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
/
Evaluation of R.F. transmitters for optimized operation of muscle stimulating implants
(USC Thesis Other)
Evaluation of R.F. transmitters for optimized operation of muscle stimulating implants
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
EVALUATION OF R.F. TRANSMITTERS FOR OPTIMIZED
OPERATION OF MUSCLE STIMULATING IMPLANTS
by
Tomonori Thomas Murakata
A Thesis Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF SCIENCE
(BIOMEDICAL ENGINEERING)
December 2003
Copyright 2003 Tomonori Thomas Murakata
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UMI Number: 1420389
INFORMATION TO USERS
The quality of this reproduction is dependent upon the quality of the copy
submitted. Broken or indistinct print, colored or poor quality illustrations and
photographs, print bleed-through, substandard margins, and improper
alignment can adversely affect reproduction.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if unauthorized
copyright material had to be removed, a note will indicate the deletion.
®
UMI
UMI Microform 1420389
Copyright 2004 by ProQuest Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
ProQuest Information and Learning Company
300 North Zeeb Road
P.O. Box 1346
Ann Arbor, Ml 48106-1346
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UNIVERSITY OF SOUTHERN CALIFORNIA
under the direction o f h K thesis committee, and
approved by all its members, has been presented to and
accepted by the Director o f Graduate and Professional
Programs, in partial fulfillment o f the requirements fo r the
degree of
Master o f Science in Biomedical Engineering
THE GRADUATE SCHOOL
UNIVERSITY PARK
LOS ANGELES, CALIFORNIA 90089-1695
This thesis, written by
Tomonori Thomas Murakata
Director
Date__Qacemher.JLZ.,...-20.03.
Thesis Committee
Chair
/U-t
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table of Contents
List o f Tables iv
List o f Figures vi
Abstract vii
Chapter 1
1.1 Introduction 1
Chapter 2
2.1 RF Transmitter Background 4
2.2 Class E Modeling Selection 11
Chapter 3
3.1 Background for Transmitter Coil Design 16
3.2 Transmitter Coil Inductance 18
3.3 Mutual Inductance Calculation 27
3.4 Determining Number of Coil Turns 32
3.5 Coil Material Test 40
3.6 Field Strength Mapping 50
Chapter 4
4.1 Modulation Test 53
4.2 Modulation Characteristic Table 55
4.3 Selection o f Capacitor Values 59
4.4 Testing of the Selected Modulation Parameters 64
4.5 Modulation Test Analysis 72
Chapter 5
5.1 Coil Shielding 75
5.2 Shielding Performance Test 76
5.3 Shielding Test Conclusion 80
ii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 6
6.1 Electromagnetic Interference Regulation 83
6.2 RF Emission Testing 85
6.3 Emission Data 88
6.4 Analysis o f Results 90
Bibliography 97
iii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
List o f Tables
1. Table 3.1: Dimensions of Coils 33
2. Table 3.2: Coil Current, Max Voltages, and Power Supply
Current Draw 39
3. Table 3.3: Comparison of 4, 5, and 6 turn coils at a lower
Voltage 39
4. Table 3.4: Coil Attributes, Dimensions and Wire Material 42
5. Table 3.5: Summary o f Coil Characteristics 43
6. Table 3.6: Key Values for Measurements 45
7. Table 4.1: Modulation Comparison Table for 120 ns and [70] ns 57
8. Table 4.2: Rise Characteristics Comparison Table for 120 ns
and [70] ns 58
9. Table 4.3: Chosen Test Configurations for 120 ns and 70 ns
pulse widths 64
10. Table 4.4: Normal Rise Time 66
11. Table 4.5 Loaded Rise Time 66
12. Table 4.6: Power of Center Frequency and Sideband 68
13. Table 4.7: Loaded Power of Center Frequency and Sideband 68
14. Table 4.8: Normal Sensitivity 71
15. Table 4.9: Loaded Sensitivity 72
16. Table 5.1: Unshielded and Shielded Coil Characteristics 77
17. Table 5.2: Shielded and Unshielded E-field Measurements 78
18. Table 5.3: Unshielded BION Sensitivity 79
iv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
19. Table 5.4: Aluminum Foil BION Sensitivity 80
20. Table 5.5: Copper Tape BION Sensitivity 80
21. Table 6.1: Normal Operation 89
22. Table 6.2: Contact with Skin 89
23. Table 6.3: Program Changes 89
24. Table 6.4: Wrapped in Aluminum Shield 90
25. Table 6.5: Tuned Transmitter Wrapped in Aluminum Shield 90
25. Table 6.6: Data converted to dBuA 94
v
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
List of Figures
1. Figure 2.1: Transmitter Circuit [Loeb et al.] 8
2. Figure 3.1: Coil Dimensions 25
3. Figure 3.2: Inductance and Q of Large Copper Coil 46
4. Figure 3.3: Inductance and Q of Large Litzwire Coil 46
5. Figure 3.4: Inductance and Q of Small Copper Coil 47
6. Figure 3.5: Inductance and Q of Small Litzwire Coil 47
7. Figure 3.6: Field Map for CDLCO1180201 51
8. Figure 4.1: Voltage and Capacitor Relation 59
9. Figure 4.2: Modulation Capacitor and Index Relation 60
10. Figure 4.3: Series and Steady (Shunt) Capacitor Relation 61
11. Figure 4.4: Frequency and Modulation Capacitor Relation 62
12. Figure 4.5: Modulation Capacitor and Rise Cycle Relation 63
13. Figure 4.6: Wet Well 69
14. Figure 6.1: Diagram of Test Setup 86
15. Figure 6.2: Test Setup 87
16. Figure 6.3: ETSI Transmitter Carrier Limits 95
vi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
A bstract
Miniature electrical stimulators containing application specific integrated
circuits (AISC) can be injected into muscles for therapeutic and functional
stimulation without the problems associated with skin contact electrodes. These
implants are externally powered through the use of a radio frequency transmitter and
inductively coupled resonant coils. In order to optimize the inherently inefficient
power transfer, design and evaluation processes were created for the transmitter
circuit and coil to achieve ideal operation. The procedures cover the design of the
transmitter coil parameters, such as the number of coil turns, and the selection of
components for the Class E amplifier circuitry used to drive the coil. Several
experiments that test the supplemental features of the transmitter were performed as
well. Since there are compromises to be made in the optimization of such RF
transmitters, the procedures were created to simplify the design of transmitter so that
they can be repeated readily.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
C h ap ter 1
1.1 Introduction
Development of neural prosthetics that aim to achieve functional electrical
stimulation (FES) of paralyzed limbs has been on the rise in the past decade.
However, FES devices are not readily available commercially because they are still
in their developing stages. The technology to stimulate muscles has met with
success, but it is the control of stimulating individual muscles in a manner to achieve
a complex movement that is proving to be difficult. Therefore the performance of
FES systems is not yet satisfactory clinically and most likely will not be for several
years. FES is promising, but it may not be the ideal candidate for many investors
because funding for research of a product depends on its commercial viability and
the investment returns of FES would not be immediate. A dilemma is created
because in order to achieve success of FES devices, more research is needed, but in
order to get industrial funding for research, the devices need to demonstrate more
capabilities.
One strategy to commercialize neuromuscular stimulation is to aim for
therapeutic electrical stimulation (TES). Just as its name implies, TES is the process
of using electricity to stimulate muscles for therapeutic reasons. Exercise is
necessary for muscles to remain strong, but it is impossible for patients with
paralysis to exercise. These patients can use TES to stimulate their muscles in order
to get the exercise the muscles require to prevent atrophy and other complications
1
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
from lack of movement. The technology for TES is available so it is considered to be
more feasible and commercially viable [Loeb and Richmond 1999]. Since TES is
basically a simpler form of FES, the commercialization and success of TES can
become a stepping stone to FES.
A prosthetic device that is capable of TES and has a promising future for FES
is the “microstimulator” implant, called the BION ™ (BlOnic Neurons), The BION
is researched and developed at the Alfred Mann Institute located in University of
Southern California. The technology behind the BION is quite remarkable. For a
device that only measures 2 mm in diameter and 16 mm in length, its features are
sophisticated. The internals o f the BION consist of a coil created by wrapping 1 mil
insulated copper wire around a ferrite core. The coil has 190 turns and it is tuned to
have a resonant frequency o f 2 MHz. The ferrite core is actually two pieces of ferrite
that surrounds a piece of printed circuit board (PCB). The PCB carries an application
specific integrated circuit (ASIC), which controls the operation of the BION. The
ASIC is powered by an external RF transmitter through inductive coupling by the 2
MHz resonant carrier frequency. The carrier frequency is amplitude-modulated to
send command data in digital form, which the ASIC in the BION decodes to control
stimulus current strength and duration. Since each BION has its own 8-bit address, a
maximum number of 256 BIONs can be operated by a single transmitter. The whole
coil assembly is then hermetically sealed inside a glass cylinder. There are
stimulation electrodes mounted on both ends of the glass cylinder that become the
stimulators when the BION is injected into a muscle by a 12 gauge needle.
2
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The current BION model only provides a stimulus to the muscles when it is
controlled by the external transmitter thus the function of the BION is limited to
TES. However, future models of the BION are being developed with better features
as clinical trials of FES by the BION have been successful. Currently a new model of
the BION is being designed to send sensing information, such as BION position and
acceleration, back to the transmitter through bi-directional telemetry. Another feature
that may be developed is the use of an internal battery. The internal battery will
allow the BION to function without the external transmitter for convenience as the
battery can be recharged by the transmitter at a later time.
The introduction provides a background of the BION, but the focus of this
paper is on the external RF transmitter that powers the BION. It was necessary to
provide the information about the BION to indicate the importance of the
communication between the BION and the transmitter. While it is not as complex as
the BION, the RF transmitter has a unique circuitry which required research for the
selection of components. Without a properly configured transmitter, the BION will
not work. Therefore several experiments were performed on the transmitter for
optimization, which became the basis of this paper. Chapter two explains how the
mechanism of the circuitry inside of the transmitter works. Chapter three talks about
the design process for the transmitter coil. Chapter four describes the experiments on
modulation o f the transmitter frequency. Chapter five summarizes the effects of
shielding on the transmitter coil. In the final chapter, the RF emissions of the BION
transmitter is tested for regulation limits.
3
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
C hapter 2
2.1 RF Transmitter Background
Like most RF transmitters, the BION RF' transmitter produces a magnetic
field to transmit energy to RF detecting devices. However, the BION RF transmitter
produces a weak magnetic field compared to a typical RF transmitter because it is
designed to induce current only into BIONs that are within close range. In other
words, the relation between the BION and the transmitter can be thought as an
inductively coupled RF transformer. It is not an ideal transformer because of the
difference in coil sizes and the lack of an iron core, but it is still a transformer. Thus
the voltage required by the BION coil for stimulation is dependent on the strength of
the magnetic field produced by the transmitter coil.
The BION transmitter contains circuitry that drives AC current flow through
the coil in order to produce a magnetic field. It is possible for the transmitter coil to
connect to the driver circuit directly, but it will be very inefficient due to power loss
from high voltage and high current [Troyk 1992]. High current is necessary in the
coil to provide a strong magnetic field strength, but not for the driving circuit.
Fortunately, the use of a properly selected capacitor in series or in parallel to the
transmitter coil can provide a resonant circuit whose resistive impedance is low
enough to operate efficiently. Even though they are both LC circuits, the operating
characteristics are different between the series and parallel connections in that each
type has its own advantage and disadvantage.
4
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
In the case of series connection, the magnitude and phase of the current flow
is equal for both the inductor and capacitor components. For optimal effect, the
current flow can be increased when the reactance of the two components is equal to
each other. Since the capacitor and inductor axe 180° out o f phase, if the magnitude
of the reactance is the same, the two cancel each other out [Thomas 1998]. This
phenomenon is called resonance and for every combination of inductors and
capacitors, there is one resonant frequency. The resonant frequency can be solved by
the following steps:
XSeries - X L - XC = 2 71 f L - 1 /(2 7t f C) = 0
27cfL = l / ( 2 7tfC)
f - 1 / (2 7 c V (LC)) (2.1)
The characteristic of the series resonant circuit is that each component has the
maximized voltage across it due to the maximized current through it, but the total
amount o f voltage in the series branch is close to zero. Therefore a device connected
to this load will not have to deal with high voltages. Unfortunately, the device will
still have a problem with high current because essentially there is a short in the
circuit.
When the capacitor is placed in parallel with the coil instead, the loading
characteristics is completely different from the series connection. In the case of the
series resonance, the combined value o f reactance and thus the impedance is at a
5
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
minimum. However, in the case of the parallel resonant circuit, the impedance is
maximized. This happens because just like in series resonance, the reactance of the
two components cancel each other and are at a minimum [Tipler 1991]. Due to the
parallel connection, the minimal reactance equates to a very large impedance as
shown below:
X Parallel = (1 / (1 / XL - 1 / XC))
= ( 1 / ( 1 / ( 2 f L ) - ( 2 7tfC)) =1/0 (ind.)
2 7 r f L = 1 / (2 7 i f C)
f = 1 / (2 7 C V (LC)) (2.2)
It is interesting to point out that equation 2.1 and 2.2 are the same, indicating
that resonance occurs at the same frequency no matter the type of connection. The
impedance does not actually go into infinity because the reactance is not completely
zero, but it shows that the impedance can become quite large. When a device is
connected to a parallel resonant load, it will not have problems with high current
because o f the high impedance unlike the series load. The parallel resonant load does
have a weakness like the series load because it will have high voltage, the opposite
of the series load.
In order to eliminate the problems of both the series and parallel resonance
circuitry, the “Class E” amplifier circuit was developed by Nathan Sokal in the
1970’s [Cantrell 1998]. The interesting feature about the Class E transmitter circuit is
6
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
that it contains both series resonance and parallel resonance, which make it possible
for the circuit to deal with both low voltage and current under certain circumstances.
The complexity of the Class E is not in its circuitry, but in the timing of voltage and
current waveforms. The transmitter driver switch must be applied to the circuit
voltage when high voltage and high current does not occur simultaneously [Sokal
2001]. The switch can be thought of as an ideal switch, meaning it has no voltage
across it when it is on and passes no current through it when it is off. This eliminates
switching losses which is a source o f power loss found in different types of RF
amplifiers [Cantrell 1998]. Due to its power efficiency, the Class E circuitry is used
for the BION RF transmitter. For clarification, the mechanism behind the “Class E”
amplifier will be explained.
The circuitry consists of a voltage supply that is connected to a choke (LI),
which is then connected by a shunt capacitor (C l) and a “series branch”, comprised
of a capacitor (C2) and the transmitter coil (L2), to ground. See Figure 2.1 below for
circuit diagram. The driver switch, in this case a MOSFET, connects the choke to
ground as well so that the shunt capacitor, the series branch and the MOSFET are in
parallel to each other. For ideal operation, the component values are selected so that
they are in both series and parallel resonance. The selection of the proper values will
be discussed in a later section. Initially the MOSFET is off or open so that there is
only a small amount of current flowing through the choke. However a change in
current occurs in the choke when the MOSFET is turned on since it essentially
creates a short to ground. When there is a change in current, energy is stored in the
7
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
choke. The energy stored turns the choke into a constant current supply during each
cycle when current would have stopped due to the switch being open [Cantrell
1998]. W ith the MOSFET turned on, the positive current flow to the series branch
reduces to zero and begins to reverse its flow towards the switch since it is the least
restrictive path to ground.
Clock
Input ■
Modulation.
lap*
D t o 's .P h I s b
Generator
-Ci
r-o
m
fl
V
-as:
C L * I
D uraS en
Flip Flop
..|
I
6 N 0
cs
m
t >
H
Teat
e ta
C l
Flip Fbp sw itc h
02]r
L 3
MOSFET*
switch
k x f l-
O sciilntor
C2
C l
T ra n s.
Coil
Z sk» crass
04 Detaetef
R 1 J S -
PSvast
Offset
Figure 2.1: Transmitter Circuit [Loeb et al.J
During this period of current reversal, the MOSFET is turned back off so that
the negative current cannot flow through the MOSFET. Without the short to ground
provided by the MOSFET, voltage is produced on the shunt capacitor and switch.
8
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The change in voltage across the shunt capacitor causes the “trapped” negative
current to flow into it instead. Once the negative current stops, the shunt capacitor
wants to discharge the voltage stored so current begins to flow back into the series
branch, resulting in the voltage on the switch to reduce. Once the voltage on the
MOSFET is zero, the switch is turned on again to pull the current into the other
direction and the cycle continues [Troyk 1992]. To summarize, when the switch is
turned on, it allows current to flow into the choke, which is then sent to the shunt
capacitor when the switch is turned off. The shunt capacitor then transfers its energy
to the series circuit. The inductor or the transmitter coil in the series circuit is able to
transmit power to the BION.
