EV Charging Cable Voltage Drop Chart 2026: Volts, Percent and Pass or Fail for Every Length, Conductor and Current

The master EV charging cable voltage drop chart 2026#

A 25m 4mm² EV charging cable drops 8.80V at 32A, which is 3.83% of a 230V supply and a pass against the 5 per cent guidance. The same run in 2.5mm² drops 14.40V, which is 6.26% and a fail, and EV Cable Hub's 2026 testing found 13.4% of UK cables exceed 5% at their own advertised length and rating.

Voltage drop is the voltage consumed by the cable's own resistance. It is spent as heat along the run and never arrives at the car, so a 230V supply feeding a 25m 2.5mm² cable at 32A presents 215.60V at the vehicle inlet. Nothing is broken when that happens. The cable is doing what copper does, and the only question worth asking is whether the amount consumed sits inside the allowance the installation is designed to.

Every figure in the table above is round-trip. Current leaves on the line conductor and returns on the neutral, so it crosses the resistance of the cable twice, and EV Cable Hub's 2026 coefficients count both conductors. This is where most published tables and online calculators quietly go wrong: a single-conductor figure halves the answer. The earth conductor carries no current in normal operation and contributes nothing.

Three relationships govern everything that follows and none of them is complicated. Drop rises in exact proportion to length, so doubling the run doubles the drop. It rises in exact proportion to current, so doubling the amps doubles the drop. And it falls as conductor area rises, close to inversely, with each step up the standard sizes taking between 33.6% and 39.7% off the figure. The fourth variable, supply voltage, does not change the drop at all. It changes only the allowance the drop has to fit inside.

The figures are stated at a 70°C conductor design basis, which is the conventional worst case rather than what a driveway actually produces. EV Cable Hub's 2026 field programme measured conductors at a mean of 44.6°C and found real drop running 8.4% below the design figure. The chart is therefore conservative, which is the correct direction for a reference chart to be wrong in, and the size of that gap is published in full further down rather than hidden.

The bottom row of the master chart is the row most people come for: the longest run each conductor size can carry at 32A while staying inside 5% of a 230V supply. It runs from 12.4m on 1.5mm² to 205.4m on 25mm². The sections below take each variable apart in turn: length, cross-section, current, supply voltage, phase count, then volts, percent, pass and fail, heat, inlet voltage, temperature, extensions, field results and cost.

One convention affects every figure on this page and is worth stating once. Voltage drop here is the drop across the charging cable between the charge point socket and the vehicle inlet. It does not include the fixed wiring behind the charge point, the supply cable to the property, or anything upstream of the origin of the installation. Those carry their own drop against the same allowance, and the whole-circuit section sets out what that does to the budget.

Voltage drop by cable length 2026#

Voltage drop rises in exact proportion to length. A 4mm² cable at 32A drops 1.06V over 3m, 3.52V over 10m and 10.56V over 30m, so every additional metre costs 0.352V on that combination and the relationship is a straight line with no curvature at all.

Length is the easiest of the four variables to reason about because the arithmetic is trivial: double the run, double the drop, every time. That makes the per-metre figure the most useful single number on this page. EV Cable Hub's 2026 measurements put it at 0.352V per metre for 4mm² at 32A, 0.576V for 2.5mm² and 0.928V for 1.5mm², so anyone who remembers one of those can compute any length in their head.

The consequence of a straight line is that the point at which each conductor size runs out of allowance is fixed and knowable in advance. At 32A on a 230V supply the 5 per cent allowance is 11.50V, and dividing that by the per-metre figure gives the compliant reach directly: 12.4m on 1.5mm², 20.0m on 2.5mm², 32.7m on 4mm² and 49.2m on 6mm². The whole purpose of a chart like this one is that the number is available before the cable is bought rather than after it is plugged in.

Twenty-five metres is where conductor size stops being a preference and becomes a requirement. At 32A a 25m cable drops 8.80V on 4mm², which is 3.83% and a pass with 2.70V of margin; 14.40V on 2.5mm², which is 6.26% and a clear fail; and 23.20V on 1.5mm², which is 10.09% and outside any reasonable reading of the guidance. That is the same physical length sold at three specifications, and only one of them delivers what the listing implies.

Lower currents move the boundary a long way out. EV Cable Hub's 2026 derivation gives 39.9m on 2.5mm² at 16A against 20.0m at 32A, and 63.9m at 10A, because the allowance is fixed in volts while the drop per metre falls with the current. A cable that is unusable at 32A on a 25m run can be entirely comfortable on the same run at 13A, which is why length questions cannot be answered without the current alongside them.

One caution about how length is counted. The figures here are the length of the cable, measured from plug to connector, not the distance from the charge point to the car. A cable that reaches from a wall unit to a vehicle parked five metres away still drops its full 25m worth if 25m is what is coiled on the driveway, because the current travels the whole conductor either way. Coiling changes nothing about voltage drop, though it changes a great deal about heat.

Our what a 15m, 20m or 25m cable actually delivers guide covers the same ground from the buying side, and the 25m charging cable range lists conductor size on every product.

