A 10k NTC (Negative Temperature Coefficient) thermistor has a nominal resistance of exactly 10,000 ohms at 25°C (77°F). As temperature rises, its resistance drops exponentially. However, the exact resistance at any other temperature depends entirely on its Beta ($\beta$) coefficient. For 90% of DIY, 3D printer, and HVAC applications, the standard is Beta 3950K. Automotive and medical sensors often use Beta 3435K or 3380K.
Below is the complete reference chart for the industry-standard 10k NTC with a $\beta$ of 3950K, followed by critical bench-level guidance on how to actually use this data without introducing measurement errors.
How to Read the 10k NTC Resistance Chart
Before jumping to the numbers, you need to understand what the columns mean and which column applies to your installation. Manufacturer datasheets typically provide three resistance values per temperature step:
- Nominal Resistance: The theoretical target calculated using the Beta parameter equation. This is the number you use in your Arduino/ESP32 code or analog circuit design.
- Min / Max Resistance (Tolerance Bounds): A standard 10k NTC has a $\pm$1% or $\pm$2% tolerance at 25°C. However, this tolerance widens at temperature extremes. At -40°C or +125°C, the physical tolerance can stretch to $\pm$5% or more. If you are building a precision incubator or a 3D printer hotend, you must design your software to expect values within the Min/Max columns, not just the Nominal column.
Every NTC chart is anchored at 25°C (298.15K). If your multimeter reads 10.2k$\Omega$ at a verified 25°C room temperature, your thermistor is well within the standard 2% tolerance. Do not attempt to "calibrate" it by adding a fixed offset in code; NTC curves are non-linear. Instead, adjust the Beta value or Steinhart-Hart coefficients in your firmware.
Complete 10k NTC Thermistor R-T Data Table (Beta 3950K)
The following table provides the resistance values for a 10k NTC thermistor with a Beta value of 3950K. Source: Calculated via the standard Beta Parameter Equation ($R_T = R_{25} \cdot e^{\beta(1/T - 1/T_{25})}$), aligning with standard TDK/EPCOS B57891S0103 and Vishay NTCLE100E3103 datasheet characteristics.
| Temperature (°C) | Nominal Resistance ($\Omega$) | Approx. Min (-2%) | Approx. Max (+2%) |
|---|---|---|---|
| -40 | 214,500 | 203,700 | 225,300 |
| -20 | 68,100 | 64,600 | 71,600 |
| 0 (Freezing) | 27,600 | 26,200 | 29,000 |
| 10 | 17,900 | 17,000 | 18,800 |
| 20 | 12,100 | 11,500 | 12,700 |
| 25 (Room) | 10,000 | 9,800 | 10,200 |
| 30 | 8,310 | 8,140 | 8,480 |
| 40 | 5,820 | 5,700 | 5,940 |
| 50 | 4,160 | 4,070 | 4,250 |
| 60 | 3,020 | 2,950 | 3,090 |
| 80 | 1,710 | 1,670 | 1,750 |
| 100 (Boiling) | 697 | 680 | 714 |
| 125 | 374 | 362 | 386 |
Bookmark this table: The highlighted rows (0°C, 25°C, 100°C) are the most frequently queried calibration points for DIY environmental chambers and 3D printer bed sensors.
Derating, Self-Heating, and What the Table Cannot Tell You
A resistance chart is only half the story. If you push too much current through a 10k NTC, the table becomes useless because the thermistor will heat itself. Here is how derating rows modify the base value and the physical limits the chart ignores.
Power Derating Limits
Thermistors have a maximum power rating, typically around 50mW at 25°C for standard epoxy-coated beads. As ambient temperature rises, you must derate the allowable power linearly. By 125°C, the maximum allowable power drops to near zero. If you exceed this, the internal heat will permanently shift the thermistor's calibration curve (a phenomenon called "thermal drift") or melt the epoxy coating.
The Self-Heating Trap (ESP32 / Arduino Builders)
The chart assumes the thermistor is in thermal equilibrium with its environment. But measuring resistance requires passing a current through the device, which generates heat ($P = I^2R$). The dissipation constant ($\delta$) tells you how many milliwatts of power will raise the thermistor's temperature by 1°C above ambient. For a standard 10k bead in still air, $\delta$ is roughly 1.5 to 2.0 mW/°C.
If you wire a 10k NTC to an ESP32 using a 1k$\Omega$ pull-up resistor to 3.3V, the current at 25°C is roughly 2.6mA. The power dissipated by the thermistor is $I^2R = (0.0026)^2 \cdot 10000 = 67$mW. With a dissipation constant of 2 mW/°C, your thermistor will self-heat by 33°C. Your ESP32 will read 58°C when the room is actually 25°C.
The Fix: Always use a 10k$\Omega$ or higher pull-up resistor (or a voltage divider with 10k/10k) for 10k NTCs on microcontrollers. This limits current to ~0.16mA, keeping self-heating well under 0.5°C.
What the Table Cannot Tell You
The R-T chart is completely blind to thermal mass and response time. It cannot tell you the Thermal Time Constant ($\tau$), which is the time required for the thermistor to reach 63.2% of a new temperature. A bare glass-bead 10k NTC might react in 1 second in stirred water, while that exact same silicon die encased in a thick brass probe for a 3D printer hotend might take 8 seconds. If your PID control loop is tuned too aggressively for a slow-response probe, your system will oscillate wildly.
10k NTC Thermistor FAQ
Can I use a 10k NTC with a 3950 Beta value interchangeably with a 3435 Beta?
No. While both will read exactly 10,000$\Omega$ at 25°C, their curves diverge rapidly outside of room temperature. At 100°C, a Beta 3950 thermistor reads roughly 697$\Omega$, while a Beta 3435 thermistor reads closer to 900$\Omega$. If you swap them without updating the Beta coefficient in your firmware's Steinhart-Hart equation, your temperature readings will be off by 10°C to 15°C at the extremes.
How do I calculate resistance for temperatures not listed in the chart?
Use the Beta Parameter Equation: $R_T = R_{25} \cdot e^{\beta(1/T - 1/T_{25})}$, where temperatures $T$ and $T_{25}$ are in Kelvin (add 273.15 to your Celsius value). For even higher precision across wide temperature ranges, use the 3-coefficient Steinhart-Hart equation, which accounts for the slight non-linear deviations that the pure Beta equation misses at extreme highs and lows.
Why is my multimeter reading slightly off from the 25°C chart value?
Three factors cause this on the bench. First, standard tolerance is $\pm$1% to $\pm$2%, meaning 9.8k to 10.2k is perfectly normal at exactly 25°C. Second, lead resistance from your multimeter probes can add 0.2$\Omega$ to 0.5$\Omega$ (negligible at 10k, but matters at 100°C where the thermistor is only 697$\Omega$). Third, the test current from the multimeter itself causes minor self-heating. For precision validation, immerse the thermistor in a stirred ice-water bath (0°C) or boiling water (100°C at sea level) rather than trusting ambient room temperature estimates.
Where can I find reliable datasheets for specific 10k NTC part numbers?
Always refer to the manufacturer's specific R-T tables rather than relying solely on generic Beta calculations. Excellent reference datasheets include the TDK/EPCOS NTC Thermistor series and the Vishay NTCLE100 series. These documents provide the exact Steinhart-Hart coefficients required for high-precision firmware implementations.
