Conceptual role of Short Range phenomenon in Statistical Thermodynamics of liquid alloys is discussed. It is shown why the popular model theories used by Calphad fall short in bringing this phenomenon to full usage. In contrast to this, presented here in details Theory of Inhomogeneous Short Range takes Short Range as basic element of formalism. Combinatorial factor provided by theory expresses the Statistical Sum of entire system as weighed average over Statistical Sums of small groups of atoms conventionally selected in the alloy. Equations of the theory are explicitly resolved in a parametric form.The theory in its tetrahedron approximation is applied, as introductory example, for the presentation the thermodynamic data of multiple two-component systems described previously in the literature by Redlich-Kister' polynomials.

The equations of the Quasichemical Theory are deduced directly from the equations of the Theory of Inhomogeneous Short Range, establishing TISR as the rigid generalization of the Quasichemical Theory.Methodological advantages of TISR are demonstrated.The Modified Quasichemical Model is critically discussed as other generalization of the same Quasichemical Theory.Multiple comparative examples are provided.

An improved combinatorial factor allows to reproduce the correct critical tem-peratures for the Ising lattices.As result, the accuracy of thermodynamic values calculated near the criticalpoint increases by the factor 2 - 3 for the most important lattice types.It is shown why one-particle (Bragg-Williams) and two-particle (Quasichemical)approximations cannot be adequately applied for description of important nuancesof interatomic interaction.The problem of negative entropy values manifested at low temperatures andtypical for model theories is discussed. It is being argued that this problem has tobe treated as exaggerated.

No abstract

Internal problems of Association Solution Model are analyzed and based on this analysis a few important modifications are introduced in the formalism and in the interpretation of the model.Reconstructed model is free from the named problems and has significant advantages from the point of view of consistency and applicability. A set of comparative examples is provided.