Enterprise Network Systems - Lecture 3
Transmission
Characteristics of Fibre Optics
The fibre carries a signal encoded beam of light by virtue of total internal reflection. TIR occurs at the interface of a transparent medium with another medium. If the transparent medium has a higher index of refraction than the surrounding medium, TIR occurs. This makes the optical fibre a waveguide for frequencies in the range 1014 to 1015 Hz, covering the visible and part of the IR spectrum.
Fig 3.1 below displays the principles of FO transmission.
|
In multimode propagation, rays of light entering the fibre at shallow angles are reflected and propagated along the fibre. Other rays at wider angles are absorbed by the surrounding material. Because there are multiple paths through the fibre, each has a different path length and this allows signal elements to spread out in time and limits the rate at which data can be accurately received. There is a smearing effect on the pulses in multimode propagation due to firstly modal dispersion in which there are different path lengths through the fibre. |
 |
At the other extreme is single-mode propagation where the radius of the core is reduced to the order of a single wavelength so only a single mode or angle may pass, the axial ray. Having only one path gives single-mode propagation superior performance over multimode as signal elements cannot spread out in time and so such distortion cannot occur. Single mode fibre SMF has the least diameter of the FO types (8 – 10 micrometres) and displays minimum loss and highest bandwidth. The small core size makes light injection difficult and the preferred light source is laser. There are critical alignment problems here, but once overcome there is neither modal dispersion nor chromatic dispersion as there is one single path through the fibre and the laser light is monochromatic. |
Fig 3.2
Single mode propagation
 |
An intermediate between multimode and single-mode is multimode graded index where the index of the refraction of the core is adjusted which has the effect of focusing the rays more efficiently than multimode. This is also known as multimode step index. |
Fig 3.3
Multimode Graded Index
Propagation
There is
also chromatic
dispersion when using LEDs as there
is a mix of light frequencies and the various components travel at
slightly
different speeds. Multimode fibre has lower bandwidth and higher loss
but
allows efficient coupling of light from cheap LEDs. LANs often use
multimode
fibre but this is almost never used in WAN applications.
This is probable the most commonly used type
of FO cable. It is easiest to make and therefore less expensive. It is
used for short to medium distances at relatively low data rates. The
core of this type of cabl elies in the 50 to 1000 micron range. It
gathers light efficiently and cheap LED sources may be used. Its
diameter means that it may be spliced more easily than the thinner
single mode step index cable, however the large number of paths for the
light means that pulse spreading can occur and this lowers the data
rate of his type of cable - modal dispersion.
This type of cable allows for several paths
for rays of light as can be seen in figure 3.3 above. The
variable RI of the core means that the light that travels closer to the
cladding travels faster due to the lower RI characteristics of the
cable. The light rays are bent smoothely and repeatedly along the
length of the cable and tend to arrive simultaneously. This results in
less modal dispersion, hence higher data rate may be carried.
The typical core diameters are 50 to 100
microns. This makes for easier splicing and cheaper light sources may
be used.
Noise on transmission lines falls into four categories:
·
Thermal noise
·
Intermodulation noise
·
Crosstalk (internal and external)
·
Impulse noise
Thermal
noise is generated when the molecules that form the structure of a
solid vibrate due to their thermal energy. The movement of these
molecules generates random noise that is difficult to filter out.
Intermodulation noise
is produced when there is some non-linearity in the transmitter,
receiver or
transmission line and is unique to the system. This non-linearity can
be caused
by component malfunction or the use of excessive signal strength.
Crosstalk occurs in systems where one carried signal interferes with another. This commonly occurs in metallic transmission media
Impulse noise occurs when an external
electromagnetic field (e.g. a bolt of lightning, poorly shielded HT
equipment etc..) induces unwanted current within the cable
carrying the electrical signal.
External Crosstalk and impulse noise do not affect FO links
so we
only need to consider thermal and intermodulation noise.
