Ground fault switch aka residual-current device

Yes. It doesn't make sense that a safety protection circuit self-reset. Thermal protection, no proplem. Overload, no problem. Stuck motion, no problem. Tingle/fault current - how would the coroner know what killed you . . . . or the next guy who wandered by?

RL

Some basics from a US perspective:

Ground Fault Circuit Breakers are not what you linked to in your first post. A Ground Fault Circuit Breaker will trip either on an overload, or on a ground fault - that is leakage other than from Hot to Neutral. A ground fault Circuit Breaker must go through the OFF position before it can be reset. These are typically found in the main panel, not as part of a receptacle. What you pictured is a Ground Fault interrupter - that is, its primary purpose is to detect leakage, and break the circuit on that condition. These devices, typically, are not as sensitive to overload as a breaker, and although they may be "dead-front" devices, they do not kill the entire circuit, just what feeds from it. So:

A device such as you linked will 'look for' leakage in what is connected to it. A faulty device or similar. And, it will then trip. But, within the box, there is still power. So, these devices are not suitable for locations where they may be exposed to water or similar, because the circuit will still be "hot" to the box.

A panel-mounted ground-fault circuit breaker will kill the entire circuit, not just *that* location. Far safer. Example: Our hot-tubs are protected by a remote safety switch with a 50A, double-pole Ground Fault Circuit Breaker - so no part of the feed remains hot (electrified) if there is a trip.

Now, consider the situation of a self-resetting device in a residential situation. From what you are describing, that device 'protects' the entire residence, not just an individual circuit: "In Spain, every house must have at least one protecting the entire house. Typically 30 mA sensitivity."

There is a trip - you go ahead and try to diagnose why there is a trip. In the midst of this, power is re-applied without warning. Crispy Critters, perhaps? Keep in mind that even the very best device will not protect you if you insert yourself into the circuit - that is, if the current is going through you to the 'correct' neutral. Such a device would be madness. Consider such a device in a non-residential setting: Where I work, we purchase primary power (13,200 volts) and are dual-fed from two primary feeds into what is called a "tie-breaker" device. Both sides feed everything - but should there be a fault on one side, the tie-breaker automatically breaks the dual-feed, isolating the fault side. ALL OF THESE DEVICES MUST BE RESET MANUALLY. And when/if this happens, the person resetting wears a 'moon suit' -

Yeah, I would be unhappy if there were a self-resetting device of unknown behavior in my system.

Peter Wieck Melrose Park, PA

Carlos refers to the device in the main breaker panel like this one:

I know some people want them because if there is a false trip when you are away for some weeks then you have to throw away your frozen food and maybe the freezer if you can't remove the bad smell. Maybe there should be a special circuit for the fridge/freezer using a rearming GFCB and the rest a normal one.

I find these false triggers are more common with newer GFCB, maybe they react too fast to any transient or spike.

Good point on the "pointer". A main-panel device is the only way to go for total-circuit safety.

Ground-fault devices seldom (very seldom, indeed) false-trip. Sure, they may be acutely sensitive to transient issues, which the typical end-user will perceive as a false-trip, but the reality is that *something* tripped it. And that *something* may be fatal if not understood and corrected. I have come across several conditions where an external receptacle got a salt-track from the hot to ground and presented apparent nuisance-trips - but, in fact, there was a problem. One had to look hard enough.

Rearming to save a freezer load of meat. Silly.

Peter Wieck Melrose Park, PA

Yes, that is the case.

If I were inside the house, I would disable the auto arming. Of course, if the fail condition remains active, it does not rearm. It is the same strategy as a person does: the thing triggers, we go to the panel and try to arm it, maybe several times.

Next step is to switch off circuits, then arm the RCD, till the problematic circuit is found.

Then we go round the house removing whatever is connected to that circuit.

Finally, if the culprit is not found, call an electrician.

There have been false triggers in that house since the installation was new.

There was a period, the first year, when they happened every day at a precise time, 6:30 AM I think it was. Investigating, the electrician found out that the substation switched the main transformer up or down a notch at that time.

