13L «45» 3&5
"u 381* 9 a. o .
xx .. .5 1..“ .. 2. ..
o“ .. a . 8 a _ in» 1.9
v.3 3m . . 5.». h
‘0“ 4-. q .4. n .3... ‘5”:
m. . . é . x. g...
“L '3 o s o u .a
9- .. so ‘ .3
s a. -A
.5... ....u w. r...“ .Y..
rb ...\.. «a . a“,
l . a .w. . .
$6.... X3 - ”U u _ V. ‘33

7 AV "flu. « h . L. :3.
. 9‘ . . 4 M Jk n fi.‘

‘v I 'C“ 0r: - o O '0
an. 4&9 \ .
{.0 o 0. ..h “a
a“ fix Mia» luw m V): v.‘
.1 a a 4 . .. .3.
u r: k 8. Ab , 3.
«iv “NJ! «to h c .1“ .5.»
‘ ‘
{to bars.

H:_:__:_itSEE::_:__:_E___:_E__ mwm

 

W ‘11mmmijlngmanu

 

LIBRARY
Mlchigan State
University

 

 

 

This is to certify that the

thesis entitled

A Radio-Frequency Pulse Generator for

UltraSOnic Studies
presented by

Walter G. Meyer

has been accepted towards fulfillment
of the requirements for

15.8. degreem P sites

in /A. ;&1LobAA¢ADHAa

Major professor

 

Date July 29: 1955

0-169

___.____.___._. _ _ ML. _ _ ._ _ ‘_ ___ _. _. _ __, A__ _ __ _.

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINE return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

we WW“

A RADIO—-FREQUENCY PULSE GENERATOR
FOR ULTRASONIC STUDIES

By
Walter Georg Mayer

A THESIS

Submitted to the School of Graduate Studies of Michigan
State University of Agriculture and Applied Science
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE
Department of Physics
1955

ACKNOWLEDGEMENT

I should like to thank Dr. E. A. Hiedemann,
who suggested the construction of the apparatus,
for his guidance throughout the experiments and
for his willingness to offer helpful suggestions

(MW

and advice.

A RADIO- FREQUENCY PULSE GENERATOR
FOR ULTRASONIC STUDIES

BY
'Walter Georg mayer

AN ABSTRACT

Submitted to the School of Graduate Studies of Michigan
State University of Agriculture and Applied Science
in partial fulfillment of the requirements
for the degree of

MASTER.OF SCIENCE

Department of Physics
1955

I ch )h- QKLLDLLAA~<buvvv

Approved-

This work describes the design-and construction of
a r-f pulse generator which can be used with either one
or two transducers to measure primarily sound velocities
and absorptions in liquids and solids. The pulse length,
repetition rate, frequency, and amplitude are variable.
Circuit diagrams and operational details are included.

A discussion of Operation principles is followed by
a report on velocity measurements made in three fatty acids
'which were suspected of showing a dispersion of sound vel-
ocity as a function of frequency and sound intensity.

EXperimental evidence is offered indicating that laur-
ic, palmitic, and stearic acid show a velocity dispersion
as a function Of sound intensity but net as a function of
tetal energy radiated into the liquid acid. The diapersion
occurs only within a range of about 15° above the respec-

tive melting points.

I. INTRODUCTION

The measurement of physical properties of materials
by ultrasonic waves has become a valuable tool in the
study of matter. There exists a number of techniques by
which ultrasonic waves can be generated and detected; how-
ever, the desired information concerning the material un-
der investigation determines to a large extent the useful-
ness of one or the other method.

The present work concerns itself only with a tech-
nique by which sound energy is propagated in pulses of a
few microseconds duration:

Progress in electronics made it possible to adapt
principles of pulse circuits to ultrasonics, and in 1940
E. Hiedemann and H. Freundl suggested an arrangement in
which a transducer was to be driven by a short, strong
electrical pulse. This pulse, transformed into ultrasonic
energy, was to be received by another crystal, amplified,
and used to activate the first transducer again. The time
lag, introduced by the medium between the two crystals,
could be measured, and the "apparent" sound velocity could
be computed from it. If the ultrasonic pulse did not meet
any obstacles, flaws, cracks, etc., in the material, the
"apparent" sound velocity was identical with the actual

sound velocity in the material. If the pulse was disturbed

by a flaw, the "apparent" sound velocity was lower than
the velocity characteristic for the material. This allow-
ed ultrasonic testing of materials.

A different procedure was patented by Firestone2 in
1942. It proved to be highly efficient and could later
be developed into the well known "Reflectoscope". In
Firestone's method a pulse travels through the sample to
be tested and is reflected by a flaw. The initial and
the reflected pulse are observed on a CRO. From the dis-
tance on the screen between the two signals and from the
time scale the location of the flaw can be found.

Extensive work on Radar and Sonar during World War II
produced more refined circuits and more accurate timing
devices which were used to some extent for ultrasonic in-
vestigations at the time of their develOpment. Teeter3
points out the various possibilities of such instruments,
and his findings are in good agreement with results obtain-
ed by other methods; Ivey et al describe similar equip-
ment for various measurementsA.

After the war a great number cf electronic improve-
ments were made available so that it became possible to
even manufacture small ultrasonic pulse generators for com-
mercial purposes. Those are mainly used for testing mater-

ials for flaws, alloys for grain size, boilers and metal

pipes for weak spots, and for similar test35’6.

