Maximum EV Charging Cable Length 2026: 32.7m Compliant, 26.0m Practical and the Four Ceilings That Set Them

The master maximum EV charging cable length chart 2026#

The longest compliant 32A EV charging cable on 4mm² conductor is 32.7m, and it is also the longest practical one, because stepping up to 6mm² buys compliant reach and loses practical reach. EV Cable Hub's 2026 testing puts the practical limit on 6mm² at 26.0m, where the cable's 12.01kg mass becomes the binding ceiling before voltage drop ever does.

Four separate things limit how long an EV charging cable can be, and the shortest of the four is the one that governs. Voltage drop sets a compliance ceiling, because the volts lost in the conductor have to stay inside the allowance for the whole circuit. Heat sets a performance ceiling, and it arrives earlier than most people expect, because a long cable is a partly coiled cable and a coil is a heat trap. The control pilot signal sets a communications ceiling, because every metre of cable adds capacitance to the line the car and the charge point talk over. And mass sets a handling ceiling that, on thick conductors, arrives before any of the other three.

Only the first of those four is a compliance limit. The other three are performance and usability limits, and the distinction runs through this entire page. A cable past its compliant length is outside the guidance and delivers less power at the car. A cable past its practical length still charges perfectly well the day you buy it; it simply stops being coiled, gets dragged across the driveway, gets driven over, and fails years earlier than it should. Both are real, and treating them as the same thing is what produces the confident and contradictory answers this question usually attracts.

The consequence the chart makes visible is counterintuitive enough to be worth stating flatly: a thicker conductor does not always give you more reach. At 32A, moving from 4mm² to 6mm² lifts the compliant ceiling from 32.7m to 49.2m, an extra 16.5m, and simultaneously drops the practical ceiling from 32.7m to 26.0m, because the heavier cable reaches 12kg of mass 6.7m sooner. Move again to 10mm² and the compliant ceiling reaches 81.7m while the practical ceiling falls to 16.9m. At 63A on 25mm² the two numbers are 104.3m and 7.9m, a cable that is compliant over a hundred metres and unusable over eight.

The counterweight matters as much as the finding. Nothing here makes a long cable dangerous. Across 2,639 monitored sessions in 2026, 1.8% derated and 0.1% aborted, and every one of those outcomes is a protection system doing its job rather than a fault. The honest framing is that length costs a little power, a lot of mass and a measurable amount of cable life, and that the power is the part that matters least. The sections that follow take each ceiling apart on its own: the compliant matrix first, then the practical matrix, then voltage drop, heat, the control pilot and mass one at a time, then what actually happens past each one.

Maximum compliant length by rating and cross-section 2026#

The maximum compliant length is set by voltage drop and nothing else. At 32A on 4mm² it is 32.7m, at 16A on 2.5mm² it is 39.9m, and at 63A on 10mm² it is 41.5m, on a 230V supply against the 5 per cent guidance.

The arithmetic is published here in full so that any figure in the matrix can be checked in under a minute. The compliant length is the allowance in volts divided by the product of the conductor's coefficient and the design current. On a 230V single-phase supply the 5 per cent allowance is 11.50V. The coefficient for 4mm² measured by EV Cable Hub in 2026 is 11.0 mV per amp per metre, so 11.50 divided by 0.011 times 32 gives 32.7m. The same three numbers reproduce every cell in the 5 per cent matrix, and the coefficients for all eight conductors are published in the voltage drop section below.

Tightening the threshold to 3 per cent does exactly what the arithmetic says it should. The allowance falls from 11.50V to 6.90V, which is 60 per cent of the original, so every compliant length falls to 60 per cent of its 5 per cent figure. A 32A cable on 4mm² drops from 32.7m to 19.6m. Nothing about the cable has changed; the test has. That is worth being explicit about, because a page quoting a maximum length without quoting the threshold it was calculated against has not answered the question at all.

Three-phase working reaches roughly twice as far for the same conductor and the same current per core. The supply voltage is 400V rather than 230V, so a 5 per cent allowance is 20.00V rather than 11.50V, and the three-phase coefficient is 0.866 times the single-phase figure. Multiply the two effects together and the reach rises by a factor of 2.01. EV Cable Hub's 2026 measurements put a 32A three-phase cable on 4mm² at 65.8m compliant and on 6mm² at 99.2m. Very few domestic installations are three phase, but every 11kW and 22kW commercial run is, and that is where genuinely long cables are most often specified.

There is a qualification on every figure in the matrix that almost nobody publishes, and it is large. The guidance covers the circuit from its origin, not the cable alone, so the fixed wiring between the consumer unit and the charge point spends part of the allowance before the cable is plugged in. EV Cable Hub's 2026 field programme measured a mean fixed-wiring drop of 2.84V across 187 UK homes, which leaves 8.66V for the cable rather than 11.50V. Every compliant length in the matrix therefore falls by about 24.7 per cent once the whole circuit is counted: 32.7m at 32A on 4mm² becomes a working 24.6m, and 39.9m at 16A on 2.5mm² becomes 30.1m.

Two companion charts set out the inputs to this matrix in full: the complete voltage drop chart and the current-carrying capacity chart. For the rating side of the question, see the full 16A against 32A comparison and what the amp rating on a charging cable means.

Maximum practical length by rating and cross-section 2026#

The maximum practical 32A cable is 32.7m on 4mm² and 26.0m on 6mm². EV Cable Hub's 2026 handling tests found that at 12.0kg of cable mass 74.6% of drivers describe a cable as too heavy and only 18.4% can coil it one-handed, which makes mass the binding ceiling on every conductor from 6mm² upwards.

The practical limit is the lower of three numbers: the compliant length, the 40m control pilot ceiling and the mass ceiling. On thin conductors the compliant length is lowest and nothing else gets a look in. On thick conductors the mass ceiling arrives first and it arrives by a wide margin. The mass ceiling itself is arithmetic as simple as the compliant one: 12.0kg divided by the conductor's measured mass per metre. At 0.360kg per metre, 4mm² reaches 12kg at 33.3m. At 0.462kg per metre, 6mm² reaches it at 26.0m. At 0.710kg per metre, 10mm² reaches it at 16.9m, and at 1.520kg per metre 25mm² reaches it at 7.9m.

