Omglobalnetwork — Practical electrical | omglobalnetwork.com
The protective conductor is chosen from the live conductor's cross-section, not from the breaker rating — and the automatic-disconnection test can force a larger size than the table suggests.
IEC 60364-5-54 divides the question into three tiers. Where the live conductors are up to 16 mm², the minimum copper protective conductor is 2.5 mm². Where the live conductors run from 16 to 35 mm², the minimum steps up to 6 mm², and from 35 to 150 mm² it steps again to 16 mm². Because the live conductors behind a 32 A breaker are almost always well under 16 mm², the 32 A circuit lands squarely in the first tier.
The reason the earth is keyed to the live conductor rather than the breaker is the physics of the fault. Fault current flows out along the live conductor and returns along the earth, so both must survive the event and both shape how much current flows. Overloads are a slower problem handled by the thermal element, which begins to release at 1.13 × In and must have released by 1.45 × In within one hour; earth faults are high-current events governed by disconnection time.
On narrow screens, swipe or scroll the plate sideways.
Those disconnection times come from IEC 60364-4-41 for TT systems: 3 s for circuits up to 32 A and 0.3 s above 32 A. The 32 A rating is therefore a boundary in its own right — the same breaker size that answers the earth-sizing question also marks where the allowed clearing time tightens by a factor of ten.
The installation-method column is the standard's built-in exception. Depending on how the cable is run, a larger protective conductor than the tier value can be required, so the tier should be treated as a floor. Skipping this column is the most common paperwork error in earth sizing.
Finally, the automatic-disconnection test can overturn the table entirely. A healthy 5 m loop of 1.5 mm² twin-and-earth measures roughly 0.5–1.2 Ω end to end, but one corroded joint can add more than 1 Ω and push the loop impedance past the tripping threshold. When the measured loop is too high for the required disconnection time, the remedy is a larger earth or a repaired joint — the table minimum is no defence. And because copper resistance rises about 0.4% per °C, a loop measured at 60 °C reads roughly 16% higher than its 20 °C value, so marginal results deserve a second look at working temperature.
Further reading