MPB back to the MET or CU?

[JohnW2]

There is, but a single building can contain multiple installations.

I think that's where the confusion (and quite probably studentspark's confusion) arises, since there is no clear definition of 'an installation' anywhere - the BS7671 definition is, to say the least, less than useful in this context!

As studentspark has intimated, is not what matters (safety-wise) is at the level of 'the dwelling' (flat, in this case), and that the important think is that the dwelling (flat) should be an equipotential zone?

In other words, is it not simply the case that any extraneous-c-p (extraneous to the flat) which enters that dwelling (flat) should be bonded to the 'local' CPCs (i.e. the 'local MET', within the flat) and that anything which looks like a bonding conductor travelling to somewhere outside of the flat (e.g. to "the building's MET", if it has one) is irrelevant (other than that, if that conductor is 'exposed' anywhere, it would constitute an extraneous-c-p which therefore needed to be 'bonded' to the local MET)?

Kind Regards, John

 

[Risteard]

I also put MEC for Main Earthing Conductor. This is not actually correct as it is THE Earthing Conductor.

Absolutely. I detest use of the term "Main Earthing Conductor" as there is only one. There is no supplementary Earthing Conductor.

I am, however, pleased that you corrected yourself before I had to.

 

[plugwash]

Lets take a step back and ask *why* BS7671 specifies minimum sizes for the bonding on a TN-C-S system.

It's clearly not about the "quality" of the equipotential zone, If it was then they would be specifying maximum resistances rather than minimum sizes.

Therefore it must be about preventing damage due to current in the bonding conductors. Either due to the conductors forming a parallel path with the PEN or due to a PEN fault diverting all neutral current from the installation via the suppliers neutral-earth bond and the bonding conductors.

Therefore I would argue that the prescribed bonding conductor size should be maintained all the way back to the supplier's earth terminal.

 

[EFLImpudence]

I am, however, pleased that you corrected yourself before I had to.

Phew!

 

[JohnW2]

Lets take a step back and ask *why* BS7671 specifies minimum sizes for the bonding on a TN-C-S system. It's clearly not about the "quality" of the equipotential zone, If it was then they would be specifying maximum resistances rather than minimum sizes.

Essentially agreed.

Therefore it must be about preventing damage due to current in the bonding conductors. Either due to the conductors forming a parallel path with the PEN or due to a PEN fault diverting all neutral current from the installation via the suppliers neutral-earth bond and the bonding conductors.

Indeed - but, assuming that there is just one DNO supply for the building (i.e. one PEN, hence one "DNO earth terminal"), that applies to the bonding conductor from the (one) DNO's earth terminal to any extraneous-c-ps which enter the building at the point where they enter the building - and those bonding conductors need to be large enough to carry (between those two points) all of the possible current due to the sort of faults which you mention.

Therefore I would argue that the prescribed bonding conductor size should be maintained all the way back to the supplier's earth terminal.

Provided that, as above, all extraneous-c-ps entering the building are bonded to the DNOs earth terminal where they enter the building by a conductor large enough to take the possible currents under fault conditions, then I can see no need/point (from the point of view of 'bonding') having some other conductor (of any CSA) going from the same "DNO's earth terminal" to a point on the extraneous-c-p downstream of where it is already adequately main-bonded ("to the DNO's earth terminal").

If some extraneous-c-p entered the building through the flat, and was not main-bonded anywhere else within the building, that would obvioulsy be different - but that would be a very unusual situation, and certainly not the situation which we (I !!) am discussing!

There obviously has to be a conductor from the DNO's earth terminal to the installation within each flat (the flat's 'local MET'), in order to provide the required fault protection (i.e. adequately low Zs in the flat) and that may well have to be bonded to conductive parts that are extraneous to the flat (and which may well be in electrical continuity with the parts that have already been main-bonded where they enter the building). In that situation, the larger one makes the CSA (hence the lower the resistance) of that path from DNO's earth terminal to flat (and any 'local bonding') the greater a proportion of currents due to supply-side faults fault currents to which you refer will go through the conductor going to the flat (in parallel with the path of the 'actual' main bonding). In that sense, one might even argue that it is better not to have a conductor of CSA appreciably larger than that needed to ensure an adequately low Zs (in the flat).

