Accuracy of Optical Pyrometry in the Range 800 °c to 4000 °c

Abstract

A pair of secondary standard tungsten strip lamps have had a luminance temperature – current calibration, in the range 800 °C to 2200 °C, at a number of national laboratories. An analysis of the calibration results confirms estimates of the accuracy of optical pyrometry in the range 800 °C to 2200 °C and supports the extension of these estimates to 4000 °C. The standard deviation uncertainty of optical pyrometry is shown to be about 1 °C at 800 °C rising to 2 °C at 2200 °C and 10 °C at 4000 °C, being about double this for the calibration of commercial pyrometers unless certain described precautions are taken.The reliability of the secondary standard lamps, when used under well-defined conditions, is confirmed and it is shown that they have a standard deviation calibration uncertainty of about 1 °C for the vacuum-type lamps in the range 800 °C to 1500 °C and 2 °C for the gas-filled lamps in the range 1500 °C to 2200 °C. Most of this uncertainty is due to primary standard optical pyrometer calibration errors. Attention is drawn to the fact that a carbon arc fulfills the requirements of a secondary luminance temperature standard at about 3514 °C.Recent determinations of the gold point and the second radiation constant indicate that the 1948 International Temperature Scale is lower than the thermodynamic scale by an amount varying from 0.8 °C at 800 °C to 12 °C at 4000 °C. This is already greater than the calibration errors of optical pyrometry and, in view of the still greater accuracies presaged by photomultipliers, a revision of the International Temperature Scale is suggested.

Theory and methods of optical pyrometry

“…The measurement of temperature utilizing optical pyrometry is based on the fact that the *This paper was presented at the Fourth Symposium on Temperature, Its Measurement and Control in Science and Industry; Columbus, Ohio; March [27][28][29][30][31]1961; sponsored by the American Institute of Physics, the Instrument Society of America, and the National Bureau of Standards.…”

“…In addition, if nonblackbodies are often used, it is well to use 5 or 6 observers to determine if any observer has a very unusual relative luminosity curve. The standard deviation of the mean effective wavelength due to luminosity factor differences has been reported [28] to be about 8 A, which in reading a strip lamp at 2300°C with a pyrometer having constant Lovejoy [7], using the results of an investigation of contrast thresholds or limens (see footnote 15) reported by Blackwell [29], calculated the expected observer standard deviation temperatures for the mean of a set of matches consisting of a bright filament to disappearance and a dark filament to disappearance for a pyrometer having a 0.05 radian exit angle, a 0.7 mm exit pupil, and a 0.038 mm actual and 10 min apparent angular filament width. The standard deviations ranged from 0.2 deg C at 1400°C to 0.3 deg C at 1063°C and increased to 1.6 deg C at 800°C,^* and agreed well with his experimental values.…”