Voltage drop by conductor cross-section 2026#

Voltage drop falls as conductor area rises. At 32A over 25m, EV Cable Hub's 2026 measurements give 14.40V on 2.5mm², 8.80V on 4mm² and 5.84V on 6mm², and each step up the standard sizes takes between 33.6% and 39.7% off the drop.

Of the four variables, cross-section is the only one the buyer can change once length and current are settled, which makes it the actionable one. A 25m run at 32A is a 25m run at 32A; the conductor inside it is a choice, and it is the choice that decides whether the same physical cable passes or fails.

The coefficient table is the source of every derived figure on this page and the derivation is published with it. EV Cable Hub measured the resistance of a single conductor at a controlled 20.0°C, doubled it because current crosses the cable twice, and multiplied by 1.1965, the measured resistance ratio of electrolytic copper between 20°C and its 70°C design operating temperature. On 4mm² that runs 4.61mΩ per metre, to 9.22mΩ round trip, to 11.03mΩ at 70°C, published as a coefficient of 11.0mV per amp per metre. Voltage drop is then the coefficient multiplied by the current and the length, divided by 1,000. Nothing on this page requires any other step.

The relationship with area is close to inverse but not exactly inverse, and the reason is worth stating because it is the difference between a measured chart and a calculated one. Nominal 6mm² has 1.5 times the area of nominal 4mm², yet the measured coefficients are 7.3 and 11.0mV per amp per metre, a ratio of 1.51. On 1.5mm² against 2.5mm² the area ratio is 1.67 and the measured coefficient ratio is 1.61. The coefficients come from conductors we put a micrometer and a strand count on, not from the number in the listing, and measured copper never lands exactly on its nominal figure.

What a step up actually buys is set out in full rather than left as a rule of thumb. Moving from 2.5mm² to 4mm² on a 25m run at 32A takes 38.9% off the drop, 179.2W off the heat and adds 12.7m of compliant reach for a mean price increase of £22 on a 10m cable in the 2026 market. Moving from 4mm² to 6mm² takes a further 33.6% off the drop and adds 16.5m of reach for £26. The delivered-power gain is small in every case (0.179kW and 0.095kW respectively), which is the honest counterweight to the compliance argument and is developed further in the cost section.

One material caveat sits under the largest size in the table. EV Cable Hub's 2026 sample contained 6 cables at 25mm² against 148 at 4.0mm², so the 25mm² coefficient carries a wider confidence interval than the rest of the column. The 35mm² and 50mm² sizes do not appear at all, because no EV charging cable in the sample used them.

Our explainer on how conductor gauge is specified and measured covers the strand-count and micrometer method behind the resistance figures, and the 2026 gauge chart converts every size here between mm² and AWG.

Voltage drop by current 2026#

Voltage drop rises in exact proportion to current. A 10m 4mm² cable drops 1.76V at 16A and 3.52V at 32A, so doubling the current doubles the drop, while the power lost as heat quadruples from 28.2W to 112.6W.

This is the most misunderstood relationship on the page, so it is worth stating twice: drop scales with current, heat scales with current squared. The first is linear and the second is quadratic, and they are frequently conflated in advice that treats a step up in amps as a proportional step up in everything else.

The practical consequence is that a cable which is comfortable at one current can be marginal at another on exactly the same run. A 25m 2.5mm² cable drops 7.20V at 16A, which is 3.13% and a pass, and 14.40V at 32A, which is 6.26% and a fail. Nothing about the cable changed. The charge point did.

EV Cable Hub's 2026 owner survey found that upgrading a charge point from 16A to 32A while keeping the existing cable is a common sequence, and the measured consequence is set out in the doubling table. On a 10m 4mm² cable the drop moves from 1.76V to 3.52V, the heat from 28.2W to 112.6W, and the steady-state conductor temperature rise from 8.1°C to 24.8°C. Delivered power rises by 98.6% rather than 100%, and the charge time for 36kWh falls from 9h 52m to 4h 58m. On 1.5mm² the same change takes the conductor temperature rise from 18.4°C to 51.6°C, which is the number that matters more than the volts.

Reading the current matrices the other way round is equally useful. At 10m, a 63A run drops 6.93V on 4mm² and 4.60V on 6mm², both comfortable passes, while the same 63A on 1.5mm² drops 18.27V and is a long way outside the allowance. At the other end, a 6A trickle drops 0.66V on a 10m 4mm² cable, which is 0.29% of a 230V supply and effectively nothing. The current a cable is actually used at matters far more than the current printed on its label.

One point on heat that the matrices make plain. Because the relationship is quadratic, the heat penalty for stepping up the current is four times the voltage penalty, and it is the heat that determines conductor temperature, which in turn determines the resistance, which feeds back into the drop. That feedback loop is measured in the temperature section and it is small. But it runs in the direction of making a marginal cable worse rather than better as the session proceeds.

Our guides to the full 16A against 32A comparison and what the amp rating on a charging cable means cover the equipment side of the same decision.

Voltage drop by supply voltage 2026#

The voltage drop in a cable does not change with supply voltage, but the allowance does. EV Cable Hub's 2026 field measurements recorded UK single-phase supply between 212.4V and 249.8V, which moves the 5 per cent allowance between 10.62V and 12.49V and the maximum compliant 4mm² run at 32A between 30.2m and 35.5m.

