This article provides a careful comparison between the electric and mechanical excitation of a tuning fork for shear force feedback in scanning probe microscopy, an analysis not found in present literature. A setup is designed and demonstrated for robust signal and noise measurements at comparable levels of physical movement of the probe. Two different signal amplification methods, combined with two excitation ways provide three possible configurations. For each method a quantitative analysis, supported by analytical elaboration and numerical simulations, is provided. Finally, it is shown that in practical circumstances electric excitation followed by detection with a transimpedance amplifier provides the best result.