UNIT
VIBRATIONSANALYS AV ELEKTRONIKENHET
Hanna Fahlgren
THESIS 2013
MECHANICAL ENGINEERING
This thesis has been carried out at the School of Engineering in Jönköping in
the subject area Mechanical Engineering. The work is a part of the Master of
Science program, Product Development and Materials Engineering. The author
takes full responsibility for opinions, conclusions and findings presented.
Scope: 30 credits
Date: 02/04/2013
Abstract
Abstract
Components in Airplanes are exposed to vibrations which can cause them to fail
due to fatigue if they are not constructed for the extreme environment. High
stresses can develop in products during resonance if parts own the same natural
frequency. The acceleration induced by one component will couple to the next
and give it an even higher acceleration and thus higher inertial displacements and
stresses.
Simulations of resonant frequencies and analyses of their couplings were
conducted on electronic units developed by SAAB Electronic Defence Systems in
Jönköping. A reinforcement proposal for one of the PCBs in one of their current
electronic units was given by the author.
The result of the simulation showed that the first resonant frequency of the
investigated PCB and the current chassis was not separated enough, and coupling
of their accelerations during resonance occurred. A rib design was developed by
the author to increase the natural frequency of the PCB, which eliminated the
coupling with the current chassis at its first natural frequency.
A new chassis was being developed by SAAB, which would be produced by
casting. Simulations of the casted chassis were also preformed and evaluations
with regard to its natural frequency were made. The casted chassis first natural
frequency was well separated with the PCB. But the weight ratio between the
casted chassis and the PCB would cause high acceleration forces during resonance
which would be transmitted to the PCB which is not preferable.
When designing components it is of great importance to know the environment in
which they will work. The engineer needs to possess good skills with respect to
the phenomena acting on the product in order to find the right input values for
simulations. One must also keep in mind that simplifications need to be made so
that simulations will not be to complex and therefore evaluation of the result must
be preformed to ensure credibility.
Keywords
ANSYS
Chassis
Coupling
Printed Circuit Board
Resonance frequency
Simulation
Transmissibility
Summary
Summary
Flygplanskomponenter utsätts för vibrationer som kan leda utmattningsproblem
om de inte är konstruerade för den extrema miljö de verkar i. Höga spänningar
kan utvecklas i produkter under resonans när komponenter har samma
egenfrekvens. Accelerationer som framkallas av en komponent under resonans
kan överföras till en närliggande enhet och ge denna en ännu högre acceleration
om deras egenfrekvenser sammanfaller. Detta medför att stora förskjutningar och
spänningar i de utsatta komponenterna vilket kan leda till brott och sprickor.
Simuleringar av resonansfrekvenser och analyser av deras kopplingar utfördes på
elektronikenheter som utvecklats av SAAB Electronic Defence Systems i
Jönköping. Ett förstärkningsförslag för ett av kretskorten i deras nuvarande
elektronikenhet togs fram av författaren.
Resultatet av simuleringen visade att den första egenfrekvensen hos det
undersökta kretskortet och det nuvarande chassit inte var tillräckligt separerade,
vilket medförde koppling av deras accelerationer vid resonans. Ett
förstärkningsförslag utvecklades av författaren för att öka kretskortets
egenfrekvens, vilket eliminerade kopplingen vid det aktuella chassiets första
egenfrekvens.
Ett nytt gjutet chassi var under utveckling av SAAB och simuleringar utfördes
också av detta chassi. Utvärderingar med avseende på dess naturliga frekvens
gjordes. Den första egenfrekvensen hos det gjutna chassit var väl separerat från
kretskortets, men viktförhållandet mellan det gjutna chassit och kretskortet skulle
medföra höga accelerationskrafter under resonans. De höga krafterna skulle då
överföras till kretskortet vilket inte är fördelaktigt.
Vid produktframtagning är det av stor vikt att känna till miljön där
komponenterna ska verka. Ingenjören behöver besitta goda kunskaper om de
fenomen som verkar på produkten för kunna ta fram rätt indatavärden för sina
simuleringar. Ingenjören måste också veta vilka förenklingar som måste göras så
att simuleringar inte blir för komplexa. Förenklingarna medför att utvärderingar av
resultatet bör utföras för att säkerställa dess trovärdighet.
Acknowledgements
Acknowledgements
I wish to express my sincere gratitude to everyone at the mechanics department at
SAAB Electronic Defence Systems in Jönköping. You all made me feel very
welcome and a part of your team.
I would also like to give special thanks to my supervisor at SAAB, Rebal Marcos,
for helping me with every little problem I had and to Fredrik Kroll, head of the
mechanics department, for giving me the opertunity to do the thesis at SAAB.
I would also like to gratefully acknowledge my supervisor at Jönköping University,
Peter Hansbo, for guiding me through my thesis work.
