Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Synthetic aperture imaging platform based on CMOS high voltage 1 to 64 multiplexer / de-multiplexer for ultrasound guided breast biopsy needle
(USC Thesis Other)
Synthetic aperture imaging platform based on CMOS high voltage 1 to 64 multiplexer / de-multiplexer for ultrasound guided breast biopsy needle
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
ii
SYNTHETIC APERTURE IMAGING PLATFORM BASED ON
CMOS HIGH VOLTAGE 1 to 64 MULTIPLEXER / DE-MULTIPLEXER
FOR ULTRASOUND GUIDED BREAST BIOPSY NEEDLE
by
Hayong Jung
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In partial Fulfillments of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(BIOMEDICAL ENGINEERING)
August 2018
Copyright 2018 Hayong Jung
iii
DEDICATION
With genuine thanks to my parents, Chansoo Jung, Misook Kim, and my lovely sister;
Sohyun Jung, and especially commemorating four my grandparents for their unconditional love
and consistent support
iv
ACKNOWLEDGEMENTS
Over the past several years I have received tremendous support and encouragement from a
great number of individuals directly or indirectly. In retrospect, it would have been impossible
for me to finish without their help. I really appreciate it.
During the years at University of Southern California, it has always been my pleasure to work
with good people and colleagues in NIH Resource Center of Medical Ultrasonic Transducer
Technology. It was a great opportunity to make me go ahead. Especially, I would like to
express my deepest appreciation to my advisor, Dr. K. Kirk Shung for his supervision and
guidance. He has always supported me with warm encouragement, thoughtful advice, and
fervent support. His guidance leads me to grateful and joyful journey. He has tough me how to
live as well as how to research. I will keep in mind your generousness, passion and patient.
I gratefully acknowledge my committee members, Dr. K. Kirk Shung, Dr. Mike Chen, Dr.
Brent Liu for their precise and sharp comments to complete my defense and dissertation. Also,
I would like to thank to all my lab members and soccer team members. Dr. Hyung Ham Kim,
Dr. Jae Yoon Hwang, Dr. Bongjin Kang, Dr. Chang Han Yoon, Dr. Chanyang Lee, Dr. Hae
Gyun Lim, Dr. Chiwoo Yoon, Dr. Nansuk Lee, Dr. Nestor E. Cabrera-Munos, Dr. Hojong Choi,
Mr. Kyosuk Koo, Mr. Sunho Moon, Kwenmo Koo and KGSA & LAKT soccer team members.
I will not forget their warm help and support and I will always remember your advice and
friendship
I would like to dedicate this dissertation to you all of you with my heart.
v
TABLE OF CONTENTS
TABLE OF CONTENTS .......................................................................................................... V
LIST OF FIGURES ............................................................................................................... VII
ABSTRACT ........................................................................................................................... XIV
CHAPTER 1. INTRODUCTION ............................................................................................. 1
1.1 HIGH FREQUENCY ULTRASOUND .................................................................................. 1
1.2 ULTRASOUND-GUIDED BIOPSY ...................................................................................... 3
1.2.1 Motivation ................................................................................................................... 4
1.2.2 Approach ..................................................................................................................... 8
1) Synthetic Aperture Imaging system ..................................................................... 9
2) Custom-designed CMOS High voltage 1-64 Mux/De-Mux .............................. 11
1.3. CONCLUSION ............................................................................................................... 12
CHAPTER 2. DESIGN OF CMOS HIGH VOLTAGE 1-64 MUX/DE-MUX ................... 13
2.1 BLOCK DIAGRAM ......................................................................................................... 13
2.2 LOW VOLTAGE 6-64 DECODER .................................................................................... 14
2.3 HIGH VOLTAGE LEVEL SHIFTER ................................................................................... 16
2.4 CMOS HIGH VOLTAGE MULTIPLEXER/DE-MULTIPLEXER ............................................ 18
2.4.1 CMOS Analog high voltage switch ................................................................... 18
2.4.2 CMOS high voltage 1-64 Mux/De-mux............................................................. 21
2.5 CLAMPED CIRCUITS FOR LATCH-UP PROTECTION ......................................................... 23
2.6 LAYOUT ...................................................................................................................... 25
2.7 DIE CHIP ...................................................................................................................... 26
2.8 PERFORMANCE ............................................................................................................ 27
2.8.1 On-resistance, insertion loss, noise figure .......................................................... 29
2.8.2 Off-isolation ....................................................................................................... 32
2.8.3 Total harmonic distortion (THD) ....................................................................... 33
CHAPTER 3. DESIGN OF 1-CHANNEL TRANCEIVER FOR SYNTHETIC
APERTURE IMAGING PLATFORM .................................................................................. 34
3.1 INTRODUCTION ............................................................................................................ 34
3.2 SIGNAL GENERATION BASED ON FPGA (FIELD-PROGRAMMABLE GATE ARRAY) ......... 36
3.2.1 Short pulse generation ........................................................................................ 38
3.2.2 Sinusoidal wave generation ................................................................................ 40
3.3 TRANSMITTER ............................................................................................................. 41
3.3.1 Unipolar Pulser ................................................................................................... 41
3.3.2 Power Amplifier ................................................................................................. 43
vi
3.4 PROTECTION CIRCUITS ................................................................................................ 43
3.4.1 Expander............................................................................................................. 43
3.4.2 Limiter ................................................................................................................ 44
3.5 RECEIVER .................................................................................................................... 45
3.6 DESIGNED MODULES ................................................................................................... 45
3.6.1 Transceiver module1 base on Class A power amplifier ..................................... 45
3.6.2 Transceiver module2 base on unipolar pulse and Class B power amplifier ...... 49
3.7 DIGITIZER .................................................................................................................... 52
CHAPTER 4 ULTRASOUND SYNTHETIC APERTURE B-MODE IMAGING ....... 53
4.1 INTRODUCTION ............................................................................................................ 53
4.1.1 Testing bench ..................................................................................................... 54
4.1.2 Data acquisition .................................................................................................. 55
4.1.3 Labview GUI ...................................................................................................... 56
4.2 B-MODE IMAGING AFTER POST PROCESSING OF SYNTHETIC APERTURE IMAGING ......... 57
4.3 CODED EXCITATION AND PULSE COMPRESSION TECHNIQUE ........................................ 63
4.4 ADDITIONAL BEAM-FORMING TO IMAGE WIDER AREA ................................................. 68
CHAPTER 5 SPECTRAL ANALYSIS FOR TISSUE CHARACTERIZATION:
INTEGRATED BACKSCATTERING COEFFICIENT(I.B.C) & ATTENUATION
SLOPE …………… ………………………… ………………………… …………………........71
5.1 INTRODUCTION ............................................................................................................ 71
5.2 INTEGRATED BACKSCATTERING COEFFICIENT IMAGING .............................................. 73
5.3 ATTENUATION SLOPE [MHZ/DB] AND Y-INTERCEPT[DB] ............................................ 77
CHAPTER 6 SUMMARY AND FUTURE WORKS ....................................................... 80
6.1 SUMMARY ................................................................................................................... 80
6.2 ACHIEVEMENT & FUTURE WORKS ............................................................................... 81
BIBLOGRAPHY ...................................................................................................................... 83
vii
LIST OF FIGURES
Figure 1. A classification scheme for acoustic waves by frequencies................................ 1
Figure 2. Ultrasound-guided biopsy ................................................................................... 3
Figure 3. Vevo 3100 (VisualSonics, Toronto, ON, Canada) ............................................. 5
Figure 4. Suggested ultrasound-guided biopsy .................................................................. 7
Figure 5. Sequence of synthetic aperture imaging ........................................................... 10
Figure 6. Delay calculation for the A-scan signal of STA Imaging ................................. 11
Figure 7. Designed high voltage multiplexer ................................................................... 12
Figure 8. Block diagram of designed circuit .................................................................... 14
Figure 9. Symbol of 6-64 decoder .................................................................................... 15
Figure 10. Symbol of High voltage level shifter ............................................................. 16
Figure 11. Schematic of level shifter ............................................................................... 17
viii
Figure 12. Schematic of source-connected analog switch ............................................... 19
Figure 13. Simulation result of the analog switch ............................................................ 20
Figure 14. Cross sectional structure of HV CMOS 0.35um(AMS H35D4D3) ............... 21
Figure 15. Top view of designed circuit .......................................................................... 22
Figure 16. HV concept with isolated LV-logic ................................................................ 23
Figure 17. Finalized potential map. .................................................................................. 24
Figure 18. Detailed structure ............................................................................................ 24
Figure 19. Layout of 6-64 decoder(left), 1-cell with high voltage level shifter and analog
switch(right) ....................................................................................................................... 25
Figure 20. Layout of core with pad-ring(6mm x 1.6mm) ................................................ 26
Figure 21. Die chip (6mm x 1.6mm) ................................................................................ 26
Figure 22. MQFP packaged chip(12mm x 12mm) .......................................................... 27
Figure 23. Testing bench .................................................................................................. 28
ix
Figure 24. turn-on mode(left), turn-off mode(right) ........................................................ 29
Figure 25. Measured on-resistance(input voltage=0.1V, 10V, 20V ) .............................. 30
Figure 26. Measured insertion Loss(input voltage=0.1V, 10V, 20V ) ............................ 30
Figure 27. Measured noise figure(input voltage=0.1V, 10V, 20V ) ................................ 31
Figure 28. Frequency vs Off-Isolation(input voltage=0.1V, 10V, 20V) ......................... 32
Figure 29 Measured total harmonic distortion. ................................................................ 33
Figure 30. Atlys FPGA board(produced by Diligent) ...................................................... 36
Figure 31. Operation blocks ............................................................................................. 37
Figure 32. PLL primitives ................................................................................................ 38
Figure 33. Schematic to make short pulse with delay line ............................................... 39
Figure 34. Input & output signals of the and gate ............................................................ 39
Figure 35. Sine wave generation with band pass filter .................................................... 40
x
Figure 36. Schematic of unipolar pulser .......................................................................... 42
Figure 37. Schematic of expander .................................................................................... 44
Figure 38. Schematic of limiter ........................................................................................ 44
Figure 39. PSA4-5043 based ultra-low noise MMIC amplifier ....................................... 45
Figure 40. schematic of class A power amplifier ............................................................. 46
Figure 41. Designed module based on the class A topology ........................................... 47
Figure 42. Power amplifier‟s gain .................................................................................... 48
Figure 43. Diagram of 2-stage pre-amplifer..................................................................... 48
Figure 44. Pre-amplifier‟s gain ........................................................................................ 49
Figure 45. Schematic of Class B push-pull power amplifier ........................................... 50
Figure 46. Gain of Class B power amplifier .................................................................... 51
Figure 47. Unipolar pulser and Class B power amp based on FPGA signal generation. . 51
xi
Figure 48. PC based digitizer(gagecard, cs122g1) ........................................................... 52
Figure 49. Post-processing of synthetic aperture imaging ............................................... 54
Figure 50. Testing bench .................................................................................................. 54
Figure 51. Data acquisition process ................................................................................. 55
Figure 52. Labview GUI(graphic user interface) to capture the signal ............................ 56
Figure 53. 30MHz linear array ......................................................................................... 57
Figure 54. Beam-formed data after post-processing(left), b-mode image of quartz(right)
............................................................................................................................................ 58
Figure 55. Beam-formed data after post-processing ........................................................ 59
Figure 56. B-mode of 25um tungsten-wire ...................................................................... 60
Figure 57. B-mode image of metal frame ........................................................................ 61
Figure 58. B-mode image of agar phantom(left, rgb scale), agar phantom(right, grayscale)
............................................................................................................................................ 62
xii
Figure 59. Chirp excitation (linear frequency modulation).............................................. 64
Figure 60. Diagram of coded excitation and pulse compression to get B-mode image ... 65
Figure 61. b-mode image from 50um tungsten wires, 1-cycle excitaion(left), coded
excitation before pulse compression(middle), coded excitation after pulse
compression(right), the given dynamic range is 20dB ....................................................... 66
Figure 62. b-mode image from 10kPa agar phantom, 1-cycle excitaion(left), coded
excitation before pulse compression(middle), coded excitation after pulse
compression(right), the given dynamic range is 20dB ....................................................... 67
Figure 63. conventional linear array beam-forming(left), additional beam-forming(right)
............................................................................................................................................ 68
Figure 64. Linear array (# of scanline: 32, total imaging range 1.6mm) ......................... 69
Figure 65. Linear array (# of scanline: 64, total imaging range 3.2mm) ......................... 69
Figure 66. Linear array (# of scanline: 96, total imaging range 4.8mm) ......................... 70
Figure 67. Fabricated 45MHz transducer ......................................................................... 73
Figure 68. pulsed-echo signal in time domain(left), Frequency response(FFT) .............. 73
xiii
Figure 69. Lateral beam profile ........................................................................................ 74
Figure 70. Axial beam profile .......................................................................................... 74
Figure 71. 2D plot of Ispta ............................................................................................... 75
Figure 72. B-mode image of pig eye ................................................................................ 76
Figure 73. Integrated backscattering coefficient image of pig eye .................................. 76
Figure 74. Attenuation slope of pig eye ........................................................................... 79
Figure 75. Attenuation y-intercept of pig eye .................................................................. 79
xiv
ABSTRACT
Ultrasound has been used as diagnostic imaging tools in medicine for a long time due to its
real-time capability and mobility as well as nonionizing radiation and safety. High frequency
ultrasound (above 20 MHz) has opened up new biomedical applications due to its fine spatial
resolution by increasing the operating frequency.
