In my view we should return to TN-S, with 4 core street mains and and with wire armouring used as the earth.

Option one. Old style mains with the wire armouring exposed and in direct contact with the general mass of earth. This approach gives a very low earth resistance since the armouring provides both a metallic earth path back to the transformer AND numerous fortuitous additional earth connections along the cable.

Option two, use the more modern type of 4 core cable with a plastic outer covering. This protects the wire armouring against rust. If selecting this option then extra connections to earth should be provided at each joint.

We could now add to that bring back TN-s campaign, some basic condition monitoring at the substation, and a little box with a SIM card  that rings in its earth leakage current readings, electrode resistances and instantaneous HV and LV load currents, for the price of digging a decent foundation, thus allowing the DNO transformer and cables to be preemptively maintained and overload to be realized before the paint blisters off the transformer.

Unfortunately the same acute constriction of the wallet that gives us the broken PEN problem in the first place probably means this is also off the table. But I agree, it would be good.

The safest alternative is probably to stop worrying about earthing altogether and adopt a similar approach to that now taken for garden appliances  outdoors, and make everything 2 pin and double insulated.

Mike.

If we have no extraneous-conductive-parts in an installation, TN-S could be more dangerous than TN-C-S,  if the PE back to the substation breaks, due to "leakage currents". Sure, commercial/industrial installations with lots of opportunistic Earthing, TN-S is definitely the correct choice ... but perhaps not always for distribution to domestic installations.

In addition, earth fault loop impedances are higher (in general) than TN-C-S.

Overall, I think each of the three usual earthing arrangements has its own advantages and disadvantages.

Conclusion is, though, perhaps BS 7671 should continue to recommend a consumer earth electrode connected to MET for TN systems.

If we have no extraneous-conductive-parts in an installation, TN-S could be more dangerous than TN-C-S,  if the PE back to the substation breaks, due to "leakage currents". Sure, commercial/industrial installations with lots of opportunistic Earthing, TN-S is definitely the correct choice ... but perhaps not always for distribution to domestic installations.

In addition, earth fault loop impedances are higher (in general) than TN-C-S.

Overall, I think each of the three usual earthing arrangements has its own advantages and disadvantages.

Conclusion is, though, perhaps BS 7671 should continue to recommend a consumer earth electrode connected to MET for TN systems.

When you say "recommend" you really mean "should" 

Do DNOs have a policy when connecting power to new properties to add an earth "pigtail" at the point of connection or where the cable exits the ground? Should they? Could be quite easy with a stainless steel braided sheath applied on site attached to the pigtail.

I think they add one where they take the joint from an existing cable, but this could be some way away from the termination.

OlympusMons said:

When you say "recommend" you really mean "should"

In standards speak, 'in a 'specification' (such as BS 7671) should' is synonymous with 'recommended', and such recommendations are not necessary for conformity to the standard. On the other hand, 'shall' is 'normative' and is necessary for conformity to the standard.