This research examines the allowable second throat cross-sectional area of exhaust diffuser in ground tests of a thrust-optimized parabolic nozzle with different expansion ratios. First, theoretical methods used to determine the minimum second throat area and the minimum starting pressure of the exhaust diffuser have been reviewed. Next, the results of small-scale experimental tests with compressed air at diffuser inlet area to second throat area ratios (Ad/Ast) of 1.6, 1.7, 1.8, 1.85, 1.9, and 2 have been presented. To describe the physical phenomena during the start and breakdown conditions, numerical simulation of the flow has been used. The results reveal that with 11% narrowing of the second throat from area ratio 1.6 to 1.8 at all nozzle expansion ratios (34.7–65), the minimum starting pressure has grown by 7–9%. Examination of the further narrowing of the second throat up to area ratio 1.85 indicates that due to temporary flow choking in the second throat, the starting pressure has grown by up to 40%. Also, at diffuser area ratios 1.9 and above, due to severe flow choking in the second throat, the diffuser never starts. Present examinations show that with a large increase in the diffuser area ratio parameter, the nozzle pressure at which the shock pattern transition occurs in the nozzle does not change significantly. Also, diffuser breakdown pressure in the test of nozzles with different expansion ratios has shown very little dependence on the diffuser area ratio parameter. Our examinations indicate that the maximum permissible area ratio reaches the upper limit of 1.8, which is very different from the values calculated by conventional theoretical methods. Further, in this research, it has been shown that the suitable correction factor in estimating the minimum starting pressure with normal shock method in the second throat has lied within 1.4–1.7. The results indicate that the correction factor has heavily depended on the diffuser area ratio parameter. Nevertheless, this factor falls within a close range using a constant diffuser area ratio in testing of a thrust optimized parabolic nozzle with various expansion ratios. Thus, for a simpler application, the area ratio-dependent mean and maximum correction factors have been introduced which have maximum error of 3.2 and 5.6%, respectively.