Here in Norway it is strictly forbidden to put up a grounded outlet without proper grounding, but if I remember right; the US NEC will not forbid to put up a grounded outlet without a ground wire if you have a GFCI.
Any body who know anything about this, or at least has any ideas?
406.4(D)(2)(c) permits a non-grounding receptacle to be replaced by a grounding type if fed thru a GFCI and clearly marked (after installation) "GFCI Protected" and "No Equipment Ground"
I always point out that you should verify there is no ground actually present before going this route. There was a period in the post WWII years where they were using an NM cable with a ground but still using non grounded receptacles (NEMA 1-15). In that case the box will be grounded but no provisions for a grounded plug cap. In that case you could use the existing ground for a NEMA 5-15 grounded receptacle. The only question is whether you should trust the "self grounding" type or pigtail a ground in from the box, I prefer the latter. The other issue is the ground in that old Romex is a reduced size but a 16ga wire should still be sufficient to clear a fault on a 15a breaker.
“Grounding” of US household electrical systems has changed quite a bit over the years, as have the common wiring methods. Pre-1960 wiring, depending on the wiring materials, often had a “ground path,” though there was no real intent to provide one. Growing up, I was actually present when the “electrical greats” debated various topics. I am sometimes amused when I see some creative soul today echo earlier ideas, thinking they came up with them. What is “grounding?” What is “grounding” supposed to accomplish? Why do we “ground” things? In my life I’ve seen major changes in theory as our understanding has improved. I’ve watched these ideas reflected in code requirements. Many of the stranger parts of our codes when it comes to grounding reflect some of these earlier beliefs.
Another puzzling aspect to our code is the recognition of two very similar wiring methods: Armored Cable and Metallic-clad Cable (AC and MC). A related method is “Flexible Metallic Conduit,: or “flex” (BX). Is the spiral metal sheath of these methods a reliable grounding path? The answer, in code, is “yes” for AC, “maybe” for flex, and “no” for MC. We can attribute the different methods to competing manufacturers. Does the metal make an effective ground path? Testing done on lightweight metal conduit (EMT) showed that even an extremely poor assembly of components would still carry a surprising amount of fault current. I guarantee that no electrician is capable of actually measuring the effectiveness of a ground path in the field.
In new work, today’s most common household wiring methods are either AC or non-metallic cable (NMC or “(Romex). Both of these require the use of a ground wire. When we add to, repair, or modify an older system that lacks a ground wire, we’re pretty much prevented from trying to add a ground. Earlier use of nearby plumbing is today defeated by the common use of plastic plumbing pipe. An old house without a ground wire will never have a ground wire. There’s no place for that third prong to go.
So how are we supposed to use modern appliances with three-prong plugs? Before we answer that, let’s ask “Why do we ground appliances?” The common answer “safety” falls flat when you realize it requires less than 1/10 amp to kill someone, and your smallest breaker / fuse is 15 amps. You’ll be well fried well before that overload ever trips.
As a compromise — a nod to reality — our rules allow such older circuits to be protected only by GFCI’s, and overlook the lack of a ground wire. The GFCI will operate in the event of a very small amount of electricity “getting lost.” This accommodation is allowed only for existing, older circuits. New circuits (complete circuits originating at the panel) must have that ground wire.
A final note: My house was built in 1957, before three-prong plugs existed. It was wired using a two-wire, cloth wrapped cable (Romex). Yet, every receptacle had ordinary 3-prong receptacles. Not a single GFCI to be found. Clearly, someone had dealt with the necessity of being able to plug in 3-prong plugs by simply changing the receptacles. The code change allowing the use of GFCI’s was in recognition of this need. The “compromise” was of limited use, as the device boxes used will be completely filled by a GFCI device, leaving no room for the necessary wires.
