H”

J

NWNWWWWilly

‘

Li‘&'!‘§2n%$€336£€ TESYINCE AS? .43. MEM‘iS Cf'F

DETECTSNG E‘JLTEGL‘fi DAMAGE 5N CEEBJaENT .‘VQCR‘MR

4mg
no N
(ID—101

Ttwsis 170-2 fins: DWI-so c»? Fix. D.
fuliCHifiAE‘t‘ STA‘EE LER-IEX’ERSWY‘
«Emu M. Dairympm
7963?

' it ~—.,
.."S‘.‘!‘:

This is to certify that the

thesis entitled

ULTRAmNIC TESTING AS A MEANS OF DETECTING
FATIGUE DAMAGE IN CEMENT MORTAR

presented by

JEAN M. DALRYMFLE

has been accepted towards fulfillment
of the requirements for

DmTCR 01" W degree iWICS

(MOM

Major professor
Date March 28, Lflé;

0-169

LIBRARY

Michigan State
University

 

ABSTRACT

ULTRASONIC TFSTINS AS A MEANS OF
DFTECTING FATIGUF DAMAGV IN CFMFNT iORTAR

by Jean M. Dalrymple

Body of Abstract

The purpose of this investigation was to detect fatigue
damage prior to fracture by means of the change in attenuation
of an ultrasonic signal. The investigation was carried out on
cement mortar beams subjected to completely reversed flexure.

An ultrasonic signal of from 112,000 to 120,000 cycles
per second was transmitted through the length of a cement mor-
tar beam which was subjected to completely reversed flexure at
the rate of 1,800 cycles per minute. The level of the output
signal was plotted against signal frequency at intervals dur-
ing the test to obtain a frequency response curve. A compos—
ite output was obtained by taking the area under this curve,
and the drop in composite output as the test progressed was
taken as an indication of fatigue damage within the mortar.

There was continuing drOp—off in composite output with
increasing numbers of cycles for a sufficiently high stress
level. For a sufficiently low level of stress, there was no
change in composite output with increasing numbers of cycles
tested. It was concluded that this method may have possibili-
ties as a means of investigating cumulative fatigue damage and

ociated phenomena.

0)
m
m

ULTRASONIC TESTING AS A MEANS OF

DETECTING FATIGUE DAMAGE IN CEMENT MORTAR

by

Jean M. Dalrymple

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OE PHILOSOPHY

Department of Applied Mechanics

1901

ACKNOWLEDGMENTS

The writer is particularly indebted to Dr. J. T. McCall
for the initial suggestion of the topic and for his help in
getting the project started; to Dr. C. A. Tatro for advice
on instrumentation; to Mr. J. W. Hoffman who, as Director of
the Engineering Experiment Station, furnished some of the
equipment used throughout the investigation and also furnished
direct financial support during a part of the program; and to
Dr. C. 0. Harris for assistance in organizing the project

results.

ii

 

INTRODUCTION

Purpose .

Bgckground

THE IKVESTIGATION .

Specimens .

Instrumentation .

Procedure .

Test Results

Conclusions

For Further u

BIBLIOGRAPHY

APPEfiDIX

f“

TABLE CE

iii

PAGE

(1‘

15

is

f\_)
O

LIST OF FIGURES

FIGURE PAGE
1. Instrumentation . . . . . . . . . . . . . . . . . . 4
2. Frequency Response Curve During Cycling . . . . . . 7
3. Frequency Response Curve Prior to Cycling . . . . . 9
4. Test History for Specimen Number One . . . . . . . . 10

50 TeSt RESUltS o o o o s o e o o s o o o e o o o o e 0 14

iv

INTRODUCTION

Pur ose. The purpose of this investigation was to
detect fatigue damage prior to fracture by means of the change
in attenuation of an ultrasonic signal. The investigation was
carried out on cement mortar beams subjected to completely

reversed flexure.

Background. The present investigation was an outgrowth

 

of work done by Dr. J. T. McCall for the Michigan State Univer-
sity Engineering Experiment Station on the fatigue of plain
concrete. The instigating idea was that if one could detect
fatigue damage before fracture and follow its progress, this
might lead to increased basic knowledge of the fatigue process,
to a better understanding of the cumulative fatigue damage con-
cept, and to further information on the "self-healing" of con-
crete. It was proposed to investigate cement mortar beams and
limestone beams separately because of the greater uniformity of
these concrete constituents, as compared to concrete beams.
Moreover, an understanding of the contribution of the individ-
ual constituents should lead to a better understanding of the
fatigue of concrete itself than could be obtained from an
investigation on concrete alone. The present investigation
represents a start towards this goal.

Current theories of fatigue and most of the literature on
fatigue are based on fatigue of metals, and rest heavily on the
phenomenon of slip. The fatigue of brittle materials has
received much less attention and is apparently pursued as a

l

separate topic. In particular, the literature on fatigue of
concrete is quite limited; a comprehensive bibliography is
given by Nordby.

The writer found nothing in the literature which would
bear directly on this investigation. However, references cited
in the BIBLIOGRAPHY are ones which were either helpful in sug—
gesting ideas and increasing general knowledge in the area of
ultrasonics or useful in developing the instrumentation, as the

annotations in the BIBLIOGRAPHY indicate.

THE INVESTIGATION

Specimens. Test specimens were cement mortar beams

3 inches square and 14% inches long. They were made of type I
portland cement and natural sand in the proportion 1:3.3 with
a water/cement ratio of 0.52. The mortar was mixed in a Lan-
caster mixer and placed in steel molds in two layers; each
layer was vibrated for five seconds. The beams were removed
from the molds after twenty-four hours and placed in the moist
room for three months. They were then removed to the testing

laboratory where they were stored until used.

Instrumentation. The identification numbers for the
equipment listed below refer to Figure l.

(l) Sonntag model SF-lU universal fatigue testing machine,
used to load the mortar beam (2) in flexure at 1,800
cycles per minute.

(3) General Ultrasonics model GU-400 (modified) ultrasonic
generator, used to apply a CW (continuous wave) signal to

(4) - arium titanate transmitting crystal (one inch in diam-
eter by one inch thick), which transferred the CW elect-
rical signal to the mortar beam in the form of a com-
pression wave.

(5) Receiving crystal (same type as the transmitting crystal),
which picked up the transmitted compression wave from the
end of the mortar beam and converted it back to an elec-
trical signal.

(6) Rectifier and filter, used to convert the transmitted

3

 

 

 

 

 

 

 

(7)

(3)

Figure l. Instrumentation

 

5
signal to DC (direct current)for the
(7) Sanborn model tO—l300.chart recorder, used to register
the level of the signal transmitted through the beam.
(8) Tektronixtywe“332 oscilloscope used to monitor input
and outpat signals and to provide a DC calibrating
voltage for the recorder.
(9) Fico model 323 vacuum tube voltmeter, used to check cal-
ibrating voltage.

