Link Budget Calculations for WiMAX network
Transcript of Link Budget Calculations for WiMAX network
Link Budgetsand Software Simulations
for Wman@Fez
Louise Berthilson, Alberto Escudero [email protected] 2008
(cc) Creative Commons Share Alike NonCommercial Attribution 3.0
Short summary: This document presents link budgets for the Wman@Fez wireless network based on typical wireless parameters from (modified) WiFi and WiMax in 5 GHz. Two simulations methods have been used; the first method is based on a simple link budget equation while the second method uses GPS coordinates and SRTM cartography in a radio simulation software (Radio Mobile).
Keywords: Link budget, Radio Mobile, Outdoor radio simulations
Table of Contents1. Introduction.......................................................................................................................22. Technical specification of selected equipment...................................................................23. Assumptions......................................................................................................................44. Site Locations....................................................................................................................45. Simplified Link budgets.....................................................................................................7
5.1 SmartBridges Nexus Pro Total Series (based IEEE 802.11a).....................................75.2 Albentia ALB200 and ARBA500 series....................................................................95.3 Conclusions from simplified link budgets..................................................................11
6. Software Simulated Link Budgets...................................................................................136.1 Radio Mobile.............................................................................................................136.2 Coverage area for 5.6 GHz........................................................................................146.3 WiMax Link budgets.................................................................................................15
6.3.1 Central hub: Tghat............................................................................................166.3.2 Central Hub: Zalagh...........................................................................................196.3.3 Internet Uplinks.................................................................................................216.3.4 Secondary hub (Tghat – Merinides Hotel).........................................................22
7. Final Conclusions............................................................................................................237.1 Equipment specifications...........................................................................................24
8. References........................................................................................................................249. Appendix 1 : GPS coordinates........................................................................................2510. Appendix 2: Link Budget Calculations..........................................................................26
10.1 The elements of a link budget.................................................................................2610.2 Transmitting side....................................................................................................26
10.2.1 Transmission power (Tx).................................................................................2610.2.2 Cable Loss........................................................................................................2710.2.3 Loss in connectors............................................................................................2710.2.4 Amplifiers.........................................................................................................2710.2.5 Antenna Gain ..................................................................................................27
10.3 Propagation Losses..................................................................................................2810.3.1 Free Space Loss................................................................................................2810.3.2 Fresnel zones....................................................................................................28
10.4 Receiver side............................................................................................................3010.4.1 Antenna Gain on receiver side.........................................................................3010.4.2 Amplifiers on receiver side...............................................................................3010.4.3 Receiver sensitivity...........................................................................................3010.4.4 Margin and SNR...............................................................................................30
10.5 Terms and Concepts................................................................................................3110.6 Other relevant calculations and approaches............................................................31
1
1. Introduction
This document presents link budgets for the Wman@Fez wireless network based on typical wireless parameters from modifiedWiFi and WiMax solutions in 5 GHz. To carry out the study, we have chosen two solutions with “average” values of transmission power, gain and receiver sensibility. The assumptions used for frequencies and throughputs are based on the prestudy “Requirements Document”. We have selected, not loosing generalization, equipment from SmartBridges (modified outdoor IEE 802.11a) and Albentia (IEEE 802.162004). For each technology, two calculations of link budgets have been performed. The first set of calculations uses simplified equations of Free Space Path Loss (FSPL). The second group of simulations results from radio simulations based on the propagation model Irregular Terrain Model (ITM) (LongleyRice) using GPS coordinates and 3D cartography.
For readers unfamiliar with the calculation method for FSPL, the document includes an appendix [Appendix 2] with an introduction to the concept of link budget.
For more information about the Irregular Terrain Model (ITM) (LongleyRice), we refer to the report “The ITS Irregular Terrain Model, v1.2.2, The Algorithm1” from the National Telecommunications and Information Administration Institute for Telecommunication Sciences.
2. Technical specification of selected equipmentThe equipments selected for the link budget study are the following:
WiFi (IEEE 802.11a)
SmartBridges AirPoint Nexus Pro Total (base station)
SmartBridges AirClient Nexus Pro Total (client unit)
WiMax (IEEE 802.162004)
Albentia ARBA500 Series (base station)
Albentia ALB200 (Backhaul client)
1 URL: http://flattop.its.bldrdoc.gov/itm/itm_alg.pdf
2
The data presented in Table 1 and Table 2 has been used to simulate the link budgets.
For WiFi, we have reference values for BPSK and 64QAM as modulation technique.
BPSK corresponds to an effective throughput of 3 Mbps at (theoretical throughput of 6 Mbps), while 64QAM provides an effective throughput of 22 Mbps (theoretical throughout of 54 Mbps).
For WiMax, we have reference values for BPSK, QPSK, 16QAM and 64QAM. WiMax vendors normally market products with effective throughput, not theoretical throughput, hence the empty spaces in the left column in Table 2.
In Table 2, we can see that both standards (802.11a and 802.162004) can provide the same efficient throughput and have the same receiver sensibility using BPSK as modulation technique.
For 64QAM, we can see differences in effective throughput although the receiver sensibility is the same for both types of equipment.
Since both technologies operate in the same frequency band and has the same receiver sensibilities for the same modulation techniques, the link budgets will be almost identical to each other.
3
Throughput and Modulation Receiver sensibility [dBm]Modulation IEEE 802.11a IEEE 802.162004
Base Client Base Client
6 Mbps 3Mbps BPSK 91 91
54 Mbps 22 Mbps 64QAM 70 69
3 Mbps BPSK 89 89
10 Mbps QPSK 83.5 83.5
20 Mbps 16QAM 77 77
30 Mbps 64QAM 70.5 70.5
Theoretical throughput
Effective throughput
Table 2: Throughput (theoretical and effective), modulation techniques and receiver sensibility for SmartBridges and Albentia equipment.
