272 Spring 2014 PDF

MIT OpenCourseWare
http://ocw.mit.edu
2.72 Elements of Mechanical Design

Spring 2009
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
Schedule
Overview
Syllabus
Project
Groups
Lab
Project
Questions
Assessment
Reading assignment for next class:
Shigley Mischke on shaft deflections & stress

Sections 3.1, 3.2, 3.4, 3.7, 3.8, 3.10 3.11
Refresh on beam bending
Refresh on 1st and 2nd order system vibrations
Martin Culpepper, All rights reserved
Purpose and pace of course

In-depth treatment of principles and practices required
to synthesize, model, design, fabricate & characterize.
APPLIED ENGINEERING
Teaching emphasis, style and grades reflect this

You will be expected to practice what you see in lecture
Reading ~ 50% of grade
2/3 semester of lectures

Info sources: Teaching staff & texts

Prof. Martin Culpepper
Required text:
Useful text:
TA: Jon Hopkins
Mechanical Engineering Design (Shigley / Mischke)

Design of Machinery (Norton)
Machinerys Handbook
Teaching philosophy
The man who sets out to carry a cat by its tail learns something that will always be useful and which never
will grow dim or doubtful. --Mark Twain
Thoughts, decisions & actions based on understanding:
Models and their associated equations are idealizations

Only perfect model is a physical embodiment
The limits/power of modeling and simulation
Mechanical system design is cost and time intensive.
Mastery of:
Concepts, principles & design processes necessary, but not sufficient

Math, physics and engineering models are necessary, but not sufficient
Practical skills and best practices are necessary, but not sufficient
The judicious use of (a), (b) and (c) is necessary
Design project
In theory there is no difference between theory and practice. In practice there is. --Yogi Berra
2.72 will focus on

(i)
(ii)
understanding concepts, principles, design process, best practices,

mathematics, physics and engineering modeling; and
rigorous application of the same to realize a complex and high quality
mechanical design.
You will learn by

(i)
(ii)
Doing
Gaining insight via interaction with staff
Project:
(i)
(ii)
(iii)
Teams of 6 work to model, design, build and characterize one lathe

You can all build copy in parallel, group must do at least one
Meeting functional requirements is critical to passing
Project: Desktop lathe
Documentation
Images and video
Take pictures/video as you go

Due in soft copy on the day their corresponding hardware/results are due
TA has camera if needed
You must keep a dedicated design notebook
Keep your ideas, calculations, and records in one organized place

Bring your notebook to all 2.72 events
Notebooks will be collected periodically, used to generate final grades
Legible and organized!
Staple or glue in loose papers, no 3-ring binders will be accepted
DO NOT take class notes in this notebook
Final report
No more than 6 pages (not including appendices)
Purpose = convince the staff that you learned & used the course material
Mechanical laboratories
Disassemble mechanical devices/assemblies
Take measurements, answer questions and reassemble

Tools will be provided
Bring your own safety glasses (we will give 1st pair)
Follow shop safety rules

Lab times
- Groups 1, 2 & 3 from 09.00pm 12.00pm
- Groups 4, 5 & 6 from 02.00am 05.00pm
Topic
1. Lathe disassembly
2. Bearing alignment
3. Transmission
4. IC Engine
Design laboratories
45 minute meetings
15 minute presentations
30 minute discussion/design review
50 inch plasma will be available
Everyone must present their part of the project

As a group:
First tell us the purpose of the meeting
Then immediately discuss Gantt chart
Details of the work to date, calculations
Have back up slides for deep dives
Parts for the lathe

Spindle
Structure
Lead screw drive
Carriage
Cross feed
Miscellaneous
Our responsibility
Your responsibility
Housing blank
Shaft blank
Headstock blank
Tail stock blank
Structure tube blank
Bearings
Preload end cap
Lead screw
Preload washers
Polymer bed blank*
Handles
Housing
End cap
Shaft Bearings
Preload mechanisms
Rails
Finished tail stock
Finished tube Finished head stock
Preload tube
Drive nut
Lead screw bearing seat
Bearing preload nut
End skirt blanks*
*May cast 3 pieces as one, stay tuned

Tool holder
Lead screw
Drive preload nut
Finished polymer bed* Finished end skirts*

Drive flexure coupling Bushing flexure coupling
Bushings
Flexure bearing
Rear flexure mount
Front flexure mount
Thrust bearing
Proper dial mount surfaces/flats on screw
Chuck
Metrology fixtures (3-ball & runout)
HSS cutting tool
Fasteners
- 20 bolts 0.50 inch long
10
Use of ME Mfg. shop

"Good plans shape good decisions. That's why good planning helps to make elusive dreams come true." --Lester Bittel
Open M-F, 8am-4pm, Clean up at 3:30pm

Use of the machine shop must be scheduled
Lab time
Monday, Tuesday and Thursday between 8am-12pm
Wednesdays and Fridays between 1pm and 4pm
10 min. late for an appointment, your appointment will be cancelled.
Process plans
2D printed, CAD drawing with dimensions and tolerances (NO sketches)

3D printed rendering of the part (e.g. screen capture from CAD)
Properly scaled DXF (see handout) on disc/e-mail to shop manager
Completed process plan table
Shop manager must sign off and then you turn into Culpepper
11
Grading
Grading:
50 %
50%
Project
In-class and take-home mini-quizzes
All materials are due by 5pm
Via e-mail to teaching assisstant, unless otherwise stated.
Email errors will not excuse late assignments.
You must understand otherwise you are dangerous
No students group may proceed w/o grade > 80% on qualifiers

Make up quizzes may be given, but course schedules wont change
Quizzes
Take-home quizzes
Reading quizzes
In-class exercises
12
Rules for collaboration

You should work together & learn from one another
What you submit MUST be your own work unless it is specified as

a group submission. In the case of group submissions, everyone
that worked on the submission must sign a cover page and
provide bullet point summaries of what you worked on and how
much of that part you did
You MUST acknowledge the contribution of others
For example, after working an assignment independently, you compare

responses with another student which alerts you to an error in your own
work which you then correct. You should state at the end of your
submission that you corrected your error on the basis of checking
responses with the other student. No credit will be lost if the response is
correct, the acknowledgment is made, and no direct copying of the other
response is involved.
13
Course Web Site

http://pcsl.mit.edu/2_72/index.html
Reading assignments
Quiz materials
Lecture notes
Software downloads
Homework downloads
14
Form groups & assign gurus

Group 1
Group 4
Group 2
Group 5
Group 3
Group 6
15
Assessments
Assessment A
On web
15 minutes
Assessment B
In class
45 minutes
Cant hurt you, only help extra credit

16
MIT OpenCourseWare
http://ocw.mit.edu

Spring 2009
2.72
Elements of
Mechanical Design
Lecture 02: Review
Intent
High-level review of undergrad
material as applied to
engineering decision making
NOT an ME redo or a how
to recitation
Import
Main goal of 2.72 is to teach
you how to integrate past
knowledge to engineer a system
Given this, how do I engineer a mechanical system?
modularsimplecomplexsystem
2.001, 2.002 2.003, 2.004 2.005, 2.006 2.007, 2.008
Use of core ME principles that you know
2.001, 2.002
2.003, 2.004
2.005, 2.006
2.007, 2.008
Impact
Understand why & how we will
use parts of ME core knowledge
Problem set Engineering
Future help
We cant use lecture time to
redo the early curriculum
BUT
We are HAPPY to help outside
of lecture IF youve tried
Schedule and reading assignment

Reading quiz
Changing from sponge to active mode
Lecture
Mechanics
Dynamics
Heat transfer
Matrix math
Hands-on
Mechanics
Dynamics
Heat transfer
Matrix math
Reading assignment
Shigley/Mischke
Sections 4.14.5: 08ish pages & Sections 5.15.5: 11ish pages
Pay special attention to examples 4.1, 4.4, 5.3 , and 5.4
Mechanics
Free body diagrams

Useful for:
Equilibrium
Stress, deflection, vibration, etc
v
v
F = 0 = ma
v
M = 0 = I
Why do we ALWAYS use free body diagrams?
Communication
Thought process
Documentation
How will we use free body diagrams?
We are dealing with complex systems

We will break problem into modules
We will model, simulate and analyze mechanical behavior
Integrate individual contributions to ascertain system behavior
Free body diagrams: Bearings/rails
11
Free body diagrams: Bearings/rails
12
Static: Head stock deformation
13
Static: Rail deformation
14
Example 1: 0 < x < a

y
Cantilever
Shear & normal

Static failure
Fatigue failure
Why do we care?: Stiffness
Forces, moments, & torques

y
Why do we care?: Stress
Displacement
Rotation
Vibration (k/m)
But, ends arent all that matters
M(0)
V(0)
v
F = 0 = F + V (0 ) V (0 ) = V ( x ) = F
v
M = 0 = + F l + M (0 ) M (0 ) = F a
b
F
M(x)
V(x)
v
F = 0 = F + V (0 ) V (0 ) = V ( x ) = F
v
M = 0 = + F (a x ) + M ( x ) M ( x ) = F (a x )
15

But, ends arent all that matters
b
F
M(x)
Relating V(x) & M(x)

V (x ) = F
M (x ) = F (x a )
V (x ) =
d
M (x )
dx
Shear moment diagrams
V(x)
v
F = 0 = F + V (0 ) V (0 ) = V ( x ) = F
v
M = 0 = + F (a x ) + M ( x ) M ( x ) = V ( x ) (a x )
V ( x ) = F
Solve statics equation

Put point of import on plots
Use V=dM/dx to generate M plot
Master before spindle materials
-Fa
M ( x ) = V ( x ) (a x ) = F (a x )
16

Stress
M(x)
M(x+dx)
M(x)
max
h
2
V(x)
v
F = 0 = F + V (0 ) V (0 ) = V ( x ) = F
y
h
v
M = 0 = +V ( x ) (a x ) M ( x ) M ( x ) = V ( x ) (a x )
b
I (x ) =
(y) = M
1
b( x ) [h ( x )]3
12
-Fa
y
c
max = M
I (x )
I (x )
max = F (a x )
max = 6
h(x )
12
2 b( x ) [h ( x )]3
F (a x )
b( x ) [h ( x )]2
x
M ( x ) = V ( x ) ( x a ) = F (a x )
=6
6Fl
bh2
F (a x )
bh 2
-6Fl
bh2
= 6
F (a x )
bh 2
17
Group work: Generate strategy for this

x
x
z
Vx
Vx
Vy
Vy
18
Dynamics
Vibration
Vibration principles
Exchange potential-kinetic energy

2nd order system model
c
m
Blocks and squiggles
What do they really mean?

Why are they important?
How will we apply this?
Multi-degree-of-freedom system
Mode shape
Resonant frequency
F(t)
k
k
m
n =
Gain
Estimate n (watch units) for:
A car suspension system
n
20
Images removed due to copyright restrictions. Please see:
http://www.hpiracing.com/graphics/kits/547/_MG_1962e.jpg
http://www.societyofrobots.com/images/mechanics_suspension_honda.gif
http://www.bose.com/images/learning/lc_susp_frontmodule.jpg
21
Vibration: MEMS device behavior

Without input shaping
With input shaping
22
Vibrations: Meso-scale device behavior
Underdamped response
Noise
Free vibration
Golda, D. S., Design of High-Speed, Meso-Scale Nanopositioners

Driven by Electromagnetic Actuators, Ph.D. Thesis, Massachusetts
Institute of Technology, 2008.
23
Vibrations: Reducing amplitude

How to change
m, k, and c?
24
So where find in reality in lathe

x
c
m
F(t)
25
Vibration: Lathe structure 1st mode
26
Vibration: Lathe structure 2nd mode
27
Heat transfer
Thermal growth errors

For uniform temperature
L = Lo T
STEEL :12L14 L = 11.5 10
m
Lo T
o
m C
m
Lo T
o
m C
m
Lo T
o
m C
ALUMINUM :6061 T 6 L = 23.6 10
POLYMER : Delrin L = 100 10

29
Convection and conduction

Convection:
q& = h Asurface (T T )
Why do we care?
Heat removal from cutting zone

Heat generation in bearings
Conduction:
q& = k Across
dT
dx
Why do we care?
Heat removal from cutting zone

Heat generation in bearings
Common k values to remember
Air
0.026 W /(moC)
12L14
51.9 W /(moC)
6061 T6
167 W /(moC)
30
Thermal resistance
Thermal resistance
T
q& =
RT
Convection
q& =
(T T )
(h Asurface )
a RT =
1
h Asurface
Conduction
k Across
dx
q& = dT
a RT =
dx
k Across
31
Biot (Bi) number

Ratio of convective to conductive heat transfer
(h AT ) convection
q&convection
hLc
a
a Bi =
q&conduction
k
T
k A
L
conduction
Why do we care?
Low Bi
High Bi
32
Example of thermal errors

For:
For:
h=
Bearing T =
Chip T =
0.1 W/(m2oC)
150 oF
180 oF
h=
Bearing T =
Chip T =
50 W/(m2oC)
150 oF
180 oF
33
Example of thermal errors

For:
For:
h=
Bearing T =
Chip T =
0.1 W/(m2oC)
150 oF
180 oF
h=
Bearing T =
Chip T =
5000 W/(m2oC)
150 oF
180 oF
34
Types of errors
Machine system perspective
System-level approach
Linking inputs and outputs
Measurement quality
Power
Material
Temperature
changes
C1 C2 C3
[Outputs]
=
C4 C5 C6
[Inputs]
C7 C8 C9
Torques
Desired outputs
- Motion, location, rotation
- Cutting forces
- Speeds
Measured outputs
- Cutting forces
- Speeds
Machine
Perceived
Error
Real
error
Actual outputs
- Cutting forces
- Speeds
Forces
Vibration
Fabrication
errors
Speeds
36
Errors.
Accuracy
The ability to
tell the truth
Repeatability
Ability to do the
same thing over &
over
Determinism
Machines obey physics!

Model understand relationships
C1 C2 C3
[Outputs] = C4 C5 C6 [Inputs]
C
7 C8 C9
Understand sensitivity
[Outputs] = J [Inputs ]
Range
Both
1st make repeatable,
then make accurate
Calibrate
Furthest extents of motion
Resolution
Smallest, reliable position change

37
Categorizing error types

Systematic errors
Inherent to the system, repeatable and may be calibrated out.
Non-systematic errors
Errors that are perceived and/or modeled to have a statistical nature

Machines are not random, there is no such thing as a random error
Consider the error for each set below
Link behavior with systematic and non-systematic errors.
A
38
Exercise
Exercise
Due Tuesday, start of class:
Lathe components
Rough sketch(es) of lathe

Annotate main components
1 page bullet point summary of where need to use:
2.001, 2.002
2.003, 2.004
2.005, 2.006
2.007, 2.008
Rules:
You may not re-use examples from lecture!
You are encouraged to ask any question!
You may work in groups, but must submit your own work
40
Group work: Generate strategy for this

x
x
z
Vx
Vx
Vy
Vy
41
MIT OpenCourseWare
http://ocw.mit.edu

Spring 2009
2.72 shaft blank exercise
The aims of this exercise are to measure and understand the (a) accuracy/repeatability of a lathe, (b) part deflections during turning, (c)
the difficulty associated with meeting tight tolerances, (d) to fabricate the blank for your shaft and (e) to measure your shafts geometry.
Names:
Group:
Total:
0
1.00
5.00
5.87
6.50
0.89
+ 0.000
0.500 0.050
1.05
+ 0.000
0.020
+ 0.020
R 0.30 0.020
Tolerances:
X.xx 0.0300
X.xxx 0.0050
X.xxxx 0.0005
0.591
+ 0.000
0.002
+ 0.020
R 0.30 0.020
All dimensions in inches; drawing not to scale
Material: 12L14 steel
Surface finish:
16 inch
Step 1: Calculate the lateral bending stiffness of the shaft when its full length is cantilevered.
Klateral:
N/m
Step 2: List 3 errors that could affect shaft dimensions during turning, identify them as systematic / non
systematic.
Error
Thermal
Systematic
Nonsystematic
X
01
02
03
Step 3: Meet with shop manager to discuss how to make your shaft
Step 4: Make a process plan (see work sheet in the Appendix of syllabus) for the shaft, obtain shop managers
approval on the plan and schedule a time with him to machine your groups shaft.
Step 5: Before you finalize the shafts geometry, you must run experiments wherein you will cut features into the
stock (make sure they wont interfere with the final geometry of the part) and measuring them. First we will
measure the diameter of 10 closely spaced plunge cuts (go 0.05 deep using the dials only) near the shafts
cantilevered end and report the mean and standard deviation. Explain how these numbers relate to the accuracy
and repeatability of the lathe.
1.
mm
2.
mm
3.
mm
4.
mm
5.
mm
6.
mm
7.
mm
8.
mm
9.
mm
10.
mm
Comments:
Step: 6: Make four plunge cuts, that are 0.05 deep, near the shafts free end such that they are on a pitch of .
Comment on the relationship between the shafts stiffness/deflection and the difference in measured diameters.
1.
2.
3.
4.
Comments:
Step 7: Make the shaft according to the print on page 1.
Step 8: Measure the final dimensions of the shaft as best as possible given the gages that you have access to.
Provide numbers that demonstrate the possible error in your measurements.
MIT OpenCourseWare
http://ocw.mit.edu

