A measured earth fault loop impedance is not compared with the maximum Zs in BS 7671's tables. Those limits hold with the conductors at their maximum operating temperature, and a test is done cold, so Appendix 3 checks the reading against 80% of the limit instead: Zs(m) ≤ 0.8 × U0 × Cmin / Ia. For a 32 A Type B circuit-breaker that is 1.09 Ω, not the 1.37 Ω printed in Table 41.3. Below: where the tables' temperatures come from, what the 0.8 does, the other correction BS 7671 allows, and why a circuit designed close to its limit can fail the 80% check on test and still comply.

The short version: Tables 41.2 to 41.4 give the maximum Zs for the circuit in service, with its line conductors at their maximum operating temperature (70 °C for PVC). Tested cold, copper has about a sixth less resistance, so Appendix 3 holds the reading to 0.8 of the tabulated value. That is one method in an informative appendix, not the requirement itself: correcting only the circuit's own conductors to their operating temperature is also allowed, and it passes some circuits that the 0.8 check fails.

What the Maximum Zs Tables Assume

Regulation 411.4.4 is the condition for automatic disconnection in a TN system: Zs × Ia ≤ U0 × Cmin, where Ia is the current that operates the protective device within the required time and Cmin is 0.95 for a supply given under the Electricity Safety, Quality and Continuity Regulations. For circuit-breakers to BS EN 60898 that is the last column of Table 41.3: 230 × 0.95 / (5In) for a Type B, / (10In) for a Type C and / (20In) for a Type D, and Regulation 411.4.202 lets the table's values be used instead of the calculation. Fuses have no formula: their limits are read from Table 41.2 for 0.4 s and Table 41.4 for 5 s.

Each of those tables has a NOTE 2, and it is the one that matters here. In Table 41.3 it says the values "should not be exceeded when: (a) the line conductors are at the appropriate maximum permitted operating temperature, as given in Table 52.2; and (b) the circuit protective conductors are at the appropriate assumed initial temperature, as given in Tables 54.2 to 54.5", and it ends: "If the conductors are at a different temperature when tested, the reading should be adjusted accordingly. See Appendix 3." The tabulated figure is a hot one:

  • Line conductors: 70 °C with thermoplastic (PVC) insulation and 90 °C with thermosetting (XLPE), from Table 52.2. NOTE 3 takes an XLPE cable back to 70 °C where it has been sized in accordance with Regulation 512.1.5 (see PVC vs XLPE).
  • A CPC inside the cable, or bunched with cables: 70 °C in a thermoplastic cable and 90 °C in a thermosetting one (Table 54.3).
  • A separate insulated CPC, not bunched with cables: 30 °C (Table 54.2).
  • Steel wire armour as the CPC: 60 °C on a 70 °C thermoplastic cable and 80 °C on a 90 °C thermosetting one (Table 54.4).

That is the circuit carrying its load. On the day it is tested, it is at the temperature of the building.

Why a Cold Reading Reads Low

Conductor resistance rises with temperature. BS 7671 itself uses an approximate resistance-temperature coefficient of 0.004 per °C at 20 °C for both copper and aluminium (Appendix 4, section 6.1, NOTE). On that basis a conductor at 70 °C has 1 + 0.004 × (70 − 20) = 1.20 times its resistance at 20 °C, and at 90 °C, 1.28 times. To bring a measured R1+R2 up to the temperature the tables assume, multiply it by:

Conductors when testedTo 70 °C (thermoplastic)To 90 °C (thermosetting)
5 °C1.2771.362
10 °C1.2501.333
15 °C1.2241.306
20 °C1.2001.280
25 °C1.1761.255

(1 + 0.004 × (tp − 20)) / (1 + 0.004 × (t − 20)), with t the conductor temperature when tested and tp the operating temperature, using the coefficient BS 7671 gives for copper and aluminium. For steel armour or conduit as the CPC, take its resistance from the manufacturer's data.

The correction belongs to the circuit's own conductors, the ones NOTE 2 names. The part of the loop outside the installation, Ze, is the supply's, and is not one of them.

What Is the 80% Rule for Zs?

It is the procedure under "Maximum earth fault loop impedance" in Appendix 3 of BS 7671, which begins: "When impedance measurements are made at ambient temperature the procedure hereinafter described may be followed to take into account the increase of resistance of the conductors with the increase of temperature due to load current." It covers TN and TT systems where an overcurrent protective device provides the automatic disconnection, and the requirements of Regulation 411.4.4, or 411.5.4 on TT, are considered to be met when:

Zs(m) ≤ 0.8 × U0 × Cmin / Ia

  • Zs(m) is the measured impedance of the earth fault current loop up to the most distant point of the relevant circuit from the origin of the installation.
  • 0.8 is "a factor to take into account the increase of resistance of the conductors with the increase of temperature due to load current".
  • U0 is the nominal line to earth voltage, 230 V.
  • Ia is the current that operates the device within the time in Table 41.1 (0.4 s at 230 V on TN, for the final circuits of Regulation 411.3.2.2) or within 5 s under Regulation 411.3.2.3.
  • Cmin is 0.95.

