The Ergun equation is commonly used to predict velocity-pressure drop behaviour in packed beds. For many particular systems, the measurements deviate significantly from this relationship. Because of this, the equation is often treated as a purely empirical relationship, and its factors as fit parameters. This paper shows that the physics-based model underlying the Ergun equation should be treated by careful investigation of the included parameters. It is shown that discrepancies between experiment and theory for particle collectives in the millimetre range and above are caused by insufficient methods to determine the specific surface area. This effect is especially large for coarse particles with very rough surfaces. As a 2D measurement method, dynamic image analysis can not provide a good measure of specific surface area. It does not account for both macroscale particle shape and microscale surface roughness, which is confirmed by an analysis of 3D particle measurements from X-ray microscopy. (C) 2020 Published by Elsevier B.V.