External garage electricity wrongly connected to neighbours supply

[EFLImpudence]

It is right Part of the N current from a load protected by a different RCD goes 'backwards' through the N side of the RCD protecting the circuit with a N-E fault, thereby creating an L/N imbalance (hence a trip, if enough imbalance) of that RCD ...



No. The full N current from load (as well as the full L current from the load) goes through the RCD protecting the circuit(s) with a load ("RCD1" in diagram above), hence no trip of that/those RCD(s)

Yes, I obviously misread your post. I thought you were saying both RCDs would trip.

This is the situation in Bernard's frequently posted diagram .

 

[JohnW2]

Yes, I obviously misread your post.

Fair enough.

I thought you were saying both RCDs would trip.

As you now realise, I obviously wasn't saying that.

The important point is that many people don't seem to realise that a load protected by a different RCD can result in tripping of an RCD protecting a circuit with an N-E fault - and, even if they do realise that, don't also realise that, if it happens, the RCD protecting the load (but not the circuit with the fault) will not trip.

 

[ericmark]

With anything other than a TN-C-S it is surely inevitable that there will be some N-E potential difference, isn't it? Absence of some N-E pd could only occur if the length of neutral conductor from transformer to installation is zero (impossible), or if no customer served by the transformer is drawing any current (incredibly unlikley)?

I agree highly unlikely, but with a split or three-phase supply, zero current can be flowing through the neutral, even when there is a load.

 

[JohnW2]

I agree highly unlikely, but with a split or three-phase supply, zero current can be flowing through the neutral, even when there is a load.

I would say that, even with a ('loaded') split-phase or 3-phase supply, it would be incredibly unlikely that the neutral current would be exactly zero, wouldn't it?

 

[SUNRAY]

It is right Part of the N current from a load protected by a different RCD goes 'backwards' through the N side of the RCD protecting the circuit with a N-E fault, thereby creating an L/N imbalance (hence a trip, if enough imbalance) of that RCD ...

View attachment 405160

No. The full N current from load (as well as the full L current from the load) goes through the RCD protecting the circuit(s) with a load ("RCD1" in diagram above), hence no trip of that/those RCD(s)

Your diagram correctly describes my earlier comments and demonstrates the most likely reason there is a PD betwix N & E.
The load you show is anywhere on the system, including the properties along the road etc, the fact you have drawn this makes me surprised you (much more than some on here) didn't understand my comment from the start, unless of course you were not aware of the situation at the time.

However; I'll add if there was no N to E PD at the point you applied the short then there must be a low resistance/impedance N to E link somewhere prior but close to RCD2, there will be no or negligable current in the fault and threfore no trip.

However#2; The same would be true if there were no link at all with N floating.

 

[SUNRAY]

I would say that, even with a ('loaded') split-phase or 3-phase supply, it would be incredibly unlikely that the neutral current would be exactly zero, wouldn't it?

I would say that's true unless all of the loads are 3 wire 3ph

 

[JohnW2]

This is the situation in Bernard's frequently posted diagram .

I should perhaps have added ...

My diagram goes further than bernard's oft-posted one. His diagram shows a N-E fault resulting in an RCD tripping when there was a lot on some other circuit protected by the same RCD.

In contrast, mine shows that there can still be a trip even if the only load(s) are on circuit(s) protected by a different RCD (different from the one protecting the circuit with the N-E fault) - which, as I said, I think a good few folk actually realise.

Indeed, as we have discussed in the past, I think one can go even further than that - since I think even load(s) in different installations (i.e. some neighbours' houses) connected to the same neutral can result in a N-E fault causing an RCD protecting its circuit to trip. In other words, the "RCD 1" (and circuits protected by it) in my diagram above could be in a different installation/premises.

 

[JohnW2]

I would say that's true unless all of the loads are 3 wire 3ph

Agreed, but this is a DIY forum and the situation you describe in never going to arise in domestic premises.

In fact, probably pretty rarely even in non-domestic premises since, even if most of the loads are 3-wire 3-phase, there would often/usually be at least some single-phase circuits/loads, wouldn't there?

 

[JohnW2]

Your diagram correctly describes my earlier comments and demonstrates the most likely reason there is a PD betwix N & E.

As I wrote to eric, and as you have now agreed, if there are any single-phase loads being supplied via the transformer, then (other than in a TN-C-S) installation, a N-E potential difference is inevitable.

I'm therefore coming to wonder whether your statement which started this whole side discussion was perhaps simply a 'statement of the obvious' - i.e. that you were talking specifically about trips due to N-E faults (you didn't say that) and that when you appeared to be saying that the presence of a N-E potential difference facilitated such RCD trips, you merely meant that an N-E fault will only cause an RCD trip if there is at least some neutral current 'somewhere' (within the same installation, or maybe even a different one sharing the neutral), manifested by a measurable N-E pd - which, I say, is obvious (and, of course, virtually always the situation which exists).