What he found then was that there was a lamp that was connected to one current limiter on (L) and to another for the (N) line. That was corrected and the regular triggers disappeared. There remained spurious triggers, like once a year, which tend to happen when one is travelling far (Murphy)

Sure one day you will forget to disable the rearming.

The only way to know if the condition remains is to reapply power, so it should rearm and trip again. Not sure if it will be so fast tripping as it was the first time. If the leak was you on the bath this is not a good idea.

Asymetric wiring is a good way to trigger the newer super-fast protectors. I know of some cases of false-trigger, they are hard to diagnose because it happens once a year. There are various opinions on the cause.

Nothing electric in the bath.

Same problem as if I am in the bath, somebody else in the kitchen, the RCD triggers, the other person resets it manually and immediately. It will not hold.

That was around 1985...

L N | | | | RCD | | | | | | | | protector | +---/ o----------------------+--- | | . | +-- | --/ o---- L | | | | | | 0 lamp | | | | | / switch | | protector | | +---/ o------ N | | | . | +-- | --/ o----------------------+-- | |

This is certainly bad wiring, but I fail to see why it would make the RCD trigger every day at 6:30, when there was a voltage surge from the transformer station. The electrician that found the problem was baffled.

Bad idea. Water pipes and water heaters are earthed together with the rest of appliances, if any has a short to ground it will be dumping the full mains voltage to the water pipes. If all is well this will trip the protector fast enough to save your life assuming you was taking a bath. Think about it, you really want to keep trying to see if it keeps saving your life again and again?? Don't forget the protector could eventually fail as well.

Also keep in mind that when "it will not hold", it is connecting the power back for some milliseconds. It has no other way to detect a leak than to apply power and check. I would not keep trying if someone is in the bath.

My theory is that asymmetric wiring will behave like an antenna, a magnetic pulse for example from lighting, strong interference or heavy machinery nearby will induce a current in one wire different than the other. Also the opposite, all transients and spikes coming from mains will be radiated more from one wire than the other resulting in an imbalance. Not to mention different stray capacitance and inductance caracteristics from the different wiring paths affecting the response to fast transients.

Once more, an observation from a US perspective.

The basics: US residential systems are, typically 120/240 volt, single phase systems with two "Hots", a neutral and a ground. Hot-to-Hot is 240 volts @ 60 HZ ac. Hot-to-Neutral is 120 Volts AC. Ideally, Neutral and Ground are at the same potential, and, again, typically, are bonded together in the main panel. The main panel is also required to have a separate ground. The main service to the utility transformer is three-wire, typically. Just the hots and the neutral. However the Utility transformer is grounded as well. As our house has a copper water service to the street, the main panel is also tied to the water service, and the water meter is bonded (ground wire bridging the water meter) which grounds the copper plumbing throughout the house. We have an hydronic heating system (black iron) and a gas stove and dryer - also fed with black iron. Both those systems are bonded to the copper plumbing. Meaning pretty much all the piping in the house is grounded, but for a few peripheral odds and ends that are fed in PEX due to upgrades and/or repairs over the years.

The first thing I did when we moved in was "balance" the panel. This with a clamp-on ammeter. So, under most conditions, there is no asymmetrical loading. 240 Volt loads are always balanced. What this means is that if we get repeat trips on a ground-fault breaker, we know we have a problem to be checked and corrected before reapplying power. We also know that any 'leak' into the plumbing will go directly to ground.

Now, differential issues. One of the discovered weaknesses in US-code wiring early on was the multiple-devices issue. Say, there are eight receptacles on a single circuit, wired in parallel, of course. The last device in the chain is after multiple splices or connections. Each one adding some tiny amount of resistance. So, there is some potential between the ground of the first device and the last. If the original electrician was sloppy, that potential can be sufficient to trip a ground-fault device. And, the end-user will perceive this as a nuisance-trip, not a problem. For the record, it is a problem and should be corrected. Not sure how things are done in Europe, but if conditions are similar, look for this as well.

