Electrical protection
There are two different things to protect, and confusing them is the costliest conceptual mistake in the trade: the installation, which is protected against overloads and short circuits, and people, who are protected against contact with current. Each has its own device, and neither replaces the other.
The principle of the thermal-magnetic breaker, the residual-current device (RCD) and grounding is developed in electrical installations. Here we go deeper into what decides a real project: earthing systems, the curves, breaking capacity, selectivity and overvoltages.
01Direct and indirect contact
Touching a live conductor. It is prevented with insulation, enclosures, barriers and clearances, and with safety extra-low voltage where appropriate.
The 30 mA RCD is an additional protection in this case: it reduces the severity, but it does not prevent the contact or replace the insulation.
Touching an exposed metal part that has become live because of an insulation fault. It is prevented with grounding of the exposed metal parts plus a device that automatically cuts the supply when the fault appears, or with double insulation.
This is the protection that is most neglected and the one that prevents the most deaths: the fault exists even if nobody notices anything, until someone touches.
| Current through the body | Effect on a person |
|---|---|
| 1 mA | Threshold of perception: tingling. |
| 10 mA | Muscle contraction: the conductor can no longer be released. |
| 30 mA | Difficulty breathing. It is the trip value of the general-purpose RCD. |
| 50 mA and more | Risk of ventricular fibrillation, which is the usual cause of death by electrocution. |
The regulatory criterion is that no exposed metal part may be more than 24 V above ground under normal humidity conditions. From this comes the condition on grounding: the resistance of the electrode multiplied by the RCD trip current must stay below that value. With a 30 mA RCD, that product is met with plenty of margin, but that does not authorize a poor ground connection: if the RCD fails, the ground is the only barrier left.
02Earthing systems
The first letter indicates how the source is grounded; the second, how the exposed metal parts of the installation are grounded.
| System | How it works | Where it is used |
|---|---|---|
| TT | Transformer neutral to ground; exposed metal parts to a separate, independent ground. The fault current returns through the ground and is low: only the RCD detects it. | It is the system of home distribution in Argentina. |
| TN-S | Exposed metal parts connected to the source neutral by a separate protective conductor. The fault is practically a short circuit: the thermal-magnetic breaker acts, and very quickly. | Industry and large installations with their own transformer. |
| TN-C | Neutral and protective functions in a single conductor. Economical, but any break in that conductor puts voltage on all the exposed metal parts. | Falling out of use in new installations. Does not allow an RCD. |
| IT | Source isolated from ground. The first fault does not produce dangerous current: a monitor signals it and it is repaired without interrupting service. | Operating rooms, processes that cannot be interrupted, vessels. |
In a TT system, the current of a ground fault has to close through the ground, passing through the resistance of the installation’s own grounding plus that of the utility’s neutral. That current is usually only a few amperes: far too little to make the thermal-magnetic breaker trip, and more than enough to kill someone. The only device that detects it is the RCD. In a TN system, on the other hand, the same fault is almost a short circuit and the thermal-magnetic breaker acts on its own.
03Curves, breaking capacity and selectivity
| Curve | Magnetic trip | For what load |
|---|---|---|
| B | 3 to 5 · In | Resistive loads and circuits with long cables: lighting, heating, outlets. |
| C | 5 to 10 · In | General use. Tolerates the starting current of small motors and switching power supplies. |
| D | 10 to 20 · In | Transformers and motors with a very high starting current. |
It is the maximum short-circuit current that the breaker can interrupt without being destroyed: 3, 4.5, 6 or 10 kA in the household and small industrial formats.
It must be greater than the prospective short-circuit current at the point of installation, which depends on the transformer power and the distance. Near the service entrance that current is high; thirty meters away and through 2.5 mm², much lower.
When a fault occurs, only the nearest protection device should act, not the upstream one. It is achieved with steps of rated current, with different curves and, in the case of RCDs, with a selective or time-delayed device at the head of the installation and the instantaneous 30 mA ones downstream.
Without selectivity, a fault in the washing machine leaves the whole house —or the whole plant— without power, and diagnosis becomes much harder.
Single-phase motor of 1.5 kW, rated current 8.5 A, starting current 6 times the rated, that is 51 A for a few seconds.
- With a 10 A curve B thermal-magnetic breaker, the magnetic trip starts between 30 and 50 A: the motor would not start, it would trip on every attempt.
- With 10 A curve C, the magnetic element acts between 50 and 100 A: the 51 A peak sits at the limit. It works, but only just.
- With 16 A curve C, the magnetic element acts between 80 and 160 A: it starts comfortably. Overload protection is provided by the thermal overload relay of the starter, set to the rated current of the motor.