Among the applications, ultrasound guided needle biopsy is one of methods to collect breast
cancer tissue. Especially, it is important to detect and to collect the targeted tissue. In this
paper, a high voltage 1 to 64 MUX / De-MUX is designed by high voltage CMOS process
(AMS H35B4D3) for the ultrasound-guided breast biopsy application to overcome the
limitation. The electronics is made up of three parts such as low voltage a 6 to 64 decoder, a
level shifter to convert from low voltage to high voltage, and 64-analog high voltage switches.
The experiment results show that the 3-dB bandwidth is over 70MHz, and it has 180 ohms of
on-resistance, -2.635 dB of insertion loss, -26.526 dB of isolation at 70MHz. Also, synthetic
imaging technique is used for the beam-forming which is suitable to the limited number of
transmitting to the elements. Also, 1-channel transceiver based on FPGA is developed and
implemented so that it supports total compact size as a synthetic aperture imaging platform. At
the post-processing, integrated backscattering coefficient is added to analyze the signal in the
frequency domain.
1
CHAPTER 1. INTRODUCTION
1.1 High Frequency Ultrasound
The definition of sound is an acoustic wave based on the mechanical operation that propagates
away through the mediums such as air or water. It generates pressure or acoustic power
according to pass the medium and it has an certain operating frequency range. From the point
of human‟s view, there are three categories of sound because there are upper and lower
frequency limitations to the human‟s ear. Human can hear the sound from 20 Hz to 20 kHz in
general. And, fig 1 shows the acoustic waves by frequencies. At the low range, human cannot
hear the infrasound which range is from 0.01Hz to 10 Hz. Also, human cannot here below 20
kHz which can be called ultrasound. Especially, the frequency range from 20 kHz to 20 MHz
are normally used to the clinical applications and this is the conventional range for the medical
imaging.[1]
Figure 1. A classification scheme for acoustic waves by frequencies. (Maldovan, 2013)
Ultrasound imaging is one of the medical imaging modalities which is widely used at the
clinical applications. Currently, ultrasound imaging system has a capability to scan the targets
2
or tissues as gray-scale images in real time and it also results in detection of blood flow which
can be represented by colored images. Also, there are additional trials and approaches to utilize
the ultrasound imaging system to the various biomedical areas such as drug delivery and
therapeutic treatment. Conventional clinical ultrasound imaging systems operate from 1 to
20MHz which has a resolution down to 1mm which depends on frequency and f-number from
the structure. And, here are the equations to calculate the spatial resolution of the ultrasound
beam width where is wave length, c is the sound speed, f
#
is f-number, Z
f
if a focal length, D
is the aperture size or diameter, and R
L
is a lateral resolution which we can call it normally
spatial resolution.
c
f c
(1)
D Z f
f
#
(2)
#
f R
L
(3)
According to the theorem, high frequency ultrasound leads better spatial resolution which it
has trade-off relationship to sacrifice the penetration depth. Utilizing the advantage of high
frequency ultrasound, new diagnostic medical application has been developed such as IVUS
(intravascular ultrasound imaging system) to see the plaque of the blood vessel, imaging eye or
skin, and small animal imaging like zebra-fish or mouse (Lockwood et al. 1996; Shung et al.,
2009; Shung, 2011)
3
1.2 Ultrasound-guided biopsy
Ultrasound-guided breast biopsy utilizes acoustic waves to image in a real time and to help
locate the biopsy needle to the targeted point which can collect the target tissue sample in order
to evaluate breast abnormalities it in vitro. It is widely used because it is less invasive than
surgical biopsy and it can also minimize the scarring after the surgery. The breast biopsy
collects the suspicious tissue in the breast and then doctors or pathologists examine the tissues.
Figure 2. Ultrasound-guided biopsy
4
1.2.1 Motivation
The ultrasound guided biopsy is widely used for a long time and the performance is
demonstrated by doctors or pathology. And, the technique enhances the biopsy procedure‟s
efficiency to image and to collect the tissues, but there still exists the false-negative due to the
sampling error during the biopsy operation. It happens even though the technicians performing
the ultrasound guided needle biopsy keep in the mind to be very careful to reduce the sampling
errors during the operations. So, thousands of the over 1.6 million women get the false-
negative diagnosis although they have obviously the tissues inside of breast in progress. This is
not the mistake or false of radiologist and pathologist, but the reason is because the current
clinical ultrasound system‟s capability has a limitation to visualize the small lesions like micro-
calcifications(<100um) which maybe be a sign of precancerous grouped cells or early breast
cancer tissues if they appear in certain patterns and are clustered together. Conventional clinical
ultrasound imaging has been regarded as a promising tool to guide percutaneous needle to the
targeted location, but the operating frequency range (<20MHz) is pretty low and it is difficult
to detect the micro-structured objects or targets including micro-calcifications. Basically, the
lateral resolution depends on the operating center frequency and the focal number to make the
beam formed from the array transducer. So, the micro-calcification‟s size is less than 100um
and the higher operating frequency (>30MHz) is required to visualize it with single element
transducer or array transducers. And, an imaging system such as Vevo 3100 (VisualSonics,
Toronto, ON, Canada), which is one of the commercial ultrasound imaging system to support
up to 70MHz, is required. The commercial system is an opened system which can manipulate
the transmitting and receiving performance and is used for the various applications.
5
Figure 3. Vevo 3100 (VisualSonics, Toronto, ON, Canada)
If we use the increased operating frequency, the product would be a powerful solution. But,
there is an additional limited requirement when we consider the needle biopsy application. This
is because the attenuation is increased by the higher operating frequency and it is not suitable to
see below several centimeter, so we need to locate the transducer at the point of lesions so that
6
the transducers get the signal much less attenuated from the target tissue nearby the biopsy
needle. And, it was proposed to integrate the transducer into the biopsy needle from Dr.
Shung‟s group recently. The designed linear array‟s target is to insert the array into the biopsy
needle (11 gauge) so that the size is quite small (1.2mm x 1.2mm) to fit the array transducers‟
size less than inner diameter(2.38mm) of the biopsy needle.
Even though the integration was successful, it still has another limitation. The bunch of cables
are required more space to connect the elements to the system outside of the needle. As more
ultrasonic array‟s elements are required to image more range, there is also a critical space
limitation due to the cables‟ thickness. To overcome the limited catheter space (<2.38 mm :
biopsy needle‟s inner diameter), a custom-built & high voltage analog multiplexer is proposed
in order to integrate array transducer into the biopsy needle and to connect the array
transducers to the outside‟s system. Fig.4 shows the suggested ultrasound-guided biopsy needle
approach with the supplementation. For enhancement of this proposal, the external linear array
still helps to locate the biopsy needle to the targeting point and the guiding external array has a
visualized limitation due to the spatial resolution with lower operating frequency and the
attenuation by increasing the operating frequency. Due to these factor, there is a possibility to
occur the false-negative biopsy operations and it results in sampling error.
To avoid and to resolve these issues, it is suggested the biopsy needle integrated with the high
frequency array and with the integrated circuit into the needle together. At first, the high
frequency internal array will help the performance enhancement because it supports to image
small region or dust. It can help to make the biopsy needle operation more precise. Also, the
huge complemented point is that it reduces the number of cables for the operation as well as
7
system‟s complexity so that we can integrate the transducer in to the needle. And, it is easier to
make the whole imaging platform because it requires just single transmitter. Lastly, there is a
analog switch and mux/de-mux for low voltage & high frequency while it is difficult to find the
commercial integrated circuit for the high frequency & high voltage mux/de-mux. So, it has a
potential to be a novel approach applied with high voltage and high frequency ultrasound
imaging platform. For these several advantages, the custom-designed integrated circuit is
proposed and designed to solve the cable‟s limitation so that it can control up to 64 array‟s
elements in this paper.
Figure 4. Suggested ultrasound-guided biopsy
8
1.2.2 Approach
There are several developed analog switches or multiplexer by companies such as Maxim
integrated, Texas Instrument, Microchip and so on. But, the problem is that there is no die
chips or packages to fit to the biopsy needle application and it requires two conditions. First
condition is that the die chip‟s size needs to be tiny so that it can be located into the biopsy
needle. So, the width or length should be less than the inner diameter (2.38mm) of the biopsy‟s
needle. Second condition is that it needs to include the structure of high voltage 1-64 mux/de-
mux which can handle the elements up to 64 or more. Even though the commercial products
which aim to 1-4, 1-8, 1-16 already exist, the size is pretty large and the size does not match to
insert into the needle or the performance like voltage or frequency is not good enough. So, it
requires the two main factors when we design the custom-built integrated circuit for the biopsy
application.
Furthermore, the structure of 1-64 multiplexer supports to transmit and receive the signal to
the single element that it can be switched to the each element fast enough step by step. This
means that it is possible to select single element for transmitting and receiving, but also it can
be utilized to change the transmitting element and the receiving element. In order to satisfy the
condition, a fast rising and falling time are required to change the mode of transmitting and
receiving to reduce the dead zone due to the switching transition time. And then, synthetic
aperture imaging beam-forming technique is selected because it is suitable to support this
mechanism and to minimize the disadvantage of single transmitting and receiving. The
technique considers the time delay from the transmitting element and receiving delay with
received signal from two elements. Each received signal is delayed and summed and one image
9
can be constructed in a result.
1) Synthetic Aperture Imaging system
Beam-forming is one of the crucial processes in the ultrasound imaging system to help the
performance. There are various beam-forming approaches according to the applications and
structures. Among the methods, a synthetic aperture imaging has the advantages to reduce the
system‟s complexity and to enhance the resolution by using dynamic focusing method when
we consider the needle biopsy application.
The synthetic aperture imaging‟s basic idea is to sum the received signals from emissions
close each other. And, Fig.5 shows the mechanism how the elements transmit and receive the
signals step by step. This approach uses the method that one array element transmits an
ultrasound pulse and all elements receive the reflected or backscattered signal. One of
advantages is that it supports the full dynamic focusing range during transmitting and receiving
in order to get the higher quality‟s ultrasound image from the higher SNR(signal to noise ratio)
as much as possible. So, the method uses a number of imaging lines as many as the number of
elements used and requires to calculate the delayed and summed signal so that it is
reconstructed as one image. Also, it enables to reduce the required channels to form the image
and it decreases the system‟s complexity and cost while the conventional system requires
several channels or hundred channels for the transmitting and receiving. From the point of the
system, this is the main reason that the synthetic aperture imaging fits to the biopsy needle
application because the synthetic aperture imaging system needs just one transmitter and
receiver for the data acquisition.
10
Figure 5. Sequence of synthetic aperture imaging
In order to calculate the delay during transmitting and receiving at the each elements, it
requires to consider the factors such as the distance(x
m
,x
n
) from transmitting element and
receiving element to the center, angles (ϴ
m
, ϴ
n
) between the element and the target. Equation.1
and equation.2 represent the transmitting delay(r
m
) and receiving delay(r
n
), respectively. The
total delay is shown as equation 3 and it is applied to the quiation.3 to sum up and to find the
final value at the each pixel. As a result, the summed echo signal A(t) can be expressed and it is
possible to reconstruct the b-mode image after post-processing.
