Main Protective Bonding to incoming Gas and Water services

I can see both angles here. You can either consider that the entire building consists of one single electrical installation, or as multiple separate installations that just happen to be physically stacked next to each other - the first akin to a big hotel the second more like a terrace of houses.

In the 'one big' installation approach you'd have one single MET for the entire building and main bonding considered at an whole building level - so if the entering pipework was insulating and had no other paths to the outside the main bonding could reasonably be omitted. Even where there is main bonding at the building entry points, it's not uncommon to see (additional) main bonding at a per-flat level - even if in BS 7671 terms it's more akin to supplementary bonding - to deal with any worries about the large distances and impedances involved between flat and building entry points.

In the 'many small' installations approach each flat would have its own MET - which may or may not be directly connected to the METs of other flats. Large buildings may have multiple risers, possibly fed from different transformers. Where the old TN-C to the flat system is still in place, there really can't be any assumption that all the METs are at about the same potential in any event. In this case with metallic pipework between flats, that pipework would be liable to transfer potentials from one flat (installation) to another, so would require main bonding.

The words of 411.3.1.2 aren't "external to the building" but "In each consumer's installation within a building".

   - Andy.

Thanks Andy, 

The second view you offered has most in most common with what is in place. The block of flats is supplied via a single TN-S 3ph supply to a dedicated intake, with bus bar riser supplying a Ryefield every third floor, which in turn feeds each flat. The containment for the lateral supplies from the Ryefields to each flat is original steel conduit (Bonded), which has been rewired within the past 10-15 years, with a separate CPC connected to an MET in each flat, and also back to a single MET within the Ryefield. There are obsolete steel Gas and Water services entering each flat (passing through adjoined properties), which most local authority owned properties have historically also been bonded back to that particular flats MET. The new water supply will be fed from a new stand alone building housing a "pumped supply" installation fed from the intake. There is also the fact that there are still properties that have an original 1960's installation within the block, in unknown states of condition. To not bond those properties at least, arguably goes against the requirements of BS7671? or doesn't it?

So every flat is connected to the same Earthing system, so the situation is exactly as I described. If a building is large enough to require separate supplies, the Earthing systems of each will be very solidly bonded together, for example in a hospital. The reason is simple, this would otherwise be an example of introducing an external potential, which is banned by 411.3.1.2.

There is a slight definition problem here Andy, which is leading to your comments, and that is that you are defining multiple electrical installations (each flat), whereas the building distribution is not in principle separate, just who is paying the bill. I do not think that BS7671 says that these are separate in any way, the metering position is not acting as some kind of break in the installation. The change from DNO (which has different rules) to BS7671 is an obvious break, but the building distribution is also covered by BS7671, so cannot logically be a different installation. The introduction of another fuseboard does not change the status of a section of an installation.

If we had a case where each flat was supplied by a separate DNO supply your argument would be possible, but I struggle to even think of such a case, in any reasonably sized block such would be very unlikely, but perhaps possible in small converted properties. In such a case compliance with 411 would be more tricky without verifying that the supplies all came from the same supply main.

All these obsolete pipes, whether bonded or not, are largely irrelevant, a fault to these is realistically extremely unlikely, and unless deliberately isolated from bonds from the past are probably Earthed, but this is very easily verified. In either case, it is a very tiny risk, and if passing through a bathroom should be bonded anyway from past connections and in any case, the area should also be RCD protected.

The description does not say if the supply is PME, but if so the E-N bond will be at the DNO head, and the installation MET is close to here also.

All this is why I said building in reference to 411.3.1.2, and I think that this wording in the new Brown Book has subtly changed to that which I said, but is probably still not fully clear, having also added consumer into the sentence. A consumer is someone supplied by the DNO, even if that is a BNO who also has consumers connected.

So every flat is connected to the same Earthing system, so the situation is exactly as I described. If a building is large enough to require separate supplies, the Earthing systems of each will be very solidly bonded together, for example in a hospital. The reason is simple, this would otherwise be an example of introducing an external potential, which is banned by 411.3.1.2.

There is a slight definition problem here Andy, which is leading to your comments, and that is that you are defining multiple electrical installations (each flat), whereas the building distribution is not in principle separate, just who is paying the bill. I do not think that BS7671 says that these are separate in any way, the metering position is not acting as some kind of break in the installation. The change from DNO (which has different rules) to BS7671 is an obvious break, but the building distribution is also covered by BS7671, so cannot logically be a different installation. The introduction of another fuseboard does not change the status of a section of an installation.

If we had a case where each flat was supplied by a separate DNO supply your argument would be possible, but I struggle to even think of such a case, in any reasonably sized block such would be very unlikely, but perhaps possible in small converted properties. In such a case compliance with 411 would be more tricky without verifying that the supplies all came from the same supply main.

All these obsolete pipes, whether bonded or not, are largely irrelevant, a fault to these is realistically extremely unlikely, and unless deliberately isolated from bonds from the past are probably Earthed, but this is very easily verified. In either case, it is a very tiny risk, and if passing through a bathroom should be bonded anyway from past connections and in any case, the area should also be RCD protected.

The description does not say if the supply is PME, but if so the E-N bond will be at the DNO head, and the installation MET is close to here also.

All this is why I said building in reference to 411.3.1.2, and I think that this wording in the new Brown Book has subtly changed to that which I said, but is probably still not fully clear, having also added consumer into the sentence. A consumer is someone supplied by the DNO, even if that is a BNO who also has consumers connected.