Voltage drop is a function of current and resistance and of nothing else. A 10m 4mm² cable at 32A drops 3.52V on a 207V supply and 3.52V on a 250V supply, identically, because neither the copper nor the current has changed. What changes is the fraction that 3.52V represents (1.70% of 207V and 1.41% of 250V), and the size of the allowance it has to fit inside.

That makes supply voltage the quietest of the four variables and the one most often left out of a specification. EV Cable Hub's 2026 field voltage programme measured 187 UK homes and found a mean of 232.4V, a median of 233.1V and a standard deviation of 7.8V across the sample. Within a single home the mean swing across the day was 5.4V, with the largest recorded swing 14.2V, and supply ran a mean of 234.6V overnight against 229.1V through the peak evening period.

The practical consequence is that identical hardware can be compliant in one house and non-compliant in another. Two homes at opposite ends of the measured range have allowances of 10.62V and 12.49V, a difference of 1.87V, which is enough to move the maximum compliant 4mm² run at 32A by more than five metres. EV Cable Hub's 2026 measurements found 4.8% of homes where a cable that is compliant at the home's mean voltage fails at its own measured minimum, which is a compliance problem created entirely by the supply and not at all by the cable.

Nine point six per cent of homes in the sample fell below 220V at some point in the day and 8.6% rose above 245V. Both tails matter, but they matter differently: the low tail eats the allowance and the high tail returns it. Anyone specifying close to the limit on a long run should be working from a measured supply voltage at the origin of the installation rather than from the nominal figure, and the difference between the two is the whole point of this section.

Three-phase supply is a separate case and gets its own section below. For the arithmetic here, the measured mean line voltage across the three-phase homes in the 2026 sample was 403.8V, and the 400V nominal figure gives a 5 per cent allowance of 20.00V against 11.50V on single phase.

Single phase voltage drop 2026#

Single-phase voltage drop counts both conductors, because the current travels out on the line and back on the neutral. EV Cable Hub's 2026 coefficients are therefore twice the single-conductor resistance, which is why a 4mm² cable drops 11.0mV per amp per metre rather than the 5.5mV a single-conductor calculation would give.

This is the correction that matters most on the page, because a large share of published tables and online calculators quietly use a single-conductor figure and understate the drop by half. A calculator that returns 4.40V for a 25m 4mm² cable at 32A is not slightly optimistic. It is wrong by a factor of two, and the correct answer is 8.80V.

The round-trip principle is not complicated. Current has to complete a circuit. It leaves the charge point on the line conductor, passes through the vehicle's onboard charger and returns on the neutral, so it crosses the full length of the cable twice and meets the resistance of two conductors rather than one. Every coefficient in the 2026 table is stated on that basis, which is why the third column of the coefficient table is the measured single-conductor resistance multiplied by two before anything else happens to it.

The earth conductor is a frequent point of confusion and the answer is unambiguous. It carries no current in normal operation, so it contributes nothing at all to voltage drop. Its cross-section matters a great deal for fault protection and not at all for the figures on this page, and EV Cable Hub's 2026 field programme confirmed the point across 2,639 monitored sessions with no measurable earth current in normal charging.

The two full matrices below cover 4mm² and 6mm², which between them account for the overwhelming majority of UK EV charging cables in the 2026 sample. They run every standard current from 10A to 63A against every standard length from 3m to 50m, so most of the questions this page is asked can be answered from one of them without any arithmetic. On 4mm² the 32A column crosses the 11.50V allowance between 30m and 40m, and the exact crossing point is 32.7m. On 6mm² the same column stays inside the allowance until the last row, where 50m reaches 11.68V, and the crossing point is 49.2m.

A note on the control lines. A Mode 3 cable carries a control pilot and a proximity conductor alongside the power cores, and neither carries load current, so neither appears in these figures. What the proximity conductor does affect is the current the cable declares itself capable of, which is the subject of our 2026 ampacity chart.

There is a second reason the round-trip figure is the right one to publish, beyond simple accuracy. A single-conductor coefficient cannot be checked against a measurement, because nobody measures the drop across one conductor of a plugged-in cable. What is measurable is the difference between the voltage at the charge point socket and the voltage at the vehicle inlet, and that difference is the round-trip figure. EV Cable Hub's 2026 field programme logged exactly those two points simultaneously across 2,639 sessions, which is what makes the comparison in the field section possible at all.

Three phase voltage drop 2026#

Three-phase voltage drop is 86.6% of the single-phase figure at the same current per core, because the line-to-line calculation uses the square root of three rather than a factor of two. On 4mm² that is 9.5mV per amp per metre against 11.0mV, a difference of 1.5mV.

The arithmetic is worth setting out openly rather than asserting. Single-phase drop counts two conductors, so the multiplier is 2. Three-phase line-to-line drop is the square root of three times the single-conductor drop, so the multiplier is 1.732. Divide one by the other and the ratio is 0.866, which is the factor applied to every single-phase coefficient in the 2026 table to produce its three-phase pair.