Spring 2009
2.72
Elements of
Mechanical Design
Lecture 03: Shafts

Reading quiz
Hand forward lathe exercise quiz
Topics
Finish matrices, errors

Shaft displacements
Stiffness exercise
Reading assignment
Shigley/Mischke
Sections 6.16.4: 10ish pages & Sections 6.76.12: 21ish pages
Pay special attention to example 6.12 (modified Goodman portion)
Deflection within
springs and shafts
Shafts, axles and rails

Shafts
Rotating, supported by bearings/bushings

Dynamic/fluctuating analysis
Axles
Non-rotating, supported by bearings/bushings

Static analysis
Rails
Non-rotating, supports bearings/bushings

Static analysis
Examples drawn from your lathe
In practice, we are concerned with

Deflection
Stiffness
Bearings and stiffness of connectivity points
Function of global shaft geometry, sometimes adjacent components
Stress
Catastrophic failure:
Ductile
Function of local shaft geometry
Brittle
Fatigue
What is of concern?
Deflection and stiffness
Beam bending models

Superposition
Load and stress analysis
Bending, shear & principle stresses

Endurance limit
Fatigue strength
Endurance modifiers
Stress concentration
Fluctuating stresses
Failure theories
Von Mises stress

Maximum shear stress
Materials
Steel vs. other materials
Aluminum
Brass
Cast iron
Important properties
Modulus
Yield stress
Is density important?
Fatigue life
CTE
Material treatment Hardening
What does hardening do the material properties

It is expensive
Affects final dimensions
You can usually design without this
Principles of stiffness: Relationships

q
q d4y
= 4
EI dx
V
d3y
= 3
EI dx
M d2y
= 2
EI dx
dy
dx
10
Modeling: General forms of equations

Lateral bending deflection (middle)
F L3
F L3
Const
=
48E I
EI
Axial deflection
RL
RR
F L
=
AE
Lateral bending angles (at ends)
F L2
F L2
ML
=
=
or
6 E I Const E I
Const E I
11
Modeling: Stiffness
Lateral bending stiffness at middle
kb
(
E I)
= 48
= Const (E I ) Ln
L3
Axial stiffness
RL
RR
AE
kA =
L
Torsional stiffness
Stepped shafts?
J G
k =
L
12
Modeling: Stiffness
These pop up in many places, memorize them
Square cross section
1 3
I = bh
12
Circular cross sections
I=
[
(d
64
J=
[
(d
32
4
4
)
(
)
d
outer
inner ]
outer
(d inner )
]
13
Principles of stiffness: Ratios

Everything deforms
Impractical to model the stiffness of everything

Mechanical devices modeled as high, medium & low stiffness elements
Stiffness ratios show what to model as high-, medium, or low stiffness
k1st
Rk =
k2nd
Stiffness ratio
l
kaxial
AE
=
l
klateral
3EI
= 3
l
AE
2
l
Rk = l = 4 2
3 EI
h
l3
Building intuition for stiffness
You cant memorize/calculate everything

Engineers must be reasonable instruments
Car suspension is easy, but flexed muscle vs. bone?
14
Principles of stiffness: Sensitivity

Cantilever
L
F
F L3
=
3 E I
1
I = b h 3
12
E b h 3

F =
4 L
1
Lm
hn
3
3
dF
d E b h
E b h
k=
=
4 L
d d 4 L
15
Superposition
You must be careful, following assumptions are needed
Cause and effect are linearly related

No coupling between loads, they are independent
Geometry of beam does not change too much during loading
Orientation of loads does not change too much during loading
Use your head, when M = 0, what is going on

Superposition is not plug and chug
You must visualize

You must think
16
Types of springs and behaviors
Springs and stiffness
kF = dF(x)/dx
Force-Displacement Curve
F(x)=F
Constant force spring

= dF(x)/dx = 0
kF
Eb-a = F (xb - xa )
F
a
Constant stiffness spring
kF
= Constant
Eb-a = 0.5 kF (xb2 - xa2 )
x
F(x)=kFx
kF
Non-linear force spring
kF
= function of x
Eb-a = F(x) dx
Images from Wikimedia Commons, http://commons.wikimedia.org
17
Non-conformal contact ball on flat

Non-conformal contacts often non-linear
Example: bearings, belleville washers, structural connections
Is anything ever perfectly conformal?

Specific case: Hertzian contact
F
80
k [N/micron]
In-plane stiffness
60
40
20
0
0
250
500
750
1000
Preload [N]
13 2 3 13
kn (F ) = Constant R E F
18
Linearization of non-linear springs

If you can linearize over the appropriate range then
you can use superposition
So how would, and when could, you do this?
R = ball radius
E = modulus of both materials (both steel)
F = contact load
dF
13 2 3 13
= kn (F ) = Constant R E F
d
19
Practical application to the lathe problem
F
Case 11 in Appendix A-9
20

y
F
x
y ( x ) AB
1
2
=
F x (11x 9l )
96EI
( )
y l
1
7 3
=
F l
2 96 EI
8
k Beam
768 E I
=
3
7
l
But, is this really what is going on?
21

y
F
x
k Beam
Vs. ?
x
k
768 E I
=
3
7
l
F
k
22
x
k
F
k
23

Or is it this?
If so, does it matter?
k
F
kbearing
k
24
1
2
F
1
F
2
k
k
25
Group work
Obtain an equation for total in terms of F, k and l
Estimate when k is important / should be considered?
What issue/scenario would cause k not to be infinite?
Look at these causes, if a stiffness is involved, would

linearity, and therefore superposition apply?
26
MIT OpenCourseWare
http://ocw.mit.edu

Spring 2009
2.72
Elements of
Mechanical Design
Lecture 04: Fatigue

Reading quiz
Announcements
Shaft due date
Shaft exercise
Goodman diagram quiz (Tuesday)
see unterhausen. "fatigue crack." March 27, 2008.

Shear-moment qualifying quiz (Tuesday)
Please
YouTube. Accessed October 28, 2009.
http://www.youtube.com/watch?v=iBuuVd0JlIM
Topics
Discuss stiffness exercises
Start fatigue
Reading
None, for Tuesday, prep for quizzes in lab time (Given lounge, top of 35)
Reading quiz
Discuss stiffness exercises

Answers
Intuition about stiffness
Carriage bearing-rail
Part
Spindle
Carriage-rail
Etc
Insight and perspective
Combined Stiffness
K/2
K/3
K/4
K/5
0
K/K
Shaft exercise
Are you on top of this!?
Fatigue part I
At what critical time in

engineering history did
fatigue became
relevant?
Why does fatigue

failure generate serious
concern?
What type of warnings

does one receive?
Fire plane wing failure

July 18, 2002 near Estes Park, Colorado
Both crew members killed
Delivered in July, 1945 to the U.S. Navy
Logged 8000+ flight hours
Investigation
NTSB found extensive fatigue
Cracks hidden from view
Please see richmondlopez13. "Plane crash caught on tape - 6."

10
December 5, 2007. YouTube.
Accessed October 14, 2009. http://www.youtube.com/watch?v=j7kr6o1s9sI
Comet airplane failures

BOAC Flight 781 crashes on 10 January 1954
Concluded fire was most likely cause

Resumed on 23 March 1954
Comet G-ALYY crashes on 8 April 1954
Pressure tests revealed fatigue

Windows to be glued & riveted, but riveted only
Square windows oval
Skin thickened
Service in 1958
11
I-35 bridge failure

By Laurie Blake, Paul Mcenroe, Pat Doyle and Tony Kennedy, Star Tribune
MnDOTs options:
Make repairs or find flaws & bolt on steel plating

Fueled emotional debate
MNDOTs action:
Image from Wikimedia Commons, http://commons.wikimedia.org
Thousands of bolt holes would weaken bridge
Launched inspection, interrupted by work on bridge surface
The state's top bridge engineer:
"We chose the inspection route.. We thought we had done all we
could, but obviously something went terribly wrong."
"Up until the late 1960s, it was thought that fatigue was not a
phenomenon you would see in bridges.
12
How much do engineers

know about fatigue?
How exact are
fatigue models?
Experimental data
80
70
60
Peak alternating bending stress S, kpsi (log)
50
40
35
30
25
Wrought
20
18
16
14
Permanent mold cast
12
10
Sand cast
8
7
6
5
103
104
105
106
107
108
109
Life N, cycles (log)

Figure by MIT OpenCourseWare. Adapted from Fig. 6-11 in Shigley & Mischke .
14
Experimental data
140
120
Endurance limit S'e, kpsi
0.5
S1e
= 0.6
Su
0.4
105 kpsi
100
80
60
40
Carbon steels
Alloy steels
Wrought irons
20
20
40
60
80
100
120
140
160
180
200
220
240
260
280
300
Tensile strength Sut, kpsi

Figure by MIT OpenCourseWare. Adapted from Fig. 6-17 in Shigley & Mischke.
15
What actions and/or

practices should be put
in place as a result?
Testing and prevention

Where life-limb-$ are important
It is your job to spec out test type and procedure

Balance of cost vs. risk
Many people listen to REAL data
Example types
Ultrasonic
Liquid penetrant
Stiffness/impulse
Eddy-current
Leaks
Visual
Few people listen to Eqxns

A choice: Job vs. safety
Images removed due to copyright restrictions.
Please see http://www.labino.com/bilder/applications/00533_RT8.jpg
http://www.riverinaairmotive.com.au/img/mpi001.jpg
On foreseeable use
Common sense
Legal
17
Where do cracks come

from?
Potential crack origins/causes

Inherent to material
Imperfections, e.g. castings
Precipitates, e.g. Al 6061 T6
Coalescing of internal dislocations
Grain boundaries
Image removed due to copyright restrictions. Please see

http://www.emeraldinsight.com/fig/2190080205007.png
Fabrication-related
Tool marks
Improper assembly, e.g. forcing (car suspension-cast materials)
Thermally induced - Weld cracks and related HAZ problems
Use-related
High stress areas

Scratches
Unintended use/damage/loading (e.g. 3 finger tight and paint lid)
19
Fatigue: Origin of problem

http://www.metallographic.com/Images/Zn-Al.jpg
http://corrosionlab.com/Failure-Analysis-Studies/Failure-Analysis-Images/
20030.SCC.304H-pipeline/20030.microstructure-ditched-grain-boundaries.jpg
and
Fig. 4b in Henderson, Donald W., et al. "The Microstructure of Sn in Near-Eutectic Sn-Ag-Cu Alloy Solder Joints and its Role
in Thermomechanical Fatigue." Journal of Materials Research 19 (June 2004): 1608-1612
or
Slide 36 in Kang, Sung K. "Near-Ternary Sn-Ag-Cu Solder Joints; Microstructure, Thermal Fatigue, and Failure Mechanisms."
Pb-Free Workshop, TMS Annual Meeting, February 2005.
20
Fatigue life review

Low cycle High cycle
Log(S)
Sut
Se
Log(103)
Log(106)
Log(N)
21
Fatigue life review

Ferrous materials/allows: Se ~ 1 000 000 10 000 000
Under ideal conditions
Non-ferrous (i.e. aluminum) generally no Se
Do we use Al in places where fatigue is important?
Aircraft
History Channel Boneyard
Low cycle High cycle
Methods
Stress
Strain
Fracture mechanics
Sut
Log(S)
Science vs. engineering
Se
Log(103)
Log(106)
Log(N)
22
Fatigue and ethical responsibility
You will be criminally

negligent if you do not
augment calculations
with TESTING for
critical fatigue
applications
Life-limb-$
23
Real life
Please see datsun_laurel. "R06 Front Carbon ARB Fatigue Test." Photobucket. Accessed October 14, 2009.
http://s23.photobucket.com/albums/b388/datsun_laurel/FSAE/?action=view&current=ARB_Test.flv
Mart n Culpepper,
i
All r ghts ireserved
24
Real life
http://video.google.com (author?)
Please see SmithersMpls. "Carbon Frame Fatigue Test." February 9, 2007. YouTube.
Martin Culpepper, All rights reserved Accessed October 14, 2009. http://www.youtube.com/watch?v=QHO_VjVhaE8
25
Real life
http://yourdailymedia.com
26
Please see motocross. "The Nitro Circus: Channel 9 Action News." March 3, 2009. YouTube. Accessed October 14, 2009.
http://www.youtube.com/watch?v=kOaYruVcvm4
What reasonable
hypotheses could one
hold for identifying
important factors?
Fatigue life modifiers

Experimental results are used to obtain modifiers
Se = ( ka kb kc kd ke k f ) S'e
Where:
ka
kb
kc
kd
ke
kf
=
=
=
=
=
=
Surface condition modification factor
Size modification factor
Load modification factor
Temperature modification factor
Reliability modification factor
Others
Se
Se
=
=
Rotary-beam test endurance limit
Predicted endurance limit for your part
28
Endurance limit depends on many factors

For ferrous materials, the following approximations may
be used for first pass design
S'e =
0.5 Sut
Sut d 200kpsi
100kpsi
Sut ! 200kpsi
700MPa Sut ! 1400MPa

This is for ideal conditions but designs are never ideal
3
7 in
16
4.94
Aluminum alloy 7075-T73

Rockwell B 85.5
25.5
10.200
Vs.
0.30 in
7
9 in R.
8
Lug (1 of 2)
Fracture
A
Figure by MIT OpenCourseWare. Adapted from Fig. 6-8 and 6-9 in Shigley & Mischke.
29
Experimental data
140
120
Endurance limit S'e, kpsi
0.5
S1e
= 0.6
Su
0.4
105 kpsi
100
80
60
40
Carbon steels
Alloy steels
Wrought irons
20
20
40
60
80
100
120
140
160
180
200
220
240
260
280
300
Tensile strength Sut, kpsi

30
Fatigue life modifiers: Surface condition

Experimental results are used to obtain modifiers
ka = a S
ut
Where:
a = function of fabrication process

b = function of fabrication process
Why does finish matter?
Surface finish
Factor a
Exponent b
Sut, kpsi
Sut, MPa
Ground
1.34
1.58
-0.085
Machined or cold-drawn
2.70
4.51
-0.265
Hot-rolled
14.4
57.7
-0.718
As-forged
39.9
272.
-0.995
Figure by MIT OpenCourseWare. Adapted from Table 6-2 in Shigley & Mischke.
31
Why would surface

condition matter?
Surface roughness review
Mean
1 L
y Ra = L 0 y dx
x
Common surface roughness (Ra in micro-inches)

Process
2000
1000
500
250
125
63
32
16
Sawing
Drilling
Milling
Turning
Grinding
Polishing
Only specify what you need & know your processes

33
Why would part size

matter?
Fatigue life modifiers: Size factor

For bending and torsion of a round bar:
kb =
0.879 d
0.107
0.910 d
0.157
for 0.11in d 2.00in

for 2.00in d 10.0in
For axial loading:
kb = 1
What if the bar is not round?
Use a 95 percent stress area
Equate volumes, length drops out
Relate cross sectional area of round and square bar
d e = 0.808h b
0.5
35
Why would the type of

loading matter?
Fatigue life modifiers: Loading factor

For bending and torsion of a round bar:
1.00
kc = 0.85
0.59
bending
axial
torsion
37
Why would temperature

matter?
Fatigue life modifiers: Temperature factor

The effect of increasing temperature
Yield strength typically decreases
May be no fatigue limit for material-temperature combos
The temperature factor
May be ESTIMATED from existing tables

Should ALWAYS BE DETERMINED EXPERIMENTALLY FOR YOUR
GIVEN MATERIAL.
Relate strength at temperature to room temp. strength
ST
kd =
S RT
39
Fatigue life modifiers: For an example steel

Temperature, oC
ST/SRT
o
Temperature, F
ST/SRT
20
1.000
70
1.000
50
1.010
100
1.008
100
150
1.020
1.025
200
300
1.020
1.024
200
1.020
400
1.018
250
1.000
500
0.995
300
350
0.975
0.943
600
700
0.963
0.927
400
0.900
800
0.872
450
0.843
900
0.797
500
550
0.768
0.672
1000
1100
0.698
0.567
600
0.549
Figure by MIT OpenCourseWare. Adapted from Table 6-4 in Shigley & Mischke.
40
Part II
Calculations
How do statistics and

probability come into
play?
Standard normal distribution, mean = 0

Standard normal distribution curve generated via the
probability distribution
Area under the curve = 1
f (z ) =
This will be covered

in the 2nd design lab
1 2
exp (z )
2
2
What if mean is not 0?

