iowens said:
Is the rule you refer to above an NEC rule ?
No, its my slight derivation of a real rule from what used to be the BS7671, a/k/a the IEE wiring regulations, which now has some European number. Chapter 46, which (if google is to be believed) starts "Every circuit shall be capable of being isolated from each of the live supply conductors; except that ... the neutral conductor in a TN-S or TN-C-S need not be switched or isolated if the neutral can be reliably be regarded as being at earth potential." In reading this, recall that I noted above that the neutral is regarded as being a live conductor (except in the USA!), and thus "each of the live supply conductors" includes the neutral.
Basically, with a TN-C-S (or TN-S) there is an assumption that there is a very limited potential difference between N and G, and thus a N/G shock is unlikely. However with the other methods there could be significant differences between ground and neutral potential, due to the high impedance of the earth path. Thus for isolation purposes, to ensure someone working on something they think is isolated and for them not be exposed to a shock hazard, the N must be switched.
iowens said:
Also, can you clarify what you mean by N & G being bonded nearby. We have a TN-C-S system, but in some cases the transformer is located in a substation, and the panelboard is 300m away.
"Nearby" usually means something like "in the same building". TN-C-S is substantially the same as The American Way, where in the service entrance one has a neutral / ground bond.
You are not entirely clear in your 300m example where the earth bond is. If the feed from the substation is four wire, and in your panel 300m away you have a N/E bond, then you've got TN-C-S. If however you have a five wire feed from the substation 300m away and no bond in your panel then you have a TN-S supply.
One of the key benefits of TN-C-S is that the effective impedance of a short to ground is very low, and thus large fault currents flow, which ensures a rapid opening of the protective device. If you are going by 240V-land regs, one has to ensure that the fault current (the "Prospective Short Circuit Current") is thusly sufficient. If it is not, one must fit an RCD (a/k/a GFCI).
This is all Pierre's fault, drifting into 'overseas' electrics