It is critical that in order to reduce power loss, the MOSFET must be turned •
on when the capacitor voltage drops to zero because any charge left in the shunt
capacitor will be shorted to ground and wasted [Sokal 2001]. That is why the driver
circuit controlling the timing of the switch, in accordance to the voltage and current
waveforms, plays an important part for the transmitter circuit. The driver circuit
consists of a flip flop which is operated by receiving feedback from the current of the
transmitter coil. The feedback is necessary to achieve optimal efficiency because the
switch needs to be turned on when the shunt capacitor voltage and the rate of change
of the voltage is zero as well [Troyk 1992]. The feedback signal is produced by a
zero cross detector on the transmitter coil current. Since the derivative of the voltage
waveform is the current, a differentiator circuit is used to phase shift the coil current
before it is sent to the zero cross detector.
9
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The zero cross detector determines when the differentiated coil current
crosses zero and sends out a digital logic high signal o f 5 V. Once the signal
produced by the zero cross detector is sent to the flip flop, the flip flop outputs a
logic high signal of 5 V to activate the MOSFET. The duration of the MOSFET
pulse is determined by the inverse output of the flip flop, which is connected to the
clear switch of the flip flop. When the clear switch is triggered, the flip flop output is
cleared. When the flip flop output is a logic high 5 V, then the inverse output would
be a logic low 0 V. If the inverse output was directly connected to the clear switch
input, the pulse would be cleared immediately after it was started. By placing a
resistor and capacitor in between the inverse output and clear input, the output signal
will have a delay due to the RC time constant. The delay in the inverse output signal
reaching the clear input is what generates a pulse. The delay can be varied by using a
potentiometer for the resistance.
Using a feedback system like the one described allows for fluctuations in
operating frequency from allowing the circuit to “de-tune” and stop functioning.
While this is a positive feature when the circuit is optimized, there is a negative side
as it allows the transmitter circuit to operate with improper capacitor and inductor
values running at the wrong frequency from the start. As stated before, improper
values will result in less optimal and less efficient operation. Therefore careful
modeling of the circuit becomes even more imperative.
10
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.2 Class E Modeling Selection
There are several parameters that need to be taken into consideration before
the component values could be selected. The first Item is the Input voltage supplied
as the source. Obviously with a higher input voltage, the output voltage can be
increased as well, but a higher input voltage will require more current draw from a
power supply [Donaldson 1983]. Because of this, optimization is critical to achieve
the highest amount of current flow to the transmitter coil while keeping the input
voltage to a minimum. An input voltage o f 12 V was decided upon as it is a
moderate sized voltage that appears to be capable to deliver enough power and also
because the driver circuit (zero cross detector) that operates the MOSFET was
designed to operate at 12 V. The second consideration is the operating frequency. A
frequency not used by other electrical equipment is required to limit interference
from those devices. A frequency o f 2 MHz was chosen because it is between the AM
and FM radio frequencies so that interference should not be a problem [Cleveland
1999]. The operating frequency must remain close to 2 MHz so that proper coupling
with the BION can occur because the coil winding of the BION is set at 2MHz.
With the constants specified, the equations expressing the relationship of the
Class E components could be used to calculate the acceptable values within the
boundary constraints. The main mechanism to evaluate is the combined series and
parallel circuit functions. If the series branch is in resonance, the branch reactance
must be approximately zero since C2 and L2 cancel each other. For the circuit to
11
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
have parallel resonance as well, the reactance of the shunt capacitor (€1) must equal
the reactance of the series branch. This suggests that the reactance o f the shunt
capacitor is approximately zero, but this is not possible. Due to this fact, the parallel
resonance and the series resonance frequencies can not be the same, but they can
have frequencies that are very close to each other. A large capacitor can be used (pF
range), so that the reactance of the shunt capacitor can become quite small. However
a problem is created when a very large value is used for the shunt capacitor.
The voltage across the shunt capacitor and MOSFET needs to be zero before
the MOSFET is switched back on for power loss reduction. A large shunt capacitor
changes the time constant, which is the factor that determines the time required for
the capacitor to charge. The time constant is the time it takes for the capacitor to rise
to 63% o f the maximum voltage so the total time required to reach the maximum
value is 5 times the time constant because each time constant increases the voltage
by a percentage. Since the cycle repeats itself every 2 MHz or 50 ps, the voltage
must rise and fall within this 50 ps time frame. If the time constant is too large, the
capacitor voltage will not drop to zero before the MOSFET is turned on. On the
other hand, if the shunt capacitor value is too small, the voltage drops to zero earlier
than the switch is turned on. Therefore to limit power loss, the lowest maximum
value for the shunt capacitor needs to be utilized for efficiency.
Altering the value for the shunt capacitor requires changing the value for the
series capacitor as well. Since the series capacitor is connected to the transmitter coil
on the series branch, the Q or quality factor o f the transmitter coil is affected. The Q
12
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
is expressed by the reactance over the resistance of the coil. When the Q is too high,
the change in switch voltage is not damped well and the voltage across the MOSFET
is below zero when the switch closes. When the Q is too low, the voltage is
excessively damped and the MOSFET voltage is above zero when it closes [Troyk
1992]. Therefore the ideal Q value or critical Q needs to be found. Q is also
important as it describes the resonance characteristic. A high Q will offer a sharp
resonance, but a narrow bandwidth while a low Q has a broad bandwidth with a less
sharp peak [Tipler 1991]. Resonance is not only used for the Class E operation, but
also for the transfer o f energy to the BION since it acts as a band-pass filter, which
accepts power for a particular frequency range while rejecting other frequencies. The
ideal coil will have a Q that is high enough to offer a sharp resonance, but at the
same time have some slack for shifts in frequency.
As one could tell by the description o f operation, the Class E circuit is very
sensitive so that the majority of the component values are dependent on each other.
By combining all o f the restrictions and limitations of the Class E operation, the
equations that define the component values can be realized. The first equation is the
voltage waveform at the shunt capacitor.
Vs - (A / 2k ) (1/2D- D2)) (V (sin2 (2 k D))) (2.3)
A is the maximum voltage amplitude at the switch, and D is the pulse duration o f the
switch. For convenience, A can be expressed as approximately twice the amount of
13
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the input voltage or 24 V. Basically the equation displays a waveform that fluctuates
between OV and 24 V to represent the change in voltage across the switch as it
undergoes Class E operation. From the switch voltage, the resistance and the
reactance o f the branches could be calculated to solve for the capacitor values.
|ZS|= V S’ /I2 (2.4)
Rc = c o e * L2 / Qi (2.5)
Xs = V(Zs2- R c2) (2.6)
C2= 1 /(© E(©E*L2- X S)) (2.7)
Ci = 1 / (©e * Xs) (2.8)
Where ©E is the operating frequency, L2 is the transmitter coil inductance, Qi is the
quality factor of the coil, and I2 is the coil current. For the BION transmitter, there is
another capacitor that is added in parallel to the shunt capacitor. This capacitor is
called the “modulation” capacitor as it allows for a modulation in output so that a
detector circuit in the BION can sense it as a digital data stream. Modulation occurs
because essentially the shunt capacitor now has two values to switch from due to the
modulation capacitor, and different values cause a difference in operation. Since it
will cause complications for the calculation of values for the original Class E circuit,
the role of the modulation capacitor will be discussed more in a later section.
Out o f the Class E components, the transmitter coil is the main factor for the
transfer of energy to the BION so the two values that are the most important is the
14
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
coil inductance and current. Out of the two, the inductance will be the most flexible
in terms o f desired value selection. The current flowing in the series branch is
somewhat limited because it is based on the coil inductance, the circuit input voltage
and MOSFET pulse width. The property of transmitter coil is interesting as it is the
one value that is changed depending on outside factors such as its shape. Due to
these reasons the other components in the Class E circuit needs to be altered
depending on changes of the transmitter coil.
15
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 3
3.1 Background for T ransm itter Coil Design
The strength of the magnetic field produced by the transmitter coil is largely
dependent on the voltage required by the BION coil to provide the desired
stimulation current to the muscle. The voltage of the BION coil is determined by
multiplying the maximum amount of current needed with the resistance between the
electrodes in the muscle where the BION is Implanted. If the desired amount of
current is 20 iuA and the electrical resistance of the muscle is estimated to be 500 Q ,
[Loeb and Richmond 2000], then the compliance voltage required for the high
voltage supplied by the coil would be 10 V according to Ohms law. The voltage
required by the BION also determines the amount of voltage necessary at the
transmitter coil since the interaction between the transmitter coil and the BION coil
is fundamentally one of a weakly coupled transformer.
In the case of a perfect transformer, which is not the case for the BION
transmitter, once the voltage level at the secondary coil is determined the amount of
voltage needed at the primary side can be solved by using the equation [Tipler 1991]
V2/V1 =(k2 *N2)/(kl *N1) (3.1)
VI & V2 are the voltages of the primary (transmitter) and secondary (BION) coil,
N1 & N2 are the number turns of the primary and secondary coil, and kl & k2 are
16
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
proportionality constants that characterize the difference in coil sizes. The fact that
the coupling is done in air and the difference in size between the BION coil and
transmitter coil, the coupling coefficient is expected to be very low. Even if a
transmitter is not a perfect transformer, the equation could have provided a possible
approximation for the transmitter coil voltage. However in the case of the BION
transmitter, there are several Issues associated with it.
The first issue is that it is not possible to solve this equation since the
transmitter coil is not yet determined so the proportionality constants are not known.
O f course all the BION coil information is known, so once the ideal transmitter coil
is created, the voltage required for the transmitter coil can be solved. The second
issue is regarding the design of the BION. It was assumed that the transmitter coil
cannot provide sufficient power for the BION to generate the required current at that
voltage level continuously due to insufficient coupling. To generate the desired
amount o f voltage, an internal capacitor is used to store charge from the induced
current. Since the output current of the BION only need to be brief pulses, the
capacitor can store charge when the BION does not have an output. Once the
capacitor is charged to the 17 V regulated voltage, it can produce a current strong
enough to stimulate the muscle. Therefore it Is assumed that the voltage generated by
the capacitor is the same as the voltage directly induced into the coil.
Since it will be difficult to model the actual amount of power transmission,
the transmitter coil will be built in terms o f desired mechanical properties instead of
electrical properties. Once it is built, the inductance o f the coil will be calculated and
17
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
measured. There will be a higher chance of improper Class E operation, but the
capacitance of the BION should provide a wide range of inductance values that is
acceptable. The amount of power transferred from the transmitter to the BION can
still be approximated by solving the mutual inductance between the transmitter and
BION. The amount of power transferred in each cycle would be incorrect, but the
amount o f power collected during the several cycles in which the BION charges can
be compared. The equation will only be an approximation, but it can determine
whether the transmitter is in the appropriate voltage range. Also to confirm its
strength, the actual magnetic field can be measured after the transmitter Is built. By
using a probe coil and connecting it to an oscilloscope, it will bepossible to detect
the amount of voltage coupling capable by the transmitter. To summarize, the design
of the transmitter coil will undergo several experiments because it will be difficult to
model it theoretically.
3.2 T ransm itter Coil Inductance
The shape and size of the transmitter coil will change the properties of the
coil. For example, the number of turns in a coil changes the length of wire used to
make the coil so the resistance of the coil will change. The number o f turns will also
influence the alignment of current flowing inside the coil and thus there will be a
difference In inductance. The shape o f the magnetic field produced by the coil, which
is necessary for coupling to the BION, will be different as well depending on the
18
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
shape o f the coil [Tipler 1991]. For these reasons, all o f the physical factors need to
be considered when designing a coil.
It was decided that the overall shape of the BION transmitter coil for one
clinical application be an elliptical shaped, spiral coil with an inner diameter of 9
inches on the long side and 4.5 inches on the short side. This “large” coil
configuration was chosen because it was the preferred style that was used in clinical
trials for preventing shoulder subluxation. The dimension o f the oval shape allows
for the coil to fit around a patient’s shoulder for better field penetration to the BION
than other coil shapes. While the shape o f the coil was decided by physical and
mechanical reasons, other factors such as the ideal number o f loops for the coil and
the material of the wire for the coil need to be determined by their operational
behavior.
When current flows inside a wire, a magnetic flux is generated around the
wire. The total magnetic flux can be increased by shaping the wire into a coil so that
the flux becomes cumulative. Induction is primarily dependent on the magnetic flux
produced by a coil. The flux of one coil to be received by another coil which induces
current to flow in the second coil. Since magnetic flux is derived from the number of
magnetic field lines in a specified area, the magnetic field must first be calculated.
The equation for magnetic field strength is:
dB = (go * I * dl *sin 0) / {An R2 ) (3.2)
19
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The equation is known as the Biot-Savart law. The po is the permeability o f free
space which is a constant o f 4% xlO'7 H/m, I dl is the current element inside of the
wire, R is the radius o f the coil, and 0 is the angle between I dl and a vector towards
the center o f the loop. The current element is directly responsible for the creation of
the magnetic field at a certain point indicated by R and 0. Thus the sum of all of the
current elements in a coil determines the total field strength [Thomas 1998].
Integrating the equation allows for the summation of the individual current elements.
The integrated form is shown by the following formula:
B = J dB = (|i0 * I * sin 0) / (4t c r2 ) * j dl (3.3)
Since the current element is the variable that is integrated, the difficulty in solving
the magnetic field equation varies depending on the shape o f the coil and location of
the field strength. For example, the magnetic field at the center of a circular coil can
be solved without much difficulty. The distance away from the current element I dl
to the center o f the coil, in other words r, would be the radius of the circle R since all
the distances would be the same Therefore 0 would be 90° at all the points, making
sin 0 equal to 1 . The integration o f dl is solved to be 2 tzR , which is the circumference
of the circle. Therefore the magnetic field of a circle is:
B = I*N* p0*R2 / (2 * (x2+R2)a(3/2)) (3.4)
20
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The value of x is used when the magnetic field strength is measured at some distance
away from the surface of the coil. Since the current element, the center of the coii
and the measured field position form a triangle, the Pythagorean theorem can be used
to calculate the distance between the current element and the measured location. In
this situation the maximum magnetic field strength is desired, so x will be 0 because
the field strength is the strongest inside of the coil. While the field strength is
strongest inside of a coil, it does not necessarily mean that the field is uniform inside
of the coil.
Unlike a circular coil, the magnetic field equation is difficult to solve in the
case for the BION transmitter coil because it is elliptical. Having an elliptical shape
causes all of the current elements, I dl, to not have a uniform distance away from the
center. In other words, R would be different for each current element. The angle 0
will also be different for each point as well, so it is not possible for sin 0 to be a
constant. The integration of dl will still be assumed to be the circumference of the
ellipse, since it will be the summation of all of the current element points. However,
there is no exact solution for finding the perimeter of an ellipse, but only equations
that approximate the length of the perimeter. O f the various approximations, the
formula created by the Indian mathematician S. Ramanujan is the most accurate
[Ramanujan 1913]. The perimeter approximation formula is:
P » n [ 3(a+b) - V((3a+b)(a+3b))] = n (a+b) [ 3 - V (4 - h )] (3.5)
21
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
In the equation, h is equal to (a-b)2/(a+b)2 . While not as accurate as the Ramanujan
equation, the perimeter could also be approximated with the perimeter equation for a
circle and using the average of the two foci for the radius.
P « 2 71 ((a+b)/2) - n (a + b) (3.6)
The difference between the two equations is minimal but the second approximation
allows for quicker calculations. Since the difference in perimeters is quite small, the
second equation will be used.
With the perimeter of the transmitter coil known, the magnetic field o f the
transmitter coil could be approximated by assuming that the r in the case of an ellipse
is the average length of the major and minor axis. The distance between the current
element and the center of the coil will be different for each element, but the distances
are bounded by the major and minor axis so the distances of each individual element
can be approximated by the average length of the two axis. The magnetic field
approximation at the center of a elliptical coil is then:
By solving the magnetic field, the magnetic flux can now be solved since it is the
product of the magnetic field and the area bounded by the generating coil. For a coil
B = I*N*|Uo*P/ (47i*((a+b)/2 )2 ) * sin 0 (3.7)
22
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
with more than one turn, the number of turns is multiplied to the initial product. Thus
the equation for magnetic flux is:
< j , = N * B * A (3.8)
The values used in the equation are number o f turns (N), magnetic field at the flux
location (B), and the elliptical coil (A). The area of an ellipse is known, unlike the
perimeter, and is given as:
A = 7i * a * b (3.9)
Where a is the length of the major foci and b is the length of the minor foci. The
magnetic flux can also be defined as the product of a coil’s inductance and the
current flowing inside the coil wire. The inductance can be thought as a
proportionality constant that defines how much magnetic flux is present for a set
amount of current.
® = L * I (3.10)
Combining the two equations for magnetic flux, the inductance of a coil
appears that it could be easily calculated from field strength. This is true when the
field strength of a coil is uniform, such as the case of a solenoid, however calculating
23
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the inductance o f an elliptical coil is difficult because the field strength is not
completely symmetric within the area o f the coil. That is why the field strength at the
center o f the coil cannot be used to calculate the inductance o f the coil. Fortunately,
it is still possible to approximate the inductance o f an elliptical coil by assuming that
an averaged field strength surrounds the whole area of the coil instead of the
divergent field strengths. Using general knowledge that the field strength of a coil is
strongest at the edges and weakest at the center, the average field strength level will
fall between the edge and the center of the coil.
