EV Charging Cable Gauge Chart 2026: mm² to AWG Both Ways, With the Conversion Error on Every Size

The master EV charging cable gauge chart 2026#

A 32A EV charging cable uses 4.0mm² conductor, which the trade labels AWG 12, a gauge that is 3.31mm², 17.3% smaller in area than the metric size it stands in for. EV Cable Hub's 2026 conductor metrology measured every conductor size used in EV charging cables in both systems, and the conventional mapping understates conductor area by a mean of 15.4%.

The chart above is the whole page in one table, and the rest of this article takes each of its columns apart. Read it left to right: the metric size, the AWG label the trade puts on that size, what that AWG actually measures, the error between the two, the AWG that is genuinely nearest by area, and then the physical and electrical properties EV Cable Hub measured for that conductor.

The two systems do not line up because they are not the same kind of number. A figure in mm² is a direct statement of the cross-sectional area of the conducting metal. AWG is not an area at all: it is a count of how many drawing dies a wire has passed through, so the numbers run backwards, the steps are geometric rather than arithmetic, and no whole gauge lands on a whole metric size. Every conversion between them therefore carries an error, and the only question is whether that error is published or hidden.

The trade convention picks the nearest AWG below the metric figure. That is conservative in the sense that it never claims a metric cable is bigger than it is, and it is simply wrong as arithmetic: on eight of the ten sizes used in EV charging cables, the conventional label and the true nearest gauge by area are different gauges. This page publishes both, with the error on each, rather than choosing one and leaving the reader to discover the gap.

The direction the conversion is read in changes the percentage, and both percentages are correct. Calling 6.0mm² an AWG 10 conductor understates the copper by 12.3%, because 5.26 is 12.3% below 6.0. Calling an AWG 10 conductor a 6.0mm² cable overstates it by 14.1%, because 6.0 is 14.1% above 5.26. Same pair of sizes, two different numbers, because the denominator moves. Any gauge chart that publishes one figure and calls it the conversion error is publishing half the answer, which is why Table 3 and Table 4 on this page run the same relationship in opposite directions.

There are exactly two sizes where the convention is generous rather than mean. EV Cable Hub's 2026 metrology recorded 0.5mm² labelled AWG 20 at +3.6% and 0.75mm² labelled AWG 18 at +9.7%, which is to say the AWG in those two cases is slightly larger than the metric size it is standing in for. Both are signal and control core sizes rather than power cores, and it is the reason a control pilot core is never the thing that goes wrong on a charging cable.

mm² and AWG side by side 2026#

No AWG size equals any metric size exactly. The closest pairing EV Cable Hub measured in 2026 is 0.5mm² against AWG 20, which differ by 3.6%, and the widest gap across the metric range is 1.0mm² against AWG 18, which differ by 17.7%.

Both reference tables in this section are complete rather than selective, because a conversion table that stops at the sizes the author expects you to need is the reason people go looking for a second one. Table 3 runs every metric size from 0.5mm² to 50mm² and gives the AWG immediately below it, the AWG immediately above it, the label the trade uses and the error that label carries. Table 4 runs every AWG from 24 to 2/0 and does the same thing in reverse.

The AWG sequence is geometric, and the multipliers are worth committing to memory because they make the whole system predictable. EV Cable Hub's 2026 measurements put one AWG step at 1.26 times the area and 1.12 times the diameter, three steps at exactly 2.00 times the area, and six steps at exactly 2.00 times the diameter. Ten steps is 10.08 times the area. That is why the numbers run backwards and why the gaps between adjacent gauges get enormous at the heavy end and trivial at the light end.

The multiplier that does not follow the pattern is the useful one. Three AWG steps double the conductor area but raise measured ampacity by only 1.55 times, and six steps quadruple the area for 2.43 times the current. Copper bought for current is bought at a worsening rate as the cable gets bigger, because heat escapes from a surface while it is generated in a volume. That relationship sets the practical ceiling on how far specifying up is worth doing, and it comes back in the cost-of-stepping-up figures later.

For a buyer the practical consequence is one-directional and it is the single most useful sentence on this page. Convert a US specification into a UK purchase using the conventional mapping and you will buy less copper than the specification called for, on every common size, every time. The gap runs from 12.3% at 6.0mm² to 17.3% at 4.0mm² and 17.7% at 1.0mm². Nothing in the convention warns you about that, because the convention was built to be safe in the other direction.

The two exceptions run the other way and both sit at the bottom of the range. EV Cable Hub's 2026 metrology put 0.5mm² against AWG 20 at +3.6% and 0.75mm² against AWG 18 at +9.7%, the only two sizes in the chart where the conventional AWG is larger than the metric size it labels. Those are the sizes used for control pilot, proximity pilot and earth continuity cores rather than for load, which is why the conversion problem documented on this page never shows up as a signalling fault.

Two rows in Table 3 are calculated rather than measured. No cable in the 214-cable sample used 35mm² or 50mm² conductor, so those rows come from the same resistivity model as the measured sizes and are published for conversion completeness. Every other figure in this section was measured on a sectioned conductor.

mm² to AWG conversion 2026#

Converting 4.0mm² to AWG 12 loses 0.690mm² of copper, 17.3% of the conductor area the metric specification called for. EV Cable Hub's 2026 chart gives the exact continuous AWG value, the true nearest whole gauge and the conventional trade label for every metric size, because the last two are different gauges on eight of the ten sizes used in EV charging cables.

The exact continuous AWG column in Table 6 is the number the drawing sequence actually puts each metric size at, before anything is rounded. It is worth looking at before choosing a conversion, because it shows how far from a whole gauge each metric size sits. 1.5mm² lands at 15.34 and 2.5mm² at 12.92, both within a third of a gauge of a whole number. 4.0mm² lands at 10.85 and 6.0mm² at 9.34. Not one of them lands on a whole gauge, which is the arithmetic reason this page exists.

