UK Scenario where lack of supplementary bonding is dangerous

[Falkus]

So I know why it is done, where it is done and how it’s done, but I’m yet to see a scenario where there is shock danger in a typical bathroom in a dwelling provided a) the main bonding is in place and b) the CPC’s of all circuits are connected together in the CU (as they will be).

Consensus is on an EICR code 2 where there is no RCD for additional protection and supplementary bonding is not in place. What is the potential danger ? What am I missing ?

 

[Pretty Mouth]

I usually find that 415.2.2 is satisfied without supplementary bonding being in place in/around the bathroom.

 

[Falkus]

I usually find that 415.2.2 is satisfied without supplementary bonding being in place in/around the bathroom.

As I said I know why and how it is done and I should have stated where it is not required- I.e where the circuits in the bathroom have additional protection from a 30mA RCD but my question was, and remains, where supplementary bonding is required, what danger exists where it is omitted, assuming all circuits have CpC and main bonding it in place ?

 

[timhoward]

where supplementary bonding is required, what danger exists where it is omitted, assuming all circuits have CpC and main bonding it in place ?

Well the assumption is really the point....
One example, someone DIY changes a light fitting with one they got from Ikea and doesn't connect the CPC's properly. Every light after it now has no CPC.
It's then possible to have two different versions of "earth" in the bathroom if someone touches a metal light fitting and an electric towel rail at the same time.
Supplementary bonding is belt and braces making sure that whatever else happens, they all have the same zero potential between them.

At least, that's my take on it. It's an interesting question.

 

[Falkus]

Well the assumption is really the point....
One example, someone DIY changes a light fitting with one they got from Ikea and doesn't connect the CPC's properly. Every light after it now has no CPC.
It's then possible to have two different versions of "earth" in the bathroom if someone touches a metal light fitting and an electric towel rail at the same time.
Supplementary bonding is belt and braces making sure that whatever else happens, they all have the same zero potential between them.

At least, that's my take on it. It's an interesting question.

It’s a reasoned response however if the metal light fitting is not earthed then touching it will not be introducing any potential will it ? If it’s faulty with a live/earth fault then yes it (metal parts of the luminaire) will be at 240v buts that’s already covered under the requirement to have a cpc connected to class 1 fittings or the use of class 2 where no cpc ?

 

[timhoward]

It’s a reasoned response however if the metal light fitting is not earthed then touching it will not be introducing any potential will it ? If it’s faulty with a live/earth fault then yes it (metal parts of the luminaire) will be at 240v buts that’s already covered under the requirement to have a cpc connected to class 1 fittings or the use of class 2 where no cpc ?

The thing is, it's 'supplementary'. It's additional protection. It's an extra safety net. We could say there's no need for all sorts of things if everything is installed perfectly and maintained perfectly.

We also need to remember that it pre-dates RCDs and it was the best way at the time to make sure that you can never have two pieces of exposed metalwork in a bathroom at different potentials to each other, when the resistance of the body is at an all time low as it's soaking wet.

It has a further role of adding possible earth paths so a phase to earth fault in any circuit in the room will cause ADS even if the faulting circuit has no upstream CPC continuity.

e.g. In the example I gave above, suppose they then like their Ikea light and install another one beyond the bathroom fitting but their twist and tape joint ends up touching the casing and the casing ends up live.
They'll turn the power back on and be none the wiser, but there's no fault protection and it's sitting there with a live case.

However in this case supplementary bonding saves the day and we'll get ADS. Without it, it will still sit there live.

Don't get me wrong, I prefer making sure every circuit in the room is RCD protected.
But I wouldn't say that supplementary bonding is pointless.

 

[Falkus]

The thing is, it's 'supplementary'. It's additional protection. It's an extra safety net. We could say there's no need for all sorts of things if everything is installed perfectly and maintained perfectly.

We also need to remember that it pre-dates RCDs and it was the best way at the time to make sure that you can never have two pieces of exposed metalwork in a bathroom at different potentials to each other, when the resistance of the body is at an all time low as it's soaking wet.

It has a further role of adding possible earth paths so a phase to earth fault in any circuit in the room will cause ADS even if the faulting circuit has no upstream CPC continuity.

e.g. In the example I gave above, suppose they then like their Ikea light and install another one beyond the bathroom fitting but their twist and tape joint ends up touching the casing and the casing ends up live.
They'll turn the power back on and be none the wiser, but there's no fault protection and it's sitting there with a live case.

However in this case supplementary bonding saves the day and we'll get ADS. Without it, it will still sit there live.