However; I'll add if there was no N to E PD at the point you applied the short then there must be a low resistance/impedance N to E link somewhere prior but close to RCD2, there will be no or negligable current in the fault and threfore no trip.

We have perhaps not thought or written enough about the meaning of "E" and, particularly, about TN-C-S.

With TN-C-S there is a very low impedance ('approaching zero') 'link' between N and E, upstream of, and usually very close to, the RCDs - and hence usually a very small ('approaching zero') potential difference between N and the 'E' within the installation (i.e. MET/CPCs). In that case, a fault between N and the installation's "E" (MET/CPCs) is presumably unlikely to result in an RCD trip, although a fault between N and 'true E' might well result in a trip (manifested, using your language, as a pd between N and true E but not between N and the installation's 'E' (MET/CPCs) ?

 

[SUNRAY]

As I wrote to eric, and as you have now agreed, if there are any single-phase loads being supplied via the transformer, then (other than in a TN-C-S) installation, a N-E potential difference is inevitable.

I'm therefore coming to wonder whether your statement which started this whole side discussion was perhaps simply a 'statement of the obvious' - i.e. that you were talking specifically about trips due to N-E faults (you didn't say that) and that when you appeared to be saying that the presence of a N-E potential difference facilitated such RCD trips, you merely meant that an N-E fault will only cause an RCD trip if there is at least some neutral current 'somewhere' (within the same installation, or maybe even a different one sharing the neutral), manifested by a measurable N-E pd - which, I say, is obvious (and, of course, virtually always the situation which exists).

With respect the whole of my comments have been based on the presence of N-E voltage and a N-E fault and even described the fault.

N-E has over 1V so some potential for tripping, It didn't take long to find a very low resistance between N&E on the garage circuit and heading out to the garage somewhat surprised to find it powered, contact with the immediate neighbour soon established the mix-up (and for that matter isolated the wet light fitting).

OK I didn't mention that nothing in either garage was running but then not relevant and at no time have disputed where neutral voltage comes from as I didn't think I needed to 'state the bleedin' obvious.

 

[ebee]

Just to add (as touched on by JohnW2) - When someone mentioned to me years ago "An Earth is an Earth is an Earth!" I replied "No it isnt!" Of course I actually meant "Not Always" .
The Earth ref in our own installation might well be somewhere around 35V or something if some heavy indudtry is nearby when compared to "True Earth".

The whole planet can be thought of as a "true Earth" but our little local portion of it might rise above that on say a sloping voltage dependant upon our distance from "Earth" of the two measurements, Hence the DNO figure of 21 ohms almost guranteed for a TT type Earth (20 ohms for the two dsting Earths and 1 ohm for their connections perhaps ??? )

We can only define where our ""Earth" is and how it compares to the rest of the planet and decide 21 ohms as best likely guarantee in those cases

 

[ericmark]

Since the earth is normally wet, would not the earth voltage be dependent on the metal used for the earth electrode?

 

[JohnW2]

Since the earth is normally wet, would not the earth voltage be dependent on the metal used for the earth electrode?

I'm not sure what you mean by "the earth voltage" - it's potential difference relative to what?

 

[JohnW2]

With respect the whole of my comments have been based on the presence of N-E voltage and a N-E fault and even described the fault.

I don't think we really have much of difference, hence there is not really any need for either of us to feel/behave 'defensively', the issue being that we have probably confused each other little.

As per what I quoted, my initial comment focussed entirely on the first sentence of what you wrote, namely ....

N-E has over 1V so some potent ial for tripping

... which appeared (but probably was not intended to be) a very general statement.

For a start, at least in my personal experience, a 1V pd between N and E is pretty low for a non-TN-C-S installation, so I was a bit confused by the fact that you seemed to be implying that it was fairly high (well, that's how I interpreted your statement). Secondly, like, as I suspect, many/most people, I tend to think of RCD trips primarily in relation to L-E faults (lets face it, a good few people don't even realise that N-E faults can cause trips), so it seemed that you were suggesting that such trips were more likely with a 'high' N-E pd - which I don't believe is the case (but something which I now realise you were probably not intending to imply).

In passing, although I didn't comment at the time, I was also rather confused by what you went on to write, namely ....

... It didn't take long to find a very low resistance between N&E on the garage circuit and heading out to the garage somewhat surprised to find it powered, contact with the immediate neighbour soon established the mix-up (and for that matter isolated the wet light fitting).

