T013

Microwave Link Planning

1 Copyright 2010 AIRCOM International
 Session 1
Introduction

2 Copyright 2010 AIRCOM International

 Course Schedule
 The day will consist of six, 45 minute sessions.
 A module may take more than one session to
complete.
 Self-assessment questions and practical examples
are included to reinforce understanding.

3 Copyright 2010 AIRCOM International

 Aims of Course
 To enable you to plan the radio elements of a point to point microwave link
against a performance requirement and to be able to predict the performance
of the link that you have planned.
 This will involve gaining an understanding of

 Antennas  Interference
 Link Budgets  Radio Propagation
 Noise  Modulation Methods
 Fading  Performance Prediction
 Diversity Techniques Methods.

4 Copyright 2010 AIRCOM International

 Why Microwave
Microwave radio links provide high speed (2 Mbps+) communication between
two points.
They are known to be:

 fast to implement
 convenient
 economic

when compared with wire-based alternatives.

5 Copyright 2010 AIRCOM International

 What does “Microwave” mean
Microwave refers to a section of the RF spectrum lying between 3 and 30 GHz. It
is also referred to as “Super High Frequency” (SHF).

300 kHz 30 MHz 3 GHz 300 GHz

LF MF HF VHF UHF SHF mm

30 kHz 3 MHz 300 MHz 30 GHz

The Microwave Band

Note that frequencies up to 40 GHz are being used for “microwave” links
although the definition suggests that this frequency is in the “millimetric”
band.

6 Copyright 2010 AIRCOM International

 System Block Diagram

Antenna

Feeder

Transceiver

The equipment layout is essentially very simple. The job of the link planner
is to specify and configure the equipment.

7 Copyright 2010 AIRCOM International

 Answers, Please!
What power
level will we
How big must receive?
the antenna
be?
At what data How high must
rate must we What will the
loss of the the antenna
send? be?
feeder be?

What should
What the transmit
frequency will power be?
we use?

How good will the

performance be?

8 Copyright 2010 AIRCOM International

 Percentage Outage

 Unlike an optical fibre or coaxial cable system, the received power level of a
microwave system will vary significantly with time.
 This is due to atmospheric effects and “hydrometeors” such as rain and snow.
 This will inevitably lead to the system suffering an outage for a small
percentage of the time.
 The link planner must be able to predict the outage periods as a percentage
on a particular system.

9 Copyright 2010 AIRCOM International

 Quantitative Analysis

 The link planner must be able to determine numerical parameters to define the
microwave system.
 The course will involve methodologies, procedures and techniques for arriving
at the correct numerical solutions.
 However, all solutions should fit in with the expectations of an intuitive
engineer.

10 Copyright 2010 AIRCOM International

 Intuitive Expectations

 If the antenna is bigger, the receive power will increase.

 If the link is longer the receive power will decrease.
 We will need a higher power to transmit a higher data rate.
 The higher the power received, the lower the percentage outage.
 The longer the feeder, the lower the receive power.

11 Copyright 2010 AIRCOM International

 Next Steps

 All the parameters affect each other in an interactive way.

 The next sections will deal with particular parameters whilst keeping one eye
on the final goal
 In the next section we shall concentrate on the antenna and methods of
predicting the receive signal power.

12 Copyright 2010 AIRCOM International

 Session 2
Antennas and The Link Budget

13 Copyright 2010 AIRCOM International

 The Microwave Antenna

 Parabolic antennas are a form of

“aperture” antenna.
 The antenna faces an incoming
electromagnetic wave that has a power
density Pd.
 The antenna converts this to a received
power Pr.

14 Copyright 2010 AIRCOM International

 The Microwave Antenna

 The “aperture” can be thought of as a E

hole through which energy passes.
 This energy is delivered to the antenna
output.. Pd

Pr  Pd Ae
Pr

15 Copyright 2010 AIRCOM International

 The Microwave Antenna

 The “effective aperture” is linked to the

physical aperture.
 For an antenna presenting a circular
cross section of diameter D when viewed
from the front D

D 2
Ae  0.6 
4

16 Copyright 2010 AIRCOM International

 The Isotropic Antenna

 A hypothetical antenna that distributes its

transmitted power equally in all
directions.
r
 As the surface area of a sphere radius r
is 4r 2 the power density produced at
a distance r is given by

Pt
Pd 
4r 2

17 Copyright 2010 AIRCOM International

 Antenna Gain

 A practical microwave antenna will

produce a higher power density by
concentrating the energy into a narrow r
beam.
 For an antenna of gain Gt, the power
density produced is, by definition

Pt Gt
Pd 
4r 2

18 Copyright 2010 AIRCOM International

 Calculating the received power
Pt Gt
Pd  Pr  Pd Ae
4r 2

Pt Gt
Pr  A
4r 2 e

 This equation allows us to calculate the

received power given the other
parameters.

19 Copyright 2010 AIRCOM International

 Calculating the received power
 Example.

Gt Ae
r
Pt
Pr

Gt  500
Ae  2 m 2
r  20000 m
Pt  0.5 watts
0.5  500  2 8
Pr   9. 95  10 watts
4 (20000) 2

20 Copyright 2010 AIRCOM International

 Antenna Characteristics
 Radiation pattern, gain, and antenna properties in general have
the same same characteristics whether the antenna is being
used as a transmitter or receiver.
 Considering the antenna as a receiver. The gain equals its
effective aperture as a multiple of the effective aperture of an
isotropic antenna.
 Aperture of isotropic antenna 4
2 G  Ae

4 2
2
Ae  G
4
 4 
G  10 log10  Ae 2  dBi
  

21 Copyright 2010 AIRCOM International

 Calculating Received Power
 Substituting Effective Aperture in terms of Gain.

Pt Gt Pt Gt Gr 2
Pr  A 
4r 2 e
4r 2 4
 
2

 Pt Gt Gr  
 4r 

Pr  Pt  Gt  Gr  20 log 10 (4 )  20 log 10 (r )  20 log 10 ( )

22 Copyright 2010 AIRCOM International

 Calculating Received Power
Pr  Pt  Gt  Gr  20 log 10 (4 )  20 log 10 (r )  20 log 10 ( )

 Changing units from metres to kilometres and from wavelength

in metres to frequency in Gigahertz:

Pr  Pt  Gt  Gr  20 log10 (4 )  20 log 10 (1000d )  20 log 10 (0.3 / f )

Pr  Pt  Gt  Gr  92.4  20 log 10 ( f )  20 log 10 (d )

Pr  Pt  Gt  Gr  FSL
FSL  92.4  20 log10  f (GHz )  20 log10 d (km)

23 Copyright 2010 AIRCOM International

 Linking Gain and Beamwidth
 A practical microwave antenna will produce
a gain by concentrating the energy into a
r 2 2 4r 2
narrow beam.

 For an antenna of gain Gt, the area 4 Gt
illuminated will be reduced compared with
that illuminated by an isotropic antenna by Gt  16
a factor equal to its gain. 2
 4
Gt
r
  230 degrees
 r Gt

24 Copyright 2010 AIRCOM International

 Linking Gain and Antenna Diameter
4 D 2
G  Ae Ae 
 2
4
D  Df 
2 2

G  
    0.3 
G  20.4  20 log 10 D  20 log 10 f
 The above equation ignores inefficiencies in the antenna
system. A more realistic equation is

G  17.5  20 log 10 D  20 log 10 f

25 Copyright 2010 AIRCOM International

 Linking Antenna Diameter and Beamwidth
 From the previous slides:

Df 
2

G 
 0.3 
230

G
230  0.3
 
Df
Df  22
• Diameter (metres) x frequency (GHz) x Beamwidth (degrees)  22

26 Copyright 2010 AIRCOM International

 EIRP
 A commonly used term - “Equivalent Isotropic Radiated Power”.

EIRP(dBm)  Pt (dBm)  Gt (dBi)

EIRP( watts )  Pt ( watts )  Gt (ratio )

27 Copyright 2010 AIRCOM International

 Feeders and Combiners
 Co-axial cable is not suitable at frequencies above about 3 GHz.
 A hollow metal tube known as “waveguide” is used over the frequency range
3 - 30 GHz.
 The size of the waveguide depends on the frequency being used and
typically has a width of 0.7.
 Usually made of copper or brass with a copper plating inner coating.
Occasionally silver plated.
 Losses typically 0.1 dB per metre. The higher the frequency, the higher the
loss.