1 2
2
sin 2
c r x r x r r
m m m
(4)
1 2
2
sin 2
c r x r x r r
n n n
(5)
n m n m
r r r
,
(6)
11
1
0
1
0
, ,
2
,
N
m
N
n
n m n m
r
c
r
y r A
(7)
Figure 6. Delay calculation for the A-scan signal of STA Imaging
2) Custom-designed CMOS High voltage 1-64 Mux/De-Mux
Fig 7 shows the proposed structure. In order to image the tissue from the array of biopsy
needle, the array needs to be mounted on the needle and to be connected to the outside of
system. As the number of ultrasonic array‟s elements is increased, there is a critical space
limitation due to the cables‟ thickness. So, this is the reason that a custom-built & high voltage
analog multiplexer is needed to overcome the limited catheter space (<2.38mm : biopsy
12
needle‟s inner diameter). This electronic‟s main purpose is that it enables to reduce the required
number of wires dramatically from 64 to 13 including 5 power wires, 7 wires for controlling
digital bits and 1 wire for transmitting and receiving the signal. So, it enables to reduce the
required cables and to increase the available elements. The integrated circuit requires the
sufficient performance such as 3-dB bandwidth, on-resistance, off-isolation, fast turn-on &
turn-off time.
Figure 7. Designed high voltage multiplexer
1.3. Conclusion
In order to defeat the conventional approach, the novel design is required and suggested to
satisfy the biopsy application. For this application, there are three critical requirements, 1)
higher operating frequency (over 20 MHz) to visualize the micro-calcification, 2) reduction of
cables‟ number, 3) 1-channel system to transmit and receive the signals.
13
CHAPTER 2. DESIGN OF CMOS HIGH
VOLTAGE 1-64 MUX/DE-MUX
2.1 Block diagram
There are three blocks which is made up of the high voltage mux/de-mux such as a low
voltage 6-64 decoder, level shifter, high voltage analog switches and Fig.8 shows the designed
block diagram. Each block requires the sufficient clamped circuit inside to protect the issues of
a latch-up and surge voltage. In order to operate the designed high voltage switches, the
controlling bits from the outside of the integrated circuit are needed for switching-on & off.
The controlling bits control the low voltage 6-64 decoder‟s output at the low voltage
level(0V~3.3V) and the additional level shifter which change the voltage level from 0V~3.3V
to -Vss V~+Vdd V is required to control the high voltage switches. The one node of high
voltage switches is connected together so that the block has one node at the one side and 64
nodes at the other side. As a result, the group of 64 high voltage analog switches looks like a 1
to 64 mux/de-mux. I/O<1:64> is connected to the each element of array and Tx. Rx can apply
the large signal and received the small echoed signal from the target. In other words, tx/rx port
can be switched to I/O<1> to I/O<64>. And, the number of total pins to the outside is reduced
down to 14 compared to the 64. So, the structure enables to minimize the number of I/O which
can require smaller numbers of cable and overcome the space limitation. This is very important
to integrate the array transducer and IC into the catheter inside together.
14
Figure 8. Block diagram of designed circuit
2.2 Low voltage 6-64 decoder
The low voltage 6-64 decoder has several ports such as control bit pins(A,B,C,D,E,F),
enable(EN) as inputs, and 64 output pins (CT_3V<1:64>). In order to design the low voltage 6-
64 decoder, a digital standard library supported by the pdk is used to optimize the fast
switching performance. In this design, it can control the 64 outputs from CT_3V<1> to
CT_3V<64> step by step with 6 control inputs so that the six inputs are able to select just one
output of sixty four outputs which is matched to the mechanism of synthetic aperture imaging .
15
The operating truth table is shown in the table.1 and it represents how to control the decoder‟s
digital output with 6 digital inputs. The decoder operates in the range of low voltage, and the
output voltage level needs to be converted to the to the high voltage level because the voltage
swing is essential to control the analog switches. So, the voltage level shifter is required after
the decoder so that it changes the low voltage input level (0V~3.3V) to high voltage voltage
level(-V
ss
~+V
dd
). It include the „enable‟ port to set up whole switch off even though some
clock is applied to the inputs(A,B,C,D,E,F).
Figure 9. Symbol of 6-64 decoder
A B C D E F EN Switch-ON
0 0 0 0 0 0 0 X
0 0 0 0 0 1 0 X
0 0 0 0 0 1 1 II<1>
0 0 0 0 1 0 1 II<2>
0 0 0 1 0 1 1 II<3>
… … … … … … … …
1 1 1 1 1 1 1 II<64>
Table 1.Truth table of decoder
16
2.3 High voltage level shifter
In general, several types of level shifters are widely used and many type of level shifter have
been proposed. Yashodhan suggested the structure to increase the speed of the level shifter and
Maziyar‟s design focused on the operation at the high voltage range and the reduction of power
consumption. But, there are the points that we need to consider the topologies to apply to my
design. The things is that James‟ structure requires many transistors and Maziyar‟s design
supports the slow rising and falling time(>5us).These are critical factors when we design the
level shifter and apply it to the ultrasound imaging system. So, the novel level shifter structure
is designed to fit to the target to minimize the number of transistor , the rising and falling time,
and layout size.
Figure 10. Symbol of High voltage level shifter
Also, it requires additional consideration for the high voltage operation. There are limitations
to the gate-source breakdown voltage and to the gate-source breakdown voltage for high
voltage transistors. The required maximum gate-source potential range is -20V/+20V based on
the pdk document.
17
Figure 11. Schematic of level shifter
The design approach has 4 different potentials such as Vdd, -Vss, 0V, 3.3V. So The level of
input signal is 0~3.3V and the final goal is to generate or to convert the voltage level to the
suitable level at the port of „CONT_SWITCH‟. In order to satisfy the conditions, the high
voltage nmos and pmos ,which have a tolerance between drain and source voltage up to 50V,
between gate and source up to 20V, are used to convert the control clock signal from low
voltage(0V~3.3V) to high voltage (-V
ss
V~+V
dd
V). At this point, the transistor‟s safe operating
voltage range should be considered to avoid breakdown limitation of the transistors carefully.
The low voltage decoder‟s output is connected to the gates of M1 and M4 and it controls the
18
node of „CON_SWITCH‟ which is the drain node of M3 and M5 which has an enough
headroom for the large voltage swing to control the analog switch. Especially, every transistor
operates without any stress of voltage limitation in the schematic and the complementary
structure help to generate the output‟s fast rising time and falling time even if the final voltage
swing is very large up to 50V.
2.4 CMOS High voltage multiplexer/de-multiplexer
2.4.1 CMOS Analog high voltage switch
In the analog and digital world, MOSFET is generally used as a switch. At the low voltage
swing, one transistor shows it performs as a good switch. But, as the input voltage is increased,
there is a critical issue at the single transistor structure. When the larger input voltage is applied
to the input at the turn-off mode, the transistor‟s pathway will be closed. But, there is an
internal diode in the structure and negative signal will go through the internal diode. To prevent
the unwanted issue, the two source connected nmos pair structure is normally used to design bi-
directional switch. In this structure, the additional diodes are shown as Fig.12 and it helps to
block the unexpected negative signal during turn-off mode. And, there are several requirements
when we design the analog switch. Also, the gate is tied together in order to control the two
transistors at the same time. As the input voltage is increased, the gate-source voltage and the
source-drain voltage limitation are the main considerable factors for the high voltage analog
switch,
19
Figure 12. Schematic of source-connected analog switch
At first, there is a break-down limitation between gate and source at the each channel of the
Mux/De-Mux. The tolerance range is from -20V to +20V from the datasheet. The input signal
goes through „SIG_I/O1‟ to „SIG_I/O2‟ back and forward during turn-on mode. The port from
the level shifter is connected to „CONT_SWTICH‟ for switching on and off the pathway.
When the voltage of „CONT_SWTICH ‟ port goes up to +20V, the available maximum output
voltage range is from -20V to +20V. This acts as „switch-on mode‟. Inversely, when the
voltage of „CONT_SWTICH ‟ port goes down to -20V, the available maximum output voltage
is close to -20V and this meaning is the voltage over -20V cannot go though and be isolated.
20
This process act s as „switch-off mode‟. Additional considerable factor is the transistor‟s on-
resistance in the turn-on mode
Figure 13. Simulation result of the analog switch
T GS ox on
V V W C L R
(8)
According to the equation 8 the on-resistance is dependent on the transistor‟s specifications.
Basically, it is required to use the transistor which has smaller length, larger width, Vgs. In the
point of the transistor‟s layout, smaller channel length is size effective but larger width leads
more layout space. Additionally, there are more and larger parasitic component by increasing
the transistor‟s size and it would reduce the operating bandwidth. So, it is very important to
consider the trade-off relationship of the transistor. Fig.13shows the simulation result of the
switch‟s turn-on & turn-off performance. Initially, the voltage of „CON_SWITCH‟ is down to -
21
20V. When the voltage is changed to +20V, it operates at the turn-on mode and the input signal
of „SIG_I/O1‟ port goes through to „SIG_I/O2‟ port. If the „CON_SWITCH‟ goes down to
again and there is the input signal, the signal pathway is blocked and there is few signal
amplitude to the „SIG_I/O2‟ port due to the switching off the transistors.
2.4.2 CMOS high voltage 1-64 Mux/De-mux
Figure 14. Cross sectional structure of HV CMOS 0.35um(AMS H35D4D3)
AMS high voltage process(H35B4D3) is selected to design the integrated circuit. It can
support the scalable high voltage LDMOS device architecture including deep n-well guarding
protection so that it operates properly from low voltage to high voltage and it isolates the
transistors and modules effectively. Also, it provides the analog and digital library based on
BSIM3 to apply the actual transistor‟s performance as much as possible. Especially, the high
voltage transistor‟s gate-source voltage limitation is up to -20V~+20V which is one of the
critical factors to design the high voltage analog switch and the multiplexer. Cadence, assura
22
(Cadence inc.) and caliber(Mentor graphics inc.) are used to simulate and to layout the
integrated circuit. After checking DRC(design rule check) and LVS(layout vs schematic), the
post-simulation is performed and applied after layout. It has 78 pads including 5 pads for power,
64 pads for array elements, 7 pads for digital I/O, and 1 pad for TX/RX. All individual pads
have the protection circuit inside to avoid the ESD issues. It is very important because the high
potential electrostatic discharge can break the metal lines and connections and it makes the chip
unavailable even though the integrated circuit is well-designed. Here, one thing is that the pad
requires certain surface area for wire bonding or flip chip bonding, and then it has a parasitic
capacitor which degrades the performance at the high frequency. So, we need to be careful
about the trade-off during designing the circuits.
Figure 15. Top view of designed circuit
23
2.5 Clamped circuits for latch-up protection
Figure 16. HV concept with isolated LV-logic
To prevent the modules which have the different operating voltage swing level, designers
need to follow the manufacturer‟s design guidance. Fig.16 shows how to set-up and to isolate
the low voltage modules from high voltage modules together. The clamping modules
(CC3/5,CC10-120) prevent from the latch-up problem which is very crucial at the high voltage
operating and the forward diode(DF_HV) is required to make the separated modules stable. In
this structure, it is possible to change the substrate‟s potential by manually and it is
manipulated to fit to the multiplexer design. For the actual design, there are three potential
ranges; -Vss~+3.3V, 0V~+3.3V,0V~+Vdd.
24
Figure 17. Finalized potential map.
Figure 18. Detailed structure
25
2.6 Layout
The layout size is one of the most importance design factors. The application supplies the
limited space to fit into the needle‟s inner diameter, so the width or length should be less than
the inner diameter. Additionally, pad-ring size should be considered as a basic frame because it
generally occupies large space including additional protective or assistant circuits inside. And,
more space margin should be better to bond the chip through wire-bonding or ball-bonding in
order to integrate the circuit on the board. In this design, the high voltage circuit requires more
area and it determines the total layout‟s size. For the high voltage protection, wider(>10um)
guarding option is selected.
Figure 19. Layout of 6-64 decoder(left), 1-cell with high voltage level shifter and analog
switch(right)
26
Figure 20. Layout of core with pad-ring(6mm x 1.6mm)
Assura(Cadence Inc.) and Caliber(Mentor graphic Inc.) are used for the layout and post-
simulation, As mentioned before, the size is critical and limited in order to fit into the needle
inner diameter. So, the high voltage switches are placed widely with 1.6mm limitation And, the
low voltage input buffers are added to the every digital input to sharpen the trigger‟s rising and
falling time
2.7 Die chip
Figure 21. Die chip (6mm x 1.6mm)
27
Fig.21 shows the die chip photo after fabrication. The size is same as the designed layout and
the thickness is 0.2mm by the CMOS process. To prevent the ESD issues, it requires to follow
the protocol to handle the integrated circuit and the pad size is 120um x 120um.
