Sensorless control of hybrid stepper motor

Today thanks to low cost and high performance DSP's, Kalman filtering (KF) becomes an efficient candidate to avoid mechanical sensors in motor control. We present in this work experimental results by using a steady state KF method to estimate the speed and rotor position for hybrid stepper motor. With this method the computing time is reduced. The Kalman gain is pre-computed from numerical simulation and introduced as a constant in the real time algorithm. The load torque is also on-line estimated by the same algorithm. At start-up the initial rotor position is detected by the impulse current method.