One of the well-known methods to share active and reactive power in microgrids, is droop control. A disadvantage of this method is that in steady state the frequency of the microgrid deviates from the nominal value, and has to be restored using a Secondary Control System (SCS). The signal obtained at the output of the SCS is transmitted using a communication channel to the generation sources in the microgrid, correcting the frequency. However, communication channels are prone to time delays which should be considered in the design of the SCS; otherwise, the operation of the microgrid could be compromised. In this paper, two new SCSs control schemes are discussed to deal with this issue: a Model Predictive Controller (MPC) and a Smith predictor based controller. The performance of both control methodologies are compared to that obtained using a conventional PI-based SCS using simulation work. Stability analysis based on small signal models and participation factors is also realised. It is concluded that in terms of robustness, the MPC has better performance.