Light that is injected into fibre optic cabling will
be attenuated or weakened as it travels
along the cable. This weakening can be due to many factors, some of
which are described below. As well as attenuation, there are many
factors that distort or affect the
travelling light signals.
Although FO is immune to external crosstalk, it can occur within the fibres themselves. The problem is known as four-wave mixing and it is caused by the refractive index of the optical fibre being nonlinear. There is interaction between channels within the fibre that creates sidebands. These sidebands cause interference between channels. The diagram below depicts the result of four-wave mixing.
Fig 3-4 Four-wave mixing
produces
unwanted sidebands
The result of this interference cannot be filtered out and forces the channels to be placed further apart. This limits the capacity of the fibre to carry data.
Another problem that FO suffers from is called chromatic dispersion. This occurs because different light frequencies travel at slightly different speeds. No matter how careful we are to produce a light of a given frequency, we cannot produce light that has solely one frequency, there is a small range within which the light frequency resides. This means that when a perfectly shaped square wave is introduced onto a fibre, some of the components will travel faster than others and some will travel slower resulting in a 'smearing out' of the pulse by the time it reaches the far end of the cable. The effect can be seen in the diagram below.
Fig 3-5 Chromatic dispersion distorts the waveform upon arrival at the destination
The effects of dispersion increase as the square of the bit rate we are sending over the channel.
This dispersion itself falls within two categories, material dispersion and waveguide dispersion. Material dispersion is described above and is due to the different frequencies in the injected light signal.
Waveguide dispersion occurs because the
refractive index of the core and that of the cladding of fibre differ
along the length of the cable, but the effect produced at the
destination is the same as depicted
in fig 3-5.
Rayleigh scattering
occurs and scatters some wavelengths more than others. It is Rayleigh
scattering that is responsible for making the sky appear to be blue as
the red end of the spectrum is scattered away more than the blue end of
the spectrum.
Light signals traveling via a fibre-optic cable are immune from electromagnetic interference (EMI) and radio-frequency interference (RFI). Lightning and high-voltage interference is also eliminated. A fibre network is best for conditions in which EMI or RFI interference is heavy or safe operation free from sparks and static is a must. This desirable property of fibre-optic cable makes it the medium of choice in industrial and biomedical networks. It is also possible to place fibre cable into natural-gas pipelines and use the pipelines as the conduit.
Linear characteristics include attenuation, chromatic dispersion (CD), polarization mode dispersion (PMD), and optical signal-to-noise ratio (OSNR).
Several factors can cause attenuation, but it is generally categorized as either intrinsic or extrinsic. Intrinsic attenuation is caused by substances inherently present in the fibre, whereas extrinsic attenuation is caused by external forces such as bending. The attenuation coefficient α is expressed in decibels per kilometer and represents the loss in decibels per kilometer of fibre.
Intrinsic attenuation results from materials inherent to the fibre. It is caused by impurities in the glass during the manufacturing process. As precise as manufacturing is, there is no way to eliminate all impurities. When a light signal hits an impurity in the fibre, one of two things occurs: It scatters or it is absorbed. Intrinsic loss can be further characterized by two components:
Material absorption
Rayleigh scattering
Material absorption occurs as a result of the
imperfection and impurities in the fibre. The most common impurity is
the
hydroxyl (OH-) molecule, which remains as a residue despite stringent
manufacturing techniques. Figure 3-12 shows the variation of
attenuation with
wavelength measured over a group of fibre-optic cable material types.
The three
principal windows of operation include the 850-nm, 1310-nm, and 1550-nm
wavelength bands. These correspond to wavelength regions in which
attenuation is
low and matched to the capability of a transmitter to generate light
efficiently
and a receiver to carry out detection.
Attenuation
Versus Wavelength
The OH- symbols indicate that at the 950-nm, 1380-nm, and 2730-nm
wavelengths, the presence of hydroxyl radicals in the cable material
causes an
increase in attenuation. These radicals result from the presence
of water
remnants that enter the fibre-optic cable material through either a
chemical
reaction in the manufacturing process or as humidity in the
environment. The
variation of attenuation with wavelength due to the water peak
for
standard, single-mode fibre-optic cable occurs mainly around 1380 nm.