Standard IEEE 802.11a IEEE 802.162004Base Client Base Client
Transmit power (dBm) 21 21 23 17
Antenna gain (dBi) 12 17 23 18
Table 1: Maximum transmit power power and (integrated) antenna gain for SmartBridges and Albentia equipment.
3. AssumptionsThe link budgets presented are based on the following assumptions:
1. The central hub of the network will either be located at Tghat or Zalagh.
2. Hotel Merinides can be used as a second hub for clients located in the eastern part of Fez with insufficient coverage from Tghat.
3. The Internet POP will either be in Arrondissemant Agdal or in the Wilaya.
4. The height of client installations is 20m (including height of building and mast).
5. The height of the central hub is 20m (including height of building and mast.
6. All links will operate within the legal limit of EIRP (1W).
7. Losses from cables and connectors are accounted for in the transmit power (Tx). This is possible since we will use very low transmit power (and high gain antennas), and hence, there is no risk of reaching the maximum transmit power.
8. A fade margin2 >10dB is required for all links3.
4. Site LocationsThe images below show two possible network topologies based on the central hub in Tghat and Zalagh.
GPS coordinates for all locations are available in Appendix 1.
2 Fade margin is “a design allowance that provides for sufficient system gain or sensitivity to accommodate expected fading, for the purpose of ensuring that the required quality of service is maintained.” Source: http://en.wikipedia.org/wiki/Fade margin
3 Informal value for “best practise” received from Albentia.
4
5
Image 1: Using Tghat as the Central hub will result in average link lengths of 6.5 km and require a sector H > 42 degrees. Hotel Merinides can be used as a second hub for clients located in the eastern part of Fez with insufficient coverage from Tghat (only for Batha in Phase 1)
6
Image 2: Using Zalagh as the Central hub results in average link lengths of 8.6km and require a sector H > 22 degrees.
SITEDistance [km]
TGHAT ZALAGHArr Agdal 5.92 7.03
Tribunal 6.76 8.84
BEC Atlas 7.18 9.04
BEC Bouremana 8.31 10.27
BEC Adarissa 7.04 10.67
BEC Dekarrat 4.94 8.57
BEC Agdal 5.81 8.35
BEC Batha 7.01 5.43
Wilaya 6.1 9.25
MediaTech Library 6.04 8.05
Max [km] 8.31 10.67
Min [km] 4.94 5.43
Average [km] 6.51 8.55
Table 3: Distances (km) between base station (BS) and subscriber station (SS) for Tghat and Zalagh as Central Hub.
5. Simplified Link budgetsThe following simulations result from using the default values from Table 1 and Table 2 and the equations presented in Appendix 2. We assume the use of builtin antennas.
5.1 SmartBridges Nexus Pro Total Series (based IEEE 802.11a)
Table 3 and 4 show the the fade margin of IEEE 802.11a (5 Ghz) for distances 110 km, alternating modulation technique and transmit power.
7
Alternative 1: EIRP 1W and BPSK modulation in both directions.
Assuming that we need a fade margin of 10 dB in each side, we can conclude that with a total transmit power of 1W (30 dBm) we can ensure up to 6 km links using BPSK (3 Mbps effective). If we want to obtain longer links but sticking to the 1W EIRP, we need to decrease the transmit power and use an external antenna in the base station (with higher gain).
Note that the shadowed rows in the Table 48 corresponds to the link lengths applicable in our case (68km).
8
Gain [dBi] Sensitivity [dBm]BASE 12 91
CLIENT 17 91
Modulation BPSK BPSKBASECLIENT CLIENTBASE
Tx [dBm] Margin [dB] Tx [dBm] Margin [dB]1 107 18 30.5 13 25.5
2 113 18 24.5 13 19.5
3 117 18 21.0 13 16.0
4 120 18 18.5 13 13.5
5 121 18 16.6 13 11.6
6 123 18 15.0 13 10.0
7 124 18 13.6 13 8.6
8 126 18 12.5 13 7.5
9 127 18 11.5 13 6.5
10 127 18 10.5 13 5.5
Distance[km]
FSPL[dB]
Table 4: Fade margin of IEEE 802.11a (5 Ghz) for Distance: 110 km, EIRP :1W and BPSK modulation.
Alternative 2: EIRP 2W (downlink) and 6.3W(uplink), corresponding to the maximum transmit power of SmartBridges base station and client unit, using 64QAM modulation in both directions (highest throughput).
We can clearly see that using 64QAM modulation technique with IEEE 802.11a equipment, operating with maximum transmit power (21dBm) and integrated antennas, stable links can only be achieved within the first kilometre.
We can conclude that using 64QAM modulation (22 Mbps effective throughput) with modified WiFi is not an option for us since we need to reach up to 68km and stay within the 1W EIRP limi
5.2 Albentia ALB200 and ARBA500 seriesTables [6,7,8] present the corresponding link budgets for the selected WiMax equipment operating in 5,4705,725 GHz.
9
Gain [dBi] Sensitivity [dBm]BASE 12 70
CLIENT 17 69
Modulation 64QAM 64QAMBASECLIENT CLIENTBASE
Tx [dBm] Margin [dB] Tx [dBm] Margin [dB]1 107 21 11.5 21 12.5
2 113 21 5.5 21 6.5
3 117 21 2.0 21 3.0
4 120 21 0.5 21 0.5
5 121 21 2.4 21 1.4
6 123 21 4.0 21 3.0
7 124 21 5.4 21 4.4
8 126 21 6.5 21 5.5
9 127 21 7.5 21 6.5
10 127 21 8.5 21 7.5
Distance[km]
FSPL[dB]
Table 5: Fade margin of IEEE 802.11a (5GHz) for Distance: 110 km, EIRP: 2W (downlink), 6.3W (uplink) and 64QAM modulation. (Maximum TX power for SmartBridges).