Table of Contents
Table of Contents
1 Introduction ............................................................................. 9
1.1 BACKGROUND ....................................................................................................................... 9
1.1.1 SAAB ........................................................................................................................... 9
1.1.2 Problem Definition .................................................................................................... 10
1.2 PURPOSE AND RESEARCH QUESTIONS ................................................................................... 11
1.3 DELIMITATIONS ................................................................................................................... 11
1.4 OUTLINE ............................................................................................................................. 12
6 References .............................................................................. 48
7 Search Terms ......................................................................... 50
8 Appendices ............................................................................. 51
List of Tables
List of Tables
List of Figures
List of Symbols
a Acceleration (m/s²)
A Area (m²)
B Rib spacing (m)
c Damping ((Ns)/m)
C Convergence (%)
CM Controller module
CW Cross wise
EDS Electronic Defence Systems
E Young’s modulus (Pa)
f Frequency (Hz) and force with time domain (N)
F Force (N)
g Acceleration of gravity (9.81 m/s²)
G Acceleration in gravity units (dimensionless)
H PCB Horizontal PCB
HPM High power module
k Stiffness ratio (dimensionless)
m Mass (kg)
n Amount of gaps (dimensionless ratio)
LW Length wise
p Percentage (%)
PCB Printed circuit board
RBE Rigid body element
Q Transmissibility (dimensionless ratio)
R Frequency ratio (dimensionless ratio)
x Displacement (m)
x Velocity (m/s)
x Acceleration (m/s²)
t Time (s) and thickness (m)
v Velocity (m/s)
V Volume (m³)
w Width (m)
W Weight (kg)
Greek symbols
ε Strain (%)
ρ Density (kg/m³)
σ Stress (Pa)
ω Angular velocity (rad/s)
Introduction
1 Introduction
The thesis was carried out at Jönköping University, School of Engineering, in the
field of Mechanical Engineering. The work was part of the master program
Product Development and Materials Engineering.
The specification of the project was made by SAAB Electronic Defence Systems
(SAAB EDS) in Jönköping, where the work was conducted. The author was
assigned to carry out vibration analyzes and improvement investigations of an
electronic unit. The electronic unit controls the redundant system of flaps and
slats in airplanes. The project was created since vibrations in the electronic unit
needed to be investigated regarding its impact on fatigue failure.
1.1 Background
SAABs business concept is to constantly develop, adopt and improve new
technology to meet the changing customer need. Their high quality products
therefore demand thorough investigations to ensure a long life in extreme
environments. The components will be exposed to high temperature changes and
vibrations which can cause high stresses and fatigue failure if it is not taken into
account. The airplane components are also exposed to high stresses and shock,
during flight and landing.
1.1.1 SAAB
SAAB offers world-leading products, services and solutions from military defence
to civil security. Their most important markets today are Europe, South Africa,
Australia and the US. They have around 13,000 employees and the yearly sale is
approximately SEK 24 billion. Their research and development account for about
20% of the sales. SAAB consists of five different business areas, which are [1]:
Aerodynamics
Dynamics
Electronic Defence Systems
Security and Defence Solutions
Support and Service
SAAB Electronic Defence Systems develop products in the area of radar and
electronic warfare. Their portfolio includes airborne, land based and naval radar,
electronic support measures and self-protection systems [1].
9
Introduction
Today’s airplanes are optimized in regard to weight, which has lead to a decrease
of the size of the airplane wings. When the airplane is up in the air at its normal
speed, the lifting area of the wing can be decreased. To increase/decrease the wing
area the airplane is constructed with slats and flaps. To ensure safe take off and
landing, a redundant system is used for this action. The additional system in
SAABs airplanes is run by an electronic motor which is controlled by an electronic
unit placed inside the airplane. The electronic unit is exposed to vibrations which
could cause some components in the PCBs to fail. Vibrations in the electronics
unit therefore needed to be investigated to eliminate further component failures.
When the thesis was conducted, the casing of the electronic unit was
manufactured from aluminum sections that were held together by rivets. A new
casing manufactured by casting was being developed to reduce manufacturing
costs.
Resonant frequencies and their couplings in the electronic unit were investigated
and ways to reduce the displacements of the components were explored. The
focus of the simulations was set on the controller module (fig. 1), since it has the
most soldered components, where fatigue and tearing could be a problem in lead
wires and solder joints.
10
Introduction
1.3 Delimitations
A product is only as strong as its weakest link and the weakest link in the
electronic unit was estimated to be the controller module since it has the most
soldered components. The two smaller PCBs have larger components that are
glued well to the PCBs. Since there have not been any noted problems with the
glued joints or cracking in the board material, the enforcement proposal will not
include these PCBs. The high power module has fewer small components than the
controller module and it has more supports holding it in place, which would make
it more tolerable. The casings and the other components in the electronic unit,
such as the transformer, the bolt in filters and the connectors are far more robust
than the PCBs, and they will not be investigated either for further enforcement.
Enforcement proposals for the electronic unit with the casted cover will not be
investigated since the design is not final. Only investigations of how it affects the
controller module will be performed.
11
Introduction
1.4 Outline
Chapter 2 will discuss how vibrations affect components in an airplane with
regard to resonant frequencies, transmissibility and displacement. It will also
discuss how the design of the chassis can affect the PCB and different fastening
technique of PCBs. Methods of decreasing displacements in PCBs, such as
dampers and stiffeners, will then be discussed. The procedure of simulation will
also be presented.
Chapter 3 will show how the gathering of input parameters and simulations of the
electronic units were performed. It will also show the method used to answer the
research questions. Analysis of the electronic units will be presented and
enforcement proposals will be stated.
Chapter 4 will show the conclusions of the simulations and evaluate the
enforcement proposal. Evaluations of the casted casing will also be presented.
Chapter 5 will discuss the method used for finding the answers to the research
questions. It will also discuss the simulation result and its reliability. The report
will be summed up and proposals for further investigations in the subject will be
presented.
12
Vibration of Electronic Equipment in an Airplane
2.1 Vibration
Vibration is when an object is subjected to an oscillating motion that causes the
body to move back and forth. If the oscillating movement is repeating itself, the
vibration is called periodic. The simplest periodic motion is usually represented by
a sinus wave (fig. 2). If the continuous motion does not repeat itself, it is called
random vibration (fig. 3). The measurement of cycles per second is called Hertz
(Hz) and the maximum displacement is called the amplitude of the vibration [2].
