L-

“g ‘lllllllllllllwill

This is to certify that the

thesis entitled

AN EVALUATION OF THE PARAMETERS
INVOLVED IN DEVELOPING A RADIOTRACER
BASED WEAR DIAGNOSTICS SYSTEM

presented by

Dick A. Barkman

has been accepted towards fulﬁllment
of the requirements for

MS degree in Mechanical Engineering

 

W4W

Major professor

7/¢/7z

Date

 

0.7639 MS U is an Afﬁrmative Action/Equal Opportunity Institution

 

 

 

 

 

 

 

LIBRARY
i Michigan State
University

 

 

 

PLACE IN RETURN BOX to remove thie checkout from your record.
TO AVOID FINES return on or before due due.

ll DATE DUE DATE DUE DATE DUE

}| l l

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

—7| l

MSU Ie An Affirmative ActioNEquel Opportunity Institution
eMMMS-M

 

AN EVALUATION OF THE PARAMETERS INVOLVED IN DEVELOPING A
RADIOTRACER BASED WEAR DIAGNOSTICS SYSTEM

By

Dick A Barkman

A THESIS

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

Master of Science

Department of Mechanical Engineering

1992

I
1’ J \
I IA 9'

.) .

u‘
/ \- .-
I

e.

.2 ..
/' 0'

ABSTRACT

AN EVALUATION OF THE PARAMETERS INVOLVED IN DEVELOPING A
RADIOTRACER BASED WEAR DIAGNOSTICS SYSTEM

By

' Dick A Barkman

Wear is an important parameter in the design and operation of
mechanical systems. Wear can cause catastrophic failure of mechanical
components, costing manufacturers and users alike due to lost production
or service time. Non-intrusive radiotracer systems provide a means of
getting wear data in a short time while the target equipment is still in
operation. The purpose of this study was to investigate the parameters
involved in developing and operating a system that monitors wear on line.
To this end, a system was designed and built to measure the piston ring
wear of an internal combustion engine. A one cylinder engine was
modiﬁed so that its oil supply could be pumped to an external reservoir
and back in a closed circuit. The external reservoir enables one to
measure radioactive material in the oil. Then, a radioactive piston ring
was installed in the engine. As the engine ran, radioactive particles were
worn off of the piston ring and mixed into the oil. Finally, a gamma ray
detection system was set up to measure the amounts of radiation in the
oil and piston ring. Wear rates were calculated from these
measurements.

Approved by:

Major Professor

ACKNOWLEDGEMENTS

To my parents, for the support, ﬁnancial and otherwise, which made it possible for me
to reach my goals.

To Dr. Schock, for ﬁnding ﬁnancial support for me where there was none, and for mak-
ing this project possible.

To Dr. Somerton and Dr. Radcliffe, for being on my commitee on such short notice.

To Dr. Ronningen, for the careful instruction on radiation, gamma ray detectors, and
numerous other subjects which made this project possible.

To Tom Stuecken, for providing the practical side of my education, and for the numer-
ous favors without which I may never have ﬁnished.

Special thanks to Bruce Henderson, Chris Heintzman, and the rest of the M.S.U. Ka-
rate Club without whom this would have been done a lot sooner.

iii

TABLE OF CONTENTS

 

Page
LIST OF TABLES .......................................................................................................... vi
LIST OF FIGURES ........................................................................................................ vii
CHAPTER 1 - INTRODUCTION .................................................................................. 1
1.1 Problem Statement. ................ - .......................... 1
1.2 Literature Survey ....................................................................................... 1
CHAPTER 2 - CONCEPTS IN RADIATION DETECTION ...................................... 4
2.1 Outline of Wear Measurement method ................................................... 4
2.2 Radiation terms and definitions ................................................................. 5
CHAPTER 3 - SYSTEM IMPLEMENTATION ......................................................... 9
3.1 Overall design ............................................................................................ 9
3.2 Individual System components ................................................................. 9
3.2.1 Test Engine ...................................................................... ' ................. 9
3.2.1.1 Dynamometer ........................................................................ 10
3.2.2 Oil System ........................................................................................ 11
3.2.3 Engine Stand. .................................................................................... 12
3.2.4 Gamma Ray Detector ...................................................................... 12
3.2.5 Oil Reservoir ..................................................................................... 15
3.2.6 Oil Reservoir Platform ...................................................................... 15
3.2.7 Exhaust Stack ................................................................................... 16
3.2.8 Exhaust Monitoring System. ........................................................... 16
3.2.9 System Environment. ....................................................................... 18
3.3 Data Acquisition ........................................................................................ 18
3.3.1 The Ampliﬁer ................................................................................... 18
3.3.2 The Analog to Digital Converter (ADC) ........................................ 18
3.3.3 The Digital Stabilizer (DS) ............................................................. 18
3.3.4 The System 100 (8100) Software ................................................... 19
3.3.5 Microsampo ...................................................................................... 19
3.3.5.1 Shape Calibrations ................................................................ 20
3.3.5.2 Peak Fitting ....... ‘. ................................................................... 20
3.3.5.3 Microsampo Problems .......................................................... 21
3.4 Calibrations ......................................................................................... 22
3.4.1 Oil Reservoir Calibration ................................................................. 22
3.4.2 Piston Ring Check in ........................................................................ 24
3.4.3 Cylinder Calibration .......................................................................... 24

iv

GIAP'IER4-THEEXPERIMENT ............................................................................ 26

4.1 Wear Calculations ..................................................................................... 26
4.2 Procedure ................................................................................................... 28
4.3 Results ....................................................................................................... 29
CHAPTER 5 - SUMMARY AND CONCLUSION ...................................................... 30
CHAPTER 6 - RECOMENDATIONS .......................................................................... 33
LIST OF REFERENCES ............................................................................................... 36
APPENDIX A - ERROR ANALYSIS ......................................................................... 39

APPENDIX B - PRODUCTION OF 78:: IONS FROM A 7Li BEAM AND A
POLYETHYLENE TARGET ............................................................ 41

APPENDIX C - EXAMPLE OF EFFICIENCY AND ACTIVITY
CALCULATIONS ............................................................................... 43

V “I

 

 

 

 

LIST OF TABLES

Page
Table C.1 Branching Ratios of Europium 152 ............................................................... 46
Table C.2 Known Intensities ......................................................................................... 46
Table C.3 Measured Intensities .................................................................................... 47
Table C.4 Efﬁciencies ..................................................................................................... 47
Table C.5 Peak Areas .................................................................................................... 49
Table C.6 Efﬁciency at C0 56 Energies ......................................................................... 49
Table C.7 Activities of Co 56 ......................................................................................... 50

LIST OF FIGURES

Page
Figme 2.1A The Marker Method ................................................................................... 53
Figure 2.113 The Tracer Method. .................................................................................... 53
Figme 2.2 Depth Calibration Curve ............................................................................... 54
Figure 2.3 The Atom ....................................................................................................... 54
Figure 2.4 Be7 Decay Path ............................................................................................. 55
Figure 2.5 Cobalt 56 Spectrum ....................................................................................... 55
Figtue 3.1 Systerm Components .................................................................................... 56
Figure 3.2 Engine Block .................................................................................................. 57
Figure 3.3 Dynamometer ................................................................................................ 57
Figure 3.4 Coupling ......................................................................................................... 58
Figure 3.5 Engine Stand ................................................................................................. 58
Figure 3.6 Gamma Ray Detector ................................................................................... 59
Figure 3.7 Auto Fill Controller ....................................................................................... 59
Figure 3.8A Detector Filling .......................................................................................... 60
Figure 3.88 Detector Full ............................................................................................... 60
Figure 3.9 Oil Reservoir .................................................................................................. 61
Figure 3.10 Reservoir-Detector Stand. ......................................................................... 61
Figure 3.11A Reservoir Platform ................................................................................... 62
Figure 3.118 Reservoir Platform .................................................................................... 62
Figure 3.12 Class W001 Filter Housing ........................................................................ 63
Figure 3.13A HEPA Filter Housing .............................................................................. 63
Figure 3.13B HEPA Filter Housing ............................................................................... 64
Figure 3.14 Test Set-Up ................................................................................................. 64
Figure 3.15 A Good Fit ................................................................................................... 65
Figure 3.16 A Too Narrow Fitting Interval ................................................................... 65
Figure 3.17A Parabolic Continuum ................................................................................ 66
Figure 3.17B Linear Continuum. .................................................................................... 66
Figure 3.18 Efﬁciency Curve at Position A .................................................................... 67
Figure 3.19 Efﬁciency Curve in Reservoir ..................................................................... 68
Figure 3.20 Efficiency Curve at Position R ..................................................................... 69
Figure 4.1 Depth Calibration Curve ............................................................................... 70
Figure 4.2 Power vs. Run Time ...................................................................................... 71
Figure 4.3 Wear Rate vs. Run Time .............................................................................. 71
Figure B. 1A 7Be Production ........................................................................................... 72
Figure B. 18 7Be Production Control .............................................................................. 72
Figure B.2 Efﬁciency Curve at Position 6 ...................................................................... 73
Figure 8.3 Efﬁciency Curve at Position 1 ...................................................................... 74
Figure B.4 Ion Distribution Curve .................................................................................. 75
Figure 3.5 Depth Calibration Curve .............................................................................. 76
Figure C.l Effect of Distance on Detector Efﬁciency ................................................... 77
Figure C.2 Typical Efﬁciency Curve ............................................................................... 78

Page

 

FigureC.3CoinBank -- .............................................. 79
Figme C. 4 Gaussian Pattern of Signal Collection ......................................................... 79
Figure C. 5 Europium 152 Spectrum ............................................................................... 80
Figme C.6 Example Efﬁciency Curve ............................................................................ 81
Figme C.7 Cobalt 56 Spectrum ...................................................................................... 82

viii

CHAPTER 1
INTRODUCTION

1.1 Problem Statement

Wear is an important parameter in the design and operation of
mechanical systems. Wear can cause catastrophic failure of mechanical
components, costing manufacturers and civilians alike in lost production
or service time. It is often cheaper to invest in wear prevention methods
than it is to repair damaged equipment. Wear prevention requires
knowing how much wear is critical, what conditions cause the most
wear, and where it happens. One way to acquire this information is to
periodically disassemble the equipment and measure the parts of
concern. This method has serious drawbacks. It causes expensive
downtime of equipment, reassembly of equipment changes the
subsequent wear pattern, and the method is limited to relatively large
amounts of wear. For industries that cannot afford the downtime of
equipment and/or need more precise measurements of wear, a better
alternative is the non-intrusive radiotracer method. This method provides
a means of getting wear data in a short time, while the target equipment
is still in operation. This method, however, is not currently commercially
available in the United States. The purpose of this study was to determine
what is involved in designing, building, and operating a system that
monitors wear on-line. In particular, this system was designed to
measure piston ring wear in an internal combustion engine by rendering a
section of the surface of a piston ring radioactive, installing it in a test
engine, and monitoring the radiation over a period of test runs. The
amount of radiation left on the ring is directly related to the amount of
wear that has taken place.