Alternative 1: 1W EIRP, using QPSK/BPSK modulation technique corresponding to a throughput of 10/3 Mbps.
Alternative 2: 1W EIRP, using QPSK modulation technique corresponding to a throughput of 10/10 Mbps.
10
Gain [dBi] Sensitivity [dBm]BASE 23 89
CLIENT 18 83.5
Modulation QPSK BPSKBASECLIENT CLIENTBASE
Tx [dBm] Margin [dB] Tx [dBm] Margin [dB]4 120 7 12.0 12 22.5
5 121 7 10.1 12 20.6
6 123 7 8.5 12 19.0
7 124 7 7.1 12 17.6
8 126 7 6.0 12 16.5
10 127 7 4.0 12 14.5
20 133 7 2.0 12 8.5
30 137 7 5.5 12 5.0
Distance[km]
FSPL[dB]
Table 6: Fade margin for WiMax equipment operating in 5.6 GHz, transmitting 1W EIRP using QPSK (10 Mbps) downlink and BPSK (3 Mbps) uplink.
Gain [dBi] Sensitivity [dBm]BASE 23 83.5
CLIENT 18 83.5
Modulation QPSK QPSKBASECLIENT CLIENTBASE
Tx [dBm] Margin [dB] Tx [dBm] Margin [dB]4 120 7 12.0 12 17.0
5 121 7 10.1 12 15.1
6 123 7 8.5 12 13.5
7 124 7 7.1 12 12.1
8 126 7 6.0 12 11.0
10 127 7 4.0 12 9.0
20 133 7 2.0 12 3.0
30 137 7 5.5 12 0.5
Distance[km]
FSPL[dB]
Table 7: Fade margins for standard WiMax equipment operating in 5.6 GHz, transmitting 1W EIRP using QPSK (10 Mbps)
We can see that both simulation sets (Alternative 1 and 2) have small margins (<10dB) in the client side, while the margin in the base station is acceptable up to 10km.
To improve the client margins, without overriding the EIRP limit of 1W, we will now simulate the use of external antennas in the client side with a gain of 28dBi and reduce the transmit power.
Alternative 3: 1W EIRP, using QPSK modulation technique (10/10 Mbps) and external antennas of 28dBi in the client side.
By increasing the gain in the client side (28dBi), we can obtain stable links with QPSK up to 10 km.
5.3 Conclusions from simplified link budgetsWe can conclude that with typical values for modified outdoor WiFi equipment and typical values for WiMax equipment, we can establish wireless links of 46 km for modified WiFi (@ 6 Mbps – max. 3 Mbps useful throughput) and links up to 8 km for WiMax (@10/3 Mbps).However, the link budgets for modified WiFi and Wimax do not differ much in throughout and link length with the same receiver sensibility and transmit power. With an external antenna of your choice, both technologies can offer the same type of link budgets.The difference between the technologies does not lie in the link budget, but in the Quality of Service, a parameter we can not visualize with these types of simulations.
11
Gain [dBi] Sensitivity [dBm]BASE 23 83.5
CLIENT 28 83.5
Modulation QPSK QPSKBASECLIENT CLIENTBASE
Tx [dBm] Margin [dB] Tx [dBm] Margin [dB]4 120 7 22.0 2 17.0
5 121 7 20.1 2 15.1
6 123 7 18.5 2 13.5
7 124 7 17.1 2 12.1
8 126 7 16.0 2 11.0
10 127 7 14.0 2 9.0
20 133 7 8.0 2 3.0
30 137 7 4.5 2 0.5
Distance[km]
FSPL[dB]
Table 8: Fade margins for standard WiMax equipment operating in 5.6 GHz, transmitting 1W EIRP using QPSK (10 Mbps) and an external antenna of 28 dBi in the client side.
The main difference between WiFi and WiMax is the media access protocol in use which provides WiMax with a great advantage in comparison to WiFi.WiMax is using TDMA as media access protocol, which, similar to GSM, provides each client an individual time slot to access the media and hence, avoiding overhead in regards to collisions.WiFi uses a contention protocol known as CSMACA to access the media. CSMA includes a backoff time which is a random value that a station must wait to access the medium once it has verified that the channel is free. By definition, a TDMA media access protocol is more efficient in terms of QoS than CSMACA, as TDMA offers less delay than CSMA, Also, it is also more efficient in terms of total transmission rate4 during of a period of time,
Another great difference between the two technologies is the enhanced MAC layer in WiMax. In comparison with WiFi, WiMax provides by default guarantees for multipath, jitter, delay, packet loss and effective throughput on a client and application basis. These QoS characteristics puts WiMax in a class of its own, as WiFi by default does not offer any QoS. However, modified WiFi can offer a certain level of QoS, but can not be compared to WiMax in this matter as the two technologies uses two completely different media access techniques which greatly impact the level of QoS. In this discussion it is also important to mention how the effective throughput for WiFi and WiMax differ as the number of clients increase. In WiFi, the total effective throughput of a link degrades severely as the number of clients increase, while with WiMax the aggregated throughput stays the same. This is a result of the different media access protocols they use. The simulations in this document are based on pointtopoint links, and hence, the different pointtomultipoint behaviours of WiMax an WiFi are not reflected in the results.
As mention previously, the throughput of a WiFi link, often presented as 11 Mbps or 54 Mbps, refers to a theoretical value based on the modulation technique used, and not to the effective throughput (Layer 2) that can be obtained.As a reference, the table below shows the effective throughput for a number of different theoretical throughputs for a pointtopoint link5. We can see that the difference between the theoretical value and the effective value decreases with the speed. For a theoretical value of 6Mbps, we should expect about 3 Mbps.