So a reading is held to 80% of the calculated maximum: for a 32 A Type B, 0.8 × 230 × 0.95 / 160 = 1.09 Ω, against the 1.37 Ω of Table 41.3. Appendix 3 is informative, and its note after the equation is short: "The above is one method of correcting for the effects of temperature difference; other methods are not precluded."

BS 7671 does not say how 0.8 was chosen. With its own coefficient, a conductor at 10 °C has 0.96 / 1.20 = 0.80 of its resistance at 70 °C, and one at 20 °C has 0.83. So for a loop made entirely of the circuit's own conductors the factor is about right: exact at 10 °C, a little cautious at 20 °C. It treats Ze the same way, which is cautious too: the more of the reading is supply, the more margin it keeps.

Maximum Measured Zs: 80% of the BS 7671 Values

For final circuits on a 0.4 s disconnection time at U0 230 V, these are the readings the Appendix 3 equation allows, worked from the same data as Tables 41.2 and 41.3:

Rating Type B (Ω) Type C (Ω) Type D (Ω) BS 88-2 fuse (Ω)
6 A5.822.911.456.24
10 A3.491.740.873.72
16 A2.181.090.541.94
20 A1.740.870.431.34
25 A1.390.690.341.03
32 A1.090.540.270.79
40 A0.870.430.210.60
50 A0.690.340.170.45
63 A0.550.270.130.35

Circuit-breakers to BS EN 60898, and the overcurrent characteristics of RCBOs to BS EN 61009-1: 0.8 × 230 × 0.95 / Ia, with Ia 5, 10 and 20 × In. BS 88-2 fuses: 0.8 × Table 41.2(a). Every figure is rounded down to 0.01 Ω, so a reading equal to it is inside the limit; 80% of the printed table, which is itself rounded, can come out 0.01 Ω higher (1.75 Ω for a 20 A Type B, where the equation gives 1.748 Ω). Indicative subset: verify against your own copy of BS 7671 and the manufacturer's data.

A Type B or C needs no second column for 5 s, because Table 41.3 gives each one value for both disconnection times. A Type D on a 5 s distribution circuit (Regulation 411.3.2.3) takes Table 41.3(c)'s 5 s row, worked from 10 × In, so its 80% figures are the Type C column. A fuse on 5 s takes Table 41.4: a 100 A BS 88-2 fuse there is 0.42 Ω, so a reading up to 0.33 Ω.

Worked Example: A 32 A Radial That Fails the 80% Rule and Still Complies

A radial circuit in 4 mm² flat twin-and-earth (copper, 70 °C thermoplastic, with the 1.5 mm² CPC it carries), 25 m long and clipped direct, is protected by a 32 A Type B circuit-breaker to BS EN 60898 on a TN-S supply. Ze is 0.80 Ω, and the reading at the origin on the day matches it. The conductor resistances are the ones the calculator holds: at 20 °C, 4.61 mΩ/m for 4 mm² and 12.10 mΩ/m for 1.5 mm², so 16.71 mΩ/m for the pair.

  • 1. The design, hot. R1+R2 at 20 °C is 16.71 × 25 / 1000 = 0.41775 Ω, and at 70 °C 0.41775 × 1.20 = 0.5013 Ω. Zs = 0.80 + 0.5013 = 1.30 Ω, inside the 1.37 Ω of Table 41.3. The circuit complies.
  • 2. The test, at about 20 °C. The continuity test gives R1+R2 of about 0.42 Ω, and the loop test reads Zs(m) of about 1.22 Ω.
  • 3. Against the table. 1.22 Ω is under 1.37 Ω, which proves nothing on its own: that limit is for the circuit hot.
  • 4. The 80% rule. 0.8 × 230 × 0.95 / 160 = 1.09 Ω. The reading of 1.22 Ω is above it, so it does not meet the Appendix 3 criterion.
  • 5. Correct the conductors only. 0.80 + 0.42 × 1.20 = 1.30 Ω, the design figure again, and inside 1.37 Ω. By this method the circuit meets Regulation 411.4.4.

Both methods start from the same readings. They disagree because the 0.8 applies to the whole loop, and two-thirds of this loop is Ze. Holding the reading to 0.8 of the limit is the same as letting all of it rise by 25% (1 / 0.8 = 1.25) before the comparison, the supply's share included. Correcting R1+R2 raises only the circuit's own conductors, the ones NOTE 2 names, by the 20% in the table above.

How much that costs depends on Ze. On this supply the 80% rule holds the cold R1+R2 to 1.09 − 0.80 = 0.29 Ω, about 17.5 m of this cable, while correcting the conductors allows 0.47 Ω, about 28 m. With Ze at 0.35 Ω, the figure usually declared for a TN-C-S supply, the two limits are 0.74 Ω and 0.85 Ω: the gap shrinks from 38% to 12%, because less of the reading is supply.

