Elements of Machin

HHH
mm HI
MBBfflffl
.m
WlfflfflHI
HSU
MM
m
I
hHHHHBI Hi HH bHrKBH HGH

fflBflfflmB
Bra HH
RHMMMlH
I5H
M,*
Hi HH 1 1HH HH H
wHBm
Mi
BH
MOW
bmmmbm
HB Hkbh
A'
v.
>
^
X*
-7*.
0'
%<
>
u
c^
\^.
^
frV
-
X*
^, v
'OO
x
'
pq.
8
1
^.
c
.0-
>
ELEMENTS
OF
MACHINE DESIGN
BY
DEXTER
Professor
S.
KIMBALL,
Sibley
A.B.
Cornell
University,
of
Machine
Design
and
Construction,
College,
Formerly
Works Manager, Stanley Electric Manufacturing Company. Member of the American Society of Mechanical Engineers.
AND
JOHN
Cornell University.
H. BARR, M.S., M.M.E.

Formerly Professor of Machine Design, Sibley College,
of the American Society of Mechanical Engineers.
Manager, Smith Premier Works.
Member
FIRST EDITION FIRST THOUSAND
NEW YORK
JOHN WILEY & SONS

London:
CHAPMAN
& HALL,
Limited
1909
Copyright 1909
BY
DEXTER
KIMBALL
and
JOHN
H.
BARR
n. K
2465*64
17
SEP
1909
Electrotyped and Printed by Publishers Printine Co..
i\M
New York,
U.
S.
A.
PREFACE
This book
College,
is
the outgrowth of the experience of the authors
in teaching Machine Design to engineering students in Sibley
Cornell
University.
It
presupposes a knowledge of
Mechanism and Mechanics of Engineering. While the former subject is a logical part of Machine Design, it may be, and usually
is,
for convenience, treated separately
and
in
advance of that
portion of the subject which treats of the proportioning of machine

parts so that they will withstand the loads applied.
logical order is usually followed in
The same
as
it is,
actual designing,
ordinarily, necessary
before proportioning the various
and convenient to outline the mechanism members.
of designing a
With the mechanism determined, the remainder of the work machine consists of two distinct parts: (a) Consideration of the energy changes in the machine, and
the
maximum
(b)
forces resulting therefrom.
Proportioning the various parts to withstand these forces.

logical procedure,
first
This
in
and the fundamental principles underseldom made clear

to
lying the
part
(a),
are
the student,
is
works of
this character;
and such information

is,
as
given on
energy transformation in machines

to special cases or types.
in general, that relating
thorough understanding of these
general principles
ful
is,
however, in most cases, essential to success-
design,
since
a consideration of the machine as a whole

consideration
of
details.
necessarily
precedes
A
is
very
brief
discussion of typical energy and force problems

fore, in
given, there-
Chapter
II, in the
hope of making
this
important matter
somewhat
clearer to the beginner.
While the treatment presented presupposes a knowledge of
Mechanics of Materials, a
straining actions
is
brief discussion of the
more important
given in Chapter III, partly to
make
the appli-
cation of the various formulae to engineering problems
somewhat
iv
PREFACE
definite,
more
assistance in selecting
and partly to present such rational theory as is of working stresses and factors of safety.
to
show why certain equations have and also to collect in concise form the more important equations relating to stress and strain with which the designer needs to be familiar. The general principles of lubrication and efficiency are Both of these are of prime importance discussed in Chapter IV. while the discussion is necessarily brief it is to the engineer; and
This discussion serves also
been selected in preference to others,
believed that the fundamental principles are fully covered.
of the book is devoted to the discussion of more important machine details, with a view of showing how the theoretical considerations and equations discussed in the first part of the work are applied and modified in practice. The treatise is, in no sense, a hand-book, neither is it a manual for the drafting room, but is a discussion of the fundamental principles of design, and only such practical data have been
The remainder
of the
some
collected as are
needed
to verify or
modify logical theory.
It is
hoped that the

tical
illustrative
numerical examples which are intro-
duced throughout the work may, in conjunction with the analy-
methods
given, suggest proper treatment of practical prob-
lems in design.
as
it
The
in
treatment of
all
topics
is
necessarily brief,
was desired
to obtain a text-book
which could be conve-
niently covered
one college year and yet present the salient
by the student as a preparation more advanced work. While intended primarily for engineering students it is hoped that it may also prove of some
features of the subject needed
and basis
for
interest to the practising designer.
It
has been the endeavor in
the preparation of the
book not only
to develop rational analytical
treatment, with due regard to constructive considerations
and
other practical limitations, but to reduce the analysis to such
forms and terms that definite numerical results can be obtained

in concrete problems.
Considerable of the matter contained in the book has already
been published, specially for the use of students in Sibley College, under the title of " Special Topics on the Design of Machine Elements," by John H. Barr, and also in " Elements of Machine
PREFACE
Design," Part
I,
by the Authors.
selves freely of "the
work
of
The writers have availed themmany others in the field, for which
due credit
is
given in the text.
The
of
authors are especially indebted to Professor G. F. Blessing
Swarthmore College, Professors W. N. Barnard, L. A. Darling,

of Sibley College, Cornell University,
all
and C. D. Albert
of
whom
also
have given instruction in the course
at various times,
to
Mr. A.
J.
Briggs, for
many
helpful suggestions
and and
criticisms.
They
will
be very grateful for further suggestions
or criticisms which will
improve the book.

D.
J.
S.
K.
H. B.
Ithaca, N. Y., June, 1909.
CONTENTS
CHAPTER
Introductory.
of Machine Design,
I
PAGE
Definitions and Fundamental Principles

i
CHAPTER
II
The Energy and Force Problem. Consideration of Machines as a Means of Modifying Energy,
CHAPTER
III
Straining Actions in Machine Elements.
Fundamental
.
Formulas for Strength and Stiffness,
-31
CHAPTER
Friction, Lubrication,
IV
96
and Efficiency,
CHAPTER V
Springs,
114
CHAPTER
Riveted Fastenings,
VI
136
CHAPTER
Screws and Screw Fastenings,
.
VII
.
.156
CHAPTER
Keys, Cotters, and Force Fits,
VIII
190
CHAPTER
vii
IX
211
Tubes, Pipes, Flues, and Thin Plates
viii
CONTENTS
CHAPTER X
PAGE
Constraining Surfaces, Sliding Surfaces, Journals, Bearings, Roller and Ball Bearings, 232
CHAPTER XI
Axles, Shafting, and Couplings,
.285
CHAPTER
XII
309
Belt, Rope, and Chain Transmission,
CHAPTER
Applications of Friction.
XIII
Friction Wheels, Friction Brakes,

350
and Clutches,
CHAPTER XIV
Toothed Gearing, Spur, Bevel and Screw Gears,
.
364
CHAPTER XV
Flywheels, Pulleys and Rotating Discs,
.
406
CHAPTER XVI
Machine Frames and Attachments,
.
428
MACHINE DESIGN
CHAPTER
The purpose
needs.
I
INTRODUCTORY
i.
of
machinery
is to
transform energy obtained

for
directly or indirectly
from natural sources into useful work

the laws that govern motion
human
force.
Useful work involves both motion and force, hence

is
the basis of
Machine Design
useful
and
The term
and
that
definite force, for
urable character.
its
work carries with it the idea of definite motion work itself is always of a definite or measAn examination of any machine will show
to
parts are so put together as to give definite constrained
motion suitable for the work

motion
is
be done.
The constrainment
of
determined by the moving parts, the stationary frame
and the nature of the connections between them. Mechanics is the science which treats of the relative motions of bodies, solid, liquid, or gaseous, and of the forces acting upon
them.
Mechanics of Machinery
is
that portion of pure mechanics
which
is
involved in the design, construction, and operation of

It
machinery.
has been noted that the consideration of a ma-
chine involves constrained motion, hence that portion of pure me-
chanics
is
mostly needed in Machine Design which deals with
stationary structures and constrained motion.

of
Mechanics
of
While the laws Machinery give us the underlying principles on

rests, their practical
which machine action
application brings in
many modifying
plication of
conditions.
Machine Design therefore may be defined as the practical apMechanics of Machinery to the design and construc-
tion of machines.
MACHINE DESIGN
A Mechanism
is
a combination of material bodies so con-
nected that motion of any
member
is
involves definite, relative,
constrained motion of the other members.

bination of mechanisms which
A mechanism or
com-
constructed not only for modify-
ing motion but also for the transmission of definite forces and for
the performance of useful
consists of one or
not necessarily
work is called a machine. A machine more mechanisms; a mechanism, however, is a machine. Many mechanisms transmit no energy
except that required to overcome their
own
frictional resistance,
and are used only
to
modify motion
as in the case of
etc.
most engi-
neering instruments, watches, models,
A brief reflection will show that the same mechanism will serve
for
different
machines
(see
any
treatise
on Kinematics) and
for a given
is
within limits the design of
the mechanism
machine
lit-
may
tle
usually be carried out, so far as motion
concerned, with
regard to the amount of energy to be transmitted.
This, of
course, does not apply to such

nors, or in general
mechanisms
as centrifugal gover-
where
inertia or other kinetic actions affect
constrainment of motion.
vided into two main parts
(i) (2)
Except for the limitations of such
cases as those just noted, the design of any
machine may be
di-
Design of the mechanism to give the required motion.

Proportioning of the parts so that they will carry the
necessary loads due to transmitting the energy, without undue

distortion or practical departure
from the required constrained
motion.
(1)
is
The
design or selection of the mechanism for a machine

is
governed by the manner in which the energy
supplied and
the character of the

plied in one
work
to
be done; for energy
may be
to
sup-
form of motion and the work may have

If
be done
mechanisms already exist which accomplish the desired result the problem is one of selection and arrangement of parts. But if a new type of machine is to be built, or a new mechanism is desired, the solution of the motion problem borders on or may indeed be of the nature of invention. While it is true that in most cases the mechanism and the
will
with quite a different one.
relative proportions of its parts
can be designed to
suit the
work
INTRODUCTION
to
it
3
in general
be done without reference to the energy transmitted,

is
necessary to
know something about
the
energy
trans-
mitted before any definite dimensions of the parts of the mechan-
ism can be
fixed,
anism
is
determined.
and frequently before the nature of the mechFurthermore, the methods and available
steam engine the
size of the cylinder
facilities
of construction control the design to a large extent.
Thus
be
be
in designing a
must
first
fixed before the length of
crank and connecting-rod can

treated apart
fixed,
and
in general while the

it is
mechanism can be
from the energy problem

stantly in mind.
(2)
necessary to keep the latter con-
The problem
of proportioning the various parts of a
ma-
chine so that they will carry their loads without excessive or un-
due deformation may conveniently be divided into two parts: (a) Solution as a whole, of the energy and force problem in
the mechanism.
(b)
Assigning of dimensions to the various parts based on
the forces acting

(a)
upon them.
When
the type and proportions of the

relative velocity of
If
been fixed the
may be
transmit
found.
is
mechanism have mechanism then the energy which the mechanism must
any point
in the
possible, in general, to find the forces act-
known,
it is
ing at any point since the law of Conservation of Energy underlies all
is
machines; or the product of velocity multiplied by force

If
constant throughout the train.
the forces acting on
machine member and the manner in which it is connected are known, these may serve as a basis for the assigning of definite
dimensions to the part.
fuller discussion of this
important
principle is given in Chapter III.
on a machine member can be deterwould seem easy to choose the material and assign proportions to it based on the laws of Mechanics, and such is the case when the stresses are simple and the conditions fully known.
(b)
If the forces acting
mined
it
Thus a machine member subjected to simple tension within known limits, can be intelligently proportioned in this manner. But in many cases the forces acting are very complex, the theoretical design is not
always
clear,
and our knowledge
of materials
MACHINE DESIGN
their laws is limited in
and
many
respects.
Recourse must there-
fore often be
made
to
judgment or
to empirical data, the result of
Even when the conditions are clear, theoretical demust always be tempered with practical modification and by The logical method of proporconstructive considerations, etc.
experience.
sign
tioning machine elements where theory

as follows
is
applicable
is,
therefore,
(a) Make as close an analysis as possible of all forces acting and proportion parts according to theoretical principles. (b) Modify such design by judgment and a consideration
of the practical production of the part.
In the case of details and unimportant parts, judgment and empirical
data are commonly the best guides. Summing up then, the logical steps in the
design of a machine
are as follows:
(I)
Selection of the mechanism.
(II)
Solution of the energy and force problem.
(III)
Design of the various machine members so they

Specification
will
not unduly distort or break under the loads carried.

(IV)
and Drawing.
and Drawing,
is the.
is
The
last step, Specification
a necessary and a powerful aid

of
important adjunct to the process of design;

to the designer's
it is
mental process and

is to
best
way
showing
the
workman what
and
also of
be done
to construct the
machine in ques-
tion,
making a record
of
what has actually been done.
It is not machine design of itself, however, as machines may be designed and built without any drawings. It is, nevertheless,
an indispensable part of the designer's equipment.

written specifications
useful but necessary.
Very often
accompanying the drawings are not only

In fact the highest
skill
on the part of the
designer
just
is
often needed to clearly
and
fully specify in writing
what
is to
be done, as the writing of specifications presup-
poses the most intimate knowledge of theory of design, and selection of materials.
From
the foregoing
it is
seen that the part of

generally
is
Machine Design
for convenience
included in
Mechanism can be and
INTRODUCTION
taught as a separate subject, and the student
is
expected to have
a knowledge of Mechanism, Mechanical Drawing, Mechanics of
Engineering, and Materials of Engineering as a preparation for

the
work contained
of
in this book.
The
chapters that follow deal
therefore with the solution of the
Energy and Force Problem,
and the Design
Machine Elements.
CHAPTER
II
THE ENERGY AND FORCE PROBLEM

2.
From
the law of Conservation of Energy
it is
known
must be
that
energy can be transformed or dissipated but not destroyed.
Therefore
all
the energy supplied to any machine
ex-
pended as either useful or lost work. Since frictional resistances, and frequently other losses, occur in all machines, the useful work done must always be less than the energy received. The useful work delivered divided by the energy received is This expression is differcalled the efficiency of the machine. ent for different machines and is evidently a fraction or less than unity. In the discussion which follows in this chapter, frictional
losses are neglected, unless otherwise stated.
Kinematic Cycle
is
made by
a machine
when
its
moving
parts start from any given set of simultaneous positions, pass
through
all
positions possible for
them
to occupy,
and ultimately
return to their original positions.
The energy received by a machine during a kinematic cycle may or may not be equal to the work done plus frictional losses. Thus the energy supplied during a number of cycles may be
some heavy moving part and then be given out during some succeeding kinematic cycle, as in the case of a punching machine with a heavy flywheel. An Energy Cycle is made by a machine when its moving
stored in
parts start from any given set of simultaneous energy conditions,
pass through a series of energy changes, and ultimately return

to their original
energy conditions.
Thus the complete mechanism of a four-stroke gas engine makes one kinematic cycle every two revolutions of the crank shaft. The slider-crank mechanism of the engine considered
separately
the crank.
makes a complete kinematic cycle every revolution of The engine makes one energy cycle every two revo6

lutions of the crank.
If a punching machine driven by a belt and running continuously, punches a hole every fourth stroke of the punch, it will be making a complete kinematic cycle every stroke and a complete energy cycle every four strokes.
Therefore, during a kinematic cycle,
Energy received = useful
work + lost work stored energy.
And
during an energy cycle,
Energy received
= useful work + lost
work.
Generally speaking, the useful work to be done and also the

character of the source of energy are
of design
is,
therefore, to select the
known and the problem mechanism which will trans-
form the motion

to
of the source of energy into the required motion,
determine the capacity of the driving device, and to proportion
the machine members.
The
and
this
proportions of any machine part depend, as regards
strength and rigidity, on the
maximum
force
it
must carry;
maximum
force
may be due
to the direct action of the
it may result from the inertia effect of some member which has a capacity for storing energy, and in such a case may be greatly in excess of any direct force that the driving deBefore this maximum force can be determined vice may deliver. for any member it is therefore necessary to make a complete
driving device, or
solution of the energy
problem including the determination of
the driving device.
A
ficient
knowledge of the quantity of energy required
desired
to do the work during a complete energy cycle is not always sufinformation upon which to base the design of the machine
its
or the capacity of
driving device.
machine may receive energy at either a uniform or variable rate and may be called upon to do work at either a uniform or Power or rate of doing work being the product variable rate.
obtained by multiplying together simultaneous values of velocity
making any energy transformations both the force and the velocity factors must be kept in mind. While the mechanism chosen may transform the motion of the
and
force,
it
follows that in
source of energy into the desired motion,
it
may
not necessarily
MACHINE DESIGN
so modify the energy as to give a distribution of force at the
point where
work
the
is
being done which exactly or even approxi-
conditions. Again, some of the moving machine parts may have to be very heavy in order to carry the required loads, and during one part of the cycle they
mately
fulfils
required
may
absorb energy, thus reducing the operating force, while at
another part of the cycle they

ing the operating force.
may
give
up energy, thus
increasentirely
Such a condition may make an
different distribution of the forces acting
on the members of the
mechanism, from that which would occur were the parts light or the motion of the machine very slow, and may materially
modify the design.
If
it
is
predetermined that some device
is
to
be used for
for giving
it
storing energy
when
the effort
is deficient,
is
in excess,
and
out
when
the effort
the capacity of the driving de-
vice need only be such as will supply during the energy cycle
an amount of energy equal

ing that cycle.
desirable and in
to the useful
work and
lost
work dur-
But in many machines such devices are not many others they cannot be applied. Two such cases may be noted, (a) In many machines under
it
continuous operation, where flywheels are not desirable,
is
found that
if
the driving device
is
proportioned so as to supply
energy at a uniform rate equal to the average rate required
throughout the energy cycle, the force at the operating point
is
sometimes greater and sometimes
less
than that required.
If
simultaneous values of the force and velocity at the working

point are multiplied together, their product
is
the rate at which
work
will
will
be done
at
the point considered.
The maximum
which work
product thus obtained
will
be the
maximum
rate at
be done and also

It is evident
at
which energy must be supplied by the

if
source.
will
that the capacity of the driving device
be greater in such a case than
based on the average rate

If
of
energy required per energy cycle.
the
driving device
under the above conditions should be too large or expensive, as

is liable to
different
of
be the case in large work, recourse must be had to a mechanism or to the use of flywheels or other means storing and redistributing energy, (b) Again, consider any

hoisting
mechanism.
Not only must

it
the driving device supply
during the cycle of operations (the raising of the load) energy

equal to the work done, but
sustain the load at any point.
the torque of the driving device
must
It is
also
be able
to start
and
evident that in such cases
on the hoisting drum, must be at least equal to that of the load, and if the torque of the driving device should be variable, its minimum torque must be equal If this to that of the load when referred to the same shaft.*
minimum
and
torque should be small compared to the
maximum,
the driving device chosen might have to be excessively large

this condition
might preclude the use of the driving device
first selected.
In any of these cases, after the form and capacity of the

driving device have been determined, the
maximum
force that
may come on any member may

It is to
also
be determined.
be noted that the choice of mechanism and the capacity

governed largely by the relative manner
it
of the driving device are in
which energy
is to
be received and work done, and
may be
well to enumerate the combinations that can occur, before apply-
ing the above principles to the discussion of illustrative problems.
In any machine under continuous operation energy

received and
(a)
may be
work may be done
in
one of the following ways:
Energy may be received

rate.
at a constant rate
and work be
and work be
done at a constant
(b)

rate.
at a constant rate
done at a variable
(c)
Energy
may
be received at a variable rate and work be
done at a constant
(d)
rate.

