The 800 kVA set at the food plant never missed a monthly test. Not once in four years. Then the grid went down on a Friday night, the plant ran on generator power for fourteen hours — and on hour nine, cylinder five quit. When we pulled the head the following week, the exhaust valve face was burned into a crescent and the seat had recessed deep enough to swallow the margin. That valve hadn’t been sealing properly for a long time. Nobody knew. There was no noise. No smoke. No warning of any kind.
The lash on that valve — the tiny gap between the rocker arm and the valve stem — had been set too tight during a “full service” two years earlier. Then the seat wore in a little more, exactly the way seats do, and the gap closed the rest of the way. The valve spent its last year never quite closing. Hot combustion gas leaked past the seat on every power stroke, cutting a groove like a torch. Fourteen hours of real load finished what two years of monthly no-load tests had started.
Here’s the thing most generator owners get backwards: a ticking valve train is annoying, but a silent one can be worse. The tick means there’s clearance. Silence can mean there isn’t any left.
I check valve lash on every diesel I overhaul or commission, and I keep a log of the numbers. Over twelve years the pattern is depressingly consistent: about half the sets I open have lash that’s drifted out of spec, and almost nobody had it on their maintenance plan. This is the most neglected thirty-minute job in the generator business. These are the five mistakes I keep finding.
This is the big one. The belief that the factory set the lash, the factory knew what it was doing, and nothing moves afterward.
Things move. Valve seats recede into the head a few thousandths at a time — normal wear, the seat hammering itself deeper with every closing event. As the seat sinks, the valve sits higher, and the gap at the rocker shrinks. On most diesel designs, lash gets tighter with age, not looser. Tighter, remember, is the quiet killer. There’s no tick to warn you. The first symptom is a burned valve, and by then you’re buying a head job.
An old hand I work with — thirty years on engines, and he’ll tell you so — says he never touched lash in twenty years and nothing ever happened. He’s not lying. His sets were mostly standby units that ran a few hundred hours a decade. Light duty forgives a lot. But put the same neglect on a prime-power set running six thousand hours a year and the math changes. Duty cycle decides whether this mistake costs you nothing or a cylinder head.
The first lash check matters most. New and freshly rebuilt engines bed their seats in during the first few hundred hours. Whatever the factory set will have moved. If your service schedule doesn’t include a lash check at the first oil-change interval after commissioning, add one. It’s the cheapest insurance in the whole manual, and half the manuals bury it in a footnote.
A rental fleet taught me this one. Their in-house tech was conscientious — genuinely one of the better ones — and he’d set the lash on a 350 kVA set right after it came back from a job, engine still warm, “to save time.” The set came back a week later running rough, down on power, popping through the intake.
He’d set the gaps hot. Metal expands. A gap that measures right on a hot engine is too tight on a cold one — and lash specs are written for a cold engine unless the manual says otherwise in big letters. He’d effectively tightened every valve on the set, and the engine told him about it the expensive way.
The other half of this mistake is the stroke. You set lash with the piston at top dead center on the compression stroke, both valves closed, the lifter sitting on the base circle of the cam. Set it on the exhaust stroke — piston also at the top, valves in overlap — and you’re measuring against a cam lobe that’s already starting to lift. Your feeler gauge lies to you. The numbers look fine. They’re fiction.
Honestly, this is a five-minute discipline problem. Cold engine. Bar it over to compression TDC on cylinder one — watch the valves, both closed, timing marks lined up. Work through the firing order. If you’re not sure which stroke you’re on, you’re not ready to touch the lock nuts. Put the wrench down and bar it around again.
Intake and exhaust valves live different lives. The exhaust valve spends every cycle with superheated gas blowing past it. It runs far hotter than the intake, grows longer with heat, and needs more cold clearance to end up at the right hot clearance. The spec card gives two numbers for a reason. They’re not suggestions.
I’ve opened engines where somebody — usually in a hurry, usually at night — set every valve to the intake number. It’s an easy mistake. The intake spec is listed first. The feeler gauge is already in your hand. And the engine will run fine afterward, which is the cruel part. It’ll run fine for a year. The exhaust valves just run a little tight, seal a little worse, burn a little sooner. You won’t connect the failure to that night for a long time, if ever.
