What sort of earthing do I have?

[plugwash]

Traditionally speaking, I would have said that (L-E) 'faults within the CU' would be incredibly unlikely,

With a plastic cased CU and no surge protection, I would agree, with a metal-cased CU and/or surge protectors I'd consider it merely "unlikely" rather than "incrediblly unlikely"

If there were "no CPCs" then OPD-based ADS would presumably also not work in most situations?

Indeed it wouldn't, which would leave the RCD detecting current flowing through the victim as the only protection in the event of a failure of basic insulation.

 

[JohnW2]

With a plastic cased CU and no surge protection, I would agree, with a metal-cased CU and/or surge protectors I'd consider it merely "unlikely" rather than "incrediblly unlikely"

Fair enough. If anything, we're only disagreeing about 'degree'

Indeed it wouldn't, which would leave the RCD detecting current flowing through the victim as the only protection in the event of a failure of basic insulation.

True, but I really didn't understand Ragnar's comment, since "if there were no CPCs" seems like a very fanciful hypothetical situation to be talking about!

I've always suspected that the reason why 415.1.2 came about was that, when RCDs appeared people became totally fixated on the 'exciting' (and novel) concept of their sometimes being able to disconnect the supply when someone got an electric shock, thereby potentially allowing some victims to survive shocks which might otherwise have been fatal. However,they became so 'fixated' on that one potential benefit of RCDs that many probably came to think of them as only having that one purpose - hence totally losing sight of the fact that they were actually 'much better' than OPDs at providing fault protection when the L-E current was not going through a human being!

The problem with OPD-based ADS is that we necessarily design circuits so that required disconnection times will (sometimes only 'just') be achieved in the event of an L-E fault of 'negligible' (aka 'zero' ) impedance, but it only needs the fault path to have an impedance of 'a few ohms' (sometimes even less) for those required disconnection times not to be achieved - hence extending the time period during which people are at risk of electric shocks. In contrast, an RCD will achieve the reequired disconnection times even when the fault path has 'considerable' impedance.

Looked at the other way around, if RCDs had 'always been there' as the recognised means of providing fault protection (ADS), I don't think anyone would have dreamed of suggesting that some other type of device, much less sensitive to L-E faults of finite impedance, should be regarded as the 'primary' provider of fault protection/ADS, would they?

 

[plugwash]

True, but I really didn't understand Ragnar's comment, since "if there were no CPCs" seems like a very fanciful hypothetical situation to be talking about!

Again, as I and Ragner have been trying to explain, modern editions of BS7671 are not written from scratch for the UK. they are based on an IEC standard modified to reflect UK practices and thinking. Even in the UK, lighting circuits with no CPCs are not unheard of, and I'm pretty sure there are places in the world where you will find power circuits with no CPCs.

Use of earth fault protection as a mitigation for missing CPCs is certainly something that is done in some countries. I'm not sure exactly how the regs are worded but I certainly knew someone in the netherlands who lived in a property where most rooms did not have CPCs, but their installation was RCD protected. In the USA, installing GFCIs (their version of RCDs) with the ground not connected to anything is a recognised way to retrofit three-pin sockets on circuits with no ground conductors (it is *not* however recognised as a protection method for new work).

The problem with OPD-based ADS is that we necessarily design circuits so that required disconnection times will (sometimes only 'just') be achieved in the event of an L-E fault of 'negligible' (aka 'zero' ) impedance, but it only needs the fault path to have an impedance of 'a few ohms' (sometimes even less) for those required disconnection times not to be achieved

Somewhat agreed, but the question is how likely is a fault that is both

* High enough impedance not to rapidly trip a OCPD.
* low enough impedance to result in a dangerous "touch voltage" on a TN supply.

OTOH on TT supplies it seems the risk of a fault that was low enough to create dangerous touch voltages, but high enough not to cause a trip was a risk they decided needed to be addressed, resulting in the requirement that the impedance to earth multipled by the RCD trip current must be less than 50V.

 

[ebee]

Perhaps it`s an age thing too, or at least an evolution thingy.
Some of us were busy wiring consumer units etc using only the humble rewireable fuse (BS 3036) for up to three different types of overcurrent protection.

Not an RCD (ELCB as was or GFCI in some places) in sight.
OK we were aware of them and the advantages if they worked as should but they were not universally common.
We enclosed some fuses and called them HBC or HRC and then MCBs started to become popular instead of the fusewire.
It was a long time before RCDs started to become more common and in differing considerations for consumer units,
that in itself went thru an evolution process and is still doing so, the RCBO perhaps viewed as a best option in many applications.
The description "Power breaker" as a stand alone socket or as a plug in adaptor was one of the widely recognised devices at times.
The old "ELCBs" had two variants, Current operated (became the RCD) and Voltage operated (got confined to history, almost) although I do remember ELCBs being considered at times and then the consideration as to the Voltage or the Current operated types.