The consequence is that a three-phase cable of the same cross-section reaches further before it meets the allowance, and the reach gain is larger than the drop reduction alone would suggest. A 25m 4mm² run at 32A per core drops 7.60V on three phase against 8.80V on single phase, a reduction of 13.6%. But the allowance moves too: 5 per cent of a 400V nominal supply is 20.00V against 11.50V on 230V. Put together, the maximum compliant 4mm² run goes from 32.7m to 65.8m, a reach multiple of 2.01, and the same multiple holds across every conductor size in the 2026 table at between 2.00 and 2.03.

Here is the honest counterweight, because this section reads much more impressively than it deserves to for a domestic audience. Only 3.7% of UK homes in EV Cable Hub's 2026 sample have a three-phase supply. This is a commercial, workplace and fleet section, not a domestic one, and a driver reading it as an argument for converting a house supply is reading it wrongly: the cost of the supply change is orders of magnitude beyond the cost of a heavier cable that solves the same problem.

Phase imbalance is the effect that three-phase calculations most often omit. Where the three phases do not draw equally, the difference returns on the neutral, and that neutral current produces additional drop. EV Cable Hub's 2026 monitoring measured a mean imbalance across all three-phase sessions producing 1.6A of neutral current at 32A, which adds 0.44V on a 25m 4mm² run and takes the total from 7.60V to 8.04V. At the worst measured band, above 12% imbalance and 1.1% of sessions, neutral current reached 6.8A and added 1.87V for a total of 9.47V. Every band remained inside the 20.00V allowance on that configuration.

One limitation to state plainly. The three-phase figures are derived from single-phase measurements by the 0.866 factor rather than measured independently on a three-phase supply at every conductor size. EV Cable Hub verified the derivation against 46 three-phase cables in 2026 and the two agreed to within 1.6%, which is close enough to publish and not close enough to leave unstated.

Voltage drop in volts 2026#

Stated in volts, a 10m 4mm² cable at 32A drops 3.52V and a 25m 2.5mm² cable at the same current drops 14.40V. EV Cable Hub's 2026 chart publishes every figure in volts first, because volts is the unit that survives a change of supply voltage and percent is not.

The volts figure is a property of the cable and the current, and of nothing else. It does not move when the supply moves, it does not move when the guidance moves, and it does not need a second number to be meaningful. That is why the complete reference table below is stated in volts and the percentage tables that follow are derived from it rather than the other way round.

Two other numbers on this page are computed directly from the volts figure, which is the practical reason to treat it as primary. Subtract it from the measured supply voltage and the result is the voltage arriving at the vehicle inlet. Multiply it by the current and the result is the power lost as heat in the cable. A 25m 4mm² cable at 32A drops 8.80V, which gives 221.20V at the inlet on a 230V supply and 281.6W of heat, and both of those follow from one measurement.

The complete reference table runs from 1m to 50m across all eight conductor sizes at 32A, which is the single most-requested combination. The short lengths at the top are there deliberately: a 1m 4mm² tail drops 0.35V, which is 0.15% of a 230V supply, and knowing that a short run is genuinely negligible is as useful as knowing that a long one is not. The step from 1m to 50m on 1.0mm² runs from 1.41V to 70.40V, which is the clearest illustration on the page of what proportionality to length actually means.

The six-combination table is the quick-reference version, and it carries the inlet voltage, the heat and the verdict alongside the volts so that a single row answers a whole specification question. Five of the six pass at 5 per cent. The sixth, 25m on 2.5mm² at 32A, drops 14.40V, delivers 215.60V at the inlet and dissipates 460.8W, and it fails. That configuration is common enough on the UK market to be worth naming every time it comes up.

Volts is also the unit that survives a change of guidance, which is not a hypothetical concern for a chart meant to be cited for years. Thresholds are conventions and conventions are revised. A table of drops in volts remains correct whatever percentage anyone chooses to apply to it, whereas a table published only as percentages of a nominal supply has to be rebuilt every time either the supply convention or the threshold moves. That is the practical case for the volts tables being the primary ones on this page and everything else being derived from them in public.

Voltage drop as a percentage 2026#

A 25m 4mm² cable at 32A drops 3.83% of a 230V supply, and the same run in 2.5mm² drops 6.26%. EV Cable Hub's 2026 chart gives the percentage for every combination, and 13.4% of UK cables measured in 2026 exceed 5% at their own advertised length and rating.

The percentage is what the guidance is written against, so it is what compliance is judged on. It is also a derived number rather than a measured one, and the thing it is derived against moves. The same 3.52V drop is 1.70% on a 216V morning and 1.41% on a 250V night in the same house, and both figures are correct. Percentages on this page are stated against 230V nominal unless a table says otherwise, and the supply voltage section gives the conversion for every other case.

Two thresholds are in common use and they answer different questions. Five per cent is the general limit applied to a circuit of this type and is the threshold every pass-and-fail cell on this page is stated against unless the table says 3 per cent. Three per cent is the tighter figure applied to lighting circuits and adopted by some specifiers as a design margin on charging circuits so that there is headroom left for future changes. Neither is a safety cliff. Both are budgets, and this page publishes the arithmetic for both so a reader can apply whichever their own specification names.

