Driving synchronous machines requires early and accurate knowledge of the absolute position of the rotor; current solutions based on resolvers, sin–cos, or absolute encoders are complex, bulky, and costly. Hence, in this work two variants of absolute rotary encoder based on the Vernier method are analyzed. One, already discussed in the literature, displays the Vernier scale across the whole circumference (full-Vernier), and the other shows half Vernier traces over the entire perimeter (half-Vernier), which is an original feature. Both the implementations are characterized by only two tracks and as many sensors: the proposed conditioning algorithms provide the absolute angular position as a function of the time delays between the wave edges generated by the two traces, thus being of easy implementation on low-cost microcontroller units (MCUs). The Vernier encoders are also compared with the state-of-the-art absolute and relative solutions, i.e., incremental, binary, and Gray-code encoders. Experimental tests are carried out to assess the accuracy of the proposed sensors. The investigation shows that: 1) the full-Vernier cannot provide, in practice, a reliable estimate of the direction of rotation and of the actual angular sector without resorting to a third sensor; 2) the half-Vernier produces a trusty measurement of the absolute angle and velocity; and 3) can give a reliable position result with less than 30° shaft turn, but it can suffer from marginal performance degradation at low velocities in conjunction with high accelerations. Compared with the Gray encoder, the half-Vernier provides a simpler and more compact hardware for a given resolution, similar to that of an incremental encoder, at the expense of a small accuracy reduction at low speed.
Read full abstract