Recent
advances in manufacturing have overcome the 1380-nm water peak and have
resulted
in zero-water-peak fibre (ZWPF). Examples of these fibers include
SMF-28e from
Corning and the Furukawa-Lucent OFS AllWave. Absorption accounts for
three
percent to five percent of fibre attenuation. This phenomenon causes a
light
signal to be absorbed by natural impurities in the glass and converted
to
vibration energy or some other form of energy such as heat. Unlike
scattering,
absorption can be limited by controlling the amount of impurities
during the
manufacturing process. Because most fibre is extremely pure, the fibre
does not
heat up because of absorption.
As light travels in the core, it interacts with the silica molecules in the core. Rayleigh scattering is the result of these elastic collisions between the light wave and the silica molecules in the fibre. Rayleigh scattering accounts for about 96 percent of attenuation in optical fibre. If the scattered light maintains an angle that supports forward travel within the core, no attenuation occurs. If the light is scattered at an angle that does not support continued forward travel, however, the light is diverted out of the core and attenuation occurs. Depending on the incident angle, some portion of the light propagates forward and the other part deviates out of the propagation path and escapes from the fibre core. Some scattered light is reflected back toward the light source. This is a property that is used in an optical time domain reflectometer (OTDR) to test fibers. The same principle applies to analyzing loss associated with localized events in the fibre, such as splices.
Short wavelengths are scattered more than longer wavelengths. Any wavelength that is below 800 nm is unusable for optical communication because attenuation due to Rayleigh scattering is high. At the same time, propagation above 1700 nm is not possible due to high losses resulting from infrared absorption.
Extrinsic attenuation can be caused by two external mechanisms: macrobending or microbending. Both cause a reduction of optical power. If a bend is imposed on an optical fibre, strain is placed on the fibre along the region that is bent. The bending strain affects the refractive index and the critical angle of the light ray in that specific area. As a result, light traveling in the core can refract out, and loss occurs.
A macrobend is a large-scale bend that is visible, and the loss is generally reversible after bends are corrected. To prevent macrobends, all optical fibre has a minimum bend radius specification that should not be exceeded. This is a restriction on how much bend a fibre can withstand before experiencing problems in optical performance or mechanical reliability.
The second extrinsic cause of attenuation is a microbend. Microbending is caused by imperfections in the cylindrical geometry of fibre during the manufacturing process. Microbending might be related to temperature, tensile stress, or crushing force. Like macrobending, microbending causes a reduction of optical power in the glass. Microbending is very localized, and the bend might not be clearly visible on inspection. With bare fibre, microbending can be reversible.
Chromatic dispersion is the spreading of a light pulse as it travels down a fibre. Light has a dual nature and can be considered from an electromagnetic wave as well as quantum perspective. This enables us to quantify it as waves as well as quantum particles. During the propagation of light, all of its spectral components propagate accordingly. These spectral components travel at different group velocities that lead to dispersion called group velocity dispersion (GVD). Dispersion resulting from GVD is termed chromatic dispersion due to its wavelength dependence. The effect of chromatic dispersion is pulse spread.
As the pulses spread, or broaden, they tend to overlap and are no longer distinguishable by the receiver as 0s and 1s. Light pulses launched close together (high data rates) that spread too much (high dispersion) result in errors and loss of information. Chromatic dispersion occurs as a result of the range of wavelengths present in the light source. Light from lasers and LEDs consists of a range of wavelengths, each of which travels at a slightly different speed. Over distance, the varying wavelength speeds cause the light pulse to spread in time. This is of most importance in single-mode applications. Modal dispersion is significant in multimode applications, in which the various modes of light traveling down the fibre arrive at the receiver at different times, causing a spreading effect. Chromatic dispersion is common at all bit rates. Chromatic dispersion can be compensated for or mitigated through the use of dispersion-shifted fibre (DSF). DSF is fibre doped with impurities that have negative dispersion characteristics. Chromatic dispersion is measured in ps/nm-km. A 1-dB power margin is typically reserved to account for the effects of chromatic dispersion.