I can but presume that the chronology of events was different from what seems to be implied by the way in which you have written that - since I can't see how you could have discovered that there was "a very low resistance between N&E on the garage circuit" before contacting the neighbour and hence finding a way of isolating the circuit. Is that correct?

There remains one thing about which I am rather confused. The argument you present (and with which I do not disagree) would seem to imply that RCD trips due to N-E faults ought to be very unlikely (if they happen at all) in TN-C-S installations - but I don't think that is the case. The N-side of the RCD in a TN-C-S installation is effectively between the (within-installation) N and E - so, given that the (within-installation) N & E are joined together, usually very close to the RCD, one (at least I!) would not expect an N-E fault on the circuit to make much difference to anything (unless there were some very large N current somewhere within the installation). What am I missing?

 

[SUNRAY]

I don't think we really have much of difference, hence there is not really any need for either of us to feel/behave 'defensively', the issue being that we have probably confused each other little.

As per what I quoted, my initial comment focussed entirely on the first sentence of what you wrote, namely ....

... which appeared (but probably was not intended to be) a very general statement.

For a start, at least in my personal experience, a 1V pd between N and E is pretty low for a non-TN-C-S installation, so I was a bit confused by the fact that you seemed to be implying that it was fairly high (well, that's how I interpreted your statement). Secondly, like, as I suspect, many/most people, I tend to think of RCD trips primarily in relation to L-E faults (lets face it, a good few people don't even realise that N-E faults can cause trips), so it seemed that you were suggesting that such trips were more likely with a 'high' N-E pd - which I don't believe is the case (but something which I now realise you were probably not intending to imply).

In passing, although I didn't comment at the time, I was also rather confused by what you went on to write, namely ....

I can but presume that the chronology of events was different from what seems to be implied by the way in which you have written that - since I can't see how you could have discovered that there was "a very low resistance between N&E on the garage circuit" before contacting the neighbour and hence finding a way of isolating the circuit. Is that correct?

The order of testing:

House CU:

With RCD off, testing on RCD O/P IR L-N showed something too low to read on the IR Robin due to connected loads. E-L showed similar unreadable low. E-N showed 1500Ω

With all MCB's off, E-N didn't change and very different readings across the individual circuits L's, in particular the garage and immersion L's showed as O/C, cooker showed ¼MΩ, ring and lights >2000Ω but too low for IR.

Clipping the leads to the N & E bars I removed the circuit N wires from the N bar one at a time to see what makes a change first was cooker - no change, next was garage, reading went to >2000Ω. Restored cooker N no change, checked IR E-N 18MΩ, switched on all MCBs except garage still 18MΩ, checked E-L also 18MΩ.

Disconnect garage L, repeated above tests no change. Reset RCD & power restored... big cheer from somewhere in the house.

Checked garage wires E-N still 1500Ω, E-N & L-N still O/C.

At that point I've proved an unexpected low resistance E-N and L is clear, (>200MΩ) on the garage circuit.

Call for customer and explain there is a fault in the garage whereupon he showed some confusion as didn't realise the garage power came from the house, discussion about is it worth going out in the rain to look, got his shoes and coat and walked over, fiddle with tricky up & over door, reaced in and automatically went to switch the light on, the fluo fitting started it's dance to fire up. Some discussion about the cable must go somewhere and seeing where it entered the garage Vs the location of the house CU I suggested it could easily have been swapped with the neighbours, he rang neighbour.

Up to that point it had taken less than 15 minutes on site.

It took longer than that to:

Wait for neighbour, (at that point we didn't know if it was going to lead anywhere) fault find and isolate, retest, close up.

Return to house to retest and reconnect, close up,

Return to garage, final check

Return to house, chat over deeds.

under 45 minutes on site plus 5 minutes each way.

There remains one thing about which I am rather confused. The argument you present (and with which I do not disagree) would seem to imply that RCD trips due to N-E faults ought to be very unlikely (if they happen at all) in TN-C-S installations - but I don't think that is the case. The N-side of the RCD in a TN-C-S installation is effectively between the (within-installation) N and E - so, given that the (within-installation) N & E are joined together, usually very close to the RCD, one (at least I!) would not expect an N-E fault on the circuit to make much difference to anything (unless there were some very large N current somewhere within the installation). What am I missing?

Assuming the N-E link is immediately before the RCD I'll place a figure of 1mΩ, assuming the faulty circuit is my garage example of say 25m of 2.5mm² so in the order of ½Ω it doesn't take too much calculation to establish that in order to get 30mA in the fault will rewuire 500x30mA or 15A so yes it will trip under a modest current

In an ideal world of course the 1mΩ N-E link would be 0Ω


However assuming there is no source side influence it can be seen it only requires a 60mA load (assuming the N-E fault is a direct short)