In short terms, what I'm really saying is simply that, as I see it, an extraneous-c-p entering a building only needs to be main-bonded (to the DNO's earth terminal) once, and that as close as possible to where it enters the building. IF that is done, then any additional bits of G/Y attached to conductive parts within one of the flats is not 'main bonding' (since that has been dealt with elsewhere), but may well be required to create an ensured equipotential zone within the flat. I think EFLI referred to it as 'supplementary bonding', but I'm not convinced that even that is the totally correct terminology.


Kind Regards, John

 

[EFLImpudence]

Don't forget that 'supplementary bonding' just means 'additional bonding' where required.

Supplementary bonding is not a 'thing' in itself.

 

[JohnW2]

Don't forget that 'supplementary bonding' just means 'additional bonding' where required. Supplementary bonding is not a 'thing' in itself.

True, and even BS7671 does not seek to define it, but various things in BS7671 seem to have resulted in many people coming to regard it as 'a thing'.

As I'm sure you understand, what I was talking about was "non-main bonding", and maybe it would be less potentially confusing if we just called it that? !

All 'bonding' seeks simply to minimise PDs and establish equipotential zones but, as plugwash has pointed out, under certain rare supply-side fault conditions, main bonding can find itself having to carry very high currents.

Kind Regards, John

 

[studentspark]

A 6mm² earth conductor may comply with that theoretical calculation, but it won't comply with most DNOs requirements or where that same conductor is used as a bonding conductor, which it very often will be.

That is the essence of my question. Is the 6mm connection also being used as a bonding connection.


Where you find 16mm2 T&E as the supply from a switch fuse to a CU. On a EICR, what code would that be?

Perhaps it would be simplest if you showed us the 'working' of your adiabatic calculation?

Well I was thinking the fuse would be a 1361 type 2 - But these are not listed in 7671 time current graphs. So I went with the BS 88-3, which is a 63 A fuse.
Though in the old switch fuses they were 60A.

With a Zdb of 0.21Ω I calculated a fault current of 1095A As this is the Zs reading I am taking this as the PEFC

To cause a 63A BS 88-3 fuse to trip in 0.1 seconds takes 820A (600A for 0.4 seconds)


So (Square root) 1095 x 1095 x 0.1 / 115 = 3.01 mm So that would be 4mm2 Protective conductor required.

I went with the 'On site measurements' (Even though this is not a real senario) because if you are testing to see if the protective conductor can withstand fault currents. you will be wanting to
test against the real world situation and not values form a book.

So I went with 0.1 for disconnection time rather than 0.4 as the PEFC suggested a fault current greater than the 820A value in Fig 3A1 7671




This all stemmed from a post I was reading on the IET forum. (It dated back to 2008, so could not ask the poster)

Not sure if I am allowed to post external links so will summarise..

The question was about is the CPC in a 16mm2 twin and Earth cable acceptable

The answer in part was...


Is it also a bonding conductor (or at least the main earthing conductor) - - If this is TNCS, your first requirement is to comply with Table 54.8 - then proceed to an adiabatic check.



1 - it is the main earthing conductor for the installation so needs to meet the requirements of the adiabatic expression or Table 54.7 if you don't want to calculate

2 - it is also a part of the bonding conductors so needs to meet the the requirements of 54.8. '

 

[JohnW2]

That is the essence of my question. Is the 6mm connection also being used as a bonding connection.

See below - if extraneous-c-ps entering the building are (as required) main-bonded close to their entry into the building, then any protective conductor going to a flat will not be being used as a main bonding conductor.

Where you find 16mm2 T&E as the supply from a switch fuse to a CU. On a EICR, what code would that be?

If, and only if, it IS being used as a main bonding conductor, that surely depends upon the adequacy of its CPC (as determined by the requirements of BS7671 {for TN-C-S}). If it's not adequate, I presume that it would probably be a C2, wouldn't it? Any additional 'rules' that DNOs may have is irrelevant to an EICR - which is undertaken in relation to BS7671.

Well I was thinking the fuse would be a 1361 type 2 - But these are not listed in 7671 time current graphs. So I went with the BS 88-3, which is a 63 A fuse. Though in the old switch fuses they were 60A.
With a Zdb of 0.21Ω I calculated a fault current of 1095A As this is the Zs reading I am taking this as the PEFC
To cause a 63A BS 88-3 fuse to trip in 0.1 seconds takes 820A (600A for 0.4 seconds)
So (Square root) 1095 x 1095 x 0.1 / 115 = 3.01 mm So that would be 4mm2 Protective conductor required.

