The solid-fluid phase behavior of mixtures is important in a variety of real situations of industrial interest, including paraffin precipitation from hydrocarbon reservoir fluids, supercritical extraction or precipitation processes, and cryogenic processes among others. In the case of binary mixtures with the presence of fluid and solid phases, such behaviors can yield different topologies in terms of critical lines, three-phase lines, critical end points and quadruple points. The classification proposed by Yamamoto et al. in 1989 for those topologies is probably the best known and most used one, besides being more complete than others when considering experimentally observed behaviors. Nevertheless, a number of publications and in particular two recent PhD theses developed in our groups, found other topologies which deviate from the types considered in Yamamoto's et al. classification. These two theses employed different modelling approaches, using equation of state models for describing fluid phases and differing essentially in considering each solid phase as either made of the heavy compound in pure state or made of a solid solution. The deviations from the types in Yamamoto's et al. classification are compiled and analyzed in this work. Moreover, a new classification is proposed for binary systems with varying degrees of asymmetry. Being motivated by the study of paraffin precipitation from reservoir fluids, this new classification, in contrast to Yamamoto's et al. classification, leaves out of consideration the lower temperature range (up to the triple point of the light component) while, on the other hand, it incorporates some possible behavior variations, in the temperature range of more practical interest, which are not accounted for in Yamamoto's et al. classification. The new classification is shown to cover all possible phase behavior evolutions in the binary series of methane or ethane or propane with a heavier n-alkane and also of carbon dioxide with a n-alkane, considering both, what is known so far in terms of experimental data, and some features predicted with different modelling approaches, which have not been confirmed yet at the experimental level.