The first time I stood next to a 2-megawatt diesel generator at commissioning, I did what every engineer does. I looked for the starter motor.
Couldn’t find one. No big cylindrical motor bolted to the flywheel housing, no fat battery cables running to it. Just clean engine block, a row of tall red bottles along the wall, and a compressor humming in the corner. The commissioning engineer caught me staring and grinned. “It’s already started twice today,” he said. “On air.”
That was twelve years ago, and I still remember the sound. Not the whine of a starter — a deep, percussive whump-whump-whump as compressed air punched the pistons down and the big engine coughed itself awake. If you’ve only ever worked with standby sets under a megawatt, this might be new territory. Past a certain size, diesel engines don’t crank on batteries anymore. They start on air.
Nothing wrong with electric starting. It’s just physics getting rude. Turning over a 2 MW engine means shoving pistons the size of dinner plates through compression. The starter motor — if you tried to fit one — would need to be enormous, and the battery bank feeding it would look like a small substation.
People do the math anyway. I’ve seen quotes where someone specced electric start on a 1.5 MW set to “keep it simple.” Four 24-volt strings, a charger the size of a suitcase, cables thick as your wrist. It works. Barely. And then winter arrives, the batteries lose a third of their cranking power overnight, and the set that tested fine in September won’t turn over in January. Batteries starve quietly — and on a big engine, they starve expensively.
Honestly? Below about a megawatt, electric start is fine and I’d never argue otherwise. But somewhere past that line, compressed air stops being exotic and starts being the obvious answer. Air doesn’t care about temperature. A bottle of air at 30 bar in January has the same punch as in July. Try saying that about a lead-acid battery.
An air-start system looks complicated on a piping diagram. In the room, it’s three things.
First, the air receivers — those tall bottles along the wall. They store starting air at around 30 bar, roughly 430 psi. The sizing rule is simple: enough stored air for several consecutive start attempts without the compressor running, because every serious spec assumes the compressor might be the thing that’s broken. On one project I worked on, the requirement was six back-to-back starts from a single charge. We tested it. The engine caught on the second.
Second, the compressor. Usually a small electric unit that tops the bottles up automatically, kicking in when pressure sags. On critical sets you’ll find two — duty and standby — because a dead compressor on a Tuesday becomes a dead generator on the Thursday the grid fails. I check compressor run hours the way other people check oil. A compressor that never runs is either perfect or broken, and you can’t tell which until you need it.
Third, the starting valves and the air distributor on the engine itself. When the controller calls for a start, the distributor routes high-pressure air into each cylinder in firing order, shoving pistons down until the engine spins fast enough for fuel injection to take over. Then the air cuts out and the engine runs on diesel like any other. The whole handoff takes seconds. It’s violent, it’s loud, and it’s beautiful.
Here’s the thing nobody tells you at handover. Compressed air systems fail in exactly two boring ways, and both are maintenance, not engineering.
Water first. Compressors squeeze moisture out of intake air, and that water collects in the receivers. Leave it and you get corrosion inside the bottles, sticky starting valves, and — my personal favorite failure — a distributor valve that works perfectly in the monthly test and jams the one morning it matters. Every receiver has a drain. Drain it. On a schedule, not when someone remembers. Automatic drain traps exist and they’re worth every penny, but I still tell customers to crack the manual drain monthly and look at what comes out. Clear water, fine. Rusty sludge, you’ve got a conversation coming.
Leaks second. Air systems leak the way old houses draft — slowly, invisibly, constantly. A weeping fitting here, a tired O-ring there, and over a week your 30 bar becomes 22, then 18. Below a certain pressure the engine still turns, just too slowly to catch, and you get the saddest sound in backup power: a big engine wheezing through start attempts it can’t finish. I once found a set whose receivers had bled down to nothing over a holiday shutdown. The fix was a five-dollar fitting. The embarrassment was free.
There’s a third, rarer one: oil carryover. A worn compressor starts passing lubricating oil into the air stream, and now your “clean” starting air is an oil mist that gums up valves. If the air coming out of the drain smells like a crankcase, the compressor needs attention, not the engine.
Big engines have a starting ritual, and the good operators treat it like religion.
Check the receiver pressure first. Not “the gauge was fine last week” — look at it now. Then the compressor: is it in auto, and when did it last run? Then drain the receivers. Then — and this is the part electric-start people forget — bar the engine over by hand or with the turning gear if it’s been sitting. A big engine that’s sat for months can have oil drained off the liners and, in the worst case, a cylinder that’s collected condensation. Starting air doesn’t forgive a hydraulic lock the way a small starter would just stall. Air will try to punch through it. Don’t let it.
Pre-lubrication matters too. Most large sets run a pre-lube pump before cranking, pushing oil to the bearings so the first revolution isn’t metal on metal. If your set has one, let it finish its cycle. The thirty seconds feel long. They’re cheap insurance.
I learned the barring-gear lesson the embarrassing way, early in my career, on a set that had sat through a long shutdown. Skipped the manual turn because we were behind schedule. The engine started — and made a noise I still hear in my sleep. Nothing broke, thankfully. A scored liner would have been a very expensive education. Now I bar first, every time, no exceptions. The schedule can wait. Liners can’t.
Honest answer: it depends on the size of the engine and the temperature of your winters.
Under a megawatt, electric start wins on simplicity and cost. Maintain the batteries — especially before cold weather — and you’ll be fine. Between one and two megawatts, it’s a judgment call; I’ve seen both work. Past two, air start is the default for a reason, and fighting it usually means fighting physics.
One more consideration people miss: long-term standby. A set that sits for months — properly preserved — keeps its air charge far longer than batteries keep theirs. Batteries self-discharge; air just sits there, waiting. When that set finally gets its start command after eighteen quiet months, I’d rather bet on the bottle than the battery.
And when you’re writing the commissioning tests — do them properly — include a start with the compressor isolated. Prove the receivers alone can do the job. That’s the test that finds the slow leak, the tired compressor, the drain nobody opened. It’s a ten-minute test. The alternative is finding out at 2 a.m., in the dark, with the grid down and a room full of people asking you why the big red engine won’t wake up.
Air start isn’t glamorous. Nobody puts it on the brochure cover. But the next time you walk past a two-megawatt set and wonder where the starter motor went — now you know. It’s not missing. It was never needed.
Email：[email protected]
Work：Monday to Friday,9:30-18:30
