I think "fundamentally flawed" is far too strong

Maybe so. What I meant to suggest is that while most safety approaches are flawed in some way or another - usually it's the case that the underlying theory is sound but it's let down by real-life implementation details - sticky contacts, weak springs, accuracy of measurements, broken wires and so on. Whereas with L-N voltage monitoring for open PEN detection, there's a pretty obvious gap even in the underlying theory (the "fundamentals" in my mind) - thus there are a (small) range of conditions where it cannot detect a broken PEN even when the whole device is in perfect condition.

Certainly the chances of having problems may be small in practice - and additional features such as latching off after an out-of-range voltage event has been detected and c.p.c. current monitoring can help to reduce the risk even further (although 722.411..4.1 doesn't appear to require either of those).

Not ideal, but offers an improvement over doing nothing.

I think that sums it up nicely!

   - Andy.

AJJewsbury said:

Whereas with L-N voltage monitoring for open PEN detection, there's a pretty obvious gap even in the underlying theory (the "fundamentals" in my mind) - thus there are a (small) range of conditions where it cannot detect a broken PEN even when the whole device is in perfect condition.

I mean there is a similar gap in RCDs, the assumption that an indefinite shock current in the range 15-30 mA is safe for the whole population, when the evidence does not really support that.* In both cases there is a probability that a perfectly functioning device does not produce the expected safety outcome, but this may be OK, depending on what that probability is.

*Our limits mostly being based on Dalziel's work, which was arrived at largely by tests on fit, young, American men. Optimal limits for women, the elderly and the very young are believed to be lower than 30 mA.

AJJewsbury said:

Whereas with L-N voltage monitoring for open PEN detection, there's a pretty obvious gap even in the underlying theory (the "fundamentals" in my mind) - thus there are a (small) range of conditions where it cannot detect a broken PEN even when the whole device is in perfect condition.

I mean there is a similar gap in RCDs, the assumption that an indefinite shock current in the range 15-30 mA is safe for the whole population, when the evidence does not really support that.* In both cases there is a probability that a perfectly functioning device does not produce the expected safety outcome, but this may be OK, depending on what that probability is.

*Our limits mostly being based on Dalziel's work, which was arrived at largely by tests on fit, young, American men. Optimal limits for women, the elderly and the very young are believed to be lower than 30 mA.

A 30 mA R.C.D. may not limit the fault current magnitude  to 30 mA or less. The fault current will be limited by the fault resistance path. The R.C.D. will though in most cases reduce the shock period duration.

Z.

Indeed, but the RCD will never disconnect a current of less than its trip threshold, between 15mA and 30mA, and that is the current that is assumed safe to flow for ever.  And as I think RCS it pointing out, there will be some folk for whom that assumption that they can stand up to 29.5mA for ever is in fact not true. Most of us in fact. But the good news is that most earth faults are not in this tricky range and are either a near-instant RCD trip, or not likely to threaten life.

Mike.