PowerPoint Presentation by 3Rl5gr4


									GE MDS Training

Radio Topics:
Antennas, Interference,
System Design, and More
Overview                              Grounding and Installation
Evaluating Radio Paths                Understanding Interference
• Availability
                                      • Interference types
• What is line of sight (LOS)?
• Fresnel zone clearance              • Filters
• Path loss                           • Other forms of mitigation
• Gains, losses, and link budgeting      – Antenna placement
Feed Lines                               – Antenna polarization
System Design                            – 9810 and TransNET radio
• RF Isolation                              features
• Frequency separation
                                         – iNET and entraNET radio
• Antenna Types
    – Antenna selection for a given
      application                        – LEDR radio features

                                                       GE MDS Training - March 8-10, 2005
This training module covers the RF Performance, Path Planning, System
Design and Interference requirement for all three MDS Technical
• Each individual exam draws at least six questions from this pool
• Each question has a point value based on difficulty
• Questions cover the entire spectrum of MDS products and applications,
   but the questions are within the context of the exam you are taking
    – You will not see, for example, Wideband questions on the Networking
    – You may see generic questions regarding one product line on another
        product line’s exam, if the concept is applicable to both product lines
         – Example: Point-to-point antenna selection

                                                               GE MDS Training - March 8-10, 2005
Evaluating Radio Paths
Evaluating Radio Paths
The fundamental goal of path evaluation is meeting the
customer’s communication needs
• Although this sounds obvious, it begs a number of questions
   – What is the application?
      – Is it point-to-point or point-multipoint?
      – Do you need continuous data throughput, or is it
   – How critical are data errors to the end user?
   – Is this a primary or back-up communication system?
The bottom line is availability
Availability is the top-level measurement of whether a radio
path meets a customer’s needs

                                                   GE MDS Training - March 8-10, 2005
Availability is defined as the amount of time during which a given system (in
our case, a radio link) is able to be used for its intended purpose
Availability is measured in percent
• Example: 99.9% availability over a one-year period is:
(86400*7*52) – 0.999 (86400*7*52) = 31449.6 seconds, or 8 hours, 44
• Similarly, 99.999% availability = 314.5 seconds, or 5 minutes, 14 seconds
Availability for links that are in series is the product of availability of each
individual link
• Example 1: 99.999% x 99.999% = 99.998% = 10 minutes, 29
• Example 2: 99.999% x 99.9% = 99.899% = 8 hours, 49 minutes/year
• The key here is that the composite availability is always worse than any of
   the individual links that make up a system
    – The corollary is that each link must have higher availability than the
       composite system in a multiple-hop application

                                                                GE MDS Training - March 8-10, 2005
Availability (2)
Availability derives from four things:
• Frequency
• Path length (which drives path loss)
• Fade margin
• Terrain and climate factors (relatively minor influences)
As a result, fade margin is not an adequate measurement; it is a first-order
approximation only!
• Fade margin is one of the factors in determining path availability
    – For constrained applications, such as 900-MHz point-multipoint
       systems with path lengths up to, say, 20 miles (32 km), fade margin is
       adequate as a system design goal. MDS recommends a 20-dB
       minimum fade margin for these applications.
    – For longer paths, higher frequencies, and more critical applications,
       understanding the path dynamics requires calculating the full impact
       of fading on availability

                                                              GE MDS Training - March 8-10, 2005
Availability (3)

For interest, calculating free-space path loss:
• αfs = 92.4 + 20 (log f) + 20 (log d)
  αfs = free-space path loss in dB
  d = path length in km
  f = frequency in GHz

                                       GE MDS Training - March 8-10, 2005
What is “Line of Sight”?
Optically, line-of-sight (LOS) is how far you can see
• Also known as the visual horizon
Radio line-of-sight is longer than optical LOS
• Radio signals bend slightly in the troposphere, the lowest
  level of the earth’s atmosphere, which extends radio range
  compared to optical LOS
• The result is that radio LOS is not the same as optical LOS—it
  is longer by about 1/3 as a rule
   – This rule is not firm, but is a good number for planning
Radio LOS is not the only factor that drives free-space path loss

