Abstract
A machine for testing turbine nozzles by the reaction method, which was described in a previous paper, was used to test a series of convergent-divergent turbine nozzles. The results of these tests, along with the test of a convergent turbine nozzle, are compared with each other and with analytical values. Two kinds of analytical values are employed, namely, the usual values obtained from an assumed isentropic expansion from inlet state to exhaust pressure, and the values obtained from the assumption that the processes in the nozzle are isentropic except for a normal shock which takes up a position in the nozzle such as to cause the stream to fill the exit area at the exhaust pressure whenever possible. This latter kind of analytical value involves no shock when the exit area can be filled at the exhaust pressure by means of isentropic processes only, or when the exhaust pressure is lowered so far that the shock has passed out of the passage. The agreement of the test results with the calculated results of this latter kind is good, and the disagreement which exists can be attributed largely to separation at the shock and to transmission of exhaust-pressure effects upstream through the boundary layer.