Please do not mistake this for an attempt to question the point of code; I'm honestly curious about the history of failures that led to the code we've got. I am also very possibly not clear on some of the hazard cases.

Once Upon A Time, household wiring was two wires, hot and neutral. I assume, but don't know, that the neutrals were all bonded to earth at the fuse box, just as they are today in a primary breaker box.

So long as the earth ground remained connected, and the neutral wire remained intact, the neutral provided a much more attractive path for electrons than any person who might make contact with hot.

If the earth ground failed, then the person might become the most attractive path to earth. If the neutral broke, ditto.

Some bad things happen, and so we switch to running three wires everywhere. Hot, Neutral, and Ground. Ground, just like Neutral, is connected to a bar bonded to earth ground (and of course connected to the service neutral).

Now, if the neutral wire busts, we still have the ground as a path to earth. If both neutral and ground bust, we're back in the original dangerous situation. And if the earth ground becomes disconnected, we're in the same situation we were in back with two wires.

So my question is: what were the failure mode that were common and destructive enough to drive the change to ground wiring? Especially since so many devices continue to have two-wire cords?

  • I'd say the frequency is 60hz us, 50hz elsewhere. Except that's when it works, not when it fails :-) Jan 16 at 16:45

1 Answer 1


In certain main-panel setups, neutral and ground are simply landed on the same bus. As such, they seem like simply two redundant wires doing the same job, that are doubled up for some reason. That is an illusion.

Actually, the neutral system and the ground system are totally separate systems tasked with totally separate jobs. They act independently, passing like ships in the night. It's not wrong that they are connected in one location, but that is an intentional bond that has a specific job to do. In fact, it could be replaced with other types of bonding, at which point the roles would become more clear.

The neutral's job is to be the normal return path for service current. It connects from the transformer and supply neutral, via the neutral bus, to each of the points-of-use which require neutral.

Ground's job is to provide safety earthing. It originates at the Grounding Electrode System (ground rods/Ufer ground), connects to the ground bus in the panel, to the ground wires of each feeder and circuit. Note that ground is not supplied by the utility.

These are separate systems. Really.

Ground was added as a "safety shield" to protect people from current faults - not to give a more favorable path (current follows every path in proportion to its conductance), but to prevent fault current from getting anywhere near anyone in the first place. This meant that a piece of equipment could be rendered vastly safer simply by having a metal enclosure.

It was all part of a much greater effort to improve electrical safety. For instance it allowed UL to require 2-prong plug appliances to have double insulation. Since if double insulation was impracticable, the manufacturer had an alternative: metal chassis and ground the chassis.

Now, there are a couple of problems which must be solved.

  • Line and neutral will be 120V or 230V from each other, but we don't want them floating (as a group) at potentially thousands of volts above ground e.g. due to a transformer leak. So to prevent that kind of floating, we want to bond one of the wires reasonably close to earth - do so, and the others will follow. The very definition of "neutral" is "the wire we picked for this bonding".
  • Current flows in loops. Human-generated stray current wants to go back to source (not earth). If current leaks from a conductor to safety earthing, we want that to efficiently return to source - preferably with plenty of conductors (very low resistance) - that way a bolted fault to ground will flow enough current to trip the circuit breaker. Doing so is called "clearing the fault". This is why outbuildings need ground wires and not simply ground rods (and using the dirt to carry current back). This function is less urgent when "whole house RCDs/GFCIs" are present; they only need to flow 5-30 milliamps to trip, and dirt can handle that much current.

So we need a bond between the ground bar and the neutral bar which can flow a lot of current to clear faults, but keeps neutral voltage near earth voltage.

We could choose, for instance, a transformer with a 1-volt secondary. That would result in a 1-volt bias between neutral and ground. A sufficiently large transformer could flow 10 KA, enough to assure instant trip of any breaker.

Or, we could choose a car battery: creating a 12-volt DC bias (but 0 volt AC bias) between neutral and ground. The car battery handles 1000A routinely when starting an engine, so 10 KA of fault current for 8 milliseconds needed for instant trip won't faze the battery. It would work.

Let's cut costs, shall we?

However, there's an even cheaper solution: a piece of copper or aluminum wire. It's cheap, passive and creates a 0-volt bias between neutral and ground. I'm particularly fond of this one, because then, I can put a clamp meter around the wire! Of course it also feeds the myth that neutral and ground are the same thing.

Now then. Builders want to cut costs at every turn. They said "Hey NFPA. Why should we have to use two completely separate bars, connected by a jumper? Why not just use one bar? It's electrically the same." And, NFPA said "Yeah, we can't really argue against that." And so that's what they do.

That's all that is. It does not make them the same thing at all.

They do different jobs.

we're in the same situation we were in back with two wires.

Not at all.

If the neutral wire breaks, what's supposed to happen is nothing. The current can no longer complete the loop. The device stops working. The user calls an electrician, who finds the loose neutral wire and repairs it. Absolutely no current flows on ground. Ground is not an "alt neutral".

If the ground wire breaks, what's supposed to happen is nothing. The current still travels the live-neutral loop as its normal route. No power goes onto ground because nothing is supposed to put power on ground*. So when the ground breaks, nobody even notices. However, a layer of safety is lost: If the ground breaks, and then, a device has a "hot-ground fault", now safety ground is energized and no one is aware until an accident occurs, or a GFCI/RCD trips.

Electrical Codes are written with "defense in depth". They don't like it when an ordinary failure causes a hazardous condition. A neutral wire break is such an ordinary failure, and that's why 3-prong dryer/range connections (bootlegging ground to neutral) were banned in 1996.

* UL is allowing certain "smart switches" and "dimmers" to function without neutral, by leaking a very small amount of current onto the ground wire. The device must be designed so no conceivable failure could allow this current to exceed a milliamp or so, and the normal service current must not exceed 0.5 milliamps.

  • 1
    Thorough explanation, but OP asks how often certain faults occur Jan 16 at 16:05
  • @Phil "So my question is: what were the failure mode that were common and destructive enough to drive the change to ground wiring? Especially since so many devices continue to have two-wire cords?". That is a yes/no question. The title seems to conflict with that. I don't know what to do with that. Should this answer be deleted for lack of completeness? shrug Jan 16 at 18:36
  • As far as I can tell, you haven't provided answer to the historical angle of the question. For ordinary household outlets, with two wire connections to (e.g.) lightbulbs, what failure modes are eliminated by all those ground wires? You explained the reason for the bonded grounds in the panel. Not 'how often faults occur' but more 'what specific types of faults are we defending against.'
    – bmargulies
    Jan 16 at 19:28
  • p.s. I never wrote not thought that neutral and ground are 'the same thing.' I wrote there are two wires between each outlet and the Same Bar in my panel, and asked for a historically-informed look at what that improves over its predecessor.
    – bmargulies
    Jan 16 at 19:44
  • @bmargulies OK added a bit after the bolded sentence. Jan 16 at 23:53

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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