Você está na página 1de 48

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.
Paper 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/

-----A
THESIS

:132919

submitt:ed to ::.he faculty of


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 deterth~

mines

f:c~::quency

of oscillation

advantage of this met.hod is the ability of


the frc~u~ncy of oscillation to change with changes in the
transd':.1cer mechanical resonant frequency.
This maintains
..r:.,.

<J~finite

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 assistance 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

AC'r

l~BSTH..

ILLUSTRATIONS

INTRODUCTION

ACKNC:vJLEDGEHENrs

LIS'r

oe
I.

rRANSDUCER

II.

CHAP~CTERISTICS

INI'riAL CONS IDEP.....~ 'I IONS


FINAL CIRCUIT

I'.J.

v.

iii
v

6
10

12

ANALYSIS


OSCILLA'l"'OR DE.SCRIP'I'ION

..

12
12
15

18

21

GENEF-'.AL DESCRIPTION

PULSE FORMING NE'IWORK

..

BIBLIOGRAPHY


APPENDIX A:
CIRCUIT PARTS LIST

RESULTS
CONCLUSIONS

VI.

ii

TRANSDUCER EQUIVALENT CIRCUIT

III.

25
26
27

29

APPENDIX C:

TRANSDUCER EQUIVALEN'r CIRCUI'r


DERIVATION

CO!v:PUTER PROGR.ll.NS

31

APP:t!NDIX D:

TR.~l\SDUCER

VI'rA

APPE:~DIX

Bs

DATA

39

42

LIST OF ILLUSTRATIONS
Figure

Page

1.

Transducer impedance characteristic

2.

Combined mechanical response of transmitting


and receiving transducers

3.

Transducer equivalent circuit


Equivalent circuit impedance characteristic

Block diagram of a simple ultrasonic generator

8
9
11

A method of effectively cancelling the


transducer shunt capacitance

7.

Block diagram of oscillator circuit

8.

Final circuit block diagram adapted to


root-locus analysis

4.
5.
6.

9.
10.

Root-locus of final circuit


Oscillator schematic diagram

11.
12.
13.

Pulse network schematic diagram


Typical output pulse waveform

11

13

13

16

Typical waveform showing pulse spacing

17
19
22
23

I.
An undesi:cable
tran~:;ducers

INTRODUCTION

prcp~rty

of crystal type ultrasonic

is the change of ch<-1.:.aeter is tics which takes

place when t11ey are subjected to temperature or loading variations.

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:hdracteristic 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 mccho.nical r1::sonant frequency.

This allovrs optin:n.t.'l1 efficiency

to be realized under all conditions.


Existing methods of ultrasonic pulse generation usually invclve driving the transducer Hith a fixed
keyed c!.::ciJ.lator.

frequ~ncy

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
frequency determining element.

th~

main

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.


quencit.~s

Four resonant frc-

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 characteristics, 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-

a sine H~ve generator while the response of a second tx-ansduc1z,r ~paced tvro feet a-vray Has obser.<_,'(:~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 frequency.
Fig. 2 sho\vS the responsrz- obtained in method three.
ducer va:3 d:rivcn

~Iith

'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.

75

Impedance
Magnitude

70

.........
,.Q

u
..........

65

-cv

'U
::3
+'
rt
~

0'1

60

~
Cl)

+10

ro

.........

'0
<ll

0..

s
1-i

55

Ol
Cl)
Q)

-10

tJ'l

Cl)

-30

r-1

-50

'0

Phase
Angle

Cl)

so

tJ'l

s::

~
(l)

Ol

-70

ro

..c
At
45

-90

0
CVl

tO

['

0
0
r-1

0
0
N

0
0
CVl

0
0

Frequency (KHz)

Fig. 1.

Transducer impedance characteristic.

40

0 db corresponds
to the 50 KHz
response.
32

24
_..._
,Q

rtJ
..._,
(!)

til

s:

16

a
til

Q)

IY.

+----+----+-+-+-

-8
0

0
CV)

--r

U)

"

I
0
0

r-f

1-

0
0

0
0

CV)

Frequency (KHz)

Fig. 2.

Cornbined :mechanical response of transmitting and receiving transducers.

0
0

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 frequency of operation is therefore 200 KHz.

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 characteristics.
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 resonance to another.
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.
1~e

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 im6
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.

