The paper presents the results of an extensive series of computer-simulation tests to determine the effects of nonlinear distortion and adjacent channel interference (ACI) on the tolerances to noise of an uncoded eight-phase-shiftkeyed (8PSK) signal and a rate−3/4 convolutionally encoded 16-ary phase-shift-keyed (CE16PSK) signal for transmitting over satellite channels. Both signals have the same theoretical spectral efficiency of 3 bit/s/Hz, but the CE16PSK signal has an asymptotic coding gain of 4 dB over the uncoded 8PSK signal. Several different earth stations are assumed to have simultaneous access to a given transponder in a satellite by frequency division multiple access, so that ACI can be introduced into the wanted channel by the channels in the immediately frequency bands. It is assumed in the study that the high-power amplifier (HPA) of each earth station may or may not introduce nonlinear distortion into the transmitted signal, but the satellite transponder is linear. The optimum performances of the systems are compared with these using a rate-3/4 convolutionally encoded 16-ary quadrature-amplitude-modulated (CE16QAM) signal and a rate-3/4 convolutionally encoded 16-ary amplitude-phase-modulated (CE16APM) signal. It has been shown that, in the absence of ACI, the CE16PSK system has advantages of 1.3 dB, 2.6 dB and 2.8 dB over the CE16APM, CE16QAM and 8PSK signals, respectively, at a bit-error rate of 10−4; and, in the presence of ACI, the advantages are 0.6 dB, 1.3 dB and 3.5 dB, respectively.