2
1 x x
1

exp
f (x ) =
x 2
2 x
f (z )
Standard normal
distribution curve
-1.1
z
43
Non-zero means in Gaussian distributions

A normal Gaussian distribution is typically observed in
fatigue behavior of parts
x = variate = x
z = transformation variate
= standard deviation
1
1 x x

f (x ) =
exp
x 2
2 x
x x
z=
x
f (z ) =
Normal
distribution curve
1 2
exp (z )
2
2
Only one table for values needed to find the area

between z values
44
Fatigue life modifiers: Reliability factor

Normal
distribution curve
x x
z=
x
Standard normal
distribution curve
-1.1
45
Example use of standard normal distribution

In a shipment of 250 connecting rods, the mean tensile
strength is 45 kpsi and the standard deviation is 5 kpsi
z 39.5
(a) Assuming a normal distribution, how many rods may be expected to

have a strength less than 39.5 kpsi?
(b) How many are expected to have a strength between 39.5kpsi and
59.5kpsi?
x x 39.5 45.0
=
=
= 1.10
x
5.0
(z 39.5 ) = ( 1.10 ) = 0.1357

Standard normal
distribution curve
N (z 39.5 ) = 250 0.1357 = 33.9 34

-1.1
x
46
Fatigue life modifiers: Reliability factor

Most strength data is reported as mean values
Standard deviations typically less than 8%, but you MUST KNOW what it
is run experiments
68% of all measurements fall within one standard deviation
95% of all measurements fall within two standard deviations
For ~ 8%
ke = 1 0.08 za
Standard normal
distribution curve
-1.1
x
47
How do we do
st
1 order fatigue
modeling/analysis?
Stress values of concern
min
max
a
m
s
Minimum stress
Maximum stress
Amplitude component = (max - min)/2
Midrange component = (max + min)/2
Steady component
R
A
Stress ratio = min / max
Amplitude ratio = a / m
Note the correction to

Va and Vm
49
1.2
Amplitude ratio Sa /Se
1.0
0.8
0.6
0.4
0.2
C
-1.2
-1.0
-0.8
-0.6
-0.4
Compression Sm /Suc
-0.2
0.2
0.4
0.6
Tension Sm /Sut
0.8
1.0
Midrange ratio
50
Fatigue diagram: Goodman

Capital S = strength!
Sa a
r=
=
Sm m
Se
Load line
Sya Sxm
+
=1
Se Sut
1
+
=
Se Sut n
Sa
Modified Goodman line
'a
0
0
Sm
Sut
51
Stress concentration and notch sensitivity

Fatigue is due to crack propagation, hence notch
sensitivity is important
Max stress
max = K f o
Kt from table A-15
max = K fs o
Notch sensitivity, q (usually 0 < q < 1) accounts for material sensitivity

Kf = 1 + q (Kt-1)
Table 6-20
Kfs = 1 + qshear(Kts-1)
Table 6-21
It is always safe to use Kt

Kt rarely > 3 for good/practical designs, but check!
52
Example
1.5 in diameter AISI 1050 cold drawn steel (Sy = 84kpsi,
Sut = 100 kpsi) withstands a tensile load that ranges
from 0 to 16000 lbf. Kf = 1.85, ka = 0.797, kb = 1, kd = 1,
kc = 0.923. (8th edition has kc = 0.85)
Modifications in example:
kc = 0.85 in 8th edition

Se = Sut in 8th edition
Example modified for 8th edition
a. Factor of safety if Va held constant
b. Factor of safety if Vm held constant
c. Factor of safety if Va/Vm = constant
Se = kakbkckdSe = kakbkckd (0.5 Sut) = 33.9kpsi
53
Part b: Va held constant

Before applying Kf
= 4.5kpsi
After applying Kf
=8.38 kpsi
33.9
Se
Va [kpsi]
8.38
Sut
75.3
Sa
75.3kpsi
=
= 8.98
n=
a 8.38kpsi
Vm [kpsi]
100
54
Part a: Vm held constant

33.9
31.1
Se
Va [kpsi]
Sut
8.38
Sa
31.1kpsi
=
= 3.71
n=
a 8.38kpsi
75.3
Vm [kpsi]
100
55
Part c: Va / Vm held constant

33.9
Se
25.3
Va [kpsi]
8.38
Sut
8.38
n=
Sa
Sm
25.3kpsi
=
=3
8.38kpsi
25.3
100
Vm [kpsi]
56
What about your shaft?

Step 1: Free body diagram
Cutting forces (2.008 person and/or next week)
Driving loads
Reaction loads
Preloads
Others OS! loads?
I can be here
Saturday to help,
if people ask!!!
Step 2: Parametric geometry & load variables

Step 3: Material properties
Step 4: Force magnitudes estimates/calculations
Step 5: Stress & fatigue
Vx, Vy, Wxy, Kt, q, Kf, Va, Vm, Vs, Vx,
In the end, SH so you should program this into excel:
In case you need to change variables there are always changes!

Optimization in excel.
57
MIT OpenCourseWare
http://ocw.mit.edu

Spring 2009
2.72
Elements of
Mechanical Design
Lecture 05: Structures

Quizzes
Quiz None
Topics
Finish fatigue
Finish HTMs in structures
Reading assignment
None
Quiz next time on HTMs
Matrix Review
What is a Matrix?
A matrix is an easy way to
represent a system of linear
equations
Linear algebra is the set of
rules that governs matrix
and vector operations
b1
b
2
Vector
a1
a
3
a2
a4
Matrix
Matrix Addition/Subtraction
You can only add or subtract matrices of the same dimension
Operations are carried out entry by entry
a1
a
3
a2 b1
+
a4 b3
(2 x 2)
a1
a
b2 a1 + b1
=
b4 a3 + b3
(2 x 2)
a2 b1 b2 a1 b1
a4

b3 b4
a3
b3
(2 x 2)
(2 x 2)
a2 + b2
a4 + b4
(2 x 2)
a2 b2
a4 b4
(2 x 2)
Matrix Multiplication
An m x n matrix times an n x p matrix produces an m x p matrix
a1
a
a2 b1
a4
b3
(2 x 2)
b2 a1b1 + a2b3
=
b4

a3b1 + a4b3
(2 x 2)
a1b2 + a2b4
a3b2 + a4b4
(2 x 2)
Matrix Properties
Notation: A, B, C = matrix , c = scalar
Cumulative Law:
A+B=B+A
Distributive Law:
c(A + B) = cA + cB
C(A + B) = CA + CB
Associative Law:
A + (B C) = (A + B) C
A(BC) = (AB)C
NOTE that AB does not equal BA !!!!!!!
Matrix Division
To divide in linear algebra we multiply each side by an inverse
matrix:
AB = C
A-1AB = A-1C
B = A-1C
Inverse matrix properties:

A-1A = AA-1 = I (The identity matrix)
(AB)-1 = B-1A-1
Structures
Machines structures
Structure = backbone = affects everything
Satisfies a multiplicity of needs
Enforcing geometric relationships (position/orientation)
Material flow and access
Reference frame
Requires first consideration and serves to link modules:
Joints (bolted/welded/etc)
Bearings
Shafts
Parts
Tools
Sensors
Actuators

http://www.clarkmachinetools.com/2003_1.jpg
10
Key issues with structural design

Machine concepts
Topology
Material properties

http://www.fortune-cnc.com/uploads/images/1600ge_series.jpg
Principles
Thermomechanical
Elastomechanics
Kinematics
Vibration
Key tools that help
Stick figures
Parametric system/part error model
Visualization of the:
Load path
Vibration modes
Thermal growth
11
Modeling: stick figures

http://americanmachinetools.com/images/diagram-lathe.jpg
1
2
12

1. Stick figures
These types of models

are idealizations of the
physical behavior. The
designer must KNOW:
2. Beam bending
3. System bend.
(a) if beam bending

assumptions are valid
(b) how to interpret
and use the results o
this type of these
models
1
2
x
13
14
Transformation
Matrices
Translational
Transformation
Matrix
Translational
Transformation
Matrix
Y
P=(A+L
P=(A,
B)1, B+L2)=(A, B)
B
L1
X
A
L2
A
X
B
1
L1
L2
16
Translational Transformation Matrix
General 2D transformation matrix
L1
L2
17
Rotational
Transformation
Matrix
Rotational
Transformation
Matrix
Y
Y
P=(A, B)
P=(A,
B)
B
X
B
X
A
18
Rotational
Transformation
Matrix
Rotational
Transformation
Matrix
Y
P=(A, B)
X
B
A = A cos + B sin
19
Rotational Transformation Matrix

Y
P=(A, B)
X
B
A = A cos + B sin
B = -A sin + B cos
20
Rotational
Transformation
Matrix
Rotational
Transformation
Matrix
Y
Y
P=(A, B)
X
B
A
cos
sin
B =
-sin
A = A cos + B sin
1
0
B = -A sin + B cos
cos
21
Rotational
Transformation
Matrix
Rotational
Transformation
Matrix
Y
(+) Counter Clockwise
General 2D rotational
matrix:
cos
sin
-sin
cos
(-) Clockwise
-sin 0
cos
sinsin
cos
cos
0
-sin
0
cos
0
0 0
0
1
X
22
Homogeneous Transformation Matrix
General 2D HTM translational and rotational matrix:
cos
sin
-sin
cos
23
HTM Applications
Simple
Beam
Example
Simple
Beam
Example:
F
L
25
Simple
Beam
Example
Simple
Beam
Example:
A
B
Y
X
26
Useful
Equations
ForceForce-deflection
Deflection Equations
d
d=
F
FL
EA
FL3
d=
3EI
FL2
=
2EI
ML2
d=
2EI
ML
=
EI
27
Simple
Beam
Example
Simple
Beam
Example:
A
B
L
A
B
L
cos
B D =
A = sin
1
-sin L
cos
FL3
D=
3EI
FL2
=
2EI
28
Drill
Press
Example
Simple
Beam
Example:
Find the HTM from a to b:
F
d
Ha =
Cross-Sectional Area of large sections =

Cross-Sectional Area of Drill Bit =
Youngs Modulus of Material =
-(D-)
Ad
E
29
Useful
Equations
d
d=
F
FL
EA
FL3
d=
3EI
FL2
=
2EI
ML2
d=
2EI
ML
=
EI
30
Drill
Press
Example
Simple
Beam
Example:
Find the HTM from a to b:
F
d
Ha =

Ad
-(D-)
FD
EAd
31
Drill
Press
Example
Simple
Beam
Example:
Find the HTM from b to c:
L
c
F
d

Hb =
cos
-sin L
sin
cos
Ad
E
32
Useful
Equations
d
d=
F
FL
EA
FL3
d=
3EI
FL2
=
2EI
ML2
d=
2EI
ML
=
EI
33
Drill
Press
Example
Simple
Beam
Example:
Find the HTM from b to c:
L
c
F
d

Ad
Hb =
cos
-sin L
sin
cos
A
FL3
=
3EI
FL2
=
2EI
34
Drill
Press
Example
Simple
Beam
Example:
Find the HTM from c to d:
L
c
F
FL

Ad
Hc =
cos
-sin
-1
sin
cos
H+2
E
35
Useful
Equations
d
d=
F
FL
EA
FL3
d=
3EI
FL2
=
2EI
ML2
d=
2EI
ML
=
EI
36
Drill
Press
Example
Simple
Beam
Example:
Find the HTM from c to d:
L
c
F
FL

Hc =
Ad
FLH
=
EI
cos
-sin
-1
sin
cos
H+2
FLH2
1 =
2EI
FH
2 =
EA
37
Drill
Press
Example
Simple
Beam
Example:
Find the HTM from d to e:
L
c
F
FL
F
d
Hd =
e
Ad
-L
E
38
Drill
Press
Example
Simple
Beam
Example:
Find the HTM from a to e:
L
c
F
d
H a = e Hd d Hc c Hb b Ha
Find the vector aVe from e to a:

0
e
aV

Ha
Ad
E
39
Method
building
systems HTM
Simplefor
Beam
Example:
Identify key nodes around the systems
structural loop
Create HTMs for each member between each
node
Multiply the members HTMs in the correct
order
40
More on HTMs
3D HTMs
For x-axis
rotation
For y-axis
rotation
For z-axis
rotation
0
1
0 cos
x
0 sin x
0
0
cos y
0
sin y
cos z
sin
z
0
sin x
cos x
0
0 sin y
1
0
0
cos y
0
sin z
cos z
0
0
Y
Z
1
X
Y
Z
For small :
cos()~1 & sin()~
0 X
0 Y
1 Z
0 1
42
HTM Rotation
Remember order of multiplication matters:
180
deg
flip
90
deg
right
90
deg
right
180
deg
flip
To combine a translation and rotation, again multiply the HTM matrices

together
Note that the order of the rotation and translation matrices does matter, so
makes sure the answer makes sense!!!
43
MIT OpenCourseWare
http://ocw.mit.edu

Spring 2009
2.72
Elements of
Mechanical Design
Lecture 06: Constraints

Quiz
Quiz HTMs
Topics
Principles of exact constraint
Bearing layout exercises
Spindle shaft constraint/bearing layout
Reading assignment
Chapter 11 in Shigley and Mischke

Read sections 11.1 11.6, 11.9
Skim sections 11.7 11.8, 11.10 11.12
This is 40ish pages, but most of it is pictures/graphs/examples
Principles of
exact constraint
Under, exact and over constraint

Constraints are fundamental to mechanical design

A mechanical designers goal is to control, i.e. to constrain, parts so that

they are where they are supposed to be.
Exact constraint:

There should be one constraint for each degree-of-freedom that is

constrained.
Under constraint
Too few constraints, part is not held in all the directions it needs to be
Over constraint

Too many constraints, some constraints may fight each other when
trying to do the same job.
Mechanical constraints
We want to learn how to model and design each
We first need to know:
How they should be used
What their functional requirements are
How they should be used and their FRs depend upon:
How they are laid out
Dos and donts
Learn to lay them out right
Use this to obtain their Functional Requirements
Then do the detailed design of each
Constraints and Degrees-of-freedom

Rigid components have 6 degrees-of-freedom
We will represent an ideal constraint as a line
6 C = R
C = # of Non-Redundant Constraints
R = # of Independent Degrees of Freedom
Example:
6 - 2constraints = 4DoFs
Rz
Courtesy John Hopkins, MIT
Constraints, compliance and motion

Exact constraint: Achieve desired motion
By applying minimum number of constraints

Arranging constraints in correct constraint topology
Adding constraints only when necessary
y
z
Ground
For now:
Stage
Start with ideal constraints

Considering small motions
Constraints = lines
Focus on rigid stage attached to ground
What do we mean by rigid?
What do we mean by constraint? - Stiffness ratios
The benefits of exact constraint

Y
X
Parts of machines are now

always the same strength and
stiffness.
Large, stiff components have a

tendency to kill their smaller
counterparts when they are
connected so that they are
forced to fight.
Loose
Exact constraint design helps to

prevent fights, therfore all your
parts live in harmony.
Binding
B
Figure by MIT OpenCourseWare.
Penalties for over constraint

Constraint examples
y
More examples:
1
Exact constraint
practice
Mechanical constraints: Some bearing types

Sliding
Bushings
etc
Radial rolling
Radial
Shallow groove
Deep groove (Conrad)
Angular contact
Tapered
11
12
Example: Carriage constraint

y
z
Assume these are

bushings as in
your lathe
13
Example cont.
y
z
Misalignment
& assembly
errors
14
Example cont.
y
z
Thermally
Induced
growth
15
Practical embodiment
y
z
Flexure that is stiff in y and z, yet compliant in x
16
Group exercise Carriage constraints

Identify the motions that you desire for the carriage and
the minimum # of constraints that are needed to yield
only these motions.
Identify the constraints from each bearing set and

determine how they act in concert to yield the desired
motions.
17
Constraint layouts
and
thermal stability
Avoiding over constraint

How to deal with the thermal growth issue
The shaft typically gets hotter than the housing because the housing has
better ability to carry heat away
Whether the outer or inner race are fixed, matters
Constraining front and rear bearings
One bearing set should be axially and radially restrained

