Electrical safety check: a stop list, not a clean bill

Heat at a connection is the mechanism, and it builds for months before anything happens

Every stop condition on this form is a proxy for the same physical event. Understanding it is what turns a list of warnings into something you can act on.

A good connection has almost no resistance. Current passes through it and essentially no energy is converted to heat. That is what a properly torqued terminal, a well-made splice and a receptacle gripping a plug all have in common.

A poor connection has a little resistance, and resistance makes heat in proportion to current squared. That squared term is why this failure is so nonlinear: a connection that is fine with a phone charger can be running visibly hot with a space heater on the same outlet, without anything having changed except the load.

Heat then makes the connection worse. Metal expands and contracts, terminals relax, oxide forms on the surfaces, and the resistance rises. More resistance produces more heat, which produces more relaxation. The process runs away slowly — over months or years — and then suddenly.

Which is why every symptom on the form is a heat symptom. The smell is insulation and plastic at temperature. The discoloration is the same thing further along.

The warmth on a faceplate is the process reaching the surface. The spark is the gap that has finally opened. They are four views of one event at four stages.

And why "it stopped smelling" is not reassurance. The smell tracks the load, not the fault. It goes away when the heater is turned off and returns with the next cold evening — by which time the connection is worse than it was.

Nothing about that is a DIY situation. Live electrical work is not a do-it-yourself subject here. Nothing on this site instructs a reader to open, probe or modify energized equipment. The remedy is a competent person with the circuit dead, finding which connection it is. Guessing at it live is how people are hurt.

What the national estimates actually say about electrical fires

Not to alarm you — the honest reading of this data is more reassuring than most safety pages, and more specific about where the risk really sits.

The headline, verbatim. CPSC: “An estimated annual average of 17,100 fires was attributable to electrical distribution equipment (e.g., installed wiring, lighting, etc.). This is 4.9 percent of the estimated annual average number of residential fires for this period.”

The same passage continues: “The annual average death estimate is 150 (6.0 percent of average annual estimated residential fire deaths); and the injury estimates averaged 680 (6.1 percent of the estimated annual average of residential fire injuries).”

Put in proportion. The same report attributes 157,100 fires a year to cooking equipment — about 45 percent of residential fires — and 37,600 to heating and cooling equipment. Electrical distribution equipment is a much smaller category than its reputation suggests.

But the ratio of deaths to fires is the part to notice. 6.0 percent of the deaths from 4.9 percent of the fires.

That is our arithmetic on CPSC’s own percentages, and our reading of why is this: a kitchen fire starts in front of somebody who is standing there.

A fire inside a wall starts where nobody is, often at night, and is well established before anything is noticed.

The chart says the same thing again. Installed wiring is the largest category by a wide margin, with receptacles, switches, cords and plugs behind it. Those are precisely the parts of an electrical system a homeowner cannot inspect — which is the honest reason this tool refuses to give anyone a clean bill.

What that means for this form. Four questions can catch a fault that has reached the surface. They cannot catch one that has not. That is a limitation of the questions, not of your house, and it is why the "no symptoms" result is worded the way it is.

A device that trips is usually a device that worked

The most common request an electrician gets after a renovation is to make a new protective device stop tripping. It is worth knowing what is being asked for.

GFCI protection, and what it is watching. NEC 210.8(A): “All 125-volt through 250-volt receptacles installed in the locations specified in 210.8(A)(1) through (A)(11) and supplied by single-phase branch circuits rated 150 volts or less to ground shall have ground-fault circuit-interrupter protection for personnel.”

It compares current going out with current coming back. A difference means some of it is returning by another route — through a wet floor, through a person — and it opens in a fraction of a second. It protects people, not equipment.

Its reach is wider than people remember. The scope is 125 V through 250 V, which is why 240 V equipment receptacles in these areas are now caught by a rule many homeowners still remember as a 120 V bathroom rule.

In the 2023 edition the qualifier "where the receptacles are installed to serve the countertop surfaces" was removed from the kitchen item, so a receptacle in a dwelling-unit kitchen is in scope whether or not it serves a countertop.

The 2023 edition also added "areas with sinks and permanent provisions for food preparation, beverage preparation, or cooking" as a location.

Somebody whose mental model is "GFCIs are for bathrooms" is working from a much older edition than the one their jurisdiction is likely to have adopted.

AFCI protection, and the different thing it watches. NEC 210.12(B) applies to “All 120-volt, single-phase, 10-, 15-, and 20-ampere branch circuits supplying outlets or devices” across 14 named categories of dwelling area, including kitchens, family rooms, dining rooms, living rooms, parlors, libraries and bedrooms.

It looks for the electrical signature of arcing — the fire mechanism described at the top of this page — rather than for a shock path.

So the two devices answer two different dangers. A GFCI would not notice a series arc in a damaged cord, because the current is still returning correctly.

An AFCI would not notice a person becoming a path to earth. Neither is a substitute for the other, which is why the Code asks for both in overlapping places.

What a trip actually means. Sometimes a genuine fault. Sometimes a wet location doing what wet locations do. Sometimes a shared neutral from an older installation that a modern device finds objectionable.

Sometimes an appliance with a small persistent leakage that was invisible before. All four are findings, and the last three are the reason "the new breaker is faulty" is usually wrong.

And the request to remove it. Asking for protection to be removed is asking to be returned to the condition the requirement exists to prevent. A good electrician will investigate the cause instead, and that investigation is the work worth paying for.