Title: Car Rooftop Antenna for Satellite Radio Reception
1Car Rooftop Antenna for Satellite Radio Reception
- Team 26
- Joe Banasiak
- Jeremy Seuring
2Introduction
- Satellite radio currently operates from 2.32 to
2.345 GHz - Two companies offer this service
- Sirius 2.32-2.3325 GHz, 12.95/month
- XM 2.3325-2.345 GHz, 9.99/month
- XM uses geostationary satellites with a line of
site of about 30 deg. above the equator - Sirius uses geosynchronous satellites (elliptical
orbits) with a line of site of about 60 deg.
above the equator - Both companies use terrestrial repeaters which
are used to carry signal where buildings block
satellite signals
3Objective
- Our goal is to create low-profile antennas for
both satellites and terrestrial repeaters - Design of a loop antenna for receiving the
vertically (linear) polarized terrestrial signal - Design of an array of microstrip patches for
receiving the LHCP satellite signal - We will incorporate phase shifting in our array
to demonstrate beam steering capabilities
Block Diagram
4Design Goals
- Frequency of operation from 2.32-2.345 GHz
- Satellite antenna gain of at least 2 dBi over
elevation angles 20 to 90 degrees - Demonstrate beam steering over elevation angles
- Satellite antenna needs to receive circular
polarization - Terrestrial antenna gain of at least -1 dBi from
0 to 20 degrees elevation - Terrestrial antenna needs to receive linear
polarization
5Terrestrial Antenna - Loop
- Loop Antenna
- The loop antenna is ideal for keeping the
design of our antennas low-profile
- Theory
- Loop antennas emit the desired endfire pattern
for loops with a diameter of roughly ?/10.
- Problems
- For our frequency, the diameter would be 1.29 cm,
this makes fabricating the antenna extremely
difficult - Wire radius needs to be small compared to loop
size
6Terrestrial Antenna - Monopole
- Solution
- We decided to compromise our low-profile design
by using a monopole antenna
- Theory
- Input reactance of dipoles is minimized when the
length of the dipole is about 0.5?, 1.5?, 2.5?,
etc. This is desired to make the antenna
efficient - Image Theory
- Introducing a ground plane caused an image of the
antenna above the ground plane to be created
below the ground plane - This allows for monopole antennas of half the
length of a dipole antenna, that look like the
dipole. - The input impedance of the monopole will be ½ the
input impedance of the dipole - The gain of the monopole will be twice the gain
of the dipole
7Terrestrial Antenaa - Monopole
L0.75?
- Typical patterns of a monopole
- We wish to have a pattern that has a wide endfire
beam, so the desired length of our monopole is
0.25?
L0.25?
L1.25?
8Terrestrial Antenna - Monopole
- Simulations Ansoft HFSS9
- We simulated our monopole antenna with different
lengths of the monopole around ?/4 to find the
best length of the monopole - The ground plane needed to be at least ?/2 on
each side to be effective, so we chose 8 cm.
9Terrestrial Antenna - Monopole
- Once simulations were complete, we plotted S11
for each monopole length and compared to find
optimal length
10Terrestrial Antenna - Monopole
- Fabrication
- Cut ground plane and drilled a hole in its center
- Place SMA connector with the probe feed and
dielectric through the hole in the ground plane - Soldered the SMA connector to the underside of
the ground plane, and cut the dielectric flush
with the top of the ground plane - Solder additional length of wire to probe, to
make the monopole longer than anticipated length - Connect monopole to Network Analyzer and cut the
length of the monopole down so S11 is minimized
in our frequency range
11Terrestrial Antenna - Monopole
- After fabricating the monopole, and looking at
S11, we determined S11 wasnt low enough - We decided making the ground plane larger may
improve S11, so we re-simulated and found that a
ground plane with each side 16cm improved S11
dramatically
Ground plane length 16 cm
Ground plane length 8 cm
12Radiation Pattern Test Set-up
- We obtained our radiation patterns using the
anechoic chamber, the HP 8510 Network Analyzer,
and 959 Spectrum v2.1 software - The anechoic chamber is a room with special
padding that absorbs electromagnetic radiation,
ensuring that the signal received by the gain
horn comes directly from the antenna - We tested our antennas by rotating them 360
degrees in three different planes - Azimuthal plane (x-y)
- Elevation angle 0 degrees plane (y-z)
- Elevation angle 90 degrees plane (x-z)
13Terrestrial Antenna - Monopole
- Simulations results vs Testing
lt- Simulated -gt
lt- Tested -gt
14Terrestrial Antenna - Monopole
- Fields, simulated vs tested
E(?)