The 0.8 check keeps margin on the whole reading, Ze included. Correcting only R1+R2 takes Ze as measured, so it relies on that figure being the one the circuit will see, and on knowing the conductor temperature when the readings were taken. Record both with the result.

What Should the Designer Put on the Schedule?

The model forms in Appendix 6 keep the two figures apart. The Generic Schedule of Circuit Details has column 12, "Maximum permitted Zs (Ω)"; the Generic Schedule of Test Results records R1+R2 in column 21 and the "Maximum measured" Zs in column 27. A footnote to column 12 says that where its value is "taken from a source other than the tabulated values given in Chapter 41", the source is stated for the circuit in the Remarks, column 31, of the Schedule of Test Results.

So column 12 is normally a hot figure and column 27 a cold one, and the two are not compared directly. For a circuit designed close to its limit, say which way the reading will be judged: 80% of column 12, or the conductor-only correction from the measured R1+R2 and the test temperature. Where there is room, leaving enough margin for the cold reading to clear 80% of column 12 avoids the question altogether.

Check It in the Calculator

The free BS 7671 cable sizing calculator works on the hot side, where the tables are. It takes R1+R2 from the conductor resistances it holds at 20 °C, multiplies by 1.20 for a thermoplastic cable or 1.28 for a thermosetting one, adds Ze, and compares the total with Table 41.3's limit for a circuit-breaker (Table 41.2 or 41.4 for a fuse) directly, with no 0.8 on top. Applying 0.8 to that figure as well would count the temperature twice.

Enter the worked example (32 A Type B, 4 mm² flat twin-and-earth, 25 m, TN-S, Ze 0.80 Ω) and its breakdown shows R1+R2 of 0.501 Ω and Zs of 1.30 Ω against a 1.37 Ω maximum. To predict the reading on test, take R1+R2 back to 20 °C (0.501 / 1.20 ≈ 0.42 Ω) and add Ze: about 1.22 Ω, above the 1.09 Ω the 80% check allows. That is the circuit to flag before the tester finds it.

Where Regulation 512.1.5 has sized an XLPE cable on the 70 °C column, the calculator follows the tables' NOTE 3 and takes its line conductors at 70 °C, a multiplier of 1.20 rather than 1.28. Its conductor resistances are indicative, from the IET On-Site Guide: check them against the cable maker's data.

For the hot limits themselves and the Ze figures for each earthing system, see earthing systems explained; for the fault current the same Zs gives the CPC, see the adiabatic equation and CPC sizing.

Frequently Asked Questions

Is the 80% rule mandatory?

No. What BS 7671 requires is that the earth fault loop impedance meets Chapter 41: Zs × Ia ≤ U0 × Cmin (Regulation 411.4.4), with Regulation 643.7.3.1 applying that to the value measured on test. The 0.8 factor is printed in Appendix 3, an informative appendix, as a way of showing that a reading taken at ambient temperature will still comply once the conductors are warm. A reading above 80% of the table has not automatically failed Chapter 41: it needs another way of showing compliance, such as correcting R1+R2 to the operating temperature.

Does the 80% rule apply to an RCD on a TT system?

Not to the limit the RCD sets. Appendix 3 describes the correction for TN and TT systems where an overcurrent protective device provides the automatic disconnection, so it does apply on TT when a circuit-breaker or fuse disconnects the fault (Regulation 411.5.4). Where an RCD gives the fault protection, Regulation 411.5.3 asks for RA × IΔn ≤ 50 V, and Table 41.5 turns that into a maximum Zs of 1,667 Ω at 30 mA and 167 Ω at 300 mA, far beyond anything conductor temperature could move. What matters there is the note to Table 41.5: an earth electrode resistance above 200 Ω may not be stable.

Can I use the manufacturer's maximum Zs instead of the BS 7671 tables?

Yes, and Appendix 3 prefers it. It says the maximum loop impedance for a given disconnection time varies between manufacturers as well as between device types, that designers should use the manufacturer's specific data wherever possible, and that the tabulated values are for when that data is absent. It calls Table 41.3's values for BS EN 60898 circuit-breakers "far more onerous". Record where the figure came from in the Remarks of the Schedule of Test Results, as the model forms ask. Whichever source sets the limit, it is a limit for the circuit in service, so a cold reading still needs the temperature correction.

Sources and verification. BS 7671:2018+A4:2026 Regulations 411.3.2.2, 411.3.2.3, 411.4.4, 411.4.202, 411.5.3, 411.5.4 and 643.7.3.1; Tables 41.1 to 41.5, 52.2 and 54.2 to 54.5; Appendix 3; Appendix 4, section 6.1; the Appendix 6 Generic Schedules of Circuit Details and Test Results. The figures on this page are an indicative subset for explanation. Verify every value against your own copy of BS 7671, and take device limits, conductor and armour resistances from the manufacturer's data for the products actually installed.

Need a Calculation Pack?

Cable sizing, protection coordination and fault level studies — delivered submission-ready.

ProDesign Cable Calculations