A pendulum with motion as described above is of no use for a continuous
demonstration of the rotation of the earth but several
simple solutions to the problem are possible.
The solution adopted by most museums is to use a much longer pendulum. This is because for equally well designed supports, the difference period dT is directly proportional to the 3/2 power of the length. A pendulum of 25 m length might have a difference period of 1000 hr and the growth in elipticity in any 8 hr day would be noticeable but probably not critical.
A second solution and the one adopted here is to introduce a Charron ring; this is the large brass ring our pendulum strikes near the end of each swing. Ideally the Charron ring would stop the pendulum completely and then release it from rest so that the motion is again like that shown in Fig. 2. Such an ideal ring would not affect the Coriolis turning of the plane of oscillation because the Coriolis force acts only while the bob is in motion and produces most of its deflection of the bob as the bob passes near its equilibrium position.
A real Charron ring however cannot stop the pendulum; it can only slow it down by a combination of impact and friction forces.
Now if the bob strikes the ring very gently, friction between bob and ring will be inadequate to reduce the tangential motion of the bob substantially and the elliptic motion will remain. If the bob is made to strike the ring very hard in an attempt to enhance the damping of the elipticity then one cannot guarantee the large impact forces do not on their own cause a deflection of the orbit and thus give a contribution to the precession of the pendulum. A compromise must be struck which leaves an elipticity in the motion with b/a ratio considerably larger than that shown in Fig. 2. As a result, the restoring force on the bob does have a substantial component perpendicular to the plane of oscillation and since this force is not precisely harmonic, it will lead to an additional deflection of the bob and an additional precession.
Now Crane's experimental discovery in 1981 was that to eliminate this residual precession one need only modify the force law to make it harmonic on average. He also noted that this could be done very simply with a pair of magnets, one on the bob and one on a fixed support underneath. The magnitude of the force modification is adjusted by changing the magnet separation.
This extra magnetic force has very little effect on the mean precession of the pendulum which is approximately we' as driven by the Coriolis force. The essential point is that elipticity driven by the anisotropy alternates between positive and negative values as the pendulum sweeps between the two axes where the frequency takes on its extreme values w1 and w2. Then, as the elipticity alternates, so too does the additional precession resulting from the anharmonic restoring forces. The total precession rate is alternately greater than and less than we' and the magnetic force is adjusted to minimize these fluctuations and allow the true Coriolis precession we' to be observed at any time of day.