The other bearing set should ONLY have radial restraint
Housing
Gear
Nut
Gear
Shaft
10.0 All rights8.0
Martin Culpepper,
reserved
Tool
Chuck
Part
4.0
19
Examples: Good or bad

Outer race fixed axially, if shaft heats is this bad?
Think about what happens to the preload
Housing
Chuck
Shaft
20

Outer race fixed axially, if shaft heats is this bad?
Think about what happens to the preload
Housing
Chuck
Shaft
This is the
back-to-back config.
It should be used
when the outer race
is not rotating
21

Assume the outer
race does not rotate
1
Housing
Sliding permitted
Gear
Nut
Spacer
Gear
Chuck
Shaft
22

Assume the outer
2
Housing
Sliding permitted
Gear
Nut
Spacer
Gear
Chuck
Shaft
23

Assume the outer
3
Housing
Gear
Nut
Gear
Chuck
Shaft
24

Assume the outer
4
Housing
Gear
Nut
Gear
Spring
Chuck
Shaft
25

Assume the outer
5
Housing
Gear
Nut
Gear
Chuck
Shaft
26

Assume the outer
6
Housing
Gear
Gear
Chuck
Shaft
27

Assume the outer
Good
Front set constrained radially and axially

Rear set constrained only radially
Housing
Sliding permitted
Gear
Nut
Spacer
Gear
Chuck
Shaft
BUT, sliding means some gap must exist and therefore

one must precision fabricate if a small gap is desired
28

Assume the outer
Bad
Front set constrained only radially

Rear set constrained only radially
Housing
Sliding permitted
Gear
Nut
Spacer
Gear
Chuck
Shaft
This design will not work if axial loads are to be applied

along both directions
29

Assume the outer
Good

Rear set constrained radially and axially
Housing
Gear
Nut
Gear
Chuck
Shaft
BUT, adding a spring increases part count/cost

Axial stiffness values (left vs. right) will be different
30

Assume the outer
BAD
Front set is constrained only axially

Rear set is constrained only axially
Housing
Gear
Nut
Gear
Spring
Chuck
Shaft
This design will not work if axial loads are to be applied

along both directions
31

Assume the outer
BAD
Front set is constrained axially and radially

Rear set is constrained axially and radially
Housing
Gear
Nut
Gear
Chuck
Shaft
At high speeds/loads, thermal growth may kill the

bearing sets
Like a double face-to-face
32

Assume the outer
Bad
Rear set constrained radially (if flexure is of proper stiffness)
Housing
Gear
Gear
Chuck
Shaft
Left bearing set is not in the back-to-back configuration

Shaft can pop out
33
Group exercise
Group exercise Spindle constraints

The spindles you have seen use tapered roller bearings.
First, sketch a layout from one of the previous lathes

and diagnose its layout
Second, generate and sketch a different way to

constrain the spindle shaft.
35
MIT OpenCourseWare
http://ocw.mit.edu

Spring 2009
2.72
Elements of
Mechanical Design
Lecture 07: Rolling
contact bearings

Quiz
Constraints
Topics
Bearing types & failure modes

Experimental results to modeling
Bearing load-life-reliability
Start spindle exercise
Reading assignment
None, work on getting bearings done
Quiz Tuesday
Bearings (conceptual)
www.renault4.co.uk/
Courtesy of Clementine's Renault 4 Garage. Used with permission.
Besides the childrens toy (sorry bout that), what do

you notice about the bearing that should be of import?
2
Loose ends
Thermal stability
Function of separation distance and contact geometry - cancellation

Will you always have bearing death if it is not thermally stable?
T
q& =
RT
RT Conv =
RT Cond
1
h Asurface
t
=
k Across
Mounting of races: Inside rotating vs. outside rotating
Friction torques
Look at the tapered roller example from lab
Rough design steps Inherently iterative

Step 1: Functional requirements
DOF
Stiffness
Lifetime/reliability
Etc
Iteration
Step 2: Bearing type/layout

Step 3: System design & mfg issues
Housing geometry
Shaft deflection
Thermal
Shaft geometry
Preload
Tolerances
Step 4: Assembly specifications

Types of rolling contact

bearings
Examples of rolling element bearings

http://www.timken.com/en-us/products/bearings/productlist/ball/PublishingImages/radialball_150.jpg
http://images.machinedesign.com/images/archive/news1300jpg_00000036367.jpg
http://www.bearingsworld.com/image/pic-cp02.jpg
http://www.rlmcor.com/image-product/23134mbw33.jpg
Elements: Rotary rolling contact bearing

Inner
Ring
Race
Outer
Ring
Race
Diameters
Outer
Bore
Ball/roller
Cage/separator
Face
Width
Bearing failure:
Causes and
failure mode
Cracks in bearings elements

Images removed due to copyright restrictions. Please see
http://www.amstedrail.com/techinfo/media/94_1f5b.jpg
http://www.amstedrail.com/techinfo/media/94_1f2d.jpg
Failure mode: Spalling

http://www.theautoist.com/Bearing5.JPG
http://www.tsb.gc.ca
Courtesy the Transportation Safety Board of Canada.
10
Causes of spalling
Spalling
Surface fatigue that occurs as a result of contact
Seeds of failure
Steel quality key to

making long-lasting
bearings
Crack growth
Inclusions
Impact
Cyclic high stress
Degradation of the lubricant
http://www.tsb.gc.ca
Courtesy the Transportation Safety Board of Canada.
Once it starts, what happens?
Minor spalling + correct problem may slow/stop

Typically increases in size with continued service
Image removed due to copyright restrictions.

Please see:
http://www.theautoist.com/Bearing5.JPG
11
Preload
Ball-flat elastomechanics: Hertz contact

Model ball-groove contacts as six balls on flats
Hertzian contacts
act as non-linear
springs
F
Hertz
1857-1894
Important relationships for ball-flat contact
kn = 2
R 2 Ee
13
Location, magnitude of max shear stress

Equivalent radius
Re =
1
1
R1major
1
R1min or
1
R2major
1
R2min or
Max
Equivalent modulus
Ee =
1
1 1 1 2
+
E1
E2
2
max depth =0.48contact radius

center of contact
Max shear stress occurs below

surface, in the member with
largest R if ball and flat of same
material
1 6 F Ee
= 0.31
2
Re
1
3
14
Sensitivity of contact stiffness to F

Preload increases stiffness
k [N/micron]
2
1
1
k
n
=
K
o

Re
3 Ee 3 Fn 3
Contact stiffness
30
Large k for small Preload
20
10
05
0
A little preload goes a long way
250
500
750
1000
Fn [N]
Classes of preload, as % of static load capacity
Heavy
Medium
Light
5%
3%
2%
15
How do you preload a bearing?

Direct
Nuts pressing directly on races
Uses compliance of contacts
Many bearings come

preloaded out of the box
Check to make sure so that
you do not add preload that
will act to overload
Internal
Oversized balls
Uses compliance of ball-race contacts
Nuts-springs
Spring in series with bearing
Primarily uses compliance of spring
Think in terms of relative stiffness because

Sensitivity of force to the displacement
16
Bearing life and

reliability
Bearing life at rated reliability

Bearing life depends on:
Load and revolutions
From experimentation: For a given R, linear log behavior
Reliability = % chance that the part will function as expected
log(F)
F L =
Where
a = 3 for ball bearings

a = 3.333 for roller bearings
log(L)
18
Regression fit to experimental data

Manufacturers provide a match of:
Cycle rating: Revolutions

Load rating: Force
For example 106 or 906 revs

Anything
that defines bearing failure for a given reliability

Given these two numbers, and:
1
Units of force
a
This is for common

reliability
Units of revolutions
F L =
Constant
This may be used to extrapolate behavior at different

loads and revs
1
1
Why C vs. F?
Ci La i = Fdesign La design
19
Design life in terms of hours or revolutions?

We can think in terms of life as time if:
rev
min
Dynamic
t [hr ]
60
= L
load
min
hr
rating
Revs at failure
1
C10 L10 a
= Fdesign
Be careful with units
1
La design
= Fdesign (t 60)
1
a
C10 (trated rated 60 ) = Fdesign (tdesign design 60)

1
a
C10 = Fdesign
1
a
(t
design
design 60)
(trated rated 60)
1
a
1
a
20
Reliability vs. life

Reliability often well-predicted via Weibull distribution
xo = minimum guaranteed value of x
= corresponds to 63.2 percentile of the variate (stochastic variable)
b = a shape parameter (controls skew, large = right)
r(x)
x x b
o
R ( x ) = exp
x

o
xo
This is for common

load on bearings
For bearings, we use this as:
L
x=
L10
Commonly used to fit experimental

data; b & come from fit
21
Relationship between load, life and reliability

BUT the catalogue never tells me what happens for.
My exact desired load
My exact desired life
Situations when I want a reliability that is different than R = 0.90
Constant reliability :
1
1
FB xB a
= =
1
FD xD a
log(F)
B
C10
R = 0.90
Constant load
x x b
2
B
o
R ( x ) = exp
x

o
D
Design
x
L
x=
LPF
log(x)
22
Relationship between load, life and reliability

Load as a function of reliability & vice versa for given:
C10
L10
R = 0.90; x = 1 when L = L10
log(F)
A
C10
R = 0.90
Design
1
a
x
D
C10 =
FD
1
b

1
xO
+
(

xO
) ln

log(x)
L
x=
LPF
23
Reliability of multi-bearing sets

What is the reliability of:
One bearing?
A spindle with two bearings?
With N bearings?
For first order design, how should individual bearing

reliability scale as a function of N?
24
How to handle combined loading

So far we have only considered radial loading
What about combined radial, Fr, and axial loading , Fa?
Use an equivalent load, Fe, that does the same amount of damage.
Fe = X i V Fr + Yi Fa
Where V = 1.2 for outer ring rotation and 1 for inner ring
This has to do with the fact that outer ring fails more often
Xi and Yi are a function of the
Axial load, Fa
Static load rating, Co
25
Other issues, Shigley/Mischke covers well

Life recommendations (hrs)
Aircraft engines
24-hour critical service
0 500
100 000
2 000
200 000
Application factors
General commercial
Moderate impact
1.1 1.3
1.5 3.0
26
Group exercise
Work on your spindle housing-shaft-bearing design
Constraint layout
Loads
Preload
Thermal stability
Cost
27
MIT OpenCourseWare
http://ocw.mit.edu

Spring 2009
2.72
Elements of
Mechanical Design
Lecture 08: Flexures

Quiz
Today: Bearing layouts (mid-class)
Thursday: Hale 6.1
Soon: Bolted joint qualifying quiz
Topics
Flexure constraints and bearings Degrees of Freedom
Reading assignment
Thursday:
Layton Hales thesis Read 2.6, 2.7, 6.1, skim rest of Chapter 6
Chapter 7 is cool to look at
Tuesday:
Read: 8.1, 8.3 8.5, 8.7, 8.9 8.11
Skim: 8.6, 8.8, 8.12
Mechanisms: Compliant vs. rigid

Rigid mechanisms
Sliding joints

http://www.physikinstrumente.com/en/primages/pi_m850_tip_i4c_o_eps.jpg
http://www.hexapods.net/images/M850Ani160-1-slow.gif
100s of nm resolution
Large range
kg load capacity
Compliant mechanisms
Motion from member compliance
Angstrom resolution
Limited range
Limited load capacity
z
Micro-scale precision machines
Static
SEM: Drs. Andras Vladar & Jason Gorman (NIST)

FIB: Dr. Konrad Jarush
(Hitachi)

Courtesy of Andras Vladar, Jason Gorman, and Konrad Jarausch. Used with permission.
Meso-scale devices: Biomedical
Two - photon endomicroscope

Lens
Scan
ning
syst
em
Lighting
Forceps
Nano-scale devices
Stator
CNT
Rigid link
Rotor
Compliant
joint
y
x
Meso-scale precision machines
Nano-scale devices
Two -photon endomicroscope
TMA
Lens
Sc a n
ning
syste
m
Lighting
Forceps
Stator
CNT
Rigid link
Rotor
Compliant
joint
y
x
10
Dip pen nanolithography on DNA arrays

What is fundamentally different?
Size Physics Fabrication

Raw materials
Surfaces vs. points or lines
http://mcf.tamu.edu/images/DPN_process.png
http://www.nanoink.net/d/Nano%20-%20Part%201_Sm_Lo-Res_240x180.wmv
http://images.iop.org/objects/nano/news/4/12/10/diagnal.jpg
250 mm
Courtesy
PI
~20 mm
~1 mm
11
Nanomanufacturing
12
Advantages of flexures
Advantages
Smooth, fine motion
Linear/elastic operation in absence
Failure modes are well understood
Monolithic or assembled
2D nature lends to 2D mfg.
Miniaturization
z y
x
Disadvantages
y
x
Accuracy and repeatability sensitive to several variables

Limited motion/stroke (usually a few to 10s % of device size)
Instabilities such as axial or transverse buckling
Dynamics
Sensitivity to tolerance
13
Elastomechanics ( & ) relationship
Elastic
= E
600
Plastic
500
Failure 80
70
60
400
, kpsi
, MPa
12L14 Steel True Stress-Strain Behavior
50
40
30
20
300
Youngs modulus, E
200
190 GPa [27.4 Mpsi]
100
Material
Titanium V
Aluminum 7075
Stainless 316
Invar - Annealed
y/E
1.00
0.70
0.09
0.19
0.00
0.01
0.02 0.03
, mm /mm
0.04
10
0.05
14
Important material properties

Nominal values
Modulus
Yield stress
Coefficient of thermal expansion
Thermal diffusivity
Density
Material property ratios

Material
Titanium V
Aluminum 7075
Stainless 316
Invar - Annealed
y/E
1.00
0.70
0.09
0.19
Normalized Values
E/
diff/CTE
0.14
0.92
1.00
1.00
0.13
0.94
0.87
0.70
Cost
3.77
1.00
3.50
5.21
15
Modules
Lever
Chevron
16
Modules cont.
Ellipse
Cantilever/flexure blade
17
Modules cont.
Flexure hinge
Torsion
Parallel four bar
Double parallel four bar
18
Module cont.: Cross flexure pivot

Deformation scale 1 : 1
75 mm
25 N
25 N
||
25 N
op
19
Review of constraint fundamentals

Rigid bodies have 6 DOF
Constraints have lines of action

C = # of linearly independent constraints
DOF = 6 C
F=6C
y
z
Ground
Stage
20
DOF in constraint-based design

A linear displacement may be visualized as a rotation
about a point which is far away
to infi
nity
21
Two principles of projective geometry

Projective geometry comes in useful here
Parallel lines intersect at infinity
Translation represented by a rotation line at a hope of infinite radius

Image courtesy of John Hopkins
MIT MS Thesis
22
Constraint fundamentals
Blandings RULE OF COMPLIMENTARY PATTERNS

Each permissible Freedom (F) is a rotation about a line and each

permissible freedom rotation line must intersect each Constraint (C)
Remember these principles of projective geometry
Parallel lines intersect at infinity

Translation represented by a rotation line at a hope of infinite radius
R = 6 C = 6 5 = 1...
so where is it?
C5
C5
C4
C4
C1
C1
C2
R1
C2
C3
C3
23
Examples
There will be a quiz on this NC

z
1
R1
R1
R2
24
Flexure bearing systems

Spherical ball joint
6C=F
R3
R2
R1
25
Blade flexure
6C=F
26
Parallel guiding mechanism

6C=F
27
Doodle hopper
6C=F
R2
R1
28
Parallel addition rules

z
What is parallel ? Elements are not in the

same load path. Loads are split between
the elements
Add constraints so where there is a
common DOF, then have mechanism DOF
Example: For instance, there are no
conflicts in displacement to z
Adapted from Layton Hales Ph.D. Thesis (MIT)

29
Series addition rules

What is series?
-Differentiate series by load path
-Shared load path = series
Series: Add DOF
Find common constraints
Follow the serial chain
Adapted from: Layton Hales Ph.D. Thesis (MIT)

30
Parallel and series systems

Redundancy does not add
Degrees of freedom
Series
Parallel
Take care of series first, define

them as single element then go
through parallel
Adapted from Layton Hales Ph.D. Thesis (MIT)
31
Accuracy
The accuracy of most flexures is sensitive to:
1. Small variations in dimensions, e.g. thickness
2. Youngs Modulus (E)
3. Time variable errors

Creep
Stress relaxation
Thermal
Dynamic/vibration
32
Repeatability
Flexures can exhibit Angstrom-level repeatability if:
Low material hysteresis

Single crystal materials useful
No dislocation motion
<< yield
Load is repeatable
Magnitude
Direction
Assembly is correct
No micro-slip
No friction in assembly
No yield during assembly
33
Accuracy and repeatability

Difficult to obtain without calibration or adjustment
Geometry
Materials
Loading
Assembly/integration
Environmental
34
Links between kinematics and elasticity

Cantilever
L
F
F L
=
3 E I
3
1
I = b h
3
12
E b h 3
F =

4 L
hn
1
L
3
3
dF
d E b h
E b h

k=
=

4 L
d d 4 L
35
Links between kinematics and elasticity

L
Cantilever
L = 0.05 L
h = 0.05 h
b = 0.05b
E (b + b ) h + h
E b h
1.05
1 = k (1+ 0.42 )
1.05
k + k =
4
4 L
0.95
L L
k = 0.42 k
36
Fabrication processes: EDM

EDM positives
Accuracy (micrometers)
3D
Surface finish (sub-micrometers)
http://www.physikinstrumente.com/en/about/images/pi_WIREEDMC_i4c_K50_eps.jpg
EDM drawbacks
Time (mm/minute)
Cost
37
Fabrication processes: Waterjet

Waterjet positives
Low force
Many materials including
brittle materials and heat
sensitive materials
Rapid (inches/min)

Images courtesy of xiaming on Flickr.
Waterjet drawbacks
Thickness limitations
Kerf limitations
Draft limitations
Accuracy ~ 125 micrometers
Images courtesy of Iansoper on Flickr.
38
Fabrication processes: Milling/cutting

Milling/cutting positives
Flexibility
Any material
Nearly any shape
Milling/cutting drawbacks
Fixturing
Compliance of parts
Work hardening
Surface damage
Image courtesy of jiskar on Flickr.

