BS 7671 gives two ways to size a circuit protective conductor: select it from Table
54.7, or calculate it with the adiabatic equation S = √(I²t) / k in
Regulation 543.1.3. The table is quick, and up to 16 mm² it asks for a CPC as
big as the line conductor. The calculation asks a narrower question: can this conductor absorb
the fault energy the protective device lets through before it gets too hot? It is the only
route by which a reduced CPC, such as the 1.5 mm² core in 2.5 mm² twin-and-earth,
complies.
The short version: Table 54.7 (Reg 543.1.4) sizes the CPC from the line
conductor. The adiabatic equation (Reg 543.1.3) sizes it from the fault: I is the earth fault
current, t the device's operating time at that current, and k comes from Tables 54.2 to 54.6
(115 for a copper CPC inside a PVC cable). A CPC that fails the table can still comply by
calculation, and the calculation should be checked at both ends of the circuit.
Two Ways to Size a CPC: Regulation 543.1.1
Regulation 543.1.1 says the cross-sectional area of every protective conductor, other than a
protective bonding conductor, shall be either calculated in accordance with
Regulation 543.1.3 or selected in accordance with Regulation 543.1.4. They are
alternatives: a CPC that satisfies either one complies.
The same regulation takes the choice away in one case. Calculation is necessary "if the
choice of cross-sectional area of line conductors has been determined by considerations of
short-circuit current and if the earth fault current is expected to be less than the
short-circuit current." Table 54.7 scales the CPC with the line conductor. Once the line
conductor has been sized for a short-circuit current, that scaling says nothing about a lower
earth fault current, which the device may take longer to clear.
Two more rules sit beside it. A protective conductor that is not part of a cable, not formed by
conduit, ducting or trunking and not inside an enclosure formed by a wiring system must be at
least 2.5 mm² copper equivalent with protection against mechanical damage, or 4 mm²
without (543.1.1). A CPC shared by several circuits is calculated for the most onerous fault
current and operating time among them, or selected from Table 54.7 against the largest line
conductor (543.1.2).
What Does Table 54.7 Require?
| Line conductor S | CPC of the same material | CPC of a different material |
| S ≤ 16 mm² | S | (k1 / k2) × S |
| 16 < S ≤ 35 mm² | 16 mm² | (k1 / k2) × 16 |
| S > 35 mm² | S / 2 | (k1 / k2) × S / 2 |
BS 7671 Table 54.7. k1 is the line conductor's k from Table 43.1; k2 is
the protective conductor's k from Tables 54.2 to 54.6. A non-standard answer is rounded up to the
next standard size (Reg 543.1.4).
For a copper circuit with a copper CPC up to 16 mm², the table asks for a CPC the same size
as the line conductor. That is why twin-and-earth with a reduced CPC needs the calculation: a
2.5 mm² cable with a 1.5 mm² CPC fails Table 54.7 outright.
What Is the Adiabatic Equation?
S = √(I²t) / k
- S is the nominal cross-sectional area of the protective conductor, in mm².
- I is the RMS fault current, in amperes, for a fault of negligible impedance, "due account being taken of the current limiting effect of the circuit impedances and the limiting capability (I²t) of that protective device".
- t is the operating time of the protective device, in seconds, at that current.
- k is a factor for the conductor material and its initial and final temperatures.
Adiabatic means the sum assumes no heat leaves the conductor during the fault: all the
energy goes into raising its temperature. That is fair for a short event and poor for a long
one, and the regulation's own note limits the equation to disconnection times not exceeding
5 s. Regulation 543.1.3 also allows the area to be obtained by reference to
BS 7454 instead, and where the formula gives a non-standard size the next larger standard size
is used.
It is easier to read squared: I²t ≤ k²S². The left side is the let-through energy
of the fault, in A²s; the right side is what the conductor can absorb. Keep that form in mind,
because it is how a device's let-through energy is compared later on.
Which k Value? Tables 54.2 to 54.6
k depends on the conductor material and on the temperatures it starts and may finish at, so
BS 7671 tabulates it by how the protective conductor is installed:
| Protective conductor | Table | 70 °C thermoplastic (PVC) | 90 °C thermosetting (XLPE) |
| Copper core in a cable, or bunched with cables | 54.3 | 115 (70 → 160 °C) | 143 (90 → 250 °C) |
| Separate insulated copper, not bunched | 54.2 | 143 (30 → 160 °C) | 176 (30 → 250 °C) |
| Aluminium core in a cable | 54.3 | 76 | 94 |
| Steel wire armour | 54.4 | 51 (60 → 200 °C) | 46 (80 → 200 °C) |
| Steel conduit, ducting, trunking | 54.5 | 47 (50 → 160 °C) | 58 (60 → 250 °C) |
BS 7671 Tables 54.2 to 54.5, with the assumed initial and final temperatures in brackets.