Contrary to initial belief, the averaged strength magnetic field does not occur
at the midpoint between the edge and the center of the coil. This is because the field
strength is stronger when it is closer to the source of the current element [Tipler
1991], It is estimated that the average field strength occurs around the coil at a
distance 1/8 of the radius R. As shown in Figure 3.1, the coil is divided at the 1/8 R
point. One side of the division will have a distance of 1/8 R and the other will have a
distance of (1-1/8) R. Due to the unsymmetrical location, the distance between the
field strength point and the coil current element is different for each element. While
each current element distance can be calculated, it will be tedious to find all the
distances separately. To reduce the amount o f calculations, the average distance of
the current elements will be computed. Since the field strength of interest is located
at a point that does not provide symmetry, it will be easier to calculate the field
strength of the two elliptical coil components separately.
24
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 3,1: Coil Dimensions
To find the average current element distance for each section of the
transmitter coil, the distance of the new minor axis at the 1/8 R location needs to be
known. This can be solved by forming a triangle using the minor axis, the line from
the minor axis to the 14 R point and the major axis from the 1 4 R point to the 1/8 R
point. The 1 4 R point is used because the elliptical coil is essential a circle until that
distance. So the distance from the minor axis to the 1 4 R point will be 1 4 R. The other
known side is on the major axis and it will be (14 - 1/8) R. With two sides of the
triangle known, the length o f the third side a.k.a. the new minor axis could be
calculated. Using the Pythagorean theorem, it was determined that the length of the
new minor axis is 0.33 R. The minor axis length will be used as a midpoint between
the two sections o f the coil to find the average current element distance.
For one section of the coil, the current element distance will be between 1/8
R and 0.33 R so the average length will be 0.22 R. The other section will have a
25
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
distance between (2 - 1/8) R and 0.33 R so the average distance will be 1.10 R. For
each section, the number of current elements will be based on their respective
perimeters. Since the smaller section cuts off at 1/8 R on the major axis, the
perimeter will be 1/ 16 of the total P. The larger section gets the remainder and has a
perimeter o f (1- 1/16) P. The magnetic field strength calculations of the two sections
can be combined to give:
B = P N * p 0*[ (1 / 16)P / (4n * D l2 ) + (1- 1/ 16)P / (4% * D22)]
B =I*N*po* [{(1/ 16) * (2uR)} / {(4tx) * ((R/8 + 0.33R) / 2)2} +
{(1-1/16) * (2tcR)} / {(4n) * ((2R -R /8 + 0.33R) / 2)2}] (3.11)
The value o f R in the equations above is the average radius of the major and minor
inner diameters o f the coil so R Is equal to (a + b) / 2. The magnetic field strength
calculated by this equation will provide a close approximation of the average field
strength value in the coll so it can be used to find the inductance of the coil. Thus the
approximate inductance of a spiral elliptical coil is:
L =N*B * A/I
L - p0N 2 * [{1/ (16)* (2tiR)} / {(4rc)* ((R/ 8 + 0.33R) / 2)2} + {(1-1/
16) * (27rR)}/{(4n)* ((2R - R/8 + 0.33R) / 2)2}] * (nab) (3.12)
26
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Once again, R is (a + b) / 2. The inductance of the transmitter coil can be solved in
order to determine the amount of power that is transferred to the BION coil. There
exists a slightly more accurate equation for a spiral coil created by Harold Wheeler
[Wheeler 1928], The equation is:
L = (N * R)2 / (8R +11 W) (3.13)
The N is the number of turns, R is the average radius so it will be (a + b) / 2 for an
ellipse, and W is the width of the coils, which changes depending on the coil turns.
Since it is more accurate, the Wheeler equation will be used for quick inductance
calculations, but it cannot be broken down to be used for analytical purposes since
the equation is based on empirical data. That is why the first equation for the
transmitter coil is necessary for the calculation o f mutual inductance between the
transmitter and the BION. The mutual inductance is needed to find the amount of
power transfer.
3.3 Mutual Inductance Calculation
The transmission of power from the transmitter to the BION is not enough to
generate a stimulus pulse directly. Fortunately the stimulation output of the BION
only need to be brief pulses, so there is enough time for a capacitor to store sufficient
charge for stimulation from the induced current. Due to size constraints of the BION,
27
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
it is not possible to place a common capacitor internally. Instead o f a separate
capacitor, the natural capacitance created between the tantalum electrode of the
BION and body fluid is used to store charge [Loeb and others 2001]. The capacitor
must be charged to the 17 ¥ compliance voltage to generate a current strong enough
to stimulate the muscle. The charging must be completed within 20 ms since the
stimulation pulse fires at a rate up to 50 Hz. Therefore the amount o f current to
recharge the capacitor within the time frame must be found.
To determine the minimum amount o f current that is necessary to recharge
the capacitor to the compliance voltage, the amount of charge that is lost from the
stimulation pulse needs to be found. Since the capacitor is charged to 17 V and the
electrical resistance of the muscle is roughly 500Q, the stimulation current is:
Istim = Vc a p / Rmuscle = 17 / 500 = 0.034 A (3.14)
The stimulation pulse duration is 100 ps, so the amount o f charge lost during
stimulation can be found using this equation [Thomas 1998]:
Istim ~~ d Q]ost / d T stjm
d Qiost = Istim * d Tstim = 0.034 * 1 0 0 x 1 0 '6 = 3.4 x 1 0 '6 (3.15)
The charge that was lost must now be recovered within the recharge time period of
20 ms. The recharge current can be solved by this equation:
28
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Irecharge d Q iost * d T recharge
(3.16)
A recharge current of 170 pA is needed for ideal operation, but a slightly lower
current of 0.1 mA should be sufficient for minimal output. In fact the recharge rate
for the BION can only be selected as 0.1 or 0.5 mA. The amount of voltage that is
required on the BION coil can be approximated to be equal to the capacitor voltage
of 17 V. Therefore to extract 170 pA from 17 V, a 100 KQ resistor is placed in
parallel with the BION coil.
The current in the BION is induced by the RF transmitter so the amount of
current the transmitter coil needs in order to induce the required current in the BION
coil must be found as well. The relationship between the two coils can be solved
through the concept of mutual inductance. In fact the equation for the voltage
relation of a transformer is a simplified version of mutual inductance calculation.
Essentially the mutual inductance equation is a variation of the self inductance
equation. For reference, the inductance o f the BION coil is that o f a solenoid
expressed as:
number of turns. The p © is the value o f permeability of free space, so under normal
L = p 0 * (N /l)2 *l*(7iR 2) (3.17)
Where 1 Is the length of the solenoid, R is the radius o f the solenoid and N is the
29
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
circumstances it would be 0.032 pH/in. However due to the fact that the BION coil
is wrapped around a ferrite rod, the permeability of the core material needs to be
used in the calculation. The ferrite rod used in the BION is listed as having an initial
permeability of 800 times po. The actual permeability is much lower than the initial
value because the permeability is dependent on the length to diameter ratio of the
ferrite rod [DeMaw 1996]. The BION ferrite rod has a ratio o f 4, making the
effective permeability 15 times po. Since N is 190 coil turns, 1 is 0.24 in, R is 0.036
in for the BION, the calculated inductance of the BION is 292.9 pH. This value is
close to the measured value of 275 pH.
With the value of the voltage across the BION coil known for the desired
current, the mutual inductance value can be used to determine the amount of current
needed for the transmitter coil. The equation is as follows:
ItranS= l / M * J V b lo n (3.18)
The value o f M is the mutual inductance. The equation for finding the transmitter
current is similar to that of the finding the current due to self inductance. This is
because the principal of mutual inductance is based on the self inductance of the two
coils involved.
Mutual inductance is possible due to the magnetic flux generated by a coil
which was previously given in Equation 3.10 as:
30
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
( j > = L * I (3.19)
L is the inductance o f the coil and I is the current flowing in it. When there are two
coils aligned next to each other, the magnetic flux of one coil can be passed through
to the other coil [Tipler 1991]. The amount of magnetic flux that passes through the
secondary coil is dependent on the mutual inductance of the two coils. The magnetic
flux generated in the secondary coil is then the combination o f the flux made by self
inductance and mutual inductance:
Where L2 is the self inductance of the secondary coil, I2 is the current in the
primary coil.
Since there is no initial current by the BION, there is no magnetic flux
produced by its own inductance. For this reason, the magnetic flux in a BION can be
defined as:
Thus the mutual inductance of BION (solenoid) coil and transmitter (spiral flat) coil
( j) 2 - L2* I2 + M 12 * II (3.20)
secondary coil, Mi 2 is the mutual inductance of both coils and fi is the current in the
( j) Mi3lon/trans Itrans ~ N bion * Btrans (3.21)
is:
31
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
M = ( j > / It = (Nb * Bt * Ab ) / It
M = m, *Nb *Nt *[{1/ (16) * (2t c R)} / {{An) * ((R/ 8 + 0.33R)/ 2)2} +
{(1-1/16) *(27cR)} / {(4t x ) *((2R- R/8 + 0.33R)/2)2}] *Ab (3.22)
Essentially the mutual inductance equation is the same as Equation 3.12 for the
transmitter coil, but the area used for the equation is that of the BION, not of the
transmitter. By solving the mutual inductance of the transmitter coil and BION coil,
it is possible to calculate how much current is induced in the BION coil by the
transmitter coil. The values required to solve the equation are 15 * 4u x 10-7 H/m or
0.48 pH/in for po, 190 coil turns for the BION, and 7r(0.036)2 in2 for the BION area.
The mutual inductance could be solved except for one unknown variable. It is the
number of coil turns for the transmitter. The higher number of turns on a coil will
produce more field strength for the same amount of current, but it becomes
problematic in maintaining a high current for a coil with a large number of turns. To
determine the ideal number of coil turns, experimenting with the Class E components
must be accomplished.
3.4 Determining Number of Coil Turns
By looking at the equation for field strength and inductance, it is obvious that
with the same current strength and dimensions, the coil with the most number of
turns would produce the highest amount of field strength. However increasing the
32
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
number o f turns of a coil will also increase the resistance of the coil [Ko 1977].
Therefore, even if the coils with different number of turns have the same current
flowing in them, they would all be at different potentials. Increasing voltage on the
transmitter coil may cause problems since it will result in an voltage increase on
critical components, such as the switching MOSFET and capacitors providing
resonance. Since the circuit is voltage regulated, a test comparing field strength
output among varying current and coil turns is performed.
Table 3.1: Dimensions of Coils
# o f Turns: 4 # of Turns: 5 # of Turns: 6 # of Turns: 7 # o f Turns: 8
Dimensions Dimensions Dimensions Dimensions Dimensions
Diam. of a „ .
ID (in):
Diam. of „ 1
a, ID (in):
Diam. of a Q
ID (in):
Diam. of a „ Q
ID (in):
Diam. of a „ „
ED (in):
Diam. o fb . .
ID (in): 4 '5
Diam. of . ,
b, ID (in):
Diam. o fb . _
4 5
ID (In):
Diam. of . _
b ID (in):
Diana, o fb . .
4 5
ID (in):
Diam. of a . „ _
OD(in): 1 0 '2
Diam. of
a, OD 10.5
(in):
Diam. of a 1 „ _
OD (in):
Diam. of a 1 n „
OD(In):
Diam. of a . 1
OD(in):
Diam. o f b _ ,
OD(in): ■
Diam. of
b, OD 6.1
(In):
Diam. o fb , 7 _
OD(in):
Diam. o f , .
bOD(in): 3
Diam. o fb 7
OD(in):
Wire Ga.: 12 Wire Ga.: 12 Wire Ga.: 12 Wire Ga.: 12 Wire Ga.: 12
Length 1Q ()
(in.):
Length 1 2 2
(m.):
Length 1 4 9
(m.):
Length .
(in.):
Length 2 0 0
(in.):
Avg Dia + „
1 4 W (in.):
Avg Dia +
1 4 W (in.):
Avg Dia + 7.62
1 4 W (in.): 5
Avg Dia +
1 4 W (in.):
Avg Dia +
1 4 W (in.):
Measured , . .
S-In.(uH):
Measured „ . Q
S-In.(uH):
Measured . „ A
S-In.(uH):
Measured 1 „ 7
S-In.(uH):
Measured _ 7 1
S-In.(uH):
Calculate _ „
S-In(uH):
Calculate „
S-In(uH):
Calculate .
S-In(uH):
Calculate . 7 ~
S-In(uH):
Calculate _
S-In(uH):
Calculate „ ...
M-In(uH):
Calculate A .n
M-In(uH): ° ' 4 9
Calculate ~
M-In(uH):
Calculate A , n
M-In(uH): ° ' 6 9
Calculate _
M-In(uH):
33
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
To find the optimal number of turns for the transmitter coil, 5 coils were
created using incremental loop numbers while keeping the inner diameters the same
(Table 3.1). The smallest coil had 4 turns, while the largest had 8 turns. It was
necessary to build the actual coil instead o f modeling it because the Class E
components are dependent of the transmitter coil. The components must be
calculated and assembled since the coils need to be attached to the transmitter
driving electronics with the circuit tuned to 2 MHz to analyze their magnetic field
characteristics.
The self inductance and the mutual inductance with the BION for each coil
were calculated and are shown in Table 3.1 as well. With the mutual inductance
known, the required amount of transmitter current can be solved by using Equation
3.18. It was determined that the ideal BION recharge current is approximately 0.1
mA of DC current with Vbion at 17 V. However the current was rectified from AC in
the BION circuit, so Vbion needs to be a sine wave which is approximated as 17 cos
(2%f * t). Since the mutual inductance values of each coil was close to each other, it
was expected that the optimal amount o f transmitter current would be close as well.
Therefore the average mutual inductance value of 0.59 pH is used for the transmitter
current calculation for all the coils:
Itrans = 1/ M * J Vb l0 n
Itrans = 1/ 0.59 x 10'6 * 17 / (2 ti * 2 x IQ6 ) sin (2n (2 x 106 )t)
Itrans ~ 2.2 sin (2k (2 X 106)t) (3.24)
34
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
According to the calculation, the BION coil needs an AC current running at 2
MHz with an amplitude of 2.2 A for the ideal amount of coupling. However, using
the Ideal amount o f current may not be necessary. The 2.2 A current Is needed when
the transmitter is recharging the BION at maximum operation, which would be 100
ps pulse widths repeating at 50 Hz. Under less than maximum operation, the 2.2 A
transmitter current is excessive. For example, If the pulse repetition rate was 5 Hz
instead of 50 Hz, the recharge current will only have to be 0.2 A. So as a
compromise, a current of 1 A (half the maximum amount) will be used in the
transmitter coil. The slight decrease in current should not cause a major problem in
transmitter while providing better use of current draw.
With the transmitter coil current known, the Class E components can be
finally solved using the Class E relation equations presented in section 3.2. The
calculation process must be redone for each coil, but for demonstration the 5 turn
coil calculation is presented, hi the case of the 5 turn coil, it has a inductance of 9.2
pH and was measured to have a Q of 122. From these two values, the transmitter coil
resistance can be solved by:
R c = ooe * L 2 / Q i
Rc = (2n * 2 x IQ6 * 9.2 x 10'6 ) / 122 = 0.947 Q (3.25)
Using the inductance value of the coil (L2 ), the current source inductor (Li) value can
be solved by the equation:
35
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Li = L2 * 3 = 29 uH
(3.26)
For Li, 33 uH will be used instead o f 29 uH because it is the closest available
inductance rating to 29 uH. Also the value of Lj does not need to be specific as long
as it is 3 to 10 times the size of L2 so it will be appropriate. The voltage across the
shunt capacitor and MOSFET is expressed as:
V s’ = (A /2 it)(l /2 D -D 2)) (3.27)
In this case, since the input voltage is 12 V, the voltage is a waveform that fluctuates
from 0 to 24 V. Thus Vs w 12 V. The impedance o f the circuit is solved as:
|Zs |-V s ’/I2 (3.28)
Both waveforms are sine waves, with the maximum voltage amplitude being 12 V
and the maximum current amplitude being 1 A, making the impedance 12 0 . The
series branch reactance is solved using the coil resistance and the circuit impedance:
Xs = V(Z 2s - R 2c) (3.29)
Since the coil resistance is small, the reactance o f the branch is 12 O as well. The
series and shunt capacitor values are solved using the reactance and the equations:
36
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
C -2 — 1 / (®e (® e * L2 — Xs))
Cj = 1 / (COE * Xs)
(3.30)
(3.31)
So in order for the 5 turn coil branch to resonate at 2 MHz, the series capacitor must
be approximately 700 pF and the shunt capacitor must be 5000 pF.