28 Copyright 2010 AIRCOM International

 Feeders and Combiners
 A diplexer is a sophisticated device that makes it possible to transmit and
receive from the same antenna.
 The received signal is sometimes the combination of two antennas.
 Combiners and splitters have an insertion loss that must be considered when
predicting the received signal level.
 All miscellaneous losses must be considered.

29 Copyright 2010 AIRCOM International

 The Link Budget

 The Link Budget is usually of the form of a table that ensures no sources of
losses or gains are forgotten.
 Expressing all powers, losses and gains in dB, dBi, dBm etc. Allows us to
simply add or subtract the relevant amounts.
 The simplified link budget equation is given below. Each element would be
arrived at by considering its constituent parts.
Received Power = EIRP - FSL + Rx antenna gain - Misc Losses

30 Copyright 2010 AIRCOM International

 The Link Budget

 Being able to determine the received power level is a significant

achievement.
 However, the question “Is this power level sufficient?” must be answered.
 To be able to answer this question requires an understanding of system
noise.

31 Copyright 2010 AIRCOM International

 Session 3
Noise Considerations
32 Copyright 2010 AIRCOM International
 Noise Considerations

 Thermal Noise forms the fundamental limitation of any

telecommunications system.
 The level of thermal noise is directly proportional to bandwidth and
absolute temperature.

Noise Power  kTB watts

 k is Boltzmann’s constant and equals 1.38x10-23 joules/kelvin.

33 Copyright 2010 AIRCOM International

 Noise Considerations

 An antenna can be thought of as a noise gathering device.

 The figure for “absolute temperature” depends on where the antenna
is looking.
 For terrestrial systems, “normal” values such as 290 K are suitable.
 For high quality satellite systems, values of T as low as 40 K are
achievable.
 Cheaper systems (e.g. Sky TV) have values of 160 K.

34 Copyright 2010 AIRCOM International

 Receiver Noise Figure and Noise
Temperature
 No radio receiver is perfect, they all add noise to the system.
 The SNR at the output of any amplifier is worse than at the input.
 This is accounted for mathematically by imagining a noise generator
at the input of the amplifier.
 This noise generator has a power output of kTeB where Te is the
noise temperature of the amplifier.

35 Copyright 2010 AIRCOM International

 Receiver Noise Figure and Noise
Temperature
kTB
G K(T+Te)BG

kTeB

 If the value of T at the input equals the “standard” temperature, To, of 290 K, then
noise at the output equals

k(To+Te)BG =kToBGF

 F is known as the noise figure of the amplifier.

36 Copyright 2010 AIRCOM International

 Receiver Noise Figure and Noise
Temperature
kToB
G k(To+Te)BG= kToBGF

kTeB
To  Te  To F
Te
F  1
To
Te  To  F  1

37 Copyright 2010 AIRCOM International

 Receiver Noise Figure and Noise
Temperature
 Using F is “convenient”. It can be expressed in dB, rather than as a
ratio.
 In dB form is represents the “amount by which the SNR gets worse”.
 However, the equation is only valid if the noise at the input equals
kToB.
 Abuse of Noise Figure is widespread. Be careful.

38 Copyright 2010 AIRCOM International

 Assessing the minimum signal level
 The error ratio experienced on a system depends on the SNR.
 We need to establish a required SNR in order to determine the minimum required
receive power (known as the receiver “threshold”).
 For example, a minimum SNR of 14 dB is required in order to deliver a BER of better
than 1x10-6. The system bandwidth is 8 MHz and the receiver noise figure is 4 dB.
 4 dB is a ratio of 2.5. Noise temperature is therefore 290(2.5-1) = 438 K.
 k(T+Te)B = 1.38x10-23(290+438)8x106 = 8.04x10-14 watts (=-101 dBm)
 SNR required of 14 dB is a ratio of 25. 25 x 8.04 x10-14 =2.0x10-12 watts or -87
dBm. This establishes the minimum signal level.
 Note: receiver manufacturers will often quote their own threshold level.

39 Copyright 2010 AIRCOM International

 Link Budget re-visited
 We have so far used the link budget to determine the received signal level.
 Establishing the minimum required signal means that we will need to adjust certain
parameters to meet this minimum.
 This may entail, for example, stipulating antenna sizes.
 Let us suppose that we have a 6 GHz system with a link length of 40 km with a
transmitter of 250 W power. The minimum received power has been determined to
be -87 dBm. Feeder losses and miscellaneous losses total 5 dB. Determine suitable
antenna sizes.

40 Copyright 2010 AIRCOM International

 Link Budget re-visited
 FSL=92.4 +20log(6) + 20 log(40) = 140 dB
 Misc Losses = 5 dB
 Tx Power = -6 dBm. Required Receive Power = -87 dBm
 Allowable losses = 81 dB.
 FSL+Misc losses = 145 dB
 Required Antenna Gains = 145 - 81 = 64 dBi.
 Each Antenna must have 32 dBi gain.
 Gain = 17.5 + 20log(diameter)+20log(frequency in GHz)
 32 = 17.5 + 15.6 +20log(diameter)
 Diameter = 0.88 m. (90 cm antennas would be suitable.)

41 Copyright 2010 AIRCOM International

 Noise Figure and Noise Temperature of
“cascaded systems”
k(T)B G1,Te1 G2,Te2

k(T+Te1)BG1 k{[(T+Te1) G1]+ Te2 } B G2

k{[(T+Te1) G1]+ Te2 } B G2 =k(T+Te)B G1 G2

Te2
T  Te1 
G1
Te2 Te3
T  Te1   .......... ...
G1 G1G2

42 Copyright 2010 AIRCOM International

 Noise Figure and Noise Temperature of
attenuators (and feeders)
k(T)B
I k(T+Te)BG

 For a matched attenuator receiving thermal noise at its input, the

noise at the output equals the noise at the input.

kTB  k T  Te BG
T  T  Te G
 Note that, for an attenuator, G will be less
T (1  G )
Te  than 1.
G

 T 1 1
G

43 Copyright 2010 AIRCOM International
 Cascaded System Example
 An antenna is connected to a receiver via a feeder of loss 2.5 dB. If
the temperature of the feeder is 290 K and the Noise Figure of the
receiver is 4 dB, determine the noise figure of the overall
combination.

• Noise Figure 4 dB. Ratio of 2.5.

• Noise temperature = 1.5x290=435 K.
• G for attenuator is 0.562. Noise temperature of
attenuator is 290(0.778)=226 K
• Overall noise temperature is 226+435/0.562=1000 K
• Overall noise figure = 1+1000/290=4.45 (6.48 dB)

44 Copyright 2010 AIRCOM International

 Cascaded System Example - Low Noise
Amplifier
 To improve the previously described situation, a Low Noise amplifier
is connected between the antenna and the feeder. This has a gain of
15 dB and a Noise Figure of 3 dB. Determine the new noise figure.

• Noise Figure 3 dB. Ratio of 2. Noise temperature = 290 K

• G for amplifier is 31.6.
• Overall noise temperature is 290+1000/31.6=321.6 K
• Overall noise figure = 1+321.6/290=2.11 (3.24 dB)

45 Copyright 2010 AIRCOM International

 Cascaded Systems - The Down-converter
 Waveguide itself becomes very lossy (~1 dB/m) as frequencies of 40
GHz are approached.
 This would lead to very poor noise performance.
 This problem is overcome by a low noise amplifier that not only
amplifies with a low noise figure but also modulates the incoming
signal with a sub-carrier that reduces the frequency to a lower value
(~1 GHz).
 Low loss coaxial cable is then used to carry the signal to the receiver.

46 Copyright 2010 AIRCOM International

 SNR Requirements
 The required Signal to Noise ratio is chiefly influenced by the modulation
scheme and the maximum permitted error ratio.
 Shannon’s and Nyquist’s Theorems provide fundamental limits.

• Shannon’s Theorem States that:

• Maximum Capacity = Bandwidth x log2(1 + SNR)
• Nyquist’s Theorem States that:
• Maximum Symbol Rate = 2 x Bandwidth

47 Copyright 2010 AIRCOM International

 SNR Requirements: Example
 Bandwidth 7 MHz, SNR 12 dB.
 Maximum Capacity = 7x106 log2(1+15.8) = 28 Mbps
 Maximum Symbol Rate = 14 Megasymbols per second
 Remember: these are theoretical maxima. It is very rare to exceed 50%
of the calculated value in practice.

