What sort of earthing do I have?

[JohnW2]

Yeah, it's not very well worded but I think the consensus is that should be read as forbidding the use of RCDs to directly provide shock protection, rather than using them as a means of ADS.

Oh - I thought the consensus was the opposite of that - namely that RCDs should be used to provide ('additional') 'direct shock protection) in addition to that (less 'direct') protection provided by OPD-based ADS ?

At least if "ADS" is taken literally, one cannot avoid the fact that RCDs, as well as OPDs, will provide ADS (whether or not with the required disconnection times). I (and I thought most people) have always assumed that the reg means that one should (ideally) not rely on RCDs to be the primary provider the ADS 'IF there is a TN-based alternative'.

It just seems perverse to me to when faced with a "not good enough" TN earth to replace it with a "worse" TT earth.

In a sense I agree, but the important thing is that if a TN earth is "not good enough" (for ADS) then, regardless of whether one just 'accepts' the under-spec ADS or installs TT rods, all circuits must be RCD protected.
What I suppose would be a little perverse would be to provide a new TT earth and then disconnect the "not good enough"TN one (particularly if it was only "not quite good enough") - if the TN 'earth' had any continuing connection to true earth, to leave it in parallel with the TT earth would improve things at least a bit (in terms of OPD-mediated disconnection times). However, it would still need to be regarded as a TT installation (particularly as regards the requirement for RCD protection of everything).

 

[plugwash]

Oh - I thought the consensus was the opposite of that - namely that RCDs should be used to provide ('additional') 'direct shock protection) in addition to that (less 'direct') protection provided by OPD-based ADS ?

Yes, additional to ADS, rather than instead of ADS.

 

[JohnW2]

Yes, additional to ADS, rather than instead of ADS.

I presume you mean OPD-based ADS? - since, as I said, and RCD provides "ADS" (in fact, dramatically more sensitively {to L-E faults} than does an OPD).

So, yes, as I said, I thought the consensus always was that (for whatever reason) the regs seems to believe that it is undesirable for primary protection against electric shocks (L-E) to be provided by an RCD 'IF there is a TN-based alternative', even though it is never worse than an OPD at doing that, and often dramatically 'better' (more sensitive).

Do you have any thoughts as to why the regs appear to have this belief? I can't imagine it's due to beliefs about relative reliability of the two types of device, given that RCDs can be easily tested, yet, in practice, OPDs can't be tested at all!!

 

[plugwash]

I presume you mean OPD-based ADS?

No, I mean EEBADs.

It's unfortunate that the authors of BS7671 (probablly as part of a process of trying to harmonise BS7671 with other IEC standards) decided to rename EEBADS to just ADS.

There are two ways RCDs can provide protection.

1. As part of EEBADs, by providing automatic disconnection of supply after current starts flowing through an earthing system.
2. On their own by disconnecting the supply after someone starts receving a shock.

I belive that 415.1.2 is trying to ban use of "2" as the sole means of protection. Interpreting it as banning "1" would mean it essentially banned TT installations completely.

 

[JohnW2]

No, I mean EEBADs. .... It's unfortunate that the authors of BS7671 (probablly as part of a process of trying to harmonise BS7671 with other IEC standards) decided to rename EEBADS to just ADS.

In the context of the current discussion (and most other contexts), putting "EEB" in front of ADS really doesn't change anything, which is perhaps why most people and organisations dropped the "EEB" many years ago. Lets' face it, allelectrical installations are effective required to have "EEB".

There are two ways RCDs can provide protection.
1. As part of EEBADs, by providing automatic disconnection of supply after current starts flowing through an earthing system.
2. On their own by disconnecting the supply after someone starts receving a shock.

Indeed so - although, as I've implied, I'm not really sure what you mean by "as part of EEBADS", given that all installations are required to have "EEB".

I belive that 415.1.2 is trying to ban use of "2" as the sole means of protection. Interpreting it as banning "1" would mean it essentially banned TT installations completely.

Yes, but you appear to be overlooking (or, for some reason, 'ignoring') the fact that, as I've said and you have essentially implied, that there is no way that one can avoid/prevent an RCD from providing both protections (1) and (2) - whereas an OPD-based system can only (see ** below) provide protection (1) - which is why I struggle to see why using just an RCD for ('L-E') fault protection should be 'forbidden' (there is obviously still a need for an OPD for overload/L-N fault protection).

[** It would be impossible to get enough current through a human being, even if dead, to trip even a B6 MCB, unless they had been 'fried' to such an extent that their body had been changed into a 'block of carbon' !! ]

In fact, in the case of an RCD [which can do both (1) and (2) ], the only difference between (1) and (2) is really quantitative - since, even in (2), the current flowing through the victim will (if they are indoors, in a building which has EEB) "flow through the earthing system.

