Download Open Frame CD/DVD Reader

P5.8.8.7
Open Frame CD/DVD Reader
4747124 EN
Table of Contents
1.0 Introduction
3
2.0 Fundamentals
4
2.1
Focusing of Laser Light
4
2.2
Diode Laser and Photodiode
7
2.3 The energy band model
2.3.1 Fermi distribution
7
8
2.4
9
Light emitting diode (LED)
2.5 Semiconductor Laser
2.5.1 Resonator and beam guidance
2.5.2 Divergence and intensity distribution
2.5.3 Polarisation
2.5.4 Spectral properties
10
11
11
11
11
2.6 Photodiodes
2.6.1 Ge and Si PIN photodiodes
12
13
2.7 Compact Disk (CD)
2.7.1 Disc Format
2.7.2 Optical Detection of Pits and Land
2.7.3 Pick - up System
2.7.4
Diffraction Grating 2.7.5 Beam Splitter
2.7.6 Quarter Wave Plate
2.7.7 Focusing Error Detection
2.7.8 Tracking error detection
14
14
16
17
17
18
18
19
20
3.0 Experimental Set-up
20
3.1
Operation
21
3.2
Measurements
21
4.0 Further Readings
22
Introduction: 
1.0 Introduction
Compact Disks revolutionised the audio world by using
truly digital stored signal reading and playback. This
technology has become possible due to the invention of
the Laser in combination with highly integrated electronics circuits. However, it should also be mentioned, that the
fine mechanics engineering plays an important role in the
success of this technology. The CD player therefore is an
example of the combination of three disciplines namely the
optics, electronics and fine mechanics. The first device for
audio recording and playback has been invented by Edison
1877. His phonograph consisted of a piece
of tin foil wrapped around a rotating cylinder. The vibration of spoken words are led
through the recording horn to a diaphragm
with attached stylus, which cuts grooves
into the tin foil. "Mary Had a little Lamb"
spoken by Edison are reported to be the
first sound record ever made.
Ten years later Emile Berliner registered his invention
of the gramophone. Instead of using a rotating cylinder,
Berliner used a rotating
disk where the vibration of
the sound was cut lateral
into a groove with constant
depth. Around 1925 electric
recording started, but the
mechanical cutting of the
grooves and the use of disks
remained. In the early 1960’s the stereo technology has
been introduced and the demand for high fidelity playback
audio systems was still growing. In the 1970’s the companies reached the limitation of analogue audio recording
without increasing dramatically the prices for home equipment. This was because quality, dynamic range and distortion are determined by the medium used. To overcome
this barrier digital recording and playback systems has
been developed which however could only be used with
tape recorder and did not satisfy the huge market for music
discs. Parallel to the development of optical video discs
which began in 1970’s the engineers began to think that
the needed bandwidth for a video disc was much higher
than for digitized audio signals. The first optical Digital
Audio Disc (DAD) has been presented in September 1977.
It took five years for the final common
specification and development of the optical Compact Disc which was launched in
October 1982 to the consumer market. The
disc size was decided upon 12 cm in order
to achieve a capacity of 74 minutes, this was
the approximate duration of Beethoven’s Ninth Symphony.
Surprisingly the initial idea of the optical video disc never
reached the success as its spin off the audio CD.
Besides the tremendous growing audio market another
field of interest came up - the Personal Computer development. Bill Gates, the owner of Microsoft stated in 1985
his vision “To put a microcomputer on every desk in
every home ...” This strategy required the introduction of
windows based easy to use software. However Microsoft
and others realised, that magnetic based storage media
neither the Floppy nor the Harddrive disk could provide
Dr. Walter Luhs - Jan. 1999, revised 2003, 2009, 2010
the needed larger storage and delivery, they started to
search for more efficient and affordable alternative media. Consequently they looked to the digital audio CD and
initiated the development of the Compact Disc Read-Only
Memory (CD-ROM). Nowadays it is standard, that every
Personal Computer is equipped with a CD-ROM reader to
install software or to retrieve information. The capacity
of the CD-ROM reaches 650 Mega Byte, enough to store
260.000 pages of Text. However, it seems that the future
requires more space, due to the same tendency as experienced with the demand for higher quality of audio systems,
the user wants more realistic movies also available for the
Personal Computer. Consequently a new generation of
CD’s is coming up, the Digital Versatile Disc (DVD) with
a theoretical capacity of up to 17 Giga Byte. Nevertheless
the optical technique to read the data will remain regardless whether an audio or data disc is used. The information
is stored like a Morse code as long and short structures on
the disk. However the structure is so small, that it cannot
be made visible by means of optical instruments since it
has a size of the wavelength of visible light. The (Fig. 1)
shows such a structure recorded with an electron microscope.
Fig. 1: Electron microscope image of the surface of a CD
Fig. 2: Data track of the CD
The arrangement of the data pits and lands is similar to the
spiral groove of the earlier audio disk cartridges. However
the distance of the “grooves” of a CD with 1.6 µm is much
more smaller than that of a disk record. Consequently the
“stylus” of a CD is formed by Laser light which can be focused down to appr. the half of its wavelength. Considering
a typical wavelength of 780 nm, a spot size of 0.4 µm can
be obtained. To detect the pits an optical pick-up system is
used as illustrated in (Fig. 3).
Page 3
Fundamentals: FOCUSING OF LASER LIGHT
LD
PBS
L2
L1
FP
A
QWP
CD
to this phenomena.
After the description of the fundamentals we will turn
back to the detailed explanation of state of the art pick-up
systems including data trace finding and error correction
loops in chapter (2.7.7) and (2.7.8).
2.0Fundamentals
2.1Focusing of Laser Light
B
PD
Fig. 3: Optical pick-up system, (A) shows the ray trace without a CD and (B) with a CD or any other reflecting surface
The light of the Laserdiode (LD) passes a polarising beam
splitter cube (PBS). The orientation of the polarisation
of the laser light is chosen in such a way, that it passes
completely the beam splitter. The lens L1 collimates the
radiation which subsequently passes a quarter wave plate
(QWP). The plate changes the polarisation state of the laser
light from linear to circular. Finally the focusing lens L2
generates a beam waist located in the focal plane FP. In
case the CD or any other reflecting surface is placed into
the focal plane (case B), the incident light will be reflected.
If the incident beam has been clockwise circular polarised,
then the reflected one will be counter clockwise circular
polarised. For the reflected beam the lens L2 now acts as
a collimator. After passing the quarter wave plate, the polarisation of the reflected beam will be converted back to
linear polarisation.
However, its linear polarisation state is turned by 90° with
respect to the incident beam. Consequently the reflected
light will be reflected from the polarising beam splitter
and diverted to the photodetector PD. Changes in the intensity of the light due to reading the data structure of the
CD are recorded by the photodiode of the detector. Before
we study the pick-up in detail, the optical components and
applied physical phenomena shall be discussed. As we see
from (Fig. 3) the pick-up system consist of a coherent light
source, the Laser diode LD, which light has to be focused
down to the smallest possible spot.
In the next chapter we will learn that a beam of laser light
cannot be treated with the methods of geometrical optics. Instead we have to treat the emission of a Laser as
Gaussian beams. After that we will discuss the properties
of the Semiconductor or Diode Laser. Due to the similarity
of generating and detecting of light by semiconductors the
subsequent chapter explains the behaviour and properties
of photo diodes.
For the separation of the incident and reflected beam an arrangement of a polarising beam splitter and quarter wave
plate is used. This arrangement is also termed as optical
diode and is used very often in optical sensor systems. In
chapter (2.7.6) the components used will be discussed in
detail.
The detection of the data structure of a CD is based on reflection and interference, thus the chapter (2.7.2) is related
Dr. Walter Luhs - Jan. 1999, revised 2003, 2009, 2010
We are facing the problem to focus a Laser beam to the
smallest possible diameter to achieve the highest data density of CD’s. The beam of the Laserdiode has to be focused
to a diameter of the order of magnitude of its own wavelength. Under these circumstances the laws of geometrical
optics fail because they anticipate parallel light beams or
plane light waves which in reality exist only in approximation.
The existence of real parallel light with plane wave fronts
allows an infinitesimal focused spot size. That means furthermore the photons have to pass a channel with zero
volume and can be localised exactly. But this would violate
the fundamental Heisenberg’s uncertainty principle.
f
Fig. 4: Focusing a beam in geometrical optics
Therefore real parallel light beams do not exist in reality
and plane wave fronts exist only at a particular point.
The reason for the failure of geometrical optics is the
fact that it has been introduced at a time where the wave
character of light was still as unknown as the possibility to describe its behaviour by Maxwell‘s equations. To
describe the propagation of light we use the general wave
equation as a result of Maxwell’s equation:

