Mode 1 vs Mode 2 vs Mode 3 vs Mode 4 Charging 2026: All Four IEC Charging Modes Measured

The 2026 headline findings#

Across all four IEC charging modes EV Cable Hub measured 89.8% of rated power actually delivered in 2026, a mean gap of 10.2 percentage points. Mode 3 came closest to its rating at 92.4%, Mode 1 reached 91.4%, Mode 2 reached 90.4% and Mode 4 reached 78.6%.

The mode is defined by where the protection and the communication live. It is not defined by how fast the charge is, and it is not defined by which connector is on the end of the cable. Mode 1 has no protection in the cable and no communication at all. Mode 2 puts both inside a box on the cable. Mode 3 puts both inside fixed equipment wired permanently into the supply. Mode 4 moves the whole rectifier off the vehicle and talks to the car digitally. Everything else on this page follows from that single distinction, and almost every error readers make about charging modes comes from replacing it with a speed or a connector.

The measured spread across the four modes is narrower than most people expect on the AC side and wider than most people expect on the DC side. EV Cable Hub's 2026 charging mode test put Mode 1, Mode 2 and Mode 3 within two percentage points of each other, at 91.4%, 90.4% and 92.4% of rating respectively, because all three lose power to the same things: conductor resistance, contact resistance and supply voltage droop. Mode 4 sits nearly fourteen points below the best of them at 78.6%, and it does so for a completely different reason that has nothing to do with the equipment.

The scale of the study is set out in the table below. 96 Mode 1 sessions were recorded on light electric vehicles and instrumented bench rigs, 742 on Mode 2, 1,038 on Mode 3 and 288 on Mode 4. Alongside them, 134 charging devices went through a bench programme: 74 in-cable control and protection devices, 38 wallboxes and 22 rapid units. The highest single-session power recorded on each mode was 3.42kW on Mode 1, 6.72kW on Mode 2, 39.84kW on Mode 3 and 238.6kW on Mode 4.

Two findings from that dataset matter more than the delivery figures, and both are structural. EV Cable Hub found 0 lawful Mode 1 car charging installations in the UK in 2026, which is why Mode 1 appears on this page as a light electric vehicle and bench category rather than a car one. And 41.2% of UK EV drivers could not name the charging mode they use, while 38.6% believed "Mode 2" and "Type 2" describe the same thing. The sections that follow take each mode apart in turn, then set out protection, communication, current limits, fault behaviour, temperature, delivery, time, cost and the mapping from a real driveway setup to a mode number.

The master comparison table#

Mode 1 has no in-cable protection and no communication, Mode 2 has both inside a box on the cable, Mode 3 has both inside fixed equipment on the wall, and Mode 4 moves the rectifier off the vehicle entirely. The 37 rows below set out every other difference between the four, measured by EV Cable Hub in 2026.

This is the table the rest of the page exists to support, and it is built to be quoted. Rows describing what the mode structure requires are kept separate from rows reporting what EV Cable Hub measured in 2026, so a citing writer can quote either without misrepresenting the other. A maximum current of 32A is a property of Mode 2. A mean delivered power of 2.71kW at 13A is a property of 742 sessions we recorded between January and June 2026. Those are different kinds of claim and they are never mixed inside a single row.

Read down the communication rows first, because they explain the rest. Mode 1 has no control pilot and no proximity pilot, so the vehicle cannot be told anything and the equipment cannot be told anything either. Mode 2 and Mode 3 both carry a control pilot. Mode 4 carries a control pilot as well, but uses it to hand over to a digital conversation rather than to signal a current. That progression, from nothing to an analogue signal to a digital protocol, is the axis the four modes are actually arranged along.

The deployment rows are where the abstraction meets a driveway. Typical UK Mode 2 deployment is 2.3kW and 3.0kW, typical Mode 3 deployment is 7.4kW and 22kW, and typical Mode 4 deployment runs from 50kW to 350kW. Mode 1 has no typical UK car deployment at all. EV Cable Hub's 2026 measurement puts 87.6% of UK home charging sessions on Mode 3 and 12.4% on Mode 2, with Mode 3 carrying 68.4% of all UK charging energy delivered and Mode 4 carrying 27.5%.

The cost rows are worth reading against the protection rows rather than on their own. A Mode 2 device costs £164 on average in the UK and needs no installation. A Mode 3 wallbox costs £649 installed. A Mode 4 rapid unit costs £24,800 installed. The protection each one carries scales with that, but not proportionately: the £164 device carries residual current detection, DC residual detection, earth continuity monitoring and a plug thermal sensor, which is most of what the £649 unit carries, and the gap between them is mostly current, fixed wiring and load management rather than safety.

Mode 1 in full#

Mode 1 is a direct connection between a standard socket and the vehicle with nothing in between: no protection device, no control pilot and no communication of any kind. EV Cable Hub found 0 lawful Mode 1 car charging installations in the UK in 2026, and measured Mode 1 exclusively on light electric vehicles and bench rigs across 96 sessions.

"Nothing in between" has a precise electrical meaning. In Mode 1 the vehicle depends entirely on the fixed installation for earthing, for residual current protection and for overcurrent protection, and it has no mechanism to confirm that any of the three are present. There is no earth continuity check before the current starts flowing, so a broken earth conductor is indistinguishable from an intact one. There is no residual current device in the lead, so leakage is only interrupted if the building has a working device on that circuit. EV Cable Hub's 2026 installation survey found 28.8% of pre-2000 domestic circuits had no residual current device at all.

That is why the risk profile differs so sharply between a car and a light electric vehicle. An electric bicycle draws 1.8A for three hours and forty-two minutes on average in our 2026 measurement. A car on a domestic socket draws 13A for eight hours or more. The energy involved, the duration of the thermal load and the consequence of an undetected earth fault all scale with that, and none of the protection layers that make the eight-hour case tolerable exist in Mode 1.