Real cable is only made in whole gauges, so every conversion rounds, and there are two defensible ways to round. Rounding to the nearest gauge by area gives the truest equivalent and is what the fifth column of Table 6 calls the safe conversion for a replacement: AWG 11 for 4.0mm², AWG 9 for 6.0mm², AWG 13 for 2.5mm². Rounding down in area gives the conventional trade label: AWG 12, AWG 10, AWG 14. The first delivers at least the copper that was specified. The second delivers less, every time.

Which rounding is correct depends on which way the specification is travelling, and this is the part that is almost never said plainly. Replacing a metric cable with an AWG one, round up in area: take AWG 11 rather than AWG 12 for a 4.0mm² specification, and accept the 4.1% overshoot rather than the 17.3% shortfall. Reading an AWG specification and buying metric, round up again. Take 6.0mm² for AWG 10 rather than 4.0mm², which is what the market already does and is the one place the convention is generous.

Table 7 puts a price on the conventional rounding in the units that matter. EV Cable Hub's 2026 measurements put the mean resistance penalty across the eight power sizes at 18.7% and the mean ampacity penalty at 11.2%. At 4.0mm² specifically, buying AWG 12 instead costs 0.690mm² of copper, raises conductor resistance by 20.9% and takes 12.4% off measured ampacity. That is a 32A cable turned into a 30A one by a rounding convention rather than by any decision anybody made.

The voltage-drop column in the same table is the one that surprises people, because it moves in the opposite direction to the others. The extra drop at 32A over 10m caused by taking the conventional gauge is 3.03V at 1.0mm² but only 0.74V at 4.0mm² and 0.33V at 6.0mm². The percentage error stays roughly constant across the range while the absolute resistance collapses, so the same rounding mistake costs a great deal on a small conductor and very little on a large one. The ampacity loss, by contrast, stays near 12% all the way up, which is why the rounding matters most as a thermal question rather than a voltage one.

AWG to mm² conversion 2026#

AWG 10, the gauge most often quoted for a 32A charging cable in US specifications, measures 5.26mm², which is 12.3% below the 6.0mm² it is conventionally described as, or, read the other way round, a label that overstates the conductor by 14.1%. EV Cable Hub's 2026 chart converts every AWG from 24 to 2/0 into metric with the error on each stated.

The arithmetic in this direction is exact and can be checked in a spreadsheet: the area in mm² is 0.012668 multiplied by 92 raised to the power of 36 minus the AWG number, all divided by 19.5. Put AWG 12 through it and it returns 3.31mm². Put AWG 10 through it and it returns 5.26mm². Every area in Table 4 comes from that expression, which is a definition rather than a measurement. AWG is a drawing sequence, so its areas are geometry. What EV Cable Hub measured is the conductor inside each of the 214 cables and how it relates to those definitions.

Both error percentages for a given pair are worth carrying, because they answer different questions. The -12.3% in Table 1 answers what a buyer loses by accepting AWG 10 against a 6.0mm² specification. The +14.1% in Table 4 answers what a seller claims by describing an AWG 10 conductor as 6.0mm². The first is a purchasing loss and the second is a labelling overstatement, and on this page they are published side by side rather than averaged into a single misleading figure.

This direction is where most of the gauge overstatement in the UK market comes from, and EV Cable Hub's 2026 sample quantifies it. A factory datasheet quotes AWG 12, an importer writes the listing as 4mm², and the cable that arrives has 3.31mm² of copper in it, 17.3% less than a UK buyer reading 4mm² would expect. Nobody in that chain has to lie for it to happen. The conversion convention does the work.

The channel breakdown makes the mechanism visible. Of the 18 UK marketplace listings converted from AWG, 61.1% overstated the gauge, by a mean of 1.62mm². Of the 18 overseas marketplace listings converted from AWG, 77.8% overstated, by a mean of 1.94mm²: a cable listed at 4.02mm² measuring 2.86mm². The equivalent figures for listings written natively in metric are far lower in every channel, which is the clearest evidence in this dataset that the conversion itself, not the seller, is the primary source of error.

The honest counterweight sits in the same table. The 78 cables bought from UK specialist retailers stated a mean of 4.34mm² and measured a mean of 4.31mm², with 2.6% overstating and a mean overstatement of 0.08mm² where it occurred. That is inside EV Cable Hub's own measurement tolerance. The problem documented on this page is concentrated, not general, and the concentration is in listings that have crossed a unit boundary.

Gauge by amperage 2026#

A 32A EV charging cable needs 4.0mm², or AWG 11 as the true equivalent, and a 16A cable needs 1.5mm², or AWG 15. EV Cable Hub's 2026 measurements give the minimum gauge for every current rating in both systems, and the conventional AWG label is one gauge smaller than the true minimum on every rating from 10A upwards.

Headroom is the column to read, because it decides whether a cable runs warm or hot rather than whether it works at all. EV Cable Hub's 2026 measurements put 4.0mm² at 35.0A against a 32A rating, which is 9.4% of headroom, and 10mm² at 63.0A against a 63A rating, which is none at all. A cable with no headroom is not unsafe on paper and it is a poor buy in practice, because the whole of its thermal margin has been spent before the first cold start, the first coiled charge or the first hot afternoon.

The 63A case is the clearest example on the chart and it deserves stating flatly: 63A on 10mm² has zero headroom on our measurements, so any 63A cable intended for continuous duty should be specified at 16mm², which measures 85.0A. The same logic applies at 6A on 0.5mm², which is also at exactly zero headroom, and at 25A on 2.5mm², where 4.0% is not enough margin to survive being coiled.

