Short answer: the earthing system decides which protective device actually works. In a TN system the fault current returns through the metallic earth path, so the loop impedance is low and an MCB or fuse can clear a fault on its own. In a TT system the fault current returns through soil, the loop impedance is high, and the breaker alone will not trip reliably — the RCD is the protective device, and the earth electrode is part of its circuit. In an IT system the first earth fault is not supposed to trip anything at all, which is why it needs insulation monitoring rather than a bigger breaker.
We export breakers, RCDs and distribution boards as an independent trading company — not a manufacturer and not an appointed distributor of any brand. This is the first question to settle on any export panel, because it changes the device list before the first item is quoted.
The five configurations in one table
Letters describe where the supply is earthed and where the installation’s exposed metalwork is earthed. The first letter is the supply side, the second is the installation side.
| System | Supply side | Installation side | Fault path | What clears a fault |
|---|---|---|---|---|
| TN-S | Earthed at the source | Separate PE conductor all the way back | Low impedance metallic return | MCB, MCCB or fuse on overcurrent |
| TN-C | Earthed at the source | PEN combined neutral and earth | Low impedance metallic return | Overcurrent device; RCD cannot be used on a PEN |
| TN-C-S | Earthed at the source | PEN splits into N and PE at the origin | Low impedance metallic return | Overcurrent device, plus RCD for additional protection after the split |
| TT | Earthed at the source | Local earth electrode, no metallic return | Through soil and the electrode | RCD is the primary protection |
| IT | Unearthed or earthed through high impedance | Local earth electrode | Very low first-fault current | Insulation monitoring; first fault alarms, does not trip |
Why TT changes the device list
On a 230 V TT circuit, an earth fault flows through the electrode resistance at the installation and the electrode resistance at the transformer. If those add up to, say, 20 Ω, the fault current is only around 11 A — enough to be lethal, not enough to trip a 32 A MCB quickly, and not even enough to trip a 16 A one with certainty. The cable does not overheat, the breaker does not see an overload, and the metalwork stays live.
The RCD fixes this by measuring current going out against current coming back, not by waiting for enough current to flow. That is why the TT device list starts with a 30 mA RCD for socket circuits and a 100 mA or 300 mA selective RCD for the incoming, with time delay so the downstream device trips first.
The electrode itself then has a simple design rule: the product of electrode resistance and RCD rated residual current must not exceed the permitted touch voltage, conventionally 50 V AC. For a 30 mA device that is a theoretical 1,667 Ω; in practice everyone designs to a few tens of ohms, because soil resistance changes with the season and the standard’s formula assumes the electrode is the only return path.
| RCD rated residual current | Electrode resistance for 50 V touch limit | Practical comment |
|---|---|---|
| 30 mA | up to 1,667 Ω theoretical | Design to a few tens of ohms, not to the limit |
| 100 mA | up to 500 Ω | Common incoming value with time delay |
| 300 mA | up to 167 Ω | Fire protection or large installations |
| 500 mA | up to 100 Ω | Where local rules permit |
TN systems: the breaker still needs a loop
TN looks easy until the loop impedance is measured. The overcurrent device trips on a magnetic or instantaneous element at a multiple of its rating — commonly 5 to 10 times for a C-curve MCB — and the fault current has to reach that value at the far end of the circuit. Long runs and small sections push the loop impedance up, the fault current down, and the disconnection time out beyond the permitted maximum for the circuit.
That is why a TN design is checked circuit by circuit, not board by board: the furthest socket is the one that fails, and it fails on a calculation, not on an inspection. Where the loop is too high, the options are a larger cable, a shorter run, or an RCD, which is permitted as additional protection and is mandatory for socket outlets in many national rules regardless of the loop.
Where RCDs fit in each system
In TN-S and after the split in TN-C-S, an RCD works normally: it compares outgoing and returning current on the conductors it monitors. It must not be installed on a PEN conductor, which is why TN-C installations cannot use one at that point — the RCD would see the neutral current of the whole installation as an imbalance.
In TT, the RCD is the main protection and the electrode is its return path. In IT, an RCD is not a substitute for insulation monitoring: the first fault produces a small capacitive current that may not reach any residual current threshold, and the point of the system is to keep running until the fault is found. Some IT designs add an RCD for the second fault, and that is a design decision made against the local code, not a default.
Mixed systems on one site
Export projects often have a TN-C-S supply at the boundary and TT wiring inside a building with its own electrode, or a generator that is IT when running islanded and TN when paralleled. The device list has to follow the system in each mode. A board that is correct on mains and dangerous on generator is a common finding in factories that added standby power after the original installation.
The practical control is a single-line diagram that states the earthing arrangement at every point where the system changes, plus the measured electrode resistance and loop impedance values written on it. Without those two numbers, the RCD choice is a guess dressed as a specification.
What to put on the purchase order
The earthing system per board; the RCD type (AC, A, F, B) and sensitivity; whether the RCD must be selective with a time delay; the number of poles and whether the neutral is switched; the rated conditional short-circuit current of the RCD combination; the electrode resistance target where TT applies; and the national rule that governs socket protection in the destination. Add the supply arrangement — mains, generator or both — because that is what decides whether the same device list is valid in every operating mode.
FAQ
Can I use an MCB instead of an RCD in a TT system?
No, not as the earth-fault protection. The loop impedance through soil is too high for the breaker to see enough current. The MCB still provides overload and short-circuit protection; the RCD provides the earth-fault protection.
Why can’t an RCD be installed on a TN-C system?
Because the neutral and protective functions share one PEN conductor. The RCD would measure the installation’s normal neutral current as an imbalance and trip. The RCD belongs after the point where the PEN is split into separate N and PE.
Does an IT system need an RCD?
Not as first-fault protection. An IT system is designed so that a single earth fault does not trip, and insulation monitoring is what detects it. Any RCD added for second-fault protection is a design decision against the local code.
What earth resistance should I aim for?
Design to the touch-voltage rule for the RCD you are using, and then improve on it for seasonal variation. Thirty to one hundred ohms is a common practical target for a small TT installation, but the number that matters is the one measured on site, in the dry season.
Related: RCCB Type AC vs Type A · NB1L vs NB2LE vs NXBLE comparison · 30 mA vs 100 mA RCD · Breaker and cable coordination · MCB trip curves B, C and D · MCB range overview · Cross-reference an existing device · Request a quotation