Since the pulse technique presents a convenient way
to make velocity and absorption measurements in liquids
and solids in a very simple and accurate manner7, the des-
ign and construction of the present instrument was under-
taken, and an attempt was made to include as many features

as possible.

II. PRINCIPLES OF OPERATION

In Fig. l a block diagram indicates the methods em—
ployed in producing and receiving the electric pulses driv-
ing the transducer and emitted by the receiving crystal.

A sharp positive pulse from the trigger circuit of the
DuMbnt Type 256-F OscilloscOpe activates the sweep cf the
sac and fires the multivibratOr. The beginning bf a rect-
angular pulse, given off by the multivibrator as the result
of the trigger pulse, is set by the repetition rate er the
CRO; its duration is adjusted within the multivibrator.

The positive-going signal from the multivibrator is inver-
ted and used to gate a variable frequency oscillator which
in turn emits r-f oscillations as long as the gate is pres-
ent. This amounts to a modulation of the rectangular pulse
by a frequency, set to coincide with that of the crystal in
use. A buffer stage couples the gated v.f.o. to the pewer
amplifier which increases the amplitude of the signal and
can be used to multiply the frequency by one, two or three.
Through a coupling device the r-f pulse is transmitted to
the transducer.

By disconnecting the gated v.f.o. and placing a c.w.
v.f.o. ahead of the buffer, the unit can be used as a con-
tinuous wave generator.

In the case of transmission Operation two antennae

couple the transmitter to the receiver, in case of reflec-

4

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

cao
INVERTER “Um- announce DCMODUL.
“BRA-[OR «- TRIGGER
earn) 7 /
a-r. BUFFER A .4ng fl//fl TUNER
oscmarorz P V ,
//
a. Transmission Method.
Imo—
MULTI~
mvamzn , VIBQATOR AMPLIFIER Der/roam.
mesa:
GATED
osaumop B ER AMPLIFIER TUNER

 

 

 

 

 

 

 

 

 

b. Reflection Method.

 

 

 

Fig. 1. Block Diagrams

 

 

 

 

 

 

tion Operation transmitter and receiver are connected dir-
ectly.

The tuner of the receiver is set to the frequency Of
the receiving crystal which is the same as the frequency
of the transmitting transducer. After the signal is det-
ected by the tuner the carrier frequency is taken Out of
the signal in the demodulating section, and only the amp-
litude and general shape of the signal, its envelope,
reaches the final amplificatiOn stage. From there the sig-
nal is fed into the "Video Input" er the CRO.

Depending on the velocity of sound in the medium bet-
ween the crystals and their seperation, the time elapsed
between the departure of the sound pulse from the trans-
mitting transducer and the arrival at the second crystal
is indicated on the screen of the CRO as a vertical deflec-
tion of the trace. The travel time of the sound through
the medium can be read on a calibrated scale at the CRO,
and depending on the choice of sweep length, this calib-
ration is accurate to 10"6 or 10"7 second, respectively.

It is possible to estimate to the nearest 10'”7 or 10-8
second, respectively.

The amplitude of deflections gives an indicatiOn of
losses the sound experiences in the medium.

For applications of reflectibn operations the receiv-

ing transducer is replaced by a reflecting surface; the

remaining crystal serves as transmitting and receiving trans-
ducer, the sound traveling twice the distance. For accur-
ate absorption measurements the transmitter has to be dis-
connected from both the transducer and receiver immediately
after the end of the activating r-f pulse. The reasons for

such a switching arrangement will be pdinted out in Sec-
tion V.

III. ELECTRONIC EQUIPMENT

The functibns of the various stages in the operation
of the apparatus have been pointed out in Section II. The
arrangement of the components is shdwn in Fig. 2, the com-
plete circuit diagrams in Fig. 9 - 12.

A. Power Supplies

The pewer supply for the pulse generator and the
transmitter consists of two sections, one for high voltage,
and one for B plus together with 6.3 volts for filament
supply, all of which is mounted on one chassis and construc-
ted according to conventional methods. Cheke inputs are
employed, a swinging choke stabilizes the high voltage, and
a variac is used to adjust the potential of the high volt-
age supply.

A separate power supply for the receiver is mounted
on the receiver chassis, its cOnstruction also follOws stan-
dard methods.

Negative biasing voltage is obtained by tapping a
bleeder connected across the last cf three RC filters which
are operated from a half wave selenium rectifier.

B. Triggered Multivibrator

In principle the pulse generator is a one-shot-multi-

vibrator activated by a positive trigger pulse from the

DuMont Type 256-F Range OscillOscope whose variable repet-

 

 

 

F18. 2.

I

 

   
 

C R O

 

 

 

Receiver

 

 

Multi-
vibrator,

Oscillator

 

Power
Amplifier

 

Fewer
Supplies

 

Electronic EQH1pment

l

zummww

k .

V1:

"‘ 1|

\

     

It.
..

ition rate determines the frequency with which rectangular
pulses are emitted by the multivibrator. The somewhat in-
volved circuit of the oscilloscope is described in the op-
eration manual supplied with the instrument.

