Solar-type cavity radiometers are instruments of the highest metrological level for measuring solar direct normal irradiance. To ensure their traceability and performance, they are periodically compared to the World Group of Standards, which realizes the World Radiometric Reference (WRR), in the International Pyrheliometer Comparisons (IPCs). Additionally, they can be characterized in an absolute way, with direct traceability to SI units and with their measurement uncertainty calculated. This paper describes the different techniques and procedures applied for the characterization and calibration of solar cavity radiometers, with the main results obtained to date in the case of an Automatic Hickey–Frieden (AHF) radiometer. Voltmeters, resistors, temperature sensors and the area of the precision apertures have been calibrated, while the effective absorptance, temperature coefficients, optical scattering and non-equivalence factor have been evaluated. The temperature dependence of the electrical current in the cavity heater has also been analysed. The resulting corrections obtained for the AHF by characterization are compatible with the WRR factors obtained by this instrument in the past IPCs. An uncertainty of 0.42% ( k = 1) has been obtained, and this paper discusses further improvements that may be able to reduce this figure to the desired expanded uncertainty of U = 0.1% ( k = 2).

In this work, the method for calculation of uncertainty of pyrheliometers’ responsivity during their outdoor calibration process in the laboratory is exposed. It is applied first for calibration of standard pyrheliometers by comparison to cavity radiometers, and after for calibration of an end-user pyrheliometer against that standard pyrheliometer. The dissemination of the WRR irradiance scale is illustrated in practice and the increasing uncertainty in the traceability chain is quantified. The way of getting traceability to both WRR scale and to SI units in the current situation, where the shift between these radiometric scales is pending to be solved, is also explained. However, the impact of this gap between scales seems to be more important for calibrations of reference Class A pyrheliometers than in the final determination of DNI irradiance, because in this case, the cumulative uncertainty is large enough as to not significantly be affected for the difference. The way to take into account different correction terms in the measurement model function, and how to compute the corresponding uncertainty, is explained too. The influence of temperature of some pyrheliometers during calibration process and the potential impact on the DNI irradiance calculated with these instruments is exemplified.