Grounding boxes in home wiring are totally safe for you?

[msuffidy]

I always assumed grounded metal boxes would save you because all the electrons would go down the wire. So I was watching a video about what is ground and it eventually got into the topic of using grounding boxes as safety for persons. It seems to me that it relies on the idea that a wire is supposed to have some degree of resistance as part of the proof. In the video, it sort of suggests the wire segment before the box is a resistor, thus the potential at the box is lower than mains. I am a bit worried about how treating the wire as a resistor has such drastic implications. I am a bit confused about how it actually works in reality, but suspect in that case the potential at the box is in fact lower than the mains potential. I drew this diagram trying to understand it without the idea the wire had resistance and came up with treating the box path as a parallel near 0 ohms resistor. Here is the video:

[msuffidy]

The answer may be like.

1) The length of the wire from the station is much higher than the grounding side thus the potential is lower at the box
2) The amperage of the near zero resistance ground path may reach the amperage limits of the supply wire and deplete the usually accepted 2nd (human) path supply.
3) You may be briefly shocked but the simultaneous ground path will ensure the trip of the circuit breaker.

Actually, theoretically in this scenario this would only happen if you were touching the box when something live hit the inside of the box, or a breaker failed to trip. In the case of the stuck breaker you would probably also be worried about an electrical fire in the wall.

[Berni]

This is why bad connections on earth wires are very bad.

Yes earth wires have a resistance to them like any other wire, so an earthed metal housing will still rise in voltage if live shorts to it. But due to the low resistances of wires and high voltages involved this will send a massive current down the earth wire(100 to 1000A). So within the next milliseconds the GFCI will trip and the fuse for that live wire will trip. This stops the current and the metal housing is at earth potential again.

If you ware holding the metal box and something else (like a metal pipe) when such a hard short happened you would likely get a brief jolt, but you would not be dead.

But if you did that with no earth connection to the box you would effectively holding a box at 110/220V and the current would flow trough your body until you let go of it. Hopefully you have a GFCI to trip and save you there and you might end up dead

[msuffidy]

I have been shocked by this and that over the years, but not at the ozzy 240. Just our 120 in Canada. I finished a basement once and did some stuff with the breakers still on without turning them off and either got straight up shocked or like rotated a wall light and it made a big spark. Around 2003 when I was renting, I got a shock when touching some appliances and found out one ot the outlets had something backwards, neutral and hot and fixed it myself. I am going to try to be more careful in the future.

[Siwastaja]

This "protective earthing for personal safety" actually has well established science and math behind it, but it's old, and in practice never worked too well.

Let's start with the terminology, get rid of "ground" as it can mean anything and it's used arbitrarily. When it comes to safety, we have no room for such inaccuracies. Either you have protective earth (PE) or you don't. PE is sacred, mixing it up with anything else is a crime worth of capital punishment. This is a mistake you just can't do.

Anyway, the old idea revolves around the short circuit current and fusing. No, it does not depend on wires having resistance, quite the opposite, it relies on the wires having small enough resistance so that the fuse blows in time. The short circuit current is actually measured in new installations as a part of acceptance, and if you are behind a long line, too small short circuit current prevents you having larger fuse sized. The idea is that if you have a short from L to PE, current will be so high that fuse blows in short enough time (tens of milliseconds) to prevent being exposed for the dangerous voltage for too long. Needless to say, low impedance of the PE is paramount. Which is also the Achilles heel of the system. Similarly, poor paths you can do nothing about (like the actual physical earth) pose an issue.

But yes, in theory, if you have a properly earthed metal box, and live wire comes loose inside and touches the box, you don't need RCD to save you. If the live wire touches poorly, with a lot of resistance, the resistive divider causes the exposed voltage to be small, and it doesn't matter if it takes longer for the fuse to blow. OTOH, if the live wire gets a good contant, causing significant voltage on the box, then the short circuit current is also large, and fuse blows so fast that even if you are touching the box while this happens, you won't die. But as you can see, this all relies on assumptions about wiring impedances and everything being done exactly right.

Now starting in 1990's, RCDs have all but saved this iffy paradigm: now any half-decent PE impedance will do, and live wire resistance also becomes irrelevant as you don't need hundreds of A of fault current to blow the fuse, tens of milliamps is ample to trig the RCD in equally short times which required 1000x more current before it!

For this reason, if safety is of concern at all, I strongly recommend you upgrade old systems with RCDs, even if legislation requires them only for new circuits.