The duration of the rectangular positive pulses from
the multivibrator is fixed by the RC-time of the grid bias
resistor and C9 and 010. Contrary to ordinary multivibrat-
era8 the duration of the negative part of the wave train can
not be changed by adjusting the grid resistor or the coup-
ling capacitors Of the second half of V4 sinbe this portion
of the cycle ends only with the arrival of a trigger pulse
from the oscilloscope. Thus C7, C8, and R9 provide the
proper bias on the grid of the second half of V4, causing
the rectangular pulses to have a flat top and a steep lead-
ing edge (Fig. 4).

R3 to R6 act as biasing resistors for the grid of the
first half of V#, keeping it below cut-off for the plate
potential employed, until a sufficiently positive trigger
pulse is applied, allowing the first half of V4 to conduct
and the same set of resistOrs to be used in determining the
duration of the rectangular pulse, which is applied to the
next stage by a cathode fellower. This method of coupling,
'whereby placing the two cathodes of V# at different oper-
ating potentials does not affect the performance of the

nualtivibrator, provides an additional amount of stability

10

which is desirable since the ratio of conducting to quies-
cent periods within one cycle is to be variable from about
1:10 to less than 1:1000 (Fig. 6).

Since it is required to supply to the transducer short
electrical pulses of controllable duration and energy level,
a number of methOds fer further medulation of the rectangular
pulse from the multivibrator are available. Probably the
simplest way Of exciting the transducer is to supply a suff-
iciently high DC pulse to the crystal. If the rise and fall
times of such a rectangular pulse are of the order of 10"7
second or less, the crystal will vibrate at its resonant
frequency. This is due to the fact that such a pulse con-
tains an infinite number of frequencies, but Fourier Anal-
ysis indicates that any one frequency contributes only a
fraction of the total energy present. Since the transducer
will be activated only in its principal mode of vibrations,
a large part of the electrical energy applied to the crystal
by a rectangular DC pulse will not be used for the product-
ion of an ultrasonic wave. It can be seen that this method
is net extremely efficient, although its application has
yielded good results9.

C. RPF GeneratOr

A much higher sound level can be obtained with the same
electrical input if the electrical energy is prepagated at
the frequency corresponding to the principal mode of vib-

rations of the crystal used fer the production of ultra-
sonic energy. There are two general approaches; either
the required frequency is generated continuously and the
output Of the generatOr connected to the transducer by means
Of an electronic switch synchronized by pulses Of desired
length; Or the pulses are emplOyed to regulate the time Of
Operation of the frequency generator, which, in this case,
is connected to the transducer without an electronic switch.
In both cases the transducer receives predetermined pulses

of one specific frequency (Fig. 3).

 

 

 

 

 

HoHNlbroior Mfi
I‘I‘l'l‘l‘l'l‘lvl' f”
. ENS-'3: ' ‘9:
W, -”——“— a ‘

 

 

 

_f1___11_ .
lawman} Lr imH ii fl
Fig. 3. Methods of Pulse Generation

Since fOr purposes of measurements the ultrasonic pulse
has to be transformed into electrical energy by means Of a
crystal whose electrical output is Of the same frequency as
that of the incident pulse, the prOcess Of amplification Of
Only the electrical crystal Output becomes difficult. With

'the frequency generator Operating On c.w. of about 100 watt

l2

there is enough radiation reaching the receiver at all
times to drown the incoming pulse in a high signal-to-
”noise ratio. Proper shielding of the receiver and the
transmitter becomes involved and the approach of an inter-
mittently operating r-f generator is preferable.
Oscillations can be produced in a number of ways, com-
monly by placing an inductance-capacitance network either
in the plate circuit or between cathode and ground of the
oscillator tube. Other forms of oscillators (phase shift,
grid-tuned, etc.) are not conveniently adaptable to modul-
ation of pulses of short duration. A shock-excited (rin-
ging) oscillator was selected since this form is now gener-
ally used in radar and similar applications involving pulse
techniqueslo. However, it was found that applying the gate
pulse, which is a rectangular positive-going pulse from the
multivibrator (Fig. 4) inverted by V5 into a negative—going

pulse (Fig. 5). to a simple one-tube ringing oscillator

EIIII
'IIIIIIIIII
‘IIIWI-I-'

V
l
I
I
I

.IIIIIIIII
IIIIIIIIII

I3I
I
AI

I
1' + a
llllllllllllE

”W

 

Fig. 4. Positive Gate

13

results in a distorted signal with an extremely long rise
time as its main undesirable characteristic. As pointed
out by Elmore and Sandsll, this difficulty is bound to
arise in such an arrangement when frequencies higher than
2 Mc are required. The problem of maintaining a fairly
stable voltage level of the oscillator output presents it-
self, and no solution to it can be found within the cir—
cuit of the oscillator if one considers that the voltage
level is the product of frequency, current, and inductance
of the oscillator. Although the frequency is inversely
preportional to the square root of the inductance, and the
current can be varied within limits, it becomes difficult
to keep the voltage reasonably constant through an extend-
ed range of frequencies.

Experiments with a slightly altered form of the ring-
ing oscillator described by Elmore and Sands in combination
with a thyratron circuit used by Easton12, also slightly
changed to be useful in this application, have proved to be
so much more suitable than other circuits tested that this
combination was incorporated in the apparatus.