Peter Wieck Melrose Park, PA

Here in Spain most last mile distribution is 230/400V at 50Hz with four wires running through the streets, one neutral and three hots (triphasic). Hot-to-neutral is 230V and hot-to-hot 400V. Most homes are connected to only one hot and neutral (230V monophasic) and it is up to the electricity company to distribute homes through their three hots to balance their systems. Inside home there is nothing to balance. In residential areas triphasic is only used for older single or dual speed lift motors and heavy air conditioners.

In older areas there are still some 115/230V networks, for example where I live. 115V was used maybe 50 years ago but it is very unlikely anyone is connected to that anymore, if someone keeps such vintage appliance there are plenity of 230-115 transformers available cheap second hand. In these areas two hots enter homes, the neutral is not even connected at the street wires (when it is, it was for 115V and currently unused).

I don't understand how that small voltage drop in the cabling causes an imbalance in current. The earth in all appliances is supposed to be fully isolated from the hot/neutral isn't it?

Not here, no.

A sensitive GF device will detect a difference between flow to the neutral and the ground, and perceive this as current going to ground.

The ground and neutral are bonded within the main panel, and there is a secondary ground feeding the main panel.

So, the full path of each circuit Neutral is from each device to the panel, where it is bonded to a local ground. From there to the Utility Transformer - which is also grounded in its location.

The full path of each circuit Ground is from each device to the panel, where it is bonded to the neutral and a local ground. It stops at the local ground.

In theory, the Neutral and the Ground should be at equal potential - thereby avoiding false trips - as that is what the GF device is looking for - current going to Ground (or somewhere), not Neutral. If the Ground and Neutral are not at equal potential - there may be something for the GF device to detect.

Peter Wieck Melrose Park, PA

Thanks for explaining. I seem to understand that the GF built in your wall plugs (we do not have them here) trips because the small voltage between neutral and ground. That does not look like the right thing it should trip on. Even if the installation has few interconnects if you plug a heavy load the drop will develop anyway.

Here in 230/400V areas there is no voltage between neutral and ground (they could also be joined at the panel) but if you short them in a plug the GFCB trips, this is because the small voltage drop developed in the cabling. Shorting them creates a parallel path and current will distribute proportionally, some current will flow through earth instead of neutral.

..............................^^^^^^^.... Not voltage. GFCI's trip on a difference in _current_ between neutral and line.

Current normally goes through the GFCI like this:

Black--------GFCI-------->--load-+ White--------GFCI--------<-------+

Current goes through the GFCI from the black (hot) wire through the load and returns through the GFCI to the white (neutral) wire. The currents (out to the load and back from the load) are compared in the GFCI and if they are the same (within about 5ma) all is well and the GFCI does not trip.

Now look at a problem where the neutral is shorted to ground on the LOAD side of the GFCI:

Black--------GFCI-------->--load-+ White--------GFCI--------<-------+ Short to ground on LOAD side---<-+

Note the arrows < (and > ) - current passes through the load and heads back toward the GFCI, but it also finds a path through the short circuit. That causes the GFCI to trip, because it detects a greater current from the hot wire than it detects returning to the neutral wire.

GFCI's have a LINE (power input side) and a LOAD (power output to the load) side, usually (always?) labelled with capital letters LINE and LOAD. That's why I used LOAD in some places.

Hope this helps. Ed

This is my view of that but Peter W apparently has another theory of GF tripping due to the wire+interconnects resistance.

Peter Wieck Melrose Park, PA

Because the incorrect wiring meant it was permanently near it's 30mA limit and a little bit of extra harmonic content pushed it over the edge?

When a 30mA RCD is prone to false triggering it can mean you are trying to feed to much stuff from it. Some leakage is normal for many devices.

Here we usually use several 30mA RCDs, maybe one for upstairs sockets, one for downstairs sockets, one for lighting or something similar, but making sure everything is still protected.

Well, it is mandatory in Spain to have at least one RCD covering the entire house, so that's the normal: one RCD at the house entry, protecting all. Some houses may have two or more, if they are really big.

But the total maximum of the house was 5 or 6 KW at the time, and normal usage was much lower. Fridge, hot water tank (1KW), washing up machine (2KW), and a water pump (0.7 KW). TV and lights.