ANVILFIELD Try FieldOS

Electrical

EV charger (EVSE) installation and commissioning guide

Mount the charger, run the branch circuit, set the load management, and commission it with a real charge instead of an energized box and a thumbs-up.

EV ChargingEVSENEC Article 625CommissioningLoad ManagementElectrical

Direct answer

Installing EVSE means mounting the charger, running the branch circuit, and commissioning it; the equipment is electric vehicle supply equipment that delivers AC or DC power to the car. Commissioning verifies the ground-fault test, a real charge, the load-management setpoint, and the network. NEC Article 625, the manufacturer's instructions, and the AHJ control.

Key takeaways

  • EVSE is a continuous load under NEC Article 625, sized at 125 percent of nameplate rated current, so a 48 A charger needs a 60 A circuit.
  • Listed EVSE carries an internal CCID ground-fault device tripping near 20 mA; stacking a 5 mA GFCI breaker upstream causes nuisance trips.
  • Hardwire any charger at 48 A and above; a NEMA 14-50 on a 50 A breaker supports only 40 A continuous and triggers the 625.54 receptacle GFCI rule.
  • A disconnecting means is required for permanently connected EVSE rated over 60 A or more than 150 V to ground per 625.43, readily accessible and lockable open.
  • Commission with the ground-fault trip test, a real charge test, measured output, a verified load-management setpoint, and network status; a green light is not commissioned.

EVSE, and what installing it actually involves

EVSE is electric vehicle supply equipment: the charger that delivers power to the car, manages the handshake with the vehicle, and provides the safety functions the connection needs. The car carries its own charger for AC; the wall unit is supply equipment, not a charger in the strict sense, which is why the trade calls it EVSE on the drawings even when everyone says charger on site.

An EVSE install is three jobs stacked on each other. There is the branch circuit and the sizing, which is its own discipline. There is the physical install: the mounting, the protection from vehicles, the wiring method outdoors. And there is commissioning, which is the part that gets skipped and the part that generates most of the callbacks. An energized box with a green light is not a commissioned charger.

This guide is the install and commissioning side. The conductor and breaker sizing math lives in the EV feeder sizing walkthrough, and the question of whether the building's service can carry the load lives in the load calculation guide. Both are cross-linked, and this guide will point you to them at the sizing decisions rather than re-running the numbers. What you get here is everything from where the charger hangs to the first car charging on it.

What is the difference between Level 1, Level 2, and DC fast charging?

The three levels differ by voltage, power, and what the building has to supply. Level 1 is a 120 V AC cord set into a standard receptacle, delivering roughly 1 to 1.9 kW, around 3 to 5 miles of range per hour. It is the charger that ships in the trunk, fine for a plug-in hybrid or an overnight top-up, and almost never the answer for a workplace or commercial site.

Level 2 is the workhorse. It runs at 208 V or 240 V AC and delivers roughly 3 to 19 kW depending on the unit and the circuit, commonly 7 to 11.5 kW at home and up to about 19.2 kW at 80 A in commercial gear. That covers most workplace, multifamily, fleet, and public charging, because it adds real range over a parking shift or an overnight without a service-class supply. When someone says EV charger on a building job, they almost always mean Level 2.

DC fast charging is a different scale of equipment. The unit takes AC in, usually three-phase at 480 V, and feeds DC straight to the battery, bypassing the car's onboard charger. Output runs from about 50 kW to 350 kW and higher, enough to take many vehicles to 80 percent in 20 minutes to an hour. The supply, the transformer, and the utility are part of the design from the first sketch, which is why DCFC gets its own section later.

LevelSupplyTypical powerRange addedWhere it fits
Level 1120 V AC, standard receptacle1 to 1.9 kW~3 to 5 mi/hrCord set, PHEV, overnight top-up
Level 2208 to 240 V AC3 to 19.2 kW (16 to 80 A)~12 to 60+ mi/hrWorkplace, multifamily, fleet, public
DC fast (DCFC)480 V three-phase AC in, DC out50 to 350+ kW~80% in 20 to 60 minCorridor, fleet, high-turnover sites

The charger is a continuous load, and where the sizing lives

Every EVSE is a continuous load under NEC Article 625, which means the conductor and the overcurrent device are sized at 125 percent of the charger's rated current. A car pulls full charging current for hours, so the circuit runs at the top of its rating with no break, and the 125 percent builds in the headroom that keeps the conductor and the terminations out of trouble. A 48 A charger is a 60 A circuit. That is the number an inspector checks first.