rate.
at a variable rate
and work be
done at a variable
3.
Case
received at
As an example of this case, where energy is a constant rate and work done at a constant rate,
(a).
consider a steam turbine running a centrifugal
pump
raising
is
water to a fixed
level.
Evidently the rate at which energy
* In certain hoisting devices friction
is
utilized to sustain the load or prevent
overhauling; this statement does not apply broadly to such cases.
IO
MACHINE DESIGN
work
is
supplied must just equal the rate at which

tional
done plus
fric-
and other
is
losses, for
any given period, and the capacity
of
the turbine
4.
very easily determined.
Case
(b).
As an example
of this case (energy received at
a constant rate and work done at a variable rate) consider the

case of a machine for punching holes in boiler plate.
Here the
driving belt can supply energy at a constant rate while the

useful work,
termittently.
which
is
of considerable magnitude, is delivered in-
If the driving belt
were designed with
sufficient
it
capacity to force the punch through the plate by direct pull
would have
to
be very
large.
The machine runs

is
idly a large
portion of the time, while the plate
being shifted, and in a
machine of this kind a device for storing energy, such as a flywheel, can be used to advantage. The total capacity of the driving belt need only be sufficient to supply, during the energy cycle, an amount of energy equal to the useful work plus the lost work. When a hole is punched the velocity of the wheel is reduced, the wheel giving up stored energy. During the time that the machine is running idly the belt can store up energy in the
up to normal. The maximum force that may be transmitted by the machine members will be based on the maximum force at the tool and will be transmitted only by the members that lie between the tool and the flywheel. As a second example of these conditions, take the design of a small shaping machine. Here the useful work is done during the forward stroke of the ram. During the return stroke frictional
flywheel
by bringing
its
velocity
resistances only are to be overcome.
The
resistance of the cut

is
during the forward stroke

limited
is
uniform and the speed of cutting
by the character
of the metal to be cut.
During the
re-
turn stroke, however, the velocity

limiting velocity depending
may be
greatly increased, the

parts, as
on the mass of the moving
these should be brought to rest at the end of the stroke without
shock.
at a
The machine
is
driven by a belt which can supply energy

is
uniform rate and, as noted above, the work
done
at a
variable rate.
Numerous mechanisms have been

tions.
devised to meet these condiSuppose a mechanism such as shown in Fig. 1 has been

selected.
II
The maximum
be done and the
length of the stroke
is
fixed
by the
work
to
or 4 inches.
length of stroke should be 3 Continuous rotary motion is imparted to the crank

it
minimum
a through the gear b of which

turn driven by the pinion
d.
c
forms a part.
is
The
gear b
is
in
which
is
rigidly attached to the shaft
On
the other end of d
a stepped pulley having diameters
el.
Diagram
Fig.
D D D
X
Z> 4
On
it
the countershaft overhead
is
a mating stepped
pulley so placed that
when
the belt
is
on the largest step of the
machine
pulley.
is
also
on the smallest step of the countershaft

is adjustable and can be moved from shown toward the centre of the crank, so
The crank
pin on a
the outer position as
that the vibrator e can be
made
to give the
ram R any
length of
12
stroke from the
MACHINE DESIGN
maximum
(20 inches in this example)
to
minimum
for
of 3 or 4 inches.
The range
of velocity of the tool

it
any length of stroke must be such that
can be lowered
to
the cutting velocity of hard cast iron or tool steel and raised to the economical cutting velocity of brass.
With the pin
in its
x
extreme outer position and the belt on the large step

speed of the
belt.
the
ram
will
is
be a
maximum
for that position of the
As the crank
drawn toward
the centre (the belt re-
ram is obvinow smaller ously decreased. If the belt is shifted to a step as D 2 the velocity of the ram will be increased, so that at any stroke It is variable speed may be obtained to suit the metal to be cut.
maining
in its original position) the velocity of the
not desirable to use a flywheel, the inertia of the moving parts

small,
is
and the problem is therefore to design the driving belt and proportion the machine members on the basis of the maximum pull which the belt may be able to exert. The mechanism transforms the uniform rotary motion of the
line shaft into the required reciprocating motion.
Consider the
crank pin
at its
extreme outward position and the belt on

for full forward stroke
The
velocity
diagram
under these condi-
tions is shown, the ordinates of the
diagram* representing the

crank length represents
velocity of the
ram
to the scale that the
the uniform velocity of the crank pin.
The diagram
is
for the
backward stroke
is
not
drawn
since
it
not needed in the
solution of the energy problem;
but
it
should in general be
drawn
to
make
sure that the change in velocity at the extreme

is
ends of the stroke

vary inversely as
not excessive.
If the belt supplies

it
energy
at a constant rate the force

its
which
can deliver
at the tool will
cutting velocity.
The
cutting resistance,
however,
is
uniform so that while the mechanism produces the

it
desired transformation in motion

of force desired.
may
not give the distribution
To
design the driving device (or belt) for such a mechanism
of
* For a full discussion of these so called quick-return mechanisms and the methods drawing velocity diagrams see "Kinematics of Machinery" by John H. Barr, " Machine Design " by Smith and Marx and " Machine Design," Part I., by F. R.
Jones.

the operating conditions of the machine
its
when
the belt has both
maximum and minimum maximum pull which a belt

''Mechanics," page 182.)
therefore
velocity
can give
must be investigated. is 7\ T 2 where 7\
The
is
the
allowable tension on the tight side of the belt.
(See Church's
is
The power*
that a belt can give out
(7\
2)
where
is
the velocity of the belt.
Since
7\
has, at all
moderate
belt speeds, a constant
maximum
from a
on the
value for a given belt, the power that a belt can deliver will vary
directly with its velocity.
The
belt receives its energy
shaft running at constant speed
and when the

it
belt is
smallest step of the countershaft cone

largest step
will
also
be on the
of the
running
at its lowest velocity,
machine cone and will in consequence be under which condition its capacity
for delivering energy is a
minimum.
required for small machine tools
'
The maximum power

approximately constant
is
at all speeds;
for since the heating effect

tool is proportional to
is
which governs the cutting capacity of the

the
work done,
it
follows that as the cutting speed
increased
the resistance of the cut
must be decreased and
vice versa, thus
keeping their product approximately constant.

designed to have sufficient capacity
stroke and the belt
ty, it will
is
If
then the belt

is
is
when
the
ram
making
full
on
lf
and hence
at the lowest belt velociIf a
have excess capacity when in any other position.

is
softer
metal
to
be cut the velocity of the
ram may be
increased,
but this can only be done by shifting the belt to a position where
its
and hence its capacity will -be greater. As before noted, the effect of moving the crank pin inward, the belt remaining in the same position, is to decrease the average
velocity
velocity of the ram.
Therefore as the stroke
is
made
shorter
the velocity of the crank, to maintain a given cutting speed,
must
be increased by shifting the belt to a smaller step of the machine

cone.
its
The
other limiting condition
is
when
the
ram
is
making
and giving a cutting velocity high enough for the softest metal to be worked. The belt should then be on the smallest diameter D 4 and hence at its highest speed.
shortest stroke
,
full
discussion of the
power transmitted by belting
is
given in chap. 12.
14
MACHINE DESIGN
An
full
is
inspection of the velocity diagram

its
stroke shows that
velocity is
ram is making a maximum when the ram

when
the
will
in
mid
position.
Neglecting friction and inertia, which here
are small, the force exerted on the
ram
be a
minimum where
belt velocity,
the velocity of the
ram
is
maximum
at
any given
is
because, for a given belt pull since no flywheel
used, force at belt

If,
velocity of belt
=
x
force at tool
velocity of tool.
there-
ram making full stroke, the capacity of the belt when running on D is made great enough to give a force at mid position of the ram equal to the required cutting force, it will
fore,
with the
have excess capacity
at
any other position; and
if
this condition
does not give too large a belt the driving device will be satisfactory.
The maximum force

be based on the
that any
member may have

belt,
to sustain will
will
maximum
torque of the
which
occur
when
this
it is
running on
for since the inertia forces are small
torque will be transmitted directly to the members, and the
resulting stresses
may be
easily
computed.
Example
Let the greatest resistance of cut
" "
"
800
20
4
lbs.
" "
"
"
"
= maximum stroke of ram = minimum stroke of ram = maximum length of crank " " minimum "
belt speed
inches.
inches.
(i
6y 2
iH
ic
"
" max. cutting speed on shortest stroke
and highest
lowest belt
= 60 ft.
ft.
per min.
stroke and
"
" max. cutting speed on full
speed
= 25
per min.
Then
in general,
linear velocity of crank
max. linear velocity of ram max. ordinate of diagram*

is
length of crank
Hence
in this
example when the ram
making
full
stroke at
lowest speed,
* In the mechanism here chosen the position of the ram for maximum velocity can be located by inspection and the value of the velocity determined without drawing the complete diagram. In general, however, the diagram must be drawn
in order to locate the
maximum
ordinate.

Linear
vel. of
15
crank
= =
cf
6--
23.5
ft.
per min.
'.
R.P.M.
of
crank
23 5
'
X *2 = X x 6J
7T
6.9.
In a similar way when the ram

at highest speed,
is
making the
ft.
'
shortest stroke
Linear velocity of crank = 42.5

Therefore, R.P.M. of crank
per min.
42 5
2
7T
X " = X ii
54.1.
Let the gear ratio be 8 to
1.
Then
the
minimum and maximum
R.P.M.
is
of shaft d
= 55.2 and
r
i_
1
i
432.8 respectively.
A
X
14" pulley
a convenient diameter for

1
.
*
Dv
1
velocity of belt
on low speed
ft.
14
7T
55.2
maximum
by the
12
= 204
If the efficiency of the
per min.
cent, the
machine be 85 per
rate of doing
work
800 - X
at this position of belt is the cutting resistance
multiplied by the
maximum
25
velocity of the ram, divided

ft.
efficiency, or
85
23,500
lbs.
per minute.
effective pull at belt
115
lbs.
approximately.
The
effective pull of single-ply belt per inch of

lbs.
width
may be
taken at 40 to 45
.'.
width of belt
= 2W"
45
of of
of of
nearly.
alike, as is
If the
cone pulleys on machine and countershaft are
the usual case in metal-working tools, then
A A
and hence,
nearly.
Max. R.P.M.
J
Machine Cone
Machine Cone Machine Cone Machine Cone
"
^Min. R.P.M.
R.P.M. R.P.M.
if
Min. A=A^| Max.

in the
'
example
A=
i 4j
A=
x4
\ \
^' 2 =
432.8
5"
MACHINE DESIGN
The maximum
force that
may
x
be applied to any
member
which machine
will
be based on the
maximum
is
torque of the driving
belt,
occurs
cone.
when the belt The difference
on
the largest step of the
in this respect
between
this case
and the
while
punching machine discussed above should be noted,

the driving
rate
for,
mechanisms of both can deliver energy at a uniform and while both do work at a variable rate, the maximum
is
load
applied in entirely different ways.
During the complete energy* cycle of the machine the total work done, neglecting friction, is equal to the length of stroke multiplied by the uniform resistance of the cut, or
800
20 = 1333
ft.
lbs.
For every cycle of the machine the
shaft d
makes 8
revolutions; hence the

if
belt could deliver
amount of energy that the work were done uniformly during one cycle
is
it
115
3370
is
ft.
lbs.
The
great as
capacity of the belt
therefore two
and one-half times as
would need to be if a device for equalizing the energy, such as a flywheel, had been used. Where a small machine is belt-driven, as in the case discussed, this added first cost is not serious. But when the power needed is great, or in such cases as
direct driving
by
electric
motor, the additional cost of a driving
device so greatly in excess of average requirements needs to be

carefully considered.
This, in fact,
is
one of the most important

desirable to
elements to be considered in fixing the size of motors needed

for direct-driven
machine
tools,
sometimes making
it
introduce a flywheel to reduce the size of motor.

5.
Case
is
(c).
One
of the best
examples of Case
(c)
where
at a
energy
received at a variable rate and

is
work
is
performed
uniform rate
since this
it
found in the reciprocating steam engine, and

is
machine
of such great
importance to the engineer

detail.
will
be discussed somewhat in
Here the energy
is
supplied in the form of steam pressure, and after cutoff occurs
and the steam expands

*The
in the cylinder the pressure falls
from
kinematic and energy cycle are, in this case, simultaneous.

the "initial" or boiler pressure to
17
somewhat above exhaust or

is
atmospheric pressure.
varying rate.
The
energy
therefore supplied at a
But the engine is required to deliver energy at The mechanism used will the driving belt at a uniform rate.
-Stroke
Fig.
3.
MACHINE DESIGN
produce the required transformation of the reciprocating motion But the of the piston into the rotary motion of the crank shaft.
distribution of the driving force in the
gential effort will not
form
it
of torque or tan-
be uniform but
will
be a
maximum
somewhere near the position at which the crank is at right angles to the connecting-rod, and it becomes zero when the crank is on The turning effort will therefore sometimes the dead centre. be greater and sometimes less than the resisting effort of the driving belt and the machine will stop unless a redistributing
device, such as a flywheel,
is
used.
The
reciprocating parts,
such as the piston and crosshead, and also the connecting-rod,
and their maximum velocity is considerable; hence the forces due to their inertia cannot be neglected. Referring to Fig. 2 (a) the crank a is required to rotate around the center O with uniform velocity and to give a uniform force
are heavy
,
The moment at the driving belt the average moment at the crank pin, hence the uniform force at the crank pin may be derived from
at the driving belt.
belt.
is
equal to
equivalent
that at the
This required driving force
at
the crank pin
may be
circle S. If
plotted radially
from the crank
circle as a base,
forming a polar
the
diagram of the required force
at the pin, as
shown by
The
crosshead
moves
at
a varying rate of speed.
velocity of the crank pin be represented
crank, the intercept

tical
Op
by the length of the made by the connecting-rod on the verthe
through
will
represent
scale.
simultaneous velocity of
plot-
the crosshead to the
same
These intercepts may be
ted at the corresponding positions of the crosshead, thus outlining the curve
whose ordinates represent the
velocity of the
crosshead at any point.
The
forces acting
upon the piston and which must be
trans-
mitted to the crank are,

(1)
The steam
The back
pressure which
is
represented at any point by

(b).
the ordinates of the curve T, Fig. 2

(2)
pressure* on the other side of the piston, actin.
* This generally amounts to 2 or 3 pounds per sq.

in
above atmospheric pressure
non-condensing engines.
ing against the steam pressure, and represented by the exhaust pressure line z z and the compression curve U.
(3)
The
inertia forces
parts.
due
to accelerating
and retarding the
heavy reciprocating
During the
first
part of the stroke these inertia forces tend to
reduce the effective pressure transmitted to the crank pin, and during the latter part they increase the effective force on the rod.
They can be represented

first
graphically by such a curve as V.
The
two curves can be found by the well-known methods of drawing indicator cards, and the third can be found either by
mathematical deduction or by graphic methods* based on the
velocity diagram.
It is believed that the analytical
method
is
the most satisfactory,
and such a method
is
presented in a suc-
ceeding
article.
If the acceleration is
acceleration
celeration,
is
also
known the force necessary to produce the known since accelerating force = mass X acat
and the force
any point (reduced
sq. in. of piston)
may be
plotted as
to pounds per shown by curve V, Fig. 2 (b).
When
the
reciprocating parts reach their
maximum
velocity
their acceleration is zero,
hence the curve of acceleration forces

very nearly at the position where
crosses the axis at a point g corresponding to the point of maxi-
mum
velocity.
This point
is
the crank and the connecting-rod are at right angles

error introduced
ratios of
and the
by assuming
this to
be so
is
small with ordinary

the recipro-
crank
to connecting-rod length.
Beyond g
cating parts are retarded, hence the inertia forces increase the
effective
crank-pin pressure from that point on.
The compression
curve (U) tends to decrease the effective pressure on the piston
and hence
pressure.
its
ordinates must be subtracted from the forward

algebraic
The
a resultant pressure curve
sum of the curves T, U, and V will give W, Fig. 2 (c), whose ordinates at any
crank
point represent the effective pressure acting on the piston rod at

that point.
This
effective pressure is transmitted to the

b.
by the connecting-rod
crank pin
The
pressure of the rod against the
may be
resolved into two components, one tangential

Machinery" by
J.
* For a full discussion of this matter see "Kinematics of Barr, page 71, paragraph 42.
H.

20
to the
MACHINE DESIGN
crank
circle
and tending
to
produce rotative motion, and

to
one radial along the crank tending

tension in the crank and
produce compression or
friction in the
main bearing.
If friction
Only the
tangential force can do useful work.
be neglected
crank must
the rate at which
work
is
done by
is
this force at the
equal the rate at which work

the curves R and W, Fig.
2 (a)
being done at the piston.

(c)
Now
and 2
respectively, give the simul-
taneous values of force and velocity at every point of the stroke.

If
such simultaneous values be multiplied together and divided

velocity of the crank
(all
by the uniform
the quotient
is
in the
proper units)
the tangential force at the pin, and this
may be
what is Curve X. These values of the tangential force can be found more easily graphically. It will be remembered that the ordinates of the velocity diagram (R) as drawn in Fig. 2 (a) represent the velocity of the crosshead to the same scale as the length of the crank
plotted radially on the crank circle as a base, thus giving
(c),
called a radial crank-effort diagram, Fig. 2
represents the velocity of the crank pin.
In Fig.
2 (c)
the connect-
ing-rod extended,
if
necessary, cuts the perpendicular through
O in the point h. Therefore O h = velocity of crosshead when O j = velocity of crank pin. Neglecting friction, the rate of
work
at the
at the
crank pin
is
equal to the rate of work at the cross-
head, hence the velocity of the crank pin multiplied by the force
crank pin
is
equal to the velocity of the crosshead multi-
plied
x
by the force
x
at the crosshead, or the tangential force
XO j
= e f xOh.
.
*
.
tangential force
= ejiXOh
b.
Lay
rm.
off
Oi =
e 1/ 1
and draw
k parallel to
ej,
Then,
Ok
-pr-r
Oi
* Therefore,
Ok =
OiXOh
(J ]
XOh u
1
tangential force.
j as
Therefore
stroke
O k may
be laid
off radially
from
an ordinate of
the required curve as

is
j k'.
The
construction for the return
performed in a similar manner.

be noted that the distribution of force as represented
is less
It will
by
this
diagram
uniform than the original curve of press-

ure at the crosshead.
21
By the conditions of the problem, howmechanism must produce a uniform turning effort at the driving belt or such as would be given by a crank-effort diaA flywheel must therefore be used to gram like S, Fig. 2 (a) store energy when the crank effort is in excess and to give out energy when the crank effort is deficient. Fig. 2 (d) shows the crank-effort diagram rectified with rectangular ordinates equal
ever, the
.
to the polar ordinates of curve
X.
The
base
YY is equal to the
circumference of the crank circle and the ordinates of the line

/
are equal to the ordinates of the required uniform crank-
effort
curve S.
Since the abscissas represent space and the ordi-
nates represent force, the areas /, K, J, I v
Kv
etc.,
represent
fly-
work.
The work
it
represented by
K+K
is
that which the

l
wheel must absorb and the area represented by I + J + I 1
which
must give up
in
one revolution.
Manifestly
+ J that I+J+I
-\1
J x must equal
A full discussion of the design of the v flywheel will be given in a later chapter.
The maximum
force that
K+K
may come upon

the engine
is
the crosshead can

It is to
be seen from an inspection

noted in this regard that
cutoff,
of the force
diagram W.
be
if
designed for variable
an indicator diagram
at late cutoff should
be drawn for
the purpose of locating this

will
not give the
maximum force, as an earlier cutoff maximum value. The method of analysis deon any member of the mechanism.
veloped above will enable the designer to determine the maxi-
mum
straining action
The
graphical method of finding the inertia curves, while con-
venient, are
open
to criticism
on account
of their inaccuracy be-
cause the tangents or sub-normals to the curve, on which these
graphic methods depend, are
difficult to construct
with accu-
racy and are at some points indeterminate.

fore,
it
In general, there-
is
thought that the following method or some similar

satisfactory.
one
is
more
Referring to Fig. 3 (page 17),
Let a
" u
R = L =
acceleration at any point,
length of crank in
"
feet,
" connecting-rod in feet,
22
MACHINE DESIGN
Let
"
N = Rev. per min.,

and
<p
angles
made with
centre line by the crank and
connecting-rod respectively at any position measured from the
crank position
r,
Let k
Let x
= distance from =
centre of crank shaft to
mid
position of
crosshead,
displacement of crosshead from mid position,
U
" v
~ =
R'
velocity of crosshead at
any point
v.
x,
"
"
t .==
o)
time elapsed corresponding to
=
k
<p
angular velocity in radians per second,
Then x +
But Lcos
= =
OB
2
BC
2
R. cos e
+L
2
cos
<p,
VL R
2
sm2
sin o
2
= #
2
J \ R
2
sin
RV n
,-.
= R
(costf
Vn sin
o)
(i)
all
Expanding the radical by the binomial theorem and omitting

error with the limiting proportions ordinarily used) equation
terms beyond the second (which can be done without appreciable

(i)
becomes,
=R
cos
in
(2)
Now x = the
distance
stroke and velocity

(2)
moved through by dx and at x =
the crosshead, from

therefore
mid
differentiating
with reference to
t
.
dx
-7-
=
=
/ _(
ic
I
sin *
+
/
sin 2 <9\^6
sin2#W0
-2^r^
+
cos2
0'\
(dd\ 2
G)
The
acceleration
a
,
dv = j; =
1
j?=- R
1
d2 x
COS0
^r)\Tt)
27zN
60
'
(4)
but
de = d/
velocity angular & ^
in
radians per sec. r
hencea =
-(-^-J^^co5^ + -^-J,
/2 7riV
COS2^\
(5)

which
is
23
the general expression for acceleration of the recipro-
cating parts.
If the
weights of parts be called
W, from Mechanics
it
is
known
that the force necessary to produce an acceleration (a) is
P = p
W a where g = 32.2
n R
in English units; therefore
-r
f2-N
^
60
M
2
\f
)
COS2 0\
cos
H
where
is
in
ft.
(6)
'
or reducing,
WrN
2 ,
(
cos
cos2 0\
)
. .
35,200
where
r is in inches.
(7)
When
when
the solution of the above expression gives a negative
result the force of inertia is acting

positive,
away from
the crank
and
toward the crank.

