3
\$\begingroup\$

I know this is a question with variants that have been asked countless times on the internet and on this site. Believe me that I have scoured through these questions and answers in search for a satisfactory answer but that I have not found one.

In your home electrical wiring, the rule appears to be that you should never connect ground to neutral anywhere else than at the main panel. The reason for that (from countless individuals/websites online) is that once you do that, you would be a using a device's ground and neutral returns in parallel, and once the ground in a system is energized, this effectively energizes all other objects connected to ground in the system.

But given that ground and neutral are already connected together at the main panel, doesn't that do exactly what is meant to be prevented with this rule? Once you have conductors (ground/neutral wires) at equal potential somewhere in the system, how does connecting them elsewhere change their relative potential?

I know that my thinking is wrong at some point, since clearly every chassis in every home isn't energized. I just don't know where my understanding goes wrong.

\$\endgroup\$
3
\$\begingroup\$

In a perfect world with conductors that have zero resistance, neutral would be always at zero volts everywhere, with respect to ground. It has current flowing through it, but that's not the same as having a nonzero voltage, which I take to be what you mean by "energized".

In this world, that's not totally true, but the point where neutral and ground are connected together is ALSO the point where there is a physical grounding rod driven into the earth. So even if some part(s) of neutral are significantly above 0V, due to high resistance, or a fault, or whatever -- the point where neutral and ground are connected should still be very close to 0V, which means that the rest of the ground wire should still be very close to 0V, as long as you don't connect it to neutral anywhere else.

If you connect neutral to ground somewhere else, away from the grounding rod, this no longer holds -- in the event that neutral were to break, current could flow through the ground wire near the additional connection between ground and neutral.

There are different systems used in other countries and situations, and the best way to set up a protective grounding system depends on what assumptions you make about failure modes. In a perfect world where the neutral conductor has zero resistance and never breaks, you wouldn't need a separate protective ground wire at all. So ultimately you need a model of what kinds of failures you expect, in order to decide what kind of protection will be most useful.

\$\endgroup\$
  • 1
    \$\begingroup\$ Thanks. I think my idealized mental model of the system included zero-resistance wires. I think my understanding will come from considering the system with gradients of voltage. \$\endgroup\$ – user2686410 May 23 at 19:33
2
\$\begingroup\$

Nowadays we have RCDs or GFCIs in the home to prevent serious electrocution should the live wire inside a piece of equipment come into contact with the chassis connection. If neutral were also bonded to the earth wire locally on the equipment, what prevents a lot of fault current flowing back through the neutral wire and convincing the RCD that “all was well” when it wasn't.

So, advocating ground and neutral being connected at equipment may not make a serious change to the potential on the neutral wire but it will divert earth current away from the earth wire and potentially render the RCD useless.

Also, earth wires are generally smaller cross section and can’t handle long term neutral current should the neutral wire fail. So, if neutral failed and full load normal current passed down the significantly lesser rated earth wire, the fuse would no longer protect the cable and a fire could result.

\$\endgroup\$
  • \$\begingroup\$ Are RCDs/GFCIs the only reason why you shouldn't connect neutral/ground at the device? I am surprised since your reason seems fine but is almost never brought up elsewhere. \$\endgroup\$ – user2686410 May 23 at 17:56
  • \$\begingroup\$ See amended answer. \$\endgroup\$ – Andy aka May 24 at 7:20
1
\$\begingroup\$

Ground is only for safety and EMI reduction and there is no good reason to do what you suggest, only bad ones due to fault detection/protection. So there is no point discussing it if you understand how it works by unbalanced currents on Line-Neutral.

Sharing neutral current with ground would make Neutral current with Line Unbalanced and thus false-trip the RCD/GFCI.

\$\endgroup\$
  • 1
    \$\begingroup\$ I am not suggesting for anyone to do anything. I am attempting to understand how it works. \$\endgroup\$ – user2686410 May 23 at 18:50
  • 1
    \$\begingroup\$ Understood. But your assumption was wrong. If it's not done, there must be a good reason. The reason is fault detection could fail. Search here for more info. electronics.stackexchange.com/search?q=gfci+rcd \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 May 23 at 19:10
1
\$\begingroup\$

A house is served by wires from a transformer, AND sits on ground that may conduct electricity from... other buildings, or solar-wind-driven currents induced in the subsoil.

Because parts of the house (reinforcing rod in concrete, plumbing pipes, metal sewage pipes, etc.) are conductive AND in contact with the ground, a random electrical cable or appliance may come into contact with grounded metal.

In order that the ground-to-wire voltage be kept in some safe limits, the transformer center-tap (which is at the midpoint of the AC voltage range) is connected to ground. This is done where it can be inspected, in the main electrical panel.

What that ground-neutral bonding prevents is the buildup of voltage in the house wiring that can exceed safe insulation ratings to the 'grounded' bits of metal that are omnipresent. Such events as lightning strikes will occasionally change the under-this-house potential to be different from the potential under a utility pole or another house; a one-ground-per-building rule is essential.

Finally, only currents INSIDE THE HOUSE can create a voltage drop between neutral and earth-ground, by resistive effects in the current-carrying wires, so your home is safe even if your neighbor has a major wiring fault.

Once you have conductors (ground/neutral wires) at equal potential somewhere in the system, how does connecting them elsewhere change their relative potential?

It isn't about the 'relative potential', it's about inspection of outlets; the test for grounding, and for power-present, could be fooled by wiring practices that make no safe ground connection to the ground pin.

\$\endgroup\$
-4
\$\begingroup\$

Neutral cable is filtered and rectified, it goes throug capacitors and then joins the chassis which is earthed. in some particular situations neutral may hopefully be safely grounded but try not exceeding 70 volts 200mA to ground or may find fusing trouble if phases suddenly changes. Essentially nothing is really neutral grounded for safety reasons but it may vary depending your region and phase changes, live wire induction and potential differential to earthing. hope that helps

\$\endgroup\$
  • 3
    \$\begingroup\$ "Neutral cable is filtered and rectified ..." In general it isn't. It is on devices using switched mode power supplies. "... it goes throug capacitors and then joins the chassis which is earthed." Neutral doesn't go through the capacitors to earth. High-frequency noise on the neutral wire is shunted to earth. "in some particular situations neutral may hopefully be safely grounded ..." No. Neutralising is done at the supply transformer or in the building supply panel, never in the equipment. "... try not exceeding 70 volts 200mA to ground ..." This is non-sensical. \$\endgroup\$ – Transistor Jun 29 at 18:26
  • 3
    \$\begingroup\$ I'm afraid that you don't understand the subject well enough to be giving advice or teaching others. Time for more study. Welcome to EE.SE. \$\endgroup\$ – Transistor Jun 29 at 18:27

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.