Scholars' Mine

Masters Theses Student Research & Creative Works

1968

An ultrasonic pulse generator using the transducer

as a frequency determining element
Donald E. Watke

Follow this and additional works at: http://scholarsmine.mst.edu/masters_theses

Part of the Electrical and Computer Engineering Commons
Department:

Recommended Citation
Watke, Donald E., "An ultrasonic pulse generator using the transducer as a frequency determining element" (1968). Masters Theses.
5224.
http://scholarsmine.mst.edu/masters_theses/5224

This Thesis - Open Access is brought to you for free and open access by Scholars' Mine. It has been accepted for inclusion in Masters Theses by an
authorized administrator of Scholars' Mine. This work is protected by U. S. Copyright Law. Unauthorized use including reproduction for redistribution
requires the permission of the copyright holder. For more information, please contact scholarsmine@mst.edu.
 l~.N UL'~'l.:U~SONIC PULSE GENERATOR USING

THE TRANSDUCER AS A FREQUENCY DETER.t"'1INING ELEI1ENT

BY

DOt~'\LD E. NA'I'KEJ I 'f t,tl/

------

THESIS

submitt:ed to ::.he faculty of :132919

rrHE UNIVERSI'I'l Ot.'' lV;ISSOURI :f\..T llOLI.A

in partial fu:!.f illr:1ent of th~:: .rt..~uir~~n1ents for the

Deg.ree of

tiJl:,Srf'ER OF SCIENC~ IN ELECTRICAL E.NGINEEHING

Rc)lla, Hissonri

1968
 ii

ABS'l"'RACT

This p-:1.:-x;r d~scrib.?S the development of a method of

ultrasonic pulse ry:meration in which the transducer deter-
mines th~ f:c~::quency of oscillation

..r:.,. advantage of this met.hod is the ability of

<J~finite

the frc~u~ncy of oscillation to change with changes in the

transd':.1cer mechanical resonant frequency. This maintains
peak efficiency at all times.

The method is demonstrated in a small sonar transmitter

but could be applied to any area involving driven ultrasonic
transducers.
 iii

ACKNO~LEDGEMENTS

The author wishes to express his gratitude to his

advisor, Dr. Norman G. Dillman, for his patience and assis-
tance throughout this Master's thesis effort.

Appreciation is expressed to Floyd s. Hall for his help

\vith the transformers used in the final circuit.
 iv

'l'l\BLE OF CO:NTENTS

Page
l~BSTH.. AC'r , ii
ACKNC:vJLEDGEHENrs
iii
LIS'r oe ILLUSTRATIONS v
I. INTRODUCTION 1
II. rRANSDUCER CHAP~CTERISTICS 2
III. TRANSDUCER EQUIVALENT CIRCUIT 6
I'.J. INI'riAL CONS IDEP.....~ 'I IONS 10
v. FINAL CIRCUIT 12
GENEF-'.AL DESCRIPTION 12
ANALYSIS .. 12
OSCILLA'l"'OR DE.SCRIP'I'ION 15
PULSE FORMING NE'IWORK 18
RESULTS .. 21
VI. CONCLUSIONS 25
BIBLIOGRAPHY 26
APPENDIX A: CIRCUIT PARTS LIST 27
APPE:~DIX Bs TRANSDUCER EQUIVALEN'r CIRCUI'r
DERIVATION 29
APPENDIX C: CO!v:PUTER PROGR.ll.NS 31
APP:t!NDIX D: TR.~l\SDUCER DATA 39
VI'rA 42
 v

LIST OF ILLUSTRATIONS

Figure Page
1. Transducer impedance characteristic 3
2. Combined mechanical response of transmitting
and receiving transducers 4
3. Transducer equivalent circuit 8
4. Equivalent circuit impedance characteristic 9
5. Block diagram of a simple ultrasonic generator 11
6. A method of effectively cancelling the
transducer shunt capacitance 11
7. Block diagram of oscillator circuit 13
8. Final circuit block diagram adapted to
root-locus analysis 13
9. Root-locus of final circuit 16
10. Oscillator schematic diagram 17
11. Pulse network schematic diagram 19
12. Typical output pulse waveform 22
13. Typical waveform showing pulse spacing 23
 1

I. INTRODUCTION

An undesi:cable prcp~rty of crystal type ultrasonic

tran~:;ducers is the change of ch<-1.:.aeter is tics which takes
place when t11ey are subjected to temperature or loading var-
iations. rhis cltunge occurs as a shift in the :cesonant
frequ8ncy of the unit such that a decrease in efficiency
results.

The purpose of this study "i.va s to develop a method of

ultrasonic pulse generation that uses the impedance c:hdrac-
teristic of the transducer to determine the frequency of
oscillation. In this v1ay, any change in transduc0r charac-
teristics would be accompanied by a corresponding chclnge in
frequency, maintaining operation at the tran3ducer's mc-
cho.nical r1::sonant frequency. This allovrs optin:n.t.'l1 efficiency
to be realized under all conditions.