E(?)
E(f)
E(f)
15Terrestrial Antenna - Monopole
- Overall specs
- Maximum Gain 4.822 dBi
- HPBW 50 degrees
- Gain greater that -1dBi over elvation angles
-12 lt ? lt 72 - Better than the specification of 0 lt ? lt 20
- Greater than 90 of power accepted over frequency
range - Input Impedance (59 j5)O
16Design of Patch Antenna Theory and Materials
- Patch antennas resonate at ?e/2, but must be
designed shorter to account for fringing fields - The type of feed and feed location will affect
the input impedance of the patch antenna - Feed location also determines types of EM
polarizations radiated and received
- Actual dimensions of patch depend effective
permittivity of the substrate - Increase in er decreases size, but increases
surface waves - Increase in dielectric thickness increases
bandwidth - Decided on er 2.2 (Rogers Duroid 5880) with
thickness 3.175 mm based on tradeoffs and
available materials
17Design of Patch Antenna Simulation and
Optimization
Initial Design
- Used an available program to find general
starting points for a matched feed position - Constructed square patch with side length of
40.8mm and feed position of 13 mm from nearest
sides (such that the feed position is located on
the diagonal of the square) - Simulated in HFSS over the range 2.3-2.6 GHz
- S11 plot shows that the patch does not reach true
resonance - VSWR is above 2 over the operating band which
results in less than 90 of the incident power
entering the patch
S11 of Original Patch
dB
Freq (GHz)
18Design of Patch Antenna Simulation and
Optimization
Optimized Design
Initial Design
- Set up an optimization in HFSS to find the best
feed point and side length for resonance and
matching - Varied side length and feed point
- Compared Smith charts, S11 plots, and VSWR plots
of the results to obtain optimized patch - Determined side length of 40.3 mm and feed
position of 15.25 mm to be optimum design - VSWR under 1.4 over entire operating band results
in over 97 of input power entering patch
- Used an available program to find general
starting points for a matched feed position - Constructed square patch with side length of
40.8mm and feed position of 13 mm from nearest
sides (such that the feed position is located on
the diagonal of the square) - Simulated in HFSS over the range 2.3-2.6 GHz
- S11 plot shows that the patch does not reach true
resonance - VSWR is above 2 over the operating band which
results in less than 90 of the incident power
entering the patch
S11 of Optimized Patch
S11 of Original Patch
Under -15 dB over entire operating band
dB
dB
Resonance at 2.33 GHz (midband)
Freq (GHz)
Freq (GHz)
19Design of Patch Array Theory
- Why choose an array?
- Arrays allow an increase in gain and beam
steering capability - Why a 3 x 3 array?
- Beam direction capabilities increase with the
number of elements, but so does complexity of
feed network. A 3 x 3 will allow us to attempt
beam steering while still having fairly simple
power combination - Array Spacing
- Half wavelength spacing allows for minimization
of sidelobes in array pattern - Array equation
- Pattern Multiplication
- Multiply the array factors and element pattern to
find radiation pattern
20Design of Patch Array Fabrication
- Build
- Weve shown single patch works, simulator ran out
of available memory when attempting to simulate
full 3 x 3 array - Used HFSS geometry to program milling machine to
mill out 3 x 3 array on Rogers Duroid 5880
dielectric - Drilled feed points and soldered a coaxial feed
to each element
- Power Combination
- Nine elements - one feed
- Used one 2 to 1 and two 8 to 1 power combiners
made available by Prof. Bernhard - Problem Power dividers are very expensive, and
most are very wideband, the ones we used worked
for 2-18 GHz (we only need 2.32-2.345 GHz) - Solution For this array to be produced the most
effective solution is to custom design a 9 way
power split.
