NTC Thermistors and Oven RTD Sensors Explained

Why this matters

Almost every board-controlled appliance on your route decides what to do from one resistance reading. The refrigerator decides whether to run the compressor, the dryer whether the heat is too high, the dishwasher whether the water is hot enough to start the wash timer, the oven when to cut the bake element. Two kinds of sensor do that job, and they move in opposite directions. A tech who carries one direction in his head and applies it to the other condemns good sensors, misses bad harnesses, and explains a fault backwards to the customer. This card covers what each sensor is, why each one sits where it does, and how to prove one with a reference temperature rather than a guess.

Two sensors, two directions

NTC THERMISTOR (refrigerator, dryer, dishwasher)
  cold ----------------------------------> hot
  high ohms                          low ohms
  resistance FALLS as temperature rises

OVEN RTD (platinum, 1000 ohm class)
  cold ----------------------------------> hot
  low ohms                          high ohms
  resistance RISES as temperature rises
  about 1000 ohms at 32 F
  about 1080 to 1100 ohms at room temp
  about 1650 to 1700 ohms at 350 F

NTC means negative temperature coefficient. The thermistor is a bead of sintered metal oxide. Warming it frees more charge carriers in the material, so it conducts better and its resistance drops, steeply and not in a straight line. Near freezing, a few degrees move the reading by hundreds or thousands of ohms, which is exactly the resolution a refrigerator needs to hold a fresh-food section within a couple of degrees.

The oven RTD (resistance temperature detector) is a platinum element. Platinum is a metal, and in a metal, warming increases the lattice vibration that scatters electrons, so resistance rises. It rises gently and nearly in a straight line, roughly 2.1 ohms per degree F on a 1,000 ohm class sensor, and platinum stays stable at bake and self-clean temperatures where an oxide thermistor would drift or fail. That is why the trade split happened: sensitivity where the range is narrow and cold, stability where the range is wide and hot.

The RTD figures in the diagram come from the platinum curve in IEC 60751, the standard most oven sensors are built to: a 1,000 ohm sensor is defined as 1,000 ohms at 32 F, and the curve gives about 1,078 ohms at 68 F, about 1,097 ohms at 77 F and about 1,670 ohms at 350 F. Many oven tech sheets publish a table that matches those values. Some ovens use a different class or add a resistor in the probe assembly, so the tech sheet for that model is the pass value, and the curve is the sanity check.

Why the tech sheet table is the only pass value

NTC thermistors are not one part. Appliance makers use thermistors with different nominal resistances and different curve steepness. Two refrigerators can both use a "10K" style sensor and still disagree by a meaningful margin at freezing, and other makers use sensors nowhere near 10K at room temperature. A reading that would be perfect on one model is a failed sensor on another. The tech sheet prints a table of resistance against temperature for that model's sensor, and that row, at the temperature you actually measured, is what the reading must match.

Read the whole row, including the temperature column. A table entry of the shape "32 F, 16,300 ohms; 77 F, 5,000 ohms" (values illustrative, not for any model) is useless if you read the sensor at 55 F in a defrosting freezer and compare it to either row. Either bring the sensor to a temperature on the table or interpolate between two rows, knowing NTC curves bend, so straight-line interpolation is only approximate across a wide gap.

The board reads it as a voltage, not ohms. Most boards put the sensor in a divider and read the voltage at the midpoint. That is why a harness fault reads as a temperature fault: a corroded pin or a chafed wire adds or bypasses resistance, and the board has no way to know it is not the sensor.

What open and shorted look like to the board

Work the direction out every time rather than memorizing a symptom list.

• NTC open (OL): infinite resistance is what an NTC reads when it is extremely cold. A refrigerator board sees a freezing-cold compartment and stops calling for cooling, so the customer reports warm food. Many boards set a sensor fault code instead.
• NTC shorted (near zero): what an NTC reads when it is extremely hot. A refrigerator board sees a hot compartment and runs the compressor without stopping, so the customer reports frozen lettuce in the fresh-food section. On a dryer, a shorted exhaust thermistor reads as overheated and the board cuts heat.
• RTD open (OL): what an RTD would read at an impossibly high temperature. The oven board treats it as a runaway and shuts down with a sensor code.
• RTD shorted or low: reads as an oven colder than it is, so the board drives the element longer. Most boards also code a reading below the plausible range.

Added series resistance (a burned or corroded connection) pushes an NTC toward reading colder and an RTD toward reading hotter. On a refrigerator that means warm; on an oven it means the board cuts heat early and the oven runs cold. That direction is the whole of the worked example below.