However, returning to (1) the quantitative difference between the protection provided by RCD and OPD is enormous, since tripping with an RCD (i.e. "ADS") will occur with a current to earth 100 times less than required to trip a B6 or about 533 times less than required to trip a B32 - so, again, I find it difficult to see why an RCD should be 'forbidden' from providing protection (1).

Kind Regards, John

 

[plugwash]

Yes, but you appear to be overlooking (or, for some reason, 'ignoring') the fact that, as I've said and you have essentially implied, that there is no way that one can avoid/prevent an RCD from providing both protections (1) and (2)

You can prevent it from providing (1) by not having an earth connection, or by having an earth connection whose impedance is too high for the RCD in question. I don't think using RCDs *instead of* earthing has ever been the "done thing" in the UK, but it certainly has in other countries. So saying that doing it is not acceptable in the UK would not be an unreasonable thing for BS7671 to do.

You can prevent it from providing (2) by having a RCD whose trip current is too high to trip on the currents likely to flow through the human body..

Indeed so - although, as I've implied, I'm not really sure what you mean by "as part of EEBADS"

I consider EEBADS to be a protection strategy where you.

1. Earth metalwork that surrounds electrical connections.
2. Bond all earthed (whether intentionally or unintentially) metalwork together.
3. Include protective devices that automatically disconnect the supply in the event of an earth fault, and perform measurements and calculations to ensure that those devices are adequate.

Most metalwork is not being touched by people most of the time, so EEBADs can, in most cases, disconnect the supply *before* someone gets a shock.

In contrast RCDs without earthing (or where earthing has failed) can only disconnect the supply *after* someone gets a shock, and only if the shock current happens to be high enough to trip said RCD.

 

[JohnW2]

You can prevent it from providing (1) by not having an earth connection, or by having an earth connection whose impedance is too high for the RCD in question.

I don't really understand what you mean by "not having an earth connection". In the absence of any earth, there can be no L-E fault currents, so no device would be able to clear anything.

I don't think using RCDs *instead of* earthing has ever been the "done thing" in the UK, but it certainly has in other countries.

Again, I don't really understand what you mean. An RCD can only work if there is some route (via 'earth') back to the supply transformer other than the N conductor. No-one is suggesting that an installation could have an RCD "instead of an earth"/

You can prevent it from providing (2) by having a RCD whose trip current is too high to trip on the currents likely to flow through the human body..

Whilst obviously true, that's surely irrelevant, isn't it, given that we are usually talking about 30 mA RCDs - and even 100 mA (or more) is far from an impossible current through a human body.

I consider EEBADS to be a protection strategy where you.
1. Earth metalwork that surrounds electrical connections.
2. Bond all earthed (whether intentionally or unintentially) metalwork together.
3. Include protective devices that automatically disconnect the supply in the event of an earth fault, and perform measurements and calculations to ensure that those devices are adequate.

Totally agreed - but why can't that protective device be an RCD, particularly given that RCDs are so much more sensitive to 'earth fault currents' than are OPDs??

Most metalwork is not being touched by people most of the time, so EEBADs can, in most cases, disconnect the supply *before* someone gets a shock.

Exactly - and if the protective device is an RCD (rather than an OPD) that "disconnection before someone gets a shock" will occur in response to a very much smaller fault current than would be the case if the device were an OPD. Hence, for example, if it is a 'gradually increasing fault' (e.g. due to moisture ingress), an RCD will operate much sooner than an OPD, thereby making it even less likely that someone will get a shock.

In contrast RCDs without earthing (or where earthing has failed) can only disconnect the supply *after* someone gets a shock, and only if the shock current happens to be high enough to trip said RCD.

Again, I really don't understand what you mean by "RCDs without earthing" ... and, as for "where earthing has failed" that would presumably prevent OPD-based ADS from working. In fact, as above, it would take a far greater degree of "earthing failure" (increase in impedance to earth) to stop an RCD working (and resulting in required disconnection time) than would be the case with OPD-reliant protection.

Indeed, it is common for the Zs of circuits to be very close to the 'maximum' that will achieve required disconnection times, so that just a slight increase in impedance of the path to earth may well be enough to result in non-compliant disconnection times, whereas it would require a dramatic increase in that impedance to prevent an RCD from providing compliant disconnection times.

 

[securespark]

I always understood that DNOs have a responsibility to maintain an earth connection where they have historically provided one.

 

[ericmark]

The problem today is a supply type can change due to the loss of grid connection, and although we test an RCD, we don't test an MCB, we know that B = 3 to 5 times thermal cutout rating, and we realise for that to flow the loop impedance needs to be under a set figure, what we don't know is if the generator can actually produce that current.

