e t a

sponse2 and the detector realization of the candela.3 these devices been operated at cryogenic tempera-tures, which allows relative standard uncertaintiesFurthermore, for applications requiring the highest accuracy, its spectral range is limited to 450–700 nm. To obtain improved accuracy and increased spectral range of the primary standard, we have developed a high-accuracy cryogenic radiometer 1HACR2, which is an electrical-substitution device that can be oper-ated near the temperature of liquid helium. In a previous paper,4 measurements of the quantum effi-ciency of silicon photodiode light-trapping detectors obtained by the use of both the HACR and the NIST Spectral Comparator Facility were compared and found to agree within the combined uncertainties of the measurements. We describe the construction, operation, and accuracy of the HACR in this paper, whereas its use in a new realization of the NIST of 0.01 % in measurements of optical power. Cryo-genic electrical substitution radiometry was first developed at NIST for temperature measurements,7 and soon thereafter for the calibration of low-temperature vacuumblackbodies.8,9 Asimilar appa-ratus was also developed at the National Physical Laboratory 1NPL2 in the United Kingdom in conjunc-tion with Oxford Instruments, first to perform a measurement of the Stefan–Boltzmann constant,10,11 and shortly thereafter for high-accuracy measure-ments of optical power that use collimated radiation.12 The NIST HACR is also based on the NPL–Oxford design. However, changes have been made in the design and operation, especially in the instrumenta-tion and computer control of the measurement pro-cess. Furthermore, we have performed an addi-tional analysis of the accuracy of the HACR, as operated at NIST. Another radiometer based on the NPL–Oxford design is in use at the Physikalisch-The authors are with the National Institute of Standards an