Two sets of oscillator coils can be placed in the cir-
cuit by a switching arrangement which also varies the induct-
ance in steps of convenient values. Two other inductances
can be coupled to the oscillator coils by the same switch in

such a manner that either L5 and L6 or L7 and L8 are in the

14

circuit. The thyratron fires after the trigger pulse from
the CRO arrives at its grid which happens before the negat-
ive gate pulse arrives at the oscillator; this difference
is due to the delay the gating pulse experiences through the
preceding circuit components. Thus when the ringing Oscil-
lator is shocked into operation the damping of its sinusoid-
al output is essentially prevented by a transfer of dis-
charge current through the cathode coil of the thyratron to
the oscillator tank circuit. With the proper values of the
cathode resistors of V6 and V8 it is possible to obtain r—f
'pulses with flat tops and rise and fall times of less than
1 microsecond. The voltage level of these pulses is about
60 volts maximum, decreasing slightly with a decrease in
pulse length. The energy delivered by the oscillator is
sufficient to drive a transducer directly.
D. Buffer and Power Amplifier

In order to expand the apparatus so that it can be
used easily for a one crystal reflection instrument it was
felt that a buffer stage and a power amplifier should be
added. The circuit used is similar to a transmitter des-
ignl3 which also allows installment of a variable frequency
oscillator preceding the buffer and power amplification
stages, enabling the operator to use the apparatus as a

standard high frequency generator (c.w.)

Operating the crystal from the modulating oscillator
yields a sound intensity great enough to make measurements
in liquids which present absorptions of relatively small
magnitudes, and in which the path of the sound is not too
long. The addition of a power amplification stage makes
it possible to overcome these limitations to some extent.
Furthermore, the correct choice of the output tank circuit
results in a frequency multiplier with which the frequency
modulating the output pulse can be increased by a factor
of two or three without decreasing the amplitude of the
pulses appreciably.

The transducer is connected to the output of the power
amplifier by a coaxial transmission line terminated in a
link coupling around the inductance of the output tank cir-
cuit.

Since during transmission Operation of the unit there
is no direct electrical connection between the transmitting
and receiving stages which would indicate on the screen of
the CEO the time of transmission of the pulse to the crys-
tal, a vertical deflection of the sweep is produced by feed—
ing a part of the signal from the transmitter into an ant—
enna which radiates enough energy to be intercepted by an-
other antenna leading into the receiver and to the vertic-'

a1 deflection of the CRO.

l6

Fig. 6 - 8 show the modulated pulse with which the
transducer is driven. The pulse repetition rate is BOO/sec,
modulating frequency 2.595 Mc, pulse length 30 sec'6.

In Fig. 6 the trace extends over 4500 microseconds,

 

Fig. 6. Trace #500 sec"6
thus two pulses arrive at the transducer. In Fig. 7 the

first 100 microseconds are shown, and in Fig. 8 the trace

MAM/mu

l

Jl‘VVVVUU‘V’Vi/Vl‘

 

Fig. 7. Trace 100 sec"6 Fig. 8. Trace 4 sec'6
pictured consists of the center portion of the pulse, the
time elapsed from beginning to end of the trace being

4 microseconds.

l7

E. Receiver

In principle the receiver is a stagger-tuned amplifier
with two inputs, one of which is the antenna whose purpose
was described in part D, the other the output of the receiv-
ing crystal. An impedance matching coaxial transmission
line connects the crystal with the input of an attenuation
box with an internal impedance of 70 ohms, connected by an—
other coaxial line to a terminating resistance of the same
value. From this resistor the signal is fed into the first
stage of the receiver.

The attenuation box has no purpose if measurements of
sound velocity are to be made, yet it is essential for cal-
culations of sound absorption. Its Operation will be des—
cribed in Section V.

The first two stages of the receiver are used for det-
ecting the signal by stepwise changing the ganged induct-
ances L11 and L12 and tuning the ganged capacitors C41 and
C49. The difference in resonant frequencies of the two
tuning networks was chosen to be about 10 Kc, variable to
a very limited extent by trimmer capacitors in parallel
with the tuning capacitors. This narrow bandwidth, the

result of stagger-tuning14

, allows an accurate tuning of
the receiver, and still compensates for small variations
of input frequency. The following pentode rectification

stage demodulates the signal and the remaining envelope

18

is fed into an inverter which is connected to a driver
stage. Two output tubes in parallel constitute the final
stage of the receiver. The output is not driven in push-
pull, the parallel tubes insure merely Operation below max-
imum rating and on the straight portion of the tube charac-
teristic curves even with large grid voltage variations on
the output stage. Two gain controls limit the amplitude
of the signal between stages.

_The output of the receiver is applied to the vertical
deflection of the CRO, either directly, or if the signal is
weak, to a video amplifier within the CRO.

l9

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

V2

 

 

 

 

 

 

 

 

 

—
L4

 

é’ofi
CD

 

L3

 

 

 

 

all
®

 

 

 

 

 

 

 

 

 

 

 

 

 

Fig. 9. Power Supplies

20

 

 

 

 

400 39106 'Q'OH

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

'—-——VW~ - 6“
§
2» ‘>° °.
39
"" L I
a -@ H1 .1
kg ”“
__ ‘ flu
Jim ’@
3 :4
_l
g 9 3 3W
3 s: | or LW
4 9 ___I_ l
U
I
8 -
0~
at “I 1% d
c U " 4%"
> > ®
" 8 —\~ 3 I , 3
" -- ._ . a
J °‘ : [7 ID I,
23
o If a
l _J
*W“ * T” 5 3
0 g! 3r ‘ t +
_QDQVOWNV“1 i _ ‘ J or ' 5'
O—-*% 3 J—s: I E
C an X It“ U o:
M TV A ‘ l +
r---- '6 ”T
b a ___q *
O O _. O O 3—,,— d ‘ “N,
u U _—
‘5:
If on) “O 1
tavern-r <>—-—~ uaeerdi 1x3 ”A4”
Fig. 10. Multivibrator

 

 

 

21

 

 

 

 

 

 

 

 

 

 

39VIKV\H9H1
T

I
A

 

~---4

 

 

 

 

(E C? .LnO'VH

 

 

)5:

_mL

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.Omz

 

 

 

 

 

 

Power Amplifier

Fig. 11.

