My dad is an Electrician and I myself am an Electronics Design Engineer, and to this day he still hasn't been able to give me a good reason for this.

Consider the two following pictures/situations - both the same case, but with neutral not tied to earth in the second. Apologies for the poor diagrams, but imagine they are sticking a fork in a plug/knife in a toaster/etc. in order to touch active.

enter image description here

In the first picture, the person gets an electric shock. Classic case. This is because there is 240VAC difference between the person's hand, and the earth at their feet. The key thing to note here is that it was the 240VAC difference that caused the shock.

In the second picture, the person is touching the active wire again - however, since earth is not tied to neutral, there is no guaranteed 240VAC difference. None. Like hooking up only 1 end of a battery to a light, this situation has no closed circuit. Thus, the only way to receive a shock is if a person held active and neutral at the same time - which you would have to be trying to kill yourself if you did that somehow (i.e. my point is, most electrical shocks are caused by active -> earth potential, not active -> neutral - and, tieing neutral to earth does nothing to prevent active -> neutral potential shocks).

Yes, earth might be floating and could be "any" potential with reference to active, and it's nice to tie it to neutral at power stations, transformer outlets, and outside our house with an earth stake so "we know" what potential it sits at. But you could make that argument that it could rise to some dangerous potential about any isolated power supply. So I don't think that's a solid argument and the only reason. On top of this, isolated transformers/power supplies are sometimes used for the sole purpose of protecting from shock - so why don't we just isolated the whole earth from our power grid? Haha.

Obviously, earthing chassis would no longer be necessary either if neutral was not tied to earth - because touching the metal casing would not be dangerous if for some reason the device became live (i.e. same as situation 2).

TL;DR: is the only reason we tie earth to neutral so that we know the ground beneath us is 0V with respect to active? Or is there some other reason?

  • 13
    \$\begingroup\$ ok, second scenario ... you poke a fork into the active terminal saying "haha, I'm not getting a shock", and as Karma would have it, your neighbour pokes a fork into the neutral terminal saying "haha, i'm ....." \$\endgroup\$ – jsotola Mar 15 '18 at 6:20
  • 1
    \$\begingroup\$ If your Dad is a lecky - then he should have a copy of the IEE regs - it’s covered in there. \$\endgroup\$ – Solar Mike Mar 15 '18 at 6:43
  • 11
    \$\begingroup\$ It would be fun if your ac would then float at 380kV \$\endgroup\$ – PlasmaHH Mar 15 '18 at 10:38
  • 4
    \$\begingroup\$ On a grid, it's likely various shorts would develop between distribution lines and earth. If the system is floating, a person grounded could be live to one or both power lines. A second short could be lethal. With mains and equipment frames grounded to earth, a short would trip a ground fault and be safer for users. \$\endgroup\$ – Optionparty Mar 15 '18 at 16:04
  • 2
    \$\begingroup\$ I don't know if anyone mentioned lightning yet. But lightning is one of the reasons why we need grounding. \$\endgroup\$ – mkeith Mar 15 '18 at 16:30

10 Answers 10


There are four reasons for grounding the neutral.

1. Grounding neutral provides a common reference for all things plugged into the power system. That makes connections between devices safe(r).

2. Without a ground, static electricity will build up to the point where arcing will occur in the switchgear causing significant loss in transmitted power, overheating, fires etc.

3. With a floating system it is possible to have a short between both in-house and neighboring systems via the ground path as shown below. Turning a light on in your house can cause a light to go on in your neighbors house too. This characteristic is highly unpredictable.


simulate this circuit – Schematic created using CircuitLab

4. Finally, by giving ground a return path to neutral, a short to the grounded chassis of an appliance causes a predictable outcome in terms of a fuse or breaker response. This provides a great deal of preemptive protection to the user.

In Summary

In a simple model it appears that not tying ground back to the neutral would be safer. However, in reality, in a distributed power system there is no guarantee of this since you have no way to know if there is some other path back to the transformer via a different route. That is, in point 3 above, you may be in danger of being electrocuted just as much as if your neutral was grounded.

In the end the other benefits of tying ground back to neutral outweigh the one possible, but unreliable, isolation benefit.

NOTE: From point 4 there is a paradigm shift in the way you need to think about the neutral-ground connection. Do not think of neutral connected to ground, but instead think of ground being connected to the neutral to allow the current from a short to ground to return to the transformer.