                                                        GE MDS Training - March 8-10, 2005
Fresnel Zone Clearance
Radio signals on an optical-LOS path are altered by objects that
infringe on the first Fresnel zone
• We look to have 60% of the first Fresnel zone unobstructed
    – This is also known as 0.6 first Fresnel zone clearance
• As long as the first Fresnel zone clearance is positive, that
   means that the radio path has (at least) optical LOS
• If there are obstructions within 60% of the first Fresnel zone,
   radio path loss increases as much as 12 dB (at the point
   when you have 0 first Fresnel zone clearance, at which point
   the radio path is at the optical LOS)
To understand the Fresnel zone concept, it helps to think of
antennas as flashlights and compare the flashlight analogy to
RF behavior

                                                    GE MDS Training - March 8-10, 2005
Fresnel Zone Regions

      F4   F2     F1

           d1                                d2

            Fn = 17.3√(n*d1*d2)/fD)
            where: d1, d2 in km, f in GHz, D in km

                                                     GE MDS Training - March 8-10, 2005
f4   f2      f1

      f4 = frequency at 450 MHz
      f2 = frequency at 900 MHz
      f1 = frequency at 1.4 GHz
                                  GE MDS Training - March 8-10, 2005
Fresnel Zone and Path Length




                               GE MDS Training - March 8-10, 2005
 Other Obstructions
Obstructions such as trees and buildings are very, very bad at 900 MHz and
higher frequencies
• The worst case is an obstruction that’s near either end of the path,
   because these obstructions block the largest portion of the RF pattern
    – Think back to the flashlight analogy
• Obstructions that are near the center of the path have the smallest
   impact on communications
• Anything solid is bad—houses, buildings, etc
• Anything with an open structure, such as towers, scaffolding, power
   distribution towers, and similar structures, usually have no measurable
   impact at 900 MHz and up
• Caveat: Keep the near field clear at all times
• Advice: Make sure that antennas clear any obstructions that are within
   100 meters of the antenna location, especially trees and buildings that
   encroach on the antenna’s path

                                                             GE MDS Training - March 8-10, 2005
Path Loss
Path loss is calculated based on frequency and distance, as
illustrated earlier
Path loss goes up by:
• 6 dB every time you double the frequency
• 20 dB each time you multiply the distance by 10
Path loss over any real path will always be near, or exceed, 100
• At 900 MHz, a 1-km path has 91.5 dB free-space loss
• At 900 MHz, a 10-km path has 111.5 dB free-space loss
   – 20 dB more than the free-space path loss at 1 km
• At 450 MHz, a 10-km path has 105.5 dB free-space loss
   – 6 dB less than the free-space loss at 900 MHz

                                                    GE MDS Training - March 8-10, 2005
Link Budget Calculation

Antenna                                                                 Antenna
Gain A(dBi)                                                             Gain B(dBi)
                                Path Loss (dB)
Feed line A                                                            Feed line B
(dB)                                                                   (dB)

TX (dBm)                                                                   RX (dBm)

      RX (dBm) = TX - Feed line A + Ant. Gain A - Path Loss + Ant. Gain B -
                 Feed line B