Keep the spec card in the logbook. Not in your head — in the logbook, next to the last readings. Intake: one number. Exhaust: a bigger number. Every time.
This is the mistake that looks like maintenance and acts like a cover-up.
A quarry’s 500 kVA set came to me with a lash record that told a story: exhaust lash on cylinder four, adjusted three times in eighteen months, each time “back in spec.” The tick kept coming back. The last tech had finally gotten it quiet by taking the clearance down to the tight end of the range. Technically in spec. Actually a disaster in progress.
The cam lobe for that cylinder was worn nearly round. With the lobe going flat, the valve wasn’t opening fully no matter what the static clearance said. Each “adjustment” had just chased the number while the real problem — lost lift — got worse. We measured lift at the pushrod with a dial indicator: barely half of spec. New cam, new lifters, problem gone. Eighteen months of feeler gauges had fixed nothing.
The lesson cost that quarry a camshaft and my afternoon: lash is a measurement, not a treatment. When a valve keeps drifting out of adjustment, stop adjusting and start asking why. Measure lift. Look at the lobe. A feeler gauge tells you the gap. It can’t tell you the cam is dying underneath it.
Same logic applies to the tick that won’t die. Loose lash ticks, sure — adjust it and move on. But a tick that returns in weeks is telling you about worn rocker bushings, a cupped pushrod seat, a lifter going soft. Adjusting the gap treats the noise. Finding the wear fixes the engine.
The clearance is only half the valve train. The other half is the parts that create it: rocker arms, shafts, bushings, pushrods, lifters. Almost nobody looks at those during a lash check, because the gauge gives you a number, and numbers feel like answers.
Roll every pushrod on a flat surface while you’re in there. A bent pushrod — from an over-rev, a stuck valve, somebody’s pry bar during the last head job — will still let you set “correct” lash at rest, then flex and wander at speed. I found one on a standby set that passed every monthly test and still fell flat on its face under real load. The pushrod had a bow you could barely see until you rolled it across the valve cover rail. Glass-smooth surface, roll, watch for wobble. Ten seconds per rod.
Wiggle the rocker arms sideways on their shaft. Excess play means worn bushings, and worn bushings mean the geometry changes under load — your careful static clearance becomes a rough guess at 1500 rpm. Check the pushrod cups and lifter faces for pitting while you’re at it. None of this takes long. It’s all right there in front of you with the cover off. The cover’s already off. That’s the whole point.
Since I’ve complained this much, here’s the procedure, short. Cold engine — sat overnight, not “cool enough.” Bar to compression TDC on cylinder one, both valves closed. Set intake and exhaust to their own specs with a feeler gauge you can actually feel drag on — not loose, not forced. Snug the lock nut to torque, then re-check, because the nut always moves the setting a touch and everyone pretends it doesn’t. Work the firing order, not just 1-2-3-4-5-6. Write every number down.
That last step is the one that turns a chore into diagnostics. A single reading tells you the gap. Three readings over three years tell you the trend: which valves are tightening, how fast, whether that exhaust valve on cylinder five is trying to tell you something. The food plant’s burned valve had been whispering for two years. Nobody wrote anything down, so nobody heard it.
And while you’re writing things down, pay attention to what the engine’s been telling you elsewhere. Valves that don’t seal leave smoke with a story to tell — learn to read the colors before you blame the fuel. Combustion gas slipping past a bad seat pressurizes the crankcase and washes oil off the cylinder walls, which is how a valve problem becomes an oil consumption problem six months later. A used oil analysis won’t name the valve, but rising iron and fuel dilution will tell you something up top is unhappy. Put the lash check on the maintenance checklist next to the oil service — it belongs there, not in a footnote. And remember that the extra heat has to go somewhere: tight exhaust valves run the whole top end hotter, which is exactly how overheating failures get their start. If your set is due — or overdue — talk to us before it teaches you the lesson itself.
Next time you lift a valve cover and hear that tick, don’t reach for the radio. Reach for the feeler gauges. And if it’s silent in there — really, completely silent on a set that hasn’t been checked in years — that’s the one that should worry you. Silence isn’t health. Sometimes it’s just the gap, gone.
Email：[email protected]
Work：Monday to Friday,9:30-18:30