1.5
0.5
Displacement (y)
Amplitude
0
0 1 2 3 4 5 6 7 8
-0.5
-1
Period
-1.5
Time (t)
13
Vibration of Electronic Equipment in an Airplane
1.5
0.5
Displacement (y)
0
0 1 2 3 4 5 6 7 8
-0.5
-1
-1.5
Time (t)
When the mass is displaced to a given position (x(t)), the force (f(t)) acting on the
system can be calculated [2]:
d 2 x(t )
f (t ) m kx(t )
dt 2 (eq. 2.1 [2])
If the force subjected to the mass is a simple harmonic excitation with the
frequency (ω), it can be described as [2]:
14
Vibration of Electronic Equipment in an Airplane
10
8
Response (kx(t)/F)
0
0 100 200 300
Frequency (Hz)
15
Vibration of Electronic Equipment in an Airplane
2.3 Transmissibility
The transmissibility (Q) of an object shows its reaction during vibration. The
greatest transmissibility occurs at the resonance frequency. The transmissibility
can be calculated if the output force ( FOut ) and the input force ( FIn ) are known as
follows [2]:
FOut
Q
FIn (eq. 2.5 [2])
For a lightly damped system the transmissibility at resonance can be calculated
from the stiffness (k), mass (m) and damping (c) using the following formula [2]:
km
Q (eq. 2.6 [2])
c
When the stiffness of an object is increased it is known that the damping is
decreased. Equation 2.6 shows that this will lead to an increased transmissibility.
Instead if the mass is increased, it is known that the damping will increase which
result in a decreased transmissibility [2].
By extensive testing it has been shown that many epoxy fiberglass PCBs, with
various edge restrains and closely spaced electronic components, have a
transmissibility at the resonance frequency ( f n ) that can be approximated by [2]:
Q fn (eq. 2.7 [2])
Equation 2.7 shows that when the natural frequency is increased, the
transmissibility is also increased.
16
Vibration of Electronic Equipment in an Airplane
17
Vibration of Electronic Equipment in an Airplane
2.6 Fasteners
If a PCB is subjected to a severe shock and vibration environment, as in an
airplane, a loosely fastened PCB will develop high acceleration loads, which will
lead to high deflection and stresses in components mounted on the PCB [2].
There exist many different types of fastening techniques such as screws, nuts,
rivets, clips and board edge guides. The methods used for ease of installation and
low cost are usually not satisfactory for severe shock and vibration environments
[2].
18
Vibration of Electronic Equipment in an Airplane
A board edge guide that grips the edge firmly is desirable since this can reduce
deflection due to edge rotation and translation, which in turn will increase the
natural frequency. Due to friction and relative motion between the PCB and the
edge guide, more energy will be dissipated during vibration, which will reduce the
transmissibility during resonance. Board edge guides are in general more costly
than other fastening techniques because of the high tolerances needed for a tight
fit and the complication of attaching the PCB to the connectors [2].
Bolted joints will experience a substantial amount of relative motion during
resonance. The relative motion may add damping to the system and reduce the
transmissibility and the natural frequency of the structure. When bolts are used to
fasten the PCB an efficient factor can be estimated. The bolted efficiency factor is
the bolts ability to hold to membranes together during vibration. For a welded
connection it is 100%, for a joint that has no connection it is 0% and for regular
screws it is usually about 25% [2].
2.7 Damping
When damping electronic equipment one must keep in mind to provide sway
space to ensure that objects do not collide. Connecting wires and couplings must
be designed to guarantee that the damped parts can move without tearing the
wires/couplings or causing them to fail due to fatigue. The design of the damping
equipment must therefore consider the specific environment where the equipment
is installed. If space is an issue, like in most airplanes, it might be more practical to
use hard mounts [2].
For low natural frequencies below 50 Hz, dampers can be quite efficient. But for
natural frequencies above 100 Hz, dampers are not very effective in reducing the
transmissibility. Most dampers are also rather sensitive to high temperature
changes because of the polymer materials that are most commonly used. At a low
temperature, the damper can become too stiff and at high temperatures, they can
become too soft. One must also keep in mind that a good vibration isolator is
often a poor shock absorber and vice versa [2].
2.8 Stiffeners
Ribs are often mounted on PCBs to increase the resonant frequency and reduce
board deflection during resonance. For high frequencies they often work better at
reducing the transmissibility than dampers. The reduced board deflection will
reduce the stresses in the electronic component lead wires [2].
Ribs are mostly fabricated from steel, aluminum or epoxy fiberglass. Caution must
be taken when using metal ribs since short circuits can occur across exposed
electrical printed circuit strips. Ribs can be bolted, riveted, soldered, cemented,
welded, or cast integrated with heat-sink plates. They can be attached around the
edges and/or on the surface in several directions depending on the type of
displacement prevented [2].
19
Vibration of Electronic Equipment in an Airplane
If the ribs are not designed properly, they will not increase the stiffness of the
PCB. The ribs should be placed so that they carry the load directly to the support
(fig. 6), if this is not possible, a secondary membrane should be added to carry the
load to the support (fig. 7) [2].
To estimate the spacing (B) of the stiffening ribs the thickness of the PCB (t) can
be used as follows [4]:
B 30t (eq. 2.19 [4])
The stiffness of the material used for the ribs should be at least equal or greater
than the stiffness of the PCB [4].