The full sizing chain, breaker to conductor to the temperature column, is worked end to end in the EV feeder sizing walkthrough, and the voltage-drop check that usually drives the conductor larger on a long parking-lot run is in the voltage drop field guide. This guide does not repeat that math. The point to carry into the install is that the wire feeding the charger is almost always one size larger than the breaker alone would suggest, because the run is long and the load is continuous.

Read the rated current off the nameplate, not the model number and not the kilowatt figure on the spec sheet. A charger advertised at 11.5 kW is a 48 A unit at 240 V, and 48 A is what the 125 percent multiplies. Size off the marketing number and the circuit comes out wrong before the conduit is bent.

Does the existing service have room for the charger?

Before anything gets mounted, find out whether the service can carry the new load, because a charger added to a full service is a service upgrade wearing a charger's price tag. The added EV load enters the building load calculation at 125 percent, or at the managed ceiling if a load-management system controls it, and the total has to fit under the service rating with everything already on it.

On an existing building, do not re-derive the whole load from nameplates. NEC 220.87 lets you use the actual measured maximum demand, taken over a year or a recorded 30-day window at 125 percent, which usually proves more spare capacity than a paper recalculation. The full method is in the load calculation guide. The short version: meter the real peak, subtract it from the service rating, and the remainder is the headroom you have for the charger.

When the headroom is short, you have three moves, and they rank by cost. Load management caps the EV load so it fits the existing service. Load shedding drops other loads while the car charges. A service upgrade is the expensive last resort. Price the managed option before anyone quotes a new service, because on most retrofits it is a fraction of the cost and it answers the same capacity problem.

EV load management and the energy management system

Load management is the single biggest lever on a multi-charger site, and the NEC builds the path for it. Where an energy management system or an automatic load management system controls the EVSE, the load on the feeder and service is the load the system permits, not the sum of every charger at 125 percent. The provision sat at 625.42 with the management-system rules in Article 750 through the 2023 NEC, and the 2026 NEC moved energy management to a new 625.48, so confirm the section and the equipment listing in the adopted edition.

The arithmetic this opens up is large. Ten 48 A chargers sized cold come to 600 A of feeder and service demand. Put them on a system limited to 200 A and the feeder is sized for 200 A, because the system will not let the bank exceed it. The chargers share the available current and throttle when several are plugged in at once, so a full lot charges slower but every stall works. The feeder math is worked in the EV feeder sizing walkthrough; the install point is that the EMS setpoint is a commissioned parameter, not a wire.

Power sharing among chargers is the same idea inside a smaller box. Several units on one circuit or one feeder divide the available current dynamically, dropping each charger's output as more cars connect. That setpoint and that sharing scheme have to be configured and then verified during commissioning, because a load-management system that was installed but never set to the right ceiling protects nothing.

Does an EV charger need GFCI protection?

EV charging needs ground-fault protection for personnel, and most of it is built into the equipment. Listed EVSE carries an internal charging-circuit interrupting device, the CCID, that trips on a ground fault, commonly at a 20 mA threshold (CCID20). That is the personnel-protection function Article 625 expects from the equipment, and a listed personnel protection system is required at 625.22. A standard GFCI, by contrast, trips at about 5 mA, which is a different device for a different purpose.

The receptacle case adds a second requirement. A receptacle installed for EV charging needs GFCI protection under 625.54, and a 125 V to 250 V receptacle in a garage also picks up the general GFCI rule at 210.8. So a plug-in EVSE on a 14-50 receptacle sits under both the receptacle GFCI requirement and the charger's own internal CCID. A hardwired EVSE is not subject to the 625.54 receptacle rule, because there is no receptacle, and it relies on its built-in ground-fault protection instead.

Here is where it bites. Stack a 5 mA GFCI breaker upstream of an EVSE that already has its own CCID and the two protective schemes fight over the same small leakage current, and the breaker nuisance-trips. This is the classic 14-50 callback. The fix on a hardwired unit is usually a standard breaker, because the charger provides the personnel protection itself, but only if the manufacturer's instructions and the AHJ agree. Do not add or remove a GFCI breaker on a hunch. Follow the listing.

