BS 7671 has some specific requirements for which we need to understand the answer to some questions:

(a) Is the LV supply from a public network or private transformer;

(b) is there a Common Bonding Network in the installation, for example for control or information technology equipment?

(c) Which country is the installation in?

so as to keep touch voltages between accessible metalwork below 50V in the event of an earth fault

There's not a requirement for main bonding to achieve 50V - in many circumstances it's rather unlikely. Even supplementary bonding is only required to limit touch voltages to <50V where the disconnection time exceeds 5s.

the protective bonding not less than half required CPC

Main bonding (i.e. to things that can introduce a potential from outside of the overall installation) is to be sized according to the protective conductor size of the installation (or building if an installation spans more than one building and it's not PME) - so the requirement is generally half the size of the building's earthing conductor, rather than half the c.p.c. of some internal circuit. You may use a c.p.c. as part of the connection between the MET and an extraneous-conductive-part if you wish, but it then needs to satisfy the requirements of a main bonding conductor as well as a c.p.c.  (Supplementary bonding sizes can be smaller, related to local circuit c.p.c.s, but then the extraneous-conductive-part would usually have already been main bonded using full size conductors elsewhere).

   - Andy.

And just for clarity - in BS 7671 terminology an installation only has one MET (Main earth terminal) - earth terminals or bars elsewhere don't get the "Main" prefix.

   - Andy.

LV supply from a private transformer.

I'm not aware there is a common bond network. so let's go with no.

This is a UK installation (England to be precise)

My misunderstanding, i thought protective bonding was to reduce touch voltages to acceptable levels - that being < 50V.

Udtwmc said:

LV supply from a private transformer.

OK, so TN-S then (assuming this is a PNB installation as discussed in GN 8)?
(I will ignore issues regarding TN-C between transformer and swtichboard,  and whether others believe this complies with Reg 8(4) of ESQCR if it really is TN-C, but if it's a private transformer earthed at the transformer, with no backup generator, then it really ought to be TN-S from the transformer.)

If it is TN-S, main protective bonding sized in accordance with Regulation 544.1.1 is half the CSA of the Earthing conductor, subject to a minimum csa of 6 sq mm and max of 25 sq mm.


BUT

If there really is a private TN-C supply circuit that also serves other buildings or parts of the installation, we are outside the realms of what is covered by BS 7671. My tendency would be to go with PME recommendations 544.1.1 (Table 54.8) for main protective bonding size in this instance, but would caution that it's tricky unless the TN-C distribution circuit has been installed to DNO rules (whilst the distribution network portion of PME is TN-C, PME has additional requirements for earthing over and above those used for TN-C in other countries).

Udtwmc said:

I'm not aware there is a common bond network. so let's go with no.

OK, reason for asking is that with CBN inductance (for functional earthing) is more important, and may need other than 25 sq mm round conductors.

I also thought about earthing of pump control if the pump is VSD, because of harmonic currents in the earthing system (high protective conductor currents). The frequencies involved may also be impacted by the armour of the SWA., and that would be a reason for providing additional copper earthing. BUt because of frequencies, if more than a couple of metres, depending on size of VSDs, round copper might not work.

Udtwmc said:

This is a UK installation (England to be precise)

Reason for asking is that, even if BS 7671 is specified, sometimes local wiring codes have additional requirements to consider.

Now, as to whether the control cabinets require main protective bonding ... they are not extraneous-conductive-parts by definition as they  are part of the electrical installation, so the answer is a clear "NO".

Udtwmc said:

My misunderstanding, i thought protective bonding was to reduce touch voltages to acceptable levels - that being < 50V.

BS 7671 does not limit touch voltages in a fault, unless supplementary protective equipotential bonding is applied (Reg 415.2) - usually only applied where ADS disconnection times can't be achieved, or in special locations (bathrooms, which follows the 50 V AC/120 V DC of 415.2, or medical locations where the voltage is reduced to 25 V AC/60 V DC).

My misunderstanding, i thought protective bonding was to reduce touch voltages to acceptable levels - that being < 50V.

and a common misunderstanding too!

