Rochester Institute of TechnologyRIT Scholar Works

Theses Thesis/Dissertation Collections

4-1-1980

Ultrasonic flaw detectionWayne Buchar

Follow this and additional works at: http://scholarworks.rit.edu/theses

This Thesis is brought to you for free and open access by the Thesis/Dissertation Collections at RIT Scholar Works. It has been accepted for inclusionin Theses by an authorized administrator of RIT Scholar Works. For more information, please contact ritscholarworks@rit.edu.

Recommended CitationBuchar, Wayne, "Ultrasonic flaw detection" (1980). Thesis. Rochester Institute of Technology. Accessed from

ULTRASONIC FLAW

DETECTION

by

Wayne A. Buchar

A thesis submitted in partial fulfillment of the requirements for the degree of Bachelor of Science in the School of Photographic Arts and Sciences in the

College of Graphic Arts and Photography of the Rochester Institute of Technology

April, 1980

Signature of the Author , ....... yy~y'n~.~u.cr~~ ............... . Photographic Science

and Instrumentation

John Carson Certified by ............................................. . Thesis Advisor

John Carson Accepted by .............................................. . Supervisor, Undergraduate Research

ULTRASONIC FLAW

DETECTION

by

Wayne A. Buchar

Submitted to the

Photographic Science and Instrumentation Division

in partial fulfillment of the requirements

for the Bachelor of Science degree

at the Rochester Institute of Technology

ABSTRACT

An ultrasonic microscope was designed and constructed

to detect surface flaw defects on materials which contain

natural discontinuities. The application of this flaw

detection system to via-fill defects of single layer

multiple layer ceramic substrates was considered. The

microscope was evaluated for its depth sensitivity and

resolution capabilities.

The results of the experiment are that the ultrasonic

system has repeatable depth sensitivities of better than

ten microns, and resolution capabilities on the order of

3.3 cycles per millimeter. The application of the flaw

detector to ceramic substrate via-fill defect inspection

was verified.

ACKNOWLEDGMENTS

I would like to thank John Carson for being my thesis

advisor- I greatly appreciate the efforts of Richard

Norman, who is responsible for the construction of the

aluminum lens and alignment collar- I would like to thank

Keramos Incorporated of Indiana for the piezoelectric

transducer used in this project. I would also like to

thank the Central Intelligence Agency for its generous

funding of this thesis.

11

TABLE OF CONTENTS

CHAPTER I

PAGE

la

A. INTRODUCTION

1. Problem

2. Past System Attempts

3. Aim

CHAPTER II .

1

3

4

5a

A. BODY *o

1. Background .

5

2 . Theory......7

3. Mechanical .

10

4 . Lens ...... 11

5. Button...... 15

6. Transducer .... 16

7. Mounting the Transducer -17

8. Water Supply18

9. Table 19

10. Test Samples 19

11. Transmitter -20

12. Receiver .... 24

111

TABLE OF CONTENTS (continued)

PAGE

13. Detection Circuit ..... 26

14. Calibration 27

B. RESULTS AND DISCUSSION 29

1. Focus 29

2. Depth Sensitivity 30

3. Resolution ....... 31

ENDNOTES FOR CHAPTER II 33

CHAPTER III 34a

A. SUMMARY AND CONCLUSIONS .... 34

LIST OF REFERENCES 36a

APPENDIX 38a

IV

LIST OF FIGURES

1. Wave Diagram ......... 9

2. Lens Construction ....... 12

3. Wave Propagation Under Lens ..... 3

4. Button 4

5. Transducer and Mount ....... 18

6. Basic Transmitter Circuit ..... 21

7. Filter Circuit 25

8. Through Focus Series ....... 29

9. The Ultrasonic Microscope ..... 39

10. The Ultrasonic Lens 40

11. Transmitted and Reflected Pulses .... 41

12. Water Supply Layout 41

13. Table 42

14. Test Sample One 43

15. Test Sample Two 43

16. Transmitter Circuit ....... 44

17. Receiver Circuit ........ 45

v

LIST OF FIGURES (continued)

18. Detection Circuit. ....... 46

19. Calibration Circuit ....... 47

20. 2220 Micrometers per Cycle, Resolution

Target at Threshold 1 48

21. 932 Micrometers per Cycle Resolution

Target at Threshold 2 (Ultrasonic) ... 49

22. 932 Micrometers per Cycle Resolution

Target (Optical) 49

23. 678 Micrometers per Cycle Resolution Target

at Threshold 2 (Ultrasonic) 50

24. 678 Micrometers per Cycle Resolution Target

(Optical) 50

25. 610 Micrometers per Cycle Resolution Target

at Threshold 2 (Ultrasonic). .... 51

26. 610 Micrometers per Cycle Resolution Target

(Optical) ......... 51

27. 398 Micrometers per Cycle Resolution Target

at Threshold 2 (Ultrasonic ..... 52

vi

LIST OF FIGURES (continued)

28. 398 Micrometers per Cycle Resolution Target

(Optical) 52

29. Scope Trace of Original Signal from the

Transmitter ........ 53

30. Scope Trace of Filtered, Amplified Signal. . 53

31. Scope Trace of Digital Signal 54

32. Scope Trace of Reference and Signal Pulses. . 54

33. Scope Trace of Transmitter Exciting Pulse . . 55

Vll

LIST OF TABLES

1. Depth Measurement Data ...... 30

2. The Lead Zirconate Titanate Transducer

K-350 (PVT5A) 38

Vlll

la

CHAPTER I

A. INTRODUCTION

1. Problem

In industry today there exists a growing need for

a device which could quickly and accurately detect

surface flaws on materials which contain small natural

discontinuities. One such example exists in the detec

tion of non-filled via-holes in a ceramic layer of a

multiple layer ceramic substrate.