∆E

n2 ∂2 E
⋅
c 2 ∂t 2
0
When we consider the technically most important case of
spherical waves propagating in the direction of z within a
small solid angle as the Laser actually does, we arrive at
the following statement for the electrical field:
 
E = E ( r, z )
with
r2
x2 + y2 + z2
In this case the solution of the wave equation provides
fields which have a Gaussian intensity distribution over
the cross-section. Therefore they are called Gaussian
beams. Such beams, especially the Gaussian fundamental mode (TEM00) are generated with preference by lasers.
But the light of any light source can be considered as the
superposition of many such Gaussian modes.
Still, the intensity of a particular mode is small with re-
Page 4
Fundamentals: FOCUSING OF LASER LIGHT
2W0
2W
R
Z
Wavefront radius of curvature
Beam diameter
spect to the total intensity of the light source. The situation
is different for the laser. Here the total light power can be
concentrated in the fundamental mode.
This is the most outstanding difference with respect to
ordinary light sources next to the monochromasy of laser
radiation. Gaussian beams behave differently from geometrical beams.
Z
0
Z
Distance z
r
Fig. 6: Radius of curvature of the wave front as a function
of the distance from the waist at z=0
0
Z
Fig. 5: Beam diameter of a Gaussian beam as fundamental
mode TEM00 and function of z.
A Gaussian beam always has a waist and its radius w
results out of the wave equation as follows:
 z
w (z ) = w0 ⋅ 1 +  
 zR 
2
2.1
w0 is the smallest beam radius at the waist and z r is the
Rayleigh length. In (Fig. 5) the course of the beam diameter as a function of z is represented. The beam propagates
within the direction of z. At the position z = z0 the beam
has the smallest radius. The beam radius increases from
here on linearly with increasing distance. Since Gaussian
beams are spherical waves we can attribute a radius of
curvature of the wave field to each point z. The radius of
curvature R can be calculated using the following relation:
R (z)
In (Fig. 7) the Rayleigh range has been marked as well
as the divergence θ in the far field, that means for z>>z0.
The graphical representation does not well inform about
the extremely small divergence of laser beams another
outstanding property of lasers.
z2
z+ r
z
Beam Diameter
Z
θ
z0
Distance z
Rayleigh Range
+zR
-zR
Fig. 7: Rayleigh range ZR and divergence θ for the far field
z>>zR
The reason for this is that the ratio of the beam diameter
with respect to z has not been normalised. Let‘s consider,
for example, a HeNe-Laser (632 nm) with a beam radius
of wo=1 mm at the exit of the laser. For the Rayleigh range
Zr we get:
This context is reflected by (Fig. 6). At z = zr the radius of
curvature has a minimum . Then R increases with 1/z if z
tends to z = 0 . For z=0 the radius of curvature is infinite.
π
3.14
Here the wave front is plane. Above the Rayleigh length
2 ⋅ z R 2 w20
2 ⋅10 6
9, 9 m
zr the radius of curvature increases linearly. This is a very
623 ⋅10 9
λ
essential statement. Due to this statement there exists a That means that within a range of nearly 10 m the beam
parallel beam only in one point of the light wave, to be can be considered as parallel. In the next example we want
precise only in its focus. But within the range
to analyse the situation where a Gaussian beam is focused
by a lens.
zr ≤ z ≤ zr
f
f
2Wo
θ
a beam can be considered as parallel or collimated in good
approximation.
2W
z
y
Fig. 8: Design of the focusing optic
The radius of the waist of the focused beam is:
Dr. Walter Luhs - Jan. 1999, revised 2003, 2009, 2010
Page 5
Fundamentals: FOCUSING OF LASER LIGHT
w0 ⋅ f ⋅ θ ⋅
w
2.2
w20 + θ 2 ⋅ z 2
The position of the waist is:
y
are used, which differ significantly in their characteristic
properties from “classical” Lasers like the HeNe-Laser,
the next chapter is addressed to the fundamentals of semiconductor laser.
z⋅f2
2
 w0 
2

z +  
 θ 
Example: The beam of a HeNe laser of 0.5 mm diameter
and of 1.5 mrad divergence is to focus by means of a lens
with a focal of 50 mm is 2 m apart from the beam waist of
the laser. We find for the beam waist:
w=
0, 5 ⋅10−3 ⋅ 0, 05 ⋅1, 5 ⋅10−3
−8
2
−6
6, 25 ⋅10 + 2, 25 ⋅10 ⋅ (2 0, 05)
= 12, 8µm
and for the position of the beam waist with respect to the
focal distance:
y
1, 95 ⋅ 0, 052
2
 0, 25 
2

