The standard NF X 10-702-1 (Determination of the opacity of the fumes in an atmosphere without air renewal) describes a fire test method of materials exposed to irradiance of a furnace, by quantifying the smoke production from material in predefined irradiance exposure conditions. The standard defines the test apparatus and the operations to be carried out in order to ensure the metrological control of test apparatus. In particular, the 2.5 W·cm-2 irradiance, which a sample is exposed to during a test, must be adjusted using a radiometer. This radiometer is calibrated against a reference calorimeter according to this standard. The aim of this paper is to describe this calibration uncertainty evaluation. It includes the following steps: evaluation of the associated uncertainty of the reference calorimeter, modelling of the calibration regression and determination of the uncertainties on the irradiance for a given output tension of the radiometer. This work allowed to globally improve the quality of calibrations, through the quantification and control of uncertainties, and by supplying complementary data enlarging the calibration range.