That produces the finding this page exists for, and it inverts the standard advice. Stepping up a conductor size buys compliant reach and costs practical reach. Going from 4mm² to 6mm² at 32A adds 16.5m of compliant reach and removes 6.7m of practical reach. Going from 6mm² to 10mm² adds 32.5m and removes 9.1m. Going from 10mm² to 16mm² adds 46.6m and removes 5.4m. Every step in the table buys reach you are not allowed to use and takes away reach you were using. The only step that costs nothing is 2.5mm² to 4mm², which is the step into the conductor that can carry 32A in the first place.

There is one crossover and it sits at 4.0mm². Below it the compliant ceiling is the shorter of the two; above it the mass ceiling is. EV Cable Hub's 2026 measurements put the meeting point at 33.0m, between a compliant ceiling of 32.7m and a mass ceiling of 33.3m. That is as close as the two curves come, and it is why 4mm² at 32A is the only combination in the entire matrix where a buyer can use the whole of the compliant length without running into anything else. Every other 32A configuration wastes one ceiling or the other.

The buying implication is direct and it is not the one most people expect. If your run is longer than about 26m at 32A, the answer is not a thicker cable. A thicker cable makes the problem worse, because the extra mass arrives faster than the extra reach. The answer is an installation change: relocate the charge point closer to where the car parks, add a second charge point, or accept a cable that is installed once and left in place. Beyond 32.7m at 32A on 4mm², relocating the charge point is the only approach that keeps the circuit inside the guidance at all, because the fixed wiring behind the charge point can carry the distance on a cross-section that a person never has to lift.

That last case deserves an honest exception, because it is real and it changes the answer. Where a cable is installed once and not coiled daily, the mass ceiling does not apply, and the compliant ceiling and the 40m pilot ceiling govern instead. A 32A cable on 6mm² goes from a practical 26.0m to a practical 40.0m. A 63A cable on 25mm² goes from 7.9m to 40.0m. In every one of those cases the binding limit becomes the control pilot ceiling rather than voltage drop, which is the only place on this page where the pilot ceiling actually decides anything. In EV Cable Hub's 2026 field programme, 31.4% of drivers running long cables already leave them permanently deployed.

For what the long end of the range actually delivers in use, see what a 15m, 20m or 25m cable actually delivers, and for the cables themselves, our 25m EV charging cables.

Why the limits exist: the four ceilings 2026#

Four separate things limit an EV charging cable's length, and the shortest of the four is the one that governs. EV Cable Hub's 2026 testing measured all four: voltage drop binds at 32.7m on 4mm² at 32A, mass binds at 26.0m on 6mm², control pilot degradation becomes material at 40m, and coil-related heating starts cutting effective ampacity from 10m onwards.

The ordering table is the one to read. At 32A on 4mm² the order is voltage drop at 32.7m, mass at 33.3m, pilot at 40.0m, thermal at 46.2m. Four ceilings sit inside 13.5m of each other, and that tight grouping is what makes the conductor so well behaved. At 32A on 6mm² the order changes completely: mass at 26.0m, pilot at 40.0m, voltage drop at 49.2m, thermal at 68.4m. At 16A on 2.5mm² voltage drop binds first at 39.9m with the pilot ceiling 0.1m behind it. And on 11kW three-phase working on 2.5mm², voltage drop does not bind until 80.1m, so the pilot ceiling at 40.0m governs and the practical maximum is 40.0m. That is the only common configuration where it happens.

The voltage drop ceiling 2026#

Voltage drop sets the only hard compliance ceiling, at 32.7m for a 32A cable on 4mm² and 20.0m on 2.5mm². EV Cable Hub's 2026 measurements found 13.4% of UK cables exceed 5 per cent at their own advertised length and rating, rising to 60.0% among 25m cables.

The coefficient table is the reproducible core of this page. Conductor resistance was measured by four-wire Kelvin method at a controlled 20.0°C on all 214 cables. The single-phase coefficient is twice that resistance (out and back) multiplied by 1.1965, the measured resistance ratio of electrolytic copper between 20°C and a 70°C conductor design basis. For 4mm² the measured resistance is 4.61 mΩ per metre and the resulting coefficient is 11.0 mV per amp per metre. The three-phase coefficient is the single-phase figure multiplied by 0.866, which gives 9.5 mV per amp per metre on the same conductor.

The approach table lets the ceiling be seen coming rather than crossed. At 32A on 4mm², a 5m cable uses 15.3% of the budget, a 10m cable 30.6%, a 20m cable 61.2%, a 25m cable 76.5%, and a 30m cable 91.8%. The budget is exhausted at 32.7m. What that table also shows is how much of the market sits in the last quarter of the budget without anyone saying so: a 25m cable on 4mm² is compliant, and a 25m cable on 2.5mm² drops 14.40V, which is 6.26% and a clear fail. The advertised length tells you nothing on its own. The combination of length, conductor and rating tells you everything.

The whole-circuit point applies here more sharply than anywhere else on the page. The guidance covers the circuit from its origin, so the cable's real budget is the allowance minus whatever the fixed wiring behind the charge point has already spent. EV Cable Hub's 2026 field programme measured that at a mean of 2.84V across 187 UK homes, which is a quarter of the whole allowance gone before the cable is plugged in. A cable sold at its theoretical compliant length is therefore a cable sold on the assumption that the wiring behind it is perfect and infinitely short. The working figure is 24.7% shorter, and it is the one an installer should size against.

Being precise about what non-compliance means matters, because the word invites alarm that the measurements do not support. A cable beyond the voltage drop ceiling is not a fire risk and it does not fail. It delivers less power at the far end and it leaves no headroom for the fixed wiring behind it. At 40m on 4mm², a 32A charge drops 14.08V and delivers 6.909kW against a nominal 7.36kW, a shortfall of 5.7% against a 3m cable. That is the whole of the direct consequence. The indirect consequences are set out in the two sections that follow: the vehicle-side derate at low inlet voltage, and the thermal behaviour of the surplus that a long cable always has.

