1. LTE Outdoor Small Cell Antenna Considerations
IBTUF– January 13, 2014
Ray Butler
VP Engineering, Commscope

2. Its ALL about Capacity!!!
Did you know that watching a video on a smartphone uses the same capacity on a network as sending 500,000 text messages simultaneously?
Paul Rasmussen.O2’s Network In Meltdown From Smartphone Usage. FierceWireless Europe 11/18/2009
[MR] Table adjusted in therms of pn, gain and length 2

3. Data Throughput Growth
[MR] Table adjusted in therms of pn, gain and length 3

4. Three Ways to Get More Capacity!!
10,000
1,000
100 10 20 25
2000
Growth factor
Spectral Efficiency
Spectrum
Number of Cells/Sectors
Growth has historically been dominated by the increase in the number of cells/sectors
Moray Rumney.Smart Cells and Wireless Capacity Growth. PowerPoint Presentation for Agilent Technologies in LTE World Summit, Posted Online May 26, 2010: August 20,

5. What Limits LTE The highest achievable data rate requires…
Claude Shannon
Shannon’s Law says…
…The Capacity of Any System is limited by the noise in the system
eNodeB
Close to the radio users experience better data rates.
The challenge is managing interference so users over the entire cell have a Great Experience
The highest achievable data rate requires…
Widest RF bandwidth radios
Highest performing RF equipment, esp BSA
Unlimited backhaul network

6. LTE Small Cell Considerations
The information presented here was gathered in a joint effort by The University of Texas at Austin and CommScope, Inc.

7. Main Objectives of Modeling
Study 3D beamforming and its impact
Determine impact of vertical directivity
Determine impact of vertical antenna pattern
Horizon
main beam 8°
42.69 𝑚
6 𝑚
Horizon
main beam
16°
20.92 𝑚
6 𝑚

8. Comparison of Topology
Traditional Grid Model
BSs are not random, have hexagon layout
BSs Density: BS/Area, R is cell radius
UE is located randomly in the network
Poisson Point Processes (PPP )
BSs are random and modeled as PPP
BSs Density: l BS/Area
UE is located at the origin point

9. Performance Comparison
Performance of fixed grid model is an upper bound
Performance of PPP model is a lower bound
Figure is from J. G. Andrews, F. Baccelli, and R. K. Ganti, “A tractable approach to coverage and rate in networks,” IEEE Trans. Commun., vol. 59, no.11, pp , Nov

10. Poisson Point Process (PPP)
White Paper discussing PPP
J. G. Andrews, Senior Member, IEEE, F. Baccelli, and R. K. Ganti, Member, IEEE, “A Tractable Approach to Coverage and Rate in Cellular Networks,” IEEE Transactions on Communications, Vol. 59, No. 11, Nov. 2011
University of Texas has developed their own propagation tool
Based on PPP
Result defines the ‘lower’ bound of predictions or is ‘pessimistic’

11. Contains macro-cell BS and small-cell BS
Our System Model
Contains macro-cell BS and small-cell BS
Base stations are modeled as PPP
User located at the origin point
main beam
macro-cell BS
small-cell BS 𝜑

12. Channel Model S – Shadow fading parameter
K – number of buildings across the direct path between
transmitter and receiver
γ – Attenuation coefficient for each building, γ<1
Gh(𝜑) – Normalized horizontal antenna gain
Gv(θ) – Normalized vertical antenna gain
Gm – Maximum antenna gain
L – Path loss
hw – Small-scale fading coefficient, Rayleigh fading

13. Macro Cell Antenna Model
Horizontal gain
Horizontal angle relative the main beam
Front back ratio
Horizontal half power beam-width
We use sectored antenna with 65 degree horizontal HPBW and 25 dB FBR for macro cell BSs.

