The Novel Optical System of Measuring the Speed of Starlight
Jingshown Wu
1
, Yen-Ru Huang
1
, Shenq-Tsong Chang
2
, Hen-Wai Tsao
1
,
San-Liang Lee
3
and Wei-Cheng Lin
2
1
Department of Electrical Engineering, Graduate Institute of Photonics and Optoelectronics, and Graduate Institute of
Communication Engineering, National Taiwan University, Taipei 10617, Taiwan
2
Instrument Technology Research Center, National Applied Research Laboratories, Hsin-Chu 30076, Taiwan
3
Department of Electronic Engineering, National Taiwan University of Science and Technology, Taipei 10617, Taiwan
Keywords: Light Speed, Optics, Astrophysics.
Abstract: We proposed a novel method and implemented an optical system accordingly to measure the speed of
starlight by using the well-known speed of light from a terrestrial source, c, as a metric basis. This system
consists of a transmitter and a receiver. The transmitter modulated starlight, terrestrial red and infrared lights
into pulses simultaneously. These pulses were detected by the distant receiver. A high speed oscilloscope is
used to record the pulses arrival times, where the terrestrial infrared pulse and the red pulse are used as the
trigger and the reference signals. During the measurement, we employed a terrestrial white light travelling
along the exact path of the starlight to calibrate the system. We found that the starlight pulses arrived at the
receiver with various degrees of delays, compared with that of the terrestrial white light pulse. The values of
the delays are likely related to the relative radial velocities of the stars. The result implies that the measured
apparent speed of starlight is not constant.
1 INTRODUCTION
The speed of light is an important physical
paremeter which is used to estimate other physical
parameters such as mass, time, space, energy, etc. In
1676, Römer investigated the eclipses of Io,
Jupiter’s nearest moon, and estimated that the speed
of light was about 214,000 km/sec. In 1728, Bradley
observed aberration of Draconi. He gave a value for
the speed of light of 301,000 km/sec. Both
measurements used light from extraterrestrial
sources. In 1849, Fizeau used a chopper and a
distant mirror to measure the speed of light on the
terrestrial. He estimated the speed of light equal to
3.153×10
8
m/sec. In 1862, Foucault employed a
rotating mirror instead of a chopper. He obtained a
value of the speed of light about 298,000 km/sec
±500 km/sec. In 1878, Michelson constructed the
famous Michelson interferometer to measure the
effect of ether on the speed of light. He concluded
that the hypothesis of stationary ether was incorrect.
During 1880 and 1882, Michelson made many series
of measurements and obtained the value 299,853
km/sec ± 60 km/sec. In the latter half of the
nineteenth century, many measurements using the
velocity of electromagnetic radiation or the ratio of
electromagnetic to electrostatic units were
conducted. The results are very similar to the
previous ones. Currently, the recommended
measured value of the speed of light on the earth, c,
is equal to 299,792.5 km/sec.
In 1905, Einstein published his special theory of
relativity based on the following two postulates: 1.
The laws of physics are the same in all inertial
frames. 2. The speed of light in vacuum is constant
regardless of any reference frame.
Dickey et al. reported observation of the time taken
by the laser light to go to the Moon and back to the
earth over the last forty years and the result implied
a decrease of the speed of light. Anderson et al.
analyzed radio tracking data from Pioneer 10/11
spacecrafts and suggested that the speed of light
might is less than the common known value of c.
In 1908, Ritz assumed that the speed of light might
be influenced by the motion of the source,
= + ,
(1)
where c is the speed of light from a resting source,
i.e. the well-known value, u is the relative velocity
of the source and the detector, c’ is the apparent
speed of light from the moving source. In this paper,
39
Wu J., Huang Y., Chang S., Tsao H., Lee S. and Lin W..
The Novel Optical System of Measuring the Speed of Starlight.
DOI: 10.5220/0004874100390044
In Proceedings of 2nd International Conference on Photonics, Optics and Laser Technology (PHOTOPTICS-2014), pages 39-44
ISBN: 978-989-758-008-6
Copyright
c
2014 SCITEPRESS (Science and Technology Publications, Lda.)
we use the star as the light source and place the
detector on the earth, so the light source and the
detector have relative motion.
