Thanks Julie could you expand a bit. I currently have TN-C-S supply to pond which on 1x rcd testing at pond disconnects within 300ms. With a PEN fault introduced at the supply, the 1x rcd test will still disconnect as fault current is taken by the parallel earth rod installed at the MET so wheres the danger of shock at the pond end via the extraneous frame with the supply disconnected?
There are actually a whole host of considerations behind the need to/to not bond , so it pays to understand these in order to understand why things are done the way they are.
So, let's imagine your house has its supply from a substation 100m away.
Ignore the line and neutral, when you consider the earth cable, that actually is at the potential of the earth at the substation - not necessarily the same as the ground potential at the location of the house.
Now imagine that your water pipes are metallic throughout, including the underground supply, which connects to your neighbour. Unfortunately your neighbour has a fault, a live conductor has contacted the water pipework.
If you were now to grab your kettle (at zero voltage due to the earth/CPC connection to the substation) and turn the tap (now at circ 230V due to the fault in your neighbour's house) you would receive this 230V!
What's worse is that this has nothing to do with your own supply, in this case, your kettle is off, so disconnecting the line or/and neutral will do nothing - the danger is coming in to your property via the pipes (an ECP)
This is the principal reason why we bond the water/gas etc at the point of entry - in this scenario, the bonding at your (and everyone else's) property safely sinks the voltage away, likely tripping the supply in your neighbour's property.
Of course there are many reasons why the pipework or local ground in general raises in voltage due to faults in supply cables maybe at high voltage etc etc this bonding keeps you safe by bringing all the potential sources of an induced voltage together at the same voltage.
However now the pipework is interconnected between properties then in the case of a fault between the substation and a group of properties (PEN fault) there could be substantial current flowing in the bonding conductor - hence the seemingly large size.
In your case, of a remote ECP the situation is kind of reversed, if one was stood on the ground local to the pool, and a fault elsewhere was to raise the potential of the ground to say 200V if you now touched the metalwork which is connected to zero volts -again you would receive 200V - and it's nothing to do with your actual 230V supply.
So bonding metallic stuff out in the wild when it doesn't have a good connection to the local ground would be a bad idea
(As an aside, this is why class 2 equipment is preferred, and why most power tools are now made class 2, when used outside if there was an induced voltage the class 1 CPC connection would present the same danger as described above)
If the metalwork does have a good connection to the local ground then the local ground will be brought to zero volts , but a substantial current could flow.
With TT you would be at that 200V when stood at the pool, but so would any local metalwork, in this case you need to ensure that contact with the earth connection from the substation is prevented.
The key issue about the whole bonding thing, is most of the time the danger isn't from your own supply, so disconnecting that won't solve anything.
Of course one still needs proper protection for your own supply in the case of a fault actually to do with your own supply.
Sorry bit of a long and boring post!