 

[Ragnar_AT]

As plugwash said, large parts of Europe used non-earthed sockets in dry rooms until fairly recent times, e.g. the Netherlands, Scandinavia and Finland. In these countries it was and partly still is considered an acceptable upgrade to fit an RCD without adding CPCs to all circuits. In other countries, like the UK, Germany and Austria, it isn‘t. In order to fulfill the requirements for fault protection you need a CPC, double insulation or PELV (e.g. the old 12 V halogen lighting setups). That’s all I take that paragraph to say.

The only reference to TT systems and fault protection I‘ve come across so far is that the Austrians amended the IEC regs to state that an RCD may not provide both fault protection and additional protection at once, i.e. in a TT installation you need two RCDs in series because the OPD can’t disconnect an earth fault due to the high impedance. The reasoning seems to be you always want some kind of backup plan in case the RCD fails.

 

[EFLImpudence]

the Austrians amended the IEC regs to state that an RCD may not provide both fault protection and additional protection at once, i.e. in a TT installation you need two RCDs in series because the OPD can’t disconnect an earth fault due to the high impedance. The reasoning seems to be you always want some kind of backup plan in case the RCD fails.

On a semantic note I would point out that "additional" simply means "extra"; it is not a specific electrical method of protection.

So if an RCD is added, whatever it does will be "extra/additional" to what was there before.

 

[EFLImpudence]

As plugwash said, large parts of Europe used non-earthed sockets in dry rooms until fairly recent times, e.g. the Netherlands, Scandinavia and Finland. In these countries it was and partly still is considered an acceptable upgrade to fit an RCD without adding CPCs to all circuits.

My own flat here in Portugal (5th floor) only has a CPC in the kitchen socket circuit.
All class 2 items and plastic pipes etc.

Being British and unused to this I added RCD protection to the other circuits.
However, as JohnW2 pointed out, if I am 'happy' to not have a CPC on these circuits then an RCD isn't likely to make much difference as there is nothing to earth me should I touch a live conductor or exposed-conductive-part.

 

[JohnW2]

Again, as I and Ragner have been trying to explain, modern editions of BS7671 are not written from scratch for the UK. they are based on an IEC standard modified to reflect UK practices and thinking. Even in the UK, lighting circuits with no CPCs are not unheard of, and I'm pretty sure there are places in the world where you will find power circuits with no CPCs.

Yes, I understand all that. However, when it came to creating UK regulations, I don't think the fact that they were 'based on an IEC standard' really excuses:

1... Having two regs which are literally contradictory - one saying that an RCD is an acceptable device for providing fault protection in a TN installation, and the other saying that an RCD alone is not acceptable​

OR​

2... Writing UK regulations which, at least in my opinion, defy electrical common sense. Yes, RCDs can do other things as well, but in terms of 'fault protection' in the normal sense (i.e. as provided by an OPD), and RCD cannot only do the same as an OPD but can, in some senses, do it better - AND can be regularly/routinely tested to confirm that satisfactory protection persist - which is not the case with an OPD.​

​

Use of earth fault protection as a mitigation for missing CPCs is certainly something that is done in some countries.

I do somewhat struggle to understand the common sense of that. In the absence of CPCs, exposed-c-ps of equipment can become live due to a fault and, if not 'earthed' via a CPC will remain live 'for ever', hence presenting 'for ever' the risk of electric shock if they touch it and simultaneously manage to touch something which is earthed. An RCD cannot clear such a fault any more than can an OPD, so the only sense in which it 'mitigates' the absence of CPCs is that, the absent CPCs having increased the risk of someone receiving a shock, it offers the possibility that the RCD may prevent the shock being fatal.

I think I would prefer the ''protection' to have made it much less likely that I would get a shock, rather than decreasing the risk of dying if I did

Somewhat agreed, but the question is how likely is a fault that is both
* High enough impedance not to rapidly trip a OCPD.
* low enough impedance to result in a dangerous "touch voltage" on a TN supply.

One can obviously argue about actual probabilities, but I would say 'fairly likely'.

The problem arises with fault impedances which are 'very low', but not quite low enough to achieve magnetic tripping of an RCD (quite likely given that some circuits often have Zs which is only just low enough for magnetic tripping the e face of a zero-impedance fault). In that situation, the 'very low' (but not quite low enough for the OPD) fault impedance could mean that (depending on relative CSAs of L conductor and CPC) not a lot less than half the supply voltage.

OTOH on TT supplies it seems the risk of a fault that was low enough to create dangerous touch voltages, but high enough not to cause a trip was a risk they decided needed to be addressed, resulting in the requirement that the impedance to earth multipled by the RCD trip current must be less than 50V.