Table 1 shows the radial velocities, magnitudes,
spectrum bands, right ascensions, declinations, and
distances of Capella, Betelgeuse, Arcturus,
Adlebaran, and Vega from the sun. The data in
Table 1 come from different references. They may
have small variation.
Figure 1 is a sketch of a celestial sphere and
positions of the stars. The orbital speed of the earth,
V
e
, is about 30 km/sec which is on the ecliptic plane.
Let V
s
denote the radial velocity of the star from the
sun and 
be the projection of the radial line of the
star on the ecliptic plane. is the angle between V
e
and 
, is the angle between V
s
and 
. So the
projected orbital velocity of the earth on V
s
is
. The relative radial velocity of the
star and the earth V
r
is
.
Figure 1: The sketch of celestial sphere, positions of
Capella, Betelgeuse, Arcturus, Aldebaran, Vega (not on
scale), the radial velocities of Capella and the earth, and θ
and φ.
2 SYSTEM DESIGN PRINCIPLE
Some of physical parameters are dimensionless,
others are dimensional and their numerical values
depend entirely on the units in which they are
defined. The speed of light is dimensional and
expressed in terms of length per unit time. To
measure the speed of light, we need a rule and a
watch. Expecially to measure the speed of light
emitting from a moving source, we encounter the
simultaneity problem. Conventionally the detector
and the moving source are considered in two
different reference frames. In 1892, H. A. Lorentz
proposed the Lorentz Transformation as follows:
If the relative motion of the two reference frames
is along their x and x’ axes, the first frame with the
space and time units x’ and t’ moves to the right
with speed relative to the second frame with space
and time units x and t, then
=
()
1
(2)
=
(

)
1
(3)
=
(4)
=
(5)
=
(
+
)
1
(6)
=
(
+
)
1
(7)
=
(8)
=
(9)
where v is the relative speed.
In our design, we will avoid using the units
transformation and simultaneity problem. We
compare the appearent speed of starlight with the
well-known value, c. Therefore, we only use the
space and time units on the earth, in other words, we
have one rule and one clock. Also because the speed
of light is about 3 × 10
m/sec which is extremely
fast, the speed of an ordinary moving light source on
the earth is much less the speed of light. To
investigate the influence of the speed of a moving
light source on the light speed measurement is a
difficult task. However the universe provides the
experimental environment.
Our measurement system consists of a
transmitter and a distant receiver. At the transmitter,
we used a telescope to collect the light from Capella,
Betelgeuse, Arcturus, Adlebaran, and Vega. These
stars are bright and have large relative radial
velocity with respect to the earth around the Spring
Equinox. Then we modulated the starlight, terrestrial
635 nm red light, and 1550 nm infrared light into
pulses simultaneously. After travelling a distance, d,
these pulses arrived at the receiver. The red light
travelled along almost the same path of the starlight.
So we were able to use the red light and the 1550 nm
infrared pulses as the reference and trigger signals.
PHOTOPTICS2014-InternationalConferenceonPhotonics,OpticsandLaserTechnology
40
We used the terrestrial white light which travelled
along exactly the same path of the starlight.
If the speed of light is constant, the travelling
time of starlight pulses and the terrestrial white light
pulses from the transmitter to the receiver should be
the same as
= /
(10)
If Ritz’s assumption is valid and the speed of
starlight has deviated from the well-known value, c,
then the time taken for the starlight pulses travelling
from the transmitter to the receiver is given by
= /(
)
(11)
where V
r
is the relative velocity of the star and the
earth.
3 SYSTEM DESCRIPTION
Based on the design principle and concept described
in the previous section, we have two different
designs: one using a rotating two-facet mirror and a
slit to modulate the continuous light beam into
pulses and the other employing a chopper.
3.1 The Optical System using a
Rotating Two-facet Mirror and a
Slit
Figure 2 shows the schematics of the system layout.
