This note presents a difficulty encountered in random vibration testing of a satellite as well as the cause. The vibration testing system did not work because of intermittent unusual vibration of the shaker. It was found to be caused by a biharmonic signal introduced in the driving signal by “3σ clipping” of nallow-band random driving signal because a resonance frequency of the mounting adapter and the shaker was twice an anti-resonance frequency and the resonance amplitude ratio was very large. The importance of carefull design of mounting adapter was renoted.