It is an interesting question - a simple series current sensing transformer and high pass filter, which is I believe all that is in the original AFDDs will be unable to tell where in the loop between substation, consumer unit and load and back to substation the faut is, as it is only looking for non-natural fast changes in current, compared to the normal rise-time of the mains waveform.

Equally the original AFDDs tripped out on motors with jumpy brushes, when ham radio gear was in use nearby, and a whole host of false alarm conditions.

It seems the new ones are rather cleverer than that, and if the John Ward and Dave Savery tests are anything to go by it seems they  now only trip out on closely spaced plumbers stop ends made from cast brass, or on graphite to brass junctions. Oh and with PVC insulation prepared with HV 'conditioning'.

Things like what low smoke cables do, and for that matter polyethylene, which apparently does not char in the same way seem to have been overlooked. They may well be a more effective alternative.

Now looking back In that thread from 2 years ago I said
The arc fault detection algorithms are something over which the maker's draw a veil, and to me that means it probably is not finished or very good, as it will be understood and <<have only >> been reviewed by very few people.


Given the total absence of any new supporting data since then, that opinion has not changed.


If the makers cannot tell us, then what we maybe need is someone like JW to reverse the polarity of the neutron flux, and see if the wobbly gap is put in the supply side it trips or not.

It is an interesting question - a simple series current sensing transformer and high pass filter, which is I believe all that is in the original AFDDs will be unable to tell where in the loop between substation, consumer unit and load and back to substation the faut is, as it is only looking for non-natural fast changes in current, compared to the normal rise-time of the mains waveform.

Equally the original AFDDs tripped out on motors with jumpy brushes, when ham radio gear was in use nearby, and a whole host of false alarm conditions.

It seems the new ones are rather cleverer than that, and if the John Ward and Dave Savery tests are anything to go by it seems they  now only trip out on closely spaced plumbers stop ends made from cast brass, or on graphite to brass junctions. Oh and with PVC insulation prepared with HV 'conditioning'.

Things like what low smoke cables do, and for that matter polyethylene, which apparently does not char in the same way seem to have been overlooked. They may well be a more effective alternative.

Now looking back In that thread from 2 years ago I said
The arc fault detection algorithms are something over which the maker's draw a veil, and to me that means it probably is not finished or very good, as it will be understood and <<have only >> been reviewed by very few people.


Given the total absence of any new supporting data since then, that opinion has not changed.


If the makers cannot tell us, then what we maybe need is someone like JW to reverse the polarity of the neutron flux, and see if the wobbly gap is put in the supply side it trips or not.