Charging protection is unusual because the load is a person's vehicle, connected outdoors, by an untrained user, for hours at a time, repeatedly, for years. Very little other domestic or commercial equipment combines all of those.
The result is a protection requirement that is stricter than the current rating alone would suggest, and a responsibility boundary between product and installation that is frequently assumed by both parties to sit with the other.
The functions that have to exist somewhere
Each of these can legitimately live in the charger or in the installation, and different markets and designs place them differently. What is not legitimate is for both parties to assume the other provides one.
- Residual current detection, including the DC component that vehicle electronics can introduce.
- Overcurrent protection sized for continuous operation rather than intermittent load.
- Protective earth continuity monitoring, so a broken earth prevents energising.
- Overvoltage and undervoltage detection, protecting both vehicle and charger.
- Over-temperature protection, particularly at the connector and the contactor.
- Welded contactor detection, so a failed switch cannot silently keep a circuit live.
Why DC residual current matters here specifically
The power electronics inside a vehicle can, under fault conditions, introduce a direct current component into the earth path. Conventional residual current devices designed for alternating current can be desensitised by that DC component, which means they may fail to trip when they are most needed.
This is why EV charging protection requires either a device rated to handle DC residual current, or detection of the DC component within the charger that trips an upstream device. Which approach is used is a design decision with cost and certification consequences.
Continuous duty changes the sizing
Most electrical loads are intermittent. A charging session runs at or near full rated current for hours without interruption, which is a duty most protection and wiring is not habitually sized for.
The consequence is thermal. Terminations that are adequate for intermittent loads can heat under continuous draw, and heat over time loosens connections, which increases resistance, which produces more heat. That is the mechanism behind a large share of charging installation failures.
Earth continuity is a precondition, not a monitor
The protective earth is what makes the whole safety concept work. A charger must verify it is intact before energising and must stop if it is lost during a session, rather than reporting it as a warning.
Where the supply earthing arrangement makes continuity uncertain, additional measures may be required. That is an installation-side determination and it varies by market and by supply type, which is why it cannot be resolved in the product alone.
Welded contactor detection
A contactor that welds closed leaves the output live when the charger believes it is off. The charger should be able to detect that its command and the actual state disagree, and respond by refusing further operation and reporting a fault.
Detecting it requires feedback from the contactor rather than assuming the command took effect, which is a hardware design decision made early and expensive to add later.
The boundary that causes incidents
The most common protection gap is not a missing function but an unowned one. The installer assumes the charger provides residual current protection of the required type; the charger assumes the installation does. Both are documented positions somewhere, and neither party read the other's.
Ask any supplier directly which protection functions are inside the unit and which the installation must provide. A supplier who cannot answer immediately has not thought about the boundary, which is itself informative.
Localization has protection consequences
Protection components sit in the safety path, which makes them the parts most constrained during any cost reduction or local sourcing exercise. A substitution here is a certification event rather than a procurement decision.
That is not an argument against localizing them eventually. It is an argument for doing so last, with full qualification, rather than first because they represent visible spend.