In our case, SmartBridges promises a maximum effective throughput of 20 Mbps for a 2 km pointtopoint link and an effective throughput of 12 Mbps for a 8 km link (much
4 Not instantaneous but over a longer period of time (kB/min for example).5 Mohammad Reza Akhavan, Study the Performance Limits of IEEE 802.11 WLANs
12
Theorectical [Mbps] Effective [Mbps]
54 30
24 17
12 10
lower than 54 Mbps)6.
Another important aspect is the spectrum efficiency, IEEE 802.11a uses a channel of 20 MHz and WiMAX in the same simulation is using 10 MHz.
Assuming that the QoS characteristics discussed above are basic requirements for the network, modified WiFi is ruled out as an alternative technology. Therefore, we will continue the simulations based on the hypotheses that we will use a WiMax solution for the Wman@Fez network.
6. Software Simulated Link BudgetsBy using GPS coordinates and 3D (SRTM) cartography of all sites, we can simulate each radio link with help of satellite maps and obtain more accurate values.
However, the satellite maps can not recreate the exact conditions of a site. The error margins of GPS simulations are larger in populated areas like cities (compared to landscape untouched by human), as large buildings are likely to block radio signals and interference with other radio equipment is likely to occur.
6.1 Radio MobileThe software used for radio simulations in this report is called Radio Mobile. Radio Mobile is a tool written by Roger Coudé (VE2DBE) which is used to predict the performance of outdoor radio systems. Although it was initially conceived for traditional amateur radio systems (voice and data communications in UHF and VHF), it allows calculations of link budgets in a wide frequency range.
Radio Mobile uses the information from terrain elevation data (digital maps) to automatically compute the path profile between an emitter and a receiver.
The software uses the ITM (Irregular Terrain Model) for radio propagation. The ITM or LongleyRice Model (1968) is a general purpose model based on electromagnetic theory and statistical analysis of terrain features. Radio Mobile implements ITM and because the model is valid in the range of 20 Mhz to 20 Ghz, the software can be used to predict the median attenuation of radio signals in WLAN and WMAN systems based on WiFi and WiMax.
Radio Mobile does not only calculates link budgets but can produce a variety of virtual maps, 3D views, stereoscopic views, and even animations.
For these simulations we have used the SRTM (Shuttle Radar Topography Mission) satellite maps which contains elevation data on a nearglobal scale and is one of the most complete highresolution digital topographic databases. SRTM is a NASA project that provides free digital topographic data. This cartography contains data samples of heights
6 SmartBridges AirPoint Nexus Pro Total: http://www.smartbridges.com/products/apnpt.asp
13
obtained by radar in intervals of 3090 meters. For Africa, the images have a precision of 3 arcs of a second or 90 meters.
6.2 Coverage area for 5.6 GHzThe images below show the coverage of radio signals in the 5.6 GHz range transmitted with 1 W (Image 4) respective 4 W(Image 5) EIRP and a gain of 23 dBi in the client side.
The table below shows the technical details for each colour of the map.
Colour Sensibility Max Sensibility Min Modulation Throughput
Red 70 74 64QAM 30 Mbps
Yellow 74 80 16QAM 20 Mbps
Green 80 86 QPSK 10 Mbps
Blue 86 90 BPSK 3 Mbps
14
Image 3: Coverage of 1W transmit power from Tghat.
We can conclude that most of our sites are located in the blue/green area of the first map (1W EIRP), which corresponds to a downlink of 10Mbps/3Mbps.
If we want to reach all sites with 16QAM or 64QAM, we need to increase the transmit power to 4W.
The notations xMbps/yMbps and xdBi/ydBi are not clear please define earlier
6.3 WiMax Link budgets
A set off WiMax link budgets have been simulated with variation in the following parameters:
1. Central hub: Tghat or Zalagh
2. Throughput (downlink/uplink): 10/3 Mbps, 10/10 Mbps
15
Image 4: Coverage of 4W transmit power from Tghat.
3. Transmit power: 1W
4. Antenna gain: 23dBi/18dBi (base station), 23dBi/28dBi (client).
5. Uplink: Arrondissement Agdal or Wilaya.
6.3.1 Central hub: Tghat
Alternative 1: 10Mbps/3Mbps asymmetric throughput, EIRP 1W, internal antennas (bs/ss: 23/18dBi),
Also in this simulation set, we see that margins are weak in the client side. To improve the client margins, without overriding the EIRP limit of 1 W, we will simulate the use of external antennas in the client side with a gain of 28 dBi and reduce transmit power.
16
Gain [dBi] Sensitivity [dBm]BASE 23 89
CLIENT 18 83.5
Modulation QPSK BPSK
Site
BASECLIENT CLIENTBASE
Tx [dBm] Margin [dB] Tx [dBm] Margin [dB]Arr Agdal 5.92 7 3.1 12 7.4
Tribunal 6.76 7 3.6 12 6.9
BEC Atlas 7.18 7 3.7 12 14.2
BEC Bouremana 8.31 7 1.9 12 8.6
BEC Adarissa 7.04 7 1.8 12 12.3
BEC Dekarrat 4.94 7 5.0 12 15.5
BEC Agdal 5.81 7 0.5 12 10.0
BEC Batha 7.01 7 1.8 12 12.3
Wilaya 6.1 7 4.6 12 15.1
MediaTech Library 5.78 7 3.4 12 13.9
Distance [km]
Table 9: Transmit power and fade margin for QPSK/BPSK, 1W EIRP using internal antennas of 23/18dBi. The simulations assume Tghat to be the central hub.
Alternative 2: 10Mbps/3Mbps asymmetric throughput, EIRP 1W, internal antenna in base station (23dBi) and external antennas in clients (28dBi).