Please see any other image of milling, such as
http://students.washington.edu/dennyt/fsae/cnc/wc_fixtplate.jpg
39
Fabrication processes: Etching

Etching positives
2 D topologies/shapes
Monolithic
Micron-level features

http://www.ee.ucla.edu/~dejan/ee115c/ucla-graphics/IBM_metal_stack.jpg
http://www.stsystems.com/uploaded_files/1101/images/scallops.jpg
Etching drawbacks
Dimensional control
Scallops
Milanovic, Veljko, et al. "Deep Reactive Ion Etching for Lateral Field Emission
Devices." IEEE Electronic Device Letters 21 (June 2000): 271-273.
Milanovic, Veljko, et al. "Micromachining Technology for Lateral Field
Emission Devices." IEEE Transactions on Electron Devices 48
(January 2001): 166-173.
Please see 371762. "How Microprocessor Work." February 14, 2009.
YouTube. Accessed October 28, 2009.
http://www.youtube.com/watch?v=loMz_l_Fpx4
40
Assembly
Stress and energy
Proper thickness of clamps and clamping load distribution

Spring washer provide force source
Fusing
Clamps members should yield before flexure

Spring washer provide force source
Surface conformity
Micro-slip is a major cause of hysteresis

Deburring and potting/bonding

Please see Fig. 8.5 and 8.6 in Smith, Stuart.
Flexures: Elements of Elastic Mechanisms.
Amsterdam, Holland: Gordon & Breach, 2000.
Misalignment = systematic errors
41
MIT OpenCourseWare
http://ocw.mit.edu

Spring 2009
2.72
Elements of
Mechanical Design
Lecture 09: Alignment

Quiz
Thursday: Hale 6.1
Soon: Bolted joint qualifying quiz
Topics
Lab notebooks
Alignment methods
Kinematic coupling grade bump = grade for use/design
Reading assignment
Read: 8.2
Examples: All in 8.2
Lab notebooks
Technical quality/quantity
Appropriate equations, codes

Units
Important results highlighted/boxed/noted/explained
Graphical quality/quantity
Appropriate sketches/pictures
Pasted CAD/etc
Archival quality
Can this be copied?
Understood by others?
Best practices
Dating and number of pages
Permanent pen
No blank spaces (X out)
Ideal alignment interface

Repeatable
Accuracy
Stiffness (sensitive?)
Load capacity
6 DOF
Perfect constraint
Lowest energy state
Metal molds
High natural frequency
Common alignment methods
Elastic averaging
Compliant kinematic Quasi-kinematic
Accuracy
Passive KC
Active kinematic
Repeatability
Passive kinematic
Accuracy & repeatability

repeatability
Active KC
Quasi-KC
Elastic C
Elastic A
Elastic B
Error
Desired Position
Error
5
Pin-hole
6 DOF
Metal molds
3 2 1 Alignment schemes
Exact constraint couplings

Exact constraint (EC):
Constraints = DOF to be constrained

Deterministic saves $
Balls (inexpensive) & grooves (more difficult to make)
There are many

types of EC couplings,
our time limits us to a
semi-focused study on
kinematic couplings
In KC design the issues are:
KNOW what is happening in the system (coupling)

MANAGE forces, deflections, stresses and friction
Balls
Tetrahedral
groove
Maxwell
V groove
Kelvin
Passive kinematic couplings

Fabricate and forget
micrometer with best practices, 10s of nm recently
What is important?
Contact forces
Contact stress
Stiffness vs. geometry
Stiffness vs. preload
Friction & settling
Thermal loading
Preload repeatability
Preload
Preload (nesting load) is the force

applied to keep the coupling
components engaged and prevent
tipping
Ball motions: Displacements
Fn
9
=
2
16 Re Ee
n _ Ball _ iA
Hertz
1857-1894
n Ball _ iA n _ Ball _ iB
Ball _ iA
Fn
9
=
2
16 Re Ee
Ball _ i = Ball _ iA + Ball _ iB
n Ball _ iB
Ball _ iB
This assumes
that the ball-ball
stiffness is > ~10x
ball-groove stiffness
Ball far-field point
B1
Ball _ iA
Ball _ iB
v
Ball _ i
Groove far-field points
B2
B3
10
Load balance: Force and moment

Preload
Error
Force balance (3 equations)
Contact
v
v
v
v
v
v
v
v
v
Frelative = 0 = (Fpreload + FError ) + (FBall _ 1 + FBall _ 2 + FBall _ 3 + FBall _ 4 + FBall _ 5 + FBall _ 6 )
Moment balance (3 equations)
) (
) (
r
r
r
r
v
v
v
v
v
v
6
M relative = i =1 M Ball _ i + M preload + M error = (rpreload Fpreload + rerror FError ) + rBall _ i FBall _ i
Goal:
1. Solve 6 equations for contact forces
2. Solve normal displacements
3. Solve relative displacements/rotations
Given geometry, materials,
preload force, error force,
solve for local distance of approach

Ball far-field point
Groove far-field points
11
Modeling round interfaces

Equivalent radius
1
R1major
1
1
R1min or
1
R2major
1
R2 min or
Equivalent modulus
Ee =
1
1 1 1 2
+
E1
E2
2
2
9
Fn
n =
2
16 Re Ee
Poissons ratio
20
10
05
0
Youngs modulus
250
0.5
500
750
1000
Fn [N]
Degree of nonlinearity
is reduced as preload
is increased
Important scaling law

k n ( n ) = 2 Re
Contact stiffness
30
k [N/micron]
Re =
Ee n
0.5
Scaling with
Matl properties
and geometry
Preload should be
repeatable in magnitude
& direction
k n (Fn ) = Constant Re 3 Ee
1
) F
12
Friction and lubrication
Magnitude
depends on
coupling
design and
test
conditions
Displacement, m
Wear in vs.
snow balling
Radial Repeatability (Unlubricated)

2
0
Number of Trials
Radial Repeatability (Lubricated)

Displacement, m
The trend of
the data is
important
Slocum, A. H., Precision Engineering, 1988: Kinematic couplings for precision fixturing
Experimental determination of repeatability and stiffness
0
Number of Trials
13

Courtesy of Elsevier, Inc., http://www.sciencedirect.com. Used with permission.
MIT OpenCourseWare
http://ocw.mit.edu

Spring 2009
2.72
Elements of
Mechanical Design
Lecture 10: Bolted
joints
Bolted joint +s and s

Good:
Low cost?
Able to be disassembled
Strong
Compatible with almost any material
Bad:
Takes up a lot of space

Micro-slip/hysteresis/damping problems
Difficult to model and control
Can require long fabrication and assembly time
Bolted joints: Their purpose

Bolted joints = connectors, impact many parts:
Stiffness
Vibration
Damping
Stability
Load capacity
Bolted joints are semi-permanent!
Max benefit obtained when it is highly preloaded, i.e. near the yield point
Threads can plastically deform/work harden
Some elements of bolted joints are not reusable
Bolted joints are used to create assemblies that resist:

(i) Tensile loads
(ii) Moments
(iii) Shear loads
Bolts are NOT meant to resist (i) (iii)
Components
Anatomy of a bolted joint

l
Grip
Ad
At
Tensile stress area
Major diameter area
tw
HBolt
Ad
L
lt
LT
t2
lt
Threaded length in grip
ld
t1
ld
HNut
tw
Unthreaded length in grip
At
Major diameter (unthreaded)
Joint components: Clamped member
Things to consider with the clamped member:

1. Stone or lap the surface (increase
stiffness)
2. Remove burs (increase joint stiffness)
3. Be sure flange surfaces are flat so bolt
does not bend
Bolted joint components: Bolt

Rolled Threads
NEVER in shear or bending
Head
Keep threads clean &

lubed to minimize losses
Shank
Cut Threads
Steel is most
common
Stress concentrations at the

root of the teeth
Fatigue crack propagation!
Exception: Shoulder bolts
~50% power to bolt head

friction
~40% power to thread

friction
~10% power to deforming

the bolt and flange
Bolted joint components: Washers

Purpose of Washers:
Spacer
Distribute load in clamped member
Reduce head-member wear
Lower coefficient of friction/losses
Lock bolt into the joint (lock washer)
Increase preload resolution (wave washer)
Bolted Joint Components: Nut
Threads do not distribute the load evenly:
What can be done to distribute the load better
1. First thread has the shortest load path

(stiffest)
1. Use a softer nut material

2. Use a bolt and nut that have different pitch
plastically
Bolts are values
used once
for with
precision applications
2.Threads
The pitch
of the boltdeform
threads and nut
to begin
threads change as they are loaded
3. Use a special nut that lengthens the load

path of the first thread
9
Stiffness
Preload
While preloading joint, are the flange & bolt springs in parallel or in series?
Series:
Same Forces
Different Displacements (stretches)
Parallel:
Same Displacements (stretches)
Different Forces
Fpreload
= Flange Compression
Kflange
Fpreload
= Bolt Stretch
KBolt
11
Preloaded joint modeled as series spring
km
kb
Need to find equivalent bolt and member stiffness

12
Bolt stiffness
Shoulder bolt/cap screw consists
of two different parts
Ad E
kd =
ld
At E
kt = *
lt
kd
kt
Threaded
Unthreaded
Each has different
Cross sectional area

Axial stiffness
The load passes through both
The effective threaded

grip length, lt*, used in the
stiffness calc is the sum of
the threaded grip length
plus three threads
They act in series

This is a series spring calculation
1
1 1
kt k d
kb = + =
kt + k d
kt k d
13
Member stiffness
dw
Pressure cone exists in the

member materials and bolt head
The clamping area at the member
interfaces depends upon
dh
25o-45o
P dz
d =
E A(z )
2
2
dw dh
A ( z ) = z tan ( ) +
2
2
Washer diameter, dw
Half-apex angle,
Bore clearance, dh
Stiffness calculation by integration

through the depth of the member
km =
Ed tan ( )
( d w d h + 2t tan ( ) ) ( d w + d h )
ln
d
d
2
t
tan
d
d
+
+
( ) ) ( w h )
( w
h
14
Loading
Tensile loads in bolted joints

Fi
Preload
External tensile load
Portion of P taken by bolt
P
tw
HBolt
Pb
L
Pm
t2
lt
Portion of P taken by members
LT
Fraction of P carried by bolt
Fraction of P carried by members
t1
ld
HNut
tw
1-C
16
Forces in the bolt and the members

When loaded with a tensile force
Most of the force is taken by the members

Very little (<15%) of the force is taken by the bolt
For most, this is counter intuitive.
P
km
kb
17
Forces in the bolt and the members

So how much does each see?
km
Pm = Portion of P taken by members

Pb = Portion of P taken by bolt
kb
kb
Pb =
P=PC
k m + kb
P = Pm + Pb
Pm = P (1 C )
Pb Pm
= =
kb k m
Fb = Pb + Fi
Fm = Pm Fi
High preload = High load capacity
P,
Fb = CP + Fi
Fm = (1 C )P Fi
What happens when joint separates?

18
Static load capacity

Typically the bolt fails first, why?
CP Fi
+
b =
At At
It is the least expensive

It is the most easily replaced
Proof load and stress
Sp = proof stress = Limiting value of b (~ 0.85 y)
Load factor (like a factor of safety)
C n P Fi
+ = Sp
At
At
n=
n > 1 ensures b < Sp
S p At Fi
CP
How high should the pre-load be?
Non-permanent: Some suggest 0.75 Fp

Permanent:
Some suggest 0.90 Fp
19
Shear resistance
When joint is in shear
Friction between the members takes the load, not the bolt
Coefficient of friction and preload are the important properties
Dowel pins or shoulder bolts should be used to resist shear
P = s Fi
20
Torque,
friction,
preload
Bolt torque and preload

How to measure
Via stretch = but impractical

Via strain = expensive built-in bolt sensor
Via torque = not ultra-repeatable but easy and most often used
Relationship between Torque and Stretch?
ETorque = E friction + Estretch

How much do you torque the bolt when tightening?
Too little = weak, compliant joint

Too much = bolt may break or the joint may bulge
Usually torque the bolt until Proof Load is reached
Continuous tightening is important:

s > k
22
Best practices
Best practices
D
H >2D
Threads should be at
least 1.5 D deep for
bolt to reliably hold a
load
24
Applications: Gasket / roller bearings

Gasket
Roller Bearings
Wave washers can reduce

tightening sensitivity to
achieve desired preload.
25
Exercise
Group exercise
The tool holder stiffness is critical to lathe accuracy.
Calculate the stiffness of the bolted joint between your

tool holder and cross slide bearing.
How does the relative stiffness of this compare with the

stiffness of other parts in the load path?
Structure
Bearings
Rails
Etc
27
Preventing Bolts from Coming Loose

How do you prevent bolts from coming loose?
1. Use the joint in a low vibration environment
2. Use bolts with fine threads (small pitch)
3. Use a large preload
4. Use materials with high coefficients of friction
5. Use Loctite on the threads
6. Use an adhesive between the bolt head and flange
7. Use lock washers
N
F
F = s*N
28
Applications: Bearing Rails
Rails
Carriage
Tighten bolts sequentially
1.5 L
L
29
Applications: Bearing Rails

Objectives:
Maximize stiffness
Decrease manufacturing cost
Maximize accuracy
Accuracy is maximized by
overlapping strain cones.
Therefore, the thicker the rail,
the few bolts are necessary.
But the rail becomes less stiff.
Same stiffness
Beware of bulging
High manufacturing
cost
Bolt spacing should be about 4x the
bolt diameter
30
MIT OpenCourseWare
http://ocw.mit.edu

Spring 2009
2.72
Elements of
Mechanical Design
Lecture 11:
Screw drives

Quiz
Bolted joint qualifying Thursday March 19th
Topics
Screw drive group activity - 90% hands on
Reading assignment
Read:
13.1 13.7
17.1 & 17.3
Skim:
Rest of Ch. 17
Screws
Convert rotary motion into linear motion:
Types of lead screws:
Sliding contact thread lead screws

Ball screws
Hydrostatic lead screws
Others
Typical screw types
Image by jgelens on Flickr.

http://news.thomasnet.com/images/large/455/455175.jpg
http://www.danahermotion.com/website/com/eng/img/product/LeadScrew2Nut.jpg
Errors
Some error sources
Misalignment
Bearings, Carriage
Geometry
Straightness
Varying pitch diameter (periodic error and backlash)
Errors in thread-contacting elements
Loads
Elastic
Vibration
Active error management
Periodic errors can be mapped

Linear position sensors
Preload why
Nut-screw:
Backlash
Use two nuts that are preloaded against each other
Use oversize rolling elements
Use a split-circumferentially clamped nut
Screw-machine:
Backlash
Buckling
Straightness
Vibration
Cross feed design exercise

Kinematic (trade off)
Loads and power (limits)
Constraints (bearing, flexure?)
Preload (Nut-screw and screw-machine)
Stress/fatigue
Errors (Causes, systematic, random)
MIT OpenCourseWare
http://ocw.mit.edu

Spring 2009
2.72
Elements of
Mechanical Design
Lecture 12:
Belt, friction, gear drives