2
It is also to
be noted that the

of
expression
W R (2xN
(
\
)
is
the centrifugal force
a weight
equal to that of the reciprocating parts concentrated at the crank
WRo
pin since centrifugal force in general
is
2
.
equal to
o
By means
of equation (7) all points
on the acceleration curve and

it is
could be found and plotted.
In general, however, the exact

sufficient to
characteristics of the curve are not essential
make the three most simple solutions as follows, and a curve drawn through the three points thus located is sufficiently accurate for all ordinary purposes.
In cases of extremely high speed be desired.

2
with small ratios of connecting-rod to crank a more accurate determination of the curve
may
When When
When
e
=
=
=
*
0,
P =
fWr N
[ V
\ (
)
35200
I+-) n'
(1
v
1 \
(8)
180,
P = + (El^!) \ /
35200
'
-) /
(L)
/
(Q )
oo or 270 *
P = (^1^1)
^
35200
I0 )
The
piston
is
not at half stroke.
24
MACHINE DESIGN
If the inertia forces are to
be combined with the steam press-
ures, as
shown graphically
in Fig. 2 (b), they
must be reduced
diagrams.
to
pounds per square inch of piston
to give correct
An example may
the following data:
serve to
make
these points clearer.
Let
it
be required to design a steam engine to deliver 150 H.P. with
Steam pressure = 90
Piston speed
stroke
lbs.
gauge.
Cutoff at y& stroke.
Ratio of crank to connecting-rod
=1
to 5.
= strokes
per minute multiplied by length of
= 640
ft.
Here something must be known about the

members.
2
(b),
size of cylinder
necessary, before definite dimensions are assigned to the various
Let a theoretical indicator card be drawn as in Fig.
neglecting for the present the inertia curve
since this
its
only tends to redistribute the energy and does not affect

quantity.
The
distance z z represents the piston travel and the
ordinates of the curve
represent piston pressures; therefore
the area between z z and the curve
represents the
work done
by the steam pressure during the stroke. In a similar way the area under curve U represents the work of compression due to back pressure. The difference of these areas is the net work
done per stroke
of piston
and the mean ordinate corresponding

proper scale the average pressure
to this area represents to the
per sq. inch on the piston during stroke.

z z
In the case given

sq. in.
= 2".
Area under
ordinate
T minus
1
area under
.875".
U = 1.75
scale
There-
fore
mean
is
.75
The
of
pressures
taken
70
= 70 lbs. .875 = 62 lbs.

1"
Therefore
mean
pressure during stroke
Let
A = area of piston. *P = mean effective pressure per L = length of stroke in feet.
sq. in.
= number of revolutions per minute. H.P. = horse power required.

Then H.P. =
2
P LA
33000
N
.
Here
P.N X L
and H.P. are known.
25
.
^^
A =
H.P.X
33000
PX2NL
be taken
at
= 150X33000 =I32 62X640
SqUare mcheS
'
or a diameter of cylinder of 13 inches.

If the stroke
about twice the diameter of the
cyl-
inder, or say 24 inches, the proportions will be good.
Hence
(a
since
2L
^ = 640, N = i6o R.P.M.

=i
ft.
ism can now be
laid out to scale.
The mechanThis has been done in Fig. 2
and
c),* the space scale being i"

stated,
As before
where
sufficient
is
the location of the three points, namely,
respectively o,
180
and 90
or 270
(Fig. 3), is
to
locate the inertia

5>
curve.
In the above example
3-5>
n =
an d
N
1
)
160.
is,
The
general expression for the inertia force
for e
o.
P=
W rN
35,200
(
1
\ + -/ = C
/
I
\ + 1
n'
where
<.
is
a constant
and here equal 4

Therefore,
3.5
X
o,
12
160
to
ô.
35,200
When
P =
30.5 (1
+ -J =
36.6
lbs.
lbs.
When When
These values serve
= =
oo,
P =
,
30.5 (-)
=6.1
180
P =
30.5 (1
-)
Fig. 2
24.4
lbs.
to locate the curve as in Fig. (2).
The
resultant of
T U and
V, curve
W,
drawn and
drawn.
to
the crank-effort diagram
X plotted.
curve can be rectified as in Fig. 2 (d) and the
(c) can now be The crank-effort mean ordinate Yl

,
The
area I +
J=K
will
be proportional to the energy
be absorbed and delivered by the flywheel.
One
inch of ordiof abscissa
nate here
= 70
lbs.
per
sq. in. of
piston
and one inch

ft.
=1
ft.
therefore one sq. in. of area
= 70
lbs.
The
Hence,
area of
if
^ = .5
sq. in.
and* area of piston
= 132
sq. in.
E = energy to be absorbed, = .5X70X132=4,620 ft.
lbs.
on
which
the
de-
sign of the flywheel can be based.
Reduced
in reproduction
about one-half.
26
MACHINE DESIGN
The maximum
pressure that can occur on the piston
is
the
initial or boiler
pressure as the ordinates of
are at
all
points
will
less
than those of T.
Hence, when running, the parts
be
subjected to less load than in starting up, ure
may be
6.
applied before inertia forces
when become
is
full boiler press-
noticeable.
at a vary-
Case D.
good example of energy supplied
ing rate and work done at a varying rate

driven air compressor.
found in a
direct-
Here the varying steam pressure

air
in the
steam cylinder
cylinder as
is
opposed by a varying
.
pressure in the air
The area of the cylinders are, shown in Fig. 5 (a) equal. The steam cylinder takes assumed to be for simplicity, steam at 80 lbs. pressure and the air compressor cylinder delivers The efficiency of the system shown is air at 100 lbs. pressure. taken at 80 per cent, and hence the area of the compressor card
is
80 per cent, of the steam card.*
If
both the pistons were
rigidly attached to the
same rod
is
it is
evident that the

air
maximum
minimum.
steam pressure
If,
will
occur where the
pressure
is
however, each cylinder
independently connected to a com-
mon shaft by means of a crank and connecting-rod mechanism, the maximum and minimum pressures of the cards may be made
by placing the crank pins at the proper In other words the mechanism may be so designed that energy will be delivered at the working point more nearly at the rate required by the work to be done. The Part of this is lost loss by friction, etc., is about 20 per cent. on the steam side and part on the air-compressor side. It can
to coincide
more
closely
angular distance apart.
be assumed, without great

loss is at a
error, that the losses
can be evenly
divided between the two slider-crank chains and also that the
uniform rate throughout the stroke.
Thus
the loss
on the steam side can be represented by the line a b, Fig. 5 (a), which reduces the effective pressure at every point by a fixed
* In the general case, where the cylinders are of different diameter and area,
the diagrams which represent
pounds per square inch of piston area would not
have a ratio equal to the
efficiency.
The mean
effective pressure of the air cylinder
multiplied by the area of the air cylinder, divided by the
mean
effective pressure of
the steam cylinder multiplied by the area of the steam cylinder, would, in this case,
equal the efficiency.

amount.
In a similar way ordinates to the line
diagram.
c
27
d increase the
effective resistance of the air
The
all
area of the diagrams
modified in this
way
will
be equal and
energy supplied will
be accounted
for.
Fig. 5 (b).
Fig. 5
(c).
*
1
-x x
Y-
Yf
V K>
\ Kl
> 'T ^
j.
Fig. 5 (d).
Since the moving parts of both slider-crank chains will be
heavy, the effect of inertia cannot be neglected.

the air
In Fig.
(6)
and steam cards are shown with the
inertia curve, the
28
friction line,
MACHINE DESIGN
and the compression curves in their correct relationFig. 5 (c) shows the resultant pressure curves, the curve pressures being plotted below the base line for conveniThe crank-effort curve of the steam cylinder is reprerepresented by Y.
ship.
of air
ence.
sented by X, and the resisting crank-effort curve of the air cyl-
inder
is
The cranks
are here placed 90
apart, the steam crank being in advance, a

in practice.
It is evident,
if
common arrangement
is
however, that this
not the most

is
advantageous angle, for

to correspond with
the point e on the air curve

5
(c),
made
on the steam curve, Fig.

be
still
the excess
and deficiency
in Fig. 5 (d),
of effort will
further reduced.
This would
This is even more clearly shown place the cranks at 45 apart. on the rectified curve of crank effort. Here the
area
Ii
K+K
the
+ Ji
revolution.
amount of energy to be absorbed and I + J + amount to be given up by the flywheel during one In the steam slider-crank mechanism the greatest
is
the
pressure
is,
as before, that
it
due
to the initial
steam pressure,
terminal air
while on the air side

pressure.
6.1.
will
be that due
to the
In the four cases discussed above the action of the maall
chine has in
instances been supposed to be continuous,

will
and
all
machines which operate continuously

classes.
belong to one of these

is
Where
the action of the
machine
intermittent or
irregular, these general solutions will not always hold
and the
design of the machine cannot be based on the energy given or
depend on the maximum force_ or maximum torque or, in other words, on the mechanical advantage which Thus the motor on an auto car has a the motor must possess.
received, but will
certain
maximum
capacity for delivering power.

at a
On
a level
road
it
can propel the car
high rate of speed, the engine
making only a few turns to every revolution of the wheels. But on a steep hill the gears must be shifted so that the engine has a greater mechanical advantage, and gives a greater torque on the axle, the engine making many revolutions to every one of the wheels. Another example of this is the case of hoisting mechanisms already discussed
somewhat
(see
article
2).
An
en-
gine or a motor might be capable of giving out energy at a rate

equal to that required to
lift
29
the load in a given time,
and
it
might be
able,
running continuously, to raise the load to the
required height.
But
its ability
to start
it
and sustain the load
at
any point will depend on whether

at that point
has a mechanical advantage
and not on
its
capacity.
Where
the torque of the
load
is
constantly changing, as in deep mine hoisting, the design
of the hoisting devices
becomes quite complicated and

It will
is
beyond
the scope of the present treatise.

that in such cases the
be noted, however,
minimum
torque of the motor or engine
must always exceed the maximum torque of the load when reThis general principle must be kept ferred to the same shaft. hoisting devices and similar machines designing in mind in
which act intermittently and slowly, or where redistributing
devices are undesirable or impossible.
6.2. Redistribution of
Energy and Inertia
Effects.
Devices
for storing and redistributing energy are very
common in transmis-
g
f
3
,1
1
A B ^m^
r~
Fig. 5
(e).
A'
lAA/W aa|
1
sion systems.
of
Thus,
in hydraulic distribution, the excess supply
power
is
stored in an accumulator,
and given out again when
the supply
is
is deficient.
In electrical distribution a storage battery
sometimes used
for the
by compressed
distribution
is
air a large reservoir is
same purpose. In transmission of power sometimes employed as a

re-
store-house of energy.
effected
In the case of a single machine, the
by compressing a gas, by using a spring, or by accelerating and retarding some heavy moving part. Thus in the steam engine the piston compresses steam in the clearance space at the end of its stroke, and the energy so absorbed is returned to
it
during the next stroke.

is
Again,
when
the energy
supplied by the steam
in excess of the effort required, the flyits
wheel absorbs the excess and thereby has
velocity (and hence
30
its
MACHINE DESIGN
kinetic energy) increased.
When
the effort
is
in excess, the
its velocity.
wheel gives up the stored energy

It
at the expense of
all
does not necessarily follow, however, that
heavy moving
parts simply redistribute the absorbed energy as useful work,

as the action
may be
a positive source of loss.
In Fig.
5 (e) let
it
be the platen of a large planing machine, and suppose

its
to
be
making
return stroke, moving from
left to right.
The
force
just necessary to slowly
move
the platen
may be
represented by
the vertical ordinates of the diagram

that a greater force
is
abed.
Suppose now,
applied, in order to hasten the operation,
so that at the position
A' the platen has been accelerated

,
till its
kinetic energy is equal to the rectangle e
g h
if
c.
Evidently the
platen will not stop at the end of the stroke
the actuating force
be removed at A', as the work of
friction
during the remainder

If,
of the stroke is less than the stored energy.
therefore, the
removed at^4' and the "driving" belt applied, the latter will slip upon the driving pulley till the excess of energy If the point A' has been propis absorbed and dissipated as heat. erly chosen the platen will just stop at the end of the stroke and the energy absorbed by the belt will equal the area / g h b. If a spring, S, were fitted to the machine, so that the work of com^ pression from the position A' to the end of the stroke just equalled
"return" belt
is
the excess kinetic energy of the platen, at that position, the return
belt could
be thrown
stroke.
off at A',
and the platen would stop

stored in the spring
at the
end of the
action
is
The energy
be returned to the platen on the forward stroke.

cylinder, Fig. 2, the energy
would then This latter
identical with that of compression in the steam-engine
under the curve

the
being returned to
the reciprocating parts on the next stroke.

this last case, that
It is to
be noted
in
even
to
if
work
of compression is not quite

latter part of the
equal to the energy

stroke, there
is
be absorbed during the

energy
is
no
loss of
(friction neglected), as
what
is
not absorbed by compression

useful effort.
absorbed at the crank pin in
CHAPTER
III
STRAINING ACTIONS IN MACHINE ELEMENTS

7.
Nature of Forces acting in Machines.

clear that
From
the fore-
machine members which transmit going chapter it is energy are subjected to forces of a varying character and intenSince the various parts of a machine must be constrained sity.
to
move
the
in fixed paths
it is
important that they should neither

strong but also
break or be distorted appreciably under the loads carried; that

is,
members must be not only
stiff.
The
pro-
portioning of machine elements as dictated by various methods

of loading is therefore
this chapter.
most important, and
will
be considered in
The
forces acting
on a machine element may be one or several
of the following:
(a)
The
useful load due to the energy transmitted.
(b)
(c)
Forces due to frictional resistances.
The weight
of the part itself or of other partSc
(d)
(e)
(f)
Inertia forces
due
to
change of
velocity.
Centrifugal or inertia forces.
Forces due to change of temperature.
(g)
Magnetic attractions, as in
forces or loads
electrical
machinery.
These
ways.
may be
applied to a machine in several
They may
and then
act steadily in
one direction; they

they
may
act in-
termittently in one direction, or they

direction
in the reverse;
may be applied first in one may be applied gradually,

is
or suddenly in the nature of a shock.
steady or dead load
is
one which
always applied steadily
in the
same
direction.
applied and removed.
A live load is one A suddenly applied

initial velocity.
31
which
load
is
is
alternately
one imposed
instantaneously but without
If the load is ap-
32
MACHINE DESIGN
blow from a
falling
plied with initial velocity as in the case of a
body, the
8.
member
is
subjected to impact.
Stress, and Strain. Since more or less elastic a machine element must change its form to some extent whenever subjected This change of form may be very small and tempoto a load. rary; it may be a permanent distortion; or if the load applied be heavy enough the element may even be ruptured. Such
Nature of Straining Actions,
all
materials of construction are
change of form, whether temporary or permanent,

strain.
is
called a
When
a machine
member
is
thus distorted under a load
certain molecular reactions, equal and opposite to the load applied,
are set
is
up within the material and
resist the
deformation.
is
Stress
the term applied to this internal reaction and
to
be clearly
distinguished from strain, stress being in the nature of a force
and
strain being a dimension.
The
results
character of the straining action and of the stress which
from a given load depend upon the direction and point of
upon the form, the posiand the arrangement of the supports of the member. A given load may produce tension, compression, shearing, flexure,
application of the load (or forces), and
tion,
or torsion or a combination of these.
pression cannot both exist at the
Of course tension and comsame time between any pair of

or, as it is often ex-
molecules.
stresses
Flexure
is
a combination of tensile and compressive
between
different sets of molecules;
pressed, in different fibres* of the

special form of shearing stress. rence of flexure and torsion it
same body.
Torsion
is
Owing
is
to the frequent occur-
convenient to treat these as
elementary forms of
stress.
The
tially
stresses
due
to tension, compression,
and
flexure are essen-
molecular actions normal to the planes separating adjacent

that
is,
sets of interacting molecules;
the stresses increase or de-
crease the distances between these molecules along lines connect-
ing them.
The primary
is
straining effect of shearing
displacement of
*
It
and torsional actions adjacent molecules, between which the stress

is
should be noted that the term fibre
used in a conventional sense when

steel.
discussing
homogeneous metals, such as iron and

acts, tangentially to the
33
In
planes separating such molecules.
uniform shear the interacting molecules

atively with a rectilinear translation.
move
or are strained rel-
In torsional action the ad-
jacent molecules each side of a plane of stress have a relative
motion or strain about an

in reality only
axis.
brief reflection will
show
that
two kinds of
strain exist, namely, elongation (con-
traction
if
negative)
and shearing.
In a similar way only two
corresponding kinds of stress are met with, namely, normal or

direct,
and tangential or shearing.

separately
as
But
for convenience
it
is
much more
tioned,
desirable to treat the special cases previously
men-
elementary
stresses.
(See
Church's
Mechanics, page 201.)
Machine members
are often subjected to combinations of these
simple stresses, as flexure and torsion.
Such
stresses are called
Compound
Stresses
is
and
will
be more
fully treated later.
When
a load
applied to a piece of material the strain which
results is a function of the load

terial involved.
is
and
of the character of the
ma-
In general for a given loading the deformation
different for different materials but constant in its relation to
stress for
any one material.
These
relations have
been deter-
mined experimentally
neering,
ject fully.
for all the ordinary materials used in engiof materials treat of the sub-
and works on mechanics
Enough
will
be inserted here to make the discussion
complete.
is tested under an increasing tensile load and by each successive load is accurately observed the relation between stress and strain can be shown graphically as at O ade Fig. 6; such a diagram is called a stress-strain diagram. If axes O X and O Y are chosen and the stresses plotted as ordinates and strains as abscissas, it will be found that up to a
If a
bar of metal
the strain caused
certain point as a, either in tension or compression, the curve so
formed
is
sensibly a straight line;
that
is,
stress is proportional
stress is released,
this relation
till
to strain.
Further,
if
at
any point below a the
the piece returns to

ceases,
its
original shape.
But above a
strain usually increases* faster than stress,

is
finally
Ordinary rubber
stress increases.
an exception
to this general rule, strain
ckcreaSi k as
34
rupture occurs.
it is
MACHINE DESIGN
If at
any point beyond a the
stress is released,
found that the piece no longer returns
to its original
dimen-
sions but has been permanently distorted.