The Class E components were calculated for their ideal performance for each
coil turn, but certain characteristics.about the electronics were kept constant in order
to measure the differences between coil responses though. The characteristics that
were kept constant are the shunt capacitor is a 3900 pF capacitor and the transmitter
coil voltage is 270 V. The value for the 3900 pF was selected because its reactance is
small at 2 MHz. The reactance is 20.4 Q, when calculated by the capacitor reactance
equation. A larger value could have been used, but the 3900 pF was readily available
and it would be sufficient to perform the coil test. Also for the comparison, the
current was regulated by controlling the pulse width of the MOSFET switch time to
test difference in components. Once again, it would seem to make more sense to
compare the coils using a constant current, but the circuit is voltage sensitive so that
is why voltage is kept constant at 270V (peak to peak). This allows the voltage from
going over 300V as the maximum value for the capacitors used to create the resonant
circuit is 300V. On a side note, 500V rated capacitors do exist so the voltage could
be raised higher for future tests. However, the voltage produced at the coil also
affects the voltage of the MOSFET o f the circuit so it is better to maintain the
voltage at a safe 270V. Using these values for driver electronics, the power supply
37
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
current draw, series capacitance, and coil current were recorded. Current was
measured by reading the difference in current draw from the power supply with the
coil driver attached and unattached. Coil current was measured with a current probe.
The field strength generated at the center of the coil is calculated as well. The center
is chosen because the field strength is at its minimum, which would be the limiting
factor. As shown in the previous section, the equation is:
B - I*N*po*27c*((a+b)/2) / (4n*((a+b)/2)2) * sin 0 (3.32)
Where I is the current flowing in the coil, N is the number of turns, a and b are the
foci of the elliptical coil and is the angle of the current element to the location of the
field strength. For convenience, inductance of the spiral elliptical coil is:
L = poN2(7iab)* [{(R/e) /(R/e +0.48R)2} +{((2eR-1)/ e) /(2R -R/e
+0.48R)2}] (3.33)
Since the field strength is dependent on both current and coil shape, the experiment
will indicate which variable has a larger effect as the current will be decreased while
the number of coil turns will be increased.
During the experiment, there was a problem with the 4 turn coil because it
could not produce 270 V with the components used in this test. The maximum
voltage the 4 turn coil was able to produce was 190 V. The limitation was created
38
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
because the MOSFET pulse width was not able to increase more than 120 ns. Due to
this, the current flow was limited thus the voltage could not go higher than 190V for
a 4 turn coil. The comparison voltage could have been set to 190 V, but then the 7
and 8 turn coils would have trouble producing a voltage that small. Therefore it was
decided that the 4, 5 and 6 turn coils be compared at 190 V to determine the
relationship between those coils. By having the two sets of voltages as comparison, it
is possible to analyze the characteristic of all the coils.
Table 3.2: Coil Current, Max Voltages, and Power Supply Current Draw
3900 pF Shunt Capacitor (a ), 270 V
# of Coil Series Current Pulse . Calculated Field
Turns Current Capacitor Draw Width Strength (@ x=0)
4 N/A llOOpF N/A N/A N/A
5 2.21 A 713 pF 190-200 mA 1 2 0 ns 7.29E-5
6 1.65 A 492 pF 170-180 mA 96 ns 6.46E-5
7 1.26 A 410 pF 160-170 mA 78 ns 5.71E-5
8 1.01 A 303 pF 160-170 mA 62 ns 5.10E-5
Table 3.3: Comparison o f 4, 5 and 6 turn coils at a lower voltage
3900 pF Shunt Capacitor @ 190 V
# o f Coil Series Current Pulse Field Strength (@
Turns Current Capacitor Draw Width x=0)
4 zVO A 1100 pF 170 mA 116 ns 5.93E-5
5 1.60 A 713 pF 140 mA 90 ns 5.28E-5
6 1.24 A 492 pF 130 mA 76 ns 4.76E-5
As shown in the tables above, the current starts to drop off as the number of
turns in each coil becomes greater. This is expected since the coil voltage stays the
39
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
same while the resistance of the coil increases due to the increase in number of turns.
Also because o f the decrease in current, the field strength gets smaller as the number
of turns gets bigger in our experiment. For this reason, the 5 turn coil produces the
highest amount of magnetic field strength when the coils voltage is set at 270 V.
When the 4 turn and 5 turn coils are compared when they are running at 170V, the 4
turn coil produces higher field strength. This means that if it is possible to raise the
current, the strongest field strength would occur with a 4 turn coil at 270V. However,
the 4 turn coil cannot produce a high current with the components used. With the
experiment’s desired conditions, the 5 turn coil produced the highest field strength.
If it is possible for the voltage level to be increased by changing the
components o f the circuit, higher current can be used for a coil with a greater number
of turns. Unlike the case for the 4 turn coil, in which there was not enough current to
generate the maximum voltage, the other coils with more turns were not current
limited. Under different circumstances, such as a higher maximum voltage, the 6 turn
coil may be able to produce more field strength than the 5 turn coil. Therefore the
optimal number of turns for a transmitter coil, must be evaluated each time the
maximum parameters of the circuit are changed.
3.5 Coil M aterial Test
There are several factors that require attention when building a coil. In the
previous section, it was determined that the number of turns on a coil has an effect
40
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
on the properties o f the magnetic field produced by the coil. In this section, the
effects due to the use of different wire materials is examined. Since the resistance of
a wire varies depending on the material used, the amount of current that flows
through should be different. This will have an effect on the Q, or the quality factor of
the coil, which determines the amount o f power transmission at resonance.
Resonance is used in the transfer energy to the transmitter to the BION since
It acts as a band-pass filter, which accepts power for a particular frequency range
while rejecting other frequencies. Q is expressed by the reactance over the resistance
of the coil and it describes the resonance characteristic. A high Q will offer a sharp
resonance, but a narrow bandwidth while a low Q has a broad bandwidth with a less
sharp peak [Tipler 1991]. The ideal coil will have a Q that Is high enough to offer a
sharp resonance, but at the same time have some slack for shifts in frequency. To
find an appropriate Q, two different wire materials were used for the experiment, 12
AWG copper wire and 14 AWG Litz wire (2400 strands of 44 gauge wire).
Litz wire is a special type of wire made with very fine, individually insulated
strands that are twisted into bigger strands. Under certain conditions, Litz wire has an
advantage over normal wire. A problem associated with normal wire is that when
there is high frequency AC current, eddy currents develop within the cross-section of
the wire. This contributes to additional resistance because the currents can only flow
close to the surface of the conductor, a condition called “skin effect” [Sullivan 1999],
In the case of Litz wire, the insulation separating the strands from each other prevent
eddy currents from circulating through the whole wire's cross-sectional area. Thus
41
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the resistance of Litz wire is less than a normal wire and the currents in each strand
are similar to each other. However, the Litz wire is much more expensive and
difficult to use compared to normal wires. The experiment will determine if Litz wire
is better to use than normal wire in the case of the BION transmitter coil.
The 14 AWG Litz wire was chosen rather than a 12 AWG Litz wire because
of its immediate availability and it was expected that the small difference in wire
diameter will not significantly affect the comparison of the wire characteristics.
Since there are only two comparisons for the experiment, it is possible for the data to
be correct only for the coil being examined. To make sure that the difference
between the copper wire and the Litz wire is consistent for all shapes, another pair of
coils with a different shape is also analyzed. The alternate coil shape is also a spiral
coil, but it is circular with an inner diameter of 3.5 inches. The two different coil
configurations and two different wire materials, which resulted in four combinations
are described in Table 3.4.
Table 3.4: Coil Attributes, Dimensions and Wire Material
Copper Coil (Large) Litzwire Coil (Large)
C opper C oil (Small)
Litzwire Coil (Small)
Inductance: 9.729 pH Inductance: 10.4 pH Inductance: 11.15 pH Inductance: 11.1 pH
Length o f Wire: 122
in
Length of Wire: 121
in
Length o f W ire: 117.2
in
Length of W ire: 110
in
Length, ID: 9 in. Length, ID: 9 in.
Length, ID: 3.5 in.
Length, ID: 3.5 in.
W idth, ID: 4.5 in. W idth, ID: 4.5 in.
W idth, ID: 3.5 in.
W idth, ID: 3.5 in.
Length, OD: 10.5 in. Length, OD: 10 in.
Length, OD: 6 in
Length, OD: 5.25 in.
Width, OD: 6 in. W idth, OD: 6 in.
W idth, OD: 6 in.
W idth, OD: 5.5 in.
# o f Turns: 5 # o f Turns: 5
# o f Turns: 8
# of Turns: 8
Wire: 12 Ga. Copper
W ire: 2400 Strand
Litzwire
W ire: 12 Ga. Copper Wire: 2400 Strand
Litzwire
42
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The coils were made to have same inner diameters, but due to the difference
in width between the Litz and the copper wire, the outer diameter of the Litz wire
coils are smaller than their copper counterparts. The 4 different coils were then
connected to the driver circuit in order to obtain the current flow in the coils. The
procedure is similar to the one used to compare the coils with different number of
turns. Once again, the shunt capacitor is a 3900 pF capacitor and the coils are
connected to the appropriate series capacitor to resonate at 2 MHz. However in this
experiment, the current through the coils will be kept constant instead of the voltage.
Since the coils are created to be the same, if the current is the same in the coil then
the field strength should be the same. The changes in voltage is of interest due to the
resistance o f the coil being different. The coil current, coil voltage, series capacitance
and power supply current draw are recorded for each coil and are listed in Table 3.5.
Table 3.5: Summary o f Coil Characteristics
C opper (Large) Litzwire (Large) Copper (Sm all) Litzwire (Small)
Inductance 9.729 pH 10.4 pH 11.15 pH 11.1 pH
Series Capacitance 713 pf 603 p f 570 p f 570 p f
M ax V oltage 246 V 262 V 264 V 264 V
Current D raw 140 mA 140 mA 160 mA 140 mA
Coil Current 2.04 A 2.04 A 1.92 A 1.96 A
27i£L Reactance 122.2582 O 130.6902 O 140.1150 0 139.4867 O
V/I R eactance 120.5882 O 128.4313 0 135.4167 0 134.6939 O
There were noticeable differences between the coils. First, the small circular
coils had higher inductance than the large coils. This was expected due to the shape
43
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
difference between the coils and from calculation. However, the difference in
inductance between the large elliptical coils was not expected, since the difference in
inductance between the small circular coils is minimal. It is believed that the
difference was caused by the tendency o f the Litz wire coil to deform into a circular
shape rather than the oval shape it was formed to be. The Teflon coating of the Litz
wire made it difficult to glue the wire together so masking tape was.used to shape the
coils. The fact that the inductance between the circular coils was similar Is consistent
with the Idea that the elliptical Litz wire coils were reshaping into circular coils.
The inductance between the two large coils were different so the reactance of
the coils were different as well. The reactance of the coils were determined using two
separate techniques. The first method is solved by using the equation:
X - 2nfL (3.34)
Where the reactance is X, the frequency in Hz is f and the Inductance is L. The
second method requires the amount o f current flowing in the coil at the operating
voltages. The reactance can be solved using:
V / I = X (3.35)
Where V is the voltage across the coil and I is the current in the coll. The first
method would be the theoretical approach as the second method Is based on actual
44
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
measurements. Reactance was calculated using both the 2%fL equation and R=V/I
equation, to see if there was a difference between theoretical and actual values. As
shown in the Table 3.5 above, the differences between the theoretical values and
actual values o f reactance are similar for all of the coils. So the difference in
reactance between the two types of is due to the distortion of the Litz wire coil rather
than a factor caused by the Litz wire material.
Judging by the operating characteristics of the four coils, there does not seem
to be much of a difference between the copper wire and Litz wire. This was expected
since the physical dimensions o f the coils are similar and the data validates that fact.
The Q, the value of interest, is expected to be different between the two wire types
and will be analyzed next. To measure Q, a Hewlett-Packard 4192a LF Impedance
Analyzer was used. The coils were tested for their inductance and Q behaviors over
the frequency range of 100 KHz to 3 MHz at intervals o f 250 KHz.
Table 3.6: Key Values for Measurements
Large C opper Coil Large Litzwire Coil Sm all Copper Coil Sm all Litzwire Coil
M fH)
Q L (pH )
Q
L (p H )
Q
L (fiH )
Q
Min:
Min:
Min: Min: Min: Min: Min: Min:
9.65 @
750 KHz
98.5 @
100 KHz
10.28 @
250
KHz
192 @
100
KHz
11.17 @
2250
KHz
101.4 @
1000
KHz
11.03 @
250
KHz
89 @ 100
KHz
M ax: Max: M ax: Max: M ax: Max: Max: Max:
9.88 @
3000
KHz
143 ©
2500
KHz
10.56 @
3000
KHz
950 @
1000
KHz
11.25 @
3000
KHz
128.7 @
250 KHz
11.2 @
3000
KHz
460 @
1250
KHz
Avg: Avg: Avg: Avg: Avg: Avg: Avg: Avg:
9.73 130.11 10.37 669 11.18 112.08 11.08 355.92
StdJDey: Std.Dev: Std.Dev: StcLDev: Std.Dev: Std.Dev: Std Dev: Std.Dev:
.07 13.3 .09 215.5 .04 8.84 .06 109.22
45
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Large Copper Coil Attributes vs. Frequency
9.95 ~ r
y =
9.9 -
9.85 -
X
3
9.8 -
®
O
9.75 -
C
2 9.7 -
o
■ o 9.65 -
S
9.6 -
9.55 -
9.5 -
0.0Q39X3 - 0.4133X2 + 8.2521 x + 95.881
R2 = 0.9491
y = 0.0038X2 - 0.04x + 9.7709
R2 = 0.9853
180
140
120
100 a
■ 80 "5
>
60 O
40
20
\V ~ \,j n>
Frequency (KHz)
Inductance
Q
-Poly. (Q)
• Poly. (Inductanc^)
rSi oP r S 3
Figure 3.2: Inductance and Q of Large Copper Coil
Large Litzwire Coil Attributes vs. Frequency
y = 0.002ix2 - 0.0068X + 10.289
10.55
0.45
10.35
10.25
70.789x + 525.1x - 274.6 2.7121X
0.15
0.9597
- 800
- 600
\ > \ 'J \ x
Frequency (KHz)
Figure 3.3: Inductance and Q of Large Litz wire Coil
Inductance 200
Q 100
Poly. (Q) 0
Poly. (Inductance)
46
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Small Copper Coil Attributes vs. Frequency
11.05
i.0977x3 + 2.5957X2 ■
R2 - 0.8711
11.25
« 11.15
y = 0.0033X2 - 0.0448X + 11.283
R2 = 0.9908
■ Inductance
♦ Q
— Poly. (Inductance),
Poly. (Q)
140
120
100
80
80
40
0
Q
^ ^ ^
Frequency (KHz)
Figure 3.4: Inductance and Q o f Small Copper Coil
Small Litzwire Coil Attributes vs. Frequncy
y = 0.9103x3 - 26.469X2 + 223.33X -119.717 500
450
x
©
G
C
2
o
3
■ D
y = 0.001 Sx2 - 0.0074X + 11.04
R2 = 0.9992
©
a
■ Inductance 100
| ♦ Q Value 50
_ i i | i___ |_ l — Poly. (Inductanqf)
✓ C s ,c\ c, J — Poly. (Q Value)
^ —
kj * 0 q p
Frequency (KHz)
Figure 3.5: Inductance and Q of Small Litz wire Coil
47
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Q Value G Value
As shown in Figure 3.2 - 3.5 above, the inductance closely follows a second
order polynomial curve for both types of wires. Overall, the inductance varied
between .25 and .5 pH for the entire frequency range for all the coils indicating that
frequency does not have a large effect on inductance. The large copper coil did have
a noticeable lower inductance than the large Litz wire coil. But as mentioned earlier,
the difference in inductance was caused by the tendency of the Litz wire coil to
deform into a circular shape rather than the oval shape it was supposed to be. The
small, circular coils had similar inductance due to the lack of deformation. To
summarize, the impedance analyzer measurements reinforces the calculated
inductance values.
There was a significant difference in Q values between the copper wire and
Litz wire. Q values are mostly dependent on the type of wire in the coil and not the
geometry o f the coil. This is illustrated by comparing the Q of the copper wire and
the Litz wire regardless of the coil shape. For both configurations, copper wire varied
only a little, approximately 20, over the entire frequency range. In comparison, the Q
for the Litz wire had fluctuated approximately 700 from the minimum to the
maximum over the frequency range for Litz wire. Since Q is determined by dividing
reactance with resistance of a coil and the reactance o f the copper and Litz wire coils
are similar, it would indicate that the resistance of the coil is different between the
coils. Litz wire is supposed to have a lower resistance compared to a standard wire
due to its high number of individually insulated wire strands so the data is consistent.
The Litz wire provides a higher Q than the copper wire.
