Very accurate positioning and low velocity tracking motion are required in devices such as surgical tools, computer disk drives, assembly robots, micro manipulators, etc. For most of these devices, performance is limited due to friction which results in position and velocity errors. In this research the reasons for this performance limitation have been investigated in detail. As a result of this study, a spectral-based friction model, the first model to accurately describe nonlinear friction (including negative damping friction) in complex mechanisms as a function of both position and velocity, is developed. This model is also the first model that can describe the contribution of each and every part of a complex mechanism to its overall friction. A better understanding of the phenomena resaponsible for the existence of nonlinear friction in complex mechanisms, achieved during the development of the spectral-based friction model, led to the successful design of two new friction compensators capable of performing precise positioning and low velocity tracking in the presence of negative damping friction. The first controller developed is a high gain PID controller, that was experimentally verified and shown to be capable of performing low velocity tracking despite negative damping friction. The second controller developed is a fuzzy logic sampled-data controller which uses narrow torque pulses to position a mechanism at a desired position despite negative damping and static friction. The proposed fuzzy logic controller, activated after a conventional controller (i.e., a high gain PID controller) brings the mechanism close to the desired position, reduces or eliminates the steady state position error created by the conventional controller. This controller is the first experimentally verified controller that is capable of positioning a mechanism at the desired position in the presence of negative damping friction.