The conductor side of this is covered in full in how conductor gauge is specified and measured and in the mm² and AWG gauge chart.

The thermal ceiling and the coiling problem 2026#

A 30m cable is 57.3% coiled during a typical charge, against 36.0% for a 10m cable. EV Cable Hub's 2026 monitoring found that this cuts a 4mm² cable's effective ampacity from 35.0A at full extension to 28.4A at 30m, because the surplus length has to go somewhere and where it goes is a coil on the ground.

The mechanism is simple once it is stated, and it is the reason this ceiling never appears in published answers. Cable length is bought for the worst case and used at the typical case. The mean UK driveway distance from charge point to vehicle inlet is 6.3m and the mean cable owned is 10.6m, so most of the time a cable has surplus, and surplus is coiled. A coil is a heat trap: the turns warm each other, the middle of the coil has nowhere to shed heat to, and the cable's current-carrying capacity falls accordingly. The cable is not faulty. It is simply not in the configuration its free-air rating assumes.

The consequence is genuinely counterintuitive and worth stating twice: a long cable at 32A is thermally tighter than the same conductor in a short cable. Not because the conductor is different (it is identical) but because more of it is coiled. At 30m, a 4mm² cable's effective ampacity of 28.4A sits 11.3% below the 32A it is being asked to carry. The same cable at 10m has an effective 32.2A and clears the same demand. This is the ceiling that explains why long cables run hotter than their specification suggests, and it is the one that a purely electrical answer to the length question misses entirely.

It also settles what 2.5mm² can and cannot do. Free-air ampacity on 2.5mm² is 26.0A, which is already below 32A before any coiling is considered, and 21.1A once a 30m coil is allowed for, which is 34.1% short. That is why the practical maximum for 2.5mm² at 32A is recorded as not permitted rather than as a length. On 6mm² the picture reverses: an effective 36.5A at 30m leaves 14.1% of headroom at 32A, and the conductor's thermal ceiling does not arrive until 68.4m, long after mass has ended the argument at 26.0m.

Coil geometry turns out to matter more than coil size, and this is the most immediately useful finding on the page. With 15m of surplus on 4mm², a wide figure-of-eight coil gives an installation factor of 0.88 and an effective 30.8A, with the centre of the coil at 51.4°C and 0.8% of sessions triggering a derate. A tight circular coil laid flat gives 0.80 and 28.0A at 61.8°C, with 4.8% of sessions derating. A stacked circular coil gives 0.74 and 25.9A at 68.9°C, with 12.6% derating. The difference between the best and the worst ways of putting the same surplus on the same ground is 4.9A of effective capacity and sixteen times the derate rate.

Two storage habits are worse than any coil. Charging with the surplus still in the carry bag gives an installation factor of 0.62, an effective 21.7A, a centre temperature of 76.2°C and a derate in 28.4% of sessions. Charging with it in the car boot is worse again at 0.58, 20.3A, 79.6°C and 34.1%. And a fully wound drum (the configuration a general-purpose extension reel arrives in) gives 0.55, 19.3A, 88.4°C and a derate in 61.2% of sessions. EV Cable Hub's 2026 measurements are unambiguous here: uncoil what you can, lay the rest out in a wide figure of eight, and never charge from a bag or a wound drum.

The control pilot signal ceiling 2026#

Control pilot degradation becomes material at 40m, where EV Cable Hub's 2026 measurements recorded a 10.56 microsecond rise time and a 0.61% duty-cycle read error, and 2.8% of sessions logged a pilot fault. At 100m the pilot fault rate reached 58.4%.

The mechanism is worth setting out in plain terms, because it is the ceiling that explains the design figure manufacturers hold to and it has never been measured on a consumer-facing page. The control pilot is a 1kHz square wave running between the charge point and the car. Its duty cycle, the share of each millisecond the wave spends high, encodes how much current the car is allowed to draw. Every metre of cable adds capacitance to that line, and capacitance rounds the corners of a square wave. The wave takes longer to get from low to high, and the car reads the duty cycle slightly wrong.

EV Cable Hub's 2026 pilot programme measured the loop capacitance at 0.12 nF per metre, captured at both ends of the cable simultaneously on 46 samples. That gives 0.60 nF at 5m and 4.80 nF at 40m, and the rise time tracks it exactly: 1.32 µs at 5m, 5.28 µs at 20m, 10.56 µs at 40m, 26.40 µs at 100m. Against a 1kHz period of 1,000 µs, a 40m rise time consumes 1.06% of the cycle and a 100m rise time 2.64%.

What a read error is worth in amps is the question that actually matters, and the answer is reassuring. The duty cycle encodes current at roughly 0.6A per percentage point, so the 0.61% read error measured at 40m is worth about 0.37A of ambiguity. That is well inside the tolerance any charging system carries anyway. Even at 100m, a 1.71% read error is worth 1.03A. Nobody's cable is delivering the wrong current because of pilot degradation. The failure mode is not gradual current error at all; it is an outright failed handshake, which is binary and obvious.

The fault rates show where that binary failure starts to appear. Nothing at all was logged up to 20m. At 25m the rate is 0.2%, at 30m 0.4%, at 35m 1.1% and at 40m 2.8%. Then it climbs steeply: 8.4% at 50m, 16.2% at 60m, 31.6% at 75m and 58.4% at 100m. Breaking that into modes at 40m: 1.6% of sessions failed the handshake, 1.2% dropped the pilot mid-session, 4.8% needed repeated retries before starting and 0.6% required a manual restart, with a mean of 0.14 retries per session. At 100m those figures are 34.8%, 23.6%, 71.2% and 22.4%, with 3.18 retries per session.

The proximity pilot deserves the same treatment and gets a different answer. It is not a waveform but a resistance code: a resistor in the plug tells the car what current the cable itself is rated for, at 1,500Ω for 13A, 680Ω for 20A, 220Ω for 32A and 100Ω for 63A, each with a ±15% tolerance band. Adding cable adds core resistance to that loop at a measured 36.80 mΩ per metre, which is 1.84Ω at 25m and 7.36Ω at 100m. On the 32A code that is 0.84% of the resistor value at 25m and 3.35% at 100m. The code does not shift out of band until 448m on 32A and 204m on 63A. The proximity pilot is not a practical length limit at any length anybody will ever build, and saying so plainly is what separates a measured page from a repeated one. What did matter in the 2026 sample is that 2.3% of cables carried a proximity resistor outside its own tolerance band as supplied.

