The digital SQUID magnetometer takes advantage of flux quantization in a superconducting loop in order to measure magnetic fields. The core element of the digital SQUID is a Josephson comparator with a superconducting antenna loop attached to one of its junctions. Evaluation of the circuit from the system's point of view requires an analysis in the frequency domain. In order to obtain a high-resolution fast Fourier transform, large datasets are necessary which are difficult to generate with transient simulation tools. In this work we derive a behavioural model for the digital SQUID in order to overcome restrictions imposed by transient simulation. By means of this model the influence of the comparator grey zone and the input loop inductance on the system performance was analysed. In order to assess the system, evaluation criteria based on the power spectral density were applied, which are commonly used for characterization of semiconductor analogue to digital converters. As a result of this study, design guidelines for an optimum antenna inductance depending on the comparator grey zone are derived, allowing us to achieve an optimum system performance in terms of noise and distortion.