If at
any point on the curve below a the

is
stress
be divided by
all
the strain a ratio
obtained which
is
is
constant for
points
of
below
a.
This ratio
If,
called
the modulus or coefficient

of elasticity is
elasticity.
therefore, this
modulus
a given material, the strain corresponding to
known for any given load may
be calculated, providing
ing to the point
a.
is
it
does not exceed the value correspond-
The
point a
called the elastic limit
and
is
well-defined in
elastic
most materials.
Cast iron has, however, no well-defined
Fig.
6.
limit
and
little
permanent elongation.
Materials of this kind are
said to be
brittle.
If sufficient tensile stress is applied to a test piece its elonga-
tion increases until finally
it
" necks
down"
at its
weakest point
and rupture occurs.

breaks
is
The
is
load per unit area under which a bar
called its ultimate strength
and the corresponding
stress
or load per unit area
called the ultimate stress.
Similar phe-
nomena
are observed
when
a piece
is
tested in compression or
It is evident that the working stress of a machine member must be less than the elastic limit if the piece is to retain permanency of form. The stress at which a member is designed to

be operated
It is to
is
35
called the
working
stress
is
and the
ratio of the
ultimate stress to the working stress
called the factor of safety.

stress
be especially noted that not only must the working
in the
member be kept below

be, shall
the value where permanent deforma-
tion takes place,

it
but also so low that the resulting strain, whatever
may
be so small as not to destroy the proper alignment

through distortion.
of the piece, or cause unnecessary friction
machine member may be amply strong enough

render
unfit for the service' desired.
to carry the load
with perfect safety, and yet distort so badly under the load as to
it
Both strength and

will dictate the
stiffness
should therefore be kept in mind in designing a machine part, as
sometimes one and sometimes the other

dimensions
to
form and
given of
be used.
exist
short discussion will

stress,
now be
the relations
which
between load,
and
strain for the
met and of their bearing on the selection of the form and size of a machine member. In this discussion it will be assumed that the load is a dead load applied without shock, and the modifying effect of suddenly applied and repeated loads will be considered after the fundamental relations between load and
cases most often
stress are established.
9. Tension.
Let p be the
stress in the section,
the load,
exists
and
the area of cross section.

is
The
relation
which
between them in simple tension
>
if
=
A
T
and
I
W
the length of the
And
be the coefficient of
total elongation
elasticity
is
member, the
given by the equation
i-fi
The
elongation per unit of length or the strain
<>
=
A
-y.
If,
then, a tension
member
is
to
be designed
to join
two
machine parts, the formula for strength dictates a piece of uniform cross section without regard
the
Hence to any particular form. most convenient or cheapest form would be used, avoiding
36
thin,
MACHINE DESIGN
wide sections where concentrated
required
is
stress at the
edge might
cause undue strain.
Suppose
racy
is
it
is
to
hold
certain limits, as
desired.
is
often the case in
the two surfaces within machine tools where accuis
If the tension
member
long
it
more than
below the
reduce
desirable,
elastic limit
though the working stress and a greater area may be necessary
may yield may be well

to
to the desired value.
Example.
10,000
to
lbs., let
Let
P = 20,000 lbs., let the allowable stress p = = 30,000,000, let = 4o", and let it be required
/
keep A within
.001".
If the
design
is
based on allowable
stress
alone,
A =
_
But for A
P = p
A =
20,000
10,000
square inches.
20,000
.001,
.PI
A
40
h,
.001
26 sq.
in.
30,000,000
In general, therefore, where tension members are of any considerable length
and
distortion under load
is
of importance, they
should be checked as above.

10.
Compression.
If
the
member under
consideration be
subjected to compression, the remarks of the last paragraph
apply equally well
if
the
member can be
is
considered a short
its
column,
i.e.,
one whose length

If longer
not greater than six times

it
least diameter.
than this
must be considered
its
as a
long column and the conditions governing
design will be
more
fully treated hereafter.
(See Art. 20.)

is
11. Shear. If the
member
subjected to simple shear the
expressions for the relations existing between the stress, area,
and load are similar
to those for tension or
>s=T
12. Torsion. If the
<
C)
member
is
subjected to a torsional stress,
the following relations exist
Let
P
a
=load applied
in pounds.
= arm of load
in inches.
37
Let / p = polar moment of inertia of the section in biquadratic

inches.
p s = shearing stress in lbs. per unit area e = distance from neutral axis to outer / = length of member in inches. = angle of deformation in radians.
at outer fibre. fibre in inches.
T twisting
s
moment
applied to
member
in inch pounds,
E = transverse coefficient of elasticity.

Then
for torsional strength in general,
Pa=T = -*
For a circular shaft of
solid section,
(D)
*-*
df) T= p^W i64
{F)
For a hollow circular section whose outside and inside diameters are d x and d 2 respectively,
For deformation under

the most
stress for a solid circular section,
which
is
common
case,
and
for a hollow circular section,
32
-
r/
8 (fli
2 )
An
inspection
of
equation
is
(Z>)
shows that the torsional
resistance for a given stress

of inertia divided
proportional to the polar
moment
outer fibre.
for
Examination
by the distance from the neutral axis to the of equations (E) and (F) shows that
strength
is
circular
sections torsional
proportional to the
third
power
of the outer diameter.

is
Equations (G) and (H)
show
that torsional deformation
inversely proportional to the

is
fourth power of the outer diameter, hence torsional stiffness

directly proportional to the fourth
power
of the outer diameter.
For a given amount
of material that section in
which
this
ma-
terial is distributed farthest
from the gravity
axis will be strongest
38
MACHINE DESIGN
stiffest
as long as the walls of the section do not become so and weak as to yield locally from other causes. The hollow circular and hollow rectangular sections, commonly called the
thin
and
"box
section," Fig.
7,
are best adapted, therefore, to resist tor-
Fig.
7.
Fig.
8.
The box section is peculiarly useful in machine construction, as many machine members must carry a combinaMachine frames may be subjected to tension, tion of stresses.
sional strains.
compression, or shearing, combined with torsion, and the box

section, while equally
good for simple
stresses,
is,
as has been
noted, vastly superior in torsion.

is
Furthermore, the box section

flexure
well adapted to resist

sides of a
combined
and
torsion.
The
box section also afford facilities for attaching auxiliary parts and its appearance is one of strength and staflat
bility.
The
thickness of the walls being thinner in hollow than
in solid
also
lies.
forms insures a better quality of metal in castings and

surface,
more skin
where the greatest strength

to
of cast iron
An
advantage not
be overlooked in some lines of work
which hollow sections can be strengthened by increasing the thickness of the walls by changing the core withis
the ease with
out changing the
external dimensions.
The
cost
of
pattern
work
little
is
about the same, in general, for hollow sections as for I
or other sections, while the

greater.
work
in the foundry
is,
in general, a
Example.
bar
circular cast iron boring bar 60 inches long
carries a solid circular boring

is
head 60 inches
in diameter.
The
subjected to a torsional
moment
of 60,000 inch
pounds

which
is
39
applied at one end.
It is
desired to keep the torsional
deflection of the tool
below
-gV"
when
the bar
is
transmitting
power through
of the tool.
its
entire length, in order to prevent chattering
What
should be the diameter of the bar
if
the
working
stress
be taken as 3,000 pounds per square inch and
be taken as 6,000,000.
For torsional strength from formula E,

d
3
60,000
X 16 = 3,000 X4 .6".
100
.\d =
For torsional
stiffness
*
r0,
=
r
-giro"
since
is
in radians
and the length
of
an arc
where
radius.
from G, a*
32
-X
6,000,000
60 X 60,000 X
5,870
-cnhr
hence
if
</
= 8.8".
It is
evident that the shaft will be amply strong
designed for
stiffness, therefore the last
value would be used.
If the section is
made hollow
Let
less
metal can be used.
In this
case either the inside or outside diameter or the ratio between
them can be assumed.
dt
4
4
whence, dS
Substituting in
= ^4" and df d = ^f d* 256 256

2
H,
"
y* 60,000x60 it-itû256 x X
6,000,000
,
-g-io
.*.
dS
8,550 and d
9.6
rr
hence d2
sq. in.,
7.2".
The
*
area of the hollow shaft
= 31.67
of a shaft
while the area of
The angular
deflection or twist
in
degrces=57.2o6X (Angular
deflection in radians).

40
the solid shaft
MACHINE DESIGN
= 60.84
sq. in., so that
with a small increase in
diameter one half the metal secures, by using the hollow section,
the
same
13.
stiffness.
Compound
Stresses.
In the
cases
of
simple loading
just discussed only
one form of
stress is
brought on the
safely based
member
on
this
and the design of the cross-section can be

stress.
When, however, the loads applied induce stresses of several kinds, it is no longer possible in general to base the design on any one stress, but regard must be had to the combinaIn many cases one or more of tion of stresses that may occur.
the stresses are so small, or their action
neglected, in designing the
is
such, that they
may
be
be borne in mind.
ber
is it
member, though they should always The stress on which the design of the memstresses.
based
may
be called the predominating or primary stress
and
may
be a simple stress or a combination of simple
The
latter will
be called a Compound
Stress.
14. Flexure.
When
beam
is
subjected to simple bending

(a)
the principal stresses that are induced are

side of the neutral axis, (b) a
a tension on one
compression on the other side of

angles to the tension and comis
the neutral axis,
and
(c)
a shearing stress which acts on every
section of the
pression.
beam
at right
Generally speaking, the shearing stress
small com-
pared with the tension or compression and can often be neglected.

It
must never be forgotten, however, and where the beam
is
designed to withstand the bending
moment
only, care should be
exercised that the sections which are subjected to a small bending
moment
are not
made
so small as to yield under shear.
The
predominating
stress will
be the tension or compression depend-
ing on the material and the form of section.
Let
When a beam is subjected to simple flexure, M = bending moment at any section in inch pounds. / = moment of inertia of section in biquadratic inches. e = distance from neutral axis to outermost fibre in inches. A = deflection at any point in inches.
P
p
load applied
in pounds.
= maximum E coefficient
stress at outer fibre in lbs. per sq. inch.
of elasticity.

41
Then
for strength, in general, within the elastic limit,
pi * M=
e
U)
Every beam when loaded
deflects
somewhat, depending on the
which it is supported, and the load applied. The curve assumed by a beam loaded within the elastic limit is called the elastic curve and is of
shape of its cross-section, the material, the
in
way
course different for different combinations of the above conditions.
The
the
general equation of the elastic curve, whatever the shape of
beam may
be, the load, or
manner
any
of support
is,
d2 - y r
(If
Xz
= ^
Ht 1
To
y,
find the particular equation for
case,
M must be expressed
The
ordinate
in terms of
x and
the expression integrated twice.
which
is
the deflection, can then be found for any value of
x and
its
greatest value is the
maximum
deflection.
This integration
in practice,
has been performed for
all
the cases usually
met with
It is to
and the
results are tabulated in
Table
I.
be noted that
for stresses
this tabulation is for
beams
of uniform section
and
within the elastic limit.
Here, as in other classes of machine
members, the design of the part may be based on strength or stiffness, depending on the conditions, and in general both should be
considered.
Example.
ends
is
steel
/ beam 20
ft.
long and supported at the

lbs.
used as a track for a crane trolley carrying 4,000
Select a standard rolled /
beam
that will carry the load with a

at the centre
deflection of not
stress of not
more than
yV
lbs.
and a
maximum
more than 8,000 From Table I,

3"
PP
48
3
4,000
240
16
whence
*
4,000
E I 48 X X 240 X 16
X
3
30,000,000
48
The
expression
X
is
20<:.
30,000,000
sometimes called the modulus of the section and

It
is
generally indicated by the letter Z.
should he noted, however, that this ex-
pression
sections.
is
applicable only to symmetrical sections as e
may have two values
for other
* /
is
termed the resisting moment.

MACHINE DESIGN
42
From handbooks on structural shapes it is found that the moment of inertia of a 12" I beam weighing 31.5 lbs. per foot is Let such a beam be chosen. Then from formula J, the 215.8.
stress p
Me 2,000X10X12X6 = = r- =
/
205
is satisfactory.
. 7,000 lbs. nearly.
The
_.
section therefore
15.
refer
Beams to beams
of
Uniform Strength.
is
The
values in Table I
all
of uniform cross-section.
In nearly
cases the
bending moment, which

if,
usually the basis of design, varies and
therefore, the
beam
is
made
strong enough at
its
most strained
it
section
and uniform
in cross-section throughout its length
will
it
have an excess of material at every other section.*

is
Sometimes
desirable to have the cross-section uniform, while in other
cases the metal can be so distributed that every section shall
have the necessary strength

more.
carefully.
to resist the
bending moment and no

forms most usually met
In the latter cases the shearing stress must be looked after
Table II gives a few
of the
and an example may make their application clear. Example. A cantilever of rectangular section 30 inches long It is to have a unicarries at its outer end a load of 1,000 lbs. form thickness. What is its vertical outline so as to have uniform strength ?
Let the thickness = b and the variable height
= y.
Table
Then
2) is
the
moment
and
this
at
any section
at a distance
(Fig. I,
Px,
must be equal
to the resisting
moment
2
of the section at
each point, hence
Px =
which
is
- = pby
pi
e
or J y
6Px pb
is at
the equation of a parabola
whose vertex
let 6
the outer
end
let
of the
beam.
lbs.
In the problem assumed
= 1.5
inches and
/>
= 4,000
other
way
graphical
Then when # = 30, y = h = 5.$". In a similar points may be found or the curve may be laid out by method. The shearing load at any point is P, and
stress increases as the cross-section of the
hence the shearing
* This of course does not cover the possible case where the effect of shearing
or other stresses
may
exceed that due to flexure.
43
beam
is
decreases.
is
When v = o,
= o, and
in
general
small, y
very small;
therefore the outer
end of the
when x member
Refer-
must be modified so as
ence will be
to safely carry the shearing stress.
made
to this again
under the section dealing with

It is to
machine attachments
(see chap. 16).
be especially noted
that these theoretical shapes are based
on certain assumptions
and unless these are observed

lines
in the design, the theoretical out-
do not apply.
thickness of the
Thus in the cantilever example above, if the beam is not kept uniform the outline for uniform
uniform
strength
is
not a parabola.
when
the thickness is not
The mistake of using a parabola is often made when I ox T

It is
sections are used instead of uniform thickness or depth.
evident, that, whatever
means
of the
may be the form of section adopted, by bending moment and shearing load the correct
will
depth of section can be found for a number of points and a

curve plotted that
strength.
16. Combined Flexure and Torsion. Let the force P, Fig. 8, upon a rod with an arm a at a distance from the support equal
answer the requirements
of
uniform
act
to
/.
Then
(a)
the stresses induced in the section close to the support
are
flexure
due
to the
bending moment PI
twisting
direct load
(b)
(c)
torsion
" "
"
"
"
Pa
and equal
to
shearing " "

is
P
due
The
to
shearing stress
usually very small
compared
to that
bending and twisting, and can be neglected; the predominatis
ing stress therefore
that
due
to the
combined action
is
of the
bending and twisting moments.

It
can be shown that
if
a bar or rod
subjected to a longi-
tudinal tensile or compressive stress
and
at the
same time
to a
shearing stress at right angles to

these stresses
its
length, the combination of
may produce
similar stresses greater than either
and acting along planes other than those along which the original
stresses act.*
* Church's " Mechanics," page 317.
44
MACHINE DESIGN
TABLE
Diagram of Loads, Bending Moments and Shear
Greatest
BEAMS OF UNIFORM SECTION

Location Greatest Location
of
Bending
Maximum
Section
Moment
Deflection
of
Shearing where Shear Force is Max.
PI
PI 3
Any
3EI
PiM-P.is
ii
Px + P2
From
CtoB
-
>oooooocpA
w
1-
VI
wl =
8EI
wi 1
+P1
ElL3
wl
8j
+P
PI
Pt 3 At
107
0.451
BtoC
EI
from
+ PI'
CtoA
wl
128
> c
3wl wl 3 185 EI
8
* B
5wl
w=load per unit length.
W= total distributed
load.
P= concentrated
load.
45
TABLE IContinued
Diagram of Loads. Bending Momenta and Shear
Bending
Moment
Greatest Location Greatest Location of of Deflection
Maximum
Shearing Force
3ection
is
where Sheor Max.
PI
VII
1** ooo
AandB
12
Equal
E]
wl
3
At
VIII
^
^s
-
wl
At At
Wl
24
ooo
-I_
EI
0.4231
from
B
T
J
t.
<g>
IX
4)
PI
4
Pi
43
BtoC
CtroA
EI
PI-2
BtoC
-3
O
Pill;
5 -^
<D I
PI
At
C
CtoA
l
3IEI
Not Max.
a j
51
CO
* O
fl
C
.-I-,
g,3
PA
PI)
BtoC
Centre
to
24
[31
EI
2
XI
8-3 1
-4lf]
DtoA
ooorfoonnA
wl
1
swr
Centre
wl
Centre
3*4EI
AorB
xii
u'=load per unit length.
W= total distributed load. P= concentrated
load.
46
MACHINE DESIGN
TABLE 1 Continued
Diagram of Loads Bending Moments and Shear
Greatest
Loca-
Greatest
Loca-
Maximum
Shearing
Stress
Bending
tion of Deflection tion of
Moment
OOOOiOOOO
<P
A
(p+fw)x
[P+4]
J
.2
a
o
-a
P
"
XIII
a
s
02
48"EI
W
2
.
k-a>-
1_.
i
C
5
Pa
DtoC
XIV
bf
B D. E
m
A
^N
9Q00 ooooofo 9
?M
Wa
J
E
CandD
S5
XVI
at
131
"Ii
PI
A.B.
and C.
PI
192
EI
noooôooo A
r
1
Hi
Aor-B
12
384
wV EI
XVII
Phi;
at-B
s3
at^C
2PIilo
3EI(It2.l|)
at
XVIII
pM.
i
>=load per unit length.
TF= total
distributed load.
P= concentrated
load.
47
TABLE
Outline
of
II
BEAMS OF UNIFORM STRENGTH

Greatest
Beam
Loca Greatest Loca- Maximum Section Bending tion of Deflection tion of Shearing where Moment Stress Shear is Max.
*'&
Is
c3
3
+3
8P1 3
PI
o
<o3
Ebh
Pi
O
11
03
h o
.j,
Plan
5
SO
3
3
a
g
fcl
6P1 3
Pi
Ebh
Diagram
of
s3
Moments and Shear

same
as I
a;
<
a>
b'i e
0,1 o,
Diagrams
as in I
>.
ipoopooooc) /
fi r3
1
:
a O
2
w
when
1]*=1
;
Diagrams
as
ajp
in No. Ill Table I
t^MI
Diagrams same
as No.
!t!
o
i
fl-P
A $
PI 3
2Ebds
Table I
Centn
as No.
Diagrams same XII Table
2i=
I
r
VII
I--.
'
Centre
'
3P1 J
SEbd^
Diagrams same
as No.
IX Table
w=\oa.d per unit length.
W= total distributed
load.
P= concentrated
load.
48
If
/
MACHINE DESIGN
be the greatest
tensile or
direct tensile or
compressive stress and
the greatest direct shearing stress applied to the bar, then the
maximum
compressive stress p due to
and
s is
given
by the following equation
p=
P.
h[t
+ VT+~4?\
p due
s
......
and
s is
(i)
and the maximum shearing
stress
I
to
V~f+^~s2
(2)
It is
evident that the numerical value of p will always exceed
thai of p a and therefore if the material used has approximately the same tensile and shearing strength the design can be safely
based on
(1).
But should the allowable shearing strength of the

is
material be less than the tensile strength, as

it
usually the case,
may happen
that the shearing stress p s as found by (2)
would
(1).
dictate a larger section than that required

If the tensile stress is
by p as found by
due
to a
bending moment and the shearthe values of
ing stress
is
due
to a twisting
moment
be found from equations J and

obtained as above in equations
(1)
respectively
5 and / can and p and p
and
(2) respectively.
is
Example.
twisting
certain section of a circular cast iron shaft
subjected to a bending
moment
of 10,000 inch lbs.
and a
moment T
,
of 60,000 inch lbs.
The
It is
allowable tensile
stress p, is 2,000 lbs. per

stress
square inch and the allowable shearing

required to design
is
1,600
lbs.
per square inch.
the cross-section of the shaft.
From
.
/,
=
s
Me =
T
/
32
M
3
Td?
16
32
X
7t
10,000
=
100,000
_ , and from D.
.
.
,
Te
p
T = = *d
16
X X 60,000
- d3
r
d3
~J
3
d
2,000
hencefrom(i),/>,
3^0,000 = -~^
and
,
since
2,000
or d
=
.
5.55"
From(2),/>
300,000
v,"
and
,
ds
since re p
-
1,600,
d3
300.000 = ^
1,600

.\d = $.8" or
49
evident
greater than that given by
(1).
It is
that the last value should be taken.
Equations
(1)
and
(2)
are general,
and applicable
to
any and
all
sections, but for circular shafts operating
under conditions that

or ideal bending
as follows.
produce both bending and twisting
it
has been found convenient
to make use of what may be called an equivalent moment which may be derived from equation (1)
Let
Let
"
M = the
t
equivalent bending which
will
produce the
maximum
direct stress p.
M=
T =
=
the bending the twisting
moment producing the direct stress moment producing the shearing stress
/.
5.
" r
radius of shaft.
From
J,
and M
//
r
=
fe
-,
r
pi
and from D,
(Since I
p
T =
=
r
= 2l
for circular or other sections for
which the mo-
ments of inertia about two perpendicular axes are equal.)

