"What size earth do I need?" has four different answers in BS 7671, because the regulations size four different conductors by four different routes. This guide gives the tables for all of them, shows the arithmetic behind the adiabatic equation, and then answers the question most sizing guides skip: how much margin the reduced earth in standard twin-and-earth actually has, and the one common UK arrangement where it nearly runs out.
Four conductors, four sizing rules
Before reaching for a table, be clear which conductor is in front of you. The terms are used loosely on site and the rules genuinely differ.
| Conductor | Where it runs | Sized by |
|---|---|---|
| Circuit protective conductor (cpc) | Within each final or distribution circuit | Adiabatic equation (543.1.3) or Table 54.7 |
| Earthing conductor | Main earthing terminal to the supply earth or electrode | 543.1, plus Table 54.1 if buried |
| Main protective bonding | MET to incoming gas, water and other services | Table 54.8 on PME, otherwise 544.1.1 |
| Supplementary bonding | Between exposed and extraneous parts in a location | 544.2, referenced to the cpc size |
Minimum earthing conductor size (Table 54.1)
Where the earthing conductor is buried in the ground, typically the run to an earth electrode on a TT installation, Table 54.1 imposes a floor that has nothing to do with fault current. It is about corrosion and mechanical damage.
| Protection | Copper | Steel |
|---|---|---|
| Protected against corrosion and mechanical damage | Sized by 543.1 — the adiabatic result, minimum 2.5 mm² Cu | |
| Protected against corrosion, not against mechanical damage | 16 mm² | 16 mm² |
| Protected against neither | 25 mm² | 50 mm² |
Above ground, an earthing conductor that is mechanically protected has a floor of 2.5 mm² copper and one that is not has a floor of 4 mm² copper, per Regulation 543.1.1. In practice a domestic TN-C-S installation with 25 mm² tails is given a 16 mm² earthing conductor, which is the Table 54.7 answer for a 25 mm² line conductor rather than anything to do with burial.
The adiabatic equation
The protective conductor has to survive the fault current for as long as the fault lasts. BS 7671 Regulation 543.1.3 expresses that as:
S = √(I² × t) ÷ k
- S — minimum cross-sectional area in mm²
- I — earth fault current in amperes, normally U₀ ÷ Zₛ
- t — the disconnection time of the protective device at that current, in seconds
- k — a material and temperature factor from Tables 54.2 to 54.6
The equation is adiabatic in the strict sense: it assumes no heat escapes the conductor during the fault. That is why it is only valid for short durations, and why BS 7671 caps its use at 5 s.
Choosing k
| Conductor | Table | k |
|---|---|---|
| Copper, 70 °C thermoplastic, core in a cable or bunched | 54.2 | 115 |
| Copper, 90 °C thermosetting, core in a cable or bunched | 54.2 | 100 |
| Aluminium, 70 °C thermoplastic, core in a cable | 54.2 | 76 |
| Copper, 70 °C thermoplastic, separate conductor | 54.3 | 143 |
| Copper, 90 °C thermosetting, separate conductor | 54.3 | 176 |
| Bare copper, no risk of damage to adjacent material | 54.3 | 228 |
The difference between 115 and 143 for the same copper conductor is entirely the assumed starting temperature. A cpc bundled inside a loaded cable is already sitting near 70 °C when the fault arrives, so it has less headroom to the 160 °C limit than a separate conductor starting at 30 °C. Using 143 for a cpc inside twin-and-earth is a common and non-conservative error.