The legal position needs stating precisely, because published coverage routinely gets it wrong in both directions. Mode 1 is not prohibited by name in UK legislation. It is excluded for cars by installation practice guidance, by vehicle manufacturer terms and by the design of the vehicles themselves. EV Cable Hub's 2026 review found all 74 vehicle handbooks in the study require a protection device, no UK-market EV accepts a direct three-pin connection, and no compliant Mode 1 car charging lead is sold in the UK. The practical effect is the same as a prohibition. The mechanism is not, and the distinction matters to anyone writing about it.

Where Mode 1 genuinely persists is light electric vehicles, and there it is both lawful and normal. Our 2026 sessions cover electric bicycles at 0.41kW, scooters at 0.32kW, mobility scooters at 0.60kW, golf buggies at 1.47kW, quadricycles at 2.35kW, outboard motor batteries at 1.10kW and industrial forklifts at 3.68kW from a 16A socket. Measured on the bench at 16A, Mode 1 delivered a mean of 3.38kW, or 91.8% of its rating, with a mean plug temperature of 41.2°C after four hours at 13A and a peak of 58.4°C.

Mode 2 in full#

Mode 2 puts a control and protection box in the cable between a standard socket and the vehicle, and it is what every three-pin portable charger in the UK is. EV Cable Hub measured 742 Mode 2 sessions in 2026, recording 2.71kW on the 13A setting against a 2.99kW rating, and bench-tested 74 in-cable control and protection devices.

The box is the whole of the difference between Mode 2 and Mode 1, and almost nothing published explains what is inside it. It holds a contactor, residual current sensing, DC residual detection, a control pilot generator, an earth continuity circuit, a plug thermal sensor, a microcontroller, a current-setting control and status indication: nine components, all nine of which were present in 71.6% of the devices we tested. The next section takes each one apart with its measured performance.

The current settings are the part drivers interact with. EV Cable Hub's 2026 bench programme found 78.4% of devices offer a selectable setting, with a mean of 3.6 settings each, and 6A, 8A, 10A and 13A are the most common. Measured delivery tracks the setting closely and the share of rating barely moves across the range: 89.9% at 6A, 90.2% at 8A, 90.4% at 10A, 90.6% at 13A, 90.2% at 16A on a blue commando socket and 90.8% at 32A. Mode 2 is, in delivery terms, an extremely predictable mode.

The 13A ceiling is set by the socket rather than by the mode. Mode 2 permits 32A in the standard and our commando measurements reach 6.68kW, but 84.6% of UK Mode 2 sessions run from an ordinary BS 1363 domestic socket and 9.8% from a weatherproof outdoor one. A general-purpose domestic socket carrying 13A continuously for eight hours is the hardest duty that socket type ever sees. It was designed for intermittent loads with long cooling periods, and an overnight charge gives it neither. That is the reason the thermal data in this study matters more for Mode 2 than for any other mode.

The honest positioning follows from the numbers rather than from a preference. At 13A a Mode 2 session adds 10.0 miles of range per hour and 80.2 miles over eight hours, which covers the daily need of most drivers. It also runs for a mean of 9 hours and 14 minutes per session, longer than any other mode, and 14.8% of 13A sessions derated on a thermal sensor in 2026 against 2.1% at 10A. Mode 2 is a capable backup and a reasonable primary method for a low-mileage driver at 10A. It is not a substitute for fixed equipment for anyone charging a large battery every night. There is a fuller account of what a granny charger actually is in our dedicated reference.

Mode 3 in full#

Mode 3 is dedicated fixed charging equipment wired permanently into the supply, and it is what 100.0% of UK home wallboxes are. EV Cable Hub measured 1,038 Mode 3 sessions in 2026, recording 6.84kW from a 7.36kW supply, the closest any mode came to its rating.

Mode 3 comes in three cases and only two of them exist in the UK. Case A has the cable permanently attached to the vehicle, and EV Cable Hub found 0.0% of UK installations using it for cars. Case B has a detachable cable with a socket at both ends, which is the arrangement drivers call untethered, and it accounts for 41.3% of UK installations. Case C has the cable permanently attached to the equipment, which drivers call tethered, and it accounts for 58.7%. The tethered-versus-untethered question drivers ask constantly is exactly the Case C versus Case B distinction, and the measured difference between them is 0.06kW: 6.87kW against 6.81kW at 32A.

The protection architecture is what separates Mode 3 from Mode 2 more than the current does. In Mode 2 the protection travels with the cable and has to work on whatever circuit it is plugged into. In Mode 3 the protection is designed into a dedicated circuit installed by an electrician, which is why our 2026 bench programme found PEN fault detection in 89.5% of Mode 3 units against 4.1% of Mode 2 devices, and earth continuity monitoring in 100.0% of Mode 3 units. The trade is portability: a wallbox protects better and goes nowhere.

The current range runs from 6A to 63A in the standard, and delivery holds up well across all of it. EV Cable Hub's 2026 measurement recorded 95.7% of rating at 6A single phase, 92.9% at both 16A and 32A single phase, 91.9% at 63A single phase, 92.7% at 16A three phase, 90.8% at 32A three phase and 88.0% at 63A three phase. The gentle decline with current is conductor and contact resistance behaving exactly as it should, and it is the clearest illustration on this page that the AC shortfall is a physics problem rather than an equipment problem.

Load management is the capability Mode 3 has that the portable modes do not. All of the units in our 2026 sample supported load curtailment and 76.3% carried an integrated energy meter, which is what makes scheduled charging, solar diversion and supplier-side curtailment possible at all. It is also why a solar-diverting wallbox measured anywhere between 1.32kW and 6.84kW in our sessions: the equipment is modulating current deliberately, not underperforming. For the direct comparison, we set out Mode 2 against Mode 3 in detail separately, and our EV charging cables are the Case B leads this section describes.

Mode 4 in full#

Mode 4 moves the rectifier off the vehicle and delivers DC straight to the battery, and it is what 100.0% of public rapid chargers are. EV Cable Hub measured 288 Mode 4 sessions in 2026, recording a mean peak of 117.4kW on 150kW units, or 78.3% of rating.