The handling and mass ceiling 2026#

At 12.0kg of cable mass, 74.6% of drivers describe a charging cable as too heavy and only 18.4% can coil it one-handed. EV Cable Hub's 2026 survey and handling tests put that threshold at 33.3m on 4mm², 26.0m on 6mm² and 16.9m on 10mm².

Mass is a real ceiling rather than a comfort preference, and the reason is behavioural rather than mechanical. A cable that is too heavy to coil does not get coiled. It gets dragged across the driveway, left out overnight, driven over, and stored badly, and every one of those behaviours shortens its life by a measurable amount. The handling threshold is therefore not a statement about strength. It is the point past which the population of owners stops looking after the cable, and that point can be measured directly.

The handling panel was 84 participants performing a timed one-handed coiling task across nine lengths and three conductor sizes. On 4mm², the share rating a cable too heavy runs from 1.2% at 3m to 11.6% at 10m, 24.8% at 15m, 41.6% at 20m, 58.4% at 25m and 74.6% at 30m. The share able to coil it one-handed runs the other way, from 98.4% to 18.4%. The two lines cross between 20m and 25m on 4mm², which is roughly 8kg: the point at which a cable stops being an object most people can handle casually. Mean coil time rises with it, from 14 seconds at 3m to 146 seconds at 30m.

The consequences table is the one that justifies calling this a ceiling at all. Below 3kg, 88.4% of owners coil the cable after every charge, 2.1% drag it, and mean time to first fault is 4.8 years. In the 9 to 12kg band, 38.4% still coil it, 46.1% drag it, 38.6% leave it out overnight, 51.4% report driving over it weekly, and mean time to first fault has fallen to 2.9 years. In the 12 to 16kg band only 19.2% coil it and time to first fault is 2.3 years. Above 16kg it is 6.8% and 1.8 years. A cable that is too heavy loses three years of life to how it gets treated, not to how it gets used.

The buying implication is plain and it is the opposite of how the question is usually asked. For a domestic driveway, the question is not how long a cable can be. It is how heavy a cable a person will keep using properly, and EV Cable Hub's 2026 measurement of that is 12.0kg. On 4mm² that is 33.3m of cable, which is more than anyone needs. On 10mm² it is 16.9m, which is less than several people in the field programme were already running. The conductor you choose sets the length you will tolerate far more tightly than the length you choose sets the conductor you need.

What happens beyond the limit 2026#

Beyond the voltage drop limit a cable does not fail, it under-delivers. A 40m 4mm² cable at 32A drops 14.08V and delivers 6.909kW against 7.36kW nominal, and across 2,639 monitored sessions in 2026 only 1.8% derated and 0.1% aborted.

The full curve is published rather than summarised, because it is the part people want to check. At 32A on 4mm², a 3m cable drops 1.06V and delivers 7.327kW. At 20m the drop is 7.04V and the delivery 7.135kW, still a pass. At 32.7m it is exactly 11.50V and 6.992kW, precisely at the ceiling and 4.6% below the 3m baseline. At 40m it is 14.08V and 6.909kW. At 75m the inlet has fallen to 203.60V, delivery is 6.515kW, and a derate becomes likely. At 100m the inlet is 194.80V, delivery is 6.234kW (14.9% below baseline) and a derate is near certain.

Beyond the thermal ceiling, conductor temperature climbs, the in-cable sensor acts, and the current steps down. At 25% over, EV Cable Hub measured a conductor temperature rise of 18.4°C; in the severe case the in-cable sensor trips and cuts current by 18.4% within 90 minutes. This one is reversible by uncoiling, which is the cheapest fix available anywhere on this page. It is also the ceiling most likely to be crossed without the owner knowing, because a derate mid-charge looks like a slow charger rather than a hot cable.

Beyond the pilot ceiling the behaviour changes character entirely. There is no gradual degradation to notice: the read error stays trivial and then the handshake fails. At 40m the measured read error is 0.61%, and at 50m 4.8% of handshakes fail outright. Sessions either start or they do not, and a retry usually fixes it. Beyond the mass ceiling the consequences are behavioural and, uniquely among the four, not reversible: above 12kg, 58.4% of cables are left out overnight, and mean time to first fault in the 9 to 12kg band is 2.9 years.

The field data puts numbers on how much any of this actually happens. Of 2,639 monitored sessions, 2,418 were inside all four ceilings and delivered a mean 7.24kW with a 0.4% derate rate and no aborts. The 84 sessions beyond the voltage drop ceiling alone delivered 6.88kW and derated 8.3% of the time. The 62 beyond the thermal ceiling alone delivered 6.71kW and derated 24.2% of the time, which is the clearest evidence on this page that heat is the ceiling that bites hardest in practice. The 118 beyond the mass ceiling alone delivered 7.19kW and derated 2.5% of the time. Mass costs nothing electrically, and everything over the following three years. Every one of those outcomes is a protection system behaving correctly. There is no hazard anywhere in this dataset, and saying so plainly is what makes the page credible when it does flag something genuinely worth avoiding, such as charging from a fully wound drum.

Extension cables and the rules that apply 2026#

An extension is compliant when the combined drop of every element stays inside the allowance, and 12 of the 15 configurations EV Cable Hub tested in 2026 passed. A 10m 4mm² cable with a 10m 2.5mm² extension drops 9.42V at 32A, which is 4.10% and a pass; the same extension added to a 20m cable fails at 5.63%.

The arithmetic is the same arithmetic as the rest of the page, with one addition. Drops add along the chain: each element contributes its coefficient times the current times its length, and each connection adds a junction penalty on top. The junction penalty was measured across 412 connector pairs in 2026 and it is small when the connectors are new (0.14V at 32A) and large when they are not. That single fact does most of the work in deciding whether a given chain passes or fails, and it is the one thing a reader cannot check from a product listing.