So my house is TN-C-S and TN-S depending on if there is a grid supply, and in the TN-S mode it relies on the earth electrode and the relay in the inverter to connect the earth and neutral together when there is no grid supply. But the earth electrode is not connected and disconnected, it simply changes name, in TN-C-S mode it is protective equipotential bonding, of an extraneous-conductive-part, and in TN-S mode it is protective equipotential bonding of an earth electrode.

At one point, it was a trick question, why can't you have an earth electrode with a TN-S system, the answer is it called extraneous-conductive-part.

 

[SimonH2]

I don't really understand what you mean by "not having an earth connection". In the absence of any earth, there can be no L-E fault currents, so no device would be able to clear anything.

I had assumed absence of a functional CPC. You can never rule out a parasitic earth path via a person ...

Totally agreed - but why can't that protective device be an RCD, particularly given that RCDs are so much more sensitive to 'earth fault currents' than are OPDs??

If you consider what's in an MCB and an RCD, there is going to be a massive difference in PFD (probability of failure on demand) - I'd guess an order or two difference. Essentially, you've a very similar trip mechanism - but the RCD adds a bunch of additional ones in the path from detection to trip. That detection path almost certainly does not have real-time integrity monitoring - so an unsafe failure of a component could occur at any time and leave the RCD inoperative, unknown until next tested (if it ever gets tested.)

So for something that is safety critical (even if it's not normally described as such), do you really want to use something with a PFD orders of magnitude higher ?

 

[plugwash]

I don't really understand what you mean by "not having an earth connection". In the absence of any earth, there can be no L-E fault currents, so no device would be able to clear anything.

The absense of an earth (intentional or otherwise) to one piece of metalwork does not imply the absence of an earth (intentional or otherwise) to other pieces of metalwork.

That said, in some countries it was the norm to have no earths at all in dry rooms, on the theory that if there is no earth there is nowhere for shock currents to go. Which worked ok until central heating, data networking and so-on came along..................

No-one is suggesting that an installation could have an RCD "instead of an earth"/

Remember, modern editions of BS7671 are based on an IEC standard (I forgot the number offhand). Things that are considered unthinkable in the UK may have been done in the recent past in other countries.

You can often tell which regs are UK additions, because they have 3 digit last numbers. The fact this has a small number suggests it's taken from the base IEC standard.

Whilst obviously true, that's surely irrelevant, isn't it, given that we are usually talking about 30 mA RCDs - and even 100 mA (or more) is far from an impossible current through a human body.

I don't think there is generally any specific limit on the RCDs used for ADS in TT systems. I could well imagine a large installation having a main RCD with a trip current measured in amps. https://crompton-metershop.co.uk/protector-trip-relays/Earth-Leakage /earth-leakage-373-elr has settings from 30mA all the way up to 10A.

Again, I really don't understand what you mean by "RCDs without earthing" ... and, as for "where earthing has failed" that would presumably prevent OPD-based ADS from working.

Indeed it would.

Ultimately I see two ways of reading 415.1.2

1. As a prohibition from using RCDs as the ADS compoent in an EEBADS system.
2. As a reminder that RCDs used as "additional protection" (aka protection when both "basic protection" and "fault protection" have failed to do their job) don't exempt one from the normal "fault protection" requirements.

Reading 1 would essentially prohibit any TT systems, so does not seem reasonable.

 

[JohnW2]

I always understood that DNOs have a responsibility to maintain an earth connection where they have historically provided one.

That's also always been my understanding.

However, I'm not sure what that has got to do with the question of why BS 7671 seemingly doesn't want RCDs to be used as the primary means of fault protection.

 

[ericmark]

I always understood that DNOs have a responsibility to maintain an earth connection where they have historically provided one.

I know of one case where that did not happen, electrician phoned DNO in hindsight should have been in writing, asked what earthing system was to be used, told TN-C-S so left tails for line, neutral, and earth. But the DNO operative it seems could not get a low enough loop impedance, so decided it should be TT, result for 4 months the house had no proper earth.

It was only after an inquiring visit, as no live tests had been made, that the electrician found the supply was now live and in use. Although in the 6 months since he had wired the house, he had got divorced and was living in a narrow boat, so of no fixed abode, and maybe his mobile was not always charged, but he never got a text message as expected to say power was connected.

In my own parent's house, I know as a young 14-year-old an error line - earth fault ruptured a 13 amp fuse, so there was an earth in the 60s, but come 2004 the only earth I could find was a GPO earth for the party line. Phoned DNO and got TN-C-S free of charge. But where did it go?

My friend, who lived in a woodland, pointed out the "Test" button on the RCD would not trip, I tried to test with my RCD test kit, and rather a high ELI, investigation found fed from another property with was TT with a ELCB-v, so not sure how the RCD ever had worked? He was not into that type of DIY, so someone claiming to be an electrician would have fitted it.