The 32A cable deserves its own paragraph because it is what most of this market buys. EV Cable Hub measured 148 cables printed as 32A in 2026, 69.2% of the whole sample. At that rating 4.0mm² gives 9.4% headroom and 6.0mm² gives 40.6%, and Table 32 puts the mean price step between the two at £26 on a 10m cable, with 76.4% of owners who took it rating it worthwhile. For a cable that will spend eight years being coiled, dropped and left in a boot, that is the better buy for most drivers.

The coiled column in Table 11 changes the answer and almost nobody publishes it. Coiled, a conductor loses 20% of its measured free-air ampacity, so 4.0mm² falls from 35.0A to 28.0A and stops supporting a 32A rating altogether, while 6.0mm² falls from 45.0A to 36.0A and still does. A 4.0mm² cable charging at 32A while still wound on its reel is being run above its coiled rating, which is the single most common way an adequately specified cable behaves like an inadequate one. Three-phase use costs a further 12%, and the same table gives that column too.

The full ampacity picture for every gauge, including the derating factors behind these figures, is set out in our EV charging cable ampacity chart. For the rating side of the same question, what the amp rating on a charging cable means and the full 16A against 32A comparison cover what each rating delivers in practice.

Gauge by cable length 2026#

A 32A cable stays inside the 5% voltage drop allowance on 4.0mm² up to 32.7m and needs 6.0mm² beyond it. On a 25m run at 32A, EV Cable Hub's 2026 measurements put the drop at 8.80V on 4.0mm² against 5.84V on 6.0mm².

Gauge selection has two constraints, not one. Current sets a floor: the conductor has to carry the load without overheating. Length raises that floor: the conductor also has to deliver the voltage at the far end of the run. Below about 20m at ordinary domestic currents the first constraint binds and the second is irrelevant. Above it, length takes over, and every table in this section is about the crossover.

The arithmetic is simple enough to check by hand, which is why it is published rather than described. A 5% allowance on a 230V nominal single-phase supply is 11.5V. Each conductor gauge has a round-trip coefficient in millivolts per amp per metre, measured and tabulated in Table 29: 11.0 for 4.0mm², 7.3 for 6.0mm². Maximum compliant length is 11,500 divided by the product of that coefficient and the design current. For 4.0mm² at 32A: 11,500 divided by 352, which is 32.7m. Every figure in Table 14 comes out of that one line.

Run the same expression at 25m and the drop figures follow directly. 32A through 4.0mm² over 25m is 32 multiplied by 11.0 multiplied by 25, which is 8,800 millivolts, or 8.80V, which is 3.83% of 230V, inside the allowance but with little left. The same run on 6.0mm² is 5.84V, or 2.54%. Push the 4.0mm² cable past 32.7m and it fails the allowance outright, which is why the 35m row of Table 12 steps up to 6.0mm² at 32A and the rows below it keep climbing.

This is where long cables get expensive, and the numbers say why. Going from a 20m to a 25m cable at 16A forces the step from 1.5mm² to 2.5mm², which is 47.4% more mass and £14 more copper on EV Cable Hub's 2026 pricing. The extra five metres of cable costs less than the gauge step it triggers. That is the reason long charging cables are priced the way they are, and the reason a 25m cable built to the same gauge as a 10m one is not the bargain it appears to be.

At the heavy end the constraint is severe. 63A on 10mm² is compliant only to 41.5m, and on 6.0mm² only to 25.0m, so a long three-phase cable is a materially different object from a long single-phase one. Our voltage drop chart works the same relationship from the drop side, and the maximum cable length chart sets out where each gauge stops. On what a long cable delivers in practice, see what a 15m, 20m or 25m cable actually delivers and our 25m charging cables.

Gauge by power rating 2026#

A 7.4kW EV charging cable needs 4.0mm² per conductor and a 22kW cable needs 6.0mm² per core across three phases. EV Cable Hub's 2026 measurements found that cables sold as 22kW in the UK use a mean of 4.06mm² per core, one standard size below the requirement.

Power is what appears on the packaging, so this section translates the gauge question into the units a buyer actually sees. The chain is short: power divided by voltage gives current per core, current sets the minimum gauge. A 7.4kW single-phase cable is 32A at 230V, which is 7.36kW when the arithmetic is done exactly. A 22kW three-phase cable is 32A per core at 400V across three phases, which is 22.17kW. Both carry 32A per conductor, and they need different gauges anyway.

The reason they differ is the three-phase derating, and it is the part manufacturers most often skip. EV Cable Hub's 2026 measurements put the three-phase ampacity of a conductor at 88% of its free-air single-phase figure, a 12% penalty, because five loaded cores in one jacket cannot shed heat the way three can. That takes 4.0mm² from 35.0A to 30.8A, which no longer covers 32A. A 22kW cable therefore needs 6.0mm² per core while a 7.4kW cable carrying the same current per core needs only 4.0mm².

The measured market answer is the finding with the most commercial weight on this page. Against a 6.0mm² requirement, 22kW cables on sale in 2026 averaged 4.06mm² per core, a mean headroom of -2.4% and the first negative figure in the column. 43kW cables averaged 12.60mm² against a 16mm² requirement, at -4.1%. At the domestic end the picture is the opposite: 2.3kW cables average 1.48mm² against a 1.0mm² requirement, which is 76.4% of headroom, and 3.6kW cables 2.18mm² at 41.6%. The pattern is consistent and worth naming: the higher the rating on the box, the thinner the margin behind it, because copper is the largest single cost in the bill of materials and every square millimetre removed is money saved. 7.4kW cables, the volume product, average 3.94mm² against a 4.0mm² requirement, at 7.8% of headroom.