Existing methods of ultrasonic pulse generation usu-

ally invclve driving the transducer Hith a fixed frequ~ncy

keyed c!.::ciJ.lator. The method developed in this paper i:l--

cludes the transducer. in a positive feedback loop in such a
manner that the transducer electrical l.mpedC\nce is th~ main
frequency determining element.
 2

II. TRhNSDDCt..~R CI-1'\RACTERISTICS

Threa .l-'>}:X-;!lco type TM-8A ultraso!1ic transducers 1 were

used to avoid obtaining one \vi th ur:.usual characteristics.
The~~c transducers are inexpt~nsive barium titanate units typ-
ical of those used in small boat sonar systems.

A typical transducer impedance characteristic is shown

in Fig. 1. The data (tabulated in Appendix D) were obtained
by driving the transducer with a sine~wave generator and
rneasurir~g voltage, current and corresponding phase shift
with an oscilloscope and current probe. Four resonant frc-
quencit.~s were f:;vident 'Iilith the p.rima.ry electrical resonance
occurring at 52.7 KHz.

SinC(-c! the transducers werQ specified for 200 KHz oper-

ation by the manufacturer, and this information appeared to
disagree with that obtained from the electrical character-
istics, the pri':tlary mechanical resonant frequency vlaS exper-
imentally determined.

'l'hre:.:: n1ethods of det:~nnin:tng m~chanical resonance were

us<?d, tv.ro of vthich involved oboerving the response of a
transducer to a d. c. pulse~ In th13 third method, on(~ trans-
ducer va:3 d:rivcn ~Iith
a sine H~ve generator while the re-
sponse of a second tx-ansduc1z,r ~paced tvro feet a-vray Has ob-
ser.<_,'(:~d on the oscilloscope. 'r~e re:::;ul ts of each method

~hov."ed 200 KHz to be t}:e pr:tma:ty ffi8chanical reso~'1ant fre-

quency. Fig. 2 sho\vS the responsrz- obtained in method three.

'l"11e apparant disagrP-~~tnent beth'cen the primary elec--

trical and mechanical resonant frequencies is probably due

-------I;;-~~~t~::nsdncers
are av~ilable from Marine Radio
Equip:nent Co.; 188 N. Hacker D:tiv-e; Chicago, Ill. 60606
at a cost of :f-9. 95 postpaid.
 3

75

Impedance
70 Magnitude

.........
,.Q 65
u
..........
-cv
'U
::3
+'
rt
~
0'1 60
~
Cl)
u +10
~
ro
'0 .........
<ll 55 Ol -10
0.. Cl)
s
1-i
Q)
H
tJ'l Phase
Cl)
-30
-
'0

Cl)
Angle

r-1
so tJ'l -50
s::
~
(l)
Ol -70
ro
..c
At
45 -90

0 0 0 0 0 0 0 0 0
N CVl ~ tO [' 0 0 0 0
r-1 N CVl ~
Frequency (KHz)

Fig. 1. Transducer impedance characteristic.

 4

40

0 db corresponds
to the 50 KHz
response.
32

24

_..._
,Q
rtJ
..._,
(!)
til 16
s:
a
til
Q)
IY.

-8 +----+----+-+-+- I 1-
0 0 0 0 0 0 0 0 0
--r 0 0 0 0
"
N CV) U)
r-f N CV) ~

Frequency (KHz)

Fig. 2. Cornbined :mechanical response of trans-

mitting and receiving transducers.
 5

to the construction of the transducer. At 52.7 KHz the

crystal is probably vibrating in a shear mode and therefore
does not couple energy to the water as efficiently as it
does when driven in a longitudinal mode. The desired fre-
quency of operation is therefore 200 KHz.
 6

III. TRANSDUCER EQUIVALENT CIRCUIT

An equivalent circuit was required for the transducer

so that an impedance function could be derived to represent
the transducer in circuit analyses. An equivalent circuit
also helped p:covide insight into the transducer's charac-
teristics.

A series RLC circuit and shunt capacitance are often

1
used to represent a ciystal near resonance However, since
the desired operating frequency of the transducer is not at
its primary electrical resonance, it was necessary to find a
model which would represent it over the entire range of
20 KHz to 300 KH:z. If a good fit could be obtained over
this range, a circuit analysis including this model should
show any tendency of the transducer to shift from one reso-
nance to another.

1~e circuit of Fig. 3 was used, letting one RLC branch

represent each resonance. The method used to determine the
element. values is described in Appendix B.

Fig. 4 is a graph of the impedance of the equivalent

circuit. The accuracy of the model is believed to be quite
good considering the range over which it was to be usable.

The circuit values were used in a computer program

(listed in Appendix C) to find the coefficients of the im-
6
pedance function. With s normalized by a factor of 10 , the
function iss

1 Narren P. Z..1ason, Piez_oelectri~ _9rystals and Their
~..P?J.iq_atiQ.!}. _to U1trasonics (Ne\-1 Yorks D. Van Nostrand
Company, 1950), P 67.
 7

N(s)
Z(s) = D(s)

N(s) = s8 + 0.159138 5 7 + 3.30227 5 6

+ 0.337821 5 5 + 3.47206 s 4 + 0.182954 s 3
+ 1.29058 5
2
+ 0.0191176 s + 0.104305

3 8
D(s) = 10- s [ 1.815 s + 0.288835 57
+ 6.15086 s 6 + 0.628086 55 + 6.61324 s4
+ 0.346611 53 + 2.50604 s2 + 0.03704 5

+ 0.0207038 J
White

1-' Q)
ro Q) 1.0
'-l_l_ '-l N '-l 0
0\ 0 VI 0
0 I
I

Shield 61
~T
1-'
1.0
0'\
1-'
N
0
N
0

1.0
1-'
0

~

'-l I
~1
0\ J
0 1-' U1 trr _1_
w

tT
Q)
w
w 1.0

Black

.Resistors in ohms, inductors

in mh, capacitors in pf.