Don't get me wrong, I prefer making sure every circuit in the room is RCD protected.
But I wouldn't say that supplementary bonding is pointless.

Unless the metalwork of the light fitting in your scenario above is actually earthed (to a cpc) then your supplementary bonding is going to make no difference whatsoever. And if it is earthed it will be at the same potential as all the other circuit Cpcs because they are all connected at the CU - which will be in close proximity to the bathroom , even in a mansion house the potential difference due to circuit length would be in the ELV class

 

[Simon47]

You've missed the key point - CPCs not connected to the MET due to a failure (e.g. the dodgy diy).
In an ideal world we'd not have O/C CPCs and all manner of other faults. But in the real world I would suggest that the JPEL64 committee decided there was evidence of sufficient faults to justify having additional protection.

 

[timhoward]

Going right back to the beginning:

I’m yet to see a scenario where there is shock danger in a typical bathroom in a dwelling provided a) the main bonding is in place and b) the CPC’s of all circuits are connected together in the CU

I agree with you that if everything is perfect, it is indeed supplementary, the clue is somewhat in the name!

Unless the metalwork of the light fitting in your scenario above is actually earthed (to a cpc) then your supplementary bonding is going to make no difference whatsoever

We seem to agree that everything isn't always perfectly installed.

And if it is earthed it will be at the same potential as all the other circuit Cpcs because they are all connected at the CU

In my scenario this isn't in fact the case as the CPC is broken by the first Ikea light upstream.

CU - Lighting circuit - break in cpc - bathroom light - cpc intact - (multiple lights) - live - earth fault
With SB from bathroom light you get ADS as there is a path from the fault to the CU
Without SB you don't and you have an especially dangerous situation in bathroom and a dangerous situation at final light.

If it isn't earthed, as you suggest:
CU - Lighting circuit - break in cpc - bathroom light - cpc broken - live - earth fault
Yes, SB is irrelevant in this case, but you also don't have a dangerous situation in the bathroom which is what the SB is trying to achieve.

Let's not argue too much - I was born in Coventry too.... ;-)

 

[Falkus]

Going right back to the beginning:

I agree with you that if everything is perfect, it is indeed supplementary, the clue is somewhat in the name!

We seem to agree that everything isn't always perfectly installed.

In my scenario this isn't in fact the case as the CPC is broken by the first Ikea light upstream.

CU - Lighting circuit - break in cpc - bathroom light - cpc intact - (multiple lights) - live - earth fault
With SB from bathroom light you get ADS as there is a path from the fault to the CU
Without SB you don't and you have an especially dangerous situation in bathroom and a dangerous situation at final light.

If it isn't earthed, as you suggest:
CU - Lighting circuit - break in cpc - bathroom light - cpc broken - live - earth fault
Yes, SB is irrelevant in this case, but you also don't have a dangerous situation in the bathroom which is what the SB is trying to achieve.

Let's not argue too much - I was born in Coventry too.... ;-)

Hey I’m not arguing and thanks you’ve clarified your scenario which I can see is a valid case and one I hadn’t thought of - in effect loss of a cpc to a class 1 accessory. If then a second fault occurs (making the metalwork live) then ADS will disconnect the circuit removing shock hazard. This is a new one to me thanks because it’s normally stated to do with voltages between extraneous conductive parts. I suspect this scenario is welcome byproduct but not the intended aim ? Because this risk is not made any greater by being in a zone, it could apply to any room couldn’t it and we’d end up supplementary bonding every room in the house ?

 

[timhoward]

@Falkus and others, out of interest how would you code this one that I’m about to do:

Small flat. 2 circuits have no rcd protection, lighting and immersion.
Someone has added an electric shower (obvious addition) and there’s no SB.

Normally I would C3 the lighting without RCD.
The bathroom light needs changing anyway as it isn’t IP rated.
If the light is the only other circuit in the room and will be class 2, I’m wavering on C3 for the lack of SB as in this case to my thinking there are too many what-if’s to warrant a C2.

Even though I see the point of SB in this particular case I’m struggling to see a compelling danger as you can’t touch anything else that is class 1.

Would you do the same?

 

[Falkus]

@Falkus and others, out of interest how would you code this one that I’m about to do:

Small flat. 2 circuits have no rcd protection, lighting and immersion.
Someone has added an electric shower (obvious addition) and there’s no SB.