Multiply equation
(1)
2
through by -, whence
P1
r
*P
2
7
,
lt
P
2
\
e
r
2
r
yi
-*
.-.--
pi
yM +
v M + T
2
2
.
(K)
In a similar manner an equivalent twisting
moment can be
deduced from
(2) thus,
2pJ
r
VM +T
2
2
. .
(KJ
The
quantities
M
x
and
for
are usually large
and the numerical

will
work involved
in solving
and
can be simplified by writing
M=x
T, where
any particular problem
be a known
quantity.
Whence
and
reduces
to,
M
K
{
y
2
T[x + v/?+"i]
....
(^2)
reduces
to,
T-Ty/j +1
(/Q
50
It is to
MACHINE DESIGN
be especially noted that
and
are equivalent
mo-
ments
in a
numerical sense only; that
is, if
a bending
moment
M and a twisting moment T are applied to a shaft,

tensile stress
t
producing a
and a shearing
stress s
respectively, then
is
a bending
moment which
will give
,
will give a stress
equal to the
maximum
a twisting
resultant
resultant tensile or compressive stress p,
and
Te
is
moment which
a stress equal to the
maximum
shearing stress ps reference being made to the same section. The application of these equations to the investigation of any
existing shaft subjected to a bending
moment
M and
a twisting
moment T
is
obvious, and
it
remains
to consider their applica-
It has been pointed out that by equation (i) does not necessarily indicate that a larger section will result from its adoption than would result from the use of equation (2). For the same obtained from K may reasons the greater numerical value of not give a larger section than would be obtained from T by applying K It is necessary therefore to determine under what
tion to the design of
new
shafts.
the greater numerical value given
conditions each should be used for designing in order that the
maximum
From
diameter of shaft shall be found in
all cases.
/,
M
d3
=
32
Whence
pnd V =V M =X
pI
r
3
32
16
....
(3)
In a similar way from
Since in any given problem
M and T are always known, M

K
s
and
can always be found from
and
(or
and
3)
and
p and p can always be assigned, the diameter of the shaft d can always be determined from both equations (3) and (4) and the larger value selected as in the problem previously solved. It is, however, desirable to know, for any
since the allowable values of
set of Conditions,
whether equation
(3)
or equation (4) will give the
greater value of d without the necessity of solving both equations.
1 ,

It is
evident that in order that equations
(3)
and
or
2
(4)
may give the
same diameter of shaft
^ must
e
equal
i
p
6
=
e
and
that for conditions other than these, either equation (3) or equation (4)
may
give the greater diameter.
It is therefore neces-
sary to
investigate the relations existing
between
and
for three sets of conditions.

(1) (2)
When
"
equations
(3)
and
(4) will
give equal values of d.
equation
(3) will
give the greatest value of d.
(1)
It
has already been shown that equations
(3)
and
(4) will
y. "
give
values of d equal H
when
P =
s
e '
2M
or
e
if
s 8
be called
then
p
2
2
T
J
2
is
VM +T M + VM +T
2
the simultaneous
the equation of a curve which expresses
all
values of
p
s
and
d.
T ~
e
for
which equations
(3)
and
(4) will
give
equal values of
The
value of either
M or T in equation
way
plot their ratio.
may
vary from zero to infinity and the most convenient

simultaneous values of
in
of plotting
If then,
and
is to
equation
(5),
the relations as given in
and
be substi-
tuted for those in
K and K
y
the equation
becomes
(6)
which
is
the equation of a curve expressing
all
the simultaneous
values of y (or
and x (or
for
which equations
and
K^
will give
It is
equal diameters of shaft.
desirable before plotting the curve to examine the limits
between which x and y may vary. It is clear that for x = o, and for T = o x= 00 hence the limits of x are o, and
,
M=o
<*>
52
MACHINE DESIGN
for
Using these same limits

that
M and
y
in equation (5)
it is
found
when
and when
M=
T=
o,
o,
M y = \ and M
=
1
and
T =
e
= T
That
is
for all materials
where the
1
ratio of allowable shearing
to tensile stress lies
between
neous values of
equal values of
is
M and T for
d.
which equations
and yi there are always simulta(3) and (4) will give

(6).
1
The
curve giving these simultaneous values
shown
(2)
in Fig. 9
and has been plotted from equation
If for
any given value of
within the limits

c f
and
/a
x
Field
r
K
2
M e~ 4U
F:
wx
-vr
eld of
K>
li-i
I- rV
.1
.2
.5
.6
.7
.8
.9
1.1
1.2
1.3 1.4 1.5 1.6 1.7 1.8 1.9
2.1 2.2
Fig. 9.
ratio of
M be
taken greater than the simultaneous value given
by the curve (or in other words if the co-ordinates chosen intersect above the curve) equation (3) will give the largest value of d.
For the value of
relatively to
can be increased only by making M greater

1
and an examination of K and K shows that inHence increases K more rapidly than it does K creasing in such cases K (or K 2 ) applies and equation (3) which is based
on them will give the largest value.

Further, for values of
P equal
to or greater
than unity, equait
P
tion (3) will also give the largest value of d.
For
has just been

this
shown
that
can never be
less
than
T
2
and only equals
53
when
M=
o.
For
all finite
it is
values of
M,
therefore,
(3)
M
is
must
(4)
be
greater than
-;
and
evident from equations

e -
and
that
for values of
= p
and
T equation M>
e
(3),
which
based on
(or
(3)
2)
will give the greatest value of d.
In a similar
l
-KIT
way
it
can be shown that for
all
simultaneous
values of
and -= which intersects below the curve and within T

P
the limits y
=i
and y =
>;
or for
is
all
materials where
Ps
is
less
than }4 } equation (4), which greatest value of d.
based on
P
x
(or
3 ),
will give the
Summary.
is
Equations
and
are the most convenient

It
forms of equivalent moments and will be used in this work.

to
be particularly noted that they are applicable only to circu-
lar or other sections
where the polar moment
of inertia is equal
to the
sum
of the rectangular
dicular axes (see page 50).

is
moments of inertia around perpenWhere the section to be considered
more complex the solution must be based on the original equations (1) and (2) in a similar manner to that employed in the example on page 49. Equation K 2 should be used where the
simultaneous values of
M intersect above the curve which and T

Ps
is
always the case whenever
Ps ->
1.
Equation
K
.
should be
used where the simultaneous values of
s !
and
P
the curve,
M intersect below T
//
which
is
always the case whenever
Ps -<|.
Example 1. An engine cylinder is i6"X24 diameter and stroke of 24"), steam pressure = 100
inch.
(piston
lbs.
16" in
per square
The
centre of the crank pin overhangs the centre of the

parallel to the axis of shaft).
main journal by 15" (measured
Assume
lbs.
that the pressure on the crank pin
may be
equal to 100
unbalanced pressure per square inch of the piston when

is
the connecting-rod
perpendicular to the crank radius.
Allow-
54
MACHINE DESIGN
ing 8,000 pounds as the
maximum
allowable direct stress and

stress,
6,400 as the
maximum
allowable shearing
compute the
diameter of the shaft.
Area
a
of piston
200
sq. inches;
radius of crank (arm of
maximum
twisting
moment)
100
= = =
12";
arm
of
bending moment
lbs.
15"
T=
200 200
15
M= M * = -=
By
X X
100
12
X X
12 15
240,000 inch
Also
300,000
6,400 p -s
1.25;
&
+
-.8--y.
.8
referring to Fig. 9
it is
seen that f or y
Ps = - =
and #
P
1.25 the ordinates intersect
above the curve, hence
should be
used.
From
2,
M
3 M d
= =
J [1.25
\/ 1.25 2
lbs.
1]
240,000
342,000 inch
17 From
/ \ (3), Vv3/ '
=
=
342,000
8,000
X 32 = X *
435
.'.
7-58^
is
Example
ing
2.
A circular cast iron shaft
subjected to a twist-
inch lbs.
moment of 250,000 inch lbs. and a bending moment of 62,500 The allowable tensile stress is 2,000 lbs. per sq. inch
stress 1,400 lbs.
and the allowable shearing

diameter of the shaft.
Determine the
Here y
A =
p
1,400
.7
and x
is
62,500
2,000
it
.25.
250,000
seen that for y
Ps - = =
.7
From
# =
the curve, Fig. 9,
and
.25 the intersection of the ordinates falls
below the curve,
hence
should be Used.
2
e
Then T = [v
d3
7T
(.25)
1]
250,000
y D
257,500
= I6X 257,500 X 1,400 9.75 inches.
Suppose, however, that
should be used.

Then,
55
J [.25
+ V(.25) +
2
-
1]
250,000
160,000
*=
.*.
32
l60 00 2,000
830
9.4 inches or .35" less than the value given
by
2.
A
9
(a)
convenient graphical solution of
and
is
shown
in Fig.
which may be used as follows:

3
For
the
make Oa =
axis.
unity;
lay off
Ob = x
to the
it
vertical
Draw
ab
extending
same beyond
scale
b
on
a
for
length
somewhat
greater than x;
2 2
ab.
= \/ Ob + Oa = Vx + hence T = a b X T. For equation K lay off be = Ob = x along Then ta + a b = ac = v + V.v + 1 Hence M = K [>c X T].
then ab
2
e
2,
the extension of
17.
Other
Formulas.
Equation K
for
is
sometimes
trans-
formed into an equivalent twisting moment.
Since in general
stress (that
it is
M
is,
=
pB
pi and
r
T =
PJ
an equal intensity of
section.
p)
T = 2M
for the
same
If therefore
con-
sidered
more
convenient to use an equivalent twisting
moment
instead of an equivalent bending

stitute for
moment
it is
allowable to subto the
(the
bending moment, equivalent

e
combined
bending and twisting moment), yi T (a twisting moment equivalent to the combined bending and twisting moments) provided
the
same allowable
.-.
direct stress is
used with
in solving for the
diameter of shaft.
r =
M + s/W+T
(K<)
56
Equations
, ,
MACHINE DESIGN
formula for
K 2 K 3 and K are all combined bending and

4
different
forms of Rankine's
twisting.
Other authorities
gives
(7)
give slightly different coefficients.
Thus Grashof
K
While others give
=
=
| M
VM
K
+ T
....
2
0.35
M + o .65 VM
2,
+ T
2
.
.
(8)
The diameter of not differ much from
shaft given by equations (7)
and
(8)
will
that given by
for
any
set of conditions,
except where the bending
moment
is
very small.
At the
limit
where the bending moment

,
is
equal to zero, Grashof's formula
gives a value of Af e 25 per cent greater than that given by

it
2.
But
may be noted
is less
that in general for
all
materials whose shearing

is is
strength
for
in
than their tensile strength (and this
the case
most materials used in engineering) that when

other words,
small
it
or,
is
when
the shearing stress predominates,
safer to use
in preference to
(7)
It will
(8)
be found that for
the range
where equations
and
be
give values greater than
K K
2,
that these values will
still
less
than those obtained from
or at least not enough greater to warrant the use of a different

in place of
,i 9
formula
K
is
Take
for
example
steel
where
.8
and x
in
which
down
close to the limit

to
where Grashof's
2.
formula gives the greatest value compared

terms of
Expressing d 3
as in equations (3)
and
3
(4),
from
K
K
y
^
N p
\
1.77
3
from
=
=
1.84 *
\ p
T
P*
from Grashof's formula

from which
it is
1.88 -J
seen that the difference between d as determined

is
by
and Grashof's formula
negligible.
The same
little
evidently
applies to equation (8) which differs but
from Grashof's.
As the value
lent bending
of x increases, the difference between these equivamoments decreases, and any variation is more than
covered by the factor of safety which must be used.

18.
57
of steamers
Combined Torsion and Compression. Propeller shafts and vertical shafts carrying considerable weight are subjected to combined twist and thrust. The span, or distance
is
between bearings,
action of the thrust
frequently so small that the shaft
may
be
considered as subjected to simple compression,

is
so far as the
concerned.
is
The
intensity of this compressive stress in such cases
in
which
If
P =
the thrust,
and d
the diameter of the (solid cir-
cular) shaft.
T =
the twisting
moment on
the shaft,
radius of the shaft,
the
polar
moment
then
of
inertia
(=
times the rectangular
moment
to T,
of inertia),
and
= =
the
2
the
intensity of shearing stress
due
T -
- *li ~ d
S
'
'
* IA. 4 I ^
l6
* d3
for solid circular shafts.
The
resultant
maximum
stresses are those
due
to the
com-
bined actions of a normal stress (compression) and a tangential

stress (shear)
as in the case of
(1)
(Art. 17);
hence equations
combined bending and twisting and (2) of the preceding article
apply and
may be
used to find the

if c
maximum
compressive or max5-
imum
shearing stress; or
be the compressive stress due to P,
be the shearing stress due to T, p c the

pressive stress,
maximum
resultant
com-
and p the
t
maximum
_j_
resultant shearing stress, then
_ c+
2
-vV 2
^ s
ancj p
/r +
4 ^ 5
58
MACHINE DESIGN
d for a given value of p c or p is much more convenient to
s
It is difficult to find the value of
from the above equations, and
it
assume a and p to
s
trial
see that they
diameter d and then check for the values of pc do not exceed the allowable compressive
and shearing
If the
stresses of the material
under consideration.
is
span of the shaft between bearings
so great that the
shaft
must be considered as a column
likely to buckle, the trial
diameter of the shaft

compressive
solving for
may be
shaft
taken so as to bring the
mean
after
stress c well
a
below the allowable value, and
pc and p the
20).
may
it
also
it
be checked as a long
is
column
stress
(Art.
In steel shafting
is
necessary usually to
apply equation (L) only, but

s
well to check the shearing
p against the allowable stress by applying (Z,,). Combined with Direct Stress. If the section XY, Fig. 10, be acted on by a force P at a distance from its
19. Flexure
gravity axis
equal to
a,
the stresses induced in the section will
be:
(a)
uniformly distributed
stress
will
due
to
the load
and
equal to
P
A.
per unit area.
This
be tensile or compressive,
depending on the direction of P.

(b)
flexural stress
due
to the
bending moment
a.
This
other.
flexural stress will be a tensile stress
on one side of the gravity
axis
which
is
at right angles to a,
and compressive on the
If the direct stress

tensile,
tensile.
induced
in the section
by the load
is is
then the flexural stress on the side toward the load

If the direct stress
induced
is is
compressive, the flexural

compressive-.
stress
on the
side
toward the load
The maxi-
mum
stress will
be the greatest algebraic sum of these combined
stresses at the outer fibres at stresses for
or F.
The
line
distribution of these
in Fig. 10,
both cases
is
shown graphically
above the
where
tensile stresses are plotted
UV
and the compress-
ive stress below;

stresses,
will
the ordinates
under
r s representing the flexural
and those under m n the show where the algebraic sum
direct stresses.
is
An
inspection
greatest.
In the case shown
the combined compressive or combined tensile stresses at
are
the greatest which
may come on
the section, depending
on the

direction of P.
will
59
This
if
is
not necessarily
so, as
a brief reflection
show
that
be located near enough
to
the reverse of
the above conditions
may
exist.
The form
of section
and location
of the gravity axis should be fixed with reference to the relative

tensile
Let
and compressive strength p' = the direct stress due
of the material used.

to P,
Let p"
Let p
= =
the tensile or compressive stress the
due
to
P a,
Y.
maximum
stress in the section at

f
X or
Then from formula A, p =

Therefore,
P ae P and from formula J,f =

,
f+
Pae*
-
(M)
Fig. io (a).
Fig. io (b).
where
used
e is the
distance from
is
to the outer fibre at either
or
Y, depending on which
is
under consideration.
central, but
If the material
equally strong in tension and compression the gravity
axis should not be far

it
from
where cast iron
is
used
is
advantageous
to
distribute
the
metal
more toward the
tension side,
thus drawing the gravity axis toward that side.

e
This increases
on the compression
side,
and hence increases the

p. 36;
*See Church's "Mechanics,"
6o
compressive
stress.
MACHINE DESIGN
It decreases e
stress.
on the tension
is
side,
and hence
decreases the tensile
Cast iron
much
stronger in
compression than in tension, and therefore a greater
moment can
be withstood by a given cross-sectional area when distributed in

this
manner.
not practicable, in general, to solve equation (M) for th&
It is
direct determination of the dimensions of a cross section to sus-
tain a given eccentric load

p,
with an assigned intensity of stress
because both A,
sions;
functions of the required dimenand with any but the simplest sections complicated funcI, e are
and
tions
result.
With
solid,
square,
or circular sections,
is
or in
general where only one dimension
unknown,
it is
possible to
reduce
M to a form which can be solved;

is
but the algebraic ex-
pression
to
a troublesome cubic equation.
The
practical
way
is
assume a trial section and check this for P or Example i, A small crane (Fig. n) has a inches. The section at m n is shown by Fig.
load corresponding to a
p.
clear swing of 28
11
(b)
Find the
maximum
fibre stress (compression) of
9,000 lbs. per square inch at n.
Pae P + I A
1-5
.-.
pAI n P_ I + Aae
3.
= 28 + 2=30 = _i_(2 X 64-
A = 2 X 4 1-5 X + 27) = 7.3

,
3.5
rp _
Example
22 inches.
7*3
9,000X3-5X7.3 =I o6olbs + 3-5X3 X 2

(Fig.
2. A punching machine Maximum force P acting
12) has a reach of
at the
punch
is
taken at
fibre
70,000
lbs.
Design the section
mn
so that the
maximum
stress at n (tension) shall be about 2,400 pounds per square inch, and check the compressive stress at m. The general form of section best adapted to this case is that shown in Fig. 12 (b). Taking the trial dimensions as in Fig. 12 (6), the neutral axis is found to be 8" from n.
.'.
22
4-
30.
It
is
also
found that
216 and
7.680.

at n,
61
P
and
P
*
=
at
70,000
70,000
30
216
325
2,200
2,525 lbs.
7,680
m,
70,000
-1-
70,000
'
30
11
216
tensile
7,680
slightly
25
+ 3> 000 =
2 > 6 75 lbs
The
stress is
greater than the limit assigned.
If this excess is
not considered permissible

is
a stronger section
must be taken. It massing the metal

compressive stress
evident that this can be accomplished by

further toward the tension side as the
still
is
very low.
Fig.
11.
Fig. 12.
The
above
tions
is
theory regarding the position of the neutral axis given

that in general use for such cases.
Recent investigabe-
have pointed out the fact that

curved beams.
this theory is not absolutely

it
rigid for
For the usual case of design

accurate.
Struts.
is
lieved that the
above
in
is sufficiently
20. Stresses
Columns or Long
When
short
bar
in
is
subjected to an axial compressive load the stress induced

is
each section
simple compression (see Art. 20), and the value

given by formula (A) or
of the stress
is
W
P.- J.
If,
however, the bar
is
more than 4
to 6 times as long as its least

will,
if
diameter, the above equation does not apply, as the bar
62
MACHINE DESIGN
will
proportioned as above, deflect laterally under the load and

ultimately break under a
and
lateral bending.
compound stress due to compression Such a member is called a column.

first
Theoretical equations for the design of columns were
developed by Euler.
Other formulae were
later
developed ex-
perimentally by Hodgkinson and Tredgold.
Gordon and Rankine
have also proposed equations

bers.
for the design of this class of
mem-
The
student
is
referred to any good treatise
on the Me-
chanics of Materials for a fuller discussion of these expressions
than can be given in
this
work.
Let
= P = I = A = P = p =
I
c '
the length of the
column
in inches,
the least radius of gyration of cross-section, the least
moment
of inertia of cross-section,
the area of the cross-section in square inches, the breaking load
the
mean
intensity
on the column in pounds, of stress under the breaking
load,
or the unit breaking load,
=P +A
c
pc =
the crushing strength of the material, or unit stress

at the yield point.
This
is
the
maximum
intensity
of stress in the
stress is
column when the mean
intensity of
',
c
n =
the factor of safety,

the working load on the
P =
p'
column
A,
in pounds,
-5-
n,
the
mean
intensity of
working
stress, or unit
working
load,
=
-sc
pi
n).
-r-
n =
This
-h
p =
the
intensity
of
working
is
stress
in
the
column
( = p
the
maximum
intensity of intensity of
stress in the
column when the mean
stress is p/.
m=
Then
a coefficient for the end conditions as shown in
Table
3.
is
Euler's formula for long columns