Fig. 3. Transducer equivalent circuit.

CD
 9

75

70 Impedance
Magnitude

65

50

45

Fig. 4. Equivalent circuit impedance characteristic.

 10

IV. INITIAL CONSIDERA'l'IONS

The block diagram of the first circuit proposed to gen-

erate ultrasonic frequencies is shown in Fig. 5. If the
ga.in of the arnplj.fier and phase of the phase-.shift network
are prop0rly adjusted 1 oscillation can take place at any of
the transducer's resonant frequencies.

A root-locus analysis of this circuit was performed

\vith a capacitor across resistor R providing the proper
phase shift. r..L'he root-locus showed a pole 'vhich remained on
the positive reu.l axis for values of gain sufficient to
cause oscillation. The locus also shoHed that the gain
\vould have to be closely controlled or the frequency would
suddenly change from 200 KHz to 53 KHz.

'I'his circuit was built and tested to gain experience

with correlating root-locus information and actual response.
The circuit could not be made to oscillate reliably due to
ampl.ifier saturation. Even with simple a.g.c. (automatic
gain control), the waveform was non-sinusoidal enough to
make the frequency unstable. The circuit also proved to be
very sensitive to gain

.Although desirable results were not obtained, this

simple circuit provided some valuable information. The real
a::::is pole in the root-locus was responsible for the ampli-
fier saturation. Proper operation requires removal of this
pole. Tr!t~ circuit also shovled that sinusoidal oscillation

m.ust :Oe maintained if the frequency is to be predictable.

 11

Phase
Amplifier shift
network

Fig. 5. Block diagram of a simple ultrasonic generator.

Output
Transformer

Fig .. 6. A method of effectively cancelling

the transducer shunt capacitance.
 12

V. FINAL CIRCUIT

GENER..i\L DESCRIPTION
One reason t.hat the undesirable root remained in the
right half s-plane in the preceeding analysis is that c ,
0
the transducer shunt capacitance, creates a zero at the ori-
gin to which this pole converges. Therefore, if the effect
of C0 were removed, a circuit might be found that would op-
erate as desired.

A method of effectively removing c is sho~vn in Fig. 6.

0
If the impedance through which feedback current If flows is
small, then the component of If due to C is cancelled by
0
cl.

This arrangement requires the driving amplifier to sup-

ply reactive volt amperes to capacitor c 1 Even though c 1
dissipates no power, the transistor dissipation is in-
creased. It is possible to eliminate this by tuning out
this capacitance and c 0 with an inductor. This inductor,
t:he transducer, and c 1 are the frequency determining cle--
ments. Th~ inductor and c
1 help select the 200 KHz reso-
nance and provide enough selectivity to prevent oscillation
at any of the transducer's other resonant frequencies. The
inductance of the output transformer can be used instead of
an additional inductor if the core material and size will
allow proper transformer operation.

ANALYSIS
The block diagram of Fig. 7 was used in the analysis
leading to the final circuit. z1 represents the transducer
impedance, z is the compensating capacitor, and z3 is the
2
inductor mentioned previously. The amplifier gain k takes
into account the turns ratio of the output transformer so
that a ratio of 1:1 may be used here.
 13

+E

1:1
I
3
z1
Amplifier
gain =k

L __ _ _ _ _ _ _ _ _ _. __ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _~

-E

Fig. 7. Block diagram of oscillator circuit.

E
G

Fig. 8. Final circuit block diagram

adapted to root-locus analysis.
 14

For a root-locus analysis of the circuit, it was ne-

cessary to reduce the circuit of Fig. 7 to that of Fig. 8
This reduction was accomplished by recognizing thats

If the input impedance of the amplifier is much less than

either z1 or z2 z

:: ki
err

I err = Il - !2

G and H in Fig. 8 may now be found.

E k
G = =

-I err
H-
E
=

A root-locus program (listed in Appendix C) was Hritten to

find the roots to 1 + kGH for various values of k.

To obtain an initial approximation for k, the above

equations were rearranged to:
 15

( k - 1 ) Yl =( k + 1 ) Y
2
+ 4Y 3

For the approximation, y 3 is chosen to resonate with 1/2 Y

2
The solution to the equation therefore is k = 1.

For the root-locus analysis, k was allowed to vary from

0 to 12. The high values were of interest to check to see
if the circuit could start to oscillate at other frequencies.
The root-locus is sho\<Tn in Fig. 9. The locus crosses into
the right half plane at a gain slightly greater than unity
and a frequency of 213 KHz. The high frequency is attri-
buted to not knowing the exact value for C0 and the fact
that small errors in the equivalent circuit for the trans-
ducer could cause a large error. The results demonstrate
however, that the method is both feasible and practical.

