The configuration of the magnetic fields is an essential ingredient of Tokamak physics. In modern day devices, the magnetic topology is normally derived from equilibrium codes, which solve the Grad-Shafranov equation with constraints imposed by the available measurements. On JET the main code used for this purpose is EFIT and the diagnostic more commonly used are external pick-up coils. Both the code and the measurements present worse performance during ELMs. To quantify this aspect, various statistical indicators, based on the values of the residuals and their probability distribution, have been defined and calculated. They all show that the quality of EFIT reconstructions is clearly better in absence of ELMs. To investigate the possible causes of the detrimental effects of ELMs on the reconstruction, the pick-up coils have been characterised individually and both the spatial distribution and time behaviour of their residuals have been analysed in detail. The coils with a faster time response are the ones reproduced less well by EFIT. The constraints of current and pressure at the separatrix have also been varied but the effects of such modifications have not resulted in decisive improvements in the quality of the reconstructions. The interpretation of this experimental evidence is not absolutely compelling but strongly indicative of deficiencies in the physic model on which JET reconstruction code is based.