Hardwired vs a NEMA 14-50 receptacle

The choice between plug-in and hardwired changes the code path, the current ceiling, and the callback rate. A plug-in EVSE on a NEMA 14-50 receptacle is fast to install and easy to swap, and the receptacle can serve as the disconnect within the rating limits. But the receptacle is a continuous-load circuit too, so a 14-50 on a 50 A breaker supports a charger set to 40 A continuous, not 50, and it brings the 625.54 receptacle GFCI requirement and the nuisance-trip risk that comes with it.

Hardwiring runs higher current, carries its own listed ground-fault protection, and avoids the receptacle GFCI conflict. It is the only practical route at 48 A and above, because the common receptacles top out below that on a continuous basis. For light-commercial, multifamily, fleet, and anything at 48 A or more, hardwire it. The higher current, the cleaner disconnect story, and the built-in protection all point the same way.

There is also a quiet quality issue. A 14-50 receptacle pulled hard every day by a charger plug it was never rated for the duty of is a heat and failure point, and the cheap receptacles back out and discolor at the terminals under that cycling. If a plug-in install is the call, use an industrial-grade receptacle rated for the duty and torque the terminations to the value on the device. For a single home charger, plug-in is fine. On a commercial site, the receptacle is usually the part you regret.

The disconnecting means

Permanently connected EVSE above the threshold needs a disconnecting means, and the rule has tightened across recent editions. A disconnect is required where the equipment is rated over 60 A or operates at more than 150 V to ground, and it has to be readily accessible and capable of being locked in the open position so a service tech can work the unit dead. That requirement lives at 625.43, which has moved across cycles, so confirm it against the adopted edition.

The 2026 edition adds an emergency-disconnect requirement for permanently connected EVSE, calling for one or more clearly identified emergency or electrical disconnects located a set distance from the equipment, reported in the range of 20 to 100 ft. The exact wording and distance are edition-specific, so do not size or place it from memory. Pull the adopted text.

For cord-and-plug EVSE within the rating and voltage limits, the plug and receptacle are allowed to serve as the disconnect, which is part of the appeal of a plug-in install on a smaller charger. Above those limits you need a real disconnect, located and rated per the edition. The practical read: a 48 A or 80 A hardwired charger gets a disconnect, a small plug-in unit may not, and the within-sight and emergency-disconnect details are exactly the kind of thing that changes between code cycles.

Mounting, location, and protection from vehicles

Mount the charger where the cord reaches the vehicle's port without crossing a drive aisle or lying on the ground, and where a car cannot drive into it. The connector and operable parts are commonly mounted with the handle reachable in the roughly 24 to 48 in range above grade, with accessible units holding the operable parts no higher than 48 in for reach. The exact height comes from the adopted edition, the manufacturer's instructions, and the accessibility standard, so confirm rather than eyeballing it.

Vehicle impact is the most common physical failure on a parking-lot install. A charger on a pedestal in front of a parking space gets hit, and a wall-mounted unit at bumper height gets hit too. Bollards are the standard protection, set to the sides of the unit so they shield it without blocking the connector, commonly held back from the equipment and spaced so a car cannot slip between them. Article 625 and the general protection-from-physical-damage rules cover the requirement; the field practice is bollards or a wheel stop on any space a vehicle pulls into nose-first.

Cord management is the other detail that separates a clean install from a callback. A connector left to drop on the ground gets run over, and the cable gets pinched in car doors and dragged through puddles. Use the unit's holster and any cable-management arm or retractor the manufacturer provides, and route the cable so it stays off the ground between uses. On a public or multifamily site, the cable that drags is the cable that fails first.

Wiring method and conduit outdoors

Most EVSE lives outdoors, on a rooftop, or in an unconditioned garage, so the wiring method has to be rated for wet and damp locations and the enclosure for the environment. Use a raceway and conductors listed for the condition, fittings rated for wet locations, and an EVSE enclosure with the right outdoor rating. Conductors in a wet raceway are treated as wet-location conductors, which is why THWN-2 is the common pull on these jobs.

The run is usually long, because the chargers sit at the far edge of a parking field while the panel sits at the building. That distance is where voltage drop enters before ampacity is satisfied, and on a continuous load it is the constraint that drives the conductor a size or two past what the breaker requires. The full drop calculation is in the voltage drop field guide; the install lesson is to check the routed length, not the plan distance, before the wire is on the order. Find the drop after the pull and the fix is a re-pull, not a bigger reel.