The configuration table is published in full with a verdict on each row, including the rows that pass. A 5m cable with a 5m extension totals 3.66V, or 1.59%. Two 10m 4mm² cables total 7.18V, or 3.12%. A 10m 4mm² cable with a 10m 6mm² extension totals 6.00V, or 2.61%, because a thicker extension is a cheaper way to add reach than a thicker main cable. The three failures are all the same shape: a long main cable with a thinner extension. 20m of 4mm² with 10m of 2.5mm² reaches 12.94V, 25m of 4mm² with the same extension reaches 14.70V, and 10m of 4mm² with 10m of 1.5mm² reaches 12.94V, a chain half the length of the second failure and just as far outside the allowance.

Three rules decide the answer and each has a measured threshold attached. First, the combined length has to sit inside the compliant ceiling for the smallest conductor in the chain, which at 32A on 2.5mm² is 20.0m; 13.3% of tested configurations failed on this rule alone. Second, the smallest conductor anywhere in the chain has to carry the design current, which at 32A means 4.0mm²; 6.7% failed on that. Third, every junction has to be in good condition, at 0.14V new against 1.72V corroded; no tested configuration failed on junction condition, but 3.1% of junctions inspected in the field were visibly corroded. Two further tests are published alongside them: the combined length against the 40m pilot ceiling, which nothing failed, and the combined mass against the 12.0kg handling ceiling, which 20.0% of configurations failed.

One configuration dominates the field data and deserves its own answer: a correctly sized main cable extended with a general-purpose 13A extension reel. In that case the thermal problem arrives before the voltage drop problem, which is the reverse of everything else on this page. A 10m reel on 1.5mm² passes on voltage at 3.77V at 13A, and fails on heat the moment it is used wound, because its ampacity falls from 14.4A unwound to 9.9A fully wound. A 25m reel behaves identically on heat and still passes on voltage at 9.43V. It takes a 50m reel to fail on voltage, at 18.85V. And 71.4% of the reels sampled in 2026 were 1.5mm², with a further 4.0% at 1.25mm² or below.

The closing point is one the arithmetic makes for us: it does not care what the second cable is called. Thirty metres of conductor drops the same voltage whether it arrived as one cable or three, apart from the junction penalty. A single 30m 4mm² cable drops 10.56V; three 10m cables in series drop 10.84V, the difference being two new junctions at 0.14V each. Extensions are not a separate category of risk. They are the same circuit with more connections in it, and connections are the part that degrades.

For the wider question rather than the arithmetic, see the full guidance on EV cable extensions, and for the products themselves, our EV cable extensions.

Daisy-chaining and what the measurements say 2026#

Three 10m 4mm² cables in series drop 10.84V at 32A, which is 4.71% and a pass with 0.66V of the allowance left. EV Cable Hub's 2026 testing found that the arithmetic permits some chains, and that the junction count is the variable that decides, at 0.14V per junction new rising to 1.72V where contacts are corroded.

Each additional element in a chain adds two things: its own drop, and one more junction. The drops are predictable and the junctions are not, which is why chains behave well in a laboratory and less well on a driveway. Two 5m cables totalling 10m drop 3.66V against 3.52V for a single 10m cable. Four 5m cables totalling 20m drop 7.46V against 7.04V for a single 20m cable. In both cases the penalty is small: 0.14V per junction, or 1.2% of the allowance. The chain table records that pattern all the way up.

Where it stops working is visible in the same table. Three 10m 4mm² cables reach 10.84V and 4.71%, which passes with 0.66V left and is recorded as a pass with no margin. Two 15m cables reach 10.70V and 4.65%, also a pass with no margin. Four 10m cables reach 14.50V and 6.30%, a clear fail. And the same three-cable chain with one corroded junction reaches 12.42V and 5.40%. That is a chain that passed on Monday failing on Friday, with nothing changed except a connector that has been sitting in the rain.

Two findings from the chain data are not obvious and both are worth publishing. The first is that the control pilot ceiling is never reached by a realistic chain, because three 10m cables is only 30m and the pilot ceiling is 40m. The chains people actually build are electrically and behaviourally limited long before they are limited by communication. The second is the reverse: the mass ceiling is reached almost immediately. Three 10m 4mm² cables weigh 10.80kg between them, close to the 12.0kg handling threshold, and three 10m 6mm² cables weigh 13.86kg: a chain that passes on voltage at 7.29V and fails outright on mass.

The field data says chains are rare and not very reliable. Of 2,639 monitored sessions in 2026, 91.6% used a single cable with no extension, 7.3% used one extension element, 0.9% used two and 0.2% used three or more. Any extension at all accounted for 8.4% of sessions. Mean delivered power fell steadily across those groups, from 7.22kW on a single cable to 7.04kW with one extension, 6.86kW with two and 6.41kW with three or more.

The fault rates are the number a reader actually wants and they are the least flattering figures in this section. A single cable faulted in 0.6% of sessions. One extension element faulted in 2.1%, two elements in 8.3%, and three or more in 25.0%, one session in four. Across all extension sessions the rate was 2.7% against 0.8% for the dataset as a whole. The subsamples at two and three elements are 24 and 4 sessions respectively, so the exact percentages should be read as indicative, but the direction is not in doubt and the mechanism is understood: every junction is a point of failure as well as a point of resistance.

Reach against delivered power: the trade-off 2026#

Every extra metre of 4mm² cable costs 11.3W of delivered power at 32A and 28 seconds on a 36kWh charge. On 6mm² the same metre costs 7.5W and 19 seconds, and on 2.5mm² it costs 18.4W and 46 seconds.

Put in the units a buyer cares about, the per-metre cost of reach is genuinely small. On 4mm² an extra metre adds 0.352V of drop, gives up 11.3W of delivered power, adds 28 seconds to a 36kWh charge, and costs 3.51 kWh and £0.28 a year at the 2026 owner survey's median of 312 charging hours. On 6mm² it is 0.234V, 7.5W, 19 seconds and £0.18 a year. On 2.5mm² it is 0.576V, 18.4W, 46 seconds and £0.45. Even on the thinnest conductor in the table, ten extra metres costs under five pounds a year.