20
Vibration of Electronic Equipment in an Airplane
When attaching the PCB to the chassis, constraints should be carefully selected.
The translational displacement can in most cases be simulated by a fixed
constraint, but the rotational displacement needs rotational spring elements since
all fixing methods will display some flexibility. The stiffness of the spring elements
can be found using a trial and error approach, tuning the resonant frequency of
the model to that of the vibration tested prototype [5].
There are different ways of modeling bolts in a simulation, such as: no bolt,
coupled bolt, RBE bolt, spider beam bolt, hybrid bolt and solid bolt. The table
below describes the modeling technique and its pros and cons (tab.1) [7].
21
Vibration of Electronic Equipment in an Airplane
Spider beam The head and nut are Less elements than Extra work required
bolt represented by line hybrid and solid bolt. to simulate head/nut
elements in a web- Tensile, bending and stiffness.
like fashion. The stud thermal loads can be
is represented by line transferred through the
elements. spider elements.
Ease in extracting results.
Hybrid bolt The head and nut are Better accuracy in Stress distribution in
simulated by solid comparison to previous the stud is not
elements and the simulation techniques. accounted for.
stud is simulated by Tensile, bending and Coupling of line
line elements. thermal loads can be elements to stud is
transferred through the required which ads
line elements. model complexity.
Full stress distribution
can be calculated in head
and nut.
Simple stud section
which result in shorter
computational time in
comparison to solid bolt.
Ease in extracting results.
Solid bolt The head, nut and Gives the most accurate The computational
stud are modeled as results. time is greater than
solids. Tensile, bending and in the previous
thermal loads can be simulations.
transferred. Contact elements
Full stress distribution in are required between
head, nut and stud can head/nut to the
be computed. flanges.
22
Vibration of Electronic Equipment in an Airplane
When modeling bolts one must take in consideration the conditions under which
the bolt will be used such as compression, tension and shear stresses or a
combination of these. This is important since the boundary conditions and
simulation technique affects the output data, the computational time, and the
accuracy of the result [7, 8].
2.9.4 Mesh
There exists several different element types that can be used when meshing an
object. When a three dimensional model is used, tetrahedral, pyramid, wedge or
hexahedral elements can be used (fig. 8). The tetrahedrally shaped element is a
good choice for complex shapes, but it causes a higher apparent (numerical)
stiffness in bending than the hexahedral element [10].
23
Vibration of Electronic Equipment in an Airplane
24
Simulation of the Electronic Unit
25
Simulation of the Electronic Unit
Three controller modules were used for the tensile test and two of the controller
modules were cut in half at opposite directions. To be able to attach the PCBs to
the tensile test machine, some components had to be removed around the
gripping area to ensure good contact.
The specimens were not pulled to their rupture point since only the modulus of
elasticity was needed. The stress strain curves from the test can be found in figure
9 and 10.
1.80E+01
1.60E+01
1.40E+01
Stress (MPa)
1.20E+01
1st half
1.00E+01
2nd half
8.00E+00 Whole PCB
6.00E+00
4.00E+00
2.00E+00
0.00E+00
0.00E+00 5.00E-02 1.00E-01 1.50E-01 2.00E-01 2.50E-01
Strain (%)
26
Simulation of the Electronic Unit
2.00E+01
1.80E+01
1.60E+01
1.40E+01
Stress (MPa)
6.00E+00
4.00E+00
2.00E+00
0.00E+00
0.00E+00 1.00E-01 2.00E-01 3.00E-01 4.00E-01 5.00E-01
Strain (%)
Two points on the curves straightest parts were used to calculate the elastic
modulus (E) and the result from the calculations (eq. 3.1) is shown in table 2.
1 2
E (eq. 3.1)
1 2
The main material of the PCB is epoxy fiber glass (FR-4) which is a woven fiber
glass that has a theoretical Young’s modulus of 24.1 GPa in the LW direction and
20.7 GPa in the CW direction. The test result shows a much smaller elastic
modulus of 42-82 MPa in the LW direction (tab.2). The test values were therefore
discarded and not used for the simulations.
Since the test values of the PCBs were discarded more detailed models had to be
computed, also taking into account the copper films. The thickness of the copper
film and the percentage of copper in the different layers were considered in the
calculations (appx B).
27
Simulation of the Electronic Unit
28
Simulation of the Electronic Unit
29
Simulation of the Electronic Unit
30
Simulation of the Electronic Unit
The PCBs were then simulated separately with their larger components and their
first natural frequencies can be found in table 4.
The maximum excitation of the controller module was found in the area where
the support pad and plate is located (fig. 13).
In the next simulation all components of the new electronic unit was added. The
coupling of the natural frequencies was investigated between the chassis and the
controller module and the first coupling was found at 286Hz (fig. 14). In the
physical test made by SAAB, the first coupling mood of the bottom plate and the
controller module was found at 295Hz.
Figure 14. Coupling between the current chassis and the controller module.
31
Simulation of the Electronic Unit
Figure 15. Coupling between the casted chassis and the controller module.
32
Simulation of the Electronic Unit
To find clamping solutions for the PCBs bench marking was used. Electronic
units from Esterline Technologies Corporation and Schroff/Pentair Technical
Products were investigated.
Figure 16 shows an electronic unit used for power distribution and control in
airplanes, helicopters, military equipment, boats and industries. It has PCB board
edge guides made from lightweight metal, which reduces weight and makes
maintenance easier [12].
33
Simulation of the Electronic Unit
Figure 16. Electronic unit with PCB board edge guides [12].
