How stuff works - Line power distribution and protective measures in Europe (consumer side)

03 Aug 2020 - tsp
Last update 07 Feb 2021

Since it seems to be a huge mystery to many people how our power systems work this article tries to provide an introducory view onto grid power supply and the used protective equipment. Note that this is not an in depth article that tries to teach how this systems work or describe them into detail. You won’t be qualified to design a power supply system or perform installation after reading this article - please refer to some professional in this case. Incorrectly used protective equipment might be as good as no protective equipment at all - so please bear in mind that one should know how to design and install such equipment safely (and you might even not be allowed to do so on your own depending on the location you live in).

3 phase power system (400V, 230V)

As of today the most common energy system used in Europe is the 3 phase alternating current system. One can imagine it as having a generator using 3 poles that are separated by 120 degrees - technically generators usually have of course more than 3 poles but always a multiple of three. The target network frequency is usually 50 Hz. There are some other systems in place - for example for railways - that operate at 16.7 Hz.

One can now imagine that a center pole of all three induction coils has been connected. The non connected sides are called clockwise L1, L2, L3 (or earlier R,S,T). They are the so called phase. The side that has been connected together is called the neutral. Why will be shown later on. The most important thing when connecting a building or machine to this power system is basically not which phase is called L1, L2 or L3 but only that they form a clockwise rotating field - so if one connects L2, L3, L1 or L3, L1, L2 in this order at the main building connection box everything’s ok too.

The basic idea is that every phase provides an amplitude of $325V$ between a phase (L) and the neutral (N). The effective value - also called root mean square - is calculated easily

[ U_{eff} = \sqrt{\frac{1}{b-a} \int_{a}^{b} U(t)^2 dt} ]

As one can see first the centerline of the squared voltages is calculated, then the square-root is calculated. One can see this as an effective or mean voltage level. This leads to the well known $230V$ between phase an neutral.

If one wants to look at the potential between two phases one has to consider the 120 degree phase shift:

[ U_{L1} = U_0 * sin(\omega * t) \\ U_{L2} = U_0 * sin(\omega * t + \frac{2\pi}{3}) \\ U_{L3} = U_0 * sin(\omega * t + \frac{4\pi}{3}) ]

The RMS can now easily be calculated as

[ U_{eff} = \sqrt{\frac{1}{T} \int_0^T (U_{L1} - U_{L2})^2 dt} ]

This leads to the well known $400V$ (or $380V$ depending on the amplitude you’re assuming) power current.

Why 3 phases and neutral?

The answer is pretty simple. If you calculate the effective sum of all three phases:

[ \int_0^T U_{L1} + U_{L2} + U_{L3} dt = 0 ]

they cancel out each other. This is only possible using three or a multiple of three phases. This is also the cause why the neutral is called that way - during optimal load all currents on the neutral wire in an 3 phase electrical system (not in your home though) should cancel out. Because of this one can build electrical power systems that do not even have a neutral wire on the grid side (see IT systems in the next chapter) or use smaller dimensions for the neutral wire than one would calculate when just adding up currents.

Network frequency - PI regulation of the power grid

TL;DR: This is not required to understand your homes electricity system but only to understand the power system at a large scale.

For the public system deviations from the targeted 50 Hz are normally kept below 0.2 Hz. This is done using a combined collection of measures. The first and most important is of course the primary regulation that tries to match the amount of primary generated power to the currently requested load. In case too much power is generated or large loads get dropped the frequency would rise, in case large loads are attached or generated power is reduced frequency would fall.

One can imagine this as having a force driving the generator that gets indirectly transferred to a motor. In case someone now adds a load to the motor the required force on the generator would increase.

The primary regulation is a proportional regulation.

Then there’s secondary and tertiary regulation. They’re usually used to do longer term regulation than primary regulation that is capable of reacting in duration counted in seconds. They both do perform some kind of integral regulation to compensate constant drift of the load centerline. In case secondary regulation stays active for a long period (for example when some prognosis was wrong, a power station failed, wind prognosis was wrong, etc.) tertiary regulation gets activated manually to make some space at the secondary regulation level or optimize costs. The manual operation of tertiary regulation as well as the prognosis play a major role on the electricity market and have a huge impact on the cost of produced electricity.