1.2 Literature Survey

Much research has been done on wear prediction and
measurement. The radionuclide technique is getting an increasng amount

of attention from industry and university researchers alike. The following
papers illustrate the wide range of application of the radionuclide
technique.

Evans [1] compared the surface layer activation (SLA) method
with a nuetron activation method for wear studies. It was found that SLA
had several advantages over the nuetron method: A much lower activity
of the sample was possible with SLA, irradiation of only the area of
interest was possible, and SLA provides a choice of radionuclides to use.

Kinsella et al [2] used the radiotracer method to study the material
removal process in the sliting of cast iron. The method was found to be
suitable for tracing wear debris in adhesive and abrasive processes.
Slitting was studied at low and high loads. At low loads all the slit cast
iron was transferred to the rim by adhesion, oxidized, and escaped from
the system as debris. At high loads the iron transferred to the rim
became hard and abrasive. In this case half the slit volume of material
became debris directly by abrasion rather than the adhesive transfer to,
and decay from , the rim, as evident in the low load case.

Fritz et al [3] performed in situ piston ring wear measurements in a
medium speed diesel engine. The research engine was a single cylinder
model run at conditions simulating the operation of line haul locomotives.
Data was taken at various speeds and loads by both the radiotracer and
radiomarker methods. The high correlation in the data from both methods
was taken to indicate that monitoring wear in situ is sufﬁcient and that
stopping the engine for direct measurements is unnecessary.

Schneider and Blossfeld [4] did a comprehensive study of the
effect of speed and power output on piston ring wear in a Detriot Deisel
"series 60" engine. The radiotracer and marker method were used with

radioactive 54Mn being the nuclide used. Rings were activated
circumferrentially and in discrete bands while wear rates were measured
at 21 different test conditions. It was determined that at constant
powerwear rates increase with decreasing speed, and at all speeds the

wear rates increased with increased power output and BMEP.

CHAPTERZ

CONCEPTS IN RADIATION DETECTION

2.1 Outline of Wear Measurement Methods

For wear analysis, two methods are currently used to render a
piston ring radioactive. The ﬁrst method, called surface layer activation
(SLA), is only useful on certian materials with atomic number greater
than 24. In SLA, a piston ring is bombarded with energetic particles,
causing nuclear reactions. These reactions change the iron to cobalt 56, a
radioactive element. A second method, which involves direct ion
implantation, has the advantage that it works on any material. In this
method, radioactive particles are implanted into the ring.

Once the radioactive ring is installed in the test engine, there are
two ways to detect piston ring wear: the marker method and the tracer
method. In the marker method (Figure 2.1 A), the level of the ring
radiation is measured directly through the engine block between intervals
of engine operation. The difference in the levels of radiation measured
before and after a run are directly related to the amount of wear on the
piston ring. In the tracer method, the radiation in the oil is related to the
amount of material worn off the ring. The tracer method (Figure 2.1 B)
has the advantages that it can be used while the engine is running and
the detector can be located some distance away from the engine. If the
detector is close to the engine, the vibration will cause system noise
which decreases the detector resolution. Tire marker method allows for
faster measurements, but this method requires that the engine be stopped
and the piston be at a ﬁxed position relative to the detector. For
comparison, both types of measurement were used in this study. The
gamma radiation emitted by the ring can penetrate most materials with
very little attenuation compared to charged particle radiation (such as

5

alpha and beta particles) and is easily identiﬁed. from background
radiation.

Once the radiation has been measured, the wear can be
determined by using the depth calibration curve shown in Figure 2.2. The
relative intensity is the ratio of the radiation measured after a period of
engine operation to the radiation measured before. This depth calibration
is produced by removing thin layers of surface material from an activated
control piston ring and measuring the difference in radiation produced.
Surface material is removed by masking around the activated region and
sand blasting. The activity remaining is then measured, while the
roundness proﬁlometry is used to determine the surface thickness
removed. Roundness proﬁlometry measures the height of the step
between masked and unmasked regions.

2.2 Radiation Terms and Definitions

Familiarity with the atom is necessary to understand how gamma
radiation is produced. The atom has a nucleus consisting of protons that
are positively charged and neutrons that have no charge. Electrons,
which are negatively charged, orbit about this nucleus. Each individual
combination of protons and neutrons is called an element, or nuclide.
Figure 2.3 represents a basic atom.

Nuclides that differ only in the number of neutrons are called
isotopes of one another. The number of protons in the atom gives each
nuclide its chemical identitity. Most nuclides are unstable (radioactive).
Radioactivity is deﬁned as the spontaneous disintegration of unstable
nuclei with the resulting emission of nuclear energy that results in the
formation of new elements. There are many modes of disintegration, or
decay, but only one shall be dicussed here: gamma ray emission.

Gamma ray emission is the production of electromagnetic
radiation by a nucleus that is reconFigureuring to reach its lowest energy
state. Radioactive nuclides give off gamma rays only at a specific pattern
of energies and intensities characteristic of that nuclide. Cobalt 56, for

6

instance, is the only nuclide to give off gamma rays at 846 keV, 1038
keV, 1237 keV, and 1771 keV, a pattern which no other nuclide shares.
The intensity of the gamma rays given off at these energies is also unique
to cobalt 56.

The activity of a radioactive sample is the instantaneous number
of atoms decaying per unit time. Activity is measured in Becquerels (Bq),
where one Bq is one disintegration per second. The curie is a unit more
commonly used. One curie is equivalent to 3.7x10"10 disintegrations per
second (d.p.s.). Only a fraction of the decays occurring give off gamma
radiation because a nucleus may decay in many different ways. The
branching ratio is the probabiltity, expressed in percents, that a nucleus,
or the excited state of a nucleus, will decay in a particular manner. This
branching ratio is given in gamma rays per 100 decays. Figure 2.4 shows
the decay paths for 7Be. There is an 89.7 percent chance that it will
decay directly to the ground state of lithium 7, and a 10.3% chance tha it
will decay to the excited state of 7Li. The excited state of 7Li will decay
to the ground state by releasing a gamma ray. The branching ratio is then
the number of gamma rays given off per 100 decays, which is 10.3.

Radioactive nuclides disintegrate spontaneously at a rate that is
dependent on the number of original atoms present and upon its decay

constant lambda, 7L . This constant A is deﬁned as the instantaneous
fraction of atoms decaying per unit time. Each nuclide has its own
characteristic decay constant. The relation between the number of atoms
left at a given time and the original number of atoms is given by the
decay equation:

N: = No*e-At (2.1)

N is the number of atoms present at some time t and N0 is the

t

original number of atoms present. The decay constant, A, is related to the
nuclides halﬂife, T1 /2 , by the equation:

I. = ln2 / T1/2. (2,2)

The halﬂife is the time it takes for the number of atoms to decay to
one half of the original amount. The instantaneous number of atoms, N,
remaining at a given time is given by:

N=A/>., (2.3)

where A is the activity. The decay- equation is important for wear
stuﬁes because the diffemce in the level of radiation being measured is
partly due to wear and partly due to natural decay. The amount of
material that has decayed must be accounted for in order to get an
accurate assesment of the wear.

Activity is measured by using a gamma ray detector. A detector
counts the number of photons given off by a radioactive source at
speciﬁc energy levels. The activity is calculated by dividing the total
number of photons given off at a speciﬁc energy by the time spent
counting. Only a fraction of the photons given off at any discrete energy
levels are gamma photons, as discussed earlier. Since only gamma
photons are being counted, the total number of photons at a speciﬁc
energy must be calculated by dividing the number of gamma photons
counted at that energy by the branching ratio for that energy. A typical
gamma ray spectrum for a count of a cobalt source is shown in Figure
2.5.

Four cobalt peaks are shown in the Figure. The activity of the
source can be calculated from any one of the peaks by the following
equation:

peak area
Activity = (2,4)
(livetime)(branching ratio)(efﬁciency)

 

The peak area is the total number of photons detected above the

background energy level at one energy. The live time is the time spent
counting. The efﬁciency is the efﬁciency of the detector and is given by:

, number of photons emitted by the source
Total Efficrency = (2.5)

number of signals generated by the detector

 

Ideally the calculated activity would be the same no matter which
peak area was used. This is not the case, however, due to uncertianties
in the counting. To approximate the true activity the mean of the
activities calculated at each peak energy is used:

2(activityi / activity erroriz)
Mean activity = . and (2.6)

2(1 / activity erroriz)

 

l
[ 2X 1 / activity errori2 )11/2

 

Mean activity error = (2.7)

CHAPTER3

SYSTEM IMPLEMENTATION

3.1 Overall design

Figure 3.1 shows a schematic of the system design. A
dynamometer, engine stand, engine, and pump are mounted on a rolling
table. The pump and engine are connected to the oil reservoir by ﬂexible
hose (not shown). The dynamometer is linked to the engine by a rubber
coupling on the crankshaft. The dynamometer has a separate controller
mounted on its own mobile stand. The detector is connected to the multi-
channel analyzer (MCA) and personal computer by long cables. Every
24 hours, the detector dewar is ﬁlled with liquid nitrogen from a 50 liter
tank. The details of each system component are given below.

