(April 2002) Placed the order for the spare parts and stainless steel
ingots. Placed the order for the lasers. Placed order for the
pyrex tube. Tooling of the base. Tooling of the suspensions for the knife-edge hinge of the pendulum.
(July 19, 2002) These are the first pictures of the clock that should
outdo the Chronolithe: it will inherit all the knowledge that
designing the Chronolithe slowly distilled into me.
This radiometric pendulum will be superior in airtightness and accuracy.
Airtightness because there will only be one single seal
at the bottom and no rod passing through like in the Chronolithe. Accuracy because it will be controlled by a laser, not infrared,
beam. The cherry on the cake is how easy it is to modify or adjust : you let air in, slide the tube upwards and that's it.
The pendulum engine is the same as for the Chronolithe: light, acting
through the radiometric effect discovered by Sir Francis
Crookes at the end of the 19th century, that no one had used to move a clock's pendulum. Briefly stated: as it passes in front of
the laser sensor the pendulum triggers halogen lamps whose light push it in phase with the oscillation. The period is adjusted by
varying the pendulum's length, the system operates as long as there is light and vacuum inside the tube. The initial swing is
produced using a magnet outside the tube. The pendulum is totally free from any outside influence. The hours, minutes and
seconds hands are driven by a step-motor connected to two small solar panels powered each time the lights turn on. On the
picture below they are barely visible, sideways between the frame rods.
Status as of October 15, 2002
The tube is 1300 mm high (51"), 200 mm wide (8"), 20 mm thick (.8").
The rod is made of invar, the bob is a stainless steel
sphere filled with lead, weighing around 4 kilos ( 9 lbs). The base comes from a bottling machine and is made of bronze. The
tubes on each side house the halogen lamps and will soon be replaced with stainless steel spheres. The frame is currently
subject to torsion up to .004mm (.2 thousandth of an inch). This is not too much of an issue and it will probably be solved by
using a magnetic stiffener that can be locked from outside. The suspension is a hardened-steel knife-edge from weighing scales.
(July 28, 2002) With this new clock the quest for isochronism will be
a quest for harmony. The problem is simply stated: the
pendulum's period should remain the same when the lights are switched off and it continues swinging out of its own inertia. The
recording below shows what happens with the Chronolithe over 1 ½ hour: the first part shows the beat duration with the
pendulum oscillating by itself, the second one with the 4 lamps operating at 5.4 volts and the last one with only 3 lamps
operating. The steady state is the first one, when nothing influences the pendulum. In the second segment we can see that the 4
halogens lamps accelerate the pendulum by about 6 microseconds per beat. In the third segment, with 25% less power, it is still
fast by about 3 microseconds. This translates into around .25 second delay per day if a lamp is turned off. These are the limits
of the Chronolithe self adjusting feature.
The challenge for the new clock is clear. It should ideally beat at the same frequency with or without power.
(August 11, 2002) The clock has been under vacuum for two days for airthightness
testing and will start working tomorrow.
The breadboard for the lamps control unit is completed, the lasers have arrived, and I'll soon be able to start the tests. The first
efforts will be focused on improving the shape of the two mica sheets that receive the light. The pendulum rests on a
hardened-steel knife-edge from precision scales. The first inertia tests under vacuum are very good: with a first swing of 2
centimetres there's enough energy for at least 24 hours. To be compared with the 6 hours of the Chronolithe under the same
(August 12, 2002) It has started! And it works so well that I do not
even dare touch it. It is radically different from the
Chronolithe, even if they have similar engines. Let's see how they are different. The Chronolithe's pendulum was sort of
bouncing on a wall of light, being pushed back every second like a ping-pong ball. In this new implementation a rod under the
pendulum cuts through a laser beam. When the beam is interrupted one way it lights one of the lamps, when it is interrupted the
other way it lights the other one. Which means that the pendulum only receives an impulse at mid-course, both ways. That's
exactly what makes self-regulation possible. Assume that one of the lamps is out of order: amplitude will necessarily decrease
and the pendulum will move slower across the laser beam. As it moves slower the lamps will be on longer, the mica sheets will
gather more heat and will therefore accelerate the pendulum. If this theory is valid amplitude should stay constant until it stalls by
lack of energy. That's what needs to be proven in the next days.
(September 14, 2002) One month of airtightness and accuracy testing.
I'll now be able to resume work: tooling of the dial, of
the solar cells stand, of the lamps housing, designing the mica sheets used to activate the pendulum.
(November 15, 2002) The design of the lamp electronic control unit is
completed and I must say it resisted fiercely until the
very end. What is most difficult is not so much to make a clock work but to simplify it to an extreme, until each and every part
is vital. Until now the clock was controlled by temporary lasers and a control unit breadboard. All this worked OK until I
decided to simplify it...
Facts: Two lasers "look" at the bob which, when it moves between the
beams and depending on which way it goes, triggers
one of the two lamps at the bottom. The alternating light of the upper lamps powers the solar panels that drive the seconds
Once under vacuum the pendulum is free of any external influence except
light, which is the only moving force of the pendulum
and hands. This clock would work just as well at the bottom of a swimming pool..
English translation by Jean-Marc Julia
Vidéo clip of the swinging pendulum