48 Copyright 2010 AIRCOM International

 Symbol Rate and Bit Rate
 Binary modulation systems such as BPSK and FSK send only one bit
per symbol.
 More sophisticated modulation schemes such as 8PSK and 16QAM
have 3 and 4 bits per symbol respectively.

BPSK 8PSK 16QAM

49 Copyright 2010 AIRCOM International

 SNR requirements of different systems

Modulation C/I for BER C/I for BER

Scheme 1 x 10-3 1 x 10-6

BPSK 7 dB 10 dB

4PSK 10 dB 14 dB

8PSK 15 dB 19 dB

16PSK 21.5 dB 24.5 dB

 The variety of C/I requirements for different modulation schemes leads to
the parameter “Energy per bit” (Eb) being used as having global
relevance.

50 Copyright 2010 AIRCOM International

 Bandwidth requirements of different
systems
Modulation Bandwidth
Scheme requirement for
8 Mbps system

BPSK 12 MHz

4PSK 6 MHz

8PSK 3 MHz

16PSK 2 MHz

51 Copyright 2010 AIRCOM International

 Signal in Noise Example
• An 8 Mbps QPSK receiver has a bandwidth of 6 MHz and requires a
SNR of requires a signal to noise ratio of 14 dB. Determine its
threshold receive level if it has a noise factor of 4 dB.

• Assuming input noise is at the level k(290)B, the effective

noise power at the input is (4 dB is a ratio of 2.5)
1.38x10-23 x 290 x 6 x 106 x 2.5 = 6 x 10-14 watts
= -102 dBm
• To deliver a signal to noise ratio of 14 dB we need a
minimum level (the “threshold”) of -88 dBm.

52 Copyright 2010 AIRCOM International

 Data Rates Carried
 The basic “building block” of digital microwave systems is a 2 Mbit/s
link that will carry, if required, 30 individual 64 kbit/s channels. The 64
kbit/s channel is the traditional “digitised speech (PCM)” channel.
 Systems are often quoted as “2x2” (i.e. 4 Mbit/s) etc..
 Further multiplexing leads to the “Pleisiosynchronous Digital
Hierarchy” where four 2 Mbit/s link form an 8 Mbit/s link, four 8 Mbit/s
form a 34 Mbit/s link and four 34 Mbit/s systems form a 140 Mbit/s link.
 The synchronous digital hierarchy (SDH) specifies higher rates of
155.52 Mbit/s and 622 Mbit/s.

53 Copyright 2010 AIRCOM International

 Session 4
Fading
54 Copyright 2010 AIRCOM International
 Fading
 Unfortunately, the strength of the received signal will vary with time,
often quite dramatically.
 The two main contributors to “fading” are:
 multipath propagation and;
 hydrometeors (e.g. rain)
 It is important to be able to predict the likely extent of fading and build
in a “margin” to allow for this in our link design.

55 Copyright 2010 AIRCOM International

 Multipath Fading
 Variations in the refractive index of the atmosphere make it possible
for the wave to propagate from transmitter to receiver via more than
one significant path.
 Constructive and destructive interference causes the signal received
to vary with time.

56 Copyright 2010 AIRCOM International

 Multipath Fading
 Multipath fading exhibits Rayleigh characteristics.
 The deeper the fade the lower the probability
 Percentage time that a fade of depth A dB is exceeded is proportional
to 10-A/10.
 ITU-R report 530-9 gives the formula for percentage as

pw  Kd 3.0
1   p 
1.2
100.033 f 0.001hL  A 10

d is the path length in km

f is the frequency in GHz
K is the “radio climactic factor”
p is path inclination in milliradians

57 Copyright 2010 AIRCOM International

 Multipath Fading
 Formulas come into the following categories:
 Deterministic
 Heuristic
 Empirical
 The multipath formula is empirical. It is based on experimental
evidence and the formula is created to fit the results.

58 Copyright 2010 AIRCOM International

 Multipath Fading
 ITU-R P.530 gives a formula for K

K  104.20.0029dN1
 dN1 can be found from ITU-R P.453-8
 values for dN1 vary between -700 and -100.
 Values for K vary between about 1.23x10-4 and 6.76x10-3

59 Copyright 2010 AIRCOM International

 Multipath Fading
dN1 ~ -200 in the United Kingdom.

K  104.20.0029200   2.4 104

60 Copyright 2010 AIRCOM International

 Multipath Fading
p is the slope of the path in milliradians

hr  he
p  d
 d is in kilometres.
 hr,e is the height of the two antennas (a.s.l.) in metres.
 For a flat path p equals zero.

61 Copyright 2010 AIRCOM International

 Multipath Fading

 For a flat path, with antennas at an elevation of 100 metres, 20 km in

length with an operating frequency of 7 GHz the probability formula
becomes:

pw  Kd 3.0
1   p 
1.2
100.033 f 0.001hL  A 10

pw  2.4 10 4  203.0 100.03370.1 A 10

 2.60 10  A 10

62 Copyright 2010 AIRCOM International

 Multipath Fading
 The formula can be used to produce a table of depth of fade against
the percentage that the fade is exceeded.

Depth of fade in dB Percentage of time

exceeded
10 0.260
15 0.082
20 0.0260
25 0.0082

It can be seen that, if 99.99% availability is required, a “fade

margin” of 24 dB would have to be designed in.

63 Copyright 2010 AIRCOM International

 Rain Fading
 Rain and other “hydrometeors” will absorb power from the
propagating electromagnetic wave and cause an additional, variable,
insertion loss. Again, a “margin” will have to be designed in to ensure
that the required availability is maintained.
 Not surprisingly, this component is very climate dependent. The
“rainfall rate exceeded for 0.01% of the time” (measured in mm/hr) is
a key parameter. Such information can be found in ITU-R P.837. The
parameter is designated R0.01.

64 Copyright 2010 AIRCOM International

 Rain Fading
 R0.01 is approximately 25 mm/hr for the UK.
 Next, ITU-R P.838 must be used to convert this to a attenuation rate
in dB/km, R.

 R  kR
 Rain attenuation is polarisation and frequency dependent
 For a flat, vertically polarised path at 7 GHz, k=0.00265, =1.312.
Hence R= 0.18 dB/km.

65 Copyright 2010 AIRCOM International

 Rain Fading
 The longer the path, and the higher the level of rainfall, the less likely it is
that it will be raining along the entire length of the path.
 This is accounted for by introducing a parameter known as the “effective
path length” that is equal to

d
1 d
d0
d 0  35e 0.015 R0.01  24
 Thus a 20 km path would have an effective length, for rainfall attenuation
purposes of 10.9 km.
 0.01% attenuation rate would be (0.18x10.9) = 2 dB.
 Insignificant compared with multipath margin (at these frequencies).

66 Copyright 2010 AIRCOM International

 Rain Fading

 For different percentages, p, the value for 0.01% can be modified

according to the formula.

Ap
 0.12 p (0.5460.043 log10 p )
A0.01

67 Copyright 2010 AIRCOM International

 Accommodating both Rain and
Multipath Fading
 Note that it would be regarded as highly unusual to add the rain and
multipath margins together.
 A more common approach would be to decide on the maximum
unavailability then build in the larger of the two calculated margins.
 The “cause of outage” requiring the lower margin would then increase
the unavailability by a very small amount.
 Rain and multipath fading would not be expected to occur
simultaneously.

68 Copyright 2010 AIRCOM International

 Selective Fading in Digital Systems

 The multipath fading that we have discussed so far caused an outage

by reducing the signal strength below the threshold.
 High error rates (hence a further “outage”) can occur in digital systems
with the signal distorted by multipath without the wideband power
necessarily reducing significantly.
 A method of predicting the unavailability due to this phenomenon is
required.
 Again ITU-R P.530 offers guidance.

69 Copyright 2010 AIRCOM International

 Selective Fading in Digital Systems
Note: relative amplitude
 I of the two paths is
given the parameter b.