1, 95 + 
 1, 5 
1, 27 mm
For this example the position y of the waist is 1.27 mm
apart from the focal point and the radius of the waist is
here 12.8 µm. From 2.1 for z >> zR we see that the divergence for the far field can be expressed as:
w(z)
w0
⋅ z ⇒ tan θ
zR
w0
zR
λ
π ⋅ w0
If we consider the case that the initial beam waist w0 is
large compared to θz of 2.2 then the equation simplifies to:
w
f ⋅θ
f⋅
λ
π ⋅ w0
For reading the data structure of the CD a focused beam
diameter of 1.7 µm or radius w=0.85 µm is required. The
wavelength of the diodelaser commonly used is 780 nm.
Two of the variables, w and λ of the above equation are
now defined. Let us start the calculation for the required
focal length f with the assumption that the initial beam diameter wo is 1 mm. For this case we obtain for the required
focal length f:
π
π
f = w ⋅ w0 ⋅ = 0.85 ⋅ 10 −6 ⋅ 0.5 ⋅ 10 −3 ⋅
≈ 1.7 mm
λ
780 ⋅ 10 −9
This result of 1.7 mm focal length is impractical since such
a lens is commonly a ball lens where the location of the
focal plane is near the surface of the lens. This is a problem because the thickness of the CD itself is 1.2 mm and
the beam has to travel this distance inside the transparent
layer of the CD. So a better choice is to increase the initial
beam diameter which can be achieved by selecting an probate value for the focal length of the lens L2 (Fig. 3).
With the knowledge of the preceding chapter the optical
design of the single beam pick-up system can be understood. However, the properties of the laser beam plays an
important role concerning its spatial intensity distribution and divergence. Because in CD - reader diode laser
Dr. Walter Luhs - Jan. 1999, revised 2003, 2009, 2010
Page 6
Fundamentals: DIODE LASER AND PHOTODIODE
Within this project we are going to use laser diodes (LD) as
well as photodiodes. Both components exploit the properties of semi conducting solid state material to convert electrical current directly into light and vice versa. Nowadays
nearly all electronically devices as transistors or integrated circuits are basing on this material and it is therefore
useful to dedicate a chapter to their fundamentals. We
want to know especially how light is generated and how
it is absorbed to understand the concept of the LD and the
photodetector.
We start with two separate atoms and describe the change
of their energy states when they are approaching each other. In a next step the number of atoms is increased and the
appearance of energy bands will be discussed. The choice
and combination of different semi conducting materials
finally decides upon whether a LED, photodetector or a
Laser diode can be realised.
2.3The energy band model
Atoms or molecules at a sufficient large distance to their
neighbours do not notice mutually their existence. They
can be considered as independent particles. Their energy
levels are not influenced by the neighbouring particles.
The behaviour will be different when the atoms are approached as it is the case within a solid body. Depending
on the type of atoms and their mutual interaction the energy states of the electrons can change in a way that they
even can abandon „their“ nucleus and move nearly freely
within the atomic structure. They are not completely free,
otherwise they could leave the atomic structure.
additional forces which are responsible for this binding. To
understand these forces we must call on quantum mechanics for help. Here we can only summarise as follows:
If atoms are mutually approached the states of the undisturbed energy levels split into energetically different states.
The number of newly created energy states are corresponding to the number of exchangeable electrons (Fig. 9).
One of the curves shows a minimum for a particular distance of the atoms. No doubt, without being forced the atoms will approach till they have acquired the minimum
of potential energy. This is also the reason for hydrogen
to occur always as molecular hydrogen H2 under normal
conditions. The second curve does not have such a distinct property. The curves distinguish in so far as for the
binding case the spins of the electrons are anti parallel. For
the nonbinding case they are parallel. It is easy to imagine that an increase of the number of atoms also increases
the number of exchangeable electrons and in consequence
also the number of newly generated energy levels. Finally
the number of energy levels is so high and so dense that we
can speak about an energy band.
2p
energy (E)
2.2Diode Laser and Photodiode
2s
1s
r0
displacement of nuclei (r)
Energy
Fig. 10: Band formation by several electrons. The most outside electrons are responsible for the equilibrium distance r0.
Distance of nuclei
Fig. 9: Potential energy due to interaction of two hydrogen
atoms
How the „free“ electrons behave and how they are organised will be the subject of the following considerations.
From the fundamentals of electrostatics we know that unequal charges attract. Therefore it is easy to imagine that
an atomic structure is formed by electrostatic forces. In
the following we will call it „crystal“. However, this model
will fail latest when we try to justify the existence of solid
Argon just by freezing it sufficiently. Since there is obviously some sort of binding within the crystal structure in
spite of the fact that inert gases are neutral there must be
Dr. Walter Luhs - Jan. 1999, revised 2003, 2009, 2010
Here it is interesting to compare the action of the electrons
with the behaviour of ambassadors:
The electrons in the most outside shell of an atom will
learn first about the approach of an unknown atom. The
“eigenfunctions” will overlap in a senseful way. One electron will leave the nucleus tentatively to enter an orbit of
the approaching atom. It may execute a few rotations and
then return to its original nucleus. If everything is OK and
the spins of the other electrons have adapted appropriate
orientation new visits are performed. Due to the visits of
these „curious“ electrons the nuclei can continue their approach. This procedure goes on till the nuclei have reached
their minimum of acceptable distance. Meanwhile it can
no more be distinguished which electron was part of which
nucleus. If there is a great number of nuclei which have approached in this way there will also be a great number of
electrons which are weakly bound to the nuclei. Still, there
is one hard and fast rule for the electrons: I will only share
my energy level by one electron with opposite spin (Pauli
principle). Serious physicists may now warn to assume
that there may be eventually male and female electrons.
But who knows.....
Page 7
Fundamentals: THE ENERGY BAND MODEL
Let‘s return to incorruptible physics.
Up to now we presumed that the atom only has one electron. With regard to the semiconductors to be discussed
later this will not be the case. Discussing the properties
of solid bodies it is sufficient to consider the valence electrons that means the most outside located electrons only as
it has been done for separated atoms. The inner electrons
bound closely to the nucleus participate with a rather small
probability in the exchange processes. Analogously to the
valence electrons of the atoms there is the valence band in
solid bodies. Its population by electrons defines essentially
the properties of the solid body. If the valence band is not
completely occupied it will be responsible for the conductivity of electrons. A valence band not completely occupied is called conduction band. If it is completely occupied
the next not completely occupied band will be called conduction band.
In the following section the question is to be answered how
the density of states of a band of electrons looks like and
on which quantities it depends. After all this information
will explain the emission or absorption of light photons
due to electron transitions inside these bands.
2.3.1Fermi distribution
In (Fig. 10) we have shown that the energy bands are the
result of the mutual interaction of the atoms. Each band
has a particular width ΔE, the magnitude of which is determined by the exchange energy and not by the number
N of the interacting atoms. Furthermore we know that the
number of energy levels within a band is determined by
the number of interacting electrons. The Pauli principle
states that such a level can only be occupied by two electrons. In this case the spins of the electrons are anti parallel. Within a band the electrons are free to move and they
have a kinetic energy. The maximum energy Emax of an
electron within a band can not pass the value ΔE since
otherwise the electron would leave the band and no longer
be a part of it.
The famous physicist Enrice Fermi calculated the theoretical energy distribution and how many electrons of energy
E ≤ Emax exist in the energy interval dE.
dn (E )
4π
3/2
2m ) ⋅
3 (
h
E
e
E −E Fermi
kT
⋅ dE
17.
+1
Emax
Emax
EFermi
T=0
EFermi
T>0
Fig. 12: Distribution of free electrons over the energy states
within a conduction band
Up to this point we anticipated that the temperature of
the solid body would be 0° K. For temperatures deviating
from this temperature we still have to respect thermodynamic aspects namely additional energy because of heat
introduced from outside.
Fermi and Dirac described this situation using statistical
methods. The electrons were treated as particles of a gas:
equal and indistinguishable. Furthermore it was presumed
that the particles obey the exclusiveness principle which
means that any two particles can not be in the same dynamic state and that the wave function of the whole system
is anti symmetrical.
Particles which satisfy these requirements are also called
fermions. Correspondingly all particles which have a spin
of 1/2 are fermions and obey the Fermi Dirac statistics.
Electrons are such particles.
Under respect of these assumptions both physicists got the
following equation for the particle density dn(E) of the
electrons within an energy interval dE:
dn (E ) 4π
E
3/2
= 3 (2m ) ⋅ E−EFermi
⋅ dE
dE
h
kT
+1
e
The above equation is illustrated by (Fig. 13).
As shown in (Fig. 12) by introduction of thermal energy
the „highest“ electrons can populate the states which are
above them.
Based on these facts we are well equipped to understand
the behaviour of solid bodies. We are going to concentrate
now our special interest on the semiconductors which will
be presented in the next chapter with the help of the previously performed considerations.
dn/dE
dE
=
The situation in (Fig. 11) shows that the band has not been
completely filled up since the Fermi energy is smaller than
the maximal possible energy. This means that this band
is a conduction band. If the Fermi energy would be equal
to the maximal energy we would have a valence band. A
transfer of this knowledge to the energy level scheme of
(Fig. 10) and a selection of the 2s band would provide the
picture of (Fig. 12)
Energy
EFermie Emax
Fig. 11: Number of electrons per unit volume V and energy
interval dE as a function of the energy E
Dr. Walter Luhs - Jan. 1999, revised 2003, 2009, 2010
Page 8
dn/dE
Fundamentals: LIGHT EMITTING DIODE (LED)
Energy
E Fermi
E max
Fig. 13: Number of electrons per unit volume and energy
interval dE as a function of the energy E, but for a temperature T > 0.
2.4Light emitting diode (LED)
If we succeed to populate the conduction band with electrons and to have a valence band which is not completely
occupied by electrons (Fig. 14) they may pass from the