[I had forgotten that we're talking about fuses, but that does make it a little easier than with MCBs, because of the shape of the curves ]

What you've done will give a 'safe' estimate of the minimum required CSA, but it will probably be a considerable over-estimate of what is actually needed.

The BS7671 curve for a 63A BS88-3 fuse stops at 0.1s but, if one assumes that the curve continues down in the way it is going at that point, by extrapolation it looks as if the disconnection time at 1,095A would be around 0.01 sec, probably a bit less. If one repeats your calculation for 0.01s it would become:

[ √(1095 x 1095 x 0.01) ] /115 = 0.95 mm²

So I went with 0.1 for disconnection time rather than 0.4 as the PEFC suggested a fault current greater than the 820A value in Fig 3A1 7671

The required CSA of the protective conductor depends upon the actual I²t (given the actual PEFC and disconnection time at that current for the OPD concerned), not the 'regulatory maximum disconnection time' you are designing for. Hence, as above, a 'proper' calculation of required protective conductor size would use a figure of around 0.01s for your PEFC and fuse, not "0.1s" (or "0.4s").

This all stemmed from a post I was reading on the IET forum. ... The answer in part was...
Is it also a bonding conductor (or at least the main earthing conductor) - - If this is TNCS, your first requirement is to comply with Table 54.8 - then proceed to an adiabatic check.
1 - it is the main earthing conductor for the installation so needs to meet the requirements of the adiabatic expression or Table 54.7 if you don't want to calculate .... 2 - it is also a part of the bonding conductors so needs to meet the the requirements of 54.8. '

As I've been trying to explain in my responses to others (particularly flameport and plugwash), in the scenario you have described, everything really depends upon whether any extraneous-c-ps (e.g. gas pipes) which enter the building are (as they should be) main-bonded to the DNO's earth terminal close to where the extraneous-c-p enters the building, then no other 'main bonding' is required, and no other protective conductors within the building are 'main bonding conductors'.

As I've said, if the gas pipework enters a flat, that will be extraneous to the flat, and therefore will need to be bonded to the flat's CPCs (at the 'local MET') in order to ensure that the flat is an equipotential zone. The 'earth conductor' from the DNO's earth terminal to the flat (necessary for fault protection) will then be in parallel with the actual main-bonding of that pipe, and therefore will carry a proportion of the current due to, say, a CNE supply-side fault.

In practice, the length of the true main bonding conductor is likely to be very much shorter (hence lower resistance) than the length of the 'earth' connection to the flat so, even if that were of the same CSA as the main bonding conductor, it's likely that only a small proportion of the fault current would flow through that conductor - as I said, rather ironically, the larger the CSA of the earth connection to the flat, the greater will be the proportion of the fault current that it will carry.

Does any of that help?

Kind Regards, John

 

[studentspark]

Thanks John. That really helps.

 

[JohnW2]

Thanks John. That really helps.

You're welcome.

I could, perhaps, have expanded a little on my comment that determining I²t (for an adiabatic calculation) is easier with fuses than with MCBs.

As I said, the I²t one needs for an adiabatic calculation is the actual I²t figure for the actual PEFC and the actual OPD protecting it.

The (t/I) curves (as those in BS7671) for fuses change fairly 'gradually' - so that, as I illustrated, one can ascertain (if necessarily by extrapolation, if the curves are 'cut off') the actual disconnection time for the PEFC one has - hence a fairly 'correct' value of I²t (at that current) can be determined.

However, for MCBs the presented (t/I) curve (as in BS7671) is essentially 'vertical' for disconnection times below about 12 seconds (corresponding to, say, a current of 5 x In, for a B-curve MCB) in other words, those curves give no real clue as to what current will result in what disconnection time below that point (i.e. when PEFC is greater than 5 x In, for a B-curve MCB). One is therefore reliant on data published by the MCB manufacturer to indicate what the actual I²t will be for any particular PEFC - curves such as (rather complicated and confusing!) this one for Wylex B-curve MCBs:

I hope that might clarify a little!

Kind Regards, John

 

[studentspark]

Thanks John