3.2 Individual System Components
3.2.1 Piston Ring

In the present study, SLA was used on the top compression ring
by Kernforschungszentrum, a cyclotron in Germany. For this type of
study the ring can rotate in its groove, therfcre it is necessary to either
activate the entire surface of the ring or to activate a portion of the
surface and pin the ring so that it cannot rotate. The ring must not rotate
when only a section is activated because the activate portion must
remain in a ﬁxed orientation relative to the gamma ray detector
positioned outside the engine. To avoid the expense of a full activation the
ring was pinned in the present study. The ring was activated in a 10 x
1.5 milirneter area 90 degrees from the ring gap as measured looking
down on the beveled surface. The ring had an activity of 7.5 microcuries.

3.2.1 Test Engine
The test engine is shown in Figure 3.2. It is a 6.2 kw (11 hp.)

.1

10

Briggs & Stratton with an aluminum block and cast iron bore. It is a
single cylinder model meant for running pumps and other industrial
equipment.

To calculate the total amount of radioactivity in the oil from a
measurement of a fraction of the oil, the oil must be homogeneously
radioactive. The location of the engine block oil ports, however, tends to
minimize mixing of the oil. These ports face each other across the
crankcase creating the possibility that return oil could bypass mixing with
the more radioactive oil and go right back to the reservoir. To eliminate
this problem a steel bafﬂe was placed in the crankcase in front of the oil
return port to circulate the oil.

Later in the study the engine compression ratio was raised in the
hopes of increasing the wear rate. Since the engine instruction manual
did not give the stock compression ratio, it was measured using the water
displacement principal to be 7.6. The head was then machined to a new
compression ratio of 9.5.

3.2.1.1 The Dynamometer

For this study, the engine speed and load had to be carefully
controlled. A 50 hp. eddy current dynamometer was used for this
purpose. It has three basic functions: to drive an engine, for either non-
ﬁring tests or starting; to carefully load the engine; and to monitor the
engine speed and torque output.

The test engine was a pull-start model. Instead of using a pull-
start for the tests, however, the dynamometer was used to crank the
engine over until it ﬁred. Once the engine ﬁred, the dynamometer stopped
driving it and began absorbing its output, monitoring the load in the
process. If a load needs to be placed on the engine, there are two ways
the dynamometer can load it: It can put a load that varies as necessary
to hold the engine at a constant speed, or it can put a constant load on
the engine while the engine speed is free to vary. The dynamometer is
shown in Figure 3.3.

11

The dynamometer shaft is connected to the engine crankshaft
through a rubber coupling, shown in Figure 3.4. This coupling protects the
dynamometer from being damaged should the engine seize up, or the
engine from being damaged by a dynamometer failure. Couplings can be
broken by putting a high enough load on the engine to stall it. This load
will be high enough to destroy, the coupling. The dynamometer operating
procedure is in the appendix.

3.2.2 Oil system

To minimize errors in radioactive tracer studies, it is important to
minimize the amount of air in the lines and to keep a constant ﬂow of oil.
Air bubbles in the line will eventually displace oil in the reservoir during a
count of its activity resulting in an erroneous calculation of total activity
and thus piston ring wear [7]. An intermittent ﬂow of oil can allow
buildup of radioactive particles at different areas in the system resulting ‘
in inhomogeneously radioactive oil. As mentioned earlier (Section 3.2.1)

homogeneous oil radioactivity is necessary for accurate calculation of
' activity.
To minimize air bubbbles the system was designed to operate at
70 psi (the optimum pressure for reducing air bubbles [7]), and a gear
type pump was used to minimize cavitation. The pump was a 1/3 hp.
adjustable relief valve model with 1/4" ports and a free ﬂow rate of 2.2
gpm of 72 degree water. It was found that cavitation could be reduced by
using 1/2" i.d. oil supply lines to the pump, as the pump seemed to draw
oil at a greater ﬂow rate than the 1/4" lines would allow. 1/4" lines were
originally used because they matched the pump port size.

In order to maintain an oil pressure of 70 psi in the reservoir and
lines a relief valve was placed downstream of the reservoir before the
return to the crankcase. Another relief valve, which is built into the pump,
allows the pump to circulate oil around its own case if the line pressure
after the pump gets too high (for example if the lines were pinched shut).

12

Brass ﬁttings are used to connect the system’s elements.
Fittings must be chosen that minimize leaks and allow the oil to ﬂow as
unrestricted as possible. Fittings can leak under heavy use, so minimizing
the number of joints is of prime importance. For example, a street elbow
can be used to replace an elbow and bushing to eliminate one joint. To
stop leakage the joints must be silver soldered. Regular solder is less
effective. For sensitive ﬁttings that cannot be soldered, such as gauges
and valves, teﬂon tape should be wrapped around the threads. Fittings
that attach directly to the pump, which has small ports, should be as large
as possible to avoid constriction of the ﬂow.

3.2.3 Engine Stand

The engine stand, shown in Figure 3.5, was built to rigidly
support the engine so that the engine crankshaft was in line with the
dynamometer shaft. It also served as a mount for the pump and relief
valve. The stand has low walls around its base to capture any oil that
escapes the engine.

3.2.4 Gamma Ray Detector

The gamma ray detector, shown in Figure 3.6, is a high purity
germanium coaxial photon detector, manufactured by EG&G Ortec,
model number Gem 90220-p-Plus. The detector contains a single crystal
of germanium, 75.8 mm in diameter and 92.9 mm in length located in the
center of the cylinder 3 mm from the end cap of the detector. This crystal
produces electron hole pairs by absorbing gamma ray photons. These
electron hole pairs are swept across a 2500 volt electric ﬁeld inducing a
current, that is sent to the preampliﬁer. Once the preampliﬁer, which is
also under the detector sheath, receives a signal from the crystal; it
boosts it and sends it on to the ampliﬁer.

This detector requires low temperatures in order to operate
with high resolution. In fact, the crystal and the preampliﬁer FET (ﬁeld

“1'

13

effect transistor) have to be kept close to liquid nitrogen temperatures in
order to reduce thermally induced noise. Thermally induced noise is
current generated by the motion of electrons that are excited to the
conduction energy level by the ambient temperature. This current is
indistinguishable from that produced by incident gamma rays. A high
voltage applied to the crystal and the preampliﬁer FET at higher than
liquid nitrogen temperatures could damage them. If the temperature gets
too high, a connection is made from the "bias shutdown" output of the
preampliﬁer to the power supply "inhibit", which will shut down the
detector. If such a shut down should occur, the detector power supply
reading will drop to zero volts. Despite this, the dial will still be set at
2500 volts, so the meter should be checked for power status, not the dial.
When such a shutdown occurs the dial should be turned to zero and the
"power on" switch turned off. The detector dewar should then be ﬁlled
with liquid nitrogen in order to cool down for at least 12 hours before the
next use.

The cryostat is the device that keeps the crystal at low
temperatures. It provides a path for heat transfer from the high vacuum
around the crystal to the liquid nitrogen in the dewar. The dewar, which
is connected to the detector, contains three liters of liquid nitrogen.
Because this nitrogen evaporates very quickly the dewar must be reﬁlled
every 24 hours in order to keep the crystal cool.

The system used to ﬁll the dewar consists of an automatic ﬁll
controller and a 50 liter tank of LN 2. The auto-ﬁll controller automatically

opens, at pre-set times, a solenoid controlled valve between the 50 liter
tank and the dewar. The tank was ﬁlled at the Michigan State University
National Superconducting Cyclotron facilities once a week during
everyday operation. When the tank is ﬁlled the relief valve will be
constantly hissing as it relieves pressure buildup. The tank has a self
pressurizing feature that creates large pressures from the nitrogen vapor
instantly. Opening the black knob on the top of the tank will activate this

Iv

14

system and may form ice on the tank. The self pressurizing function is
normally not needed as the relief valve can be used to get a higher
pressure, although it takes longer for the pressure to build up.

The auto ﬁll controller, shown in Figure 3.7, is located in the
power supply bin with the rest of the electronic equipment. The display
on top of the controller shows the amount of time left until the next ﬁll.
The green light bulb labeled "ﬁlling" below the l.e.d. display lights up
when the detector is being ﬁlled. The button labelled "manual" opens the
solenoid valve no matter where in the countdown process the . controller
is. The norm/ inhibit switch allows normal operation when set to norm
and when set to inhibit prevents the valve from being opened either
manually or automatically. The maximum time the controller holds the
valve open can be set by dip switches on the side (which can only be
seen when the controller is removed from the bin). If this time is
exceeded the detector will activate an alarm which sounds until the
alarm reset button is pushed. The preset time should not be less than the
time it takes for the dewar to ﬁll. When ﬁlling is in progress the vent line
sends out a stream of vapors. When the dewar becomes full, liquid comes
out the vent line (Figure 3.8 A & B). The controller uses a temperature
sensor in the dewar vent line to tell when the dewar is full. When the
liquid strikes the sensor, the sensor resistance is reduced. When this
happens the controller closes the valve. The sensitivity of the controller
can be adjusted, but it must be done with care. If it becomes too sensitive
the valve will be shut off by just the presence of vapor in the line. If it is
adjusted too much the other way the liquid itself will not cause the valve
to shut and the liquid will keep venting until the 50 liter tank is empty.

Power can only be applied to the detector when it is cool. This
should be at least 12 hours after ﬁlling the dewar from a warm state. To
do this, it is necessary to turn on the detector power supply, unlock the
main power dial, and slowly turn the dial, thus increasing the voltage until
a positive 2500 volts is reached. This should be done at a rate of 100
volts per second. Only a low current power supply (micro amp level)

H“

15

should be used. The present study used a Canberra model 3105, 0-5kv
power supply that could deliver up to 100 mA of current.