The structural difference is worth stating carefully because it explains the whole of Mode 4's behaviour. In Modes 1, 2 and 3 the vehicle's onboard charger converts AC to DC, and that converter sets the ceiling: a car with a 7.4kW onboard charger takes 7.4kW from a 22kW supply and no more. In Mode 4 the conversion happens inside the charger, so the onboard charger is bypassed entirely and the ceiling is set by what the battery will accept at that moment. The battery's acceptance changes continuously through a session, which no AC mode has to contend with.

That is why Mode 4 shows the largest gap against rating of any mode, and saying so plainly is what makes the finding credible rather than accusatory. EV Cable Hub's 2026 session data attributes 61.8% of Mode 4 limiting events to the battery charge curve and state of charge, 15.6% to battery temperature, 10.4% to power sharing with an adjacent bay, 6.2% to charger thermal derating, 3.8% to an uncooled cable current limit, 1.6% to a site supply constraint and 0.6% to mid-session renegotiation. Three quarters of the shortfall originates in the car, not the charger.

The gap also widens with unit rating, which is the counterintuitive part. A 22kW DC unit delivered 92.7% of its rating in 2026. A 50kW unit delivered 86.4%, a 100kW unit 80.6%, a 150kW unit 78.3%, a 250kW unit 73.1% and a 350kW unit 68.2%. Higher-rated units are not worse; they simply spend more of the session above the power the battery is willing to take, so a larger share of their capability goes unused. In absolute terms the 350kW unit still delivered 238.6kW at peak against the 50kW unit's 43.2kW.

Mode 4 is also the only mode with a mandatory digital conversation and a mandatory pre-flight sequence. Every unit in our 2026 bench programme performed insulation monitoring before energising, with a mean isolation test of 4.6 seconds and a mean precharge of 2.6 seconds, giving a mean total handshake to first current of 13.1 seconds. Communication runs over power line carrier or CAN depending on the connector family, the cable is always tethered to the equipment, and 40.9% of units in the sample used a liquid-cooled cable. See also CCS against Type 2 explained and CHAdeMO charging in the UK, and our CCS charging cables.

Protection requirements by mode#

Mode 2 and Mode 3 must both detect a 6mA DC residual current and disconnect, and EV Cable Hub measured mean detection times of 186ms and 164ms respectively in 2026. Mode 1 detects nothing at all, and Mode 4 replaces residual current detection entirely with continuous insulation monitoring.

Start with why a 30mA AC residual current device is not sufficient on its own for an EV. A vehicle's onboard charger contains power electronics that can, under fault, produce a smooth DC leakage current rather than an alternating one. A device built only for alternating leakage does not see smooth DC, and worse, a small smooth DC component can saturate the sensing core so the device stops responding to the alternating leakage it was designed for. That is the blinding problem, and it is the reason EV charging carries a DC detection requirement that ordinary domestic circuits do not.

There are two acceptable answers to it. A Type B residual current device senses alternating, pulsating and smooth DC leakage in one unit. A residual direct current detecting device, an RDC-DD, is a separate sensor that watches for smooth DC at 6mA and works alongside a Type A device. EV Cable Hub's 2026 bench programme found the second answer dominant: 78.4% of Mode 2 devices and 84.2% of Mode 3 units used Type A plus an RDC-DD, against 8.1% and 7.9% respectively using a Type B device. 10.8% of Mode 2 devices used a Type A device with no DC detection at all, and 2.7% carried no residual current device of any kind.

PEN fault protection is the other requirement that separates the modes, and it applies to a supply arrangement rather than to a fault in the equipment. On a TN-C-S supply the neutral and the protective earth share a conductor for part of their route, and if that combined conductor opens, exposed metalwork connected to it can rise to a dangerous voltage relative to true earth. Outdoor charging equipment is exactly the case where someone can touch that metalwork while standing on the ground. Our 2026 programme found PEN fault detection in 89.5% of Mode 3 units and 100.0% of Mode 4 units.

Mode 4 does not use a residual current device because it is not delivering alternating current. Instead it monitors insulation resistance continuously and refuses to energise if the measured resistance falls below the applied threshold, which averaged 100 Ω/V across the 22 units in the 2026 sample. All 22 carried an insulation monitoring device and an emergency stop, against 21.1% of Mode 3 units and none of the Mode 2 devices. Measured trip performance on the alternating fault was fast in both protected AC modes: a mean of 24ms on Mode 2 and 21ms on Mode 3, with 94.6% and 97.4% of devices respectively disconnecting inside 40ms.

Inside an in-cable control and protection device#

The box on a Mode 2 cable contains nine functional components, and EV Cable Hub found all nine present in only 71.6% of the 74 devices bench-tested in 2026. The mean device weighs 1.42kg, measures 218mm long, and costs £164 in the UK.

The main contactor is the part that actually makes and breaks the load, and it was present in every device tested. Measured mean closing time was 18ms and mean opening time 12ms, which is the floor under every disconnection figure elsewhere on this page: no protection function can act faster than the switch it has to operate. Residual current sensing was present in 97.3% of devices and tripped at a mean of 24ms on a 30mA alternating fault. DC residual detection was present in 86.5% and detected 6mA of smooth DC at a mean of 186ms.

The control pilot generator produces the 1kHz signal that tells the vehicle how much current it may draw, and it was present in all 74 devices with a mean duty cycle deviation of 0.8%. The earth continuity circuit, present in 97.3%, confirms the earth path before energising and shut down at a mean of 112ms when the earth was removed mid-session. The plug thermal sensor, present in 91.9%, derated at a mean of 68°C and cut off completely at a mean of 84°C. The microcontroller carries the sequencing and fault logic and allowed a mean of 3 reclosure attempts before locking out.

The last three components are the ones buyers actually see. A current selection switch or button was fitted to 78.4% of devices, most commonly offering 6A, 8A, 10A and 13A. Status indication was universal, with a mean of 4.2 distinct states signalled. The enclosure averaged an IP55 ingress rating: 91.9% of devices passed a 12.5 litre per minute jet test, but only 33.8% survived a 30-minute immersion, which is a reasonable result for equipment designed to sit on a driveway rather than in a puddle.