        Fade Margin (dB) = RX - RX Threshold

                                                              GE MDS Training - March 8-10, 2005
Gains, Losses and Link Budgeting
Path calculations consist of a simple algebraic sum of gains and losses,
including these items:
• Transmit power in dBm (starting point)
• Transmit-side filter loss (negative, if present)
• Transmit-end feed line loss in dB* (negative)
• Transmit-end antenna gain in dB (positive)
• Path loss (negative)
• Excess path loss as a result of inadequate first Fresnel zone clearance
• Receive-end antenna gain (positive)
• Receive-end feed-line loss* (negative)
• Receive-end filter loss (negative, if present)
• What’s left at the end of this chain is the RSS
• The difference between the RSS and the radio’s specified sensitivity is the
   fade margin
*See note on next page regarding connector loss
                                                              GE MDS Training - March 8-10, 2005
Ranking Impacts on Link Success
In order from highest to lowest, these are the key factors once
you know the required availability
• Path distance
    – It’s either short enough to work or it’s not
• Lack of local obstructions at either end of the path
• First Fresnel zone clearance
    – Significant impact on availability, especially for longer
• Lack of interference (moderate to strong)
• Antenna gain
• Tower height
• Feed line length
• Lack of interference (low-level or infrequently occurring)
                                                    GE MDS Training - March 8-10, 2005
Dealing with Long Links
Sometimes paths are just too long for good availability; here are
some ways to address that situation, from lowest cost to
• Live with a low fade margin if performance is acceptable
• Use the highest-gain antennas that the supports can
   accommodate ($)
    – Not always practical in spread-spectrum systems
• Lease space on a taller tower to get more height ($R)
    – Link to the nearby tower via radio to limit recurring cost
• Use a leased line to a site that has a better path to the far
   end ($$R)
• Find a repeater location, install a tower, and link through it
   via radio ($$$$)
    – No recurring cost, but very high initial cost
                                                    GE MDS Training - March 8-10, 2005
Feed Lines
Feed Lines
Feed line loss is published by each feed line manufacturer
• Often specified in dB per 100 feet or dB per 100 meters—be careful to use the
   same units of measurement when determining actual losses
Choose an appropriate feed line for the length of your cable, frequency of
operation, and application
• Master radios and point-to-point applications require better cables
    – The cable runs tend to be long, and these sites are critical to system
        performance—don’t skimp here if you can avoid it
• Small, inexpensive cable is often suitable for remote sites even if the cable run
   is relatively long
    – Base your decision on the required availability and the path length
• Watch your budget—feed lines and connectors are very expensive items!
Connector Loss
• Properly installed connectors have practically no loss and do not need to be
   taken into account in path calculations. In reality, properly installed
   connectors have losses in the range of 0.01 dB
• The same caveat applies to lightning protection devices
• Install connectors with care!

                                                                  GE MDS Training - March 8-10, 2005
System Design
 System Design and RF Isolation
Focus on RF topics
• Transmitter separation—frequency
   – Always separate transmitters at adjacent sites by at least two
      channels (preferably four channels)
   – Always configure the transmitters either low or high (in frequency) on
      both links at a given site
       – This keeps the frequencies adjacent, which provides the maximum
          duplexer and receiver rejection of co-located systems
• Antenna separation—distance
   – Choose spacings that provide at least 60 dB (more is better) of
      isolation between co-located radios
       – This requires about 3 meters at 900 MHz
   – Use vertical separation for vertically polarized antennas and
      horizontal separation for horizontally polarized antennas
       – Horizontal separation for vertically polarized antennas is the worst
          case—never do this!

                                                              GE MDS Training - March 8-10, 2005
System Design and RF Isolation (2)
• Polarization diversity—for interference rejection
   – Alternating polarization between adjacent links results in 20-25 dB of
      RF isolation
• For spread-spectrum systems—frequency separation
   – Use as little overlapping spectrum as possible at a single site
   – Take advantage of radio features to maximize radio performance in
      an interference-prone environment
       – These features are covered later for each product
• For RF isolation in any system, use the least transmitter power that
  provides the RF performance your customer expects
   – System design is much easier when the power levels are reasonable;
      less isolation is required for satisfactory operation

                                                            GE MDS Training - March 8-10, 2005
System Design and RF Isolation (3)
FCC Part 15 regulations
• In the US, the FCC limits radiated power for devices at operate in the 902-
  928 MHz band using spread-spectrum technologies
   – Examples: MDS 9810, entraNET 900, iNET 900
• EIRP is capped at 36 dBm, or 4 W
• EIRP is the algebraic sum of
   – The radio’s power output
   – All the gains and losses in the transmitter’s RF path, including the
• Example with a 1-W radio:
   – 30 dBm transmitter – 2 dB feed line loss + 6 dBi antenna gain gives 34
      dBm EIRP
• Be very comfortable calculating EIRP
   – You may have to convert between watts and dBm, so refresh yourself
      on that as well

