re Earth problem. A quicker way (well when I do it) Disconnect obviously (before some brite spark points out the need) Break the earth at some point, ideally the DB. (You could try at any socket but there may be covert earthing loops), connect ONE cpc to live. Check in which other sockets E-L resistance is low. Those that are are on one leg upto the break. For confidence disconnect the first cpc and connect the second cpc then check again. All other sockets should be on the second leg. So the break is between the two legs. In concept this is similar to the Zs test mentioned above but I find that Zs is very variable anyway.
The RCD trip is conceptually nothng to do with the broken earth, unless the broken cable is touching the nutral. Tracking the fault can vary from easy to ***** difficult. If you have a solid nutral to earth low resistance, say 1 ohm or less then you are in luck. Disconnect both leve and nutral at DB and progress around circuit measuring the E-N resistance. the fault will be somewhere close to the lowest resistance. If the N-E resistance is high, more than a few ohms, then its back to the binary chop menthod given by other responders.