Normally I would C3 the lighting without RCD.
The bathroom light needs changing anyway as it isn’t IP rated.
If the light is the only other circuit in the room and will be class 2, I’m wavering on C3 for the lack of SB as in this case to my thinking there are too many what-if’s to warrant a C2.

Even though I see the point of SB in this particular case I’m struggling to see a compelling danger as you can’t touch anything else that is class 1.

Would you do the same?

Well firstly (assuming this is an EICR that you are about to “do”) I would code the EICR on what is the “now” situation not what it will be after you have changed the light or other remediation. Secondly this is the bit I’m trying to rationalise myself because if it’s a C3 then there is no danger or potential danger which would imply supplementary bonding has no bearing on the installation being satisfactory for continued use.

 

[Pretty Mouth]

@Falkus and others, out of interest how would you code this one that I’m about to do:

Small flat. 2 circuits have no rcd protection, lighting and immersion.
Someone has added an electric shower (obvious addition) and there’s no SB.

Normally I would C3 the lighting without RCD.
The bathroom light needs changing anyway as it isn’t IP rated.
If the light is the only other circuit in the room and will be class 2, I’m wavering on C3 for the lack of SB as in this case to my thinking there are too many what-if’s to warrant a C2.

Even though I see the point of SB in this particular case I’m struggling to see a compelling danger as you can’t touch anything else that is class 1.

Would you do the same?

Can pipework be touched at the same time as the shower?

Edit: I usually just C2 a shower without RCD protection, based on MI's

 

[timhoward]

Can pipework be touched at the same time as the shower?

I meant to cover that, in this case answer is no. Shower feed is plastic through the back, the only thing that came close was the bath taps but I checked and that has plastic plumbing too. Sink has a copper cold pipe but is completely out of reach (and also behind the pedestal)

 

[timhoward]

This is a new one to me thanks because it’s normally stated to do with voltages between extraneous conductive parts. I suspect this scenario is welcome byproduct but not the intended aim

I think in the broad sense the aim is indeed to ensure there is no potential difference between two things you can touch when you are soaking wet. Whether that is two versions of 'earth' or 'earth' and 230v to my thinking the same principle applies.

Well firstly (assuming this is an EICR that you are about to “do”) I would code the EICR on what is the “now” situation not what it will be after you have changed the light or other remediation

I take your point and many would do that. I have a working relationship with this letting agency that minor things I 'just fix'. A real bug bear of mine is an unsatisfactory EICR because there's a sticker missing on the consumer unit! I honestly think it works better for everyone to just resolve things that are easily resolvable.

Secondly this is the bit I’m trying to rationalise myself because if it’s a C3 then there is no danger or potential danger which would imply supplementary bonding has no bearing on the installation being satisfactory for continued use.

I think we have to look at each case on it's own merits.
If there was a radiator, or a heated towel rail, or anything else with exposed or extraneous conductive parts I wouldn't be suggesting this. In this case the supplementary bonding would be between a shower, a class 2 light fitting and a class 2 extractor fan. So it honestly doesn't really do a lot!

 

[Pretty Mouth]

I meant to cover that, in this case answer is no. Shower feed is plastic through the back, the only thing that came close was the bath taps but I checked and that has plastic plumbing too. Sink has a copper cold pipe but is completely out of reach (and also behind the pedestal)

C3 for the lack of sup bonding. I think this is suggested in BPG4, but check that.

C2 for the no RCD for the shower, but that is my personal opinion.

 

[timhoward]

C2 for the no RCD for the shower, but that is my personal opinion.

Sorry, Shower is on the RCD side of the board. Just lights and immersion are not.
So regs issues are ack of RCD for domestic lighting, and lack of supplementary bonding as not every circuit in room has RCD protection.
It's a proteus board so a couple of RCBO's wouldn't be a bad plan if I can sell the concept.

 

[Falkus]

I take your point and many would do that. I have a working relationship with this letting agency that minor things I 'just fix'. A real bug bear of mine is an unsatisfactory EICR because there's a sticker missing on the consumer unit! I honestly think it works better for everyone to just resolve things that are easily resolvable.

I know where you’re coming from with that but unless it’s a code 2 then I’m on an issue is not going to result in an unsatisfactory EICR and if indeed you remedy issues on the fly they won’t be reported against as being an observation on the report anyway.