Figure 17. Robust electronic unit with modular clamshell PCB inserts [13].
34
Simulation of the Electronic Unit
The simulation showed that the board edge guides where not enough to give the
controller module a reasonable first natural frequency. Ribs were therefore added
to the controller module to increase the natural frequency. Equation 2.17 was used
to find proper spacing between the ribs.
B 30t 30 * 2 60mm (eq. 3.3)
The number of ribs (n) was calculated using the width of the board (w).
n w / B 180 / 60 3 (eq. 3.4)
The simulation was first run setting the width of the ribs to 8mm, changing the
thickness of the ribs to approach the aimed frequency of 618Hz (fig. 19).
35
Simulation of the Electronic Unit
1000
900
800
700
Frequency (Hz)
600
500
400
300
200
100
0
0 1 2 3 4 5 6 7 8 9
Rib thickness (mm)
Figure 19. Board edge guides with ribs, frequency – rib thickness.
Figure 19 shows linearity up to 745Hz where a dip occurs. The dip can be
explained by the change of mode shape at this frequency. Since the relation
between the frequency and the rib thickness seems linear at the aimed frequency,
the linearity assumption can be used to calculate the thickness of the ribs.
dt ( f a f 0 ) (8 0)(618 162)
t rib 4.86 5mm (eq. 3.5)
df (912 162)
36
Simulation of the Electronic Unit
When trying to implement the board edge guide design in the electronic unit it
was found that it was not possible to use the idea. The board edge guides would
collide with the connectors if the controller module would be placed in its current
position. The position of the controller module could not be altered since there
were components connecting the controller module with the high power module.
The high power module could not be repositioned either, since it was connected
to the heat sinks. The repositioning of the controller module would therefore
affect the heat sinks, which would have to be redesigned. The round capacitors,
which were connected to the high power module, would also collide with the
bottom plate and would have to be replaced with similar smaller components. The
above mentioned collisions are illustrated in figure 20 below. The board edge
guides where discarded as a solution to the enforcement problem of the electronic
unit since it could not be fitted in the current design.
Figure 20. Collision of parts when trying to reposition the controller module.
37
Simulation of the Electronic Unit
38
Simulation of the Electronic Unit
The first natural frequency for different thicknesses of the grids was simulated and
the result of the simulations is shown in figure 24. As can be seen from the
diagram the relation between the thickness and the frequency is linear for the
second and the third design. The nonlinearity of the first design comes from a
change in the mode shape, where two of the edges start oscillating instead of the
midsection.
1800
1600
1400
1200
Frequency (Hz)
1000 Design 1
Design 2
800 Design 3
600
400
200
0
0 50 100 150 200
Added weight (g)
Since all designs are linear at the aimed frequency of 618 Hz, the linearity
assumption can be used for the thickness calculations. Using the linearity, the
thickness (t) of the aimed frequency ( f a ) can be calculated as follows.
dt ( f a f 0 ) (5 0)(618 399)
t1 1.55 2mm (eq. 3.6)
df (1107 399)
dt ( f a f 0 ) (5 0)(618 399)
t2 1.07 2mm (eq. 3.7)
df (1426 399)
dt ( f a f 0 ) (5 0)(618 399)
t3 0.88 1mm (eq. 3.8)
df (1647 399)
The weight was then calculated using the rib designs and the thicknesses from
calculations 4.6-4.8. The weight of the first design was 27g, the second design had
a weight of 54g and the third designs weight was 32g.
39
Simulation of the Electronic Unit
40
Simulation of the Electronic Unit
The first natural frequency was found for different thicknesses of the grid. The
relation between the thickness and the first natural frequency is linear to 1187Hz
where the mode shape changes, as can be seen from the diagram (fig. 26).
1400
1200
1000
Freqeuncy (Hz)
800
600
400
200
0
0 1 2 3 4 5 6
Rib thickness (mm)
The thickness of the ribs ( t rib ) was than calculated using the linearity assumption:
dt ( f a f 0 ) (3.5 0)(618 399)
t rib 1.11 2mm (eq. 3.9)
df (1087 399)
The incorporated component rib design was inserted in the current electronic unit
and a simulation was run. The simulation ran till 416 Hz and showed one
excitation of the controller module at 310Hz when the chassis and the high power
module started to oscillate. The reason for the oscillation could have to come
from the fact that the controller module has a higher natural frequency than the
chassis that is attached to the controller module through the spacers. The first
natural frequency of the controller module was therefore used to calculate the new
minimum thickness of the ribs.
dt ( f a f 0 ) (3.5 0)(736 399)
t rib 1.71 2mm (eq. 3.10)
df (1087 399)
The calculations gave the same minimum thickness of the ribs that had already
been proven to fail. The aimed-for frequency used for the calculations in 3.9
belonged to the envelope that transmitted the vibrations to the bottom plate. The
author decided to use the first natural frequency of the bottom plate instead,
which was 549Hz.
41
Simulation of the Electronic Unit
The weight of the grid was calculated from the thickness of the ribs and found to
be 68g. With the removal of the support plate and material at the spacers, the
additional weight of the enforcement solution was calculated to 19g.
42
Findings and Analysis
43
Findings and Analysis
44
Discussion and Conclusions
45
Discussion and Conclusions
The mesh used for the simulation most probably added numerical stiffness to the
simulation. If the mesh from the mesh refinement study could have been used,
this would have given a more accurate result. This limitation had to be made to
ensure that the simulation was able to run.
