I'm paraphrasing something I read here a while back, but someone made the comment that it requires an enormous amount of complex signal processing equipment to reliably detect arcing signatures.
So somehow, the manufacturers managed to miniaturize all the processing equipment in a single pole breaker case, manufacture it in Mexico, and sell it for $39.95. Seems plausible, right? :roll:
If you consider the computing power that can fit into a pocket sized device - yes it is plausible. After all that is the heart of the AFCI - a small processor.
Will it detect and respond to series and parallel arcing? Yes. Will it detect and respond to
all series and parallel arcing? That is too complicated to answer. For one thing, all arcing will include arcing would be considered normal operation of some loads.
I once read that part of the reason one brand of AFCI will respond differently to same load because they all basically need to have "arc signatures" to which they will respond to essentially programmed into them. That means if a particular signature is not in the data base, then it will ignore it. If a normal operation should contain a signature that is in the data base then it will have unwanted tripping. I have been told there is no actual data base, but this analogy still has some validity. I think for these devices to truly work, we need to be able to tell them a particular condition is an acceptable condition. Of course if that happens many will simply override every tripping condition instead of checking to see if there really is a problem, kind of like how for years when people replaced 15 and 20 amp plug fuses with 30's - they did not blow as often.
I don't know what the solution is. I think the concept of AFCI is a positive one, they are probably far from meeting the objectives though.
And yes a high resistance connection is not an arcing fault. It is just a heater located in a place where it doesn't belong. It may turn into an arcing fault, but damage will have been done first.