What Fails First in an EV Charger: Four Weak Points

Ask an operator which parts of a charger break and the answer is rarely a list. It is a story about one site where the fans clogged, another where a contactor welded, and a third where the connector would not latch. Underneath those stories is a consistent order, because the parts that fail first are the parts whose stress accumulates fastest.

That order is worth knowing for two different reasons. For an operator it predicts what to inspect and what to keep in stock. For a buyer it reveals design decisions that are otherwise invisible: where the heat goes, how often the contactor operates, and whether the parts that wear are replaceable on their own.

Why Failure Order Is Predictable

Components fail through different mechanisms, and each mechanism has its own clock. Some wear with temperature, some with operating cycles, some with mechanical handling, and some simply with time in service regardless of load.

A charger is a machine that spends most of its life waiting and a small part of it working hard. The parts stressed by waiting — capacitors held at temperature, seals exposed to sun, electronics running at idle — age continuously. The parts stressed by working — contactors, semiconductors, connector pins — age in bursts. Whichever clock runs out first defines the service history of that model.

In practice the first clock to run out is most often thermal, and the parts most sensitive to it are the ones holding energy.

Electrolytic Capacitors and Temperature

Electrolytic capacitors store energy on the DC bus and smooth the ripple that the conversion stages produce. Inside them, a liquid electrolyte slowly escapes or degrades, and the component's capacitance falls while its internal resistance rises. The process accelerates with temperature, and it does so steeply rather than gently: a modest rise in operating temperature shortens life by a large factor.

The temperature that matters is the one at the capacitor, not the one in the weather forecast. Inside a cabinet, the air may be considerably hotter than outside, and inside a module it is hotter still. A unit with a poor air path can run its capacitors far above ambient even when the site looks cool.

This is where cabinet architecture becomes a purchasing consideration. Separating the control compartment from the power modules keeps heat-sensitive electronics away from the hottest air, and it also keeps the two serviceable independently. Units in this class often place the controller and the power stage in separate compartments for exactly this reason; the arrangement is visible in the layout and worth asking about when two products look otherwise equivalent.

The early symptom is not a failure. It is a charger that reaches its rated output more slowly, or that begins to limit output on hot afternoons while behaving normally in the morning.

Fans, Filters and Airflow

Fans fail in two ways and both are gradual. The bearing wears, so the fan turns more slowly or becomes noisy. And the intake filter loads with dust, pollen and road grit, so less air moves for the same fan speed.

Neither shows up as a fault code. Both show up as temperature, which shows up as derating, which shows up as a complaint about slow charging in summer. The chain is long enough that the fan is rarely suspected.

This is the argument for treating filters as a scheduled item rather than a repair. A filter replaced on a calendar costs little; a filter left until the cabinet derates has already shortened the life of every capacitor downstream of it. Sites differ enormously in how fast this happens: a coastal or industrial location loads a filter far faster than a clean suburban car park, and a maintenance interval that suits one is wrong for the other.

It is also the argument for looking at how a cabinet is arranged to cool. A design that pulls ambient air through a single large filter and past every heat source puts the electronics in the dirtiest and hottest air. A design with separate paths, or with the control section sealed and cooled indirectly, removes that exposure at the cost of some complexity.

Contactors and Switching Cycles

A contactor is a relay built for large currents. Every time it closes or opens under load, an arc erodes the contacts a little. On the DC side the arc is harder to extinguish than on AC, because direct current does not pass through zero, so the wear per operation is greater.

The consequence is that contactor life is measured in operations rather than in years, and the number of operations depends on how the charger is controlled. A unit that opens and closes its contactor on every small change of a load management signal accumulates cycles far faster than one that closes once per session and stays closed until the session ends.

This is one of the less visible differences between two chargers with the same hardware. The control logic decides how often the mechanical parts move, and control logic is rarely on a datasheet. It is, however, something an operator can infer from a session log that records contactor events, or from a supplier who can describe the switching policy.

The symptom of a worn contactor is intermittent and frustrating: sessions that fail to start, or that abort shortly after starting, with no consistent pattern. Because it is intermittent, it is often misattributed to the vehicle or to the connector.

Connector Latches and Cable Strain

The connector is the only part of the charger that the public touches, and it is treated accordingly. It is dropped on the ground, dragged across kerbs, driven over, pulled out at an angle and forced when the latch does not release cleanly.

The latch is the first casualty. Once it weakens, the connector seats imperfectly, contact resistance rises, and the pin gets hot. A hot pin eventually produces a session that starts and then aborts, or a connector body that discolours. The failure has moved from a mechanical annoyance to an electrical fault, and the second is much more expensive than the first.

Design choices change the cost of this. A connector head that can be replaced independently of the cable turns a worn latch into a small maintenance item. A connector moulded onto the cable turns it into a cable replacement. Strain relief at the point where the cable leaves the housing has a similar effect: it moves the bending away from the conductors and the coolant path, if there is one.

The physical layout matters too. A cable that hangs with its full weight on the connector, or that lies on the ground at every session, accumulates damage far faster than one supported by a retractor or stowed above ground.

Designing for Derating Instead of Peak

There is a design philosophy that treats the rated output as the normal operating point and another that treats it as a ceiling rarely reached. The second produces a machine that lives longer, and the difference is not a component upgrade. It is a decision about how hard to run the parts that age.

Running below the maximum relieves every thermal mechanism at once. Capacitors run cooler, fans run slower, semiconductors see lower junction temperatures, and the cabinet's own heat load falls. The cost is a lower peak figure for the same bill of materials, or a larger bill of materials for the same peak figure.

This is why a derating curve is a more informative document than a headline rating. It states the conditions under which the unit backs off, and by how much. A unit that holds its rating up to a high ambient temperature is telling you it was designed with headroom. A unit that starts backing off early is telling you the opposite, and the consequences show up years later in the parts described above.

For an operator, the practical reading is that a slightly lower sustained output may deliver more energy over a year than a higher peak that only appears in winter. For a buyer comparing two products, the question to ask is not which one reaches a higher figure, but under what conditions either one stops reaching it. The 60 kW DC station publishes its operating envelope alongside its rated output for that reason.

None of this replaces a maintenance regime. It changes what the regime should look at. Filters and coolant levels are calendar items. Contactor behaviour and summer derating are trends to watch rather than faults to fix. Connector condition is a daily observation that a site attendant can make in seconds and a monthly report cannot.

The maintenance side of this picture is treated separately in the note on EV charger maintenance, which covers routines and records rather than the physics behind them.