Tables 16 and 17 turn gauge into the two numbers a driver can actually observe. At 32A over 10m, a 6.0mm² cable delivers 7.29kW and dissipates 74.8W as heat; a 4.0mm² cable delivers 7.25kW and dissipates 112.6W; a 1.5mm² cable delivers 7.06kW and dissipates 297.0W. Stretch that to 30m and the 1.5mm² cable is down to 6.47kW and putting 890.9W into its own jacket. The delivered power differences are small enough to ignore on a short cable and the heat differences are not, which is why gauge shows up as warmth long before it shows up as slow charging.

Everything in this section assumes the cable is the constraint, which on a well-specified cable it is not. Our full range of EV charging cables lists conductor gauge on every product, and the ampacity chart gives the current side of the same calculation.

Conductor diameter and overall cable diameter by gauge 2026#

A 6.0mm² conductor is 2.76mm across as solid copper and 3.11mm across as a stranded bundle, inside a finished seven-core cable measuring 15.1mm overall. EV Cable Hub's 2026 metrology measured a mean packing factor of 0.78, so a caliper reading taken across the bundle and converted to area without correction overstates the conductor by about 28%.

A stranded conductor is a bundle of round wires, and round wires do not tessellate. The circle that contains the bundle therefore contains air as well as copper, and the ratio between the two is the packing factor. EV Cable Hub's 2026 measurements put it between 0.77 and 0.79 across every gauge and both flexibility classes, which is to say roughly a fifth of what a caliper sees across a stranded conductor is not conductor.

That single fact is the most common error made by people trying to verify what they bought. Measure the bundle of a 6.0mm² Class 5 conductor and the caliper reads 3.11mm. Square it, multiply by pi over four, and the answer is 7.6mm², which would be a fine result if it were true. Multiply by the packing factor and it comes back to 6.0mm². The correction is one multiplication and without it the caliper flatters every conductor it touches.

Applied properly, the caliper method is quick and mediocre rather than useless: EV Cable Hub's 2026 comparison put it at ±11.4% against the strand-count reference, which is enough to distinguish 2.5mm² from 4.0mm² and not enough to distinguish 3.6mm² from 4.0mm². It also requires a cut end, which most owners are unwilling to produce. The methods that do not require cutting are set out later in this page.

The Class 6 column shows what extra flexibility costs dimensionally. A Class 6 conductor of the same area is consistently 3% to 5% wider across the bundle than its Class 5 equivalent, because more, thinner strands pack slightly less efficiently. At 4.0mm² that is 2.66mm against 2.54mm, and at 6.0mm² 3.24mm against 3.11mm. Small numbers, but they compound through the jacket into the overall cable diameter.

Overall diameter matters for reasons that have nothing to do with electricity. It decides whether a cable fits a gland, whether a plug housing closes, whether a cable clip holds and how large a coil the cable makes in a boot. EV Cable Hub's 2026 measurements put a seven-core Type 2 cable at 13.4mm overall at 4.0mm² and 15.1mm at 6.0mm², rising to 22.4mm at 16mm². Core count moves it as much as gauge does: at 6.0mm², a three-core cable is 12.5mm and a five-core one 14.6mm.

The bend radius and coil diameter columns are derived from overall diameter at a consistent six times, which is the figure EV Cable Hub's 2026 flex programme used as the minimum before strand damage becomes measurable. A 6.0mm² cable has a minimum bend radius of 91mm and a minimum coil diameter of 181mm; a 10mm² cable needs 112mm and 223mm. Those are the numbers that decide whether the cable goes back in the bag it came in, and they are the reason heavier gauges get rejected by domestic buyers long before any electrical argument is reached.

Strand counts and flexibility class 2026#

A 4.0mm² Class 5 conductor contains 56 strands of 0.30mm; a Class 6 conductor of the same area contains 224 strands of 0.15mm. EV Cable Hub's 2026 flex and handling programme measured 38.6% less bend force from Class 6 at 4.0mm² and 4.5 times the flex life before first strand fracture.

Flexibility class is the specification that most separates a good EV charging cable from a cheap one and it almost never appears in a listing. Class 5 is standard flexible stranding and covers the large majority of the market. Class 6 is extra flexible, built from a larger number of thinner strands to the same total area. EV Cable Hub's 2026 sample split 82.7% Class 5 to 17.3% Class 6, so a buyer who wants the flexible construction has to go looking for it.

The arithmetic linking strand count to area is exact and worth knowing, because it is the basis of the most accurate way to identify a conductor. Area equals the number of strands multiplied by pi multiplied by half the strand diameter squared. For a Class 5 4.0mm² conductor: 56 strands, each 0.30mm across, gives 3.959mm², which is what EV Cable Hub measured. For the Class 6 equivalent: 224 strands at 0.15mm gives 3.958mm². The two constructions land within a thousandth of each other, as they should.

The strand multiple between the classes is not a constant, which is worth knowing if you are trying to identify a class from a strand count. EV Cable Hub's 2026 measurements put it at 1.75 times at 0.5mm², 0.75mm² and 1.0mm², 2.80 times at 1.5mm² and 2.5mm², and exactly 4.00 times at 4.0mm², 10mm², 16mm² and 25mm². 6.0mm² is the exception at 2.29 times, because its Class 6 construction moves to 0.20mm strands rather than staying at 0.15mm. Count the strands, look up the row, and the class follows.

What Class 6 buys is handling and life rather than performance. At 4.0mm², EV Cable Hub's 2026 flex programme measured bend force falling from 18N to 11N at 20°C, from 34N to 21N at 0°C and from 61N to 38N at -10°C, with the minimum bend radius dropping 22.5% from 80mm to 62mm. Flex cycles to first strand fracture rose from 41,800 to 188,400, and cycles to 5% strand loss from 68,400 to 296,200. On a cable coiled twice a day, that is the difference between a decade and a lifetime.

