  
 
|
FSO Comparisons
Free space optical communications is now
established as a viable approach for addressing the emerging broadband
access market and its “last mile” bottleneck.. These robust systems, which
establish communication links by transmitting laser beams directly through
the atmosphere, have matured to the point that mass-produced models are
now available. Optical wireless systems offer many features, principal
among them being low start-up and operational costs, rapid deployment, and
high fibre-like bandwidths. These systems are compatible with a wide range
of applications and markets, and they are sufficiently flexible as to be
easily implemented using a variety of different architectures. Because of
these features, market projections indicate healthy growth for optical
wireless sales. Although simple to deploy, optical wireless transceivers
are sophisticated devices.
The many sub-systems require a multi-faceted approach to system
engineering that balances the variables to produce the optimum mix. A
working knowledge of the issues faced by an optical wireless system
engineer provides a foundation for understanding the differences between
the various systems available. This paper aims to examine the many
elements considered by the system engineer when designing a product so
that the buyer can ask those same questions about the systems they are
evaluating for purchase.
Which Wavelength?
Currently available Free Space Optics (FSO) hardware can be classified
into two categories depending on the operating wavelength – systems that
operate near 800 nm and those that operate near 1550 nm. Each vendor
manages to make huge claims that their own chosen wavelength is "best".
But as we point out below, it's actually "real world link margin", not
"market-eering" that matters.
Contrary to claims, there are only a few compelling reasons for selecting
1550 nm Free Space Optics (FSO) systems, and many against. One
argument in favour is about laser eye safety, but ignores the effect of
increased transmit aperture used by a competing 980nm solution from a
vendor such as CableFree. There is reduced solar background radiation at
1550nm, but the receiver devices are much less sensitive than the enhanced
silicon at 800-980nm, completely negating any advantage. A spurious
argument is "compatibility with existing technology infrastructure" -
well, as the signal is always regenerated before going into the network,
with an O-E-O conversion, that argument is a complete dead duck.
The wavelength choice alone is misleading. FSO is all about usable
"real-world" link margin. A properly-designed 980nm system is proven
to be far better than a supposed "carrier class" 1550nm system that
suffers reduced output power and insensitive receivers.
The true argument for 1550nm is about the existence of EDFAs, optical
amplifiers which can boost transmit signals to whole watts of power, and
the existence of DWDM components which enable multi-channel multi-gigabit
systems for the future. But the cost penalty associated with 1550nm
makes it inappropriate for a "today" customer needing E1/T1 -> Gigabit
under 4km.
Eye-Safety
Laser beams with wavelengths in the range of 400 to 1400 nm emit light
that passes through the cornea and lens and is focused onto a tiny spot on
the retina while wavelengths above 1400 nm are absorbed by the cornea and
lens, and do not focus onto the retina, as illustrated in Figure 1. It is
possible to design eye-safe laser transmitters at both the 800 nm and 1550
nm wavelengths but the allowable safe laser power is about fifty times
higher at 1550 nm. However, this argument ignores the effect of aperture.
A small-aperture 1550nm system is actually less safe than a large aperture
980nm system, a fact that 1550nm proponents neatly ignore.
Atmospheric Attenuation
Carrier-class Free Space Optics (FSO) systems must be designed to
accommodate heavy atmospheric attenuation, particularly by fog. Although
longer wavelengths are favored in haze and light fog, under conditions of
very low visibility this long-wavelength advantage does not apply. The
argument is about usable link margin, NOT wavelength.
Optical Receivers
There are a number of factors to consider when examining the effectiveness
of the receiver in an FSO system; these include the type of detector used,
the sensitivity rating and size of the detector, the size and design of
the receiver optics, and the operating wavelength itself. In order to
correctly assess the efficiency of the overall system, one must also take
into account the number and power of the lasers being used to generate the
signal.
Types of optical detectors used in FSO equipment come in two basic types:
PIN and APD. The PIN detector is a lower cost detector that has no
internal gain, while the APD is a more expensive but also more sensitive
detector with internal gain. The Benefits of using APD over PIN technology
will vary, but real-world results indicate the benefits to be an
improvement in sensitivity of approximately 4x that of a PIN detector.