conduction band to the valence band. That way a photon is
generated. By absorption of a photon the inverse process
is also possible.
spectral distribution
of electrons
Energy
population density
of electrons
0
transition
created photon
population density
of holes
spectral
distribution
of holes
0
population density (dn/dE)
Fig. 14: Densities of states and spectral distributions
If one succeeds to establish a population inversion as
shown in the figure above and care has been taken that a
once created photon will not absorbed again such a band
structure can be used as semiconductor laser. History has
shown that the LED was the forerunner for diode laser.
The understanding of the photon emission and absorption
process for a particular band structure disclosed the way
for the production of laser diodes
Attention must be drawn to the fact that, until now, we
only discussed a semiconductor consisting of one type of
atoms. Consequently the situation shown in (Fig. 14) is, at
least for this type of direct semiconductor, only fictitious.
Light emission can only be created for very short intervals
of time and can therefore not be taken into consideration
for the realisation of a LED. By doping the basic semiconductor material we can create band structures with
different properties. A very simple example may be the
semiconductor diode where the basic material, germanium or silicon, is converted into p or n conducting material
Dr. Walter Luhs - Jan. 1999, revised 2003, 2009, 2010
Al2O3 contact
P
Zn doped
N
GaAs0.6 P0.4
x=0.4
N
GaAs1-xPx buffer layer
x=0
N
GaAs substrate
Fig. 15: Structure of a red emitting GaAsP-LED
CONDUCTION
BAND
VALENCE BAND
using suitable donators and acceptors. By the connection
of the doted materials a barrier (also called active zone)
is formed. It will be responsible for the properties of the
element. Silicon is mainly used for highly integrated electronic circuits while ZnS is chosen as fluorescent semiconductor for TV screens. As light emitting diodes and laser
diodes so called mixed semiconductors like AlGaAs are
in use. Mixed semiconductors can be obtained whenever
within the semiconductors of valence three or five individual atoms are replaced by others of the same group of
the periodical system. The most important mixed semiconductor is aluminium gallium arsenide (AlGaAs), where
a portion of the gallium atoms has been replaced by aluminium atoms.
This type of semiconductor can only be produced by a fall
out as thin crystal layer, the so called epitaxy layer, on host
crystals. To perform this stress free it is important that the
lattice structure of the host crystal coincides fairly well
with the lattices structure of AlGaAs (lattice matching).
This is the case for GaAs substrate crystals of any concentration regarding the Al and Ga atoms within the epitaxy
layer. In that way the combination of AlGaAs epitaxy layers and GaAs substrates offers an ideal possibility to influence the position of the band edges and the properties
of the transitions by variation of the portions of Ga or Al.
The (Fig. 15) shows as an example the structure of a red
emitting GaAsP (Gallium Arsenide Phosphide) LED. An
extremely pure GaAs single crystal is used as substrate. To
achieve a high electron photon conversion efficiency the
active pn layer must have a uniform undisturbed lattice
geometry. For best lattice matching a buffer layer with a
concentration gradient is deposited in such a way on top
of the GaAs substrate that a smooth transition to the subsequent active layer is achieved. On top of the n doped
active layer a p doped layer is placed with Zn as acceptor.
Finally an Al2O3 layer is attached for the later bonding of a
gold wire to be contacted to the leads of the LED housing.
The light is generated inside the pn junction and emits in
all directions even into the structure itself. To improve the
amount of usable light state of the art LED’s are equipped
with an additional reflector.
As already mentioned the emission wavelength depends
on the mixtures of different materials. These are mainly
elements of the III and V valency group. The table below
gives a brief overview which material is used for the generation of different emission wavelengths of LED’s.
Page 9
Fundamentals: SEMICONDUCTOR LASER
Colour
Red
Orange
Amber
Yellow
Green
Blue green
Turquoise
Blue
chip material
AlInGaP
AlInGaP
AlInGaP
GaP
GaP
GaN
GaN
GaN
λ (nm)
635
609
592
570
565
520
495
465
Iv (mCd)
900
1,300
1,300
160
140
1,200
2,000
325
Remarkable for the time being is the existence of blue emitting LED’s. As already mentioned that LED’s has been in
almost every case the forerunner for the laser diode, so
also here. The first blue emitting laser diodes are available now and it can be seen that in near future they will
substitute the nowadays used laser diodes for CD-ROM
reader which are emitting around 780 nm. As we also have
learned in chapter (2.1) the smallest possible beam waist of
a laser beam depends besides other parameter mainly its
wavelength. The shorter the wavelength, the smaller the
size of the data structure of the CD-ROM can be, thus resulting in an increase of the storage capacity. Smaller pits
of the CD-ROM are also reducing the lateral distance of
the data tracks increasing even more the storage capacity.
Considering the wavelength ratio of 450 nm / 780 nm the
storage capacity could be increased by a factor of 3.
At this point we see, that the diode laser used plays a key
role for the CD-ROM technology and which is why we
focus in the next chapter on the history and properties of
this key component, the diode laser.
Active Zone
p - GaAs
n - GaAs
Injection Current
Fig. 16: Simple laser diode around 1962, working at 70 K
and with 100 kA/cm2 in the pulse mode.
In the course of the following years the threshold could
be lowered to 60 kA/cm2 by improving the crystals but
only the use of a hetero - transition (Bell Labs. and RCA Labs.) brought the „break-through“ in 1968. The threshold
could be lowered to 8 kA/cm2 and working in the pulse
mode at room temperature was possible (Fig. 17).
Active Zone
p - GaAlAs
n - GaAs
Injection Current
Fig. 17: Simple hetero - structured laser around 1968,
working at 8 kA/cm2 in pulse mode at room temperature.
In this concept a layer of p conducting GaAlAs is brought
on the p layer of the pn transition of GaAs. The slightly
higher
band gap of GaAlAs compared to GaAs ensures
As simple as it may seem, it took about 20 years until peothat
a
potential
barrier is created between both materials in
ple had acquired the necessary technology of coating una
way
that
charge
carriers accumulate here and the formader extremely pure conditions. It all began in 1962 with
tion
of
inversion
is
increased respectively the laser threshthe first laser diode, just two years after Maiman had demold
is
remarkably
lowered
to 8 kA/cm2 .
onstrated the first functional ruby laser. In the course of
1962 three different groups reported more or less simultaneously the realisation of GaAs diode lasers.
2.5Semiconductor Laser
1. 2. 3. R. N. Hall M. I. Nathan T. M. Quist General Electric
IBM
MIT
The first laser was basically made of highly doted GaAs
(Fig. 2.20). A threshold current of 100 kA/cm2 was needed
since the GaAs material of those days was not by far as
good as it is today regarding the losses within the crystal.
Because of thermal conditions the laser could only work at
70 0K and in the pulsed mode.
created
Photon
Transition
n - AlGaAs
n - GaAs
p - AlGaAs
Fig. 18: Energy band diagram of a N n P - double hetero
structure.
The next step in development was the attachment of a similar layer on the n-side of the crystal. That way the threshold could be lowered once again in 1970. Now it decreased
to about 1 kA/cm2. Until today nearly all commercially
sold laser diodes are built up on the double hetero structure principle (Fig. 19) and (Fig. 20).
Dr. Walter Luhs - Jan. 1999, revised 2003, 2009, 2010
Page 10
Fundamentals: SEMICONDUCTOR LASER
AuZn contact
Oxide Layer
p-AlGaAs
n-AlGaAs
GaAs active zone
n-GaAs substrate
AuGe contact
p-AlGaAs
n-AlGaAs
Fig. 19: „Buried“ hetero structure. The active zone has been
buried between some layers which ensure an optimal beam
guidance in the zone.
10-30° depending on the type of laser diode.
If the beams are extended geometrically into the active
medium the horizontal beams will have another apparent
point of origin as the vertical beams. The difference between the points of origin is called astigmatic difference
(Fig. 21). It amounts to about 10 µm for the so called index
guided diodes whereas for the gain guided diodes these
values are appreciably higher.
Modern diodes are mostly index guided diodes. This
means that the laser beam is forced not to leave the resonator laterally by attaching lateral layers of higher refractive
index to the active zone. At the gain guided diodes the
injection current is forced to pass along a small path (about
2-3 µm width).
2.5.1Resonator and beam guidance
Astigmatic difference δε
As already mentioned at the beginning the diode laser difθ
fers from the „classical“ lasers in the dimensions of the
δε
resonator and in the propagation of the beam. For the diode
lasers the active material represents the resonator at the
same time. Furthermore the ratio of the resonator length L
(300 µm ) to the wavelength λ (820 nm) 366. For a HeNeLaser ( λ = 632 nm ) with a typical resonator length of 20
cm this ratio is 3108. Considering additionally the lateral
dimensions of the resonator we get a ratio of 12.5 for the P
θ
diode lasers with a typical width of 10 µm for the active
zone. With capillary diameters of the He-Ne tubes of about
1 mm one gets a value of 1582. This already indicates that
the beam characteristics of the laser diode will distinguish
Fig. 21: Astigmatic difference de
significantly from „classical“ lasers.
In this way the direction of the amplification (which is proportional to the current flux) and the laser radiation are
2.5.2Divergence and intensity distribudetermined.
tion
At the gain guided diodes the formation of curved wave
fronts within the resonator is disadvantageous since they
simulate spherical mirrors. In this case higher injection
currents provoke transversal modes which will not appear
θ
in index guided diodes because of the plane wave fronts.
Laser diodes with intensity profiles following a Gaussian
curve and a beam profile which is only limited by diffraction are called Diffraction Limited Lasers (DFL).
They represent the most „civilised“ diode lasers. For the
time being they are only available for powers up to 200
mW. High power diode lasers as used, for example, to
P
θ
pump Nd YAG lasers partially have very fissured nearly
rectangular intensity profiles.
Fig. 20: Elliptical beam profile of a diffraction limited laser
diode in the far field (some meters).
2.5.3Polarisation
It is understandable that the laser radiation of the diodes
has a distinct direction of polarisation, since the height of
Not only the beam guidance but also the size of the la- the exit window is 4 times and the width 12.5 times larger
ser mirrors influences the beam geometry. Generally for than the wavelength. Because of the fraction of spontaconventional lasers the mirrors are very large compared neous emission the light of the laser diode also contains
with the beam diameter. The laser mirror (crystal front components oscillating in the vertical direction.
face area of the active zone) of the laser diodes has a size The ratio of polarisation, P to P , depends on the output
⊥