OSCILLATOR DESCRIPTION
The schematic diagram of the oscillator is shown in
Fig. 10. Capacitor c was chosen to be equal to the total
3
effective shunt capacitance of the transducer at 200 KHz.
Inductor L was chosen to resonate with the series combi-
1
nation of the transducer capacitance and c 3

In use, the transducer would be connected also to a re-

ceiver so that it is desirable for one terminal to be con-
nected to ground. This is made possible by taking the feed-
back current from the centertap of the output transformer
T2.

star-ting ,..,ith the assumption that the circuit is not

oscillat.:ing (point A essentially grounded), if point A is
suddenly raised above ground by the pulse forming net\"lork,
 16

lO
lO
0
"
0
0
r-1
.t.>~
lO lO II II II - 1. 50

.-1 " ~ ~ ~
II
0
II
X I
.
""'
-;
~

1.25

1.oo
X

X
0.75
s normalized
by 10 6

o.so

0.25

Real ,, I o.o
axis -+- II
lO M r-1 \.0 ~ N
"
0 0

0

0

0
0
0
0
0
0
0

0 0 0 0 0
I I I I 0 0 0
I I I

Fig. 9. Root-locus of final circuit.

 T2
Transducer
Dl

AI ''' I I or I II II I I -

I I I I '-"\ I I ,..._, ,..._, Ll

D3
D2
~ c1
I I I
02

To pulse
forming
network

Fig. 10. Oscillator schematic diagram. 1-'

'-J
 18

diode o 3 will conduct and point A will be held at about

+0.7 volt. This voltage will be coupled to the bases of
transistors 0 1 and 0 2 through the input transformer T bias-
1
ing 0 1 and 0 2 to class A operation. Capacitor c 1 is charged
to about 0.3 volt by the resulting emitter current. As re-
generation takes place, the input signal grows until the
bias is insufficient to maintain class A operation, at which
time the circuit begins operating class AB. When this oc-
curs, capacitor c 1 begins to charge to a higher voltage.
When this voltage tries to exceed +0.7 volt, the input cur-
rent is shorted to ground by diodes n , o2 and n 3 Only
1
enough current will flow into the transistors to keep c
1
charged to +0.7 volt. The output current is therefore con-
trolled by this a.g.c. action and the circuit oscillates at
a frequency de-termined by the transducer and L 1 and c 3 The
output current level and hence the output vol t:age can be
controlled by resistor R1 This resistor determines how
much current must flow through the transistors before a.g.c.
limiting begins. If point A is suddenly grounded, bias is
removed from o1 and 0 2 and with the input current shorted to
ground by the diodes, the circuit quickly stops o~;cillating.

PULSE FORMING NETWORK

Typical requirements for a pulse generator call for a
pulse width of 0.5 to 1.0 msec with a pulse spacing of 40 to
50 msec.

The schematic diagram of the pulse forming network is

given in F'ig. 11. Transistor o3 along with R 2 , R3 and c 4
fonn a conventional unijunction oscillator with R2 and c 4
determining the pulse repetition rate. Transistors 0 4 and
o are connected in a monostable multivibrator configura-
5
tion. An advantage of this configuration is that no current
is draHn by e5.ther transistor between pulses.
 +6

R3
R2 ? I ~

Q4
< Ra <
Rg

Q3 I <..
I
R6
Hp cs

c4 I 1 l as
- -
R7

To
oscillator

Fig. ll. Pulse network schematic diagram. ......

\0
 20

vlhen the base of 0 receives a negative pulse from the

4
unijunct.:.ion transistor, 0 turns on and the voltage at its
4
collector rises toward +6 volts. This in turn starts tran-
sistor a 5 conducting, causing its collector voltage to drop
to"~;vard zero. c 5 begins to charge tov1ard +6 volts with the
indicated polarity. When the voltage across c 5 becomes
large enough for insufficient current to flow to keep o
4
conducting, o begins to turn off and regenerative action
4
quickly causes both o and o 5 to revert to their off states.
4
R and c detennine the pulse width.
9 5

output is taken from the junction of resistors R4 and

Rs '\vhich are chosen to supply sufficient current to the os-
cillator circuit. Between pulses, this point is essentially
at ground and during the pulse, it is at +0.7 volt due to
diode n in the oscillator.
3
 21

RESULTS
The element values for the oscillator and pulse forming
netvlork are tabulated in Appendix A. The circuit has the
follo\Ting characteristics z

d.c. supply voltage 5 - 8 volts

d.c. current 320 rna maximum

7 rna average

po~,rerdelivered to 0.75 watt peak

1 2
transducer ' 0.015 watt average

1
frequency 199 KHz

pulse width Oo8 msec

pulse spacing 45 msec

Fig. 12 shows the voltage pulse across the transducer.

The right half of the trace is the reflection from the end
of the vlater tank. The overshoot is due to the time it
takes capacitor c1 to charge before the a.g.c. begins oper-
ating properly. Ringing of the transducer is evident at tl1e
end of the pulne but it dies out in 0.1 msec. Thus the
Hould be ready to "listen" for echo signals
tra.n~:!ducer

1 msec after the beginning of the pulse. This restricts

the minimum range of the device to about 2.5 feet.

