Since 2018, Central Research Institute of Korea Hydro & Nuclear Power (KHNP–CRI) has been operating an X-ray irradiation system with a maximum voltage of 160 kV and 320 kV X-ray tube to test personal dosimeters in accordance with ANSI N13.11-2009 “Personnel Dosimetry Performance- Criteria for Testing”. This standard requires that dosimeters for the photon category testing be irradiated with the X-ray beams appropriate to the ISO beam quality requirements. KHNP-CRI has implemented the fourteen X-ray reference radiation beams in compliance with ISO-4037-1, 2, and 3. When installing the X-ray irradiation system, KHNP-CRI evaluated the uncertainties of dose conversion coefficients for deep and shallow doses, based on “Catalogue of X-ray spectra and their characteristic data – ISO and DIN radiation qualities, therapy and diagnostic radiation qualities, unfiltered X-ray spectra” published by Physikalisch Technische Bundesanstalt (PTB). A CdTe detector (X-123, AMPTEK) with disk type collimators made of tungsten was used to acquire X-ray spectra. The detector was located at 1 m from the center of the target material in the Xray tubes. Six uncertainty factors for the dose conversion coefficients for the fourteen X-ray beams were chosen as follows; the minimum and maximum cut-off energies Emin and Emax, the air density (ρ), the accuracy of the high-voltage of the X-ray tube, statistics of the pulse height spectra and the unfolding method. For example, uncertainty of each quantity for a HK30 beam was calculated to be 0.3%, 2.32%, 0.19%, 1.25%, and 0.13%, and 0.18%, respectively. The combined standard uncertainty for the deep dose conversion coefficient of the HK30 beam was calculated to be 2.67%. The coverage factor corresponding to a 95 percent confidence interval was obtained as k = 1.8 using a Monte Carlo method, which is slightly lower the coverage factor of k = 1.95 for a Gaussian distribution. This seems to result from that two dominant uncertainties, the unfolding uncertainty and minimum cut-off energy uncertainty, follow a rectangular distribution.