of about 10 µm x 2 µm, through which the laser beam has power since for higher laser
power the ratio of spontane„to squeeze“ itself. Diffraction effects will be the conse- ous to stimulated emission is changing (Fig. 23).
quence and lead to elliptical beam profiles (Fig. 20).
The polarisation is parallel to the „junction plane“, that is 2.5.4Spectral properties
the plane which is passed by the injection current perpendicularly. The divergence angles θ⊥ and θ differ by about Another property of the diode laser is the dependence of
Dr. Walter Luhs - Jan. 1999, revised 2003, 2009, 2010
Page 11
Fundamentals: PHOTODIODES
Temperature →
Fig. 22: Emission wavelength as a function of the temperature of the laser diode and hysteresis.
The wavelength increases with increasing temperature.
The reason for this is that the refractive index and the
length of the active zone, respectively the resonator, increase with increasing temperature. Beyond a certain temperature the mode does not fit any more into the resonator
and another mode which faces more favourable conditions
will start to oscillate. As the spectral distance between two
successive modes is very large for the extremely short resonator (typical 300 µm ), the jump is about 0.3 nanometer.
Lowering the temperature gets the laser jumping back in
his wavelength. After this the laser must not be necessarily
in the departing mode.
Laser Threshold
Laser Power
Spontaneous
Emission
"LED"
Induced and
spontaneous
Emission
In regard to „classical“ lasers the light of a diode laser contains a remarkable high fraction of non-coherent „LED“
radiation. For currents underneath the laser threshold the
spontaneous emission is dominant. Stimulated emission is
responsible for the strong increase above the laser threshold. The threshold current can be determined by the point
of intersection of the extrapolated characteristic lines of
the initial and of the lasing range. The rounding off of the
characteristic line is the result of spontaneous emission. It
also is the cause for the oscillation of several modes next
to the threshold. At higher currents the mode spectrum
becomes more and more clean.
Laser
+ LED
LED
Injection Current
Fig. 24: Output power of the laser diode as a function of the
injection current
Another important part of the CD-ROM reader is the
photodiode which converts the intensity variation of the
pick-up head into electrical signals. Since the reading
speed is desired to be as high as possible, the right detector among a variety must be selected. The next chapter
therefore will discuss the fundamentals and properties of
semiconductor photodetectors.
2.6Photodiodes
T1
T2 >T1
Injection Current
Fig. 23: Laser power versus injection current with the temperature T as parameter
Applications anticipating the tuning ability of the laser
diode should therefore be performed within a jump-free
range of the characteristic line (Fig. 22).
A similar behaviour is observed for the variation of the
injection current and in consequence for the laser output
power. Here the change in wavelength is mainly the result
Dr. Walter Luhs - Jan. 1999, revised 2003, 2009, 2010
of an increase in the refractive index which again is influenced by the higher electron density in the active zone. A
higher output power provokes also a higher loss of heat
and an increase in temperature of the active zone.
Output Power
Wavelength →
its wavelength on the temperature (about 0.25 nm/°K) and
on the injection current (about 0.05 nm/mA). Users who
need a well defined wavelength have to adjust temperature
and injection current in a way that the wavelength remains
constant. By changing the temperature the wavelength of
the laser radiation can be altered.
Semiconductor pn transitions with a band gap of Eg are
suitable for the detection of optical radiation if the energy
Ep of the photons is equal or greater than the band gap. In
this case an incident photon can stimulate an electron to
pass from the valence band to the conduction band (Fig.
25). Here three types of events are possible:
A An electron of the valence band in the p-zone is stimulated and enters the p-zone of the conduction band.
Because of the external electric field due to the voltage
V it will diffuse through the pn- junction area into the
n-zone and contributes to the external current passing
the resistor R L unless it recombines in the p-zone.
B If an electron of the pn- junction is hit by a photon
the generated hole will migrate into the p-zone and
the electron into the n-zone. The drift of both charges
through the pn-junction causes a current impulse. The
Page 12
Fundamentals: PHOTODIODES
duration of the impulse depends on the drift speed and
on the length of the pn-zone.
C The case is similar to case A. The hole migrates due
to the presence of the external field into the p-zone or
recombines in the n-zone.
V
-
RL
+
4. Avalanche currents are flows of electrons which appear
at high electric field strengths, if, for example, a high
voltage is applied to the photodiode
All these effects contribute to the dark current iD in such a
way that finally the characteristic curve of the diode can be
expressed as follows:
 q⋅U D
is e kT