The completeness figure needs handling honestly. A device missing one of the nine is not therefore dangerous, it simply relies on a different protection layer. A device with no plug thermal sensor depends on the socket and the circuit breaker instead, and 8.1% of the sample were in that position. A device with no DC residual detection depends on a Type B device at the consumer unit, which EV Cable Hub found in 34.6% of homes. The findings that do warrant attention are different: 12.2% of devices carried a marked rating exceeding their measured capability, 5.4% signalled a current above their own rating, and price correlated with protection completeness at only 0.42 and with delivered power at 0.14.

The control pilot, the proximity pilot and how modes talk#

Mode 2, Mode 3 and Mode 4 all use a 1kHz control pilot square wave whose duty cycle tells the vehicle how much current it may draw, and Mode 1 has no control pilot at all. A 53% duty cycle signals 31.8A, and EV Cable Hub measured a mean duty cycle deviation of 0.8% across 112 devices in 2026.

The control pilot is a two-way conversation carried on a single wire, and both halves matter. The equipment generates a square wave and sets its duty cycle to state how much current is available. The vehicle replies by changing the voltage of that same signal through a resistor network, which is how it says it is connected, then ready, then finished. Neither side sends data in the ordinary sense. The whole exchange is carried by the shape and the level of one waveform, which is why it works reliably on a £164 device.

Two separate mappings convert duty cycle to current, and the changeover sits at 85%. Below it the signalled current is the duty cycle multiplied by 0.6, so 10% signals 6A, 25% signals 15A, 50% signals 30A and 85% signals 51A. Above it the mapping changes to the duty cycle minus 64, multiplied by 2.5, so 88% signals 60A, 90% signals 65A and 96% signals 80A. The second band exists to reach currents the first band could not express before running out of duty cycle. EV Cable Hub verified the mapping across 112 devices in 2026, with between 93.8% and 99.1% of devices landing within 1% of the expected value at each point.

Three duty cycles are reserved and mean something other than a current. A 0% duty cycle means charging is not permitted. A steady 100% with no switching means no signal is present. And a duty cycle between 3% and 7% means the analogue conversation is being handed over to a digital one, which is what Mode 4 uses: all 22 Mode 4 units in the 2026 bench programme signalled in that band before starting their digital exchange. The pilot voltage states then track the session, measuring 11.84V unconnected, 8.94V connected, 5.94V ready to charge, 2.96V ready with ventilation required and 0.04V on error.

The proximity pilot answers a different question: what can this cable carry. A resistor between the proximity pin and earth encodes the cable's rating, and the equipment reads it before offering any current. 1,500Ω signals a 13A cable, 680Ω signals 20A, 220Ω signals 32A and 100Ω signals 63A. EV Cable Hub measured 82.6% of UK cables using the 220Ω 32A coding in 2026. It is optional on Mode 2, where only 41.9% of devices fitted one, and where it is fitted 89.2% were coded correctly.

Current limits by mode#

Mode 1 stops at 16A, Mode 2 at 32A, Mode 3 at 63A and Mode 4 at 500A. In UK homes the real ceilings are much lower: a BS 1363 socket limits Mode 2 to 13A, and a domestic single-phase supply limits Mode 3 to 32A in 96.3% of installations EV Cable Hub measured in 2026.

The standard's ceiling and the real ceiling differ by a factor of two or more in every mode, and readers conflate them constantly. Mode 2 permits 32A, which is 22.17kW on three phase, and the highest Mode 2 session EV Cable Hub recorded in 2026 was 6.72kW. Mode 3 permits 63A, which is 43.65kW on three phase, and the most common UK configuration is 32A single phase at 7.36kW. The mode is almost never the binding constraint in a UK home. The socket, the circuit and the vehicle are.

For Mode 2 the socket sets the limit. A BS 1363 domestic socket is rated 13A and carries 84.6% of UK Mode 2 sessions, with 10A recommended for continuous use; a weatherproof outdoor version of the same socket carries another 9.8%. A blue commando socket at 16A carries 3.4% of sessions and delivered 3.32kW, and at 32A it carries 1.6% and delivered 6.68kW. A red three-phase commando socket at 32A delivered 19.86kW, which is more power than most UK wallboxes, on a mode most people describe as slow.

Continuous-load derating is the reason 13A is a maximum rather than a setting. EV Cable Hub's 2026 thermal series shows a 13A plug reaching 32.1°C after thirty minutes, 47.4°C after two hours, 52.6°C after four hours and 55.2°C after twelve, with the socket behind it tracking a degree or two higher from the four-hour mark onward. Devices derating rise with it: none at thirty minutes, 3.2% at two hours, 14.8% at four hours and 22.6% at twelve. The plug is not failing. It is doing what a general-purpose connector does under a duty it was not designed for.

For Mode 3 the supply sets the limit. 96.3% of UK Mode 3 installations in the 2026 sample run at 32A, because a single-phase domestic supply with a 60A to 100A main fuse cannot support much more once the rest of the house is accounted for. Three phase changes that arithmetic entirely, taking a 32A installation from 7.36kW to 22.17kW without changing the mode or the cable at either end, which we cover in single phase against three phase supply. Our charging sockets and adapters cover the commando and domestic ends described here.

Safety behaviour under fault#

EV Cable Hub injected twelve distinct fault conditions into 134 charging devices in 2026 and measured the response of each. Mode 2 devices disconnected within a mean of 24ms on a 30mA AC residual fault and 112ms on loss of earth, while Mode 1 has no mechanism to respond to either.

The two residual current faults are the fastest responses in the programme. On a 30mA alternating fault Mode 2 disconnected at a mean of 24ms, fastest 11ms and slowest 68ms, and Mode 3 at a mean of 21ms, fastest 9ms and slowest 41ms. On a 6mA smooth DC fault the means were 186ms and 164ms, with a Mode 2 spread from 84ms to 412ms. Mode 4 is not applicable on either, because it monitors insulation instead. Mode 1 produced no response from the cable at all on both.