Tx Rx

 The resilience of receivers to such distortion is measured by means of

introducing a two-ray system whereby the delay and relative strength
of the second signal can be adjusted.
 Attenuation is adjusted for a number of values of  so that the pre-
decided minimum value of BER is reached.
 The result is a set of “signature curves”.

70 Copyright 2010 AIRCOM International

 Selective Fading in Digital Systems
1-b
Contours for BER of 10-6

0.4
=32 ns

0.3 =16 ns
0.2
=8 ns
0.1

-4 -2 0.0 2 4 Notch offset (MHz)

 Measurements produce the above “signature curves”.

 For a fixed BER the relative strength of second path depends on the
delay and the notch position.

71 Copyright 2010 AIRCOM International

 Selective Fading in Digital Systems
1-b
Contours for BER of 10-6

0.4
=32 ns

0.3 =16 ns
0.2
=8 ns
0.1

-4 -2 0.0 2 4 Notch offset (MHz)

 From a set of curves the parameters, signature width, signature depth

and reference delay can be obtained. These can be quoted by the
manufacturer.

72 Copyright 2010 AIRCOM International

 Minimum phase and Non-minimum phase
Note: relative amplitude

 of the two paths is

given the parameter b.

Tx I Rx

 A slightly different set of curves is produced if the stronger signal is

delayed. This is known as the “non-minimum phase” configuration.
 The same parameters must be measured for the minimum and non-
minimum phase configurations.

73 Copyright 2010 AIRCOM International

 Determining the outage probability due to
selective fading
 Step 1: estimate the mean time delay on the path

1.3
 d
 m  0.7  ns
 50 

 Step 2: estimate the “multipath activity factor”, for the path.

0.2 P0 0.75
  1 e

P0 
Kd 3.0
1   p  1.2
100.033 f 0.001hL
100

74 Copyright 2010 AIRCOM International

 Determining the outage probability due to
selective fading
 Step 3: Obtain values for signature width (W), signature depth (B dB)
and reference delay  from the manufacturers data.

 Step 4: Calculate the outage probability Ps.

  m2  m2 
Ps  2.15 WM 10  BM 20
 WNM 10  BNM 20 
  r ,M  r , NM 
 

75 Copyright 2010 AIRCOM International

 Determining the outage probability due to
selective fading - example
 Considering a 20 km, 7 GHz link at 100 m altitude as before. (Hence
we can take K to be 2.4 x 10-4) Steps 1 and 2:

P0  2.4 104  203.0 100.03370.001100  100  0.0260

0.20.026 0.75
  1 e  0.013

 Step 3: From manufacturers details

WM=WNM=0.008 GHz (it seems we have an 8 MHz

system here); BM=BNM=5 dB; r = 4 ns.

76 Copyright 2010 AIRCOM International

 Determining the outage probability due to
selective fading - example
 Step 4: Calculate the probability of outage.

 5 20 0.2
2
5 20 0.2 
2
PS  2.15  0.0130.008 10   0.008 10  
 4 4 
 2.5 106

 Note that this probability is dependent on path length, frequency and

bandwidth, but NOT on received signal level.

77 Copyright 2010 AIRCOM International

 Atmospheric Absorption
 Resonances with oxygen and water molecules lead to energy being
absorbed in a frequency dependent way by the atmosphere. This adds
to the path loss.
 Atmospheric absorption is not, strictly speaking, an example of fading
as it is a constant loss. Nevertheless it is necessary to design a margin
into the link in order to compensate for such absorption.
 Atmospheric absorption is negligible below 10 GHz, rising to
approximately 0.1 dB/km at 20 GHz. It is approximately 0.1 dB/km
between 20 GHz and 40 GHz apart from a resonant peak of 0.2 dB/km
at approximately 24 GHz.

78 Copyright 2010 AIRCOM International

 Atmospheric Absorption
20

10

 Graph showing losses due to water

vapour and oxygen absorption.
1
Total atmospheric absorption is
obtained by summing the two

LOSS dB/Km
losses.
0.1

r
ou
p
va
r
e
at
W
0.01
g en
Oxy

1 10 100
FREQUENCY GHz.

Additional Loss Due To Atmospheric Content.

79 Copyright 2010 AIRCOM International

 Estimating Link Performance
 Now we appreciate the fading mechanisms and their effects, we can look
again at our 7 GHz, 20 km system. Suppose we use a 100 milliwatt (20
dBm) transmitter. The threshold is assumed to be -87 dBm with
miscellaneous losses amounting to 5 dB. 60 cm antennas are used.
 Step 1: Estimate antenna gains to be 17.5+20log(0.6)+20log(7) = 30 dBi
 Step 2: Free space loss = 92.4+20 log(20)+20 log(7)=135 dB
 Step 3: calculate unfaded receive level to be
 20-5-135+30+30=-60 dBm

80 Copyright 2010 AIRCOM International

 Estimating Link Performance - multipath
fading
 As the frequency is below 10 GHz, atmospheric absorption can be
ignored.
 The unfaded receive level can be seen to be 27 dB above the threshold.
This gives us a “fade margin” of 27 dB.
 We have previously derived a formula
A
pW  2.60 10 10
For a link of this length and frequency
 For A = 27 dB, pW is found to be 5.19x10-3%

81 Copyright 2010 AIRCOM International

 Estimating Link Performance - rain fading
 We have previously shown that the rain fading margin for a 20 km, 7
GHz path for 0.01% of the time is 2 dB.
 Although the likelihood is that rain fading can be ignored, we can
determine the percentage outage given a fade of 27 dB from the formula
Ap
 0.12 p 0.5460.043 log10 p 
p 0.5460.043 log10 p   112.5
A0.01
 For a value of Ap of 27 dB,
 Examining this equation it is found that the outage will be far less than
0.001%, which is the valid range of the equation. We can therefore
ignore outages due to rain fading.

82 Copyright 2010 AIRCOM International

 Estimating Link Performance - selective
fading

 We have previously shown that the selective fading outage probability for
a 20 km, 7 GHz, 8 MHz bandwidth path is 0.0025%. This is not affected
by the received power level.
 Summing the outages, we would predict a total outage of 0.0052
+0.00025 = 0.0055%.

83 Copyright 2010 AIRCOM International

 What’s next?

 We have obtained encouraging estimates of outage. The link, if

implemented, would provide a high quality service.
 However, we must be able to accommodate situations where the initial
prediction is for an unsatisfactory performance.
 Diversity techniques can be used to improve the performance.
 The next session reviews and analyses diversity improvement methods.

84 Copyright 2010 AIRCOM International

 Session 5
Diversity Techniques
85 Copyright 2010 AIRCOM International
 Diversity Techniques
 Our 20 km, 7 GHz, 8 MHz bandwidth link just meets the 0.01%
unavailability requirement.
 It is sensible to assume that, if we made the path longer, or increased
the bandwidth, or increased the operating frequency, we would
struggle to meet the requirements.
 Sometimes it is possible to improve the situation by increasing the
transmit power, or antenna size.

 Occasionally, these steps alone are not sufficient.

86 Copyright 2010 AIRCOM International

 Diversity Techniques
 Diversity basically relies on establishing more than one link and selecting the
best performing link at any one time or, ideally, combining the outputs from
the two links to provide the optimum output.
 Suppose we had estimated the unavailability to be 1% on a particular link.
 If we established a separate, but virtually identical, link that would also have
a 1% unavailability.
 The probability of both links being simultaneously unavailable could be
calculated to be 1%x1%=0.01%.

87 Copyright 2010 AIRCOM International

 Diversity Techniques
 Performing the calculation described would be valid only if the two links
established were independent of each other (zero correlation between fading
characteristics).
 However, as they are very similar links between the same two points, one
would intuitively expect there to be correlation between the two links.

88 Copyright 2010 AIRCOM International

 Diversity Techniques - most common
types of diversity systems
 Space diversity:- two receive antennas (usually one above the other) at each
end.
 Frequency diversity:- effectively two transceivers at separate frequencies
passing the same information over the same antenna.
 Polarisation diversity:- transmitting the same information via two orthogonal
feeders.
 Angle diversity:- usually achieved by having two separate feedhorns near the
focus of the antenna, each providing a different radiation pattern.

