The Engine Runs Fine. There’s No Power.

Tuesday, 11:40 p.m. The foreman’s voice on the phone had that particular edge people get when the concrete trucks are already rolling and the batch plant is dark. “Generator’s running,” he said. “Sounds perfect. But there’s no power. Nothing.” I drove forty minutes out there expecting a burned-out alternator. It took me four minutes to find the real problem, and I never opened a single panel to do it.

That call is the whole story of no-voltage faults in one night. The engine runs fine. The set sounds happy. The meters read zero. And almost everybody — including some technicians who should know better — starts the diagnosis at the wrong end and spends the customer’s money finding their way back.

Here’s the thing. A diesel generator that cranks, starts, and runs at rated speed but makes no electricity is almost never suffering from one mysterious disease. It’s suffering from one of about six ordinary ones. The trick is checking them in an order that costs you nothing until it has to.

Measure at the alternator, not at the panel

Mistake number one is where you put your meter probes. People measure at the distribution panel — or worse, they plug a drill into a socket at the far end of a site cable and declare the generator dead. Don’t. Put your probes on the alternator’s output terminals first: L1, L2, L3 to neutral, right where the power leaves the machine.

Why there? Because everything between the alternator and that drill is somebody else’s problem: breakers, cables, connections, a site electrician’s creative wiring. If you’ve got 400 volts at the terminals and nothing at the panel, the generator is innocent. Walk downstream. I’ve seen grown men price a replacement AVR because a site cable got sliced by an excavator. The alternator was fine. The cable was in two pieces.

And the breaker. Check the main breaker before you touch anything clever. Tripped, open, or switched off for maintenance and never switched back — I keep a mental tally, and the breaker accounts for more “dead generator” calls than any single electrical component on the set. That 11:40 p.m. batch plant? Main breaker tripped on a faulted site cable. I reset it, it held, the plant lit up. Four minutes. The foreman looked like he wanted to laugh and hit something at the same time.

Honestly, this is the step people skip because it feels too simple to be the answer. That’s exactly why it’s usually the answer.

Zero volts at the terminals means the excitation gave up

Okay. Breaker closed, terminals still dead. Now you’re allowed to think about the alternator itself — but only about one small corner of it.

A synchronous alternator doesn’t make voltage by itself. It needs a small DC current fed into its field winding — the excitation — to build the magnetic field that the spinning rotor turns into AC power. No field current, no voltage. It’s that direct. Most no-voltage faults are excitation faults, and excitation faults are nearly all cheap.

Work this little chain in order. First, the fuses. The automatic voltage regulator — the little box that meters out field current — gets its power and its sensing voltage through fuses. A blown sensing fuse blinds the AVR; a blown power fuse starves it. Either way the output is zero volts and the part that failed costs less than a sandwich. Pull them, look at them, meter them. Don’t eyeball a fuse and call it good. I’ve been burned by that.

Second, the field circuit. With the set stopped and locked out, check continuity through the field winding and the exciter. An open field winding is rare, but a broken wire at a terminal, a connector shaken loose by vibration, a brush worn to nothing on a brushed exciter — those are Tuesday-afternoon failures. Vibration is the quiet killer of every connection on a generator. Treat every terminal as a suspect.

Third — and this is the one that gets the sets that sat too long — residual magnetism. Here’s a piece of physics the brochure never mentions: the AVR needs a tiny bit of leftover magnetism in the rotor iron to bootstrap itself. A few volts, just enough to start the feedback loop. Let a set sit six or eight months, especially after a shock load or a short circuit at the end of its last run, and that leftover magnetism can fade to nothing. The set starts, runs beautifully, and makes exactly zero volts. Forever. Until someone fixes it in about ninety seconds.

The ninety-second fix everyone is afraid to try

Field flashing. The name sounds more dramatic than the job. You give the field winding a brief kiss of DC from a battery — twelve volts, two or three seconds — to re-magnetize the rotor iron. Then the AVR can see something, the loop starts, and the voltage climbs back to normal like nothing happened.

