back on the 'surge'

In big handfuls, yes, actually almost any wiring has the potential to have currents induced in it in a way that causes volts to pile up in a bad way somewhere. The chances of those induced signals being large enough to cause problems, enough that they need to be clipped or filtered out in some way depends very much on what is being protected from what. There is no magic about the DNO wiring, except there is a lot more of it and you have less control of where it goes, so it makes a bigger ‘antenna’ and may also share with sources of hostile waveforms on other premises..

So what raises or lowers the level of risk -  elevated wiring is more exposed than buried, and as Graham points out in the limit you could be providing the earth path for a lightning conductor.

That is not the average garden light however, unless perhaps it is a floodlighting tower.

Things with a larger thermal mass, transformers, motors etc do much better against surges then electronics, generally. So be less worried about an outside light spurred off the supply to a simple pump, than to one with a complex computer control.

In your ascii- art the ** is indeed one place you may add additional suppression. But it can only clamp the voltage on LNE to be close to each other.  On its own it cannot prevent currents that flow in on one side of the building and out the other. LNE may be close to each other at the origin and also the exit of a large building, but the two locations may be kV apart for a few microseconds during an impulse event. Despite the odd journal article to the contrary a fatter CPC between the 2 locations does not help much, as there is not time for the electrons to get to the other end and back and realise they are connected.

The problem is that the answer is very layout dependent, and there is a lot of advice out there that is only good in a limited set of circumstances.

On a 230V system, most cases are on the edge of not mattering - which is why folk are only bothering about it now, though there is more vulnerable kit out there than a few decades ago, so the price of failure is slowly getting higher..

Putting real numbers on a situation is very hard without an awful lot of information, most of which you never have.

Mike.

In big handfuls, yes, actually almost any wiring has the potential to have currents induced in it in a way that causes volts to pile up in a bad way somewhere. The chances of those induced signals being large enough to cause problems, enough that they need to be clipped or filtered out in some way depends very much on what is being protected from what. There is no magic about the DNO wiring, except there is a lot more of it and you have less control of where it goes, so it makes a bigger ‘antenna’ and may also share with sources of hostile waveforms on other premises..

So what raises or lowers the level of risk -  elevated wiring is more exposed than buried, and as Graham points out in the limit you could be providing the earth path for a lightning conductor.

That is not the average garden light however, unless perhaps it is a floodlighting tower.

Things with a larger thermal mass, transformers, motors etc do much better against surges then electronics, generally. So be less worried about an outside light spurred off the supply to a simple pump, than to one with a complex computer control.

In your ascii- art the ** is indeed one place you may add additional suppression. But it can only clamp the voltage on LNE to be close to each other.  On its own it cannot prevent currents that flow in on one side of the building and out the other. LNE may be close to each other at the origin and also the exit of a large building, but the two locations may be kV apart for a few microseconds during an impulse event. Despite the odd journal article to the contrary a fatter CPC between the 2 locations does not help much, as there is not time for the electrons to get to the other end and back and realise they are connected.

The problem is that the answer is very layout dependent, and there is a lot of advice out there that is only good in a limited set of circumstances.

On a 230V system, most cases are on the edge of not mattering - which is why folk are only bothering about it now, though there is more vulnerable kit out there than a few decades ago, so the price of failure is slowly getting higher..

Putting real numbers on a situation is very hard without an awful lot of information, most of which you never have.

Mike.