2.8 Performance
Figure 22. MQFP packaged chip(12mm x 12mm)
In order to test the performance, the chip is packaged with MQFN which can support up to
100pins with gold wire-bonding connections. The package‟s size is 12mm x 12mm and it is
mounted on the PCB(print circuit board). PC communicates with FPGA board in real time
which purpose is to control 7 digital input pins of the integrated circuit simultaneously. Power
pins are connected to the commercial power supply to provide 5 different potentials
(Vdd,Vss,0V, 3.3V) for the analog circuit and the digital logic circuits. And, BTH 150-01-F-D-
A(Samtec Inc.) adapter is attached to connect the board and the array.
28
Figure 23. Testing bench
For the digital control, a FPGA board (Atlys, Diligent Inc.) is selected because it support fast
clock speed up to 400MHz and it also has the communication protocol such as USB, UART,
I2C and so on. From fig.23, the FPGA board controls the operation to select the channel. It has
6-control bit and 1-enable bit to control 64 channels with 3.3V clock input with internal input
buffer. +VddV, -VssV, +3.3V, 0V are applied for the chip operation and the maximum input is
limited by the range of +VddV~-VssV. The testing condition is set up likely to the transducer,
so 380ohm & 10pF, which are expected equivalent circuit‟s components, are attached to the
load and the voltage is applied up to 36V for the test. The commercial function generator
(AFG3252,Tektronics) and power amplifier(ar,amplifier research,10WD1000) are used to
generate the large signal. And, the signal is capture by the oscilloscope(TDS5052,Tektronics)
29
for the initial testing.
2.8.1 On-resistance, insertion loss, noise figure
Figure 24. turn-on mode(left), turn-off mode(right)
Fig.24 shows the measured results during turn-on & turn-of off mode. The input signal
operates at 60MHz with 36V and the output signal is 24.4V through the one channel switch.
This is because on-resistance is shown during the turn-on mode and the attenuation is happened.
During turn-off mode, the pathway needs to be closed ideally. As the operating frequency is
increased, however, it is increased the effect of the additional parasitic capacitances which are
from the transistor‟s Cds, and internal diode‟s capacitance. This is the reason why there is a
certain amplitude, 1.7V, from the output evenly in the turn-off mode. Additionally, the on-
resistance is measured by change the input amplitude with fixed frequency. Fig.25 & Fig.26
shows the insertion loss and on-resistance with input variation (0.1V,10V,20V).
30
Figure 25. Measured on-resistance(input voltage=0.1V, 10V, 20V )
Figure 26. Measured insertion Loss(input voltage=0.1V, 10V, 20V )
10
0
10
1
10
2
50
100
150
200
250
300
350
400
Frequency[MHz]
On-Resistance[ohms]
Frequency[MHz] vs On-Resistance[ohms]
Input Voltage=0.1V
Input Voltage=10V
Input Voltage=20V
10
0
10
1
10
2
-5
-4.5
-4
-3.5
-3
-2.5
-2
-1.5
-1
Frequency[MHz]
Insertions Loss[dB]
Frequency[MHz] vs Insertion Loss[dB]
Input Voltage=0.1V
Input Voltage=10V
Input Voltage=20V
31
Figure 27. Measured noise figure(input voltage=0.1V, 10V, 20V )
It represents that the on-resistance is increased up as the input amplitude is larger and it is
relative to the nmos property. And, fig.26 says that there is change of off-isolation with the
different input voltage.
G kTB
S G
kTB
S
SNR
SNR
F
output
input
1
2 2
(8)
It can be derived with an input signal to noise ratio(SNR
input
) and an output signal to noise
ratios(SNR
output
). It includes the terms of signal power(S), flank constant(k), temperature(T),
bandwidth(B). But it is cancelled out and the noise factor is expressed as the inversed gain(G).
And, a noise figure is the noise factor which is expressed in the dB scale as shown as Eq. 8.
32
2.8.2 Off-isolation
Figure 28. Frequency vs Off-Isolation(input voltage=0.1V, 10V, 20V)
There is no critical change with different input voltage amplitude. The tendency is that the
off-isolation is gradually worse as the frequency goes up. This is because the structure of the
analog switch which includes the internal diodes inside and the diodes‟ impedance is getting
smaller. So, additional circuits or structures are good solution to enhance the performance at
the higher frequency. A considerable factor is that the peak to peak amplitude is used to
measure the off-isolation and the signal‟s spectrum during turn-off is larger than 100MHz at
least. This means that there is little effect from the isolated signal because the operating
frequency is pretty low compared to the spectrum‟s frequency range.
10 20 30 40 50 60 70 80 90 100 110 120
-55
-50
-45
-40
-35
-30
-25
-20
-15
-10
Frequency[MHz]
Off-Isolation[dB]
Frequency[MHz] vs Off-Isolation[dB]
Input Voltage=1V
Input Voltage=5V
Input Voltage=20V
33
2.8.3 Total harmonic distortion (THD)
Figure 29 Measured total harmonic distortion.
Additionally, a total harmonic distortion(THD) is a considerable factor to evaluate an analog
switch or multiplexer. Fiq.16 shows the measured total harmonic distortion(THD). THD is
defined as the squared root of sum of all the squared harmonics‟ magnitude divided by
fundamental‟s magnitude as shown as Eq. 9. And, the result shows that there is a variation with
frequency and input voltage amplitude and it can go up to 0.05%.
% 100
1
5 . 0
2 2
4
2
3
2
2
a
a a a a
THD
N
(9)
34
CHAPTER 3. DESIGN OF 1-CHANNEL
TRANCEIVER FOR SYNTHETIC APERTURE
IMAGING PLATFORM
3.1 Introduction
Ultrasound wave can propagate in tissues such as water, human‟s body and so on. In order to
generate the signal and to apply the signal to the transducer, ultrasonic transducers are used
because it can convert from electrical energy to mechanical energy, vice versa. Normally, there
are two type of signal which are used to the ultrasound imaging system such as pulsed signal
and continuous signal. A pulsed signal is generally used to image with same transmitting and
receiving transducer. Also, it has an advantage that the axial resolution is better when smaller
number of cycles is applied. On the other hand, continuous wave is used when the transmitting
transducer and the receiving transducer are different. For these cases, a large input signal(>10V)
is applied to the transducer to get the large pulse-echoed signal and a pulse with large output
amplitude or a power amplifier with function generator are used to apply the signals. Also, the
pulsed-echo signal‟s amplitude is very tiny from the tissues and the pre-amplifier is required as
a low noise amplifier. There are commercialized product like panametric5900(olympus),which
can generate a short negative pulse(amplitude:-100V, 3-dB bandwidth:200MHz), and
ar1000(amplifier research),which operates from 1 MHz to 1000 MHz and can generate up to
100V
pp
.These show the good performance but there are two limitations for the medical
35
ultrasound imaging system. First, the commercial power amplifier has a good linearity in
normal because the power amplifier‟s topology is selected as class A, but there is an amplified
noise level(>20mV) even though the signal is not applied and it cannot be ignored. This is very
important because we usually analyze the received signal after transmitting. In other words, the
system gets the received signal which is added with unwanted noise signal as well. So, input
signal is required to have low noise level as well as large amplitude ideally and it require other
approaches to reduce or to remove the noise level. Secondly, the commercial products are too
bulky and expensive in order to integrate as a system. Also, it is not easy to set- up multiple
channels for the arrays. So, the 1-channel cost-effective transceiver for synthetic aperture
imaging is proposed in this chapter. It has an unipolar pulser and a Class B power amplifier to
support the imaging approach. To make a sinusoidal single cycle or multiple cycles,
DAC(digital-to-analog converter) is widely used. But, FPGA(field-programmable gate array)
and band pass filter are used to generate the input signal instead of DAC(digital-to-analog
converter) or commercial function generator in order to reduce the system‟s complexity and
power consumption,.
36
3.2 Signal Generation based on FPGA (field-programmable gate
array)
In general, a DAC(digital-to-analog converter) or commercial function generator are used to
generate the signals. It has a capability to make the arbitrary waveform including chirp signal
but it requires more complicated electronics and it consumes more power to generate the signal
according to the bit resolution and sampling rate. In this section, the signal generation without
any function generator or DAC is used. It makes just a positive short pulse or square
wave(0V~3.3V) but the electronics change the output to be amplified as a short large
amplitude(amplitude>48V,3-dB>120MHz) or a sinusoidal waveform(amplitude:48V
pp
,3-
dB<200MHz).
Figure 30. Atlys FPGA board(produced by Diligent)
37
Figure 31. Operation blocks
For the system design, atlys(produced by Diligent) is selective. The Atlys circuit board is a
complete, ready-to use digital circuit development platform based on a Xilinx Spartan-6 LX45
with speed grade level - 3. The large FPGA and on-board collection of high-end peripherals
including Gbit Ethernet, HDMI Video, 128MByte 16-bit DDR2 memory, and USB and audio
ports make the Atlys board an ideal host for a wide range of digital systems, including
embedded processor designs based on Xilinx's MicroBlaze. Atlys is compatible with all Xilinx
CAD tools, including ChipScope, EDK, and the free ISE WebPack™, so designs can be
completed at no extra cost. And, it can generate the clock with PLL up to 1040 MHz which is
good enough to make the short pulse. As a result, UART communication is selective to control
the board and VHDL is used to program the signals.
38
Figure 32. PLL primitives
3.2.1 Short pulse generation
The board has a 100MHz CMOS oscillator as a reference clock. And, the board can support
two PLL and one DCM to generate the clock with PLL up to 1040 MHz. And, there is one
method to generate a short pulse. It uses one input and delayed input with and gate and the and
gate output will show the overlapped time area between the input and the delayed input,. Fig.
33 shows the schematic how to make the short pulse with delay line. And, fig.34 is the result of
the schematic where t
d
is the overlapped time area and V
pp
is the amplitude of the output
voltage. So, it is possible to generate the short pulse up to
39
Figure 33. Schematic to make short pulse with delay line
Figure 34. Input & output signals of the and gate
40
3.2.2 Sinusoidal wave generation
FPGA can generate digital signal which can be called as pulse or square wave. It has a certain
frequency and amplitude (0~3.3V) but it is not enough to apply to the transducer. So, the signal
amplification and deformation are required. In order to amplify the signal easily, the signal is
converted from square wave to sine wave. To make sufficient waveform, a band pass filter is
used because square wave includes several harmonics but it is possible to get the sine wave to
filter out the harmonics except fundamental frequency.
Figure 35. Sine wave generation with band pass filter
41
3.3 Transmitter
The generated short pulse and sinusoidal waveform have good enough shapes but it requires
larger amplitude because the FPGA board only support up to 3.3V. So, the higher amplitude is
necessary to get the larger pulse-echoed signal through the transducer. To enhance the echo
signal‟s amplitude, there are two approach to apply the large amplitude‟s waveform : pulser,
power amplifier with function generator.
3.3.1 Unipolar Pulser
An unipolar pulser are generally used to the ultrasound applications. Normally, it has negative
short pulse because the transducer‟s outside is connected to the ground and the negative spark
generate the waveform forward from the transducer. There are several products such as
5800PR(Olympus), 5900PR(Olympus), 5627RPP-1(Olympus), DPR300(JSR), DPR500(JSR),
USC-UT350(Ultratek) according to the operating frequency and the generated amplitude. Short
pulse can cover from low frequency to the certain higher frequency and it is a good solution if
the customer does not know the operating frequency exactly. But, it is bulky and expensive and
it is not suitable to the certain applications which require the multiple transmitters. So, this is
the reason to design and to build the custom-designed unipolar pulser. And, the performance is
shown as fig. 36 and table.2. For this design, two power nmos transistors(M1,M2) are used to
amplify the input signal into output because it required the high supply voltage(V1:48V
DC
).
Two different voltages (V2,V3) are applied from the regulators and the resistors(R1, R5) make
the stable biasing condition. Capacitors (C1,C2, C3, C4) are used for two reasons, impedance
matching and blocking the dc voltage. Also, the resistors (R2,R3,R4) keep the circuit‟s stability.
42
Finally, the input voltage amplitude is amplified up to 48V.
