Nuisance RCD Tripping and Lightning Damage in a TN-C-S Office Building

No - 4.8 Ω is not the fault, and chasing 1 Ω is money spent on the wrong problem.

• In a TN-C-S system the disconnection time is achieved by the protective conductor and the overcurrent device, not by the electrode resistance. The electrode's job is to reference the installation and to dissipate lightning current; there is no maximum value for the electrode in a TN system.
• Even if this were a TT installation, the touch-voltage criterion 50 V/IΔn gives RA ≤ 50/0.30 = 167 Ω for the 300 mA device; 4.8 Ω is two orders of magnitude better than required.
• For the lightning protection system, IEC 62305-3 recommends a low earthing resistance, ideally below 10 Ω at power frequency; 4.8 Ω already satisfies that.

The real value of the 4.8 Ω reading is also questionable: with the ring bonded to the foundation reinforcement and to eight down conductors, the measurement is dominated by parallel paths through the structure, and the fall-of-potential result depends heavily on the probe spacing and on whether the utility neutral and the down conductors were disconnected during the test. It should be re-measured with those paths isolated before any number is quoted in a report.

Recommendation to the client: stop the rod-driving exercise. Redirect the effort to the two items that actually caused the failures - the neutral–earth link downstream and the missing coordinated SPD and bonding - and record the electrode resistance as part of the LPS documentation, not as a fault to be corrected.

RCD tripping. The link between the neutral and earth bars at the fifth-floor sub-board creates a parallel return path for neutral current through the earth bar, the structure and the ring electrode. Because that path bypasses the 300 mA device's toroid, the residual current it senses is a real difference between phase and neutral - nuisance tripping that grows with load, which matches the evening pattern. The fix is to remove the link and keep N and PE strictly separate downstream of the main earthing terminal, verify with a clamp meter that no current flows in the main earthing conductor under load, and re-test the RCD with a ramp test and at half-rated residual current. In a TN-C-S system the PEN is split only once, at the origin, and never recombined.

Lightning damage. Several defects combine:

1. No Type 1 SPD at the origin of the installation where the line enters the building - the surge had no first-stage diversion.
2. Only Type 2 devices at a sub-board, with no coordinated Type 3 protection at the IT rack or BMS controller, the two items actually destroyed.
3. Connecting leads of 2.5 m in 6 mm²: the inductive voltage across long leads adds several kilovolts to the let-through voltage at the equipment, so the SPD clamps the surge at the board but the electronics still see the remainder. Leads must be as short as practicable, well under 0.5 m, straight, and twisted together.
4. No equipotential bonding of the IT rack, raised floor and BMS panels to the lightning protection earth; differences in potential between the LPS and the services are what damages electronics.

Specification: Type 1 (Iimp per the calculated risk, typically 12.5 kA 10/350 per pole) at the main board with 16 mm² Cu connecting conductors, Type 2 (In ≥ 20 kA 8/20) at sub-boards with ≥ 6 mm² Cu, Type 3 at the IT rack, coordinated by manufacturer's energy tables, each with a suitable backup fuse or MCB and a status indicator. Bond all incoming metallic services and the rack at a single equipotential bar.

Earth electrode resistance:

• Method: fall-of-potential (three-point) with the current probe at least 5 × the electrode's largest dimension away and the potential probe stepped at 52 %, 62 % and 72 % of that distance; results within a few percent of each other indicate a valid reading.
• Precaution: disconnect the down conductors and the utility neutral from the installation under test, or the result simply reflects the bonded structure. A clamp-on tester is only valid where a single, isolated earth path exists and must not be used here.
• Criterion: no maximum for the TN system; record the value and confirm it is consistent with the LPS design (a low-frequency value below 10 Ω is the usual target).

Bonding continuity:

• Measure between the main earthing terminal and each bonded part (riser, rack, BMS panel, pipework): the reading must be low and consistent with the calculated resistance of the protective conductor run - in practice a fraction of an ohm on short runs. Confirm the N–PE link at the fifth-floor board has been removed and that no current flows in the main earthing conductor.

RCD verification:

• Ramp test to confirm the 300 mA device operates between 0.5 and 1 × IΔn, and the 30 mA devices within 300 ms at IΔn; record the trip times.

SPD verification:

• Visual: correct type, rating, status window green, backup device present and correct, lead lengths measured and recorded.
• Electrical: insulation resistance and continuity of the SPD's connecting conductors; note that insulation testing must be done with the SPD disconnected or with a test voltage below the SPD's Uc to avoid damaging it.
• Documentation: record the SPD let-through values and the manufacturer's coordination table in the handover file.