The behaviour of an N-E fault is complex and can seem unpredictable. There are paths of different resistances to both neutral and earth for current from the faulty circuit, from other circuits and even from neutrals of other installations when there's no load on this one. How each of the currents divides can be hard to predict; sometimes they will cause nuisance tripping when there is heavy load on, sometimes just failure to test, sometimes the RCD won't reset at all.
What they should do consistently, though, is fail insulation tests, so I am a little surprised that wasn't picked up first.
What they should do consistently, though, is fail insulation tests, so I am a little surprised that wasn't picked up first.
