BY JIM STEWART
We will build a temperature gauge for the range 0°C to 50°C using a thermistor
in an analog circuit. Why use a thermistor? Two words: temperature transducers.
Along the way, we will look at various ideas such as how to linearize a thermistor,
the effect of thermistor self-heating, and zero and span adjust in a gauge.
Thermistors are inherently non-linear. When
using one as a temperature transducer, there
are two basic approaches to converting its
resistance to the corresponding temperature:
B Value Maximum Heat Dissipation Thermal (°K) Power (W) Constant (m W/°C) Time Constant (seconds)
4100 4. 5 20
■ FIGURE 1
• Digital: Use software and a look-up table to
translate resistance to temperature, or solve the equations
to get T from the measured R.
• Analog: Use a circuit to linearize the resistance vs.
A commonly-used equation for thermistor resistance
as a function of temperature is:
R(T) = R -1 -1 0 exp[ β(T - T0 ) ]
■ FIGURE 2
T is the temperature being measured while T0 is the
reference temperature (usually 25°C). R is the measured
resistance while R0 is the resistance at T0. T, T0, and the
parameter β (Beta) are all in kelvins (K). Beta (sometimes
written as B) is a characteristic of the material used to make
the thermistor and is given on the device’s datasheet.
The thermistor used is Digi-Key part number PNT119-
ND. Specs for it are shown in Figure 1. The part was
chosen because its disc format makes it more rugged than
a smaller bead type. Also, its thermal time-constant of 20
seconds means it will not respond to short puffs of air.
A similar device with the same R0 and β is part 21T10K
from Electronix Express.
The datasheet gives the resistance at 25°C as 10K.
We will need the resistances at 50°C and 0°C so we will
calculate them here:
50°C = 323.15 kelvins
25°C = 298.15K
0°C = 273.15K
R 50°C = 10K × exp[4100 × (1/323 - 1/298)] = 3.45K
R 0°C = 10K × exp[4100 × (1/273 - 1/298)] = 35.2K
Actually, a C language program (RXCALC) was used
to do the calculations.
A linear function (Figure 2) has the equation y = mx +
b where b is the y intercept and m is the slope. Both b and
m are constants. In instrumentation terms, b is the offset