[JohanH]

For this reason, if safety is of concern at all, I strongly recommend you upgrade old systems with RCDs, even if legislation requires them only for new circuits.

Some older sections of our house don't have RCDs (GFCI for American friends). I've added one for my work room/home lab. Highly recommended to retrofit if possible. There are also dapters to put in the outlet if you don't have the possibility to upgrade the whole system.

[ajb]

The short circuit current is actually measured in new installations as a part of acceptance, and if you are behind a long line, too small short circuit current prevents you having larger fuse sized. The idea is that if you have a short from L to PE, current will be so high that fuse blows in short enough time (tens of milliseconds) to prevent being exposed for the dangerous voltage for too long.

And the flipside of this is that with a very low total impedance providing a very high short circuit current ("Available Fault Current" is the term now used in the US) you need overcurrent protection devices with a higher interruption rating, and all of the equipment in the circuit must be capable of withstanding that current for the time it takes the breaker to trip.  Even standard residential circuit breakers will have interruption ratings given in tens of thousands of amps for this reason, in industrial settings this is even more important due to the higher powers available to begin with. 

Going back to that first bit: "The short circuit current is actually measured in new installations as a part of acceptance", does that include residential settings?  I don't think that's typical for residential installations anywhere in the US (except maybe apartment blocks or the like), but then our lower voltages mean the power to your house always comes from a relatively small transformer somewhere nearby, which I'm sure simplifies the analysis. 

[Siwastaja]

Yeah, here the available fault current needs to be measured as part of acceptance even just in residential settings (during commissioning, of course, no need to retroactively check old installations) but here everything involves bureaucracy and as a result, costs arm and leg, and this doesn't mean the paperwork actually matches reality, no one cares as long as the crimes are committed by someone who has the right papers to do the job (i.e., membership of the club of those who are allowed to do anything with no risk of consequences). But the worst thing you can do is to make a correct installation without the papers.

As you say, this fault current thinking leads into expensive oversizing of everything, but that isn't a bad thing necessarily, with the addition of RCDs the system works actually very well.

Here in our TN-C-S three-phase distribution system one of the bigger risks is the failure of the Neutral wire. There are no proper safeguards against this, and hence we have these exciting procedures such as: "if you see some of your lights being dim and some very bright, walk carefully not touching anything, get a nonconductive object and use it to turn off the power from main distribution box." How many actually know this? This happens because at disconnection of PEN, unbalanced loads connected to phases work in series to shift the floating PEN node, which is now different from the physical Earth. But because PE and N are connected together, every grounded product turns into a deathtrap! The only way to deal with this is to try make wiring reliable. Which obviously doesn't work out in countryside and air wires. So we have these failures, and people usually do not die because of luck, and we try not to think about it too much.

[themadhippy]

The short circuit current is actually measured in new installations as a part of acceptance", does that include residential settings?

In the uk  its required although we call it prospective fault current,theirs two version. prospective short circuit current,either live to neutral or phase to phase  and prospective earth fault current.The highest value is the one you record.

How many actually know this?

in the uk more than did pre 2006 when some clever bureaucrat decided it was a good idea to change the wiring colours and swap the colours of phase 3 and neutral around

[KaneTW]

Of course sometimes you cannot install RCDs (such as with 3L-PE filtered VFDs), but you also have to deal with fun things such as high earth leakage currents and making sure your PE is rock solid (usually by having two separate PE connections or a >16mm2 PE)...

[Siwastaja]

Do remember that even a 200mA RCD is a huge improvement over no RCD at all. It is not officially for protecting from shock, but if you compare it to only relying on short circuit currents to protect from shock, it's still an improvement.

[jonpaul]

Hello: A few points: Commercial and industrial equipment is tested for resistance from chassis to mains cable earth at plug, with some hy-pot and grounds resistance testers. UL, TUV, VDE, ETL etc all require the resistance testing at high currents eg 25 A.
Depending on the regulatory environment and equipment rating the acceptable resistance varies.

In power distribution and transmission, every tower, pole , protector and transformer has an earch ground via rods or a grid.

A major risk it dry soli so the resistance increases. IEEE and ERPI research proves correlation between lightning damage and ground circuit resistance.

In RF engineering I learned long ago the proverb: "There is NO such thing as a ground"

Bon chance


Jon

[KaneTW]

Yeah, you'd have to use a 300mA Type-B RCD. But does this provide an actual increase in safety in a properly designed system? It's not enough to protect a person from shock, so it's only useful as fire protection. For non-critical use cases I don't think it's worth it (and the VDE agrees), as it'd require an extremely low probability PE failure to trigger the RCD but not the circuit breaker.