17
Gain [dBi] Sensitivity [dBm]BASE 23 89
CLIENT 28 83.5
Modulation QPSK BPSK
Site
BASECLIENT CLIENTBASE
Tx [dBm] Margin [dB] Tx [dBm] Margin [dB]Arr Agdal 5.92 7 6.9 2 7.4
Tribunal 6.76 7 6.4 2 6.9
BEC Atlas 7.18 7 13.7 2 14.2
BEC Bouremana 8.31 7 8.1 2 8.6
BEC Adarissa 7.04 7 11.8 2 12.3
BEC Dekarrat 4.94 7 15.0 2 15.5
BEC Agdal 5.81 7 9.5 2 10.0
BEC Batha 7.01 7 11.8 2 12.3
Wilaya 6.1 7 14.6 2 15.1
MediaTech Library 5.78 7 13.4 2 13.9
Distance [km]
Table 10: Transmit power and fade margin for QPSK/BPSK, 1W EIRP using internal antenna in the base station (23dBi) and external antenna in Client (28dBi). The simulations assume Tghat to be the central hub.
Alternative 3: 10Mbps symmetric throughput, EIRP 1W, using internal antenna in base station (23dBi) and external antennas in clients (28dBi).
Although some margins are < 10dB, field trials are needed to verify the expected performance of this configuration.
18
Gain [dBi] Sensitivity [dBm]BASE 23 83.5
CLIENT 28 83.5
Modulation QPSK BPSK
Site
BASECLIENT CLIENTBASE
Tx [dBm] Margin [dB] Tx [dBm] Margin [dB]Arr Agdal 5.92 7 6.9 2 1.9
Tribunal 6.76 7 6.4 2 1.4
BEC Atlas 7.18 7 13.7 2 8.7
BEC Bouremana 8.31 7 8.1 2 3.1
BEC Adarissa 7.04 7 11.8 2 6.8
BEC Dekarrat 4.94 7 15.0 2 10.0
BEC Agdal 5.81 7 9.5 2 4.5
BEC Batha 7.01 7 11.8 2 6.8
Wilaya 6.1 7 14.6 2 9.6
MediaTech Library 5.78 7 13.4 2 8.4
Distance [km]
Table 11: Transmit power and fade margin for QPSK/QPSK, 1W EIRP using internal antenna in the base station (23dBi) and external antenna in Client (28dBi). The simulations assume Tghat to be the central hub.
6.3.2 Central Hub: Zalagh
Alternative 1: 10Mbps/3Mbps asymmetric throughput, EIRP 1W, internal antennas (bs/ss: 23/18dBi).
19
Gain [dBi] Sensitivity [dBm]BASE 23 89
CLIENT 18 83.5
Modulation QPSK BPSK
Site
BASECLIENT CLIENTBASE
Tx [dBm] Margin [dB] Tx [dBm] Margin [dB]
Arr Agdal 5.92 7 0.8 12 11.3
Tribunal 6.76 7 1.3 12 11.8
BEC Atlas 7.18 7 0.8 12 11.3
BEC Bouremana 8.31 7 4.0 12 6.5
BEC Adarissa 7.04 7 0.0 12 10.5
BEC Dekarrat 4.94 7 3.9 12 6.6
BEC Agdal 5.81 7 4.9 12 5.6
BEC Batha 7.01 7 5.7 12 16.2
Wilaya 6.1 7 0.1 12 10.6
MediaTech Library 8.05 7 1.1 12 11.6
Distance [km]
Table 12: Transmit power and fade margin for QPSK/BPSK, 1W EIRP using internal antennas of 23/18dBi. The simulations assume Zalagh to be the central hub.
Alternative 2: 10Mbps/3Mbps asymmetric throughput, EIRP 1W, internal antenna in base station (23dBi) and external antennas in clients (28dBi).
20
Gain [dBi] Sensitivity [dBm]BASE 23 89
CLIENT 28 83.5
Modulation QPSK BPSK
Site
BASECLIENT CLIENTBASE
Tx [dBm] Margin [dB] Tx [dBm] Margin [dB]Arr Agdal 5.92 7 10.8 2 11.3
Tribunal 6.76 7 11.3 2 11.8
BEC Atlas 7.18 7 10.8 2 11.3
BEC Bouremana 8.31 7 6.0 2 6.5
BEC Adarissa 7.04 7 10.0 2 10.5
BEC Dekarrat 4.94 7 6.1 2 6.6
BEC Agdal 5.81 7 11.8 2 12.3
BEC Batha 7.01 7 15.7 2 16.2
Wilaya 6.1 7 10.1 2 10.6
MediaTech Library 8.05 7 12.2 2 12.7
Distance [km]
Table 13: Transmit power and fade margin for QPSK/BPSK, 1W EIRP using internal antenna in the base station (23dBi) and external antenna in Client (28dBi). The simulations assume Zalagh to be the central hub.
Alternative 3: 10Mbps symmetric throughput, EIRP 1W, using internal antenna in base station (23dBi) and external antennas in clients (28dBi).
6.3.3 Internet Uplinks
A set of simulations for the uplink of the network have been performed based on the following assumptions:
1. The central hub of the network will either be at Tghat or Zalagh.
2. In the case of using Tghat as central hub, a second hub will be needed to reach client in the eastern part of Fez (the old town). This hub will be located at Hotel Mirinides.
3. The Internet PoP will either bee in Arrondissement Agdal or in Wilaya.
However, it is not excluded that the Internet PoP will be located at Tghat, providing connectivity from Wana's core mobile network. In that case, the Internet Uplink, simulated in this section, will no be required and both Arrondissement Agdal and the Wilaya will be connected to the central hub as Clients7.
