The Battery Tested Fine. The Generator Still Wouldn’t Crank.

The phone rang at 2:40 on a Tuesday morning. Night guard at a packaging plant outside the city. “The generator won’t start,” he said. Except it would. Sort of. It cranked — that slow, labored groan, like the starter was thinking about whether it really wanted to.

I knew the set. 400 kVA, 24 volts, two batteries in series, both replaced three months earlier. The charger light was green. On paper, everything was fine. On site, the engine turned over like it was swimming in glue.

Here’s the thing nobody tells you about starting problems: a slow crank with a healthy battery is almost never the battery’s fault. It’s the road between the battery and the starter. And that road was stealing volts before the current ever reached the motor.

The Battery Is New. Why Does It Still Crank Like It’s Half Dead?

A starter motor doesn’t read the battery’s label. It doesn’t care that you bought the expensive AGM pair last quarter. It cares about one thing: how many volts actually show up at its own terminals while it’s trying to turn hundreds of kilograms of cold iron.

Every connection between the battery post and the starter is a small resistor. The terminal clamp. The lug crimp. The disconnect switch. The ground strap bolted to the frame. Clean and tight, each one costs you a tenth of a volt or so. Corroded, loose, or bolted over fresh paint — and yes, I’ve seen ground lugs bolted down over a fresh coat of paint — each one can eat half a volt or more.

Add them up. On that packaging plant’s 24V set, I put the meter on the starter terminal during cranking: 16.8 volts. The batteries were delivering 25 at their posts. More than eight volts lost in the wiring. Eight. The starter was getting a fraction of what it needed, so it turned like it was tired. It wasn’t tired. It was starved.

And the ground side is where most people stop looking. They shine up the red cable and ignore the black one. The ground strap from engine block to frame, frame to battery negative — that path carries exactly the same current as the positive side. A corroded ground bolt steals just as happily as a corroded positive terminal. On this set, 2.4 volts of the loss was sitting in one ground strap bolt. It looked tight. It wasn’t.

I pulled the strap, and the faces that were supposed to be metal-to-metal had a gray-green film between them — corrosion you can’t see until you separate the joint. Twenty minutes with a wire brush, both faces shiny, bolt back on, torqued properly. Next crank: 21.5 volts at the starter. The engine caught before I’d finished exhaling. Zero parts. Twenty minutes. The guard made me instant coffee from a kettle that had clearly seen things.

Bigger Batteries Make Bad Connections Hotter, Not Better

This is the question I get most often, and honestly, it’s the most expensive wrong answer in starting systems.

A bigger battery pushes more current through the same bad joint. For a while, the crank sounds better. You declare victory. But the joint is still a resistor, and now it’s a resistor with more current going through it, which means more heat, which means it corrodes faster. You haven’t fixed anything. You’ve put the problem on a payment plan.

I’ve walked into plants that replaced their batteries twice in fourteen months. Two invoices, same slow crank. Nobody touched the lugs, because the lugs “looked fine.” I measured 1.9 volts of drop across a single terminal clamp — and the clamp was warm to the touch after a long crank, which is resistance talking. Warmth is the confession. New batteries didn’t fix the crank. They just made the bad joint the most expensive heater in the building. If your batteries keep dying young, start with how the batteries themselves are maintained — but don’t stop there.

A veteran tech I respect disagrees with me on the fix. His rule: when a set cranks slow, upsize the cable one gauge and move on — thicker copper, problem gone. My rule: the copper was never the problem. A fat cable on a dirty lug is jewelry. The joint is the resistor, not the wire. We argue about this over noodles every few months. He’s right that oversized cable forgives a lot of sins. I’m right that the sin is still there. Pick your side — but at least know which one you’re paying for.

One more lie to retire: the green light on the charger. It tells you the charger is happy, not that current can reach the starter. A charger can hold a battery at a perfect float voltage (the charger deserves its own checkup) while a corroded terminal quietly eats two volts on the way out. Green light, dead crank. I’ve seen it more times than I can count.

Five Minutes With a Cheap Multimeter

You don’t need a lab. You need a multimeter, a helper to hit the crank button, and five minutes. This is the test that ends the guessing.

Set the meter to DC volts. For the positive side: one probe on the battery positive POST — the lead post itself, not the clamp around it — and the other probe on the starter motor’s main terminal. Crank. Read the number while it’s cranking. That’s your drop on the positive side. Measuring at the post instead of the clamp matters, because the clamp-to-post interface is itself a suspect joint. Don’t let it off the hook.

Then the ground side: one probe on the battery negative post, the other on a clean bare spot on the engine block. Crank again. Read it.

Here’s the short version of what the numbers mean:

• Under 0.5V per side — that’s a healthy circuit. If it still won’t crank right, the problem is elsewhere; run the full no-start sequence instead.
• 0.5 to 1V — suspicious. Clean it at the next maintenance window, before it gets worse.
• Over 1V on either side — found your thief. Fix it now, not next quarter.

Keep the whole loop under about 1.5V on a 12V system, under roughly 3V on a 24V set. And do the test on a cold morning if you can — resistance gets worse when it’s cold, and thick cold oil makes the starter work harder for every amp it draws. (Winter starting and cold oil are the other half of this story.)

No helper? There’s a quick-and-dirty version. Clamp a set of jumper cables directly from the battery posts to the starter terminals, bypassing the whole cable run, and crank. If it suddenly spins like new, the cable run is guilty and the battery is innocent. Takes two minutes. I’ve condemned more cable runs with borrowed jumper cables than with any fancy tool.

The fix is boring, which is why people skip it. Take the joint apart. Wire-brush both faces until they’re actually shiny — the mating surfaces, not just the outside. Put it back together and torque it to spec. Lead battery posts crack if you gorilla them; I split one in 2019 and paid for the replacement battery out of my own embarrassment, so I use a torque wrench now. Then — and only then — smear a thin coat of dielectric grease or petroleum jelly over the outside of the joint to keep air and moisture out. Grease goes on top after tightening, not between the faces. Between the faces it’s an insulator wearing a disguise.

And check the ground strap from engine to frame while you’re there. It’s the joint everyone forgets, and it carries every amp the starter draws.

So the next time your set cranks slow and the batteries test fine, don’t order new batteries first. Grab the meter. The battery is innocent until the meter says otherwise. And if you want a second opinion on the whole starting chain, talk to us — we’ve chased this particular thief through a lot of engine rooms.

That packaging plant’s guard still calls me when the monthly test runs. The set cranks like it means it now. He still makes terrible instant coffee. I drink it anyway. Some fixes you remember because of the machine. That one I remember because of the kettle.