Outdoor and rooftop conduit also runs hot and often packs several charger circuits in one raceway, so ambient correction and conduit-fill derating both bite. A conduit baking on a roof in summer sits well above the table's reference ambient, and the correction knocks the allowed ampacity down. The derating math is in the feeder guide. Account for it at takeoff, because a conductor sized cold can fall short once the corrections land on a hot, full pipe.

What is OCPP and when do I need a networked charger?

OCPP is the Open Charge Point Protocol, the open language a networked charger speaks to a back-end management system. It lets a charger from one maker report status, take start and stop commands, accept a current limit, and feed usage data to a management platform from a different maker. Where billing, access control, usage data, or remote load management are part of the job, the charger is networked and OCPP is how it talks to the back end.

A networked charger connects over cellular, Wi-Fi, or a wired LAN, and the connection is part of the commission, not an afterthought the owner sorts out later. The unit has to associate with the management platform, the SIM or the Wi-Fi credentials have to be loaded, the access and payment rules have to be set, and the charger has to appear online and controllable in the platform. A charger that delivers power but never came online on the network is half-installed, and on a paid site it is generating no revenue.

Not every site needs a networked charger. A single fleet charger behind a gate or a home unit often runs fine as a standalone device with no back end. The network earns its place where you need to bill users, restrict access, meter consumption per port, or share load across a bank under central control. Decide that early, because it drives the charger model, the connectivity at the location, and a real chunk of the commissioning time.

The connector: J1772, CCS, NACS, and CHAdeMO

The connector decides which vehicles a station serves, so it is a site decision before it is an install detail. SAE J1772 is the standard AC connector for Level 1 and Level 2 in North America, and a Level 2 charger almost always carries a J1772 handle or a NACS handle. The vehicle's onboard charger does the AC-to-DC conversion, so the AC connector only has to pass power and the control-pilot signal.

DC fast charging uses a different family. CCS, the Combined Charging System, is J1772 with two added DC pins, and it has been the common North American DC standard. NACS, standardized by SAE as J3400 in a 2023 technical information report, is the Tesla-origin connector that most automakers are now adopting, which is reshaping what a new DC site should install. CHAdeMO, an early DC standard, is being phased out in North America. On a new DCFC site, confirm the connector mix against the vehicles the site actually serves and the funding program's requirements, because that is moving fast.

The practical call: a Level 2 install follows the charger you specified, J1772 or NACS, and an adapter covers the gap for most drivers. A DC site is a bigger commitment, and getting the connector mix wrong strands the equipment for the vehicles that show up. Tie the connector choice to the fleet or the public traffic, not to whatever was on the shelf.

Commissioning the charger, not just energizing it

Most EV callbacks are not the equipment. They are the commissioning nobody finished. An energized charger with a green light has been tested for nothing except that it powers on. Commissioning is the sequence that proves it charges a car, trips on a ground fault, holds its load-management ceiling, and reports to the network, and it is the part that earns the closeout.

Work it as a sequence. Verify the circuit is dead, confirm terminations are torqued to the value on the device, then energize and confirm the unit powers up and self-tests clean. Run the ground-fault test the way the manufacturer specifies, because the CCID is the personnel protection and a charger that will not trip on a fault is a charger that has to come back off the wall. Then put a real load on it: charge an actual vehicle, or use an EVSE test adapter that simulates the vehicle handshake and draws current, and confirm the unit delivers the current it should at the voltage it should.

The parts that get skipped are the configured ones. Confirm the load-management setpoint is actually set and that the charger throttles when it should, not just that the EMS is installed. Bring the unit online on the network, confirm it appears and is controllable in the management platform, and run a test transaction through the billing and access flow if the site charges users. Measure the delivered power and compare it to the rating. Then record all of it. A commissioning record that shows the trip test, the charge test, the setpoint, and the network status is what backs the install when a charger acts up six months out.

  • Verify the circuit dead, then confirm all terminations torqued to the device value before energizing.
  • Energize and confirm the unit powers up and passes its internal self-test clean.
  • Run the manufacturer's ground-fault (CCID) test and confirm the unit trips as specified.
  • Charge a real vehicle or use an EVSE test adapter that draws current through the full handshake.
  • Measure delivered voltage, current, and power, and compare against the charger's rating.
  • Confirm the load-management setpoint is set and the charger throttles when the ceiling is reached.
  • Bring the charger online on the network and confirm it shows up and is controllable in the platform.
  • Run a test transaction through the billing and access flow on any site that charges users.
  • Record the trip test, the charge test, the measured output, the setpoint, and the network status.