89 Copyright 2010 AIRCOM International

 Diversity Techniques - estimating
improvement
 The ITU provide a recommended method of estimating the improvement
provided by a diversity technique.
 Essentially, this involves estimating the degree of correlation between the
fading of the two links.
 The term “Improvement Factor” (I) is used where
I  p( A)
pd ( A)
 p(A) is the probability of a fade without diversity; pd(A) is the probability with
diversity.

90 Copyright 2010 AIRCOM International

 Diversity Techniques - space diversity
 ITU-R P.530 gives the following equation for the improvement factor.

  
I  1  exp  0.04  S 0.87 f 0.12 d 0.48 p01.04 10 AV  10
where
p0  multipath occurence factor (%)

 S is the vertical separation in metres. V is the difference in gain between the

Tx and Rx antennas (usually zero).

91 Copyright 2010 AIRCOM International

 Diversity Techniques - space diversity
 In our original link, we predicted a multipath (non-selective) outage of
0.008% for a margin of 25 dB. To make the situation more realistic for
diversity purposes, let’s assume that the transmit power was reduced so as
to make the fade margin 15 dB. That would give an outage probability of
0.08%.
 Therefore the relevant parameters are: A=15; f=7; d=20; p0=2.5. If the
antennas are separated by 5 metres the improvement factor is
  
I  1  exp  0.04  50.87 7 0.12 200.48 2.5 1.04 1015 10
 5.95
 Thus the outage probability with diversity would be expected to be 0.013%.

92 Copyright 2010 AIRCOM International

 Diversity Techniques - space diversity
 The equation was produced by examining data produced on links covering
the frequency range 2 - 11 GHz; path lengths 43 - 240 km and antenna
separations of 3 to 23 metres.
 Care must be taken when operating outside these parameters. However, the
equation doesn’t immediately “collapse” and the general rules hold:
 The bigger the separation the bigger the improvement
 The longer the path length the bigger the improvement
 Improvement factor is not very frequency-dependent exhibiting a
slight decrease with increasing frequency.

93 Copyright 2010 AIRCOM International

 Diversity Techniques - space diversity

 The equations considered so far have dealt with the “non-selective” fading
aspects of the unprotected system.
 A separate procedure must be followed to determine the new outage
probability for the selective fading.
 These two must then be summed in order to obtain the new outage estimate.

94 Copyright 2010 AIRCOM International

 Diversity Techniques - other methods

 ITU-R P.530 describes equivalent procedures for estimating the improvement

factor for Frequency, Angle and Polarisation diversity techniques.

95 Copyright 2010 AIRCOM International

 Diversity Techniques - combining
methods

 Greater improvement can be obtained by implementing more than one

technique; e.g. frequency and space diversity.

Rx f1 Tx

f2
f1

96 Copyright 2010 AIRCOM International

 Session 6
Interference Issues
97 Copyright 2010 AIRCOM International
 Interference Issues
 Interference is a problem because it “de-sensitises” the receiver.
 It does this by effectively raising the noise floor.
 Remembering our 8 MHz bandwidth system, we calculated a threshold
of -87 dBm by deducing that the noise floor was -101 dBm and that the
SNR requirement was 14 dB.
 If interference adds to this noise floor, then the threshold will be raised
and fade margins reduced.

98 Copyright 2010 AIRCOM International

 Interference Issues - adding powers
 In order to add powers it is necessary to convert from dBm to milliwatts.
 X dBm = 10X/10 milliwatts
 X dBm + Y dBm = 10log10(10X/10 + 10Y/10) dBm
 E.g. if an interfering signal of -98 dBm is added to the noise floor of -101
dBm, the resultant power level is 10log10(10-9.8 + 10-10.1) = -96.2
dBm
 The noise floor has effectively increased by 4.8 dB, making the new
threshold -82.2 dBm.

99 Copyright 2010 AIRCOM International

 Interference Issues : co-channel and
adjacent channel interference.
 The spectrum is divided into “slots” often referred to as “channels”. The
width of each slot determines the bandwidth of the system.

Used Channel
• Possible channel allocations
Adjacent Channels
for a 7GHz system.

MHz
6.992 7.000 7.008 7.016 7.024 7.032 7.040 7.048

• Interference within the bandwidth of the channel

being used is known as “co-channel”. The slots
either side are known as “adjacent channels.
100 Copyright 2010 AIRCOM International
 Interference Issues : co-channel and
adjacent channel interference.
 Co-channel interference is the most serious.
 Adjacent channel interference is reduced by the selectivity of the filter at
the receiver. Typically, it will be attenuated by 20 dB.
 Interference at frequencies outside this region will be attenuated further
and is less likely to pose a threat to the system.

101 Copyright 2010 AIRCOM International

 Interference Issues : possible scenarios
 Off beam gain of a parabolic antenna is typically 45 dB down on main
beam (Effective gain of -10 dBi). This makes high interference levels
unlikely.
 Terrestrial microwave links are more likely to cause interference to
satellite systems than be victims themselves. This has licencing
implications.

102 Copyright 2010 AIRCOM International

 Interference Issues : possible scenarios
 Multi-hop paths present a possible interference problem because of
“overshoot”.

• The effect can be reduced by using orthogonal

polarisations on consecutive hops and/or by
changing the direction between consecutive hops
by more than the antenna beamwidth.

103 Copyright 2010 AIRCOM International

 Interference Issues : reduction
technique
• Offsetting the direction of the hops.

Interfering antennas no
longer “look at” each
other.

104 Copyright 2010 AIRCOM International

 Interference Issues : possible scenarios
 Microwave transmission systems often have a “hub”.
 This hub receives signals from many different links.

105 Copyright 2010 AIRCOM International

 Interference Issues : reduction
techniques
 High performance antennas can be purchased. These are less
susceptible to “off-beam” interference.
 Frequency planning of the duplex links can also help alleviate problems

106 Copyright 2010 AIRCOM International

 Interference Issues : anomalous
propagation
 Terrestrial microwave systems are very much “line of sight” systems.
The signal tends not to propagate over the horizon.
 However, on rare occasions, interference occurs from distant systems
under conditions known as “ducting”.
 Ducting falls into a category of propagation conditions referred to as
“anomalous” (“highly unusual”; “noticeably different”).

107 Copyright 2010 AIRCOM International

 Interference Issues : ducting

Normal conditions: no interference threat

Anomalous conditions: interference threat

108 Copyright 2010 AIRCOM International

 Interference Issues : ducting

 Ducting occurs when the refractive index of the atmosphere varies with
height in such a way as to cause the radio wave to bend with the
curvature of the earth. Receive levels within a few dB of “free space
levels” can be experienced on trans-horizon paths.

109 Copyright 2010 AIRCOM International

 Interference Issues : Intermodulation
products

 No amplifier is perfectly linear. For an input vi, the output is generally:

v0  avi  bv i 2  cvi 3  dvi 4  .....
 The “even numbered” terms are out of band (harmonics), the “odd
numbered” terms are “in band” and therefore more serious.

110 Copyright 2010 AIRCOM International

 Interference Issues : Intermodulation
products
• If a number of signals at different frequency are combined within an
amplifier, the third, fifth and seventh order terms produce an interesting
effect.

Original Signals

Intermodulation Products

111 Copyright 2010 AIRCOM International

 Interference Issues : Intermodulation
products
 If a broadband receiver is receiving multiple carriers, two dominant signals
can severely interfere with a third carrier.
 If two signals at the input to an amplifier are at f1 and f2, the most
damaging intermodulation products will be at 2 f2 - f1 and 2 f1 - f2.
 A weak signal at these frequencies will be interfered with.
 Lesser effects occur at 3 f2 -2 f1 and 3 f1 -2 f2.

112 Copyright 2010 AIRCOM International

 Intermodulation products (example)

 A broadband receiver receives two signals. One at 10.02 GHz and another
at 10.035 GHz. Determine the frequencies of the four most dominant
intermodulation products.
 2 f2 - f1 = 10.050 GHz
 2 f1 - f2 = 10.005 GHz
 3 f2 - 2f1 = 10.065 GHz
 3 f1 - 2f2 = 9.990 GHz

113 Copyright 2010 AIRCOM International

 Session 7
Repeatered Systems
114 Copyright 2010 AIRCOM International
 Repeatered Systems

 Severe difficulties occur attempting to establish single hops greater than

about 50 km due to both fading and visibility problems.
 Longer paths require repeaters.
 Shorter paths with visibility problems will also require repeaters.