Fig. 13 is an oscillogram sho-:.;ing pulse 3pacing. A

s_?acing of 45 msec is sufficient for a maximum ran(JC of ap-
proximately 100 feet.

1
1~1e actual value is dependent upon the transducer.
2 This is po\-Ter delivered to the 200 KHz mode only;
not thQ total tranducer volt amperes.
 22

Vertica l 20 volts/div .
Horizontal 20 usec/div.

Fig . 12. Typical output pulse 't-laveform .

 23

vertica l 20 volts/div .
nori zonta l 5 msec/div .

Fig . 13 . Typical waveform showing pu l se spac ing .

 Stability of the circuit is quite good and a supply
volt.:age variation of 5 to 8 volts can be tolerated without
shifting frequency appreciably. Should the unit be used
consistently with supplies of less than 6 volts, the output
transformer should be rewound with a higher turns ratio.
This would maintain proper a.g.c. action and prevent voltage
limiting in the transistors which causes the frequency to
decrease.

Should a transducer be encountered with a shorter cable

than the 12 foot length supplied, the smaller cable capac-
itance can be compensated for by changing the value of c
3
This capacitor could be replaced by one of lesser value in
parallel vJi th a small trimmer. A neon bulb could be connec-
ted across the output transformer and c 3 tuned for minimum
brilliance. This is an easy method of initially adjusting
the unit for optimum operation and is based on the fact that
the amplifier is essentially a current source due to the
a. g. c. Since the impedance of the transducer ts rninimn."ll
. at
resonance, and it is being dr.iven by a constant current,
minimum brilliance of the neon bulb indicates proper op-
eration.
 25

VI. CONCLUSIONS

The t:hree main points of interest regarding the per-

formance of the circuit are&

1. The transducer helps determine th9 frequency

of oscillation allovTing the unit to track
changes in the transducer's resonant fre-
quency.

2. The circuit efficiency is high with an av-

erage battery drain of only 7 rna.

3. Substantial power output and good pulse shape

are obtained with a relatively simple circuit.

rhe ability of this circuit to adapt to changes in the

transducer's characteristics was best demonstrated by im-
mersing the transducer into water so that intimate contact
'tvith the water vTas not obtained. Since the resonant fre-
quency of the transducer in air .is higher tha:t i t is in
".vo.ter 11 the circuit should sho\ an increase in frequency d11e
to thc decreased loading. Results shovied trje frequency of
oscilJ.ation to be 3 KHz higher than when the face of the
transducer was soaped to insure intimate contact.

Although the circuit considered here was oriented to-

Tvlard G~H1 in a small boat sonar system, the rc,ethod of ultra-
soDic frequency generation could be used for any application
involvin~:r an ultrasonic transducer.
 26

BIBLIOGRAPHY

KINSLER, L. E. and A. R. FREY (1962) Fundamentals of

Acoustics. 2nd ed., ~iley, New York,
pp. 333-378.

YASON, W. P. (1950) Piezoelectric C1ystals and Their

Application to Ultrasonics. Van Nostrand,
New York, 508 P
MASON, w. P. (1942) Electromechanical Tr~nsducers and Wave
Filters. Van Nostrand, New York, pp. 195-
215.

I1ITCHELL, w. F. (1959) Transistorized Fish Finder.

Electronics vJorld, August 1959, PP 42-44,
108, 110.
SINGLE, H. c. (1960) Portable Depth Finder for Small
Boats. Electronics, Vol. 33, no. 6,
Feb. 5, 1960, pp. 50-51.
 27

APPENDIX A

CIRCUI'r PAR1'S LIST

Resistors ( Oh.'US)
Rl 2.2
R2 7.5 K
R3 300
R4, Rs 200
R6 3.3 K
R7 2.2 K
Ra 1.0 K
Rg 330

Capacitors (uf except as noted)

cl 10 e1ectro1ytlc
c2 55 electrolytic
c3 1390 pf
c4 6.8 electrolytic
cs 1 electrolyt.ic

Inductors
700 uh

Transj_stors
Fairchild 2N3569
RCA 2N2160 uniju:nction
TI 2N3703

TI 1N658
 28

'rrc-.nsformers
Primary - 30 turns; secondary - 10
turns each half. bifiler wound on an
Indiana General core no. CF102 of
0-6 type material.
'f Primary - 9 turns each half. bifiler
2
wound; secondary - 75 turns each half.
bifiler wound on an Indiana General
core no. CF102 of 0-6 type material.
 29

APPENDIX B

'rRl\NSDUCER EQUIVALENT CIRCUIT DERIVATION

Since no existing analytical method was known for find-

ing the equivalent circuit, a set of starting values was ob-
tained which was then modified by trial and error until
acceptable results were achieved.

The initial approximation was made that the Q's of each

resonance v1ere sufficiently high that the series resistors
could be temporarily neglected. The transducer impedance
could be expressed asa

2 2 2 2 2 2 2 2
(s +w1 )(s +w 3 )(s +w )(s +w )
Z(s) - 5 7

where the odd subscripted w's are the frequencies of the

irnpedance minima and the even subscripted w's are the fre-
quencies of the impedance maxima.

