Upload
others
View
5
Download
0
Embed Size (px)
Citation preview
Sensor and Simulation Notes
Note 257
October 1976 .
A Study of Charge Probe Response ona Junctionof Unequal Radii Thin Wires
Santiago SingarayarB. M. Duff
University of Mississippi CLEAREDFORPil&LICRELEME
p@39 5%+97
Abstract
The experimental measurement of charge density induced on conducting
surfaces often makes use of a short monopole as the charge probe. Measure-
ments on structures containing junctions of unequal radii cylinders have
been hampered by incomplete knowledge of the probe characteristics. An,
investigation of charge probe response as related to the radius of the
cylinder on which it is mounted is reported here. An experimental determin-
ation of the ratio of probe response when mounted on two different radii
cylinders has been conducted using a coaxial line geometry with a step
change in radius of the inner conductor. In addition, a theoretical solution
in numerical computations of the probe response in relation to the radius of
the cylinder upon which it is mounted has been conducted. The results of
these studies are presented in the following report.
o
TABLE OF CONTENTSN
LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . . . .“.3~
CHAPTERw
1. INTRODUCTION . . . . . . . . . . . . . . . . . . . . . ...5
II. MEASUREMENT TECHNIQUE . . . . . . . . . . . . . . . . . . . @*
lntroduc.tLon . . . . . . . . . . . . . . . . . . . . . . 8
Analysis of Coaxial Transmission Line . . . . . . . . . . 12
Description of the Measurement System . . . . . . . . . . 1.6
Determination of Probe Calibration Factor . . . . . . . . 21
Curve Fitting: Modified Prony’s Method . . . . . . . . . 2~”
111. NUMERICAL ANALYSIS. . . . . . . . . . . . . . . . . . . . .28
Charge Probe as a Scatterer: Theory and Solution . . . . 28 m
Formulation of Straight Wire Integral Equation . . , . . 29
Charge Probewith50flLoad . . . . . . . . . . . . . ..32
Numerical Solukion Technique . . . . . . . . . . . . . . 33
Expansion of Current and Formulation of Matrix . . . . . 35
Determination of Probe Calibration Factor , . . . . . . . 39
IV. PRESENTATION OF DATA AND CONCLUSION . . . . . . . . . . . .42
Introduction . . . . . . . . . . . . . . . . .’. . . . .4z
Measured Data for the Charge Distribution . . . . . . . .42
Probe Calibration Factor . . . . . . . . . . . . . . . .~2
Conclusions. . . . . ... . . . . . . . . . . . . . . . . .64
“
2
1
oLIST OF FIGURES
FIGURE*
2.1
8 2,2
2.3
2.4
2..5
2.6
3.1
3.2
3.3
3.4
3.5
3.6
4.1
4.2
4.3
4.4
4*5
4.6
Charge Probe . . . . . . . . . . . . . . . . . . . . . . . . .
I/p Variation as a function of (p-a) . . . . . . . . . . . . .
Coaxial System with Step Radius System . . . . . . . . . . . .
(a) Zero Order . . . . . . . . . . . . . . . . . . . . . . .
(b) Approximate actual distributions of Charge per unit
length and Radial E field in Coaxial Line . . . . . . . .
Coaxial Transmission Line . . . . . . . . . . . . . . . .
Block Diagram of Measurement System . . . . . . . . . . .
Charge Probe on a Charged Conductor . . . . * . . . . . .
Charge ProbeasaScatterer . . . . . . . . . . . . . . .
Lumped LoadZ1oadati= 0 . . . . . . . .. . . . . . . . .
Triangles in Piecewise Linear Testing Scheme . . . . . . .
Pulse Arrangement for Approximating current on Straight
Wire. . . . . . . . . . . . (l. . . . . . . . . . . . . .
Relative locations of-Pulses (Bases Set) and
Triangles (Testing Set) on Straight Wire . . . . . . . . .
Maximum Charge Condition - Probe Height 0.10”, al = 5/16”
Medium Charge Condition - Probe Height 0.10”, al = 5/16” .
Minimum Charge Condition - Probe Height 0.10”, al = 5/16”
Max5.mumCharge Condition - Probe Height 0.08”, al = 5/16”
Medium Charge Condition Probe Height 0.08”, al = 5/16” . .
Minimum Ch&rge Condition - Probe Height 0.08”, al = 5/16”
. .
. .
. .
. .
. .
. .
. .
. .
. .
. .
● ☛
✎ ✎
✎ ✎
✎✌
PAGE
9
11
13
14
14
15
27
30
31
31
34
38
40
43
44
45
46
47
48
3
-!- =
.
FIGURE
4.7
4,8
4.9
4.10
4.11
4.12
4.13
4.14
4.15
4.16
4.17
4.18
4.19
4.20
Maximum Charge Condition - Probe Height 0.06”, al = 5/16” .
Medium Charge Condition - Probe Height 0.06”, al = 5/16?’. .
Minimum Charge Condition - Probe Height 0.06”, al = 5/16” .
Maximum Charge Condition - Probe Height 0.10”, al = 3/161’ .
Medium Charge Condition - Probe Height 0.10”, al = 3/16”. .
Minimum Charge Condition - Probe Height O.10~’,al = 3/16” .
Maximum Charge Condition - Probe Height 0.08”, al = 3/16” .
Medium Charge Conditicm - Probe Height 0.08”, al = 3/16”. .
Minimum Charge Condition - Probe Height 0.08”, al = 3/16” .
Maximum Charge Condition - Probe Height 0.06”, al = 3/16” .
Medium Charge Condition - Probe Height 0.06”, al = 3/16!’. ,
Minimum Charge Condition - Probe Height 0,06”, al = 3/16” .
Probe Calibration Factor for 5/16’~to 1/8” ratio of radii .
Probe Calibration Factor for 3/16” to 1/8” ratio of radii .