                                                              GE MDS Training - March 8-10, 2005
Watts-dBm Conversion

Converting between dBm and EIRP is simple if you
remember four simple rules:
1) 30 dBm is 1 W
2) 3 dB is a halving/doubling of power
3) 10 dB is a ten-times increase/decrease in power
4) 1 dB is about a 20% change in power
 Based on these rules, you can move in any increment
  you choose starting from 1 W output

                                          GE MDS Training - March 8-10, 2005
Antenna Types
Choose antennas based on your application
• Point-multipoint
• Point-to-point
• Sector or quadrant
• Multipoint-to-multipoint
• High-gain omnidirectional (omni)
• Low-gain omni
• High-gain directional (Paraflector, dish)
• Low-gain directional (Yagi)
• Medium-gain, semi-directional (panel or quadrant)
• Special purpose—horizontally polarized omni

                                                      GE MDS Training - March 8-10, 2005
Selecting Antennas
Typical applications
• Licensed PTP link: dishes or Paraflectors at each end
    – Sometimes a dish at one end and a smaller antenna at the other end
       is used to reduce wind load at one site
    – Be sure to comply with your country’s regulatory requirements
• Typical MAS system: High-gain omni at master; Yagis at remotes
    – In interference situations, a horizontally polarized omni can be used,
       and the Yagis can be mounted for horizontal polarization
• Typical spread-spectrum systems:
    – PTP links—Yagis at both ends
        – Higher-gain antennas are possible, but controlling power is
    – MAS—omni at master site and Yagis at remotes

                                                              GE MDS Training - March 8-10, 2005
Selecting Antennas (2)
• Point-multipoint, peer-to-peer systems
   – Small omnis at each location, or small Yagis (preferred)
   – Select the antenna with the narrowest beamwidth that provides
     acceptable system operation; this prevents interference and provides
     more signal strength
      – A 3-element Yagi offers about 6 dBi gain, which is comparable to a
         1.5-meter-long omni, but provides a much better pattern and
         about 100 degrees of beamwidth
• Any appropriate application except PTP: In interference situations, panel
  antennas are often used if limited RF coverage is needed
   – Typical panel antenna beamwidths: 90° to 140°
   – Panel antennas are relatively expensive
   – Stacked (phased) Yagi antenna arrays provide a less expensive

                                                            GE MDS Training - March 8-10, 2005
Grounding and Installation
Grounding and Installation
Grounding—follow the manufacturer’s instructions
• Surge suppressors, antennas, and feed lines all have recommended
   grounding practices—follow them
• Local electrical codes drive installation practices
    – These take precedence over manufacturer instructions
• Towers and buildings should have their own ground systems
• The NEC (in the US) requires that these are bonded together
    – Consult the NEC or a licensed electrician for details
• Have an engineer or a licensed electrician design and install or supervise
   the installation of your sites to ensure that grounding practices are
    – Damage from improper grounding and lightning strikes are not
       covered under MDS or other vendors’ warranties

                                                             GE MDS Training - March 8-10, 2005
Understanding Interference
Interference comes in many forms
• Self-interference
• Intermodulation interference
• Harmonics
• Fundamental overload
Troubleshooting interference is a valuable skill that is not easily learned or
taught, but includes:
• Good test equipment and the knowledge to use it properly
• Experience
• Knowledge of various mitigation techniques so that you can choose the
   most appropriate one(s) to solve a particular problem

                                                               GE MDS Training - March 8-10, 2005
Understanding Interference

Tools for identifying interference in a MAS
Spectrum analyzer for identifying adjacent
channel and co-channel signals
Serial Protocol Analyzer to identify errors in
data stream
Orderwire handset to listen to On-the-Air
channel for noise and other radios

                                       GE MDS Training - March 8-10, 2005
Understanding Interference

Intermodulation Interference
Products produced by mixing of two signals,
at each sum and difference frequency (2f1-f2,
2f2-f1, etc)
Sources of intermodulation products
Rusty bolt syndrome
Mixing in transmitters