C , the total shunt capacitance, was found by evalu-

o
ating the function at 20 KHz which \Tas far enough away from
the resonant frequencies that they had a negligible effect.
Y ( s) v.ras then expanded by partial fractions. ':"Che values of
inductance and capacitance for each branch were determined
from thi.s expansion. Starting values for the series resis-
tors vrere determined from the value of the minimmn impedance
at each resonance.

A computer program {listed in Appenuix C) was written

t.o compute the impedance of the model at 100 frequencies.
An iterative procedure ,.,as then followed which involved
making specific changes in parts of the model to fit the
measured characteristic. The possible changes and their
 30

corres~~nding effects were1

1. Changing C0 - This alters the spacing between

the minimum and maximum points of all reso-
nances. Increasing c 0 decreases this spacing.

2. Increasing L 1 and decreasing c 1 by the same

factor - This increases the Q of the partic-
ular resonance causing larger impedance var-
iations.

3. Changing c or L individually - This changes

1 1
the frequency at which the resonance occurs

.'1\fter
. each change, the computer program ~;as used to
assess the results. The values found in Fig. 3 were be-
lieved to be near optimum for this model.
 31

APPENDIX C

COHPU'.rER PROGRAMS

The large ru~ount of mathematical manipulation involved

in the development was handled with the aid of an IBH System
360 model 50 computer at the University of Hissouri at Rolla
Computer Science Center. The follo~1ing routines were Hri t-
ten in Fortran IV language.
 32

PROGIW..M ro CALCULATE THE IMPEDANCE OF THE EQUIVALENT CIRCUIT

DIXENSION R(5,3) ,FF(500) ,Zt-1(500) ,ZA(500) ,FA(lO) ,NA.(lO),

lFD(lO)
COMPLEX CI,YT,ZT
CI=(O.O,l.O}
READ,((R(I,J),J=l,3),I=l,5)
READ,N, (FA( I) ,NA(I) ,FD(I), I=l,N)
K=l
DO 10 L=l,N
PI=3.14159
F=FA(L)*lOOO.O
M=NA(L)
FDA=FD(L)*lOOO.O
DO 10 LA=l,M
W=2.0*PI*F
1 YT=O.O
DO 3 I=l,S
3 YT=l.O/(R(I,l)+Vl*R(I,2}*CI+l.O/(W*R(I,3)\-CI))+YT
FF(K)=F/1000.0
ZT=l.O/YT
ZM(K)=CABS(ZT)
ZA (K) =A TAN2 { A.IM.AG ( ZT) , REAL ( ZT) ) *57. 29 58
K=K+l
10 F=F+FDA
K=K-1
~"lRITE ( 3, 50) (FF (I), ZH (I) , ZA (I), I=l ,K)
50 FORMAT(lOX,Fl5.3,Fl0~3,F7.1)
STOP
END
 31

f'ROGRA.H TO CALCULA 'l'E TRANSDUCER IMPEDANCE FUNCTION

REAL L(S)
DIMENSION R(S),C(S)
DOUBLE PRECISION T(5,12),A{5,12),D(5,12) 1
B(l2)
CO.MMON A, T, B
R~~D,{R(I),L(I),C(I),I=l,S)
DO 1 I=l,S
DO 1 J=l,l2
A( I, J) :..::0.0
1 D( I I J) =0 0
A(l,l)=l.O
D(1$>2):.=:C(1)
DO 2 I::;2,5
A(I,l)=l.O/L(I)*C(I))
A(I,2)=R(I)/L(I)
Jl.. ( I I 3 ) = 1 0
2 D(I,2)=l.O/L(I)
DO 3 I==l,l2
3 B(I)=O.O
CALL POL(D,l,2,3,4,5)
CALL COM
CALL POL(D,2,1,3,4,5)
CALL COM
CALL POL(D,3,1,2,4,5)
CALL COM
CALL POL(D,4,1,2,3,5)
CA.LL COM
CALL POL(D,5,1,2,3,4)
CALL COM
CALL POL(A,1,2,3,4,5)
ViRITE(3,10)
10 F'O .."UIJ1\ T( '1 TRANSDUCER INPEDANCE EQUATION' I I' COEFF TC IENT
lS OF POVlERS OI!""' S 014' Z( S) 'I/l.8X 1 11 NUHERA TOR' 1 21X, 0 DENOH
2INATOR 'I)
DO 15 I=l,l2
a==r-1
1 5 V.1 RITE ( 3 , 2 0 ) J 1 r ( 4 , I ) 1 B ( I )
20 FORNAT{' S**',I2,2D30.16/)
STOP
END

SUBROUTINE POL(H,MH,HAl,HA2,HA3,MA4)
DOUBLE PRECISION H(5,12),T(5,12),A(5,12),B(l2)
CO.HHON A, T, B
CALL POLM(H,3,MH,A,3,MAl,NT,l)
CALL POLM( T, NT, 1 ,A, 3 ,J-!A2, NT, 2)
CALL FOL!-1 ( T, NT, 2 ,A, 3, tv".wl\3, NT 1 3)
CALL POLM(T,NT,3,A,3,MA4,NT,4)
RETURN
END
 34

