Thanks for the comments.

You say you do not map fire melts. But "fire melts" are also effective in de-energizing that particular branch circuit downstream from the point of the severing. So don't they also provide information, namely arcing that is found downstream of a "fire melt" on the same circuit must have occurred prior to the "fire melt"?

Regarding the detection of non-severing arcs, I can see that they obviously show the conductor was energized at that location at the time of the arcing, but beyond that they don't help as they may have left the entire circuit still energized (since they didn't sever) or they may have left the entire circuit de-energized by tripping the OCPD.

Regarding the general guidance that "the arcing furthest from the power happened earlier", isn't the most important variable the route that the fire took as it grew, was later suppressed and finally extinguished (which is what you are trying to determine in the first place with the arc survey)? To give an extreme counter example that I have actually seen, a fire started in a HVAC component in a residence. That fire progressed throughout the cold and hot air ventilation paths and caused arcing (both severing and non-severing) on the opposite side of the house two floors above the basement point of origin, even though the breaker panel approximately 10' away from the point of origin was never attacked (something like 29 out of 32 breakers were tripped although the panel and breakers were intact and still functional). The arcing furthest from the power guidance would have indicated the point of origin on the wrong floor and wrong side of the house. Fortunately, the other fire patterns made the point of origin clear (which wouldn't have been the case in a more severe fire that would have left the entire interior attacked but still amenable to arc survey analysis).

Regards,

Mike Learmonth