The whole code 2 versus code 3 area is indeed very grey. All electrical equipment is potentially dangerous by its nature. I think it should be code 1: as is, immediate danger - requires isolation/locking off, code 2: improvements REQUIRED to improve safety (e.g add RCD protection, single insulation visible, damaged accessories, ) 3: improvements RECOMMENDED to improve safety or to comply with current regs, e.g add SPD, add AFDD, replace Type AC with type A, replace plastic CU under stairs etc

 

[Falkus]

Sorry, Shower is on the RCD side of the board. Just lights and immersion are not.
So regs issues are ack of RCD for domestic lighting, and lack of supplementary bonding as not every circuit in room has RCD protection.
It's a proteus board so a couple of RCBO's wouldn't be a bad plan if I can sell the concept.

I’ve taken the view on rental properties to err on side of code 2 versus code 3 on the basis that there is an increased likelihood from tenants misusing , damaging or disrespecting equipment and / or not reporting or paying to have it fixed having done so. Basically I’m taking into context the environment of the installation as well as the technical elements and its age. Landlords have a duty of care to ensure the safety for their tenants and if that means upgrading or replacing a consumer unit that does not have RCD protection even though that may technically fit within code 3 world . After all for say £600 quid or a CU swap with SPD and RCBO’s is that not a good investment for the next 20 years ? It would be replacing something that is really beyond its serviceable life in any case

 

[Pretty Mouth]

My take on supplementary bonding:

I don't think it's intended as a backup CPC, although it can achieve that. Its full name "supplementary protective equipotential bonding", and that it includes extraneous parts, leads me to conclude that its main goal is equal potential. The formula we need to meet is:

R <= 50V/Ia (Ia= operating current of protective device)

When a fault to earth occurs, a circuit is completed through which a current flows. As long as this current flows, there will be a voltage drop between the site of the fault and the MET, so there will be a potential difference between the site of the fault, and all metalwork connected to the MET. The aim of supplementary bonding, as I understand it, it to keep that potential difference to no more than 50V.


Example: electric shower without RCD protection, on a B40. Extraneous pipework in touching distance, connected to the MET via main bonding.

Fault of negligible impedance occurs at shower, 200A flows. There is a voltage drop between the shower and the MET, and therefore a potential difference between the shower and the nearby pipework.

50/200 = 0.25ohms

As long as the resistance between the shower and the pipework is 0.25 ohms or less, then the voltage between them will be no more than 50V.


The above example is for a fault that caused instantaneous disconnection. Supposing a tenuous fault occurred that didn't cause instantaneous disconnection. Let's use a fault current of 199 ohms, as this is right on the edge:

50/199 = 0.251 ohms

So as the fault current gets lower, the max resistance between the shower and pipework to keep the potential down to 50V gets higher. If you meet the formula in 415.2.2, then any fault will either instantly disconnect, or will be kept to 50V or lower.

 

[Falkus]

My take on supplementary bonding:

I don't think it's intended as a backup CPC, although it can achieve that. Its full name "supplementary protective equipotential bonding", and that it includes extraneous parts, leads me to conclude that its main goal is equal potential. The formula we need to meet is:

R <= 50V/Ia (Ia= operating current of protective device)

When a fault to earth occurs, a circuit is completed through which a current flows. As long as this current flows, there will be a voltage drop between the site of the fault and the MET, so there will be a potential difference between the site of the fault, and all metalwork connected to the MET. The aim of supplementary bonding, as I understand it, it to keep that potential difference to no more than 50V.


Example: electric shower without RCD protection, on a B40. Extraneous pipework in touching distance, connected to the MET via main bonding.

Fault of negligible impedance occurs at shower, 200A flows. There is a voltage drop between the shower and the MET, and therefore a potential difference between the shower and the nearby pipework.

50/200 = 0.25ohms

As long as the resistance between the shower and the pipework is 0.25 ohms or less, then the voltage between them will be no more than 50V.


The above example is for a fault that caused instantaneous disconnection. Supposing a tenuous fault occurred that didn't cause instantaneous disconnection. Let's use a fault current of 199 ohms, as this is right on the edge:

50/199 = 0.251 ohms

So as the fault current gets lower, the max resistance between the shower and pipework to keep the potential down to 50V gets higher. If you meet the formula in 415.2.2, then any fault will either instantly disconnect, or will be kept to 50V or lower.

The 50V rule makes sense. In any case a fault current (not overload) will take out the OCPD in 0.4s provided max Zs has been met. So either a short or Ipf of 200ohm and the OPCD surrenders instantly or you’ve got up to 0.4 seconds of pain between 50V and 240V on a linear scale with reference to the resistance between the extraneous conductive part to the MET and the fault current ?