It must have depended on where you lived. My 1953 house had 2 prong NEMA 1-15 receptacles but there was a ground present. The Romex with the 16ga bare wire in it. I was just a kid who was learning about electricity when I pigtailed NEMA 5-15s everywhere and later when I got my first bug eye tester, they were all good. Later in life IBM gave us ECOS testers that did test the grounding integrity under load and verified <1 ohm ground impedance. It also wanted something greater than 0 ohms to catch bootleg grounds. The purpose of the ground is to create a very low voltage delta between the case of your appliance and the surrounding "earth" reference. That is only as good as the integrity of the grounding electrode and the grounding path in the branch circuit. It is also why NFPA finally decided in 1996 that we should not be using the neutral to ground things. There will be voltage drop in that neutral that shows up as a voltage rise at the appliance.
The major difference between AC and MC cable is the little bonding strip in the bundle that shunts out the wraps in the armor and prevents it from becoming a choke. That choke effect is also what limits FMC as a ground path to fairly short runs. I would still use a green wire ground tho, no matter how short the run is. The relatively new "MC Lite" has a much larger bonding conductor inside and is OK for grounding and you are not required to connect that bonding conductor in the box. I completely agree retrofitting just about anything but the normal 5-15 in a box can be challenging if they were sized to the normal "2x the largest conductor" deduction for the device. I am sure it has been proposed but I think the oversize devices should get a different fill specification. I ran into this trying to put USB receptacles in. One work around that doesn't look horrible is to use a surface raceway box over the regular box as a box extender. When I was working for the state one of my inspections was to test the integrity of WWII era EMT installed in a building they wanted to use as a hospital. They wanted to be sure that old EMT was sufficient for the redundant ground path. I was surprised but every one I tested with my Ecos had less than 1 ohm of impedance. I was glad I wasn't the one snaking in the new green wire tho.
One other thing the NEC allows [ 250.130(C) ]is running a independent ground wire, such as a #10 AWG green, back to the grounding electrode system. This can be a good option when the ampacity of the branch circuit is sufficient for the load and you just need a grounding type receptacle, such as a old 3 wire NM 10/3 cable for a dryer can be converted to a 4-wire NEMA 14-30.
Lots of voodoo around grounding. When dealing with receptacles and utilization equipment, the most important aspect is not that the metal is connected to the earth. The important thing is that it is connected to the service neutral so that an OCPD will open if a fault occurs. Being connected to earth is just a side effect because neutral is connected to earth.
Sometimes grounding is more harmful than good. Consider a generator insulated from the earth. It will just have a receptacle on it that has a ground terminal bonded to neutral. Suppose you plug in a metal power tool. If the tool is ungrounded, there is no chance for the user to get shocked unless you get in the circuit between the line conductors. But if the tool is grounded, you basically put one of the line conductors in your hand.
That's what I'm saying. You can't get shocked by a ground fault on an ungrounded tool plugged into an ungrounded generator. But if you use a three wire cord that has a EGC bonded to the neutral in the generator, you have now forced yourself into contact with a wire referenced to the neutral.
So what is the fault path, even if you touched the hot? (Assuming you were not touching the ground/neutral). You are sitting in the dirt, the generator is up in the air. There is no fault path touching any single conductor. That same theory is at work in ungrounded delta that you might see in a critical application. You also see it in the little plastic box generators like the 2kw Honda. They may not bond the neutral and they get away with it because the generator is insulated. The ground at that point is antenna. OSHA says you have to put the bonding strap back. Any metal frame generator like you see on the job site is assumed to be potentially grounded by incidental contact so they do bond the neutral. If you connect to a ground electrode, you are an SDS. Rubber tire generators are like that Honda but 250.34 still says they bond the neutral. Bonding the neutral really only causes the breaker to trip but if the generator is truly isolated a phase to tool handle short is not going to kill you. Just don't cut the ground pin off and use it on a grounded system,
I see that this thread has developed quite much. I hope we may aggree in that electricution is dangerous! If you are insulated from ground, or other poles in the system you should be pretty safe. If the Generator (or transformer) is ungrounded, and well isulated from ground, both wires N an L will have a capacitive connection to ground, and the wires are twisted, they will have approxematly the came capacitive value, even with a looooooooong wire. The voltage measured between the wires and ground will be close to the half of the value measured between the wires. The current that may be possible to get between a wire and ground may increase bye the amount of wires in the system, and at a point it will be dangerous to touch ground and a wire. A GFCI close to where you are will protect you pretty good. This is one of the risks with the Norwegian system. 230V and no Neutral.