I use this house with its known results to test my testers, so if something happened to my earth I would know, but unless some work is done, why would anyone test? Had my dad not been fitting a wet room, I would not have tested his earth, but again, care and repair from council had fitted a socket for his doorbell. Why did they not find no earth?

How long does it take to plug in a tester?

 

[JohnW2]

The absense of an earth (intentional or otherwise) to one piece of metalwork does not imply the absence of an earth (intentional or otherwise) to other pieces of metalwork.

Of course it doesn't, but if there is a fault from L to 'one piece of metalwork which isn't earthed', then neither an OPD nor an RCD would operate, regardless of what other metalwork was or was not earthed.

I don't think there is generally any specific limit on the RCDs used for ADS in TT systems. I could well imagine a large installation having a main RCD with a trip current measured in amps. https://crompton-metershop.co.uk/protector-trip-relays/Earth-Leakage /earth-leakage-373-elr has settings from 30mA all the way up to 10A.

Well, for a start, this is a DIY forum where we are (should!) essentially be talking only about domestic installations. In any event ...

411.5 TT system
.......
411.5.3 Where an RCD is used for fault protection, the following conditions shall be fulfilled:
(i) The disconnection time shall be that required by Regulation 411.3.2.2 or 411.3.2.4, and
(ii) RA × IΔn ≤ 50 V
where:
RA is the sum of the resistances of the earth electrode and the protective conductor connecting it to the
exposed-conductive-parts (in ohms)
IΔn is the rated residual operating current of the RCD.

... the effect of which is, when an RCD is used for fault protection in a TT system, to limit IΔn (albeit way above the devices IΔn) - e.g. 500 mA in the case of a 100 Ω TT earth.

In practice, of course, the RCD would have an IΔn of 30 mA, in order that it also provided 'additional' ('direct personal') protection, which majorly 'trumps' the much more modest requirement of 411.5.3.

However, regardless of the above reg, my first reaction to your point would be to answer "So what?" - I don't really see why you would (should) have a problem with ADS (say, for a 32A circuit) being provided by a device which would trip with an L-E fault current of 10A when you appear to think that it would be 'better' to have it provided by a device which could require 160A to trip !¬

Ultimately I see two ways of reading 415.1.2
1. As a prohibition from using RCDs as the ADS compoent in an EEBADS system.

As before, I don't really understand what you mean by an "EEBADS system" - since 'EEB' is required in all installations, any system with provides 'ADS' will also be 'EEBADS' in any compliant installation.

2. As a reminder that RCDs used as "additional protection" (aka protection when both "basic protection" and "fault protection" have failed to do their job) don't exempt one from the normal "fault protection" requirements.

OK - but, as I've asked, why can't those "fault protection requirements" be satisfied by an RCD - which essentially does the same (but, in some senses, "much better') than an OPD in this respect?

Reading 1 would essentially prohibit any TT systems, so does not seem reasonable.

I observed at the very start that 415.1.2 appears to be 'incorrect', in that it seemingly fails to acknowledge that one has no choice but to rely on RCDs for fault protection in TT installations (and, taking 'practicalities' into account, even some TN ones). That is inconsistent with

411 PROTECTIVE MEASURE: AUTOMATIC DISCONNECTION OF SUPPLY
.....
411.5 TT system
....
411.5.2 One or more of the following types of protective device shall be used, the former being preferred:
(i) An RCD
(ii) An overcurrent protective device.
NOTE 1: An appropriate overcurrent protective device may be used for fault protection provided a suitably low value of Zs is
permanently and reliably assured.

Furthermore 411.4.5 explicitly says that an RCD may be used to provide ADS in a TN installation - so there are multiple inconsistences within the regs.

411 PROTECTIVE MEASURE: AUTOMATIC DISCONNECTION OF SUPPLY
.....
411.4 TN system
....
411.4.5 The following types of protective device may be used for fault protection:
(i) An overcurrent protective device
(ii) An RCD

 

[JohnW2]

I always understood that DNOs have a responsibility to maintain an earth connection where they have historically provided one.

I know of one case where that did not happen .....

From what you go on to write, it sounds as if this may have been a 'new supply' - was that the case? If so, then if the DNO had not previously 'provided an earth', then they would presumably be under no obligation to provide or 'maintain one'?

.... electrician phoned DNO in hindsight should have been in writing, asked what earthing system was to be used, told TN-C-S so left tails for line, neutral, and earth. But the DNO operative it seems could not get a low enough loop impedance, so decided it should be TT, result for 4 months the house had no proper earth. ........ It was only after an inquiring visit, as no live tests had been made, that the electrician found the supply was now live and in use. Although in the 6 months since he had wired the house ....

It sounds as if the DNO person who did that was more than a little naughty. Even if he/she was not able to provide a TN earth (because of ELI), he/she should surely not have made an installation live when they knew that it currently had no earth at all ?