3.2.5 Oil Reservoir

The purpose of the oil reservoir is to contain a fraction of the
system oil in such a way as to enable efﬁcient counting of the radiation in
the oil. Because the counting time is high when the amount of radiation
being measured is low, the reservoir was built to contain a large ﬁaction
of the oil to reduce counting times. It was also designed to surround the
detector with oil so the geometric efﬁciency would be high.

Figure 3.9 shows the reservoir, a double coil of copper tubing into
which the detector slides. Because of its close proximity to the detector
the reservoir had to be designed so that it would not heat the detector.
For this purpose two layers of galvanized steel sandwiching a layer of
ﬁberglass insulation were placed on the inside. For additional protection a
small blower was used to circulate air between the detector and and the
insides of the insulation.

To accurately compare one radiation measurement from the
reservoir to the next, the reservoir and detector must be in the same
relative position for each measurement. This was ensured by the
reservoir-detector stand as shown in Figure 3.10 . The coil slips inside the
restraining bars and is bolted in while the detector base slides into the
carefully positioned grooves. A metal stop keeps the detector in this
reference position for measurement.

3.2.6 Oil Reservoir Platform

A location was needed to keep the coil and detector so that
they would not be accidentally damaged. The location also had to have
enough room for 260 lbs of lead brick used to shield the detector from the
direct path of photons from the engine block. To accomodate the hose
used in this system and the lack of space in the engine lab, a rolling table
was made to hold the coil, detector, and lead bricks. The oil reservoir

16

platform is shown in Figures 3.11A and B.

The reservoir detector stand bolts down to the base while lead
shielding is bolted down next to it, slightly offset to keep the center of
mass near the geometric center. The shielding and coil were placed on
the lower shelf so that the center of gravity would be low and the risk of
the table falling over would be minimized.

3.2.7 Exhaust Stack

Safety regulations dictate that the errrission of exhaust take
place where it cannot return into a building containing people. Therefore
an exhaust stack was built to channel the exhaust to a safe height, 15
feet above the roof of the Engine Research Laboratory. The stack was
fabricated out of an aluminum tube. Brackets were made to hold the tube
against the side of the bulding, and a short length of tube was welded
onto the lower end giving it an L shape for ease of connection to the
exhaust ﬂex tube running from the engine. A steel cap was attached to
the top to keep rain from entering the tube.

3.2.8 Exhaust Monitoring System

Testing a radioactive piston ring in a running engine creates the
possibility that some of the radioactive particles will be swept out the
exhaust. If so, this may pose an environmental hazard. One way to deal
with this hazard is to pump air into the exhaust to dilute the concentration
down to acceptable levels. This is an expensive and unelegant solution
because the total amount of hazardous material put into the air remains
the same. To fully understand the issue, the amount of radiation in the
exhaust was measured during the piston ring break-in period (where the
wear is the highest).

This test consisted of placing ﬁlters in the exhaust line to trap
particulates. To this end, three ﬁlter systems were designed, but only one
was used. The ﬁrst two ﬁlters were of glass wool, and the third filter was

17

a HEPA ﬁlter normally used for gas masks. The glass wool seemed to be
the ideal ﬁlter but it was discovered that there was no convenient way of
determining the efﬁciency in trapping particulates. The HEPA ﬁlter,
however, had a well known efﬁciency of 99%. The disadvantage of this
ﬁlter was that it was too small and too efﬁcient to be placed directly in
the exhaust ﬂow without stalling the engine. The problem was solved by
drawing off a portion of the exhaust gas through the HEPA ﬁlter with a
vacuum pump. '

The glass wool used for the ﬁrst two ﬁlters had to completely
ﬁlter the exhaust. The container shown below was built to hold the glass
wool against the pressure of the exhaust gases while not allowing any to
escape to the outside air. The outer shell was built from an ordinary truck
mufﬂer having 3" diameter inlet and exhaust ports. Steel grating was
welded inside to support the glass wool and a spacer was built to
separate the ﬁlters from each other and to allow removal of the ﬁlters
when the test was ﬁnished. In testing there was some difﬁculty
establishing the right amount of glass wool. Too little resulted in melting
of the wool and too much blocked the ﬂow. Another problem was that
the wool became saturated with water if the outside temperature was
below the dew point. The glass wool ﬁlter housing is shown in Figure
3.12. In the end these ﬁlters were not used.

The HEPA ﬁlter was mounted in the case shown in Figure
3.13A and B. Since exhaust water condensation was a problem a water
trap was installed upstream of the ﬁlter. The trap was connected to the
glass wool container (which was still in-line though not used) by copper
tubing. The vacuum pump drew a ﬂow rate of 0.8 gpm.

The engine was run for 15 minutes, this being enough time for
radiation to collect on the ﬁlter if the level was indeed hazardous.
Subsequent tests of the ﬁlter showed no radiation greater than
background. Tests of the trapped condensed water also showed no
radiation greater than background. Based on these results, it was judged

“l

18

safe to run the engine without diluting the exhaust.
3.2.9 System Environment

The dynamometer was rolled outside on rails where the noise
of operation would not disturb everyone in the building (Figure 3.14). It
was shielded from the elements by a canvas tent. The detector and oil
reservoir remained inside the lab along with the dynamometer controller
and other electronics.

3.3 Data Acquisition

The data aquisition system includes the detector (discussed
earlier), and the Multi Channel Analyzer, which consists of the Ampliﬁer,
the Analog to Digital Converter, and the Digital Stabilizer.

3.3.1 The Ampliﬁer

The ampliﬁer takes poorly deﬁned pulses from the preampliﬁer and
smooths them into near gaussian shapes to ensure that the signals
entering the ADC are consistent.

3.3.2 The Analog to Digital Converter (ADC)

The ADC takes the input voltage pulses from the ampliﬁer and
converts them to clean digital signals that the computer can read. An
input of 0 to +10 volts is divided into a number of channels speciﬁed by
the user. The higher the number of channels, the greater the resolution.
For this study 8192 channels were used. The input pulses are placed into
channels according to energy; higher energies pulses are placed in higher
channels.

3.3.3 The Digital Stabilizer (DS)

The Digital Stabilizer receives signals from the ADC and ensures
that the ADC places pulses of a particular energy in the same channel
every time. While it does this automatically it needs information from the
operator to do the job correctly. The operator needs to give the stabilizer
two reference peak channels that it should stabilize about, one high

.1-

19

energy and one low energy. If this is not done, the pulses that should be
close to a particular peak energy channel will be sent to channels much
further away, broading peaks or even creating false peaks.

3.3.4 The System 100 (8100) software

Signals from the detector are sent through the multi channel
analer (MCA) to the personal computer. The 8100 software enables
data collection, display, and ﬁling. Data is displayed on the screen with
channels on the horizontal axis and number of counts on the vertical,
allowing easy energy calibrations. 8100 allows easy aquisition of peak
areas off the main screen.

3.3.5 Microsampo

Microsampo is a software package designed to calculate
accurately peak areas, activities, and efﬁciencies from data collected by
System 100. This program has many ﬂaws however and is very user
unfriendly. Details of the operation are given in the manuals but an

overview of its operation and discussion of some important points is given
below.

General Operating Procedure
1) Input data ﬁles from the System 100
2) Program the energy calibration used for the 8100 spectra

3) Perform a peak search - this tells Microsampo to go through
the spectrum and identify all the peaks that are signiﬁcantly
above the level of background radiation.

4) Calibrate peak shapes. This is the calculation of shape
parameters used to speed up the next step, peak ﬁtting.

5) Fit peaks. this is the calculation of shape functions, based on
parameters calculated previously, which enable the calculation
of peak areas.

20

3.3.5.1 Shape Calibration

Shape calibration is the creation of shape parameters to speed
up ﬁtting. Shape functions, consisting of gaussian centers and exponential
tails, are matched to a few good peaks (a good peak is one that has a
gaussian shape). The parameters derived from these functions are peak
width (cw), which is the standard deviation in channels of the gaussian;
the lower tail parameter (cl), which is the distance in channels from the
centroid to the point where the lower tail connects to the gaussian; and
the upper tail parameter (ch), which is the distance from the centroid to
the point where the high tail connects to the gaussian. The parameters
derived from these good peaks are then used in the peak ﬁtting process
for all of the peaks.

Microsampo requires the user to go through the spectrum peak
by peak, examine the agreement between function and actual peak
shape, and either accept or reject the parameters derived from this
match. If the match is not acceptable; the user can reject it, change
some ﬁtting options, and try a rematch. The only matching options are the
high and low end of the match intervals. A good match looks like Figure
3.15. A match in which the interval should be widened is shown in Figure
3.16.

3.3.5 .2 Peak Fitting

Peak ﬁtting is the process of putting together the shape
functions to approximate the actual peak shape and calculating the area
of the peak. The computer calculates the peak areas by integrating under
the shape function curves and subtracting the background continuum.
Microsampo reports the peak area, area error, and the goodness of ﬁt.
The user has two options available to control how Microsampo draws
the background continuum and two more to control how it puts the
functions together. The continuum is drawn either linearly or
parabolically, as shown in Figure 3.17A and B. The shape functions are

21

ﬁtted either linearly or nonlinearly. In the linear ﬁt, the centroid of the
peak is matched to the gaussian functions centroid and the functions are
constructed from the match. Nonlinear ﬁtting means that the cenroids are
allowed to vary as the shape functions are put together. The user must
choose the combination that produces the lowest error through a trial and
error method. The combinations are summarized below:

1) linear continuum - linear ﬁt

2) linear continuum - nonlinear ﬁt

3) parabolic continuum - linear ﬁt

4) parabolic continuum - nonlinear ﬁt

Another important option is the interval width. After ﬁnding the
ﬁt-continuum combination for least errors, the errors can sometimes be
reduced even further by manipulating the ﬁtting interval width. Trial and
error is required here but a little experience will show which peaks can
be improved and which cannot.

