Pilot Light That Will Not Stay Lit in Kansas City

A pilot that lights, burns steadily for ten or twenty seconds, and then dies the moment the knob is released is the most recognizable failure in the whole appliance. It is also one of the very few where the diagnosis is largely settled by counting seconds.

What follows explains the tiny electrical circuit that holds the fuel open, then works outward through the things that interfere with it: a blocked intake screen, a poor draft, and finally the control itself. Anyone weighing this repair against the age of the unit will find the general case for and against KC water heater repair set out on the front page.

The tiny circuit that holds the gas open

A standing pilot is not merely a flame. It is one half of a self-checking safety system, and the other half is a thermocouple, a probe made of two dissimilar metals joined at the tip. Heat that junction and it generates a voltage measured in thousandths of a volt.

That voltage feeds an electromagnet inside the gas control. The magnet holds a small valve open, and that valve is the only thing letting fuel reach the pilot at all. No household current is involved anywhere. The flame is quite literally holding its own supply open.

The logic is deliberate. Lose the flame for any reason and the probe cools, the voltage collapses, the magnet releases, and fuel stops within seconds. That is why a draft blowing out a pilot does not fill a basement with unburned gas.

A healthy probe produces somewhere between twenty five and thirty five millivolts with nothing connected to it. A tired one may still manage ten or fifteen, which is enough to look alive on a meter and nowhere near enough to hold the magnet. That in-between state is what produces the classic complaint.

Count the seconds after you release the knob

Immediate loss, within two or three seconds, points at a probe tip sitting outside the flame envelope or a connection at the control that has worked loose. Position matters here: the tip belongs in the hottest part of the pilot flame, not beside it.

Loss after twenty or thirty seconds is the marginal case, and the popular account of it has the physics backwards. Output from that junction climbs as the tip heats, it does not fall away. What a delayed dropout reports is a probe whose best reading still lands below what the electromagnet needs to hold on its own.

The knob explains the delay. Pressing it down bypasses the magnet mechanically and keeps fuel at the pilot no matter what the probe is producing. Let it go and the armature coasts for a few seconds on the little pull available before it releases. A tip that is not sitting fully in the flame envelope reaches the same ending by another route, since it never gets hot enough to make full output at all, and that one is worth repositioning before it is condemned. A probe reading weak in the correct position gets replaced rather than adjusted, because it only gets worse from there.

A pilot that holds for hours and then drops out overnight is a different animal altogether. Look at draft and at whatever changed in the house recently.

Lint, dust, and the screen around the base

Nearly every gas storage heater built in the last two decades draws its combustion air through a screen ringing the bottom of the jacket. It exists so that a flame inside the chamber cannot reach flammable vapor out in the room, and it is a maintenance item that almost nobody maintains.

A partly blocked screen, packed solid with dryer lint in a laundry room or with clippings and pet hair in a garage, starves the burner compartment of air. The pilot then burns small and yellow instead of crisp and blue, the probe never gets properly hot, and a perfectly good thermocouple gets replaced by someone chasing the wrong part.

Vacuuming from outside helps and rarely finishes the job, since the deposit builds on the inner face. Clearing it properly means opening the compartment, which is where a service visit earns its money.

Newer controls use a thermopile and a board

Units built in the last fifteen years often swap the single probe for a thermopile, which is a stack of the same junctions wired in series. Stacking them raises output to several hundred millivolts, enough to run a small control board, a status light, and a temperature dial with no household current at all.

That blinking light near the base is a diagnostic display on those models. Manufacturers publish their own flash patterns, and the pattern distinguishes a low thermopile reading from a tripped limit or a failed sensor. Count the flashes and the pauses before calling, because the pattern is worth far more than any description of the symptom.

Some designs carry both, a thermocouple for the safety magnet and a thermopile for the electronics. Replacing one and not the other leaves an appliance that lights and then behaves oddly, which is a common reason for a second visit. These parts are also model specific, and a universal probe off a hardware shelf does not substitute for a manufacturer assembly.

Draft, backdraft, and the flue

A pilot that dies on windy nights and behaves perfectly in still weather is reporting a venting condition rather than a fuel problem. Air moving the wrong way down a flue extinguishes a small flame directly.

Backdraft often has causes sitting well outside the appliance. A powerful kitchen or bath exhaust fan running in a tight house pulls replacement air from wherever it can find it, and a water heater flue is frequently the easiest path available. A clothes dryer running in the same closed room does the same thing.

Shared chimneys are a common arrangement across this metro, and they turn awkward when the furnace beside the appliance is replaced by a high efficiency model venting through a wall. The old flue is then oversized for the water heater alone, exhaust cools before it can rise, and draft collapses.

A small mirror or a lit match held near the draft hood shows which way air is travelling within seconds of the burner firing. Smoke pushed back toward you instead of drawn up the vent means combustion products are entering the room, and what they carry is carbon monoxide, which gives no smell and no color to warn anybody.

That finding outranks the pilot complaint completely. Turn the control dial to Off if it is on your way past, get everyone out of the building, and make the call from outside. A carbon monoxide alarm already sounding changes who you are calling. That one is emergency services rather than a plumbing dispatch, and the fire department clears the house before any repair gets scheduled.

When the control body is the answer

The combination gas control is the last suspect rather than the first, because it is the expensive one. It reaches the top of the list by elimination: a good millivolt reading, a clean screen, a flue drafting properly, and a pilot that still refuses to stay on.

Controls fail with a magnet too weak to hold at normal voltage, or with an internal pilot passage blocked by debris carried down the supply line. Either way the assembly is replaced whole, since the internals are not serviceable in the field.

Age decides the outcome at this point. On a unit under eight years old the control is worth fitting. Past twelve, that money is going into a lining that is already thin, and it belongs toward a new appliance instead.

Questions about pilot light

What does a thermocouple replacement cost?

It is among the least expensive repairs on a gas appliance, and the part itself is a fraction of the visit. What you pay for is the diagnosis and the access, since the burner assembly usually has to come out to reach the mounting bracket. Ask for the millivolt reading that justified the swap.

Can I change the probe myself?

Mechanically it is a simple part. The risk is not the swap, it is disturbing the pilot tube and the main burner fitting on the way in and out, which is where gas leaks get created. Anyone proceeding should soap test every joint they loosened before relighting anything.

Why did this start right after a windy storm?

Weather that shifts pressure across a chimney can blow a standing pilot out, and a marginal probe that had been holding on will not restart afterward. The storm exposed a part already near the end of its life rather than causing the failure itself.

The pilot holds but the burner never fires.

That is a separate fault from this page. Proving the pilot and admitting fuel to the main burner are two different functions of the same control, and an appliance stuck between them usually has a thermostat or control problem rather than a probe problem.

Should I keep relighting it each time it goes out?

Once, following the label printed on the jacket. A pilot that will not hold after a single careful attempt is telling you something is broken, and repeated relighting only vents more fuel into the room. Turn the dial to Off and leave it there until someone looks at it.