Discrimination (RCDs)

[Nick Atty]

A couple of threads recently about RCDs have mentioned the "discrimination" problem - if you have multiple RCDs in series, then you cannot guarantee that any particular one will fail if you have a fault at the end of the circuit.

Now the recommended way of providing a shore line connection (ie, a mains input) to a canal boat is through an on-board RCD.

Obviously, if you plug that into an external socket which is RCD protected (which seems highly likely) then you end up with exactly this situation. So what exactly is the problem with discrimination: is it just fault finding, or what? I suspect the idea here is that the safety provided when there isn't an RCD anywhere else outweighs the discrimination problem when there is.

[John Rumm]

You may or you may not... if it was just the single socket, or for that matter the single circuit feeding the socket that was protected by the RCD then you could argue that you have sufficent discrimination.

The requirement to discriminate is simply to try and ensure that when a fault causes a protective device to trip, the effects of this trip are reasonably localised. The cannonical example is when a RCD trips on a power circuit, it does not also cause the lighting circuit to get de-energised (hence saving you from electrocution but causing you to fall downstairs and break your neck!). That is why use of "whole house" RCDs is depricated.