With the present thrust of energy conservation measures in residential, commercial, and industrial sectors, refrigerant systems have been the object of considerable study. Manufacturers of equipment as well as end users are seeking ways of reducing the operating cost of air-conditioning units and other energy transfer systems by increasing the efficiency of individual components. As the system modeler knows, a change in efficiency of a single component will cause the entire unit to seek a new equilibrium operating statepoint requiring a complete set of balance calculations to analyze operation; over the ambient range. Refrigerant systems in their simplest form use two heat exchangers, a compression device, and a throttling mechanism. The throttling mechanisms most commonly used are thermostatic expansion valves or capillary tubes. These two devices cause radically different system behavior characteristics and modeling the flow of refrigerant in-throttling devices has long been a critical part of systems analysis. The purpose of this paper is to present a mathematical, iterative approach to modeling, the flashing flow in capillary tubes from either liquid or two-phase inlet to either liquid or two-phase outlet, and show the variables which must be converged to generate balance calculations for refrigerant systems.Considering the abundance of pressure drop relationships available in the literature, choosing one which best models capillary flow is no mean task. An exhaustive search for two-phase correlations resulted in the choice shown in this paper because of its completeness-in all aspects of closed loop analysis. If the-method-appears simple, it is. But matching a methodology to a set ofrelationships which follow the well accepted ASHRAE curves is the contributionof this paper. As researchers have long known, capillary tube depressurizationis difficult to model because such factors as tube spiral curvature and tubeorientation may effect the flow parameters. This paper affords no clear-cutsolutions to universal modeling of these parameters, but offers a method whereappropriate factors may be considered in matching computer predictions to experimentaldata.