Wilkinson Power Divider
Can extend 2-way split to either binary tree or a
N-way split
21Design of Patch Array Test Results
Network analyzer results for patch resonance and
coupling
- Typical S11 -15 dB at 2.3325 GHz
- Typical coupling for adjacent patches -17 dB
- Typical coupling for diagonal patches -23 dB
(Reflection)
(Adjacent Coupling)
(Diagonal Coupling)
- More than 90 power accepted over entire
operating band - 97 power accepted at 2.3325 GHz
22Design of Patch Array Test Results
Max Gain 10.2 dBi Axial Ratio 1.4 dB Side Lobe
Level -5 dBi HPBW 30 degrees
E(?)
E(f)
23Design of Phase Shifter Theory
- Original Plan To purchase phase shifters for the
antenna elements to demonstrate beam steering
ability - Problem Phase shifters are very expensive and we
would need 9 of them. Most phase shifters cover
a large bandwidth - Solution Build phase shifting lines for 3
discrete steering angles
- Chosen Steering Angles
- ? 70, f 0 (Along the x-axis)
- ? 30, f 90 (Along the y-axis)
- ? 70, f 30 (Off-axis)
The necessary element phasing can be calculated
with the following equation
knm -k(ndsin(?)cos(f)mdsin(?)sin(f))
- n is the element position along the x-direction
- m is the element position along the y-direction
- d ?o/2
- The delay lines were designed with lengths to
match the calculated delays - Converted from delay length in radians to
centimeters - Originally set reference element line length to
?e - HFSS gave inaccurate results due to length of
line, redesigned lines with reference length ?e
/2
24Design of Phase Shifter Simulation Fabrication
- Simulated phase lines in HFSS, made necessary
corrections - After final simulations all phase lines within 1
degree of desired phase length - Fabricated the lines with three lines to a feed
and left connection gaps so we could solder the
corresponding line for each beam steering
direction
25Design of Phase Shifter Testing
X-axis lines
Network analyzer results
Desired Phase Delay
Y-axis lines
Network analyzer results
Desired Phase Delay
Off-axis lines
Network analyzer results
Desired Phase Delay
- X-direction lines in proper range or off by
similar relative phase - Y-direction lines differ more, but are off by
relatively similar values - Off-axis results far from desired ? did not test
radiation pattern
26Design of Phase Shifter Testing X - Direction
Radiation Patterns in dBi
Max Gain 8.2 dBi Axial Ratio 2.1 dB Side Lobe
Level 0 dBi HPBW 34 degrees
E(?)
E(f)
27Design of Phase Shifter Testing Y - Direction
Max Gain 9.4 dBi Axial Ratio 2.1 dB Side Lobe
Level -5 dBi HPBW 35 degrees
E(?)
E(f)
28Function Test
- We ordered an XM radio tuner and connected our
antennas to it - The microstrip array antenna received well
- The array needed to be pointed toward the south
to pick up a signal. - On our demo day, we were located at Everitts
shipping doors. We noticed that we were able to
receive a signal when the array was pointed
toward Talbot Lab, showing that the satellite
signal was being reflected. - Monopole antenna performed well
- We were able to pick up a signal, with almost any
orientation of the antenna on a clear day when
not surrounded by buildings. This demonstrated
the monopoles large beam width.
29Cost Analysis
30Recommendations
- Build narrow band power dividers to reduce the
cost of the antennas - Steer beam by mechanical means rather than using
phase shifters - To steer beam effectively, need many different
phase shifts - Would allow smaller array or single element
- Design different layout of patches so that beam
steering is not needed - Requires circuitry to select the appropriate
patch - Gain of each patch needs to be large enough
- Encase antenna in material such as radome to
shrink the size - Construct housing for antenna for protection
Thanks to Prof. Bernhards group for materials
and guidance Special thanks to Greg Huff