The argument comes back again to "usable
real-world link margin". If the vendor has used a cheap detector, it
has to be made up somewhere else. A factor of 4, or 6dB, means 4x
the transmit power, or half the distance.
The size of the receiver optics is also
important; a larger area receive optic contributes to reducing errors due
to scintillation. Scintillation is atmospheric turbulence due to solar
loading and natural convection, causing temporally and spatially varying
refractive index changes in the air. As a laser beam propagates through
the atmosphere, there is a time-varying intensity at the receiver due to
this phenomenon; this is referred to as 'scintillation'. This is quite
similar to the apparent twinkling of the stars or distant city lights,
which is due to the same effect. The result is that an FSO communications
receiver can experience error bursts due to surges and fades in the
receive signal strength. One way to combat this scintillation effect, and
thus improve the error-rate performance, is to use a large aperture
receiver. A collecting aperture that is much larger than the spatial scale
of the scintillation provides an averaging effect of the localized surges
and fades, thus improving the error rate. This large-aperture approach is
more effective for scintillation reduction than multiple smaller
apertures, which perform less averaging at each lens. Another way to
mitigate the effects of scintillation is to use multiple transmitters,
each of which takes a slightly different path through the atmosphere,
which also contributes an averaging effect. The net result is that a
properly designed system can defeat scintillation impairments.
The operating wavelength of an FSO system also contributes to the
performance of the receiver. It is generally true that high-quality
photodiodes at both 800nm and 1550nm achieve comparable quantum
efficiencies.
However, the sensitivity of a 1550nm receiver is generally lower due to
noise floor. Again, it all comes down to "usable real-world link margin".
Commercial Infrastructure
1550 nm wavelength range is actually NOT the most commonly specified for
terrestrial fiber-based optical communications - most LAN ports using
fibre, and new automotive fibre systems all use 850nm, which is cheaper.
Ignoring such bogus arguments, and thinking about pace of technology
developments, most of it is in terms of reducing cost and increasing reach
and bandwidth. In fibre systems, 1550nm is used for long haul, as
dispersion is lower (not true for FSO though - there is no advantage)
850nm devices are used for shorter fibre links, and are much cheaper.
Performance – Transmit Power & Receiver Sensitivity
Free Space Optics (FSO) products performance can be characterized by
four main parameters (for a given data rate):
• Total transmitted power
• Transmitting beamwidth
• Receiving optics collecting area
• Receiver sensitivity
High transmitted power may be achieved by using erbium doped fibre
amplifiers, or by non-coherently combining multiple lower cost
semiconductor lasers. Narrow transmitting beamwidth (a.k.a. high antenna
gain) can be achieved on a limited basis for fixed-pointed units, with the
minimum beamwidth large enough to accommodate building sway and wind
loading. Much narrower beams can be achieved with an actively pointed
system, which includes an angle tracker and fast steering mirror (or
gimbal). Ideally the angle tracker operates on the communication beam, so
no separate tracking beacon is required. Larger receiving optics captures
a larger fraction of the total transmitted power, up to terminal cost,
volume and weight limitations. And high receiver sensitivity can be
achieved by using small, low-capacitance photodetectors, circuitry which
compensates for detector capacitance, or using detectors with internal
gain mechanisms, such as APDs. APD receivers can provide 5-10 dB
improvement over PIN detectors, albeit with increased parts cost and a
more complex high voltage bias circuit. These four parameters allow links
to travel over longer distance, penetrate lower visibility fog, or both.
In addition, Free Space Optics (FSO) receivers must be designed to be
tolerant to scintillation, i.e. have rapid response to changing signal
levels and high dynamic range in the front end, so that the fluctuations
can be removed in the later stage limiting amplifier or AGC. Poorly
designed Free Space Optics (FSO) receivers may have a constant background
error rate due to scintillation, rather than perfect zero error
performance.
Fixed-Pointing or Active Tracking/Pointing?
Another element of Free Space Optics (FSO) system design that must be
considered by a prudent buyer is the challenge of maintaining sufficiently
accurate pointing stability. A number of Free Space Optics (FSO) systems
employ an active pointing-stabilization approach, which represents an
effective approach for addressing this challenge. However, the cost,
complexity, and reliability issues associated with active-pointing
approach can be avoided in some applications (particularly for shorter
ranges and lower data rates) by utilizing the fixed-pointed approach
schematically shown in the figure.