i
n-zone
A
conduction band

1

C
iD
i Ph
This current i passes the load resistor R L and provokes the
voltage drop Ua, which finally represents the signal.
i
B
incident
photons
i Ph
Ud
Eg band gap
Ub
RL
p-zone
Ua
valence band
pn-junction
Fig. 25: Absorption of a photon with subsequent transition of the stimulated electron from the valence band to the
conduction band
Only electrons from the pn - junction (case B) or near its
boundary (area of diffusion, case A and C) contribute to
the external current due to stimulation by photons. All others will recombine within their area. In the utmost case
one elementary charge q can be created for each incoming photon. As already mentioned, not every photon will
create in the average a current impulse. In this context the
production rate G, leading to an average current <iPh> is
defined as follows:
q⋅G
i Ph
At a light energy of P0 a number of
Po
w
Ub
is
Ud
P1
P2
Fig. 26: Characteristic curve of a photodiode in the photoconductive mode
A good detector for optical communication technology is
characterised by the fact that it is very fast (up to the GHz
range) and that it has a high quantum efficiency which
means that it is very sensitive. Depending on the wavelength range which has to be covered by the detector one
uses silicon or germanium semiconductor material for the
construction of the detectors.
2.6.1Ge and Si PIN photodiodes
photons will hit the detector as w is just the energy of
To have absorption of a photon at all, its energy has to fit
a single photon. But only that fraction of photons is coninto the band structure of the material under consideration.
verted into current pulses which is absorbed in the area of
From the condition
the pn-zone. This fraction may be called ηP0 , where η is
hc
called quantum efficiency. The number of generated curE ph ω hν
≥ EG
rent pulses or the production rate therefore will be
λ
G
η
⋅ P0
ω
and the average photo current:
i Ph
η⋅q
⋅ P0
ω
One recognises that for large wavelengths the energy of
the photon may no more be sufficient „ to lift“ the electron
in a way that it passes the band gap. For smaller wavelengths one has to respect that the conduction band and
also the valence band have upper edges which is followed
by a band gap.
Because of processes which are typical for semiconductors there is already a current flowing even if there are no
photons entering the detector. This current is called „dark“
current and has four reasons:
1. Diffusion current, it is created because of statistical
oscillations of the charge carriers within the diffusion
area
2. Regeneration or recombination current, it is generated
by random generation and annihilation of holes
3. Surface currents, which are hardly avoidable since the
ideal insulator does not exist
Dr. Walter Luhs - Jan. 1999, revised 2003, 2009, 2010
Page 13
Fundamentals: COMPACT DISK (CD)
CD looks like and how it can be read by the optical pick-up
system.
100
Ge
80
2.7Compact Disk (CD)
0.8
1.0
1.2
1.4
1.6
1.8
Wavelength µm
Fig. 27: Spectral sensitivity curves of Si and Ge
photodiodes
2.7.1Disc Format
12
0
m
m
15
ad
O
ut
m
m
Photons having energies exceeding the upper limit of the
conduction band can no more be absorbed. The wavelength
of the applied light source decides which detector material
is to be used. For wavelengths above 1 µm up to 1.5 µm
Germanium is recommended. Below these values Silicon
detectors are used. In the present experiment LED’s of 520
and 630 nm wavelength are is applied. Therefore a silicon
detector is used. To get a high quantum efficiency not a PN
but a PIN detector has been chosen.
a
0.6
at
0
0.4
D
20
In
40
With the experience from the times of the disk record
player where the size of the disks has been internationally
standardised, the manufacturer of the new CD-ROM media agreed to establish similar standards for the CD-ROM
itself as well as for the data structure. Every CD should
work world-wide independently of the producer of the CD
or the CD reader. This in fact was the greatest challenge
during the development of this new technology. In the so
called “Red Book” the inventors of the CD’s Philips and
Sony published 1982 the complete format including the
physical size as well as the data structure of the CD-DA
(Digital Audio).
ad
60
Le
Relative sensitivity %
Si
1.2
SiO2
Le
Photon
Anti reflex
coating
i
RL
p
Signal
n
A
Fig. 29: International standard for CD size
η
(1 R)(1 e
α ⋅d
)e
α ⋅d p
Protective
Lacquer
coating
10-30 µm
5 µm
Based on the standard of the red book, the standard for
computer the CD-ROM’s (Read Only Memory) has been
+
defined in the “Yellow Book”. The next (Fig. 30) shows
Fig. 28: Construction of a PIN detector
the magnified detail A. The metallized pit surface is
embedded
between a protective lacquer coating and the
Contrary to a detector with a simple pn-layer this type of
transparent
disc substrate which is usually made from
detector has an intrinsic conducting layer inserted in bePolycarbonate.
tween the p and n layer. Therefore the name PIN diode.
The reason for this is to enlarge the pn junction area which
increases the probability of absorption of a photon and the
Printing Ink
generation of a current impulse, e.g. the quantum efficiency. The quantum efficiency for such an arrangement is:
R is the Fresnel reflection of the Si or Ge surface which is
hit by the photons, α is the coefficient of absorption, d the
Metalised
thickness of the intrinsic zone and dp the thickness of the p
Pit surface
layer. By attachment of a reflex reducing layer on the upper
0.05 - 0.1 µm
Transparent
Pit
side of the p layer R can get a value of less than 1%. Since
Disc Substrate
α⋅dp is anyhow <<1, the thickness of the intrinsic layer Fig. 30: Magnified cross section of the CD
should be chosen as large as possible. The consequence of
this is that the drift time rises and the limiting frequency Independently on the kind of CD the reading process alof the detector is reduced. In so far a compromise between ways starts from the inner of the disk beginning with the
high quantum efficiency and high limiting frequency has “Lead-In” (Fig. 29) with a width of app. 5 mm and contains
a kind of table of contents of the data. Next to it the data
to be made.
Now we collected all informations about the optical key range with a width of max. 33 mm follows. The Lead-Out
components of a CD ROM reader and within the next area is located directly behind the end of the actual data
chapter we will give the details how the data structure of a area and contains a kind of “end of file” information.
The digital encoded signal is pressed into the CD as pits.
Dr. Walter Luhs - Jan. 1999, revised 2003, 2009, 2010
Page 14
Fundamentals: COMPACT DISK (CD)
0.5 µm
3T - 11 T
Looking from top of the CD where the label is printed,
the signature appears as pits, whereas the laser beam
which contacts the CD from the bottom side “sees” bumps.
Between the pits the surface is flat and termed as land. The
pits are arranged on a spiral which starts from the inner
centre (Lead In) to the outer part of the CD (Lead out) with
a pitch of 1.6 µm. Unspiraled the length of the track would
be 5.600 m !
1.7 µm
0.833 - 3.054 µm
focused laser beam
Fig. 31: Arrangement and size of the pits.
One can suppose, that due to the existence of pits and land
the binary information is directly encoded as zero or one,
but the principle of a rotating disk will not allow this simple principle. Instead the change from a pit to a land will
recognised as logical one and the duration of detecting
whether a pit or land will be interpreted as logical zero. By
the way, it is the same principle as used by magnetic harddrive disks, where changes of the magnetic flux are acting
like pits and lands.
7
00000111 00100100000000
8
00001000
01001001000000
9
00001001 10000001000000
10
00001010
10010001000000
a nd so on.
One aspect has not been solved by using the EFM modulation, the problem when a byte with a leading logical
one follows a byte which ends with a logical one. For this
purpose three merging bits are attached to each byte for
the separation of the bytes. Finally each byte now has a
length of 17 bits. A block of 24 bytes represents a frame,
the smallest information block of the CD. However some
more information is attached to a frame before the composition forms a data block. At the beginning of the frame
a synchronisation pattern consisting of 17 bits is attached
which informs the reader, that a new frame is coming up.
The synchronisation pattern is followed by a control byte
of 17 bits and the 24 data bytes. The frame is closed by an
error correction code consisting of 8 bytes each consisting
of 17 bits. In summary the smallest logical data block or
frame consists of 588 bits.
Bits
27
1 ⋅ 17
24 ⋅ 17
8 ⋅ 17
588
Synchronisation pattern
Control byte
Data
Error correction
Bits of a frame
On the next higher level the frames are organised by so
called sectors whereby each one contains 98 frames. A
sector is organised in such a way that on one side the data
of the frames and on the other side the error correction and
control bytes are arranged
3234
2352
882
0 0 1 0 0 0 0 0 1 0 0 0 1 0 0 01 0 0 0 0 0 0 0 0 01 0 0 0 0 01 0 0 0 0 01 0 0 0 0 0 0 01 0 0
Data 2352 Byte
EDC/ECC
#1
392 Byte
EDC/ECC
#2
392 Byte
Control
Bytes
98 Byte
Fig. 32: Transition from Land to Pit generates a logical one
24
4
4
1
It is obviously that with the above algorithm it is not possible to generate a 11 sequence, since a logical one appears
only at a pit to land or land to pit transition. Therefore a
..
logical one requires always a following logical zero. In
..
..
practise actually it has been shown, that for a secure detection of a sequence of two logical ones in minimum two
98 Frames
logical zeroes should between them. For this reason the
Fig. 33: Arrangement of frames inside a sector
EFM modulation or encoding is used, which converts an
The structure of a sector (Fig. 33) is the genuine format for
eight bit long byte into a fourteen bit long one.
all audio CD’s (CD-DA). Each sector is assigned to be read
in 1/75 second, that means a CD drive reads 75 sectors per
Decimal Binary
EFM
second. Considering that a common CD player operates
0
00000000 01001000100000
at a sample rate of 44.1 kHz with 16 bit samples and two
1
00000001 10000100000000
channels, 1.411.200 bits per second are read in one second,
2
00000010 10010001000000
corresponding to 18.816 bit per 1/75 second. This value
3
00000011 10001000100000
divided by 8 bits results in 2352 bytes, the same amount
4
00000100 01000100000000
stored in one sector.
5
00000101 00000100010000
Because the sectors have to be played back in a well de
6
00000110 00010000100000
fined time its addressing is done by a time stamp as for exDr. Walter Luhs - Jan. 1999, revised 2003, 2009, 2010
Page 15
Fundamentals: COMPACT DISK (CD)
width of the pits is in the range of the beam waist of a focused diode laser beam at 780 nm. In all cases where an
object has nearly the same dimension of the wavelength
of the light with which the object interacts, the diffraction
phenomena are dominating instead of geometrical considerations.
Incident light
Intensity Distribution
φ
Slit
b
ample 22/48/56, meaning the 56 sector of the 48 second of
the minute 22. Further global information for the addressing of sound tracks are kept inside the 98 control bytes of
a sector. From the first bit of the control bytes the so called
sub-channel P is generated which contains the information
about the beginning of a sound track. The sub-channel Q
which is formed by the second bits informs about the position of the sector with respect to the CD. A sub-channel
Q of a sector within the Lead-In area of the CD contains
special information like the table of contents from which
also the number of records can be concluded. The residual
6 bits of the control bytes are used for the generation of
synchronisation signals to control the drive speed.
The error correction bytes (2*392) which are stored in
each sector are based on an algorithm which is defined in
the red book and already applied on the hardware level of
the audio CD or CD-ROM reader to ensure an error free
transmission of the sectors. This algorithm stems from the
widely used Reed-Solomon-Code, which is slightly modified and termed as Cross-Interleave-Reed-Solomon-Code
or in its short form CIRC. The probability that a data error
is not detected is 10 -8 using this algorithm. That means that
from 100 million bits one erroneous one can slip through
the error trap. For audio applications this rate is quite sufficient, whereas for computer programs the rate cannot be
accepted since crashing a program can result in severe
problems. Therefore CD-ROM’s are operating with a sector format which is slightly modified as it is defined in the
red book. The format of the CD-ROM is defined in the yellow book which bases on the red one, however it describes
the storage of computer data on audio CD’s.
To improve the error detection and correction rate more
information is stored in the sectors by reducing the amount
of effective data from 2352 to 2048, which anyhow is a
more suitable value for computer data processing.
Fig. 35: Diffraction at a slit with a width of b
This becomes clear if one looks at the bottom surface of a
CD, one notices the occurrence of rainbow colours. Beside
that the CD also acts as a mirror. Thus we can conclude,
that a CD acts as a reflecting grating. If no pits where
printed to the CD it would act only as a simple mirror. To
understand why the back reflected intensity of the incident
laser beam is changed when it hits a bump (pit) we have to
recall the diffraction at a slit.
Actually the CD has no slits, but the description of a slit
with a mirror behind it is equivalent. In a first step we neglect the height of the bump. From basic physics textbooks
we learn that the intensity distribution of the light behind
a slit can be written as:
3234
2352
Synch Header
12 byte 12 byte
Data 2048 byte
I (φ )  b2 ⋅
882
EDC
4 byte
free
ECC
8 byte 276 byte
EDC/ECC
#2
392 byte
EDC/ECC
#1
392 byte
Control
Bytes
98 byte
Fig. 34: Sector format of a CD-ROM Mode 1
The yellow book defines two different kinds of CD-ROM
sector formats, the mode 1 and 2. Apparently almost all
CD-ROM’s for computer applications are of the mode 1
type so that we do not go into details of the mode 2 type.
At the beginning of a sector additional 12 synchronisation
bytes and a 4 byte header have been added. From the header the reader can recognise which mode is used and which
address (minute/second/number) the sector has.
Some other sector formats of CD’s are existing, but a detailed description would stress the frame of this project.
For more information please see the chapter “Further
Reading”.
After understanding the way how and in which format
data are arranged on a CD we will discuss in the following
chapter how the data are read by the CD drive.
2.7.2Optical Detection of Pits and Land
 π⋅b