PAGE
49
50
51
52
53
54
55
56
57
58
59
60
61
63
.
o
●
●
0
The
antennas
standing
measurement of
and scatterers
Chaptez I
INTRODUCTION
the current and
is important in
charge distributions on
order to gain an under-
of their electromagnetic behavior and to aid in developing
and checking of theoretical solutions, The basic techniques for\
making these measurements have been established for some time and
used by many investigators. However, during a recent experimental
project, to be described below it became obvious that more accurate
knowledge of the response characteristics of charge probes was
necessary. The resulting investigat-ionof charge probe response
forms the basis of this thesis,
The experimental project which leadsto the investigat~on re-
ported here, concerned the measurement of
induced on structures containing junction
radii. The purpose of that investigation
charge distribution
of thin wires of unequal
was to experimentally
determine the appropriate condition on charge which should
applied in theoretical solution of this type of structure.
A theoretical analysis and discussion of the junction
be
has been
presented by T.T. Wu and R.W.P. King [3]. The probe used for the
measurement of charge distribution consisted of.a very short mono-.
pole mounted on the surface of.the cylindrical structure. For
the experimental determinati.onof the charge-condition at the
5
-junction,an accurate knowledge of the probe response as a function
of the radius of the cylinder upon which it is mounted is necessary.
A charge probe mounted on a cylindrical structure has been analysed
theoretically by R.W.P. King [1]
However, these investigations do
solutions for the probe response
and experimentally by Whiteside [2].
not provide sufficiently accurate
as a function of the radius of the
cylindrical stru~tures. An independent experimental and theo~’etical
investigation of the charge probe response has been conducted, the
results of which are presented in this thesis.
Chapter II deals with the measurement technique by which the
probe calibration factor can be determined. This probe calibration
factor has been determined for 3 different probe heights for 2
junctions of thin wires of 5/16” to 1/8” radii and 3/16” to 1/8”
radii. To make sure that the probe calibration factor is the same
for all charge conditions at the junction, the probe calibration
factor has been determined for 3 different charge conditions namely
,
8
maximum, minimum, and a condition in between maximum
The experimental data was analysed by using modified.
A numerical solution for the current induced in
and minimum.
Prony’s method.
the probe
has been carried out using moment method techniques [4,5]. The
analysis models the probe as a monopole receiving antenna perpen-
dicular to an infinite plane condcctor, but includes the incident
~ field variation which is to be expected when the probe is mounted
on a cylinder. Chapter 111 deals with the methods used for this
numerical analysis.
6
Finally, in Chapter IV, the experimental data, the results and
the conclusion are presented.
w
.
●
CHAPTER 11
MEASUREMENT TECHNIQUE
(i) Introduction
The measurement of the surface charge density on a conductor may
be related to ameasurementof ~ field by using the boundary condition,
fi.cE = rjo
(2.1)
where &o
is the relative permittivity, 6 is the unit normal vector to.
tfiesurface of
surface charge
charge density
the test structure under consideration and q is the
density. It is therefore possible to obtain the surface
by using a sensor whose output is proportional to the
normal component of the E field at the surface. The most common method
uses a short monopole receiving antenna, as shown in Fig. (2.l),which
is perpendicular to the surface of the test structure. The signal
induced in the load is however related to the ~ field along the entire
length of this monopole rather than simply the field at the surface.
If the normal component of the ~ field is constant over the length of
the probe there is no problem. But generally this is not the case.
In order to relate the monopole probe response to the charge density
on the test structure it is therefore necessary to have some knowledge
of the manner in which ~ varies over a short distance away from the
surface.
.
8
w
‘8 f
;.
Coaxial cable
-.
f 1
--L_+--—.—. .—.! .—. —. —.—. -(!9* .—.—I*(a) Probe carriage for smaller radius tube.
a *
.—— —7T’—
. — . —. —.— . —.t“ ‘—1t1
.
(b) Probe carriage for
—.—. e ,’.—— .1 .—1!Ilarger radius tube.
Fig. (2.1) CHARGE PROBE.
I
The present work considers only test structures composed of
electrically thin conducting cylinders. The field produced by a
1uniformly charged infinite cylinder is known to vary as —. If the
P
test structure (a cylindrical antenna) is of electrically thin radius
and the probe is electrically short, then ~ over the length of the
probe is controlled primarily by the local charge distribution. The
field then becomes.essentially the same as that of an infinite cylinder
and ~ is very nearly
near discontinuicies
of this variation on
1proportional to –. This assumption is not valid
P
of the conducting surface, however. The effect
the response of the probe depends on the radius
of the test structure. To illustrate this effect, Fig. (2.2) presents
~plotted as a function of (p-a) for three different cylinder radii,
For a given probe height, the variation of EP
over the length of the
probe is less when the probe is mounted on a large radius cylinder.
The effect of the: variation ntaybe reduced by reducing the
probe height. However, the signal induced across the load is also
reduced and so the signal to noise ratio. A second effect of changing
the radius of the test structure is that the image plane for the
monopole probe is of different radius of curvature. The fact that the
charge probe is of finite size and it responds to fields away from the
suiface of the test structure makes it necessary to determine a probe
calibration factor as a function of cylinder radius in an exact manner.
To determine the probe calibration factor experimentally, a system
containing cylinders of different radii and for which the variation of
the charge density as a function of cylinder radius can be predicted
o
t . *
e
,. wP (b
●
0.04 0.06 0.08 0.10 0.120.0 0.02
Fig. (2.2) ; VARIATION AS A
(p-a) in inches.
FUNCTION OF (p-a).
1’
I
is needed. The system chosen was a coaxial line with step change in
the radius of the inner conductor. This system is shown in Fig. (2.3)
and has been analysed by Marcuvitz[6]. The TEM mode ~ field dis-
tribution of the above system is shown in Fig. (2.4). The ~ field is*
uniform and proportional to ~ inside the coaxial line except near theP
junction. The local effect near the junction can be replaced by a
shunt capacitance.so that the Kirchoff~s current law is satisf~ed.
At the junctton the voltage is continuous making it easy to analyse
the charge condition at the step in the following way.
(ii) Analysis of Coaxial Transmission Line
Fig. (2.5) illustrates a coaxial transmission line with a TEM
mode propagation. In cylindrical coordinates the two dimensional
Laplace equation is,
1 32@~*(p?&=-= oP a~2
For a potential function independent of angular coordinate+,
eq. (2.2) reduces to
The solution of equation (2.3) may be expressed as,
(2.2)
(2.3)
-(2.4)
●
Imposing the bo!.mdaryconditions @ = VO at p = a and @ = O at p = b,
eq. (2.3) yields the final solution for the potential as,*
12
● 1
e
1,,
I
PLJ
.
