The Michelson Velocity Of Light Experiment

We are indebted to Mr. E. C. Nichols of the Department of Instrument Design at the Mount Wilson Observatory, Pasadena, for the accompanying description of the latest experiment to determine the velocity of light in a vacuum.—Editor.

HE following is a brief description of the Michelson velocity of light experiment in vacuum. This experiment, made at the Irvine ranch near Santa Ana, California, during the period from May, 1930, to July, 1931, by A. A. Michelson of the University of Chicago, F. G. Pease of the Mount Wilson Observatory, and F. Pearson of the University of Chicago, was an endeavor to determine more accurately the velocity of light. The investigation was carried out jointly by the Mount Wilson Observatory of the Carnegie Institution of Washington and the University of Chicago.

In this experiment two systems of measurement were used : the first, that of time; the second, that of distance. In the case of the latter measurement it was felt that the direct measurement of a short base line without the additional triangulation might yield a higher order of accuracy. This was accomplished with a vacuum tube one mile long, the light traversing it eight or ten times. The base line was measured by Commander Garner of the United States Coast and Geodetic Survey. The results of this measurement have a probable error of ±0.47 mm. or one part in 3,400,000.


Two clocks were used in the time measurement. One was a ship's chronometer beating seconds on a relay and omitting every 59th second. The rate of the chronometer changed frequently, rarely remaining the same for 24 hours. The other clock was a constant-frequency oscillator of General Radio make, controlled by an oscillating quartz crystal, the period of which was increased by two multivibrators. The seconds relay and the syncro-clock of the constant-frequency oscillator were operated on a shaft driven by a unipolar motor. The unipolar motor was in turn operated by the multivibrators.

The rate of the constant-frequency oscillator was decidedly more constant than the ship's chronometer. Comparisons were made on a chronograph m having two ink pens operated by relays. The chronograph was driven by a synchronous motor and had a peripheral speed of one inch per second. Time signals were also recorded on the chronograph. These signals were received from Arlington four times a day on 1700 meters.


The mile-long three-foot diameter vacuum tube consisted of 60-foot lengths of riveted and soldered corrugated galvanized pipe No. 14 gauge, joined with rubber balloon tire inner tubes and cemented to the pipe ends with rubber cement. At each end two steel tanks were included in the tube to house the mirrors and their controls for the optical system. These tanks were fabricated from ^g-inch steel plate and welded. They rested on base plates of the same material and were sealed with a lead wire and hydroseal; no bolts were necessary. All joints were painted with several coats of Glyptal. Not a single machined surface was necessary in the entire vacuum container, which had a volume of 40,000 cubic feel and resisted a total collapsing pressure of 53,000 tons.

The mirror mountings and their controls in the tanks were supported independently on separate concrete piers by steel columns extending up through openings in the base of the tanks. These openings were sealed off by rubber sleeves. All adjustments to the mirrors inside the tube were made with small motors operated by remote control through a motor generator Selsyn system operated from the head station at the south end of the tube. Two Kinney vacuum pumps having a total capacity of 450 cubic feet of free air per minute were used to evacuate the tube. A vacuum of 0.5 millimeters was obtained.

THE OPTICAL SYSTEM Light from an arc lamp was imaged by a condensing lens onto a slit. The light coming through the slit passed above a small right-angle prism and onto the upper half of one of the faces of the 32-sided rotating mirror. This mirror rotated at approximately 500 revolutions per second. From the rotating mirror the light was reflected through a plane-parallel window into the tank to a diagonal flat mirror and then at right angles to a 50-foot focus concave mirror, which changed the light into a parallel beam. From the concave mirror the light passed over a 22-inch diameter flat mirror and fell upon another 22-inch flat one mile away at the north end of the tube. Thence the light was reflected nine times back and forth the length of the tuhe between the two 22-inch mirrors, finally emerging through the window in the tank over the same path but slightly lower and striking the lower half of the rotating mirror on a face adjacent to the one from which it was originally reflected. From this face the light was reflected into the small right-angle prism and thence onto the cross wire and was observed in the eyepiece. The single vertical cross wire was mounted in a micrometer which had divisions reading to 0.001 inch.

The rotating mirror was driven by a small compressed air turbine mounted directly on the mirror spindle.


In the null method used, the light emerged from one face of the rotating mirror and was received on the adjacent face. As the mirror started rotating the image gradually passed from the field of view, later to reappear from the opposite side of the field as the mirror approached its proper speed. The rotating mirror was brought into synchronism with a tuning fork whose period of vibration had to be measured. The slight angle in which the return beam differed from 1/32 revolution was measured with the micrometer. The distance remained fixed. The time interval to be measured, therefore, was that during which the rotating mirror turned from one face to the next, plus or minus a small angle observed in the eyepiece.

The period of the fork was then determined by stroboscopic methods in terms of free-pendulum beats. As the rotating mirror accelerated, light from a 6-volt lamp was reflected from a small mirror on the tuning fork onto a polished face of the nut clamping the rotating mirror to its spindle. As the mirror continued to accelerate, the image from the tuning fork passed through a series of vibrating and stationary states to a final stationary stale for which the beats heard between the fork and the rotating mirror ceased. At this point a second observer made a setting on the return image formed by the light traversing the tube and read off the micrometer. A reversal of the direction of rotation of the mirror eliminated any necessity for making zero readings.

In checking the tuning fork with the pendulum, light from a small lamp was focused on a narrow slit and passed into the pendulum case, whence it was reflected by a small mirror on the pendulum and focused on the edge of the tuning fork. When the fork vibrated, flashes of light from the mirror on the pendulum illuminated the fork in various positions, thus producing a series

The standard-frequency assembly used in the Michelson experiment was similar to the one shown above, except that one of the multivibrators was replaced by an amplifier for the 0.1-second pulses from the synero-clock

of saw-tooth images. When the fork period was an exact multiple of the pendulum, the images appeared stationary. When the periods differed, the teeth appeared to travel across the field of view. The period of the fork in terms of the free pendulum could he determined from the number of flashes shown in traveling from one tooth to the next.


The determination of the period of the free pendulum in terms of true time was done in two steps. First a comparison was made between the beats of the pendulum and a flash box, operated by the constant-frequency oscillator or by the ship's chronometer. The second comparison was made between the chronograph records of second-marks from the chronometer and the time signals from Arlington. Light from the flash box was reflected from a small fixed mirror inside the pendulum case and from a small mirror placed on the axis of the pendulum. These two reflections returned to a transparent scale in the flash box where the time of their coincidence could be observed.

Accuracy in determining the pendulum period depended upon the precise operation of the flash box shutter. The superior performance of the General Radio Standard-Frequency Assembly in operating the shutter was a very great advantage. — E. C. Nichols.

Was this article helpful?

0 0

Post a comment