⋅ sin φ
sin 2 
 λ

 π⋅b

⋅ sin φ


λ
2
If the width of the slit b is large compared to the wavelength λ of the incident beam we will obtain an intensity
distribution as shown in (Fig. 36) A. When we are going
to decrease the width of the slit towards b=λ we will get an
intensity distribution like (Fig. 36) B. For the case of the
CD we can conclude that due to the now broad intensity
distribution only a small part of the back reflected light
will enter the aperture of the pick-up optics resulting in a
reduction of the reflected intensity. The light which shines
beside the bump is simply reflected back and enters without change of direction and intensity the pick-up optics.
Both the reflected light from the bump as well the light
around it are in phase. If we introduce a phase shift of 90°
(λ/2) between both light beams destructive interference
occurs resulting in a further decrease of the back reflected
light intensity. For this reason the height of the reflecting
bump is chosen to be λ/4.
We already know that the information stored on a CD is
encoded by the length of the individual pits and lands. The
Dr. Walter Luhs - Jan. 1999, revised 2003, 2009, 2010
Page 16
Fundamentals: COMPACT DISK (CD)
piece of optical, mechanical and electronic engineering.
Meanwhile this optical stylus has been improved in such
a way that secure reading of CD’s in connection with electronic buffer circuits is even possible in automobiles or as
portable “CD walkman”.
Intensity
A
LD
PBS
DG
L1 QWP
L2
CD
B
-80
-60
-40
-20
0
20
40
60
80
Diffraction angle φ [deg.]
Fig. 36: Intensity distribution for a width of the slit
A) b = 6 λ and B) b = λ
ACD
Fig. 38: Three beam pick-up system
λ/4
At the beginning of the CD era the pick-up systems conIn summary the change of back reflected intensity when sisted of so called single beam arrangements. Nowadays
the laser beam hits a bump is due to diffraction and in- almost all reader or player are using three beams as shown
terference. In the next (Fig. 37) the situation is shown to in the (Fig. 38). During the reading process the pick-up
is moved from the inner to the outer of the CD whereby
illuminate the phenomena of interference.
an electromagnetic correction system keeps it on the data
track. Single beam units are required to move the pick-up
on a non linear curve in order to stay in the right orientation to the data track. Due to the improvement of the
correction system and the use of a three beam stylus this
movement is possible now along a straight line which
reduces the mechanical expenditure. In general the pickup system consists of a laserdiode (LD) which emits at a
wavelength of 780 nm, the diffraction grating (DG) which
generates the required three beams out of the single beam
A
of the laserdiode, the polarising beam splitter (PBS), the
B
collimating lens (L1), the quarter wave plate (QWP) and
bump
finally the optics L2 which acts as focusing as well as pickup optics. Before we going to analyse the function of the
pick-up system we will discuss the so far not explained
components like the diffraction grating, the beam splitter
Focused Laser spot and the quarter wave plate.
2.7.4 Diffraction Grating
2.7.3Pick - up System
Intensity distribution
s
Incident light
φ
The fig. 2.40 shows the simplified situation of interference.
A part of the incident beam (A) is diffracted whereas another part is reflected from the surface of the CD. Since the
bump has a height of λ/4 resulting in a total phase shift of
λ/2 of the beam B versus A. Actually the above (Fig. 37)
is not quite correct, because the amplitude of a reflected
wave is always zero at the reflecting surface. To give a correct description one has to use the time scale rather than a
local description. But a graphical presentation in the time
domain would suffer for understanding.
Furthermore it should be noted, that beam A and B do not
travel along a common path which however is necessary
for interference. But due to the diffraction which causes a
local intensity distribution one always will find rays fulfilling the demand for a common path.
In (2.7.4) we already discussed the diffraction of light
when it interacts with a single slit.
b
Fig. 37: Interference of reflected beam A and B having a
phase shift of λ/2 to each other.
Fig. 39: Fig. 2.42: Diffraction at a grating with two slits
having a width of b and a distance of s
The pick-up of a CD reader is an ingenious master- If we are going to increase the number p of slits the intenDr. Walter Luhs - Jan. 1999, revised 2003, 2009, 2010
Page 17
Fundamentals: COMPACT DISK (CD)
sity distribution behind the slit can be expressed as:
tally polarised one (P) will be transmitted (case B of (Fig.
41)). A beam with a 45° polarisation can be considered as
a superposition of two beams with S and P polarisation

 π⋅b

 π⋅p
sin 2 
⋅ sin φ sin 2 
⋅ s ⋅ sin φ
and therefore it will be separated spatially into its initial