A multiple layer ceramic substrate consists of

thirty layers of thin ceramic sheets assembled into a

package containing an integrated three-dimensional net

work of electrical current passageways.

A single multiple layer ceramic substrate can re

place thousands of conductive circuits used to transport

electrical signals between separate integrated circuit

chips. The manufacturing process to produce these three-

dimensional multiple layer ceramic substrates begins by

casting a ceramic material into large sheets about two

hundred and fifty microns thick. The large castings

are cut into one hundred and seventy millimeter square

sheets. These single ceramic sheets are referred to as

greensheets, as they are made of a ceramic material

which has not yet been kiln fired. Fifteen thousand,

one hundred and twenty-five micron diameter holes are

precisely punched into the greensheets at various pre

determined locations. A thin metal mask containing the

desired surface land-patterns for each individual

ceramic layer level type, is aligned over the greensheet.

A metallic past is spread over the mask, being forced

down through the mask, filling the punched via-holes,

and simultaneously producing a screened image of the

mask on the surface of the greensheet. The mask is

removed, and the greensheet is dried in an oven to

harden the paste. The greensheet is then sent to

inspection. When the thirty desired greensheet layers

for a particular multiple layer ceramic substrate are

produced and inspected, they are stacked, in proper

order, on top of one another. The assembled packages

are put under high pressure and placed into a kiln and

fired. The thirty layers cohese and the metallic paste

flows. The previously screened land patterns internal

to the substrate become the horizontal passageways for

the electrical current and the filled via-holes become

the vertical electrical connections between each previous

layer and the surfaces of the multiple layer ceramic

substrate. The integrated circuit chips are then sol

dered on top of the substrate.

The problem concerned with here arises at the in

spection step after screening. The bottom of the green

sheets must be inspected for via-hole fill. If a

via-hole is not full to specifications, that is, if the

metallic paste is farther than 40 microns from the bottom

of the greensheet, it must be repaired. If the via-hole

is not filled it will cause an open circuit in the

final multiple layer ceramic substrate. Presently,

inspection is done manually. The amount of personnel

required to perform this 100% inspection is tremendous,

and is a major contributing cost of the process. The

future of the multiple layer ceramic substrate demands

a mechanical via-defect detection system.

2. Past System Attempts

A system designed for detection of non-filled vias

by Intec Corporation of Connecticut consisted of a

scanning Helium-Neon laser, photomultiplier tube receiver,

and a single direction moving table. The laser was

mounted25

from normal incidence at a distance of18"

from the table below. The photomultiplier tube receiver

was mounted at a75

incident angle at a distance of 16".

The laser beam scanned in the x-direction while the

table moved in the y-direction below. The greensheet,

located on the table, bottom side up, was scanned in

this manner. It was theorized that if a non-filled via

was located under the incident laser beam, a reflection

would occur from the inside wall of the via. The photo

multiplier tube receiver would be triggered from the

reflection and the defect would be found. Due to the

properties of the ceramic substrate, little specular

reflection reached the photomultiplier tube. On the

contrary, specular reflections were received from the

metallic particles in the past of non-defective vias.

It was also found that via edge characteristics, when

the via holes were punched, produced undesirable

effects. The system failed the detection requirement

specifications necessary. The system described, at

maximum threshold, produced 96% detection of via hole

defects, with 50% overkill per sheet.

Aim

The application of non-destructive ultrasonic

defect detection as a means of detecting via-fill

defects in a ceramic substrate layer is considered

here. The scope of this thesis encompasses the design

and verification of a depth measuring ultrasonic

microscope and its application to the via-fill

detection problem.

5a

CHAPTER II

A . BODY

1. Background

Sound waves, perceivable by the average human ear,

range from about twenty to twenty-thousand cycles per

second. Sound waves of higher frequencies are referred

to as ultrasonic waves. Ultrasonic waves have been

produced as high as three gigihertz. This frequency

corresponds to a wavelength of 520 manometers in water,

slightly shorter than 550 manometers, the optical wave

length at the center of the visible spectrum (green

light).

Scientists in the past forty years have experi

mented with these ultrasonic waves. Many applications

such as sonar, flaw detection, and microscopy have been

developed .

In the above applications of ultrasonic waves, the

system set-ups are similar. They all make use of ultra

sonic waves on the pulse-echo mode. Basically the design

consists of a transmitter which produces an electrical

signal that causes a transducer to oscillate, producing

the ultrasonic waves. These waves propagate through the

medium in contact with the transducer until a boundary

is reached. "When a wave traveling through one material

impinges on a boundary between it and a second medium,

6

part of the energy travels forward as one wave through

the second medium while part is reflected back into the

first medium, usually with a phasechange"

(Ref. 1).

The amount of energy reflected is a function of density

and wave velocity at the time of intersection. This

characteristic is referred to as the specific acoustical

impedance of the medium. The amplitude of the reflected

wave is related to the incident wave by: (Ref. 2.)

R-i-

RQAr = -

R1+

R2

where: L =

p C,1 l i

R2=

P2C2

p= density of material

C =

velocity in medium

A = ratio between reflected and incidentr

amplitudes .

