As I responded to Spark123, I'm not sure that one should get so focusssed on the reg's definition of an extraneous-c-p that one loses sight of the whole point of main equipotential bonding. It goes without saying that armour which is not connected at either end is not a practical situation - it's simply a stage in the thought process.I do not think you can consider the consequences of swa armour which is not connected to earth at either end as that would defeat the purpose of having the armour. As, has been said, because it is part of the electrical system it therefore is/can not be considered extraneous and so is earthed to requirements, the same as conduit, trays or trunking etc.
In most situations, one would have no reason to specifically 'suspect' that the armour may have come in contact with the ground - one would simply have an awareness that such a situation might arise at any time, without any indication that it had happened. Safety measures have to be put in place before faults arise - not after they have been detected!! To take your suggestion to its ultimate (ridiculous) conclusion, one would only have to install any main bonding if one 'suspected' (and confirmed by testing) that, due to a fault, the incoming 'earth' potential had risen to appreciably above true earth potential!I suppose if you suspect that the armour is in contact with the ground it could/should be tested and the appropriate action taken to remedy or remove any danger.
Once it is earthed (connected to a CPC), all of the exposed-conductive parts of the installation (and any bonded metalwork) will be, to a lesser or greater extent, 'connected parts'.The simplest of which may be to ensure that no part of the armour and its connected parts are touchable metal.
As I keep saying, it all depends upon how large a CSA of the connection (whatever one calls it) between SWA armour (or any other 'extraneous-c-p') and MET is really (rather than per regs) required in order to reliably achieve the required equipotentiality - if a CSA possibly as low as 1mm² is considered adequate (which it probably would, provided it survived) then, electrically speaking, there is not an issue. As always, the issue which people discuss (and the reason for the CSA of MPBs we use) is the extremely unlikely one (almost, but not quite,'hypothetical') of a very high current flowing in the 'bonding' conductor (the conductor joining the path to true earth with the supply 'earth') due to a fault, with the possiblity of melting a low-CSA conductor or, at least, producing significant PDs across it.
As I've said, if it were (for whatever reason) armour of an unused cable that was terminated in an accessible metal box, you might feel that, technically speaking, it required main bonding at the point of entry. The question therefore arises as to how large a CSA connection between that box and the MET would make you feel that 'additional' main bonding was not required.I do not see bonding at the point of entry to be sensible.
That's really not a very appropriate analogy. It's obviously a very special case, in that it is the very conductor to which one bonds everything else. I'm not even sure what you would bond it to - itself?! A better, allbeit very hypothetical (and unlikely) analogy would be if a supply pipe entered the premises and was immediately connected to some electrically-powered Class I something. That would put the pipework in continuity with the installation's CPCs but I don't think many people would believe that this removed the need for proper main bonding, would they?This would be akin to you bonding your earthing conductor at its point of entry (ignoring the fact that you have other better earthing through pipes).
Kind Regards, John