According to this approach, the transmitted beam is broadened
significantly beyond its near-perfect minimum beam divergence angle, and
the receiver field of view is broadened to a comparable extent. The
broadening of the transmitted beam and receiver field of view leads to
large pointing/alignment tolerances and a very low probability of building
motion being of sufficient magnitude to take the link down. Well
engineered hardware exploits this approach of designing for loose
alignment tolerances. Therefore, it is possible to perform initial
alignment of the transceivers at opposite ends of the link during
installation and then leave them unattended for many years of reliable
service.
Practical experience shows that systems
with beamwidths less than 5 milli-radians are unreliable without tracking,
as buildings move more than that; systems above 5mrad are reliable without
tracking.
Network Protocol – Transparency is best
For carriers today the issue of interoperability of systems within
their multi-faceted networks made up of both legacy and next generation
networks is crucial. Most Free Space Optics (FSO) systems currently
available are physical layer devices that act the same way as fiber optic
cables and receivers and are therefore able to work with all protocols
while not being limited to any of them. Most designers of Free Space
Optics (FSO) systems have opted for a protocol ‘transparent’ approach for
both deployment flexibility and cost-reduction. Should a customer wish to
add such switching functionality to networks incorporating physical layer
products there are many switches available on the market, all of which
will interoperate with a physical layer device.
Reliability in FSO products
Every customer wants to know the expected failure rate of the equipment
they are investing in, for outdoor or industrial applications the
ruggedness of a system becomes even more important. A system can be
engineered and designed for exceptional reliability.
Engineering a product for long-life includes selecting top-quality,
long-life components from reliable vendors. Telecom grade components are
preferred, as are low-stress electronics. The system must also be designed
to maintain an optimum operating environment for the selected components
and sub-systems. A rugged, environmentally-sealed housing is the first
defence of a system against the elements. Appropriate heating and cooling
mechanisms should be also in place in order to maintain optimum
temperature and humidity within the device. In addition, a system design
that incorporates a mechanism for reducing laser power during clear
weather will extend the life of the laser drivers and the product itself.
Active cooling of each laser will further enhance the lifespan of these
relatively expensive sub-systems. If these considerations are taken into
account, the system should have an impressive MTBF (mean time before
failure).
Performance - Transmit Power & Receiver Sensitivity
Free Space Optics (FSO) products performance can be characterized by four
main parameters (for a given data rate):
- Total transmitted power
- Transmitting beamwidth
- Receiving optics collecting area
- Receiver sensitivity
A figure of merit (FOM) can be used to compare competing systems, based on
the basic physics of this equation:
Figure of Merit = (Power*Diameter2)/(Divergence2*Sensitivity); where
Power = Laser power in milliwatts
Diameter = effective diameter in cm (excluding any obscuration losses)
Divergence = beam divergence in millirad
Sensitivity = receiver sensitivity in nanowatts
High transmitted power may be achieved by using erbium doped fibre
amplifiers, or by non-coherently combining multiple lower cost
semiconductor lasers. Narrow transmitting beamwidth (a.k.a. high antenna
gain) can be achieved on a limited basis for fixed-pointed units, with the
minimum beamwidth large enough to accommodate building sway and wind
loading. Much narrower beams can be achieved with an actively pointed
system, which includes an angle tracker and fast steering mirror (or
gimbal). Ideally the angle tracker operates on the communication beam, so
no separate tracking beacon is required. Larger receiving optics captures
a larger fraction of the total transmitted power, up to terminal cost,
volume and weight limitations. And high receiver sensitivity can be
achieved by using small, low-capacitance photodetectors, circuitry which
compensates for detector capacitance, or using detectors with internal
gain mechanisms, such as APDs. APD receivers can provide 5-10 dB
improvement over PIN detectors, albeit with increased parts cost and a
more complex high voltage bias circuit. These four parameters allow links
to travel over longer distance, penetrate lower visibility fog, or both.