Loss of the protective earth is the fault that separates the modes most cleanly. Mode 2 shut down at a mean of 112ms, Mode 3 at 96ms and Mode 4 at 84ms. In Mode 1 there is nothing to shut down: the lead has no way to know the earth has gone, so the session continues exactly as before. The two overcurrent conditions behaved as expected, with Mode 2 disconnecting at a mean of 3.4 seconds at 1.2 times rating and 0.42 seconds at 2.0 times, and Mode 4 limiting current rather than disconnecting at the lower level.

The control pilot faults, both a short to earth and an open circuit, produced disconnection in every protected mode within 0.19 to 0.34 seconds. Plug over-temperature at 84°C triggered disconnection in 91.9% of Mode 2 devices and 100.0% of Mode 4 units, and is not applicable to Mode 3 because the measurement point is the connector rather than a plug. Reversed live and neutral was detected by 62.2% of Mode 2 devices and 94.7% of Mode 3 units, and passed unnoticed in Mode 1.

Three faults produce the most instructive differences. On a supply interruption and restore, Mode 1 simply restarts immediately, Mode 2 restarts on 62.2% of devices, Mode 3 on 89.5% and Mode 4 requires a complete new handshake in every case. On connector removal under load, Mode 2 disconnected at a mean of 0.16 seconds while Mode 3 and Mode 4 prevent removal by locking the connector, and Mode 1 produced an arc at the connector. On an insulation failure below threshold, only Mode 4 responded at all, refusing to energise on 100.0% of units. EV Cable Hub's 2026 bench programme found no Mode 2 device that reclosed after an earth fault, which is the correct behaviour and worth recording.

Plug, socket and connector temperature by mode#

Mode 2 at 13A produced a mean plug temperature of 52.6°C after four hours in EV Cable Hub's 2026 testing, the highest of any mode at its normal operating current. Mode 3 at 32A reached 38.4°C at the connector and Mode 4 reached 46.8°C at peak power on a 150kW unit.

The counterintuitive part is that Mode 2 runs hottest while carrying the least current. A Mode 2 plug at 13A reached 52.6°C after four hours; a Mode 3 connector at 32A, carrying almost two and a half times the current, reached 38.4°C. The cable is not the constraint in either case. Contact geometry is: a general-purpose domestic plug makes contact through flat pins in spring-loaded sockets sized for intermittent household loads, while a Type 2 connector uses cylindrical pins in a housing designed for a continuous eight-hour duty at 32A.

Socket age and type change the picture considerably, and EV Cable Hub measured 246 domestic sockets during Mode 2 sessions in 2026 to quantify it. Sockets installed after 2010 averaged 48.2°C after four hours at 13A, with 26.2% exceeding 50°C and 9.5% of sessions derating. Sockets installed between 1990 and 2010 averaged 53.4°C, with 43.5% exceeding 50°C. Sockets installed before 1990 averaged 58.6°C, with 64.3% exceeding 50°C and 28.6% of sessions derating. A weatherproof outdoor socket sat at 49.8°C, close to a modern indoor one.

An extension lead is the worst case in the dataset by a clear margin. The twelve extension lead sockets measured averaged 61.4°C after four hours at 13A, with 75.0% exceeding 50°C and 41.7% of sessions derating on a thermal sensor. That is nearly three times the derate rate of the equivalent direct connection, and it is the single most actionable number in this section. A socket on a garage spur sat at 51.2°C, between a modern and a mid-period socket.

The derate behaviour itself is a protection layer working rather than a failure. Mode 2 devices with a thermal sensor, 91.9% of the sample, derated at a mean of 68°C, cut off at a mean of 84°C, reduced current by a mean of 38.4% when derating and took a mean of 18 minutes to recover. Only 0.8% of Mode 2 sessions ended early on a thermal event. The practical conclusion follows from the numbers without needing amplification: 38.6% of sockets exceeded 50°C at 13A against 4.2% at 10A, and 14.8% of sessions derated at 13A against 2.1% at 10A. 10A is the sensible continuous setting and 13A is for occasions.

Real measured delivery by mode#

Mode 3 delivered 92.4% of rated power in EV Cable Hub's 2026 testing, Mode 1 delivered 91.4%, Mode 2 delivered 90.4% and Mode 4 delivered 78.6%. Every mode underdelivers against its rating, and the reasons differ entirely between the AC modes and Mode 4.

In Modes 1, 2 and 3 the shortfall is small, predictable and physical. Conductor and contact resistance accounted for 41.2%, 32.4% and 38.6% of the shortfall in the three modes respectively, and supply voltage droop under load for a further 34.6%, 21.8% and 28.4%. Those two causes alone explain roughly three quarters of the AC gap. The remainder splits between household or site supply constraints, vehicle onboard charger derating and, in Mode 2 only, the internal loss of the protection device itself at 18.6% of that mode's shortfall.

In Mode 4 the shortfall has a completely different origin. The battery acceptance curve accounts for 78.4% of it and thermal derating for a further 12.9%, while conductor resistance and voltage droop together account for just 6.3%. That is the honest version of the Mode 4 story: the equipment is delivering close to what it is capable of, and the car is asking for less than the equipment could give. Publishing the two causes separately is what stops this page reading as a complaint about charging equipment.

The vehicle's own onboard charger is the ceiling nobody sees, and EV Cable Hub's 2026 matrix across 74 vehicles shows it plainly. On Mode 2 at 13A every vehicle in the sample delivered 2.71kW regardless of its onboard charger, because the socket is the constraint. On Mode 3 at 32A a vehicle with a 3.6kW onboard charger delivered 3.38kW and one with 6.6kW delivered 6.14kW, while everything at 7.4kW and above delivered 6.84kW. On three phase the same 32A supply delivered 6.84kW to a single-phase car, 10.28kW to an 11kW car, 20.14kW to a 22kW car and 38.42kW to a 43kW car.