 

[timhoward]

@Pretty Mouth thanks for taking the time to write a textbook answer with good examples.
With so many installations having Dual RCD boards or RCBO boards it's scary how easy it is to forget the details of regs that don't need to be regularly considered day in and day out.

 

[Pretty Mouth]

The 50V rule makes sense. In any case a fault current (not overload) will take out the OCPD in 0.4s provided max Zs has been met. So either a short or Ipf of 200ohm and the OPCD surrenders instantly or you’ve got up to 0.4 seconds of pain between 50V and 240V on a linear scale with reference to the resistance between the extraneous conductive part to the MET and the fault current ?

I think the point of sup bonding is when a fault occurs that doesn't produce enough current to operate the device in the required time, so the victim could be subjected to a voltage for longer than 0.4 seconds. I assume that the sorts of faults that it deals with are considered tolerable in general, but not in the high-risk situations that SB is required (eg bathrooms).

I suppose this would be faults where there is not a direct metal-to-metal short. Perhaps a heating element that has corroded and is connecting to earth through moisture may be of higher impedance, may take some time to disconnect. Or perhaps if there was a problem with low supply voltage.

 

[Pretty Mouth]

@Pretty Mouth thanks for taking the time to write a textbook answer with good examples.
With so many installations having Dual RCD boards or RCBO boards it's scary how easy it is to forget the details of regs that don't need to be regularly considered day in and day out.

You're welcome. I found that it isn't very well explained in the textbooks that I have read, plus it is difficult to get the head around anyway, which may be why many of us aren't 100% on it.

 

[Falkus]

I think the point of sup bonding is when a fault occurs that doesn't produce enough current to operate the device in the required time, so the victim could be subjected to a voltage for longer than 0.4 seconds. I assume that the sorts of faults that it deals with are considered tolerable in general, but not in the high-risk situations that SB is required (eg bathrooms).

I suppose this would be faults where there is not a direct metal-to-metal short. Perhaps a heating element that has corroded and is connecting to earth through moisture may be of higher impedance, may take some time to disconnect. Or perhaps if there was a problem with low supply voltage.

I’m still not convinced that with the parallel paths of the ECP and CPCin a dwelling there would be any significant PD take a standard 3bed semi with the CU under the stairs , this is no further than the loft or airing cupboard In practice from the bathroom in many cases

 

[Pretty Mouth]

I’m still not convinced that with the parallel paths of the ECP and CPCin a dwelling there would be any significant PD take a standard 3bed semi with the CU under the stairs , this is no further than the loft or airing cupboard In practice from the bathroom in many cases

I agree, as in post #2

 

[mainline]

@Falkus and others, out of interest how would you code this one that I’m about to do:

The bathroom light needs changing anyway as it isn’t IP rated.

Must be a very low ceiling to code it for not being ip rated

 

[nicebutdim]

Must be a very low ceiling to code it for not being ip rated

Or a light which drops below 2.25m. I see plenty of pendant sets fitted in new build bathrooms.

 

[mainline]

Or a light which drops below 2.25m. I see plenty of pendant sets fitted in new build bathrooms.

There are plenty of ip rated pendant lights

 

[timhoward]

Must be a very low ceiling to code it for not being ip rated

It is, and I'm not exactly short of reasons to condemn this light!

 

[mainline]

It is, and I'm not exactly short of reasons to condemn this light!

View attachment 116244

I don't think that can be classed as a light any more, more like a jumble of wires hanging out of the ceiling

 

[nicebutdim]

There are plenty of ip rated pendant lights

And there are many more IP20 rated pendant sets hanging in bathrooms.

 

[oscar21]

Surely one of the most important parts of supplementary bonding is to bond the radiator and hot tap/pipe to the cold one. Both the hot tap and radiator pipes could well have a voltage introduced onto them via things like the boiler electrics or immersion heater, the cold could well be isolated from all other pipework yet be at earth potential due to the cold supply pipe coming out of the ground, you could well have 230V across the two taps.

The hot/cold/radiator pipes were the only things that needed to be bonded when I started out and of course there wasn't any RCD on immersion and lights back then, only sockets, so it made sense. It was only later that we had to bond the lights/towel rad/shower etc along with the pipework, this made less sense to me than doing the pipework as everything was effectively bonded together at the consumer unit anyway. I get that a 4mm earth will have lower resistance than the cpc but you were allowed to use the cpc of the thing in the bathroom you were bonding anyway, you could bond the electric towel rad at the spur outside, not directly to the rad and the 1mm cpc acted as the bond, bizarre really some of the regs they come up with.