Polarization mode dispersion (PMD) is caused by asymmetric distortions to the fibre from a perfect cylindrical geometry. The fibre is not truly a cylindrical waveguide, but it can be best described as an imperfect cylinder with physical dimensions that are not perfectly constant. The mechanical stress exerted upon the fibre due to extrinsically induced bends and stresses caused during cabling, deployment, and splicing as well as the imperfections resulting from the manufacturing process are the reasons for the variations in the cylindrical geometry.
Single-mode optical fibre and components support one
fundamental mode, which
consists of two orthogonal polarization modes. This asymmetry
introduces small
refractive index differences for the two polarization states. This
characteristic is known as birefringence. Birefringence causes
one
polarization mode to travel faster than the other, resulting in a
difference in
the propagation time, which is called the differential group delay
(DGD).
DGD is the unit that is used to describe PMD. DGD is typically
measured in
picoseconds. A fibre that acquires birefringence causes a propagating
pulse to
lose the balance between the polarization components. This leads to a
stage in
which different polarization components travel at different velocities,
creating
a pulse spread
PMD is not an issue at low bit rates but becomes an issue at bit rates in excess of 5 Gbps. PMD is noticeable at high bit rates and is a significant source of impairment for ultra-long-haul systems. PMD compensation can be achieved by using PMD compensators that contain dispersion-maintaining fibers with degrees of birefringence in them. The introduced birefringence negates the effects of PMD over a length of transmission. For error-free transmission, PMD compensation is a useful technique for long-haul and metropolitan-area networks running at bit rates greater than 10 Gbps. Note in Figure 3-13 that the DGD is the difference between Z1 and Z2. The PMD value of the fibre is the mean value over time or frequency of the DGD and is represented as ps/ km. A 0.5-dB power margin is typically reserved to account for the effects of PMD at high bit rates.
Polarization dependent loss (PDL) refers to the difference in the maximum and minimum variation in transmission or insertion loss of an optical device over all states of polarization (SOP) and is expressed in decibels. A typical PDL for a simple optical connector is less than .05 dB and varies from component to component. Typically, the PDL for an optical add/drop multiplexer (OADM) is around 0.3 dB. The complete polarization characterization of optical signals and components can be determined using an optical polarization analyzer.
The optical signal-to-noise ratio (OSNR) specifies the ratio of the net signal power to the net noise power and thus identifies the quality of the signal. Attenuation can be compensated for by amplifying the optical signal. However, optical amplifiers amplify the signal as well as the noise. Over time and distance, the receivers cannot distinguish the signal from the noise, and the signal is completely lost. Regeneration helps mitigate these undesirable effects before they can render the system unusable and ensures that the signal can be detected at the receiver. Optical amplifiers add a certain amount of noise to the channel. Active devices, such as lasers, also add noise. Passive devices, such as taps and the fibre, can also add noise components. In the calculation of system design, however, optical amplifier noise is considered the predominant source for OSNR penalty and degradation.
OSNR is an important and fundamental system design consideration. Another parameter considered by designers is the Q-factor. The Q-factor, a function of the OSNR, provides a qualitative description of the receiver performance. The Q-factor suggests the minimum signal-to-noise ratio (SNR) required to obtain a specific BER for a given signal. OSNR is measured in decibels. The higher the bit rate, the higher the OSNR ratio required. For OC-192 transmissions, the OSNR should be at least 27 to 31 dB compared to 18 to 21 dB for OC-48.
Nonlinear characteristics include self-phase modulation (SPM), cross-phase modulation (XPM), four-wave mixing (FWM), stimulated Raman scattering (SRS), and stimulated Brillouin scattering (SBS).
