Ask an electrical engineer to name the failure mode behind a large share of electrical losses, and the answer is unglamorous: a connection that was not tight. Not a design flaw, not an exotic component failure, not a lightning strike. A conductor and a terminal that were supposed to be joined under a specified torque and were not, or were once, and loosened over years of service.
Loose connections matter to claims professionals for two reasons. First, they are among the most common physical findings in electrical failure investigations. Second, they carry liability implications, because a loose connection almost always traces back to a human act: an installation that missed the torque specification, a maintenance program that never checked, or a modification that disturbed a connection and left it degraded.
Why a loose connection becomes a fire hazard
Electricity flowing through a connection encounters resistance. In a properly made connection that resistance is tiny and the heat it produces is negligible. Loosen the connection and the contact area shrinks, resistance climbs, and the connection begins generating heat in proportion to the square of the current flowing through it. Double the current through a degraded connection and the heat roughly quadruples.
The process feeds on itself. Heat oxidizes the contact surfaces, and oxide layers conduct poorly, which raises resistance further, which produces more heat. Thermal cycling from daily load changes expands and contracts the metal, working the connection looser. Over months or years, a connection that started marginally loose can progress to glowing temperatures, hot enough to carbonize insulation and ignite nearby combustibles, all while drawing current levels that never trip a breaker. That last point is worth underlining for anyone evaluating a claim: overcurrent protection is designed to detect too much current, and a glowing connection is not an overcurrent condition. The breaker performing exactly as designed is fully consistent with a connection failure.
Where they hide
Certain locations account for a disproportionate share of connection failures: breaker and panel lugs, particularly on high-load circuits such as HVAC, ranges, dryers, and EV chargers; neutral connections, which carry current continuously and whose failure can produce damaging overvoltages on connected equipment; aluminum branch circuit terminations in structures of a certain era, where dissimilar metal and oxidation issues compound the problem; splices in junction boxes, especially those made during renovations or additions; and utility-side connections at the weatherhead, meter, and service drop, which raises the separate question of whose equipment failed.
That last category matters for subrogation. A failed connection on the utility side of the service point is a different claim than a failed connection in the customer's panel, and the physical evidence usually makes the distinction clear if it is preserved.
What the evidence looks like
In panel and equipment examinations, the terminations are the first place I look, because they are so often where the story is written. A connection that failed thermally leaves a signature: localized melting and damage concentrated at the termination, oxidation and discoloration of the conductor surfaces, insulation damage most severe at the connection point and diminishing with distance, and often deformation of the terminal hardware itself. That damage pattern, concentrated at a connection and diminishing away from it, is different from what external fire attack produces and different from what an overload distributed along a circuit produces. Reading that distinction is a substantial part of electrical failure analysis, and it is why the terminal, the conductor, and the surrounding components need to be preserved together rather than cut apart during cleanup.
Timeline evidence matters too. A connection failure is usually a slow process. Occupant reports of flickering lights, warm outlet covers, buzzing at the panel, or intermittent equipment problems in the weeks before a loss are consistent with a degrading connection and are worth documenting in recorded statements before memories fade.
The liability question
Because connections are made and maintained by people, connection failures raise attribution questions that pure product failures do not. Was the termination torqued to specification at installation? Manufacturers publish torque values, and industry standards address workmanship. Did a later modification disturb the connection? Was the connection inspected as part of any maintenance obligation? Did the load on the circuit change, for example by adding equipment, in a way that pushed a marginal connection over the edge?
These questions have documentary answers: permits, invoices, maintenance records, inspection reports. Pairing the physical evidence with the paper trail is what turns "a connection failed" into "this party's work failed, and here is the evidence."
The smallest defects in an electrical system are often the ones that end up in litigation. A quarter turn of a screw, left undone years earlier, is a finding an engineer can identify, document, and defend.