The piston came out of that 500 kVA engine looking like a candle left too close to a flame. The center of the crown had sagged. The edges were still sharp. And the exhaust valves — I held one up for the site manager — had been pulled down into little trumpets. Tuliped, the old mechanics call it. I told him the words nobody wants to hear: this didn’t happen last week. This happened over months. Right under your nose.
Here’s the thing. That generator had an exhaust gas temperature gauge. A pyrometer, thermocouple screwed into the manifold, needle sitting right there on the panel. For two years it had been pinned past the red zone every time the set ran hard, and nobody wrote the number down. Nobody asked what it meant. The engine was screaming in the only language it has, and the room was empty.
Exhaust gas temperature — EGT, if you want the shop shorthand — is just leftover energy. Fuel that finished burning in the cylinder leaves as power. Fuel that finished burning late leaves as heat in the exhaust. So the number on that gauge is really a confession: how cleanly is this engine turning diesel into work?
A healthy engine at a healthy load has a steady EGT. It climbs as load climbs — that’s normal, that’s physics. What isn’t normal is the number creeping up month after month at the same load. Or one reading that sits way above where the engine maker says it should. Roughly speaking, continuous turbo-inlet EGT on most industrial diesels lives somewhere in the 600°C neighborhood. Your engine’s datasheet is the boss here, not my memory. But whatever the ceiling is, the gauge exists to tell you when you’re living near it.
I’ve started telling customers: oil pressure tells you the engine is alive. EGT tells you whether it’s happy.
High EGT always means the same thing underneath — too much heat leaving through the exhaust instead of becoming crankshaft work. The causes are a short, honest list:
Notice what’s not on the list: bad luck. EGT doesn’t drift for no reason.
This is the part most people miss. If every cylinder reads hot, you’ve got a system problem — air, timing, overload, the whole engine breathing wrong. But if one cylinder reads hot and the rest are fine? That’s not the engine. That’s that cylinder.
I carry an infrared thermometer for exactly this. Shoot each exhaust port at the manifold and you get a per-cylinder EGT map in about ninety seconds. On a six-cylinder set last year, five ports read in a tight band and number four sat roughly 130°C above the rest. Swapped the injector with number two. The hot spot moved with the injector. Ten minutes, one wrench, diagnosis done. A $400 injector, caught before it melted a $4,000 piston.
Old-timers did this with a wet finger on the manifold — the port that sizzled fastest was the hot one. Don’t do that. Buy the IR gun. They’re cheap now, and your fingerprints will thank you.
A pyrometer before the turbocharger and one after it are reading two different realities. The turbo eats a huge amount of heat — that’s literally what drives it — so a post-turbo probe reads roughly 100 to 200°C lower than a pre-turbo one on the same engine. Neither is wrong. But you can’t compare your post-turbo number against someone’s pre-turbo spec and conclude anything.
This is why I tell people the trend matters more than the absolute. Log the number at the same load, same ambient, once a month. A reading that climbs 50 degrees over a year at identical load is the engine tapping you on the shoulder. A single scary number with no history is just a number.
Set the alarm from the datasheet, not from vibes. And wire it to something someone actually looks at — an unlogged alarm is decoration.
Honest answer: not always. A 100 kVA standby that runs fifty hours a year on test cycles? The money probably goes further on the basics — fuel, batteries, coolant. I’m not going to sell you a gauge for a machine that barely runs.
But prime-power sets, big standby sets, rental fleet units, anything where the engine actually works for a living — yes. And here’s the part people forget: during load-bank testing and commissioning, even a temporary probe earns its keep. It’s the fastest way to confirm the set is truly taking rated load and not just spinning impressively. I clamp a probe on for every commissioning I supervise. Takes twenty minutes. Has saved me from signing off on two underperforming sets that “sounded fine.”
Sounded fine. That’s what the site manager said about the 500 kVA with the melted piston, too.
That rebuild cost him an in-frame overhaul, a new set of valves, and three weeks of rental power. The pyrometer that would have warned him costs less than the crane they used to lift the cylinder head. I still think about that dead needle every time I walk past a panel. Write the number down. It’s the cheapest engine insurance you’ll ever buy. And if you need a hand speccing the right set for the load in the first place, talk to us — we’d rather sell you the right generator once than sell you parts forever.
Email：[email protected]
Work：Monday to Friday,9:30-18:30