The cumulative version tells the same story. At 32A on 4mm², a 3m cable delivers 7.327kW and takes 4h 55m for a 36kWh charge. A 10m cable delivers 7.247kW and takes 4h 58m. A 25m cable delivers 7.078kW and takes 5h 05m. A cable at the full 32.7m compliant reach delivers 6.992kW and takes 5h 09m. That is fourteen minutes longer than the shortest cable in the table, across the whole of the compliant range, for a power penalty of 4.6%.

So the case for buying only the length you need is not about power, and pages that argue it on power alone are arguing from the weakest available evidence. The real costs of length are mass, handling and cable life, and the same table shows all three moving far faster than delivered power does. Going from 10m to 30m costs 3.1 percentage points of delivered power. It also takes mass from 3.60kg to 10.80kg, takes the share of drivers rating the cable too heavy from 11.6% to 74.6%, takes mean time to first fault from 4.4 years to 2.6 years, and takes price from £124 to £324.

The counter-case is real and the same tables carry it, which is why they are published whole. A cable that will not reach is worth nothing whatever its delivered power. The share of owners reporting their cable is too short runs at 38.4% for a 3m cable, 21.6% at 5m and 8.1% at 10m, falling to 1.1% at 25m and 0.6% at 30m. Under-buying is a far more common mistake than over-buying, and it is the one that cannot be fixed by putting the cable away properly. Anyone reading this page as an argument for the shortest possible cable has read it backwards.

The full range, with the conductor stated on every listing, is at our EV charging cables.

Maximum length by cable on sale in the UK 2026#

The longest EV charging cable in EV Cable Hub's 2026 UK sample was 30m, and 80.0% of the 30m cables sampled exceeded 5 per cent voltage drop at their advertised rating. Only 4.0% of UK orders were for a cable of 25m or longer.

The market is concentrated at the short end and thins out fast. 10m is the most common length at 34.2% of 2026 orders, followed by 5m at 22.8% and 15m at 18.6%. Together those three account for 75.6% of everything sold. 7.5m takes 9.4%, 20m takes 6.9%, 3m takes 4.1%, 25m takes 3.1% and 30m takes 0.9%. Cables of 20m and above are 10.9% of orders and cables of 25m and above are 4.0%. The long end of this page's chart describes a very small share of the market and a disproportionate share of its problems.

The compliance pattern follows the same shape in reverse. Nothing at 3m, 5m, 7.5m or 10m failed the 5 per cent test. 5.9% of 15m cables failed, 19.0% of 20m cables, 60.0% of 25m cables and 80.0% of 30m cables. That is not a coincidence of sampling; it is a direct consequence of what the mean measured conductor does as advertised length rises. It sits at 3.86mm² to 3.94mm² across the short lengths, rises to 4.02mm² at 15m and 4.18mm² at 20m, then drops to 3.64mm² at 25m before recovering to 4.42mm² at 30m.

That dip at 25m is the single most useful thing in this section. At 32A, a 25m cable needs 4.0mm² and the market is supplying an average of 3.64mm². A 25m cable on 4mm² drops 8.80V, which is 3.83% and a comfortable pass. The same cable on 2.5mm² drops 14.40V, which is 6.26% and a clear fail. The difference between a good 25m cable and a bad one is one number that most listings do not print, and the failure rate at that length (60.0%) is what happens when buyers cannot see it.

The honest counterweight is that the failures are not evenly distributed and it would be misleading to describe this as a market-wide problem. Splitting the 41 cables of 20m and above by channel, UK specialist EV retailers averaged 5.24mm² of conductor with an 11.1% failure rate, and 72.2% of them stepped the conductor up at 25m. UK general electrical retailers averaged 4.38mm² and failed 37.5% of the time. UK sellers on online marketplaces averaged 3.68mm² and failed 66.7% of the time. Overseas marketplace sellers averaged 3.02mm², failed 83.3% of the time, and not one of them stepped the conductor up at 25m.

That does not make a long cable a bad purchase. It makes an unspecified long cable a bad purchase. A 25m cable on 4.0mm² is compliant at 32A with 2.70V of the 11.50V allowance to spare, weighs 9.00kg, and delivers 7.078kW. It is a perfectly good product for the 3.1% of buyers whose parking genuinely needs it. The test is whether the listing states the cross-section, and whether the stated cross-section is what the length and rating require.

Maximum length by use case 2026#

The mean UK driveway distance from charge point to vehicle inlet is 6.3m, and the mean cable owned is 10.6m, leaving 4.3m of surplus coiled on the ground during a typical charge. EV Cable Hub's 2026 survey gives a recommended length for each of nine use cases, and in five of them, covering 81.4% of drivers, it is 10m or less.

The use case table turns the whole chart into a purchase. A garage with the charge point beside the car measures 2.4m and takes a 5m cable. A short driveway with the charge point on the house wall measures 4.2m and takes 7.5m. A standard driveway with the car nose-in measures 5.1m and takes 7.5m. A standard driveway with the car nose-out, or with the inlet on the far side, measures 6.4m and takes 10m. At 22.8% of drivers that is the single most common situation in the survey. A long driveway or two-car layout measures 7.8m and also takes 10m.

Above those five the recommendations step up quickly, because the parking position stops being fixed. A detached garage or a charge point on the side of the house measures 9.1m and takes 15m. Shared parking or a communal bay measures 11.4m and takes 20m. Rear or lane access measures 12.8m and takes 25m. A yard, farm or commercial run measures 18.6m and takes 30m on 6.0mm² conductor, left permanently in place rather than coiled after every charge. Those four cases together are 18.6% of drivers, and they are the only ones on this page where a cable beyond 15m is the right answer.

The margin rule that emerges from the survey is simple and it has a measured optimum. Take the straight-line distance from the charge point to where the inlet sits when the car is parked as it normally is, then add 2.5m. Owners whose surplus landed in the 2.5m to 3.5m band reported the highest satisfaction in the 2026 survey, with 84.1% saying the length was about right, only 3.1% saying it was too short and 11.8% saying it was too heavy. That band also holds 24.1% of owners, the largest single group.

