Ground Penetrating Radar - Shawna Jones, Meghan McGinn, Nicholas Riordan
Ground Penetrating Radar - Shawna Jones, Meghan McGinn, Nicholas Riordan
Ground Penetrating Radar - Shawna Jones, Meghan McGinn, Nicholas Riordan
by
Shawna Jones,
Meghan McGinn,
and
Nicholas Riordan
ECE 345
Julio Urbina
April 30, 2002
Project Number 17
ii
Abstract
This report will cover the design, build, and testing of a ground-penetrating radar system. The
goal of the project is to be able to map underground utilities in 3-D dimensional space in order to
avoid damage. The system is able to be broken down into separate components which are able to
be built and tested separately. It uses a frequency modulated continuous wave radar system
which allows the signal to be sent and received by the same system. Unlike conventional
systems, there is no restriction on the pulse length and does not involve any switching circuits to
assure the transmit and receive signals do not interfere. Although the entire system was not able
to be completed as one unit, the process of designing and building the components was a
valuable experience.
iii
Table of Contents
i.
ii.
iii.
Title Page.........................................i
Abstract........................................ii
Table of Contents....................................iii
1.
2.
Introduction......................................1
Design Procedure.....................................2
2.1. Frequency Modulation...............................2
2.2. FMCW Radar Subsystem..............................3
2.3. Antenna Design.....................................5
Design Details..................................7
3.1. Frequency Modulation..............................7
3.2 . FMCW Design Details.............................8
3.3. Antenna Design Details..........................10
Design Verification....................................11
4.1. Frequency Modulator..............................11
4.2. FMCW Testing....................................11
4.3. Antenna Testing...................................12
Costs...........................................14
Conclusion.....................................15
6.1. Accomplishments................................15
6.2. Uncertainties....................................15
6.3. Future Work/Alternatives................................15
3.
4.
5.
6.
References..............................................16
1. Introduction
In order to put together a ground penetrating radar system, the transmission characteristics need
to be determined. Once these are determined, the system seems to be relatively easy to hook up. There
are many choices which will affect how our design proceeds, two most pressing are the antenna and the
frequency range. In order to determine this we have spoken with multiple professors to gain insight and
direction. They have been able to provide us with some alternatives to our original idea. The following is
the modified ground penetrating radar system block diagram.
Modulator
FMCW
Antenna
2. Design Procedure
2.1 Frequency Modulation
When designing the frequency modulator to supply the modulated signal to the antenna there
were a few very specific design decisions that were taken into account. First is the functionality of the
Radar system. We inevitably wanted a device that was portable and could be attached to a large piece of
machinery or used by a worker to detect an underground pipe. This meant that is should be relatively
small in size. We also needed a device that was capable of frequencies that were in the desired frequency
range, from 600 to 1000MHz. There were really two options that were available. There was the use of a
Digital Signal Processor (DSP) through the use of FSK and the use of a VCO. The DSP that was
available had a sampling rate of 44.1 kHz, which would be unable to achieve the desired frequency
output. It was also not as portable as a VCO chip. For these reasons the VCO was chosen. It provided
the portability desired and the performance required
From here a VCO that fulfilled the frequency requirement was needed. The most appropriate
model was found from ON Semiconductors. It had a frequency range up to 1100 MHz and had the
ability to work in a variable or fixed frequency mode. To oscillate appropriately the VCO required a
Capacitor Inductor tank. To operate in the variable frequency mode the capacitor was implemented
using a tuning diode in which the capacitance can be varied with the voltage input. The basic circuit
design can be seen in Fig. 2.1.
The value of the capacitance that is required can be found from equation 2.1.
(2.1)
From Eqs. (2.1) the desired frequency range dictates the capacitance values that are needed for the L C
tank, the Inductance remains constant. The voltage input to the tuning diode varied the capacitance of
3
the tuning diode. Fig. 2.2 shows the capacitance output in relation to the voltage input to the tuning
diode.
The VCO was also designed to operate in fixed frequency mode. Fig 2.3 and Fig 2.4 show the
typical frequency output with respect to capacitance and inductance. These two values can be varied to
operate in the correct frequency.
