The biggest problem that I am having right now is the issue if sustaining arcs with enough energy to start a fire even exist at the voltages found in a dwelling unit. Recent information such as the UL report cited in this thread suggest that such arcs do not exist. It is my opinion that most fires that are of electrical origin are cause by poor connections that can produce high temperatures at low current levels. Sure there will be the cases of over heated undersized cords and things like that, but I doubt those things are major causes.
If arcs at that voltage are self sustaining, then why are we not required to have GFP protection on the larger 208/120Y systems like we are required to have on the 480/277Y systems?
And why is arc flash not a problem on 120 volts as it is on 480/277? H'mmmmm.
There have been so many articles written on series arcs as well as glowing arcs that can cause fires just because they can last a long time, and even a AFCI that has a more sensitive detection system but is gated so good arcs will be ignored should be able to extinguish these, but UL is so bent on the premise that the AFCI must extinguish the arc within the first 1/2 cycle so they wont allow a gated detection and series arcs will for ever go undetectable since it is impossible to distinguish between a good arc and a bad arc other then the time duration of most good arc's (switch contacts) being very short.
It doesn't take an engineer to understand that the signature of any arc is a complex chaotic random form of many waves of electrical energy, a good arc as well as a bad arc will produce this same form of chaos that is totally random, the only thing that will change this is the amount of load being drawn across the arcing point and the amount of voltage, but this will apply to both good arcs as well as bad arcs, so how do we make a good arc appear different to the detection device, placing resistors would alter the arcs but that would waste energy, how about placing caps across the good arc, that could work but would only shorten the time it took to extinguish the arc which would shorten up the gate time of detection on an AFCI if UL would allow it, throw inductive loads into the circuit and it all changes again, so having a background in electronics I can see the problems of trying to detect a series arc and how hard it is to detect the bad arc from a good arc.
I think the manufactures thought they could get this solved by the time the AFCI's hit the market but someone didn't think through it enough to realize how hard it was going to be, and they got caught with their pants down by the time they realized how hard it would be to tell the difference between a good arc and a bad arc.
The thing is that they have tried to put a band-aid on it by including basically a RCD or GFP into their design to cover up their short comings of detecting a series arc, if anyone remembers back when they were first introduced into the market the GFP was never disclosed, it was a held secret until we as electricians discovered it because of the problems that mimic a regular GFCI H'mm why was this? now with the combo ones to me they are just using a faster acting breaker to remove the power from a parallel fault faster (I have no fact to this, just my opinion, maybe they will somewhat detect an arcing parallel fault at a higher current level to act faster but I have not tested the trip curves of these new breakers) but the fact remains that they still will not detect a series arc by itself, of course they claim it will once the arc goes on long enough to turn into a parallel arc but that is nothing but a play of words as any fast acting breaker or GFP will also remove power from the circuit once it develops into a parallel arc, but that is not where the fire starts, the fire starts from the series arcs as well as the glowing arcs, just think about it, if a series arc has to evolve to a parallel fault to trip an AFCI think of the heat that has already been generated to melt this much insulation??? it doesn't take much thought to realize that if a fire is possible it will have already begun by the time the AFCI trips.
So as far as the technology working I don't see it, until they can get it to detect a series or glowing arc it will not add to the safety and we will not see any results saying so, because a GFP and faster acting breakers can do the same thing that the AFCI's can now do as they are now designed, the RCD system that Europe has had in use for some years probably has done more for the safety then the AFCI's have ever done, but even they come up short in a few points, with out a EGC in a cord they are not effective for protecting after the receptacle, also since the RCD is in the mains (main breaker) the whole house goes dark rather then just having them on the branch circuits, as well as no extra protection for line to line or line to neutral faults, as far as for faster acting breakers goes well we all know the problem with motor and transformer loads and a faster acting breaker just doesn't work in these cases.
Sorry for the ramble but most know my feelings on this subject and it has been going on as far back as when they were first put into the code, and my short experience with them in 2002 until we removed them from our state code the problems they caused and the cost for our customers because of these problems without no foreseen added safety when it was discovered that they would not in-fact detect a series arc as they were first claimed by the manufactures and UL, if they would have worked I would have been all over it to install them, but I don't go for smoke and mirrors much less snake oil as it turned out.
I would have loved for them to work, but as I noted above relying on a series arc to burn long enough to cause a parallel fault is not a reasonable excuse to now say they will detect a series arc because if a fire is possible at this point in the wiring then it would have already started.