In addition, Free Space Optics (FSO) receivers must be designed to be
tolerant to scintillation, i.e. have rapid response to changing signal
levels and high dynamic range in the front end, so that the fluctuations
can be removed in the later stage limiting amplifier or AGC. Poorly
designed Free Space Optics (FSO) receivers may have a constant background
error rate due to scintillation, rather than perfect zero error
performance.
Qualification & Testing
All necessary certifications for Laser Safety (CFR, ANSI, IEC), Electrical
Safety (CSA, UL, EN) and Electromagnetic Compatibility (EMC) must be met.
However, many additional tests may be performed to ensure excellent
performance in all conditions. Laser eye safety is classified by the
International Electrotechnical Commission (IEC), which is the
international standards body for all fields of electrotechnology. While
the IEC is an advisory agency, its guidelines are adopted by the
regulatory agencies in most of the world’s countries.
A laser transmitter which is safe when viewed by the eye is designated IEC
Class 1M. Extensive testing should be performed to properly qualify the
system for highly reliable performance in all outdoor environments.
Testing can include, rain tests wind-driven rain tests, pressure tests,
dew point checks, vibration tests, wind tests both for survivability and
pointing stability in high-winds and underwater immersion tests to ensure
seal integrity against water leakage. The testing can also include
humidity tests and salt-corrosion tests to ensure the unit is suitable for
marine environments. All tests, when possible, should be conducted with
the units in operation so that functionality can be measured throughout
the test rather than simply assuring the survivability of the unit. In
addition to the environmental tests, accelerated life tests can be
conducted to ensure a long operating life. In addition, critical
sub-systems can be tested separately in even more extreme temperatures. In
addition, testing should be done to certify the system to various
international electromagnetic standards for products. These tests include
radiated emissions, conducted emissions, and also radiated susceptibility
and conducted susceptibility.
The emissions tests are required to ensure the Free Space Optics (FSO)
terminal and its power supply will not interfere with other systems. The
susceptibility tests go one step further, and help ensure that the Free
Space Optics (FSO) equipment will not be affected by adverse
electromagnetic environments from nearby equipment, such as rooftop air
conditioning systems, microwave and cellular transmitters, etc. As well,
tests for harmonic current emissions, and immunity to intense magnetic
fields, voltage fluctuations and interruptions, electrostatic discharge,
and lightning surge impulses can also be included.
Field Tests & Availability
The final proof of the viability of any broadband access approach,
including optical wireless, is the successful conclusion of rigorous
field-tests. Ideally, such field tests should include operation 24 hours
per day, 7 days per week. The most convincing tests are those in which
weather conditions vary widely during the tests, and include periods of
steady drizzle, heavy driving rain, snow, and various degrees of fog.
Independent Tests
Another way to assess the field-worthiness of Free Space Optics (FSO)
hardware is to consult an independent assessment of the performance. These
objective assessors will test a device in a number of ways. Some are
focused on performance and ease of use, they will try to install a unit
direct from the box, evaluating manuals and procedures along the way. They
can also attempt to impede transmission with various ‘obstacles’ through
which the communications must pass. As well, tests that rate the
reliability of a device in MTBF (mean time before failure) can be
performed. Some of the most stringent standards in the industry are
maintained by Telcordia/Bellcore to establish guidelines for carrier-class
products. In general independent tests validate manufacturer’s claims of
reliability and performance.
Installation
Free Space Optics (FSO) installation hardware should be designed both for
ease of setup and alignment, as well as providing a stable, rugged mount
which maintains alignment under environmentally stressing conditions and
aging. For example, a yoke assembly mounts on a vertical pole, and allows
the optical head to be easily dropped into place. Coarse angular
adjustment can be provided using rifle sighting scopes. The signal can
then be maximized by viewing on a personal computer display (or
voltmeter), and adjusting push-pull screws, which offer both fine
alignment and lock-down capability. The unit should be designed to
maintain its initial alignment over its operational life – for
fixed-pointed systems the beam should be broad enough to withstand
building motion on tall skyscrapers and to operate in high-winds; for
active-pointed systems the system should be sensitive enough to respond to
sudden events such as high-winds or earthquakes.
