The synchronization between wireless sensor nodes represents a big issue in all low-cost systems in which each node performs measurements using its own sensors and stores locally the measured values. In this paper, we will analyze in detail the implementation limits concerning the simplest synchronization strategy that can be implemented for Wireless Sensor Networks (WSNs). In fact, we consider a type of star network in which a master periodically sends a synchronization message to the slaves belonging to its own subnet. The considerations made are referred to commercial low-cost microprocessors and System-on-Chips (SoCs). Solutions too simplistic lead to large synchronization errors or excessive power consumption; correct approaches require continuous calibration by the slaves and are therefore impractical.