115 Copyright 2010 AIRCOM International

 Repeatered Systems

 Longer paths require repeaters.

 Shorter paths with visibility problems will also require repeaters.

116 Copyright 2010 AIRCOM International

 Active Repeatered Systems

 Active repeaters have a transceiver at each repeater station,

demodulating and re-transmitting the message.
 On digital systems the BER on the entire system is approximately the
sum of the individual BER’s.
 On analogue systems, the noise will accumulate, causing serious
problems.

117 Copyright 2010 AIRCOM International

 Active Repeatered Systems (Analogue)
Clean Signal Noisy Signal Noisier Signal

 The Signal to Noise ratio on a point to point link is calculated assuming

that the signal is “clean” when it leaves the transmitter.
 On the second hop the signal will be noisy as it leaves the transmitter.
Noise accumulates from hop to hop.
 Analogue systems are vastly inferior to digital systems when multi-hops
are considered.

118 Copyright 2010 AIRCOM International

 Active Repeatered Systems (Analysis)
Clean Signal Noisy Signal Noisier Signal

 Repeatered digital microwave systems can be analysed by regarding

each hop as an individual single hop system.
 The total unavailability can be approximated to be the sum of the
individual unavailabilites (provided that the individual unavailabilities are
fractions of a percent).

119 Copyright 2010 AIRCOM International

 Passive Repeatered Systems

 On short, obstructed links, it is possible to avoid the expense of an active

repeater and, instead use a passive repeater.
 The diagram shows a back to back antenna configuration of a passive
repeater.
 Total path loss is the sum of the individual hops.

120 Copyright 2010 AIRCOM International

 Passive Repeatered Systems (Example)
3 km 3 km f = 14 GHz

 Example: A 14 GHz microwave system is carried over a 6 km path. The

path is obstructed at its mid point and a passive repeater is installed.
The antennas used have a 1.2 m diameter. Estimate the path loss and
compare with that of a single hop of the same length.

121 Copyright 2010 AIRCOM International

 Passive Repeatered Systems (Solution)
3 km 3 km f = 14 GHz

 Antenna Gain ~ 17.5 +20log(1.2)+20log(14)= 42 dBi

 FSL (3 km) = 92.4+20log(3)+20log(14)=124.9 dB
 Loss per hop = 124.9 - 84 = 40.9 dB
 Total loss = 81.8 dB
 For a single (6 km) hop, FSL = 130.9 dB. Path loss = 130.9 - 84 = 46.9 dB.
 Passive repeaters increase the path loss substantially.

122 Copyright 2010 AIRCOM International

 Passive Repeatered Systems (Reflectors)

 As an alternative to back-to-back antenna systems, “billboard reflectors” can

be used as passive repeaters. These simply reflect the signal from one
antenna to the other.
 Gain of the repeater depends on its size, the frequency of operation and the
angle between the paths.

123 Copyright 2010 AIRCOM International

 Reflector Systems (Analysis)


FSL1 FSL2

 For a reflector of surface area A, the gain is given by:

 G = 42.8 + 40 log f(GHz) + 20 log A (m2) + 20 log [cos (/2)] dB
 Overall free space path loss is then FSL1 + FSL2 - G where FSL1 and FSL2
are the losses of the individual parts of the path.

124 Copyright 2010 AIRCOM International

 Reflector Systems (Example)

f = 14 GHz

120 degrees
3 km 3 km

 Considering a 6 km, 14 GHz path as before with 1.2 m antennas, determine the
size of billboard required to limit the path loss to 81.8 dB.

125 Copyright 2010 AIRCOM International

 Reflector Systems (Example)

f = 14 GHz

120 degrees
3 km 3 km

 FSL1 = FSL2 = 124.9 dB.

 Path loss = 124.9 + 124.9 - 42 - 42 - G = 81.8 dBi
 G = 84 dB = 42.8 + 40 log 14 + 20 log A (m2) + 20 log [cos (60)]
 1.4 dB = 20 log A
 A = 1.2 square metres.

126 Copyright 2010 AIRCOM International

 Reflector Systems (Example)
1.2 m2

5 km 120 degrees 1 km

 Comparison with non-symmetric split.

 G = 84 dB
 Path loss = 129.3 + 115.3 - 42 - 42 - 84 = 76.6 dBi (compared with 81.8 dBi)
 Conclusion is that placing the reflector near one of the sites is advantageous.
 Limitation occurs when it is so close to one end that path loss equals that of a
single hop (always check to ensure your prediction for path loss is greater than
that for a single hop).

127 Copyright 2010 AIRCOM International

 Reflector Systems (Double Reflectors)

 Where the angle between the paths is greater than about 130 degrees, the gain
of the antenna reduces noticeably (120 degrees is the “-6 dB angle”; 130
degrees is the “-7.5 dB angle).
 Double reflector systems can be used for greater angles.

128 Copyright 2010 AIRCOM International

 Double Reflectors (Analysis)

15

 Provided adequate clearance is provided (the 15 clearance shown is taken as

sufficient), the gain of the double reflector is approximately equal to the gain of
the smaller of the two.
 If the direction of propagation is changed at the reflector then each reflector will
change the direction of propagation by a different amount.

129 Copyright 2010 AIRCOM International

 Double Reflectors (Analysis)

2 
 = 2+180 - 1
1

 Remember
 G = 42.8 + 40 log f(GHz) + 20 log A (m2) + 20 log [cos (/2)] dB
 Compute G for both reflectors and take the smaller of the two.

130 Copyright 2010 AIRCOM International

 Double Reflectors (Optimisation)
 G = 42.8 + 40 log f(GHz) + 20 log A (m2) +
20 log [cos (/2)] dB
 2 and 1 should be as small as possible.  = 2+180 - 1
 E.g. if  has to be 160 degrees. 2 = 20
degrees and 1 = 40 degrees will be a
better solution than 2 = 60 degrees and 1
= 80 degrees.
 However, the smaller the angle the harder it
2 
is to ensure that the one reflector does not
obstruct the other.
1

131 Copyright 2010 AIRCOM International

 Session 8
Clearance Issues
132 Copyright 2010 AIRCOM International
 Clearance Requirements

 Microwave links are very much “line of sight” systems.

 If anything obstructs the path, the signal strength becomes drastically
reduced.
 A “clearance requirement” needs to be established for each link.
 This will include analysing both “earth bulge” and terrain obstructions.

133 Copyright 2010 AIRCOM International

 Clearance Requirements

Clearance

Earth Bulge

 We need to be able to calculate the “earth bulge”.

 Then, the terrain data needs to be extracted from mapping information.

134 Copyright 2010 AIRCOM International

 Earth Bulge
h h - earth bulge
d1 R - earth radius
d2
d1,2 - distances from hop ends
 Using the law of intersecting cords
2R d1d 2  2 Rh
d1d 2
h
2R
 If h is required in metres and R, d are in
kilometres:
1000d1d 2
h
2R

135 Copyright 2010 AIRCOM International

 Earth Bulge
1000d1d 2
h
2R
 Earth bulge is a maximum where d1=d2=d/2.
 Then the earth bulge = 125d 2 Taking the earth radius to be 6373
km:
R

Path Length (km) Max Earth Bulge (m)

10 2.0
20 7.9
30 17.7
40 31.4
50 49.0
60 70.6

136 Copyright 2010 AIRCOM International

 Modified Earth Radius
 Radio signals will continue slightly beyond the horizon. This is because the
refractive index of the atmosphere tends to reduce with height causing the radio
wave to bend in the direction of curvature of the earth.
 Thus the effect of the earth bulge does not have as big an effect as first
calculated.
 The effective earth bulge can be calculated by assuming the earth’s radius is
larger than its physical value.

Visible horizon
Radio horizon

137 Copyright 2010 AIRCOM International

 Modified Earth Radius
 The actual Earth’s radius is multiplied by a factor given the value k (often referred
to as the k-factor).
 For a “standard atmosphere”, k = 1.33 reducing the effective earth bulge to 0.75 of
its calculated value.
 k varies with atmospheric conditions.

k = 0.66

k = 1.0
k = 1.33
k = 2.0
k = 4.0
k=

138 Copyright 2010 AIRCOM International

 Variability of k-Factor
 Just as we need to know the extent of multipath fading for, say, 0.1% of the time,
we also need to know the minimum value of the k-factor of the same percentage
in order to establish the necessary clearance.
 The value exceeded for 99.9% of the time depends on the climate and on the path
length (as very anomalous atmospheric structures will tend not to occur over large
distances simultaneously.
k

0.9

Value of k exceeded for

0.7
99.9% of the worst month.