(10) Hewlett-Packard model 5238 electronic counter, used to
measure frequency of signal output of the ultrasonic
generator.

(ll) Triplett model L30 multimeter, used to monitor trans-
mitting crystal current.

(12) Meter rectifier.

(13) Thermometers.

Procedure. Basically, the experimental method consisted
of converting a CW electrical signal into a compression wave
which was applied to one end of the mortar beam, converting
the transmitted compression wave at the other end of the beam
again to an electrical signal, and recording the level of this
output signal. Since the Sanborn recorder could not respond to
a signal in the frequency range used, the CW output signal was
rectified to present a DC signal to the recorder. A decrease
in the level of the rectified output was taken as an indication
of fatigue damage in the mortar beam. Rather than depend upon

the rectified output level corresponding to a single input fre-

quency (for reasons discussed in the ARRFNDIX, page 20), a com-
posite output based on a range of input frequencies of from
ll2,000 to 120,000 cycles per second was used as follows: The
CW signal input to the beam was varied over the desired fre-
quency range. At intervals in this range the corresponding rec-
tified output level was recorded on the Sanborn chart. Then the
rectified output was plotted against input signal frequency to
give a frequency response curve, as shown in Figures 2 and 3.
The area between the frequency response curve and the Sanborn
zero level was used as the composite output. This composite
output may be considered to be a type of average output over a
range of frequencies, and it was the change in composite output
which was actually used as an indicator of fatigue damage. Fre-
quency response curves were plotted before the cycling was
started and periodically during the testing of each specimen.
The data for the latter curves were taken while the fatigue
machine was running, and required about fifteen minutes, or
27,000 cycles, of running time. The data were always taken by
starting at the high frequency end of the frequency range used.
When the mortar beam broke, the fatigue machine, which had a

built-in cycle counter, was automatically shut off.

Test Results. Results of the test are based primarily

 

on three specimens. Figures 2 through 5, which are described
below, summarize the test results on these three specimens.
Figure 2 is a frequency response curve taken on specimen

number three at 3,800 cycles. The variation of rectified output

8
with input signal frequency shown on this curve is typical of

response curves which were made while the cycling was in pro-

U)

gres .
Figure 3 is a frequency response curve taken on specimen
number three prior to the start of cycling. The deep dip in the
curve at A in the figure is the most pronounced feature which
distinguishes this from the later curve, Figure 2. This same

feature, the exa gerated dip in rectified output level, was

J

noted also on specimen number two prior to the start of cycling.
It was not evident on the corresponding response curve for spec-
imen number one, perhaps because for this specimen a narrower
range of frequencies was used for the response curves than for
those of specimens two and three.

Figure 1 presents the test results for specimen number
one. The upper graph in this figure shows how the composite
output varied as the number of cycles increased. The lower
graph shows how stress level was varied during the test on this
specimen. The scale for number of cycles tested is the same for
both graphs so that one can be compared to the other directly.
The dashed line at 946 psi (pounds per square inch) represents
the rupture strength as obtained from the average of seven beams
tested statically with third-point loading. A chronological
description of the test on specimen number one follows:

The time at which the fatigue machine was started was
taken as zero cycles. PriOT to this time, frequency response
curves were made at intervals to check the reliability of the

instrumentation. The composite output obtained before starting

 

. l

C ,

Std

701‘.

P .

-8

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

     

 

 

 

 

 

 

 

 

 

 

 

0 . . . . r0. . _ . 0. 0 L . . I” .v 0 T . . L . e . .7» 0.. - 0W0 . .1 H . >0” w ML..- “L

.-.d 1 x. .- ... . . .1. ..L .. L. L. ... ..0.... .. ..n.H._.. . , .. L3H...»
... . . .0 . 1 . .. . L .. 0.. . 4.. u._.. x . .u .‘0 00 . 0 .- ..u.-n!v .

h . .. ..-.L: ..L .. :.... ...:Leae ....eLLm-pm-L-E.....r-ELL- . -.: ...-....

5 I...»- .03 ”-..1 .- 01-010LL001-.10. -%1.-0.0...10;0.. ...-044440.011». ..L .- L .-.-L ... - . .. .. .H . T L..»; O

2 .0...-_L.. a L..; ”KT. ..LL - .- L L :...L.L L .....Lw-L..... . 1 L L. ...0.00L0.0.0..-2

- . .. 1 - . . . . L. .. .. -..H0.0 -..-L110? .010? 7 .10 -00 .0 .0...- . . .-

0 VHHIHIHIW. H0. 0000.v0. . .. .%~ 0 .4004 07.01.1444 $1010.”). T00 7007‘ - .... .u L . . ... ... 0.... 0 ..N..4H 5

2 r..w-....-...H-. 0.0. - 40.; .L . . ...-L....Ya. ...w... . .. . . H .H 4.~. 1. M - ... .7000. 0w 0 -000.lw1.714 ...... JAIN-.HIHW 0|
5 ...-HILL-.. L LIL-1.1.704 001“...--..L-..0M-_0.Tx.30-q 0. 1.407031.- .1-17-«0- «N0 . ...- . “L..-.0?

1 ..-.L.n.... ..L ......-.-. L. L. L..... . . . .. ... . . L.. L L. .. ...J....0..L-0..00
....-0.-_.J.. .n- - . .L . L..: .-.-L0 . 1Y0 -0.0.. .-.-4-.dU..-000-.- U000: U..-0.-- ..12 a .L-n
wit“. -.:-701.404. ....L-I-l .0... 7.73 .... ... .. .L. . . TjJ . L «f L .J .L -.-00+0-.005

