Cable Sizing · Updated 31 July 2026 · For contractors and project engineers
A cable big enough to carry the current can still be too small. By the time the supply reaches the far end of a long run the voltage has sagged, and the equipment there sees less than it was designed for. BS 7671 puts a limit on that sag — and the limit is not a single number. It depends on what the circuit serves, where the supply comes from, and how long the wiring system is. Here is how the rule actually reads, how to calculate the drop, and the two places designs most often go wrong.
Set a circuit up and watch the voltage profile from the origin to the equipment. Drag the length past 100 m and the limit line itself moves. Switch the supply to a sub-main feed and the two legs start sharing one budget — which is the failure mode this page exists to explain.
Regulation 525.1 requires the voltage at the terminals of any equipment to be no less than the lower limit of that equipment's own standard. Regulation 525.201 says that requirement is met if the drop does not exceed the values in Appendix 4, and those values live in Table 4Ab. There are two rows, and which one applies depends on where the supply comes from, not on the size of the installation.
| Supply | Lighting | Other uses | At 230 V | At 400 V |
|---|---|---|---|---|
| (a) Direct from a public LV distribution system | 3% | 5% | 6.9 V / 11.5 V | 12 V / 20 V |
| (b) From a private LV supply * | 6% | 8% | 13.8 V / 18.4 V | 24 V / 32 V |
* The footnote to row (b) is the part that gets missed: the drop within each final circuit must still not exceed the row (a) values. A private supply — your own transformer, a generator, a site supply — buys headroom for the distribution, because you control the source voltage. It does not relax the final circuit.
Beneath the table sits a sentence almost no online calculator implements. Where the wiring systems of the installation are longer than 100 m, the permitted drop may be increased by 0.005% per metre beyond 100 m, and that increase may not exceed 0.5%. So the 5% limit becomes 5.25% at 150 m and 5.5% at 200 m — and then stops, because 100 m beyond the threshold at 0.005%/m is exactly the 0.5% cap. Past 200 m the limit never moves again.
| Wiring system length | Allowance | Permitted drop |
|---|---|---|
| 100 m or less | — | 5.00% |
| 120 m | +0.10% | 5.10% |
| 150 m | +0.25% | 5.25% |
| 200 m | +0.50% | 5.50% |
| 300 m and beyond | +0.50% (capped) | 5.50% |
It is a small allowance, but on a long run it is often the difference between one cable size and the next — and it is defensible, because it is in the standard. Claiming it means stating the wiring-system length you based it on.
Regulation 525.202 names both ends: from the origin of the installation to the terminals of the equipment. Not from the distribution board. Not to the accessory. That single sentence is what makes voltage drop cumulative, and it is where most failures come from — because each cable was checked on its own and nobody added them up.
In the figure above the sub-main spends 2.92% and the final circuit 2.87%. Each is comfortably inside 5%. Together they are 5.79%, and the installation does not comply. The socket at the end of that circuit does not care which cable spent the budget.
Two practical consequences follow. First, a sub-main is worth uprating even when it passes, because every circuit downstream inherits whatever it spends — one bigger sub-main is usually cheaper than five bigger final circuits. Second, a change to a sub-main silently changes every circuit it feeds, which is exactly the bookkeeping that hand calculations and spreadsheets lose track of when a design moves.
The current used is the design current Ib — the demand after diversity, not the rating of the protective device. Sizing voltage drop on In when the real load is half of it is a common and expensive way to over-specify a cable.
Every cable size has a tabulated voltage-drop figure in millivolts per ampere per metre (mV/A/m) in BS 7671 Appendix 4. The drop is:
Vd = (mV/A/m × Ib × L) / 1000
with Ib the design current in amps and L the route length in metres — the length of the run, not the length of conductor in it; the tabulated values already account for the return path. Divide by the nominal voltage and multiply by 100 for the percentage. For a three-phase circuit take the value from the three- or four-core column and express the result against 400 V, not 230 V: a common error is to compute a three-phase drop and then divide it by the phase voltage, which overstates the percentage by about 73%.