People treat this like open-heart surgery. It isn’t. But there are two rules, and the second one is where the money gets lost. Rule one: disconnect the AVR’s field leads first. You do not want battery current shoved backward into the regulator. Mark the leads — F+ and F-, or X and XX on some sets — take a photo with your phone, then disconnect. Rule two: mind the polarity. Flash it backward and you can wipe out what little magnetism was left, or sulk the AVR into a fault state. Positive to F+, negative to F-, two seconds, done. Reconnect, start, watch the meters.

I keep a pair of test leads with alligator clips in my bag for exactly this. A rental 200 kVA came back from a festival season last year, sat in the yard through winter, and showed up on a spring job making zero volts. The depot crew had already priced a new AVR. I flashed the field in the parking lot. Voltage came up to 400 before I’d finished coiling the leads. Parts cost: nothing. That set ran the whole job without another hiccup.

That said — if you’re flashing the field every month, something else is wrong. Losing residual after long storage is normal. Losing it chronically means the AVR’s build-up circuit is tired or a diode is leaking in the exciter, and then you’re into real component diagnosis instead of a party trick.

The five ways people turn a cheap fix into an expensive one

This is the part the old hands at the depot argue with me about. Their rule is simple: no voltage, swap the AVR. “It’s always the AVR,” one of them tells every apprentice. My logbook disagrees. Last year I wrote up twenty-three no-voltage calls. Eleven were excitation components — AVR, diodes, sensing-circuit fuses. Twelve were breakers, wiring, loose connections, and lost residual magnetism. So no, it isn’t always the AVR. Barely half the time. Here’s how the other half of the money gets wasted.

Buying the AVR before measuring anything. A decent AVR costs real money, and half the ones that get swapped were fine. Measure the field circuit first. If the field winding is open, a new AVR won’t fix it. If a fuse is blown, a new AVR won’t fix it. The meter is free. The part isn’t.

Flashing the field with the AVR still connected. I said it above and I’ll say it again, because I’ve seen the aftermath: a regulator with its output stage cooked by a direct battery hit. Now you genuinely need that new AVR, plus the embarrassment of explaining why.

Ignoring the rotating diodes. On a brushless set, the exciter’s AC output gets rectified by diodes spinning on the rotor shaft. One shorted diode drags the whole excitation down — voltage collapses under load, or never builds at all. Ten minutes to check with the set stopped, if you know where the diode assembly lives. Nobody checks. Everybody buys the AVR first.

Forgetting the speed. This one’s sneaky. The AVR can only regulate voltage if the engine is actually at rated speed. A set lugging at 1,350 rpm instead of 1,500 because of a fuel or air problem will show low voltage that no regulator on earth can fix. Glance at the frequency meter before you touch anything electrical. Fifty hertz means the engine is doing its job. Forty-five means your problem is fuel, air, or the governor — not the alternator.

The wiggle-and-walk-away. Voltage comes back after you jiggled a connector, so you pack up and leave. The loose connection is still loose. It will fail again at 2 a.m., in the rain, when the customer needs the set most — and then it’s a mystery shutdown instead of a five-minute fix. Find it, clean it, tighten it. If the terminal is heat-damaged, replace it. A fix you can’t explain isn’t a fix.

When the alternator really is dead

Sometimes it isn’t cheap. Shorted windings, a cooked rotor, insulation broken down to the frame — it happens, usually after years of heat, moisture, or a failed space heater in a damp room. But here’s my rule: you don’t get to condemn the alternator until the excitation chain has been proven innocent, one link at a time, with the meter.

The confirmation is a megger test on the windings — insulation resistance, the same check you’d run on any suspect machine. Windings read clean and the field circuit is continuous? The alternator is almost certainly fine, and your fault is still hiding in the excitation. Insulation failed? Then — and only then — you’re talking about a rewind or a replacement, and that’s a capital decision, not a field repair.

Either way, you got there by elimination, not by guessing with the parts catalog. So the next time a set runs beautifully and makes nothing: start at the breaker, work the excitation chain, and keep your hand off the parts shelf until the meter tells you which part earned it. Most of the time the generator isn’t broken. It’s just waiting for someone to check the simple things first. The batch plant foreman still tells that story, by the way — usually right after reminding me I drove forty minutes for a four-minute fix. If your set is pulling this trick right now and you’d rather have experienced eyes on it, get in touch — we’ve seen this movie before.