Note that regulation usually means to switch on or off large generators or loads like power stations (gas power stations and reservoir power station being really fast, nuclear being the slowest with regulation periods in hours so they’re normally only used to provide base power) or pumping water in reservoir power stations to consume excessive power.

In Europe there exists some quartiary regulation that’s done by Swissgrid. Since the power systems frequency can be used to provide a timing source for cheap synchronous clocks which has been really common before the invention of DCF, NTP and GPS based clocks the time tracked via network frequency is compared to correct UTC. In case there is a deviation of more than 20 seconds the system tries to increase or lower the centerline of the network frequency by about 10 mHz to compensate for that error. This is also done by requesting subtle load and generation changes inside a huge PI regulation loop.

How earth potential is handled (IT, TN and TT system)

There are three different basic power supply systems that all differ in how neutral and ground potential is handled. What is the ground potential? The basic idea is that our earth, all buildings, etc. are not perfect insulators - and on the other hand they can collect charges like an capacitor. To counter that effect two measures are typically used: The electrical bonding in which different components are simply connected via low resistance paths (i.e. wires) so the electrical potential equalizes. This is for example used to connect all pipes, metal cases, metal bathtubs, steel reinforcement, etc. to a common potential.

Then there is the grounding in which one connects potential literally to a large rod in the ground that should provide a low resistance path to earth. Normally bonding and grounding are connected together - but one has to differ between protective earth and ground to not interfere with RCD operation (i.e. the bonding used in your home is not allowed to be connected to earth directly - more on that later).

As one can imagine electrical bonding is already a safety feature - it prevents electrical shock in case a potential is connected to any touchable surface or pipe and it prevents electrical shock by charge buildup due to capacitive coupling.

So how are earth an neutral handled on the network side? There are basically 3 different systems in place. Note that you cannot decide which kind of network you’re building when connecting to a public grid - this has to match your grid provider.

Isole Terre (IT) system

This is the most basic system in which only the three phases are provided by the network provider. Earth is connected locally to a ground rod but never ever is connected to the grid. Such systems normally are pretty small and not commonly seen. The generator can be but hasn’t to be grounded on it’s own.

Terre Terre (TT) system

This is more commonly seen up to the main building connection box. In this case the network operator supplied three phases (L1, L2, L3) and a neutral connector. The generator as well as your building are connected to earth.

Terre Neutre (TN) system

This is the most commonly seen system inside buildings. There is a distinction between TN-C (Terre Neutre Combine), TN-C-S (Terre Neutre Combine Separe) and TN-S (Terre Neutre Separe) systems. They mainly differ in the way protective earth is connected to the network and how the building is grounded.

In TN-C the source is grounded to earth and the network supplier provides 4 wires - the three phases and a neutral that also doubles as protective earth (called PEN). On the consumer side PEN is connected to neutral side of the network as well as to the PE connection. This is the common setup that one sees in main building attachment boxes supplied from the grid operator.

In TN-C-S the network provider routes PEN as described for TN-C to your house attachment box and separates PE and N wires there. They’re routed separately to your main distribution closet - and are attached to a local grounding rod also at the main attachment box.

The TN-S systems (the part seen after the separation done for TN-C-S) goes a step further and separates the PE and N wires at the thought source and routes them as separate wires to your main attachment box.

As one sees the basic idea is the same everywhere and one can come into contact with multiple different systems depending on the location of one’s building. Up till the main connection box one most of the time sees an TN-C system consisting of 4 wires (L1, L2, L3, PEN). Inside the main attachment box one separates PE and N - they are both connected directly to PEN - and does attach a local ground connection - so one see’s a TN-C-S system. Inside one’s distribution cabinets one then sees separate PE and N wires all the time so one sees a TN-S system.

The basic difference between these systems is the point at which PE and N are separated or if this separation is seen from the point of view of the local system - and of course where the earth potential is connected. The idea is basically always the same.