0.5

0.3
10 20 40 80 Path length (km)

139 Copyright 2010 AIRCOM International

 The Fresnel Parameter
 The amount of clearance required depends on the path length, the position of the
obstruction along the path and the frequency of operation.
 The Fresnel Parameter links these together to give a universally applicable
parameter.

b c
h

d1 d2

b  c  d1  d 2

• There exists a value of h such that

b  c  d1  d 2  
2

140 Copyright 2010 AIRCOM International

 The Fresnel Parameter
b
h
c • There exists a value of h
b  c  d1  d 2  
such that
d1 d2
2

 The locus of points for which this is true form an ellipsoid in three dimensions
known as the “First Fresnel Zone” and the values of h at points along the path are
known as the F1 values.
 If h<<(d1+d2) then F1 in metres is given approximately by
d1d 2
F1  17.3
f d1  d 2 
• f is in GHz, d1, d2 are measured in kilometres.

141 Copyright 2010 AIRCOM International

 The Fresnel Ellipsoid

F1

d1 d2

d1d 2
F1  17.3
f d1  d 2 

142 Copyright 2010 AIRCOM International

 Clearance Requirements

 The antennas should be sufficiently high to meet the more onerous of the
following requirements.
 For k = 1.33, clearance of 1.0 F1 should be obtained.
 For k = “minimum exceeded for 99.9% of the time”, clearance of 0.3 F1
should be obtained if the obstacle is rounded or zero if there is a sharp
single isolated obstacle.

143 Copyright 2010 AIRCOM International

 Clearance Requirements (example)

 A 30 km, 14 GHz path has an isolated obstacle 12 metres in height at a distance

of 13 km from one end.

13 km

12 m

30 km

144 Copyright 2010 AIRCOM International

 Clearance Requirements (example)
 k = 1.33. Earth Bulge at 13 km from one end = 17.7/1.33 = 13.3 m
 Add 12 m obstacle height to give 25.3 m in total.
 F1 at 13 km from one end = 12.5 metres
 Total required clearance 37.8 metres.
 Each antenna should be 37.8 metres in height.

13 km

12 m

30 km

145 Copyright 2010 AIRCOM International

 Clearance Requirements (example)
 k = “minimum exceeded for 99.9% of the time” = 0.69. Earth Bulge at 13 km from
one end = 17.7/0.69 = 25.7 m
 Add 12 m obstacle height to give 37.7 m in total.
 Zero F1 clearance required as obstacle is isolated.
 Each antenna should be 37.7 metres in height.
 37.8 metres calculation should be used.

13 km

12 m
30 km

146 Copyright 2010 AIRCOM International

 Diffraction loss.
 If the required clearance cannot be obtained, then an additional path loss will be
incurred.
 This loss can be estimated using the following techniques:
 Knife edge obstacle
 Obtain Fresnel parameter,v from
 (all parameters in the same units) 2 1 1
 Then predict path loss in dB vh   
  d1 d 2 

 Note: valid for v > -0.7

Path loss  6.9  20 log  v  0.1 1  v  0.1dB

2

147 Copyright 2010 AIRCOM International

 Diffraction loss (example)
 A 30 km, 14 GHz link is obstructed at its midpoint by a knife-edge obstacle such
that the value of h is 5 metres. Determine the diffraction loss incurred.

2 1 1 
vh   
  d1 d 2 
2  1 1 
5   
0.0214  15000 15000 
 0.558
Path loss  6.9  20 log  v  0.1  1  v  0.1
2

 6.9  20 log 0.558  0.1  1  0.558  0.1

2

 10.75 dB

148 Copyright 2010 AIRCOM International

 Diffraction over “average terrain”
 The description “knife-edge” may not apply to a particular obstacle.
 An approximate formula for average terrain exists based on the “normalized
clearance” expressed as a multiple of F1.
 Path loss ~ 10 - 20 h/ F1 dB.
 Note: valid for values of h larger than F1 (obstructed paths only).

149 Copyright 2010 AIRCOM International

 Diffraction over “average terrain”
(example)
 A 30 km, 14 GHz path propagates over “average terrain” of height equal to the
height of the base of the antenna towers.
 The antennas are 15 m above ground level.
 Determine the diffraction loss when the k-factor is 0.7.

150 Copyright 2010 AIRCOM International

 Diffraction over “average terrain”
(solution)
 Earth Bulge =
125d 2 kR  25.6 m

d1d 2
F1  17.3
f d1  d 2 
 12.6 m
 Obstruction = 25.6 - 15 = 10.6 metres
 h/F1 = -0.84
 Diffraction Loss = 10 - 20x(-0.84) = 27 dB

151 Copyright 2010 AIRCOM International

 Fading due to Ground Reflections

Diffracted Rays

Reflected Ray

 Multipath caused by ground reflection can cause severe fades.

 Smooth ground causes more severe fading than rough ground.

152 Copyright 2010 AIRCOM International

 Fading due to Ground Reflections

• As the path length difference between the reflected and direct ray alters,
“constructive” and “destructive” interference is experienced. Destructive
interference can cause a severe reduction in signal strength.

153 Copyright 2010 AIRCOM International

 Fading due to Ground Reflections

Direct Ray
Constructive
Interference
Reflected Ray

Resultant

154 Copyright 2010 AIRCOM International

 Fading due to Ground Reflections

Direct Ray
Destructive
Interference
Reflected Ray

Resultant

155 Copyright 2010 AIRCOM International

 Protecting against Reflection Fades

 The effect is that an “interference pattern” develops in which the strength

of the received signal varies with height.

“Null” or “Trough”
“Peak”

156 Copyright 2010 AIRCOM International

 Fading due to Ground Reflections

 ITU-R P.530-10 gives a “more

accurate” equation for determining
3.90.003dN1 0.42
the climatic factor K: K  10 sa

4
Approximation:K  2.4 10
sa  6 K  2.4 104
sa  12 K  1.8 104
sa  24 K  1.3 104
sa  42 K  1.0 104

157 Copyright 2010 AIRCOM International

 The Rayleigh Criterion
 The phase difference  between two
4s sin 
rays reflecting from two different    42sf sin 
surfaces separated by distance s is 
given by


s
42sf sin   0.1 specular
42sf sin   10 diffuse
0.1  42sf sin   10 transition

158 Copyright 2010 AIRCOM International

 Protecting against Reflection Fades

 Use vertical polarisation – the reflection coefficient is lower.

 That causes the reflected ray to be weaker.

159 Copyright 2010 AIRCOM International

 Protecting against Reflection Fades
 Shielding of Reflection point
 Checks must be made to ensure that shielding occurs throughout
the range of k-factors that will be experienced.

160 Copyright 2010 AIRCOM International

 Protecting against Reflection Fades
 Moving of Reflection point to poorer
reflecting surface

161 Copyright 2010 AIRCOM International

 Protecting against Reflection Fades
• Optimum choice of antenna heights.

“Null” or “Trough”

“Peak”

162 Copyright 2010 AIRCOM International

 Protecting against Reflection Fades

 A problem – the interference pattern is not stationary. Changes in earth

curvature k-factor and variations in the tide cause the pattern to move
with time.

163 Copyright 2010 AIRCOM International

 Protecting against Reflection Fades
 The pattern moves because the path
length difference changes.
d /2 d /2

• If the reflection point is at

the midpoint, then (by

 
Pythagoras):
d  2 d / 2   h  d / 2 
2 2

164 Copyright 2010 AIRCOM International

 Protecting against Reflection Fades
 Example: d  20 km; k  1.33
 Antenna heights: 25 m a.s.l.
 Ground heights vary from 6 m below s.l. to 4 m a.s.l.

d /2 d /2

165 Copyright 2010 AIRCOM International

 Protecting against Reflection Fades
 Earth Bulge: 125d 2 / kR  5.9 m
 At 6 m below sea level path length
difference is:

 
d  2 100002  25.12  10000  0.063 m
 At 4 m above sea level path length
difference is:

 
d  2 100002  15.12  10000  0.023 m

 Variation is 0.040 metres.

