Abstract
A new method for temperature scale calculation of a radiation thermometer with four-point blackbody calibration data and knowledge of the shape of the spectral responsivity was developed. This method was applied to numerical models of narrow-band and wide-band infrared radiation thermometers and it was compared with conventional characteristic equations.
The developed characteristic equation has an array representing the shape of the spectral responsivity and four constants: the array Sk, the gain coefficient C3, the offset D3, the center wavelength λ0 and the wavelength step d. The wavelength of Sk is given as λ0+k·d(k=-n, …, -1, 0, 1, …, n). Sk is obtained from relative data of the spectral responsivity or rough estimation on the shape of the spectral responsivity. Then the calculation of the four constants is made from four-point blackbody calibration data. When the offset can be measured, necessary blackbody calibration data are reduced to three points.
Spectral responsivity data of a germanium photodiode were used as the numerical model of a wide-band radiation thermometer. The product of the responsivity data and spectral transmittance data of a narrowband-pass filter of an actual radiation thermometer was used as the numerical model of a narrow-band radiation thermometer.
In the case of the wide-band model, the deviation of the calculated temperature scale from the model was less than ±1°C in the range from 140°C to 1500°C when the calibration points were 231.928°C, 660.323°C and 1084.62°C, and the used shape was a Gaussian distribution. The deviation was mostly less than 1/2 of the deviation of the best conventional equation in the interpolation range and it was 1/6 at 1800°C.