In many situations the the voltage at the point of the fault (i.e. on the exposed-conductive-part) can be around half the supply voltage (for TN) (even higher if reduced c.s.a. c.p.c.s are used) - and if R2 is large compared with Ze (as is often the case) main bonding can only make a small improvement in the voltage difference between the exposed- and extraneous-conductive-parts. Even that can be reduced if the exposed-conductive-part itself has a low resistance to Earth or the supply PE elsewhere (e.g. connected to metallic water or gas mains). Hence the emphasis on short disconnection times (0.4s) even within the equipotential zone. While BS 7671 puts requirements on the c.s.a. of main bonding conductors there's no limit on their length - and hence no limit on their impedance - and thus no limit on the voltage that can develop along their length when carrying a given fault current. So main bonding is really a matter of doing the best you can, and anything is better than nothing, but there are no specific guarantees. Main bonding also serves a purpose for shunting diverted N currents around thin c.p.c.s. in PME systems and can sometimes be very effective in TT systems.

Supplementary bonding is a slightly different kettle of fish in that there is a prescribed impedance that they must be below - which is derived from 50V and the fault current needed to open the protective device within 5s (Ia) - but there's no guarantee that the fault current will actually be that low - on the contrary we're normally careful to ensure that earth fault currents exceed what's needed to open a protective device quickly - and even if we have a rare case where the end of the circuit doesn't meet disconnection times, faults can happen anywhere and a fault closer to the supply will inevitably have a higher fault current. So what happens when the fault current exceeds Ia? Then even with supplementary bonding touch voltages can exceed 50V. So how does that work to prevent fatal shocks? Well in general protective device that take up to 5s to open (i.e. fuses or MCBs, rather than RCDs) open progressively more quickly when the fault current increases. Meeting the 5s requirement for a current that produces 50V normally means you're on a curve that'll disconnect within 0.8s when the current is high enough to produce 120V and 0.4 when it could produce 230V. So it's a kind of hybrid solution that doesn't require disconnection at all when faults currents are low (and the resulting touch voltage ≤50V) but falls back on a kind of sliding scale ADS if the current increases above that.

In normal situations you don't have to worry about such details, but it can pay to have them in the back of your mind for the occasional weird setup - e.g. if you're relying on time-delay RCDs for ADS with exceptionally long fixed delays.

   - Andy.

That nomenclature is fine for a "classic" domestic setup with the earthing conductor connected to a 4-way terminal strip beside the meter and continuing to the CU; with gas and water main bonding. However, with utilities coming in plastic pipes, all we have is the earth bar in the CU, which I suppose could be the "main" earth terminal even though it is the only one.

By contrast, once you have separate buildings with separate distribution circuits, and earth terminals in one or more of them, which is the MET? And what do you call the terminals in each building? In a sense, at least in the sense which is commonly understood, each is a MET. "Building earth terminal" and "earth marshalling terminal" or even "building earth marshalling terminal" are alternatives, but they do not particularly appeal to me.

AJJewsbury said:

And just for clarity - in BS 7671 terminology an installation only has one MET (Main earth terminal) - earth terminals or bars elsewhere don't get the "Main" prefix.

I think I agree with Chris Pearson , because main protective bonding connects extraneous-conductive-parts to the MET, and Regulation 411.3.1.2 requires main protective bonding to be applied to each building.

So, there's possibly only one LV MET per building (in some small buildings like control buildings, substations, etc., this may well be a bonding ring conductor) ... in fact, in very large buildings with multiple private transformers, it may well be that there are a number of MET's, which are connected together with a common bonding network of a design relevant to the design of the building and installation(s) contained in it.

For distinct buildings, I prefer the term Building Earth Marshalling Terminal (BEMT) for subsidiary buildings, extraneous-conductive-parts in the remote building being connected to the MET via the building's submain protective conductor, although I can see the argument for calling it a MET as well. Certainly where the outbuilding has a distinct earthing system (e.g. TT'd from a TN main building) calling it a MET makes sense.  But the point I was trying to make was that it can be confusing to label earth bars in each control cabinet within the same building as METs (as in the OP's diagrams), especially when trying to interpret the wording of BS 7671. (Maybe I was being too subtle for once)

  - Andy.