7 Arrondissement Agdal will require higher throughput than a regular client to accommodate the eGovernment traffic from the Data centre.
21
Gain [dBi] Sensitivity [dBm]BASE 23 83.5
CLIENT 28 83.5
Modulation QPSK BPSK
Site
BASECLIENT CLIENTBASE
Tx [dBm] Margin [dB] Tx [dBm] Margin [dB]Arr Agdal 5.92 7 10.8 2 5.8
Tribunal 6.76 7 11.3 2 6.3
BEC Atlas 7.18 7 10.8 2 5.8
BEC Bouremana 8.31 7 6.0 2 1.0
BEC Adarissa 7.04 7 10.0 2 5.0
BEC Dekarrat 4.94 7 6.1 2 1.1
BEC Agdal 5.81 7 11.8 2 6.8
BEC Batha 7.01 7 15.7 2 10.7
Wilaya 6.1 7 10.1 2 5.1
MediaTech Library 8.05 7 12.2 2 7.2
Distance [km]
Table 14: Transmit power and fade margin for QPSK/QPSK, 1W EIRP using internal antenna in the base station (23dBi) and external antenna in Client (28dBi). The simulations assume Zalagh to be the central hub.
For each combination of Internet POP and central hub (4 combinations), two different sets of simulations has been carried out.
The table below shows the details for each simulation and the margins obtained in each side. The average margin has been used as an indicator on how suitable a certain combination of sites would be for acting as the uplink.
Using Tghat as the central hub has previously proven to be convenient due to its proximity to the other sites in comparison to Zalagh's position.
From these simulations, it looks like the Wilaya is a better alternative than Arrondissement Agdal, assuming Tghat to be used as central hub.
However, after conducting sites surveys at both sites, we are convinced that Arrondissement Agdal will be an excellent location for the uplink. From its flat rooftop (approx 10m high), we have a clear line of sight to Tghat without any obstructions.
6.3.4 Secondary hub (Tghat – Merinides Hotel)
Assuming that Tghat is the central hub, some clients located in the eastern part of the city, in and around the old town, are situated in a dip and will not have a clear line of sight to Tghat due to surrounding buildings.
For these cases, we have simulated a secondary hub located at Hotel Merinides, north of the old town.
The table below shows a 1W symmetric link using 28dBi gain antennas in both ends. Simulations are done for both 3/3Mbps (BPSK) and 10/10Mbps (QPSK). The results shows that a margin > 10dB can be obtain for both cases.
22
BASE HUBBASE HUB
Arr Agdal 7 23 7.4 89 2 28 6.9 83.5 7.2
Arr Agdal 2 28 12.4 89 2 28 6.9 89 9.7
Wilaya 7 23 15.1 89 2 28 14.6 83.5 14.9
Wilaya 2 28 20.1 89 2 28 14.6 89 17.4
Arr Agdal 7 23 11.3 89 2 28 10.8 83.5 11.1
Arr Agdal 2 28 16.3 89 2 28 10.8 89 13.6
Wilaya 7 23 10.6 89 2 28 10.1 83.5 10.4
Wilaya 2 28 15.6 89 2 28 10.1 89 12.9
Tx [dBm]
Gain [dBi]
Margin [dB]
Sensitivity [dBm]
Tx [dBm]
Gain [dBi]
Margin[dB]
Sensitivity [dBm]
Avg Margin [dB]
Tghat
Zalagh
Table 15: Simulation sets for the uplink using Arrondissement Agdal and Wilaya as Internet PoP and Tghat and Zalagh as central hub.
7. Final ConclusionsFollowing the regulatory requirements for 5 Ghz of a maximum of 1 W EIRP, the results of the simulations can be summarized as follows:
1. Tghat is the optimal location for the hub base station
2. The WMAN network should be designed assuming two different modulation techniques: QPSK (downlink) and BPSK (uplink). 10 Mbps total downlink capacity and 3 Mbps total uplink capacity
3. High gain parabolic external antennas (28 dBi) should be used in the clients. Low transmit power (not bigger than 5 dBm) should be used in the clients.
4. A panel antenna of 23 dBi gain and 4560 degrees aperture should be used in the main base station. A transmit power not bigger than 10 dBm should be used in the base station
5. According to the radio simulations, the Wilaya seems to be the best location for the uplink. However, site surveys to both sites reveal that Arrondissement Agdal is the better alternative as a clear lineofsite between its rooftop and Tghat can be obtained.
6. A dedicated pointtopoint (Arrondissement AgdalTghat) needs to be implemented using a separate 10 Mhz channel. High gain parabolic external antennas (28 dBi) should be used in both sides of the link with a transmit power not bigger than 5 dBm
7. Licence for two 10 MHz channels needs to be obtained from ANRT to accommodate for the backhaul link (from Internet PoP to Central hub) and the access network (Central hub to Clients). In the case of the Internet PoP being located at Tghat, it is still a good idea to separate the access network from the Arrondissement Agdal – Tghat, since that link needs more capacity to accommodate the eGovernment requests.
23
TGHAT – MERINIDES
2 28 BPSK 3/3 89 19.2
2 28 QPSK 10/10 83.5 13.7
Tx [dBm]
Gain [dBi]
Modulation technique
Throughput [Mbps]
Sensitivity [dBm]
Margin [dB]
Table 16: Simulations of 1W symmetric links (3/3Mbps and 10/10Mbps) using external antennas of 28dBi for the distance Tghat – Merinides shows working links with fade margins of 13.7/19.2 dB.