I ..-... .L ....L L. 71L“ ..L. v.61...» .L-LLLLL-Yit 3-. -...0000. .. .
L “.0 -. .. . .0. L . 70%.!h0001f‘. I10 +1. ..0000..LL14.! wi00010. .rdm . . - .. 0.+-7- 0 . “-000.01”..- 4..IY.10 — . . v. . . - . . m . . a.
0.0-1- i . . . . LL. . _ . .. -..-,- ..
.. - -... ...L. L....- .. .. ... . .0 L 00 L ....
...L. a- L. -. ..L Lat .L ...L. 00 .1,- 171 AL 0 .L- ....1- .d L. .... L 7.50m ..
.L . L .. ..7. .L ._ . .. .. .. LL. . . L... . . . ... .00..
. . L . .. L .. ..L .+ L n. . . L _ L0 .L . .00 0 1
- . L..... _ .. .- ... ... _ 0 L. I 000-0.... -0 0-0.. 00 L. 10.. - 07.0-. . T1 LL . .5
5 0.10 ”I 7... 0-1-01L10- ..LL-.-10.0.0.0.-L00.0- .... - -0 -L 1.00.7 0004.00-17. 07-11-MMLI 10 .. .1..T .. L. 0JL L L L. .... ..H . O
0. H .L n . L .. . L . L .. ...HL. M .. w . L L .. L b . ._0 0 -00 .0.0 .L 00.- an
o L. L L 01L1pv-0rf-f..r.v010ff aw.-0-01J; 1011- L:0L-< 0L-0d L. L ;
27....0-..00--.L.--0...d.-00 . L. L .. .. . L. . . L L L L L 00.0 0400 .....zl.--_.0%.-- 0.0.0:...LL-LL5
’0 . - u.. .L . . v m n O L. w . IIN.0.H00740-7.|i0 . 0...“.70- w.- -.u00.000 L 0 0%. I! .0.-..r9. m w H . . II
I.. It‘l- 04...l0|ltl 0|! . 00! 00. 0| ‘7 ...... '00 .-ll 0.7 0 . ..0l 1|. \ l...0404!l L . . a c . . . -
00. L 00“. . +1...J.~ a. L . a I _ v _ L L _ L. .W . _ H lu0 01“. I L. {my 000-. O
n . L. . . . L . . . L047 0- .000 .00 '00.-L0 0. 00. 001L. 14.7 -. 0 0.0 14 . .L _ . .(I
0.00-.-.10LL .- ...--.L.-.0r. 0-L- . 0.... 14-00 .11 1m 4 L L ... L L .L L L £1,pr
- L. L L . L . . . L L ... L L - _ - e3 0,. 0.. -L . L---.-L--d- ...
. . . 100-. 0 0 0‘0! 100 .0 ..0- 000! I00 v -. 0 0 J . . . . . . . . .
VJ. 4-04 0 00!.0 -. o - .400.- 000000 0L. 0100 .10 L .W . L» C . . L . u _ H a .L .. . L . - . . . . L. . . . I 00.0000Hl' .
. L #.L_L . L L L. L. L . L 0- . L L 7717141§7L5110T77-r T7
00. . _1--.--0-_------.-- - ...-1... - 0 .. . a _ . L L _ ._ .. . L
10.. ,. T01. . . .L . L L L N 1 ...H . a .. . _ L L . . L - L. _ .L L L1-- 5
L . L . L . .L L L . L L.. r L. L . r0.-:L- -.L. 1 .---1.. 0 0.0 01.71. -. 0--..-0.--..-L. ,- 0-1. :...-000.001 11-».01.0.. -..0 ..00 . .. 2
50- .LL L..-TL 1. -... J..--.._.- ... ..:»..01-L.-. 10 0.---0..-m.-....a. 7.00.1; - .. L L . ... L.. .. L. . L ..LL... . .L O 1. o
2 a L . . . .L _ L . .. _ .L . . . . . L L L L. .LL-L0-0.-:.00;---0L0-000-.:-0L -0-0. 00. ..w.:
.- _ L . . L . L L e1- L---:e-0-00 0:000L0-¢--e- - 0-a;0.. . - L . ; L .. .D 5_
O 1-714VI 4-h0.la...1070 0lr0-040m!004.4 v..0!.040- .0040..L 0. H 0 00.4. m L L h r _ L . n L - fl L . . . LL 5
~04 m .- I -. a .L . v ....m L h . L .L ..- L v. .- 0.0.40! 1 0 01.71- 0.00070 .1 I $000700 ..I400 0”. ..r. I 00 H . 07“ . - ~. ....TW0J7400L Cl
. L . . .. 3-..0; .0 - .- .-0....-0-L0- - .00..»0- _-0. ----40.00....- ...-,-..-.L-H0-.0.. ..J- --L .. . L ... . .. . . . . ..
m -001-70..-+.-t-L. . L.. L L L. L . L. 7 e004 .. . L .L .L.-L.L-.-0:70no
fi- L L. . . . . L L . L . _ L L L L . d - .Vlhl.” + ----L01--.0-.0..-0- 00 - --.-0-- .. -..-0. 1L . lL . L .-
o .. 0.”.00-4. ”0 007%.». .-0- -0 IL.- l.. % ...“: r0—6 ! ”000.0 .0 - .4 3“ 1000 o 007 .00l...¥. 00.4.0 ewll. L :.f/ . L W ... . .m . , h- ” 4. .F 00
0.00% . . .L L. . L L L . L .. _ L r. . .- - . L . L 8. .-.
. H a .. 04“ ,. F 1’ ' L7 L01- 4 - L 1Fvl01 *wr d0b|0 70 e d. . J .. e 100‘. .. . . . .- T .. m“ H.
.P 1» r f . - - a . L L L. . L L _ L . -0.L - : - 0 -. -. 0 .. .. -1. L.- 1 L - - .- - .... I.
- L L a L-CL- -rzfi.&}0L.L T: L-? Lz-ETP L rrL. . L L .1}
L.- - - .- L.-.-.- - - . 0.-.-0- « L . . . . . P. . r0 . . 1....
...L L. L A _ . . L . . L L L ,. L _ . L . -- -00kt1-110 ...... 0..- ...-7.00.0..- 00..0 710:.- 0. . #01
L . L . . _ L- , 00..-..L-f-01.. 0 .-1T- woe--0110 00-0..»- -0010-00 00-0 0-1-0 10 0.. ..0- ..0- ..--.0 L. 100 _ .. .. . L L . L
. . I.0!1!0|Y-0|00..00|I0!T44000..10-0 01.0-00.4 .* m L L . b L _ I. . . L .1 .
r! .......... 4.407 L . _ . U A. . L .L . n _ . L L . 40—. . . . . . L L ”L.-
. L L . . L . . L LAW L L L .. L . L. ...- L . ....0 . .-.... .- .-.. 00....ML...
. . L . L L .. L 4 _ ._ . L ...... ...--. ..ezt-Lf . r. _ L . ... n
s . -0...-L: Li--. . L.- ..0 -- .. . ..L .0. .0. 0. .. L L L . . . . h
. . . . L . L . L.--..:-T- ..- -.-0 ...... L 0 . 000001. . _ . .,
O ..u 0' 70b 7.4. “-0 I. .0. .. . 0 4 4 4 1.4“ ~ . -.- H . .rl't‘dl!‘ . L . L
3 L , _L L _ L L _ L L L. L aH.. .0;274011L.10Lm
r) 7 0 0'. - . . . l4. 1 0 0-. 0 l 00 . 0.0 I c J 4 | I m0 ~ 40! 0| -” 0 - 01 00 0 _ . L. 414407, . . .
0L- -. L L L , L .. L L L A. L L L L L L . .L. -. -.-. 70.007.700.170. 000007- 11000 - h.0.-Lm
L . . L a 7LT90LL0.-T 2L7f+-a-L.-1- J- L171TJI L. . . L
m ..L .0 -7 3.0 .L. -L ._ . o. . ._ L . . . L L L 7000-00-0 01100! L -.01 - L. -3...
f L . .. _. u L m . .. L . .m L 0. L 0 00 .- 00 - ”-41.0! 0 L ..4 0 . ... . o. 0 000 .. 0 4 0 L4 0 u. a . r.
_ ‘ L L .. L 0 .... .0-.. 0 1.». .. .. - 00.0 - . . “- . a _ _ . L L L L
5 0 -0 .-...0- .. .L - -..-.- L... -4 ...... m . L L L . L .L L L . L L _ .. _ L0 _ . .0 0.-- Off-L
.. W L. L L fl L a. b M. 0m 7 07 4. 01 70. J! 0! r0 70 b0. .0 fl . 0 .0 M
H . L . 47 L . _ 1 < L . .. . .
L a .- -1 L - Mu. L 6 . ...u L w L O L a... L L o L L 0 L...
,V. L . 3 LL _L 7 _. L , L- 78 ;L.L 2-L.L fiIi-Llid..1£
S“... L. _ «u%---..- L. - - p.52.- . .00.... L--... 0. - ...-.-.. - L.. . L. L _ L L L L L L LO
by .4-4 [070 .51.} .0. . s - . .t .0 0 -. 0L . . . . L .— . ..10 o 0 ..-.| ... - .
2 L. .L0 L L L «Luce pea, wznmzo.awwm@0EoD-t - tw--t-L ..-:r -L :n 0e?-. . .L . .2
0.01-70.0- . - L - . .r. , . L . L L L H .. L . L L . L . L . . . L . L 3
2 .-L . L L L .L L L L . L 1L..-” - L L00L...-,1., -071 .-..-. -- ...-- L--;L - -.L- 0 LIL-.0 1.
...-1...-- . 0L- . - _ L---L. ..L .. - _ . _. L H L _ . . L ‘ L . L _ L L L---L0...
L . L . _ H. . . L L L L ..L- -..L L...- L- -1- - .. -_.. .. .0. -.L. ... L.- - L - L- 0.. ... ... L0
0114441....7..+.00.000 . 9'0 0 0 w: 0! k. 00 0 .0 W. 0 .—. .. o . “0 Id . .. L I -. * L. .L. _ . u L L L L H — .. a H
0 0a . , . _ . L _ . L , L L L -. -L.0... -L. .- L --.-01....
_ L0 L L.. L L L L 37---- L- L.. L. . - - .- L. ._.- T. L700 L- - L L. _ . L . .
5 J10- 01*..-!-0--..0.00.l.-70 fl- 0010M0 w--$ .0 1. 0. 4-0-0 .... . L . L L H ._ m u . n . b o . _ . L 40-..4 .40..0 LI 7.0.4. 0|“
L. w. _ . L . 1W0 . L - . r r r .. 0.07-7“:- Ir 7 4 r 004" III!" Ill.'0l.I..-.l .
wlwr — h h .-..10> L'I'llLIII 00. 0.4.4II.04.'!.I.0‘ 04.710