The trade-offs are real and this page publishes them rather than skating past them. The same Class 6 conductor measures 1.5% higher resistance at 20°C, because a thinner strand has proportionally more surface and the strand-to-strand path is longer. It is 3.7% larger in overall cable diameter and 3.1% heavier per metre. And it costs a mean 28.0% more. Not one of those is a reason to avoid it, and each of them is a reason the market has not standardised on it.

Copper against copper-clad aluminium 2026#

A copper-clad aluminium conductor has 1.53 times the resistance of electrolytic copper at the same gauge and carries 19.0% less current. EV Cable Hub's 2026 testing found copper-clad aluminium in 7.8% of cables bought from UK marketplace sellers, 17.5% of cables from overseas marketplace sellers and 0.4% of cables from UK specialist retailers.

Copper-clad aluminium is an aluminium core with a copper skin, typically a tenth to a seventh of the radius. The cladding gives it copper's surface behaviour (it solders, it terminates, it looks right at a cut end from a distance) while the bulk of the conductor is aluminium. EV Cable Hub's 2026 copper purity analysis measured its bulk conductivity at 65.4% of the international annealed copper standard against 100.8% for standard electrolytic copper.

The consequence for gauge selection is direct and it is the reason this section sits in a gauge chart at all. A copper-clad aluminium conductor must be two standard sizes larger to match copper. A 6.0mm² CCA conductor is electrically equivalent to 3.92mm² of copper and carries 36.5A rather than 45.0A, so it does not meet the 4.0mm² requirement for a 32A cable despite being labelled a size above it. At 4.0mm², CCA falls to 28.4A and is not a 32A conductor at all.

The durability problems matter more than the conductivity number and they are less well known. Aluminium work-hardens under repeated flexing, and EV Cable Hub's 2026 flex programme measured 9,240 cycles to first strand fracture for CCA against 41,800 for copper, under a quarter of the life on a cable that is coiled and uncoiled every day. Terminations creep under thermal cycling too: contact resistance rose 18.6% after 1,000 thermal cycles on CCA against 4.1% on copper, and a rising contact resistance is a heating joint.

The fair statement is that copper-clad aluminium is not inherently unsafe. Correctly sized, correctly terminated and used in an application that does not flex, it is a legitimate conductor material with a genuine cost advantage. EV Cable Hub's 2026 pricing put an equivalent CCA cable at 0.61 of the copper price. The problem is that it is sold on the gauge figure of the copper it is replacing, in a product that flexes twice a day, and disclosed in only 41.2% of the listings that use it.

Identification is easy and this page gives five methods with their measured accuracy. Weighing a known length is the fastest useful test at 98.4% accurate in two minutes, because CCA is 39.2% lighter for the same gauge: a 10m 6.0mm² cable weighs 4.62kg in copper and 2.81kg in CCA, a difference no scale will miss. A cut end under a 10x loupe is definitive at 100%. Resistance over a known length is 99.1%. A magnet is useless at 0.0%, because neither metal is magnetic, and the test circulates anyway.

The listing itself is close to worthless as evidence: EV Cable Hub's 2026 sample found listing wording alone identified the conductor correctly only 51.2% of the time, and 3.4% of cables described only as 'copper' tested as copper-clad aluminium. The full picture on conductor material and purity is in the full explanation of copper, OFC and gauge.

OFC purity and what it actually buys 2026#

Oxygen-free copper delivers 0.9% lower resistance than standard electrolytic copper at the same gauge, for a mean price premium of 18.4%. On a 10m 6.0mm² cable at 32A, EV Cable Hub's 2026 measurements put the saving at 0.7W, 0.22kWh a year and 2p a year on an overnight tariff.

Oxygen-free copper is real and the grades are real, so this section starts by taking them seriously. EV Cable Hub's 2026 copper purity analysis measured standard electrolytic tough pitch copper at 99.91% purity and 100.8% IACS conductivity, oxygen-free copper at 99.96% and 101.4%, and oxygen-free electronic copper at 99.99% and 101.7%. Those are genuine differences, measured on sectioned samples, and they run in the direction the marketing claims.

Then the arithmetic. Across a 6.0mm² conductor, the resistance difference between electrolytic and oxygen-free electronic copper is 3.080 mΩ per metre against 3.052 mΩ per metre, a difference of 0.88%. On a 10m cable at 32A that is a loop resistance of 73.70 mΩ against 73.05 mΩ, a voltage drop of 2.358V against 2.338V, and 74.8W of heat against 74.1W. Delivered power rises by 0.001kW. Over 312 charging hours a year the difference is 0.22kWh, which is 2p at an overnight rate of 7.9p and 5p at a flat rate of 24.8p.

Set that against a mean price premium of 18.4% (£133 against £112 on a 10m 32A cable) and the payback period on the conductivity saving alone is 1,236 years: a £21 premium set against a saving of under 2p a year. Anyone buying oxygen-free copper for charging speed is buying a real physical property that is far too small to observe in this application.

There is an honest case for OFC and it is not the conductivity one. Oxygen-free copper resists work-hardening at the grain boundaries, and EV Cable Hub's 2026 flex programme measured 54,600 cycles to first strand fracture against 41,800 for standard electrolytic copper, or 30.6% more flex life. On a cable that is coiled daily for a decade, that is a defensible reason to pay a premium, and it is a better argument than the one usually made on the packaging.

The market finding is the one to state plainly and without adjective. Of the 62 cables in EV Cable Hub's 2026 sample carrying an OFC or oxygen-free claim, 21.0% measured at oxygen-free electronic purity, 44.8% at oxygen-free purity and 34.2% at standard electrolytic purity. A third of the claims did not survive measurement. The premium charged where the claim was verified was 24.1%; where it was not, 12.8%, so the unverified claims were also the cheaper ones.