Phase modulation of an optical signal by itself is known as self-phase modulation (SPM). SPM is primarily due to the self-modulation of the pulses. Generally, SPM occurs in single-wavelength systems. At high bit rates, however, SPM tends to cancel dispersion. SPM increases with high signal power levels. In fibre plant design, a strong input signal helps overcome linear attenuation and dispersion losses. However, consideration must be given to receiver saturation and to nonlinear effects such as SPM, which occurs with high signal levels. SPM results in phase shift and a nonlinear pulse spread. As the pulses spread, they tend to overlap and are no longer distinguishable by the receiver. The acceptable norm in system design to counter the SPM effect is to take into account a power penalty that can be assumed equal to the negative effect posed by XPM. A 0.5-dB power margin is typically reserved to account for the effects of SPM at high bit rates and power levels.
Cross-phase modulation (XPM) is a nonlinear effect that limits system performance in wavelength-division multiplexed (WDM) systems. XPM is the phase modulation of a signal caused by an adjacent signal within the same fibre. XPM is related to the combination (dispersion/effective area). CPM results from the different carrier frequencies of independent channels, including the associated phase shifts on one another. The induced phase shift is due to the walkover effect, whereby two pulses at different bit rates or with different group velocities walk across each other. As a result, the slower pulse sees the walkover and induces a phase shift. The total phase shift depends on the net power of all the channels and on the bit output of the channels. Maximum phase shift is produced when bits belonging to high-powered adjacent channels walk across each other.
XPM can be mitigated by carefully selecting unequal bit rates for adjacent WDM channels. XPM, in particular, is severe in long-haul WDM networks, and the acceptable norm in system design to counter this effect is to take into account a power penalty that can be assumed equal to the negative effect posed by XPM. A 0.5-dB power margin is typically reserved to account for the effects of XPM in WDM fibre systems.
FWM can be compared to the intermodulation distortion in standard electrical systems. When three wavelengths (λ1, λ 2, and λ 3) interact in a nonlinear medium, they give rise to a fourth wavelength (λ 4), which is formed by the scattering of the three incident photons, producing the fourth photon. This effect is known as four-wave mixing (FWM) and is a fibre-optic characteristic that affects WDM systems.
The effects of FWM are pronounced with decreased channel spacing of wavelengths and at high signal power levels. High chromatic dispersion also increases FWM effects. FWM also causes interchannel cross-talk effects for equally spaced WDM channels. FWM can be mitigated by using uneven channel spacing in WDM systems or nonzero dispersion-shifted fibre (NZDSF). A 0.5-dB power margin is typically reserved to account for the effects of FWM in WDM systems.
When light propagates through a medium, the photons interact with silica molecules during propagation. The photons also interact with themselves and cause scattering effects, such as stimulated Raman scattering (SRS), in the forward and reverse directions of propagation along the fibre. This results in a sporadic distribution of energy in a random direction.
SRS refers to lower wavelengths pumping up the amplitude of higher wavelengths, which results in the higher wavelengths suppressing signals from the lower wavelengths. One way to mitigate the effects of SRS is to lower the input power. In SRS, a low-wavelength wave called Stoke's wave is generated due to the scattering of energy. This wave amplifies the higher wavelengths. The gain obtained by using such a wave forms the basis of Raman amplification. The Raman gain can extend most of the operating band (C- and L-band) for WDM networks. SRS is pronounced at high bit rates and high power levels. The margin design requirement to account for SRS/SBS is 0.5 dB.
Stimulated Brillouin scattering (SBS) is due to the acoustic properties of photon interaction with the medium. When light propagates through a medium, the photons interact with silica molecules during propagation. The photons also interact with themselves and cause scattering effects such as SBS in the reverse direction of propagation along the fibre. In SBS, a low-wavelength wave called Stoke's wave is generated due to the scattering of energy. This wave amplifies the higher wavelengths. The gain obtained by using such a wave forms the basis of Brillouin amplification. The Brillouin gain peaks in a narrow peak near the C-band. SBS is pronounced at high bit rates and high power levels. The margin design requirement to account for SRS/SBS is 0.5 dB.