At the transmitter, for the starlight path, we used the
one-meter telescope of the Lulin Observatory to
collect the starlight. One end of a five-meter fiber
with core diameter of 68 μm and the numerical
aperture about 0.0729 was placed at the focal point
of the telescope to guide the starlight. The other end
of the fiber was fixed at the focal point of the off-
axis parabolic mirror, P1, which made the ray
collimating. It was then reflected by the rotating
mirror, M1, and was incident upon the off-axis
parabolic mirror, P2. A 100 μm wide slit was located
at the focal point of P2 and another off-axis
parabolic mirror, P3. When the rotating mirror M1
spun, the light would scan across the slit to produce
light pulses. The pulses were reflected by P3 to the
planar mirrors, M2 and M3, where M3 was located
at the Tungpu Hostel about 2,147 m from the Lulin
Observatory. The collimating ray from M3 travelled
2,155 m back to the Lulin Observatory to reach the
30 cm off-axis parabolic mirror, P4. A photo-
multiplier tube (PMT) was placed at the focal point
of P4. Therefore, the total travel distance, d, was
4,302 m.
For the path of the reference red light, a laser
with a wavelength of 635 nm and a 4 μm single
mode fiber pigtail was used as the reference light
source. The end of the fiber was located at the focal
point of the lens, L1. The collimating ray from L1
was incident to the rotating mirror, M1, and then the
lens, L2, with a standard 62.5 μm multimode fiber
Figure 2: The schematics of the optical system using rotating mirror.
9um SMF
(3085 m)
TheNovelOpticalSystemofMeasuringtheSpeedofStarlight
41
connected a 9 μm single mode fiber located at the
focal point. The 62.5 μm fiber acted as a slit and
guided the pulses. The total length of this fiber link
was 63 meters, which separated the red light and
starlight pulses by about 300 nsec on the
oscilloscope screen when M1 spun at 17,929 rpm.
The other end of the single mode fiber of the link
was placed at the focal point of the lens, L3. Then
the collimating ray from L3 was combined with the
starlight by a beam combiner BS1. Thereafter the
red light and the starlight pulses travelling along the
same path were received by the PMT to convert
them into electrical pulses which could then be
recorded by the oscilloscope. For the trigger signal,
a 1550 nm laser with an Erbium doped fiber
amplifier and a 9 μm single mode fiber pigtail was
used as the trigger source. The end of the pigtail
fiber was placed at the focal point of the lens, T1.
The collimating ray from T1 was incident to M1 and
then the lens, T2. At the focal point of T2, we had a
standard 62.5 μm multimode fiber which acted as a
slit and guided the pulses through the 3085-meter
single mode fiber to the receiver. A photodetector
was employed to convert the infrared pulses to
electrical pulses which were used as the trigger
signal for the oscilloscope.
3.2 The Optical System Employing a
Chopper
When we use the rotating two-facet mirror and a slit
as a modulator, the reflected beam from the mirror
M1 scans over the slit to product the pulses while
the rotating mirror spun. The spindle which drives
the rotating mirror may have wobble and the angular
velocity deviation. Here we propose the optical
system using a chopper as the modulator. In this
system the light beam is fixed. When the chopper
spins the light beam is modulated into pulses. Figure
3 is the schematics of the optical system using a
chopper. The starlight is collected by the telescope
and guided by the 68 μm fiber whose output end is
placed at the focal point of P1. The reflected
collimating beam is incident to M1, M12 and P2.
The focal point of P2 and P3 is at the slit of the
chopper. The reflected collimating beam from P3
then is incident to the planner mirror M2 and
transmits over 4.3 km to P4. The PMT is located at
the focal point of P4. The 1550 infrared is
collimated by the lens L1 and then combined by the
beam splitter, BS1, with the starlight main path.
After the chopper, part of the 1550 nm light is
separated by the P90 beam splitter and through the
lens L2 and incident to the multimode fiber
connected to the 3.8 km single mode fiber delay line.
At the end of the single mode delay line, the
1550 nm optical/electrical converter converts the
Figure 3: The schematics of the optical system using chopper.
PHOTOPTICS2014-InternationalConferenceonPhotonics,OpticsandLaserTechnology
42
optical signal to the electrical signal. The rest part of
1550 nm travels along with the starlight and reaches
the receiver where the beam splitter BS2 and an
Avalanche Photon-Diode are used to separate the
1550 nm light and converts to the electrical signal.
When the chopper spins, the starlight and the
1550 nm infrared are modulated into pulses
simultaneously. The two 1550 nm pulses at the
receiver can be used as the trigger and the reference
signals. In this system, the chopper is a very
important element, which must be light weight and
easy to have dynamically balanced. We have used
titanium alloy and composite carbon fiber to
fabricate the chopper. However the preparation is
still on-going.