The constraints used when bounding the components in the simulation showed
some limitation, since only fixed and no separation boundary constraints were
available for modal simulation. In reality, friction and contact pressure also affect
the connections. At low acceleration levels the bounded constraint can represent
the friction and contact pressure, but at high acceleration levels on the
components they would separate in reality, which the simulation constraints do
not allow. For high acceleration levels the result of the simulation might therefore
not be as reliable.
The author did not use springs in the attachment of the screws in the simulation
since she did not find the input values for the composite components to be fully
reliable. Trying different stiffness of the springs could have yielded a result much
closer to the tests made by SAAB. However, the reliability of the spring stiffness
would have been poor, since this would merely cover up the error from the
composite components.
Despite all uncertainties in the simulations, they showed that couplings exist,
which was supported by the physical tests made by SAAB, making the result
reliable. The reliability of the enforcement solution and the simulation of the
casted electronic unit were ensured by the usage of the same method as for the
current electronic unit, which had been proven to be reliable. The validity of the
result was therefore considered to have been meet by the author.
46
Discussion and Conclusions
5.3 Conclusions
One must always keep in mind when working with simulation software that the
numerical results can’t fully represent reality. Simplifications need to be used to
ensure that the simulations will run smoothly and the input data needs to be
thoroughly investigated, since wrong input data will lead to inaccurate results.
The reliability of the work conducted was ensured by comparison with physical
test values and based by theory regarding the vibration phenomena.
The author recommends that more vibration tests should be preformed to find
the transmissibility of the components and their natural frequencies when
separated from the chassis. Transmissibilities in the complete electronic units
should also be found. The data can then be used to calculate the accretion forces
acting on the components and displacements and stresses can thus be found.
Doing so for the current electronic unit and the enforced electronic unit can verify
the benefit of the enforcement design.
More work should also be put in to the casted casing before releasing its final
design to ensure that it will not cause any damage to its components, due to its
development of high acceleration forces.
47
References
6 References
[1] SAAB Group, “Electronic Defense Systems,” SAAB Group, 2010. [Online].
Available: http://www.saabgroup.com/en/About-Saab/Company-
profile/Organisation/Electronic-Defence-Systems/. [Accessed: August 28,
2012].
[6] Y. Wu, Sensitivity-based Finite Element Model Updating Methods with Application to
Electronic Equipmen. PhD [Dissertation]. Mons: Faculté Polytechnique de
Mons, 1999. [Online]. Available:
http://mecara.fpms.ac.be/Theses/YunXinWu_phd.pdf. [Accessed:
September 27, 2012].
[7] J. Montgomery, Methods for modeling bolts in the bolted joint, Orlando, FL:
Siemens Westinghouse Power Corporation, 2002. [Online]. Available:
http://ainastran.org/staticassets/ANSYS/staticassets/resourcelibrary/conf
paper/2002-Int-ANSYS-Conf-38.PDF. [Accessed: November 8, 2012].
[8] J. Kim, J.-C. Yoon, B.-S. Kang, “Finite element analysis and modeling of
structure with bolted joints,” Applied Mathematical Modelling, vol. 2007,
no.31, pp. 895-911, March 2006, [Online] Available: SienceDirect,
http://www.fml.t.u-
tokyo.ac.jp/~izumi/Bolt/WG/fem_bolted_joints_modelling_Korea.pdf.
[Accessed: November 8, 2012].
48
References
[10] SAS IP Inc, ANSYS 13.0 Help, Canonsburg: SAS IP Inc, 2010.
49
Search Terms
7 Search Terms
acceleration ..................................... 17
amplitude ........................................ 13
aspect ratio ...................................... 24
board edge guides ...................... 18, 35
bolted joints ..................................... 19
bounded........................................... 23
chassis ............................................. 18
controller module ............................ 10
coupling .................................... 31, 32
damping .......................................... 19
density............................................. III
displacement ................................... 17
element types ................................... 23
fastener............................................ 18
flaps ................................................ 10
high power module .......................... 10
horizontal PCB ................................ 10
low natural frequencies .................... 19
modes of vibration ........................... 16
no separation ................................... 23
octave rule ....................................... 18
periodic vibration ............................ 13
physical test ............................... 29, 31
random vibration ............................. 13
resonant frequency .......................... 15
ribs .......................................19, 37, 40
SAAB................................................ 9
SAAB EDS ....................................... 9
slats ................................................. 10
smearing technique .......................... 21
tensile test ....................................... 26
transmissibility ................................ 16
weight ratio ............................... 18, 33
vertical PCB .................................... 10
vibration .......................................... 13
Young’s modulus ...................... 29, 30
50
Appendices
8 Appendices
APPENDIX A – PARTS INCLUDED IN THE SIMULATION ............................................... I
APPENDIX B – YOUNG’S MODULUS CALCULATIONS OF PCBS ................................. II
APPENDIX C – DENSITY CALCULATIONS ................................................................III
APPENDIX D – CONVERGENCE STUDY ................................................................... IV
APPENDIX E – CONVERGENCE CALCULATIONS ................................................. XXIV
APPENDIX F – ELEMENT SIZE ............................................................................ XXV
APPENDIX G – SIMULATION OF THE OLD ELECTRONIC UNIT ........................... XXVII
APPENDIX H – SIMULATION OF THE CURRENT ELECTRONIC UNIT .................. XXVIII
APPENDIX I – WEIGHT OF COMPONENTS .......................................................... XXIX
51
Appendix A – Parts Included in the Simulation
I
Appendix B – Young’s Modulus Calculations of PCBs
II
Appendix C – Density Calculations
III
Appendix D – Convergence Study
Bolt in filter
37500.00
36500.00
35500.00
34500.00
Frequency (Hz)
33500.00
32500.00
31500.00
30500.00
29500.00
28500.00
0.00 200.00 400.00 600.00 800.00 1000.00 1200.00 1400.00
No. elements
Bottom plate
4400.00
4300.00
4200.00
4100.00
Frequency (Hz)
4000.00
3900.00
3800.00
3700.00
3600.00
3500.00
0.00 10000.0 20000.0 30000.0 40000.0 50000.0 60000.0 70000.0 80000.0 90000.0 100000.