What the equation asks for
Minimum S in mm² for a copper cpc inside a cable, k = 115. Every figure below is computed from the equation above, not transcribed.
| Fault current | t = 0.1 s | t = 0.4 s | t = 1 s | t = 5 s |
|---|---|---|---|---|
| 100 A | 0.27 | 0.55 | 0.87 | 1.94 |
| 200 A | 0.55 | 1.10 | 1.74 | 3.89 |
| 300 A | 0.82 | 1.65 | 2.61 | 5.83 |
| 500 A | 1.37 | 2.75 | 4.35 | 9.72 |
| 800 A | 2.20 | 4.40 | 6.96 | 15.56 |
| 1,200 A | 3.30 | 6.60 | 10.43 | 23.33 |
| 1,600 A | 4.40 | 8.80 | 13.91 | 31.11 |
| 2,000 A | 5.50 | 11.00 | 17.39 | 38.89 |
| 3,000 A | 8.25 | 16.50 | 26.09 | 58.33 |
Read down a column rather than across a row and the shape of the problem becomes obvious: time matters far more than current. Quadrupling the fault current doubles the required area. Quadrupling the disconnection time does exactly the same. A fault that clears in 0.1 s at 3,000 A needs less copper than one that lingers for 5 s at 800 A.
Table 54.7: the simplified method
Where you do not want to calculate, Table 54.7 gives a protective conductor size directly from the line conductor, provided both are the same material.
| Line conductor S | Minimum protective conductor |
|---|---|
| S ≤ 16 mm² | S (the same size) |
| 16 < S ≤ 35 mm² | 16 mm² |
| S > 35 mm² | S ÷ 2 |
Table 54.7 is a permitted simplification, not a mandatory floor. Regulation 543.1.1 offers the calculation and the table as alternatives. That distinction matters immediately, because the most common cable in the country does not satisfy the table.
The reduced cpc in twin-and-earth
Standard 6242Y twin-and-earth carries a bare cpc smaller than its line conductors. On Table 54.7 alone, every size from 1.5 mm² upwards fails. The cable is entirely compliant, because the adiabatic route is used instead.
The table below gives the cpc actually present in each size, what Table 54.7 would have asked for, and the largest fault current that cpc can carry for a given duration, from I = S × k ÷ √t with k = 115.
| T&E size | cpc | Table 54.7 would want | Max I at 0.1 s | Max I at 0.4 s | Max I at 5 s |
|---|---|---|---|---|---|
| 1 mm² | 1 mm² | 1 mm² | 364 A | 182 A | 51 A |
| 1.5 mm² | 1 mm² | 1.5 mm² | 364 A | 182 A | 51 A |
| 2.5 mm² | 1.5 mm² | 2.5 mm² | 545 A | 273 A | 77 A |
| 4 mm² | 1.5 mm² | 4 mm² | 545 A | 273 A | 77 A |
| 6 mm² | 2.5 mm² | 6 mm² | 909 A | 455 A | 129 A |
| 10 mm² | 4 mm² | 10 mm² | 1,455 A | 727 A | 206 A |
| 16 mm² | 6 mm² | 16 mm² | 2,182 A | 1,091 A | 309 A |
A 2.5 mm² twin-and-earth will carry 545 A for 0.1 swithout the earth conductor exceeding its temperature limit. Domestic prospective earth fault currents are rarely a third of that.
Why the cpc almost never binds on an MCB circuit
Here is the part that a table alone will not tell you. The maximum earth fault loop impedance permitted for 0.4 s disconnection on a type B MCB is set by the instantaneous trip threshold of 5 × Iₙ, with the 0.95 voltage factor applied: Zₛ = (230 × 0.95) ÷ (5 × Iₙ). At that impedance the fault current is exactly 5 × Iₙ, the MCB is in its magnetic region, and the fault clears in well under 0.1 s.
