Tradefox

Tradefox - Renforcer les compétences

Apprendre la construction

BS 7671 Guide · Section 3

Protection against electric shock

Every circuit you install needs two layers of shock protection: one that stops people touching live parts, and one that limits harm when something fails. This section explains how Chapter 41 delivers both, and what that means for the readings you take and the RCDs you fit.

Two layers: basic and fault protection

Think of shock protection as two separate jobs. Basic protection stops anyone touching a live part in normal use. Insulation, barriers and enclosures do this job.

Fault protection covers the case where basic protection has already failed. A loose line conductor touches a metal case, and the case becomes live. Something then has to limit the touch voltage or how long it lasts.

Chapter 41 sets out protective measures that each provide both layers: automatic disconnection of supply (Section 411), double or reinforced insulation (412), electrical separation (413) and SELV or PELV (414). Additional protection (Section 415) sits on top as a backup: a 30 mA RCD or supplementary bonding.

Regulation references: Chapter 41, Sections 410 to 415

Automatic disconnection of supply

Every exposed metal part is connected to earth by a protective conductor. If line faults to that metal, a large current flows round the earth fault loop and the protective device disconnects before the touch voltage can cause harm.

That needs sound earthing and bonding, a low enough loop impedance, and a device fast enough at the resulting current. Earthing arrangements and bonding are covered in Earthing and bonding.

Maximum disconnection times

Table 41.1 applies to final circuits up to 63 A that feed one or more socket-outlets, and to final circuits up to 32 A that feed only fixed equipment. Everything else, including sub-mains, gets a longer allowance. For a 230 V supply to earth:

CircuitTN systemTT system
Final circuit covered by Table 41.10.4 s0.2 s
Distribution circuit, or a larger final circuit5 s1 s

TT gets the tighter times because the fault current returns through the ground, so the touch voltage on exposed metal is usually higher. In a TT installation where an overcurrent device gives disconnection and all extraneous-conductive-parts are bonded, the TN times may be used instead. Check the conditions in the current edition before relying on this.

Regulation references: 411.3.2, Table 41.1

Maximum Zs and the 80% check

Tables 41.2 to 41.4 give the maximum earth fault loop impedance (Zs) for common fuses and circuit-breakers. Each value is worked out from three things: the nominal voltage to earth (230 V), a voltage factor Cmin of 0.95 to allow for the supply running low, and the current that makes the device trip within the required time.

Those tabulated values assume the conductors are at their maximum operating temperature, which is 70 °C for standard thermoplastic cable. When you test, the cable is usually near room temperature, so its resistance is lower. A reading that only just scrapes under the table value could fail once the cable warms up under load.

The usual rule of thumb, found in Appendix 3, is to compare your measured Zs against 80% of the tabulated maximum. If you are under that figure, the circuit will still be compliant when hot.

Worked example 1: B32 ring final

  1. A Type B breaker trips magnetically at 5 times its rating, so Ia = 5 × 32 = 160 A.
  2. Maximum Zs = (230 × 0.95) ÷ 160 = 1.37 Ω. This matches the table value.
  3. Rule-of-thumb limit = 1.37 × 0.8 = about 1.10 Ω.
  4. You measure 0.86 Ω at the furthest socket. That is comfortably under 1.10 Ω, so it passes.

Worked example 2: C20 radial to a workshop machine

  1. A Type C breaker trips magnetically at up to 10 times its rating, so Ia = 10 × 20 = 200 A.
  2. Maximum Zs = (230 × 0.95) ÷ 200 = about 1.09 Ω.
  3. Rule-of-thumb limit = 1.09 × 0.8 = about 0.87 Ω.
  4. You measure 0.95 Ω. That is under the table value but over the 80% figure, so treat it as a fail. Options include a larger CPC, a shorter route, a Type B device if the load allows, or RCD fault protection.
Watch out: swapping a Type B for a Type C of the same rating halves the maximum Zs. Never upgrade a breaker curve to cure nuisance tripping without re-checking Zs.
On site: these examples show the method. For real jobs, use the values in Tables 41.2 to 41.4 and the Appendix 3 guidance in your copy of the current edition, and follow the device manufacturer’s data where it differs.

TT installations

On TT, electrode resistance is usually too high for an overcurrent device, so an RCD provides fault protection. Electrode resistance multiplied by the RCD’s rated residual current must not exceed 50 V: about 1,667 Ω for 30 mA in theory. Above 200 Ω, readings may not stay stable through the seasons, so aim well below.

