Unknown reason why the frequency spectrum of a "knocking force" (impulse) gets some spike

During my experiment (vibration analysis), I deliver a knocking force from a shaker (and the shaker itself has an impulse or an half-sine wave signal). The system automatically calculates with the FFT algorithm and gives me the following graph for the input force signal. Note that I was investigating different parameters, and this is not the proper experiment yet. Though the input signal would be probably kept the same for the main experiment.

My question is that: why do we spikes here, in the frequency spectrum?

Thanks for your help.