[Berni]

Yeah, you'd have to use a 300mA Type-B RCD. But does this provide an actual increase in safety in a properly designed system? It's not enough to protect a person from shock, so it's only useful as fire protection. For non-critical use cases I don't think it's worth it (and the VDE agrees), as it'd require an extremely low probability PE failure to trigger the RCD but not the circuit breaker.

Here it is standard to have a low sensitivity 300mA RCD for the entire house and a separate high sensitivity RCD for the bathroom. You can have issues with false trips if you try to use a high sensitivity RCD for the whole house because the leakage currents can add up.

The 300mA RCDs are quite high when it comes to shock indeed, but can still save you in situations where you grab something live while having a good contact with something else metal. Even the high sensitivity RCDs will give you a hefty shock before the switch has time to react and break the power. But this does not make them useless. The 300mA RCDs do trip quite easily when mains stuff gets wet. Like an outdoor fixture getting water inside it, someone pouring liquid on a extension cord etc.. it also does not need a good low resistance earth path to trip a breaker. Without a RCD something developing a high resistance path to earth will generally burn more and more until the path gets carbonised enough to let a serious current flow, often setting things on fire as the carbon path starts to glow. If you are lucky the flame might form a arc to live in order to make enough current flow to trip the breaker quickly enough before everything is on fire.

So it is very good practice to have a RCD in some form.

[Siwastaja]

Here it's 30mA for the whole house. (Well, nothing prevents you from using multiple, with smaller chance for false positives, but guess how many electricians do that, these things cost like 30€! Professional work being expensive does not mean it shows in the results. )

Refrigerators and freezers can be legally wired to separate outlets without RCDs. If not, false positives would be really awkward.

The American way of multiple 6mA circuits offers even better safety, although that being said, 30mA is pretty good at preventing death. 300mA - not so much, and it's not meant for that, but if you have to choose between taking the shock without RCD, or with a 300mA RCD, the choice is obvious!

[KaneTW]

Oh, I'm not against RCDs in general. We have 30mA RCDs on every subcircuit, and they did in fact detect a medium-impedance short to ground (Riso on the order of 10kohm).

Just arguing that in some applications an RCD is more trouble than it's worth. They're limited, and nothing else should be on that subcircuit, but still.

[Jester]

A small note on ground resistance. Years ago I was part of a team that were developing a 240V power meter for utilities that was to be installed at the pole mounted transformer. We designed the meter with a suitable HRC fuse. The company was sold to investors and they were looking to cut manufacturing costs and suggested we replace the fuse with a narrow trace on the PCB. Their “engineer” trained who knows where explained to us that if a fault occurred the trace would simply melt and everything would be fine. We disagreed. In the end the investors agreed to run a test. Fortunately we had access to a high voltage, high current testing facility and they agreed to run the test. We placed a wire wrap style wire just down circuit from the narrow trace. The testing facility records these test on video. The device went off a bit like a hand grenade. We repeated the test with the HRC fuse and and it was a non event.

Their “engineer” did not seem to grasp the concept that the considerable fault current available at the transformer terminals with very little resistance ( no long line from the transformer to the load) makes a rather significant difference in the actual fault current. That’s why for example a 1A HRC fuses has 200kA breaking capability.

Their “engineer” also seemed to miss the point that yes the trace will melt, however the arc will not be quenched and all this will occur VERY quickly as in explosive quickly.

[Siwastaja]

A small note on ground resistance. Years ago I was part of a team that were developing a 240V power meter for utilities that was to be installed at the pole mounted transformer. We designed the meter with a suitable HRC fuse. The company was sold to investors and they were looking to cut manufacturing costs and suggested we replace the fuse with a narrow trace on the PCB. Their “engineer” trained who knows where explained to us that if a fault occurred the trace would simply melt and everything would be fine. We disagreed. In the end the investors agreed to run a test. Fortunately we had access to a high voltage, high current testing facility and they agreed to run the test. We placed a wire wrap style wire just down circuit from the narrow trace. The testing facility records these test on video. The device went off a bit like a hand grenade. We repeated the test with the HRC fuse and and it was a non event.

I have seen disastrous management like this. When it's a con "engineer", it's even worse. That's exactly why we have regulations and standards - you don't need to waste time demonstrating that illegal design is dangerous . That's why it's illegal, end of story. Even if the end result of your story was good, a lot of time was wasted when all that was needed is to management believe in their own experts (the actual engineers) saying this is against the standards and dangerous.