The main difference and root cause why one normally separated protective earth and neutral in the main house attachment closure (or has to do so by local regulations) is that using a combined PEN can be more dangerous than having N and PE separated. Why is this the case? Under normal operation one uses 230V effective voltage between each phase and the neutral. In case the PEN wire is broken on the supplier side all loads are connected to each other via the N wires inside the building - so they’re connected in series between two phases (400V). The potential at a connector is then mainly determined by load resistance so in worst case one can have 400V connected directly to your power outlet which might destroy devices and provide a hazard to people interacting with the system. More severe residual current detection devices and breakers (RCD) do not work anymore with broken earth connection. When having separated PE and N the probability of both of them being broken is lower so in case the N wire is broken overvoltage protection - and protective equipment - still works. Because of this separation between N and PE is normally done at the earliest possible point.

Protective equipment

Miniature circuit breaker (MCB / LS)

Circuit breakers limit the current inside the protected circuit. Note that they do not trip at the voltage specified but obey a specific curve that determines how long it takes them to trip at a given current. This curve is normally specified by a letter like A, B or C. The other rating MCBs have is the current they’re rated for. A B13 or B16 breaker for example is having a B timing curve with a 13A or 16A limit. Note these devices are not there to protect humans but to reduce fire hazard. They normally trip in case of short circuits between phase and neutral which would form a low resistance connection and thus in theory infinite current. Though infinite current is not possible wires can only have a limited carrying capacity because current heats them up. If one operates wires at too high current this can lead to sever damage and is one of the most common fire causes up to day. To counter that one has to add MCBs in front of each circuit loop that match the wire diameter (or are rated smaller than the used wire diameter). The diameters and methods to route wires are regulated - please do not violate these regulations since they take much detail into account.

Normally one sees a MCB tripping in case of either a damaged cable (short circuit) or a damaged device. The other common cause in which an MCB trips is overload - if you attach too many devices to a power network the specific circuit trips.

Note that normally protection is not directly hierarchical. Of course upstream fuses are rated higher than downstream fuses but they’re normally not dimensioned as the sum of the downstream fuses. If you have 10 16A circuit breakers your fuses inside the main attachment box are normally not rated 160A but they’re selected according to a given parallelity factor or a given booked maximum current. Because of this it’s entirely possible that upstream fuses blow even if the downstream MCBs are not.

There are basically two types of MCBs: Ones that only switch the phase and ones that switch phase and neutral. The switching of the neutral is required for some areas like bathrooms and is a good idea also in the general case.

Residual current operated circuit breaker (RCCB / RCD / FI)

The residual current operated circuit breaker is a protective device that protects human beings due to insulation fault. It’s a simple system that sums up all current flowing through the device. In case the sum deviates too large from zero the device breaks - normally with currents as low as 25 to 50 mA. The protective effect happens because of protective earth. The RCCB basically only sums up current between L1, L2, L3 and N. Protective Earth is routed around the RCCB and never has to be connected to the neutral loop downstream of the RCCB. In case any insulation fault happens or someone touches a live wire current flows through earth back to the network (since N and PE are connected in your house attachment box). Since current does not flow through the RCCB the sum of currents is not zero any more and the device trips.

The summation of currents is done for non electronic monitoring devices in a pretty simply way: All four wires are routed through a transformer ring. Another wire that senses the sum current is wound onto the transformer ring - the induced current is tripping the circuit breaker in case it’s reaching a given threshold. Depending on the system design this transformer can be sensitive to AC current only (type AC) or also to pulsed direct current (type A). Type AC should not be used for most installations anymore since it’s design also prevents function in case of overlayed direct current which can drive the transformer core into saturation and thus prevent imbalances from triggering the switch - because of this they’re also forbidden in some European countries, for example in Germany. Type A RCDs are the most commonly seen as they’re capable of protecting against most failure scenarios seen in household and most industrial settings - and they’re pretty cheap to manufacture. Note that type A RCDs may stop to function in case of too large phase shift - this is also another reason why it’s not allowed to operate devices that do too much phase shifting (like a huge number of fluorescent lamps in series, huge capacitors, etc.) without power factor correction directly on the network.