The single most useful point an installer takes from this page is that the budget is not the cable's alone. The guidance covers the whole circuit from the origin of the installation, so the fixed wiring between the consumer unit and the charge point consumes part of the same allowance before the cable is plugged in. EV Cable Hub's 2026 modelling uses a representative fixed-wiring drop of 2.84V, and on that basis a 25m 4mm² cable at 32A consumes 76.5% of the 11.50V budget on its own and takes the whole circuit to 11.64V, which is 5.06% and a fail, despite the cable alone passing comfortably at 3.83%.

That is the finding worth reading twice. A cable at 3.83% is a pass in isolation and a fail in a real circuit with ordinary fixed wiring behind it. On 25m of 2.5mm² the cable consumes 125.2% of the budget before the fixed wiring is counted at all, which is why that configuration is the one this page keeps returning to. Moving to 30m on 6mm² brings the whole circuit back to 9.85V and 4.28%, which passes with the fixed wiring included.

The two percentage matrices below cover 32A and 16A across every conductor size, so the figure can be read off directly rather than computed. On 4mm² at 32A the 5 per cent line falls between 30m and 40m; on 2.5mm² it falls between 15m and 20m; and on 1.0mm² it has already been crossed by 10m.

Pass and fail against the 5 per cent guidance 2026#

A 4mm² cable at 32A passes the 5 per cent guidance up to 32.7m and fails beyond it. Every pass and fail on this page is stated against a named threshold rather than as a blanket verdict, and on EV Cable Hub's 2026 sample 13.4% of cables exceed 5% at their own advertised length and rating.

A cell in the tables below reads PASS or FAIL against one specific threshold on one specific supply voltage, and it is worth being precise about what that does and does not mean. It means the drop is inside or outside 5 per cent of 230V, which is 11.50V. It does not mean the cable is safe or unsafe, approved or unapproved, or that a different specification would reach the same verdict. Change the threshold to 3 per cent, or the supply to 240V, and some cells move. Both alternatives are published on this page so the reader can apply their own.

What failing actually means in practice is narrower than the word suggests. It is not a hazard in itself. It is a cable that will not deliver its rated power at the far end, a circuit with no headroom left for the fixed wiring behind it, and a conductor running hotter than it needs to for the same job. On a 30m 1.5mm² run at 32A, which drops 27.84V, the practical outcome measured in the field was not a fault but a vehicle quietly settling at 16A and 3.46kW instead of the 7.36kW it asked for.

The failure rate across the market is the finding with the most commercial weight, and it is concentrated almost entirely at the long end. EV Cable Hub's 2026 conductor resistance programme measured 214 cables and found none failing at 3m, 5m, 7.5m or 10m at their own advertised rating. At 15m, 5.9% failed. At 20m, 19.0%. At 25m, 60.0% of the 15 cables sampled failed, and at 30m, 80.0% of the 5 sampled did. Across the whole sample the figure is 13.4% at 5 per cent and 41.6% at 3 per cent.

The mean measured conductor in the 25m group was 3.64mm², the smallest of any length group and smaller than the 3.86mm² sample mean. That is the mechanism behind the failure rate: the cables that most need a larger conductor are, on average, sold with a smaller one. Every 25m cable built on 2.5mm² fails at 32A, because 14.40V exceeds 11.50V by a margin no measurement tolerance closes, and the worst single drop recorded at a cable's own advertised rating was 28.16V.

The maximum compliant length tables are the inverse view and the one most installers use. They give, for every conductor size and every standard current, the longest run that stays inside the allowance: 32.7m for 4mm² at 32A, 20.0m for 2.5mm², 49.2m for 6mm², and the same grid again at 3 per cent where 4mm² at 32A comes back to 19.6m. Read the 3 per cent grid before specifying anything close to the 5 per cent limit.

Our guide to what a 15m, 20m or 25m cable actually delivers covers the buying decision behind these figures, the 25m cable range states conductor size on every listing, and our 2026 maximum cable length study takes the same limits from the ampacity and standards side.

Power lost as heat in watts 2026#

A 25m 4mm² cable at 32A dissipates 281.6W as heat along its length. EV Cable Hub's 2026 measurements show heat scaling with the square of current, so the same cable at 16A loses 70.4W rather than half of 281.6W.

The relationship is simple and worth stating in both forms. Watts lost equals voltage drop multiplied by current. It equals, equivalently, current squared multiplied by the loop resistance of the cable. The first form is easier to use from this page's tables; the second explains why the number moves so fast.

The practical consequences run in a chain. The heat raises the conductor temperature. The higher conductor temperature raises the copper's resistance. The higher resistance raises the voltage drop. And the higher drop lowers the voltage arriving at the car, so delivered power falls slightly as the session proceeds. Every step of that chain is measured in the temperature section below, and the total effect on a correctly sized cable is small: a 0.13% fall in delivered power on a 10m 4mm² run at 32A between a cold start and thermal steady state.