22

«L

 

01’

>md

 

 

 

 

 

 

 

 

 

 

 

 

 

D

 

 

 

 

 

 

 

 

w,

u}
7

 

 

 

 

 

 

 

 

 

JV

 

 

 

0‘63 493!” 0' 123 3A Q—‘

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

no

 

U
x08 I
mm»! <-“I

NOILV

 

 

 

 

 

__l
inéNl

Receiver

Fig. 12.

23

F. List of Components

 

Multivibrator:
Resistances (R).
3 680K
4 1.2Meg
5 120K
6 300K var.
7,13 10K 5W
8 25K 10W
9 220K
10 6.8K
11,14,18 39K
12 68K
15 25K
16,17 2.2K
19 81K
20 330K
21,24 400 var.
22 470K
23 5.1K
33,34 450 SW
35 20K 10W
36 8K 5W

gower Supplies:

Resistances‘TR).
1 43K 200W
2 130K 50W

Inductances (L).

 

Capacitors (C).

5 lSmmf

659,10 24mmf

7,12,17,18 .01

8 .001

11 .047

13 .0015

14 920mmf

15 .02

16 300mmf var.

32-33-34 50-40-30 350V

35 250mmf

Tubes (V). Inductances (L).
6SN7 5 9 microh

5 6A0? 6 5.5 microh

6 884 7 17 microh
6J5 8 5 microh

Capacitors (C).

1,2 4 1000V ele.
3 l6 450V ele.
4 20 450V ele.

Transformers (T).

RFC l 2.5 mh

Tubes (V).
1 5Y3
2,3 866A

1 5/25 hy 1 375-0-375, 5v-3amp. 6.3v-1.8amp
2 20 hy 2 Variac
3 5 hy 3 2150-0-2150
4 l4 hy 4 2.5V-10amp
gower Amplifier:
Resistances_(R). Capacitors (0).
25,27 25K 19,30,31 .005
26 220K 20 .05
28,29,32 100 5w 21,22 .1
30 25K 100w adj. 23 300mmf var.
31 20K 100W 24 .0015
.001 2000V

Transformer (T).
5 6 OBV-I'l 08amp

25,27,28,29
26

24

145-145 var.

Tubes (V). Inductances (L).

9 6AG7 9 28 microh1
10 807 10 3 turns 4g" diam. 1/8" spacing
ll VR 150 14 38 microh
RFC 2,3,4 2.5 mh
RFC 5,6 7.0 mh
Receiver:
Resistances (R). Capacitors (C). .
37,42,52,55 2.2Meg 37 lOOmmf
38 220 38 500mmf
39,44,58 10K 39,40,44,48 .001
40,45 81K 41,49 365mmf ganged var.
41,47 500K var. . 42 .05
43 470 43 .0015
46 30K 45 24 150V electrol.
48 39K 46 1.0
49 220K 47,50 .01
50 25K 51 800mmf
51 3.1K 52,53,54 16 450V electrol.
53 500K
54.59 470K
56 2.5K 5W’ Inductances (L).
57,60 450 5w 11,12 38 microh
61 45K 10w 13 14 hy
62 25K 10W RFC 7 2.5 mh
63 70
Tubes (V). Transformer (T).
12,13,14 6A0? 6 350-0-350, 5V-3amp, 6.3V-1.2amp
15 6SJ7
l5 VR 105
17 VR 150
18 80
19,20 6L6

All values of resistances in ohms, unless otherwise
Specified, all values of capacitors in microfarad, unless
otherwise specified.

25

IV.‘MECHANICAL EQUIPMENT

The sound source consists of a crystal which is sup-
ported by a holder of conventional design. For maximum
output the crystal is air-backed. In all experiments des-
cribod here a barium titanate disk was used.

Fig. 13 shows the arrangement for the reflection meth-
od. The tank holding the liquid is stationary, the left
vertical wall is parallel with the crystal face and serves
as the reflecting surface. The crystal holder is supported
above the tank and can be moved through the liquid by a
vernier, calibrated to 0.1 mm.

If the instrument is used for transmission measure-
ments, a second transducer, identical with the first one,
is mounted on the supports on the left between the thermo-
meter and the stirring motor.

For experiments with liquids of temperatures higher
than 50° a Pyrex dish replaces the glass tank, and a metal
reflector is submerged in the liquid if the reflection
method is employed.

The crystal holder can be turned sufficiently around
its three axes to hold the crystal in any necessary pos-

ition.

26

 

 

 

Crystal Mounting .

Fig. 13.

27

V. MEASUREMENTS

As pointed out already measurements can be made either
by the transmission or the reflection method.