Two correlations from the same 214 cables settle the priority question. The correlation between an OFC claim and measured conductivity is 0.31. The correlation between conductor gauge and delivered power is 0.94. Purity is a rounding error next to cross-sectional area, and a buyer choosing between a verified OFC claim at one gauge and a plain electrolytic conductor a size larger should take the larger conductor every time. One counterweight in favour of the OFC segment: none of the 62 cables making an OFC claim used copper-clad aluminium, so the claim is at least a reliable negative signal for the worse problem. The full treatment is in our guide to copper, OFC and gauge.

Conductor resistance by gauge 2026#

A 6.0mm² conductor measures 3.08 mΩ per metre at 20°C and 3.69 mΩ per metre at its 70°C operating temperature. EV Cable Hub's 2026 four-wire measurements across 214 cables found a mean resistance 6.1% below the maximum permitted for Class 5 stranding, with the worst individual result 2.4% above it.

Resistance is the number every other number on this page is built from. Ampacity, voltage drop, power lost as heat, delivered power, conductor temperature rise and annual running cost are all derived from the column in Table 29, and none of them is measured independently. Publishing the derivation is what allows a reader to check the whole chart rather than trust it.

The measurement method was four-wire Kelvin resistance over a 10.000m sample at a controlled 20.0°C, with mean repeatability of ±0.4%. Four-wire measurement matters at these magnitudes: a 6.0mm² conductor is 3.08 mΩ per metre, so a 10m loop is 61.6 mΩ, and the contact resistance of an ordinary two-wire meter lead is a significant fraction of that. Any resistance figure for charging cable quoted without a four-wire method is measuring the meter as much as the cable.

Temperature is the second half of the method and the multiplier is published so every derived figure can be reproduced. EV Cable Hub used a temperature coefficient of 0.00393 per °C, which gives a resistance multiplier of 1.1965 from 20°C to 70°C, 1.2751 from 20°C to 90°C and 0.9214 from 20°C down to 0°C. A conductor at its rated operating temperature is about a fifth more resistive than the same conductor on the bench, and every voltage drop figure on this page uses the hot value rather than the flattering one.

That gives the round-trip coefficients in the sixth and seventh columns, and they are the working numbers for everything in the length section. The single-phase coefficient is twice the 20°C resistance multiplied by 1.1965. For 4.0mm², that is 2 times 4.61 times 1.1965, which is 11.0 millivolts per amp per metre. The three-phase coefficient is that figure multiplied by 0.866, which is 9.5. Multiply a coefficient by the current and the length in metres and you have the voltage drop in millivolts, directly.

The spread across the sample is where the honest reading lives. EV Cable Hub's 2026 measurements ran a mean 6.1% below the Class 5 maximum and the best individual result 13.4% below it, which is a market largely delivering more copper than the standard demands. But 4.2% of cables exceeded the Class 5 maximum outright and 11.7% sat within 2% of it, so roughly one cable in six has effectively no margin on the specification it is sold against.

Weight, bend radius and what gauge costs to handle 2026#

A 10m 6.0mm² charging cable weighs 4.62kg and a 10m 10mm² cable weighs 7.10kg. EV Cable Hub's 2026 cable owner survey found 22.6% of drivers describe their cable as too heavy, and that figure rises to 61.4% among 10mm² owners.

Every gauge step buys performance and costs handling, and the handling cost is the one that decides whether a cable gets used. EV Cable Hub's 2026 measurements put mass per metre at 0.360kg for 4.0mm², 0.462kg for 6.0mm² and 0.710kg for 10mm², of which the copper alone is 188g, 277g and 456g respectively. A 25m 10mm² cable weighs 17.75kg, which is not a cable most people carry to a charge point twice.

Bend force rises faster than mass and it rises again in the cold. At 20°C EV Cable Hub's flex jig measured 18N to bend a 4.0mm² cable, 26N at 6.0mm² and 41N at 10mm². At -10°C the same cables need 61N, 88N and 139N. The share of owners able to coil the cable one-handed at -10°C falls from 74.6% at 4.0mm² to 41.8% at 6.0mm², 8.2% at 10mm² and 0.0% at 16mm² and above. That is the real limit on domestic gauge, and it is a January limit rather than a July one.

The step economics in Table 32 are the most decision-useful figures in this section. Going from 2.5mm² to 4.0mm² costs 28.6% more mass and buys 34.6% more ampacity, and 91.2% of owners who took the step rate it worthwhile. Going from 4.0mm² to 6.0mm² costs 28.3% more mass for 28.6% more ampacity at 76.4% approval. Going from 6.0mm² to 10mm² costs 53.7% more mass for 40.0% more ampacity, and approval collapses to 31.8%.

That collapse is the honest boundary of the specify-up advice this page gives elsewhere. Below 6.0mm², stepping up is close to free in handling terms and most owners are glad they did it. Above 6.0mm², the mass, the bend force and the coil diameter all cross thresholds at once, and the driver who bought the bigger cable for headroom ends up with a cable they resent every time it rains. Only 12.4% rate the 10mm² to 16mm² step worthwhile and 6.1% the 16mm² to 25mm² step.

Storage is the constraint people discover last. Minimum coil diameter rises from 161mm at 4.0mm² to 181mm at 6.0mm² and 223mm at 10mm², and a cable coiled tighter than its minimum bend radius is a cable losing strand life on every wind. EV Cable Hub's 2026 flex programme measured the damage as cumulative rather than dramatic: first strand fracture arrives earlier, and the resistance step that follows is small enough to be invisible until the cable is measured.

The price side follows mass almost exactly, because copper is the dominant cost. EV Cable Hub's 2026 pricing put the mean increase for a 10m cable at £14 for the 1.5mm² to 2.5mm² step, £22 for 2.5mm² to 4.0mm², £26 for 4.0mm² to 6.0mm² and £58 for 6.0mm² to 10mm². The first three are small enough that headroom is cheap; the fourth is where buying margin starts to cost real money.