Figure 36. Schematic of unipolar pulser
Specification
Type Unipolar Pulser
Output[V] -48V
-3dB bandwidth 120MHz
-6dB bandwidth 175MHz
Table 2. Specification of the designed unipolar pulser
43
3.3.2 Power Amplifier
There are several types of power amplifiers according to the bias condition and the matching
technique. And, there are trade-off relationships to the properties. For example, a class A power
amplifier has a linearity but I also has disadvantages such as a permanent high noise level from
the output and low power consumption efficiency. On the other hand, Class F supports a low
power dissipation and filters out the odd harmonics with the impedance matching but it has a
low linearity. So, it is very important to select the suitable power amplifier to the certain
applications.
3.4 Protection Circuits
Transmitter‟s output is connected to the input of the pre-amplifier, but the large amplitude
cannot allow to the pre-amplifier‟s input because of the transistor‟s property limitation. So,
additionally electronics are required to protect the transmitter and receiver each other.
3.4.1 Expander
Normally, two high speed recovery diodes are used to design expander so that it pass through
the large voltage swing from the transmitter but the small signal(<0.7Vpp) cannot go through
the expander because of the knee voltage. So, the expander is the disconnected path to the
small signal‟s view and it makes the stable operation. Additionally, its role is to decrease the
noise level relatively compared to the large signal swing. These are the reason why general
ultrasound transceiver includes expander circuit inside. And, fig.37 shows that expander‟s
schematic.
44
Figure 37. Schematic of expander
3.4.2 Limiter
While the expander protects the transmitter, it is not enough to protect the pre-amplifier‟s
input from the transceiver. At least, it recommends to clipping the input voltage amplitude up
3~5V for the safety issue. In Fig.38, D1 clamps the voltage level down to -0.7V while D2
clamps the voltage level up to +0.7V. As a result, the available voltage swing becomes between
-0.7V and +0.7V so that the voltage level is not harmful to the pre-amplifier‟s input.
Figure 38. Schematic of limiter
45
3.5 Receiver
The pulse-echoed signal is generally very small and it is hard to detect with digitizer or
oscilloscope directly. So, a small signal amplifier is required to magnify the received signal. In
this case, we need to think about two things such as sufficient gain and low noise figure. Noise
figure is dominant on the first stage mainly, so the correct choice of pre-amplifier is needed.
Among the commercialized product, PSA4-5043+(minicircuit inc.) is chosen because it has
high enough gain at the wide bandwidth and the noise figure is less than 1 dB.
Figure 39. PSA4-5043 based ultra-low noise MMIC amplifier
3.6 Designed Modules
For the general ultrasound applications, there are two main things we need to consider. First,
it requires a wide-band operating frequency range because this supports a good amplification
without less signal distortion. Also, the noise level from the power amplifier should be as small
as possible. This is the reason why it affects to the echoed-signal after the large signal
transmission and it results in the worse image quality. So, the two specifications are considered
in this design.
3.6.1 Transceiver module1 base on Class A power amplifier
For these design, two power nmos transistors(M1,M2) are used to amplify the input signal
into output because it required the high supply voltage(V1:28V
DC
). Two different voltages
46
(V2,V3) are applied from the regulators and the resistors(R1, R5) make the stable biasing
condition. Capacitors (C1,C2, C3, C4) are used for two reasons, impedance matching and
blocking the dc voltage. Also, the resistors (R2,R3,R4) keep the circuit‟s stability. Finally, the
input voltage amplitude is amplified up to 56V ideally because the inductive load capable to
make the ac voltage swing twice than the supply voltage. Capacitors (C1,C2,C3,C4) are used
for impedance matching and blocking dc voltage.
Figure 40. schematic of class A power amplifier
47
Figure 41. Designed module based on the class A topology
Pin Description
1 Tx Input
2 Tx Output
3 Connector to the transducer
4 Rx output
5 Power Connector
Table 3. Pin description
48
Figure 42. Power amplifier‟s gain
Figure 43. Diagram of 2-stage pre-amplifer
In order to get enough gain, 2-stage cascade structure is chosen. It has 5V supply voltage and
the capacitors(C1,C2,C3)‟ main role is to block the dc level between the modules‟ input and
output. And, the resistors(R1,R2) control the current consumption and it control the amplifier‟s
gain. Fig.44 shows the measured gain performance which has 31~33.7dB below 40~120MHz.
49
Figure 44. Pre-amplifier‟s gain
3.6.2 Transceiver module2 base on unipolar pulse and Class B power
amplifier
In this design, a push-pull structure is used to amplifier the input amplitude with rare noise
level. The upper structure and the lower structure divide the input signal into two parts, positive
amplitude and negative amplitude, and then amplify separately. After TR1 transformer, the
input signal is divided into two out-phase signals and the transistors, M1 &M2, amplify them.
Because the divided signals have a negative amplitude, transformers(TR2, TR3) are added to
flip the signal in to positive amplitude. Thrtough M3 and M4, the same process is applied and
the final two signals are mixed with the transform,TR4. The supply voltages (V3,V4,V5,V7)
are same as 48V so that the maximum output voltage is 48V with 50 ohms. The advantage is it
has rare noise level compared to the class A power amplifier as well as no power dissipation
without input. So, it is very stable during long operation. And, it can operate over 200MHz.
50
Figure 45. Schematic of Class B push-pull power amplifier
51
Figure 46. Gain of Class B power amplifier
Figure 47. Unipolar pulser and Class B power amp based on FPGA signal generation.
52
3.7 Digitizer
In general, ADC(analog to digital converter) is used to record the received signal. The ideal
ADC for the ultrasound applications needs high resolution and high sampling rate with very
low total harmonic distortion and good SNR. In the point of image, a high resolution(>8 bits) is
good enough because it helps to recognize the boundary sensitively and also it reduces the
minimum noise level which determines the minimum detectable level. According to the
Nyquist theorem, the sampling frequency should twice than the fundamental frequency at least
and three time or four times of the center frequency is sufficient to avoid the issues.
In order to satisfy the requirements, a 12- bit digitizer(model name:cs122g1) is selected. It has
1channel input which can support the sampling frequency up to 2GS/s. Its bandwidth is
500MHz and performs up to 9.5ENOB at 200MHz. And, it supports user interface to
communicate by matlab, labview, and c/c#.
Figure 48. PC based digitizer(gagecard, cs122g1)
53
CHAPTER 4 ULTRASOUND SYNTHETIC
APERTURE B-MODE IMAGING
4.1 Introduction
B-mode image is the most general image from the ultrasound imaging systems. It can show
the tissues and organs with received signal‟s amplitude strength which can be represented by
gray scale. In the B-mode images, the black region is the tissue which the received signal is
relatively low while the white section means that the area which it has the stronger received
amplitude. Fig.49 shows the steps how to get the b-mode image. At first, the ultrasound
imaging system requires the pulse echoed signals from the single element transducer or array
transducers. Normally, a high bit resolution digitizer such as ADC (analog to digital converter)
oscilloscope is required to get the signals. It needs enough bandwidth and sampling frequency
to receive the signals without any distortion or aliasing. A post-processing enables to make
proper and enhanced images with the signals from the transducers. At this step, the signals are
delayed and summed to focus dynamically at the every pixel points and the summed signal
makes a final image after envelope detection and log compression. And, the post-processing is
performed with MATLAB(mathworks).
54
Figure 49. Post-processing of synthetic aperture imaging
4.1.1 Testing bench
The custom-built modules are integrated together to show the performance. It is made up of
30MHz linear array, FPGA board, Class-B power amplifier and the custom-designed high
voltage mux/de-mux. And, an agar phantom is use for the imaging.
Figure 50. Testing bench
55
4.1.2 Data acquisition
Basically, PC controls the output pins of the FPGA board so that it generates the seven trigger
signals(<3.3V). And, it is switched to the large voltage at the stage of voltage level shifter. As
the pc controls the integrated circuit, it leads the high voltage switches opened and closed
continuously. By the transmitting and receiving sequence, the pulsed-echoed data is recorded
by digitizer(Gagecard, CS121). Finally, it is saves as listed files and the
software,Matlab(Mathworks), do the post-processing to make B-mode images.
Figure 51. Data acquisition process
56
4.1.3 Labview GUI
In order to control the setting environment, labview(national instrument) is selected. It is a
system design platform and development environment for visual programming language, so it
supports a powerful user-interface. To capture the data, it has functions 1) to select the number
of element which control the high voltage multiplexer with FPGA board, 2) to average the
received signal in order to reduce the noise level, 3) to filter out with FIR band pass filter. And,
fig.52 shows the designed GUI with labview.
Figure 52. Labview GUI(graphic user interface) to capture the signal
57
4.2 B-mode imaging after post processing of synthetic aperture
imaging
For the testing, 30MHz linear array is used. It has 256 elements but 64 elements are just
selected to test the system. Testing condition is to input 30MHz signal with 24V and to amplify
the received signal with additional external amplifier so that the total gain is 50dB, to set up the
running average the factor of 10 in order to reduce the noise level
Figure 53. 30MHz linear array
Input Specifications
Center Frequency 30 MHz
-6-dB bandwidth 45%~55%
Pitch 50um
Number of Array 256
# of Average 10
Table 4. Specification of 30MHz linear array transducer
58
1) Quartz
Normally, quartz is used as a perfect reflector and the pulse-echoed signal can be considered
as an impulse response of the transducer. In this experiment, 8 elements are selected to make an
image. And, fig. 54 shows the beam-formed data and b-mode image of the quartz target. The
array‟s pitch and the data sampling frequency are calculated and the data is acquired after
delayed and summed process. It shows that the signal‟s amplitude is highest at the surface of
the quartz and the reflected signal‟s amplitude are relatively high compare to the noise level
because it reflects most of signal at the surface.
Figure 54. Beam-formed data after post-processing(left), b-mode image of quartz(right)
3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Original Image
Depth[mm]
Envelope Value
Original Image
Lateral[mm]
Axial[mm]
-0.2 -0.1 0 0.1
5
5.5
6
6.5
7
7.5
8
59
2) Wire phantom
After checking the algorithm with the quartz target, a 25um tungsten wire is used to determine
the spatial resolutions such as axial and lateral resolution. And, fig.55 shows the envelope
values after post-processing. The dynamic range is given to 30dB. From the calculation of the -
6dB range from the peak point, the results show that the lateral resolution is 392um and axial
resolution is 90um.
Figure 55. Beam-formed data after post-processing
6.8 7 7.2 7.4 7.6 7.8 8 8.2 8.4
-50
-40
-30
-20
-10
0
10
20
Original Image
Depth[mm]
Envelope Value
60
Figure 56. B-mode of 25um tungsten-wire
Original Image
Lateral[mm]
Axial[mm]
0 0.5 1 1.5
7
7.5
8
8.5
9
9.5
10
61
3) Metal Frame
Additionally, a metal frame is used to verify the performance. It is a good structure to check
the synthetic aperture‟s artificial because the transmitting signal is reflected by the angled
structure. Left figure of Fig.57 shows the frame‟s boundary and the right figure represents the
b-mode image from the metal frame. It cannot show the bottom boundary near 7mm with
dynamic range=26. While it shows the bottom boundary with dynamic range=36, but there is
an additional artificial image near 5mm.
Figure 57. B-mode image of metal frame
62
4) Agar Phantom
The biological tissue has been made and mimicked by various materials, which have shown
certain similarity and property compared to the human tissues. Agar phantom is one of the
general mimic phantoms to the ultrasound imaging system verification. It is acceptable fidelity
to real tissue, cost-effective and easy to calculate the phantom‟s stiffness. It is made up of agar
power, silica power as a scatterer, water, and n-propanol to make 10kpa with a phantom recipe.
According to the results, the system is capable to image the tissues which are close to the
human tissues.
Figure 58. B-mode image of agar phantom(left, rgb scale), agar phantom(right, grayscale)
63
4.3 Coded excitation and pulse compression technique
In the ultrasound imaging, there is a trade-off between penetration depth and resolution. Also,
the penetration depth can be improved to increase the input pulse‟s amplitude or the pulse
duration. However, there are limitations to increase two factors: 1) most of commercial
ultrasound imaging systems reach to generate the peak amplitude which lead the maximum
mechanical index, 2) the longer pulse duration gives better penetration depth but it makes the
axial resolution worse. So, the coded excitation and pulse compression technique was
suggested to compensate two factors.
