Friday, 4:40 p.m. The load bank had been humming for nine minutes when the smell hit us — that sweet-acrid stink of cooking varnish. Then the smoke, thin and gray, curling out of the alternator vents. The operator killed the breaker. Nine minutes. That’s how long it took a freshly rewound 500 kVA stator to tell us the rewind was garbage.
Here’s the thing nobody tells you when your generator’s alternator dies: the winding isn’t the expensive part. The decision is.
A burned stator winding forces you into one of two camps. Rewind it — strip out the copper, wind new coils, bake it, varnish it, and put the old iron back to work. Or replace the whole alternator with a new one. Rewind shops will tell you rewinding saves you half. Alternator dealers will tell you a rewind is a ticking clock. Both of them are selling something.
I’ve been the guy standing in front of that burned stator more times than I care to count. Let me walk you through how I’d actually make the call.
First: “the winding burned” is a symptom, not a diagnosis. A stator winding is copper coils tucked into the slots of a laminated steel core — the iron part that almost never wears out. Between the copper and the iron sits insulation: slot liners, phase paper, and the varnish that got vacuum-baked into everything when the machine was built. That insulation has a temperature rating — Class F (155°C) or Class H (180°C) on most industrial sets.
When insulation fails, it fails for a reason. Moisture, for one. I’ve written before about what standby dampness does to windings — a megger reading that drops a little every year until one hot afternoon under load, a weak spot arcs over and cooks. Overload does it too: run a 400 kVA set at 450 for a few summers and the varnish ages a decade. Or contamination — oil mist, carbon dust, salt air — that turns insulation into a slow fuse.
So before anyone talks price, you need three answers. One: how bad is the damage — one shorted turn, or a whole phase cooked black? Two: what killed it — because whatever killed it will kill the next winding too. Three: what’s the condition of the iron. Everything else is noise.
The steel core of a stator doesn’t age the way copper does. If the laminations are straight, the core hasn’t been cooked into a hot spot, and the frame and end brackets are intact, the iron is worth keeping. A quality rewind — new Class H insulation system, vacuum pressure impregnation (VPI, where varnish gets pulled into every gap under vacuum instead of just dripped on), proper bracing on the end windings — gives you a machine that’s as good as it was built. Sometimes better, if the original build was a budget job.
And the money is real. On a 250–500 kVA set, a proper rewind usually lands at 30 to 50 percent of a new alternator’s price. Lead time helps too: a good rewind shop turns it around in two to four weeks. A new alternator from a name-brand line can be six to twelve weeks out if it isn’t sitting in a warehouse somewhere. When the outage clock is ticking, that gap matters more than the invoice.
On one of my projects, a hospital set with a 400 kVA alternator — the winding died on a Friday night test, because of course it did — the rewind took nineteen days and cost about forty percent of the replacement quote. That machine has run clean for six years since. Rewinds can absolutely work. The rewind isn’t the problem. The rewinder is.
Now the other side. There are machines I’d never rewind, and the reasons are boring and specific:
That Friday-afternoon 500 kVA I mentioned? The shop had skipped the VPI step and dipped it instead. Under load, a loose end winding vibrated, chafed through the thinner varnish, and arced. Nine minutes. The rewind was cheap because it skipped the expensive parts of the job — which is the part the quote never spells out.
If you remember one thing from this article, make it this: ask for a core-loss test. Also called a loop test — they wrap a few turns of cable around the stator core, push flux through it, and measure the watts per pound the iron burns off as heat. It takes an hour. It tells you whether the iron is still good.
Almost nobody orders it. Why? Because the answer can kill the job. A rewinder who knows the core is cooked would rather not know, and a buyer who wants the cheap option would rather not hear it.
I once watched a loop test on a twenty-year-old core come back forty percent over the manufacturer’s loss figure. The rewinder — an honest one, to his credit — told the owner straight: it’ll run, but it’ll run fifteen degrees hotter than design for the rest of its life, and insulation life roughly halves for every ten degrees you add. The owner replaced the alternator. That test cost a few hundred dollars and saved a five-figure mistake.
Get it in writing: megger and polarization index (PI) readings before and after, surge comparison test on the new coils, hipot at the right level for the voltage class, and the loop test number. If the shop waves any of that off, walk.
Here’s the comparison the way I’d actually run it on a whiteboard.
Rewind: roughly a third to half the price of new. Two to four weeks. You keep your frame, your mounting, your wiring — no mechanical surprises on reinstall. But you inherit whatever life the iron has left, and the warranty is usually one year, parts and labor, from a local shop.
Replace: full price, longer wait, and a day or two of extra fitting work — new alternators don’t always bolt up to old engine adapters without a fuss. In return you get a full factory warranty, a known insulation system, and a machine with zero hours on the clock. If the old set is fifteen years old and this is its second rewind, replacement starts looking less like spending and more like honesty.
The breakpoint I use: if the rewind quote is under half of new, the core tests clean, and the shop does VPI with a documented test package — rewind. If any two of those three fail, replace. And if a dealer tells you rewinding is always a false economy while quoting you a new alternator with a fat margin baked in, now you know why he said it. When you’re staring at a real quote and want a second pair of eyes on the numbers, that’s exactly the kind of call we take.
One more trap, since we’re being honest: if the voltage was wandering before the failure, make sure it was the winding and not the AVR. I’ve seen owners pay for a rewind when the regulator was the patient all along. A dead winding has a smell and a megger reading. A dying AVR has neither.
Forget the brochure talk. When the stator is sitting on the shop floor with the end bells off, there’s one question that sorts the whole mess out: what’s the loop-test reading on the core?
If nobody measured it, nobody actually knows whether the iron is worth saving. And a decision built on “the quote was cheaper” is how you end up smelling varnish on a Friday afternoon, nine minutes into a load bank run, wondering why the cheap fix cost you a second shutdown.
Next time someone tells you a rewind is always the smart money, ask them that one question. The pause before they answer tells you everything you need to know.
Email：[email protected]
Work：Monday to Friday,9:30-18:30