Quiz
Bolted joint qualifying Thursday March 19th
Topics
Belts
Friction drives
Gear kinematics
Reading assignment
Read:
14.1 14.7
Skim:
Rest of Ch. 14
Topic 1:
Belt Drives
Belt Drives
Why Belts?
Torque/speed conversion
Cheap, easy to design
Easy maintenance
Elasticity can provide damping, shock absorption
Image by dtwright on Flickr.
Keep in mind
Speeds generally 2500-6500 ft/min
Performance decreases with age

Please see:
Image by v6stang on Flickr.
http://www.tejasthumpcycles.com/Parts/primaryclutch/3.35-inch-harley-Street-Belt-Drive.jpg
http://www.al-jazirah.com.sa/cars/topics/serpentine_belt.jpg
Belt Construction and Profiles

Many flavors
Flat is cheapest, natural clutch

Vee allows higher torques
Synchronous for timing
Usually composite structure
Rubber/synthetic surface for friction

Steel cords for tensile strength
Belt Drive Geometry

Driven
Pulley
Slack Side
d2
1
d1
Driving
Pulley
vbelt
Tight Side
Belt Drive Geometry
2
1
dspan
dcenter
Contact Angle Geometry

2
d2
1
1 d1
dspan
dcenter
d2 d1
1 = 2sin
2dcenter
1
d2 d1
2 = +2sin
2dcenter
1
Belt Geometry
2
d2
1
1 d1
dspan
dcenter
d2 d1
dspan = d
2
center
Lbelt = 4d
2
center
(d2 d1 ) + 1 (d11 +d22 )

2
2
Drive Kinematics
2
d2
1
1 d1
dspan
dcenter
d1
d2
vb = 1 = 2
2
2
d1 2
=
d 2 1
10
Elastomechanics
Elastomechanics torque transmission
Kinematics speed transmission
Link belt preload to torque transmission
Proceeding analysis is for flat/round belt

Driven
Pulley
Slack Side
d2
1
d1
Driving
Pulley
vbelt
Tight Side
11
Free Body Diagram

y
x
dS
F
dN
F+dF
dN
d/2
Tensile force (F)

Normal force (N)
Friction force (N)
Centrifugal force (S)
12
Force Balance
y
dS
x
F
Using small angle approx:
dN
F+dF
dN
d/2
d
d
Fy = 0 = (F + dF ) F
+ dN + dS
2
2
Fd = dN + dS
Fx = 0 = dN F + (F + dF )
dN = dF
13
Obtaining Differential Eq
y
dS
x
F
Let m be belt mass/unit length
dN
F+dF
dN
d/2
d 2
dS = m d
2
Combining these red eqns:
2
d
dF = Fd m 2 d
2
2
dF
d
F = m 2
d
2
14
Belt Tension to Torque

Let the difference in tension between the loose side (F2) and the
tight side (F1) be related to torque (T)
T
F1 F2 =
d
2
F1
Solve the previous integral over contact angle and apply F1 and
F2 as b.c.s and then do a page of algebra:
contact
Te
+1
Ftension = contact
de
1
2e contact
d 2
F1 = m + Ftension contact
e
+1
2
F2
2
d 2
F2 = m + Ftension contact
e
+1
2
Used to find stresses

in belt!!!
15
Practical Design Issues

Pulley/Sheave profile
Which is right?
Manufacturer lifetime eqs
Belt Creep (loss of load capacity)

Lifetime in cycles
Idler Pulley Design
Catenary eqs deflection to tension

Large systems need more than 1
Idler
IDL
ALT
P/S
Water
pump
& fan
Idler
Crank
Images by v6stang on Flickr.
16
Practice problem
Delta 15-231 Drill Press
1725 RPM Motor (3/4 hp)
450 to 4700 RPM operation
Assume 0.3 m shaft separation
What is max torque at drill bit?
What size belt?
Roughly what tension?

http://www.rockler.com/rso_images/Delta/15-231-01-500.jpg
17
Topic 2:
Friction Drives
18
Friction Drives
Why Friction Drives?
Linear Rotary Motion

Low backlash/deadband
Can be nm-resolution

http://www.beachrobot.com/images/bata-football.jpg
http://www.borbollametrology.com/PRODUCTOS1/Wenzel/
WENZELHorizontal-ArmCMMRSPlus-RSDPlus_files/rsplus.jpg
Keep in mind
Preload bearing selection

Low stiffness and damping
Needs to be clean
Low drive force
19
Friction Drive Anatomy

Motor and
Transmission/Coupling
Drive Roller
Drive Bar
Concerned with:
Linear Resolution
Output Force
Max Roller Preload
Backup
Rollers
Axial Stiffness
20
Drive Kinematics/Force Output

Kinematics found from no slip cylinder on flat
bar
d wheel
=
2
dwheel
vbar
d wheel
= wheel
Force output found from static analysis
Either motor or friction limited
Foutput
2Twheel
=
d wheel
where Foutput Fpreload
21
Maximum Preload
1
1
Ee =
+
Ebar
Ewheel
2
wheel
2
bar
Variable Definitions
1
1
+
Re =
d
r
wheel
crown
2
3Fpreload Re
acontact =
2Ee
For metals:
max =
3 y
2
1
3
Shear Stress Equation
acontactEe 1 + 2 wheel 2
wheel =
+ (1 + wheel ) 2(1 + wheel )
2Re
2
9
Fpreload , max =
3
Re
2
16 3 max
1+ 2 wheel 2
3E
+ (1+ wheel )
2(1+ wheel )
2
9
2
e
22
Axial Stiffness
1
1
1
1
kaxial =
+
+
+
kshaft ktorsion ktangential kbar
2
d
wheel
4ae Ee
ktangential =
(2 )(1+ )
k shaft =
k torsion =
k bar =
3 Ed
4
shaft
3
4L
4
Gd wheel
32 L
EA c , bar
L
23
Friction Drives
Proper Design leads to
Pure radial bearing loads
Axial drive bar motion only
Drive performance linked to motor/transmission
Torque ripple
Angular resolution

http://www.borbollametrology.com/PRODUCTOS1/Wenzel/WENZELHorizontal-ArmCMMRSPlus-RSDPlus_files/rsplus.jpg
24
Topic 3:
Gear Kinematics
25
Gear Drives
Why Gears?
Torque/speed conversion
Can transfer large torques
Can run at low speeds
Large reductions in small package
Image from robbie1 on Flickr.
Keep in mind
Requires careful design

Attention to tooth loads, profile
Image from jbardinphoto on Flickr.

http://elecon.nlihost.com/img/gear-train-backlash-and-contact-pattern-checking.jpg
http://www.cydgears.com.cn/products/Planetarygeartrain/planetarygeartrain.jpg
26
Gear Types and Purposes

Spur Gears
Parallel shafts
Simple shape easy design, low $$$
Tooth shape errors noise
No thrust loads from tooth engagement
Helical Gears
Gradual tooth engagement low noise

Shafts may or may not be parallel
Thrust loads from teeth reaction forces
Tooth-tooth contact pushes gears apart
27
Gear Types and Purposes

Bevel Gears
Connect two intersecting shafts
Straight or helical teeth
Worm Gears
Low transmission ratios
Pinion is typically input (Why?)
Teeth sliding high friction losses
Rack and Pinion
Rotary Linear motion

Helical or straight rack teeth
Pinion, m2
F (t)
k
k
Rack, m1
Viscous damping, c
Rack & Pinion
28
Tooth Profile Impacts Kinematics

Want constant speed output
Conjugate action = constant angular velocity ratio

Key to conjugate action is to use an involute tooth profile
Output speed of gear train
Real involute/gear
Ideal involute/gear
out, [rpm]
Non or poor involute
time [sec]
29
Instantaneous Velocity and Pitch

Model as rolling cylinders (no slip condition):
v v v v v
v = 1 r1 = 2 r2
1 r2
=
2 r1
Model gears as two pitch circles
Contact at pitch point

Pitch Circles Meet @ Pitch Pt.
r1
r2
r1
r2
1
v
30
Instantaneous Velocity and Pitch

Meshing gears must have same pitch
-Ng = # of teeth, Dp = Pitch circle diameter
Diametral pitch, PD:
Circular pitch, PC:
PD =
PC =
Ng
Dp
D p
Ng
=
PD
31
Drawing the Involute Profile

Gear is specified by
diametral pitch and
pressure angle,
Pitch Point
Pitch Circle
Base Circle
DB/2
DP/2
http://upload.wikimedia.org/wikipedia/commons/c/c2/Involute_wheel.gif
DB = DP cos
32
Drawing the Involute Profile

Pitch Point
Pitch Circle
L3
L2
3
DB
Ln = n
2
2
1
L1
Base Circle
DB/2
33
Transmission Ratio for Serial Gears

Gear train
Power in: Tin y in
Transmission ratio for elements in series:
From pitch equation: P1 =
Power out: Tout y out
out
TR = ( proper sign )
in
N1 N 2
=
= P2
D1 D 2
D1 N1 2
=
=
D 2 N 2 1
11
For Large Serial Drive Trains:
Productof drivingteeth
TR = ( proper sign)
Productof driven teeth
34

Serial trains:
Example 1:
Product of driving teeth

Product of driven teeth
TR = ?
in
out
Example 2:
driven
drive
driven
drive
TR = ?
driven
drive
in
out
35

Example 3: Integral gears in serial gear trains
What is TR? Gear 1 = input and 5 = output
Product of driving teeth

Product of driven teeth
Gear - 1
N1 = 9
5
4
Gear - 2
N2 = 38
Gear - 3
N3 = 9
Gear - 4
N4 = 67
Gear - 5
N5 = 33
36
Planetary Gear Trains

Planetary gear trains are very common
Very small/large TRs in a compact mechanism
Terminology:
Ring
gear
Planet
Arm
Planet
gear
Planet
gear
Planet
Arm
Sun
gear
Planet
gear
37
Planetary Gear Train Animation

How do we find the transmission
ratio?
http://www.cydgears.com.cn/products/Planetarygeartrain/
planetarygeartrain.jpg
Train 1
Ring
gear
Arm
Planet
gear
Sun
Train 2
Ring
gear
Arm
Planet
gear
Sun
38
Planetary Gear Train TR

Sun Gear
Planet Gear
If we make the arm

stationary, than this is
a serial gear train:
Ring Gear
ra ring arm
=
= TR
sa sun arm
N sun N planet
N sun
TR =
N planet N ring
N ring
Arm
pa
sa
planet arm
sun arm
= TR
N sun
TR =
N planet
39
Planetary Gear Train Example

Sun Gear
Planet Gear
Ring Gear
If the sun gear is the

input, and the ring
gear is held fixed:
0 arm
ra
=
= TR
sa sun arm
Arm
N sun N planet
N sun
TR =
=
N planet N ring
N ring
TR
output = arm =
sun
TR 1
40
Case Study: Cordless Screwdriver

Given: Shaft TSH (SH) find motor TM (SH)
Geometry dominates relative speed (Relationship due to TR)
2 Unknowns: TM and M with 2 Equations:
Transmission ratio links input and output speeds

Energy balance links speeds and torques
41
Example: DC Motor shaft
T( ) = TS 1
NL
P() obtained from P() = T() y
T():
Motor torque-speed curve
T()
( 0 , TS )
Speed at maximum power output:

(NL , 0 )
P( ) = T( ) = TS
NL
PMAX =
PMAX
NL
2
NL
= TS
P(
)
Motor power curve
PMAX
(NL , 0 )
PMAX
42
Example: Screw driver shaft

A = Motor shaft torque-speed curve
What is the torque-speed curve for the screw driver?
T()
B
Train ratio = 1/81
A
C
SCREW DRIVER SHAFT
TSH, SH
MOTOR SHAFT
TM, M
Electric
Motor
GT-1
GT-2
Screw Driver Shaft
System boundary
GEAR train # 1
GEAR train # 2
43
Example: Screw driver shaft

P()
C = Motor shaft power curve

What is the power-speed curve for the screw driver?
Train ratio = 1/81
C
E
SCREW DRIVER SHAFT
TSH, SH
MOTOR SHAFT
TM, M
Electric
Motor
GT-1
GT-2
Screw Driver Shaft
System boundary
GEAR train # 1
GEAR train # 2
44
MIT OpenCourseWare
http://ocw.mit.edu

Spring 2009
2.72
Elements of
Mechanical Design
Lecture 13:
Gear failure prevention

Quiz
Activity at end: Optional/Extra credit

Please see images of very large and very small gears, such as:
Topics
http://mems.sandia.gov/gallery/images_gears_and_transmissions.html
http://www.cage-gear.com/large_gear_cutting.htm
Gear lifetime/selection
Reading assignment
None!
Selection vs. design of gears

It is rare to custom DESIGN a gear.
Many gear selection programs
Anybody can read S/M and plug in #s
BASIC considerations to select gears:
Ensure geometric compatibility (e.g. equal pitch and same type)
Avoid low-cycle failure (e.g. root stress)
Avoid high-cycle failure (e.g. pitting)
Focus on what is important

A failure
How to model the gear teeth

W
Wr
Wt
Wt
rf
t
a
x
Gear
manufacturing
Gear manufacturing - Hobbing
Please see digtos. "GEAR HOBBING M20/32 CITIZEN CINCOM."

February 17, 2008. YouTube. Accessed October 26, 2009. http://www.youtube.com/watch?v=fR2duvm3lPo
Gear manufacturing - Shaping
Please see rolvon. "Gear Cutting." May 16, 2008. YouTube. Accessed October 26, 2009. http://www.youtube.com/watch?v=xF9CjluRFJ4
Selection vs. design of gears

Why do we care about gear tooth surface finish
What affects the finish on the gear surfaces?
How good could it be?
How much would it cost?
Why do we care about the tooth geometry at the root
What affects the quality of the fillet at the root?
How good could it be?
How much would it cost?
10
Perspective
Failure modes
Tooth bending/shear
Contact failure
Science modeling
Engineering modeling
American Gear Manufacturers Association (AGMA)
Example
Single pressure angle
Full-depth teeth
Others
11
Calculating stresses
bending
Pd K m K B
= Wt K o K v K s
(U.S. units)
F J
1.9
Dynamic factor, kv
1.8
Qv = 5
1.7
Qv = 6
1.6
Qv = 7
1.5
1.4
Qv = 8
Qv = 9
1.3
Qv = 10
1.2
Qv = 11
1.1
1.0
"Very accurate gearing"

0
2000
4000
6000
8000
10,000
Pitch line velocity, Vt, ft/min

contact = C p
Km C f
(Wt K o K v K s )
(U.S. units)
dp F I
Incredibly uninteresting, plug-chug & non-scientific

12
Gear failure
at the root
Bending
Basic stress calculation

Stress near the tooth root, model tooth as a cantilever
Mc
=
I
Wt P
=
FY
6 Wt L
=
F t2
Root
P: Diametral pitch
Y: Lewis form factor
~ to for = 20o
f( # of teeth )
Conservative:
Implies that one tooth carries the load

Heaviest load occurs mid-tooth
14
Basic stress calculation

Stress near the tooth root, model tooth as a cantilever
M c
=
I
Wt P
=
FY
6 Wt L
=
F t2
Wt
Wr
Wt
P: Diametral pitch
Y: Lewis form factor
~ to for = 20o
f( # of teeth )
rf
a
x
t


http://www.oilanalysis.com/Backup/200101/Gear3.jpg
Conservative:
Implies that one tooth carries the load

Heaviest load occurs mid-tooth
15
Dynamic effects
How to incorporate dynamic effects
One way of addressing
a + V
K
v =
V = pitch line velocity
Kv depends on fab
For rough estimates
Wt P
= KV
FY
This is for English units, for SI is different
16
Allowable bending stress
Allowable bending stress number, St kpsi
These types of plots are associated with conditions
50
Grade 2
St = 102 HB + 16 400 psi
40
Grade 1
St = 77.3 HB + 12 800 psi
30
St = t . Hb + Ct
20
10
150
200
250
300
350
Brinell hardness, HB
400
450
17

all
S t YN
=
(U.S. units)
S F KT K R
all
St YN
=
(SI units)
S F Y YZ
Elements of the equations:
St
YN
KT
KR
SF
Stress cycle life factor
Temperature factors
Reliability factors
AGMA factor of safety
Allowable stresses for:
Unidirectional loading
10 million stress cycles
99 percent reliability
18
Gear failure
at the surface
Fatigue
High cycle failure: Pitting