Network Management & Monitoring
For modern carrier networks, a network management and monitoring
capability with an easy-to-use graphical interface is essential. Many Free
Space Optics (FSO) products are physical layer devices, i.e. completely
analogous to fiber optic cables and transceivers, so the network interface
that monitors the active devices is easily provided over a separate CAT5
cable or RS-232 port, and is preferably IP-addressable. Some carriers
prefer the network interface to be combined with the optical data stream,
either with a separate channel, or using inter-element data bits in a
SONET data stream. Sending SNMP over a SONET data communication channel
requires additional Layer 3 equipment, either built into the Free Space
Optics (FSO) terminal, or in the equipment room. However it is provided,
the SNMP interface should allow the monitoring of the status of active
components health with enough fidelity to predict problems before they
become a system outage.
Useful status indicators include:
received signal strength
transmitter power settings (bias, modulation currents) for each, laser
temperature of each of the lasers
temperature of key interior locations
interior humidity
four power supply voltages and currents
TE cooler controller currents
clock recovery status
network signal status, and
extensive historical logging capability.
Product mix – one size does not fit all
The Free Space Optics (FSO) market segments along three primary
characteristics: cost, data rate, and link length. Because of the cost
sensitivity, it is imprudent to attempt to design one product to cover the
entire market space, since that will generate a high-priced product that
is more expensive than a customer can justify for the large fraction of
the Free Space Optics (FSO) market that comprises the less-challenging
applications. In order to provide the customer with the greatest value,
then, it is essential to offer a product mix that recognizes “one size
does NOT fit all.” Each of several key application areas demands a
complementary set of features that just meets the demands of that
application area without adding additional costs associated with
non-essential performance features.
Production capacity and Volume
For carriers, volume manufacturing is a key requirement, as mainstream
deployment in a worldwide carrier network includes thousands of network
elements. A high volume production facility for Free Space Optics (FSO)
should ideally be designed specifically to produce Free Space Optics (FSO)
products, as Free Space Optics (FSO) is unique from any other fixed
wireless or fiber optic products. Layout and process flow should be
optimized for product families, for example fixed-pointed and
actively-pointed products may be better suited toward separate production
lines. Procurement processes should be compatible with just-in-time parts
delivery, which requires coordination and good working relationships with
all major suppliers. The capability to operate on double (or even triple)
shifts provides maximum return on fixed assets, as well as flexibility to
adjust to changing market conditions.
Cost - Is the Price Right
While cost is always a consideration when procuring telecom products, many
buyers are interested in obtaining the best value proposition in the
medium to low cost range. For example, higher performance, with little
extra cost penalty, often provides the best value. The key factors that
affect cost are system design (i.e. choice of components and their
configuration), minimization of manual labor (especially for optical
alignment), and volume manufacturing to reduce procurement costs and
amortize non-recurring costs. When this philosophy is consistently applied
to product design, the overall result can be very high incremental
performance with low incremental cost.
References - Sites and Customers
For many prospective buyers an appropriate reference builds the most
confidence in a company and their products. Many manufacturers offer case
studies, demo sites or reference customers that can be contacted. Some
Free Space Optics (FSO) companies have a worldwide installed base in the
hundreds of units. Case studies that represent carrier-class,
mission-critical links will be of most value to carriers wishing to deploy
the system. A solid field-testing program can determine how well the
devices will function in real-world deployments. See FSO @ Work around the
globe ».
Conclusions
The potential for Free-space optical networking to solve communications
bottlenecks is making it a popular option for reliable, broadband access.
A thorough examination of the issues affecting the design of these
sophisticated systems is a useful tool when evaluating Free Space Optics
(FSO) systems for purchase. Systems that incorporate the most beneficial
features, are well-engineered, and thoroughly tested will be top
performers and provide the best value.
To find out more, please visit:
FSO Primer
FSO Technology
FSO Guide
FSO Comparisons
FSO Laser Guide
FSO Products
FSO Applications
FSO Glossary
FSO FAQ
FSO Technology FAQ
FSO Metro App
|
Copyright
2000 CableFree Solutions Limited
Free Space Optics (FSO), Optical Wireless, Infrared
Fixed Wireless Access,
Wireless
Broadband, Laser
Download
Adobe's Acrobat Reader to view pdf files
 |
|
|