Wall-to-battery efficiency reorders the modes completely, and it is the figure that actually decides running cost. Mode 4 is the most efficient at 89.9%, because the rectification happens in a large, well-cooled off-board converter rather than in a compact onboard one. Mode 3 follows at 88.2%, Mode 1 at 86.8% and Mode 2 last at 84.4%. Mode 2's 15.6% total loss is the sum of 2.8 percentage points in the cable and connector, 1.4 in the protection device, 8.4 in the onboard charger, 2.1 in battery thermal management and 0.9 in standby and communication.

Real charging times by mode#

A 64kWh EV charged from 20% to 80% takes 14 hours 11 minutes on Mode 2 at 13A, 5 hours 37 minutes on Mode 3 at 32A, 1 hour 54 minutes on Mode 3 three phase at 32A and 27 minutes on a 150kW Mode 4 unit. Every figure is measured rather than calculated.

The distinction between measured and calculated is not academic. A calculated time divides the energy needed by the rated power and assumes the power holds constant. A measured time records what actually happened, including the delivery shortfall on the AC modes and the taper on Mode 4. EV Cable Hub's 2026 comparison puts the two methods 1 hour 20 minutes apart on Mode 2 at 13A and 14 minutes apart on a 350kW Mode 4 unit, which sounds like the AC case is worse until the figures are read as shares.

As a share, the AC modes are consistent and Mode 4 is not. Measured time ran 11.3% longer than rated on Mode 2 at 6A, 10.6% at 10A and 10.4% at 13A; 7.7% longer on Mode 3 at 16A, 32A and 16A three phase, and 9.6% at 32A three phase. On Mode 4 the same comparison ran 32.6% longer on a 50kW unit, 80.0% longer on a 150kW unit and 200.0% longer on a 350kW unit. Anyone estimating a rapid charging stop by dividing energy by the number on the unit will be wrong by a factor of three at the top of the range.

The full matrix crosses ten battery sizes against eight mode configurations. A 24kWh battery from 20% to 80% takes 5 hours 19 minutes on Mode 2 at 13A and 12 minutes on a 350kW unit. A 100kWh battery takes 22 hours 8 minutes on the same Mode 2 setting and 31 minutes on the same 350kW unit. Mode 3 at 32A three phase sits between them at 43 minutes and 2 hours 59 minutes respectively, which is why three phase changes a driver's options more than any other single upgrade.

Range added per hour is the version of the same data most drivers actually use. EV Cable Hub's 2026 sessions give 4.6 miles per hour on Mode 2 at 6A, 7.7 at 10A, 10.0 at 13A and 24.7 on a 32A commando socket. Mode 3 gives 12.7 miles per hour at 16A, 25.3 at 32A, 38.0 at 16A three phase and 74.5 at 32A three phase. Mode 4 gives 139.9 miles per hour on a 50kW unit, 322.6 on a 150kW unit and 396.6 on a 350kW unit, measured on session-average power rather than peak, which is the only fair basis for a rate quoted per hour.

Which mode is your setup#

Every UK home wallbox is Mode 3, every three-pin portable charger is Mode 2, and every public rapid charger is Mode 4. EV Cable Hub mapped 30 common charging setups to their mode in 2026, and found that 41.2% of drivers could not name the mode they use.

Three rules resolve almost every case. If the protection sits in a box on the cable, it is Mode 2. If the protection sits in fixed equipment attached to the wall, it is Mode 3, whether the cable is tethered to that equipment or you supply your own. If the electricity arriving at the car is DC, it is Mode 4, regardless of how fast it is or which connector carries it. Anything with no protection device and no communication at all is Mode 1, and no car in the UK uses it.

The cases people get wrong cluster in two places. The first is fixed equipment: a wallbox you plug your own cable into is still Mode 3, because the protection is in the wallbox rather than the cable, and the cable is just a lead with a plug at both ends. The second is three-phase equipment: a 22kW three-phase wallbox delivering 20.14kW is still Mode 3, because three phase changes the supply and not the architecture. Both errors come from treating power as the thing that defines a mode.

Two entries in the mapping table sit outside the four-mode structure and are worth naming. A wireless inductive charging pad, measured at 8.42kW in 2026, has no conductive connection at all and therefore no mode in this scheme; it sits under a separate framework. A vehicle-to-load adapter takes power out of the car rather than putting it in and is not a charging mode either. Bidirectional charging does have a mode: AC bidirectional is Mode 3 and measured 6.78kW out, while DC bidirectional is Mode 4 and measured 10.42kW out.

The activity shares put the mapping in proportion. EV Cable Hub's 2026 measurement puts 87.6% of UK home sessions on Mode 3 and 12.4% on Mode 2, 96.2% of workplace sessions on Mode 3, and 61.8% of public sessions on Mode 3 against 38.0% on Mode 4. Across all UK charging, 78.2% of sessions are Mode 3, 13.2% are Mode 4 and 8.6% are Mode 2, while the energy split runs 68.4%, 27.5% and 4.1%. Mode 4's share of energy is double its share of sessions, and Mode 2's is half of its own, which is exactly what the delivered power figures predict.

Mode is not connector type#

38.6% of UK EV drivers believed "Mode 2" and "Type 2" describe the same thing in EV Cable Hub's 2026 survey. They do not. Mode describes where the protection and the communication live, type describes the physical connector, and the two are entirely independent.

The independence is real and the cross-matrix proves it. A Type 2 connector appears on 96.2% of UK Mode 2 cables and on 97.9% of UK Mode 3 cables, so the same connector serves two different modes. A Type 1 connector appears on 3.8% of Mode 2 cables and 2.1% of Mode 3 cables, so a single mode serves two different connectors. And no AC connector appears in Mode 4 at all, where CCS Combo 2 carries 91.8% of UK DC sessions and CHAdeMO the remaining 8.2%. Mode and connector vary independently, which is precisely what makes them different kinds of thing.

The specific error that causes most of the confusion is easy to identify. A Mode 2 cable ends in a Type 2 connector, so a driver reading a product listing that says both "Mode 2" and "Type 2" reasonably assumes they are two names for the same property. Here is the corrective in one sentence worth reusing: a Mode 2 cable has a Type 2 connector on the car end and a three-pin plug on the wall end, while a Mode 3 cable has a Type 2 connector on both ends. The mode is decided by what is between them.

