The problem of ellipticity in the motion of the ordinary Foucault pendulum and the error caused by it are discussed. A simple method of slightly modifying the force–displacement relation in such a way that precession does not result from ellipticity is presented. The practical range of length of the Foucault pendulum is extended downward substantially. As experimental verification, results from a 70‐cm pendulum, mounted as a ’’wall clock’’ are given. Rates of turning from daily readings show a 2% rms deviation from the mean, and the long‐term average is within 2% of the rate calculated for the local latitude.