4I P’
Charge Probe I $dvable Short “1a ( .--, --- ---I
II
a
1---- ---- ---
I 4/
iJ _ .If.—.—.—.—-—.—“z
I----- ---- -- ---,I ------ ---
/ 4
Step radius system/ I Coaxial system ~
Ito#
t1
I,,
;1 ,1
1’
Fig. (2.3) COAXIAL SYSTEM WITH STEP RADIUS SYSTEM.
F.b
●
m“
------- .
---- ---
.— --- -
-------- --
------ --
-“
------ -
.z----.----
.
r
.
---’--------—
-----------
---------- --
----- ----”
---- ---- -
+n>
4.
t
+
+
+4.
+
+
+
+
+—
,------ ---
----- --
-.-...+_ --
----% ---
-------
..-.-------m-------
-. . .------ -
--‘. ------ -
-- --~------
-------- ----
----- ----
[
r
------ ---
----- --- -.----- -------- ___
---—- ------- ---------- -_-,----- -------- -------- --—_ .
1
Iv!-cr~”
I
_____ ____
—--- --- ---
------- ----
--- ------ _.
---- - ----
1I
i1
1IIIII1
.a>.!
I
I1III
III
;T
*NN
0w ;
P.-—-- ---
----- -------- --
----- --~
--- -— --”
---- —-
---- ----
-- -----
----- ---
---- --- -.- —- --- ----
----- ----
---- -----
----- ---- ---
---- ------ -
N&nN g
ri
-,,, ,
i
●
●
----
---
,)-
k field
~ field_—
Fig. (2.5) Coaxial transmission line.
an associated charge density which differs from that of a TEM mode as
indicated in Fig. (2.4)(b).
The voltage
different radii,
at the junction.
the junction are
is continuous at the junction but because of the
a difference in the magnitude of charge density exists
The ratios of linear and surface charge densities at
given by L
L .LL1 L,-=— .
(iii) Description of the Measurement System
A block diagram of the measurement system
(2.9)(a)
(2.9)(b)
.
.
is given in Fig. (2.6).”
The commercial instruments used in the system are given in the block
diagram itself.
The power at 300 MHz is fed to the amplifier. The amplifier sig-
nal is fed to the coaxial system through a balanced feed to avoid
higher order mode propagation. Matching of the amplifier output to
the coaxial system is done through a single stub tuner. The reference
signal to the vector voltmeter reference channel is taken from the
amplifier outpu~ with proper attenuation. The signal to the frequency
meter is also is taken from the amplifier output with proper attenua-
tion. The signal from the charge probe is fed to the test channel
16
o
~ field of TEM mode in +Z direction,
where k is the propagation constant.
Using eq. (2.5) ~ may be expressed as,
The surface density of
at z=O may be obtained
eq. (2.1) as,
‘O 2Pe-jkz——In(&) p
al
chafge ?IIon an inner conductor
from eq. (2.7) and the boundary
(2.5)
(2.6)
(2.7)
of radius a,,1.
condition
(2.8)
-.Equation (2.8) can be written as,
q~ = c1‘o
where Cl is the capacitance per unit length and ql = 2na1Tllis the
linear charge density. The same.relationship applies to the inner
conductor of radius a2“
It must be remembered that the above analysis assumes that only
TEM mode fields exi’stand that the junction is represented by a lumped
shunt.capacitance. The effect of the step change of radius is actually
distributed.over some short distance on either side of–the junction with.
#
17
of the vector voltmeter. The power level is monitored by the broad-
band sampling meter to compensate for the drift in power level.
The magnitude or the phase of the signal is switched to the.
digital voltmeter by means of the relay actuator. The calculator
samples this output 5 to 10 times and takes the average to minimize .
the effect of the minute-fluctuations in the probe signal. The magni-
tude and phase quantities are stored in the calculator for lat~r
plotting and for further use.
The stepped radius coaxial line was
brass tubing. The three different inner
constructed from cylindrical
conductor radii 0.3125’!
(0.79375 cm), 0.1875’’(0.47625cm), and 0.125’’(0.3175cm) were selected
as being available readily. Two inner conductor systems were con-
structed one composed of 0.3125T’and 0.1251’radius tubing, the other
of 0.1875” and 0.125’!radius tubing.
was used as the outer conductor. The
conductor was filled with styrofoam.
A tube of 1.0” inside diameter
space between inner and outer
A slot of 1/19’’(0.13368cm) was
cut axially in the inner conductor tubing to allow movement of the
charge probe along the structure. The probe carriage was designed
to slide inside of the inner conductor tube and extend through the
slot to the surface of the tube so that the structure is electrically
continuous. The charge probe was constructed on a probe carriage
using a semi rigid coaxial cable. The charge measurement was con-
ducted on two different radii tubes. If two different probes had been
used it would have been necessary that they be constructed identically.
.
0
.-
@Othe~ise, it would have become necessary to calibrate them indepen-
dently. To avoid this problem, a single charge probe with two different
probe carriages was used. The wall thickness of the tubes had to be
taken into consideration while making the probe carriage. For example.
for the 0.3125” radius tube wall thickness was 0.06’’(0.1524cm) and the
0.125” radius tube was 0.04”. The probe carriage fit exactly inside.
of the tube and through the slot to the surface so that the charge
probe was effectively mounted on the surface of the inner conductor
tube. The cable used for making .the probe was UT 20 manufactured by
Uniform Tubes, Inc., with the following specifications.
Impedance: 50 Ohms
o
Outside diameter: 0.023’’(0.05842 cm)
Center conductor diameter: 0.0045’’(0.011.43cm)
The outer–conductor of–the cable was soldered to
cut at the surface. The dielectric insulator on
was removed so that the inner conductor extended
the probe carriage and
the inner conductor
beyond the surface to
form the monopole antenna. It was necessary that this monopole be
straight and perpendicular to the test structure. The signal from the
probe was carried by a coaxial cable contained inside of the inner
conductor tube. Care was taken to see that stray pickup did not affect
the signal by shielding the signal cables. A 1/8” diameter tube
attached to the probe carriage was used both to protect the signal
cable and to provide a means of positioning the probe.
aThe
actual E
measured
charge density measurement was made at intervals of 1 cm. The
field distribution near the step due to local effect made the
result unreliable in this region. Moreover the charge probe
was mounted at the center of the probe carriage which made it impos-
sible to reach the step in the large radius tube. For analysis of the
data, afew data points were eliminated near the junction to eliminate
the local effect and the data was analysed by using a modofied Pronyfs
method to fit curves to the TEM mode distribution. This procedure
will be discussed separately.