Images removed due to copyright restrictions. Please see any photos of surface pitting in gears, such as:
http://2.bp.blogspot.com/_tBh5ORa6LOk/R8UaDR3pgVI/AAAAAAAAAGE/DikmlvWPS84/s1600-h/pitting.gif
http://commons.wikimedia.org/wiki/File:Roue_creuse_03.jpg
20
Avoiding high cycle failure: Stress variables

Equivalent modulus
1
Ee =
2
2
1 v1 1 v21
+
E1
E2
Half contact width
2Wt d1 d 2
b=
L Ee (d1 + d 2 )
= 0.333
Watch out! The book

switches meaning of F here
Maximum contact pressure
2Wt
q=
bL
Image from Wikimedia Commons, http://commons.wikimedia.org
21
Allowable contact stress [ANSI/AGMA 2001-D04 and 2101-D04]
Allowable contact stress number, Sc
1000 lb/in2
S c = c H b + Cc
175
Grade 2
Sc = 349 HB + 34 300 psi
150
125
Grade 1
Sc = 322 HB + 29 100 psi
100
75
150
200
250
300
350
400
450
Brinell hardness, HB
22
Allowable contact stress

c,all
SC Z N C H
=
(U .S . units )
S H KT K R
c,all
S C Z N ZW
=
(SI units)
S H Y YZ
Elements of the equations:
SC
ZN
CH
KT
KR
SH
Allowable contact stress
Stress cycle life factor
Hardness ratio factors for pitting resistance
Temperature factors
Reliability factors
AGMA factor of safety
Allowable stresses for:
Unidirectional loading
10 million stress cycles
99 percent reliability
23
Exercise
Gears
General machine design
Activity: Refit lathes for the mfg. shop

Study the lathes in the shop
1. What types of failures do we have?
2. Calcs/sims/tests need to augment Shigley/Mischke:
Gearing
Belts
Friction elements
3. Worst case consequence of these kinds of failures:
Pitting failure
Tool break/failure
Fatigure
25
Exercise
Windmill gear boxes
Windmill failures: Catastrophic
Please see mrturbodk. "windmill failure." February 28, 2008. YouTube. Accessed
October 26, 2009.
http://www.youtube.com/watch?v=TmM3KQnFmXs
27
Wind energy overview: Lakawona

http://i124.photobucket.com/albums/p7/NBBooks/WTGTurbinesGettingLargerSM.jpg
http://i124.photobucket.com/albums/p7/NBBooks/WTGUSWindResources.jpg
28
Images removed due to copyright restrictions. Please see p. 19 in

http://www.clipperwind.com/pdf/liberty_brochure.pdf
29
Activity: Extra credit - As a group

You are tasked to build a windfarm off Cape Cod
1. Shigley/Mischke is not perfectly suited to cover gear
needs in this application. Why/how?
2. What calculations/simulations/tests would you do to
augment Shigley/Mischke?
3. What happens if you have pitting failure and what
would you do about it?
4. What happens if you have failure at a tooth root and
what would you do about it?
30
MIT OpenCourseWare
http://ocw.mit.edu

Spring 2009
Reading and plans

Shigley-Mischke sections
None
Today: Actuators: Hydraulic and Electromagnetic
Energy transfer and scale

Hydraulic / fluidic
DC Permanent magnet motors
Perhaps. wrap up of MEMS
Friction-based machines
Purpose:
Do work at a given rate, Energy - Power
Physics: Energy and mass conservation/balances
Characteristics of import
Load
Speed
Bandwidth
Cost
And there can be other issues of import

http://www.onefunsite.com/images/donkey.jpg
Consequences
Please see trigirl. Crane Drops Steamroller on Car! May 8, 2007. LiveVideo. Accessed November 25, 2009.
http://www.livevideo.com/video/16A18C6512B945C29547A8658E890AF1/crane-drops-steamroller-on-car.aspx
An unpleasant (I hope) example
Common actuators for mechanical systems

Biological
Pneumatic/Hydraulic
Electromagnetic
Electrostatic
Piezo
Thermal
Biological
People powered machines

Energy
Power
Load
Speed
Bandwidth
Why is it important to understand what humans can do?

Hydraulics
Basic principles
Sub-system design
Pump
Motor
Output shaft
2=?
Tp = ?
Dp = in3/rev
Dm = 2 in3/rev
p = 100 rpm
Apiston = 1 in2
2 = 100 rpm
Nm = 40
vp = ?
N2 = ?
p1
Pump
p2
p3
Motor
p4
Flow direction
Reservoir
10
Examples: Real but practical ;) ?

http://darkdiamond.net/wp-content/uploads/2006/08/115638952415_hugegundam1.JPG
http://gizmodo.com/
11
Other less than practical examples
Please see HydraulicGuitar. Hydraulic Guitar. September 10, 2006. YouTube.

Accessed November 25, 2009. http://www.youtube.com/watch?v=Elt1XriaQXU
12
Please see any video of a hydraulic low rider assembly.
13
Please see arefadib. The Flying Steamroller. October 17, 2006. YouTube.
Accessed November 25, 2009. http://www.youtube.com/watch?v=sKGRRIiR5xA
14
Hydraulic systems in machines

Advantage:
High force/torque and routing of power
Disadvantage:
Leaking and wear due to contaminants
Liquids & gases in fluid-based machinery
Hydraulics:
Pneumatics:
Fluid is a liquid
Fluid is a gas
15
Example: Piston pump doing work

Hydraulic machines can be used to do work
Load on the system extracts energy from the liquid

Pressure increases between the input and output components
Pressure is used to do work on internal parts of hydraulic devices
Power is input/extracted via shaft (motor) or rod (cylinder)
Power out
0
Motor
Cylinder
Pressure gauge
16
Volume flow rate and displacement

Displacement (D)
Displacement = volume of fluid moved / cycle

Cycle = rotation (drill pump) or stroke (cylinder)
Q = volume moved per unit time
D
Volume
Cycle(s)
f
Cycles
*
second
=
=
Q
Volume
second
F is the frequency of a machines cycle

o For hydraulic pumps, f = speed of the shaft = /(2)
o For hydraulic motors, f = speed of the shaft = /(2)
o For cylinders,
f = strokes/second
[rad/s]
[rad/s]
f [Hertz]
17
Displacement: Physical example

D = volume pumped per cycle
1 Cycle = expansion + contraction
rinitial
r final
Vinitial
Vinitial
4
= (rinitial )3
3
Vfinal =
Vfinal
Displacement = D = Vinitial Vfinal =
4
r final
3
)3
4
(rinitial )3 r final
3
)3 ]
18
Incompressibility
Incompressible fluid:
Compressible fluid:
19
Why is incompressibility important?

Mass balances
The mass density (m) of fluids changes with pressure (p)
Compressible fluids: exhibit large (m) for small (p)
If (m) is large, it is possible to store significant mass in a machine
This complicates our analysis
m in = m out
d
+ mstored
dt
20
Why is incompressibility important?

Energy balances
All fluids store energy when compressed (similar to a spring)
Compressible fluids store A LOT of energy (think balloons!!)
Stored energy complicates analysis (calculating can be difficult)
Ein = [Eout ] + Estored
21
Example: Locked piston positions

Incompressible fluid:
Compressible fluid:
22
Incompressibility
Bulk modulus: Measures of resistance to volume
=
dp
dV
initial
For small (incremental) changes in volume :
p
V
V
initial
Example: Fluid in tube exposed to pressure increase

V
V final
Vinitial
23
Incompressibility
Hydraulics, pneumatics and incompressibility
Low , usually compressible

High , usually incompressible
Pneumatics = gas:
Hydraulics = liquid:
What makes a good assumption?
Depends on the error you are willing to live with
Example: Incompressibility of water (e.g. H2O)

H 2O = 2.2 109
N
m2
= 3.2 105
(V V ) p
lbf
in 2
Example for H 2O where p = 2500 psi, V
= 0.006 = 0.6%
Is this OK?
24
Volume flow rate, Q

Link between mass flow rate & volume flow rate:
Q = time rate of volume flow through a hydraulic system
From mass conservation

From 8.01
Qi =
m i
mi
m in
m in
mi Ai v i
mi
m out
mout
d
(Vstored ) ~ 0
dt
d
mstored
d
+ dt
Qin = Qout + (Vstored )
dt
mstored
Mass densities are equal and cancel out of equation if fluid is incompressible
For incompressible flow in a pipe : Ai n vi n = Qi n = Qout = Aout v out
25
Vane pumps
Series of vanes extending radially from rotating core
Vanes can slide in/out or deform depending upon design
How a sliding vane pump works:

Step 1: Fluid enters when volume between vanes is increasing
Step 2: Fluid travels when volume between vanes does not change
Step 3: Fluid exits at when volume between vanes is decreasing

http://pumpschool.com/images/vnsteps.gif
credit: pumpschool.com
26
Pump types: Piston

How it works:
Step 1: Piston forces fluid out during initial stroke

Step 2: Valves change fluid path (only allows flow into pump)
Step 3: Piston recharged with fluid, cycle starts again
What is the displacement?
27
Pump types: Piston

How it works:
Step 1: Piston forces fluid out during initial stroke

Step 2: Valves change fluid path (only allows flow into pump)
Step 3: Piston recharged with fluid, cycle starts again

http://www.animatedsoftware.com/pics/pumps/wobble.gif
http://www.flexicad.com/bilder/Rhino_Galerie/Kolpenpumpe.jpg
www.animatedsoftware.com/pumpglos/wobble.htm
http://gallery.mcneel.com/fullsize/11155.jpg
28
Pump types: External gear pump

Only one gear is driven, the other spins free
Which way does the flow go?
Step 1: Fluid comes in at ?

Step 2: Fluid travels through ?
Step 3: Fluid exits at ?
What is the displacement?
29
Pump types: External gear pump
Bushing Bearings
O-Ring
a) Pump Body
a) Gear and Shaft
30
Hydraulics
Exercise
Competition: Pump
Form group
In 10 minutes, make best estimate of gear pump displacement

Hand in answer/analysis at end of exercise (with all names)
Sketches, calculations, etc must be handed in before bell sounds
91
R10
10
H1
R5
H2
10
50
10
R10
13 5
32
H7
13
5
H5
R14
70
R14
H3
H4
26
10
H6
12
10
Gear Pump Cavity Plate with Dimensions

All Dimensions in mm
10
50
28
14
R10
10
6
8
Minutes
32
Hydraulics
power
Power example: Pump at steady state

Tshaft
Ain
pin
Aout
Pump
xin
pout
xout
Ai n vi n = Qi n = Qout = Aout v out

34
Example: Pump at steady state

Work done on
fluid via shaft input
Pressure force does work
on fluid entering pump
Tshaft
Ain
pin
Aout
Pump
vin
Pump does work on

Fluid exiting pump
pout
v out
35
Example: Pump at steady state

Pin = [Pout ] +
d
(Estored ) Pinlet + Pshaft = [Poutlet + Ploss ] + d (Estored )
dt
dt
Pinlet = [Fin ] vin = [ pi n Ai n ] vin
Tshaft
Pshaft = Tshaft shaft

Poutlet = [Fout ] v out = [ pout Aout ] v out
Ain
pin
Aout
Pump
pout
If can assume A&B, Ploss & d(E)/dt can be neglected
A.
B.
d
(Estored ) <<< Pin Pout
dt
vin
v out
Ploss <<< Pin Pout
Substituting into energy balance (top equation on sheet)
[ pin Ain ] vin + Tshaft shaft = [ [ pout Aout ] vout + ~ 0]+ ~ 0
36

[ pin Ain ] vin + Tshaft shaft = [ [ pout Aout ] vout + ~ 0]+ ~ 0
Ai n vi n = Qi n = Qout = Aout v out
Qi n = Qout = Q
Tshaft
[ pout pin ] Q = Tshaft shaft

Ain
pin
Pump
vin
Aout
pout
v out
37
Hydraulics
System example

Dm = 0.5 in3/rev
Dp = 0.5 in /rev
Tm = ?
Tp = 10 in-lbf
m = ?
p = 1000 rpm
2
p1
p2
Pump
p1 = 10 psi
p3
Motor
p2 = ? psi
p3 = 10 psi
Density =
Reservoir
39
DC permanent
magnet motors
DC Permanent magnet motor

T vs. for Black & Decker Screw Driver
4.0
T [N-m] = - 0.012 [N-m/rpm] * + 3.678 [N-m]
Torque [N-m]
3.0
2.0
1.0
0.0
0
50
100
150
200
250
300
350
[ rpm ]

T( ) = TStall 1
NL
rpm
Data
310
210
190
167
148
121
67
33
0
T
Nm
Data
0.00
0.90
1.08
1.47
2.04
2.30
2.92
3.32
3.60
T
Nm
Fitted
-0.09
1.11
1.34
1.63
1.85
2.17
2.83
3.23
3.63
% Error
N/A
23
24
10
-9
-6
-3
-3
1
41
Understanding the model

y
Simple 1 loop model

x
T,
r
FE
1
V1
V4
Low loss in wires

Steady state
Single loop
No ferrous cores
Snap shot with loop plan in
the y-z plane
FB
i
Assumptions
FB
Goal: understand trends
FE
42
Point we will study

Points we are studying
D
+TMAX
T=0
-TMAX
Torque curve of simple loop
Side view of simple loop
43
Forces
y
Force on wire
x
T,
r
L points in direction of current flow

L
FE
MAGNETIC FLUX DENSITY, B
FB
V1
CONDUCTOR / WIRE
CURRENT, iwire
FB
i
V4
K
K K
FB = i (L B )
FE
4
MAGNETIC FORCE
Lorentz Force
L
K
K
K K
FE = q E + q v B
44
Torque inducing forces on wire

y
Force on wire
x
T,
r
FB
V1
V4
K
K K
T = 2 (r FB )
90o in
y-z plane
T = 2 r (i L B ) sin ( r F )
T = 2r
L
4
Torque at = 0
FB
i
K
K K
FB = i (L B )
(V1 V4 ) Battery
(V1 V4 ) Battery
LB
T R
=
2r L B
45
dV
E=
dx
Lorentz force
y
Force
K dueK to E & B K
K
FE = q E + q (v B )
T, E = v B sin ( v B )
x
z
r
FE
1
V1
3
i
V4
FE
4
Wire 1-2
(V2 V1 )
L
= (r ) B
Wire 2-3
K K
v B not along r
V2 = V3
Wire 3-4
(V4 V3 )
L
= (r ) B
46
Induced voltage due to rotation

y
V due to rotation,
x
T,
(V2 V1 )
L
r
FE
1
(V4 V3 )
L
= (r ) B
V2 = V3
= (r ) B
V1
3
i
V4
FE
4
(V4 V1 )
L
= 2 (r ) B
(V1 V4 ) = 2 (r ) B L
47
Total voltage
y
V due to rotation,
x
T,
r
FE
FB
V1
3
FB
i
V4
(V1 V4 ) = 2 (r ) B L
V due to battery
Total potential diff.
(V1 V4 ) Battery
T R
=
2r L B
V = (V1 V4 ) Battery + (V1 V4 )

FE
T R
V =
2 (r ) B L
2r L B
48
Torque relationship
y

x
T,
r
FE
V1
3
FB
i
V4
FE
4
Ohms law
V = i R
FB
T R
V =
2 (r ) B L
2r L B

iR =
T R
2 (r ) B L
2r L B
T- relationship
L
4 r 2 L2 B 2
T = 2i L r B
R
49
Torque relationship cont.

y
T- relationship
x
T,
r
FE
V1
V4
T- relationship
FB
i
Stall torque
TStall = 2 i L r B
FB
4 r 2 L2 B 2
T = 2i L r B
T = TStall
FE
4 r 2 L2 B 2
50
Torque relationship cont.

y
T- relationship
x
T,
r
FE
FB
Motor
constant
T vs. for Black & Decker Screw Driver
4.0
FE
4
Tstall
T [N-m] = - 0.012 [N-m/rpm] * + 3.678 [N-m]
T = TStall
3.0
Torque [N-m]
FB
Loop
constant
T- curve
V1
V4
T = TStall
4 r 2 L2 B 2
TStall
NL
2.0
1.0
NL
0.0
0
50
100
150
200
[ rpm ]
250
300
350
51
Scaling
Follow up on
micro-actuator lecture
Electrostatics
How does electrostatic physics scale?
UE =
o L L V
2 z
How does ratio of FElectric scale to FBody?
FElectric 1
~
FBody
L
What does this mean for MuSS interaction?
What happens if you downsize each by factor of 10?
53
Electrostatics
U Electric z =
o L L V
2 z
FElectric z
dU
=
dz
Fbody = V
FElectric z =
o L2 V 2
2 z2
FElectric 1
~
FBody
L
Table removed due to copyright restrictions. Please see

http://www.sizes.com/built/clean_rooms.htm
54
Semi-intuitive example
Cooling
Heating
55
Thermal behavior
How does thermal physics scale (small Bi #)?
h A
t

V c
T Tinf
=
=
inf Tinitial Tinf
h L Convection
Bi =
~
k
Conduction
56
Thermal behavior
How does thermal physics scale?
dT
h A (T Tinf ) = c V
dt
h A
t

V c
hL
Bi =
k
T Tinf
=
=
inf Tinitial Tinf
V c
h A
Is this a good or a bad thing for MEMS actuators?