Ordinarily the actual distance between the crystals)
(or between crystal and reflector) is not known. The trans-
ducer is placed somewhere in the liquid, the reading of the
mm vernier is recorded for this particular position, and the
movable marker on the sweep of the CRO is placed on the be-
ginning of the vertical deflection representing the first
reflection. The time correponding to the position of the
transducer is read on the sweep delay scale of the CRO.

Then the crystal is moved some distance, either towards the
second crystal (or reflector) or away from it. The readings
are repeated for the second position, and the differences
between the distances and the times are used to calculate
the velocity of sound through the medium.

In order to determine the velocity in solids only one
reading is taken, and the actual distance between the crys-
tals (or crystal and reflecting surface) has to be known.

Fig. 14 and 15 show the difference for two readings in
water. The crystal was moved through 50.0 mm; thus the
difference between the second deflections represents the
time reqired for the sound to travel 100.0 mm (reflection
method). The first deflection is caused by the initial

28

pulse activating the transducer.

 

Fig. 14. Short Path Fig. 15. Long Path
Reflection Reflection

If the absorption of the medium is not too high multip-
le reflections become visible, and it is possible to use the
second or third reflections for measurements.

Fig. 16 shows a pattern obtained by using the trans-
mission method. The first deflection is the amplified an-

tenna pick-up, the second is the signal caused by the sound

 

Fig. 16. Transmission
Pattern

29

pulse reaching the receiving crystal. Some of the sound
energy is reflected from the holder of the receiving trans-
ducer, travels back to the transmitting crystal, is reflect-
ed there and finally reaches the receiving crystal. Thus
the third deflection is the trace of the sound pulse after
it has crossed the medium between the crystals three times,
the fourth deflection after five crossings etc.

As pointed out previously measurements of absorption
meet a fair amount of difficulties. Yet it is not impos-
sible to obtain results, numerically too low, yet indic-
ative. The attenuation box is placed between the receiv-
ing transducer and the input of the receiver, thus the in-
put can be attenuated in steps 0f 1 db to a total of 41 db.
Measurements are made by observing the negative amplitude
of one of the reflections, preferably the first one on ac-
count of its relatively high negative deflection, and by
noting the position of its lowest point. Then the crystal
is moved closer to the second crystal (or reflector), and
the deflection of the same echo becomes more negative
(longer). Attenuating the receiver input decreases the
amplitude, and the amount of attenuation needed to bring
the deeper deflection up to the level of the first setting
gives an indication of losses due to increased distance of
the sound path.

Galt15 has used essentially the same method of obtain-

50

ing absorption coefficients, yet there are three main dif-
ficulties to be considered. First, since the output of the
transmitter is tuned to the crystal frequency, the output
stage will resonate with the crystal if it is activated by
the sound pulse reaching it from the reflector. This means
that not all the electrical energy generated by the crystal
reaches the receiver but part of it is consumed in the out-
put stmge of the transmitter. This loss varies with the
impedance match between crystal and transmitter and between
crystal and receiver input. Fig. 17 shows an extreme case

where enough power reaches the output tank circuit to be

 

Fig. 17. Vibrations in-
duced in Transmitter

visible on the CRO. This unwanted situation can be used for
measurements of sound velocity in a very simple manner, name-
ly by not having a receiver at all, and using the tank cir-
cuit as detecting device.

The second difficulty arises because the cross section

31

of the wave train constituting the sound pulse varies with
the distance from the source, thus not the same cross sec-
tion is intercepted by the second crystal (or the first af-
ter reflection) at all distances, and as long as not the en-
tire pulse is utilized in making absorption measurements,
the measurement will be inaccurate.

The third difficulty is caused by properties of the elec-
tronic circuit. If attenuations are made to decrease the
signal amplitude reaching the receiver, then the overshoot
of the receiver amplifier caused by a strong initial pulse
also decreases, and the base line, which is an.exponentially
decreasing sweep from the end of the initial pulse, approach-
es a more mearly horizontal position. Thus the pulse traces
originating on the base line show a deflection minimum lower
than they would have without attenuation and with a more

curved base line. Fig. 18 illustrates this source of error.

 

 

 

 

 

e - . __ _V w - - i
cite/imam
_ - __ T __ _.
1‘ {561’ any 2 ”4.561507; afar Mada/(2m

Fig. 18. Effects of Attenuation

Therefore only velocity measurements are included here

since they do not depend on above mentioned phenomena.

32

VI. EXPERIMENTAL RESULTS

A. Calibration

Before any velocity readings of the substances to be
examined could be taken the instrument had to be calibrated
and checked for consistency; therefore a few preliminary
measurements were made in liquids with a known change of
sound velocity as a function of temperature. Then sound
velocities were measured in liquids which are not suspected
to exhibit any abnormal behavior, like tap water and motor
oil. Fig. 19 shows some 0f the results, the readings ob-

 

 

 

 

 

 

 

 

 

 

 

 

/560 M90
7 Wafer -\.:
§~ , egg-1 L 0
$540 1400 s
e / “3‘
g , 5.4.5 40 oh; \E
m . a
534“” .Sfliiaflad \\\\\\%\\\ [snag
§ 3
a K 5%

HML r /240

20 40 60 80
7enyxya£afle WC

Fig. 19. Sound Velocity in Water and Oil

33

tained for distilled water did not differ by more than
0.1% from the values given by Masonl6. S.A.E. 20 motor
oil was compared with S.A.E. 40 oil which was examined by

Melchor and Petrauskasl7.