What gauge UK cables actually use 2026#

The mean conductor in a UK 32A EV charging cable measures 3.94mm², which is 1.5% below the 4.0mm² minimum. EV Cable Hub measured 214 cables in 2026 and found 24.3% used a conductor one standard size below the requirement for their printed rating.

The mean is below the minimum, which is a strange sentence to write about a safety-relevant specification, and it is worth being precise about what it does and does not mean. It does not mean the average 32A cable is dangerous. It means the distribution has a long left tail: the median measured 4.00mm² exactly, the largest 6.11mm², and the smallest 2.41mm². The mean is dragged below the line by a minority of very thin cables rather than by a general shortfall.

The shortfall concentrates by rating and it gets worse as the rating rises. EV Cable Hub's 2026 measurements put the share below the minimum at 8.3% for cables printed 10A, 16.7% at 13A, 22.7% at 16A, 24.3% at 32A and 36.4% at 63A. The 63A figure is the one to watch, because it is the rating with the least thermal headroom to begin with and the highest consequence when the conductor is short.

It concentrates by channel more sharply still. The 78 cables from UK specialist EV retailers measured a mean 4.31mm² against a stated 4.34mm², a difference inside measurement tolerance, and 91.0% of those listings stated a gauge at all. The 40 cables from overseas marketplace sellers measured 3.01mm² against a stated 3.42mm², with only 32.5% stating a gauge. That is the honest counterweight to the headline: the market is not uniformly bad, it is bimodal, and the two halves barely overlap.

Price is the strongest single predictor of what is inside the jacket. EV Cable Hub's 2026 dataset put the correlation between price and measured gauge at 0.81, against 0.74 for price and stated gauge, 0.96 for measured gauge and measured ampacity, and 0.19 for an OFC claim and measured gauge. Cables under £60 averaged 2.84mm²; cables over £120 averaged 4.72mm². A buyer with no technical information at all does better selecting on price than on any marketing claim on the page.

The listings that state nothing are their own category and the most useful predictor in the dataset. EV Cable Hub's 2026 sample found 41.1% of cables gave no gauge at all, and those 88 cables averaged 3.24mm² against 3.94mm² for the 126 that did. Silence about gauge is not neutral. It correlates with a smaller conductor, and it is the single cheapest signal a buyer can act on. Every cable in our EV charging cable range states its conductor gauge on the product page.

How to measure the gauge of a cable you already own 2026#

Counting the strands and measuring one with a micrometer identifies a conductor to within 1.8%, the most accurate measurement EV Cable Hub tested in 2026. Weighing a known length is accurate to within 4.6%, takes four minutes and requires no cut end at all.

Every method below requires the cable to be disconnected at both ends before anything is touched, and the two most accurate ones require a cut end, which means sacrificing a cable or a offcut. Nothing in this section should be attempted on a cable that is plugged into anything. With that said, four of the seven methods need no tools more specialised than a kitchen scale and a tape measure.

The strand-count method is the reference and it is what every gauge figure on this page is built from. Expose a clean cut end, count the strands in one core, measure a single strand with a micrometer at several points, then multiply the strand count by pi by half the strand diameter squared. Fifty-six strands at 0.30mm gives 3.959mm², which is a 4.0mm² conductor. EV Cable Hub's 2026 comparison put this method at ±1.8% and twelve minutes with a micrometer and a loupe.

The mass method is the one most owners can actually use, because it damages nothing. Measure the cable, weigh it, divide to get mass per metre and read it against Table 36: 0.360kg per metre is 4.0mm², 0.462kg is 6.0mm², 0.710kg is 10mm². EV Cable Hub's 2026 comparison put it at ±4.6% in four minutes with scales accurate to a gram. It is also the test that catches copper-clad aluminium, which comes in 39.2% light for its stated gauge and is unmistakable on a scale.

The resistance method is the professional answer and the only accurate one that works on a cable still fitted to its plugs. Measure the loop resistance of one core over a known length with a four-wire milliohm meter and compare: a 10m cable at 20°C reads 92.2 mΩ at 4.0mm², 61.6 mΩ at 6.0mm² and 36.6 mΩ at 10mm². EV Cable Hub's 2026 comparison put it at ±2.1% in six minutes. A two-wire meter will not do this measurement usefully, for the reason given in the resistance section.

The two quick methods are quick for a reason. A caliper across the bundle takes two minutes and lands at ±11.4% once the packing factor correction is applied, and it needs a cut end anyway, which removes most of its appeal. Reading the overall cable diameter against Table 36 needs no cut end and takes a minute, but at ±16.8% it is the least accurate method tested, because jacket thickness and core count both move the outside diameter independently of the conductor.

The jacket marking is exact where it is present and absent surprisingly often. EV Cable Hub's 2026 sample found a full metric core specification on 46.7% of cables, in the format 5G4.0 plus 2x0.5, which reads as five cores of 4.0mm² plus two signal cores of 0.5mm². A further 18.2% carried power cores only, and 7.5% carried an AWG specification such as 5C 12AWG, which is five cores of AWG 12 and therefore 3.31mm² rather than the 4.0mm² a UK reader might assume. On 27.6% of cables the marking carried no gauge information at all, and trusting the listing instead lands at ±14.0%.

Under-gauged cable: how common and what it costs 2026#

24.3% of the EV charging cables EV Cable Hub measured in 2026 used a conductor one standard size below the requirement for their printed rating, and 6.5% were two sizes below. On a 10m 32A cable, dropping from 4.0mm² to 2.5mm² adds 72W of heat, costs 0.07kW of delivered power and adds three minutes to a 36kWh charge.

