154 BELL SYSTEM TECHNICAL JOURNAL 



ACCLRACY OF THE ClOCK 



So far wc have not considered the possibility of error in the clock 

 as a factor. Of course, the fork cannot keep better time than the 

 clock which controls it. 



The clock used at present was made by L. Leroy and Co., Paris, 

 electrically driven and beating half seconds. The drive consists of 

 an electric circuit including a single primary cell mounted in the 

 clock, a driving coil and a contact which is closed by the escapement 

 wheel for approximately .1 second in each second. Attached to the 

 lower end of the pendulum is a steel bar which moves into the driv- 

 ing coil as the pendulum oscillates. The electrical impulse is so 

 timed that the driving coil gives the pendulum a slight pull as it is 

 entering the coil. This impulse is sufficient to keep the pendulum 

 oscillating. An additional contact on the clock is used to furnish 

 an electrical impulse for timing purposes. 



Time records of the clock have been kept over a period of several 

 months and the rate has been found to be constiint to about one-half 

 second a day, which is better than 1 part in 150,000. Since this 

 accuracy is not verj' much greater than the precision with which the 

 fork keeps in step, any further accuracy will require refinements in 

 the clock itself. With this object in view, an investigation was made 

 of the possibility of obtaining greater accuracy from the existing 

 clock. 



Errors are of two kinds. First, if the timing contact is obtained 

 by the operation of the escapement wheel, there may be a cyclic 

 variation in the length of time between successive impulses extend- 

 ing over one revolution of the wheel, (1 minute) even though the 

 pendulum keeps perfect time. This has been found to be the case 

 in some of the best clocks in the country. This error can be over- 

 come by taking the contact direct from the pendulum. The con- 

 tact we are using at the present time is of this type obtained from 

 the pendulum by means of a photo-electric cell. 



The optical system is shown on Fig. 4. Light from the source A 

 is concentrated on the mirror, which in turn reflects it on to the photo- 

 electric cell. When the pendulum passes through the center of its 

 stroke, it momentarily cuts ofT this beam of light. This causes a 

 large increase in the resistance of the photo-electric cell, the change 

 taking place almost instantaneously. 



Referring to the diagram of connections on Fig. 4, the potential 

 of battery B is disided almost equally between the photo-electric 

 cell and the grid of the tube if the grid leak is made approximately 

 equal to the resistance of the cell when exposed to the light. This 



