> Is it still a recommended practice on the US to rely on overcurrent protection to protect against current leakage? It's an horrible practice, as it requires s

> Is it still a recommended practice on the US to rely on overcurrent protection to protect against current leakage? It's an horrible practice, as it requires several amperes of leakage before triggering. Yes. I don’t agree with it, personally — I think that there should be standard ground fault trip curves that are functions of circuit breaker rating such that they coordinate properly, and that they should be used at least on all circuits that extend outside of a building and really on most circuits, period. This would add costs, and it would detect and thus prevent a lot of dubious wiring practices and outright errors. But you should still bond all your “ground wires” together so that even if these breakers fail to trip you have a degree of protection. (And breakers do fail.) Under current NEC (the US code), when you supply a building with a feeder (i.e. from another building or similar structure), you must provide it with all the current carrying wires and separate ground, and you must ground it to the Earth. In any case, if you live in a place where your utility supplies power above ground, and a utility pole or line is hit by lightning, its ground is most certainly solidly connected to your house. So you can’t really pretend that your house’s ground is connected only to the Earth near your house and that it won’t let lightning in. Hmm, I wonder if you can take the ground wire running between two buildings and carefully insulate it and give it a wrap or three around a piece of steel to add some inductance and reduce lightning currents that might otherwise go through it. I’m not saying this is wise…