For conductors larger than 16 mm² the tables stop giving a single value and split it into a resistive part (mV/A/m)r and a reactive part (mV/A/m)x, because the conductor's inductance is no longer negligible next to its resistance. Appendix 4 §6.2(b) combines them with the load's power factor:
(mV/A/m) = (mV/A/m)r × cos θ + (mV/A/m)x × sin θ
where cos θ is the power factor. The single (z) figure tabulated for large cables assumes unity power factor and is the pessimistic case for the resistive component; at a poor power factor the reactive term matters and the answer changes. Below 16 mm² the single value is used directly and this refinement does not arise.
A working subset of the Appendix 4 tables for the cables that come up most often. The first figure in each cell is the single-phase (two-core) value, the second the three-phase (three- or four-core) value. Aluminium is not tabulated below 16 mm².
| Cable | 1 mm² | 1.5 mm² | 2.5 mm² | 4 mm² | 6 mm² | 10 mm² | 16 mm² | 25 mm² | 35 mm² | 50 mm² | 70 mm² | 95 mm² |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| PVC 70 °C copper | 44 / 38 | 29 / 25 | 18 / 15 | 11 / 9.5 | 7.3 / 6.4 | 4.4 / 3.8 | 2.8 / 2.4 | 1.75 / 1.5 | 1.25 / 1.1 | 0.93 / 0.8 | 0.63 / 0.55 | 0.46 / 0.41 |
| XLPE 90 °C copper | 46 / 40 | 31 / 27 | 19 / 16 | 12 / 10 | 7.9 / 6.8 | 4.7 / 4 | 2.9 / 2.5 | 1.85 / 1.6 | 1.35 / 1.15 | 1 / 0.86 | 0.68 / 0.59 | 0.5 / 0.44 |
| PVC 70 °C aluminium | — | — | — | — | — | — | 4.5 / 3.9 | 2.9 / 2.5 | 2.1 / 1.8 | 1.55 / 1.35 | 1.05 / 0.92 | 0.77 / 0.68 |
Indicative values for the reference methods these tools cover — verify every figure against your own copy of BS 7671 Appendix 4 for the actual cable, installation method and reference table.
Voltage drop is the odd one out among the checks in a cable calculation. Current-carrying capacity can be rescued by a better installation method; earth-fault loop impedance can be rescued by a more sensitive device. Voltage drop has only three levers — a bigger conductor, a shorter route, or less current — and no protective device, however good, changes the answer.
The relationship is close to inverse: mV/A/m falls by roughly 60% each time you go up two standard sizes, so going from 2.5 to 4 mm² cuts the drop by about 39%, and 2.5 to 6 mm² by about 59%. That is why the fix is almost always "one or two sizes up" rather than a redesign — but on a long run, one size up on a sub-main beats one size up on every circuit behind it.
The limits are not arbitrary caution. Equipment behaves measurably worse on low voltage, and the two headline cases behave very differently — which is precisely why lighting gets the tighter number:
Table 4Ab is a steady-state limit. Regulation 525.203 accepts a larger drop for motor starting and other high-inrush conditions, provided the voltage stays within the limits of the relevant equipment standard. A direct-on-line motor drawing six or seven times full-load current for a few seconds will breach the steady-state figure comfortably, and that is expected — but you have to check it against what the motor and its contactor will actually tolerate, not simply ignore it. The two checks answer different questions and you need both.
The same logic applies in reverse to harmonics: the drop should include any effect due to harmonic currents, which on a load with significant third harmonic means the neutral is carrying current the fundamental calculation never accounted for.
Voltage drop is one of the five checks in a compliant cable calculation, alongside current-carrying capacity (with its derating factors), earth-fault loop impedance, the adiabatic check on the protective conductor, and breaking capacity against the prospective fault level. It is the check most likely to be the binding constraint on a long run, and the one least forgiving of a design change, because it is cumulative and network-wide.
Our free BS 7671 cable sizing calculator runs all five for a distribution board and prints a calculation report, and the kW to amps calculator will get you from a load in kilowatts to the design current this page starts from.