The field test: a known reference, not a guess

Safety first. Unplug the appliance or switch off and lock the breaker, then verify absence of voltage at the sensor connector with a meter proven on a known live source before and after, per 'Appliance Electrical Safety Check Before Service'. Resistance readings are only taken dead; a live sensor circuit carries the board's sense voltage and will throw the reading.

Ice water for an NTC. Fill a cup with crushed ice, top it with water, stir, and let it sit a minute so the water settles at 32 F. Disconnect the sensor, put the bulb in the slurry without wetting the connector, stir, and read after the reading stops moving, usually within a minute or two. Compare with the table's 32 F row. A sensor that matches at 32 F and moves smoothly as it warms in your hand is a good sensor.

Room temperature for an RTD. An oven sensor rarely has an ice-water value worth chasing in the field, and it does not need one. With the oven cold for at least an hour, read the kitchen air with a thermometer near the oven, read the sensor at its own disconnected plug, and compare with the table at that temperature. On a 1,000 ohm class sensor in a kitchen between 68 F and 77 F, expect roughly 1,078 to 1,097 ohms.

Then read again at the board. The sensor at its own plug proves the sensor. The same reading taken at the harness plug pulled off the board, sensor still connected and harness in circuit, proves the path the board actually sees. The two must agree within a few ohms. A gap between them is a harness fault and the sensor is good.

Worked example: an oven that runs cold with a good sensor

The complaint. A wall oven set to 350 F bakes slowly. The customer's oven thermometer, placed center rack, settles around 310 F after a full preheat. No fault code.

Step 1, sensor at its plug. Oven cold overnight, kitchen at 75 F by the tech's thermometer. Breaker off, locked, verified dead at the sensor connector with a proven meter. Sensor unplugged at the oven back and read across its two pins: 1,093 ohms. The tech sheet table for this model gives about 1,090 ohms at 75 F. The sensor passes.

Step 2, through the harness at the board. Sensor plugged back in, harness plug pulled off the board. At that plug, read across the two sensor pins: 1,178 ohms.

Step 3, the subtraction. 1,178 minus 1,093 leaves 85 ohms that exist somewhere between the board and the sensor, in series. The sensor did not add them; the harness did.

Step 4, what 85 ohms is worth. On this sensor's slope of about 2.1 ohms per degree F, 85 ohms is about 40 F of error. The board believes the cavity is about 40 F hotter than it really is, cuts the element about 40 F early, and the oven settles near 310 F on a 350 F set point. The direction checks: added resistance on an RTD reads hot, a board that thinks the oven is hot underheats it, and the complaint is a cold oven. The number checks: 350 minus 310 is 40, which matches.

Step 5, finding it. With the breaker still off and locked, the tech unplugs the inline connector behind the oven where the sensor lead meets the cavity harness. One female terminal is darkened and loose on its pin. Reading across that connection with the halves mated shows most of the missing ohms. The terminal is replaced with a high-temperature crimp terminal of the type the maker specifies, and the connector seated.

Step 6, the finish test. At the harness plug, the sensor now reads 1,096 ohms against 1,093 at the plug, a 3 ohm difference, within normal lead resistance. Harness plugged back onto the board, breaker on, the oven preheats to 350 F, and after the cycle settles the center-rack thermometer reads within the maker's stated band, about 345 F on this call.

The tech who skipped step 2 would have read a good sensor, assumed a calibration drift, used the board's temperature offset to add 40 F, and left. That masks the fault for a while, and a loose high-temperature terminal keeps heating and corroding until it opens, at which point the oven shuts down with a sensor code and the customer pays for a second visit.

What would change the call

• A sensor that fails only hot. An RTD with a cracked element can read correctly at room temperature and go erratic at bake heat. If the cold readings pass and the oven still swings, read the sensor signal at the board during a bake at the live test point the tech sheet names (usually a voltage), compare with the sheet's value for the cavity temperature, and watch for jumps.
• A refrigerator with frost on the sensor. A thermistor clipped to the evaporator reads whatever the coil is doing. Let it reach a known temperature before comparing, or the "fault" is just an unfinished defrost.
• Dual sensors in one housing. Some dishwashers combine the thermistor with a turbidity sensor in one assembly. Only the two thermistor pins named on the diagram are read as a thermistor; the rest are not.
• A table that does not match the curve. If the tech sheet's RTD values sit far from the platinum curve, believe the sheet, not the curve. The maker may use a different sensor class.

References

• Manufacturer tech sheet resistance-versus-temperature table for the specific model; the only pass value on this card
• IEC 60751, the industrial platinum resistance thermometer standard whose curve the 1,000 ohm at 32 F oven sensors follow
• See related: Appliance Schematic Symbols and Wire Color Abbreviations
• See related: How to Test an Oven Temperature Sensor and Its Harness
• See related: Appliance Electrical Safety Check Before Service