2
7T
EI
It is to
be especially noted that Euler's equation is rational and

elasticity.
deduced from the theory of
The
coefficient
m is
also

rational
will
63
and applicable
to other
forms of column
is strictly
formula.'.
As
be shown
later, the
equation
applicable only to
very long columns.
Very short compression members, of ductile material, fail under stresses corresponding to, or only slightly in excess of, the
TABLE
'VALUES OF
III
m FOR DIFFERENT FiND

CASE.II
"Pin Ended"
CONDITIONS
CASE. IV
"Square
CASE
CASE
III
Fixed at one
end, free at other.
Both, ends free,
"Pin & Square' One end nixed,

the other
guided.
but guided.
Ended" Both ends

fixed.
=.\
9/i
m=4'
'>///////////
apparent elastic
limit,
or yield point;
it
for
when
this stress is
reached the metal flows, although
does not actually break.
Very long columns may approximate the resistance as given by

Euler's
formula.
Columns
flexure
If
of
lengths
intermediate
between
compression members which yield by simple crushing and those
which
fail
by pure
stronger than the latter.

perfectly
weaker than the former and a column is initially exactly straight,

are
to
homogeneous, and subjected

is, if it is
an absolutely concentric
to
load (that
an ideal column) there seems
be no reason
why
its
strength should diminish rapidly with an increase of
length, other conditions remaining the same.
However, even an ideal very long column would reach the condition of unstable equilibrium
when subjected
is
to a certain critical
If the
load (the greatest load consistent with stability).

increased beyond this limit and a deflection
load
is
caused in any way,
the deflection will increase until the stress
due
to flexure pro-
duces failure of the column.
If a
deflection is caused while the
column
is
under a load
less
than this greatest load consistent
64
with
stability,
MACHINE DESIGN
the elasticity of the material tends to
make
the
column regain its normal form. Initial defects in the form or structure of a column or eccentric application of load tend to produce such a deflection; hence long
struts fail
under smaller loads

Or, in other words,
than short struts of similar material and cross-section, for the

ideal conditions are not realized in practice.
for equal safety
under a given load long columns must have a
greater
stress.*
terial,
cross-section,
Even
in
columns
and lower mean, or nominal, working of moderate length, if of ductile maif
the flow at the yield point causes buckling.
Merriman
as one
says that
the length of a compression

its least
only from four to six times
"diameter,"
it
member be may be treated

Johnson gives
as
which
will yield
by simple compression.
limits within
I
-f-
which the Euler formula should not be applied
p=
150 for pin-ended, and

of the
200 for square-ended columns.

different limits;
Other authorities give somewhat

agree that most
but nearly
all
columns
in ordinary structures
and ma-
chines are intermediate between simple compression
members
and those
great
to
which Euler's formulae apply.

formulae proposed.
is
There have been a

graphical represenin Fig. 13.
many column
tation of several of these formulae
shown
In this
diagram, abscissas represent ratios of the length of column to

the least radius of gyration of the cross-section, and the ordinates
represent the nominal (mean) intensity of compressive stress. Or,
x =
I -*-
\/ /
A, and y
p\
= P -
A.
The diagram is drawn for the ultimate resistance of pin-ended columns with a material having a crushing resistance, p (yield point) of 36,000 pounds per square inch, and a modulus of elasc
value of p f c is 36,000 for a very short compression member, and it is evident that a long column
ticity,
E, of 29,400,000.
The
could not be expected to have a greater strength; hence no
for-
mula should be used which would

*
give a value of
'
in excess of
Owing to the flexure of the long column, the stress is not uniform across the The maximum intensity of stress must be kept within the compressive strength of the material; hence the mean stress is less than for shorter compression members, in which the mean stress is more nearly equal to the maximum.
section.

the crushing resistance p c
.
65
will ap-
Referring to the diagram,
it
pear that the Euler formula (represented by the curve
EE E
1
2)
cannot apply
in
to
pin-ended columns
90.
If
(of this particular material)
which
I -*-
<
columns with a
ratio of
to p less
than
this limit yielded
by simple crushing, and those with a greater

formula, the straight line
ratio of
to p followed Euler's
FF
and
the curve
It is
F E E
1
would give the laws

p
for all lengths of columns.
not reasonable to expect such an abrupt change of law in

(/
-s-
passing this limit

of
90)
and, as already stated, columns

stress considerably less
moderate length
fail
under a mean
than
the simple crushing resistance of the material; or the strength

F
F,
!
>K l\
30
"T^k
A\
>i
^
20
^\B
R,N sRs
\H
sVi
V
\ ^K *\
X v^L?
f^
F
x = 1^-P
1
10
100
200
300
Fig. 13.
of
columns
is
inversely as
some function
of the length divided
by the "least diameter." Mr. Thomas H. Johnson has developed a formula which is based on the assumption that the strength of the column may be taken inversely as / -s- p. This expression is
tc
lc
(1)
in
which the
coefficient k has the value,
66
This formula
13.
It will
is
MACHINE DESIGN
represented by the straight line
is
THJ
in Fig.
be noted that this line
tangent to the Euler curve at
J v and the equation of the latter is to be used, should the columns exceed the length corresponding to this point of tangency
(/
+p >
150).
This expression
It is
is
very simple, after k has been
determined.
very convenient in making a large
number
of
computations for columns of any one material, and

in structural
it is
employed
work
to a considerable extent.
It
does not appear
to
have any advantage, on the ground of simplicity, when some
particular value of k does not apply to several computations.
For determination of nominal working

above)
f
stress,
pj
(as
computed
pc
-r-
may be divided by a suitable factor of safety, n. Or if n = p' the expression may be put in the following form for
}
direct
computation of mean working
stress.
,-L-*l-,-tJ_*i.L > 3?wj E n np

3
:
(2)
of
the
Professor J. B. Johnson has derived a formula from the results very careful experiments of Considere and Tetmajer.
is:
His formula
*-*-$*<
for
(3)
pin-ended columns.
The
is
curve
FB J
(Fig. 13) represents
this expression.
This curve
its
a parabola tangent to the Euler
curve,
and with
vertex in the axis of ordinates at F, the direct

material.
crushing stress of the
For columns having

to
-t-
p
x
greater than the value corresponding to the point of tangency
(should such be used), the Euler formula
is
be employed.
it
The formula
It gives
of Professor Johnson's
is
empirical, but
agrees
remarkably well with very refined experiments on breaking loads.

considerably higher values for allowable stress than other
generally accepted formulas, probably because
more
refined tests,
or
based upon from removed upon conditions further

it is
those in practice.
Professor Johnson says ("Materials of Construction," pages
301-302) that both Bauschinger and Tetmajer "mounted their
'

columns with cone or knife-edge bearings
axis,
67
at the
computed gravity
while
M. Considere mounted
his with lateral screw adjust-
ments, and arranged a very delicate electric contact at the side

so as to indicate a lateral deflection as small as 0.001
mm.
He
then applied moderate loads to the columns and adjusted the end
bearings until they stood under such loads rigidly vertical, with
no
lateral
It
movement whatever."*
this
would appear that
precaution tends to
make
the test one
of the material
and
not of a long strut;
for the eccentricity of
the load (relative to the nominal geometric axis) compensates, in a
measure, for the lack of homogeneity of the material.

correction been
if
Had
the
made under
greater load, the results of the tests,
would probably be still nearer the line F F iy and the difference between these test columns and columns as used in practice would be greater, requiring a higher contingency
plotted in Fig. 13,
factor in the latter for safety.
value of p e ' (as computed from the above form of Johnson's expression) should be
stress, the
For determining the working
divided by a suitable factor of safety n.
Or, the formula
may be
put in the following form for computing nominal working stress:
'--to'The Rankine
or
It
Gordon formula
(see
Church's "Mechanics,"
pages 372-376) has been extensively used for columns.
may
be expressed as follows
''
The above formula

strength of columns.
is
^- 77T7(Lf m
v
(i
,5>
'
'
based upon experiments on the breaking

coefficient P
it
The
is
purely empirical, and
this fact limits its usefulness, for

to
leaves
much
uncertainty as
how this coefficient should be modified for materials different from those which have been actually tested as columns. The
*
this
"This precaution is essential to a perfect test of the material. way can other sources of weakness be eliminated." [J. B. J.]
Only
in
68
MACHINE DESIGN
intensity of
f
mean
working
stress,
f
,
might be inferred by divi
ing p c by n, or the expression can be written
p
i
1 > (-) + \p/

m
\p
(6)
but
it is
not entirely satisfactory to assume the action for stresses
within the elastic limit, from the results of tests for breaking
strength.
The form
is
of the
Rankine expression
his "
is
rational,
but
the coefficient P
not.
Professor Merriman says, in

:
Mechanics
of Materials,"
page (129) "Several attempts have been made to establish a formula for columns which shall be theoretically correct. Thei most successful attempt is that of Ritter, who, in 1873, proposed
. .
the formula
m E
it
"The form
formula,
. .
.
of this
formula
is
the
same
as that of Rankine's
it
but
it
deserves a special
name because
by finding
com-
pletes the deduction of the latter formula
for p a value
which
is
closely correct
when
the stress
elastic limit
"
c
The above
notation
is
p does not exceed the changed to agree with'

-*c
that previously used in this. article. of safety.
The ratio p
this
is
the factor
For ultimate strength,
formula might be written:
mr?E \p
but the
first
2
form
13)
(eq.
is
N)
is
the
more important.
The
curve
R T R
x
(Fig.
the graphical representation of the last
expression, eq. iV 1 .*
Merriman
P
"*"
c
gives the Euler formula for a factor of safety of

is
P>
which
* Professor
Ritter.
Merriman developed equation Ni independently, but

the 1897
later
than
He
gives Ritter sole credit for the formula in
edition of his
"Mechanics
of Materials."

[Failure occurs
if
69
>_ p c
The
Ritter formula (eq.
N) reduces
to this last expression for
columns so long that the term unity

strictly
in the
denominator
is
negligible;
infinity.
speaking, this
is
only
the case
when
I -s-
Professor
mathematically, that the two curves,

tangent to each other
If
is
/-=-/>
EE E
x
Merriman also shows, T R 2 are 2 and R

X
,
when
-h
infinity.
= o,
the Ritter formula reduces to
p'=P-A, which
it
the ordinary formula for short compression
members.
gives the
The
lies
facts that this
formula
is
rational in form, that
correct values at the limits
wholly within the
= ooand l + p = o, and that it boundary F F E E (Fig. 13), all justify its

/-^-/>
x x
use,
and
it
will
be adopted in
this
work.
It will
be noted from
closely
Fig. 13 that the Ritter

for the
and Rankine formulas agree very
material taken for illustration;
but the fact that the
curve of the latter crosses the Euler curve near the right-hand
limit of the
diagram indicates that
its
constant
/?
is
not theo-
retically correct.
Exception
may
it
be taken to the use of the Ritter formula for

involves the use of the stress at the elastic
cast iron, since

limit,
and the
coefficient of elasticity,
definite fixed values for cast iron.

plies to the use of
both of which have no But the same criticism apelastic
any rational formula founded on the

iron
is
theory,
as
far
as cast
concerned.
Thus
the expres-
sions for deflection in simple

iron,
beams contain
which, for cast

Since cast-iron
in
may
it
vary from 15,000,000 to 20,000,000.
columns designed simply for strength are very rare

design
it
machine
therefore seems best to use the formula since otherwise
fulfils all
needs better than any other.
If
it is
desired to design a cast-iron

Pc
column with
great accuracy
values of
J-= -2 E
may be
taken which will give results in ac-
cordance with experiment and which practically transforms the

equation into Rankine's formula. ^
iron
If
m- r~= E
1
,
P L
q.
then for cast-
columns with
fixed
ends q
=
,000 5,0-
for
one end fixed and the
7o
other free but guided q
A
MACHINE DESIGN
I. 7 8
,
and
for
both ends free but guided
In addition the student should consult treatises on

5,000
the strength of materials treating fully of this subject.
All of the above formulas give the value of the
stress (p e
f
mean
ultimate
-h
=P
4c
A), or the
mean working
stress (p'
= P
A),
corresponding to a
maximum
working
stress p, respectively.
ultimate stress pc or a maximum However, the ordinary problem
Stress in 1000 Lbs.^jy

10
12
14
/
'/, //
A/ / / //
/ k
s*
X
H=4
Y
1
Jf
//
.
IV.
11
/
'//
i//t
%VAV 9 i I w \//B
1
10
//
v/ x A/ / // 7/ '/ys %>
III.
U=h
f
+p
"N
n.M-i
I.M =u
x=
20
40 60
80
100
120
140
Fig. 13 (a).
of design is to assign proper dimensions for the
member under
the given load.
It is
not practicable to solve directly, for the
area in such expressions as those given in this article as p' (or p) and p are both functions of the area of the cross-section. It is
usual to assume a section somewhat larger than that

for simple crushing,
demanded
and then
to
check for the ultimate load P,
or the working load P'.
Professor
W. N. Barnard
has devised
a diagram which
steel or
is
very convenient for these computations for

It is
wrought-iron columns.
shown, to a reduced
scale,
in Fig. 13 (a).
The
four curves are for the four end conditions

given in Table III, page 63.
They
are plotted for a

inch,
maximum
of
working
stress of 10,000
fy
'
pounds per square
and a value
a E
__
OOO
which
is
an average value for
steel.
The
curves
29,500,000
should not, however, be used for cast iron, wood, or other materials
where the
ratio
t-
will give values far different
from the
above, but such cases may be solved directly by equation N. They may be used for any other stress by proceeding as follows: Divide / by this Assume a trial cross-section, which fixes p.
value of p; take this quotient on the lower scale and pass directly
upward
to the proper curve for the given
end conditions;
then
is
pass horizontally to that one of the radiating diagonals which
numbered
point pass
to
correspond with the selected stress; from this

to the horizontal scale at the top of the
f
last
upward
diagram,
)
where the value of the unit load or mean working

read
off.*
stress (p
is
If this value of
agrees sufficiently well with the

trial area,
quotient of the load divided by the
the section
may be
considered as satisfactory.
In the case of a square-ended column, or when the supporting

action of the ends
is
equal in
all
possible planes of flexure,
it is
sufficient to take the least radius of gyration of the section;
or
to take p for the axis about
which the section
is
weakest.
In
case of a pin-ended column, as a connecting-rod, the cylindrical
supporting pins
make
it
equivalent to a square-ended column
against flexure in the plane of the axes of the pins, provided these
bear symmetrically with reference to the axis of the column

the
while
column
is
pin-ended with reference
to a
plane perpendicular
to the axes of the pins.
If the cross-section of
such a column
has equal dimensions in these two planes (circular, square sections,

etc.),
the
column need only be computed

I section, etc.),
it
for the latter

(elliptical,
plane.
If the
pin-ended column has an oblong section
rectangular but not square,
may be weaker
in
*
in
The method
is
of using the
diagram
is
indicated by the arrows, for an example
which /-^/>=8o and the
maximum working
stress= 14,000 (pin-ended).
In this
case, p'
found to be about 7,900.
72
either
of
MACHINE DESIGN
these
two
for
planes,
to
notwithstanding
the
difference
in
end conditions
relative
them;
and
it
may be
section
necesis
sary to compute
both planes, unless the
ob-
viously stronger in one of them.
If a rectangular, or elliptical,
in which the dimension in the plane more than one-half the dimension in the plane perpendicular to the pins, it will suffice to compute as a pin-ended column against flexure in the latter plane, and
column has a
the
section
of
pins
is
vice versa.
In the preceding discussion, the various formulae have been

given both for breaking and for working loads.
The Euler and

hence
Ritter formulae are derived from the theory of elasticity;
these are proper for computations pertaining to working loads, in
which the
stress
should never exceed the elastic limit.*
It
does not follow that these two rational formulas will agree with
experiments on the ultimate resistance of columns or for materials
which do not follow Hooke's law

strain.
of proportionality of stress to
These expressions
formulae.
are, in this respect, like the

J.
common
beam
Such formulae as Rankine's and
B. Johnson's,
derived from tests of ultimate resistance of columns, are, for

similar reasons, less rigidly applicable to working loads
stresses.
and
is
Example.
feet
The
is
connecting-rod
of
steam engine
per
the
long and
subjected to a load of 20,000 lbs.

stress
is
If the
maximum
rod.
lbs.
allowable
of
9,000
lbs.
sq.
in.,
deter-
mine the diameter

per
sq.
in.
a circular section
the
at
centre
of
the
Take = 30,000,000, and

Hence
elastic
limit
^ = 36,000
The rod may be
considered
a pin-ended
column.
If the
m = i.
20
rod were designed as a short column, the required area
would be
inches;
A =
it
000
9,000
is
2.2 sq.
ins.
or
a diameter
of
ifj-
and
evident that for a long column the diameter

this.
must be greater than

*
Assume 2]/ 2 inches
as a trial diameter.
The Euler formula
is
not applicable for practical applications, except for
quite long columns.

