This paper presents the results of an experimental investigation of steady flow through a pair of exhaust poppet valves. An account is given of the gas exchange process on engines which use poppet valves and the reason why pressure losses should be kept to a minimum is explained. Tests carried out on the cylinder head of a uniflow two-stroke cycle engine are described following a brief description of the apparatus used. The results of a simple analysis of incompressible flow are also given. It is shown that the two previous models of flow through a valve, namely the sudden enlargement and constant static pressure, both give unrealistic pressure losses for large area ratios, i.e. at high valve lifts. A new model is introduced which leads to realistic pressure losses at small and large area ratios, i.e. at low and high valve lifts. Effective areas for the present tests are calculated on the basis of the constant pressure model, and details of calculation of pressure losses are outlined. The blockage effect caused by placing the exhaust valves near the cylinder wall is given in the discussion of the test results. This is zero for 0 < l/d < 0·08, but reaches a maximum blockage of 10 per cent at l/d = 0·28. With unrestricted twin valves the effective area is about twice that of a single valve up to l/d = 0·18 with a progressively larger effective area at lifts up to 13 per cent higher at l/d = 0·4. A comparison is also made with other data readily available. The pressure losses determined from the tests were analysed using a parameter derived in the simple theory. The parameter used is found to be almost independent of pressure ratio and the results are presented by means of this pressure loss parameter as a function of valve lift. The representation provides a quantitative method of comparing the performance of a given configuration of valve and port. On this basis the twin poppet valves are shown to give a slightly higher pressure loss than a single valve.