14. Macro Cell Antenna Model
Vertical gain
vertical pattern we use
vertical pattern from CommScope

15. Macro Cell Antenna Model
Vertical gain
θ – Negative elevation angle relative to horizontal plane
θtilt – Main beam down tilt angle
Bv – Vertical half-power beamwidth
Fv – Side lobe level relative the max gain of main beam
Horizon
Out-of-cell interference
main beam
𝜃tilt 𝜃

16. Small Cell Antenna Model
Dipole antennas:
1 element
2 elements
4 elements
2.15 dBi
78°
-3 dB
+3 dB
39°
-3 dB
+6 dB
19.5°
-3 dB

17. Small Cell Antenna Model
Dipole antennas:

18. Small Cell Antenna Model
Real 2-elements dipole antenna:
Vertical pattern:
Dimensions:
Length: mm | 25.0 in
Outer Diameter: 38.1 mm | 1.5 in
Net Weight : 1.8 kg | 4.0 lb

19. Small Cell Antenna Model
Quasi-omni antenna
Connect 3 sectored antennas to create one "quasi" omni antenna
Horizontal pattern:

20. Small Cell Antenna Model
Horizontal pattern:
Horizontal pattern used in analysis:
Generated using 3 sectored antenna with 73 degree horizontal HPBW and 25 dB FBR.
Horizontal pattern of actual antennas
(Red and Blue lines denote the +/- slants of the dual pol antenna)

21. Small Cell Antenna Model
Vertical pattern of quasi-omni antenna:
Vertical pattern we use: 14 degree vertical HPBW, 16 dB SLL

22. Urban Macro to UE Outdoor Pico to UE Path Loss Model
R – BS-UE separation in kilometers
Carrier frequency is 2 GHz

23. Small Cell Antenna Model
Study focus
Determine impact of vertical directivity
Determine impact of vertical antenna pattern
Horizon
main beam 8°
42.69 𝑚
6 𝑚
Horizon
main beam
16°
20.92 𝑚
6 𝑚

24. Simulation Settings Parameter Value Parameter Value
Power of macro cell BS
20 W
Macro cell BS density
2.05/km2
Height of macro cell BS
30 m
HPBWh of macro cell
65°
FBRh of macro cell
25 dB
Downtilt of macro cell
10°
HPBWv of macro cell 7°
SLLv of macro cell
18 dB
Gm of macro cell BS
18 dBi
Parameter
Value
Power of small cell BS
2 W
Height of small cell BS
6 m
Gm of dipole small cell antenna with 78° HPBW
2.15 dBi
Gm of dipole small cell antenna with 39° HPBW
5.15 dBi
Gm of dipole small cell antenna with 19.5° HPBW
8.15 dBi
Gm of Real 2-elements dipole small cell antenna
Gm of quasi omni small cell antenna
10.2 dBi

25. Simulation Settings Parameter Value
HPBWv of quasi omni small cell antenna
14°
SLLv of quasi omni small cell antenna
16 dB
Downtilt of small cell
8° and 16°
Attenuation coefficient γ
-40 dB
Building density to macro-cell BS density ratio ρ 15
Average building height
15 m
Average building length
25 m

26. Simulation Results
Comparison of coverage probability performance of different small cell antenna pattern, θtilt = 8°, λ2 = 15 λ1
With down tilt, the quasi omni antenna performs better.

27. Simulation Results
Comparison of coverage probability performance of different small cell antenna pattern, θtilt = 16°, λ2 = 15 λ1
With down tilt, the quasi omni antenna performs better.
Coverage probability increases with the decrease in antenna beam-width.

28. Area spectral efficiency (bps/Hz/km2)
Simulation Results
Comparison of Area of Spectral Efficiency (ASE) of different small cell antenna pattern with λ2 =15 λ1
Cases
Area spectral efficiency (bps/Hz/km2)
No tilt
8° tilt
16° tilt
1-tier network contains only macro tier BSs
14.66
2-tier network
Dipole
HPBWv = 78°
61.30 --
HPBWv = 39°
60.20
62.28
65.25
Real 2 elements dipole
58.41
61.80
69.29
HPBWv = 19.5°
52.84
62.67
74.35
Quasi-omni
HPBWv = 14°
47.94
62.91
82.00