48
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Unfortunately, the Litz wire Q may be too high and prove to be not ideal for
this application. The Q can also be expressed as the resonance frequency over the
resonance width, as shown below:
Q = f0 / A f (3.36)
Having a high Q like 700 means that at a resonant frequency of 2MHz, the
transmitter only has a bandwidth of 3 KHz. For the copper wire, the Q is much lower
at 130 so the bandwidth increases to 20 KHz. Not only is the Litz wire Q high, it is
also unstable as it is high for some frequencies and as low as the copper wire Q for
other frequencies. An unstable Q will have an effect on the sensitivity of the
coupling between the transmitter and BION. During the experiment, we came across
an article in the April 1987 issue of "The Lowdown" in which the author, Mitchell
Lee states that "Litz wire fails to be useful at frequencies above 2 MHz, as
capacitance effects begin to dominate and the relative size of a single conductor with
similar loss characteristics becomes small enough to be practical". Since we are
driving the coils at 2MHz, the advantages of using Litz wire instead of copper wire
may not be as great. In fact, Litz wire's disadvantages (difficult and expensive to
obtain) may overcome its advantages. Due to this reason, copper wire is chosen as
the optimal wire type for the large spiral coil. Perhaps for applications where a lower
frequency and a different coil configuration are used, a Litz coil may prove to be
better than a copper coil.
49
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.6 Field Strength M apping
The magnetic field strength produced by the transmitter coil differs
depending on the location and distance away from the coil. Field strength changes
need to be taken into consideration because when the field strength is too weak, the
BION is not able to produce the amount of current it was programmed to make. To
better understand the changes of the magnetic field strength affecting the operation
of the BION, the field strength at various locations around the coil was recorded to
have a better picture of the field shape. A sense coil was used to measure the voltage
induced on a BION coil by the transmitter coil. The sense coil is a 1 cm diameter
loop created from a coaxial cable By connecting the coaxial cable to an oscilloscope,
the magnetic field picked up by the coil can be displayed as voltage. By altering the
voltage levels at the primary coil, the voltage levels at the secondary coil will be
changed as well. By recording the voltages at both the primary and secondary coils,
it becomes possible to find the coupling coefficient. This makes it possible to
confirm the necessary primary coil voltage for the desired BION coil voltage.
The probe coil was positioned to be parallel to the plane of the coil since that
provides the best coupling of the field. The field strength was measured at different
locations inside the coil at a distance 0 cm and 5 cm away in order to create a 3
dimensional field shape. The transmitter used in this experiment had the serial
number CDLCO1180201. The field strength measurements are shown in Figure 3.6
50
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
above. The numbers on the top are the field strength values at 0 cm and the shaded
numbers on the bottom are the field strength values 5 cm away from the coil.
Coil Center
Height = 0 cm
Height = 5 cm (Shaded, lower number)
Figure 3.6: Field Map for CDLC01180201
According to the measurements, the field strength produced by the elliptical
coil was much higher around the edges than the center when the probe was 0 cm
away from the coil. However when the probe coil was 5 cm away, the field strength
around the center points were higher than the values on the edges. This indicates that
the outer field strength produced by the coil makes an ellipsoid shape because at the
edges, the probe has to be closer to the coil in order to match the field strength that is
at the center o f the coil at 5 cm.
51
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The field strength map shows that the field strength changes less in the
middle of the coil than the outer sections of the coil. Therefore, it is better if the
BIONs are tested around the center of coil. BIONs may work well at the outer edges
o f the coil at 0 cm since the field strengths are high, but they would not work as well
at further distances away from the coil since the field strengths are considerably
weaker at the outer edges. For consistent operation it is better for the BION to be
located as close to the coil as possible. In the case for shoulder subluxation
treatment, the coil can be placed around the shoulder so that the BION is located on
the axis of the coil.
It is important to note that the magnetic field that was recorded is an
arbitrary value that is dependent on the size of the pickup probe coil. In other words,
the field strength voltage will be different from those absorbed by the BION coil
because they have different shapes. Therefore the field strength values recorded are
meant to be used as indicators. By having these indication values, it is possible to
compare the performance of different BIONs at the same field strength levels.
52
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 4
4.1 M odulation Test
It is important that characteristics such as the carrier frequency and field
strength do not deviate from the norm, because a BION will not work unless there is
proper coupling between the transmitter and BION. However in some cases, even if
the BION appeared to be receiving enough power for operation and the transmitter
frequency was correct, the BION was not operational. After some experimentation, it
was observed that the modulation index/shape of the magnetic field produced by the
transmitter has a significant effect on the operation of the BION.
As previously stated, the transmitter that provides power to a BION also
transmits data to a BION on the 2 MHz carrier signal. The data is encoded in bit
form (digital logic) by amplitude modulation, where each bit is represented by 16
cycles on the main carrier signal. A bit is determined to be 0 or 1 depending on
whether there is a change in state between the first 8 cycles and the last 8 cycles. The
bit is 1 when there is a transition from the voltage level normally representing a
value of 1 to the voltage level representing a value of 0 and vice versa. This type of
encoding is called Manchester encoding after the University o f Manchester, where
the first recorded use of it occurred in the late 1940's [Forster 2000], Manchester
encoding is a safety measure as it guarantees that the mean carrier amplitude is
independent of the data values being transmitted. From this data, a processor chip in
the BION determines the strength and duration of the BION’s output current.
53
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
However, the Manchester encoding can still fail under certain modulation conditions
thus these incidents were investigated.
There are several modulation properties that need to be investigated to
understand the effects of modulation on the transmitter and when the BIONs stop
responding to the transmitter. The first property to check is the frequency change in
the transmitter coil during modulation. The modulation capacitor is added in parallel
with the shunt capacitor so additional capacitance alters the main transmitting
frequency. It could be problematic if the frequency changes considerably since the
transmitter circuit is quite sensitive. The second property to look for is the rise and
fall time of the modulation. The modulation in the transmitter output responds
gradually because of it high Q, so it may not settle to a stable level in the available 8
cycles for each state. This can be determined by comparing the number of cycles in
the allocated time frame. The third property to analyze is the effects of capacitive
loading when the coil is adjacent to a conductive material such as the body. It has
been noticed that some coil drivers that were able to operate a BION were not able to
operate the same BION when a hand loads them.
To examine the differences in operation due to the different modulation
characteristics, an experiment was created that tests the interaction and sensitivity of
a BION with discretely configured transmitters. The setup consists of a BION in a
special enclosure filled with saline to simulate the environment o f a muscle. The
enclosure, nicknamed the “wet well”, is connected to an oscilloscope so that when a
BION is operating correctly a signal could be observed. Details of the wet well will
54
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
be described later. Since a BION generates current, the wet well has a resistor
soldered on it for voltage measurement. The resistor has a value of 470 O to simulate
the resistance of the muscle. The transmitter is placed on a height adjustable stand,
so that it could be placed at different distances from the BION. ■
In order for the results to not be affected by other variables, the modulation
test requires the transmitter circuit to be modified In a way that the center frequency
and peak field strength stays the same while the modulation index changes. This is
complicated because of the highly intertwined circuit used in the BION transmitter.
Changing the modulation capacitor will affect both the center frequency and field
strength. In order to keep the other variables constant, the shunt (C l), modulation
(C2), and series (C3-C5) capacitors on the circuit need to be changed accordingly.
4.2 Modulation Characteristic Table
To help understand the effects o f modulation, a table consisting of various
combination of capacitors was created. Table 4.1 shows the voltage and modulation
characteristics of the capacitor combinations. The table can also be used to determine
if the pulse width of the MOSFET has an effect on modulation. This is because a low
pulse width provides smooth Class E operation resulting in a clean sinusoidal voltage
waveform. As the pulse width becomes greater, the Class E operation becomes rough
and the voltage waveform starts to distort. There could be complications when
modulation is applied to the rough waveform. For this reason, the values for both
55
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
high (120ns) and low (70 ns) pulse widths are included in the table. The modulation
table was created using the coil driver with the serial number CDLC01180207.
The values that are listed in the table are the maximum and minimum voltage
o f the transmitter coil, modulation index, rise cycle, rise time, and frequency shift
during modulation. The maximum and minimum voltages are measured using an
oscilloscope and connecting the probe to the two ends of the coil. The modulation
index is calculated using the equation:
V - V
y max r M IN
^UAX + VmIN (4.1)
V max is the peak to peak of the maximum voltage amplitude. Vmtn is the peak to
peak of the minimum voltage amplitude. Rise cycle is measured by counting the
number of clock cycles (peaks) it takes the signal to rise from the minimum voltage
to the maximum voltage. The rise time is the amount of time it takes for the
minimum voltage to rise to the maximum voltage. The necessity in recording both
the rise cycle and rise time is to determine the shift in frequency shift during the act
of modulation. The modulation capacitance is added in parallel to the shunt
capacitance, so there is a slight change in center frequency when the modulation is
switched on. By dividing the rise cycle with the rise time, the frequency of the cycle
can be solved. This value can be used to compare any differences from the carrier
frequency of 2 MHz. It is not known whether this change in frequency might be a
cause of change in BION response with different modulation index.
56
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 4.1 : Modulation Comparison Table for 120 ns and [70ns]
Initial values Shunt Cap. (pF)
820 1300 2000 2700 3300 3900
Series Cap. (pF) 1232 [1360] 922 [1030] 802 [900] 735 [780] 713 [746] 702 [713]
Initial Volt. (V) 87 [60] 111 [81] 150 [110] 225 [128] 262 [145] 294 [159]
Min volt (V)
820 1300
Shunt (pF)
2000 2700 3300 3900 Modulation (pf)
330 77.2 [60.4] 110.0 [81.4] 151 [112] 214 [122] 259 [147] 280 [148]
430 77.2 [61.0] 110.0 [82.0] 154 [114] 214 [148] 262 [137] 286 [148]
680 77.2 [60.0] 110.0 [81.6] 154 [114] 210 [137] 262 [136] 286 [148]
820 77.6 [57.8] 109.0 [84.4] 152 [116] 212 [123] 265 [151] 296 [156]
1100 76.0 [55.4] 113.0 [81.4] 147 [112] 210 [142] 267 [145] 299 [158]
1300 78.0 [59.4] 110.0 [81.2] 145 [112] 211 [130] 266 [145] 289 [153]
1500 69.6 [58.4] 115.0 [82.0] 148 [111] 212 [132] 272 [152] 295 [160]
2000 62.0 [57.6] 114.0 [83.2] 150 [110] 214 [136] 272 [145] 305 [168]
2400 74.8 [59.2] 110.0 [81.6] 143 [112] 214 [130] 274 [155] 299 [168]
2700 70.0 [56.8] 111.0 [81.2] 141 [110] 217 [136] 274 [154] 301 [172]
Max volt (V)
820 1300
Shunt (pF)
2000 2700 3300 3900 Modulation (pf)
330 82.4 [64.4] 116.0 [88.4] 160 [120] 222 [129] 269 [157] 284 [161]
430 84.0 [65.0] 117.0 [90.0] 165 [122] 225 [156] 269 [157] 292 [166]
680 89.2 [67.0] 124.0 [96.4] 181 [124] 240 [146] 282 [146] 292 [159]
820 90.0 [70.0] 130 [102] 188 [126] 246 [134] 282 [168] 305 [178]
1100 96.8 [73.4] 137 [106] 203 [121] 263 [155] 299 [162] 312 [183]
1300 102.0 [75.4] 148 [108] 208 [124] 266 [143] 290 [174] 303 [175]
1500 106.0 [76.4] 164 [104] 220 [125] 269 [147] 299 [171] 310 [180]
2000 119.0 [82.0] 175 [109] 240 [131] 281 [158] 313 [169] 327 [190]
2400 130.0 [82.0] 188 [112] 251 [141] 291 [157] 320 [182] 329 [188]
2700 137.0 [86.4] 198 [117] 260 [142] 296 [165] 324 [184] 332 [193]
Mod index(%) Shunt (pF)
Modulation (pf) 820 1300 2000 2700 3300 3900
330 3.26 [3 21] 2.65 [4.12] 2.89 [3.45] 1.83 [2.79] 1.89 [3.29] 0.71 [4.21]
430 4.22 [3 17] 3.08 [4.65] 3.45 [3.39] 2.51 [2.63] 1.32 [6.80] 1.04 [5.73]
680 7.2
1 [5
51] 5.98 [8.31] 8.06 [4.20] 6.67 [3.18] 3.68 [3.55] 1.04 [3.58]
820 7.40 [9 55] 8.79 [9.44] 10.59 [4.13] 7.42 [4.28] 3.11 [5.33] 1.50 [6.59]
1100 12.04
[13 98] 9.60 [13.13] 16.00 [3.86] 11.21 [4.38] 5.65 [5-54] 2.13 [7.33]
1300 13.33
[11
87] 14.73 [14.16] 17.85 [5.08] 11.53 [4.76] 4.32 [9.09] 2.36 [6.71]
1500 20.73 [13
35] 17.56 [11.83] 19.57 [5-93] 11.85 [5.38] 4.73 [5.88] 2.48 [5.88]
2000 31.49 [17 48] 21.11 [13.42] 23.08 [8.71] 13.54 [7-48] 7.01 [7.64] 3.48 [6.15]
2400 26.95 [16 15] 26.17 [15.70] 27.41 [11.46] 15.25 [9.41] 7.74 [8.01] 4.78 [5.62]
2700 32.37 [20 67] 28.16 [18.06] 29.68 [12.70] 15.40 [9.63] 8.36 [8.88] 4.90 [5.75]
57
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 4.2: Rise Characteristics Comparison Table for 120 ns and [70ns]
Rise cycles
820 1300
Shunt (pF)
2000 2700 3300 3900
Modulation
(pf)
330 4 [3] 2 [3] 7 [3] 5 [2] 3 [5] 3 [5]
430 3 [3] 3 [3] 6 [4] 6 [3] 4 [6] 4 [7]
680 3 [4] 3 [4] 5 [6] 6 [3] 5 [6] 5 [7]
820 3 [5] 4 [4] 5 [7] 9 [3] 7 [8] 4 [10]
1100 3 [6] 4 [6] 6 [3] 11 [6] 11 [7] 5 [13]
1300 4 [7] 5 [8] 6 [4] 11 [5] 12 [7] - 5 [13]
1500 4 [3] 5 [7] 7 [5]
11 [5]
13 [8] 5 [13]
2000 4 [4] 5 [8] 9 [6] 11 [5] 13 [10] 7 [14]
2400 4 [3] 6 [5] 9 [10] 12 [5] 11 [14] 9 [14]
2700 5 [4] 6 [5] 8 [11] 12 [6] 9 [12] 10 [13]
Rise Time
(us) Shunt (pF)
Modulation
(Pf)
820 1300 2000 2700 3300 3900
330 2.08 [1.54] 1.02 [1.56] 3.54 [1.56] 2.48 [1.00] 1.50 [2.52] 1.50 [2.48]
430 1.60 [1.60] 1.54 [1.56] 3.04 [2.08] 2.98 [1.50] 2.00 [3.00] 2.00 [3.48]
680 1.64 [2.20] 1.56 [2.12] 2.56 [3.12] 3.04 [1.54] 2.52 [3.04] 2.52 [3.52]
820 1.66 [2.80] 2.10 [2.14] 2.58 [3.72] 4.56 [1.54] 3.52 [4.10] 2.02 [5.04]
1100 1.68 [3.46] 2.12 [3.24] 3.12 [1.60] 5.56 [3.10] 5.54 [3.56] 2.52 [6.52]
1300 2.28 [4.06] 2.66 [4.36] 3.12 [2.12] 5.60 [2.62] 6.06 [3.56] 2.52 [6.60]
1500 2.32 [1.80] 2.70 [3.84] 3.64 [2.68] 5.60 [2.62] 6.58 [4.16] 2.54 [6.64]
2000 2.38 [2.42] 2.74 [4.40] 4.76 [3.28] 5.64 [2.62] 6.60 [5.16] 3.56 [7.16]
2400 2.42 [1.86] 3.32 [2.80] 4.78 [5.52] 6.18 [2.64] 5.62 [7.24] 4.58 [7.16]
2700 3.06 [2.48] 3.34 [2.86] 4.28 [6.06] 6.20 [3.20] 4.62 [6.24] 5.10 [6.64]
Cycle/Time
(MHz) Shunt (pF)
Modulation
(pf) 820 1300 2000 2700 3300 3900
330 1.92 [1.95] 1.96 [1.92] 1.98 [1.92] 2.02 [2.00] 2.00 [1.98] 2.00 [2.02]
430 1.88 [1.88] 1.95 [1.92] 1.97 [1.92] 2.01 [2.00] 2.00 [2.00] 2.00 [2.01]
680 1.83 [1.82] 1.92 [1.89] 1.95 [1.92] 1.97 [1.95] 1.98 [1.97] 1.98 [1.99]
820 1.81 [1.79] 1.90 [1.87] 1.94 [1.88] 1.97 [1.95] 1.99 [1.95] 1.98 [1.98]
1100 1.79 [1.73] 1.89 [1.85] 1.92 [1.88] 1.98 [1.94] 1.99 [1.97] 1.98 [1.99]
1300 1.75 [1.72] 1.88 [1.83] 1.92 [1.89] 1.96 [1.91] 1.98 [1.97] 1.98 [1.97]
1500 1.72 [1.67] 1.85 [1.82] 1.92 [1.87] 1.96 [1.91] 1.98 [1.92] 1.97 [1.96]
2000 1.68 [1.65] 1.82 [1.82] 1.89 [1.83] 1.95 [1.93] 1.97 [1.94] 1.97 [1.96]
2400 1.65 [1.61] 1.81 [1.79] 1.88 [1.81] 1.94 [1.89] 1.96 [1.93] 1.97 [1.96]
2700 1.63 [1.61] 1.80 [1.75] 1.87 [1.82] 1.94 [1.88] 1.95 [1.92] 1.96 [1.96]
58
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 4.1 and 4.2 provides a good range of possible combinations for the
shunt, modulation, and series capacitance values to tune the coil driver for different
characteristics. For this comparison, the main parameter of interest is the amount of
modulation so the capacitor combination that provides the largest acceptable range in
modulation index is desired for testing. To select the appropriate capacitor
combination, the table values require analysis.
