How to Test NTC Thermistor with Multimeter
As an Amazon Associate we earn from qualifying purchases.
How to test NTC thermistor with multimeter: look up the part’s resistance-temperature table, measure the part’s temperature and its resistance, and check that the reading falls inside the tolerance band for that temperature. A second reading at a different temperature confirms the slope. This guide works the arithmetic for one 10 kilohm series, shows how the band widens away from 25 degrees C, and explains why an ordinary thermometer limits what a multimeter can prove.
Electrical safety first
Work on mains circuits can kill. OSHA’s fact sheet on working safely with electricity says that if the ground path is broken, fault current may travel through a worker’s body, causing electrical burns or death. A meter’s CAT rating must match the circuit, since Fluke’s category-rating page says equipment with a rating that does not match the task can explode if used improperly. Panel and service work belongs to a licensed electrician; OSHA’s rule for general industry allows only qualified persons to work on parts that have not been de-energized.
This guide stays on the bench. It covers an NTC thermistor that is loose, or a sensor that is out of the equipment and not connected to any power source, checked in the ohms function only. It gives no steps for refrigerators, ovens, water heaters or other mains-powered appliances. If the listing for your meter does not state a CAT rating, do not treat the meter as safe for household mains.
Quick answer
- Find the part’s table: R25, the B value and the resistance at the temperature you will test.
- Read the part’s resistance and the temperature of the part itself at the same moment.
- Pass the reading if it lands inside the tolerance band for that temperature, which we calculate below.
- Confirm the slope with a second temperature, for example an ice-water bath at 0 degrees C.
What the datasheet gives you
The Vishay application note on NTC thermistors names the main parameters. R25 is the resistance at 25 degrees C (298.15 K), which the note calls a convenient reference point for thermistors. The B value is a material constant in kelvin that sets the slope of the resistance-temperature curve. To a first approximation the note gives the curve as R(T) = R25 × exp[B × (1/T − 1/298.15)], with T in kelvin. It adds that B varies somewhat with temperature, which is why it is defined between 25 and 85 degrees C as B25/85.
Tolerance comes in two parts. The note says tolerances on R25 come mainly from variations in ceramic manufacture and chip dimensions, and tolerance on B comes mainly from material composition and sintering. The NTCLE100E3 datasheet shows how to combine them. With X as the R25 tolerance and Y as the resistance deviation due to B tolerance, the complete deviation is Z = (1 + X) × (1 + Y) − 1, and the sheet also gives Z ≈ X + Y as a quick estimate. That sheet lists R25 tolerance options of 2, 3 and 5 percent for the 10,000 ohm part, and a table of nominal resistance by temperature, which is our working data below. Datasheets for other parts differ, so use the sheet for your own part. Our guide to ohm meters covers meters built for resistance work.
Work out the pass band
The table below uses the 10,000 ohm column of the NTCLE100E3 sheet at the 5 percent R25 grade. The nominal value and the B-tolerance column are quoted from the sheet. The band is our own arithmetic from the sheet’s formula: the high limit is nominal × (1 + Z) and the low limit is nominal × (1 − Z′), where Z′ = 1 − (1 − X) × (1 − Y). The sheet gives only the upward formula, so the low limit is our extension by symmetry.
| Temperature (°C) | Nominal (ohms) | Deviation due to B tolerance (%) | Low limit (ohms) | High limit (ohms) |
|---|---|---|---|---|
| 0 | 32,554 | 0.92 | 30,642 | 34,496 |
| 10 | 19,872 | 0.53 | 18,778 | 20,976 |
| 20 | 12,488 | 0.17 | 11,843 | 13,135 |
| 25 | 10,000 | 0.00 | 9,500 | 10,500 |
| 30 | 8,059 | 0.17 | 7,643 | 8,476 |
| 40 | 5,330 | 0.48 | 5,039 | 5,623 |
| 50 | 3,605 | 0.77 | 3,398 | 3,814 |
A worked row shows the method. At 0 degrees C, X = 5 percent and Y = 0.92 percent, so the high limit is 1.05 × 1.0092 = 1.05966 and the low limit uses 0.95 × 0.9908 = 0.94126. Multiply the nominal 32,554 by each and you get 34,496 and 30,642. The sheet’s own example reaches the same Z of 5.966 percent at 0 degrees C and says a 10 kilohm part can show 32.55 kilohms anywhere between −1.17 and +1.17 degrees C. In other words, a part inside its band is consistent with the table to within about a degree, which is as far as a multimeter test can go.