Multifamily, workplace, and fleet sites

Multi-charger sites change the job from one circuit to a system, and the load and the metering are where it lives. Each charger still gets its own branch circuit at 125 percent of its rating, but the feeder ahead of them is sized to the load-management ceiling, not the sum, which is the whole reason load management belongs on these jobs. The feeder math, managed versus unmanaged, is worked in the EV feeder sizing walkthrough.

Metering separates these sites by who pays for the power. A workplace may eat the cost and meter nothing per port. A multifamily site usually needs per-port metering or a networked charger that bills each resident, because the building owner is not absorbing every tenant's charging. A fleet depot meters for cost accounting and to manage the depot's peak. Decide the metering model before you pick the charger, because it drives whether the unit has to be networked and revenue-grade.

Demand charges are the cost that surprises the owner. A bank of chargers that all pull full current at once spikes the building's peak demand, and on a commercial utility tariff the demand charge can dwarf the energy cost. This is the second reason load management pays for itself: capping and staggering the charging holds the demand peak down, not just the feeder size. Raise it early, because the owner who learns about demand charges on the first bill blames the installer.

DC fast charger specifics

DC fast charging is service-class power equipment, and it gets designed with the utility from the first sketch. The unit draws tens to low hundreds of kilowatts, usually three-phase at 480 V, often enough that the site needs a new transformer, a dedicated service, or a utility upgrade. The utility coordination, the make-ready, and the lead time on the transformer are the long poles, not the conduit. Get the utility into the conversation before the design is final, because their timeline drives the schedule.

The continuous-load logic still holds. The AC supply conductors and overcurrent device are sized at 125 percent of the equipment's rated input current under Article 625, and the three-phase voltage drop on the supply uses the 1.732 factor from the voltage drop field guide. The trap is sizing off the output kilowatts. A DC charger's output power is not its AC input amps, so pull the listed input rating from the equipment data, not the marketing kilowatts, or the supply comes out wrong before the first conduit is bent.

DCFC units make heat, and many are liquid-cooled or carry serious forced-air cooling for the power electronics and sometimes the cable on high-current units. The cooling system is part of the install and the maintenance the owner inherits: clearances for airflow, a coolant loop to service, filters to keep clear. A DC unit that overheats throttles its output, so a charger that charges slowly on a hot afternoon is often a cooling problem, not a power problem.

Future-proofing and make-ready

The cheapest charger is the one you wire for today but build the path for tomorrow. Make-ready is the practice of installing the conduit, the panel capacity, and sometimes the conductors for future chargers while the trenches are open and the walls are off, so adding the next charger is a connection instead of a construction project. The industry splits this into EV-capable (conduit and capacity in place), EV-ready (circuit run to the space), and EV-installed (charger mounted and live).

The economics are simple and the regret is common. Trenching a parking lot once for ten conduits costs a fraction of trenching ten times, and leaving panel and service capacity for a known future bank is far cheaper than a service upgrade after the lot is paved. If the site plan or the local code calls for a percentage of spaces to be EV-capable, that is the floor, not the target. Build for where the load is going, not where it is on opening day.

Make-ready is also where the load calculation and the service decision pay off. If the building load calc shows headroom for four chargers but the lot will eventually want twelve, the make-ready plan should reserve the path and the management scheme for twelve, even if four go live now. The load calculation guide covers proving that headroom. The install lesson is to leave the conduit and the capacity, because the second trip is the expensive one.

Keeping it running after turnover

An EVSE is outdoor equipment handled by the public, so it wears in ways a panel never does, and the owner inherits that maintenance the day the job closes. Walk them through it at handoff or the first failure is a callback dressed as a defect. The connector and the cable take the most abuse: the contacts wear and corrode, the cable gets run over and dragged, and the handle gets dropped, so they are the parts that fail first on a busy site.

Build a short maintenance rhythm into the closeout. The ground-fault function should be tested on the manufacturer's schedule, because the CCID is the personnel protection and an untested trip device is a trip device you are trusting blind. The connector and cable get a visual check for cracked insulation, bent pins, and corrosion. The enclosure seals and gaskets keep water out, so a cracked seal is a future failure. On a networked site, the connectivity and the back-end status need watching, because a charger that quietly drops off the network stops billing and nobody notices until the revenue report.