Once all the peaks have been ﬁt, a table of peak areas is
produced. This peak ﬁt report should be sent to the printer because peak
ﬁt results cannot be saved to disk. Once areas have been calculated,
Sampo allows you to calculate efﬁciencies and activities, but its routines
are very user unfriendly and not accurate. In fact, only the peak areas
were used in any of the work for the present study.

3.3.5.3 Microsampo Problems

Many times during use of Sampo, and totally without warning or
a pattern, the message "Insert write protect, hit any key to continue" will
appear. If it happens consecutively, it means the program will crash after
several more repetitions of the message. No reason for this has been
discovered.

Another problem with Sampo is that peak ﬁt results cannot be

22

saved to disk for later use or modiﬁcation. This means that to rexamine
the data set the peaks will have to be ﬁt again, requiring an additional
hour of time. This lack of proper documentation is one of Sampo’s
greatest ﬂaws. ‘

3.4 Calibrations
The activity of any sample measured by the detector is given by

the formula:

peak area
Activity =

 

(livetime)(branching ratio)(efﬁciency)

The peak areas and livetirnes (livetime being the time spent
counting radiaton) are readily measured, the branching ratios are known,
but the efﬁciency is not. As discussed in appendix C, to get an efﬁciency
curve for a set position relative to the detector a known source must be
placed in this position and measured. The ratio of the measured
intensities to the actual (known) values is the efﬁciency at that position.
For this study, accurate counts of the piston ring and oil had to be made
repeatedly. To- make this possible, efﬁciency curves were generated for
ﬁxed reference positions near the engine block and oil reservoir. An
efficiency curve was also generated for a general reference position
where the ring could be counted before and after the test.

3.4.1 Oil Reservoir Calibration

Before the activity in the oil can be calculated, an efﬁciency curve
must exist for the detector in its ﬁxed position within the reservoir. This
means that a known source of radiation must be placed within the
reservoir and measured. Only a source in a water solution could be put in
the reservoir and completely removed without contaminating the
reservoir. ThCl4 was chosen because it is a naturally occurring

23

radioactive substance, it has well known peaks, it is available in a water
soluble powder form that had a low enough activity to be shipped as a
non radioactive substance.

The calibration procedure is outlined below: .
1) Count the ThCl4 in a ﬁxed position, called position A, near the

detector while it is in powder form and as close to a point
source shape as possible.

2) Measure a known source at position A to create an efﬁciency
curve. Use this curve to calculate the ThCl4’s activity.

3) Mix the ThCl4 in a water solution and put it in the reservoir.
Count the reservoir.
Step 1: 5 mg of ThCl4 in a glass bottle was placed 25 cm away

from the detector on the detector centerline. The distance was measured
from the detector to the powder inside the bottle. This was called position
A. The ThCl4 was counted.

Step 2: The known point source ( a mixed standard containing
the nuclides antimony 125, tellurium 125, europium 154 and 155), was
placed at position A. The source was counted.

Step 3: After some experimentation a vacuum pump was used
to draw air out of the reservoir resulting in water being drawn in to
replace the air. The line from the reservoir to the pump. was long and
coiled so that when water was observed exiting the reservoir there
would be time to shut off the pump before the water reached the. pump.
Experimentation showed that if the water in the beaker dropped to a
level near that of the end of the intake hose air would enter the lines.
Thus a large enough volume of water was used so that when the
reservoir was ﬁlled and the pump was shut off, the level of water in the
beaker was higher than the hose end. The reservoir volume is 1260 ml.

24

A white sheet was placed over the lab table. 17 ml of thcl4 ' was
placed in a 1300 rrrl of water. To this 3 ml of food coloring was added so
that any spills could be easily located on the white sheet. The total
volume of solution was 1320 rrrl. The vacuum pump was used to get the
solution into the reservoir. Once the reservoir was full the small amount
of solution still in the lines was put back in the beaker. The volume
fraction of thc14 solution in the reservoir to the total volume of solution

was .9545 i 4.8%. The reservoir was then counted.

The efﬁciency curve shown in Figure 3.18, which was generated
using the mixed point source, was used to obtain the activity of the thcl4.
The curve shown in Figure 3.19 is the efﬁciency curve for the solution in
the reservoir itself.

3.4.2 Piston Ring Check In

To obtain a measurement of the initial piston ring activity a
special platform was built. This platform enables exact repositioning of
the piston ring every time a measurement is made, such as before and
after the testing. This position is called position R. Three known sources
were in position R. The sources were a combination europium standard
(the same one used for the coil calibration), a steel foil activated with co-
56 that was supplied with the piston ring from KFK, and a sample of En-
152. These sources were then counted to get an efﬁciency curve for this
position. The curve is shown in Figure 3.20. The rings activity was
measured to be 7.5 microcuries.

3.4.3 Cylinder Calibration

For the direct radiation measurements necessary in the Marker
Method an efﬁciency calibration was necessary in the cylinder of the
engine. To this end two point sources were placed at the same position in
the cylinder and measured. Because of the error in the placement of the
sources the resulting effeciency curves were signiﬁcantly different from
each other. This was becuase the sources were approximately one inch
away from the detector when being measured. At this distance even

25

very small innacuracies in the positioning of sources can have a very
large effect on efﬁciency. Because there was no way to reposition the
sources with any greater accuracy the tests were run without an
efficiency calibration for this position. The relative intensities are all that
is necessary for the wear calculations in the marker method, and these
were measured directly.

CHAPTER4

THE EXPERIMENT

4.1 Calculation of Ring Wear

The depth calibration proﬁle shown in Figure 4.1 was provided with
the ring from KFK. The relative intensity is the number of gamma
photons per second given off by the ring after a test period divided by the
amount of gamma photons given off before the test period. One way to
calculate this is to make a direct count of the piston ring activity before a
given test run, getting Abeforc , then count the ring after the test run,

getting Aafter' The relative intensity is then:

Relative intensity = Aafter / Abefore (4,1)

Wear, in micrometers, is then calculated from the line ﬁt to the
depth calibration:

Wear = 54.5*( 1 - Relative Intensity) (4.2)

Another way to calculate the relative intensity is from the amount
of radiation in the oil. The relative intensity after the nth test run is given
by:

Relative Intensity = [Amg - Acne») / [Amg - Adlai-1)] (4.3)

Equation 4.2 is then used to calculate the wear. Note that in this
case Aring is the original activity of the ring decay adjusted to the

26

27

activity it had at the time of the oil count:

Aring = A0 * exp(1112*decay time / halﬂife) (4.4)

Aoil is the activity counted in the oil reservoir multiplied by the

ratio of total oil volume to reservoir volume:

A - (total volume / reservoir volume)(A

ou- <45)

reservoir)

It is assumed that the volume ratio is constant. This may not
always be the case due to changes in oil temperature, oil losses, and
cavitation induced air bubbles. The error in the wear is calculated as
follows:

Wear error = [ 2“‘(Aﬁng error)2 + (A (n) error)2

oil

4» (Aoil(n'1) error)2 ] 1/ 2 (4'6)

Where the Aoil error is:

Aoil error = [ (volume ratio error)2 +

(oil reservoir activity error)2 ] 1’2 (4-7)

Calculation of wear rates is then just the nth measurement of wear
divided by the length of the nth test run:

28

Wear rate = wear(n) / Run time(n) (4.8)

Assuming no error in the measurement of time the error in the
wear rate is equal to the error in the wear.

4.2 The Procedure

Since the wear rates in the test engine were completely
unknown before the study, the procedure had to take into account the
possibility that all of the radiation could be worn off in a very short time,
even a matter of minutes. To this end, the tests were started at low
speed and low load to minimize wear. Whenever counts of the piston
ring were necessary, the detector was set on a steel platform bolted to
the engine stand. This platform ensured that the detector centerline was
at the same height as the piston ring when the piston was at top dead
center. The platform also allowed the detector to be held close to the
block for high efﬁciency. To count the oil, the detector was placed in the
reservoir—detector stand described earlier.

The test procedure is outlined below:

1) Run the engine at set speed and load for a set time.
2) Stop engine, count the ring.

3) Pump the oil into the reservoir and count the oil.

Note that the oil pump was not run continuously. There were two
reasons for this. The ﬁrst was that the pump had begun to leak. Replacing
the pump was tried but no pump that matched the existing hose size
could be found within the time constraints of the study. The second
reason was that if the pump and engine were run simultaneously the oil
activity would be increasing while it was being counted. This was due to
the fact that the time for the detector to count the low activity in the oil
was initially very large. The data from the oil was used to calculate wear
rates during the test in order to determine at what speed and load levels
the engine should be run. Figure 4.2 shows the power levels the engine

 

29

wasrun at.

The engine was run at its stock compression ratio for the ﬁrst 23
hours of the test. Afterwards, the compression ratio was increased. from
7.6 to 9.5 in the hopes of increasing the wear rate. The engine was then
run for 18 more hours at the same power level to see the effect of this
modiﬁcation. Next, a friction reducing oil additive was added to the oil.
Immediately after this addition, the engine head cracked, therefor no data
as to the effect of the additive was obtained. Then engine was then
motored for 7 hours to see the effect of very low loads. The data from
the motored tests seem to be ﬂawed, however, because during a move of
the test equipment to a new site; a wire in the detector broke, and the
reservoir suffered repositoning.

4.3 Results

Figure 4.3 is a plot of the engine wear rates vs. cumulative run
time. In the ﬁrst hour of the test, data was taken at 15 minute intervals;
and a clear break in period of high wear rates was observed. After this
peak, the wear rates stayed fairly constant despite variations in engine
power and speed. the Another break-in period is evident when the
compression ratio was raised. This break-in did not reach the high wear
rates of the ﬁrst peak, however. The wear rates steady out for about 18
hours after the second break-in period, during which time the power is
held constant. After this period, the head cracked and motored tests
were begun. The data for the motored tests is ﬂawed, however, because
of a broken detector wire and movement of the oil reservoir from its
reference position.

intel-

’~ *<-—-

 

 

CHAPTER 5
SUMMARY & CONCLUSIONS

Parameters affecting the calculation of wear are listed below.