0 0 0 0 0 0 0 0 0 00
No. elements
IV
Appendix D – Convergence Study
Envelope
5000.00
4500.00
4000.00
3500.00
Frequency (Hz)
3000.00
2500.00
2000.00
1500.00
1000.00
500.00
0.00
0.00 5000.00 10000.00 15000.00 20000.00 25000.00 30000.00 35000.00 40000.00
No. elements
Fixing plate
45000.00
44500.00
44000.00
Frequency (Hz)
43500.00
43000.00
42500.00
42000.00
41500.00
41000.00
0.00 200.00 400.00 600.00 800.00 1000.00 1200.00 1400.00 1600.00
No. elements
V
Appendix D – Convergence Study
Front wall
40000.00
35000.00
30000.00
25000.00
No. elements
20000.00
15000.00
10000.00
5000.00
0.00
0.00 5000.00 10000.00 15000.00 20000.00 25000.00
Frequency (Hz)
46000.00
45000.00
44000.00
Frequency (Hz)
43000.00
42000.00
41000.00
40000.00
39000.00
38000.00
37000.00
0.00 5000.00 10000.0 15000.0 20000.0 25000.0 30000.0 35000.0 40000.0 45000.0 50000.0
0 0 0 0 0 0 0 0 0
No. elements
VI
Appendix D – Convergence Study
336.00
334.00
332.00
Frequency (Hz)
330.00
328.00
326.00
324.00
322.00
320.00
318.00
0.00 200.00 400.00 600.00 800.00 1000.00 1200.00 1400.00
No. elements
Casted casing
1500.00
1450.00
1400.00
Frequency (Hz)
1350.00
1300.00
1250.00
1200.00
1150.00
1100.00
25000.00 35000.00 45000.00 55000.00 65000.00 75000.00 85000.00
No. elements
VII
Appendix D – Convergence Study
37400.00
37200.00
37000.00
Frequency (Hz)
36800.00
36600.00
36400.00
36200.00
36000.00
35800.00
0.00 1000.00 2000.00 3000.00 4000.00 5000.00 6000.00 7000.00 8000.00
No. elements
Controller module
370.00
360.00
350.00
Frequency (Hz)
340.00
330.00
320.00
310.00
0.00 5000.00 10000.00 15000.00 20000.00 25000.00 30000.00
No. elemnents
VIII
Appendix D – Convergence Study
Heat sink
34000.00
33000.00
32000.00
31000.00
Frequency (Hz)
30000.00
29000.00
28000.00
27000.00
26000.00
25000.00
0.00 2000.00 4000.00 6000.00 8000.00 10000.0 12000.0 14000.0 16000.0 18000.0 20000.0
0 0 0 0 0 0
No. elements
620.00
618.00
616.00
Frequency (Hz)
614.00
612.00
610.00
608.00
606.00
604.00
0.00 1000.00 2000.00 3000.00 4000.00 5000.00 6000.00 7000.00 8000.00 9000.00
No. elements
IX
Appendix D – Convergence Study
Horizontal PCB
1380.00
1360.00
1340.00
Frequency (Hz)
1320.00
1300.00
1280.00
1260.00
1240.00
1220.00
0.00 2000.00 4000.00 6000.00 8000.00 10000.00 12000.00
No. elements
410.00
400.00
Frequency (Hz)
390.00
380.00
370.00
360.00
350.00
0.00 5000.00 10000.00 15000.00 20000.00 25000.00
No. elements
X
Appendix D – Convergence Study
370.00
360.00
Frequency (Hz)
350.00
340.00
330.00
320.00
310.00
0.00 5000.00 10000.00 15000.00 20000.00 25000.00
No. elements
420.00
410.00
Frequency (Hz)
400.00
390.00
380.00
370.00
360.00
350.00
0.00 5000.00 10000.00 15000.00 20000.00 25000.00
No. elements
XI
Appendix D – Convergence Study
Large screw
180000.00
170000.00
160000.00
Frequency (Hz)
150000.00
140000.00
130000.00
120000.00
0.00 20000.00 40000.00 60000.00 80000.00 100000.00 120000.00 140000.00
No. elements
Medium connector
390.00
380.00
370.00
Frequency (Hz)
360.00
350.00
340.00
330.00
0.00 2000.00 4000.00 6000.00 8000.00 10000.00 12000.00 14000.00 16000.00 18000.00
No. elements
XII
Appendix D – Convergence Study
Rear side
12000.00
10000.00
8000.00
No. elements
6000.00
4000.00
2000.00
0.00
0.00 5000.00 10000.00 15000.00 20000.00 25000.00
Frequency (Hz)
Round capacitor
20900.00
20800.00
20700.00
Frequency (Hz)
20600.00
20500.00
20400.00
20300.00
20200.00
0.00 50.00 100.00 150.00 200.00 250.00 300.00 350.00 400.00 450.00
No. elements
XIII
Appendix D – Convergence Study
48800.00
48700.00
48600.00
Frequncy (Hz)
48500.00
48400.00
48300.00
48200.00
48100.00
48000.00
47900.00
0.00 500.00 1000.00 1500.00 2000.00 2500.00 3000.00 3500.00 4000.00
No. elements
26600.00
26400.00
26200.00
Frequency (Hz)