[Berni]

Yeah this story is the reason why we have standards and laws that force you to follow them.

There are times where engineers explain the proper way of doing something, but the management sees the whole thing as just unnecessary cost. Particularly if there are people in management that know a little bit of the field, as that gives them confidence to push some of the stupid ideas, yet not know enough of the field to realize it is a bad idea.

[Siwastaja]

Also why protective laws for whistleblowers exist. Not here of course, but in many countries which value things like that.

[Shock]

The video was mainly to illustrate that most of the current flow is through the path with the least resistance, but current still flows in parallel paths. The 0.001 path is basically a dead short so when the guy grabs onto the grounded chassis and is grounded through his feet or whatever, just like if measuring down the live wire in a functioning circuit he would see very little voltage drop.

I think he made the 1ohm wire to match his 1k resistance and make the math look good. But in reality he won't be standing on a perfect ground either so his resistance could be even higher. His appliance will likely have a fuse and the mains or rcd will trip, but it doesn't mean he gets away without feeling it.

Edited for clarity.

[Shock]

I re-watched RSDs first video and it reminded me why I never watched them fully, he flipped to electron flow mid video and started increasing negative voltage, it got a little squirrely afterwards with terms like "backed up voltage". At that point I had an aneurysm and was forced to close the video.

I'll admit though tackling it unscripted is hard work.

[mag_therm]

Jonpaul said: "There is NO such thing as a ground"

I agree,also suggest careful calculations before using "Earth Ground Rods " for safety grounds.
( I saw a lot of discussion about "good" ground rods on internet forums, in my opinion, usually doubtful)

Refer to Dwight "Electrical Coils and Conductors" , 1945,  Chapter 10, where he did a method for ground earth rod resistance based on capacitor analogy,  which is still used these days, re-published with ISO units.
 Similar calculations can be done in FEM. (I did some)
Chapter 10 Eqn 15 for the resistance between two vertical ground rods is given in the old CGS (physics) units:
In ISO units following the conversion is by  this reference:
https://blog.nvent.com/calculating-on-ground-electrode-resistance-of-a-single-rod-ground-electrode-design-principles-and-testing/

R = ( rho / 2*pi* L) * [ln (4*L/ radius) -1 ]
where L is driven rod length [metre]
 L is also the distance apart of two rods , between which the resistance is calculated
radius is of the rod [metre]
rho is soil resistivity  [ Ohm/metre], reciprocal of the more common conductivity values

Example for Michigan USA where the soil rarely dries , using a map of national measurements done in 1975:
 https://www.fcc.gov/media/radio/m3-ground-conductivity-map
conductivity : 8 milli Siemen/metre   : rho = 125 Ohm.metre

L is the driven rod length, and there are 2 of them at distance L apart, L = 2 metre

substituting:
R = 38 Ohm That is the DC resistance between the two rods. 50/60 Hz resistance is similar.

In residential and industrial locations , the distance between the utility PCC will be greater than mentioned above.
But optimistically assuming 2 metre, the short circuit currents through the protection device via "Earth Ground Rods"
for a " 1 phase Bolted Fault" can be calculated:

120 Volt : Isc = 3.2 Amp
240 Volt Isc = 6.4 Amp
4160 Volt single phase ( common in older USA cities) : Isc = 109 Amp

13000 Volt : Isc = 342 Amp

The above is why, in my opinion, there must always be a properly sized metallic earth/ground conductor between PCC source and point of ground fault.

[Jester]

Around 1980 I worked for Manitoba Hydro, they generate most of the power for the province about 500 miles north of the border (Nelson river) where they convert it to DC and then send it South along two conductors instead of three. Conversion station near Winnipeg than converts it back to conventional three phase ac.  I think they generate about 6000MW now. Normally it operates bipole, however they can take one side down and use a single conductor plus ground. Apparently the ground path has much lower resistance than the actual conductor, if memory serves me ~0.1 Ohm vs. 5 Ohms? They have an array of ground rods at either end. Now that’s a serious ground!

[Berni]

Dang that must be some serious grounding there to use as a conductor.

I am guessing the one line operation is for backup or maintenance. Since id imagine sending DC trough the ground rods would not make them happy in the long term. The lucky side would have some nice corrosion protection from the DC current while the other side gets eaten away by corrosion at an accelerated rate.