 

 

ll
the fatigue machine is plotted as 100 per cent at zero cycles.
The fatigue machine was run first at a 200 psi stress level
(completely reversed bending) for 13.5 million cycles. There
was no significant change in output during this time.

The stress level was then increased to 450 psi for 4.0
million cycles. During this period there was a significant
decrease in rectified output level and a flattening of the fre-
quency response curve (that is, less variation in rectified out—
put level over the frequency range used).

After a very definite downward trend in composite output
was established at 450 psi, the stress level was reduced to 200
psi for 1.5 million cycles. The frequency response curves dur-
ing this period were significantly flatter and at a lower level
than those taken during the previous 200 psi run. No change in
level or shape of the frequency response curves occurred during
this 1.5 million cycles.

Tne fatigue machine was then shut off for a four-day
period (at B on Figure 4). This period of time is plotted as
an equivalent number of cycles on the figure. At the start of
the rest period, the frequency response curve was comparable in
both shape and level of rectified output to that of the 200 psi
run just preceding. At the end of the rest period, the output
level had increased significantly and the response curve was
less flat.

The stress level was again set to 200 psi-~for 10.8 mil-
lion cycles. The shape of the frequency response curve retained

some resemblance to that at the end of the rest period. However,

l2
by the end of 3.3 million cycles, the curve had flattened out
somewhat and the rectified output level was slightly lower than

at the end of the rest period. No additional changes were noted

during the rest of this period.

)

Next, the stress level was increased to 300 psi for 20_.5
million cycles. Initially, the level of the frequency response
curve was slightly greater and the curve was less flat than at
the end of the preceding 200 psi run—-more like the shape of the
response curve obtained during the original 200 psi run. The
frequency response curves taken during this period were remark-
ably consistent both as to output level and as to shape. Over
the whole 202.5 million cycles there was a slight decrease in
output level. However, for any portion of this period compar-
able in length to other runs, the change in output level was
imperceptible.

The stress level was then increased in steps,

30.5 million cycles at 450 psi

3.2 million cycles at 480 psi

4.9 million cycles at 500 psi

34.0 million cycles at 520 psi
with no significant change noted in the output.

At the next step, when the stress level was increased to
540 psi for 20.7 million cycles, there seemed to be a slight
drop in the level of the rectified output at the lower frequen-
cies (below 117,000 cycles per second).

During the rest of the test on specimen number one:

30.0 million cycles at 560 psi

13

29.7 million cycles at 580 psi

98.2 million cycles at 000 psi

44.2 million cycles at o20 psi

23.9 million cycles at 640 psi

4.4 million cycles at ooO psi (fracture)
the output level on the low frequency side of the response curve
continued to drop off. Because of this, at the end of the 600
psi run, the frequency response curve was significantly differ-
ent in shape from that at the end of the 520 psi run. It should
be noted that this change took place over a span of 178 million
cycles, whereas there was a greater change in 4.o million cycles
during the first run at 450 psi before the rest period. How-
ever, when viewed on the semi-logarithmic plot of Figure 5, the
lepe of the curve (rate of change of composite output) during
the l78 million cycles appears to be nearly the same as that dur-
ing the 4.o million cycles.

Tne specimen broke approximately in the center of the
beam after a total of 5b2,910,000 cycles, while running at the
660 psi stress level.