73
Then A =
4.9, p
=- = ^ o
4
60",
whence
in
JV
0,000
/
1
+
I
36,000
6O
4,3
2
7T
30,000,000
,\P = 4,300X4.9 = 21,070
lbs.
which
is
little
more than the
required load and the section will
fulfil
the requirements.
The
student should also follow the solution through

21. Eccentric
on the diagram.
In the preced-
Loading of Long Columns.

it
ing discussion of columns
has been assumed that the load has

is
been applied
axially.
This
obviously the best
way
of applying
at a dis-
the load, but cases often occur
where
it
must be applied
tance a from the axis of the column.

is
In such a case the column
said to carry an eccentric load,

If the length of the
eccentricity.
and the arm a is called the column be less than 4 or 6
times
its least
diameter, that
is, if
the ratio
be
less
than about
in para-
25, the
member may be
and formula
treated by the
method outlined
p
graph
If,
(10)
M will apply or
Pae P = + I A
-
however, the column be longer than 4 to 6 times

it
its least
diameter,
can no longer be assumed that the direct stress

uniformly distributed over the section, as
P
A.
it
due
to the load is
has been shown by the discussion on long columns that such

not the case.
is
In addition,
that the
if
the load
is
applied eccentrically,
it
it is
obvious
if
column
will deflect
somewhat more than

This
will
would
the
load were applied axially.

the original lever
flection.
have the
effect of
,
adding to
arm a an
therefore
additional
amount
due
to this de-
The
column
(a)
stresses
acting
on an eccentrically loaded
the
are
compressive stress p such as would be induced
if
load were axial.
74
(b)
MACHINE DESIGN
flexural stress
p 2 due
,
to the eccentricity
{a
and propor-
tional to the
bending moment
+ )
For the
first
from
Ritter's
formula (N)
and
for the second
from
(/)
(a
P h=
Therefore the
+
1
) e
P (a+
a) e
A?
is
maximum
compressive stress* in the section
pT
For columns whose
stresses
ratio
A
of
fl\ 2
/
(a+a)
e -\
is less
than ioo, and working
such as must be used in machine design, the deflection

neglected.
may be
For columns longer than
this,
or where the
stress is necessarily high, a

elasticity.
can be determined by the theory of
see
For a full discussion of the manner of computation Merriman's "Mechanics of Materials," 1905 edition, page For the ordinary cases of machine design this refinement 217.
omitted.
may be
Example.
of 10 inches.
circular
to carry a transformer
wooden pole 30 feet high is required weighing 800 pounds, with an eccentricity
the diameter at the middle in order
What must be
due
Let
that the stress
to this load shall not
exceed 500 pounds per
square inch?
1,500,000.
p = 3,000 pounds
c
.
per square inch and
E=
Also
m=
4
(See Table III.)
The
stress
induced on the convex side of an eccentrically loaded column

but
will
may
be either
cave side.
tensile or compressive,
always be
less
than the stress on the con-
For materials whose
elastic strength is
is
about the same in either tension

If,
or compression the stress on the convex side
of no importance.
however, the
whose tensile strength is much less than its compressive strength, the character and magnitude of the stress on the convex side should be investigated. If e' be the distance from the neutral
column
is
made
of a material, such as cast iron,
axis to the outer fibre
on the convex
side,
then the stress (p) on the convex side
is,
A
if
[_
^~E \p)
is
^~J
2
1S
P SItIVe
e Str6SS
1S
t6n "
sile;
negative the stress
is
compressive.
75
Assumed
a diameter of 8".
Then
/>
= 2 and A = 50
/^6o\
2
Whence p r =
800
50
r L
+
1
X
r
?,ooo 6
'
2
7T
(*) + X 1,500,000 V 2 /
-1
10X4I ^ 4 J
800
|_i
26
+ 10J
= 592 pounds
If this excess is
per square inch.
considered too great, a second approximation
can be made.
22. Stress
metals expand
Due to Change of Temperature. Practically all when heated, and contract again when cooled.
a bar expands per unit of length, for a rise
is
The amount which
of one degree in temperature,
called
its
coefficient of linear
expansion, and will be denoted by C.

values of
The
following table gives
for various substances for
one degree Fahrenheit

.0000074
.0000065
Hard
Steel
C=
C=
Soft Steel
Cast Iron
C=
C=
stresses are
;
.0000062
Wrought Iron
If
.0000068
it
a bar of metal
is
held at the ends, so as to prevent
from
in it which are same as though the bar had been compressed, or elongated, an amount corresponding to its expansion or contraction due to the change in
expanding or contracting,
produced
called temperature stresses
the effect being the
temperature.
Let
change
stress
in
temperature in degrees.
Let p
Since
induced per unit area.

t
E =
stress
P.j, n *j? = -.-. E p = C

C
/
strain
Example.
bar of wrought iron 2" square

its
is
raised to a
it
temperature of 100 degrees above
normal.
If held so that
it,
cannot expand, what stress will be induced in
and what
force
must oppose
it
to prevent
expansion?
76
MACHINE DESIGN
Let
E =
X
30,000,000
=C E =
t
.0000068
100
30,000,000
20,400
lbs.
and the
total
opposing force
P =
23. Resilience.
P will be 20,400 X 4 = 81,600 lbs.

all
In
the previous discussions on the vari-
ous straining actions to which a

applied without
member may be
or impulse.
subjected,
it
has been assumed that the load was a simple dead load and
initial
velocity
But,
as
it
already
pointed out, the load
may be
applied impulsively; or
may be
applied in any way, and removed and applied again and again
repeatedly.
The
application of a load in an impulsive manner,
or the repeated application of a load, does not affect the character of
the straining action, but does affect the amount of stress or
strain.
In order to more clearly discuss the

it
effect of
impulsive
effect of
loading
will
be necessary to consider the straining
load somewhat more fully; the discussion of repeated loads will
be given in a succeeding
If
section.
a material
is
distorted
doing a certain amount of

If the
by a straining action, it is capable of work as it recovers its original form.

elastic strain, this
deformation does not exceed the

is
amount
elastic
of
work
equal to the work done upon the material in producing

If the material is strained
such deformation.
limit, it returns
beyond the
in
work only equal

set, is
to that
expended
producing
elastic
deformation; and the energy required to cause the plastic

not recovered, as
it is
deformation, or
not stored but has

of form.
been expended
in producing such
permanent change
etc.,
Ordinary springs
illustrate the first case;
the shaping of ductile

are processes in
metals by forging, rolling, wire-drawing,
which nearly
all
of the energy is
expended
in
producing perma-
nent deformation.
The work
the
required to produce a strain in a

If the strain
member
is
is
called
work
of deformation.
produced
equal to
is
the deformation at the true elastic limit, the energy expended

called elastic resilience*
*
is
If the piece is
ruptured, the energy

elastic resilience
When
the term resilience
is
used without qualifying context,
to
be understood.

expended
II
77
in
breaking
it
is
called total
work
of deformation.
a d
e (Fig. 6) is
the stress-strain diagram for a given mate-
rial,
the area
a a' represents the elastic resilience, and
a d
e e'
represents the total

material.
work
of deformation per cubic inch of the
In such materials as have well-marked elastic limits (proportionality

line
between
stress
and
strain
through a definite range) the
Oa is a sensibly Oaa' = âa'xOa'; or,
straight
line,
and the
elastic
resilience
the elastic resilience equals the elastic strain
(Oa') multiplied by one-half the elastic stress
(%
aa').
The area
Oadee' equals the base (Oe') multiplied by the

of the curve Oade;
or, if
the quotient of
mean ordinate (y) this mean ordinate of

called k,
the curve divided
by the maximum ordinate be

stress.
the
work
action
of deformation equals the ultimate strain multiplied
by k
times the
maximum
It is
evident that for a straining
beyond the
curve
elastic limit, k
>
X and k
<
1.
The
former.
OADEE'
represents the stress-strain diagram of a
material having higher elastic and ultimate strength than the
The
greater inclination of the elastic line (OA) with
the axis of strain

of elasticity,
as
(OX) shows, in the second case, a higher modulus this modulus equals the elastic stress divided
In the
first
the by J
elastic
strain.
case
E =
x
a'
7-. in
the
sec-
Oa'
ond
case,
E =
2
A A'
OA'
diagram
The
rials
stress-strain
OADEE'
and
shows that
of
two mate-
one
still
may have
have
curve
both the higher elastic and ultimate strength,

less total
and
tion.
less elastic resilience
work
of deforma-
If the
a" d" e"
is
the stress-strain diagram of a

of
elasticity
third material
first), it
(having a modulus
similar to the
appears that
this third material possesses greater elastic
resilience,
but
less total
work
of deformation than the
first.
A
more
comparison of these
illustrative stress-strain
diagrams
(for
quite different materials) also shows that, for a given stress, the
ductile, less rigid material
may have
the greater resilience.
Hence, when a member must absorb considerable energy, as in
78
MACHINE DESIGN
weak
yielding material
case of severe shock, a comparatively
may
be safer than a stronger,
stiffer
material.
This
is
frequently
similar to
recognized in drawing specifications.
The
principle
is
that involved in the use of springs to avoid
undue
stress
from
shock.
In fact springs
differ
from the so-called

load.
rigid
members
only in the degree of distortions under loads, or in having

greater resilience for a given
If a material is strained
much
f
maximum
beyond
its elastic limit,
as to a
(Fig.
upon removal of the load it will be found to have such a permanent set as O O'. Upon again applying load, its elastic curve will be 0' a'; but beyond the point a' its stress-strain diagram will fall in with the curve which would have been produced by continuing the first test (i.e., a'de). Similarly, if loaded to a" the permanent set is O O", and upon again applying load, the stress-strain diagram becomes O" a" d e. The
13 b),
,
elastic limit a" of the overstrained material is evidently higher
than the original

deformation,
elastic limit, a;
e, is
while the original total work of
a d
considerably greater than the total
work
of deformation of the overstrained material,

effects of strain
O"
a" d
e.
The
beyond the
elastic limit are
thus seen to be
I.
Elevation of the elastic strength and increase of the elastic
resilience.
II.
Reduction of the
facts
total
work
of deformation.
to re-
These
have an important influence on resistance
peated shock.
overstraining
The above noted
elevation of the elastic limit
can usually be largely or
by wholly removed by
It will
annealing.
24. Suddenly applied Load, Impact, Shock.
perhaps
be well
the
to first consider the general case of

initial velocity;
a load impinging on
member, with an
fall
this velocity (v) correspondh.
ing to a free
through the height
For
simplicity, the dis-
cussion will be confined to a load producing a tensile stress; but

the formulae will apply equally well to uniform compressive
and
shearing stresses, and

torsion
all
except
(5)
apply directly to cases of
and
flexure.
79
W=
d
static
value of load applied to member.
h = height
corresponding to velocity with which load
is
applied.
= total distortion of member due to impulsive load. p = maximum intensity of resulting stress. A = area of cross-section of the member. P = p A = total max. stress due to load as applied suddenly. X = total distortion of member due to static load, W.
x= ft ^r X
&=a
(for convenience).
its
constant;
value
is
E. L.
is
exceeded k
K > y
2
if
E. L.
is
not passed; but

i.
if
and k
<
1
1
--W-'--*
.
.
-J^^-
Fig. 13
(c).
Fig. 13 (b).
The energy
k Po.
to
be absorbed by the
is
member due
the
to the impulsive
is
application of the load
(h + d);
work
of deformation
(See preceding article, Resilience.)

I.
Case
Maximum Stress within Elastic Limit.

+
>>)
W (h
n = P
ps
=HP&
(1)
d:X::P:W
w
=
(2)
2Wh W + _
2
2W
. px
2W
(3)

80
2
MACHINE DESIGN
2 W h =T- +2WP
.'.P 2
.'.P
= W(l +
2X)
A/I
2X)
(4)
P
d
+ -J=-J(l+Vl A
=
A (1
P_
(5)
2 x)
(6)
The
limit.
elongation at the elastic limit equals

of elasticity
F+E,
in
which
E=
modulus
If
and F = intensity
of stress at the elastic
(A
L = length of - L) (F - E)
:
the
:
member,
A)
:
(W +
F;
.'.
WL
-^
E.
(7)
As
will
is
small for metals (except in the forms of springs) a
moderate impinging velocity

be evident that
A
may produce
very severe
stress.
It
and
8 are directly
proportional to the length
of the
member; hence
is
the stress produced
by a given
velocity of
impact (height h)
sible.
reduced by using as long a
member
as pos-
If the
ity,
load
is
applied instantaneously, but without
initial veloc-
h = o and x = o; whence
p =
w
p
(1
+ vTTo) =
2
(30
*=a=-T__
= + V + Case II. Maximum Stress Beyond the Elastic Limit,
d
<s0
2 A
x (1
o)
(60
If the
maximum
tion (1) is
stress exceeds the elastic limit, the constant k of
equaexact
between
% and
(see Art. 23, Resilience),
and
its
value cannot be determined in the absence of the stress-strain dia-
gram
is
for the particular material.
Thus
(Fig. 13 c),
{h +
8),
represented by the rectangle
m n ca;
work
is
and
this area
must equal
resili-
the area
ence,
a b
c;
the latter being greater than the elastic

less
a',
and
than the
total
of deformation
O ade e',
in this illustration.
When
(a)
the stress-strain diagram

:
known, the following prob-
lems can be readily solved
Determination of the velocity of impinging of a given load

stress, or strain.
(or
corresponding value of h) to produce a given

(b)
Determination of the load which

Determination of the
will
produce any parvelocity.
ticular stress, or strain,

(c)
when impinging with a given

stress, or strain,
produced by a given
to
load impinging with a given velocity.
Let the work of deformation corresponding

stress,
the
d.
known
If the
or strain,
in
(a)
is
and
(b),
be called
R=k P
and member, let be the load per unit of sectional area; W the height due the velocity of impinging divided by the total acting length of the member; 8' the distortion per unit of length of the member due to impulsive load; and R r the resilience for unit of volume, or the modulus of resilience.
stress-strain
diagram
for stress per unit of sectional area
strain per unit of length of the
(a)
W
w
:
(h'
V)
kp
d'
R'.
h'
J-,
(7)
W-
'
= JPTs
The
solution of this problem
is
(8)
(c)
not quite so definite, in

it
the general case, as the preceding;
but
can be easily accom-
plished, graphically, with sufficient accuracy.

(Fig. 13 c)
it
Draw
the line g q
(indefinitely), parallel to
f
,
and
at a distance
from
equal to
W\
take out the
value of o, the shaded area
area/zg = gOt. Whatever the <$'; hence the unO c qfi gO =
shaded area under the

few
stress-strain curve
must equal
hf
trials will suffice to locate
give/i
qf =
m O =
11
t
the limiting line b q c which will
h'.
The
limit
to
is
case in which the
maximum
stress is within the elastic

it is
by
far the
most important, as
almost always desired
keep the
maximum
intensity of stress,
-f-
A, within the
elastic limit,
especially as
every overstrain (beyond this limit)

total resilience (see Fig.
raises the elastic limit

13).
and decreases the
The
effect of a
is
shock which strains a

its
member beyond
for
the
elastic
limit
to
reduce
margin of safety
its
subsequent
similar loads,
because of reduction in
cause the
ultimate resilience.
Numerous
actions
successive reductions of the total resilience

finally
by such
may
6
member
to
break under a load
which
it
has often previously sustained.
82
MACHINE DESIGN
No
doubt many cases of
failure
is
effects just discussed;
but there
can be accounted for by the another and quite different

is
kind of deterioration of material, which

ing article.
treated in the follow-
Dr. Thurston has shown that the prolonged application of a
dead load may produce rupture, in time, with an intensity of stress considerably below the ordinary static ultimate strength
but above the
ble time
is
elastic stress.
It is well
known
that an appreciait
necessary for a ductile metal to flow, as

is
does flow
when
show
its
section
changed under
stress;
hence, a test piece will
by quickly applying the load more slowly, provided the application of load is not so rapid as to become impulsive. The kind of failure which is the subject of the next topic is due to a real permanent deterioration of the metal, and it is due to distinctly different causes from those mentioned above.
greater apparent strength
it
than by applying
25.
It
On
the Peculiar Action of Live Load.
Fatigue of Metals.
has been found by experience and experiment, that materials

to
which are subjected
continuous variation of load cannot be

if
depended upon
to resist as great stress as they will carry
applied
but once, or only a few times.
When
the load
is
suddenly applied,
and frequently repeated, the decline

endurance
of strength or of the
power
of
may perhaps be
ascribed, in part at least, to the elevaof the ultimate resilience, as

this cause,
tion of the elastic limit
and reduction
discussed in Art. 24.
But apart from
with repeated
loads, even in the absence of appreciable shock, a decided de-
terioration of the material very frequently occurs.
This
effect
has been called the Fatigue of Materials, although some authorities

restrict this
term
to the
kind of deterioration already referred to as
the simple result of a decrease of resilience.
The term
fatigue
implies a weakening of the material due to a general change of

structure.
It
was formerly supposed
that the repeated variation
of stress caused such change of the general structure, possibly
owing
to
slight
departure from perfect elasticity under stress
much below
crystalline
that ordinarily designated as the elastic limit.

this view; this
The
but
appearance of the fracture sustained

tests of pieces
numerous
from a member ruptured in
way,

(taken as near as possible to the break)
fracture,
,
S3
and
it is
difficult to
test pieces
show such crystalline reconcile the normal appearance and

fail to
behavior of such
structure.
with the theory of general change of

that every piece
theory which has been largely accepted
is
of metal contains
innumerable minute flaws or imperfections,
often originally too small to be detected by ordinary means.
These
" micro-flaws " tend to extend across the section .under variation
of stress,
and may,
in time,
reduce the net sound section so greatly
that the intensity of stress in the fibres
which remain
intact be-
comes equal to the normal breaking strength of the material. Professor Johnson suggests: "the gradual fracture of metals" as a more appropriate term than " fatigue." Many men of large practical experience still prefer wrought iron to mild steel for various members which are subject to constantly reversing stress.
It
is
probable that the prejudice against
steel is largely the
more sensitive material; and the product of the best steel makers of to-day is much stronger and more reliable than wrought iron. However, it is just possible that the very lack of homogeneity in wrought iron renders it safer under varying stress {other things being equal) as the fibres are more or less separated by the streaks of slag, and a flaw is less apt to extend across the entire section than it is in the continuous structure of steel. Wrought iron may be likened to a wire rope, in which a fracture in one wire does
result
of unskilful
manipulation of
this
not directly extend to adjacent wires.
The "gradual
seems
to
fracture" through extension of "micro-flaws"
accord with the observed facts more closely than the
older theory of general change of structure.
In the
American Machinist
(Sept. 27, 1906) will
be found an
account of recent researches tending to show that metals are
made up of grains, each grain consisting of many crystals, and that when deformation takes place in a metal these crystals move
relatively to
each other along

is
gliding planes."
If the stress
producing such sliding

the gliding planes
series of
repeated often enough the contact at

into a crack or
weakens and finally passes cracks which extend across the section.
84
MACHINE DESIGN
The
theory of the subject
is,
as yet, too incomplete to permit
of derivation of rational formulae to account for the effects of re-
peated
is
live loads;
and
if
the "micro-flaw" theory
is
correct, it
not probable that such rational analysis can ever be satisfacAll of the formulae that have been derived for computation of
torily applied.
breaking strength under
known
variations of load, or stress, are

fit
empirical ones which have been adjusted to

ally
the experiment-
determined
facts.
Consult: Johnson's "Materials of Construction."
Merriman's "Mechanics of Materials."
Unwin's "Testing of Materials."
Weyrauch
Steel."
(Du
Bois)
" Structure
of
Iron
and
Experiment has shown that the breaking strength under

peated loading, or the "carrying strength,"
is
re-
a function of the
magnitude of the variation

tions
of stress
and
of the
number
this
of repetiis
of
such
varying
stress.
Furthermore,
function
different for different materials;
tions
rials,
on record which go
to
and there are authentic observashow that, as between different mateIn

is
the one with the higher static breaking strength does not
always possess the greater endurance under repeated loading.

general, however, the carrying strength under repeated loads
function of the static strength.
The
number
(b)
allowable working stress usually depends upon

of applications of the load.
(a)
The
This should be considered
as indefinite, or practically infinite, in
many machine members, The range of load. This is frequently either from zero to maximum; or between equal plus and minus values, (c) The
breaking strength or the
first
static
elastic strength.
The
systematic experiments upon the effect of repeated
loading were conducted by

for example,
sile
Wohler [1859
to 1870].
He
found,
that
a bar of wrought iron,
subjected to ten-
stress
varying from zero to the
maximum, was ruptured
by:

800 repetitions from o
107,000
" "
" "
to 52,800 lbs. per sq. in.
85
o to 48,000
o to 39,000
"
"
450,000
10,140,000
"
to
35,000
Merriman, page 191.
up to something less than the elastic limit an indefinite number of times (several millions) before rupture occurred; but with complete reversal of stress, or alternate equal and opposite stresses, (tension and compression), it could be broken, by a sufficient
It
was
found that the stress could be varied from zero
60000
- - ' '^'^^jS
40000^
s'yp>^
^'^^ y/
D
^,~Z*^*^
S>
20000
r
I
/
O
Tension Compression.
/
number
of
20000
Fig. 13 (d).
applications,
when
the
maximum
to
stress
was only
about one-half
to two-thirds the stress at the elastic limit.
A number of efforts
of
have been made
deduce from the experto the design
iments of Wohler, formulae which could be applied
machine members
is
(see
Unwin, page
it is
36).
One
it
of the best of
easily applied
these formulae
that of Professor
Johnson as
is
to all cases that will arise;
simpler than most of those previ-
ously proposed; and
it is
probably as reliable as any yet offered.

