is measuring the wrong air.
VRLA, AGM, gel — the maintenance-free batteries everyone loves. On normal float they recombine almost all the gas internally. Almost. Overcharge them, cook them, or let one slide into thermal runaway, and the pressure valves open. That is the design working as intended. A sealed battery that could never vent would be a pressure vessel with no relief; the valves are the relief.
So the “sealed” label changes the odds, not the physics. A bulging VRLA case is a battery telling you it is venting or about to. I have seen people keep a bulging battery in service because “it’s sealed, it can’t leak gas.” Honestly, that sentence scares me more than the bulge.
And here is a detail that fools people: hydrogen itself is odorless. A cleaner once told me a battery room “smelled funny,” and what she smelled was the sulfur stink of abused cells — not the hydrogen. By the time abused batteries smell, the odorless gas has been there a while. Do not trust your nose. Your nose cannot detect 4% hydrogen. Nothing in your body can.
People picture the batteries sparking. Loose terminal, big arc, boom. That happens — usually while someone is wrenching on a live bank, which is its own special category of mistake. But in my experience the ignition source is almost always something else in the room.
The ATS, for one. Those big contactors throw an arc every time they transfer, and transfer switches share rooms with batteries more often than anyone admits. (Here is how an automatic transfer switch actually works.) The room light switch. A relay inside the charger itself. The engine’s own electrics, if the batteries sit inside the canopy next to the control panel.
The ugliest layout I ever saw: a battery rack directly under the room’s only light fixture. Every time someone flipped the switch to check the batteries, an arc happened inches above the gas layer’s favorite parking spot. I asked the owner how long it had been like that. “Since installation. Twelve years.” Twelve years of flipping a light switch under a hydrogen ceiling. Look, not much in electrical rooms rattles me anymore. That one did.
I am not going to hand you a code book. The codes exist — IEEE 1635 covers ventilation for stationary batteries, and NFPA 110 has things to say about battery installations — and your engineer should read them. But most battery rooms I see were never engineered. They accreted. Somebody put batteries where there was space, and nobody revisited the decision.
So here is the five-minute monthly check I do, the same one I would hand to any facility manager:
Five minutes. A meter most techs already own. That is the whole program.
Dead batteries get all the attention because they cause the failures everyone sees — the no-start on test day, the frantic call. (Troubleshooting a generator that will not start usually ends at the batteries.) Hydrogen gets none of the attention, because its failure mode is a room you cannot walk into anymore. Twelve years around these machines, and the batteries have tried to kill exactly one engine I know of — and it was not by going dead. If you have a battery room, go look at the ceiling. That is where the story is. (Talk to us if you want a second pair of eyes on yours.)
Email：[email protected]
Work：Monday to Friday,9:30-18:30