1) Method of measuring radiation

A) The marker method requires much less time to measure
radiation with less than 5% peak area error than the tracer
method does. The marker method is unable to measure a
difference in the level of radiation that is smaller than
approximately 1%. This results ill more time being required,
compared to the tracer method, to estimate wear.

 

B) The tracer method was used to detect the wear rates in this
study while the marker method was unable to detect any
wear; This was becuase the total radiation wont into the oil
was below 1% of the initial radiation in the ring. Thus, while
the tracer method requires more instrumentation, it is more
sensitive than the marker method and estimates of wear rates
can be made with the very ﬁrst measurement. Typical
counting times to reduce errors to 5% for the tracer method,
under the present test conditions, are approximately 50
minutes.

2) Length of test runs: Shorter run lengths will allow a better estimate
of the shape of the wear rate curve. Every test run, however,
requires a 50 minute reservoir measurement, so for short run lengths
most of the time will be spent measuring rather than running.

3) Precision of detector positioning: Since the efﬁciency curves
necessary for activity (and thus wear) calculations are generated for
only one ﬁxed position relative to the detector, the source of radiation
must be in the same position every time a measurement is made.
The detector is extremely sensitive to sample to sample variations in

30

 

4)

31

this position. At a distance of 1 inch from the detector, a variation in
the reference position of 1/8 inch is enough to completely invalidate
the efﬁciency curve.

Oil volume: The accuracy of the tracer method depends on the
accuracy in the measurement of the total oil volume. This volume
can change over time due to the burning of oil by the engine and oil
leaks in the lines. Although the reservoir volume is constant the
volume of oil in it is not, due to the presence of air bubbles. The mass
of oil in the reservoir will not always be the same even if the volume
is constant due to temperature induced density variations.

5) Intrinsic detector parameters:

6)

7)

A) The detector has been known to shut down, for unknown
reasons, during a measurement. When this happens, the
software clock does not stop, meaning that a greater counting
time is indicated than actually happened. A comparison of
spectrum background peaks can be used to check for detector
shutdown. Background peaks from data taken in the same
position should have the same intensity. Large variations in the
intensity may indicate detector shutdown.

B) Any warming of the detector will reduce its resolution. There is
no indicator that this is happening, however, so care must be
taken to keep the detector cool. The effective dewar holding
time is 30 hours so conscientious ﬁlling is necessary.

Test Engine: Very small loads had large effects on engine speed,
severely limiting the range of loads and power that could be studied.
For friction and wear testing of diesel and gasoline powered
automotive engines the radiotracer technique would appear to be
ideal.

No hazard exists from the exhaust given off by the engine run in this
study. The ﬁnal activity of the oil is below the legal limit, which is 2

 

 

naturally occurring radioactive substance and thus is not regulated.

 

32
microcuries per gram, set by the department of Transportation for
disposal as a radioactive substance. The thorium compound is a
I...

 

 

CHAPTER6

RECOMMENDATIONS

1) Detector:

A)

B)

C)

D)

The detector must always remain cool. It must not be placed
too close to the engine block or allowed to be in the reservoir
unless the cooling system is on.

Detector cables should have crimp-on BNC connectors. The
screw-on variety come loose and cause short-circuits.
Detector cables should not be looped because this will cause
system noise.

A pulser should be used to double check the detectors
indicated livetimes.

A more precise method of securing the detector in position
should be used. This method should eliminate the metal straps
that hold the detector to its base, because these have a small
amount of play in them.

2) Environment: A test cell is needed to shield the experiment from wind
and rain. A dedicated cell is needed to conduct wear tests of this

type.

3) Sources should be tested before being used to generate efﬁciency
curves.

4) Oil System:
A) Very careful measurement of the system oil should be made

using a graduated cylinder or similar measuring device. A
means of measuring the oil volume while tests are in progress
should be developed. This method should be more accurate

33

 

 

34

titan the dipstick.

B) The oil system as a whole, (the engine, the pump, the reservoir,
and the oil lines), should be calibrated by using a radioactive
source to quantify the effect of air bubbles on the volume
calculations. The radiation measured in the reservoir divided
by the total radiation should ideally be equal to the ratio of
reservoir volume to the total system volume. The amount by
which these numbers differ will indicate the errors in the
process, much of which may be due to air bubbles.

C) Minimizing the number of joints in the plumbing and soldering
all joints is a necessity.

5) A test engine that was designed to operate at a wide range of speeds
and loads should be used .

6) New software must be purchased or written that meets the following
criteria:
A) Must be able to back up and print out all data and plots.

B) Must be able to communicate with each other, if multiple codes
are used.

C) Must be able to calculate wear rates directly from spectral
data.

7) Once the tests are started there should be no shutdown of the
detector or movement of the oil reservoir.

8) One addition that should be made is a plexi glass shield to cover the
engine assembly. This would keep oil from being sprayed across the
room if the oil hose burns through where it is in contact with the
engine.

 

.q.’ ‘ﬂ

 

REFERENCES

 

36

LIST OF REFERENCES

l. R. Evans.."Radioisotope Methods for measuring Engine Wear: A
Thin Layer Activation Method for the Measurement of Cam Follower
Wear and its Comparison with a Neutron Activation Method". Shell
Research Ltd., Thornton Research Centre, Great Britian, Nov 26, 1979

2. EH. Kinsella and TH. Childs.."A Study of Wear in the Friction
Cutting of Cast Iron Using Thin Layer Activation", Wear, 66 (1981), 241-
255

3. S.G. Fritz and G.R.Cataldi "In Situ Piston Ring Wear
Measurements in a Medium Speed Diesel Engine", Lubrication
Engineering, Vol 46,6, 365-370

4. Eric W. Schneider and Daniel H. Blossfeld. "Effect of Speed and
Power Output on Piston Ring Wear in a Diesel Engine", Intemational
Congress and Exposition, Detroit, MI, February 29- March 4, 1988, SAE
880672

5. S.A.E1bakr, 1.]. Van Heerden, WJ. McDonald, G.C.Neilson,
"Measurements of Nuetron Angular Distribution From the 7li(p,n)7Be
Reaction". Nuclear Instruments and Methods 105, (1972) 519-523, North
Holland Publishing Co.

6. Burkhardt T., McHarris W., Lloyd 1., Rachel T., Schock H.,
Grummon D, Shalek R, Ronningen R, Mallory M, "Developments in High
Energy Ion Implantation For Wear Studies", Michigan State University
Status Report, MSUERL-90-5, 1990

 

 

37

7. Personal Communication

8. Ziegler J.F., Biersack J.D., Littrnark A., "The Stopping Power of
Ions in Solids", Pergamon Press Inc, Maxwell House, Fairview Park,
Elrnsford, N.Y. 10523 USA

9. EG&G Ortec, "Quick List of Nuclear Decay", Academic Press

10. Lindeburg M. Engineer in Training Review Manual, Sixth
Edition, Professional Publications, Inc, Belmont, CA 94002

11. Bevington P.R, "Data Reduction and Error Analysis for the
Physical Sciences", McGraw Hill Book Company

 

 

APPENDICES

 

APPENDIX A

ERROR ANALYSIS

Statistical uncertianties arise from counting a number of events ( such as
emission of gamma photons) per unit time interval. The uncertianty is not
due to imprecision in measuring the time interval or inexactness in
counting the number, but due to the fact that the number of events
ﬂuctuate randomly from sample to sample. The uncertainty in counting a
number of events per unit time interval is given by:

du/u=N1’2/N.

where N is the number of events. This formula is used by the computer
software to calculate the peak area error.

 

Other uncertainties come from tabulated values of branching ratios,
reference values of calibration sources, and almost every measurement
that is made. The propagation of these uncertainties is very important
when determining wear rates. The following equations show the general
form of error propagation in algebraic equations.

x = au 1 bv : 6x2 / x2 = azou2 + b2°v2 + 2abouv2 (Al)
x = iauv : 0x2 / x2 = 0‘12] 112 + ovz / v2 + 20uv2 / uv (A.2)

x=iaulvz ox2/x2=ou2/u2+ov2/v2-2ouv2/uv (A3)

For the present calculations second order terms can be neglected. Thus
the activity and activity error is calculated by the following equations:

peakarea

Activity = I (A .4)
(livetime)(branching ratio)(efﬁciency) ’

 

39

 

40

Activity error = [ (area error)2 + (efﬁciency error)2] “2 (A5)

where the error in livetirne and the branching ratios have been neglected
as they are very small compared to the other errors. Since it is necessary
to average activity values the following formula must be used to weight
the average by the individual activity errors:

X = 2(Xi / Oiz) / 2(1 / 0'3), (A.6)

where xi is the individual measurement, and (Si is the individual

measurement error. The uncertainty in this mean is given by:

0112 = 1 /2(1 loiz). (A.7)

 

 

APPENDIX B

PRODUCTION OF 7BE IONS FROM A 7L1 BEAM AND
A POLYETHYLENE TARGET

The Michigan State University National Superconducting
Cyclotron is studying methods of producing 7be ions for wear studies.
These ions would be implanted into components whose wear properties
are being studied using the radiotracer or radiomarker method. This ion
implantation technique would replace the presently used surface layer
activation technique that can only be used on materials with atomic
number greater than 24. The ability to do ion implantation on campus
would greatly reduce the costs of wear studies and eliminate all the
hazard and paperwork of shipping radioactive components overseas.