The energy of the wave is proportional to the ampli

tude squared, as long as the wave travels in the same

medium. (Ref. 2. )

Therefore ;

R_

RPll

"

op1c1

+p2c2

where: R = reflected energy

R0= incident energy

In the case of an air-solid boundary, almost 100% of the

incident energy is reflected.

The energy passing through the boundary proceeds

until another boundary is encountered. The reflected

wave propagates back through the medium in which it

came and impinges on the transducer. The transducer

oscillates producing an electric pulse. The transducer

at this time is coupled with the receiver. The time

between the pulse is sent out and the reflection received

is directly related to the thickness of the material in

which the wave travels. The wave travels the distance

twice, this must be considered. As the object is passed

under the transducer the reflections are electronically

deciphered and a location versus depth image is displayed.

2 . Theory

The system is designed in this thesis utilizes the

pulse-echo mode of ultrasonics, applied to surface flaw

detection. The set-up consists of a piezoelectric Lead

Zirconate Titanate ceramic transducer of diameter 0.5

centimeters and fundamental frequency of 4 megahertz.

The transducer is epoxied to the flat end of an aluminum

rod. The other end of the rod contains a hemispherical

lens surface. The lens surface is located over the

8

ceramic substrate and submerged in water. At first,

the transmitter is connected to the transducer. The

transmitter produces a minus 200 volt pulse across the

plates of the transducer causing the piezoelectric

transducer to resonate at its fundamental frequency

of 4 megahertz. The planar ultrasonic waves produced

propagate through the aluminum at a velocity of 5200

meters per second until they impinge upon the hemi

spherical lens surface. A reflection occurs at this

aluminum-water interface. The reflected wave travels

back to the transducer. At this time the transducer

is electronically coupled to the receiver. The ultra

sonic energy contained in these reflected waves causes

the transducer to oscillate, producing an electrical

differential. This voltage pulse produces a time

reference to which the final reflected signal is

compared. The portion of the ultrasonic energy contained

in the plane waves transmitted at the aluminum-water

boundary, is focused, detraction limited, to the surface

of the ceramic substrate.

"Because the velocity of sound is much greater in

aluminum than in water, an ideal lens with a single

sherical surface can be used without spherical

aberration."

(Ref. 2.)

Reflected spherical waves from the surface of the

ceramic and metallic paste are collected by the same

9

lens, converted back to planar waves, and are transmitted

back to the transducer, which is still connected to the

receiver.

Electronics separate the returning pulse from the

reference pulse. The time difference between these two

pulses is proportional to the distance the wave traveled

through the water beneath the lens, plus the return time

in the aluminum lens barrel. This time interval is con

verted to distance by originally setting a"zero"

depth

with the electronics of the detector circuit. Changes

in signal propagation time can then be referenced to this

zero setting, assuming that the returning pulse is re

flected within the depth of focus of the lens, and be

converted to absolute depth.

BRASS

TUBING

WATER

/TRANSDUCER

i \ULTASONIG

-WAVE5

5U&5TATE

Figure 1. Wave Diagram

10

The ultrasonic microscope is also designed to be

operated in a"go-no-go"

arrangement. The allowable

change in propagation time of the returning signal can

be set, and depth distances greater than this level will

cause the internal electronics to signal a defect. In

this way the surface of the ceramic substrate layer can

be scanned and defective vias can be located and

repaired.

Mechanical

"Most conventional ultrasonic systems use several

feet of shielded cable to connect the transducer and

relatedelectronics."

(Ref. 3.) "The capacitance of

the cable typically degrades the rise time and the

amplitude of the pulse across the transducer, and

degrades the bandwidth of the received echos. The

cable is often a source of noise, which decreases the

effective dynamic range of thereceiver."

(Ref. 4.)

In order to relieve these difficulties the com

ponent layout of the microscope is such that the

distance the excitation pulse travels through the coax

cable is less than one inch in length.

In order to obtain the desired set-up, the

receiver and transmitter were developed small enough

to fit in the focusing barrel of the microscope. The

detection circuit, however, is located in the head.

11

The amplified signals travel from the receiver to the

detection circuit through shielded cable.

The body of the microscope was constructed from

1/4 inch Plexiglass and all joints were connected with

an acrylic solvent. Plexiglass was chosen as it is

non-conductive (electrical), waterproof, and has a low

specific acoustical resistance. The microscope was

attached to a sheet of particle-board, at its base.

Focusing adjustment of the ultrasonic spherical waves

is accomplished by the fine focusing mechanism attached

to the focusing barrel of the ultrasonic microscope.

(_See Appendix, Figure 9.)

Lens

The ultrasonic lens was constructed from an aluminum

rod of radius one inch. The rod was turned down on a

machine lathe to a diameter of 20 millimeters. The rod

was then cut to 12.5 millimeters in length. One end of

the cylinder was turned flat while the corners were

beveled on the other end leaving a 65 millimeter plateau

in the center of the cylinder.

12

P/ //

Lf STEEL

\ /BALI-// PRCS5

t^ ;

V )

Figure 2. Lens Construction

The concave hemispherical lens surface was created

by pressing a steel ball of 3/8 inch diameter into the

surface of the plateau. Alignment of the steel ball

was obtained by an alignment collar- The collar was an

aluminum disk into which a 1/4 inch diameter hole was

drilled. The back end of the disk had a recessed center

of diameter slightly greater than the diameter of the

aluminum lens barrel. This indentation allowed the

collar to be positioned over the end of the aluminum

13

barrel. The ball was then positioned in the center of

the collar, in the hole provided. This allowed slight

contact between the steel ball and the lens barrel

beneath. The assembly was then placed in a press. The

ball was forced into the top of the lens barrel slightly.