Values are for conductors up to 300 mm²; above 300 mm² the thermoplastic values in
Tables 54.2 and 54.3 are lower (copper 133 and 103). Bare conductors are in Table 54.6.
The row to notice is copper. The same conductor is k = 143 run as a separate single and
k = 115 as the CPC inside a PVC cable. The metal is identical; the difference is the
starting temperature. Inside the cable the CPC is assumed to start at the 70 °C the live
conductors run at, so it has 90 degrees of headroom to 160 °C rather than 130.
Where Do I and t Come From?
I is the earth fault current through the CPC. At the far end of the circuit it
is at its lowest: I = U0 / Zs, with Zs the
earth fault loop impedance to that point.
t is read off the protective device's time/current characteristic at that
current. BS 7671 Appendix 3 prints the curves, drawn for the slowest operating times the device
is allowed:
- At 5 × In a Type B circuit-breaker operates within 0.1 s: Figure 3A4 gives that one current for every time from 0.1 s to 5 s. A Type C needs 10 × In (Figure 3A5) and a Type D 20 × In (Figure 3A6, which drops to 10 × In for 5 s).
- Above that current the breaker is in its instantaneous region, and the figure's table ends at 0.1 s. Each figure says what to do next: for prospective fault currents above those giving instantaneous operation, "refer to the manufacturer's let-through energy I²t data", and "the application of the rules of Chapter 43 should take into account both minimum and maximum fault current conditions."
Where the circuit-breaker itself disconnects the earth fault, as it does on a TN supply, a first
pass can take t = 0.1 s, but only once you have checked that the minimum fault
current reaches the instantaneous threshold:
5 × In for a Type B, 10 × In for a Type C and
20 × In for a Type D. At or above it, 0.1 s is the top of the
instantaneous band, so it overstates the time rather than understating it. Below it the breaker
is on the slower part of its curve, and t is whatever the characteristic gives at your current:
Regulation 543.1.3 defines t as the operating time "corresponding to the fault current I
amperes".
For a Type B or C, meeting Table 41.3 settles it, because those Zs limits are worked
from the threshold itself: 230 × 0.95 / (5In) and
230 × 0.95 / (10In). For a Type D on a circuit allowed 5 s
(Regulation 411.3.2.3) it does not. Table 41.3(c) gives a 20 A Type D up to 1.09 Ω for
5 s, worked from 10 × In, so the fault current can be as low as
200 A, and Figure 3A6 allows the breaker 5 s at 200 A. A copper CPC in a PVC cable
then needs √(200² × 5) / 115 = 3.89 mm², where 0.1 s
would have said 0.55 mm².
A fuse has no instantaneous band, so its t is always read at your current, and the answer moves
a long way (see below). On a TT supply an RCD normally clears the fault instead of the breaker,
and t comes from the RCD; that case follows the worked example.
Worked Example: 2.5 mm² Twin-and-Earth on a 20 A Type B
A radial circuit in 2.5 mm² flat twin-and-earth (copper, 70 °C thermoplastic, with a
1.5 mm² CPC) is protected by a 20 A Type B circuit-breaker to BS EN 60898, on a TN
supply with U0 = 230 V. Take the earth fault loop impedance at the far end as
Zs = 0.935 Ω: for instance Ze = 0.35 Ω plus
0.585 Ω of R1+R2 for a 25 m run, which is what the calculator
works out for that cable.
- 1. Table 54.7. Line conductor 2.5 mm², CPC of the same material, so the table wants a 2.5 mm² CPC. The cable has 1.5 mm²: calculate.
- 2. Fault current. I = U0 / Zs = 230 / 0.935 = 246 A.
- 3. Time. 246 A is more than the 100 A (5 × 20 A, Figure 3A4) that operates the breaker within 0.1 s, so t = 0.1 s.
- 4. k. A copper CPC inside a 70 °C thermoplastic cable is Table 54.3: k = 115.
- 5. Minimum area. S = √(246² × 0.1) / 115 = 77.8 / 115 = 0.68 mm², against 1.5 mm² fitted. It complies.
Two checks sit behind those steps. Zs of 0.935 Ω is inside the 2.19 Ω that
Table 41.3 allows a 20 A Type B. And the breaker really is in its instantaneous region:
even at the minimum voltage factor Cmin = 0.95 the current is
0.95 × 230 / 0.935 = 234 A, more than twice the 100 A that
Figure 3A4 gives for operation within 0.1 s.
In energy terms the fault lets through at most 246² × 0.1 ≈ 6,050 A²s,
and the CPC can absorb k²S² = 115² × 1.5² = 29,756 A²s, almost
five times as much. The reduced CPC complies by calculation; Table 54.7 alone would have
rejected it.