4 MEASUREMENT RESULTS
We used the optical system using the rotating mirror
and the slit to perform the measurement in
November 2009, March 2010, and January 2011. In
order to minimize the error due to deviation of the
spindle speed, we carefully aligned the parabolic
mirrors and the lenses P1, P2, T1, T2, L1 and L2,
such that the three beams from P1, T1, and L1, after
reflected from the rotating mirror M1, were
simultaneously incident to the slit after P2, the focal
points of T2 and L2, respectively.
Because the optical power of the starlight was
very low (in the order of a few nanowatts or less),
the PMT operated in the photon counting mode
which generated spikes instead of full waveforms.
The maximum amplitudes of dark current and
thermal noises were 0.024 volts and very small
compared to the spikes and trigger pulses. To avoid
accumulation of noises, we first choose a threshold
at 0.025 volts and set all recorded data smaller than
the threshold to zero. If there are spikes in the frame,
we classify it as a valid frame, otherwise we discard
it. To reconstruct the starlight and the red light pulse
waveforms, taking a moving average of length 100
on the trigger pulse of each valid frame, we calculate
the centroid (center of gravity) by the weighted time
average method, i.e.
centroid i i i
X XY Y=
∑∑
, where
X
i
is the sampling time and Y
i
is the amplitude
which is larger than 20% of the peak value. Then we
adjust the time axes of the frames by aligning the
centroids of the trigger pulses for 2,000 valid frames
and simultaneously accumulate spikes to form the
waveforms of the starlight and the red light pulses.
Because the PMT operated in the photon counting
mode, we only take the peak value of the spike
during the accumulation process. We apply Gaussian
fitting to the red light waveform to obtain the
centroid. After finishing this process for the entire
set of starlight measurement frames, we align the
centroids of the fitted red light pulses to reconstruct
the complete waveforms of the starlight and the red
light pulses. We follow the same procedure to
reconstruct the pulse waveforms of the terrestrial
white light and the red light.
Next, we again use Gaussian fitting to estimate
the centroids of the starlight, white light, and red
light pulses. We applied different fitting methods
and obtained a similar result.
In the spring of 2013, we used the similar setup
as shown in Figure 2 to measure the speed of
starlight, where we omitted the red light and used
the 1550 nm as the trigger and the reference signals.
We obtained the similar result as the previous ones.
Table 2 summarizes the average delays of the
starlight measured in 2010, 2011, 2012, and 2013.
Note that if the starlight pulses arrive at the receiver
earlier than the terrestrial white light pulses, the
delay value is negative, e.g. the Vega pulses.
The pulses of Adlebaran, Capella, and
Betelgeuse have positive delays and that of Vega
and Arcturus have negative delays. As shown in
Table 1, the relative radial velocities of Adlebaran,
Capella, and Betelgeuse are positive, but that of
Vega and Arcturus are negative. Because we had
laid the equipment on the floor of the Lulin
Observatory, the ambient temperature and the
relative humidity made the floor contraction or
expansion which affected the measurement. The
error of delays of Arcturus in 2011, Vega in 2011,
and Capella in 2012 are large.
5 CONCLUSIONS
In this paper, we presented a novel method to
measure the speeds of starlight. This method
compares the travelling times of these starlights and
the local white light from the transmitter to the
receiver. Such that physical unit transformation,
clock synchronization and definitions of dimension
units problems can be avoid. This system utilizes the
existing telescope of the observatory, the orbiting
speed of the earth, and the radial velocities of stars.
Comparing the measured apparent speeds of
Adlebaran, Capella, Betelgeuse, Arcturus, and Vega
with the well-known speed of light from a rest
source, c, we find that Adlebaran, Capella, and
Betelgeuse have positive delay, while Vega and
Arcturus have negative delay. Note that Adlebaran,
TheNovelOpticalSystemofMeasuringtheSpeedofStarlight
43
Table 1: The data of Capella, Betelgeuse, Arcturus, Adlebaran, and Vega.
