This chapter considers the application of iterative learning control to a rotary electro-mechanical system. The system has a non-minimum phase nature, a situation that is known to cause difficulties due to the inherent properties of non-minimum phase zeros (Chaps. 4 and 5). The slow convergence behaviour described in those chapters is analysed and verified experimentally and followed by detailed numerical simulations to illustrate the convergence performance of the unconstrained algorithms as well as their robustness against model uncertainties and noise/disturbances.