Figure 5 is a semi-logarithmic plot of composite output
versus number of cycles tested, for three specimens. The plot
for specimen number one is from the same data as that for Fig-
ure 4, and hence the variation in stress level on this specimen
must be kept in mind when comparing the curves shown in Figure 5.
Specimen number two was run at a constant stress level of 550
psi and broke at 38,000 cycles tested. Only one frequency

response curve was taken after the start of cycling because of

 

 

 

   

 

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

     

 

 

 

 

 

 

 

 

 

 

 

 

 

. . 4 4 .l- 1
ll 1111-1 _ - . 4.. . . . H
.11 11411- 4 . . . . 4 .
1- 1....411114-17 . :44 4,144 . .
4 _ .4 .4 H 4 .. 4 ., 4 _ 4 4 . .4 .4 _ . . .
.H. -, :4 r44.4..4. .. . 4.... Y4 .1 4
4..- 1 . rafff- 4,. I44...:%-4... Z.;.Fi11:.i§4:.
. . . ..-T4 . 11H. es4; ..:4_- . s4r.44.. 3;-;e g :.:..-.+1144 .1.. .. .
w 1 4 H . ..... .- . . . _ H. ..4. : H . H . 4 ...H . . H H F1 19..I..H...311|-.....311T-.1..-1 H 4 H . . H. H... 44 .. . H -
H..; H4. H. H..-::H . H . . . 4k-4 H .. 41.. ..H H. H1. Hr.“ 141-1211- .11 1. .-. 4.1.4.141 4.1 H H .. ...H...» 4...... H . . H ..H. H .. r; .. . . 1!
.. :... .... . . .. ...1... - . 1.11 111-1 I... . .. . 4 _ . .4.. . 4 . 4 . 4 ... _ .
-.:14.:-.r.-_ - 4....4ge....1.4 .14 W e-.HmoaHm.4 . reHze.-;1-:4-. .. . . 1% ... . .oaya . .
H. H . _ : . H. H . H .. .. 7? ..:. HH, 4 _ ..MH. ”Hp .. M4 :4... a 4 -1---11
. H H-.H...:H ...-1.4.1.... - 4. . . . we 4 . . 4 H. ...-H +1.4 H.....4 ..H . m4. ifl m . . m. .4. 4 4 . , .- 013-14
4 4 H4 .1 ..H.-1 - H.. ..H. H. H. .4. H .. .H... . q . .(Hfilr-H .. .. .. . .. H - ..44 4 4 m H _ 4
H H. .4 q a .. . H ..... . .. .. H 4 H .H H. H.......H ...w %H H ..H..1.. -... n H4 H H #14 o H .14.... . .4 9.4 4 4 .
..-§-. 4- s. ....- 43; .4 .--.i-zx..4 . . .... .u-:;;:.. . . . .. 4 . _ .
4 1 4 ... .H.. .n't.:.‘Yl... . m“ H. 9 wOW.H-.w-Z.H ..H. ...-H.. H . r...H ..1....H... .....H.. 11-. Ln. mwijH .H .H.-H .4 . H....-H . H . H. . H. .H T. 4 H . H . 4 H
H.}.fii: ..4... u . 4 .. H H . . 1. 4 ”.H.. - . ... . .4. .:H . H.. ..:.r H. 4.. . . . H H .H ... . .... .1 - ..H 1 .
714-.H 4 .4 4. .744u.e-w@4 a.mwm.wmwe1 4 . ..-.H-r .:1-_.-...:;. - .:.f+;H;1-4+H . ..
-4. - .T-- ;.5 :... 4443. :.:; .4.. ..4:.:e;. .441T+sr1F1-z-1. ...1 .- . .
H. H....H . H H ... H .4.. w. . .4 H .. H .H. ...1..? ....H.. H... 4.. ...111_?.1 11. +1 u 1.1 Hlolluf 4 . . H..... 14. . .4 LL. . .19.},1 ..H....ItJILIJ H .
4 ,4 H. -.H.. __ .. ..-.. .. H.. . 4_H._4, .. . --. 4-. .-41..4-.-HJJHH.--.. . .4. . . ._ . , 4 H-..... ,4 .. .1.-.4.. ......1....-....H..4....4.. ...-1--. . . .
H“- H. H ._1.11Hv-.111V...1.¥1....(1..o9119H1. .9191..H.111-9 o H... r. .. . H ......H H... . ..H....o. H ... H .Hl‘lllll 1 l 91.11%»... .
H141fi,..HH Hz. _ 14111114111211 1.1.4
4.4. _ H 4.. . ...-.1.
. 4 .4 . . 1.... .
. 4H