4.3 Selection of Capacitor Values
Figure 4.1 is derived from the data above and it shows that the un-modulated
coil voltage increases as the shunt capacitance increases. It also shows the voltage
increases as the pulse width increases. Modulation is added to the Initial voltage, so
it is possible to set a desired voltage by choosing the appropriate shunt value.
350
300
S 250
200
0 >
s
> 150
1
c
1 0 0
50
0 H -
Unmodulated V oltage V s. Steady State Capacitance at 2MHz
600
♦ 70us
■ 120us
B
♦
m
♦
1100 1600 2100 2600
S tead y State Cap. (pf)
3100 3600 4100
Figure 4.1: Voltage and Capacitor Relation
59
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The next data set that was plotted was the modulation index based on the
modulation capacitance value. The graph is shown in Figure 4.2 below. Basically the
modulation index linearly rises as the modulation value increases. There are some
points that are not linear and it is not clear whether that was caused by errors in
measurement or caused by electrical characteristics. It also shows that a low shunt
value combined with a high modulation value will cause a very high modulation
index. This is consistent with the fact that the modulation value is added to the shunt
value when the modulation MOSFET is turned on. With the data from Figure 4.1 and
4.2, it is now possible to predict the initial voltage and the amount of modulation.
Modulation Index vs. Modulation Capacitor at Different Steady State Values
3 2 0
s
! .
❖ 820
01300
A 2000
X2700
X3300
j # 3900
l 6 ^
♦ a
• * • m
i * *
m 9
500 1000 1500 2000 2500 3000
Modulation Cap. (pf)
25
ss
B 1 5
3
S
§ 10
t "
A
$ • •
t
&m
A
m
S x x
a
x v
X X
• • •
♦
A
1 “
x X
X *
• •
0 500 1000 1500 2000 2500 3000
Modulation Cap (pf)
Figure 4.2: Modulation Capacitor and Index Relation
60
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The next process is to test if there is a change in frequency when the
modulation capacitance is added to the shunt value. Figure 4.3 shows that as the
shunt value changes, the series capacitance needs to be changed as well in order to
maintain the coil driver to operate at 2 MHz.
Series vs. Steady State Capacitance at 2MHz
1600 -
1400 -
1200 -
1 1000 -
d
0 800 1
1 i
S 600 - j
to I
400 - ]
!
200 -
0 -
0 500 1000 1500 2000 2500 3000 3500 4000 4500
Steady State Cap. (uf)
Figure 4.3: Series and Steady (Shunt) Capacitor Relation
Since the series capacitance stays the same when the modulation cap is added, the
frequency must shift. Figure 4.4 shows the changes in frequency as the modulation
capacitance value increases. There is a greater shift in frequency when the shunt
value is low and the modulation value is high. This shows that a BION failure can be
a result of frequency shift and not too much modulation as previously thought.
61
♦ 70us
H120us
I i a
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
However for the capacitor values used in the current coil driver, the frequency does
not shift as much so this is unlikely a problem.
F req u en cy v s. M odulation Cap. a t Different S tead y S tate V alues
70 u s P ulse
2.10
2.00 -
X a • ®
i 4 v y • 9 w w
■ X X X
^901 - A X *
♦ ■ i * a a
H
X
S
O 1.80
1.70
1.60
1.50 -
x
#820
B 1300
A 2000
X 2700
X 3300
• 3900
A A
B
B I
♦ ♦
120 u s P ulse
2.10
2.00
_ 1'90
£
S
o 1.80
C
©
§ ■
£
u.
1.70
1.60
1.50
♦ B A A A
* f
X X
A A
m
0 1000 2000 3000
M odulation C ap. (pf)
0 1000 2000 3000
Modulation Cap. (pf)
Figure 4.4: Frequency and Modulation Capacitor Relation
The characteristic of the magnetic field that is assumed to have the most effect on the
operation of the BION is the rise time of the modulation. Since the coding for the
BION to operate is encoded in Manchester encoding, if the time for the modulation
to reach its maximum voltage is too slow, it misses the allocated pulse width for the
code. In other words, the BION cannot operate because it receives the wrong code
62
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
due to the inability of the coil driver to quickly switch its logic states. Therefore the
rise time of the modulation must be completed within the given time of 8 cycles. The
rise time data is shown in Figure 4.5.
Number of Rise Cycies vs. Modulation Cap. at Different Steady State Cap.
14
12
©
£
3
Z
10
♦ 820
■ 1300
*2000
X2700
X3300
• 3900
70 us Pulse
• m
• ® • e
x
A
X A
x ax ■
9 mA X * H
XX ■ A
m ♦ xx x m m
A E M A ^ ^
m xx a ♦ ♦
x
x
0 1000 2000 3000
Modulation Cap. (pf)
16
14
12
jjj 10
O
S
£
£
3
120 us Puise
X X
X X X
X X X X X
•
X A m X
A
A X A •
X X A A B m
X ■X mmu B ❖
m ■ ♦ A
m B # ♦
0 1000 2000 3000
Modulation Cap. (pf)
Figure 4.5: Modulation Capacitor and Rise Cycle Relation
Based on the charts and analysis, the 2700 pF shunt capacitor would be the best
selection. This is because for both the 70 ns and 120 ns data, the modulation
percentage can be varied from theoretically acceptable (3% to 8%) to not acceptable
by using different sized modulation capacitors. The smaller shunt capacitor values
63
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
also provide a range of modulation indices, but the operating frequency shifts
dramatically when the modulation capacitor is added. The larger shunt capacitor
values, like the 3900 pF capacitor, may actually be more ideal for production
because it offers the lowest impedance with a decent amount of modulation.
However the larger capacitor values could not produce a range in their modulation
indices, which is a problem for this experiment. That is why the 2700pF shunt
capacitor is chosen and the exact component combination selected for this
comparison are:
Table 4.3: Chosen Test Configurations for 120 ns and 70 ns pulse widths
Coil Driver
Modulation
Cap. (pF)
Shunt/Shunt Cap.
(pF)
Series Cap.
(pF)
Frequency
(MHz)
Modulation
(%)
CDLC01180207 680 2700 735 1.9525 6.67
CDLC01180207 1100 2700 735 1.9435 11.21
Pulse Width 120 ns
Coil Driver
Modulation
Cap. (pF)
Shunt/Shunt Cap.
(PF)
Series Cap.
(PF)
Frequency
(MHz)
Modulation
<% )
CDLC01180207 1300 2700 780 1.9570 4.76
CDLC01180207 2400 2700 780 1.9400 9.41
Pulse Width 70 ns
4.4 Testing of the Selected M odulation Param eters
Since the characteristics of the transmitter changes by using different
modulation capacitors, the selected capacitors will provide a controlled environment
for the experiment. By performing the BION sensitivity test, which basically
64
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
measures the output of the BION, the right modulation conditions for proper
operation can be determined. The inconsistencies between similarly tuned
transmitters should become clear as well. Before the sensitivity test is performed
however, the differences in modulation characteristics need to be evaluated further. It
is simple to observe when a BION stops responding but it is a much difficult task to
understand whether the BION failure was a result of the slope/shape of modulation,
the frequency distribution or possible loading effects.
The selected component values are ideal for this experiment because of their
relation of the modulation index and the modulation rise time. For the majority of
case when the modulation index is large, the rise time would be large as well, so it
would be difficult to determine if the modulation index or the rise time is the cause
of BION operation failure. In the case o f the values shown in table below, there is a
combination that has a high modulation index with high rise time and one
combination that has a high modulation with a low rise time. The individual effect of
high modulation index and high rise time is able to be solved through this
comparison.
To simplify m atters, the characteristics of the transmitter coil when it is
loaded will be shown with the unloaded values for each of the following
comparisons. This is because loading the coil would affect the characteristics of the
other modulation parameters and the differences between unloaded and loaded
values should be noted. The loading of the coil is achieved by touching the coil with
65
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
a hand as it would be the closest simulation of a coil driver placed on a patient’s
body.
Table 4.4: Normal Rise Time
Coil Driver
Pulse Width
(ns)
Max Voltage
(V)
Min Voltage
(V)
Rise Time
(us)
Rise Cycle
(cycles)
01180207
(6.67%)
120 240 210 3.04 6
01180207
(11.21%)
120 263 210 5.56 11
01180207
(4.76%)
70 143 130 2.62 5
01180207
(9.41%)
70 157 130 2.62 5
Table 4.5: Loaded Rise Time
Coil
Driver
Pulse Width
(ns)
Max Voltage
(V)
Min Voltage
(V)
Rise Time
(us)
Rise Cycle
(cycles)
Modulation
Change(%)
01180207
(6.67%)
120 211 198 2.56 5 3.18
01180207
(11.21%)
120 239 195 4.10 8 10.13
01180207
(4.76%)
70 130 120 2.10 4 3.18
01180207
(9.41%)
70 140 122 2.12 4 6.87
There may be a problem with the rise time of the 120ns pulse width, 11%
modulation index transmitter since it is 5.56 jis. The length of each cycle is 0.5 ps,
so the time it takes to complete a state is 8 cycles or 4.00 ps. Therefore the rise time
should be lower than 4.00 ps because anything greater would mean the encoding bit
will extend into the time of the following encoding bit. It is expected that if rise time
is a factor, the transmitter with this configuration would not able to control the
66
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
BION. The rise time of the 6.67% modulation index transmitter should work. When
the transmitter pulse width was set to 70%, another set of modulation levels were
formed. One modulation was at 4.76% and one was at 9.41%. In this case, both
modulation rise times are less than the 4.00 jus cutoff, so the BION should operate if
modulation index is not a determining factor.
To analyze the importance of the frequency changes due to modulation, the
spectral distribution at different modulation indices was measured. When a
transmitter carrier has modulation, the center frequency is not the only frequency
with power. During modulation, side bands are created and they consume some of
the total energy. The spectral distribution test will determine the amount of energy
lost in the side bands under different modulation indices. This is achieved by
connecting a coil probe to the spectrum analyzer to pick up the magnetic field
produced by the transmitter. The values recorded are shown in Table 4.6 below. An
interesting point to note is the frequency measured by the spectrum analyzer is
slightly off from the frequency measured by the oscilloscope. After some
investigation, it was found that the probe of the oscilloscope adds some capacitance
to the series capacitance and it shifts the whole frequency. This may appear to cause
a problem but the frequency is only off by 0.01 MHz or 10 KHz so it does not have a
noticeable effect. The data from the power spectrum does show that the peak power
occurs at the center frequency and that there is a side band that occurs +/-30 KHz
away from the center frequency. This is a clear indication that is modulation is
having an effect of the carrier.
67
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 4.6: Power of Center Frequency and Sideband
Coil
Driver
Pulse
Width (ns)
Center Freq.
(MHz)
Center
Magni. (d.B.)
- fl (MHz)/
Mag. (d.B.)
+fl (MHz)/
Mag. (d.B)
-£2 (MHz) /
Mag. (d.B.)
+£2 (MHz) /
Mag. (d.B.)
01180207
(6.67%)
120 1.9435 -24.3050 1.9135/
-41.1208
1.9742/
-43.2610
1.85350/
-44.8911
2.0342 /
-48.0562
01180207
(11.21%)
120 1.9525 -22.0522 1.9225 /
-43.9213
1.9832/
-45.9841
1.8617/
-46.839
2.0440 /
-52.2345
01180207
(4.76%)
70 1.940 -23.3620 1.910/
-38.5962
1.971 /
-42.3860
1.849/
-41.7088
2.031 /
-44.2875
01180207
(9.41%)
70 1.957 -26.5461 1.927/
-36.2335
1.987/
-36.9423
1.865 /
-40.1835
2.049 /
-39.9586
Table 4.7: Loaded Power of Center Frequency and Sideband
Coil
Driver
Pulse
Width (ns)
Center Freq.
(MHz)
Center
Magni. (d.B.)
- fl (MHz)/
Mag. (d.B.)
4-fl (MHz)/
Mag. (d.B)
-fl (MHz) /
Mag. (d.B.)
+f2 (MHz) /
Mag. (d.B.)
01180207
(6.67 ->
3.18%)
120 1.9390 -24.0949 1.9090/
-41.2121
1.9690/
-42.7793
1.8483 /
-45.1039
2.0298 /
-47.9063
01180207
(11.21->
10.13%)
120 1.9488 -22.0751 1.9180/
-43.8917
1.9788/
-45.7126
1.8580/
-47.0422
2.0395 /
-52.1568
01180207
(4.76 ->
3.18%)
70 1.930 -22.225 1.8997/
-37.190
1.9605/
-40.054
1.8390/
-39.599
2.0205 /
-42.762
01180207
(9.41 ->
6.87%)
70 1.947 -27.170 1.9174/
-36.070
1.9770/
-36.790
1.8561 /
-40.110
2.0380/
-40.305
With the characteristics of the modulation evaluated, the BION sensitivity
test can be performed to observe under what conditions does the BION stops
responding. The test is accomplished by using the wet well apparatus. Basically the
68
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
wet well Is a vial with a modified lid that contains a small cylinder and two
electrodes. The BION is inserted into this small cylinder and is held in place by a
rubber washer. When the BION is inserted, it creates a space in the cylinder that is
separate from the remaining vial space. The two electrodes are placed so that each
electrode has contact with only one space. By filling these two spaces with saline to
act as a medium between the BION electrodes and the wet well electrodes, the
potential difference across the two spaces can be measured using a multi-meter or
oscilloscope. Normally there would be no potential difference, but when the BION is
activated and commanded by the transmitter, stimulation pulse current will be
generated from the BION which will cause a voltage change. By observing the
voltage level In the wet well, one can determine if the BION is properly working or
not. The wet well is shown in Figure 4.6.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The transmitter circuit is designed to amplitude modulate the carrier
frequency to transmit data, but it is actually a device called the Personal Trainer (PT)
that sends the data signals to the transmitter circuit for modulation. The Personal
Trainer can be programmed by a PC to send specific commands to the receiving
BION. A specific BION output current magnitude and duration can be programmed
to the BION so confirmation of the output values can be used to make sure that the
BION has indeed received the correct code. For this experiment, the BION is
programmed to produce 14 mA stimulus currents that are 100 us pulse-widths at a
repetition rate of 4 Hz. The recharge current is set to 100 uA.
The sensitivity of the transmitter is determined by the response of the BION
at different distances. The first step is to place the transmitter at a distance 0 cm
away from the BION, which Is positioned in the middle of the coil. The field strength
received by the BION from the transmitter coil is at its strongest at this distance.
Therefore the BION will have enough power to output a current that It is
programmed to produce. The first value that is recorded Is this maximum pulse
output generated by the BION.
The next step is to gradually increase the distance between the transmitter
and BION. As the distance between the transmitter and BION is increased, one of
two things is expected to occur. In one case, the BION may still receive the correct
data modulation, but it may not receive enough power to produce the programmed
current. So it will produce a current that is lower in amplitude, but will still have the
correct current duration. To indicate when the BION is not receiving enough power,
70
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the field strength at which the BION stops producing the maximum pulse is
measured. The field strength is recorded using the 1 cm diameter coil probe.
The second case that will cause the BION to stop responding is when the
modulation in field strength can not be detected and/or there is not enough power for
the BION to respond. This generally occurs when the distance between the BION
and transmitter is large. Thus the final value that is measured is the field strength at
which the BION stops responding or when it only produces 5% of the maximum
pulse. This value tells when the BION stopped functioning because of data
transmission a.k.a. modulation problems. The sensitivity test will be performed for
all the capacitor combinations when the transmitter Is unloaded and loaded.