For the STM?
57
Fluidics
How do fluid-based physical phenomena scale?
r 4 p
Q=
8 L
Q = U r 2
D
L
8 U
p =
L
2
r
High pressure change over narrow flow paths
58
Fluidics
Reynolds number
D
L
Re =
U D
D = 50 m
Ratio of inertial forces to viscous forces

U = 500 m/s
L = 1000 m
ReAir and ReH2O << 1
What does this mean:
Heavily damped
Limits response time (ms vs. s)
59
MIT OpenCourseWare
http://ocw.mit.edu

Spring 2009
2.72
Elements of
Mechanical Design
Lecture 16: Dynamics
and damping

Quiz
None
Topics
Vibration physics
Connection to real world
Activity
Reading assignment
Skim last gear reading assignment (gear selection)
Resonance
Basic Physics

Exchange potential-kinetic energy

Energy transfer with loss
PE
Loss
KE
F(t)
Modeling

2nd order system model

Spring mass damper
Differential equations, Laplace

Transforms
Why Do We Care

Critical to understanding motion of

structures- desired and undesired
Generally not steady state
Location error
Large forces, high fatigue
k
n =
m
Gain
n
Vibrations - Input
Oscillation of System
Why Categorize
Forced
Different causes
Different solutions
Input Form
Forced Steady State

Command Signal
Electrical (60 Hz)
Free Transient
Impacts
t
A
Free
t
Vibrations - Source
Source
Undriven
Disturbance
Spectrum
109
Disturbance
Driven
1013
Device Motors
Rotating Components
1017
102 1
102 104
Undriven
Electrical (60 Hz)
People (2 Hz)
Cars (10 Hz)
Nearby Equipment
Driven Disturbance
Command
Step Response
System
t
Command
Vibration
Response
5
Vibrations: Command vs. Disturbance
Underdamped response
Noise
Free vibration
Golda, D. S., Design of High-Speed, Meso-Scale Nanopositioners

Driven by Electromagnetic Actuators, Ph.D. Thesis, Massachusetts
Institute of Technology, 2008.
Vibrations - Example
Example
Chinook
Identify:
Mode
Form
Source
Response
Forced
Free
109
Undriven Disturbance
Spectrum
1013
1017
102
102 104
Driven Disturbance
A
System
Response
Attenuating Vibrations
Change System
Mass, stiffness, damping

Adjust mode shapes
Change Inputs
Command: Input Modulation, Feedback

Disturbance, reduce:
undriven vibrations, e.g. optical table
driven vibrations alter device structure (damping on motors, etc.)
F
t
1
ms2+bs+k
Modulating Command Vibrations

Change
m, k, c
Input shaping
Behavior
Regimes
(1) Low Frequency (<n)

(2) Resonance (n)
(3) High Frequency (>n)
Example Spring/Mass demo
= 0.05
Amplification factor (output/input)
10
= 0.1
= 0.2
= 0.3
= 0.4
= 0.5
= 0.6
= 0.7
1st order system
= 0.8
= 0.9
= 2.0
= 1.0
-1
-1
/n
10
1
1
x
= 2
=
F ms + bs + k k ( )
10
Behavior Low Frequency

Frequency Response Regimes

(2) Resonance (n)
Low Frequency (<n)
System tracks commands

Ideal operating range
High disturbance rejection
= 0.05
x 1
F k
10
= 0.1
= 0.2
= 0.3
= 0.4
= 0.5
= 0.6
= 0.7
1st order system
= 0.8
= 0.9
= 2.0
Resonance Frequency (n)
System Response >> command

-1
-1
keff
Disturbances will cause very large response
Quality factor = magnitude of peak Figure by MIT OpenCourseWare.
Damping = Q
= 1.0
/n
10
x
1
= 2
F ms + bs + k
High Frequency (n)
System Response << command

11
Behavior - Resonance
10

(2) Resonance (n)
Low Frequency (<n)

x Q
<

k F1 k
keff
Quality factor = magnitude of peak
Damping = Q
= 0.05
= 0.1
= 0.2
= 0.3
= 0.4
= 0.5
= 0.6
= 0.7
1st order system
= 0.8
= 0.9
= 2.0
= 1.0
-1
-1
/n
10
x
1
= 2
F ms + bs + k
High Frequency (n)

12
Behavior High Frequency


(2) Resonance (n)
= 0.05
10
Low Frequency (<n)

= 0.1
= 0.2
= 0.3
= 0.4
= 0.5
= 0.6
= 0.7
1st order system
= 0.8
= 0.9
= 2.0
= 1.0

keff
Quality factor = magnitude of peak
Damping = Q
High Frequency (n)

Poor disturbance rejection
-1
-1
/n
10
x
1
= 2
F ms + bs + k
1
x
F m 2
13
Constitutive Relations
Relevant equations
Damping ratio
Gain
Quality factor
c
=
2 km
k
n =
m
GP =
d = n 1
1
2 1 2
1
<
2
1
Q=
2 (1 2 )
14
Application of Theory
Relate Variables to Actual Parameters
Vibrational Mode
Mass
Stiffness
Damping
Transfer between Model and Reality
Iterative
Start simple (1 mass, 1 spring)
Add complexity
Limits
Example building (video)
x
k
F(t)
Iteration
Mode?
k?
m?
c?
15
Strategies for damping

Material
Pros and cons of each
Grain boundary
Internal lattice
Viscoelastic (elastomers/goo)
Viscous
Air
Fluid
Electromagnetic
Friction
Active
Combinations
Sponge
16
Example: Couette flow relationships

Relevant equations
dx
=
dy
dx
x
F = A = A
= A
dy
h
F = cx
(F, x )
c=
h
17
Exercise (see next page too)

Perform a frequency analysis of the part
Develop & prove (FEA) how to increase nat. freq. via geometry change
Any constraints you might have? Geometry changes cant be unbounded
Explain effect of your change on vibration amplitude (relative to outer
base) at given , via sketches & plots

Xtra credit, assume:

Flexure is contained between two parallel plates (on top and bottom)
Viscous air damping in the gaps on both sides
1 micron gap between the flexure sides and plates
Elaborate on how well flexure is damped (dont just use intuition)

Useful equations (c = damping coefficient, k = stiffness, m = mass)
c
=
2 km
Gain at peak amplitude
Frequency at peak with max gain
Damping ratio
p = n 1 2
Gp=
1
2
4 4
4
18
Flexure
Flexure design
top
Constrained on 4 sides
Bottom
See this diagram for extra credit:
Top plate
Flexure
Bottom plate
19
Multiple Resonances
Gain
10
Center Stage Modes
10
10
Measured
Model
Fit
-2
10
Phase Lag (deg)
Z-Axis
10
Frequency (Hz)
10
0
-100
-200
-300
10
10
Frequency (Hz)
10
20
MIT OpenCourseWare
http://ocw.mit.edu

Spring 2009
2.72
Elements of
Mechanical Design
Lecture 18:
Friction-based elements
Reading and plans

Reading:
14.1 14.7
16.2, 16.6, 16.9
Today:
Friction-based elements
Friction-based machine elements

Purpose:
Bring two bodies to same relative speed
Friction forces do the work
Force/Torque and Mass/Inertia are in play
Used in many types of ubiquitous machines
Drum brakes
Disk brakes
Clutches
etc
Clutches
www.tech.plymouth.ac.uk/sme/desnotes/clutlabe.jpg
Courtesy of David Grieve. Used with permission.
Drum brakes
http://www.carbibles.com/images/drum-double.jpg
http://peugeot.mainspot.net/glossary/14_drum_brake_assembly.jpg
Can you imagine how much

fun it is to put this together
What would be problematic?
Disk brakes
Images courtesy of phaty on flickr and Wikimedia Commons, http://commons.wikimedia.org
What are the holes for?

Perspective on heat, back of the envelope

Assume you need to stop a 2 ton truck: 65 to 0 mph;
how much energy must be dissipated as heat/sound?
Assuming that this happens so fast that the steel

components of the brake absorb the heat (assume
little energy goes into sound/vibration) before it is
taken away by convection or conduction what max
magnitude of T are we looking at? (e.g. worst case)
Really need modeling software to do this properly but
rough analysis is enlightening
Types of actuation
Pneumatic/hydraulic (e.g. pistons)
High force (hydraulic)
Inexpensive (pneumatic)
Maintenance (hardware/leaks & fluid)
Magnetic (e.g. solenoids)
Low maintenance
Fast reaction time
Ease of control
Mechanical (e.g. lever)
Moderate force
Moderate maintenance
The best thing to do is to make

a matrix of FRs and DPs and
then select from that. Vendors
are usually very helpful in filling
out the matrix.
8
Many variations
Purpose is to cover
general fundamentals
so you can extend to
specific cases
Issues in play
Materials
Physics
Friction-based machines
Issues of concern in engineering of these devices:
Force
Torque
Energy loss
Temperature
Practical design/performance criteria:
Torque
Friction
Wear/longevity
Failure criteria:
Maximum temperature
Maximum pressure
There is no cook book formula, you

must KNOW the application and then
prioritize what is important.
11
Assumptions
Pressure:
Distribution, typically assume simple shapes for first order
Relationship to deformation is linear
Contact area vs. actuation area -> rigidity
Relative rigidity of:
Friction material
Independence of:
Backing material
Opposed surface
Sometimes the
vendor or OEM will
have the info you
need, usually geek
engineers have it.
Material properties as function of temperature

Pressure and deformation of friction material
Coefficient of friction & pressure
Coefficient of friction & deformation of friction material
12
Materials
Desire:
High friction
Constant friction
Inert
Wear resistance
Flexibility

http://www.aa1car.com/library/brake_pads.jpg
http://www.aa1car.com/library/brake_dust.jpg
http://www.mooseutilities.com/showImage.jsp?class_id=499&image_type=fullsize&rank=100
Sintered metal
f ~ 0.30 (dry)
Pmax ~ 500 psi
Tmax ~ 930 F
Asbestos composites
f ~ 0.35 (dry)
Pmax ~ 700 psi
Tmax ~ 800ish F
13
Servicing brakes
If you are interested, you can walk through a brake repair at:
http://www.diy-brake-repair.com/how-to-change-brake-pads.html
14
Drum-based
Friction
Elements
Basic model of inner drum brake

Red =
friction material
a = max pressure
1

a
16
Pressure distribution
If we assume:
Shoe and drum rigid relative to material

p ~ Compression
f Temperature or compression
Red =
friction material
Then when pad rotates by , pad:
Compression ( ) = r ( ) sin ( )
Can prove this via math (Shigley) but observation works
17
Pressure distribution
2
So if
Compression ( ) = r ( ) sin ( )
Then
Compression ( )
r ( ) sin ( )
p
( )
=
=
Compression ( a ) r ( ) sin ( a ) p (
a )
And.
p ( )
pa
=
sin ( ) sin ( a )
This is ultra-useful as we need to know the pressure

profile in order to integrate and find moment/torque
18
For small included angle

p ( )
pa
=
sin ( ) sin ( a )
2
F
1 2
19
For large included angle

p ( )
pa
=
sin ( ) sin ( a )
2
F
Better performance
but pad costs $$$!!!
1
20
Where to put
brake material?
Best bang for the buck!
Where to put the brake material
Cost metric?
Torque/$
Envelope metric
Torque/V
Max T metric
Torquemax
Why not just get a KA actuator?
22
Modeling
Behavior
Drum-type example
Total torque applied to DRUM

T = f r dN
f dN
dN
T = f r [ p (b r d )]
c
r - a cos()
pa
T=f r
sin ( ) (b d )
sin ( a )
1
2
cos(1 ) cos( 2 )
T = f b r pa
sin ( a )
2
24
Might it be self energizing? Look at the SHOE

Two moment components
f dN
dN
Shigley switches: Torque to Moment
From normal load

2
pa
2
(
MN = b r a
sin
) d
sin ( a ) 1
c
r - a cos()
When will the friction

torque pull the show
into the drum?
From friction load

2
pa
Mf = f br
sin ( )(r a cos( )) d
sin ( a ) 1
25
So it can be self energizing if

What do we mean?
f dN
Rotation + geometry = self-engage
Trick = geometry condition
How to go about this?
Moments on the shoe
M = F c M f + M N = 0
F c = MN M f F =
dN
c
r - a cos()
MN M f
c
You can change geometry to avoid or leverage this
a=?
c=?
26
Wear
Modeling and estimating wear

Sliding force:
Fs = f p A
Work by sliding force over displacement S
Ws =
(f
p A) (S ) = ( f p A) (v t )
Material volume removed ~ work
w = wear in linear units
w A = K p A (v t ) w = K p (v t )
Analogous to material removal rate in metal cutting
28
Exercise
disc-type
brakes and clutches
Basic model of disc clutch/break
2
1
30
Uniform wear vs. uniform pressure

Why
Uniform wear p v = constant?
Longevity
p ( r ) = pa (a ra )
Uniform pressure?
Performance
p = pa
Calculating torque/moment, general form:
2 ro
T=
r dN
1 ri
31
Activity: Disc clutch

Bossman hears about new constant wear brake
technology and asks you for a back of the envelope
engineering assessment of clutch performance for:
(a) constant pressure and;
(b) constant wear.
routter
=12in
rinner = 6in
You must explain the pros/cons and relay the

implications of making this design change via
qualitative & quantitative means. What do you tell
him? Ratios and analogy might be helpful here
32
MIT OpenCourseWare
http://ocw.mit.edu

Spring 2009
Appendix: Process plan chart template
Process plan
Date approved by:
Step
00
01
02
03
04
05
06
07
08
09
10
11
12
13
14
Group No.:
Part name:
Group members in attendance:
Task & questions
Machine
Date accepted by Culpepper:
Tooling
Measurement

272 Spring 2014 PDF

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

272 Spring 2014 PDF

Uploaded by

Copyright:

Available Formats

MIT OpenCourseWare

2.72 Elements of Mechanical Design

Shigley Mischke on shaft deflections & stress

Martin Culpepper, All rights reserved

Purpose and pace of course

Teaching emphasis, style and grades reflect this

Reading ~ 50% of grade

2/3 semester of lectures

Info sources: Teaching staff & texts

TA: Jon Hopkins

Mechanical Engineering Design (Shigley / Mischke)

Martin Culpepper, All rights reserved

Thoughts, decisions & actions based on understanding:

Models and their associated equations are idealizations

Concepts, principles & design processes necessary, but not sufficient

Martin Culpepper, All rights reserved

2.72 will focus on

understanding concepts, principles, design process, best practices,

You will learn by

Teams of 6 work to model, design, build and characterize one lathe

Martin Culpepper, All rights reserved

Project: Desktop lathe

Martin Culpepper, All rights reserved

Take pictures/video as you go

You must keep a dedicated design notebook

Keep your ideas, calculations, and records in one organized place

No more than 6 pages (not including appendices)

Martin Culpepper, All rights reserved

Take measurements, answer questions and reassemble

Follow shop safety rules

30 minute discussion/design review

50 inch plasma will be available

Everyone must present their part of the project

First tell us the purpose of the meeting

Then immediately discuss Gantt chart

Details of the work to date, calculations

Have back up slides for deep dives

Martin Culpepper, All rights reserved

Parts for the lathe

*May cast 3 pieces as one, stay tuned

Drive preload nut

Finished polymer bed* Finished end skirts*

- 20 bolts 0.50 inch long

- 20 bolts 0.75 inch long

- 20 bolts 1.00 inch long

Martin Culpepper, All rights reserved

Use of ME Mfg. shop

Open M-F, 8am-4pm, Clean up at 3:30pm

2D printed, CAD drawing with dimensions and tolerances (NO sketches)

Martin Culpepper, All rights reserved

In-class and take-home mini-quizzes

All materials are due by 5pm

Via e-mail to teaching assisstant, unless otherwise stated.

Email errors will not excuse late assignments.

You must understand otherwise you are dangerous

No students group may proceed w/o grade > 80% on qualifiers

Martin Culpepper, All rights reserved

Rules for collaboration

What you submit MUST be your own work unless it is specified as

You MUST acknowledge the contribution of others

For example, after working an assignment independently, you compare

Martin Culpepper, All rights reserved