The coded excitation has been used in radar for a long time to increase the signal to noise
ratio(SNR) and it is applied to the ultrasound imaging system successfully. The basic principle
is same between ultrasound and radar but the propagating mediums are different. The
advantage is that we can keep the excitation peak amplitude and the axial resolution as well as
increase the penetration depth because the transmitted pulse‟s energy cannot be increased
beyond a certain threshold level for the medical ultrasound application. So, And, O‟Donnell‟s
group already showed that the transmitted energy can be increased up to 15~20dB by using
coded excitation before the limitation of intensities (I
sppa
, I
spta
), mechanical index and other
factors. After the demonstration, there are many approaches to enhance and to utilize the
technique for ultrasound applications. One of the approaches is to use the technique for the
elastography so that we can get the clearer images from the target tissues. Other approach is to
improve the SNR as well as the axial resolution by using mismatched matched filter. So, the
technique results in the improvement of signal to noise ratio(SNR), contrast to noise
ratio(CNR), axial resolution. And, the lateral resolution can be enhanced when we select proper
64
tapering window for the coded signal.
Figure 59. Chirp excitation (linear frequency modulation)
One of method to increase the time duration-bandwidth product of a waveform is to use the
frequency modulated signal. And, a typical chirp signal is a linear frequency modulation
waveform that the frequency is changed linearly with time t, upper frequency f
2
, lower
frequency f
1
, bandwidth B =f
2
-f
1
, time duration T, tapering window function w(t), coded signal
C(t). And, the chirp signal is expressed by the following equation
( ) ( ) (
) (10)
In order to decode the received signal, the matched filter is used and it can be expressed
( ) ∫ ( ) ( )
(11)
65
where is a time shift. Especially, the signal to noise ratio(SNR) improvement is calculated
by the two factors such as time duration T, band width B which is called time-bandwidth
product(TBP). The decoded signal‟s SNR is TB times higher than the un-modulated pulse.
Also, the output signal‟s time duration means the axil resolution and it depends on the
bandwidth. So, the wider bandwidth results in the better axial resolution.
Figure 60. Diagram of coded excitation and pulse compression to get B-mode image
Fig.60 shows the process of the coded excitation and pulse compression. Basically, the coded
signal is applied to the transducers instead of the conventional short pulse. And, the additional
66
different thing is to use the matched filter after the data acquisition. The matched filter is the
optimal linear filter for maximizing the signal to noise ratio(SNR) in the environment with
additional noise. The basic concept is that we can measure the reflected signal and can look for
similarity with what we transmitted.
Figure 61. b-mode image from 50um tungsten wires, 1-cycle excitaion(left), coded excitation
before pulse compression(middle), coded excitation after pulse compression(right), the given
dynamic range is 20dB
Even though the given dynamic range is 20dB to the three figure, the signal with 1-cycle
excitation is pretty weak. One interesting factor is that the signal looks strong with coded
excitation technique but there is no dramatical change of the SNR between the conventional
67
excitation and coded excitation before pulse compression. However, the image quality and the
SNR is enhanced after pulse compression which factor is increased from 14.77dB to 31.53 dB.
And, there is about 15dB improvement as calculation.
Figure 62. b-mode image from 10kPa agar phantom, 1-cycle excitaion(left), coded excitation
before pulse compression(middle), coded excitation after pulse compression(right), the given
dynamic range is 20dB
Also, 10kPa agar phantom is used for the additional demonstration. The given dynamic range
is same as the previous images. The received signals are delayed summed but the signal to
noise ratio is not high enough with 1-cycle excitation. So, this is the reason why there are
speckle patterns but the image quality is pretty bad and it is hard to recognize the tissue. Bu
using the coded excitation, the top layer and bottom of the agar phantom is shown at least but
the image quality is still not enough. After the pulse compression, we can see the clear speckle
68
pattern and the top layer of the agar phantom because the signal to noise ratio and contrast to
noise ration are improved compared to the previous two images. So, it shows that the beam-
forming and the pulse compression are properly performed..
4.4 Additional beam-forming to image wider area
The beam-forming for the conventional linear array is performed but there is still the
limitation. There are limited scanlines for the linear array and it has narrow imaging range
compare to the phased array. However, the acoustic wave propagates not only to the straight
but also to the all directions. So, the additional imaging area is calculated with the received
signals. Figure. 63 shows the different beam-forming approach. Left figure represents the
conventional beam-forming which right figure shows the additional beam-forming to see wider
imaging area
Figure 63. conventional linear array beam-forming(left), additional beam-forming(right)
69
Figure 64. Linear array (# of scanline: 32, total imaging range 1.6mm)
Fig.64 shows the image with conventional approach. In order to get the b-mode image, 36
elements, which supports 36 scanlines, are used and the total imaging range is 1.6mm. This
means that we can get the image only from the area below the array. But, it is possible to show
the additional area if the scanlines are increased up to 64 or 96. And, fig.65 and fig.66 show the
results with the novel approach.
Figure 65. Linear array (# of scanline: 64, total imaging range 3.2mm)
70
Figure 66. Linear array (# of scanline: 96, total imaging range 4.8mm)
71
CHAPTER 5 SPECTRAL ANALYSIS FOR
TISSUE CHARACTERIZATION: INTEGRATED
BACKSCATTERING COEFFICIENT (I.B.C) &
ATTENUATION SLOPE
5.1 Introduction
The precise biological tissue characterization has been a hot topic to the ultrasound imaging.
The properties give stronger information and contrast dependent on the biological structure and
tissue type. Among several approaches, ultrasound spectral analysis in the frequency domain
gives more information from the received signals or images. The calibration process reduces
the system‟s artifacts and noise and gives the spectral quantitative measurement result. The
calculated integrated backscattering coefficient can be used as a new data set instead of the
conventional b-mode images.
The integrated backscatter coefficient uses a frequency domain averaging technique and it
reduces the effects of frequency dependent fluctuations due to inhomogeneous scattering
distributions in tissues. The calculated coefficient is defined as the ratio of the backscatter
energy in the frequency domain from a scattering particle or volume and the reference signal‟s
energy from the flat quartz or glass reflector which are regarded as the perfect reflectors over
the certain bandwidth.
72
f f
f
c
c
df
f R
f V
f
IB
2
1
log 10
10
(12)
Where R(f) and V(f) are the spectrum of the reference and sample signals, fc is the center
frequency of the transducer and f is the bandwidth of R(f).
73
5.2 Integrated backscattering coefficient imaging
To demonstrate it, a 44.7MHz transducer is selected. It has 88% band width and the axial
resolution=42um and the lateral resolution=80um, f#=2.4. It is made up of LiNbO
3
as a main
piezo-electric material and 1mm backing layer(E-solder), matching layer with parylene coating.
Figure 67. Fabricated 45MHz transducer
Figure 68. pulsed-echo signal in time domain(left), Frequency response(FFT)
74
Figure 69. Lateral beam profile
Figure 70. Axial beam profile
0
0.2
0.4
0.6
0.8
1
1.2
-150 -100 -50 0 50 100 150
Lateral beam profile
[um]
0
0.2
0.4
0.6
0.8
1
1.2
-1500 -1000 -500 0 500 1000 1500
Axial beam profile
[um]
75
Figure 71. 2D plot of Ispta
Pig eye is used for the demonstration. The data is obtained from UBM system with single
element transducer. The scan-line step size is 20um to see the target in detailed and the
sampling frequency is 2GHz. In the b-mode image, it shows the cornea, iris clearly, but the
integrated backscattering represents the certain speckle pattern at the bottom boundary of
cornea (lateral:3mm, axial:4mm). It is more obvious compare to the b-mode image and it looks
like small dots. Also, there are many small speckle patterns in the cornea. And, the image will
be better if we image the pig eye with array transducers because the single element transducer
has the functionality to fix the focus near 5mm.
-150
-90
-30
30
90
150
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
-150
-130
-110
-90
-70
-50
-30
-10
10
30
50
70
90
110
130
150
Ispta[mW/cm^2]
0-0.2 0.2-0.4 0.4-0.6 0.6-0.8 0.8-1 1-1.2 1.2-1.4 1.4-1.6
x[um]
y[um]
76
Figure 72. B-mode image of pig eye
Figure 73. Integrated backscattering coefficient image of pig eye
B-mode Image
Lateral[mm]
Axial[mm]
0 1 2 3 4
0
1
2
3
4
5
6
5
10
15
20
25
30
35
40
45
50
Integrated backscattering coefficient Image
Lateral[mm]
Axial[mm]
0 1 2 3 4
0
1
2
3
4
5
6 -100
-90
-80
-70
-60
-50
-40
77
5.3 Attenuation slope [MHz/dB] and y-intercept[dB]
There are two ways to lose the acoustic energy-absorption and scattering. Although
ultrasound absorption may be quantified by direct measurement the amount of energy
converted into thermal energy, most of experiments measure the total ultrasonic attenuation
including the scattering effects. And, an attenuation is the one of the most important factor we
need to consider in the ultrasound physics. This is because it decreases the ultrasound beam‟s
intensity of pressure or received echoed amplitude as traveling through the medium. And, table
5 shows the attenuation coefficient to the different tissues
Body Tissue Attenuation Coefficient
Water 0.002
Blood 0.18
Brain 0.6
Breast 0.75
Fat 0.48~0.63
Liver 0.5~0.94
Kidney 1.0
Muscle 1.3~3.3
Tendon 4.7
Bone, cortical 5~6.9
Bone, trabecular 9.94
Table 5. Tissues‟ attenuation coefficient
78
And, the scattering contribution can be ignored below 10MHz while the scattering scales
become more considerable factor as the operating frequency is increased.
Attenuation coefficient represents that how strongly ultrasound wave‟s amplitude decreases as
a function of frequency at the certain tissues. And, it can be derived as equation.13
Attenuation= α[dB/MHz ㆍ cm] ㆍ l[cm] ㆍ f[MHz]
(13)
It shows that the frequency of the incident ultrasound beam linearly depends on the
propagation length, attenuation coefficient, and attenuation and the attenuation coefficient vary
with different tissues And, the logarithmic compressed spectral backscatter coefficient is
defined as equation.14.
( )[ ]
| ( )|
| ( )|
(14)
y= a
1
f+a
0
(15)
From the equation.14, a
1
(attenuation slope) and a
2
(0Hz y-intercept) are taken as shown as
equiation.15. through linear regression. And, the attenuation slope represents the attenuation
coefficient. Fig74 and fig75 show the results of attenuation slope and attenuation. The result
shows that the surface of cornea & lens have the 0.2dB/MHz attenuation slope. And, y-
intercept are shown as fig.75, and the range is from -10 to +12.
79
Figure 74. Attenuation slope of pig eye
Figure 75. Attenuation y-intercept of pig eye
Attenuation slope[dB/MHz]
Lateral[mm]
Axial[mm]
0 1 2 3 4
0
1
2
3
4
5
6
-0.4
-0.3
-0.2
-0.1
0
0.1
0.2
Attenuation intercept[dB]
Lateral[mm]
Axial[mm]
0 1 2 3 4
0
1
2
3
4
5
6 -10
-5
0
5
10
80
CHAPTER 6 SUMMARY AND FUTURE
WORKS
6.1 Summary
To overcome the limitation of the cable‟s space limitation and system‟s complexity, a high
voltage CMOS 1-64 multiplexer/de-multiplexer is proposed and designed in this paper. It can
support high voltage analog input signal with fast switching on & off time. Also, it operates at
the high frequency range up to 70MHz which is the 3-dB bandwidth. The most important factor
is that the chip size is mainly considered in order to fit into the biopsy needle. The performance
is verified with 50um tungsten wire and agar phantom so that it shows the sufficiency of the
integrated circuit. To enhance the image quality, coded excitation technique and beam-forming
for additional imaging area are used. But, the result is degraded than simulation because there
are additional parasitic capacitance and resistance from the layout and pcb board. Additionally,
the high voltage CMOS transistors are protected each other with deep n-well guard-ring, the
parasitic diode shows at the high voltage operation and it also degrade the performance. So, on-
chip measurement is required without wire-bonding and packaging to evaluate the exact
characteristic and SOI-CMOS structure would be better options if we want to enhance the
performance.
81
6.2 Achievement & future works
So far, synthetic aperture imaging platform is designed base on the custom-designed high
voltage multiplexer and the performance is verified with b-mode image. Furthermore, spectral
analysis is done with integrated backscattering coefficient image and attenuation slope.
However, there are still several points to be enhanced as a brilliant imaging platform. And,
there are three possible future works which are required to enhance the platform‟s potential and
capability.
1) Optimization of the system
Currently, it is possible to get the signals and to analyze it after getting it at the post-
processing. But, it takes long time to get the final b-mode image, so the optimization of
algorithm and implementation are needed. This is because the number of data to make one b-
mode image is very large and it takes the long time for the processing. So, a fast-data
processing based on the GPU and parallel computing is one of the solutions. The different thing
is that it will perform based on the hardware instead of software. After this, next step will be
doppler imaging based on the real-time processing.