Take the reading at room temperature
Fluke’s guide to measuring resistance gives the meter side. Turn off power and discharge any capacitor, set the dial to ohms, expect OL until the leads touch the part, and note that best results come when the component is removed from the circuit. It adds that the human body becomes a parallel resistance path, so keep your fingers off the metal tips. Use the range button for a fixed range and note the unit shown beside the digits. Between 20 and 30 degrees C the nominal values run from about 8 to 12.5 kilohms, and at 0 degrees C the high limit reaches 34,496 ohms, so choose a range that tops out above that.
The temperature must be the part’s own, not the air in the next room. Let the part sit where you take the thermometer reading until the resistance stops changing. At 20 degrees C the band is 11,843 to 13,135 ohms, and at 30 degrees C it is 7,643 to 8,476 ohms. The nominal value falls from 12,488 to 8,059 ohms between those two temperatures, so a 10 degree error in the temperature you assume would move the expected value by about a third. Even 1 degree matters: the sheet puts the slope at −4.38 percent per degree at 25 degrees C. See also the digital ohmmeter page if you are choosing a meter with a suitable range.
Confirm the slope at a second temperature
One reading proves the part against one point. A second temperature tests the slope. The Walla Walla University lab calibrates a thermistor from readings in an ice bath and a boiling bath, and says solving for B needs resistance at two known temperatures. The Notre Dame lab uses the same two-point idea for the Steinhart equation. For a home check, skip the boiling bath and use the ice point alone: add ice and water to a cup with the minimum water needed to fill the gaps, let it sit, immerse a sealed probe, and stir while the reading settles, as the Walla Walla page describes. The Notre Dame handout asks for much more solid ice than liquid water. The pass band at 0 degrees C is 30,642 to 34,496 ohms for this series.
A bare coated part is a different case. The NTCLE100E3 datasheet says the coating has no specified insulation properties, so do not assume that part can sit in water. A refrigerator training page from Fred’s Appliance Academy tests its sealed thermistor in ice water and sets a service rule of replacing the part if it is open, shorted, or out of range by more than 10 percent of the manufacturer’s chart. That 10 percent is the rule of one procedure, and our band is tighter because it comes from the datasheet.
Estimate B from two readings
You can turn two readings into a B value. Rearranging the Vishay equation gives B = ln(R1 ÷ R2) ÷ (1/T1 − 1/T2), with T in kelvin. To show the method without inventing measurements, apply it to the sheet’s nominal values for 0 and 25 degrees C: ln(32,554 ÷ 10,000) = 1.180, and 1/273.15 − 1/298.15 = 0.000307, so B = 1.180 ÷ 0.000307 = 3,845 K. That is 3.3 percent below the 3,977 K the same sheet lists for B25/85. The gap is expected, because the Vishay note says B varies somewhat with temperature. As a check, the 85 degree C value of 1,070 ohms in the same table gives ln(10,000 ÷ 1,070) ÷ (1/298.15 − 1/358.15) = 3,977 K, so the formula and the table agree.
Real readings bring errors. Our arithmetic shows two examples. If the room is 1 degree warmer than you assumed, B comes out at about 3,709 K, 3.5 percent lower. If the 25 degree reading alone is 5 percent high, perhaps because the part had not settled, B comes out near 3,686 K, and 5 percent low gives about 4,012 K. Those swings of roughly 4 percent are far larger than the 0.75 percent B tolerance on the sheet, so a home B estimate is a sanity check, not a verdict.
Check the meter’s own heating
Current from the meter warms the part. Vishay Dale’s selection guide defines the dissipation constant as the power needed to raise the thermistor 1 degree C above its environment, and says the measuring current must be small enough to produce negligible self-heating error. The NTCLE100E3 sheet gives a dissipation factor of 7 mW per K for information. The pages opened do not state any meter’s test current, so take an assumed value for scale: our own assumption of 0.1 mA through 10,000 ohms. Power is I² × R = (0.0001)² × 10,000 = 0.0001 W, or 0.1 mW. Divide by 7 mW per K and the rise is about 0.014 K, which moves the resistance by 0.014 × 4.38 = 0.06 percent. For that part and that assumed current the effect is small, but it depends on the meter and the part, so check your manual.
Reading the results
| What you see | What it may mean |
|---|---|
| Inside the band at room temperature and at 0 degrees C | The part agrees with its table at both points |
| Inside the band at one temperature only | A temperature error, a settling problem, or a part with a different B |
| Outside the band by a similar percentage at both points | R25 may have shifted; Vishay’s note lists R25 and B shifts after long exposure at limit conditions |
| OL, or near zero | Open element, or a short or parallel path |
| Reading wanders as you hold the leads | Finger contact or body heat on the part |
Mistakes to avoid
- Comparing against R25 at the wrong temperature. Near 25 degrees C a single degree changes the value by about 4.4 percent.