Measurements were made for probe heights of 0.10’’(0.2540cm) to
0.06’’(0.1524cm) at an interval of 0.02’’(0.0508cm). The charge
condition at the junction was varied by using the movable short and
the experiment was carried out for three charge conditions at the
junction, namely charge maximum, charge minimum and a case in between
maximum and minimum. The experiment was repeated for consistency.
Even though frequency drift was observed, the change in electric length
of the test structure was of only of the order of ~ 0.00IA which was
considered negligible. Any change in the power output
later was
broadband
Care
compensated for
sampling meter.
w’astaken while
probe carriage fits flush
by monitoring the oscillator
making the probe to see that
with the test structure. To
of the oscil-
output using a
the top of the
achieve good
elect~ical contact between the probe carriage and the test structure,
.
0
silver conducting grease (Eccoshield, manufactured by Emerson & Cuming)
20
o was applied along the sides of the probe carriage and the sidewallsof the probe. The signal carrying probe tube was coated with an
absorber (Eccosorb by Emerson & Cuming) to avoid the possibility that
. the probe tube would become the center conductor of a TEM structure
inside of the inner conductor of the test structure and give rise to
.propagation and unwanted resonances.
After obtaining the extrapolated value of the probe reading at the
junction the probe calibration factor was obtained as described in the
following section.
(iv) Determination of Probe Calibration Factor
Let PI and P2 be the probe reading obtained by extrapolating the
measured data to the junction of–the large and small tubes of radii a1
0
-.
*
and a .2
Let N1 and N2 be the normalizing factors.
‘1 = Nlrj
‘2= N2?12
therefore
~1 = ‘1 ‘2.—~ ‘2 ‘1
nlfrom eq. (2,9)(b) the ratio of— is given by,
~2
a2—
al
(2.10)
(2.11)
21
N,‘HIe
(v)
probe calibration factor, $ is then given by,QL
Curve Fitting: Modified Prony~s Method
,
(2.12)
The experimental data obtained has magnitude and phase so that it
is a complex quantity. Curve fitting to the experimental data is done
by using a modified Prony’s method. Prony’s method [7] fits a set of
data to a series of exponential. By assuming a function of the form
f(z) = Aleyz + A2e-yz
because of the TEM mode transmission,
the exponential series of the general
Y= (a +
only two terms
Prony’s method
j $)
(2.13)
are selected from
with
(2.14)
.
.
.
where a is the attenuation constant, B is the propagation constant.
= A e(a+j6)z+A2e-(a+j6)’f(z) ~ (2.15)
where Al and A are complex constants, A2
~, A2, a, 6 are to be determined.
Let Az be the spacing between the data points and
~yAz‘1 =
(~.16)
-yA!z .
‘2=e (2.17)
220
then
Assume U1 and U2 are
f ~+l(!dz) = A1u1k+A2u2k
two roots of
U2+(%U+UO=01
where aoand a
1are unknown.
Equation (2.18) can be expanded to
‘1=A1+A2
‘2 = ‘cl 4“A2U2
2 2
‘3 = ‘I”l ‘-‘2U2
‘k= A1ulk--l+A2U;-1
(2.18)
(2.19)
(2.18)(a)
(2.18)(b)
(2.18)(c)
(2.18)(d)
Since the above set of equations satisfy equation (2.19)
‘3 + ‘2al ‘E‘Iao =o
‘4 + ‘3al ‘p‘2ao = 0
The unknowns an, a, can be determined from the above equations. Onceu L
a.‘ al are‘nom’the roots Ul, U2 can be determined from equations(2.19). Once U1 and U2 are known, y may be obtained from equations
(2.16) or (2.17) or,
Y = ln(ul)/Az
-y = kn(u2)/Az
23
(2.20)
(2.21)
●
With y known Al and A2
Since the roots U1 and
stituting in (2.19) we
are then determined from (2.18}(a) and (2.18)(b).
1
%!are reciprocal by definition, U2 = ~ . Sub-1get
2‘1 ‘W+ao= 0
~2+ct ~+ao=o1U2
‘2
J.
(2.22)
(2.23)21 + alu2 ‘%U2 = 0
Comparing (2.19j and (2.23) we obtain,
ao=l
U*+alu+l= o (2.24)
If we determine al, then the roots can be found easily in the following
way,
fl+f2~1+f3=o ; f2a1 = -(fl + f3)
f2+f3c11+f4=o ; ‘3% =-(f2 + f4j
‘3+f4%+f5=0; f4cx1= -(f3 + f5)
With three datapoints one can easily determine al from which U1 and U2
can be determined.
The method descrj>ed so far is not satisfactory for the set of
data’where experimental noise is present. A larger number of datapoints
24 ●
used and a least square curve fitting technique isis employed.
3M=N, where NLet MAz be the spacing between datapoints. When
is —,.th_etotal_nurnber_ofdata_po_ints,
6f ti+f = om+l 2m+l
fl +
‘2 +
f3 -1-
fm +
; fm+la=-(fl+ f2m+1)
.
; f~+2~ =-(f2 + f )2m+2f~,a + f
2m+2 =o
fm+3‘+f2m+3=0 ; ‘m+3
~ =-(f3 + f )2m+3
f2ma -1-f3m = o .$ f2mcl=-(fm + f3m).,
In general
[F]cx= -(G)
.[FTF]cx= -[FT] [G] whereo
[F’] is the transpose matrix of-[F]
Therefore u = - [FTF]-l [FT] [G] (2.25)
The residues Al and A2 can be determined in the following way
‘1 = Al -1-A2
A ;yAz ●1
A e-yAz2‘2 =
i
A1(eyAz)N-l +A2(e-yAz)N-lfN =
[F] = [G] [A]
[A] = [GTG]-l [F][GT] (2.26)
25
0
.-
with Al, A2, Y1 and Y2 thus determined, a continuous curve may then
be computed from equation (2.15) and extrapolated to the junction. The
value of P and P in equation (2.12) were obtained by this procedure.1 2 .
a
.