  • \$\begingroup\$ Electrical codes allow for the use of three-prong GFCI receptacles in cases where a grounding conductor is unavailable, provided they are marked "no equipment ground". On such receptacles, the ground lead is supposed to be disconnected. I would think that would create a danger, though, that a short between hot and exposed metal would go undetected unless or until it is touched by something with a path to ground. It would seem safer to have a special GFCI with some kind of monitoring circuitry on the non-connected ground pin, but I've never heard of such. \$\endgroup\$ – supercat Mar 15 '18 at 22:58
  • 2
    \$\begingroup\$ I think that rule (which only applies to old installations) is a lesser of two evils type of thing. People are going to use their class 1 appliances one way or another and it's better that when they do they have GFCI protection to at least limit the duration of the electric shock. \$\endgroup\$ – Peter Green Mar 16 '18 at 14:39
  • 1
    \$\begingroup\$ Except that dirt isn't nearly as good a conductor as everyone thinks it is. That's why they go to all the trouble to mine copper. It can flow enough current to bite you, but not to return any productive amount of current. They optimistically hope for 25 ohms, the same resistance as three miles of common 12AWG branch-circuit wire. \$\endgroup\$ – Harper Mar 16 '18 at 17:47
  • 1
    \$\begingroup\$ In addition to the above I am not sure how much benefit you'd get from a residual-current circuit breaker without grounding. Also in individual cases where a floating circuit is desired for single devices it can be achieved with a transformer. \$\endgroup\$ – kasperd Mar 17 '18 at 23:11
  • 1
    \$\begingroup\$ @Fattie that is correct. \$\endgroup\$ – Trevor_G Mar 18 '18 at 14:14

What you're talking about is an isolated system. I have an extended treatise on it here. In an isolated system, "the first ground fault is free" (and becomes the neutral-ground bond). This is the idea you are promoting.

The problem is the second one. Unless you have maintenance staff actively doing isolation testing and chasing down and eliminating that first ground fault, it will fail silently, undetected, and lay in wait. So you're right back in the same predicament, only now, you have no idea whether hot or neutral will be lethal to you today.

There is also the fallacy that you have discovered one use-case where your idea is better, but you're failing to consider all the other use-cases. The NFPA does, and considers them all in balance, and develops best practices that will save the most lives and houses. That is literally their job, being the National Fire Prevention Association.

Also an isolated system doesn't work unless you have your own transformer, because the entire system must be under common maintenance so you can assure it remains isolated. I have the luxury of having my own transformer. I have run it as an "isolated system" by accident (faulty neutral-ground bond). The "first ground fault" indeed failed silently and caught me unawares. I discovered this after de-energizing a circuit and pulling the wires off an outlet. I flashed hot to earth just to make sure the circuit was off, and this re-lit the circuit! What??? Turns out on an unrelated circuit, hot had faulted to ground. Ground was 120V from neutral everywhere in the system even on circuits which were turned off! That's super bad, and just the kind of nonsense that happens on isolated systems that aren't competently maintained. Failing silent is BAD.

I will say this: it was a good validation test for the previous work, which was a complete rewiring of a site which had dozens of serious defects.

  • 1
    \$\begingroup\$ Re: your last sentence, some people call it 'homeowner' wiring (untrained people trying to do the work themselves). I found some of that also. Now you know a new swearword! \$\endgroup\$ – user56384 Mar 16 '18 at 17:22
  • 1
    \$\begingroup\$ I think this is a good argument in the case of my solar query above. But there is a certain logical failing: After 1:N faults, the system is exactly as dangerous as the grounded N system is after 0 faults. Of course I live in a 3 wire country, and metal cases will be grounded regardless of whether N is. (so 2 faults don't result in a hazard - a 3rd ground wire fault is required) \$\endgroup\$ – Henry Crun Mar 16 '18 at 22:32
  • \$\begingroup\$ @harper In your treatise you mention "By the way, many panels have an "electrician's outlet" a few inches away from the panel." Is this controlled by the main-switch? Or can you still use it when the main-switch is off? (I rather like the idea, and must check the wiring regs to see if that is legal here) \$\endgroup\$ – Henry Crun Mar 16 '18 at 22:35
  • \$\begingroup\$ @Henry not exactly as dangerous. By causing the first fault in an engineered way, you can min-max for minimum life hazard, unlike my ugly random failure. That minmaxing gets done by the NFPA, it is pretty much their one job. The electrician's outlet is on a regular branch circuit breaker all by itself. No bypassing safety equipment, it's legal no question. \$\endgroup\$ – Harper Mar 17 '18 at 2:49
  • 1
    \$\begingroup\$ @HenryCrun that presumes all such faults are of equal probability. The nasty one (LC) is far less probable, because UL insists that equipment be built to make it so. Without a UL listing it can't be used in fixed wiring and won't be sold at reputable sellers (Amazon Marketplace is not a reputable seller). \$\endgroup\$ – Harper Mar 21 '18 at 18:00

On an IT-network, where both lines on the socket are live, the GFCI wouldn't work on a single fault.
Which has benefits in in some high continuity systems (eg: operation rooms), a single fault doesn't turn everything off.
But you will need to actively monitor for single faults using insulation monitoring.