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1101 v9.1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. . .414 .HTH .:4 .
H _ ...-.:.... .H ... -. H.. ..H ... H . . 4. . .H H _ . .. . - . \L/ .
.:. 444.4._...; _ .. 4 _ . . w
. ..4 _ H. H .H _. 4 4 . . . H..... .. H Emu .OHHmWw. 00-M- .“ -mul .HHU>UH. . .. ”...-.-Hl-L- 1......r9- .
. . . _ . .1.... _ . . .. . .4 1.-. . _. .
.f. 4 . . 4.. . _ H 4 ...1.... . .. .. . . 1.. . H ... 4 .
4. . 4 . i- ... ...- 4. H . 4 a4 . . ..- ..--..1.....- 4a ”.4.-4.1... 3 4: 4:35:44 n. 4 . . . ....
.. .. . H H H...... .1-. H. _ .... 4 H H:...H H. .........v. .. o. 1. . 1 1 H p 4 H4 » H. H H. ..H.. 4 . .H b may? m0dé>wu ...WD. .. H .H H . . . h. .. .
H.. H..: .... _. . H 4 _ 4 . 4 _ 4 .H.-....-. H1111....H.: .-.- . HzH H 4. H.. .. 4 . 4 .. . . . 6&9 H .. .:.. . . 4 H. .4 H . _ 04H .wo111.111
.. 4.1.4. H-.. .4. . .-1...-4; 11.11-. 4 .. H. . . H. .. _ 4 H..... ..H ....4. . 4 .. 4. .. . 4 . . . . _ , ix 4
H 4 H..-1911 ..:-91.; -1.... . . ..4.... . 4 . . . :14. _ . ...- ... . .H. H. . .... . 4 .1 1 . . . ..
..1...-4.1.41. .H.... 4 4 .. . H H4 H742... .. . ,. ..:H 4 4..-H 4.1411. -. ... .. ... . 4 H .1. .104. CmEHomQ. .4 .. . _. .
H. H . :....4 ..-.4 4 . . . . . 4.4 H: H.. ..., . 4 .4 .. .4. .4. .. H H 4. H 4 34-11 11141.7. . . m. : Q? Q4 . . . .4 to .
4.41%. 4:.+:-4 ,H. ....H4H;H4 .4._H .. . . .1t H4 41. .H . . 4... . 4 . 1 .04.:me4om t .. 4-. . w .
_ 4 4 ... . _ _ 1111-11 4 _ .. .4 4 .. . . . . 4
. . . .. . .4 ... . 4 . .H .. 4 . . ..4... H ._ 11f111|1 .. .4 4. . . H..... . . . . . . . C . ... f 4 . 4 . HI . 1
_:HHH.....1 ... .;11I1-... 44 1w4- fl 7.4, .. ..H... ..H_:4.. 4 . . . . . - .oz.cme4omqm.1H w11-11:+.-,w. .
H444 . ... . . . 4H-f4-4-- 4. . .. . 44. .4 .t-e.... a. ..... ... . ..: . 1-- __ . .. .
44. 4H.4.-r.,4. . ~4_.4. 4 4....” .4 .e. . H.4-.. - ,4 .-1... 114111.. . .4 _ . . . - .
.-HH;4 .. ..-4 r.4 . . . 44.44 4 ..-.H . . ...H4_..---.-e-....: . .. .. . 4. 4 4 4 4 4 . 44 4. ”..H. H , . .4.
H . .4 4-4 - . . .44 4 4 .. ps-zr. . 4-H_ H;:e.. 4.. . .... . 4 ._ . _ . 4... . 1.. .1 .
H.. .. 4 4 H . ..H. H H _ ....H.1.1-1... . .rl. .1.-14.1.8.1 _ 4. . H. H H.- H H H. . H . . . . . . . . . .44 4 4 H 1.. -.:—11111-9 11
r ..... .H 4 . . - 1 - . .l 1- _ . . 4 4 . _. . . 4. . . _ . . 4 . 1.01 -.Ia -flJIlflioo x .
H.....r ”NH: 4-.....1..1:__1.: ..--41.1.1. H . . 4. H .H. .H. 4. H .4 . . ..4 4.. . H H.. 4 . . 4 . H . . . L._1 “- 111- 111. .1.-114 . T1 4. . -4 .. .4 9
H H .H H.. H.. ‘H... .H. . H H H .H H.H .H.... . . 4 .. H _ H . H . .. 1 9 I 1 7.11.0... .f-Z‘ooti H H . H . . ....H h . H H
.4.... . .. .. . .4 . .. . 4. 4 H.°. Q 1.- .. . 41 .H. 1.....- ....- 4 4 .. H. , . H . . H 4. .4... . . m. .
H H H H {H.I. fl... 4. . .. . 4 . H. H. H. H. H 11H1111.1°- 101 1-- 1.- H H H _ 4.41. H . - . .. . H H ...41..... _. 4 H . .. H . . hw 4
.1.. . ._ . H . _ q . . . I- 1-911.191. :... 4 H . . _ . H ... . .4 . . .-.... . 1. 1 .
.4 41 H .H.. . .. H - H 1.111.. . 11. 111---111111114..11H11_.H.° . H....H 4 .H. .4 .H.. ... . 1... . H . H... 4- . ..H . 4 . - 41.1 1.1- .. .... . H
.. . 1. _ 4 4.4 . . . . .:.... . . . 4 ,
_:HH+HL.:49 .4 . fieH ...1eH .4:.. W 37111 .:. . _ .4tH 1H .4 H gr+.- .
H . .H. .. .. ’. ..H. . 4 4 .. H . .4 H .H--.1... .1 1.. ... . 4 «1911.7 ._ .. 4 . . 4 1 H .4. H .H. . .:.. _ H .. . 4 :Hi... ..
..-H. 43.4 .. ..r T11estir13 4 . .H44 H4 . .. . ._.44.e-4;;4 _ . oe-1111I1
- ..1...- 1 ..I .. ... . .. . .. .. . _ , .. 4. . 1
H H.. 419-.H1119+.131111H. 911.1H-1- ..1 4. H. . .4 H.. . . H . . 41H H H ... H .. . .. .4. - ... . 1 H.-._ H . H . 4. H ‘111110111111'1919- 1 cl 4 :.H ... . . . 4
.-H.-.4 H 4 .. H . . . H H- . H H . ... H..-..HH . 4 4 H . 4 H L . ,1 . H . _ , . 4 .4.
-f-..-. . ...1i. 4.: :E.:- J; 111 11. .-. . 4 . . - .
_ .H. H. -. . H . . .... ...H. . _ 0011. 9'4 r . . .. . . .. H . . ._ . ... ... .. . . 4 H... .o1 .. o .
r44--:...-w.:+: .. ;H4_ . - 4 fi1. .EHH,+.rs- .4 : g, . H4. H4 ._4- H . ., . 4 . p;:-..
H.. H. H.. .. .1 . . . H 1 111 4H H . H .H. H . . . . H .1.-.H.} 4. .H ...H H .4. 4 .. H .. . . H . . H.1._~. H. H . .1.... 4H .. 4.. 4 H @1191. TI... 1-H ... . .
. _ 44. - 4 . . H... 4 . . H-.. H ...... ... .H. .. ..AH-H. H.. .. H.. ... .. .4 - .- . .1.... 4 4 fill.Y.1-r11 1H1119411.-$.11-v. H. . . .. .. .
4. :.:14;--:.+..;:4 1.44.-:-.- .;H ..e ..: -H-.4.:4;._44 . -.-.: .-111H:1-... ... 4,.. a . .. 1 4
-..:If?13+i .:2. .14 ...... 7H. .3. . . .15t-11éi11- 44...? r.:. . z .?.;E. .
HxH f . .. ...... ... ‘ H H. .H1H 4.... . 4 H ..H. 1H. 1 . 4 q . ..I ..4 HT.H . H.I1.. . .. . .. . H . . H..:H ...H .H H.. .. .4. H 4.. .. ...H . . H... :::-H . . .. H . . H
...... .H. ... . ... H.. . 4 . . 4 _ e ... .. _ _ ..1... . ... .. - . ..H. . 2.4.1.7.-..4H . . l 1H1 » I 31 4
4..:..H.. . . - . .- 4 . 71-....- H.111H111 . . ......H. .. .. .4. - 4. H ... .. .H..- H- .- H..-.. .H. .....H. . .. 1. .4. H. ..H ...1..: 4...... H..-.H. H 11H 4 . id H
.H . . 4 . - o . 9 q 7 ... ... .4. . H9... ...- .- .. .7... . ... .H H H.. 4 -.+ H P H . H r r 4» ll . .9. H1 .9 . . 24.1.. . .. o
4 .: . .4. . . . ........ 4. .. .. . : :..-4.. ...-.4.. .. - 14. . . _ . . .4. .. . . , . . .
:1 .H..1.-: .-: ......rfe-.. .1.. .e :.. .:1H 4 . - . ... .1z-4. .-.-_ 4;:.. r. .1-;.. -:. :.-.. 4 .
.E- -. .41.. . .m... 1H.+ .1e..n . . . :-.1e_-::H.H,.4 A... :- .-H;:1..T---. 4 4 .
-5 %:.T .:T.J¥..;.%€ t.:.:..4...§§;. TWJ%%3T4. H..-.1 - 4 44W T r-..: e.WW:. .H
.4... -. .. . ...-..H. ..... H.. .. .H.. .. . . H . . . . .. .. .4.-5.4 ... . ..H. .- . . .- . . .. . . . . .. H- .1lr1.1.H.1-.- ..111. - ...H . . .. .. 1 H. . .. H....H . . . 44... . 1. .. ...1...... H. .. . 4
. ., w ...-.4 . H ..r 4- 4 .14--..- -_ ... -. . . .-.-.:.. 4.. .44 . .. . .. 4-.. . ._ 4 H .
... 9. ¢ 4 . .9.. . . . . . » . > . a . 14‘ ...1 .... H . .1. .. . . H - H J.- . - .... 1 4 . 4. .... I1- ..-
.. -4---4. ...--41-... . .4.. ..H. .. .. .4 . 4 H .. 4. . .. H. ....-. .. .. 4,. .-..11..--‘.-14,-.--. .
y L . 1.9. . . o .:.-...1.... ...1. H .. . . .. . . . . H..-1H1}... ... .... .... . ... :... .... . .1.... r 1 1 1!
. .. . , ...1...-..H. .... .. . . . . . 4
4;-iz_s;. _ ..1ftii- . .:H:e..H .- _ 4 1-...;4 2
--. .....gsr -... :.é.. .-.--..-;-+;p...H: H.. 1.:: . 1.. . ..11114 4
..... . 21:4 : - 4.. . 4. :4 .. -
...“.HM....... 4. . H. H. . . 11.11 .hW-H1Hlfll; . 2
H .1.