Vd — voltage drop, the difference between the voltage at the origin and at the equipment terminals.
mV/A/m — the tabulated drop per ampere per metre of run, from BS 7671 Appendix 4.
Ib — design current: the demand the circuit is actually intended to carry, after diversity.
In — the rated current of the protective device, which is not what voltage drop is calculated on.
U0 — nominal line-to-neutral voltage, 230 V in the UK.
U — nominal line-to-line voltage, 400 V, the reference for a three-phase drop.
Origin of the installation — where the supply enters and the installation begins; one end of every voltage-drop measurement.
Sub-main — a distribution circuit feeding another board rather than a final load.
cos θ — power factor, needed once the conductor is larger than 16 mm² and reactance matters.
For an installation supplied directly from the public low-voltage network, BS 7671 Appendix 4 Table 4Ab gives 3% for lighting and 5% for other uses, measured from the origin of the installation to the terminals of the equipment. At 230 V single-phase that is 6.9 V and 11.5 V; at 400 V three-phase, 12 V and 20 V.
Voltage drop = (mV/A/m × Ib × L) / 1000, where mV/A/m is the cable's tabulated value from BS 7671 Appendix 4, Ib is the design current in amps, and L is the route length in metres. Divide the result by the nominal voltage and multiply by 100 to get the percentage. For a three-phase circuit use the three- or four-core column and express the answer against 400 V, not 230 V.
From the origin of the installation to the terminals of the equipment, so it is cumulative. Regulation 525.202 is explicit about the two end points. The drop across a sub-main and the final circuit it feeds must be added together and the total kept within the limit — which is why a final circuit that passes comfortably on its own can still put the installation over.
Yes. Where the wiring system of the installation is longer than 100 m, Table 4Ab allows the limits to be increased by 0.005% per metre beyond 100 m, up to a maximum increase of 0.5%. So a 5% limit becomes 5.25% at 150 m and 5.5% at 200 m, after which the allowance stops growing because it has hit its cap.
Yes. Where the installation is fed from a private LV supply — its own transformer or a generator rather than the public network — Table 4Ab allows 6% for lighting and 8% for other uses. But the footnote to that row still requires the drop within each final circuit not to exceed the 3% and 5% values, so the extra headroom is for the distribution, not for the final circuit.
Equipment receives less voltage than it was designed for. Induction motor torque falls roughly as the square of the voltage, so a 5% drop costs about 10% of torque, and motors draw more current to compensate, which heats them. Filament lamp output falls faster still. Electronic equipment usually copes but its input current rises. Long-term the risk is nuisance tripping, overheating and shortened equipment life rather than an immediate failure.
Increase the conductor size, shorten the cable route, move the distribution board nearer the load centre, or split the load across more circuits. On a network it is often cheaper to uprate one sub-main than several final circuits, because every circuit downstream inherits the sub-main's drop. Check the assumptions too — is the full design current really continuous, and is the route as long as assumed?
Because above 16 mm² the conductor's inductive reactance becomes significant next to its resistance, so the drop depends on the load's power factor. Appendix 4 section 6.2(b) gives the combined value as (mV/A/m)r × cos θ + (mV/A/m)x × sin θ, where cos θ is the power factor. Below 16 mm² the reactance is small enough that the single tabulated z value is used directly.
The limits are for normal steady-state service. Regulation 525.203 accepts that a greater drop is allowed for motor starting and other high-inrush currents, provided the voltage stays within the limits specified in the relevant equipment standard. In practice you check the running condition against Table 4Ab and check starting separately against what the motor and its contactor will actually tolerate.
Sources and verification. BS 7671:2018+A4:2026 Regulations 525.1, 525.201, 525.202 and 525.203; Appendix 4 §6.2 and §6.4 and Table 4Ab for the limits; Appendix 4 Tables 4D2B, 4E2B, 4H2B and 4J2B for the mV/A/m values. The figures on this page are an indicative subset for explanation — verify every value against your own copy of BS 7671 for the actual cable, installation method and reference table, and remember the drop should include any effect due to harmonic currents.