In drive designs with write data rates >1 Gb/s, transitions sent to the write head collide at 1T with reflections from earlier transitions. This increases the need for write-driver voltage headroom, efficient interconnects, and efficient magnetic materials. Design alternative studies in this regime require detailed models of all components in the write process. We modeled nonlinear bit transition shift (NLTS) and spectral signal-to-noise ratio (SNR) measurements using our preamp/interconnect/writer/media model and compared it to measurements of the actual devices that were modeled. We describe methods for reconciling write driver models with varying head and interconnect impedance, verification performed with scanning Kerr effect microscopy (SKEM), and magnetic models with instantaneous nonlinear effects included.