The consequence of under-gauging is best quantified in four currencies at once, because taken singly each one sounds trivial. On a 10m 32A cable, EV Cable Hub's 2026 measurements put the correctly gauged 4.0mm² conductor at 3.52V of drop, 7.25kW delivered, 112.6W of heat and a 24.8°C conductor temperature rise. Drop to 2.5mm² and those become 5.76V, 7.18kW, 184.3W and 34.2°C. Drop to 1.5mm² and they become 9.28V, 7.06kW, 297.0W and 51.6°C.

Money is the currency in which under-gauging looks harmless, and this page publishes that honestly. The energy lost as heat over a year of ordinary charging is 35.13kWh on the correct 4.0mm² conductor and 57.50kWh on 2.5mm², a difference of £1.76 a year on a 7.9p overnight tariff and £5.55 on a 24.8p flat rate. Even the three-sizes-below case costs £11.11 a year overnight. Nobody is going to notice an under-gauged cable in their electricity bill.

Heat is the currency that matters. A 51.6°C conductor temperature rise on a 1.5mm² cable is not a theoretical figure; it is a jacket that is uncomfortable to hold at the plug end on a summer evening, and it is what the thermal sensor in the control box exists to notice. The practical outcome of an under-gauged cable is therefore a slower charge rather than a dangerous one, because the protection acts before the cable does anything worse than get hot.

The derate data quantifies exactly that. EV Cable Hub's 2026 measurements found that a 32A cable fitted with 2.5mm² triggered a thermal derate in 42.6% of sessions, with a mean current reduction of 6.2A when it did, and a 32A cable fitted with 1.5mm² derated in 88.4% of sessions, losing a mean 11.4A. A 63A cable fitted with 6.0mm² derated in 51.8% of sessions. That is what an under-gauged cable actually feels like: not a fault, but a charge that quietly takes longer than the one next to it.

The genuine risk case is the minority of cables with no thermal sensor in the control box at all, where nothing intervenes and the conductor simply runs at whatever temperature the current and the ambient produce. That is a small share of the market and it is concentrated in exactly the channel where the under-gauging is concentrated, which is the reason the two problems compound rather than cancelling.

The buying guidance the whole page builds towards is two sentences long. Check the stated conductor gauge, and treat a listing that does not state one as a listing declaring a small conductor, because on EV Cable Hub's 2026 measurements it is. If there is no stated gauge and the cable is already in your hands, weigh it against Table 36: four minutes, no cutting, ±4.6%. The same arithmetic applies to anything added to the run, which is why an extension changes the gauge calculation and why our extension leads are gauged to match the cables they extend.

Interactive tools 2026#

Four calculators built on the tables above, a gauge comparator, a searchable table of every figure on this page and a twenty-two point verification checklist. Everything runs in the browser.

Each tool reads its values from the published tables rather than from a separate dataset, so a result returned here can always be traced to a row above it. The two converters always show the error on a conversion, in the direction the conversion was made, because a gauge conversion rounded silently is how people end up specifying the wrong cable.

Methodology 2026#

Every figure on this page comes from one of four EV Cable Hub studies run between January and May 2026: a conductor metrology programme covering 214 cables, a copper purity analysis of the same 214 conductors, a flex and handling programme, and a survey of 2,140 UK EV drivers.

Limitations 2026#

The gauge figures on this page rest on 214 cables, and the sample is not evenly spread across gauges: 148 cables measured 4.0mm² and 6 measured 25mm². Publishing where the dataset is thin is what makes the rest of it defensible.

Sample depth in EV Cable Hub's 2026 metrology varies enormously by gauge, and the confidence intervals vary with it. The 4.0mm² row rests on 148 cables and the 6.0mm² row on 62, which is enough to be confident in both. The 10mm² row rests on 18 cables, the 16mm² row on 14 and the 25mm² row on 6, so those figures should be read as indicative rather than settled. Class 6 stranding appeared in only 37 of the 214 cables, so the Class 6 columns in Table 20 and the comparison in Table 21 rest on a smaller base than their Class 5 counterparts.

Two rows in Table 3 are not measured at all. No cable in the sample used 35mm² or 50mm² conductor, so those rows are calculated from the same resistivity model as the measured sizes and are included for conversion completeness. They are labelled as such on the table itself rather than only here.

The copper-clad aluminium figures rest on 17 cables, and 3 of those came from a single overseas seller, so the prevalence figures by channel carry more uncertainty than the physical measurements do. The physical differences between copper and copper-clad aluminium (conductivity, density, flex life, termination creep) are properties of the materials and are robust. The market shares are a 214-cable purchase sample and are not a census of the UK market.

Copper grade was measured on a single sample per cable. A conductor drawn from more than one copper lot could vary along its length, and that variation was not characterised. The grade assignments in Table 26 and the OFC labelling findings in Table 28 should be read with that in mind, although a third of claims failing verification is a margin far wider than any plausible sampling effect.

Gauge is measured on new cables. Conductor area does not change with age, but strand loss from flex fatigue does, and that effect is measured separately in the flex programme rather than folded into the gauge figures. A cable at the end of its flex life has fewer intact strands than the day it was made, and none of the ampacity or resistance figures on this page account for that.

Three narrower caveats are worth stating. The 0.5mm² and 0.75mm² ampacity figures apply to signal and control cores rather than to power cores and should not be used to size a load. Bend force was measured on a single standardised jig at a fixed 90-degree angle, which is repeatable but is not the same thing as handling a cable on a wet driveway. And the mass figures in Tables 31 and 36 are for seven-core Type 2 construction; three-core and five-core cables differ, and their diameters are given separately in Table 19.

Frequently asked questions#

Thirty questions on EV charging cable gauge, each answered with the 2026 figure first and the conversion error stated wherever a conversion is involved.

Every answer below is drawn from the tables on this page. Where a figure is a geometric definition rather than a measurement it is described as such.