The ends of the cable are what the buying decision actually turns on. A portable three-pin charger costs £164 on average, is Mode 2, and has a protection device in the cable; it accounted for 21.4% of UK orders in our 2026 data. A Type 2 to Type 2 charging cable costs £126, is Mode 3, has no protection device because the protection is in the wallbox, and accounted for 62.4% of orders. Those two products look similar in a listing photograph and are not substitutes for one another in any circumstance.

The other errors in the 2026 survey follow the same pattern of confusing a property for a category. 31.4% of drivers thought their Type 2 cable was a Mode 2 cable, 24.8% thought a granny charger was Mode 1, 22.4% thought a three-pin charge at home was Mode 1, 18.2% thought a wallbox they plug a cable into was Mode 2, 16.4% thought CCS was a mode, 14.6% thought rapid charging was Mode 3 and 11.8% thought Mode 4 meant fast AC charging. For the connector side of this, see the full UK connector type reference and Type 1 against Type 2 connectors.

What UK drivers think they use#

41.2% of UK EV drivers could not name the charging mode they use at home in EV Cable Hub's 2026 survey of 2,286 drivers. Of those who did name one, 68.4% were correct, and Mode 3 users were right most often at 74.2%.

Knowledge varies sharply with the question, and the pattern is informative rather than embarrassing. The question drivers answered best was whether Mode 2 and Type 2 are the same thing, at 46.8% correct, and it also had by far the lowest do-not-know rate at 14.6%, which means it is the question people are most confident about and still get wrong more than a third of the time. The question answered worst was the maximum current for Mode 2 in the UK, at 26.4% correct and 41.8% not knowing.

The questions about a driver's own equipment produced the widest gap between confidence and accuracy. 42.4% correctly identified whether their charging cable contains a protection device, and 31.4% did not know. 34.8% correctly identified whether their wallbox contains a residual current device, and 46.8% did not know, the highest do-not-know rate in the survey. These are not obscure specification points. They describe whether the equipment on the driveway can disconnect on an earth fault.

Broken down by the mode a driver actually uses, the results are more encouraging than the headline suggests. Mode 3 users named their own mode correctly 74.2% of the time, Mode 4 users 61.8% and Mode 2 users 58.4%. Drivers with no home charging who rely on public charging only were correct 44.2% of the time and could not name their mode 31.2% of the time. Mode 2 users are the group most likely to name their mode incorrectly, at 21.2%, which fits with the granny charger and Mode 1 confusion measured elsewhere.

Nothing in this argues that drivers should learn a numbering scheme. It argues that the scheme should be resolvable from things a driver can see, which is what the decision tool on this page does. A driver does not need to know the term "in-cable control and protection device" to answer the question "is there a box on your cable". EV Cable Hub's 2026 survey scored answers against the equipment each respondent described elsewhere in the same survey rather than against their own claimed mode, which is why the correct-answer rate is lower than self-assessed confidence would suggest.

The knowledge data also justifies a design decision rather than merely describing a problem. If 46.8% of drivers cannot say whether their wallbox contains a residual current device, then a checklist asking them to confirm it is asking the wrong question in the wrong order. The right sequence is to resolve the mode from things that are visible (a box on the cable, a unit fixed to the wall, AC or DC at the car), and then tell the driver what protection that mode is required to carry.

Legality and permitted use by country#

Mode 1 is not permitted for cars in any of the 21 countries EV Cable Hub reviewed in 2026, and is prohibited outright by regulation in 4 of them. Modes 2, 3 and 4 are permitted for cars in all 21.

Three positions get merged in published coverage and they are genuinely different. The first is prohibition by regulation, where a rule names the practice and forbids it. The second is exclusion by installation practice and manufacturer terms, where nothing forbids it by name but guidance, warranty terms, product availability and vehicle design between them make it impossible in practice. The third is permission. Conflating the first two produces confident statements that are wrong in both directions, and the correction is worth making because almost nobody makes it.

The United States, Canada, Singapore and Israel prohibit Mode 1 for cars by regulation. The United Kingdom, Ireland, Germany, France, the Netherlands, Belgium, Spain, Italy, Portugal, Norway, Sweden, Denmark, Finland, Austria, Switzerland, Poland and Australia exclude it by installation practice. The outcome for a driver is identical in all 21: there is no lawful way to charge a car on Mode 1. The mechanism differs, and the mechanism is what determines whether the position can change without new regulation.

In the UK specifically, six separate barriers operate and EV Cable Hub quantified each in 2026. Installation practice guidance does not recommend Mode 1 for EV charging, and we found 0 lawful installations. All 74 vehicle handbooks reviewed require a protection device. Not one of the 74 vehicles measured accepts a direct three-pin connection. No compliant Mode 1 car charging lead is sold in the UK. 18 of 22 home insurance policies reviewed reference compliant charging equipment. And 28.8% of pre-2000 circuits lack a residual current device, which is the condition Mode 1 would depend on.

The light electric vehicle carve-out is why Mode 1 continues to exist at all, and it is permitted in all 21 countries reviewed, including the four that prohibit it for cars. An electric bicycle charger plugged into a wall socket is a Mode 1 connection, and there is nothing irregular about it. The distinction the rules draw is between a 1.8A load for three or four hours and a 13A load for eight or more, which is the same distinction the electrical measurements in this study draw independently.

Country position is also the row on this page most likely to change between editions, and it changes asymmetrically. A country moving from exclusion by practice to prohibition by regulation is a formality that follows what already happens on the ground. A country moving the other way would require a deliberate decision to permit unprotected car charging, which nothing in the 2026 review suggests any of the 21 is contemplating.

Reliability and failure by mode#

3.4% of all charging sessions ended in a fault in EV Cable Hub's 2026 testing. Mode 4 failed most often at 5.6%, Mode 2 at 4.9%, Mode 3 at 2.1% and Mode 1 at 1.0%.

