The defrost circuit: heater, thermistor, and why frost returns

Frost is normal. Frost you can see is not.

Every refrigerator makes frost on purpose, in one place, out of sight. The evaporator is the coldest surface in the cabinet, so moisture gives itself up there first, carried in from the room every time a door opens and drawn out of the food between times. That coil is expected to grow a layer of white between defrosts, and a heater is built into it to melt the layer away on a schedule. What is not normal is frost you can see from the kitchen: on the drawer fronts, on the packages, in a sheet across the back wall. That is the coil telling you it is already full. A Sub-Zero adds one detail worth knowing first. Dual refrigeration gives the fresh-food side and the freezer their own sealed systems and their own evaporators, with no air passing between the compartments, so each side defrosts on its own schedule. Heavy frost on one side proves nothing about the other.

• Normal: an even white coating on the evaporator behind the rear panel, gone after the next cycle
• Not normal: frost on the compartment walls, the drawer runners or the food packages
• Not normal: frost back within a week or two of a full manual defrost

The four parts that make a defrost happen

A defrost is not a component, it is a small circuit with four jobs in it. The heater is an element bonded to or threaded through the evaporator, and its only task is to put enough warmth into the coil to turn frost into water. The termination sensor watches the coil and cuts the heat once it has climbed past freezing, so the compartment is never warmed a second longer than it must be; on the electronic generations that sensor is a thermistor, on older units a bimetal thermostat that opens at a set point. The control decides when a cycle is due. Older built-ins ran that from a timer counting clock hours whether the appliance needed it or not; electronic controls count compressor run time and door openings instead, so a kitchen opened thirty times a day defrosts more often than a house standing empty. The drain then carries the melt water out through a trough at the base of the coil, down a tube, into a pan in the machine compartment where condenser heat evaporates it. Remove any one of the four and the frost stays.

• Defrost heater: melts frost off the evaporator
• Termination thermistor or thermostat: ends the heat once the coil is above freezing
• Control board or timer: decides when a defrost is due
• Drain trough, tube and evaporation pan: carry the water out of the compartment

How each one fails, and what it looks like from the kitchen

A heater that has failed open is the most common of the four and the quietest. Nothing announces it. The compartment goes on cooling, the coil frosts a little heavier each day, airflow falls away, and the complaint that eventually arrives is temperature rather than frost, because the display reports from a sensor that is not sitting in the blocked airstream. A thermistor drifted warm makes the control believe the coil has already come up to temperature, so the cycle ends almost as soon as it starts; one drifted cold leaves the heat on too long, and you get soft frozen food and water that never fully refreezes. Control faults are the least common, though the relay that switches the heater does fail and looks exactly like a dead heater until both are metered. The drain is the one that fools people. It ices at the trough or partway down the tube, the water has nowhere to go, some refreezes on the bottom of the coil and the rest works forward and out under the door. The cycle ran. The ice did not leave.

• Heater open: heavy even frost, falling airflow, temperature complaints first
• Sensor drifted warm: short cycles, frost that never fully clears
• Sensor drifted cold: soft frozen food and water in the compartment
• Drain iced: frost returns and water appears on the floor

The tests, and why they are taken at the coil

The heater is metered across its terminals for continuity, and an open reading condemns it on the spot. The thermistor is measured against its resistance curve at a known temperature rather than judged by feel, and a glass of ice water is the reference every technician carries because it is a repeatable thirty-two degrees; a sensor that does not read close to its published value there has moved, whatever the display says. Then force a cycle and stand through it: the coil should clear completely, the trough should end up empty, and warm water poured down the drain should appear in the pan. Temperature gets read at the coil and the air outlet with a probe, not off the front panel, which reports one sensor in one location and is a poor witness to what happens behind the interior wall. The one absolute is how ice comes off. Heat and time, doors open or warm air kept moving, never anything metal: a screwdriver through the thin aluminum of an evaporator turns a modest parts job into a sealed system repair, and that is the one mistake nobody can take back.

• Continuity across the defrost heater terminals
• Thermistor resistance checked in ice water against its published curve
• A forced cycle watched start to finish, not assumed
• Warm water down the drain to prove it flows through to the pan
• Coil and outlet air temperature taken with a probe

The half of the problem that is not the circuit

A defrost circuit in perfect health can still be out-run by the load put on it, and that is worth ruling out before anyone opens the cabinet. The seal is first. Close a strip of paper in the door and pull it out; where it slides free with no drag, that stretch of gasket has taken a set or lost magnet strength and the cabinet is drawing room air around the clock. Then the habits: a drawer packed a fraction too full holds the door open by a millimeter and nobody notices, and a week of loading warm groceries from a hot car will frost a coil heavily and honestly. Humidity matters more than people expect, and it is one of the few things that genuinely splits this service area in two, since marine layer mornings at Dana Point and San Clemente put far more moisture into every door opening than the same habit does inland at Ladera Ranch. The test costs nothing: clear the frost, close everything properly, give it two weeks. Frost that stays away was a load problem. Frost back in seven days is the circuit.

What is cheap here, what is expensive, and what waiting costs

A frost complaint usually lands at the inexpensive end of the parts list. A defrost heater and a thermistor are both modest parts, and most of what you pay for is access and patience rather than the component: unloading the compartment, taking down the interior rear panel, thawing a coil properly instead of quickly, then proving the cycle. A control board is a different order of money, which is why it should be the last thing anyone reaches for. The genuinely expensive repair here is the sealed system, and a frost problem almost never lives there. What changes the arithmetic is time. A coil left encased for months blocks the air so completely that the evaporator fan spends its life pushing against ice, runs hot and joins the bill. A drain left frozen through a season puts water under a cabinet standing on a subfloor you cannot dry out afterward. Frost stays cheap only while it is still just frost.