Your Generator’s Hour Meter Is Probably Lying. The Filters Aren’t.

“A five-year-old 500 kVA, three thousand hours, half the price of new. Should I send the deposit today?”

That was the call. Tuesday afternoon, a buyer I’d never met, excited about the deal of the year. I told him what I tell everyone: don’t wire anything until somebody who isn’t selling you the machine has looked at it. He paid for my fuel, so the next morning I drove two hours to a laydown yard to look at three thousand hours.

The set was clean. Fresh paint on the canopy. The controller display said 3,012 hours. The seller smiled like a man holding four aces.

I popped the side door and pulled the oil filter. Date code stamped on the can: 2019. The coolant filter next to it was stamped the same year. The canopy latches were worn silver where hands had grabbed them thousands of times, and the battery cables were dated 2018.

Three thousand hours, my foot. That machine had lived a whole life the meter knew nothing about.

The number everybody trusts — and what it’s supposed to mean

I get it. The hour meter is the one number that feels objective. No opinions, no marketing. Just a count.

Buyers use it three ways. First, for maintenance scheduling — oil and filters typically every 250 to 500 hours, valve checks at the big milestones, the whole routine that keeps a standby set alive. Second, as a wear estimate — a set with 2,000 hours should have more life left than one with 18,000, and the price reflects it. Third, as a resale argument: “low hours” is the most profitable phrase in the used generator business.

All of that works. If the number is real. If the number means what you think it means.

Those are two big ifs.

What an hour meter actually counts

Here’s the thing most buyers never think about: the meter counts engine running time. That’s it. It doesn’t know what the engine was doing during those hours.

An hour at 80% load and an hour at 10% load both add exactly one to the count. But the first hour wears the engine honestly. The second one lets it rot from the inside — wet stacking, cylinder glazing, carbon packing into the turbo. A 3,000-hour set that idled its whole life can be in worse shape than a 10,000-hour set that worked for a living.

The meter also doesn’t count starts. And starts are brutal. Every cold start is the hardest ten seconds of an engine’s week — dry bearings, oil pressure building from zero, the starter dragging the crankshaft around. A standby set that started 400 times and ran 500 hours has lived harder than its hours suggest. The meter never mentions the starts.

Three kinds of hour meters, three kinds of trust

Quick field guide. The old mechanical counters with the rolling number drums — vibration and moisture kill them, and when they die somebody fits a new one that starts at zero. The standalone digital meters — battery-backed, and they quietly stop counting when the internal battery gives up years before the engine does. And the controller-integrated count in modern Deep Sea or ComAp panels — survives everything except a controller swap or five minutes with the configuration software.

Three designs. Three different ways for the number to part company with reality.

Five ways the number lies to you

I don’t mean the meter is broken. I mean the number on the display and the truth about the machine are often two different things.

1. The meter got replaced. When a mechanical counter dies, the new one starts at zero. Nobody engraves the old reading on the canopy. I’ve seen “1,200-hour” sets with ten-year-old radiators.
2. The digital number got reset. The hour count lives in controller software, and with a laptop it can be changed in about five minutes. I’m not saying every low-hour set was tampered with. I’m saying the capability sits right there in the service menu, and the used market runs on trust.
3. The controller got replaced, and the hours went with it. Controllers die from lightning, from charger spikes, from old age. A replacement starts at zero hours. The engine doesn’t. Not dishonest — just how the hardware works. But the buyer reading the display doesn’t know the controller is two years younger than the engine.
4. The meter was wired wrong. I’ve found meters wired to key-switch power instead of an oil-pressure or run signal. Key on, meter runs — engine silent. A tech doing weekly inspections with the key on for ten minutes each visit adds roughly nine hours a year of nothing. Over a decade, it adds up.
5. The number is honest and still useless. I once inspected a set with 1,800 genuine hours. Ninety percent were unloaded exercise runs — twenty minutes a month, no load bank, for eight years. The exhaust stack was wet with slobber, the turbo was coked, the injectors were weeping. Eighteen hundred honest hours of slow murder.

What I check instead of the display

I’m not saying ignore the meter. I’m saying it’s the start of the investigation, not the verdict. Here’s my actual routine when somebody asks me to bless a used set.

I read the date codes. Every spin-on filter — oil, fuel, coolant — carries a manufacturing date code. Filters get changed with the oil, so the dates tell you when the last real service happened. A “3,000-hour” set wearing 2019 filters hasn’t been properly serviced since 2019, and the hours between then and now are anybody’s guess. If the seller lets me, I cut one oil filter open. The pleats tell a story no display can.

I pull an oil sample. Better yet, I ask for the trend — three samples over time showing iron climbing, fuel dilution, coolant sneaking in. One sample is a rumor. Three samples are a biography.

I check the batteries’ install dates. Batteries get dated when they’re fitted. If the “3,000-hour” set is on its third set of batteries with 2021 date codes, that machine has been starting and sitting a lot longer than the meter admits.

And I insist on a load bank test. Always. No meter can tell you the set makes rated power, holds voltage when a motor starts, or survives two hours at full load without the coolant temperature climbing. A load bank can. It’s the closest thing to truth in this business.

My rule, after twelve years of looking at other people’s generators

An old-timer I used to work with swore by hour meters. “Hours are hours,” he’d say. “You can’t argue with a number.”

You absolutely can argue with a number. You just have to know what the number leaves out.

My rule now: hours times load factor times start count. A set with 5,000 hours at 70% load and 200 starts is a healthier machine than one with 2,000 hours at 15% load and 900 starts. The meter gives you one of those three numbers. The other two, you dig for.

So the next time a seller shows you a display reading like it’s a certificate of health, smile, nod — and then open the side door. Read the filter dates. Smell the oil. Ask when it last saw a load bank. And if you’re looking at a used set and want a second pair of eyes before the deposit goes out, that’s exactly what we’re here for — talk to us before you buy, not after.

The meter gives you a number. The machine gives you the truth. Read the machine.