N(s)

Z(s)

N(s)

= s8

D(s)

+ 0.159138 5 7 + 3.30227 5 6

+ 0.337821 5 5 + 3.47206 s 4 + 0.182954 s 3


+ 1.29058 5

D(s)

+ 0.0191176 s + 0.104305

3
8
= 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

White

'-l_l_
0\

ro

'-l

1-'

Q)

Q)

1.0

'-l

VI

1-'
1.0
0'\

1-'
N

~1

1-'

61
~T

Shield

'-l
0\

1-'

U1

trr

1.0

tT

Q)

_1_

1.0

Black
.Resistors in ohms, inductors
in mh, capacitors in pf.

Fig. 3.

Transducer equivalent circuit.


CD

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 generate 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 amplifier 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
shift
network

Amplifier

Fig. 5.

Block diagram of a simple ultrasonic generator.

Output
Transformer

Fig .. 6.

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 origin to which this pole converges. Therefore, if the effect
of C0 were removed, a circuit might be found that would operate 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 supply reactive volt amperes to capacitor c 1 Even though c 1
dissipates no power, the transistor dissipation is increased. 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 resonance 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

z1

Amplifier
gain

=k

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

-E

Fig. 7.

Block diagram of oscillator circuit.

Fig. 8.

Final circuit block diagram


adapted to root-locus analysis.

14

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

z2 z

or

:: ki

err

I err = Il -

!2

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

H-

-I err
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

+ 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


and a frequency of 213 KHz.
The high frequency
buted to not knowing the exact value for C0 and
that small errors in the equivalent circuit for
ducer could cause a large error.

than unity
is attrithe fact
the trans-

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 combi1
nation of the transducer capacitance and c 3
In use, the transducer would be connected also to a receiver so that it is desirable for one terminal to be connected to ground.
This is made possible by taking the feedback 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

"

r-1

II

II

II

0
lO

lO

.-1

"

II

II

.t.>~

- 1. 50

""'

-;

1.25

1.oo
X

0.75

s normalized
by 10 6

o.so

0.25

Real
axis

-+-

"

,,

lO

0
I

Fig. 9.

II

r-1
0

o.o

I
\.0
0
0

0
0

0
I

0
0
0

Root-locus of final circuit.

T2

Transducer

Dl

AI

'''
I

or
I

II II

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

,..._, Ll

D3
D2

I I

c1
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 bias1
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 regeneration 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 occurs, capacitor c 1 begins to charge to a higher voltage.
When this voltage tries to exceed +0.7 volt, the input current 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 controlled 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 configura5
tion. An advantage of this configuration is that no current
is draHn by e5.ther transistor between pulses.

+6

R3
R2

<

Q4

Q3

c4

<..

<

Ra

Rg

Hp

cs

R6

as

R7

To
oscillator

Fig. ll.

Pulse network schematic diagram.

......
\0

20

vlhen the base of 0

receives a negative pulse from the


unijunct.:.ion transistor, 0 turns on and the voltage at its
4
collector rises toward +6 volts. This in turn starts transistor 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.
5
9
4

output is taken from the junction of resistors R4 and


Rs '\vhich are chosen to supply sufficient current to the oscillator 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
7

delivered to
1 2
transducer '

po~,rer

frequency

pulse width
pulse spacing

rna maximum
rna average

0.75 watt peak


0.015 watt average
199

KHz

Oo8

msec

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 operating 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
1 msec after the beginning of the pulse.
This restricts
the minimum range of the device to about 2.5 feet.
tra.n~:!ducer

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

s_?acing of 45 msec is sufficient for a maximum ran(JC of approximately 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
Horizontal

Fig . 12.

20 volts/div .
20 usec/div.

Typical output pulse 't-laveform .

23

vertica l
nori zonta l

Fig . 13 .

20 volts/div .
5 msec/div .

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 capacitance 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 connected 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 operation.

25

VI.

CONCLUSIONS

The t:hree main points of interest regarding the performance of the circuit are&
1.

The transducer helps determine th9 frequency


of oscillation allovTing the unit to track
changes in the transducer's resonant frequency.

2.

The circuit efficiency is high with an average 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 immersing 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 toTvlard 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.

MASON,

w.

I1ITCHELL,

P.

(1950) Piezoelectric C1ystals and Their


Application to Ultrasonics. Van Nostrand,
New York, 508 P
(1942) Electromechanical Tr~nsducers and Wave
Filters. Van Nostrand, New York, pp. 195215.