On an undersized cable the same chain produces numbers that are hard to ignore. A 25m 2.5mm² cable at 32A dissipates 460.8W continuously along a cable lying on a driveway, and a 25m 1.5mm² cable dissipates 742.4W. Those are space-heater figures being produced by something with no fan, no thermostat and frequently a coil still in it. The voltage drop verdict on those configurations is a fail, but the heat is the more immediately physical objection.

The heat-against-current table makes the quadratic relationship legible in one column. On any conductor, moving from 10A to 20A multiplies the heat by 4.00, to 32A by 10.24, to 40A by 16.00 and to 63A by 39.69. Nothing about the cable changes across that row. Only the current does, and the heat moves by nearly forty times while the voltage drop moves by just over six.

The energy figure follows directly: watts multiplied by hours gives kilowatt hours, and the annual totals at EV Cable Hub's measured median of 312 charging hours are given in the cost section. What that section argues, and this one supports, is that the heat matters more than the money. A conductor running 51.6°C above ambient is ageing its own insulation, and that is a cable-life argument rather than an electricity-bill argument.

Voltage at the vehicle inlet and what the car does about it 2026#

A 25m 2.5mm² cable at 32A delivers 215.60V at the vehicle inlet from a 230V supply. EV Cable Hub's 2026 field monitoring recorded 12.4% of vehicles derating in the 207V to 210V band at the inlet, rising to 91.4% between 190V and 195V.

The car does not measure the cable. It measures the voltage arriving at its own inlet, and below a threshold it reduces the current it draws in order to protect its onboard charger. That is the mechanism by which a voltage drop figure on a chart turns into something a driver actually experiences, and it is the reason the inlet voltage table is on this page at all.

Derating is gradual rather than binary and it begins higher than most people expect. EV Cable Hub's 2026 monitoring recorded no derating at all above 220V. Between 220V and 225V, 0.4% of sessions saw a reduction. Between 215V and 220V it was 2.1%, between 210V and 215V 6.8%, and between 207V and 210V 12.4% with a mean current reduction of 3.1A. Below 200V the behaviour changes character: 68.2% of sessions between 195V and 200V derated, 91.4% between 190V and 195V, and 98.6% below 190V, where the mean settled current was 17.4A against the 32A demanded.

The effect is self-limiting, which is why it produces a slower charge rather than a failure. As the vehicle reduces its current, the voltage drop falls in exact proportion, so the inlet voltage recovers and the system settles at a lower current. The measured sequence on a 30m 1.5mm² cable at 32A runs from an initial demand producing 27.84V of drop and 202.16V at the inlet, through three derate steps, to a settled state at 16A with 13.92V of drop and 216.08V at the inlet, delivering 3.46kW against the 7.36kW demanded. Across the 2026 sessions the mean time to settle was 4 minutes 12 seconds, over a mean of 2.8 steps of 5.7A each.

Session aborts sit at the far end of the same distribution and they are rare. Below 190V at the inlet, 46.2% of sessions aborted; between 190V and 195V, 18.6% did; above 207V the figure never exceeded 0.2%. Across the whole 2026 field programme, inlet voltage fell below 207V in 2.1% of sessions and below 195V in 0.4%, and 0.1% of sessions aborted on low inlet voltage. The worst case in the data is the predictable one: a low supply voltage combined with a long undersized cable, where the two deficits add.

One reading caution on the derating table. Thresholds vary by manufacturer and by model year, and the figures here are aggregate behaviour across 41 vehicle models rather than a threshold that applies to any individual car. A specific vehicle may derate earlier or later than the aggregate, and the only way to know is to measure the inlet voltage on that vehicle.

The self-limiting behaviour has one consequence that is easy to miss and worth stating for anyone diagnosing a slow charge. Because the system settles rather than fails, an undersized cable does not announce itself. The car charges, the app reports a session in progress, and the only symptom is a rate lower than expected, which most drivers attribute to the vehicle, the charge point or the weather. EV Cable Hub's 2026 field data suggests the cable is a more common explanation than any of those, and the check takes one measurement: the voltage at the inlet against the voltage at the charge point.

Voltage drop and conductor temperature 2026#

Voltage drop rises as the conductor warms. On a 10m 4mm² cable at 32A it rose from 2.94V at a 20°C conductor to 3.23V at a steady 44.8°C, an increase of 9.9%, and the mean rise from cold start to thermal steady state across EV Cable Hub's 2026 sample was 12.8%.

Copper's resistance rises with temperature at a measured 0.393% per degree Celsius, and every voltage drop figure is therefore a figure at a stated temperature or it is meaningless. This page states every drop at a 70°C conductor design basis, which is the conventional worst case for a cable of this construction and the temperature the 1.1965 multiplier in the coefficient table converts to.

The multiplier table lets any figure on this page be restated at any conductor temperature without going back to first principles. At 20°C the multiplier against the design basis is 0.836, so the 25m 4mm² run at 32A drops 7.36V rather than 8.80V. At 0°C it is 0.770 and the same run drops 6.78V, which is 23.0% below the design figure. At 90°C the multiplier is 1.066 and the drop is 9.38V. The full column runs from -10°C to 90°C in the table below.