The collar was then removed, and the steel ball, placed

in the pre-formed indentation, was pressed the remainder

of the way into the aluminum producing the hemispherical

lens surface (see Appendix, Figure 10).

The focal length of an ultrasonic lens is given by

(Ref. 3. )

f = CR0/(l-y)][l- |y2(l-cos aQ)]

where : Rn= radius of curvature

=.476 cm.

Vw 1440nnn

=

5200=

-27? C'm-

AIj

a.= half angle

=.737 cm.

the focal length of the described lens becomes:

f =[.476/l-.277][l- |(.277)2(l-cos(.737)]

f =. 65 cm.

14

The f -number of the lens is given by:

.65

.65

= 1

where d is the diameter of the lens. The length of the

the aluminum lens barrel was determined by realizing

that multiple-internal reflections would exist from the

lens surface and the transducer. The returning plane

wave signal from the ceramic must reach the transducer

in absence of such reflections (see Appendix, Figure 11)

The beveled edge on the bottom of the lens barrel was

designed to reflect unwanted scattered ultrasonic

reflections away from the lens surface (see Figure 3).

Figure 3. Wave Propagation Under Lens.

15

5. Button

The button is the device which supports the ultra

sonic lens assembly and provides isolation between stray

ultrasonic radiation and the sensitive electronics of

the receiver. The button consists of a hard rubber

cylinder (a rubber stopper was used) of radius 5 centi

meters and of length 4 centimeters. A 5 millimeter

diameter hole was drilled through the center of the

rubber cylinder to allow passage of the coax connecting

cable. A hole of 18 millimeters in diameter and 20 mill

imeters deep was cut into the bottom of the rubber button

to accept the lens and transducer assembly (see Figure 4),

rubber

BUTTOM

LENS

Figure 4. Button,

16

6. Transducer

The transducer chosen was a Lead Zirconate Titanate

piezoelectric crystal, type PVT5A. The crystal was

chosen for its wide range of applications and its avail

ability (see Appendix, Table 2). The fundamental

frequency of 4 megahertz was chosen because many

computer clock crystals oscillate at this frequency,

an important consideration had another type of trans

mitter, other than the one chosen, been used. The four-

megahertz frequency also corresponds to a wavelength of

360 micrometers in water. "A 375 megahertz scanning

acoustical microscope built by Rolf D. Weglein of Hughes

Research Laboratories has a measured resolution of 1.67

micrometers."

(Ref. 5.) This value corresponds to a

spatial frequency of 3340 cycles per millimeter, the

wavelength of the 375 megahertz waves in water- The

set-up described in this thesis is very similar to Rolf

Weglein'

s, therefore defects of 180 micrometers in size

should be able to be detected with this system. Although

the diameter of the via holes in the ceramic substrate

layer are smaller than 180 micrometers (125 micrometers),

the application of ultrasonics to the problem of via-

hole defect detection can still be shown. A transducer

of higher frequency, six megahertz, would be necessary

to resolve the via holes.

17

7. Mounting the Transducer

The piezoelectric transducer was obtained without

the leads attached. The transducer surfaces, which were

plated by Keramos Incorporated of Indiana, with silver,

were first polished with a very fine abrasive. The

transducer was attached to the back of the lens barrel

using an epoxy resin. The resin was mixed with a fine

aluminum powder. The mixture was then applied to the

center of the flat end of the lens barrel. The amuminum

powder was used not only to help match the acoustical

impedance of the transducer to that of the lens, but

also to act as a conduction medium between the bottom

plate of the transducer and the lens surface.

The ground lead of the coax connecting cable was

connected to the outside of the lens barrel. The signal

lead, consisting of a short length of fine stranded

wire, was carefully tack-soldered to the top plate of

the transducer. The other end of the signal lead was

soldered to the center wire of the coax connecting cable.

A strain relief line was attached between the lens and

the coax cable to prevent damage to the transducer -

The lens-transducer assembly was then inserted into the

bottom of the button and tack epoxied. (See Figure 5.)

18

GROUNDLEAD

RELIEF LINE

t- N\"T RAW5DUCER

LENS

Figure 5. Transducer and Mount

8. Water Supply

Water is necessary between the hemispherical lens

surface and the ceramic substrate as described previously.

Water is transported to the lens and object via four

brass tubing ducts located between the lens and the

sample. The 5/32 inch brass tubing jets pass through

the head of the microscope to the top where they are

connected to a four-terminal fishtank air-valve,

attached to the body of the microscope. The valve

controls the amount of water flow of each individual

jet on the sample. Water is fed to the valve input

19

by gravity from a basin located above the microscope.

(See Appendix, Figure 12.)

9. Table

The table was constructed from clear 1/4 inch

thick Plexiglass. The table consists of two main

sections, a large stationary rectangular tray and a

smaller square tray. The smaller tray slides, in one

direction, within the rectangular tray. A micrometer

position indicator is attached to the moving tray.

The ceramic substrate layer can be placed in the small

tray and can thus be scanned, in one direction, under

the lens located above. (See Appendix, Figure 13.)

10. Test Samples

Samples were made to determine the depth measure

ment accuracy, repeatability, and resolution, of the

ultrasonic microscope.