26000.00
25800.00
25600.00
25400.00
25200.00
25000.00
24800.00
24600.00
0.00 5000.00 10000.00 15000.00 20000.00 25000.00 30000.00
No. elements
XIV
Appendix D – Convergence Study
Section wall
14000.00
12000.00
10000.00
Frequency (Hz)
8000.00
6000.00
4000.00
2000.00
0.00
0.00 2000.00 4000.00 6000.00 8000.00 10000.00 12000.00 14000.00 16000.00 18000.00
No. elements
3800.00
3700.00
3600.00
Frequency (Hz)
3500.00
3400.00
3300.00
3200.00
3100.00
3000.00
0.00 500.00 1000.00 1500.00 2000.00 2500.00 3000.00 3500.00 4000.00 4500.00 5000.00
No. elements
XV
Appendix D – Convergence Study
Small connector
290.00
285.00
280.00
Frequency (Hz)
275.00
270.00
265.00
260.00
255.00
250.00
0.00 2000.00 4000.00 6000.00 8000.00 10000.00 12000.00
No. elements
Small screw
160000.00
150000.00
140000.00
Frequency (Hz)
130000.00
120000.00
110000.00
100000.00
0.00 2000.00 4000.00 6000.00 8000.00 10000.00 12000.00 14000.00
No. elements
XVI
Appendix D – Convergence Study
60000.00
59000.00
58000.00
Farquency (Hz)
57000.00
56000.00
55000.00
54000.00
53000.00
52000.00
51000.00
50000.00
0.00 1000.00 2000.00 3000.00 4000.00 5000.00 6000.00
No. elements
25600.00
25400.00
Frequency (Hz)
25200.00
25000.00
24800.00
24600.00
24400.00
0.00 1000.00 2000.00 3000.00 4000.00 5000.00 6000.00 7000.00 8000.00 9000.00 10000.0
0
No. elements
XVII
Appendix D – Convergence Study
25600.00
25400.00
Frequency (Hz)
25200.00
25000.00
24800.00
24600.00
24400.00
0.00 1000.00 2000.00 3000.00 4000.00 5000.00 6000.00
No. elements
1.25E+05
1.20E+05
Frequency (Hz)
1.15E+05
1.10E+05
1.05E+05
1.00E+05
0.00 5000.00 10000.00 15000.00 20000.00 25000.00
No. elements
XVIII
Appendix D – Convergence Study
Square capacitor
1300.00
1290.00
1280.00
1270.00
Frequency (Hz)
1260.00
1250.00
1240.00
1230.00
1220.00
1210.00
1200.00
0.00 1000.00 2000.00 3000.00 4000.00 5000.00 6000.00
No. elements
44300.00
Frequency (Hz)
43800.00
43300.00
42800.00
42300.00
0.00 500.00 1000.00 1500.00 2000.00 2500.00 3000.00
No. elements
XIX
Appendix D – Convergence Study
1060.00
1055.00
Frequency (Hz)
1050.00
1045.00
1040.00
1035.00
1030.00
1025.00
0.00 500.00 1000.00 1500.00 2000.00 2500.00 3000.00 3500.00 4000.00 4500.00
No. elements
1770.00
1760.00
Frequency (Hz)
1750.00
1740.00
1730.00
1720.00
1710.00
0.00 1000.00 2000.00 3000.00 4000.00 5000.00 6000.00 7000.00
No. elements
XX
Appendix D – Convergence Study
3960.00
3940.00
Frequency (Hz)
3920.00
3900.00
3880.00
3860.00
3840.00
3820.00
0.00 5000.00 10000.00 15000.00 20000.00 25000.00 30000.00 35000.00
No. elements
318.00
317.00
Frequency (Hz)
316.00
315.00
314.00
313.00
312.00
311.00
0.00 2000.00 4000.00 6000.00 8000.00 10000.00 12000.00
No. elements
XXI
Appendix D – Convergence Study
Transformer
9000.00
8000.00
7000.00
6000.00
Frequency (Hz)
5000.00
4000.00
3000.00
2000.00
1000.00
0.00
0.00 10000.00 20000.00 30000.00 40000.00 50000.00 60000.00 70000.00
No. elements
T-support
22000.00
21800.00
21600.00
21400.00
Frequency (Hz)
21200.00
21000.00
20800.00
20600.00
20400.00
20200.00
20000.00
0.00 500.00 1000.00 1500.00 2000.00 2500.00 3000.00 3500.00
No. elements
XXII
Appendix D – Convergence Study
Vertical PCB
8000.00
7000.00
6000.00
Frequency (Hz)
5000.00
4000.00
3000.00
2000.00
1000.00
0.00
0.00 5000.00 10000.00 15000.00 20000.00 25000.00
No. elements
XXIII
Appendix E – Convergence Calculations
XXIV
Appendix F – Element Size
XXV
Appendix F – Element Size
The error values marked green are within the convergence limit of 5% and the
error values marked orange are outside of the same limit.
XXVI
Appendix G – Simulation of the Old Electronic Unit
XXVII
Appendix H – Simulation of the Current Electronic Unit
XXVIII
Appendix I – Weight of Components
XXIX