.
26
:
. ‘1
‘1
I‘1,!II,:
1’,,
II Nd
,’,,
II ,/
nFREQUENCY r SINGLEMETER STUBHP 5345A TUNERBROADBANDSAMPLINGVOLTMETER ~HP 3406A1
50L?TEEf
1 5ofJ
r--
HP 778A DUALDIRECTIONAL COUPLE
OSCILLATOR AMPLIFIER
GR 1362 BOONTON 230A ---- -. ---- -.
i 4 50QTEE
=4MAG.
VECTORVOLTMETERHP 8405A
H&E
I I & 50Q I4 1
11% I FEEDS CHARGE PROBE
1AII f
REI?.
PEST.1
poWER
DIVIDERa- ! COAXIAL MOVABLE SHORr7
~1’STVM
‘3A -E!ZI”FFig. (2.6) BLOCK DIAGW OF MEASUREMENT SYSTEM.
“uPLOTTERHP 9862A
#
CHAPTER 111
NUMERICAL ANALYSIS
.
(i) Charge Probe as a Scatterer: Theory and Solution
The preceding chapter discussed the experimental determination of
the response of a charge probe. A theoretical analysis and numerical
solution for the probe response and calibration factor are presented
in the present chapter. For the purpose of this analysis, the probe
is modeled as a monopole scatterer perpendicular to an infinite plane
conductor and planar
Chapter II the major
the probe is mounted
image theory is employed.
effect of the cylindrical
1is the ~ variation of the
As discussed in
conductor upon which
~ field produced byIJ
the charge on the test structure. In the treatment of the probe as a
scatterer the correct ~ dependence of the incident field is imposed.
It is recognized, however, that some error is introduced by using
planar image theory rather than the Green’s function for a source ex-
terior to a cylindrical conductor.
From the specification given in Chapter 11, for the cable from
which the probe is constructed, the electrical radius, ka, is found to
be .00036 for an operating frequency of 300 MHZ. It is therefore
appropriate to make the usual thin wire approximations.
.
a
.,
28
(ii) Formulation of Straight Wire Integral Equation
The charge probe of length ~ slides in a charged conductor of
radius a~ in a homogeneous medium characterised by (u, e, u = O) is-.
shown in Fig. (3.1). The z axis of a cylindrical coordinate system
lies at the center of the charged conductor as shown in Fig. (3.1).
The probe is perpendicular to the charged conductor and extends frcm
PH=
al ‘0 p.= al+z”For convenience a coordinate
used with a new variable < = (p-al) measured along
shown in
The
ES(C)on
where Ac
—J-
Fig. (3.2).
current and charge on the probe produce an:,
the probe surface which has a L component,
— ~;(~) ..5U!! (t
k2 3
9
0
4PI,
slot.- . ..- ------ ---
Coaxial Line 1
1-. —.- ----- . . . —--+
Charged Conductorz
Fig. (3.1) CHARGE PROBE ON A CHARGED CONDUCTOR.
*.I t
.
e
,’
I
a
L.-lF-J
+H/2
o
III
-H/2 i
II‘1
‘:
1III
I
Fig. (3.2) CHARGE PROBE AS
+H/2
0
-H/2
I
.
50S250Q
:
.> A
7
+H/2 1
Ak
-H/2
Fig. (3.,3) LUMPED LOAD ‘LOAD AT C=O.
*
‘,
‘,
t “1’
‘x”
I
where a = radius of probe, and @ = angle variable’of a cylindrical
coordinate system with C as the axis.
Applying the boundary condition that the tangential electrical
field on the probe surface be zero, we obtain,
+0 +E;(G) = Oand therefore
@ = -E;(C)
Thus the integro-differential equation becomes,
(,iii) Charge Probe with 50Q Load
(3.4)
The charge probe sees the 50 0 input impedance of the coaxial
cable, so that loading on the probe should be considered. For a
lumped load of impedance Zload inserted at L = O as illustrated in
Fig. (2.3), the current IC in ZIoad causes a scalar potential drop
A@g across the load of
The approximate electric field on the surface of
which must be added to the Es is,
the load element ,
To achieve more useful representation for the electric field due
to the IC in Zload
, we look upon the effect of the load at K = O as a
32
—
local behavior at C = O and write,
.
In the case of a.scatt$rerz
As explained in the
tional to ~ where p
(3.5)
beginning E;(3) has been selected to be propor-
= (~+al). Equation (3.5) with (3.2) is the final
form of the integral equation for the probe current. Equation (3.5)
has been solved numerically by the method described in the next section.
(iv) Numerical Solution Technique
In this section a technique for solving the integro-differential
equation is presented. This technique uses piecewise linear testing
and pulse expansion. One performs the testing of the equation to be
solved with elements t
,
ww-
C-m c-m+l c-m-l-2
-H/2
Fig. (3.4) TRIANGLES IN PIECEWISE LINEAR TESTING SCHEME.
e,..,,....._,-----------.....-.———...II II
o L—___ . . . . . . -t .,f
0. . . ... ,— I
@
(B
‘(3.6)(a,))AC+Z ~oadIc(o)8(c),t><
E; ,m tm >
.where the scalar product is
-- ..:-+, $=J+H:2,*(3.6)(b)
type considered here where
.
-{Jreal testing set t ofmFor atin(c)
the subdomain
is zero outside
. .Clz
Equation (3.6) becomes
“(cm-l,Cm+l)
!-
<
Jm+l
.=
E;(c) tin(c)dc
c =Cm-l
A A-1 C- gm ISetting tm =Am =A
in the above, the equation becomes
(3.7)
(3.8)
(Lm)
‘r
+
%1
+-“JIG(0)
35
with xn=O~l,~2,~3,~ 4,.., ~(M-l) - (3.10)(a}
and
For a function f sufficiently
one employs the approximation
%-f-l
(3.10)(b).
smooth over the interval (Gm-l,~til)
The final equation becomes
+-AZ I (0)60m= AE; (Qm)load G
where 6 is the kronecker delta
&=~{
1, m=nO, m#n
m= 0, ~1, -& 2,~ 3, +4,...,& (M-1)—
(3.11)
(3.12) ● -
(v) Expansion of Current and Formulation of Matrix
The current Icis represented by expanding in known expansion
functions each weighted with an unknown coefficient.
coefficients ultimately determine I . In terms of aL
uPn , as the known functions and the complex unknown
These unknown
pulse basis set,
coefficients In?
the current becomes
.