198765 4

3

2

15
the relatively short running time. However, this was consistent
with the other specimens tested, with respect to drop in compos-

ite output. Specimen number three was run at a constant

(I)
(‘1‘
'1
D
'3
10

level of 450 psi and had not broken at the time of writing.

Figures 4 and 5 are quantitative in a relative way; that
is, the trend is as shown. In the present investigation there
were two items which detracted from the data as an absolute
measure, and which could be corrected for in any subsequent
investigation:

1. The nonlinearity of the Sanborn deflection, due to biasing
the zero level off-center on the chart. Correcting this
would cause a greater fall—off in composite output than
Figures 4 and 5 indicate.

2. The calibrating voltage for the Sanborn recorder decreased
during the test on specimen number one and was not discov-
ered until the next specimen was being set up. Correcting
this would probably result in a more pronounced decrease

in composite output during the latter part of the run on

specimen number one.

Conclusions. On the basis of Figures 4 and 5, the follow-
ing conclusions were drawn:

1. A definite decrease in composite output occurred with
increasing cycles tested at a sufficiently high stress
level.

2. At a sufficiently low stress level, there was no change

in composite output with increasing cycles tested.

3.

J

A possible evidence of "healing" is indicated after the
rest period at B on Figure 4.

Figure 4 indicates that after damage at a medium stress
level and subsequent rest, increased damage may not
accrue until a higher stress level than that at which

damage originally occurred.

For Further Study. Some of the questions raised below

 

might be answered directly by a person well versed in three~

dimensional wave propagation theory. Nevertheless, the writer

deems the following items worthy of further consideration:

1.

DevelOp a method of isolating the location of fatigue dam-
age indicated by the method presented.

Eliminate standing wave effects, possibly by means of
white noise or by pulse methods, to facilitate item (1).
DetErmine the cause of or explanation for the sharp dip in
the frequency response curve at A on Figure 3.

Determine the cause of a random lateral shift in the fre-
quency response curve of anywhere from two to one hundred
cycles per second. This did not appear to be due to tem-
perature or to any change in instrumentation.

Eliminate the modulation effect on the output signal due
to the 1,800 rpm cycling of the fatigue machine. This
might be explained in terms of a nonlinear Operation of
the barium titanate crystal.

Determine the direct physical explanation for the drop in

the level of the output signal.

10.

I7
Investigate the possibility of applying dislocation theory
to fatigue damage in mortar and concrete.
In conjunction with item (7), investigate any connection
between energy loss, dislocation theory, and drOp in out-
put level.
DevelOp the method presented to the point where it can be
used to answer questions concerning cumulative fatigue and
undei- and over-stressing with a reasonable certainty.
Continue this type of investigation with limestone and

with plain concrete as suggested in the INTRODUCTION.

BIBLIOGRAPHY

Carlin, Benson. Ultrasonics. New York: McGraw-Hill, 1949.

 

A textbook covering the entire field of utrasonics, with
special emphasis on engineering applications. Gives many
details on instrumentation.

Desch, C. H., D. O. Sproule, and W. J. Dawson. "The Detection
of Cracks in Steel by Means of Supersonic Waves," Journal 9f
the Iron and Steel Institute, vol. 153 (1946), pp. 319-328.
Theoretical considerations and instrumentation required
for, and typical results from, using pulsed ultrasonic
system for flaw detection.

Heuter, Theodor F., and R. H. Bolt. Sonics. New York: Wiley,
1955
Textbook on ultrasonics in general, more theoretical
than Carlin.

Jones, R. "A Method of Studying the Formation of Cracks in a
Material Subjected to Stress," British Journal 9f Applied
Ph sics, vol. 3 (1952), pp. 229-232.
Describes use of pulsed ultrasonic system to detect
cracks in concrete as a result of static strength tests.

McCall, John T. "Probability of Fatigue Failure of Plain
Concrete," Journal 2: the American Concrete Institute,
vol. 30 (1958), pp. 233—244.

Previous work which led to the present investigation.

Nordby, Gene M. "Fatigue of Concrete--A Review of Research,"
Journal gj the American Concrete Institute, vol. 30
(1958), pp. 191-219.

Comprehensive review and bibliography of previous work
on fatigue of concrete.

Patton, R. G., and P. Hatfield. "Ultrasonic Techniques in the
Rubber Industry," Electronic Engineering, vol. 24
(1952), pp. 522-525.
Use of low frequency (50—250 kilocycles per second) CW
signal for flaw detection in rubber.

 

Sittel, Karl, Marvin Herman, and Robert C. Good, Jr. Early
Detection 9f Fatigue in Alumingm Alloys by Ultrasonics.
Wright Air Development Center Tech. Report 58-35, ASTIA
Document No. 155558 (1958)

Theoretical discussion of various possible means of
fatigue detection. Discussion of an efficient type_of
fatigue experiment, from the standpoint of early detection.
Good bibliography on fatigue and ultrasonics.

 

18

19

Symposium pp Ultrasonic lestin , ASTM Spec. Tech. Pub. No.

 

(1951)
Applications of ultrasonic methods, principally to
various types of flaw detection.

APPENDIX

EYPEPIMFNTAL TECHNIQUES
The first of the following sections presents an overall
view of the develOpment of the method used, citing the major
difficulties encountered. Subsequent sections give more details

on instrumentation components.

Evolution. The original idea was that one could set the
ultrasonic generator to an arbitrary frequency (one which would
give a large output), transmit a signal through the beam, and
use the signal attenuation as a measure of fatigue damage. The
first preliminary runs were rather discouraging in that there
seemed to be no regularity to observed changes in output signal.
By a laborious process of questioning, checking, and double-
checking, three factors eventually emerged as contributing to
the randomness of the signal changes: (1) crystal contact with
the specimen, (2) signal frequency, and (3) temperature.

Contact between crystal and specimen was originally main-
tained by pressure, through a grease film at the surface of con-
tact. This was all right for static checks; but not for
dynamic, long-term tests. Variations in signal due to change in
crystal contact could not be avoided. This difficulty was over-
come by cementing the crystals to the specimen.

