Leon (technical)

Eddy Clock and clock

Short (less than 1 metre) eddy current pendulums are rare throughout the world, as the problems encountered in their construction quickly become insurmountable. Long pendulums are infinitely easier to build in comparison. The first problem is the suspension. It must be perfect, otherwise the pendulum will choose the most favourable plane and stay there. The steel wire cannot have any internal tensions and must be perfectly spherical. The fastening should also be perfect: the classic small precision chucks with 4 jaws are not enough, as they still need to be carefully polished. The second problem is the precession of the ellipse. This phenomenon is only very slight on large pendulums and is not taken into account during their construction. On pendulums of less than 2 metres, it causes greater forces than the eddy effect. It must therefore be cancelled, or at least reduced. One way of doing this is to interpose a Charron ring on the upper third of the wire, the inside diameter of which should be perfectly polished. A short pendulum cannot function without a system for reducing the precession of the ellipse. All these phenomena put together make a Foucault pendulum of less than 1 metre almost a miracle. The time spent on machining and assembly will not exceed one tenth of the total time, all the rest being lost in adjustments. And since each new adjustment requires a minimum wait of 16 hours to see the results…

This pendulum has an oscillation period set to one second. I use this phenomenon to turn the seconds hand, which drives the minutes and hours. The fact of using a Foucault pendulum as a clock imposes other constraints: 1) to be able to regulate precisely the duration of the period 2) to compensate the variations of length of the pendulum due to the differences of temperature 3) to stabilize the amplitude to respect the isochronism. The first Foucault pendulum I designed(Pestoline) used an ingenious but perfectible system: the pendulum wire slid through a hole and could be adjusted upstream, in the same way as a guitar string is tuned. But shortening or lengthening it also changed the position of the magnet in relation to the coil, thus varying the strength of the impulse. This made fine-tuning the duration of the amplitude a sport. With this model, the wire does not move: it is the hole that can be raised or lowered. This system even allows the period of the balance to be adjusted while it is moving without disturbing its course, an option not found on conventional clocks. Temperature compensation is provided by a bimetal located above the assembly. The problems inherent in the machining of the Charron ring were solved by the use of a ruby. The roundness is thus perfect and the wear zero.

But why two separate dials? The one on the left indicates the time given by the pendulum, and the other is a quartz clock which serves as a reference here. So at last we have a clock that can satisfy everyone: people in a hurry will always read the right-hand dial because it is more accurate, poets will enjoy the left-hand dial (because what we really like in people as in things, isn’t it precisely the imperfection?) and scientists will carefully read the average of the two. For example, if the second hand of a pendulum is a little behind one moment, it will then catch up a quarter of a revolution later. This is due to the exentration of the Charron ring, and the average of these fluctuations is zero. So this clock cannot be set like a normal clock, because any eccentricity of the Charron ring causes an acceleration of the passage of the pendulum, which will invariably be followed by a slowing down of the same amplitude 1/4 of a revolution later. The average time of half a revolution must therefore be set. This will vary according to the latitude where the clock is located. In my workshop it is 16 hours 34 minutes and 14 seconds.

The first tests reveal an inaccuracy (perfectible by adjusting the height of the balance) of less than one second per day. Those who want to consult the data below can ask me. (These data can only be opened with Bryan Mumford’s MicroSet software)

A short working eddy current pendulum is almost a miracle: it must be installed permanently. It is not possible to take it and run it without further ado in several different places, because the level settings are very precise and do not need to be changed once the pendulum is working properly. These settings can take up to two weeks to set up.

Failures and prototypes

Not all my tests were successful, far from it! Three prototypes will be developed below. What they have in common is that I thought they would all work on the first try… but they never worked.

1) Gimbal suspension

This is a gimbal suspension consisting of two Berkel balance knives resting on two ball bearing cages, allowing it to move in all directions. While this type of suspension is best for long pendulums, it is very difficult to use on a short pendulum because the suspension must be perfectly balanced. To do this, the bottom of the suspension must be significantly lighter than the top, so that the suspension straightens itself out. Any differences in level can then be corrected with small weights until the suspension is vertical. Another way to adjust the levels is to place two small magnets on the two knives of the suspension. The adjustment is made by moving the magnet away from the centre so that the centre of gravity moves. When this is done, the axle of the balance can be screwed in and testing can begin. This prototype never worked properly.

2) Tungsten wire suspension

This suspension uses a 0.05mm thick tungsten wire. It is almost invisible, and anyone who sees the pendulum move thinks that the pendulum is levitating. I had chosen such a thin wire to try to avoid any internal tension that could influence the rotation of the pendulum. Unexpectedly, and without my knowing why, this pendulum never worked. All attempts to adjust the levels and rotation of the suspension were unsuccessful.

3) Ball suspension

More prototypes that never worked! In the first picture, you can see that the suspension is a 1mm tungsten carbide ball resting on a sapphire bed. The suspension is free to go in any direction. Practically, it never turned as it should have. This prototype oscillated for 9 months without any adjustment being able to influence it

In the photo below, you can see another test using the same principle. This is a rolling ball on a steel plate. The attempts were just as unsuccessful as the tungsten ball prototype.
The adjustments of the suspension balances are very similar to those of the gimbal. The first thing to do is to remove the complete balance and make the suspension as light as possible. The bottom of the suspension must be significantly heavier so that it will lift itself. If it does not rise perfectly vertically, the balances must be changed by removing weight from the brass ring. This is done by simply grinding the ring on the side where it leans (you can see the marks on the photo)