The satisfaction curve either side of that optimum is steep in both directions, which is what makes the rule worth following rather than approximating. Owners with less than 0.5m of surplus reported 42.6% too short. Nearly half of them found the cable unusable in at least one parking position. Owners with 0.5m to 1.5m still reported 18.4% too short. On the other side, owners with 5m to 8m of surplus reported 38.4% too heavy and only 46.2% still coiling the cable after every charge, and owners with more than 12m of surplus reported 71.4% too heavy and 18.4% coiling it. Buying two sizes up is not a safety margin. It is a different problem.

The full range is at our EV charging cables, and the long end at 25m EV charging cables.

The longest cables measured in the field 2026#

The longest cable in EV Cable Hub's 2026 field programme was 30m and it delivered 7.02kW at 32A on 4mm². Three homes ran cables at 30m, and all three were partly coiled during every monitored session, at a mean coiled share of 57.3%.

This is the section where the chart meets reality, and the first thing to say about it is how small the long end of the sample is. Of 187 monitored homes, 33 ran a cable of 20m or longer, 14 ran 25m or longer and 3 ran 30m. Nobody ran anything longer than 30m as a single cable, and the longest total run recorded anywhere in the programme was 40m, reached with an extension. Every figure in this section rests on those subsamples and should be read as indicative rather than precise.

Delivered power behaved exactly as the arithmetic said it would, which is the most reassuring result on this page. Mean delivered power at 32A was 7.24kW on 10m cables, 7.11kW at 20m, 7.06kW at 25m and 7.02kW at 30m. That is a spread of 0.22kW, or 3.0%, across twenty metres of extra cable. Nobody in the field programme was losing a meaningful amount of charging speed to cable length, and any owner who thinks their long cable is charging slowly is almost certainly looking at a thermal derate rather than a voltage drop.

The thermal picture is where the long-cable homes genuinely differed from the rest. Coiled share ran at 54.5% for 20m cables, 54.4% at 25m and 57.3% at 30m, against 36.0% for a 10m cable. Sessions triggering a thermal derate ran at 0.4% for 10m cables, 2.8% at 20m, 4.1% at 25m and 6.2% at 30m, a fifteen-fold increase across the range. Mean conductor temperature on 30m cables at 32A was 48.6°C. Pilot faults were almost absent, at 0.2% of sessions at 25m and 0.4% at 30m, which is consistent with a ceiling that does not become material until 40m.

Asked whether they would buy the same length again, 68.4% of long cable owners said yes, 24.6% would buy shorter and 7.0% would buy longer. The more interesting answer is the one that cuts across all three: 41.2% would buy a thicker conductor. That is a group who have concluded from experience that their cable is working near a limit, and on 4mm² at 25m or 30m they are right about the electrical margin and wrong about the consequence, because a thicker conductor at the same length would have taken them past the handling ceiling instead.

The handling data from those homes closes the argument the rest of the page makes from the survey. Only 34.8% of long cable owners coiled the cable after every charge, and 31.4% left it permanently deployed. Mean time to first fault was 3.1 years for cables of 20m and over against 4.4 years for cables under 20m. The long cables in this programme were not failing electrically, were not failing thermally in any way their owners noticed, and were wearing out roughly fifteen months sooner than the short ones because of how they were being stored.

• Homes running cables at 20m or longer: 33 of 187
• Homes running cables at 25m or longer: 14 of 187
• Homes running cables at 30m: 3 of 187
• Longest cable recorded in the field programme: 30 m
• Longest total run recorded including an extension: 40 m
• Mean delivered power, 10m cables at 32A: 7.24 kW
• Mean delivered power, 20m cables at 32A: 7.11 kW
• Mean delivered power, 25m cables at 32A: 7.06 kW
• Mean delivered power, 30m cables at 32A: 7.02 kW
• Mean coiled share, 20m cables: 54.5%
• Mean coiled share, 25m cables: 54.4%
• Mean coiled share, 30m cables: 57.3%
• Sessions triggering a thermal derate, 10m cables: 0.4%
• Sessions triggering a thermal derate, 20m cables: 2.8%
• Sessions triggering a thermal derate, 25m cables: 4.1%
• Sessions triggering a thermal derate, 30m cables: 6.2%
• Sessions logging a pilot fault, 25m cables: 0.2%
• Sessions logging a pilot fault, 30m cables: 0.4%
• Mean conductor temperature, 30m cables at 32A: 48.6 °C
• Long cable owners who would buy the same length again: 68.4%
• Long cable owners who would buy shorter: 24.6%
• Long cable owners who would buy longer: 7.0%
• Long cable owners who would buy a thicker conductor: 41.2%
• Long cable owners citing a second parking space as the reason: 38.6%
• Citing rear or lane access: 21.4%
• Citing a shared or communal bay: 18.2%
• Citing a detached garage: 12.4%
• Citing commercial or yard use: 9.4%
• Long cable owners who leave the cable permanently deployed: 31.4%
• Long cable owners who coil it after every charge: 34.8%
• Mean time to first fault, cables 20m and over: 3.1 years
• Mean time to first fault, cables under 20m: 4.4 years

What over-length costs 2026#

The mean UK driver carries 4.3m more cable than they use, which costs £38 in purchase price, 1.55kg in mass, 48.6W in delivered power and two minutes on a 36kWh charge. EV Cable Hub's 2026 figures put the annual electricity cost of that surplus at £1.20 on an overnight tariff.

Priced out in every available currency, the surplus is a mixed picture and it is worth being clear about which parts of it matter. The electricity is trivial: 1.51V of extra drop, 48.6W of delivered power given up, 15.2 kWh a year at the median 312 charging hours, £1.20 a year at an overnight 7.9p and £3.76 a year at a 24.8p flat rate. Two minutes on a 36kWh charge. If the argument against over-buying were an electricity argument, there would be no argument.

The purchase price is not trivial. The mean 4.3m of surplus costs £38 up front on 4mm², which on a £124 10m cable is nearly a third again. Across the length table the price gradient is steeper than any of the electrical ones: £74 at 3m, £124 at 10m, £168 at 15m, £214 at 20m, £268 at 25m and £324 at 30m. Every step up the stocked sizes costs between £40 and £60, and each one buys reach that the measured driveway distances say most buyers do not need.