Star Capella Betelgeuse Arcturus Vega Adlebaran
Apparent Magnitude 0.91/0.76 (B-V) 0.58 (V) -0.04 (V) 0.03 (V) 0.85 (V)
Stellar Classification G8 III / GI III M2 Iab K2 III A0 V K5 III
Right Ascension 5h16m41s 5h55m10s 14h15m39s 18h36m56s 4h35m55s
Declination +45˚59’52” +7˚24’25 +19˚10’56” +38˚47’01” +16˚30’33”
Radial Velocity V
s
29.65 21.91 -5.19 -13.9 54.26
Distance ( light year / parse)
42.5 ± 0.5 /
13.04 ± 0.03
643 ± 146 /
197 ± 45
36.7 ± 0.3 /
11.24 ± 0.09
25.3 ± 0.1 /
7.76 ± 0.03
65 ± 0.1 /
20.0 ± 0.4
Relative Radial Velocity ( km/sec )
( March 18, 2010 )
56.47 49.56 -20.77 -27.10 82.43
Table 2: The average delays of starlight measured in 2010, 2011, 2012 and 2013.
Delays (ns)
2010
2011
2012
2013
Adlebaran
2.40
Capella
2.28
2.41
4.57
1.20
Betelgeuse
1.19
1.80
Arcturus
-0.45
-4.01
-0.19
Vega
-1.35
-7.40
-1.20
Capella, and Betelgeuse have positive relative radial
velocity, Vega and Arcturus have negative relative
radial velocity, i.e. Adlebaran, Capella, and
Betelgeuse are leaving away from the earth and
Vega and Arcturus are approaching to the earth. The
result implies the measured apparent speed of
starlight likely relates to the relative motion of the
source and the detector.
ACKNOWLEDGEMENTS
The authors are grateful for Professor Wen-Ping
Chen, Director Hung-Chin Lin, Miss Hui-Ting Tsao
and the staff of the Lulin Observatory of National
Central University to provide the facilities and the
necessary help. This work was supported in part by
Excellent Research Projects of National Taiwan
University and the Nation Science Council, Taiwan,
under Grants 98R0062-06, NSC 100-2221-E-002-
035- and NSC 101-2221-E-002-002-.
REFERENCES
Froome, K. D., and Essen, L., 1969. The Velocity of Light
of Radio Waves, Academic Press, New York, USA.
Sanders, J. H., 1965. The Velocity of Light, Pergamon
Press Ltd., Oxford, UK.
French, A. P., 1968. Special Relativity, W. W. Norton &
Company Inc., New York, USA.
Dickey, J. O., Bender, P. L., Faller, J. E., Newhall, X. X.,
Ricklefs, R. L., Ries, J. G., Shelus, P. J., Veillet, C.,
Whipple, A. L., Wiant, J. R., Williams, J. G., and
Yoder, C. F., 1994. Science 265, 482.
Anderson, J. D., Laing, P. A., Lau, E. L., Liu, A. S., Nieto,
M. M., and Turyshev, S. G., 1998. Phys. Rev. Lett. 81,
2858.
Ritz, W., 1908. Amn.de Chim, et de Phys. 13, 145-275.
Torres, G., Claret, A., and Young, P. A., 2009. Astrophys.
J. 700, 1349.
Wilson, R. E., 1963. General Catalogue of Stellar Radial
Velocities, Carnegie Institution of Washington
Publication, Washington, D. C., USA.
SIMBAD Astronomical Database, from http://simbad.u-
strasbg.fr/simbad/.
Wu, J., Chang, S. -T., Tsao, H. -W., Huang, Y. –R., Lee, S.
-L., et al., 2012. Testing the Constancy of the velocity
of Light. Am. Phy. Soc. April Meeting. Atlanta, USA.
Huang, Y. -R., Wu, J., Lin, W. -C., Chang, S. -T., Lee, S. -
L., Tsao, H. -W., et. al., 2012. An Optical System of
Comparing the Speeds of Lights from Moving Stars.
ODF 12. Saint-Petersburg, Russia.
Wu, J., Chang, S. -T., Tsao, H. -W., Huang, Y. -R., Lee, S.
-L., et. al., 2013. A Method of Comparing the Speed
of Starlight and the Speed of Light from a Terrestrial
Source. SPIE Optics + Photonics 2013. San Diego,
USA.
Huang, Y. -R., Wu, J., Chang, S. -T., Tsao, H. -W., Lee, S.
-L., and Lin, W. -C., 2013. Measurement of
Differences and Relativity between Speeds of Light
from Various Stars. APPC 12. Chiba, Japan.
PHOTOPTICS2014-InternationalConferenceonPhotonics,OpticsandLaserTechnology
44