36
J
,
0= —
(0
“
IC(i) = ~ InPn(G)n=-N
(3.13)
The pulse arrangement for approximating the current on straight wire
is shown in Fig. (3.5)..
The vector potential can be written as
o
where
‘c!:)= f InALn(L)n=-N.
J
.H2
‘ALn(c) =% pn(C’)K(C-C’)dC’
G’=-H/2
(3.1.4)
(3.15)
@ a partial vector potential contributed by the pulse basis element
Pn(c)* In convenience, the following quantities are defined:
z =mn
+
Vm =
In terms of.the above definition equation
N~ ZmnIn = V
mn=-N m=O, +1, +2,. ... ~(1)l)——
37
(3.16)(a)
(3.16)(b)
(3.17)
LJm
I I 1,1,
IL(O
‘“w
G-N C-N+l ‘-N+2
-H/2
Lb
-.””.-
A
‘\ $.J+\
‘N-2 ‘N-1 ‘N
Fig. (3.5) PULSE ARRAIJGEKENTFOR APPROXIMATING CURRENT ON STRAIGHT WIRE.
,, ,, 1,,
+H/2
*
@
ends
The boundary condition that the current goes to zero at the
is incorporated by directly setting p_N = pN = O. Equation
(3.17) can
Therefore
be conveniently written in matrix form as,
‘“[Z][1] = [V] (3.18)
[1] = [Z]-l[V]
The total scheme of testing antiexpansion is shown in Fig. (3.6).
(vi) Deterrni.nationof Probe Calibration Factor
Once the current induced in the load can be found as a function
of a the probe calibration factor can be determined as follows.1’
Let 11 be the current induced in the load when the probe is mounted
on the larger radius tube and I~ be the current induced in the load
when the probe is mounted on the smaller radius tube, then relating
these to surface charge density
11 = ‘1’h
12‘:‘2%
If the linear density of charge is the same for both tubes then
q~ =q2
.
nl a2-=’ (—)‘2 al
39 L
(3.19)
(3.20)
.
4(9TestingTriangl I_A
C-5 C-4
-H/2
‘Opo
*-2
&‘-3
‘-3 c-z
k!!L
Note: m=~=o, ~l,~z,...
L
L-J A
13 “’
a“‘4 15=~ (Ha~f_pulse)‘3 ‘4
@
From eq. (3.19) and (3.20~, ‘1the probe calibration factor ~ isN 2
obtained as,
‘1 11 al—=— ._ (3.22)‘2 12 a2
The computed value of this ratio for different probe heights for two
different radii of cylinders are shown in Fig. (4.19) and (4.20).
41
CHAPTER IV
PRESENTATION OF DATA AND CONCLUSION
(i) Introduction
Chapter 11 and 111 have discussed the measurement technique and
the numerical analysis respectively. It is the purpose of the
present chapter to present the data resulting from both approaches
and discuss these results.
(ii) Measured Data for the Charge Distribution
As explained in Chapter 11 on measurement the charge density
measurement was made for 3 charge conditions at the junction near
maximum near minimum and a condition in between minimum and maximum.
The figures (4.1) to (4.9) present the charge distribution of
the step radius system with al= 5/16” and a~= 1/8” for 3 different. .
probe heights 0.10”, 0.08”, and 0.06”. For each probe height,
measured data for 3 different charge conditions at the -junctionare
presented along with the curve determined from the modified Prony’s
method for the magnitude of the measured data. Similarly the
Figures (4.10) to (4.18) present the data for the step radius system
with a~= 3/16” and a2= 1/8”.
(iii) Probe Calibration Factor
The probe calibration factor (N1/N2) as a function of probe
height as computed from the measured data is shown in Fig. (4.19) for
al= 5/16” and a2= 1/8”. The theoretical probe calibration factor
42
:
4-u
,,
CHFIRGE PROBE RESPONSE: PROBE HEIGHT 0.10”c1 PHFtSE 91 = 5/16” 92 = 1/8”x MRGNITUDE
o ,—.+ CURVE FIT
0.00 -50.00 -io.oo .10.00 30.00 :X10”
Z(MMI
Fig. (4.1) MAXIMUM CHARGE CONDITION - PROBE HEIGHT = O.1O”, al= 5/16”
o(3
G’N,
o0
G-+Hod
x’
fg-(3.
gel .LLJC3
o-‘?LIJo~~a
sL’
o0
&N
!00
. ..- ,-.
11,-
1
— —
tIII
Fig. (4.2)
CHRRGE PROBE RESPONSE:cl PHRSEx MFIGNITUIIE
o CURVE FIT
(9
PROBE HEIGHT 0.10”91 = 5/16n 82 s 1/8”
0.00
MEDIUM CHARGE
-30”00 -’20.00 . lb.oo 3i?.oo)kl~i
Z[MMjCONDITION - PROBE HEIGHT = 0.10”, al= 5/16”
5
.—..._ %. . ~
I
,.) ,
:. ,1
I
,1’
,’
1,
CHflRGE PROBE RESPONSE:
% MRGNITUDE
‘z CURVE FIT
o0
G’
Q
PROBE HEIGHT 0.10”fll = 5/16” W = 1/8”
.,
0.00 -30.00 -io.oo - lb.oo 3tl.oo 5
o’0 1,
>4 -iN
,1
00.
.s:00
)KltllZ(rfrl)
Fig. (4.3) MINIMUM CHARGE CONDITION - PROBE HEIGHT = 0.10”, a,= 5/16”.L .—.