29. Average Area Throughput (Gbps/km2)
Simulation Results
Comparison of average area throughput with λ2 =15 λ1 and 20 MHz bandwidth
Cases
Average Area Throughput (Gbps/km2)
No tilt
8° tilt
16° tilt
1-tier network contains only macro tier BSs
0.29
2-tier network
Dipole
HPBWv = 78°
1.23 --
HPBWv = 39°
1.20
1.25
1.30
Real 2 elements dipole
1.17
1.24
1.39
HPBWv = 19.5°
1.06
1.49
Quasi-omni
HPBWv = 14°
0.96
1.26
1.64

30. Throughput gain over 2 elements no tilt dipole, λ2 =15 λ1
Simulation Results
Throughput gain over 2 elements no tilt dipole, λ2 =15 λ1
Cases
Throughput Gain
No tilt
8° tilt
16° tilt
Dipole
HPBWv = 78°
5.13%
HPBWv = 39°
2.56%
6.84%
11.11%
Real 2 elements dipole
0.00%
5.98%
18.80%
HPBWv = 19.5°
-9.40%
27.35%
Quasi-omni
HPBWv = 14°
-17.95%
7.69%
40.17%

31. Simulation Results of Part 3
Comparison of coverage probability performance of different small cell BS power for quasi omni antenna, λ2 = 15 λ1
Coverage probability does not increase much as small cell BS power increases from 2W to 5W when the down tilt is small.

32. Simulation Results RF Power
Comparison of ASE of different small cell antenna pattern and power with λ2 =15 λ1
Cases
Area spectral efficiency (bps/Hz/km2)
No tilt
8° tilt
16° tilt 2W 5W
Dipole
HPBWv = 78°
61.30
61.91 --
HPBWv = 39°
60.20
60.62
62.28
61.66
65.25
64.52
Real 2 elements dipole
58.41
59.08
61.80
61.35
69.29
69.52
HPBWv = 19.5°
52.84
53.02
62.67
62.22
74.35
73.82
Quasi-omni
HPBWv = 14°
47.94
48.39
62.91
63.32
82.00
82.70

33. Simulation Results With 5W RF Power
Comparison of ASE of different small cell antenna pattern and power with λ2 =15 λ1
Cases
Area spectral efficiency (bps/Hz/km2)
No tilt
8° tilt
16° tilt 2W 5W
Dipole
HPBWv = 78°
4.95%
4.75%
HPBWv = 39°
3.06%
2.57%
6.63%
4.33%
11.71%
9.17%
Real 2 elements dipole
0.00%
5.80%
3.81%
18.63%
17.63%
HPBWv = 19.5°
-9.54%
-10.29%
7.29%
5.28%
27.29%
24.91%
Quasi-omni
HPBWv = 14°
-17.93%
-18.12%
7.70%
7.14%
40.39%
39.93%

34. The Initial Study of Metro Cells in Terms of Area Spectral Efficiency (Univ. Texas)
AREA SPECTRAL EFFICIENCY (bps/Hertz/km2)*
EFFECTIVE ASE & GAIN OVER 2-ELEMENT STACKED OMNI BASED ON DEGREES TILT
NO TILT
GAIN
8°TILT
16°TILT
Baseline: Flat LTE Network of eNodeBs Only --
14.66
Het Net with Metro Cells Included
1-Element Dipole Omni
78° HPBWv
61.30
2-Element Dipole Omni
39° HPBWv
60.20
Quasi-Omni
14° HPBWv
47.94
-20.4%
62.91
4.5%
82.00
36.2%
* Based on University of Texas Austin PPP model, assumes 1900MHz with RF power held constant at 2W and 5 metro cells per sector
Quasi-Omni Offers ~35% Improvement in Network Expense

35. Value Proposition
Comparison of costs to add 40% more sties vs. adding a more expensive antenna

36. Conventional Planning Tool Analysis
Real 2-element dipole – HPBWv = 39°
Quasi Omni – HPBWv = 7°
15 random BS
Antenna Height = 7.62 m (25 ft.)
60 watt PA
Each PA connected to 3 sectored Quasi-Omni antenna
Allows for each ‘sector’ to have independent tilt