4
Input
Power Divider
Mixer
Transmit
Band pass
Filter
LNA
Receive
Output Signal
Figure 2.5 Modified FMCW Radar Subsystem
The power divider provides the majority of the signal to the transmitting antenna, the remaining
portion is sent on to the mixer. The transmitted signal is detected by the receiving antenna and then
amplified by a low-noise amplifier to counter-effect the attenuation in the ground. The amplified signal
is then filtered to leave only the frequency range of 500-1100 MHz. Once filter, the signal is coherently
demodulated by the mixer the signal has completed the FMCW radar subsystem portion of the ground
penetrating radar system.
The mixer and the low-noise amplifier were obtained through Maxim Semiconductors. These
parts were hard to find in the desired frequency range, but once found their operating characteristics
were attractive for this project. Table 2.1 contains specifications for the MAX2640 Low-Noise Amplifier
at 900MHz.
Table 2.1 MAX2640 HIGHLIGHTED SPECIFICATIONS
Operating Frequency (MHz) Noise Figure
Gain
Reverse Isolation
MAX2640
400 to 1500
.9 dB
15.1 dB
40 dB
The specifications for the MAX2682 Down converter Mixer are given in Table 2.2:
Table 2.2 MAX2682 HIGHLIGHTED SPECIFICATIONS
Input (MHz) LO (MHz) IF (MHz) Noise Figure (at 900 MHz)
MAX2682 400 to 2500 400 to 2500 10 to 500
6.3 dB
The power divider is a 3-port device designed to provide 85% power to the transmitting antenna
and the remaining 15% into the mixer. These values were chosen arbitrarily as it could have been any
ratio desired. Since the power level returning to the receiving antenna can vary based on ground
conditions, more power was desired to operate the mixer. The power divider was based on the Wilkinson
Power divider. There was a resistor present to be able to combine power which was eliminated from the
design as this would only be working as a divider.
Zm1
Port 2
Z01
R2
Z0
Port 1
Zm2
Z02
Port 3
R3
C
B
3. Design Details
3.1
Frequency Modulation
When designing the finalized circuit we wanted a continuous sine wave that varied linearly from
600 to 1000 MHz. The modulating signal used is a saw-tooth function that varies over one time period
from 600 to 1000. The carrier signal is a sine wave. These along with the modulated output signal can be
seen in Fig 3.1.
Frequency (MHz)
1000
Carrier Signal
900
800
700
600
0.5
1.5
2.5
3.5
1
0.5
Time (ns)
0
-0.5
Modulating Signal
-1
10
12
14
16
18
20
Time (ns)
Modulated Signal
Figure 3. 1. Frequency modulation signals
The circuit design from the manufacturer called for a 40 nH inductor. This value was not
available in the Parts Shop; the smallest value available was a 3.1 uH inductor. Also, as can be seen in
Fig. 2.3 and Fig 2.4 a smaller inductance would allow a larger capacitor in the tank, hence the value of
10 nH was chosen because with a 5 pF capacitor the frequency output would be 700 MHz, which is
within our frequency range. Since a piece of wire has a 10 nH inductance associated with it a wire was
used for the inductor in the L-C tank. Equation 2.1 was then used to find the correct values of the
capacitance for the desired frequency output. The inductance was fixed at 10 nH.
From here the exact capacitance was found for both the maximum and minimum frequencies.
The exact values are in Table 3.1 below.
8
Table 3. 1 Capacitance and Frequency
Frequency (MHz) Capacitance (pF)
600
7.04
1000
2.53
These values were rounded to 3 pF and 7 pF. Therefore the actual frequency output was 918
MHz on the high end and 601 MHz on the low end. The next step was to use Fig. 2 to find the voltage
input that would provide the desired capacitance. Table 3.2 contains all the calculated values.
Table 3. 2 Frequency output and Voltage input for Capacitance Values
Capacitance (pF)
3.0
7.0
Frequency (MHz)
918.9
601.5
The finalized circuit can be seen in Fig. 3.2 is the circuit for the variable frequency output. For a
fixed frequency output the Tuning diode is replaced with a capacitor.