7.1 Equipment specifications
CPEs: Heavy duty antenna (710 degrees) 28 dBi, TX: 5 dBm (downlink QPSK, uplink BPSK)
BS: Heavy duty panel (4560 degrees) 23 dBi, TX: 10 dBm (downlink QPSK, uplink)
PtP uplink: Heavy duty antenna (710 degrees) 28 dBi, 5 dBm (downlink QPSK, uplink QPSK)
8. References[1] Simulating WiMAX System Performance, Kostas Stefanou, Dr. Yorgos E. Stratakos, Dr. Rodoula Makri, Professor Nicolaos Uzunoglu, ED Online ID #17724, December 2007
URL: http://www.mwrf.com/Articles/Index.cfm?ArticleID=17724&pg=2
[2] Techwise Wireless FamilyAir Link Series
URL: www.btc.co.id/images/download/Techwise_Wireless_Family.pdf
[3] Study the Performance Limits of IEEE 802.11 WLANs, Mohammad Reza Akhavan
URL: http://epubl.ltu.se/16530187/2006/36/LTUPBEX0636SE.pdf
24
9. Appendix 1 : GPS coordinates
25
Role Site Latitude LongitudeCentral hub (1) Hill Tghat 34°04'09.73"N 5°03'21.37"W
Central hub (2) Merinides hotel 34°04'09.58"N 4°58'55.78"W
Central hub (3) Hill Zalagh 34°06'23.16"N 4°57'49.19"W
Uplink Arrondissement Agdal 34°02'57,35"N 4°59'47,02"W
Client BEC Atlas 34°01'45,30"N 4°59'41,51"W
Client BEC Bouremana 34°01'01.02"N 4°59'29.42"W
Client BEC Adarissa 34°01'06.65"N 5°00'37.25"W
Client BEC Dekarrat 34°02'29.78"N 5°00'50.69"W
Client BEC Agdal 34°02'21,23"N 5°00'15,62"W
Client BEC Batha 34°03'34,90"N 4°58'50,68"W
Client Wilaya Building 34°02'08.61"N 5°00'09.95"W
Client Tribunal 34°01'55,52"N 4°59'52,57"W
Client MediaTech Library 34°02'30,10"N 5°00'10.10"W
Future client Bab Elkhokha 34°03'37,50"N 4°57'55,60"W
Future client Blida 34°04'07,10"N 4°58'13,50"W
Future client Lamteyine 34°04'03,90"N 4°58'34,80"W
Future client Andalous 34°03'28,60"N 4°58'23,00"W
Future client Lbouaanania 34°03'49,80"N 4°58'56,00"W
10. Appendix 2: Link Budget CalculationsA wireless link budget for a point to point radio link is the accounting of all of the gains and losses from the radio transmitter (source of the radio signal), through cables, connectors and free air to the receiver. Estimating the value of the “power” in the different parts of the radio link is necessary to be able to make the best design and the most adequate choice of equipment.
10.1 The elements of a link budgetThe elements can be broken down into 3 main parts:
1. Transmitting side with effective transmit power2. Propagation part with propagation losses3. Receiving side with effective receiving sensibility
A complete radio link budget is simply the sum of all contributions (in decibels) across the tree main parts of a transmission path. All positive values are gain and all negative values are losses.
Transmitter power [dBm] Cable TX loss [dB] + Antenna TX gain [dBi] Free Space Path Loss [dB] + Antenna RX gain [dBi] Cable RX loss [dB] = Margin Receiver Sensitivity [dBm]
The following section presents each element of a radio link budget.
10.2 Transmitting side
10.2.1 Transmission power (Tx)
The transmit power is the power output of the radio card. The upper limit depends on regulatory limits and therefore on country/region and point in time.
26
Image 5: The full transmission path from transmitter to the receiver.
The EIRP (Equivalent isotropically radiated power or Effective isotropic radiated power is the amount of power that would have to be emitted by an isotropic antenna to produce the peak power density observed in the direction of maximum antenna gain.
For example, let us assume a Maximum EIRP of 1 Watt (30 dBm) and no cable loss. By default from the regulatory point of view this scenarios can be considered equivalent:
(1) transmitter of 20 dBm (100 mW) and an antenna of 10 dBi gain.(2) transmitter of 3 dBm (2 mW) and an antenna of 27 dBi gain(3) transmitter of 30 dBm (1 W) and an isotropic antenna of 0 dBi gain. (this antenna does not exist in reality)
10.2.2 Cable Loss
Losses in the radio signal will take place in the cables that connect the transmitter and the receiver to the antennas. The losses depend on the type of cable and frequency of operation and are normally measured in dB/m or dB/foot.
Cable losses are very much dependent on frequency. As a rule of the thumb, you can count with double amount of loss in cables operating in the 5,4 GHz range in comparison with the 2,4 GHz range.
10.2.3 Loss in connectors
Allow at least 0.25 dB (loss) for each connector in your cabling. This value applies for properly made connectors while badly soldered DIY connectors will imply higher loss. Check data sheets for losses at your frequency range and for your connector type. Additionally, lightning arrestors that typically are used between antennas and the radio gear behind them, should be budgeted for 1 dB loss.
10.2.4 Amplifiers
Optionally, amplifiers can be used to compensate for cable loss or for other reasons as signal boosting. In general, the use of amplifiers should be seen as a last option. Intelligently optimized antennas and high sensitivity in the receiver are better than brute force amplification.
10.2.5 Antenna Gain
A typical antenna gain ranges from 2 dBi (simple integrated antenna) to 5 dBi (standard omni directional) up to 2530 dBi (parabolic). Keep in mind that are many ways in which the nominal gain of an antenna might be reduced. Losses can take place due to a number of different reasons, mainly related to incorrect installation (tilt losses, polarization losses). That means that you can only expect to get the full gain out of the antenna if it is optimally installed.
27
10.3 Propagation LossesThe propagation losses are related to all attenuation of the signal that takes place when the signal has left the transmitting antenna until it reaches the receiving antenna.