One reaction used to produce 7be ions is:
p(7Li, 7Be)n (B.1)

Figure B.1 shows the set-up used to test this reactions production
efﬁciency. The 7Li beam produces 7Be ions when it strikes the
polyethylene targets. The 7Be ions have momentum in the direction of
the 7Li beam so a stack of aluminum foils is needed to stop them. The
number and distribution of the 7Be ions in the aluminum tells us how
many ions were produced and their energies. This was only true,
however, if no 7Be ions were produced by the 7Li beam striking the
aluminum foils on its way through. To test this possibility a control stack
of aluminum foils were struck with the 7Li beam and counted for any
traces of 7Be.

The testing procedure is as follows:

1) An efﬁciency calibration is done for the position where the foils
will be counted, called position 1.

2) A count is made of each stack (stack 1 consisting of the

41

 

42

polyethylene targets and 15 Al foils, stack 2 consisting of the second 15
foils)

3) Stack l was disassembled and each foil counted separately, the
foils being mounted on cardboard sheets.

4) Stack 2 was disassembled and each foil counted separately.

Two sources of eu-152 were used to get the efﬁciency calibration
curve for this test. One, the strong source, had an activity of 14.5
rrricrocuries, while the other, the weak source, had an activity that was
unknown. To detect the small amount of beryllium in the foils they had to
be placed close to the detector to reduce the errors. The strong source
however could not be placed close to the detector because it would
cause a great deal of deadtirne in the electronics. To get around this, the
strong source was used to get an efﬁciency curve at a position further
away from the detector, position 2, enabling the calculation of the activity
of the weaker source. Once the activity of the weaker source was
known it was placed in position 1, the position very near the detector, and
used to generate an efﬁciency curve there. These two curve are shown
in Figures B.2 and B3.

The table below shows the data taken during the foil counts. This
data was used to calculate the distribution of 7Be ions in the aluminum
foils, this relation being plotted in Figure BA. The data was also used to
generate a depth calibration plot of the type received with the piston ring,
as shown in Figure B.5.

 

 

APPENDIX C

EXAMPLE CALCULATION OF EFFICIENCY AND ACTIVITIES

As stated before the total efﬁciency of the detector is given by:

number of photons striking the detector
Total Efﬁciency = (CI)

number of photons emitted by the source

 

"‘7
The total efﬁciency is the product of the geometric efﬁciency,
which is given by:

, . number of rays striking the detector
Geometric Efﬁcrency =

 

C.2
number of rays emitted by the source ( )

and the intrinsic efﬁciency, which is given by:

number of signals generated by the detector
Intrinsic Efﬁciency = (C3)

number of rays striking the detector

 

The geometric efﬁciency is inversly proportional to the distance
between the detector and source. Because less rays strike the detector
when it is further away from the source, as shown in Figure C.l. When
the gamma ray detector is struck by a photon, it gives off a current pulse
called a signal. The intrinsic efﬁciency reﬂects the fact that not every
photon that strikes the detector causes a signal to be generated. In fact
lower energy photons have a greater chance of generating a signal than
do higher energy photons. An efficiency curve must be generated that
shows efﬁciency as a function of energy. This curve would be valid for
only one set position (one ﬁxed distance between source and detector)

43

44

due to the geometric efﬁciency. Such a curve is shown in Figure C.2
Above approximately 200 keV, there is a nearly linear relationship
between the log of the energy and the log of the efﬁciency.

Before any efﬁciency curves can be generated the cemputer
must be able to distinguish between the energies of the signals it is
receiving. Figure C.3 shows a simple analogy of how the computer
receives and stores signals. Pictured is a coin bank. Coins roll down the
slot and fall into the ﬁrst bin that is large enough to hold them. If it is
American currency being put into the bank then one can easily deduce
which slot holds which coins because the sizes of dimes, nickels, etc., are
well known. If however an unknown currency is put in, then the value of
the coins in any particular bin is unknown, only their size relative to one
another is known from the bins they fall into. The multi-channel analyzer
(MCA) and computer function in a similar manner with the signals
generated by the detector. Signals are received by the MCA and put in
bins called channels. The signals generated by the higher energy photons,
like the larger coins, fall into channels further to the right than the signals
generated by the lower energy photons. So, although the signals are
sorted by energy, the computer cannot label any particular channel as
any particular energy. What is needed to identify the energies is called
an energy calibration.

Before energy-channel calibration can be discussed, however, an
important difference between the bank analogy and the computer system
must be recognized. In the bank the number of coins in any bin is known
exactly. In the computer, however, the placement of signals into the right
channels is not completely accurate. In fact, the signals become arrayed
in a gaussian pattern about the channel that corresponds to their energy.
To get the total number of signals for one energy, a count must be made
of the signals in the channel that corresponds to that peak energy, and in
the channels around it. Figure C.4 shows a gaussian shaped peak,
channels being on the horizontal axis and number of counts on the
vertical. Integrating under the gaussian curve gives the peak area, which

 

 

M18 1
mm it
of con
lower

tents

torts;
harm
elimpl
Fight
baclgn
peaks 1
wi’hin
clear i

ii tl
Till“.

Til

8‘8

Ila}:
N017;

ll

45

is the total number of counts, N, above background at that energy. The
error in this area is approximately equal to the square root of the number
of counts divided by the number of counts, so the greater the area the
lower the error. To rrrinirnize errors it is then necessary to have many
counts and thus long counting times.

Area error = N1/2 / N , ‘ (C4)

To enable the computer to recognize which channels
correspond to which energy an energy calibration must be performed. A
known source is placed in front of the detector and measured. An
example spectrum taken from a count of europium 152 is shown below in
Figure C .5. Examination of the spectrum shows several peaks above the
background. The speciﬁc energies at which europium has its strongest
peaks must be determined from tabulated data and located by their size
within the spectrum shown. Then, since the channels of these peaks are
clear from the spectrum, and their energies have been determined from a
table; linear interpolation between these channels allows the calculation
of the energies of every other channel. Once this energy-channel
relationship is known, the identiﬁcation of any unknown source is easily
accomplished. The energies of the biggest peaks in the spectrum are
identiﬁed and compared to tabulated values to determine what nuclide
produces them. Recall every nuclide has a characteristic peak energy
pattern.

Once the channel-energy relationship is known, an efﬁciency curve
can be generated. To generate an efﬁciency curve, a known source must
be measured at some ﬁxed position from the detector. Since the source
activity and identity are known, the exact number of gamma photons per
second (intensitiy) it emits at its characteristic energies is known. For
example, if our known source is europium 152 we have the following
information:

liable 2,1 Bmchrg' g Ratios Qf Europim 152

 

 

Energies Branching Ratio
[keV] [Gammas per 100 decays]
121.78 .284
344.28 29.58
1112 - 13.56
1408 20.85

Since its activity is known, for example in this case 9 nricrocuries
(333000 d.p.s.), intensities can be calculated:

intensity (gammas per second) = activity * branching ratio

 

 

WW
Energies Intensity
[keV] [Garnmas per second]
121.78 94572
344.28 98501
1 1 12 45155
1408 69430

If the source is counted for 1000, seconds the following data might
be the result:

47

Table C.3 Measured Intensities

 

 

 

Erlel'gies Intensity

[keV] [Gammas per second]
121.78 260410

344.28 199550

ll 12 54280

1408 72087

Dividing the peak areas by the livetime gives the number of
photons counted per second at each energy. The livetime in this case
was 1000 seconds. The total efﬁciency is then the number of counts per
second divided by the number of gamma photons per second.

Efﬁciency = (counts / second) / (photons / second)

 

 

W
Energies Efﬁciency
[keV] [counts per photon]
121.78 .002754
344.28 .002026
1 1 12 .001202
1408 .001038

A log-log plot of this is shoWn in Figure C6.

The error in the efﬁciencies is calculated using the equation

48

Efficiency error = [ (area error)2 + (activity error)2 +

(branching ratio error)2 ] 1’2

The area error is given by the computer with the peak area. The
activity error is the uncertainty of the source activity. The errors in the
branching ratios are typically tabulated, but for the most part they are so
small that they can be neglected.

Once the errors in the efﬁciencies are known, a line is ﬁt through
the points, the ﬁt being weighted by the errors. A computer routine given
in Reference [1] is used to do this. The ﬁt allows the calculation of
efﬁciencies for energies that are between those of the europium peaks.
The ﬁt error is derived from the efﬁciency errors.

Once the efﬁciency curve is known for some detector position, it
can be used it ﬁnd the activity of an unknown source placed in this
position. The activity is given by the equation below

peak area

 

Activity =
(livetime)(branching ratio)(efﬁciency)

Livetime is the amount of time the computer actually spent
counting. It is real time minus the time the computer shuts down to
process signals. This shutdown time is called the deadtirne. Livetime =
realtirne - deadtirne. Dividing by the total efﬁciency gives the total
number of gamma photons, called intensities, emitted by the source at
that energy. Dividing by the branching ratio, which is the intensities per
100 disintegrations for this energy, the total number of disintegrations per
second is found. This is the measurement of activity. This activity
calculation will be done for the major peaks in a spectrum and the error
weighted mean of the results will be the activity, as shown in chapter 2.

49

Example: Using a source of Eu-152, an energy and efﬁciency
calibration is created as shown in Figure. Next, the unknown source is
placed in the position of known efﬁciency and is counted for 1000 see,
resulting in the spectrum shown in Figure C.7 below. By comparing the
energies of the four largest peaks in the spectrum below, ( 846, 1037,
1238, 1771 keV), with tabulated values, one can identify the nuclide as
Co-56. It has the following branching ratios (from tables) and peak areas
(from the count data):

Table C.5 Peak Areas

Energies Branching Ratio Peak Areas Area error

 

 

 

 

[keV] [gammas [100 decays] [counts] [% error]
846 .9996 121590 0.3
1037 .1016 15613 0.9
1238 .6697 63573 0.4
1771 .1551 11053 1.0

The efﬁciency of the detector at these energies is then calculated
from the line ﬁt to the efﬁciency curve on Figure C.6.