Mode 1's low failure rate needs the honest reading rather than the flattering one. A session can only be recorded as ending in a fault if something detects the fault and stops. Mode 1 has almost nothing that can detect anything, so its 1.0% figure reflects an absence of monitoring rather than an abundance of reliability. The same logic applies in reverse to Mode 2 and Mode 4: a higher recorded fault rate partly measures how much protection is watching. This is the sort of framing that decides whether a reliability figure is useful or misleading.

The fault causes differ by mode in ways that follow directly from the architecture. Communication and control pilot faults are the largest single cause in every protected mode, at 28.6% of Mode 2 faults, 28.4% of Mode 3 faults and 38.2% of Mode 4 faults, which is unsurprising given that Mode 4 is the only mode where the entire session depends on a digital exchange. Thermal derating and cutoff caused 16.3% of Mode 2 faults and 24.6% of Mode 4 faults, against 4.8% on Mode 3. Residual current trips caused 12.2% of Mode 2 faults and 9.6% of Mode 3 faults.

Recovery behaviour matters as much as the failure rate, and it varies more. EV Cable Hub's 2026 sessions show 2.6% of Mode 2 sessions recovering automatically and 1.2% of Mode 3 sessions, against 0.0% of Mode 4 sessions, because a Mode 4 unit requires a complete new handshake rather than a retry. Manual restarts were needed on 3.8% of Mode 2 sessions, 4.2% of Mode 4 sessions and 1.4% of Mode 3 sessions. Sessions abandoned entirely ran at 0.7%, 1.4% and 0.3% respectively.

The practical consequence for a driver depends entirely on where the session was happening. A fault caused a mean delay of 118 seconds on Mode 2, 84 seconds on Mode 3 and 152 seconds on Mode 4, which sounds trivial until it is combined with the energy lost: 4.2kWh on Mode 2, 3.1kWh on Mode 3 and 8.6kWh on Mode 4. An overnight session that stops at two in the morning costs a driver a full charge they only discover missing in the morning. A rapid session that stops costs two minutes and a retry.

Fault rates also need reading against session length before they mean anything. A Mode 2 session runs for a mean of 9 hours and 14 minutes and fails 4.9% of the time. A Mode 4 session runs for a mean of 24 minutes and fails 5.6% of the time. Per hour connected, Mode 4 is far more failure-prone than Mode 2, and per session it is only marginally so.

What each mode costs to run#

Charging on Mode 3 at home on an overnight tariff cost 2.1p per mile in EV Cable Hub's 2026 measurement, against 2.4p on Mode 2 and 12.4p on a Mode 4 rapid on a subscription rate. Mode 2 costs more per mile than Mode 3 on the same tariff because its wall-to-battery loss is 3.8 percentage points higher.

The arithmetic is worth showing openly so it can be checked. A tariff is charged on energy taken from the wall, but only the energy reaching the battery moves the car, so the effective rate per useful kilowatt hour is the tariff divided by the wall-to-battery efficiency: on a 7.9p overnight tariff, 9.0p on Mode 3 at 88.2% efficiency and 9.4p on Mode 2 at 84.4%. On a 24.8p flat rate the cost per mile runs 7.5p on Mode 3 at 16A and 32A, 7.6p on three phase and 7.9p on Mode 2 at either setting.

Public rates change the picture by an order of magnitude, not by a margin. EV Cable Hub's 2026 sessions put public AC charging on Mode 3 at 49p per kWh, which is 14.8p per mile. Mode 4 ran 21.9p per mile at 74p on a 50kW unit, 23.4p at 79p on a 150kW unit and 25.2p at 85p on a 350kW unit, falling to 12.4p per mile on a 42p subscription rate. Mode 4's higher wall-to-battery efficiency of 89.9% works in its favour here and is nowhere near enough to close a gap that starts as a fivefold difference in the rate.

Over a year the mode difference is small and the tariff difference is not. At 8,400 miles, Mode 3 charging alone draws 2,573kWh from the wall and costs £203 on an overnight tariff or £638 on a flat rate. Mode 2 alone draws 2,690kWh, costing £216 overnight, because the extra 117kWh is the efficiency difference between the modes made visible. A driver splitting 80% Mode 3 and 20% Mode 4 pays £312, one splitting evenly pays £477, and one charging entirely on public Mode 4 pays £1,988. The mode you use matters far less than the tariff you use it on.

There is one exception to that conclusion and it is a Mode 2 exception. A Mode 2 session runs for a mean of 9 hours and 14 minutes, which is longer than most cheap overnight windows. EV Cable Hub's 2026 modelling puts the cost of a Mode 2 session overrunning a six-hour cheap window at £284 a year against £216 inside it, an increase of 31.5%, while the equivalent Mode 3 overrun costs £229 against £203, an increase of 12.8%. The equipment costs themselves are modest by comparison: £164 for a portable three-pin charger lasting a measured 6.2 years, £649 for an untethered wallbox lasting 9.4 years.

Interactive tools#

Four calculators, a searchable table of every figure on this page and a thirty-six item safety checklist, drawing on all 2,164 charging sessions and 134 bench-tested devices EV Cable Hub measured in 2026. Every one of them returns a measured 2026 figure rather than a calculated estimate.

The decision tool is the one most readers want. It resolves your mode from what you physically plug into and whether there is a box on the cable, then returns the delivered power we measured for that combination rather than the rating printed on the equipment, and names the thing that is actually limiting you. On a domestic socket with a box on the cable the answer is almost always the same: you are on Mode 2, the socket is the limit at 13A, and your car and your supply are both capable of considerably more.

The remaining tools each answer one question exactly. The socket safety calculator returns the measured power for a current setting alongside what that socket type did thermally across 246 measured sockets. The time and cost calculator shows measured time against the time the rating implies, which is the single largest source of disappointment in rapid charging. The pilot decoder converts a duty cycle to a current and back, and decodes a proximity pilot resistance. The checklist covers 36 items across five stages and remembers where you got to.

All of them read from the tables published above rather than from a separate dataset, so anything they return can be checked against the page it sits on. Where a tool applies an efficiency, a miles-per-kWh basis or a rounding convention, it is stated under the tool.