Table 4.8: Normal Sensitivity
Coil
Driver
Pulse
Width
BION Pulse
Strength @ 0 cm
(100% Operation)
Minimum Coil Field
Strength for 100%
Operation
Coil Field Strength
@ Dead point (5%)
(ns) (V) (mV) (mV)
01180207
(6.67%)
120 5.70 66.2 49.8
01180207
(11.21%)
120 No Pulse N/A N/A
01180207
(4.76%)
70 6.94 42.4 31.2
01180207
(9.41%)
70 7.10 39.2 32.8
71
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 4.9: Loaded Sensitivity
Coil
Driver
Pulse
Width
BION Pulse
Strength @ 0 cm
(100 %
Operation)
Minimum Coil Field
Strength for 100%
Operation
Coil Field Strength
@ Dead point (5%)
(ns) (V) (mV) (mV)
01180207
(6.67 ->
3.18%)
120 5.70 96.2 49.8
01180207
(1 1.21->
10.13%)
120 6.62 94.6 46.6
01180207
(4.76 ->
3.18%)
70 6.94 54.4 32.4
01180207
(9.41->
6.87%)
70 7.06 52.8 32.8
4.5 M odulation Test Analysis
The results of the sensitivity test show that there is definitely a difference in
BION operation due to modulation characteristics. The BION responded to all of the
transmitters except for the 120 ns pulse width, 11% modulation index transmitter.
Since the BION did respond to the 70 ns pulse width, 9.4 % modulation index
transmitter, this suggests that a high modulation index does not cause encoding
problems. It is interesting to note that when the coil was loaded by a hand, the BION
responded to all of the transmitters including the one that failed unloaded. This
indicates that loading the coil does have an effect on the operation of the transmitter.
Therefore the other remaining factors must be looked upon to determine what
changed in the modulation.
72
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Having the transmitter frequency tuned to the natural frequency of the BION
is important because an improperly tuned frequency will not provide enough power
to the BION through inductive coupling. It was expected that when the modulation
capacitance is added, the center transmitter frequency changes and possibly becomes
a source of the problem. By using the power spectrum analyzer, it was determined
the majority of the power was at the center frequency so there is no problem with the
BION not getting enough power. It is shown that there is change in frequency, but it
is very slight, so the BION is able to operate. In fact when the transmitter coils were
loaded, the transmitter frequency shifted even more, but the BION still continued to
operate. Therefore the change in frequency is not the problem as there appears to be
more leeway with frequency shifts.
Thus the only remaining factor involved with modulation, that could cause
problems with data transmission, is the modulation rise time. In the case of the 120
1 1s, 11% modulation index transmitter, the modulation rise time was much longer
than the other values at 5.56 ps. It was previously stated that a rise time longer than
4.00 ps would most likely cause problems because a data bit would take too much
time to switch logic positions and that would affect the following bit value. An
interesting point that strengthens the argument for the modulation rise time being the
main culprit is that when the coil was loaded, the BION fired. Looking at the
experimental data, when the coil was loaded the rise time went down to 4.10 us,
which is a acceptable value in theory. Some of the other modulation characteristics
were changed, but it was shown that the other values do not have much of an effect
73
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
on the BION operation. Thus the experiment shows that the most important factor of
modulation that determines whether a BION operates or not is the rise time of the
modulation.
74
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
C hapter 5
5.1 Coil Shielding
With the proper selection of the capacitor values for the Class E amplifier
circuit, the transmitter will operating quite efficiently. Further optimization on the
circuit may be unnecessary but the emission of the magnetic field by the coil could
still be improved. When the transmitter produces a magnetic field to transfer power
to a BION by inductive coupling, an electric field is emitted as well. The electric
field is unwanted because it can interfere with the operation of the BION. The
problem is caused by the effect o f the electric field’s capacitance coupling between
the transmitter coil and the BION. The capacitance caused by the coupling adds to
the total capacitance of the circuit and this will cause the resonant frequency to
distort from the ideal 2 MHz.
The capacitive effect can be minimized by placing a shield made out of any
conductive foil (such as aluminum) around the coil of the transmitter which is
connected to ground. By adding the grounded layer between the two coils, a potential
difference cannot be created. The shield should cover the majority of the transmitter
coil for the effective shielding, but there should be a break in the shield to prevent it
from forming a full loop around the coil. A shield that forms a full loop will have a
problem of eddy currents which can generate its own magnetic field and cause
problems [Moreland 1999]. This experiment is performed to see how well the coil
driver transmitter produces a magnetic field while it is wrapped in aluminum. It has
75
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
been demonstrated that the electric field will be reduced considerably while the
magnetic field will have little or no changes at all [Lu 2000].
5.2 Shielding Performance Test
An experiment similar to the one performed for the modulation experiment
will be used to analyze the performance of the shielded transmitter coil. The
operation of the BION is observed as the distance between the transmitter and BION
is increased, but in this case the difference in the sensitivity of the BION responding
to an unshielded coil and the sensitivity of the BION responding to a shielded coil is
analyzed. The Personal Trainer is programmed to produce 14mA currents of 100 us
pulse widths one again. Along with the electric field probe, thel cm diameter probe
coil is used to measure the magnetic field strength received at the BION.
To make the results of the experiment more substantial, the testing will be
performed on two different types of shielding materials. The two materials are
common household aluminum foil and copper tape made for shielding. The
aluminum foil was initially cut to wraparound the transmitter coil. To reduce the
amount of eddy currents, the aluminum was cut in spaced intervals. Essentially the
shield could be described as a connection of multiple “fingers” wrapping around the
coil. The copper tape is easier to apply as a shield since it was created as a shield.
The copper tape was taped all around the transmitter coil with even intervals and
with a thin line of tape, all the individual slices were connected.
76
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
It was observed that adding the shield on to the transmitter coil altered the
characteristic of the coil. The values are listed in Table 5.1 below. The frequency of
the carrier frequency was slightly higher than the original value. An explanation for
this change in frequency is the shield acts as a capacitor that is connected in parallel
with the coil. A capacitor in parallel with the coil is equivalent to a capacitor in series
with the series resonance capacitors. To correct the frequency shift, the steady state
and series capacitance were changed to bring the carrier frequency back to its
original value. However the current flowing through the coil needs to be consistent
with the original value, since the magnetic field strength is dependent on the current
through the coil. The corrected capacitor values are shown below.
Table 5.1: Unshielded and Shielded Coil Characteristics
Unshielded Aluminum Aluminum Copper
Foil Shield Foil Shield Tape Shield
w/o tuning w/ tuning w/ tuning
Shunt Cap. (pF) 2700 2700 3300 3300
Modulation Cap. (pF) 620 620 820 820
Series Cap. (pF) 735 735 735 702
Frequency (MHz) 1.99 2.05 2.007 2.01
Coil Voltage (V) 260 260 232 254
Coil Current (A) 2.24 2.33 2.24 2.24
Drive Pulse (ns) 96 96 96 96
H-field in coil center
102 98 110 114
@ 0 cm (mV)
The magnetic field does not decrease due to the shield so theoretically the
transmitter should be able to continue with normal operation. In fact, the magnetic
77
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
field strength increases with the shield. The electric field still needs to be measured
in order to determine if the shield is a success or not. An oscilloscope probe will be
used to measure the electric field produced by the transmitter coil. By positioning the
ground clip next to the measurement electrode, the oscilloscope will measure no
potential difference between the two points, ha this configuration, the probe will
detect the potential difference generated as it enters the electric field. The electric
field measurements are meant only to be used as a value for comparison between the
shielded and unshielded transmitter coil, just like how the magnetic field measured
by the 1cm diameter coil probe was used for comparison purposes only.
Table 5.2: Shielded and Unshielded E-field Measurements
Distance (cm) Unshielded E-
field (mV)
Aluminum Foil
Shield E-Field
(mV)
Copper Tape
Shield E-field
(mV)
0 282 104 160
5 160 62 113
10 142 37.6 82
15 114 20.2 44
20 60 14.2 31
The results of the E-field measurements prove that the shielding is effective,
For both the aluminum and copper shields, the electric field is dramatically reduced
for all distances away from the coil. It is interesting to note that the reduction of the
electric field by the aluminum shield is far greater than the reduction by the copper
tape. This suggests that aluminum is a better material to use as a shield, but the
78
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
sensitivity of the BION with both the aluminum and copper shielded transmitter
must be tested before the final determination is made.
The BION sensitivity test is similar to the one performed for the modulation
experiment. To summarize, the BION is inserted in the wet well, which is connected
to an oscilloscope for output measurements. Place the wet well and the transmitter on
the height adjustable test stand. For consistency, the Personal Trainer is programmed
to produce 14mA currents of 100 us pulse widths. Adjust the height of the test stand
to record the operation of the BION at various distances away from the transmitter.
The BION should perform the best when the coil driver is placed close to it. Record
the maximum output of the BION and the required field strength to power the BION
for 100% operation. Raise the test stand until the BION stops operating at 100 %
output and record the field strength. At this point, the BION output will linearly
decrease as the distance is increased. Raise the test stand even more until the BION
stops responding as this will be the dead point. For specific details on the procedures,
refer to the previous chapter.
Table 5.3: Unshielded BION Sensitivity
BION
Serial
Number
[Address]
Maximum
Field
Strength ■
for 100%
Output
(mV)
BION 100%
Output
Voltage (V)
[Output
Current
(mA)]
Minimum
Field
Strength
for 100%
Output
(mV)
Field
Strength
at Dead
point
(mV)
Distance
between
BION &
Coil at Dead
point (cm)
K620 [20] 102 4.38 [18.25] 50.0 44.4 6
K563 [21] 102 4.56 [19.00] 67.6 67.6 3.75
K619 [17] 102 4.20 [17.50] 52.0 43.2 6
79
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 5.4: Aluminum Foil BION Sensitivity
BION
Serial
Number
[Address]
Maximum
Field
Strength
for 100%
Output
(mV)
BION 100%
Output
Voltage (V)
[Output
Current
(mA)]
Minimum
Field
Strength
for 100%
Output
(mV)
Field
Strength
at Dead
point
(mV)
Distance
between
BION &
Coil at Dead
point (cm)
K620 [20] 110 4.36 [18.17] 58.0 44.0 6
K563 [21] 110 4.60 [19.17] 70.0 70.0 3.5
K619 [17] 110 4.20 [17.50] 57.6 57.6 5
Table 5.5: Copper Tape BION Sensitivity
BION
Serial
Number
[Address]
Maximum
Field
Strength
for 100%
Output
(mV)
BION 100%
Output
Voltage (V)
[Output
Current
(mA)]
Minimum
Field
Strength
for 100%
Output
(mV)
Field
Strength
at Dead
point
(mV)
Distance
between
BION &
Coil at Dead
point (cm)
K620 [20] 114 4.36 [18.17] 75.4 51.0 5.5
K563 [21] 114 4.60 [19.17] 81.2 58.4 5
K619 [17] 114 4.20 [17.50] 74.2 47.0 6
5.3 Shielding Test Conclusion
Looking at the data of the shielded and unshielded coils, it is observed that
the BION responses are similar with some differences. All of the transmitters were
able to activate the BION at maximum output and they all had similar ranges In
terms of when the BION stopped responding. However for most of the BIONs
operated by the unshielded transmitter, they operated at 100% output much longer
before the output started to decrease compared to the BIONs powered by the
80
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
shielded transmitter coil. In other words, both BION sets had a similar range in field
strengths, but the BION operating characteristics were different. The discrepancies in
BION operation may be due to the difference in BIONs since some BIONs are more
sensitive than others. However the overall performance of the BIONs are similar
between the shielded and unshielded coils so the difference should not be BION
related.
As mentioned earlier, the magnetic field strength was slightly increased once
the shielding was placed. However, it was expected that the field strength would be
the same as the unshielded coil since the current flowing into the coil was adjusted to
be the same. The only logical explanation for the difference would be that the
magnetic field shape is slightly altered when the shield is in place. By looking at the
data, the shielded transmitter coils do generate a little bit less magnetic field strength
than the unshielded transmitter coil at the same distance away. Tlie shield seems to
strengthen the magnetic field strength at close proximity, but weakens as the distance
increases. The reason why the range of both the unshielded and shielded transmitter
are similar, is because the digital dead point occurs due to effects of modulation.
The use of a shield on the transmitter coil for optimization is questionable.
While the electric field produced by the transmitter was reduced considerably by
both copper and aluminum shields, the shields did not improve the functionality of
the BIONs. In fact for a certain range, the BIONs responded with less current output
than when it was powered by an unshielded coll. However, even if the performance
were to decrease, the reduction of electric field emissions may be more beneficial
81
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
when it conies to EMI regulation. If a shield were to be used, the aluminum shield
would be the better choice as the maximum output operating range of the copper
shield is more limited than the alumiiumi shield and the electric field was reduced
more by the aluminum foil.
82
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
C hapter 6
6.1 Electromagnetic Interference Regulation
RF transmission is an useful technique for power distribution, but the
emission of RF transmitters can cause potential problems. For example, since there
are many devices that require the operation of RF, interference among other devices
operating with RF of similar frequencies is a possibility. Another problem is that RF
emissions with high energy content may produce biological damage. However, the
energy of RF electromagnetic field is not great enough to cause ionization or thermal
damage to biological material [Cleveland 1999]. Due to the potential problems of
unregulated RF emissions, the government has developed regulations regarding the
amount of power an RF transmitter could produce for certain frequencies. Therefore
the RF emission of the current BION transmitter must be evaluated in order to
determine if it is within acceptable range.
For the evaluation, the acceptable value of RF emission by the BION
transmitter Is based on the standards from the European Telecommunications
Standards Institute. The experiment procedure is according to the specifications
listed in the ETSI document, EN 300 330-1. Since the BION RF transmitter operates
at a carrier frequency of 2 MHz, the transmitter is categorized as Product Class 2
according to Section 7.1.4 of the ETSI electromagnetic compatibility test document
(EN 300 330-1 V I.3.1). Product Class 2 Is defined as an inductive loop coil
transmitter, like Product Class 1, but allows field customization of the loop antenna.
Customization is only allowed according to the manufacturers antenna design rules
83
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
published in the equipment manual. The restrictions that apply to this Product Class
are 9 kHz to 30 MHz frequency range, Loop antenna area < 30 mA 2 and the length of
any antenna loop element shall be A/4 < {15!f where f is in MHz) or < 30 m
whichever is shorter.
ETSI describes a standard setup for outdoor testing of transmitters less than
30 MHz, which is the test that will be performed on the BION transmitter. Hie test
from section, ETSI EN 300 330-1 Vl.3.1 (2001-04), is as follows:
A .l Test sites and general arrangements for measurements
involving the use of radiated fields
A .1.1 Outdoor test site
The outdoor test site shall be on a reasonably level surface or
ground. For measurements below 30MHz no artificial ground
plane shall be used. A non-conducting support, capable of
rotation through 360° in the horizontal plane, shall be used to
support the test sample in its standard position, at 1 m above
ground. The test site shall be large enough to allow the
erection of a measuring or transmitting antenna at a distance of
10 m or optionally 30 m.
A .1.1.1 Standard Position
— For equipment with a rigid antenna, the antenna shall be
vertical.
A .1.2 Test Antenna
A .1.2.1 Below 30 MFIz
A calibrated loop antenna shall be used to detect the field
strength from the test sample. The antenna shall be supported
in the vertical plane and rotated about the vertical axis. The
lowest point of the loop shall be 1 m above ground level.
84
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The procedures listed in the ETSI standards are very thorough and it also
includes procedures for extreme test environments, in terms of temperature, humidity
and power source. However the evaluation test will only be performed under normal
test conditions because of the difficulty setting up the experiment under the extreme
conditions. The normal condition temperature range is +15°C to +35°C and the
normal relative humidity range is 20 % to 75 %. The normal condition test voltage is
12V since it is the nominal voltage for which the equipment was designed.
According to the ETSI, the frequency of the test power source corresponding to AC
should be between 49 Hz and 51 Hz, but the test will be performed in the US, so it
will be 60 Hz.
6.2 RF Emission Testing
It is imperative that the RE emission testing follow the directions listed
above carefully. The measurements could easily be corrupted by outside sources of
RF, so the experiment must be performed in a large open field where there is nothing
that would cause interference. In this case, a soccer field was used since it provided
enough empty space from buildings. The top view of the soccer field and location of
the test equipment is shown In Figure 6.1 below.
85
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fence
Spectrum
in™ Analyzer
-385.4 m
125 ft
Transmitter
22.80 m
Building
Figure 6.1: Diagram o f Test Setup
The field is large enough to provide 125 ft of empty space on one axis of the
field and 75 ft on the other axis. There are no power lines running next to the field as
well so there should be no sources of interference to the. To properly setup the
experiment, the transmitter and the receiving antenna are placed 10 m away from
each other. The test stands holding the receiving antenna and the transmitter must be
assembled so that they are both 1 m above the ground. The transmitter coil also
needs to be positioned in parallel with the antenna for optimal reception so the coil
was taped to stand perpendicularly to the ground. The antenna has a threaded hole at
the bottom of its stand so a screw was used to secure the antenna to the test stand. A
diagram of the test setup is shown in Figure 6.2.