SUBROUTINE POLM(R,NR,MR,S,NS,HS,NT,MT)
DOUBLE PRECISION R(5,12),S(5,12),T(5,12),A(5,12),B(5,1
12)
C0!1110N A, T, B
NT=NR+NS-1
DO 1 I=l,12
1 T(MT,I)=O.O
DO 2 I=l,NR
P=R(MR,I)
DO 2 J=l,NS
2 'l'(HT, I+J-1) =P*S (HS, J) +T(MT, I+J-1)
RETURN
END
 35

PROGRD,H 'IO CALCULATE ROOT-LOCUS POINTS

R&-1\L !<E, KF , KK, KKI , I<KN, N1 ( 10 )

DIHENSION AAP.(l2) ,AA(12) ,AB(l2) ,BA(l2) ,BB{l2) ,BAM(l2),
1COZ(23),COP(23),COD(23),CODR(23),ROOT:K(23),ROOTI(23),
2KKN(1l),Dl(10)
~lRITE { 3,15)
15 FORMAT( '1 ')
READ,(Nl{I),I=l,lO),{Dl(I),I=1,9),RSTRT,RINC,NUHR,
lKI, KF, DEL'fA
NA=l
NB=lO
.Ml\=9
MB=O
BB(1)=1.0
R=RSTRT
DO 113 IR=1,NUMR
CALL SPEC{R,Nl,Dl,AA,AB,BA)
~-1RITE( 3,190) R
190 FORFJAT( 'OBEGINNING OF SET R =' ,H'6.0/
1 ' ROOT LOCUS PROGRAM')
CALL AXB(AA,2,BA,lO,AAA,11)
\'JRI'l,E { 3,16) (AAA (I), I=l, 11), (AB (I), I=1, 11)
16 FORHAT ( 'ONUHERA'IOR' ,5El8.7/lOX,6E18.7/'0DENOiv1INATOR',
15E18.7/12X,6E18.7)
YJ:<I=KI
MA2=M..~+1
105 ~vRITE ( 3, 200)
200 FORl/';A T ( lHO, 4HGAIN, 24X, 5HROOTS/20X, 4HREAL, 14X,
19HIMAGINARY)
KK=I<KI
vJHI'l'E ( 3, 210) KK
210 FO.RI'IA1'(1HO,Ell.4)
DO 106 J=1 ,MA2
BAM ( .J) =I<K*BA ( J)
106 CONTINUE
NA2=NA+1
MA2=!-1A+l
MN2=r,J.\2+NA2-l
C.~~LL AXE ( AA, NA2 , BAM, i-1A2, COZ, HN2)
N'B3=NB+l
MB3=HB+l
MN3=MB3+NB3-l
C'\I...L AXB(AB,NB3,BB,MB3,COP,HN3)
IF{MN3-MN2)107,108,108
107 HN4=MN2
GO TO 109
108 MN4=MN3
109 CALL ADD(COZ,MN2,COP,MN3,COD,MN4)
.t-1N'5~MN4-l
CALL BAIR{COD,MNS)
 36

112 CON'l'INUE
Kl<I==KKI+DELTA
IF(KKI-KF)105,113,ll3
113 R==R+RINC
114 STOP
END
SUBROUTINE AXB(A,N,B,M,C,NN)
DIMENSION A(N),B(M),C{NN)
DO 105 K=1,NN
C(K)=O.O
I=1
100 J=K+1-I
IF(J-1)105,101,101
101 IF(J-M)l02,102,104
102 IF(I-N)103,103,105
103 C(K)=C(K)+A(I)*B(J}
104 I=I+1
GO TO 100
105 CON'I'INUE
RETURN
END
SUBROUTI~~ ADD(A,M,B,N,C,NN)
DIMENSION A(M),B{N),C(NN)
IF(M-N)l00,102,104
100 J=H+1
DO 101 I=J, N
C{I)=B(I)
101 CONTINUE
102 DO 103 I=1,M
103 C{I)=A(I)+B(I)
GO TO 107
104 DO 105 I=1,N
105 C(I)=A(I)+B(I)
J::;.;N+1
DO 106 I=J,l-i
106 C(I)=A(I)
107 RE'I'URN
END
SUBROUTINE P.AIR(A,N)
DIMENSION A{24),B(24),C{24)
I''l=N+1
IF(M-3)500,230,255
255 P=O.O
o=o.o
IF(A3)256,257,256
256 P=A(2)/A(3)
Q=A(1)/A(3)
2 57 NA 'l"=M- 2
B(M)=A(M)
180 B(M-1)=A(M-l)-P*B(M)
 37