So the worst permitted loop impedance produces the smallest fault current the device will still clear quickly. Both sides of the adiabatic equation shrink together, and the cpc ends up with a great deal of headroom.
| Type B MCB | Max Zₛ (0.4 s) | I at that Zₛ | S needed at 0.1 s | Typical T&E | cpc present | Headroom |
|---|---|---|---|---|---|---|
| 6 A | 7.28 Ω | 30 A | 0.08 mm² | 1 mm² | 1 mm² | 12.1x |
| 10 A | 4.37 Ω | 50 A | 0.14 mm² | 1 mm² | 1 mm² | 7.3x |
| 16 A | 2.73 Ω | 80 A | 0.22 mm² | 1.5 mm² | 1 mm² | 4.5x |
| 20 A | 2.19 Ω | 100 A | 0.27 mm² | 2.5 mm² | 1.5 mm² | 5.5x |
| 32 A | 1.37 Ω | 160 A | 0.44 mm² | 6 mm² | 2.5 mm² | 5.7x |
| 40 A | 1.09 Ω | 200 A | 0.55 mm² | 10 mm² | 4 mm² | 7.3x |
| 50 A | 0.87 Ω | 250 A | 0.69 mm² | 16 mm² | 6 mm² | 8.7x |
Every domestic MCB circuit in that table has between four and twelve times the copper the adiabatic equation requires. The practical conclusion is that on an MCB-protected circuit the maximum Zₛ check bites long before the adiabatic check does. If your loop impedance passes, your reduced cpc has already passed with room to spare, and there is no need to calculate it again. Type C devices trip at 10 × Iₙ rather than 5, which halves the permitted Zₛ and doubles the fault current, but doubling the current only doubles the required area and the headroom remains comfortable.
The case where it genuinely is close
The exception is a 5 s disconnection time protected by a fuse, which is a combination that keeps the fault current moderate and the duration long — the worst quadrant of the table above.
Take a 32 A BS 88-3 fuse on a distribution circuit at its 5 s maximum Zₛ of 1.75 Ω. The fault current is 218.5 ÷ 1.75 = 125 A, and it flows for the full 5 s:
S = √(125² × 5) ÷ 115 = 2.43 mm²
The cpc in 6 mm² twin-and-earth is 2.5 mm². That is a margin ofunder three per cent. Change the insulation to 90 °C thermosetting and k drops from 115 to 100, the requirement rises to 2.79 mm², and the cable no longer complies. This is the one arrangement in common UK use where the reduced cpc is a real constraint rather than a theoretical one, and it is worth calculating rather than assuming.
Main protective bonding (Table 54.8)
Bonding conductors are not sized by fault current at all. On a PME (TN-C-S) supply, Table 54.8 sizes them against the supply PEN conductor.
| Supply PEN conductor | Main bonding conductor (copper) |
|---|---|
| ≤ 35 mm² | 10 mm² |
| > 35 to 50 mm² | 16 mm² |
| > 50 to 95 mm² | 25 mm² |
| > 95 to 150 mm² | 35 mm² |
| > 150 mm² | 50 mm² |
Almost every UK domestic property sits in the first row, which is why 10 mm² green-and-yellow is the default answer for gas and water bonding. On a supply that is not PME, Regulation 544.1.1 applies instead: not less than half the cross-sectional area of the earthing conductor, subject to a minimum of 6 mm² and with no requirement to exceed 25 mm² copper.
Steel wire armour as the cpc
The armour of an SWA cable is frequently used as the circuit protective conductor, and it usually has ample cross-sectional area — but the calculation needs two numbers this page deliberately will not guess. The armour CSA is specific to the cable construction and must come from the manufacturer's data sheet, not from a generic table, because it varies between two-core, three-core and four-core cables of the same conductor size. The k value comes from BS 7671 Table 54.4 and is much lower than copper. Check both against the actual cable before relying on the armour, and note that the armour also has to satisfy the loop impedance requirement, where its higher resistance is the more usual limitation.
NEC equipment grounding conductors
Under the NEC the equivalent conductor is the equipment grounding conductor, and the approach is different: it is sized from the rating of the overcurrent device rather than from a fault calculation. NEC Table 250.122:
| OCPD rating | Copper EGC | Approx. metric |
|---|---|---|
| 15 A | 14 AWG | 2.08 mm² |
| 20 A | 12 AWG | 3.31 mm² |
| 30 to 60 A | 10 AWG | 5.26 mm² |
| 100 A | 8 AWG | 8.37 mm² |
| 200 A | 6 AWG | 13.3 mm² |
| 300 A | 4 AWG | 21.1 mm² |
NEC 250.122(B) also requires the EGC to be increased proportionally wherever the ungrounded conductors have been upsized above the minimum, which most often happens when a long run has been upsized for voltage drop. Uprate the phase conductors and the ground has to follow.