The first sample consisted of an array of various

thicknesses of metal strips arranged on a stainless

steel base (see Appendix, Table 14). The thicknesses

of these materials were previously known. Optical

microscope depth measurement readings were also conducted

on these samples.

The second sample, to test the resolution of the

microscope, consisted of a three-dimensional bar target

20

of varying spatial frequency (see Appendix, Figure 15).

The bars of the resolution target were cut from aluminum

tape 125 microns thick. The bars, of various width,

were carefully positioned under an optical microscope,

on a stainless steel base. The spatial frequencies of

the bars were optically measured. Plots of height versus

location, of the bars, were generated (see Appendix,

Figures 20-28).

11. Transmitter

"There are two different methods commonly used for

exciting ultrasonic transducers in the pulse echo mode.

The first involves the generation of a medium voltage,

pulsed radio frequency sinusoid using a gated oscillator,

and the second involves the generation of a high voltage

spike by a sudden discharge from a capacitor charged

to several hundredvolts."

(Ref. 6.) The second method,

referred to as shock excitation, was the method used.

The basic circuit is shown in Figure 6. (See following

page. )

21

Figure 6. Basic transmitter Circuit.

The circuit consists of a D.C. voltage source V,

s

storage capacitor C, damping resistor R, and silicon

controlled rectifier SCR1. Assuming SCRl is initally

open, capacitor C is charged to supply voltage Vs

through resistor R . SCRl is triggered at its gate bys

an input pulse allowing C to discharge through SCI to

ground potential. A pulse voltage then occurs across

the damping resistor R. When the charge in capacitor

C has dissipated, SCRl turns off as the current through

it has dropped below the necessary holding current.

The pulse circuit is initiated by an input pulse

at the gate of the SCR. The SCR experiences a finite

turn-on time before its anode reaches ground potential.

22

The voltage across the damping resistor is given by:

-V -t/RC

VQ= -S RC(l-e ).

on

As can be seen from the above equation, as the turn-on

time of the SCR increases the output voltage V"n decreases.

"From the frequency domain viewpoint, the shorter the

turn-on time of the SCR, the more higher frequencies

are contained in the falling edge of the output pulse

voltage and the easier it is to excite proportionally

higher frequencytransducers."

(Ref. 7.) One of the

features of using an SCR circuit as described here is

that there is no need for any external circuit to turn

off the SCR. The SCR is designed such that it turns

off when the magnitude of the anode current falls

below the required holding current I., thus:

V

IT< Ih

s

where I, is the SCR holding current. It is also desirableh

in pulse-echo systems to have the damping resistor R

small so that the energy of the pulse is dissipated

quickly to allow echos from close flaws to be received.

"The lowest value of damping resistance depends essen

tially on the maximum pulse current that the SCR can

deliver." (Ref. 8.) Therefore:

23

V

R .

=*

min Imax

The average electrical power delivered to the transducer

is given by:

1 9P = CV f^AVE 2 s

1

where f is the pulse repetition frequency of the

transmitter.

The transmitter circuit selected was designed by

J. G. Okyere and A. J. Cousin of the University of

Toronto (see Appendix, Figure 16). The circuit basically

consists of a 555 integrated circuit timer running in an

astable mode at a repetition rate of 1000 cycles per

second, making up the rate generator. The inverter

(CD4049) and power transistor (2N2270) drive the dual

secondary transformer, making up the pulse amplifier.

The two separate secondary windings provide isolation

between the gates of the two series connected SCR's .

"In order to obtain minimum turn-on time of the series

connected SCR's, it is desirable not only to trigger

the gates with a strong drive but also to produce a

triggering pulse with a fast risetime."

(Ref. 9.)

"To accomplish this, the slowly rising pulse produced

at the secondary of the pulse transformer is shaped to

give a fast rise-time using a pulse sharpening circuit

first introduced by GeneralElectric."

(Ref. 10.)

24

The SCR 2N4203 was used because of its high peak

blocking voltage (700v), its fast turn-on time (100ns)

and its high current switching capability (200 Amps).

The voltage across the damping resistor of 33 ohms

becomes:

_200

v-6 -100/(33)(.012E-6)

V0=

TM(ns)(33)(.012E

(1-e

VQ=-177 volts.

By selecting one of four damping resistors, output

pulses of different amplitude and duration can be

produced.

12. Receiver

The receiver circuit operates continuously when the

microscope is turned on (see Appendix, Figure 17). The

return (echo) signals are converted to electrical signals

by the piezoelectric transducer. The receiver is

directly coupled to the transducer at all times. Un

fortunately, the allowable input base voltage of trans

istor T, is only a few volts, and thus some kind of

protection must be provided against the transmitter

200 volt exciting pulse. Diodes Dl and D2 limit the

amplitude of this "mainbang"

spike and prevent damage

to the receiver. The circuits consisting of Tl, T2 and

T3 amplify the 4 megahertz signal and filter out all

25

other signals. Transistor T4 and associated components

also serve to amplify the 4 megahertz signal, but the

amount of amplification is determined by the setting of

sensitivity control Rl. The amplification of the

receiver is also varied by the input D.C. bias voltage.

The range of the variable voltage regulated by integrated

circuit 723 and resistor R2, is from 2-7.3 volts D.C.

Filtration of the 4-megahertz input signal of

unwanted signals is accomplished by using two pass band

filters, Fl and F2 . The basic circuit is shown in

Figure 7.

L G

o--^m

r\U>

SK

O- 1 1 r\W 1 : w

Figure 7. Filter Circuit.

