Electric water heater tripping the GFCI

Your electric water heater trips the GFCI the second it calls for heat, and resetting it buys you maybe an hour before it goes again. That is not the same failure as a plain breaker that trips on load. A GFCI is watching for current sneaking off to ground, and it acts on an imbalance of just 5 milliamperes. Something in that tank is leaking a whisper of current, and the fix starts with finding where.

GFCI trip is not a breaker trip

People use the words loosely, so start here. A standard circuit breaker is an amp counter. It trips when the current through it climbs past its rating, say 30 amps, and holds there long enough to overheat. Two elements firing at once, a shorted wire, a seized part pulling locked-rotor current: that is overload territory, and that is the story in the reset-keeps-tripping guide.

A GFCI does something else entirely. It compares the current going out on the hot leg to the current coming back on the neutral. On a healthy circuit those match to the milliamp. When they do not, current is escaping the intended loop, which almost always means it found a path to ground, through the tank, through the water, or through a person. The instant that leak passes 5 mA, the GFCI opens. It does not care about total amps. A heater can be drawing well under its rated load and still trip a GFCI, because the trigger is the leak, not the size.

Why does a water heater sit on a GFCI at all when most have run on plain breakers for decades? Newer code sections have pushed ground-fault protection onto more 240V appliance circuits, and any heater installed in a garage, basement, or crawlspace where water and bare concrete meet is a prime candidate. So the protection is doing exactly its job when it trips. The nuisance is real, but the device is honest: there is a leak, and these are the usual sources, ranked by how often they turn out to be the culprit.

Moisture: the leak most people miss

Water and electricity make a ground fault the easy way. It does not take a puddle. A film of damp mineral residue laid across two terminals inside the top wiring box gives current a low-resistance detour straight to the grounded tank shell, and 5 mA is a tiny amount of current to lose that way. The GFCI reads the imbalance and opens.

Where does the water come from? A weeping gasket under a heating element is the classic source. The element screws or bolts into the tank through a rubber or fiber gasket, and once that gasket ages or the element loosens, a slow seep runs down the element flange and pools around the terminals right where the live connections sit. Condensation is the other one, common on a heater in an unconditioned garage or a damp crawlspace, where warm tank surfaces sweat and drip into the junction box below. Even a roof or plumbing leak overhead that finds the top of the tank will do it.

The test is blunt and it is free. Cut the power, open the top cover and both element panels, and look. Water beads, rust trails, white crusty film across terminals, a corroded green ground screw: any of these is your answer. Dry it all thoroughly, a hair dryer on low speeds this up, reseat or replace a leaking gasket, and reset the GFCI. Watch it through a few full heat cycles. If it holds, you were chasing water, not a dead part. If it trips again dry, move to the elements.

The cracked element: current straight to the tank

A heating element is a coil of resistance wire sealed inside a metal sheath. Water surrounds the sheath but never touches the coil, that is the whole design. Heat cycling expands and contracts the metal thousands of times, and sediment packed around the lower element makes it run hotter than it should. Over time a hairline crack opens in the sheath. Water seeps in, reaches the live coil, and now current has a direct road to the grounded tank. On a plain breaker that might trip only on a hard short. On a GFCI it trips at the first 5 mA of leak, which is why a marginal element that limped along for months finally starts kicking the GFCI.

Testing for it is a five-minute job with a multimeter. Power off, always confirm dead. Pull one lead from each element so you are reading the element alone. Set the meter to its highest resistance range. Put one probe on an element terminal and the other on bare tank metal or the ground screw. A healthy element shows open, no path. Any reading at all, and certainly anything below a few hundred kilohms, means the sheath is cracked and leaking to ground. That element is done. While you are there, read across the two element terminals too: a normal 4500W element reads somewhere around 12 to 13 ohms, and open there means the element burned out separately. The lower element takes this hit far more often than the upper, because it lives buried in the sediment layer at the bottom of the tank.

When the tank tests clean but it still trips

Sometimes the box is dry, both elements read open to ground, and the thing still trips. Two culprits are left, and both are less common than the first two.

The first is a nuisance trip from the wiring itself. Every long cable run has a little natural capacitance, and a GFCI protecting a heater at the far end of a long home run can see enough tiny leakage across that distance to trip even with no real fault present. A neutral that got shared or crossed with another circuit downstream does the same thing, because now return current splits and the GFCI sees an imbalance that is not a ground fault at all. Confirm the heater neutral and ground each land only on their own terminals and nowhere else. If they do and it still trips, the run length or a bootleg neutral is an electrician's call.

The second is the GFCI device itself. They do wear out, and a tired one can trip on nothing or refuse to reset. It happens, but it is the rarest of the four, so it goes last. Chasing a "bad breaker" before you have checked the tank is how people replace a perfectly good GFCI and still have the problem an hour later. Only after the box, both elements, and the wiring all check clean does swapping in a fresh GFCI breaker of the same amperage make sense. If the new one holds, the old one was worn. If it trips again, the fault is still in the tank and you missed it, usually a lower element that reads marginal rather than dead open.

The order that actually saves time

Work it in the order the bars suggest, cheapest and most likely first. Power off, then look and smell for moisture in the box and under the element gaskets, because water is the single most common cause and costs nothing to check. Dry, reseal, retest. Still tripping dry? Ohm each element to the tank and pull the one that reads a leak. Elements clean too? Now look at the neutral and the run. Only when all of that is genuinely clean do you touch the GFCI itself. Skipping to the last step first is the mistake that turns a one-hour fix into a weekend and a wasted part.