Now the other end of the circuit
The far end gives the minimum fault current. The note on Figure 3A4 asks for minimum and
maximum fault current conditions when the rules of Chapter 43 are applied, and the CPC carries
the same earth fault, so it deserves the same look. Near the board, here at the origin of the
installation, the loop impedance is little more than
Ze, and with the maximum voltage
factor Cmax = 1.1 the fault current approaches
1.1 × 230 / 0.35 ≈ 723 A. Hold the time at 0.1 s and the
1.5 mm² CPC only proves out up to
I = kS / √t = 115 × 1.5 / √0.1 ≈ 545 A.
That does not make the circuit wrong. 723 A is more than seven times the 100 A that
gives instantaneous operation. Figure 3A4's table stops at 0.1 s, and its note sends you to
the manufacturer's let-through energy for currents like this. The comparison is the one Regulation
434.5.2 spells out for the live conductors, and Regulation 543.1.3's own definition of I points
the CPC the same way: the breaker's let-through I²t at that current has to be below k²S², here
29,756 A²s. Read that figure from the manufacturer's data for the breaker you are fitting.
The same circuit on a TT supply
The 0.1 s above assumes the circuit-breaker disconnects the earth fault, as it does on a TN
supply. On a TT supply that is normally the job of an RCD (Regulation 411.5.2 prefers one
there), often the residual-current part of an RCBO. Regulation 411.5.3 requires
RA × IΔn ≤ 50 V, and Table 41.5 gives the
matching Zs limit: 1,667 Ω for 30 mA. Protect the example circuit with a
20 A Type B RCBO of 30 mA instead, and take the loop impedance at the origin, earth
electrode included, as Ze = 100 Ω. Zs is then
100.59 Ω, well inside the limit, and the fault current is
230 / 100.59 ≈ 2.29 A. That is far below the 100 A that operates
the MCB element within 0.1 s, and below its 20 A rating, so the MCB element plays no
part and the 0.1 s shortcut does not apply.
On TT the disconnection relies on the residual-current characteristic, so the t of Regulation
543.1.3 is the RCD's operating time at the fault current. Regulation 411.5.3(a) requires
disconnection within 0.2 s at 230 V for the final circuits of Regulation 411.3.2.2
(Table 41.1), and within 1 s for other circuits (Regulation 411.3.2.4), so taking that time
as t errs on the safe side. With so small a current the CPC is not in doubt: even 1 s gives
√(2.29² × 1) / 115 ≈ 0.02 mm². What changes is where t comes
from: the RCD, not the breaker's instantaneous band.
What Changes With a Fuse?
A fuse has no instantaneous band. Its operating time climbs steeply as the current falls,
steeply enough that I²t goes up as the fault gets smaller. Reading Figure 3A3(a) for a
32 A BS 88-2 fuse:
| Fault current | Operates within | I²t | Minimum CPC at k = 115 |
| 310 A | 0.1 s | 9,610 A²s | 0.85 mm² |
| 220 A | 0.4 s | 19,360 A²s | 1.21 mm² |
| 125 A | 5 s | 78,125 A²s | 2.43 mm² |
For a fuse the lowest fault current, at the far end with the highest Zs, is the one
that governs. At the edge of its Zs limit, a distribution circuit allowed 5 s
needs about twice the CPC area of a final circuit held to 0.4 s on the same fuse. The
calculator does not read a fuse's time off a typical curve. It takes the energy the fuse can let
through from BS 7671's own figures for its rating: Tables 41.2 and 41.4 give the current that
clears it in 0.4 s and in 5 s, the slowest a compliant fuse may be. Between those
points it follows the line joining them; beyond the faster one it holds the energy at that
point's value, because a cartridge fuse's let-through only falls as the current rises (a
rewireable fuse, which does not limit current, is never taken below a half cycle); and it never
takes more than the disconnection time you set would give. Above 200 A the tables stop, so
a CPC that fails only on the full disconnection time is marked for review rather than failed.
Armour, Conduit and Trunking as the CPC
Regulation 543.2.1 lets the protective conductor be a metal covering, "for example, the
sheath, screen or armouring of a cable", or metal conduit, trunking or another enclosure.
The method does not change; k and the area do. Steel wire armour is k = 51 on a
70 °C thermoplastic cable and 46 on a 90 °C thermosetting one (Table 54.4), against 115
and 143 for a copper core. Through Table 54.7 that becomes explicit: a CPC of a different material
is k1/k2 times the same-material figure, which is 115/51 ≈ 2.25 for a
copper PVC cable with steel armour and 143/46 ≈ 3.1 for an XLPE one.
The armour's cross-sectional area is a property of the cable, so take it from the manufacturer's
data rather than assuming it. Where armour, conduit or trunking is the CPC, Regulation 543.2.7
still wants the earthing terminal of each accessory connected by a separate protective
conductor to an earthing terminal in its box or enclosure.