Sizing it on site
The Voltix earthing and grounding conductor calculator applies the adiabatic equation and the Table 54.7 lookup for any circuit. For the fault current, use the measured Zₛ from your loop test — the earth fault loop impedance guide covers how to obtain and correct that reading, and the BS 7671 cable sizing guide covers the line conductors the cpc has to accompany.
Standards and assumptions.Written against BS 7671:2018+A2:2022, NEC 2023 and the values in the tables named above. Every computed figure on this page — the adiabatic areas, the withstand currents and the maximum Zₛ values — is generated from the equations shown, using U₀ = 230 V with the 0.95 voltage factor of Regulation 411.4.200 where maximum Zₛ is calculated. Twin-and-earth cpc sizes are the standard 6242Y constructions. This is reference material for qualified people and not a substitute for the published standard or for design verification on the installation in front of you.
Last reviewed September 2026.
Frequently Asked Questions
- What is the minimum earthing conductor size in the UK?
- It depends which conductor you mean. A buried earthing conductor to an earth electrode has a floor set by BS 7671 Table 54.1: 16 mm2 copper if it is protected against corrosion but not against mechanical damage, and 25 mm2 copper if it is protected against neither. An earthing conductor that is mechanically protected and protected against corrosion is sized by the adiabatic equation instead, with an absolute minimum of 2.5 mm2 copper where it is mechanically protected and 4 mm2 where it is not. A main protective bonding conductor on a PME supply is never less than 10 mm2 copper.
- Is the reduced earth in twin-and-earth cable big enough?
- On a domestic circuit protected by a type B MCB, almost always yes, and with a large margin. At the maximum permitted Zs for 0.4 s disconnection the MCB is operating in its instantaneous region, so the fault lasts about 0.1 s and the adiabatic equation asks for well under 1 mm2. A 2.5 mm2 twin-and-earth carries a 1.5 mm2 cpc, roughly five times the requirement. The margin narrows sharply on circuits with a 5 s disconnection time protected by a fuse.
- When does the earth conductor actually fail the check?
- When the disconnection time is long rather than when the fault current is high. A 32 A BS 88-3 fuse at its 5 s maximum Zs of 1.75 ohms passes about 125 A for 5 s, which needs 2.43 mm2 of copper. The cpc in 6 mm2 twin-and-earth is 2.5 mm2. That is a margin of under three per cent, and it is the one common UK arrangement where the reduced cpc is genuinely close to the limit.
- Which k value should I use in the adiabatic equation?
- k comes from the conductor material, its insulation and whether the protective conductor is a core inside a cable or a separate conductor. BS 7671 Table 54.2 covers a protective conductor incorporated in a cable or bunched with cables: copper with 70 degree C thermoplastic insulation gives k = 115. Table 54.3 covers a separate protective conductor not bunched with other cables: the same copper conductor gives k = 143 because it starts at 30 degree C rather than 70 degree C. Bare copper where there is no risk of damaging adjacent material gives k = 228. Aluminium inside a cable gives k = 76.
- Does Table 54.7 apply to twin-and-earth cable?
- No, and this is the point people most often miss. Table 54.7 would require a 2.5 mm2 line conductor to have a 2.5 mm2 protective conductor, and standard twin-and-earth carries 1.5 mm2. Table 54.7 is an optional simplification, not a mandatory minimum. Where the cpc is smaller than the table value, BS 7671 Regulation 543.1.1 permits the adiabatic calculation of 543.1.3 to be used instead, and for twin-and-earth on an MCB it passes comfortably.