To determine patient doses or doses to the medical staff, Monte Carlo calculations are frequently applied. In these kinds of calculations the x-ray tube is often simplified to make the calculations faster. The purpose of this study is to investigate the influence of simplifications in the Monte Carlo set-up of the x-ray tube on the observed differences between measurements and calculations in the scattered field. At a distance of 50 and 100 cm from the focal spot, air kerma calculations are done for different angles from -90° to 90° from the central beam axis in steps of 15° with the Monte Carlo software code MCNP-X. Different calculations were performed where each time a component of the simulated x-ray tube (collimator, filters, etc) or the environment (walls) is included. Scattered doses are also measured with thermoluminescent dosemeters. For the most simplified geometry of the x-ray tube, measurements are on average 70% larger than the calculated results. A much better agreement with the measurements is observed for more realistic calculations. The current work applies to a particular source in the SCK•CEN calibration laboratory; therefore the obtained results are representative and relevant for studies in calibration laboratories. As clinical sources have more shielding material and as in real imaging situations the scatter generated at the patient is much larger than the scatter leaking from the source, the results of this study have a limited impact on the wider field of clinical dosimetry.