                                      GE MDS Training - March 8-10, 2005
RF Filtering
Many options exist; the most common ones:
• Band-pass filtering
   – Ideal for spread-spectrum applications and fixed-frequency
      applications with extremely aggressive or multiple-frequency
   – Expensive in terms of both cost and rack space for fixed-frequency
      situations; very inexpensive in both realms for spread spectrum
   – The only option for transceivers, since you cannot separate the
      transmit and receive paths—therefore the filter is in-line in transmit
      and receive
• Notch filtering
   – Removes the offending frequency or band of frequencies
       – Removing a band of frequencies is possible with stagger-tuned
          notch filters
   – Relatively inexpensive
   – Suitable for many fixed-frequency applications

                                                               GE MDS Training - March 8-10, 2005
When Filtering is Appropriate
Situations that call for filters:
• Strong licensed service interfering with your application
    – Frequent occurrence at 900 MHz, with paging band from 929-931 MHz
        – Paging transmitters run as much as 1.5 kW into gain antennas in
            this range
        – Can saturate front ends or cause intermodulation interference
    – Calls for notch or band-pass filtering, depending on frequency
       separation from desired signal, and on amplitude of interferer(s)
• Nearby signals in-band
    – Calls for band-pass filtering
• Unknown, transient interference
    – If it’s known to be off-channel, band-pass filters are the most certain
       way to eliminate this type of interference

                                                             GE MDS Training - March 8-10, 2005
Filter Placement

Usually filters go in the receive path, if possible
• Non-redundant radio: Between the duplexer and the
• Redundant radio: Between the duplexer and the
  receive splitter
For transceivers (band-pass filters)
• Place in-line with the radio
• Use a short jumper and place the filter between the
  radio and the surge suppressor/lightning arrestor
                                           GE MDS Training - March 8-10, 2005
Spread Spectrum Band-Pass Filter
For the 900-MHz band, MDS offers a small inline filter
• Covers 902-927 MHz
• Same footprint as iNET, entraNET AP radios;
  somewhat thinner
• N connectors for input and output
• Inexpensive
• Must skip the top frequency zone when using this
  filter—filter rolls off by at least 6 dB in top zone
   – Zone 10 in iNET
   – Zone 8 in entraNET, TransNET and 9810
                                            GE MDS Training - March 8-10, 2005
Other Methods of Interference Mitigation
Radio features
• Almost always the best choice, since they’re included in the product
• May cause performance impacts such as increased latency (delay)
• Appropriate for very strong desired signals that can still meet availability
   expectations with lower signal strength
• Most appropriate for radio systems where the transmit and receive paths
   are separate
    – Not transceivers such as x710, x810, iNET, entraNET, TransNET
    – LEDR, x790, P70/P20 (depending on application)

                                                               GE MDS Training - March 8-10, 2005
Interference Mitigation (2)
Antenna Polarization
• Almost everything in typical MDS applications is vertically polarized
• Almost every licensed service in or near the same bands MDS radios use is also
   vertically polarized
• Changing to horizontal polarization is low-hanging fruit for interference mitigation
   in many applications
    – Adjacent-channel and co-channel interference
    – Strong adjacent-band interference
    – Spectrum sharing with spread-spectrum systems
         – Example: Point-to-point spread-spectrum link feeding point-multipoint
            system—use horizontal polarization in PTP portion
• Horizontal polarization isn’t always the best solution
    – Horizontally polarized omnidirectional antennas are very expensive (at least
       three times the cost of vertically polarized Omnis)
    – Labor to convert a system to horizontal polarization can be expensive
    – In such cases, you may want to use a filter instead; consider total cost of the
       solution as well as the long-term performance impacts