The blunt version: the chargers that fail early are the ones nobody maintains. A connector check and a trip test on a schedule cost minutes. A cooked connector or a failed personnel-protection device on a public charger costs a service call and, on the protection side, a real hazard.

What to document at closeout

An EVSE install you cannot trace from nameplate to live charge falls apart the moment a charger trips or the service gets loaded up. The closeout record has to chain from the charger nameplate to the live charge, so the next person can see that the sizing, the protection, the load management, and the commissioning were all done and proven.

Capture the charger make and model and rated current, the circuit it sits on, the load-management scheme and setpoint, the ground-fault and disconnect details, and the network and billing status. The table below is the minimum a defensible EV closeout carries, and it doubles as the commissioning record the owner hands to whoever services the unit next.

Field to recordWhy it matters
Charger make, model, level, rated current/kWDrives the 125% sizing and the whole chain
Branch circuit: OCPD, conductor, routed lengthProves the circuit was sized and the drop checked
Load management scheme and setpointJustifies a feeder smaller than full simultaneous load
GFCI / CCID and personnel protection resultShows the trip test passed at commissioning
Disconnecting means and locationCloseout and the next service tech's safety
Network: platform, connectivity, billing statusProves the charger came online and bills
Commissioning: charge test and measured outputShows it actually charged a car, not just powered on

Common mistakes

  • Sizing the circuit on the charger's running amps or its kilowatt rating instead of 125 percent of the nameplate current.
  • Skipping load management and either tripping the service or paying for a service upgrade the site did not need.
  • Stacking a GFCI breaker on an EVSE that already has its own CCID, then chasing nuisance trips instead of hardwiring it.
  • Using a 14-50 receptacle without GFCI protection where 625.54 requires it, or using a cheap receptacle that backs out under daily plug cycling.
  • Sizing the conductor for ampacity and finding the voltage drop too high after the wire is already in the pipe on a long lot run.
  • Mounting the charger with no bollard or wheel stop on a space a car pulls into nose-first, then replacing a unit that got hit.
  • Energizing the charger and calling it done without the ground-fault test, the charge test, the setpoint, or the network commission.
  • Sizing a DC fast charger's supply off the output kilowatts instead of the listed AC input current.

Field checklist

0 of 9 complete

Want this checklist to run itself on every job — with photo proof and a signed record crews can hand the customer? That's FieldOS.

Standards and references

NEC, NFPA 70, is the framework. EV supply equipment lives in Article 625, which classifies the charger as a continuous load and sets the 125 percent sizing for the conductor and the overcurrent device. The load-management provision sat at 625.42 pointing to Article 750 through 2023 and moved to 625.48 in the 2026 NEC, the disconnecting means at 625.43 with the 2026 edition adding an emergency-disconnect requirement, the receptacle GFCI requirement at 625.54, and the listed personnel protection system at 625.22. The numbering in Article 625 has shifted across cycles, so confirm every one of these against the edition the jurisdiction has adopted.

Adjacent rules govern the rest of the circuit. General receptacle GFCI is at 210.8, the existing-service metered load method is at 220.87, conductor ampacity comes from the ampacity tables at the 75 C terminal column with corrections in 310.15, terminal temperature limits are at 110.14, and the equipment grounding conductor is sized from 250.122, including the proportional upsize when phase conductors grow for voltage drop. The deep math for all of these is in the EV feeder sizing walkthrough and the load calculation guide.

Outside the NEC, SAE J1772 governs the AC connector and SAE J3400 standardizes the NACS connector, OCPP is the open networking protocol between chargers and management platforms, and accessibility follows the applicable ADA and Access Board guidance for EV charging spaces. Federally funded sites carry program requirements such as NEVI, which can mandate connector standards and installer certification. The EVSE listing under UL and the manufacturer's instructions can impose tighter requirements that do govern, so cite the standard that controls the point and let the project specification and the AHJ override the rule of thumb.

Units, terms, and conversions

EV charging carries its own vocabulary, and the same equipment reads differently across a charger spec sheet, a one-line, and a permit set.

The charger is the EVSE, electric vehicle supply equipment, and its rated current is the number the 125 percent rule multiplies. Power is given in kilowatts (kW) on the spec sheet, which is voltage times current, so an 11.5 kW Level 2 unit is 48 A at 240 V. Level 1 is 120 V AC, Level 2 is 208 V to 240 V AC, and DC fast charging takes three-phase AC in and delivers DC out. CCID is the charger's built-in ground-fault device, OCPP is the networking protocol, and EVEMS is the energy management system that caps the load.