26

Series resonance is established at the fundamental

frequency of 4 megahertz between inductor L and

capacitor C. At frequencies higher or lower than

resonance, the impedance of the L-C combination increases

over that of R, resulting in a voltage drop across R.

Only the 4 megahertz signal passes to the next ampli

fication stage.

13. Detection Circuit

The amplified 4 megahertz signals are sent from

the receiver to the input of the detector circuit (see

Appendix, Figure 18). The analog signal is converted to

digital by Schmitt trigger IC2. A threshold is set so

that only the reference signal, and the ultrasonic

signals reflected from the ceramic substrate, are passed.

Noise from the original transmitter pulse is removed by

the logical AND performed by IC3A between the output of

the Schmitt trigger and the output pulse from IC1. IC1

is triggered by the astable 555 integrated circuit timer

of the transmitter. The output of IC3A is fed into

monostable multivibrator IC6, and converted into two

individual pulses. These pulses are separated by IC3

and IC5 resulting in only the reflected signal pulse

being outputted by IC3. This signal is further sharpened

by IC2 and is fed into IC4B. At the same time the

inverted output of IC5 is logically AND with the output

27

of IC6 resulting in only the reference pulse at the out

put of IC4A.

The output of IC4A, or reference pulse, then

triggers IC7 producing a pulse of variable time dura

tion. This output pulse is logically AND with the signal

pulse from IC2 at IC4B. If an output pulse results, IC8

produces a pulse of one millisecond duration. The high

state produced by IC8 causes light emitting diode LI to

glow.

If the result of the AND performed at IC4B does

not output a pulse, IC8 will not be triggered and LI

will not glow. This will be an indication that the

signal pulse occurred, after the output of IC7 went low.

The length of the pulse duration of IC7 can be increased

by increasing the resistance of VRl . This is done until

LI just begins to glow.

14. Calibration

Calibration of the ultrasonic microscope to

absolute depth distances is accomplished with selector

range switch VR2, and variable resistor VRl. As men

tioned in the detection analysis section, monostable

multivibrator IC7 produces a pulse of variable time

duration related to the capacitance across pins 10 and

11, and variable resistance VRl. To obtain precise

and repeatable depth measurements, the change in pulse

28

duration of IC7 by adjusting VRl should be small. To

allow measurements to be made over a large range,

selector switch VR2 is coupled in series with potentio

meter VRl. Thus, VR2 becomes a fine adjustment at that

range. The three ranges available can also be adjusted

by potentiometers VR3, VR4 and VR5 (see Appendix, Fig

ure 19). To obtain a depth measurement the microscope

is focused by setting the zero on the top surface of

the substrate layer. When in focus, L.E.D., LI glows.

A change in depth is indicated when LI fails to glow.

The resistance of variable resistor VRl is increased

until LI just glows. If L2 continues to remain off,

the selector switch VR2 is rotated to position two.

VRl is again rotated until LI glows. This procedure

is continued, if necessary, until the limit of the

instrument is reached. This will occur when the

electrical signal falls below the threshold set by the

detection circuit, due to the finite depth of focus of

the microscope. The variable resistance dial of VRl is

calibrated in micrometers. When the L.E.D. begins to

glow, a reading of depth is taken from the position of

VRl, and range selector switch VR2 . The dials of VRl

and VR2 were calibrated using an oscilloscope realizing

that the velocity of sound in water is 1440 meters per

second.

29

B. RESULTS AND DISCUSSION

Focus

The signal to noise ratio was small for some

substances such as plastics, wood and rubber. This

is of no surprise as the amount of reflection at a

boundary is proportional to the difference in the

sound velocities at the interface. Ceramic, similar

to that of ceramic substrate layers, produces very

low signal to noise ratio, about 1.5 to 1. The

metallic paste, however, has similar reflection

properties to that of aluminum, about 5 to 1. The

signal to noise ratio is also a function of focus.

The signal to noise ratio of an ultrasonic wave

reflecting from a stainless steel boundary, as a

function of propagation time is shown in Figure 8.

Figure 8. Through Focus Series

30

The threshold of the microscope was set at about

the 4 to 2, signal to noise, level, utilizing a depth

of focus of about 1.5 microseconds, corresponding to

a depth distance of two millimeters.

2. Depth Sensitivity

Depth measurements from the first sample are

shown in Table 1.

TABLE 1

Depth Measurement Data

1

Trial

2 3 X S Known

OpticallyMeasured

205 190 200 198.3 7.6 203 193.0

140 135 140 138.3 2.9 152 177.8

115 115 120 116.7 2.9 127 121.2

100 90 90 93.3 5.8 102 109.2

75 75 75 75.0 0.0 76 78.3

60 60 65 61.7 2.9 38 50.8

The repeatability of the microscope in measuring

depth was very good. Accurate samples, however, were

a problem. The exact heights of the different levels

were difficult to measure optically. The microscope

was, however, able to measure 40 to 125 micrometer

31

depths with repeatability error less than six percent,

required for defect detection on a ceramic substrate

layer .

3. Resolution

The resolution of the system as described was no

better than 0.5 cycles per millimeter (see Appendix,

Figure 20). This value was extremely lower than the

2.8 cycles per millimeter predicted. This difference

was attributed to the threshold setting of the

detection circuit. It was found that the amplitude

of the returning signal fell greatly when the edge of

a bar was encountered. The amplitude of the resulting

signal was too low to trigger the detection circuit.