                                                                    GE MDS Training - March 8-10, 2005
 Interference Mitigation (3)
Antenna Placement
• Most of this is common-sense
    – Antennas should always be separated from each other by a reasonable
      distance, and not pointed at each other
    – Omnidirectional antennas should always be vertically separated to achieve RF
      isolation—vertical alignment is important
    – Antennas on a water tank or building should be shadowed from each other
      structure to provide some RF isolation
• Think about the worst-case placements
    – Horizontally polarized antennas, vertically separated
    – Directional antennas pointed at each other
    – Omnidirectional antennas in the same vertical plane (horizontally separated)
    – Avoid these situations or you invite problems!
Attenuation and Reducing Power
• Both methods can be used to improve co-location interference

                                                                  GE MDS Training - March 8-10, 2005
Radio-Specific Features
The following pages cover the use of MDS radio features to combat
interference and improve system performance in interference situations
• System performance impacts are also discussed
• The list is not exhaustive with regard to implementation options, but
   covers all of the options available for each product to help resolve and
   prevent interference issues
• Radio features can be used in combination with:
    – External filtering
    – Antenna selection
    – Antenna polarization
    – Other creative ways of optimizing system performance
    – No single cure or single approach works for all situations

                                                               GE MDS Training - March 8-10, 2005
Interference Mitigation: 9810
Send count
• Increases the number of times each data packet is transmitted over the
• Increases latency dramatically, but has a similar improvement in data
• The shorter the hoptime, the less likely interference is to affect any given
• Increases latency somewhat
Zone Skipping
• First choice
• Can skip all but one zone; allows for maximum co-location and RF
  isolation between systems

                                                               GE MDS Training - March 8-10, 2005
Interference Mitigation: TransNET
Repeat Count
• Sets the number of times each packet is transmitted, like send count in 9810
• Increases latency significantly
Retry Count
• Determines how many times an errored packet is retransmitted over the air
• Efficient; doesn’t retransmit unless radio detects an error
• Impacts latency only in high-error situations
Forward Error Correction (FEC)
• Eliminates residual errors; can significantly improve interference if not severe
• Significant throughput decrease
Dwell Time
• Shorter dwell times decrease probability of interference affecting a given packet
• Increases latency somewhat
Zone Skipping
• Almost always the first option to choose—no performance impact
• Can skip up to four zones (half the band)
                                                                   GE MDS Training - March 8-10, 2005
 Interference Mitigation: entraNET
Broadcast Repeat Count
• Significantly decreases throughput; retransmits every broadcast and
   multicast packet
Unicast Retry Count
• Increasing this can eliminate most non-severe interference effects
• Minimal negative impact on system performance unless channel is very
   noisy or has significant interference
Dwell Time
• Shorter dwell times decrease probability of interference affecting a given
• 7 ms provides best interference immunity, but least throughput and
  longest latency
Zone Skipping
• Almost always the first option to choose—no performance impact
• Can skip up to four zones (half the band) to avoid or prevent interference
                                                              GE MDS Training - March 8-10, 2005
Interference Mitigation: iNET
Zone Skipping
• Can skip up to three zones (30% of the band)
• Almost always the first option to choose—no
  performance impact
Switch from 512 to 256 kbps
• Significantly more robust modem; half the
Enable compression
• Send packets in fewer hops; reduced chance of

                                           GE MDS Training - March 8-10, 2005
Interference Mitigation: iNET (2)
Dwell Time
• Shorter is better for interference rejection, but
  latency goes up and throughput goes down with
  shorter dwell times
• Dwell time is packet-length-dependent
   – If you are sending large packets, the dwell time
     stretches to accommodate them without
     fragmentation, so the benefit of dwell time is less
     in an iNET system than in other MDS radios
Retry count is fixed in iNET
• Unlike other radios, this isn’t adjustable

                                               GE MDS Training - March 8-10, 2005
Interference Mitigation: LEDR
Increase Interleave
• Increases latency, approximately linearly (doubling the
   Interleave doubles the latency)
• Latency is a function of:
    – Data rate (inversely proportional)
    – Interleave setting (directly proportional)
All other interference effects are handled by fixed features
• Adaptive equalizer
-OR- require external filters to eliminate

                                                    GE MDS Training - March 8-10, 2005

To top