 λ

 λ
components (case C of (Fig. 41)). Some manufacturers are
I (φ ) 
⋅
2
π


2
using beam splitter plates instead of cubes, however such
π
b
⋅


sin  ⋅ s ⋅ sin φ
⋅ sin φ



a
plate causes an additional deviation parallel with respect
λ

λ
to the incident beam. As we have learned in the (2.5.3) the
emission of a laserdiode is polarised. In the pick-up system
In (Fig. 40) this function is represented. For a given wave- the laserdiode is arranged in such a way that the laser light
length λ the grating can be optimised in such a way that the completely passes the polarising beam splitter cube before
intensity distribution now consists of three distinct maxi- it is collimated by means of the lens L1. The now parallel
ma forming at least the desired three beam configuration. beams passes in a next step the quarter wave plate which
The so formed beams are passing as next the polarising is manufactured from crystalline quartz.
beam splitter(PBS). A shown in the (Fig. 41) the splitter is
manufactured as cube which actually consists of two ce- 2.7.6Quarter Wave Plate
mented right angle prisms whereby the hypothenuse face
Such a material exhibits an amazing effect. At a certain
of one prism is covered with a dielectric coating.
thickness, a multiple of a quarter wavelength, the plate
converts linear polarised light into circular polarised light
when the beam enters the plate under an angle of 45° with
respect to the optical crystal axis.
Z
x
Intensity
45˚
Incident beam
-0.3
-0.2
-0.1
-0.0
0.1
0.2
0.3
Diffraction angle φ [deg.]
Fig. 40: Three beam intensity distribution generated by a
for this purpose designed grating
Quarter Wave Plate
Fig. 42: Quarter wave plate converts linear to circular
polarised light
2.7.5Beam Splitter
The design of the coating determines the behaviour if
light is divided whether with respect to its intensity (neutral beam splitter) or polarisation. Furthermore the design controls also the ratio of the deviated beam intensity.
Commonly this ratio is 1, that means that the intensity of
the transmitted or deflected beam is the same.
S
P
S
P
A
y
P
S
B
Crystalline quartz has two distinct optical orientation
which differ in the index of refraction. One of this orientation is termed as ordinary whereas the other one as extra
ordinary, because light which travels in this direction is retarded against that one which travels in the ordinary direction. If the light or the plate is orientated in such a way that
the incident light beam and the optical axis of the crystal
are forming an angle of 45° then one half (S) of the beam
travels along the ordinary and the other half (P) along the
extra ordinary direction. As result the leaving light consist
now of two orthogonally polarised beams travelling in the
same direction each in the other. In addition the P light
has a retardation or phase shift phase with respect to the
S component. If the thickness of the plate is adjusted to a
quarter or a multiple of it of the wavelength of the light, the
resulting polarisation is now circular as shown in (Fig. 43).
45˚
C
Fig. 41: Polarising beam splitter cube
If a vertically polarised beam (S) enters the cube it will be
totally deflected (case A of (Fig. 41)) whereas a horizonDr. Walter Luhs - Jan. 1999, revised 2003, 2009, 2010
Page 18
Fundamentals: COMPACT DISK (CD)
x
x
z
z
y
y
right circular
left circular
Fig. 43: Circular polarised light
The inverse process happens when circular polarised light
passes a quarter wave plate it is converted to linear polarised light. This behaviour of quarter wave plate is exploited to separate incident beam from the CD reflected
beam. To understand this we will trace the light path of the
pick-up system step by step.
The light which leaves the laserdiode is P polarised and
passes completely the polarising beam splitter cube (PBS).
After passing the quarter wave plate (QWP) it is converted
into right circular polarisation and is focused by means of
the lens L2 onto the pits and lands of the CD. When it is
reflected it travels back and appears to be now left circular
polarised. When it passes the quarter wave plate it is converted back to linear polarised light, but - and this is important - it is converted to S polarised light because it was
left circular polarised with respect to the incident beam
due to the reflection. And now comes the trick: S polarised
light will be deflected by the polarising beam splitter cube
and therefore spatially separated from the incident beam
and deviated to the photodiodes ACB.
The beam coming from the CD via the beam splitter cube
passes a cylindrical lens. Such a lens acts only on the horizontal rays of the beam as a focusing lens whereas the
vertical ones are not affected. Assuming a round shaped
beam enters the cylindrical lens two distinct beam waists
will occur, one for the horizontal part and one for the vertical part of the beam. In case of a perfect lens the beam
shape at this two focal locations would be a horizontal and
a vertical line respectively. At a certain location the beam
will have a round shape. Before and after this point the
beam has either a vertical or horizontal elliptical shape.
The change in the shape of the beam is used to determine
the position of the focal spot on the surface of the CD by
means of a 4 quadrant photodiode array as shown in (Fig.
45)
a
d
b
c
too far S < 0
just right S = 0
too near S > 0
S = (a+c) - (b+d)
Fig. 45: Four quadrant photodetector to determine the location of the focus with respect to the CD
The four quadrant detector is positioned in the beam path
in such a way that when the focus of the reading beam just
hits the reflecting land area of the CD the circular shape is
imaged on the detector. If the CD moves apart from this
position the shape of the image changes as shown in the
(Fig. 44). The signal S is used as input signal for the focal
2.7.7Focusing Error Detection
control loop which consists of the control electronics and
a
moving coil on which the focusing lens is attached (Fig.
As we will learn in the next chapter the photodiode C de46).
To obtain the data signal DS the sum of the detectors
tects the light and dark transitions caused by the reading of
like
DS=a+b+c+d
is formed.
a pit or land whereas A and B will be used to control the
pick-up that it stays always in a proper position to the data
Focusing
Pole shoe
track. In addition the photodetector C is used to generate
lens
Permanent
a signal to keep the focal point of the laser light on the CD.
Magnet
Before this can be done we have to insert a cylindrical
lens into the path between the beam splitter cube and the
photodetectors which is an essential need for the focus
control.
Moving coil
horizontal
elliptical
round
cylindrical
lens
vertical
elliptical
beam coming from
beam splitter cube
Fig. 46: Moving coil system
The control loop for the correct focusing is one part of the
entire control system. The other one is control system for
the data tracking. It has to be made sure that the pick-up
system exactly follows the data track from the beginning
at the inner to the outer circumference. For this purpose
the pick-up system is mounted onto a sledge which is driven by a motor.
Fig. 44: Beam shaping with a cylindrical lens
Dr. Walter Luhs - Jan. 1999, revised 2003, 2009, 2010
Page 19
Experimental Set-up: COMPACT DISK (CD)
3.0Experimental Set-up
2.7.8Tracking error detection
If we recall that the tracks have a lateral distance of 1.6 µm
and a width of 0.5 µm a pure mechanical tracking system
having this accuracy would be two expensive and too large.
Therefore a mechanical system is used for the coarse and a
lateral moving coil for the fine tuning. In (Fig. 46) such a
combined moving coil for vertical and lateral movements
is shown. It consists in this version of two permanent magnets each having a pair of pole shoes and four coils with
an oval shape allowing to be moved due to electromagnetic forces vertically as well as laterally. All four coils are
fixed to a common base plate to which the focusing lens
is attached. The error correcting signal for the tracking is
generated by the additional two beams neighboured to the
main beam.
The equipment consists of the open frame CD-reader
mounted on a base plate (Fig. 49) and the attached control
unit (Fig. 50) which gives access to the relevant signals of
the CD reader. For the operation a IBM compatible PC is
required. The PC must have a USB input. For the connection of the CD reader a USB cable is provided
Due to rapid changes of available models in the market, the
actual reader may differ from the following figures.
approx. 20 µm
Fig. 49: Open frame CD reader mounted on a base plate
A
B
mini USB
connector
C
Fig. 47: Tracking error signal generation
In case of B of (Fig. 47) the signal of both side beams are
equal and their difference yields zero. This is the desired
kind of operation.
a
e
d
power “ON
switch
Fig. 50: Rear side of the reader showing the control box and
the connectors
At the rear panel of the control unit the CD reader is connected by means of the provided USB cable to the computer. The low power voltage (12/5 VDC) are connected via
the provided wall plug supply. The CD-ROM is switched
on by the power switch (Fig. 50).
b
c
f
T=e-f
Fig. 48: Fig. 2.51: Two additional detectors e and f are used
to generate the track error signal
Case A and C are showing the situation where the centre
beam is not in line with the track. One of the side beam
shines on land whereas the other on pits and lands thus the
difference is either greater or smaller zero. The signal T is
the difference of the intensity of the two lateral beams and
is used as input signal for the lateral control loop.
Dr. Walter Luhs - Jan. 1999, revised 2003, 2009, 2010
Fig. 51: Control unit front side
At the front panel of the control unit the following signals
are available:
Page 20
Experimental Set-up: OPERATION
PICK UP SIGNALS:
3.1Operation
When all connections have been done, start the PC. Once
the operating system has been completely loaded you may
check if the PC has recognised the CD reader. For this purpose open the “My Computer” folder and check if the CD
reader is available. To insert the provided audio CD press
FOCUS ERROR
the eject button at the front panel of the reader or click on
The four quadrant detector signal according to (Fig.
“Eject” menu item of the property folder of the reader.
45) indicating the state of the focusing error signal.
TRACK ERROR
This signal corresponds to the tracking error signal
formed by the detector e and f.
OUTPUT SIGNALS:
ANALOG LEFT
Audio signal of the left channel. This signal is only
present if the analog audio of the used computer is
activated, or the digital transfer deactivated see also
(Fig. 52)
ANALOG RIGHT
Audio signal of the right channel. This signal is only
present if the analog audio of the used computer is
activated, or the digital transfer deactivated see also
(Fig. 52)
DIGITAL TTL
Audio digital TTL signal, sometimes also termed S/
PDIF signal. Many modern PC CD-ROM drives have
a two pin digital output connector in the back of the
drive and they sometimes call that interface S/PDIF.
The electrical signal which comes from it is not exactly what is described in S/PDIF specifications, however the data format is exactly the same, but the signal
is TTL level signal instead of the normal 1Vpp signal.
Fig. 52: Digital audio must de unchecked in order to activate the analog audio transfer
Dr. Walter Luhs - Jan. 1999, revised 2003, 2009, 2010
3.2Measurements
Connect an oscilloscope to the desired output BNC connectors at the front panel of the control unit. Please note
that the signal displayed on the oscilloscope may look different depending on the type of oscilloscope used.
Amplitude
OPTICAL DETECTOR
This output shows the analogue intensity variations
caused by the reading of pits and lands.
0
1
2
3
4
5
6
7
8
9
10
11
Period τ (3τ − 11τ)
Fig. 53: Eye Pattern Structure
If one uses an analogue Oscilloscope with a bandwidth of
appr. 100 MHz and connecting the OPTICAL DETECTOR
signal to one channel of the oscilloscope a display like
(Fig. 53) will be shown. The width of pits and lands are
varying in a range of 3T - 11T and are arriving in sequential. If the persistence of the oscilloscope is long compared
to the periods of the intensity variation a pattern like (Fig.
53) results.
Fig. 54: Lower track signal from “OPTICAL DETECTOR”
and upper track “AUDIO TTL”
If the same signal is connected to a digital oscilloscope
with a sampling rate of 1 giga sample per second, the signal can be resolved as shown in the (Fig. 54).
Page 21
Further Readings: MEASUREMENTS
Fig. 55: Left and right analog audio output signal
The (Fig. 55) shows the oscilloscope track of the analog
audio output of the recorded test signal. It can be seen, that
two different frequencies are used and at a closer look the
digitising steps become visible.
Changes in focusing or tracking error signals can be observed by introducing disturbance due to knocking on the
table. If a not any be longer used CD is available, defects
can be generated by bending or locally induced thermal
deformation and the influence of the error correction signal observed.
4.0Further Readings
Ken C. Pohlmann
The Compact Disc Handbook, ISBN 0-89579-300-8
Chris Sherman
CD-ROM Handbook, ISBN 0-07-056693-3
Dr. Walter Luhs - Jan. 1999, revised 2003, 2009, 2010
Page 22
WWW.LD-DIDACTIC.COM
LD DIDACTIC distributes
its products and solutions
under the brand LEYBOLD