Since they work based on transformers RCDs of type AC and A are usually omnidirectional - it doesn’t matter if power feed goes in on the top and exits on the bottom or the other way round.

There are currently three additional RCD types available - namely B, B+ and F. Type B and B+ are triggering electronically and not electromechanically any more. They do provide protection against non pulsed discrete current, alternating current and work independent of phase shift between voltage and current. Currently they’re not seen that often since they do require a substantial amount of logic that raises cost. They’re often seen in laboratories though where they’re also required by regulations in some countries. Type B+ also protects against higher frequency failure currents up to 20 kHz. Note that since type B and B+ RCDs use electronic components they’re usually not omnidirectional any more. Refer to the manufacturers instruction when in doubt.

Type F is the most seldom seen RCD - it can be seen as an extension of a type A residual current detection device that’s also sensitive up to 1 kHz of failure currents and is used in case of variable frequency drives since they’ve higher frequency effects on the power grid.

Note that these devices should be tested regularly. Testing is done via an integrated test button that connects phase on one side to neutral on the other side of the switching equipment. One should do this test at least once a year to detect mechanical fault.

Residual current monitoring (RCM)

RCM is a method not seen in households. The difference to RCCBs is that residual current is only monitored and detected. This is often seen in data centers and sometimes in industry. This allows one to monitor slow degrading of insulations or capacitors and replace equipment without unplanned large scale outages.

Selectivity when using residual current circuit breakers

A question that often arises is what happens when one uses residual current detection devices in series. This is when selectivity gets relevant. All RCCBs operate at a given tripping current with a specified minimum and maximum time constant. When building systems serially one can either aim partial selectivity at which one selects upstream RCDs to be at least two times slower than downstream RCDs or absolute selectivity by which downstream RCDs maximum reaction time is lower than the upstreams minimum reaction time.

Note that leakage current of course builds up with wire length and the number of machines (especially heavy machinery) connected so one also has to take that into account when designing the protection system.

NH fuses in the house connection box

Technically these do not belong to the consumer side. NH fuses are normally used in the house connection box in front of the electricity meter. Sometimes additional fuses - many times enclosed in round packaging that can be screwed into their receptible - are used in front of meters but after NH fuses in multi-apartment buildings so that tripping the fuses in front of the metering device does not trip the whole building and such that working on cabling for a single apartment does not require shutting down the whole building.

NH fuses should not be handled by non-professionals. There is special equipment that allows one to pick them on their conductive parts even while under power and that protects the wearer from and current as well as from arcs. Normally regulations require one to wear special gloves (sometimes integrated into the tooling) as well as a face shield and sometimes an insulating skirt when working with these fuses. Note that most of the time they’re contained in an plumbed compartment so access is only allowed to professionals from your electrical network operator or electricity provider.

Arc fault detection device (AFDD)

These are also called fire protection switches in some case. AFDDs try to electronically detect the build up of arc flashes. This is possible due to electronics monitoring the frequency spectrum of the flowing current. Since arcs of squeezed or partially broken cables are one of the most common fire sources in households these devices are already mandatory in some European countries like the United Kingdom. It’s a good idea to use AFDDs - but of course they’re costly. Note that they do not detect an arc immediately but most of the time fast enough to prevent an fire outbreak.

Combined devices

Of course all of the mentioned protective equipment might also be combined and applied in multiple levels. To make this somewhat easier there are some combined RCD and miniature circuit breakers available that are way more costly than traditional RCDs and mostly type A but allow one to secure every loop with it’s own RCD. They’re mostly used as a second level RCD beyond a selective upstream one.

This article is tagged: How stuff works, Tutorial, Power grid

Data protection policy

Dipl.-Ing. Thomas Spielauer, Wien (

This webpage is also available via TOR at http://jugujbrirx3irwyx.onion/

Valid HTML 4.01 Strict Powered by FreeBSD IPv6 support