That signal frequency affected output was obvious from
the start; but the subtleties involved were not so obvious.

Any frequency above 120,000 cycles per second was attenuated so

20

21
severely in being transmitted through the beam that it was
planned originally to use a constant signal frequency in the
vicinity Of 120,000 cycles per second which would give a large
enough signal to be [Jsable with the available instrumentation,
although it was recognized that a higher frequency would be more
desirable from the standpoint of "flaw" detection. Due to crys-
tal frequency response, specimen geometry, crystal location on
the specimen, and possibly other factors, peaks were obtained in
the output signal level at various frequencies in the range from
112,000 to 120,000 cycles per second. The one which gave the
highest peak was picked as the operating frequency. However, it
was found that the ultrasonic generator changed frequency as a
result of change in ambient temperature. Moreover, the effect
on the level of the output signal depended upon whether the
original Operating frequency had been at a true peak or at one
side of the peak; the effect on signal output if the original
Operating frequency had been below that of the peak selected
was just Opposite to the effect Obtained if it had been a higher
frequency than that Of the peak. This, plus the fact that one
division on the tuning dial Of the generator was equivalent to
about a 2000 cycle per second change in frequency, led to the
addition of trimmers on the generator tuning condenser so that
frequency could be precisely set, and to the introduction of the
electronic counter into the instrumentation so that frequency
could be accurately measured. This was still not adequate, how-
ever, and so the frequency response curve was introduced. It

was found that even at no load and constant temperature, if two

TO

2

‘_I

frequency response curves were made at different times. tn 1

“"7

1D

might coincide if one were superimposed on the other, yet be

shifted laterally from each other on the frequency scale. It wa

(f)

thus seen that checking at one particular frequency was uselss,
and the frequency response curve (see page O for a description)
was adopted as a means of comparison. The area under the curve
\mas taken arbitrarily as a quantitative measure Of output level. fsnzrf:
Aside from the frequency change noted above, temperature
affected output level in yet another way. The output Of the
crystal itself was temperature dependent. Before the frequency

response curve was introduced, a futile attempt was made at tem-

 

perature compensation by using four crystals (two on the test
beam, two on an unloaded beam) in a bridge arrangement. The idea
of using a thermistor arrangement for temperature compensation
was considered and discarded also. The complexity of factors
affecting the frequency response curve indicated that it would

be next to impossible to have all the components in a temperature
compensating network sufficiently well-matched as to give con-
sistent results. Accordingly, the temperature effect was avoided
by having the ambient temperature stabilized the same each time

a frequency response curve was run.

Crystals. The following factors represent major consider-
ations in using the barium titanate crystals.
1. Coupling. Lubriseal, a stopcock grease, was used to
couple the crystals to the beams for all static checks. For

the dynamic checks, three types of adhesive were tried:

INA

23
a) Lepages contact cement, which gave too much attenuation;
b) Duco cement, which would not maintain a proper bond cor-
sistently; and

c) Pliobond, which proved to be satisfactory.
Pliobond was applied to both the specimen and to the crystal.
The crystal was immediately pressed firmly against the specimen
so as to make a thin adhesive bond at the interface, and posi-
tioned while the adhesive was still reasonably fluid. The crys-
tal was held in position under pressure, usually overnight. The
Pliobond joint gave consistent results throughout the cycling,
and yet the crystal was easily removed by applying a steady
shear force across the joint.

2. Maximum current. If too much power is applied to a barium

titanate crystal, it will heat up and lose its polarization.
The output then drops to a negligible value. After some prelim-
inary checks, twenty-five milliamperes of crystal current was
taken as the maximum safe value. The multimeter, item (11) in
Figure l, was used to ensure that twenty-five milliamperes Of
crystal current was not exceeded when making an initial set-up
on a specimen. The final figure used during the test was set
well below twenty-five milliamperes. The multimeter, by monitor-
ing the transmitting crystal current, was also used to indicate
that the adhesive bond was consistent during the test. 0ne
crystal which did overheat in the early stages of the investiga-
tion was repolarized satisfactorily by the application of 4,500
volts DC overnight. Incidentally, the multimeter is shown

mounted away from the fatigue machine, since vibration would

 

24
damage the meter movement bearings.

3. Electrical contact. Several methods of making electrical
contact with the face of the crystal which was against the spec-
imen were tried. These included (a) two coats of silver paint
an the end face of the specimen, with a lead fastened to the
coating for electrical contact, and (b) a piece of brass shim
stock sandwiched between the crystal and specimen, with a portion
brought out for electrical connection. The best method, however,
was to cement a tiny piece of brass shim stock to the side of
the crystal adjacent to the face which would contact the speci~
men, carefully solder a lead to this piece of shim stock, then
COat the connection with silver paint and paint a strip to the
face of the crystal which contacted the specimen. The crystal
was then cemented directly to the specimen. The crystals are
furnished with an evaporated coating on each face; if this
becomes worn, it can be replaced satisfactorily by silver paint.
DuPont conductive silver, number 4817, was used with good
results.

4. Shielding. The crystal was covered by a copper cylinder
to eliminate pick-up from stray electrical fields present in the
laboratory.

5. Location. Each crystal was located on the neutral axis
of the end face of the specimen in such a position that a maxi-
mum signal output was obtained. The position was determined by
trial and error. It may well be that different positions would

be Optimum at different frequencies; this was not investigated.

IA‘ ,—

z
I
...

£1)

It was noted that the output level varied gre tly for differer+
positions on the end faces; and also, that for a given position
of the crystals on the end faces, the output level could be
altered significantly by placing another crystal or metal washer
or the like at various locations on the sides of the specimen

(coupled through a grease film). All of this indicates a pro-

nounced standing wave effect in the specimen. F3.» 92

ggneratgr. The ultrasonic generator as received had a
very "dirty" output signal. It seemed desirable to clean this r

up; and it was accomplished by

 

l. increasing the filtering on the bias power supply,
2. separating the variable power supply of the oscillator
plates from that of the screen grids, and
3. disconnecting the rectifier cooling motor (not needed for
the low power used).
In addition, half the power amplifier tubes were removed in order
to bring the low power operation (required because of the crys-

tal limitations) to a stable condition.

Recording. The Sanborn recorder was not too satisfactory
for long-term testing because of excessive zero drift and change
in gain. This was compensated for by adjusting zero level and

calibrating before each frequency response curve was taken.

Miscellaneous. Several times during the investigation
either a very fine wire grid cemented to--or a silver paint grid

painted on--the sides of the specimen was used in an effort to

2o
see if fine cracks developed before fracture. Any interruption
in the continuity of the grid stopped an auxiliary cycle counter
which was synchronized with the cycle counter on the fatigue
machine. No indication of surface cracks prior to fracture

were obtained by this method, however.