mapj1:

Hmm. Now BS7671 is not a legal requirement, and there are plenty of situations were it is not applied, and installations can be perfectly safe.  The DNO side of the wiring comes to mind as the most obvious that most folk will have come accross, but as one who has worked in both university and commercial R and D establishments,  I can say that there are several situations where the RCD and for that matter the MCB/fusing requirements of the regs are simply inappropriate, and a properly engineered approach specific to the situation is needed. It is not sensible to demand that the regulations are modified to cover all possible cases.

If this was an industrial setting with a difficult load, I'd be well on the way to  risk assessing the RCD out of circuit,  or fitting an earth fault relay set to a high level, and using  cables with earthed armour or braids, and perhaps  additional bonding to cases of class 1 equipment, to give an equivalent protection. By this I mean that more than 2 credible and independent faults must occur to reach a dangerous state. In such a case, the paperwork is not quite the BS7671, but incudes the design authority calculations and description documents and risk assessment, that shows how the credible faults are safely mitigated.


This however is more the case of an enthusiastic amateur, and I agree this is trickier. First eliminate the possibility there is a real fault. Then confirm it is earth leakage that is causing the problem.  Could the conduit take another set of cables, and split the load over two circuits ? This adds a slight risk if there is ever a  very odd fault occurs between the two circuits, but should halve the stress on any one RCD.

You may decide to remove the RCD (or change it for one that is not providing protection at <30mA) but this needs a savvy customer who understands what the increase in risk actually means, and if you belong to a part P notification scheme that insists you must follow BS7671, then you may not be able to.

It is of course no more dangerous than any number of installations in current use, problems arise if there is a faulty load connected, and a person is connected between L and E, and the RCD should trip and it does not.


 

mapj1:

Hmm. Now BS7671 is not a legal requirement, and there are plenty of situations were it is not applied, and installations can be perfectly safe.  The DNO side of the wiring comes to mind as the most obvious that most folk will have come accross, but as one who has worked in both university and commercial R and D establishments,  I can say that there are several situations where the RCD and for that matter the MCB/fusing requirements of the regs are simply inappropriate, and a properly engineered approach specific to the situation is needed. It is not sensible to demand that the regulations are modified to cover all possible cases.

If this was an industrial setting with a difficult load, I'd be well on the way to  risk assessing the RCD out of circuit,  or fitting an earth fault relay set to a high level, and using  cables with earthed armour or braids, and perhaps  additional bonding to cases of class 1 equipment, to give an equivalent protection. By this I mean that more than 2 credible and independent faults must occur to reach a dangerous state. In such a case, the paperwork is not quite the BS7671, but incudes the design authority calculations and description documents and risk assessment, that shows how the credible faults are safely mitigated.


This however is more the case of an enthusiastic amateur, and I agree this is trickier. First eliminate the possibility there is a real fault. Then confirm it is earth leakage that is causing the problem.  Could the conduit take another set of cables, and split the load over two circuits ? This adds a slight risk if there is ever a  very odd fault occurs between the two circuits, but should halve the stress on any one RCD.

You may decide to remove the RCD (or change it for one that is not providing protection at <30mA) but this needs a savvy customer who understands what the increase in risk actually means, and if you belong to a part P notification scheme that insists you must follow BS7671, then you may not be able to.

It is of course no more dangerous than any number of installations in current use, problems arise if there is a faulty load connected, and a person is connected between L and E, and the RCD should trip and it does not.