To relieve this situation the threshold level was

lowered. Unwanted noise from internal reflections

in the lens broke the new threshold level, upsetting

the present detection curcuit . The resolution of the

microscope, however, increased drastically. Measure

ments at this threshold level were done directly from

the oscilloscope cathode ray tube. Bars of spatial

frequency as high as 2.5 cycles per millimeter were

resolved (see Appendix, Tables 21 - 24). The detection

circuit can easily be modified to remove the unwanted

noise (see Appendix, Figure 18). The microscope could

then be operated, independent of the oscilloscope, at

32

this new threshold level. Comparisons between the

optical microscope measurements, and the ultrasonic

microscope measurements, reveal many differences

(see Appendix, Figures 21-28). It should be

pointed out that two totally different wavelengths,

and wave types, are at work here.

33

ENDNOTES FOR CHAPTER II

1. Carlin, Ultrasonics (N.Y., N.Y., 1960), pp. 9.

2, 5. Maugh, Science, vol. 201, 2220 (1978).

3. Baum, Tr . Am. Acad. Opthth. Otalaryn vol. 69,

943, (1965).

4. Myers, Thumin, Feldman, deSantis, Lupo,

Ultrasonics, Mar., 87 (1972).

6, 7, 8. Okyere, Cousin, Ultrasonics, Mar., 81-82,(1979).

9, 10. SCR Manual, 5th Ed. (Syracuse, N.Y., 1972).

34a

CHAPTER III

34

A. SUMMARY AND CONCLUSIONS

The ultrasonic microscope system described in this

thesis is capable of accurately and repeatably detecting

depth differences of 25 to 2000 micrometers. The system

possesses resolution capabilities of up to 2.5 cycles per

millimeter, corresponding to the wavelength of the ultra

sound in the medium in which the waves propagate. Appli

cation of ultrasonic flaw detection as a means of

detecting surface flaw defects has been shown.

Defects in ceramic substrate layers, caused by non

filled vias can be found. Considerations must be made,

however, if this instrument is to be directly applied to

multiple layer ceramic substrate manufacture. Due to the

characteristics of the ceramic, little reflection occurs

at the water-ceramic interface. A strong reflection does

occur from the metallic paste in the vias. The system

described, after being modified to operate at 6 megahertz,

could be operated in the"go-no-go"

mode. Signals would

only be detected from the vias. If a via depth should be

below the specified 40 micrometer distance, the returning

signal would be detected, and the defect would be found.

A means of calibration check would have to be considered

to assure that the reference setting is correct in

35

relation to the ceramic substrate height. Many methods

could be utilized to accomplish this. A method of two-

dimensional scanning would also have to be developed.

The application of ultrasonics to via-fill defect

detection on a single ceramic layer of a multiple layer

ceramic substrate does exist. The ultimate ultrasonic

tool, making use of those considerations mentioned,

could be developed, and the via-fill detection problem,

of multiple layer ceramic substrate manufacturing, solved.

36a

LIST OF REFERENCES

36

LIST OF REFERENCES

REF. #

20 Beaver, W., Dameron,D.

,and Macovski, IEEE on

Sonics and Ultrasonics. Su-24. 4 (1977)235-236.

2 Carlin, Ultrasonics. New York: McGraw-Hill Pub.

Co., Inc. (1960).

21 Chart Recording Depth Sounder. Heath Company,Benton Harbor, Mich. (1976).

7 Cheney, S. P., Lees, S., Gerhard, Jr., Kranx, P. R.,

Ultrasonics. 11 (1973) 111-113.

15 Devey, P. Wells. Science America. 238 (1978)98-104.

13 Glickson. Basic Ultrasonics. J. F. Rider, Publ,Inc.

,N.Y. (1960).

19 Kossoff, G., Robinson, D., Garrett, W.

J. Acoust. Soc. Am. 44.5 (1968) 1310-1318.

11 Mason. Piezoelectric Crystals and Their

Application to Ultrasonics. D. Van Nostrand

Co .,Inc .

18 Mattiat, 0. J. Acoust. Soc Am. 25.2 (1953) 291-

296.

1 Maugh, T. H.,III. Science. 201 (1978) 2220-1113.

12 Myers, G. H., Thumin, A., Feldman, S., deSantis, G.

,

Lupo, F. T. Ultrasonics. 20 (1972) 87-89.

10 Olson. Acoustical Engineering. Princeton, N.J..:

D. Van Nostrand Co., Inc., (1964).

4 Okyere, J. G., Cousin, A. J. Ultrasonics . 17

(1979) 81-84.

37

REF. #

3, 5 Penttinen, A., Luukkala, M. J. Phys . D. : Appl.

Phys., 9 (1976) 2547-57.

8, Luukkala, M. Ultrasonics. 15 (1977)

205-210.

9 Quate, C. F., Lemons, R. A. Science. 188 (1975)905-911.

16. Scientific American. 241 (1979) 62-70,

14 Seti, W. W. Acoustics. New York, N.Y. : McGraw^

Hill Book Pub. Co., Inc. (1971).

17 Smith, W. M. R., Awojobi, A. 0. Ultrasonics.

Jan. (1979) 20-26.

22 Turabian, K. A Manual for Writers. Chicago, 111.

University of Chicago Press, (1973).

12 Wilson. An Introduction to Scientific Research.

New York, N.Y. : McGraw-Hill Book Co,, Inc.,(1952).