B. Measurements

It was hoped that a pulse generator might be an ins-
trument delicate enough to show experimentally whether some
liquids cause a dispersion of sound velocity which is not
due to temperature changes but to either changes of frequen-
cy or changes in sound intensity.

Bergmann18 lists a number of researchers who conducted
experiments designed to show the existence of velocity dis-
persion; only fractions of 0.1% dispersion could be detected.
Hueter and Bo1t19 state that at attainable frequencies the
dispersion is ordinarily so small that it is not measurable.
Yet experimental evidence of dispersion was offered by Mas-
on et aleO who investigated polymers of isobutylene. Also
Ghurevitch21 mentions experiments performed in 1938 proving
that acetic and formic acid show frequency-depending dis-
persion. Other workers found a dispersion of sound veloc-
ity in rubber4 and a few other substances with high molec-
ular weights22.

However, none of the papers mentioned demonstrate any
velocity dispersion as a function of sound intensity, a1-

thoueh it was suspected23. until Wade. Simbo, and Ode21+

34

published their findings of frequency dependence of sound
velocity in fatty acids, including one case of intensity
dependence.

Thus it was considered interesting to investigate the
dependence of sound velocity on sound intensity in lauric,
palmitic, and stearic acid. Fig. 20- 22 show the results
of this investigation. The sound frequency used in all ex-
periments was 2.595 M0, the transducer a disk of BaTi03
which has the pr0perty of decreasing its output intensity
at higher temperatures, yet in the range covered this de-
crease is smallls. The only influence a change of BaTiO3
can have on the results would show as slightly too low vel-
OCity readings in the vicinity 0f 810 and above.

It is possible to repeat the experiments at different
frequencies without replacing the transducer since BaTiO3
responds to all frequencies within a wide range25; however,
this change to a different frequency would quite naturally
involve a change in the intensity of the sound output, and
since it was found that sound velocity depends on both the
frequency and intensity, it would be impossible to differ-
entiate between the additive effects of both changes. This
impossibility arises from the fact that the sound pulses

are so short that no instrument could be found to measure

the actual intensity of a single pulse.

 

I280 X

 

60

 

a
/

 

N
O

 

.5on Velocity m/sec.

IZOO

 

 

80
\.

 

 

 

 

 

 

 

40 .50 60 70
Tempe/draft °C
Pea/A {o — peak daring Voltages;

x 90 volt
o 70 volt
A éETwafi‘

Fig. 20. Velocity Dispersion in
Laurie Acid

36

 

60‘

 

0\
Q

 

 

Sou/7d Veloazv 07/566.
8
0
/

8
2"

 

 

 

 

 

 

 

 

.MMO \\
60:
60 70 60
@mperczfare °C
0 90 volts
x 70 volts

Fig. 21. Velocity Dispersion in
Palmitic Acid

37

 

 

 

/’// ,
7

k)
0

 

I200 \\

 

.Somd Veloci'é/ m/sec‘.

so \\

 

 

 

 

 

 

 

 

 

60 \__+
60 70 so 90
femperalzzre ‘0
x 90 v0 [1‘5
0 70 volts
A 55 volts

Fig. 22. Velocity Dispersion in
Stearic Acid

38

By using the same transducer at a fixed frequency it
becomes convenient to express the relative intensity chan-
ges in peak-to—peak voltages with which the crystal is driv-
en. Different crystals will not yield the same sound inten-
sity, even if supplied with the same voltage. Therefore no
attempt was made to confirm wada's results concerning frequ-
ency dependence.

Fig. 23 indicates the results of an investigation made
to determine whether the velocity dispersion in fatty acids
is due to only frequency and intensity changes. Palmitic
acid was used since it shows the greatest dispersion.

Points L on the curve were obtained by transmitting sound
pulses of 47 microseconds duration, points 8 by using pulses
of 19 microseconds; in both cases the driving voltage rem-‘
ained the same. Considering that the ratio between the
pulse length is about 2.5, and comparing it with data in
Fig. 22 where the intensity ratio between the two curves is
about 1.65, the results seem to indicate that velocity dis-
persion is not due to a change in total energy (pulse length)
radiated into the medium but only to amplitude variations
(pulse height), the frequency remaining constant.

The dependence of sound dispersion on sound intensity
does not fit into the general scheme Of a linear relaxation
theory. For constant temperature Nada observed a dispersion

curve showing a resonance type behavior: the sound velocity

39

 

 

/

 

\

.S

\

\
1
E5

 

Sound Veloaly 07/586

 

 

 

 

 

 

 

/20
\
L
' \
//80
60 7O 80
fifiqpenmmvr”%7
Rzlmlflc Add 80 volts
Pulse Length: .3 . .. /9/u¢ec.
L ...... 47/usac.

Fig. 23. Effects Of Increased
Pulse Length

40

increases with increasing frequency to a peak value and then
decreases again. This special type of dispersion has been
predicted in the unified relaxatiOn theory given by Hiedemann
and Spence26 for the case that a substance is present in two
different phases. This agrees with the explanation offered
by Wada that near the melting point fatty acids may contain
certain groups with a higher state of order than the rest.
By means of a rather formal approaeh‘wada succeeded in der-
iving a dispersion formula predicting a dependence of dis-
persion on sound intensity. Wade. himself, however, pointed
out that this theory does not give a physical explanation

of the relaxation.mechanism involved.