The Same Equation on the Line Conductors
Regulation 434.5.2 applies the same physics to the live conductors: a fault "shall be
interrupted within a time such that the fault current does not cause the permitted limiting
temperature of any conductor or cable to be exceeded", with t = k²S² / I²
and k from Table 43.1 (115 for copper with 70 °C thermoplastic insulation, 143 for
90 °C thermosetting). It is a separate check from the CPC's, run at the short-circuit current
rather than the earth fault current. For current-limiting devices on faults shorter than
0.1 s it is judged on let-through energy: k²S² greater than the device's I²t. Our
cable sizing guide explains why that
check is easy to drop and when it decides the cable.
Check It in the Calculator
The free BS 7671 cable sizing calculator
runs this check on every circuit as its CPC step. It works out Zs from Ze
and the run, takes I = U0 / Zs, uses
t = 0.1 s for an MCB or RCBO, the bounded energy above for a fuse, and the
disconnection time you set for an MCCB. It takes k from Tables 54.3 to 54.5 for the CPC you choose, and passes the CPC if it meets
either the adiabatic minimum or Table 54.7. On a TN supply the 0.1 s rests on its
Zs check. The calculator works its Zs limit for a breaker from the
instantaneous threshold, using 20 × In for a Type D even on a 5 s
circuit, which is stricter than Table 41.3(c). So on TN a circuit that passes its Zs
check is in the instantaneous region. If the Zs check fails, the CPC step's
0.1 s does not hold either, and t has to come from the curve. On a TT supply the calculator
takes its Zs limit from the RCD instead, 50 / IΔn as in
Table 41.5, so a pass says nothing about the breaker's instantaneous region, and the CPC step's
0.1 s for an RCBO is not assured either: take t from the RCD, as in the TT case above.
Enter the worked example (20 A Type B,
2.5 mm² copper PVC with a 1.5 mm² CPC, 25 m, Ze 0.35 Ω) and its
breakdown shows R1+R2 of 0.585 Ω, Zs of 0.94 Ω against
a 2.19 Ω maximum, and a CPC step of 246 A, 0.1 s, k 115 and 0.68 mm²
against 1.5 mm². It works at the far end only, so the board-end let-through check above is
still yours to make.
For the fault current at the board, use the
fault level calculator and the
guide to the two voltage
factors. For where the Zs limits come from, see
earthing systems explained.
Frequently Asked Questions
What k value do I use for a twin-and-earth CPC?
115. A copper protective conductor inside a 70 °C thermoplastic (PVC) cable is covered by Table 54.3, which assumes it starts at 70 °C and may reach 160 °C, giving k = 115 (103 above 300 mm²). Inside a 90 °C thermosetting (XLPE) cable it is 143. The same copper run as a separate insulated conductor, not bunched with cables, starts at 30 °C under Table 54.2 and gets k = 143 with PVC insulation.
Does 2.5 mm² twin and earth with a 1.5 mm² CPC comply?
Not by Table 54.7, which asks for a CPC the same size as the line conductor up to 16 mm². It can comply by calculation. On a TN supply, with a 20 A Type B circuit-breaker and a far-end Zs of 0.935 Ω, the fault current is 246 A, above the 100 A at which the breaker operates within 0.1 s, so S = √(246² × 0.1) / 115 = 0.68 mm², well under 1.5 mm². Check the board end as well, where the fault current is higher: there the breaker's let-through energy from the manufacturer has to be below k²S² = 29,756 A²s.
Can the steel wire armour of an SWA cable be the CPC?
Yes. Regulation 543.2.1 lists the armouring of a cable as a protective conductor. Its k comes from Table 54.4: 51 for steel armour on a 70 °C thermoplastic cable and 46 on a 90 °C thermosetting one, against 115 and 143 for a copper core. Through Table 54.7 a CPC of a different material is k1/k2 times the same-material figure, so steel armour on a copper PVC cable up to 16 mm² needs 115/51, about 2.25 times the line conductor's area. Take the armour's actual cross-sectional area from the cable manufacturer's data.
Sources and verification. BS 7671:2018+A4:2026 Regulations 543.1.1 to
543.1.4, 543.2.1, 543.2.7, 434.5.2, 411.3.2.2 to 411.3.2.4, 411.5.2 and 411.5.3; Tables 54.2 to
54.7, 43.1, 41.1, 41.3 and 41.5; Appendix 3 Figures 3A3(a), 3A4, 3A5 and 3A6. The figures on
this page are an indicative subset for explanation.
Verify every value against your own copy of BS 7671, and take let-through energy and armour
cross-sectional areas from the manufacturer's data for the products actually installed.