2) Coded-excitation technique
Synthetic aperture imaging technique support the 64 elements beam-forming but it has a dis
advantage that the SNR(signal to noise ratio) is relatively low. For this reason, additional
technique to enhance the SNR is desirable. One of suggesting method is a coded-excitation and
82
pulse compression technique with linear frequency modulation. It increases the pulse duration
by using coded excitation which increase the total transmitted energy and allow for the
minimization of the transmitted peak power. But, the elongating the pulse duration results in
the negative effect of decrease the axial resolution of the image. In order to avoid the axial
resolution issue, a pulse compression technique should be used after applying the coded signal
excitation. The critical advantage is to improve the echo signal‟s SNR(signal to noise ratio) by
increasing the time bandwidth of the coded signal. In other words, it enables to increase the
penetration depth. Also, there are several method to use the tapering window for the
degradation of side lobe and enhancement of lateral resolution. So, it will help to get better
image quality.
3) Spectral analysis for tissue characterization
B-mode image uses the strength of the received signal and it represents the degree as a gray-
scale level. So, it shows the information in the time domain. And, this is the reason to add the
integrated backscattering and attenuation slope imaging technique so that more information can
be extracted in the frequency domain. And, there are many different approaches to analyze the
data base on the statistical models to extract the distribution degree. Also, tissue‟s stiffness is
the potential aim to analyze the tissues‟ properties in detailed by using ARFI(acoustic radiation
force imaging) and shear wave imaging.
83
4) Actual clinical application.
After finishing the suggested implementations, the platform should be applied to the actual
clinical application. The best option is to collect breast cancer tissue sample and to image with
the high frequency linear transducer and the platform. In order to make sure whether the
analysis is correct or not, the comparison with the histology and the spectral analysis result is
suggested. Also, the different cancer tissues are required to analyze and to compare it for more
precise examination.
BIBLOGRAPHY
K.K. Shung, “Diagnostic Ultrasound: Imaging and Blood Flow
Measurements,” Boca Raton, Fl : CRC Press, 2006.
Baret, Jean-Christophe, Oliver J. Miller, Valerie Taly, Michaël Ryckelynck, Abdeslam El-
Harrak, Lucas Frenz, Christian Rick, et al. 2009. Fluorescence-activated droplet sorting
(FADS): Efficient microfluidic cell sorting based on enzymatic activity. Lab on a Chip 9 (13):
1850-1858.
T. Szabo, “Diagnostic Ultrasound Imaging: Inside Out,” New York:.
Elsevier, 2004.
84
Youk JH, Kim EK, Kim MJ, Lee JY, Oh KK. Missed breast cancers at US-guided core needle
biopsy: how to reduce them. RadioGraphics 2007; 27:79-94
Bommannan, D., Hirohisa Okuyama, Paul Stauffer, and Richard H. Guy. 1992a. Sonophoresis.
I. the use of high-frequency ultrasound to enhance transdermal drug delivery. Pharmaceutical
Research 9 (4): 559-564.
Bommannan, D., Gopinathan K. Menon, Hirohisa Okuyama, Peter M. Elias, and Richard
H. Guy. 1992b. Sonophoresis. II. examination of the mechanism(s) of ultrasound- enhanced
transdermal drug delivery. Pharmaceutical Research 9 (8): 1043-1047.
D. Karakitsos et al., “Real-time ultrasound-guided catheterisation of the internal jugular vein:
A prospective comparison with the landmark technique in critical care patients,” Crit. Care, vol.
10, no. 6, p. R162,2006
Cannata, Jonathan M., Timothy A. Ritter, Wo-Hsing Chen, Ronald H. Silverman, and K.
Kirk Shung. 2003. Design of efficient, broadband single-element (20-80 MHz) ultrasonic
transducers for medical imaging applications. Ultrasonics, Ferroelectrics and Frequency
Control, IEEE Transactions on 50 (11): 1548-1557.
F. J. Herth, R. Eberhardt, P. Vilmann, M. Krasnik, and A. Ernst, “Realtime endobronchial
85
ultrasound guided transbronchial needle aspiration for sampling mediastinal lymph nodes,”
Thorax, vol. 61, no. 9, pp. 795?798, 2006.
D. B. Williams et al., “Endoscopic ultrasound guided fine needle aspiration biopsy: A large
single centre experience,” Gut, vol. 44, no. 5,pp. 720?726, 1999
Feleppa, Ernest J., Andrew Kalisz, Jone B. Sokil-Melgar, Frederic L. Lizzi, Tian Liu, Angel
L. Rosado, Mary C. Shao, et al. 1996. Typing of prostate tissue by ultrasonic spectrum
analysis. Ultrasonics, Ferroelectrics, and Frequency Control, IEEE Transactions on 43 (4):
609-619.
Feleppa, Ernest J., Tian Liu, Andrew Kalisz, Mary C Shao, Neil Fleshner, Victor Reuter, and
William R. Fair. 1997. Ultrasonic spectral-parameter imaging of the prostate. International
Journal of Imaging Systems and Technology 8 (1): 11-25.
Foster, F. S., M. Y. Zhang, Y. Q. Zhou, G. Liu, J. Mehi, E. Cherin, K. A. Harasiewicz, et al.
2002. A new ultrasound instrument for in vivo microimaging of mice. Ultrasound in Medicine
& Biology 28 (9): 1165-1172.
Franke, T., S. Braunmüller, L. Schmid, A. Wixforth, and D. A. Weitz. 2010. Surface acoustic
wave actuated cell sorting (SAWACS). Lab on a Chip 10 (6): 789-794.
86
Cummins, T., Eliahoo, P., & Shung, K. K. (2016). High-Frequency Ultrasound Array
Designed for Ultrasound-Guided Breast Biopsy. IEEE Transactions on Ultrasonics,
Ferroelectrics, and Frequency Control, 63(6), 817-827. doi:10.1109/tuffc.2016.2548993
Ultrasound Imaging. (2011). InTech.
Trots, I., Nowicki, A., Lewandowski, M., & Tasinkevych, Y. (2011). Synthetic Aperture
Method in Ultrasound Imaging. Ultrasound Imaging. doi:10.5772/15986
Nock, L., & Trahey, G. (n.d.). Synthetic aperture imaging in medical ultrasound with
correction for motion artifacts. IEEE Symposium on Ultrasonics.
doi:10.1109/ultsym.1990.171638
Daher, N., & Yen, J. (2006). 2-D array for 3-D ultrasound imaging using synthetic aperture
techniques. IEEE Transactions on Ultrasonics, Ferroelectrics and Frequency Control, 53(5),
912-924. doi:10.1109/tuffc.2006.1632682
Ahmedhassan, M. (2014). Synthetic Aperture Ultrasound Image Reconstruction. International
Journal of Computer Applications, 95(3), 17-22. doi:10.5120/16574-6259
Hwang, Jae Youn, Jungwoo Lee, Changyang Lee, Anette Jakob, Robert Lemor, Lali K.
Medina-Kauwe, and K. Kirk Shung. 2012. Fluorescence response of human HER2+ cancer-
and MCF-12F normal cells to 200 MHz ultrasound microbeam stimulation: A preliminary
87
study of membrane permeability variation. Ultrasonics 52 (7) (9): 803-808.
Hwang, Jae Youn, Nan Sook Lee, Changyang Lee, Kwok Ho Lam, Hyung Ham Kim,
Jonghye Woo, Ming-Yi Lin, et al. 2013. Investigating contactless high frequency ultrasound
microbeam stimulation for determination of invasion potential of breast cancer cells.
Biotechnology and Bioengineering 110 (10): 2697-2705.
Hwang, Jae Youn, Changyang Lee, Kwok Ho Lam, Hyung Ham Kim, Jungwoo Lee, and
K. Kirk Shung. 2014. Cell membrane deformation induced by a fibronectin-coated
polystyrene microbead in a 200-MHz acoustic trap. Ultrasonics, Ferroelectrics, and
Frequency Control, IEEE Transactions on 61 (3): 399-406.
Jasaitiene, D., S. Valiukeviciene, G. Linkeviciute, R. Raisutis, E. Jasiuniene, and R. Kazys.
2011. Principles of high-frequency ultrasonography for investigation of skin pathology.
Journal of the European Academy of Dermatology and Venereology : JEADV 25 (4): 375-382.
Jeong, Jong Seob, Jung Woo Lee, Chang Yang Lee, Shia Yen Teh, Abraham Lee, and K. Kirk
Shung. 2011. Particle manipulation in a microfluidic channel using acoustic trap. Biomedical
Microdevices 13 (4): 779-788.
Jönsson, Henrik, Cecilia Holm, Andreas Nilsson, Filip Petersson, Per Johnsson, Thomas
Laurell, Metabolism and Endocrinology Diabetes, et al. 2004. Particle separation using
88
ultrasound can radically reduce embolic load to brain after cardiac surgery. The Annals of
Thoracic Surgery 78 (5): 1572-1577.
Synthetic Aperture Radar (SAR) Imaging Basics. (2011). Synthetic Aperture Radar
Polarimetry, 1-22. doi:10.1002/9781118116104.ch1
Tasinkevych, Y., Trots, I., Nowicki, A. and Lewin, P. (2012). Modified synthetic transmit
aperture algorithm for ultrasound imaging. Ultrasonics, 52(2), pp.333-342.
Rastello, T., Haas, C., Vray, D., Krueger, M., Schroeder, K., Brusseau, E., . . . Ermert, H.
(n.d.). A new Fourier-based multistatic synthetic aperture focusing technique for intravascular
ultrasound imaging. 1998 IEEE Ultrasonics Symposium. Proceedings (Cat. No. 98CH36102).
doi:10.1109/ultsym.1998.765282
Kuo, I. Y., and K. Kirk Shung. 1994. High frequency ultrasonic backscatter from erythrocyte
suspension. Biomedical Engineering, IEEE Transactions on 41 (1): 29-34.
Tran, C., Kawaguchi, H., & Sakurai, T. (2005). Low-power high-speed level shifter design for
block-level dynamic voltage scaling environment. 2005 International Conference on Integrated
Circuit Design and Technology, 2005. ICICDT 2005. doi:10.1109/icicdt.2005.1502637
Lee, Changyang, Tae-Jin Kim, Jungwoo Lee, Jae Youn Hwang, A. Jakob, R. Lemor,
89
YingXiao Wang, and K. Kirk Shung. 2012a. Ultrasonic stimulation of single bovine aortic
endothelial cells at 1GHz. Ultrasonics Symposium (IUS), 2012 IEEE International, : 21-23
Lee, Changyang, Jungwoo Lee, Hyung Ham Kim, Shia-Yen Teh, Abraham Lee, In- Young
Chung, Jae Yeong Park, and K. Kirk Shung. 2012b. Microfluidic droplet sorting with a high
frequency ultrasound beam. Lab on a Chip 12 (15): 2736-2742.
Lee, Changyang, Jong Seob Jeong, Jae Youn Hwang, Jungwoo Lee, and K. Kirk Shung. 2014.
Non-contact multi-particle annular patterning and manipulation with ultrasound microbeam.
Applied Physics Letters 104 (24): 244107,244107-3.
Lee, Jungwoo, Shia-Yen Teh, Abraham Lee, Hyung Ham Kim, Changyang Lee, and K. Kirk
Shung. 2009a. Single beam acoustic trapping. Applied Physics Letters 95 (7): 073701,073701-3.
Lee, Jungwoo, Changyang Lee, and K. Kirk Shung. 2009b. Calibration of acoustic trapping
forces by fluid drag forces. Ultrasonics Symposium (IUS), 2009 IEEE International, :410-413.
Lee, Jungwoo, Shia-Yen Teh, Abraham Lee, Hyung Ham Kim, Changyang Lee, and K. Kirk
Shung. 2010a. Transverse acoustic trapping using a gaussian focused ultrasound. Ultrasound in
Medicine & Biology 36 (2) (2): 350-355.
Lee, Jungwoo, Changyang Lee, and K. Kirk Shung. 2010b. Calibration of sound forces in
90
acoustic traps. Ultrasonics, Ferroelectrics and Frequency Control, IEEE Transactions on 57
(10): 2305-2310.