Last edited by dsk; 03/31/2405:49 AM. Reason: spelling
The service outlets in IBM mainframe machines were derived from an ungrounded secondary so on a scope, the two ungrounded conductors were roughly 60v above ground but the EGC was connected normally. Touching either conductor might give you a slight tingle but barely even noticeable. Certainly not enough to operate a GFCI. In the 360 days and before, the design specs were that no neutral was even brought to computer room panels. All loads were line to line, either 240 or 208, depending on the service.(delta or wye). The insulated neutral bus was used for the isolated ground that was only tied to the GEC in the service disconnect enclosure. All of that IG foolishness went away in our design specs in the early 70s when the 370 line was introduced. The legend that it actually fixed anything went on long after that. It turned out the "noise" they were trying to isolate was actually coming from the switching power supplies in our machines, not elevators fluorescent ballasts or any of the other usual suspects. I am not sure we ever admitted it tho. Anyone with an AM radio figured it out pretty fast.
I'm not arguing against any of that. I think my point is being missed because I'm not good at explaining things. My point, ultimately, is grounding things doesn't necessarily make them safer, much as ungrounded things doesn't necessarily make them dangerous. I sometimes get calls from customers about home inspections that note ungrounded receptacles. As much as I would like to say just rewire the house, honesty wins. I told one recently that it be ideal to have new wiring, but the reality is that it would be really hard to get shocked on a carpeted floor inside a house.
Add to that most equipment doesn't have a ground pin anyway. I lived the first 15 year of my life with 2 pin (NEMA 1-50) receptacles and never really even noticed.
My point, ultimately, is grounding things doesn't necessarily make them safer, much as ungrounded things doesn't necessarily make them dangerous.
One of the problems grounding the metal frames of residential lighting and appliances was supposed to address is Residual Current Leakage (RCD) or 'ground fault' conditions where a metal housing could be elevated to a voltage where touching it with wet hands could result in more than the amount of 'let go' current running thru a person.
In Europe and Japan they have more types of grounding systems than Canada & the US. But no matter what system they use they all employ RCD detection in all the branch circuit breakers. Given the RCD thresholds being equal in both systems there may not be much difference between grounded and ungrounded systems.
With 'standard' US and Canadian breakers that have no RCD protection a grounded system would be safer, as that offers a default RCD level thru the equipment grounding wire only.
For example if you have say a 5ma/25/ms RCD protection (AKA GFCI) gounded and ungrounded systems are the same and I think a 2-prong NEMA 1- or 2- receptacles would be just as safe as 5- and 6-
Interestingly Siemens and Eaton both manufacture a 'GFPE' breaker that has a 30ma/2000ms trip level, per some obscure super outdated UL standard. They can be identified by their 'red' test buttons.
If UL or CSA updated the standard for breakers to the more modern IEC standards for RCD 30ma/25ms they might be a good option for 2-prong NEMA 1- and 2- ungrounded receptacles and 240V equipment. But for now in the US the only RCD is a 5ma/25/ms UL Class A GFCI, which seems to nuisance trip due to the low 5ma trip level for wet locations.
However since per the NEC article 110.7, no proper functioning circuit should have "short circuits, ground faults, or any connections to ground other than as required or permitted elsewhere in this Code." thus not more the 30ma flowing on the Equipment ground (Earth Leakage) I am not sure why this GFPE breaker has not replaced the standard breaker (non GFCI ) where a standard breaker is permitted. At least in dry locations or fixed equipment in dwelling units.
I am familiar with the RCD systems and when NFPA gets done, I think our solution will be better but more expensive. Between the GFCI and AFCI that incorporates GFPE protection, every circuit will be protected. The main advantage is, when you get a ground fault, the whole house doesn't go black. That might be an Easter Egg hunt most of us would like to avoid. In most cases you will be in eyesight of a GFCI you tripped..