That is the logic of every motor start: the thermal-magnetic breaker protects the cable and the thermal overload relay protects the motor.
04Overvoltages
| Origin | What it produces |
|---|---|
| Lightning strike | Pulses of kilovolts and microseconds, direct or induced in the lines. They destroy electronics throughout the installation. |
| Switching | When inductive circuits —motors, transformers— are opened, spikes appear that age the insulation and disturb the equipment. |
| Loss of neutral | The most destructive fault in distribution: without a neutral, the loads end up in series between phases and some receive up to 380 V. |
They are installed in the distribution board, between the live conductors and ground. On a surge they conduct the energy to ground and limit the residual voltage to a value the equipment tolerates. They are classified as type 1 for direct strikes, type 2 for general protection of the board and type 3 next to the sensitive equipment.
They need a low-impedance ground connection and short leads: an SPD with half a meter of cable to ground loses a good part of its effectiveness.
A relay that measures the mains voltage and disconnects the installation when it leaves a range —for example 190 to 250 V— with timed automatic reconnection. It is what saves the equipment from loss of neutral, which no thermal-magnetic breaker or RCD detects.
05In the lab
With a test instrument, measure the actual trip current and time of 30 mA and 300 mA RCDs. Compare with the test button, which only verifies the mechanism. Also try an old RCD: the difference in times is usually revealing.
On a training board with a grounded metal housing, cause a controlled insulation fault through a resistor. Measure the fault current and the voltage of the housing with respect to ground, with and without the protective conductor connected. The comparison explains on its own why the PE is not optional.
Subject curve B and C thermal-magnetic breakers of the same rated current to increasing overloads and record the trip time. Plot on a log-log scale and compare with the manufacturer’s curve.
Measure the resistance of a ground rod with an earth resistance tester using the three-point method. Repeat after wetting the soil and after adding a second rod in parallel. Verify how much the value really drops.
06Common mistakes
| Mistake | Consequence |
|---|---|
| Believing the thermal-magnetic breaker protects people | In a TT system, the ground fault current is too low to make it trip. Only the RCD protects. |
| RCD without a protective conductor | The fault finds no path to ground and the RCD only acts once someone touches: it is too late. |
| Bypassing the RCD “because it keeps tripping” | It trips because there is a real fault. Bypassing it removes the only protection for people. |
| Curve B on a motor circuit | The starting current makes the magnetic element trip: the motor never starts. |
| Insufficient breaking capacity | On a short circuit, the breaker cannot interrupt and is destroyed, with arcing and flying debris. |
| Whole board with identical protection devices | There is no selectivity: a fault in one appliance leaves the entire installation without power. |
| SPD with a long ground lead | The inductance of the cable raises the residual voltage and the arrester stops doing its job. |
| Joining neutral and ground downstream of the board | Currents flow through the protective conductor and the RCD trips for no apparent reason. |
07Self-assessment
What is the difference between direct and indirect contact?
Direct contact is touching a part that normally is live; indirect contact is touching an exposed metal part that should not be live and is, because of an insulation fault.
Why is the thermal-magnetic breaker not enough in a TT system?
Because the ground fault current closes through the ground and is only a few amperes: too low for the breaker and enough to be lethal. Only the RCD detects it.
What do the letters in the TT, TN and IT systems indicate?
The first, how the source is grounded; the second, how the exposed metal parts of the installation are grounded.
What current is considered dangerous for a person?
From 30 mA there is serious risk, and above 50 mA the danger of ventricular fibrillation appears. That is why RCDs for protecting people are 30 mA.
What is the difference between a B and a C curve?
The magnetic trip threshold: 3 to 5 times the rated current for B and 5 to 10 times for C. C tolerates starting currents; B protects long cables better.
What is breaking capacity and why does it matter?
The maximum short-circuit current the breaker can interrupt without being destroyed. It must exceed the prospective short-circuit current at that point of the installation.
How is selectivity achieved between RCDs?
With a selective or time-delayed RCD of higher rated sensitivity at the head of the installation and the instantaneous 30 mA ones on the final circuits.
What protects the motor and what protects the cable during a start?
The thermal overload relay, set to the rated current, protects the motor; the thermal-magnetic breaker protects the cable and clears the short circuit.
What happens when the neutral is lost in a three-phase network?
The single-phase loads end up in series between phases and some receive voltages well above 220 V. Neither the thermal-magnetic breaker nor the RCD detects it: an overvoltage relay is needed.
Why does an SPD need short ground connections?
Because with a microsecond pulse, the inductance of the cable produces a significant drop that adds to the residual voltage. With long cables the arrester protects much less.