Instead, we ground neutral so that even on a single fault the protection mechanisms work. We call this a TT-network.

It has nothing to do with touch safety. SELV (safety extra low voltage 42V) is for wet areas and touch safety.

  • 3
    \$\begingroup\$ Thanks for making the clear mention of the different networks. It is important to realize that this is not always the case, and there are situations where we want everything to be "floating" so we can actually tolerate certain faults without causing downtime. \$\endgroup\$ – Joren Vaes Mar 15 '18 at 7:36
  • \$\begingroup\$ You may want to update the IT link to the English Wikipedia. \$\endgroup\$ – MSalters Mar 16 '18 at 10:43
  • \$\begingroup\$ @MSalters There is no such page. But the german version has an illustration. \$\endgroup\$ – Jeroen3 Mar 16 '18 at 10:59
  • 2
    \$\begingroup\$ Well, the page listing the multiple systems has an English version: Earthing system. And yes, mentioning earthing systems is a must on questions like these. \$\endgroup\$ – Mast Mar 16 '18 at 19:31
  • \$\begingroup\$ Here we call it an isolated system. IT networks are used to surf Facebook. \$\endgroup\$ – Harper Mar 21 '18 at 18:02

As Neil has pointed out, the big picture is that you are part of a big electricity network, and if it wasn't grounded somewhere, the whole damn thing would float high - perhaps to $lightning volts.

Your second question "Wouldn't it be safer to just float it", becomes a very interesting question when you have a local, unconnected solar power system. The electrical regs (here) oblige you to ground N, but really, that is just making it unsaferer.

This is a topic we (installing solar power) have argued about at some length without a good conclusion.

  • 2
    \$\begingroup\$ Some networks are actually kept floating. By doing so, you can tolerate certain faults (such as one of the lines connecting to ground) without suffering downtime. This is called an IT network, and is discussed in the answer by @Jeroen3 \$\endgroup\$ – Joren Vaes Mar 15 '18 at 7:38
  • 1
    \$\begingroup\$ Indeed, but the question is the merits of it vs grounded N when you are not connected to a network \$\endgroup\$ – Henry Crun Mar 15 '18 at 7:46
  • \$\begingroup\$ Solar panels tend to be outdoors, so some kind of fault out there (line falling on the frames or something) could bring the fault inside. For that reason, at least the frames should be grounded. Then consider the amount of isolation between the frame and the solar power wiring... Shoot, just ground it. \$\endgroup\$ – user56384 Mar 16 '18 at 17:29
  • \$\begingroup\$ Same with uninterruptible power supplies - they can easily be turned into a floating, powerful, deadly AC source that no safety device in your house wiring is going to protect you from. First thing to to make anyone inexperienced in a server room well aware of. \$\endgroup\$ – rackandboneman Mar 16 '18 at 19:06
  • \$\begingroup\$ @nocomprende Not really true: there are no overhead wires - thats why solar. Panels have to be grounded for lightning though. Some systems have hi DC voltages, and ground issues arise here to - but hv dc is way scarier. \$\endgroup\$ – Henry Crun Mar 16 '18 at 21:59

In the TV lab we prescribed the use of an isolation transformer to galvanically separate our device under test from the mains. This made the TV safe to touch, with ONE hand. It also made the TV safe to test, i.e. to connect the ground of your oscilloscope to the circuit. But when you connect a grounded 'scope to a floating circuit, it becomes grounded again, and in principle unsafe to touch !

To get to the point, we had a law that it is forbidden to connect a power strip to an isolation transformer. Use one transformer per device. Otherwise it becomes too easy to touch two devices and find out the hard way that one is "hot" relative to the other. You cannot galvanically separate a whole building and expect the circuit to remain floating and safe.