Yes, but that merely requires that (for a 30mA RCD) the TT earth have an impedance less than 1667Ω, which is always going to be the case if it's compliant. In this case, it doesn't really matter what the fault impedance is, since, whilst just 'a fraction of an Ohm' can be enough to prevent magnetic tripping of an MCB, it would require 'a good few Ohms' to make any noticeable difference in a TT installation !

 

[JohnW2]

As plugwash said, large parts of Europe used non-earthed sockets in dry rooms until fairly recent times, e.g. the Netherlands, Scandinavia and Finland. In these countries it was and partly still is considered an acceptable upgrade to fit an RCD without adding CPCs to all circuits. In other countries, like the UK, Germany and Austria, it isn‘t.

I'm really not sure that practices and regulations in other countries should really affect how sensible UK regulations are in relation to the UK

In order to fulfill the requirements for fault protection you need a CPC, double insulation or PELV (e.g. the old 12 V halogen lighting setups). That’s all I take that paragraph to say.

I'm not too sure which 'paragraph' you think says that?

The only reference to TT systems and fault protection I‘ve come across so far is that the Austrians amended the IEC regs to state that an RCD may not provide both fault protection and additional protection at once,

Well, as EFLI has pointed out, if one takes the (I would say 'silly') terminology literally, then that becomes a self-fulfilling truth, since no form of protection can be "additional" to itself

However, what one can say is that an RCD does provide some degree of 'personal' ('additional') protection to people receiving a shock as well as providing protection against the sort of L-E faults than OPD-based ADS can deal with.
has to be the truth

i.e. in a TT installation you need two RCDs in series because the OPD can’t disconnect an earth fault due to the high impedance. The reasoning seems to be you always want some kind of backup plan in case the RCD fails.

Hmmm - so why does one not also need 'some kind of backup' in case the OPD fails (particularly given that, unlike RCDs, OPDs cannot be 'tested) ??

 

[SimonH2]

Hmmm - so why does one not also need 'some kind of backup' in case the OPD fails (particularly given that, unlike RCDs, OPDs cannot be 'tested) ??

As I wrote in post 25 above, MCBs have very little to go wrong, while RCDs have much more to go wrong. These days we tend to consider them "reliable", but I suspect they were far less so back when the regs were written.
Plus, you need overcurrent protection anyway because the RCD does nothing for overload or L-N faults.

Then there's a discussion about practicality. I suspect we'd be having interesting discussions if the regs required two MCBs in series.

 

[JohnW2]

As I wrote in post 25 above, MCBs have very little to go wrong, while RCDs have much more to go wrong.

You did, but they certainly can 'stick', and it's even possible that the contacts could become welded together. In the absence of a practical way of testing them, we really know nothing about the reliability of MCBs in service.

.... Plus, you need overcurrent protection anyway because the RCD does nothing for overload or L-N faults.

Indeed. That's why every circuit in every installation (TN or TT) has to have an OPD. In a TT installation (in which it's extremely unlikely that an OPD would provide adequate fault protection), we are stuck with having to rely on an RCD for that protection (despite BS7671 saying one shouldn't do that ). What we are talking about is a circuit in a TN installation which has a Zs slightly too high for an OPD to provide the required disconnection times (but hopefully low enough for L-N faults!).

 

[Ross1978]

OK - Exciting update follows. I see I have given you guys plenty of talking points on the regs!

National Grid guy came round this morning and had a look inside their box (pic below).

The earth lead coming from their box is not connected internally to anything. If it was connected that metal strap you can see would be vertical and connected to the copper braids you can see.

So our earth lead serves no purpose. I asked if it could be connected, and he said he's not allowed to do that. Apparently if the supply was going direct into a building, it would be connected, however as they are treating our installation as a "sub main" they didn't connect it when the cabinet was installed.

I've gone back to National Grid people to ask if this can be connected as spark no. 2 says this will be a better earth than we have currently.

So I think I am right in saying that as it stands, this is basically a "TT" type of arrangement as far as our building is concerned.

(added pics of the board below too for completeness, this is 80m away from the meter)

 

[SimonH2]

You did, but they certainly can 'stick', and it's even possible that the contacts could become welded together. In the absence of a practical way of testing them, we really know nothing about the reliability of MCBs in service.

Err, so what does the lever on the front do ?
It will exercise at least parts of the mechanism. Assuming these work, there is very little left to go wrong.

 

[JohnW2]

Err, so what does the lever on the front do ? ... It will exercise at least parts of the mechanism. Assuming these work, there is very little left to go wrong.

I'm far from convinced that's a very foolproof way of 'testing' an MCB but, even if it were, how many people (even those undertaking EICRs) regularly operate the levers of all their MCBs to see what happens?

CUs do (should) bear labels indicating that RCDs should be 'tested regularly', but as regards MCBs ...... !