1,
‘k-0
1’
0
CHftRGEax
9 .—.*
RESPONSES PROBE HEIGHT 0.08”PHflSE RI = 5/16” f%? s 1/8”MRGNITUIIECURVE FIT
1
0=00 -30 ●OO -io.oo lt).oo 30.00 5M@
Z(Mtl)
Fig. (4.4) ?4AXl141JlfCHARGE CONDITION - PROBE HEIGHT = 0.08”, al==5/16”
0
c)SJ
;0LLJo
:
CHFIRGE PROBE RESPONSE: PROBE HEIGHT 0.08” I
c) PHFISE ,Rl = 5/16” f12 = 1/8”x tlRGNITUllE I
o* CURVE FIT ~
I Io’0
.
I
m.l--m -d)
,!
-1-*-~w
,,I!’
>’w
-1
-1’0-INx “.-f=
0
I
t
(-J-
0.GoLIJ0
,,
(3
-c---l h.1
Io-:WIom ,i I
-—,
oc)
t4- 1.
1I
1 1 I
0.00 -30.00 -Lo’-ou-- .10.00 30.00 50’!003KmL
i!trml
Fig. (4.5) MEDIUM CHARGE CONDITION - PROBE HEIGHT = 0.08”, al= 5/16”
I
(4
mal-(-o
LLJ
00
2’.
PROBE RESPONSE: PROBE HEIGHT 0.08”PHI?SE 91 = 5/16” 82 = 1/$”MflGNITUOECURVE FIT
Fig. (4.6) MT.NIMUM
a
0.00 -30.00 -10.00 10.00 30.00
1 1,
IEIQ1Z(MMI
CONDITION - PROBE HEIGHT = 0.08”, al= 5/16”
m a
q ‘,!1
:
.c-W
b .
:
.
,1
CHflRGE PROBE RESPONSE: PROBE HEIGHT 0.06”a PtifW3E 91 = 5/16” fl12s 1/$”% I’IRGNITUDE
~,~
-1(!)
I I
m
.,
,,
In
on
09
~c5I.LJo
0-
31fll)1Z?(MM]
,,
I
Fig. (4.7) MAXIMUM CHARGE CONDITION - PROBE HEIGHT = 0.06”, al= 5/16”
I
u
CHRRGE PROBE RESPONSE: PROBE HEIGHT 0.06”Q PHRSE 81 = 5/16” W? = 1/8”x MRGNITUDE
cl_v
CURVE FIT
c?)
0.00 -!30 .00 -ioaoo . lb.oo 30.00 53E101
Z(Mtll
J?ig. (4,8) MEDIUM CHARGE CONDITION - PROBE HEIGHT = 0.06”, al” 5/16”
O . ,
ao
&N
o0
●
r-l
x
Q
o-%W
1!00
*
..
i .:
CHRRGE PROBE RESPONSE: PROBE HEIGHT 0.06”a PHfiSE fll = 5/16” ff2 s 1/8-# MFIGNITUDiE
o* CURVE FI’T 1:
‘o
-OJE“-0
IJJo
~o
(2)
o
0
IS”
(3
(3
c1
I
v K0.00 -30.00 -10.00 10*OO 30.00 &c
XIO1
Z[MM)
Fig. (4.9) MINIMUM CHARGE CONDITION - PROBE HEIGHT = 0.06”, al= 5/16”
I
o
CH9RGE PROBE RESPONSE:(9 PHRSEM tlflGNITUOE
o
N— CURVE FIT
Fig. (4.10)
,
V@
PROBE HEIGHT 0.10”fal = 3/16” R2 = 1/8”
Q
II 1
0.00 -50 ●OO -io.oo .10.00 30=00 53E101
Z(MMI
MAXIMUM CXMRGE CONDITION - PROBE HEIGHT = 0.10”, al= 3/16”
@!
II II . 1,
ml .. : *
.,
CHRRGE PROBE RESPONSE: PROBE HEIGHT
c1 PHflSE “ ,91 = 3/16”
% Mf!GNITUOE
o*lo- 1
92”= 1/8”
E CURVE FIT I
0.00 -30.00 -io=ooI I
10*OO 30.00 5IE1(I1
Z(MMI
I?ig, (4.11) MEDIUM C~GE CONDITION - PROBE HEIGHT = 0.10”, a~= 3/16”
,’!
0’<0:
.&do4
N
0
O(J-J
Q
o
0.
,R1!00
I
—-- n..
WI.P-
Ct!FIRGE PROBE RESPONSE: PROBE HEIGHT 0.10”
a PHRSE 91 = 3/16” F12 = 1/8”
x MRGNITUDEo@J — CURVE FIT
t?)
@)
;0.00 -30.00 -io.oo lb=oo 30.00 Ixmr
Z[MM)
Fig. (4.12) MINIMUM CHARGE CONDITION - PROBE HEIGHT = 0.10”, al= 3/16”
,=*
, ,
I ,,I
m’.+ ●
:
1!
CHflRGE PROBE RESPONSE: PROBE HEIGHT 0.08”c) PHflSE .Ifll = 3/16n 92 = Ii/8”% MRGNITUDE
oN CURVE FIT
-1
wo
o
0.
0
c)
0.00 -30.00 -io.oo lU.00 30.00 5MIO1
ZIMMI
1’,
‘1
o~iIOi!
0-
Dc)
.
DN
!00
Fig. (4.13) MAXIMUM CHARGE CONDITION - PROBE HEIGHT = 0.08”, al= 3/16”
,
●
CH13RGE PROBE RESPONSESc1 PHWSEx flRGNITUilE
aml CURVE FIT
PROBE HEIGHT 0.08”RI = 3/16” R2 = lf8”
c)
c1
Q
C9
(3
0.00 -30.00 -iOaOO .10.00 30.00*IQ’
2(flM)
Fig. (4.14) MEDIUM CHARGE CONDITION - PROBE HEIGHT = 0.08”, al= 3/16”
* , 1.
. w
m-J
:
CHf3RGE PROBE RESPC?NSEX PROBE HEIGHT 0.08”c) PHf7SE 91 = 3/16” 92 = 1/8”x tlfiGNITUDE
~R— CURVE FIT .
.CJ
-
Coml-.-l”C!3°
‘>n-1
wo3
I~l.g
b-i.
