PESTOLINE (Combination Foucault pendulum and clock)

Foucault pendulums are usually long, situated in tall buildings, with the heaviest possible bob (up to 150 feet high, and 600
pounds). The period of these devices can be as long as 10 seconds. But what about short pendulums? For a long time I had
contemplated designing a clock based on the Foucault pendulum principle. It had to show the earth rotation while indicating the
time, latitude and hemisphere. To do this it must be one meter long to achieve a one-second oscillation. The first prototypes
started operating in 2000, with trials of axial, magnetic, universal joint, and pin vise suspensions. The only thing I learned is that
a short Foucault pendulum is much more difficult to design than a long one. Then the Chronolithe became my priority  and I
postponed the project.
 
 
 Designing a Foucault pendulum is easier if it is over 2 meters high (6 feet). But mine demands a 1-meter length to achieve a 1-second beat. Only a few people in the world have been able to design short pendulums that operated correctly. I'd like to mention here the pioneer work of Haym Kruglak and Stanley Steele in 1984 and H. Richard Crane en 1981. The two main issues with short pendulums are:  1) suspension;  2) elliptic precession. Suspension is by far the most important:  if it is less than perfect the pendulum will only swing in its "easiest" plane. One  solution is to cast the suspension fixture around the wire but because of the high temperature involved this will often alter  the wire characteristics. A commonly used solution is to grip the wire with a precision pin vise. But even in this case the pendulum will pick a privileged position and will stabilize in a fixed plane. That's because  commonly available vises all have small imperfections. Richard Crane would use a lathe to polish the jaws of the pin vise with very fine abrasive. Under this condition only would the pendulum work correctly. The other problem is how you cancel elliptic precession. This issue only exists for small pendulums and gets worse as the length shortens. This unwanted phenomenon can be somewhat minimised by placing a perfectly tooled and polished Charron ring at one tenth of the total pendulum length starting from the top.

Here is the first prototype that really worked, to be used as a test bench for building the other ones.  Once all the tests were
over it was also intended to end up in my kitchen since I had no clock at home...

The suspension has three functions: hold the pendulum, guide the pendulum wire with very high accuracy (1/100th centimetre)
and allow adjusting the pendulum length even while it operates. This last function is almost never found in pendulums but it
should prove very handy afterwards. I gave up the pin vise system as too difficult to tool with my limited capacity and allowing
no adjustment once set. Instead of being gripped by jaws the .18 mm steel wire goes through a hole, winds itself around a
pulley and is secured at the bottom clamp. A screw bends the wire between the pulley and the clamp therefore adjusting the
length of the pendulum. I can even add a bimetallic temperature compensation if needed.

Let's see now how  this pendulum is different from those that were built so far, i.e. the fact that it also serves as a clock. As far
as I know this has only been attempted once by H. Richard Crane (University of Michigan). His pendulum would show the
passing of time by reading the pendulum's oscillation plane on a graduated circle.  However his pendulum is not really a clock in
that it needs an external time reference, a timer that stops its operation 6 hours per day. By contrast my pendulum does both
functions at the same time: true clock and Foucault pendulum. Which triggers interesting questions like, for example: how do
you compensate for temperature since you cannot use invar (invar being magnetic you quickly end up with a compass); or how
do you show the time of the day with the pendulum since its scale is not 24-hour based? And when I have solved all these
problems I still will have to face the biggest one: very soon my cat will want to try his paw at Huygens' law.
 
Many people still believe that a Foucault pendulum rotates at the rate of one turn per 24 hours. This is only true at the North pole, clockwise, and at the South pole, anticlockwise. As the latitude nears the Equator the rotation rate gets slower: in Paris it is 32 hours/turn, at the Equator it is infinite. Using an additional clock-controlled electro-magnet, Crane stopped the pendulum rotation  6 hours per night  (between midnight and 6).

Some facts. Magnetically driven pendulum; the electromagnet is activated upon detection of the bob. Brass for almost all parts
to eliminate magnetic interferences. Hours, minutes and seconds are shown on the face, hours are also shown on the bottom.
The disk at the upper third is a safety device to stop the bob from falling should the wire break. It only happened once, by my
fault, but the amount of swearing on this occasion convinced me that it the device was worth making.

Amplitude is stabilised through another of Léon Foucault's discoveries: the brake. It is implemented here as the precisely
dimensioned, non magnetic, graduated ring over which the bob ends its swing. As the magnet sweeps over it it creates eddy
currents (discovered by Foucault) that slow it down if it swings too fast.

A roller-coaster, that's what the recording of the first test of Prestoline looked like. I ran this test in my workshop as the length
of the pendulum was not adjusted, the Charron ring was not centered and I was working around all day long. Still, a number of
interesting things can be seen. Over the course of two days cyclical disruptions happen three times, with a period of 16.5 hours.
That's the time it takes for the pendulum to oscillate twice in a given plane (half  the time it takes for a complete revolution). The
disruptions therefore come from the Charron ring: it is not perfectly centered and some machining defects are still visible. The
same test done without the ring shows no disruption at all but the swing then becomes elliptic and the Foucault effect is barely
visible.

After a few adjustment of centering and balance here is what the computer shows.

(December 8, 2002) This graph shows that Pestoline is 7.8 seconds fast per day, which is easily taken care of by lowering the
bob a little. But more important the succession of peaks (one for each half-revolution) is more even. There still remain recurring
disruptions in the valleys, pointing to a small mechanical defect in the wire or in the suspension. That's how I use this clock as a
test bench to improve the next one. With this type of clock you do not have to trade timing accuracy for Foucault effect,
Prestoline shows that both can be fully achieved.

(July 2003) Pestoline is rewarded with the first prize at  the Kinetic Art Organization worldwide contest in the "engineering
ingenuity" category.

Those who have the MicroSet sofware by Bryan Mumford can donwload test data by clicking here, they will be able to see
what happens at the heart of a Foucault pendulum, down to a millionth of a second, disruptions caused by drafts, extraneous
oscillations, etc. This is a 5-day  long sample. Those who are interested in how to adjust a Foucault pendulum can learn about
the process by clicking here. This is advanced data meant for those who want to build their own pendulum.

Please email or call me if you want to order. The minimum manufacturing delay is three months. Please specify the
latitude at which the pendulum is to be used, to allow customisation of the hour ring. A full custom design based on
your own specifications is also possible.

I'd like to thank those who helped me during the design and implementation of this pendulum. The Gaillard jewellery in Sion, H. Richard Crane for his pioneer work,  and above all Bob Holmström (editor of  the Science Horological Newsletter) without whom I would probably  have completed this work.

English translation by Jean-Marc Julia

HOME English  HOME (French)