The paper presents a method to estimate the inverse of the rotor time constant of the induction motor. The estimation is done using a sliding mode observer under the assumption that the stationary frame fluxes are known. These fluxes are first obtained using a voltage model observer; they are also used for field orientation. With measured voltages, currents, known fluxes and speed, the rotor time constant is estimated using a pseudo-MRAS sliding mode observer with dual feedback terms. The paper shows the design of this observer — this works well under ideal conditions. However, if the speed is inaccurate or if the magnetizing inductance saturates, the estimation accuracy suffers. The paper develops a model for the saturated induction motor and, using the equivalent controls that correspond to the sliding mode terms, attempts to estimate the saturation level — however, it is found that this is not possible.