Regulation references: 411.4, 411.5.3, Tables 41.2 to 41.5, Appendix 3

Additional protection: where 30 mA RCDs are required

A 30 mA RCD won’t prevent a shock, but it cuts the current fast enough to make direct contact far less likely to kill: a cut flex, or a drill through a cable. It is a backup, never the only protective measure (Regulation 415.1.2).

  • Socket-outlets up to 32 A (Regulation 411.3.3): need a 30 mA RCD where ordinary persons, children or vulnerable people are likely to use them, and in other locations too. The same applies to mobile equipment up to 32 A used outdoors.
  • Domestic lighting (Regulation 411.3.4): AC final circuits supplying luminaires in dwellings need a 30 mA RCD.
  • Cables concealed in walls and partitions (Regulation 522.6): cables less than 50 mm deep must run in a safe zone with a 30 mA RCD to Regulation 415.1.1, or have mechanical protection under Regulation 522.6.204. In walls with metal parts, the RCD or 522.6.204 protection is needed whatever the depth. A4 brings these options together in the new Table 52.1. For more detail, see Cable selection and sizing.
  • Special locations add their own RCD rules. See Special locations.

Omitting RCDs on non-domestic sockets

Outside dwellings, there is one narrow exception. A documented risk assessment can decide that an RCD is not needed on certain sockets. It only covers sockets that ordinary persons are not likely to use. Typical examples are a dedicated server-rack outlet or a socket in a locked plant room. The client or duty-holder owns that decision. A person who is skilled in electrical work should be involved, and the assessment should go with the certificate. Cost is never a valid reason.

Watch out: the exception never applies in a dwelling, or to mobile equipment used outdoors. If you can’t produce the risk assessment, the RCD goes in.

Regulation references: 411.3.3, 411.3.4, 415.1, 522.6, 522.6.204, Table 52.1

Choosing the right RCD type

Modern loads rarely leak a neat sine wave. Electronics, inverters and chargers can produce DC or high-frequency residual currents. These can saturate the core of a basic RCD so that it fails to trip. Regulation 531.3.3 requires you to match the RCD type to what the load can produce.

TypeDetectsTypical use
ACSinusoidal AC onlyOnly fixed equipment known to produce no DC component
AAC plus pulsating DCThe sensible default for most final circuits
FAs Type A, plus mixed-frequency residual currentsSingle-phase variable-speed drives, such as some washing machines, pumps and heat-pump units
BAs Type F, plus smooth DCThree-phase inverters and drives, and some EV and PV set-ups where the maker requires it

Always check the equipment instructions for the RCD type the manufacturer requires. See EV, solar, batteries and heat pumps.

RCDs in series and blinding

A4 makes a point many installers had missed: a load can blind an RCD further upstream, not just the one directly protecting it. Smooth DC from an inverter on a Type B device can still disable a Type A RCD above it. So every RCD in the chain must suit the residual currents expected downstream:

  • No Type AC upstream of Type A, F or B.
  • No Type A upstream of Type F or B.
  • No Type F upstream of Type B.

The one exception: a Type AC, A or F RCD may sit upstream of a Type B where the RCD manufacturer specifically declares the arrangement suitable. This comes from an informative note to Regulation 531.3.3, so check it and Figure A53.2 in your copy.

The new Figure A53.2 shows these combinations. Remember that stacking two non-delay 30 mA RCDs gives no selectivity anyway. If you need an upstream RCD, use a time-delayed (S-type) device and check its type as well.

On site: before adding an EV charger or inverter to an existing board, look at what sits upstream. A whole-board Type AC or Type A RCCB may now be the wrong choice.

Regulation references: 531.3.3, Figure A53.2

Other protective measures

SELV and PELV

Section 414 limits the voltage so that a shock can’t do serious harm. Extra-low voltage means no more than 50 V AC or 120 V ripple-free DC. The supply must come from a safety source, such as a safety isolating transformer, with proper separation from higher-voltage circuits.

The difference is earthing. SELV circuits are not connected to earth. PELV circuits are earthed, often for functional reasons. Some special locations set lower limits.

Double insulation, and Class I and Class II

Class I equipment has basic insulation plus an earthed metal case. It depends on your CPC and on ADS. Class II equipment has double or reinforced insulation and no earth connection, so it carries its own fault protection. Class III equipment is designed to run from SELV.

Where a whole installation or circuit relies on double or reinforced insulation alone (Section 412), it must be under effective supervision so nobody fits Class I equipment later. That rarely happens in houses. Run a CPC to every point anyway: today’s plastic fitting may be swapped for a metal one tomorrow.

Regulation references: 412, 413, 414

Next: Inspection and testing et Certificates and EICR codes.

Lien secret