What the conductor actually reaches is a separate measurement and it is well short of the design basis. EV Cable Hub's 2026 bench testing recorded a 4mm² conductor at 32A settling 24.8°C above a 20°C ambient, giving a conductor temperature of 44.8°C and a real coefficient of 10.12mV per amp per metre against the 11.0mV design figure, a design conservatism of 8.0% on that combination. On 6mm² at 32A the conservatism is 9.6% and on 2.5mm² it is 6.6%. Driven to their full rated current rather than 32A, the same conductors run hotter and the conservatism widens to between 13.1% and 16.9%.

Publishing that gap is what makes the chart credible rather than making it look wrong. A reference chart stated at a worst-case temperature will always read high against a field measurement, and the useful thing is to know by how much. The answer for this dataset is between 6.6% and 9.6% at 32A on the common sizes, and 8.4% across the field programme as a whole.

The session table shows the effect arriving in real time. On a 10m 4mm² cable at 32A the conductor moves from 20.0°C to 30.4°C in the first fifteen minutes, to 40.4°C by the first hour, and settles at 44.8°C by the four-hour mark. The voltage drop follows it from 2.94V to 3.23V and delivered power falls from 7.266kW to 7.257kW, a change of 0.13%. That is the honest scale of the effect on a correctly sized cable: real, measurable, and far too small for a driver to notice.

Voltage drop in extension and adaptor setups 2026#

Voltage drop adds along a chain. A 10m 4mm² cable plus a 10m 2.5mm² extension at 32A drops 9.42V in total against 3.52V for the cable alone, and EV Cable Hub's 2026 testing measured an additional 0.14V at each junction on a new connector pair, rising to 0.31V after 5,000 mating cycles.

The rule is additive and there is no hidden term. Add the drop of the first cable, the drop of the extension, and the drop of each junction between them, and the total is the number to test against the allowance. Everything in the configuration table below is that sum, computed from the same coefficients as the rest of the page.

This section will be read by people looking for permission, so the framing has to be honest in both directions. Some extension configurations are comfortably compliant. A 10m 4mm² cable extended with a 10m 4mm² lead drops 7.18V in total, which is 3.12% and a clear pass. Extended with a 10m 6mm² lead it drops 6.00V. Three 10m 4mm² cables in series drop 10.84V, which is 4.71% and a pass with essentially no margin left. The deciding factors are the extension's conductor size and the combined length, not the fact of extending.

Others are not compliant and the arithmetic says so plainly. A 20m 4mm² cable with a 10m 2.5mm² extension reaches 12.94V and 5.63%, a fail. A 10m 4mm² cable with a 10m 1.5mm² extension reaches the same 12.94V and fails for the same reason from the opposite direction: the thin section dominates regardless of where in the chain it sits. A 15m 4mm² cable with a 10m 2.5mm² extension lands on 11.18V and 4.86%, which passes with 0.32V in hand and no allowance left for the fixed wiring behind it.

The junction measurements are the part of this section that is not published anywhere else. EV Cable Hub's 2026 connector programme measured contact resistance across the full current-carrying path of a mated pair rather than a single contact, on 412 pairs in eight condition categories. A new factory-clean pair measures 2.19mΩ and adds 0.14V at 32A. After 5,000 mating cycles it is 4.84mΩ and 0.31V; after 10,000 it is 7.19mΩ and 0.46V. Light surface oxidation, which was present on 12.4% of the junctions sampled, measures 8.75mΩ and adds 0.56V. Visible contact corrosion measures 26.88mΩ and adds 1.72V, water ingress 2.64V, and a bent or displaced contact 4.00V. That is nearly thirteen times the penalty of a connector worn through 5,000 mating cycles, from a fault that is visible to anyone who looks.

The failure case that appears most often in the field data is not an EV extension at all. It is a correctly sized main cable plugged into a general-purpose 13A extension reel, where the conductor is 1.5mm² in 71.4% of the reels sampled and the reel is left wound. On voltage drop alone a 10m reel at 13A drops 3.77V and passes; a 25m reel drops 9.43V and passes; only at 50m does the voltage verdict fail at 18.85V. But the thermal problem arrives long before the voltage problem does: a fully wound 1.5mm² reel carries 9.9A against 14.4A unwound, so a 13A draw is already outside the reel's own capacity on a 10m drum that comfortably passes on volts. That is the case where a pass on this page is emphatically not a pass overall, and it is the reason the reel guidance is a thermal one rather than a voltage one.

Our full guidance on whether you can use an EV cable extension covers the safety and standards side, and the extension range lists conductor size on every product.

Voltage drop measured in the field 2026#

Field-measured voltage drop came in 8.4% below the design figure across 187 UK homes in 2026, at a mean of 4.62V against a predicted 5.04V. The gap is explained by conductor temperature: measured conductors averaged 44.6°C against the 70°C design basis used throughout this chart.

This is the credibility section. A reference chart that has never been checked against a measurement is an assertion, and the useful thing about checking is not that the chart turns out to be right but that the size and direction of its error become known.

The direction is consistent across every length group. At 3m the measured drop ran 9.4% below design, at 10m 8.2% below, at 25m 7.8% below and at 30m 7.4% below. The gap narrows slightly as cables get longer because longer cables run marginally hotter (mean conductor temperature rose from 42.1°C on the 3m group to 47.4°C on the 30m group), which moves the real coefficient closer to the design one. The 30m subsample is 3 homes, so that row should be read as indicative rather than settled.