86
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Receiving
Antenna
Transmitter
Jk
1 m T
▼
r
Uround levef
< ►
10 (optionally 30) m
Figure 6.2: Test Setup
The antenna is a specialized antenna built by a company called A.R.A. for the
purpose o f performing RF emission tests. The specifications of the antenna areas
follows. Frequency Range: 10 kHz - 30 MHz; Impedance: 50 O; Input Power: 5
Watts; Antenna Factor: -9.9 dB/m @ 2 MHz; Loop Size: 24” square. The emissions
that are detected by the antenna is processed by connecting a spectrum analyzer.
Since the carrier frequency of the BION transmitter is 2 MHz, the power spectrum
around the frequencies of 2 MHz is the point of interest. The power amplitude
display will need to be adjusted by changing the offset and division accordingly once
the transmitter is turned on using the SCALE REF menu. The orientation of the
transmitter and antenna were checked and adjusted so that the maximum power can
be received.
Once the test setup was completed, the emission characteristics of the BION
transmitter were measured and recorded. The values o f importance are the power
87
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
amplitude distribution among the center frequency and the side bands generated due
to the modulation of the center carrier. The measurements were repeated for several
other situations that the transmitter may face to determine if there are any changes in
emission. The first situation is the effects of loading the transmitter by having a
person hold or touch the coil driver to simulate a patient carrying the coil driver.
Another situation that may cause differences in the output is when the Personal
Trainer operating the transmitter undergoes several program changes. While this is
assumed to have little or no changes, the codes for the different program will
modulate the carrier frequency slightly different so it is a possibility. The final
situation that will be tested is the presence of the aluminum shield. In the previous
chapter, the aluminum shield reduced the electric field generated by the coil driver so
measuring the actual power reduction using a spectrum analyzer to see if the
reduction is necessary will be interesting.
6.3 Emission Data
The BION transmitter that is used for the testing is the standard large
elliptical, spiral coil. The dimensions o f the inner diameter are 9 x 4.5 inches. It has 5
turns and the current that is running through the coil is 2.3 A. The maximum voltage
at the terminals of the coil is 260 V and modulates to 240 V. This configuration is
the one being used for clinical trials so the test transmitter will provide a glimpse of
what is being used in real applications.
88
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 6.1: Normal Operation
Time Temperature Humidity
6:25 PM 79 F 46%
Coil
Driver
Number
Center
Frequency
(MHz)
Center
Magnitude
(dBm)
+Side
band
Freq.
(MHz)
+Side
band
Magn.
(dBm)
-Side
band
Freq.
(MHz)
-Side
band
Magn.
(dBm)
CDLC01
180207
2.027 -71 2.058 -100 1.995 -100
Table 6.2: Contact with Skin
Time Temperature Humidity
6:35 PM 77 F 48%
Coil
Driver
Number
Center
Frequency
(MHz)
Center
Magnitude
(dBm)
+Side
band
Freq.
(MHz)
+Side
band
Magn.
(dBm)
-Side
band
Freq.
(MHz)
-Side
band
Magn.
(dBm)
CDLC01
180207
2.018 -70 2.050 -89 1.988 -87
Table 6.3: Program Changes
Time Temperature Humidity
6:28 PM 79 F 47%
Coil
Driver
Number
Center
Frequency
(MHz)
Center
Magnitude
(dBm)
+Side
band
Freq.
(MHz)
+Side
band
Magn.
(dBm)
-Side
band
Freq.
(MHz)
-Side
band
Magn.
(dBm)
CDLC01
180207
2.027 -71 2.058 -100 1.995 -100
89
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 6.4: Wrapped in Aluminum Shield
Time Temperature Humidity
6:30 PM 79 F 47%
Coil
Driver
Number
Center
Frequency
(MHz)
Center
Magnitude
(dBm)
+SIde
band
Freq.
(MHz)
+Side
band
Magn.
(dB)
-Side
band
Freq.
(MHz)
-Side
band
Magn.
(dBm)
CDLC01
180207
2.104 -72 2.133 -100 2.074 -100
Table 6.5: Tuned Transmitter Wrapped in Aluminum Shield
Time Temperature Humidity
6:40 PM 78 F 47%
Coil
Driver
Number
Center
Frequency
(MHz)
Center
Magnitude
(dBm)
+Side
band
Freq.
(MHz)
+Side
band
Magn.
(dBm)
-Side
band
Freq.
(MHz)
-Side
band
Magn.
(dBm)
CDLCOl
180201
2.012 -73 2.040 -100 1.984 -100
6.4 Analysis of Results
The measurements for the experiment started in the evening, just as the
temperature started to fall so all tests comply within the allowed range of
temperature and humidity. All of the electrical power for the devices came from a
standard US outlet running 110V AC at 60 Hz. The transmitter that was used for the
testing was a standard large coil with the serial number CDLC01180207. The
spectrum analyzer was set to read at 5 dBm/div with the reference level set at -40
90
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
dBm. Measuring the large flat coil under normal operation.;, it was observed that the
power spectrum magnitude fluctuated slightly. Therefore the magnitude values have
an uncertainty of +/- 2 dBm, while the frequency uncertainty is +/- 0.002 MHz. The
normal operation data will be used as a baseline to compare the data from the
different tests.
The contact with skin test showed a difference from a normal operation
BION transmitter. The magnitude of the center frequency slightly dropped but the
side band amplitude rose considerably. The frequency o f the carrier dropped as well.
This indicates that the modulation of the coil driver has increased. An explanation
for this occurrence is that the skin acted as an additional capacitor in series with the
inductor coil. If the capacitance is increased, the frequency becomes lower than 2
MHz similar to when the skin touched the coil. In order to “tune” the coll driver back
to 2 MHz, the capacitor values in series with the inductor coil need to be adjusted to
compensate for the skin capacitance. It was also observed that by lowering the
frequency, the free running voltage o f the coil drops slightly. That will help the
modulation capacitance have a greater effect on the steady state capacitance and
cause a higher modulation index. This explains why the center frequency magnitude
drops, while the side band magnitudes increases.
There were no noticeable differences in the different program test mainly
because they all run on the same carrier frequency. The different programs send
different codes, so the carrier frequency will modulate differently depending on the
code. Since the side bands are generated by the modulation wave, it would seem that
91
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the different programs would alter the side bands. However the codes are similar in
that the duration of the modulation can only be 8 cycles or 16 cycles long. Therefore
even though the modulation occurs at different times, the modulation frequency stays
the same so there are no differences between different programs.
The power output of the aluminum shielded transmitter was slightly lower
than the unshielded transmitter. However there is a considerable amount of
difference in frequency between them. The frequency of the shielded transmitter is
higher, suggesting that the inductance or the capacitance in series with the inductor
has changed in the circuit. From this test alone the cause of the lower power output
cannot be determined whether it is due to the shield or due to the frequency shift.
The shift in frequency was expected, so another transmitter was that was tuned to 2
MHz after shielding was tested as well. Since the tuned shielded coil did have a
lower power output than the de-tuned shielded coil, the shield has an effect on the
power output. Although the shield lowers the power, the importance of the shield
still needs to be determined. If the unshielded BION transmitter passed the regulation
limit, then there is no need for the shield. Therefore the recorded measurements were
compared to the standards.
The data is measured in dBm so the values need to be converted to dBfiA in
order to compare with the values o f ETSI. The following equations provide the
conversion from dBm to dbV.
92
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
P = 10 * logic (p /p o ) po = 1 mW (6.1)
U = 20 * log1 0 (u/uo) uo= 1 fiV (6.2)
p = u2 / Zc (6.3)
For the standard operating coil driver, the output was -70 dBm. Using Equation 6.1,
the power in mW is IE-10 mW. Using Equation 6.3 with the impedance of the
receiver being 50 Q, the power is converted to pV. Then using Equation 6.2, the
conversion to dBpV can be solved. In this case, -70dBm is 36.897 dBpV. From
dBfiV, the value needs to be converted to dBV using Equation 6.4:
With the measurement in dBV, the antenna calibration factor can be added to the
value. The equation for this conversion is:
Since the Antenna calibration value is -9.9 dB/m for a 2MHz frequency, that value
will be added to the -83.0103 dBV value, which gives a measurement of -92.9103
dB A/m. Converting to dBpA/m we get 27.0897 dBpA/m. The converted values are
listed in Table 6.6.
U = 20 * logio (u / u0 ) (6.4)
H(dB A/m) = AFH (dBAV-1 m-1) + VO (dBV) (6.5)
93
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 6.6: Data converted to dBuA
Trans
mitter
Condition
Center
Freq.
(MHz)
Center
Magn.
(dBm)
Center
Magn.
(dBuA
/m)
+Side
band
Magn.
(dBm)
+Side
band
Magn.
(dBuA
/m)
-Side
band
Magn.
(dBm)
-Side
band
Magn.
(dBuA
Im)
Normal 2.027 -71 26.089 -100 -2.910 -100 -2.910
Skin
Contact
2.018 -70 27.089 -89 8.089 -87 10.089
Diff.
Programs
2.027 -71 26.089 -100 -2.910 -100 -2.910
Alum.
Shield
2.104 -72 25.089 -100 -2.910 -100 -2.910
Tuned
Alum. Sh.
2.012 -73 24.089 -100 -2.910 -100 -2.910
According to the ETSIEN 300 330-1 V I.3.1 (2001-04) 40 Annex B, the
emission of 26.0897 dBjxA/m by the normal condition transmitter is higher than the
transmitter carrier limits at 2 MHz. As shown in the carrier limits graph shown in
Figure 6.3, the allowed limit is 20 dBpA/m at 2 MHz. Even the aluminum shielded
transmitter coil produces 24.0897 dBpA/m. It Is not by much, but the emissions is
higher than the allowed amount according to ETSI.
94
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
A nnex B {normative]:
T r a n s f e r carrier limits
Distance of 10 m
iilh
70 ■
J ....— • " !
r— .. ........ ....
■yp
60
f ,
| 40-
£ 30-
i
X
20-
10 \
0
0.0
-mi
!» * -“
% ,
f
k -
*-
"1
kr
01 0.01 0.1 1 10 1C
Frequency (MHz) Large Coil
* 26.0897 dBuA/m 2 MHz
Figure 6.3: ETSI Transmitter Carrier Limits
Based on this experiment following the ETSI test format, the standard large
flat coil configuration transmits -71 dBm or 26.0897 dBpA/m to an antenna 10
meters away. The skin contact lowered the output as well as the frequency, while the
aluminum shield increased the output and the frequency. However, the tests show
that the output levels of the BION transmitter remain roughly around the same
magnitude. W ien the transmitter output values are compared to the ETSI
95
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
specifications, to see if it corresponds, they were found to be slightly higher than the
transmitter carrier limits based on Annex B. The easiest way to combat this problem
is to reduce the if the voltage and current through the coil of the transmitter. It would
limit the range of the transmitter, but the transmission should fall under the limit
curve.
The ETSI procedures used in this experiment provides an indication of how
close the H-field emission of BION transmitter is to the regulated amount.
Unfortunately, the ETSI procedures for this experiment did not include steps to
determine whether the E-field emission of the BION transmitter is within regulation.
From the coil shielding experiments in Chapter 5, it was determined that the E-field
was reduced with a shield. However, the E-field emission comparison was made at
close distances away from the coil so the E-field characteristic may be entirely
different when it is measured from greater distances. Therefore along with the H-
field measurements, the E-field emissions will need to be measured soon to
determine compliance.
96
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
B ibliography
Cantrell B. 1998. How to Design a Class-E Transmitter for Your LowFER Beacon.
Retrieved May 25, 2003, from
http://www.greerguitars.com/k3pgp/Notebook/Wd5cvg/Classetx/ciassetx.litm
Cleveland RE, Ulcek JL. 1999. Questions and Answers about Biological Effects and
Potential Hazards of RE Electromagnetic Fields. Washington, DC: FCC Office of
Engineering & Technology. Bulletin. 50 p.
DeMaw D. 1996. Ferromagnetic Core Design & Application Handbook. New Jersey:
Prentice Hall. 256 p.
Donaldson NDN, Perkins TA. 1983. Analysis of resonant coupled coils in the design
of radio frequency transcutaneous links. Medical & Biological Engineering &
Computing 22(9):612-627
Forster R. 2000. Manchester encoding: opposing definitions resolved. Engineering
and Science Engineering Journal 9(6): 278-280
Ko WH, Liang SP, Fung CDF. 1977. Design of radio-frequency powered coils for
implant instruments. Medical & Biological Engineering & Computing 15(11):634-
640
Lee M. 1987. Broadband Longwave Receiving Loop. The Lowdown 7(4):23-27
Loeb GE, Richmond FJR. 2000. BION Implants for Therapeutic and Functional
Electrical Stimulation. In: Chapin JK, Moxon KA, Gaal G, editors. Neural
Prostheses for Restoration of Sensor and Motor Function. Boca Raton: CRC. p 75-
101 .
Loeb GE, and Richmond FJR. 1999. FES or TES: How to start an industry?
Proceedings of the 4th Annual Conference o f the International Functional Electrical
Stimulation Society; 1999: p 169-172.
Loeb GE, Peck RA, Moore WH, Hood K. 2001. BION System for Distributed
Neural Prosthetic Interfaces. Journal of Medical Engineering and Physics 23(6);9-32
Lu J, Wong F. 2000. High Frequency Coaxial Transformer with Faraday Shield.
Digests of the 2000 Intermag Conference; 2000: p FE-09.
97
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Moreland CW. 1999. BFO Theory. Retrieved May 25, 2003, from
http://www.thunting.com/cgi-bin/geotech/pages/common/
Ramanujan S. 1913. Modular equations and approximations to 7 t. Quarterly Journal
of Pure and Applied Mathematics 45:350-372
Sokal NO. 2001. Class-E Power Amplifiers. QEX 2G(1):9~2G
Sullivan CR. 1999. Optimal Choice for Number of Strands in a Litz-wire
Transformer Winding. IEEE Transactions on Power Electronics 14(3):283-292
Thomas RE, Rosa AJ. 1998. The Analysis and Design of Linear Circuits. New
Jersey: Prentice Hall. 966 p.
Tippler PA. 1991. Physics for Scientists and Engineers. New York: Worth. 1425 p.
Troyk PR, Schwan MAK. 1992. Closed-Loop Class E Transcutaneous Power and
Data Link for Microlmplants. IEEE Transactions on Biomedical Eng. 39(6):589-599
Wheeler HA. 1928. Simple Inductance Formulas for Radio Coils. Proceedings o f the
I.R.E. 15(10): 1398-1400
98
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Effects of prenatal cocaine exposure in quantitative sleep measures in infants
PDF
Head injury biomechanics: Quantification of head injury measures in rear-end motor vehicle collisions
PDF
Characteristics and properties of modified gelatin cross-linked with saline for tissue engineering applications
PDF
Contact pressures in the distal radioulnar joint as a function of radial malunion
PDF
A model of upper airway dynamics in obstructive sleep apnea syndrome
PDF
Estimation of optical fluorescent impulse response kernel in combined time and wavelength spaces
PDF
Comparisons of deconvolution algorithms in pharmacokinetic analysis
PDF
A model of cardiorespiratory autoregulation in obstructive sleep apnea
PDF
Cellular kinetic models of the antiviral agent (R)-9-(2-phosphonylmethoxypropyl)adenine (PMPA)
PDF
Destructive and non-destructive approaches for quantifying the effects of a collagen cross-linking reagent on the fatigue resistance of human intervertebral disc
PDF
Fourier analysis of the Einthoven resultant EKG vector
PDF
Biological materials investigation by atomic force microscope (AFM)
PDF
Design of a portable infrared spectrometer: application to the noninvasive measurement of glucose
PDF
A multimodal screen reader for the visually impaired
PDF
Development of ceramic-to-metal package for BION microstimulator
PDF
A finite element model of the forefoot region of ankle foot orthoses fabricated with advanced composite materials
PDF
An open ear canal sound delivery system
PDF
Bayesian estimation using Markov chain Monte Carlo methods in pharmacokinetic system analysis
PDF
Empirical evaluation of neuromusculoskeletal models for functional electrical stimulation of the lower limb
PDF
Finite element analysis of the effects of stem geometry, surface finish and cement viscoelasticity on debonding and subsidence of total hip prosthesis
Asset Metadata
Creator
Murakata, Tomonori Thomas
(author)
Core Title
Evaluation of R.F. transmitters for optimized operation of muscle stimulating implants
School
Graduate School
Degree
Master of Science
Degree Program
Biomedical Engineering
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
engineering, biomedical,OAI-PMH Harvest
Language
English
Contributor
Digitized by ProQuest
(provenance)
Advisor
[illegible] (
committee chair
), Khoo, Michael C.K. (
committee member
), Yamashiro, Stanley (
committee member
)
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c16-312622
Unique identifier
UC11327957
Identifier
1420389.pdf (filename),usctheses-c16-312622 (legacy record id)
Legacy Identifier
1420389.pdf
Dmrecord
312622
Document Type
Thesis
Rights
Murakata, Tomonori Thomas
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 au...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA
Tags
engineering, biomedical