The current precision reached by lattice QCD calculations of low-energy hadronic observables, requires not only the introduction of electromagnetic corrections, but also control over all the potential systematic uncertainties introduced by the lattice version of QED. Introducing a massive photon as an infrared regulator in lattice QED, provides a well defined theory, dubbed QED M , amenable to numerical evaluation [1]. The photon mass is removed through extrapolation. In this contribution we scrutinise aspects of QED M such as the presence and fate of the zero modes contributions and we describe the determination of the photon mass corrections in finite and infinite volume. We demonstrate that the required extrapolations are well controlled using numerical data obtained on two ensembles which only differ in volume.