My experience is limited to the Fairchild Metro (SA-227) w/ Garret TPE 331 engines and direct drive Dowty-Rotol props. On our models, we actually didn't have the typical Master/Slave system found on most free turbine turboprops.
On the Metro, each governor had a magnetic RPM sensor that fed into the Sync Controller (The"box"). Each prop governor also had a "Biasing Coil" integrated into the speeder spring/fly-weight assembly so that when an electrical current was ran through the coil, it would cause an "artificial under-speed" condition = fly-weights move in condition (I imagine by either adding a little bit of electrical resistance that the flyweights would have to overcome or by magnetically attracting the flyweights inward ever so slightly). When this condition is induced, normal governor operation will respond as usual by increasing prop RPM until a new higher RPM equilibrium is established. Each prop RPM is fed into the sync controller and the sync controller controls which governor's coil is "biased."
With this type of system, a prop's RPM can only be INCREASED so the sync controller will increase the lower RPM prop to match the higher RPM prop. In other words, there is no master/slave prop, which ever is lower gets raised. Since my knowledge of normal sync systems is purely academic, I imagine this system is preferred for direct drive props because if the prop slows down, so does the engine and I imagine engineers didn't want the prop sync to be able to reduce engine speed. (Note: our training material describes this idea by stating that "The Propeller Synchrophaser will never override the RPM selected with the speed levers.")
Like other prop sync systems, the props needed to be within a certain RPM of each other in order to be allowed to work and ours was a max of 1.4% RPM difference between prop/engine speeds.