The VCO and Tuning diode were delivered as surface mount parts and were soldered on a
Printed Circuit board after which the circuit could be completed.
3.2
FMCW Design Details
The impedances of all lines to be used in the circuit were calculated. Once this was complete, the
impedances were mapped into the lengths and widths which would be used in the microstrip circuit.
These equations can be found in Pozar [162,163].
Table 3.3 Summary of Line Characteristics
Impedance() Length(mm) Width(mm)
Z02
100
75.18
2.1
Z01
17.5
70.4
28.4
R2
21
N/A
N/A
9
R3
Zm1
Zm2
Stub A (gk=1)
Stub B (gk=2)
120
32.4
77.5
36.7
18.35
N/A
71.93
74.6
71.93
70.4
N/A
13.3
3.5
11.3
26.8
Input
Band pass Filter
Port 1
Port 3
Mixer
Output
10
11
4. Design Verification
4.1
Frequency Modulator
To test the effectiveness of the Frequency Modulator the HP Spectral Analyzer was used. The
spectral analyzer measured the output frequency. A spike on the screen indicated the power of the
frequency that was outputted. The actual frequency that was outputted by the finished modulator was
lower than expected, approximately 520 MHz and would not modulate very well. The output ended up
being a fixed frequency. The output to the Spectral Analyzer can be seen in Figure 4.1.
500
520
Frequency (MHz)
540
The reason for the lower frequency output than expected was due to the use of a wire as the 10
nH inductor for the L-C tank. The L-C tank required a capacitor and an inductor with a high Q-factor.
Meaning that it needed devices that were reliable at the high oscillating frequencies in the VCO. A wire
does not have a high enough Q-factor for this device. By the time this was discovered it was too late to
order an inductor for the circuit but given more time that is what would have been done. Instead an
inductor was used for testing purposes to make sure the modulation would occur. Table 4.1 bellow
shows the frequency output expect from calculations and the actual frequency output for different
capacitance values and a 3.1 uF inductor. The frequencies actually outputted were relatively close to the
expected values.
Table 4. 1 Comparison of Measured and calculated Frequency output
Capacitance (pF)
6.3
5
3.3
4.2
Calculated Frequency
Output (MHz)
36.01
40.43
49.76
FMCW Testing
The operational band of the band pass filter is from 675-875 MHz. This is where the
transmission of the signal and the reflection is above and below, respectively, 3 dB. The desired band
was from 500 to 1000 MHz. This leaves an error of 60%. Since designing this component, new
techniques were learned to expand the bandwidth. There is one issue within the operational band that
12
was not within the original design. Starting at approximately 750 MHz and extending 100 MHz, the
reflection level increases and the transmission decreases. This would not allow for proper operation of
the circuit. This effect comes from the center stub in the filter. The stub is relatively wide which does not
allow it to appear one quarter wavelength from the outer stubs. The filters operation relies on exact
phase shifts to produce the desired response. In order to fix this in future designs, tapped stubs are an
option. The effect of the tapped stubs reduces board space and can reduce the effective electrical length
as a shortened shorted stub and an open appear in the place of the original stub.
The power divider did not work as intended. The best response is 47dB down from the original
level of .7 dB which was designed in the 85% power arm. In order for the mixer to work, it needed the
signal from the power divider which was attenuated beyond use. In order to solve this problem, a new
design using coupled line filters would be more effective, the basic layout of which can be seen in
Figure 4.2. The desired 85% power could be output on the thru line and the remaining power can be
coupled into the second line (see figure below). Coupled line filters also provide DC isolation , which at
the lower frequencies involved this would be desired.
Input
85% power output
15% power output
Figure 4.2
One of the mixer chips was also damaged while placing a capacitor on the board. The damage was such
the chip was unusable. A replacement was ordered but did not arrive in time to be able to complete the
circuit. The testing of the band pass filter and the power divider was done on an Agilent Network
Analyzer. The calibration scheme was a full 2-port SOLT which was automated by the analyzer. Once
the calibration was completed, the circuit was attached and the measurements were done over the
frequency range.