Besides inadvertent grounding through some device there is also leakage current to ground, through capacitors. Your computer has a galvanically separated power supply, so it is safe to touch. But there is a C between the primary ground and the secondary to short-circuit the EMI of the SMPS. If the ground is not connected and you touch the case then the 50-60 Hz current through that C (and the C of the transformer) gives you a tingle. Connect 10 such devices with 10 Cs together without explicitly grounding any of them and that tingle becomes a shock. That's why you should use an outlet with ground for modern electronic devices. [edit: added schematic from another thread Henry Crun]


  • \$\begingroup\$ Thanks for explaining the tingle I got as a teenager caressing my shiny new stereo system receiver! Ah, the good old days. \$\endgroup\$ – user56384 Mar 16 '18 at 17:26
  • \$\begingroup\$ There is also a (significant) capacitance from primary -> secondary of overlap wound transformers. Low leakage transformers have side-by-side wound bobbins, with thick bobbin plastic between windings. Alternatively, Farady shields are a grounded foil between the primary and secondary. (so the C is to ground) \$\endgroup\$ – Henry Crun Mar 16 '18 at 21:50

The main reason is to blow protective fuses, to ensure that the fault current is sufficient for that purpose. However it also helps limit voltage excursions in 3 phase distribution.

Live to chassis ground is a common fault. Without neutral being bonded to earth, no significant current would flow to blow the fuse and disconnect the live.

Consider a 3 phase local distribution transformer, 240v phase to N, 415v between phases. If a live to ground fault grounded the red phase, then N would become 240v to ground, and blue and yellow phases would become 415v to ground, putting more stress on the insulation in all the other properties taking their single phase supply from the same transformer.


One word answer: Predictability.

Sometimes, it's better for a network to be predictable than to "sometimes" or "usually" be safer/cheaper/better in some other way. Predictability makes global safety/efficiency/effectiveness possible, since it simplifies use of the network and design of things plugged into it. You solve problems once, instead of at every implementation.

  • \$\begingroup\$ Better the devil you know, than the devil you don't know. \$\endgroup\$ – user56384 Mar 16 '18 at 17:27

Here in Australia we have what is called an MEN system. Multiple Earth Neutral, the IEC describes the MEN system as a TN-C-S system (Terra Neutral Combined Seperate) which is a fancy way of saying; the neutral and earth conductor are functionally and physically the same conductor between the star point of the distribution transformer and the point of supply, which will be at the consumer's property.

It is at the point of supply that the combined conductor seperates into two physical conductors, the neutral and earth. The main earthing terminal is then connected to the greater mass of the earth via the main earthing conductor and an earth stake. This process is repeated at every property and thus forms part of a system we call the PME system (Protective Multiple Earth) .

The reason the the PME system is straight forward, the further you get from the transformer the greater potential rises on the neutral conductor with respect to earth. The PME system allows the the voltage rise to disapate to earth at each property and thus keeps the neutral voltage consistantly low. By keeping the neutral voltage at as close to ground potential as possible affords use a good reference voltage and a means to mitigate voltage differences appearing between exposed conductive parts of equipment and extraneous conductive parts through equipotential bonding.

Having an earthing conductor allows for automatic disconnection of supply in the event of a short circuit to earth fault by fault current a low impedance path sufficient enough to operate the circuit protective device.

Fault current always wants to find its way back to the origin (the transformer).

So to answer your question; earthing is actually a very complex part of any distribution system and forms an integral part of the protection devices by allowing them to function as they have been designed to. The earthing conductor doesn't get enough credit for what it does!!!


Power transformer with live and neutral wires

This power pole outside my house shows an advantage of having the neutral wire grounded. The live wire is located by itself at the highest and safest location, while the neutral wire is further down on the pole.


The point of earthing devices is that if (realistically when) there is a short to a part you can touch a circuit is made, current flows rapidly for a very short period of time, and then the over-current protection on the branch trips alerting an onlooker that there is a problem. Neutral is tied to ground to that the potential danger of a short can be sensed and protected against. I think a better example of the importance is a toaster that is shorting mains to it's chassis. If it's grounded the circuit breaker pops every time you plug in the toaster and you either fix it or get a new one. If the toaster isn't grounded mains potential is sitting there at the chassis of the toaster waiting for you to complete the circuit to earth (like touching the toaster with one hand and sink with the other). The second situation leaves you in significant danger. If the outlet is not protected with a GFCI you might see several amps flow through you for several ms before a traditional magnetic breaker trips. This is more than enough to do serious damage and/or kill depending on the path. If neutral is not tied to ground then there is no certainty that a short will trip the over-current protection.


protected by clabacchio Mar 21 '18 at 8:51

Thank you for your interest in this question. Because it has attracted low-quality or spam answers that had to be removed, posting an answer now requires 10 reputation on this site (the association bonus does not count).

Would you like to answer one of these unanswered questions instead?

Not the answer you're looking for? Browse other questions tagged or ask your own question.