10.3.1 Free Space Loss
The majority of the power of a radio signal will be lost in the air. Even in perfect vacuum, a radio wave loses some of its energy since (according to the Huygens Principle) some energy is always radiated in directions other than our link axis. Note that this has nothing to do with air, fog, rain or any other influence that will further add losses.
The Free Space Loss (FSL) measures the power loss in free space without any kind of obstacles. The radio signal weakens in free space due to expansion into a spherical surface.
FSL is proportional to the square of the distance and also proportional to the square of the radio frequency. In decibel, that results in the following equation:
FSL(dB) = 20log10(d) + 20log10(f) + K
d = distance
f = frequency
K = constant that depends on the units used for d and f
If d is measured in meters, f in Hz and the link uses isotropic antennas, the formula is:
FSL(dB) = 20log10(d) + 20log10(f) − 147.5
Distance [km]
915 Mhz 2,4 Ghz 5,8GHz
1 92 dB 100 dB 108 dB
10 112 dB 120 dB 128 dB
100 132 dB 140 dB 148 dB
Image 13: Free Space Loss (FSL) in dB for a set of distances and frequencies.
These values are theoretical values and can very well differ from your measurements. The term “free space” is never quite so “free”, and the losses can many times be larger due to terrain influences and climatic conditions.
10.3.2 Fresnel zones
Taking our starting point in the Huygens principle, we can calculate the zones, the space around an axis that relevantly contributes to the transfer of power from source to destination.
Based on this, we can then find out what the minimum distance of an obstacle (e.g. a
28
building or a hill) from our axis should be.
Most wireless professionals work with an approach that demands that the first Fresnel zone be unobstructed, although one might be more demanding. Others demand a radius containing 60% of the total power unobstructed.
The formula below calculates the first Fresnel zone.
r=17,32∗ d1∗d2 / d∗f
d1= distance to obstacle from transmitterd2 = distance to obstacle from receiverd = distance [km]f = frequency [Ghz]r = radius [m]If the obstacle is situated in the middle (d1=d2), the formula can be simplified as:
r=17,32∗ d /4f
The radio containing 60% of the total power can be calculated as:
r 60percent =10,4∗d /4f
Distance[km] 915 Mhz 2,4 Ghz 5,8 GHz Height [m] (rel. earth*)
1 9 6 4 0,02
10 29 18 11 2
100 90 56 36 200
Image 15: Radius [m] for the first Fresnel zone
29
Image 14: Fresnel zones
The “relative earth height” describes the “hill” that the earth curvature creates between two points.
10.4 Receiver sideThe calculations are almost identical to the ones at the transmitter side.
10.4.1 Antenna Gain on receiver side
See Antenna Gain on transmitter side.
10.4.2 Amplifiers on receiver side
The calculation and the principles are the same as on the transmitting side. Again, amplification is not a recommended method unless all other options have been considered and it is really needed, e.g. for cable loss compensation.
10.4.3 Receiver sensitivity
The sensitivity of a receiver is a parameter that deserve special attention as it indicates the minimum value of power that is needed to successfully decode/extract “logical bits” and achieve a certain bit rate.
The lower the sensitivity is, the better radio receiver. A typical value is 82 dBm for a 11 Mbps link and 94 dBm for a 1 Mbps link. A 10 dB difference here (which easily can be found between different cards) is just as important as a 10 dB gain that might be won by the use of amplifiers or bigger antennas.
10.4.4 Margin and SNR
Although the signal received in the receiver is bigger than the sensitivity, it is also needed a certain margin between noise and signal to achieve a certain data bit rate.
The relationship between noise and signal is measured by the signal to noise ratio or SNR. A typical requirement of SNR is 16 dB for a 11 Mbps connection and 4 dB for the lower speed of 1 Mbps.
In situations where there is very little noise, the radio link is first limited by the sensitivity of the receiver. In urban areas where there are many radio links operating, it is common to see high levels of noise (as big as 92 dBm). In those scenarios, the radio link is limited by the need of a high received signal to satisfy the Signal to Noise (SNR) condition.
Signal to Noise Ratio [dB] = 10 * Log10 (Signal Power [W] / Noise Power [W])
In normal conditions without any other source in the 2.4 Ghz band and without industrial noise, the noise level is around 100 dBm.
30
10.5 Terms and ConceptsThere are a number of terms and concepts you will meet when dealing with radio link calculations.
System operating marginThis values tells you the difference between the signal value and the sensibility.
EIRP (Effective Isotropic Radiated Power)The Effective IsotropicallyRadiated Power (EIRP) or the Maximum Radiated Power is regulated by the national radio regulatory authority. It specifies the maximum power that is legally permitted to be send out to the free air in a specific country/area. The legal limit in Europe is normally 100 mW. In some very concrete scenarios (pointtopoint links) and in some countries outside of Europe, the maximum allowed radiated power is 1 4 W.
EIRP is a measure of the effective output of a system and is expressed as equivalent to an isotropically radiating system. In simple words, this parameter tells you how strong you are allowed to send your signal in the free air.
The Radiated Power is the result of subtracting power losses in the cable and connectors to the Transmitter Power and adding the relative “gain” of the antenna.
Radiated Power (dBm) =
Transmitter Power (dBm) – Losses from cable and connectors (dB) + Antenna Gain (dBi).
10.6 Other relevant calculations and approachesIn addition to the elements considered so far, we have to take into account correction factors due to terrain and building structures, climatic factors, and many others. All of these are very empirical by nature.
You will find them under terms like rain fade, urban fade, terrain fade, with a lot of different approaches to calculating them properly. However, there are limits to these theories as the factors that can NOT be easily calculated or predicted, are normally the ones that decide whether a link works or not.
In a long distance link, factors like rain, fog and even a changing in vegetation conditions can easily contribute to 15 dB in losses.
31