What this means practically is that the chart is conservative, which is the correct direction for a chart to be wrong in. A cable that passes on this page will pass in reality with margin to spare. The reverse is emphatically not true: a cable that fails on this page is not rescued by the field gap, because the failures are large and the gap is small. A 25m 2.5mm² cable at 32A is 25% over the allowance on design figures and still 15% over it after the full 8.4% field correction.

The seasonal breakdown shows the gap widening in winter, exactly as the temperature coefficient predicts. Across 2,639 sessions the monthly mean measured drop ran from 4.24V in January, 15.9% below design at a mean conductor temperature of 29.4°C, to 4.84V in July, 4.0% below design at 44.2°C. The full-year mean across all sessions was 4.49V and 10.9% below design. Individual sessions spread wider still: the widest single gap recorded was 18.4% below design in January and the narrowest 1.2% below in July.

One line in the summary table deserves more attention than it usually gets: 1.8% of sessions measured a drop above the design figure rather than below it. Those are the cases where something other than the cable is contributing, such as a warm ambient, a hot conductor already loaded, or a junction in the chain, and they are the reason the chart is stated at a worst case rather than at the field mean. Highest conductor temperature recorded in the field was 71.4°C, which is above the design basis this whole chart is built on.

Two further readings of the field data are worth making explicit. First, the gap between measured and design drop is not evidence that the design basis is wrong; it is evidence that it is a worst case, which is what a design basis is for. Second, the gap is not a licence to specify closer to the limit, because it is a mean and the distribution around it includes 1.8% of sessions on the wrong side of the design figure. A specification that consumes the field gap as margin has no margin left for the sessions that do not behave like the mean.

What voltage drop actually costs 2026#

On a correctly sized cable, voltage drop costs very little. A 10m 6mm² cable at 32A wastes 23.3kWh a year, worth £1.84 on an overnight tariff, and EV Cable Hub's 2026 figures show the cost only becoming material on long undersized runs, where a 25m 1.5mm² cable wastes 231.6kWh.

The arithmetic is open. Watts lost multiplied by annual charging hours gives kilowatt hours, and kilowatt hours multiplied by the unit rate gives the money. The median driver in EV Cable Hub's 2026 owner survey charged for 312 hours in the year, and both cost tables below use that figure with a 7.9p overnight rate and a 24.8p flat rate observed in 2026.

On a 25m run at 32A the annual cost of the drop is £4.61 on 6mm², £6.94 on 4mm² and £11.36 on 2.5mm² at the overnight rate. Stepping from 4mm² to 6mm² therefore saves £2.33 a year. On the flat rate the same step saves £7.34. By length on 4mm², the cost runs from £0.83 a year at 3m to £13.88 at 50m.

Here is the point that matters, and it runs against the commercial interest of anyone selling cable. Voltage drop is a compliance problem and a heat problem far more than it is a money problem, and anyone selling a cable upgrade on the electricity saving alone is overstating it substantially. A £2.33 annual saving against a £26 mean price difference between a 10m 4mm² and a 10m 6mm² cable takes eleven years to repay, by which point the cable has been replaced.

The reasons to size up that do hold are compliance, delivered power at the far end, conductor temperature and cable life. A 25m 4mm² cable delivers 7.078kW against 7.173kW on 6mm², which adds four minutes to a 36kWh charge. A 25m 2.5mm² cable delivers 6.899kW and adds twelve minutes, and its conductor runs 34.2°C above ambient against 18.1°C on 6mm². The time cost is trivial; the temperature is not, because it is the thing that ages the insulation.

The worst configurations are where the money does start to register. A 25m 1.5mm² cable at 32A wastes 231.6kWh a year, which is £18.30 on an overnight tariff and £57.44 on a flat rate, and it fails the 5 per cent guidance by a wide margin while doing it. That is the only part of this dataset where the electricity cost is a reasonable argument on its own, and it applies to a configuration that should not have been specified for other reasons first.

Our full EV charging cable range lists measured conductor size on every product, and the 2026 ampacity chart gives the current rating side of the same specification.

It is worth putting the money figures next to the price of the cable itself, because that is the comparison a buyer is actually making. The mean 2026 price step from 10m of 4mm² to 10m of 6mm² was £26, and the annual electricity saving from that step on a 25m run is £2.33 at the overnight rate. On the flat rate it is £7.34 and the arithmetic looks better, but it still does not make the electricity the reason to buy. State the reason honestly and it holds: the heavier conductor is bought for compliance, for the power that arrives at the far end and for a conductor that runs cooler for longer.

Interactive tools 2026#

Three calculators built on the 2026 coefficient table, a conductor comparator, a searchable table of every figure on this page and a twenty-six point compliance checklist. Everything runs in the browser and nothing leaves it.

Each tool draws on the tables above rather than on a separate dataset. The defaults in every calculator reproduce the published figures exactly, which is deliberate: if a tool and a table on the same page disagree, one of them is wrong, and the only way to know they agree is to be able to check.