Table C.6 Efﬁciencies at Cobalt 56 Energies

Energies Efﬁciency Efﬁciency Error
ﬂ [counts per photon] [% error]

 

 

846 .0009099 0.2
1037 .0008356 0.2
1238 .000776 0.2
1771 .00066815 0.2

From here the activity can be calculated from each peak area

50

peak area
Activity =

 

(livetime)(branching ratio)(efﬁciency)

Activity error = [ (area error)2 + (efﬁciency error)2] “2

This formula for activity error assumes negligible error in the
branching ratio and livetime.

So the activities and errors are

Table C.7 Activities of Cobalt 56

 

 

 

Energies , Activity Activity Error
[keV] [Decays per second] [% error]
846 133683 0.36
1037 183905 0.92
1238 122329 0.45
1771 106658 1.02

Ideally all these values would be the actual activity and thus all the
same. As it is, an error weighted mean of the above values will
approximate the actual activity:
2(activityi / activity erroriz)
Mean activity =

 

2(1 / activity erroriz)

51

1

 

Mean activity error =

[ X( 1 / activity errori2 )]1/|2

52

FIGURES

53

piston ring

 

 

 

 

 

 

 

 

----- v :
: —: I
gamma ray detector
engine

 

Figure 2.1A The Marker Method

. , lead shield
piston rrng . .
orl reservorr

é;

 

 

OOOOOOOOOO

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

‘ﬁ L_,
I gamma ray detector
engine i“
7’F
oil lines

Figure 2.1B The Tracer Method

54

CALIBQATION'CUQVE

 

INTENSITY ---

-—- Rel.

 

"VIC?” : OW

88°70! “-80

MI at!!!“ : ear-snort

be?! 0' “the?!“ : In 00 8”!

one 0' alumni m u I”!
“avenue W : I".

lens-n : Deer" aerrv. who're:
. : rareunorehe
g.- corn ﬁr- Oedlue

 

 

to 20 30 40 so so 70 so
----- DEPTH ~---- [pm]

Proton

 

Neutron

Figure 2.3 The Atom

 

55

 

Figure 2.4 Be7 Decay Path

 

 

 

:IIﬁI

4“ g .

4 .W 11 l Iﬁ

in p
1 II M m N I“ M in 1111

Figure 2.5 Cobalt-56 Spectrum

 

 

56

 

 

 

 

 

 

engine dynamometer
coupling
dynamometer
pump controller
engine stand 0%

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

oil reservoir

 

computer I

 

 

 

 

 

 

 

 

 

detector

 

 

 

 

 

 

liquid

nitrogen

tank

 

Figure 3.1 System Components

 

 

57

 

Figure 3.2 Engine Block

 

 

Figure 3.3 Dynamometer

 

Figure 3.4 COUpling

if}.

 

58

mug-WMVMM....L.. .,,~__ w “a *

.g...«.%,_ m“

"V - WM"

59

 

 

Figure 3.7 Auto Fill Controller

 

Figure 3.88 Detector Full

61

 

Figure 3.9 Oil Reservoir

 

Figure 3.10 Reservoir-Detector Stand

62

 

 

Figure 3.11A Reservoir Platform

 

Figure 3113 Reservoir Platform

63

 

Figure 3.13A HEPA Filter Housing

 

rm. 2*“ '.~.- wwsa

     

We».

 

Figure 3.14 Test Set-Up

 

 

 

 

rIIIII .Iit. .- Ti

1 t ‘:

r . "

I ' ‘

C 3 . 1

0 9 ‘ t

u

n 1 r ‘
t - r

s . 2 «
' I

’1? 7 .4 1

E i I. 3

41M“! W. 3

WW

Channels

Figure 3.15 A Good Fit

 

 

 

rIItII 1‘. ' 7
, I t .
i . . 1
. ' I .
I
C i Q 1
o
u I I ‘
n i . 1
t I
s i 7 ' 4
' r
r . . r
I r
r .
l . t 1
t r“ m It. ‘1
I
r”; 4
III IlII
Channels

Figure 3.16 A Too Narrow Fitting Interval

 

unseen

 

 

 

hm .
r I e
I
r H -
I
r ‘ 1
I
.m .
i
D . I
i
b I I
’ .. ' m '
til I -
r I I I . . I d
i *
h . I“ * . . I
’ II t ﬁt“ ‘
“I“ ' I a
mnIuIrhrrrIr I 'rullll I I I
II 135
Channels

Figure 3.17A Parabolic Continuum

 

anal-son

 

 

pun '

r ' d

r- II "
I

D A I

I

in "'

D . * d
t

b ' "

r . H

.n .

* I
r I I I Q ‘
f i '
r- * . u
g I § ’ ‘“ 1
t‘
P“ I I:
W
111 135

Channels

Figure 3.17B Linear Continuum

67

efficiency vs. energy ct position A

 

0.0100

IIIII
.1

0.0010

efficiency

 

0.0001

 

L l A 1 l L

I T

l

[[111

 

 

100
energy kev
log(ett)= -0.279541 tlog(energy) +

Figure 3.18 Efﬁciency Curve at Position A

-2.37772

1000

68

efficiency vs. energy in reservoir

 

 

 

 

0.100 .. ' I
>\
o
s d
“g 0.010: ..
-.-. : -
n- d
0.001 r r r r r i . l l
1000
energy keV
Iog(eit)= -0.590455 clog(energy) + -0.391218

Figure 3.19 Efﬁciency Curve in Reservoir

efficiency

69

efficiency vs. energy at position R
I ' ' I *1 r r j

 

0.0100

0.0010

 

0.0001

 

M J 1 l 1 1 l 1 1

 

 

1000
energy keV
Iog(eff)= -O.418068 olog(energy) + —1.81715

Figure 3.20 Efﬁciency Curve at Position R

H

INTENSITY ---

on---- n.’ .

70

CALIBQATION'CUQVE

 

 

 

Figure 4.1 Depth Calibration Curve

1
“YEW“ : .WANIW
IDOT” : “-00
MI MATERIAL : CAST-IMO
D"! 0' ACTIVATIW : m 0. I”!
05?. or “AW“: m II I”!
ACTIVATIW W t “I.
“I“. : um ACTIV. DWILI
.8 0 . 3 '" 'I‘TU‘ II“
enc- Ibun Igc- UJJIVI
.6“
‘ db
1
O
O
9
§
9
o . .
o
o - A f“ I" *
0 10 20 3O 40 50 60 7O
----- DEPTH ----- Kym)

 

”mac-o

71

 

 

 

 

 

 

 

 

 

 

 

 

0‘
O

8 7.5 hp.
6.5 hp.
6 _—|_
" 5.2 hp.
1' t
4
{rm-'24 hp.
2
0
0 10 20 ' 30 4o 50
HOURS RUN
Figure 4.2 Power vs. Run TIme
500 E ''''''' I """"" W W r i rrrrrrrrr I rrrrrrrr 4
.3
N
M 300, compression ratio 3
H E increase 3
O E 1 E
U 200: i
R E 3
100: l -3
i t m
0 .rr. lrnrrrrrr L ........ ~ :1
0 10 20 so 40
HOURS RUN

Figure 4.3 Wear Rate vs. Run Time

 

 

 

72

 

 

3, Whitman
mneam f\1/

polygtgheygene 15 Al foils

Figure B.1A 7Be Production

7Li Beam \

Figure B.1B 7Be Production Control

 

 

73

efficiency vs. energy at position 6

 

01310

   

efficiency

 

 

0.001 . r . t . . . . l
1000

 

energy keV
log(efi)= —0.476254 tiog(energy) + -1.02603

Figure B.2 Efﬁciency Curve at Position 6

efficiency

0.10

0.01

74

efficiency vs. energy at position 1

 

 

L

I

 

 

energy keV

log(eff)= -O.489068 ulog(energy) + -0.438379

Figure B.3 Efﬁciency Curve at Position 1

1 000

 

 

5' of 7be ions per 5 micron foil

2.5)(109

2.0x109

L5x10

10x10

5.0x10

0

75'

. ii Pf 7ble ions vs Implantation depth
. . I ,

 

T

T T T I I T T T I T T T T I T T T T I T T T

 

r l L 1

T

l l l l l l l 1 l l l l l l .I I 1 l 1 I l I l J_

 

 

O

20 40
Implantation depth (microns)

Figure B.4 Ion Distribution Curve

0)
0

rr

 

 

activity remaining (dps)

100

76

percent activity remaining vs. surface removed
T T I r I I I I r

 

80-

60'-

4o—

20—

 

1 A l L J J. I

T

 

 

20 40
Surface removed (microns), in 5 micron increments

Figure B.5 Depth Calibration Curve

60

 

 

77

 

 

 

 

 

 

 

|_J

100m Higthﬁciency

 

 

   

 

 

  
 

 

 

 

 

 

 

l" 13 uu J

Low Efﬁciency

 

Figure C.1 Effect of Distance on Detector Efﬁciency

efficiency

0.010

0.001

78

' efficiency vs. energy

 

 

 

L

 

energy keV
Iog(eff)= —0.476254 slog(energy) +

Figure C.2 Typical Efficiency Curve

1 000

-1.02603

 

 

79

 

A
\2/

 

 

 

 

 

 

 

 

 

0500
QC)

 

Figure C.3 Coin Bank

 

 

 

 

 

 

 

 

 

 

 

 

Figure C.4 Gaussian Pattern of Signal Collection

mast-son

80

 

 

l

mmmummmuu’

Channels

Ill

Figure C.5 Europium 152 Spectrum

 

efficiency

0.010

0.001

81

 

 

 

1 1 L

l

 

energy keV
iog(eff)= . -0.476254 ‘log(energy) +

Figure C.6 Example Efficiency Curve

1 000

-1.02603

 

want-:00

82

 

:Mf :n m:

P I
4“ g . .
. f
I : d

5. z
'- ' :
I ‘ z i ; l
'4 d it 1 ‘ ‘
ill

 

 

 

11mmmmtmmm

Channels

Figure C.7 Cobalt 56 Spectrum

"Ilillllllliﬂﬂiliillr