4.3
Antenna Testing
Testing was performed using an HP 8510 Network Analyzer. A calibration was made from 5001000 MHz. Both the transmit and receive antennas were tested. First, The Smith Chart plot was
observed to determine how well the impedance of the antenna was matched to the feed line. The length
of the balun was changed until the best match was found. This process was repeated for the other
antenna. The next step was to verify that the antenna was resonating within the 600-1000 MHz range.
This was done with the Log Mag plot. The Log Mag and Smith Charts are shown in Figure 4.3.
Some alternative structures to the bow-tie would be the TEM Horn or Yagi antenna. The TEM
Horn allows for better directivity of the signal so that less power is lost by being dissipated into the air.
Unfortunately, the size of the structure would probably restrict its use for this application. The Yagi
antenna is unidirectional and the beam width is fairly narrow which would allow for better detection of
pipes. The one possible drawback to using the Yagi antenna is that it may not cover the necessary 6001000 MHz range.
Another improvement in the design of the antennas would have been to test the antenna
impedance more accurately. Testing in this case was done as the antenna was radiating into air. A better
method, if possible, would be to perform these tests while the antenna was radiating into the ground.
The ground, since it is not a loss-free medium, may change the impedance of the antenna.
13
14
5. Costs
Parts: VCO
Tuning Diode
Etching Solution
Brass Sheets
Plexiglas, cables
Total Parts
$15.00
$10.00
$10.00
$60.00
$10.00 __
$105.00
$
$
Total: Parts
Labor
$31,720.00
Total
$31,825.00
105.00
100.00
120.00
$31,500.00
$31,720.00
15
6. Conclusion
6.1
Accomplishments
There are a few very important accomplishments from this project. First, given how little any of
us knew about radar systems before beginning we have learned a great deal. The first half of the project
was simply learning information that was essential to complete its design. On top of learning about
building our individual parts we all learned how a radar system essentially works. The specific FMCW
radar that we chose to design is a relatively new technology and this was another large accomplishment
for us, to be able to adapt a newer technology to out project.
The other major accomplishment we made was to have parts that were mostly working. The
antenna was resonating in the correct frequency range, the frequency output was a little low but within
the resonating frequencies of the antenna and the band pass filter was working. Though all the parts
were unable to come together there were aspects that functioned properly.
6.2
Uncertainties
There are a few issues that we faced while designing this project that ultimately affected the
outcome. The biggest issue was the inability to test the three parts together. We had everything ready to
test together, including boxes with pipe and dirt to send the signal into and receive it from. This was
mainly due to the equipment we needed to test the return signal was unable to be moved and then a part
was damaged just prior to the demo. Had we had a little more time it would have come together to be
tested. The other major issues are the size of the antenna and the lower than desired frequency output of
the modulator. The larger sized antenna makes the model impossible to be placed on a piece of
machinery, as originally desired. The higher frequency would not allow the pipes to actually be seen, but
these could also have been improved with more time.
6.3
Future Work/Alternatives
This project is an excellent project to be carried on. There are a few improvements that could be made.
Different types of antennas could be used. This might improve directivity for less power loss. There may
also be better ways to test the performance of the antennas. For the modulator an inductor and a
capacitor with a high Q-factor would bring the frequency range up. For the processing of the signal
finding methods to improve the band pass filter and the power divider could be explored. All in all this
project has been very challenging and others would benefit from improving our design in the future.
16
References
1. D. J. Daniels, Surface Penetrating Radar. London : Institution of Electrical Engineers, 1996.
2. GeoRadar Inc. How GPR Works February 2002, http://www.georadar.com/howitwrk.htm.
3. Jonathan D. Fredrick, A Novel Single Card FMCW Radar Transceiver with On Board Monopulse
Processing, M.S. Thesis, University of California, Los Angeles, CA, 2000.
http://www.ee.ucla.edu/students/archive/fredrick_ms.pdf.
4. Electrical Engineering Training Series, Book 12, February 2002,
http://www.tpub.com/neets/book12.
5. Simon Haykins, Communication Systems. New York: John Willey & Sons, Inc. 2001, 88-182.