- Treating the nominal table as a hard limit. The sheet states tolerances, and we apply them above.
- Using a different part’s chart. Each series and R25 value has its own table.
- Measuring in a circuit. The Fluke 87 manual says the meter shows the total resistance through all possible paths between the probes.
How we chose
This guide draws on pages opened on 2026-10-21: the Vishay application note on NTC thermistors, the Vishay NTCLE100E3 datasheet, the Vishay Dale selection guide, lab handouts from Walla Walla University and Notre Dame, a Fred’s Appliance Academy page, the Fluke resistance guide and the Fluke 87 manual, two OSHA pages and a Fluke category-rating page. We chose one 10 kilohm series with a full table so the arithmetic could be traced to the sheet, and we label every calculation as ours. The licensed-electrician advice is a site rule, since OSHA speaks of qualified persons.
What we could not verify
The pages opened do not give a test current for any multimeter, so the self-heating example rests on an assumption we state. They do not say how accurate a home thermometer is, which limits any band check. The low limit of our band is an extension of the sheet’s one-sided formula. The B-value estimate was run on table values, not on readings from a part. The 10 kilohm series is not every NTC thermistor, and glass-bead or probe types have their own sheets. Nothing in them describes a safe way to immerse a bare coated bead.
Frequently asked questions
What should a 10k NTC thermistor read at room temperature?
About 10,000 ohms at 25 degrees C, and more below that. At 20 degrees C the Vishay table gives 12,488 ohms. Allow the R25 tolerance and the B tolerance on top.
How much can an NTC thermistor differ from its table?
For the 5 percent grade of the NTCLE100E3 10 kilohm part, the band is 5 percent at 25 degrees C and about 6 percent at 0 degrees C by the sheet’s formula. Other grades and parts differ.
Can I test an NTC thermistor without a thermometer?
You can check direction by warming and cooling the part, but you cannot check the value. Without a known temperature, a reading has nothing to be compared against.
Why do the labs use two temperatures?
Two known temperatures let them solve for B, as the Walla Walla lab does with an ice bath and a boiling bath. One reading fixes one point on the curve.
Does the multimeter heat the thermistor?
A little. The size depends on the test current, which the pages opened do not state for any meter. With an assumed 0.1 mA the example above gives a rise of about 0.014 K.
Bottom line
Judge an NTC reading against the table for the temperature the part is actually at, allow the tolerance band, and use a second temperature to confirm the slope.
Sources
- Vishay BCcomponents, NTC Thermistors application note (doc 29053): https://www.vishay.com/docs/29053/ntcappnote.pdf (opened 2026-10-21)
- Vishay BCcomponents, NTCLE100E3 datasheet (doc 29049): https://www.vishay.com/docs/29049/ntcle100.pdf (opened 2026-10-21)
- Vishay Dale, Selecting NTC Thermistors: https://www.vishay.com/docs/33001/seltherm.pdf (opened 2026-10-21)
- Walla Walla University, ENGR 325 Lab 7 Measuring temperature: https://gab.wallawalla.edu/~larry.aamodt/engr384/lab7_21.pdf (opened 2026-10-21)
- University of Notre Dame, AME 21216 Experiment A3 Sensor Calibration: https://prumbach.nd.edu/ame20216/a3/S26_A3_procedure.pdf (opened 2026-10-21)
- Fred’s Appliance Academy, Testing a Sub-Zero Refrigerator Thermistor using a Glass of Ice Water: https://academy.fredsappliance.com/video/testing-a-sub-zero-refrigerator-thermistor-using-a-glass-of-ice-water/ (opened 2026-10-21)
- Fluke, How to Measure Resistance with a Digital Multimeter: https://www.fluke.com/en-us/learn/blog/digital-multimeters/how-to-measure-resistance (opened 2026-10-21)
- Fluke, 87 True RMS Multimeter Users Manual: https://assets.fluke.com/manuals/87______umeng0800.pdf (opened 2026-10-21)
- OSHA, Working Safely with Electricity fact sheet: https://www.osha.gov/sites/default/files/publications/OSHA3942.pdf (opened 2026-10-21)
- OSHA, 29 CFR 1910.333 selection and use of work practices: https://www.osha.gov/laws-regs/regulations/standardnumber/1910/1910.333 (opened 2026-10-21)
- Fluke, IEC Category Ratings: Use the Right Tools for the Job: https://www.fluke.com/en-us/learn/blog/safety/iec-category-ratings-use-the-right-tools-for-the-job (opened 2026-10-21)
As an Amazon Associate we earn from qualifying purchases.