EVSE
Electric vehicle supply equipment, the charger, a continuous load under NEC Article 625
Level 1 / Level 2 / DCFC
120 V AC; 208 to 240 V AC; and DC fast charging fed three-phase, delivering DC to the battery
CCID
Charging-circuit interrupting device, the EVSE's built-in ground-fault protection, commonly tripping at 20 mA
EMS / EVEMS
Energy management system that limits EVSE load so the feeder and service need not carry full simultaneous demand (625.42)
OCPP
Open Charge Point Protocol, the open language a networked charger speaks to a management platform
Make-ready
Conduit, panel, and capacity installed for future chargers; EV-capable, EV-ready, or EV-installed
J1772 / J3400
SAE AC charging connector standard; J3400 standardizes the NACS connector for AC and DC

Related tools

Calculators and readiness checks for this work

Compare your options

FAQ

What is the difference between Level 1, Level 2, and DC fast charging?

Level 1 is 120 V AC at about 1 to 1.9 kW, slow enough for overnight top-ups. Level 2 is 208 to 240 V AC at roughly 3 to 19 kW, the workhorse for workplace, multifamily, and commercial sites. DC fast charging feeds three-phase power and delivers 50 to 350+ kW straight to the battery.

Does an EV charger need GFCI protection?

Listed EVSE carries its own ground-fault protection, a CCID that trips around 20 mA. A receptacle installed for EV charging also requires GFCI under NEC 625.54, and a garage receptacle picks up 210.8. A hardwired charger relies on its built-in protection and generally does not need a GFCI breaker unless the manufacturer or AHJ requires it.

What is EV load management?

EV load management caps the charging load so the feeder and service do not have to carry every charger at full current at once. Under NEC 625.42, an energy management system lets you size the feeder to the managed ceiling instead of the sum, and chargers share and throttle current as more cars plug in.

Hardwired vs plug-in EV charger: which should I install?

Hardwire any charger at 48 A or above and most commercial work, since it runs higher current, carries its own ground-fault protection, and avoids the receptacle GFCI nuisance trip. A plug-in 14-50 install is fine for a single home charger but limited to 40 A continuous and exposed to the GFCI conflict and receptacle wear.

Why does my NEMA 14-50 EV charger keep tripping the GFCI breaker?

The charger already has its own ground-fault device (CCID), and a 5 mA GFCI breaker upstream reacts to the same small leakage, so the two double-trip on harmless current. The common fix is hardwiring the unit so the charger provides the personnel protection, where the manufacturer's instructions and the AHJ allow it.

How do you commission an EV charger?

Energize after confirming torqued terminations, run the manufacturer's ground-fault test, then charge a real vehicle or an EVSE test adapter and measure the delivered output. Verify the load-management setpoint, bring a networked unit online in its platform, run a test transaction, and record it all. A green light alone is not a commissioned charger.

Do I need a disconnect for an EV charger?

Through the 2023 NEC a disconnecting means was required for permanently connected EVSE rated over 60 A or more than 150 V to ground; the 2026 edition broadened that requirement, so confirm the trigger against the adopted edition. It must be readily accessible and lockable open under NEC 625.43. The 2026 edition adds an emergency disconnect for permanently connected units. A small cord-and-plug charger can use the plug as its disconnect within the rating limits.

What is OCPP and do I need a networked charger?

OCPP is the open protocol a networked charger uses to talk to a back-end management platform for status, control, billing, and load management. You need a networked charger where you bill users, restrict access, meter per port, or share load centrally. A single fleet or home charger often runs fine standalone.

How do I add an EV charger to an existing service without an upgrade?

Use NEC 220.87 to find the real existing demand from a metered peak, subtract it from the service rating, and check the remaining headroom. The charger enters at 125 percent. If it does not fit, load management that caps the EV load is usually far cheaper than a service upgrade, so price it first.

How high should an EV charger be mounted?

The connector and operable parts are commonly mounted in the roughly 24 to 48 in range above grade, with accessible units holding operable parts no higher than 48 in for reach. The exact height comes from the adopted code, the manufacturer's instructions, and the accessibility standard, so confirm it rather than eyeballing the bracket.

People also ask

Codes cited in this guide

This guide is written and reviewed against the published standards below. Always confirm the current adopted edition with the authority having jurisdiction.