38a

APPENDIX

38

APPENDIX

TABLE 2

The Lead Zirconate Titanate Transducer

K-350 (PVT5A)

Physical Properties

Density (g.cc) 7.7

Curie Temp (C) 360

Mechanical Q 75

Dielectric Properties

Dielectric Constant 1700

Piezoelectric Properties

K31 35%

K33 70%

K15 65%

D31170 (x 20~12m/v)

D33350 (x 10~12m/v)

D1C- 575 (x 10~12m/v)

G01 11 (x 10~3v.m/NEW)'31

'33

}15

G0 25 (x 10 3v.m/NEW)

G 37 (x 10_3v.m/NEW)

APPENDIX

39

a

o

o

CO

ou

o

o

c

oCO

u

-p

rH

CD

1

CD

hi)H

APPENDIX

2

5

40

of

<

oo

11

J _ _L

T

t1

\i>

*

I 1 4

t ^<in

CO

V-

z

o

U-

. ii

1

T

CO

CD

O

H

C

ow

d

u

p

0

En

O

i-l

CD

U

UJO

cn

APPENDIX

41

VOLTA6E

SPIKE

REFERENCE

PULSE

INTERNA-L

REFLECTIONS

SIGNAL

PULSE

\8 \Z

//SECONDS

16 20

Figure 11. Transmitted and Reflected Pulses,

I.EN3

Figure 12. Water Supply Layout

APPENDIX

42

CD

r-f

,Q

dEh

CO

CD

&DH

Pm

43

APPENDIX

Figure 14. Test Sample One.

932/"^/ctcL6

>78//<n/cYCL6

6IO,um / CVCLE

338 /UmlcIcva-e.'

22 20/jm/iCYCLE

ALUMINUM

Figure 15. Test Sample Two.

APPENDIX

44

2

s

U.

4.

O

<0

VsAAA^-

lA/WVV-

S

ri

-/www

-5D*T

p-^WAA^-i

O

ii

5%

r-fc

<5

_MV\r-

rf^

8A

5

hJWWV

4.3

HI'

o

uH

u

CD

P

P

Cfi

C

oJ

Eh

CXI

CD

MH

Ph

C4

o

o

I vvwv- _i/W^r-

19 ol

r-it.

r

APPENDIX

45

>

0D

e<l

o

(I vWV !?

3- <?

-VVV^ *

KH'

\AAA

u.

o .

o-

f-!S

it

4||^.

|-Wv^i/^^rlsN^A-< '

8-

-s/Wv 1>

+=3-

3O

H

o

CD

>iH

CD

O

CD

T-l

CD

U

bfiH

Ph

APPENDIX

46

M 5fcO t~ \y4

W^ JX.

rci

7

4

3

OOI/J

'^

A 5^B

LCIC3

7

4

0

0

6

C

IC4

7

40

0 y

3

V

-d

1C5

7

I

i

o "

-<.001//

n

n.

n

IC6

7

4i

2

n_

I5K

X

-42

ICT7

4i

l

Vt

I VVV-

T0U-

521//

T805

Y.

f&v

Y2tY

Figure 18. Detection Circuit.

47

APPENDIX

Figure 19. Calibration Circuit

48

APPENDIX

r\pA

110-

M 1 1

0

1

1 1

\

1

T

\5

1

1 1

1

1

1

1

1

1

1

3o-

1

I

1

| 1 11

' "

I L 1 l '450 WICROrrtETERS perdwision

1 T

Figure 20. 2220 Micrometers per Cycle,

Resolution Target at Threshold 1.

49

APPENDIX

Figure 21. 932 Micrometers per Cycle Resolution

Target at Threshold 2 (Ultrasonic)

Figure 22. 932 Micrometers per Cycle

Resolution Target (Optical)

50

APPENDIX

110+

i

C

0

!>&--

T =*450 mnicn0m6Te<l& Pe OWi-SION

Figure 23. 678 Micrometers per Cycle Resolution

Target at Threshold 2 (Ultrasonic)

Figure 24. 678 Micrometers per Cycle

Resolution Target (Optical)

51

APPENDIX

Figure 25. 610 Micrometers per Cycle Resolution

Target at Threshold 2 (Ultrasonic)

Figure 26. 610 Micrometers per Cycle

Resolution Target (Optical)

52

APPENDIX

45o miCRomeTeo-s P6R Division

Figure 27. 398 Micrometers per Cycle Resolution

Target at Threshold 2 (Ultrasonic)

Figure 28, 398 Micrometers per Cycle

Resolution Target (Optical)

.2S

pen

DiV.

53

APPENDIX

REFERENCE /SIGNAL

flEar?

HI

in

Bffl?*J*MHM

2>US /DlV.

Figure 29. Scope Trace of Original Signal from

the Transducer.

REFERENCb /5IGNAL

HHImHilsHHHH

T^BSSSESTl

0

a v

PER

DW,

\am1II inun

2jjS ,/ p\\/

Figure 30. Scope Trace of Filtered,

Amplified Signal.

54

APPENDIX

NoisE REFERENCE S16NAL

o-

ZV

PER

pi v.

HHfeMBMMM

2^5 / DlV.

Figure 31. Scope Trace of Digital Signal

REFERENCE SiSA'AL

2^5/DW.

Figure 32. Scope Trace of Reference

and Signal Pulses.

55

APPENDIX

o-

PERHmmmm

m

DiV.

50On5/D\V.

Figure 33. Scope Trace of Transmitter

Exciting Pulse.