166 Copyright 2010 AIRCOM International

 Protecting against Reflection Fades
 It is this variation of path length difference that is significant. A
variation by a full wavelength means that any fixed antenna will
experience a “null”.
 40 mm corresponds to a frequency of 7.5 GHz.
 At frequencies of 7.5 GHz and above it would be impossible to
position an antenna so as to avoid the occurrence of a null.

d /2 d /2

167 Copyright 2010 AIRCOM International

 Protecting against Reflection Fades
 In such circumstances, diversity would be required.
 For example, if two antennas were placed at different heights,
the likelihood of a null occurring simultaneously at both
antennas is virtually zero.
 Frequency Diversity can also be used to similar effect.

75d
fh

168 Copyright 2010 AIRCOM International

 Using Field Measurements

 If the reflection coefficient of the ground is low, or the reflecting surface

is very rough, the difference between the peak and the trough will be
very small.
 One advantage of reflection fades is that the reflecting surface is always
there (unlike atmospheric ducts).
 It is therefore possible to measure the variation of signal strength with
height and assess the seriousness of the problem.
 The difference between the peak and the null indicates the likely depth
of reflection fading that will be experienced.

169 Copyright 2010 AIRCOM International

 Session 9
Unavailability and Performance Objectives

170 Copyright 2010 AIRCOM International

 Definitions

 Unavailability: System “not working” for 10 consecutive seconds.

 “Not working” defined as BER worse that 1 x 10-3.
 Outages: Exist for less than 10 seconds and the system is still regarded
as “available” (even though the user cannot access it).
 Outages are subject to “performance objectives”.

171 Copyright 2010 AIRCOM International

 Unavailability

 Unavailability can be caused by either:

 Propagation problems
 Multipath
 Diffraction Loss
 Ducting
 Rain
 Equipment Failure

172 Copyright 2010 AIRCOM International

 Propagation Problems
 Multipath problems: unlikely as multipath outages tend to be short-lived
(much less than 10 seconds).
 Diffraction loss: obeying clearance rules should avoid this.
 Ducting: generally restricted to well-known geographical regions. Can be
combatted with space diversity.
 Rain: the most likely cause of “propagation related” unavailability at high
(10 GHz+) microwave frequencies.

173 Copyright 2010 AIRCOM International

 Equipment Problems
 Definitions
 MTBF: Mean Time Between Failures (usually several thousand
hours)
 MTTR: Mean Time To Restore.
 Availability
MTBF
A 100%
MTBF  MTTR
 Unavailabity
U  100  A

174 Copyright 2010 AIRCOM International

 Equipment Problems
 Example:
 For a single transceiver and associated equipment
 MTBF = 50,000 hours
 MTTR = 6 hours
 Link MTBF = (Terminal MTBF) x 0.5 = 25,000 hours
25000
A 100%  99.976%
25000  6
U  0.024%
 If we have a 12 hop link the total unavailability = 12x0.024=0.29%
(approximately 25 hours per year).

175 Copyright 2010 AIRCOM International

 Equipment Problems: Hot Standby
 A “hot standby” is a duplicate system permanently powered up and ready
to replace the active system should a fault occur.
 Should a transmitter fail, for example, a replacement is switched into its
place. This can occur in as short a time as 20 ms.
 With MTBF as long as 50000 hours, unavailability due to transmitter or
receiver failure becomes negligible.

176 Copyright 2010 AIRCOM International

 Implementing Hot Standby
 Transmitter Hot Standby.
 Cannot have both connected simultaneously.
 An RF switch is required to connect the Hot Standby to the
antenna system in the event of the Main Transmitter failing.

Main
Transmitt
er

Hot
Standby

177 Copyright 2010 AIRCOM International

 Implementing Hot Standby
 Receiver Hot Standby.
 It is possible to connect two receivers to the antenna system
simultaneously, via a coupler.

Main
Receiver

Hot
Standby

178 Copyright 2010 AIRCOM International

 Receiver Hot Standby: Coupler Analysis
 A symmetrical coupler will have an insertion loss of at least 3 dB (usually
nearer 4 dB) that must be accounted for in the link budget.
 Asymmetrical couplers can put more insertion loss in the standby leg and
less in the main leg.

From Antenna To Main From Antenna To Main

To Standby To Standby

Symmetrical Coupler Asymmetrical Coupler

179 Copyright 2010 AIRCOM International

 Receiver Hot Standby: Coupler Analysis
 Suppose a coupler produces a 1 dB resistive loss.
 That means that, if we have 10 nW at the input, we will have a total of 8
nW at the output.
 If this is divided equally, each arm of the coupler will receive 4 nW,
equivalent to a loss of 10log10(2.5)=4dB.
 Alternatively, one arm could receive 7.27 nW and the other arm 0.727 nW.
 The losses would then be 1.4 dB and 11.4 dB respectively.
 The choice of having only 1.4 dB loss “permanently” and an extra 10 dB
degradation of fade margin during standby periods is argued to be
superior to having 4 dB loss in both “main” and “ standby” modes.

180 Copyright 2010 AIRCOM International

 Unavailability Objectives
 ITU-T G.821 divides a hypothetical long-distance channel into “High
Grade”, “Medium Grade” and “Local Grade” services.
 Objectives for High Grade circuits of length L, where L is between 280 km
and 2500 km are:


A  100  0.3  L
2500
%
 Local Grade (e.g. GSM interconnect) objectives proposed vary between
0.01% and 0.2%. This affects repair philosophy.

181 Copyright 2010 AIRCOM International

 Performance Standards
 These standards define the required system performance when it is
available.
 Causes of degradation in performance:
 Multipath Fading (as previously analysed)
 Background Errors (Gaussian noise has no absolute maximum
value and, hence some errors will occur)
 Wind (causes misalignment of antennas)

182 Copyright 2010 AIRCOM International

 Performance Criteria: High Capacity
Services
 ES: Errored-second; any 1 second period in which an error occurs.
 Not an appropriate measure when 1 second can contain several million bits.
Instead, a block of data is considered and new terms are introduced.
 EBR: Errored Block Ratio; refers to blocks containing one or more errors.
Block size is specified for each system rate.
 ESR: Errored Second Ratio; A 1-second period that contains one or more
errored blocks.
 SESR: Severely Errored Second Ratio; A 1-second period that contains
greater than 30% or errored blocks.
 BBE: Background Block Error; An errored block not occurring as part of an
SES.
 In-service measurements of block errors is possible.

183 Copyright 2010 AIRCOM International

 Performance Criteria

 Local Grade (including links between cellular sites)

 SESR should not exceed 0.00015 during the worst month.
 ESR should not exceed 0.012 during the worst month

184 Copyright 2010 AIRCOM International

 Linking SNR, BER, ES, ESR and SESR
 We have seen that SNR affects the BER. This will in turn affect the other
parameters.
 As an example, consider the procedure to predict the SESR caused by rain
attenuation.
 Step 1: for the system under consideration use ITU-R P.530-9 to estimate the
BER that will result is SES (BERSES)
 Step 2: calculate the receive level without rain attenuation and hence
calculate the rain attenuation margin.
 Step 3: calculate the annual time percentage that the rain attenuation will
exceed the margin.
 Step 4: translate this to a worst month percentage (see ITU-R P.841)

185 Copyright 2010 AIRCOM International

 Rain-induced SESR (example)
 Consider an 18 GHz, 2 Mbit/s (VC-12) link located in Italy. Link length is 5 km.
 Step 1: BER to cause SESR is 4.0x10-4.
 Step 2: use manufacturers data to ascertain that BER of 4.0x10-4 corresponds
to receive level of -95 dBm. Calculate free space receive level to be -66 dBm.
Margin is therefore 29 dB.
 Step 3: R0.01 = 40 mm/hr. R = 8.1 dB/km (see section 4 Ques.2). Effective
path length = 3.97 km. Rain attenuation = 32 dB. We need percentage of
time that rain attenuation will be 29 dB. This is given by
Ap
 0.12 p (0.5460.043 log10 p )
A0.01
 By trial and error this is found to be 0.013%
 Step 4: Convert to worst month.

186 Copyright 2010 AIRCOM International