~ 0“(3
o0
&’
‘o
I1’
(
)c
I I I I
0.00 -30.00 -10.00 10.00 30.00 5*101
Z(MMJ
Fig. (4.15) MINIMUM CHARGE CONDITION - PROBE HEIGHT = 0.08”, al= 3/16”
~
,,‘1
uw
aa
.
*
-.
I
,1.,
CHRRGE PRi3BE RESPONSE: PROBE HEIGHT 0.06’a PHFISE 91 = 3/16” f12 = 1/8”x MRGNITUDE
ww CURVE FIT
4
m
. . I
?%lO1Z(MM1
- PROBE HEIGHT = 0,06”, al= 3/16”Fig. (4.16) MAXIMUM CHARGE CONDITION I I
●b +
.—,“
,
* ..
I
II I
1,,, I
,!z
:
CHflRGE PROBE RESPONSE: PROBE HEIGHTa PHfWEx MRGNITUDE
oN CURVE FIT
III
I
,,
.0
n
mml--d
.go
>-
-1.-l1--1:z“-0
I.lJn
~o
C5
o0
.0
91 = ,3/16=0.06”
R2 = 1/0”
c1
I v1 I 1 I
0.00 -30.00 -10.00 .10.00 30900 53K10X
Z(MMJ
Fig. (4.17) MEDIUM CHARGE CONDITION - PROBE HEIGHT = 0.06”, al= 3/16”
.0N
o0
.
,=
1!00
,,
‘1,
I ,)
Fig.
.
1.’1’. .1, .
(4.18)
<
PROBE RESPONSE: PROBE HEIGHT 0.06”PHRSE R1 = 3/16” F12 = 1/8”
tli3GNITUOEC3N CURVE FIT
Q
(!)
MH!
i I I i0.00 -30.00 -10”00 .10.00 30.00
MINIMUM CHARGE
MI(Y2(MMJ
CONDITION - PROBE HEIGHT = 0.06”, al= 3/16”
o“
C20
●
CJ(N
.
c)-?Ll
r2-
08
:
1.4
1.32
1.24
N1/N2
1.08
1.0
I
,! I, I Charge Condition at the’1 :,
I Junction ~1,
1,
4~I I I
JI l,’
PROBE HEIGHT IN INCHES
MinimumMedium
Maximum
Theoretical
,,,,II
,,
I
.. ,
0.04 0.06 0.08 0.10 0.12
Fig (4.19) PROBE CALIBRATION FACTOR FOR 5/16” to 1/8” RATIO OF RADII
obtained by the technique explained in Chapter III is also shown in
Fig. (4.19). Similarly Fig. (4.20) presents the probe calibration
factor for al = 3/16” and a2 = 1/8”.
Examination of Figs. (4.19) and (4.20) reveals differences of about
8% in the experimental values of N1/N2 for different charge conditions
are the junction. With the assumption that the probe responds to the
local charge density at its location on the cylinder N1/N2 should be a
function of probe height and cylinder radii only. The observed effect
of the position of the junction with respect to the standing wave pattern
is believed to result from inaccuracies in the extrapolation of the data
to the,junction. The extrapolation,
is accomplished by fitting the curve
data 5s available over a distance of
the accuracy in determination of the
limited. For these reasons the data
and hence determination of N /N12
to the measured data. Since the
only A/2 each side of the junction
parameters for this curve fs
obtained for the case of maximum
charge at the junction is the most reliable since the slope of the curve
in the extrapolation region is near zero. In future use of these experi-
mentally determined probe calibration factors greatest weight will, there-
fore, be given to the values determined for the maximum charge condition.
Examination of Figs. (4.19) and (4.20) also reveals a consistent
difference between the measured and theoretical values of N1/N2. The
reasons for this difference are not fully understood; however, it must
be remembered that the theoretical solution is based on a planar image
theory. The effects of the cylindrical geometry are the most likely
62
..= .-
m w : ●✌‘e
1.25
I
I 1.20
1.15
N1/N2
1.10
1.05
I
),,
!
Charge Condition at;theJunction
I
I
PROBE HEIGHT IN INCHES
1 * #
Medium
Minimum
Maximum
Theoretical
0.04 0.06 0.08 0.10 0.12
Fig. (4.20) PROBE CALIBRATION FACTOR FOR 3/16” tO 1/8” RATIO OF RADII
.
●source of the observed differences between experimental and theoretical
values. Until more accurate data and theoretical solutions are available,
the experimental value of N1/N2 will be used as the
(iv) Conclusions
The results of this investigation have clearly
effect of cylinder radius on the response of charge
most reliable.
demonstrated the
probe and have taken
a considerable step toward better understanding of charge measurements.
The probe calibration factors obtained can now be used to correct charge—
. distribution measurements on structures containing junctions of unequal
radii cylinders.
Further work, however, is needed for complete understanding of the
response of monopole probes on nonplanar surfaces. In”particular, an.—
effort should be made to include the cylindrical image plane geometry o
into the theoretical treatment presented here. In addition, other probe
configurations should be examined and evaluated with respect to their
response on nonplanar surfaces.
64
—
[El
[2]
[3]
[4]
[5]
[6]
[7]
REFERENCES
R.W.P. King, “Quasi-Stationary and Nonstationary Currents inElectric Circuits,” pp. 267-270, Encyclop-ediaof Physics,Edited by S. Flfigge,Vol. XVI: Electric Fields and Waves,Springer-Verlag, 1958.
H. Whiteside, “Electromagnetic Field Probes,”Technical Report SO. 377, Harvard University,setts, 1962.
T.T. Wu and R.W.P. King, “The Tapered Antenna
Cruft LaboratoryCambridge, Massachu-
and its Applicationto the Junction Problem for Thin Wires,” IEEE Transactions onAntenna and Propagation, Vol. AT-24, pp. 42-45, January 1976, andalso Interact~on Note 69, January 1976.
C.M. Butler, K.R. Umashankar and C.E. Smith, “Theoretical Analysisof Biconical Antenna,” report submitted to the U.S. Army,STFL4TCOM,1975.
R.F. Barrington, Field Computation by Moment Methods,Company, New York, 1968,
Macmillan
N, Marcuvitz, Waveguide Handbook, McGraw-Hill Book Company,New York, 1957.
Hildebrand. Introduction to Numerical Analvsis. McGraw-HillBook Company, New York, 1956.
65
!-