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
R2
R3
R4, Rs
R6
R7

Ra
Rg

2.2
7.5 K
300
200
3.3 K
2.2 K
1.0 K
330

Capacitors (uf except as noted)


10 e1ectro1ytlc
cl
55 electrolytic
c2
c3

1390

c4

6.8

cs

pf
electrolytic
electrolyt.ic

Inductors
700

uh

Transj_stors
Fairchild
RCA
TI

2N3569
2N2160
2N3703

TI

1N658

uniju:nction

28

'rrc-.nsformers
Primary - 30 turns; secondary - 10
turns each half. bifiler wound on an
Indiana General core no. CF102 of
'f

0-6 type material.


Primary - 9 turns each half. bifiler
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 finding the equivalent circuit, a set of starting values was obtained 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

Z(s) -

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

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


irnpedance minima and the even subscripted w's are the frequencies of the impedance maxima.
C , the total shunt capacitance, was found by evaluo
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 resistors 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 resonances. Increasing c 0 decreases this spacing.

2.

Increasing L 1 and decreasing c 1 by the same


factor - This increases the Q of the particular resonance causing larger impedance variations.

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 believed 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 tten 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)
CO.MMON A, T, B

B(l2)

R~~D,{R(I),L(I),C(I),I=l,S)

DO 1
DO 1

I=l,S
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

~vRITE ( 3, 200)
200 FORl/';A T ( lHO, 4HGAIN, 24X, 5HROOTS/20X, 4HREAL, 14X,
19HIMAGINARY)

105

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

100
101
102
103
104
105

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
J=K+1-I
IF(J-1)105,101,101
IF(J-M)l02,102,104
IF(I-N)103,103,105
C(K)=C(K)+A(I)*B(J}
I=I+1
GO TO 100
CON'I'INUE
RETURN

END
SUBROUTI~~

100
101
102
103
104
105

ADD(A,M,B,N,C,NN)
DIMENSION A(M),B{N),C(NN)
IF(M-N)l00,102,104
J=H+1
DO 101 I=J, N
C{I)=B(I)
CONTINUE
DO 103 I=1,M
C{I)=A(I)+B(I)
GO TO 107
DO 105 I=1,N
C(I)=A(I)+B(I)
J::;.;N+1
DO 106 I=J,l-i
C(I)=A(I)

106
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

250

260
265
235
270

275

300
305
240
230

400

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)
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
IF(ABS(DELQ)-1.E-5)265,265,180
E==l.O
IF(P*P-4.0*Q*E)270,275,275
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
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
11=M-2
IF(M-2)500,400,305
DO 240 I=l,M
A(I)=--=B(I+2)
IF(M-3)230,230,255
E=A(3)
P=A(2)
Q==A(l)
GO TO 235
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
CA~L
CA~L
CA~L
CA~L

ADP(N2,2,N3,1,AC1,2)
AXB(Dl,9,D4,2,AC2,10)
AXB(AC1,2,AC2,10,AC3,11)
b%B(Nl,lO,D4,2,AC1,11)
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
(KHz)

Voltage
(volts p-p)

Current
(rna p-p)

10

20.0

2.2

20

20.0

4.2

40
50
52.7
53.8

20.0
20.0
20.0
20.0

54.2

20.0
20 .. 0

8.8
14
27
10
4.0
5.6
10
13

55
60

70
80
90
100
110
120
125
126
126.7
130
140
141.7
150
155
160
170
1.80
190
200
204

20.0
20.0
20.0
20.0
20.0
20.0
20.0
20.0
20.0
20.0
20.0
20.0
20.0
20.0
20.0
20.0
20.0
20.0

15.5
17
20
20
28
33
28
21.5
25.5
32
30
36
34
28
32
35

20.0
20.0

44
52.4

20.0

44

Phase Angle
(degrees)

-90
-90
-90
-85
-46
+10
-32
-75
-85
-85
-85
-85
-85
-85
-80

-70
-55
-65
-80
-75
.. 75
-75

-75
-70
-75
-75
.. 73
-60
-35
c

~'{

40

Frequency
(KHz)

Voltage
(volts p-p)

Current
(rna p-p)
33
36
24
31
38

Phase Angle
(degrees)

207
210

20.0

220

20.0

230

20.0

250

20.0

300

20.0

so

400

20.0

72

-80
-80
-80

500

20.0

92

-80

20.0

-45
-45
-65
-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
(KHz)

Voltage
(volts p-p)

50

0.02

53

0.22

55

0.04

100

0.02
0.09

125
170

0.15
0.40

195
199.5

1.00
1.60

208

0.60

214

o.so

220
230

0.50
0.10

300

0.06

150

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

Você também pode gostar