It may be significant that recent work27 has proved the
existence of a liquid crystal state in derivatives of the
fatty acids (sodium stearate etc.). Investigations of
ultrasonic dispersions in liquid crystals will show if these
effects can be used as indicatOrs for the liquid crystal

state 0r also for a lower state of order.

41

VII. SUMMARI

The usefulness of a radio-frequency pulse generator
was demonstrated. The possibility to change various char-
acteristics of sound output allows a great number of in-
vestigations.to be carried out with the instrument. It
was shown that the apparatus is not limited to sound vel—
ocity measurements in liquids which.were described.

DiSpersion of sound velocity in three fatty acids was
shown to exist; it was found that this phenomenon is a func-
tion of sound intensity and not of total sound energy rad-

iated into the medium.

42

2.
3.

4.

10.

11.

VIII. BIBLIOGRAPHY

Hiedemann, E. "Wellenausbreitung in festen KBrpern,
Ultraschall", Fiat Review of German Science, Physics
of Solids (1947).

Firestone, F. A. U.S. Patent 2,280,226 (1942).

Teeter, Jr., C. E. "AbsOrption of Supersonic waves in
Liquids", J. Acoust. Soc. Am., 18, 488 (1947).

Ivey, D. G., Mrowca, B. 1., Guth, a. “Propagation of
Ultrasonic Bulk waves in High.Polymers , . Appl.
Phyflo, 20, 486 (1949).

Lutsch, A. "ZerstBrungsfreie Prflfung durch.Uberschall
mit dem Laufzeit-Echoverfahren", Arch. Eisenhttw.,

23! 57 (1952).

Seemann, H. J., Bentz, W. "Untersuchungen.flber den Ein-
fluss des Gefflges auf die Extinktion von Ultraschall-
wellen in metallischen Steffen", Z. Metallk., 45,

663 (1953)-

Lutsch, A. "Eine einfaehe Methods zur Messung der elast-
ischen Konstanten mit Hilfe von Ultraschallimpulsen”,
Z. ang. Phys., 4, 166 (1952).

Termann, F.E., Pettit, J. M. "Electronic Measurements",
MeGraw—Hill, New York (1952).

Ficken, G. W. "A Sing-Around Circuit for the Measurement
of Sound Velocity , Master's Thesis, Michigan State
College (1953).

"Radar Electronic Fundamentals", NAVSHIPS 900,016, Navy
Department, Washington, D.C. (1944).

Elmore, W.C., Sands, M. "Electronics", McGraw-Hill,
New York (1949).

Easton, A. "Pulse Modulated Oscillator", Electronics,
MflPCh 1947.

"Radio Amateur's Handbook", 32nd cd., ARRL, west Hart-
ford, Conn. (1955).

Martin, Jr., T. L. "Ultrahi h Frequency Engineering",
Prentice Hall, New York 1950).

43

15.

16.

17.

18.

19.

20.

21.

22.

23.

27.

Galt, J. K. "Mechanical Properties of Alkali—Halide
Crystals", M. I. T. RLE Techn. Report 45 (1947).

Mason, W. P. "PiezOelectrie Crystals and Their Applic-
ation to Ultrasonic", Van Nostrand, New York (1950).

Melchor, J. L., Petrauskas, A. A. "Ultrasonic Studies
of Polymethyl Methacrylate", Ind. Eng. Chem.,
44, 716 (1952).

Bergmann, L. "Der Ultraschall und seine Anwenduni in
Wissensehaft und Technik", Hirzel, Zfirich (l9 9)

Hueter, F., Bolt, R. H. "Sonics", John Wiley & Sons,
New York (1955).

Mason, W. P., Baker, W. 0., MCSkimin, H. J., Heiss, J. H.
"Mechanical Properties of Long Chain Melecule
Liquids at Ultrasonic Frequencies", Phys. Rev.,

73, 1074 (1948).

Ghurevitch, S. B. "Absorption of Ultra-Acoustic Waves
in Liquids", C. R. (Doklady) Acad. Sci. URSS,
EV. 17 (1947)-

Ballou, J. w., Smith, J. c. "D amic Measurements of

Polymer Physical Properties', J. Appl. Phys.,
20. 493 (1949).

Seidl, F. "Schallabsorption in beschalltem Transforma-
toren81", Colloquium Over Ultrasonore Trillingen,
Koninklijke Vlaamse Acad. voor Wetenschappen van
Belgié (1951).

Wade, Y., Simbo, 5., Oda, M. "Dispersion of Ultrasonic
Velocity in the Liquid Fatty Acids", J. Phys. Soc.
Japan, 5, 345 (1950).

Schaaffs W. "Sehallgeschwindigkeit und Molekfllstruktur
in Flflssigxeiten", Erg. ex. Naturw., xxv, 109 (1951).

Hiedemann, E., Spence, R. D. "Zu einer einheitlichen
Theorie der Relaxationserscheinungen", Z. f. Phys.,
133. 109 (1952)-

Moses, H. A. "The Proton Resonance Absorption in

Liquid Crystals", Master's Thesis, Michigan State
College (1953).

44