C(M)==B(M)
C(M-1)==B(M-1)-P*C(M)
DO 250 K==2,NAT
N3==!1-K+l
N4=H-K+2
N2=M-K
B(N2)=A(N2)-P*B(N3)-Q*B(N4)
250 C(N2)=B(N2)-P*C(N3)-Q*C(N4)
B{1)=A(1)-Q*B(3)
DELP=(B(2)*(C(3)+P*C(4))-B(1)*C(4})
DELP=DELP/(C(3)*(C(3)+P*C(4))-(C(2)-B(2)+P*C(3))*C(4))
DELQ=(C(3)*B(1)-{C(2)-B{2)+P*C(3))*B(2))
DELQ=DELQ/(C{3)*(C(3)+P*C(4))-(C(2)-B(2)+P*C(3))*C(4))
P==P+DELP
Q==Q+DELQ
IF(ABS(DELP)-1.E-5)260,260,180
260 IF(ABS(DELQ)-1.E-5)265,265,180
265 E==l.O
235 IF(P*P-4.0*Q*E)270,275,275
270 X==-P/(2.0*E)
YA==SQRT(ABS(P*P-4.0*Q*E))/(2.0*E)
YB=-YA
WRITE(3,105)X,YA
~lRITE( 3,105 )X, YB
GO TO 300
275 X=-P/(2.0*E)+SQRT(P*P-4.0*Q*E)/(2.0*E)
Y=O.O
v1RITE (3, 105) X, Y
X==-P/(2.0*E)-SQRT(P*P-4.0*Q*E)/(2.0*E)
WRITS(3,105)X,Y
300 11=M-2
IF(M-2)500,400,305
305 DO 240 I=l,M
240 A(I)=--=B(I+2)
IF(M-3)230,230,255
230 E=A(3)
P=A(2)
Q==A(l)
GO TO 235
400 X==-B{3)/B(4)
Y=O.O
~-J.;~rrE ( 3,105) x, Y
500 COtY.CINUE
105 FOH..t-iAT(14X,2E18.6)
RETURN
END
SUBROUTINE SPEC(R,N1,Dl,D4,AB,BA)
f{l~~AL Nl ( 10) , N2 ( 3) , N3 ( 3)
DIMENSION D1{10),D4(12),AB(l2),BA(l2),AC1(20),
1AC2(20),AC3(20)
N2(1)=0.0
 38

N2(:2)~1SSO.OE-6
N3 ( l. ) :;1 0 /R
D4(l.)::::O.O
D4.(2)::::7oo.o
CA~L
ADP(N2,2,N3,1,AC1,2)
CA~L
AXB(Dl,9,D4,2,AC2,10)
CA~L
AXB(AC1,2,AC2,10,AC3,11)
CA~L
b%B(Nl,lO,D4,2,AC1,11)
CA~L
ADP(AC1,11,AC3,11,AC2,11)
DO 2 :t::::l,9
2 AC3(!)::::4.0*Dl(I)
CA~L ADP(AC2,11,AC3,9,AB,11)
PO 1 J>-=1, 10
1 P.Cl(:t):::-Nl(I)
CALL ~DD(N2,2,N3,l,AC2,2)
C~LL AXB(D1,9,AC2,2,AC3,10)
CALL .ADD ( A.C 3 , 1 0 , ACl , 1 0 , BA, 10 )
RB~tJRN
END

~:t
'1
.,l'"
 39

APPENDIX D

TRAl\SDUCER IMPEDANCE DATA

Frequency Voltage Current Phase Angle

(KHz) (volts p-p) (rna p-p) (degrees)
10 20.0 2.2 -90
20 20.0 4.2 -90
40 20.0 8.8 -90
50 20.0 14 -85
52.7 20.0 27 -46
53.8 20.0 10 +10
54.2 20.0 4.0 -32
55 20 .. 0 5.6 -75
60 20.0 10 -85
70 20.0 13 -85
80 20.0 15.5 -85
90 20.0 17 -85
100 20.0 20 -85
110 20.0 20 -85
120 20.0 28 -80
125 20.0 33 -70
126 20.0 28 -55
126.7 20.0 21.5 -65
130 20.0 25.5 -80
140 20.0 32 -75
141.7 20.0 30 .. 75
150 20.0 36 -75
155 20.0 34 -75
160 20.0 28 -70
170 20.0 32 -75
1.80 20.0 35 -75
190 20.0 44 .. 73
200 20.0 52.4 -60
204 20.0 44 -35

~'{ l
 40

Frequency Voltage Current Phase Angle

(KHz) (volts p-p) (rna p-p) (degrees)
207 20.0 33 -45
210 20.0 36 -45
220 20.0 24 -65
230 20.0 31 -80
250 20.0 38 -80
300 20.0 so -80
400 20.0 72 -80
500 20.0 92 -80
 41

'l'RANSDUCER HECHANICAL RESPONSE DATA

Voltage into transducer #1 10 volts p-p

Voltage values below are across transducer #2, loaded
with 1000 or.ms and spaced 2 feet from transducer #1.

Frequency Voltage
(KHz) (volts p-p)
50 0.02
53 0.22
55 0.04
100 0.02
125 0.09
150 0.15
170 0.40
195 1.00
199.5 1.60
208 0.60
214 o.so
220 0.50
230 0.10
300 0.06
 42

VITA

The author vlas born on February 26, 194-1 in Kansas

City, !1issouri. There, he graduated from Southeast High
School in 1962. In the fall of that year he enrolled at the
University of Missouri at Rolla and received the Bachelor of
Science degree in Electrical Engineering in January, 1966.

Mr. Watke has been enrolled in the Graduate School of

the University of Missouri at Rolla since September, 1966.
During that time he has held an !IDEA Title IV fellowship.

The author is a member of Sigma Xi, Pr'i Kappa Phi, Tau

Beta Pi, Eta Kappa Nn and Phi Eta Sigma.

132~119