Lee, Jungwoo, Changyang Lee, Hyung Ham Kim, Anette Jakob, Robert Lemor, Shia- Yen
Teh, Abraham Lee, and K. Kirk Shung. 2011a. Targeted cell immobilization by ultrasound
microbeam. Biotechnology and Bioengineering 108 (7): 1643-1650.
Lee, Jungwoo, Jin Ho Chang, Jong Seob Jeong, Changyang Lee, Shia-Yen Teh, Abraham Lee,
and K. Kirk Shung. 2011b. Backscattering measurement from a single microdroplet.
Ultrasonics, Ferroelectrics, and Frequency Control, IEEE Transactions on 58 (4): 874- 879.
Lee, S. J., and S. Y. Lee. 2004. Micro total analysis system (micro-TAS) in biotechnology.
Applied Microbiology and Biotechnology 64 (3): 289-299.
Lizzi, Frederic L., Michael A. Laviola, and D. Jackson Coleman. 1976. Tissue signature
characterization utilizing frequency domain analysis. Ultrasonics Symposium (IUS), 1976
IEEE International, : 714-719.
Lizzi, Frederic L., Michael Greenebaum, Ernest J. Feleppa, Marek Elbaum, and D. Jackson
Coleman. 1983. Theoretical framework for spectrum analysis in ultrasonic tissue
characterization. The Journal of the Acoustical Society of America 73 (4): 1366-1373.
Lizzi, Frederic L., Michael Ostromogilsky, Ernest J. Feleppa, Mary C. Rorke, and Mykola M.
91
Yaremko. 1987. Relationship of ultrasonic spectral parameters to features of tissue
microstructure. Ultrasonics, Ferroelectrics and Frequency Control, IEEE Transactions on 50
(3): 319-329.
Pan, D., Li, H., & Wilamowski, B. (n.d.). A low voltage to high voltage level shifter circuit
for MEMS application. Proceedings of the 15th Biennial University/Government/ Industry
Microelectronics Symposium (Cat. No.03CH37488). doi:10.1109/ugim.2003.1225712
Garcia, J. C., Montiel-Nelson, J. A., & Nooshabadi, S. (2009). High performance CMOS dual
supply level shifter for a 0.5V input and 1V output in standard 1.2V 65nm technology process.
2009 9th International Symposium on Communications and Information Technology.
doi:10.1109/iscit.2009.5340988
Lizzi, Frederic L., Michael Astor, Tian Liu, Cheri Deng, D. Jackson Coleman, and Ronald H.
Silverman. 1997. Ultrasonic spectrum analysis for tissue assays and therapy evaluation.
International Journal of Imaging Systems and Technology 8 (1): 3-10.
Lizzi, Frederic L., Ernest J. Feleppa, S. Kaisar Alam, and Cheri X. Deng. 2003. Ultrasonic
spectrum analysis for tissue evaluation. Pattern Recognition Letters 24 (4): 637-658.
Lizzi, Frederic L., S. Kaisar Alam, Samuel Mikaelian, Paul Lee, and Ernest J. Feleppa. 2006.
On the statistics of ultrasonic spectral parameters. Ultrasound in Medicine & Biology 32 (11):
92
1671-1685.
Lockwood, Geoffrey R., Daniel H. Turnbull, and F. Stuart Foster. 1993. High frequency (>20
MHz) spherically shaped ceramic transducers. Ultrasonics Symposium (IUS), 1993 IEEE
International, : 495-498.
Lockwood, Geoffrey R., Daniel H. Turnbull, and F. Stuart Foster. 1994. Fabrication of high
frequency spherically shaped ceramic transducers. Ultrasonics, Ferroelectrics and Frequency
Control, IEEE Transactions on 41 (2): 231-235.
Lockwood, G. R., D. H. Turnball, D. A. Christopher, and F. S. Foster. 1996. Beyond 30 MHz
[applications of high-frequency ultrasound imaging]. IEEE Engineering in Medicine and
Biology Magazine 15 (6): 60-71.
Manneberg, Otto, S. Melker Hagsäter, Jessica Svennebring, Hans M. Hertz, Jörg P. Kutter,
Henrik Bruus, and Martin Wiklund. 2009. Spatial confinement of ultrasonic force fields in
microfluidic channels. Ultrasonics 49 (1): 112-119.
Serneels, B., Steyaert, M. and Dehaene, W. (2008). A high speed, low voltage to high voltage
level shifter in standard 1.2 V 0.13 μm CMOS. Analog Integrated Circuits and Signal
Processing, 55(1), pp.85-91.
93
Moghe, Y., Lehmann, T., & Piessens, T. (2011). Nanosecond Delay Floating High Voltage
Level Shifters in a 0.35um HV-CMOS Technology. IEEE Journal of Solid-State Circuits, 46(2),
485-497. doi:10.1109/jssc.2010.2091322
Khorasani, M., Berg, L. V., Marshall, P., Zargham, M., Gaudet, V., Elliott, D., & Martel, S.
(2008). Low-power static and dynamic high-voltage CMOS level-shifter circuits. 2008 IEEE
International Symposium on Circuits and Systems. doi:10.1109/iscas.2008.4541825
High voltage analog switch, by Jimes Lei(2016, Jul 26). US9401659 B2
Bi-directional high voltage analog switch having source to source connected field effect
transistors, by Basil Weir(1986, Jun 17). US4595847 A
Vittoz, E. (1985). The Design of High-Performance Analog Circuits on Digital CMOS Chips.
IEEE Journal of Solid-State Circuits, 20(3), 657-665. doi:10.1109/jssc.1985.1052365
Digitally controlled high-voltage analog switch array for medical ultrasound applications in
thin-layer silicon-on-insulator process
O'Donnell, M., D. Bauwens, J. W. Mimbs, and J. G. Miller. 1979. Broadband integrated
backscatter: An approach to spatially localized tissue characterization in vivo. Ultrasonics
Symposium (IUS), 1979 IEEE International, : 175-178.
94
KIM, S., LEE, Y., CHUNG, H. and JANG, Y. (2011). A Bootstrapped Analog Switch with
Constant On-Resistance. IEICE Transactions on Electronics, E94-C(6), pp.1069-1071.
Park, Jinhyoung, Jungwoo Lee, Sien Ting Lau, Changyang Lee, Ying Huang, Ching-Ling
Lien, and K. Kirk Shung. 2012. Acoustic radiation force impulse (ARFI) imaging of zebrafish
embryo by high-frequency coded excitation sequence. Annals of Biomedical Engineering 40
(4): 907-915.
Partanen, Ari, Charles Mougenot, and Teuvo Vaara. 2009. Feasibility of agar-silica phantoms
in quality assurance of MRgHIFU. AIP Conference Proceedings 1113, : 296- 300.
Passmann, Christian, and Helmut Ermert. 1996. A 100-MHz ultrasound imaging system for
dermatologic and ophthalmologic diagnostics. Ultrasonics, Ferroelectrics and Frequency
Control, IEEE Transactions on 43 (4): 545-552.
Petersson, Filip, Andreas Nilsson, Cecilia Holm, Henrik Jonsson, and Thomas Laurell. 2004.
Separation of lipids from blood utilizing ultrasonic standing waves in microfluidic channels.
The Analyst 129 (10): 938-943.
Solomon, A., Castellazzi, A., Hoyland, R., Agyakwa, P., Li, J., & Johnson, C. (2009). A
highly integrated high-voltage bi-directional switch. 2009 International Semiconductor Device
Research Symposium. doi:10.1109/isdrs.2009.5378177
95
Sugimoto, Y., Kakitani, H. and Tsurumi, H. (1995). Low-voltage, high-speed, and high-
precision current-mode sample-and-hold circuit. Electronics and Communications in Japan
(Part II: Electronics), 78(7), pp.68-81.
Plummer, J., & Meindl, J. (1976). A monolithic 200-V CMOS analog switch. 1976 IEEE
International Solid-State Circuits Conference. Digest of Technical Papers.
doi:10.1109/isscc.1976.1155469
Shung, K. Kirk. 2006. Diagnostic Ultrasound: Imaging and Blood Flow Measurements. Boca
Raton, FL: CRC Press.
Shung, K. Kirk, Jonathan Cannata, Qifa Zhou, and Jungwoo Lee. 2009. High frequency
ultrasound: A new frontier for ultrasound. 2009 Annual International Conference of the IEEE
Engineering in Medicine and Biology Society 2009, : 1953-1955.
Shung, K. Kirk. 2011. Diagnostic ultrasound: Past, present, and future. Journal of Medical
and Biological Engineering 31 (6): 371-374.
Varadan, Vasundara V., and Vijay K. Varadan. 1979. Low-frequency expansions for acoustic
wave scattering using Waterman‟s T-matrix method. The Journal of the Acoustical Society of
America 66 (2): 586-589.
96
Waterman, P. C. 1969. New formulation of acoustic scattering. The Journal of the Acoustical
Society of America 45 (6): 1417-1429.
Yuan, Y. W., and K. Kirk Shung. 1986. The effect of focusing on ultrasonic backscatter
measurements. Ultrasonic Imaging 8 (2): 121-130.
Abstract (if available)
Abstract
Ultrasound has been used as diagnostic imaging tools in medicine for a long time due to its real-time capability and mobility as well as nonionizing radiation and safety. High frequency ultrasound (above 20 MHz) has opened up new biomedical applications due to its fine spatial resolution by increasing the operating frequency. Among the applications, ultrasound guided needle biopsy is one of methods to collect breast cancer tissue. Especially, it is important to detect and to collect the targeted tissue. In this paper, a high voltage 1 to 64 MUX / De-MUX is designed by high voltage CMOS process (AMS H35B4D3) for the ultrasound-guided breast biopsy application to overcome the limitation. The electronics is made up of three parts such as low voltage a 6 to 64 decoder, a level shifter to convert from low voltage to high voltage, and 64-analog high voltage switches. The experiment results show that the 3-dB bandwidth is over 70MHz, and it has 180 ohms of on-resistance, -2.635 dB of insertion loss, -26.526 dB of isolation at 70MHz. Also, synthetic imaging technique is used for the beam-forming which is suitable to the limited number of transmitting to the elements. Also, 1-channel transceiver based on FPGA is developed and implemented so that it supports total compact size as a synthetic aperture imaging platform. At the post-processing, integrated backscattering coefficient is added to analyze the signal in the frequency domain.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
High frequency ultrasonic imaging for the development of breast biopsy needle with miniature ultrasonic transducer array
PDF
High frequency ultrasound array for ultrasound-guided breast biopsy
PDF
High-frequency ultrasound array-based imaging system for biomedical applications
PDF
Array transducers for high frequency ultrasound imaging
PDF
Single-cell analysis with high frequency ultrasound
PDF
Microfluidic cell sorting with a high frequency ultrasound beam
PDF
Intravascular imaging on high-frequency ultrasound combined with optical modalities
PDF
Quantification of cellular properties using ultra-high frequency single-beam acoustic tweezer
PDF
High frequency ultrasonic phased array system and its applications
PDF
Configurable imaging platform for super-harmonic contrast-enhanced ultrasound imaging
PDF
Highly integrated 2D ultrasonic arrays and electronics for modular large apertures
PDF
Novel beamforming with dual-layer array transducers for 3-D ultrasound imaging
PDF
Audio and ultrasound MEMS built on PZT substrate for liquid and airborne applications
Asset Metadata
Creator
Jung, Hayong
(author)
Core Title
Synthetic aperture imaging platform based on CMOS high voltage 1 to 64 multiplexer / de-multiplexer for ultrasound guided breast biopsy needle
School
Viterbi School of Engineering
Degree
Doctor of Philosophy
Degree Program
Biomedical Engineering
Publication Date
08/09/2018
Defense Date
05/02/2018
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
beam-forming,biopsy needle,breast cancer,high frequency,high voltage switch,OAI-PMH Harvest,power amplifier,pulser,synthetic aperture imaging,ultrasound
Format
application/pdf
(imt)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Shung, Koping Kirk (
committee chair
), Chen, Shuo-Wei (
committee member
), Liu, Brent (
committee member
)
Creator Email
hayong@usc.edu,kwjhy35@gmail.com
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c89-60671
Unique identifier
UC11670524
Identifier
etd-JungHayong-6663.pdf (filename),usctheses-c89-60671 (legacy record id)
Legacy Identifier
etd-JungHayong-6663.pdf
Dmrecord
60671
Document Type
Dissertation
Format
application/pdf (imt)
Rights
Jung, Hayong
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Tags
beam-forming
biopsy needle
breast cancer
high frequency
high voltage switch
power amplifier
pulser
synthetic aperture imaging
ultrasound