37. Conventional Planning Tool
Real 2-Element Dipole – Single Fixed Tilt (0°)
Quasi-Omni – Optimized Tilts

38. SIR Over Studied Area

39. Tilt Settings of Planning Tool Optimization
Tilt Setting (degrees) 1 2 3 4 5 6 7 8 9 10
# of sectors at indicated tilt setting 34 17 11 12 13
Shows significance of using a more sophisticated antenna
By adjusting the tilts of the various ‘sectors’ of the quasi-omni, compensation for variances in terrain, site placement, and other can be made
Emphasizes the importance of controlling the interference experienced
Using down tilt to limit out of cell coverage
Narrower beamwidth antennas to better control where RF is transmitted

40. Validating the Value Proposition and Net Pricing through RF Network Design
A cooperative study to determine how each of the following impacts networks and subscribers (e.g., ASE, SINR and node count)
Power
Frequency
Antenna pattern
Beam vertical directivity (tilt)
Placement of null zones
Site sectorization

41. Coverage and Key Performance Metrics Quasi-Omni Achieves Best Performance in Hot-zone
PERCENTAGE OF COVERAGE AREAS EXPERIENCING GREATER THAN -105 dBm RSRP
LOCATION
TOTAL COUNT
eNodeB ALONE
1W OMNI
5W OMNI
1W QUASI
5W QUASI
On Street Locations
24,562
74.5%
76.0%
79.5%
78.7%
80.1%
In-Building Locations
30,226
31.4%
25.6%
30.9%
29.6%
31.6%
Courtyard Locations
2,980
27.2%
3.8%
10.3%
6.0%
Site Count -- 36
149
107
113 77
Average Downlink Throughput (kbps)
375
2,100
1,690
2,700
2,090
Improvement above Macro Alone
460%
351%
620%
457%
Quasi-omni offers 28% reduction in site count over omni (backhaul and rental costs alone can easily be about $ per month per pole)
Quasi-omni offers 24% improvement in average user throughput
Detailed LTE Link Budget Analysis Complete

42. Impact of vertical beam-width
Summary
Impact of down tilt
Both coverage probability and ASE of the heterogeneous network are improved with the introduction of small cell BS antenna down tilt.
Both coverage probability and ASE increase when the down tilt of small cell BS antennas increases.
Impact of vertical beam-width
With no small cell BS antenna down tilt, ASE decreases as the vertical beam-width of small cell BS antenna decreases.
With small cell BS antenna down tilt, both coverage probability and ASE increase when the vertical beam-width of small cell BS antenna decreases.

43. Proposed Field Trial
Commscope proposes a field trial to quantify benefits of antenna pattern improvements
Horizontal and vertical patterns, including effects of tilt
Trial could utilize macro-sites
Better availability and performance history than metro-sites
Looking for an isolated cluster with 10 – 15 sites
Trial duration would be ~8 weeks
Specifics of trial on the following slides

44. Field Trial – Azimuth Beam Parameters
Sector Power Ratio
The angular span between the half-power (-3 dB) points measured on the cut of the antenna’s main lobe radiation pattern
Actual beamwidths >65 can be problematic to network performance
Trend is to narrower beamwidths.
Smaller SPR indicates a higher performing antenna.
It is a measure of how much energy is radiated outside of the sector.
SPR is analytically determined from measured antenna range pattern data.
Andrew recommends less than 2%

45. Field Trial – Elevation Beam Parameters
With mechanical tilt of 8 degrees, antenna blooms to 93 degrees from no tilt beamwidth of 65 degrees
The trial would demonstrate the benefits of EL Beamwidth and tilt
Demonstrate also the effects of mechanical tilt on AZ beamwidth

46. Trial –High Level Methodology
Two week baseline period
Network based KPI’s baselined
UE based data performance
Replace antennas
Optimize tilts, parameters
Repeat two week test
Network and UE based
Compile date, create report and conclusions

47. Thank you!
Q & A 47