The relationship between surplus and cable life is the strongest finding in this section. Owners whose surplus was under 1.5m had a mean cable mass of 2.42kg, coiled it after every charge 89.4% of the time, reported driving over it weekly 16.2% of the time, and had a mean time to first fault of 4.8 years, with 2.1% of cables failing within three years. Owners with 5m to 8m of surplus were at 6.12kg, 46.2%, 44.6%, 3.6 years and 7.2%. Owners with more than 12m of surplus were at 10.96kg, 18.4%, 64.8%, 2.6 years and 16.8%. The cable does not wear out because it is long. It wears out because a long cable stops being put away.

Where the run is genuinely long, the table of approaches is the answer rather than a longer cable. Up to 26m at 32A a standard 4.0mm² cable is the right tool and delivers between 7.25kW and 7.07kW across that range. From 26m to 32.7m a 4.0mm² cable left permanently deployed still works, delivers 7.00kW and weighs 10.8kg, which 74.6% of drivers would call too heavy to coil daily. From 26m to 40m, 6.0mm² left in place delivers 7.14kW at 18.5kg. Beyond 32.7m the honest answer is to relocate the charge point, because the fixed wiring carries the distance without anybody lifting it, and beyond 49m the cable is simply the wrong tool for the job.

The whole page therefore closes on three sentences. Buy for the measured distance plus a 2.5m working margin, which puts most drivers between 7.5m and 15m. Step the conductor up rather than the length where a long run is genuinely needed, and accept that above 4mm² each step up costs practical reach. And if the run is beyond 26m at 32A, treat it as an installation problem rather than a cable problem, because that is what the four ceilings say it is.

For the long end specifically, see what a 15m, 20m or 25m cable actually delivers and 25m EV charging cables.

Interactive tools 2026#

Four tools built on the 2026 dataset, plus a searchable table of every figure on this page and a twenty-four point decision checklist that remembers where you got to. Everything runs in the browser.

Each tool draws on the tables above rather than on a separate dataset, and each one names the tables it reconciles with. Where a tool and a table disagree the table is right, so the defaults have been set to reproduce the published rows exactly.

Methodology 2026#

Every figure on this page comes from one of five EV Cable Hub programmes conducted in 2026: 214 cables bench-tested for length limits, 46 measured for control pilot integrity, a handling panel of 84 participants, 187 UK homes monitored across 2,639 charging sessions, and a survey of 2,140 UK EV drivers.

Limitations 2026#

The single largest caveat on this page is that the 12.0kg mass ceiling is a behavioural threshold rather than a physical one, and the second is that only 3 of 187 monitored homes ran a cable at 30m.

A reference chart that admits its boundaries is cited more often than one that does not, so the boundaries are published here rather than buried. Three of them matter more than the rest. The mass ceiling is a measurement of what people will tolerate, not of what a cable can withstand, and tolerance varies. The long end of the field sample is very small. And the pilot figures above 30m come from a bench rather than a driveway, because there were no driveways to take them from.

Two of the limitations run in the direction that makes this page conservative rather than optimistic, which is worth stating explicitly. Voltage drop figures use a 70°C conductor design basis, while field-measured drop came in 8.4% below that, so the compliant lengths published here are shorter than a real cable at real temperatures would justify. And the pilot measurements above 30m include the joints used to build the test lengths, which can only make the numbers worse than a continuous cable would be. Where the data errs, it errs towards caution.

One limitation runs the other way and should be read carefully. The left permanently deployed column removes the mass ceiling, which is correct as far as the handling data goes, but the 2026 programme did not quantify weather exposure, vehicle crossings or trip hazard for cables left out. That column therefore describes an electrical and communications ceiling on a cable nobody has to lift, and it is not a recommendation to leave a cable on a driveway indefinitely.

Finally, the market figures are a snapshot. Prices are UK retail as observed in 2026 and will move; annual costs assume 312 charging hours a year, the median in the 2026 owner survey, and a household charging twice as often will see twice the figure. The 25mm² sample is 6 cables against 148 at 4.0mm², so the thickest conductor in every table carries much wider error than the most common one. Not one of those caveats changes the four ceilings or the order in which they bind, which is the part of this page designed to hold from one edition to the next.

• The 12.0kg mass ceiling is a behavioural threshold from an 84-participant handling panel, not a physical limit. It is the mass at which 74.6% of participants described a cable as too heavy and only 18.4% could coil it one-handed. A stronger or more motivated user will exceed it comfortably
• The 40m pilot ceiling is where degradation becomes material, at a 2.8% session fault rate, not where communication stops. Faults were recorded from 25m onwards at 0.2%
• Pilot fault rates above 30m come from bench sessions rather than field observation, because only 3 of 187 field homes ran a cable at 30m and none ran longer
• The 30m field subsample is 3 homes and the 25m subsample is 14, so both should be read as indicative
• Voltage drop figures use a 70°C conductor design basis. Field-measured drop came in 8.4% below that figure, so the compliant lengths on this page are conservative
• The fixed-wiring drop of 2.84V is a single representative figure from 187 homes rather than a measured distribution, and a house with a long run from the consumer unit to the charge point will have more
• Cables longer than 30m were built up from measured 10m sections joined for test purposes. The joints were bypassed in the resistance measurement but the pilot measurements include them, so pilot figures above 30m carry an unquantified pessimistic bias
• Coiled share figures come from photographic evidence and on-site verification at 94 of 187 homes; the remaining 93 rely on self-report
• The 25mm² sample is 6 cables against 148 at 4.0mm²
• Price figures are UK retail as observed in 2026 and will move
• Annual cost figures assume 312 charging hours a year, which is the median in our 2026 owner survey
• The left permanently deployed case removes the mass ceiling but does not account for weather exposure, vehicle crossings or trip hazard, none of which were quantified in 2026
• Use case distances are means across self-selected categories and the within-category spread is wide

Frequently asked questions 2026#

Thirty-two questions on EV charging cable length, each answered with the 2026 figure first and the compliant limit kept separate from the practical one.

Every answer below is drawn from the tables on this page. Where a figure is a behavioural threshold rather than a physical one, it is described as such.