Launhardt's for
Two
formulae which have been very generally accepted for

are:
computing the probable carrying strength
86
MACHINE DESIGN
varying stress of one kind only, and Weyrauch's for stress which
changes
60,000
sign.
Suppose a material
lbs.
to
have a
static ultimate strength
of
per sq.
in.
If the
minimum
unit strength be plotted
as a straight line,
A OB
is,
(Fig. 13 d), the locus of

is
the
maximum
tensile
unit stress, from the
Launhardt formula,
for example,
the broken curve
from
to
D.
.That
when
the
minimum
stress is 12,500, the
maximum
tensile carrying stress
about 40,000;
indefinite
or the material could be expected to

of loadings
if
would be stand an
In a
number
the range of stress did not
exceed 15,000 to 40,000 pounds per square inch in tension.

similar way, the
broken curve from
to
is
the locus of maxiof
mum
line
tension,
from the Weyrauch formula, when the locus

compression)
line
is
minimum
stress (negative tension, or

It will
the straight
O.
appear that the straight
CDB
it
agrees
fairly well
with these two curves.
Inasmuch
is
as
seems un-
reasonable to expect an abrupt change of law

stress passes
when
the
minimum
for
through zero, and as there
no rational basis
Launhardt and Weyrauch formulae, it appears reasonable to adopt the upper straight line as the locus of the maximum stress.
the
Owing
to the discrepancies in the observations
(which must be
expected from the probable cause of the deterioration of the

metal), this straight line
may be
accepted as representing the

line.
law as accurately as could be expected of any empirical
These
are, in substance, the
reasons given by Professor Johnson

line
for basing his
formula on the straight
C D B.
For
full dis-
cussion and derivation of the following formula, see Johnson's

" Materials of Construction," pages 545-547.
Let p2
maximum intensity of stress. p = minimum intensity of stress. u = ultimate (static) intensity of stress.
x
Then
in general
h=
As
~p[
(l)
the expressions contain the ratio of the
minimum
to maxi-
mum intensities of stress,
instead of their difference, they are ap-

plicable
for
87
when
the area of cross-section of the
member is unknown;
same
as the
whatever
this area, the ratio of the stresses is the
ratio of the loads
producing these
stresses.
In substituting values
of
p and p 2 care must be taken

x
,
to use
proper signs;
is
thus,
if
tension
is
taken as positive, compression
negative;
is
or, if the
stress varies
/>,
between tension and compression p 2
positive
and
is
negative.
x
For dead load, p
= p
P2
For repeated load when p
o,
p =
-
.,A-;-X.
For complete reversal
of load,
!.....
p
x
(3)
-^ pz
(
^T^-"^*"*"
2
y
2
y
2
P2
The
three special cases (2), (3),

in designing,
of.
and
(4),
are those most
com(1)
monly met with

Example.
is
but the
general expression
should not be lost sight
bar of
steel,
whose ultimate
static tensile
strength
70,000 lbs. per sq. inch, is subjected to a
repeated load whose

value.
minimum maximum
number
value
is
one half the
maximum
What
is
the
stress that
can be carried by the bar for an indefinite

P
of repetitions?
to the load
Since the stress will be proportional
p1
Hence substituting
....
in equation (1),
p.,
A
l
x
I
70,000
P2
47,000.
2
It
is
2p
to
be noted that the allowable
maximum
if
stress is
above the
original elastic limit of

to
most
steel,
and
the piece were designed
be stressed
to
47,000
lbs. the
result
would be that the
first
application of the load would raise the elastic limit to that value.
88
MACHINE DESIGN
of in
But the piece would take permanent set and be in most cases no further use. A factor of safety must therefore be used
order that the
limit.
maximum
stress
may be
well below the elastic
The
rials;
experiments of Wohler, and his successor in the
field,
Baushinger, were conducted on a very limited variety of mateso that while the above discussion points out
what may be
expected in a general
ficiently
factor of
way from most materials, they are not sufconclusive to make it possible to pick out the exact safety to be used in all cases. They do, however, throw
on the apparently high factors of safety which must for which no other satisfactory explanaof Safety.
much
light
sometimes be used, and

tion has
26.
been found.
The Factor
The
preceding paragraphs
(arti-
cles 9 to 26)
have considered the
effect that different
methods
stress
is
of applying the load will
have on a member, and the relations

the load
which
exist
between a given dead load and the resulting

It
and
strain.
has been shown in Art. 24 that
if
applied suddenly the resulting stress and strain will be twice as

great as for a dead load.
And
finally in
Art. 25
it
has been
shown
that the
maximum
stress that
can with safety be induced

It
repeatedly in a member, will depend on the range of stress.
would seem as though a member designed

these logical theories
in
accordance with
would be
satisfactory.
But
it
must be
very
remembered
incomplete;
that these theories are not absolute, that the inis still
formation regarding the characteristics of materials

that flaws
and hidden defects always exist; and finally that there is always danger of accidental overloading. In addition, it is generally essential that a machine member be not only strong enough to avoid breaking under the regular
maximum working
permanent
if it
load, but also that
it
shall not receive
set;
for a
machine member ordinarily becomes useless

it
takes such set after
has been given the required form.
In
many
cases a temporary strain, even considerably below that

limit,
corresponding to the elastic
would
seriously impair the
accuracy of operation;
and
in such cases the
member
often reIt
quires great excess of strength to secure sufficient rigidity.

follows, therefore,
89
from these considerations that

the
if
the design of
a machine
member were based on
maximum
allowable stress,
as indicated
fied
by Wohler's experiments (such
stress
being modi-
by the theory of suddenly applied loading, should it be would be no margin to allow for the uncertainties and unknown defects enumerated above; and in many cases leave no assurance that the elastic limit would not be expresent), there
ceeded.
theories
So that while stresses fixed in accordance with these
form a good
basis, they
must
in general
be reduced by
means
of a factor of safety so that the
working
stress is
enough
lower to provide for these uncertainties.
The
factor of safety
is
generally defined as the quotient of the
ultimate static strength divided by the working stress.
con-
sideration of Wohler's experiments shows that such a definition

is
misleading.
For a factor of safety of
2,
for instance,
might
with
be perfectly safe for a dead load; but

stress in
for a repeated load
one direction
it
would leave no margin

factor of safety
at all for continto
gencies.
The apparent
and the
would seem
be a
better term,
real factor of safety
quotient of the carrying strength, or

as given
may be defined as the maximum allowable stress

stress.
by Wohler's experiments, divided by the working

too often applied,
is
The
and, as
factor of safety has been called the "factor of ignorance,"

it
is.
it
perhaps
little
else.
Thus
shall
very often
it is
specified that all the
members
of a
machine
be designed with a certain fixed factor of safety without regard

to the conditions
act.
under which the various members

applied in this
It is
may have
is,
to
factor of safety
manner
generally
speaking, a factor of ignorance.

of safety will always retain
probable that the factor

it
an element of ignorance, for
can
hardly be hoped that the powers of analysis will ever permit the
prediction of the exact effect of every possible straining action,
due
to regular service
and accident.
Neither can
it
be expected
that the
methods
of manufacture,
and inspection,
will
become
so
perfect as to eliminate or
in materials or
measure precisely every possible defect
workmanship.
(at
But a careful study

least
of the condito the effects
tions of each particular case
and a proper attention
which may be weighed
approximately) should, with
90
the knowledge
fairly
MACHINE DESIGN
now
is
to
be had, enable the designer
to
make
accurate application of the factor of safety, an intelligent

the most important part of design.
choice of which
Most
of the formulae of
Mechanics which are applicable
to the
design of machine members, are based on theoretical treatment
induced by the action of given forces within the upon the member under consideration; and the theoretical conclusions so reached are amply verified by practical experiment. When, therefore, the conditions under which the member is to work can be analyzed, and the laws of Mechanics applied to its design, such methods as outlined in this chapter
of the stresses
elastic
limit
are perfectly rational,

gencies.
if
intelligent allowance is
made
for contin-
Many machine members,

and
less satisfactory
however, are subjected to

that analysis cannot be
such a complicated system of

strictly applied,
stress
approximations or assump-
tions are unavoidable in the present state of knowledge.
When
such
is
the case, the designer must either base the design on the
stress, if
predominating
there
is
such, allowing such a margin or
factor of safety as experience or experiment
may
show, to pro-
vide for the minor uncertain stresses; or,
if
the case considered

to
be beyond such treatment, recourse must be had
empirical
methods or judgment.
as a
(See Art.
i.)
While therefore mathematical treatment of any case
will serve
good guide
to correct proportions,
such treatment must

other branches
always be tempered with judgment,
a high development of
all
which
is
necessary to successful design, as in
of engineering.
While,
also,
no
fixed
rules
for
selecting
the
factor
of
safety can be laid
down, a knowledge of Wohler's experiments,

suddenly applied loads, will greatly aid the
and the
to
effect of
designer in the matter.
Thus when
it is
known
that the load
is
be a dead load, an apparent factor of safety of 3 will, for wrought iron, or steel, bring the working stress well below the
elastic limit
and allow something
for contingencies.
If,
however,
to a
the load be a repeated load, the stress varying
from zero
maximum
must
tensile stress, the apparent factor of safety for steel
at least
be
5, to
allow a good margin below the elastic limit;

and
in either case,
if
91
in addition the load is to

2 to
be suddenly applied,
these factors
must be multiplied by
insure safety.
Example.
steel
beam
is
subjected to a suddenly applied
load which alternately induces an equal tensile and compressive

stress;
if
the ultimate strength be 60,000 lbs. per sq.
in.,
what
apparent factor of safety should be used, and what

real factor of safety?
will
be the
Since the stress
is
a reversed one, the
maximum
allowable
stress or carrying strength is
of
by Wohler's experiments one-third If the working stress is oneultimate strength or 20,000 lbs.
it
half of this value or 10,000 lbs.,
will leave a
good margin
(10,000)
for
contingencies, disregarding the impulsive effect.

is
But the load
applied suddenly and, by Art. 24, this value

2,
must
be again divided by
sq. in
making the working
stress 5,000 lbs. per

is
Therefore the apparent factor of safety
60,000
12,
5,000 20,000
while the real factor based on Wohler's law

If the
is
4-
5,000
member should have

or
if
to
work under extremely trying

still
conditions,
shock or other stresses which could not be

further
analyzed were present, this value might have to be

reduced.
TABLE
IV
FACTORS OF SAFETY
Repeated
Stress in One Direction.
Repeated Reversed Stress.

*T3
Character of
Material.
Dead
Load.
Sag
(A
Sag
CO
Wrought
or
Iron, Steel,
other
Ductile
3
5
Metals
Cast Iron, or other Brittle Metals
IO
12
12
10
20
Table 4 contains factors of safety such as are used
in practice
and which agree

are, of course,
fairly well
with the foregoing theory.
They
average values and must be used with judgment;
92
MACHINE DESIGN
its
but in the absence of trained judgment, or as an aid to

velopment, they
de-
may be found
useful.
Table
contains values of the ultimate strengths and elastic
limits of the materials
most used
in engineering.
They,
also, are
average values such as the designer must use in the absence of

exact information regarding the material to be employed, and in
general such exact information

It
is
lacking.
may be
observed that an increased factor of safety
may
not
If
always in the case of cast metals give a stronger member.

results, the gain in strength
the increased dimensions give sections so thick that sponginess
may be
negative; and
when
internal
pressure, such as
it
is
found in hydraulic work,
is
often necessary to
is to be withstood, do with a smaller factor of safety to
insure soundness.

of
93
o o o
o"
Elasticity.
o o
o"
ooo'ooo'oi
Transverse
Coefficient
o o
ooo'ooo'oi
ooo'ooo'oi
ooo'ooo'oi
ooo'ooo'oi
ooo'ooo'oi
pT
o o o
o"
o^
o^
CM
ooooooooooooooo ooooooooooooooo ooooooooooooooo d d d d d d d d d d d d d d ooooooooooooooo

o"
o o o
o"
to
.
oo"
o^ o"
o^
o^
o^
POPOfOPOPOPOrO<N POM
o"
o"
o*
o^ o"
o^ o"
o^
o^
o^
o^
o^
o^
o^
o^
o^
i^
io o" m" ^t
P
!3
o o o
OO"
o o
o^
N
P)
s H J Z a =s H
CO
6.
o o
o^
o o
o^
oo"
P)
o o
o^
"3"
o o
o^
"TT
o o
o^
o o
o^
o o
o^ n"
o o o
o"
o o o
o"
o
PO
o o
o H
t/2
E o
vo
O" no"
rO PO LO NO t^
oo
po
o
'35
o o o
o"
oooooooooooo oooooooooooo
o^ o"
c S
Eh
d d d d PO^l-^-iOLONO PO 00 PONO
p*"
o^
o^
o^
o^
o^
o^
o^
o^
o^
o o
O^ o"
cs
o o
O^
o^
o^
iO
no"
o o
oo"
po rC
d\
no" no"
io to
3 o
CO
ooooooooo ooooooooo
o N o o
o"
o"
P
o" oo" <t 't
n "OÔ
o^
o_ o"
fO oo" OO 00 sf
q^ o"
o^
o^
o o o
o"
o o
o^ pT
o 2
H
Pi (-
a H <
O
d
o o
o o
o o o
NO
o"
o o
o^
o o
o^ pT
to pT ON ^t"
|3
73
ooooooooooooooooooo ooooooooooooooooooo
o o
o_
o^
o^
o^
o^
o^
o^
o^
o^
o^
c y
r-
o"
pi
iO to
i/> it,
fO o" C> fO CO* o" no" 00 00 O ii NO n
o" o" VOO
o^ pT
o^ hT
fO
o^ o"
o^
o^
io o"
roiooo n
o^ o"
in
<
<fi
c
:
i
c c -
i
c
r> ~
,2
R
c c h
c
: _:
B a
b
s-
c
p
c
1/
u
01
1
)-
1/
L u c
G
DC
_
1
7
p
s
u
i c c
c
p
c
1/ 5
NO
J;
'
J 1
c -
7
b
s
c
c
c
"a
"c pc
"a
2
'Z
C - 7 7 u u u i c -a
a a
-
7 3
_t
C/3
C
C
c7
7 7 7
9.
1 7
L.
"
~z
N c c -
g L
7 H <
94
MACHINE DESIGN
TABLE
Character of Stress or Strain.
VI
Formula.
A
B c
Stress in
Ten. or
Comp
P
Pl A- AE
Strain in Ten. or
Comp
D
E
F
Torsional Stress,
Solid
Cir-
Pa
-T-
Palp
e
10
Torsional Stress, Hollow Cir-
Pa
4 -'.*> r _A-(<.
IOfli
Torsional Strain, Solid Circular Shaft Torsional Strain, Hollow Circular Shaft
Deflection in Bending
Stress
6=
e __
2Tl
TrEsd*
32
H
I
TrEaidS
See Table
I.
d<f)
due
to
Flexure
M
e
= tL
e
See Table
I.
K
K,
Combined Bend'g and Twist'g

"
2
K
K
L
"
"
M = }4M+y V M*+7 J T =V M M = }4[x+V X -\-l]T r =rV^ +

2
2
-\2
Combined Torsion and Compression
*-iV|/Vf,+fi ]
*-7*f"+*r
Lx
Combined Torsion and Compression
M
N
Combined Flexure and Direct

Stress
>'
P=
F
A+
Pae
I
m
O
Eccentric
7r
E\
'
P J
Loading of Long
Columns
*.
rf, A
L
1
(i\', ('+)'l
p"
'
m*>E\p)
95
TABLE' VII
PROPERTIES OF SECTIONS
Shape
of Section.
Moment
of
Inertia.
I
Modulus
of
Section.
I
Square of Radius of
Gyration.
"Eolar
Moment, ofi
Inertia.
p'-A
.09SD 3
TTd*
_
.049d
7TD
D
.16
ttd 4
32
61
IS
m^\
12
D 2+d2
32
7T(D 4-d 4 )
32
D
2
BH
*-BIhlk
t..
::.
B:HCB a+H.2 )
12
12
[BH?-bl. 3 ]
g^rf-"a
j^ f
12
BH 3 bh ] [bh bh. J
3
i[BH^-bh 3]
^[BH 3-bh3]
12
rBH 3 -bh 3 1 [BH bhj
b h.
i[BH -bh ]
3
3
^[BH -bh uH
3 3 1 ~BH -bh 12|_BH-bh J

~|
|
hH-
_ (BH 2-bh~)-4BH bh (H-h)

)
fcS
= (B H -bh 1
2
B-
b-
-1 BH bh (H-h) 2 12(BH-bh)
2
)
6(BH 2 -bh 2 2 2 2 I _ (BH -bh ) 4BHbh(H-h) e2 6(BH 2 -2bhH+bh 2

ei
)
^[bH'+Bh
BH
^[bHW]
e!
I
12(bH + Bh)
tffej
* b-H
24
!
^BH
12
e2
7rBH
64.
ttBH-
tt(BH 3+HB
64.
Eh
CHAPTER
IV
GENERAL THEORY OF FRICTION, LUBRICATION, AND

EFFICIENCY
27. Friction in General.
When
two
solid surfaces are held in
contact by any appreciable force, any effort tending to

relatively to
tially to
move them
This
resist-
each other
is
met by a
resisting force acting tangen-
the surface of separation of the two bodies.

is
ance to relative motion
due
to the interlocking of the
minute de-
pressions and elevations which exist even in the smoothest surfaces
and will, of course, vary with different properties of materials and different qualities of finish. Thus, unsurfaced cast iron will show a very great resistance to relative motion, while two hardened and ground surfaces of steel will move over each other
with
much more
ease.
If the
two surfaces are very carefully

will,
fitted together
without any foreign matter in between, they
in the case of
still
many
substances, adhere firmly together, which

If oils
further increases the resistance to relative motion.
or lubricants of any kind are interposed between the surfaces,
the
resistance to relative motion
is,
to a considerable extent,
overcome.
This tendency
feature
to resist relative
motion
is
sometimes a desirable
and sometimes
not.
In the case of bearing and rubbing
surfaces generally, such frictional resistances result in loss of
power and should be reduced

frictional resistance is of
to a
minimum;
while in the case
of friction clutches, brake straps, keys, screw fastenings, etc.,
great utility
and every
effort is
made
the
to insure its presence.
The laws
of friction,
and the manner
of their application therefore, are of prime importance to
engineer.
These laws are

though
considerable
at present rather
imperfectly
understood,
It
experimental
work has been done.
has been found that
many
of the older theories

96
based on ex-
GENERAL THEORY OF FRICTION

perimental work are
true
97
of
only
for
the
range
conditions
covered by the experiments, and that conditions different from

these
show
entirely different results.
The
ratio of frictional resistance F, to the

;
called the coefficient of friction
or
if
this ratio
normal load P, is be denoted
by/, then for flat surfaces,
==
F or F
= f P.
In the case of circular surfaces, such as journals and bearings,

the distribution of the normal pressure
is
variable
and dependtogether.
ent
on the manner
it
in
which the surfaces are
fitted
In such cases
is
customary for convenience
to define the co-
efficient of friction in a similar
way
as in
flat
surfaces,
and con-
sider that
it
has special values for circular surfaces.
The
co/.
efficient of friction for circular surfaces will
be denoted by
Hence
as before,
intensity of
F =
P.
surfaces, as before
it
The
normal pressure on circular
stated, is difficult of accurate determination
and
is
therefore
customary
to take as the
normal pressure the intensity of presOr,

if
sure per unit of projected area.
diameter of shaft,
and
=
c
length of bearing, then the intensity of pressure per
unit of projected area,
w =
-.
The energy absorbed by

into heat which
is
frictional resistance is transformed
conducted away by conduction and radiation

kinds of bearings, by water
of friction
is
to the air, or, in the case of certain
circulation or other means.

fore
flat
The work
of energy
is
if
often there-
an important factor
plates the foot
in the
design of rubbing surfaces.
For
is
is
pounds
absorbed
per minute
E = fP
in
T, where
the velocity
in feet
per minute and
pounds.
For circular surfaces,
be the number of revolu-
tions per minute,
and d
^'
the diameter of shaft in inches,
E =
ldNP =
12
26lS
Np
For
other forms of surfaces sec Kent's "Engineer's Pocketbook," page 938,

40.
and Thurston's "Friction and Lost Work," page

7
98
If
MACHINE DESIGN
then /or
\i
be known, for any pair of rubbing surfaces, the
frictional resistance
and the energy absorbed
for
any load
P may
be calculated.
tally for
many
Values of/ and n have been obtained experimenof the materials and conditions met with in engistill
neering, but the data so far available are
incomplete.
The
consideration of the laws of friction, as applied to
ma-
chinery, naturally divides itself into two parts.

(a) (b)
Friction of
Dry or Unlubricated
Unlubricated
Surfaces.
Friction of Lubricated Surfaces.

of
28. Friction
Surfaces.
The experiments
of Morin, Rennie,
Coulomb* and many
others, furnish the fol-
lowing laws for dry or very slightly lubricated surfaces.

(1)
The The
frictional resistance is
approximately proportional to
the normal load.

(2)
frictional resistance is approximately independent of
the extent of the surfaces.

(3)
The
frictional resistance, except at very
low speeds, de-
creases as the velocity increases.
was formerly supposed that an abrupt change took place the value of / when the body passed from a state of motion
It
in
to
one of
of rest
rest.
is
It
seems now, however, that while the
coefficient
in general greater than that of motion, the

is
change in
value
is
gradual and the value at rest
not far different from
that at very slow motion.

of
As the
velocity increases, the value
materially decreases and this must be taken account of

friction is
in
designing machinery where
involved.
Unfortuis
nately the information regarding high or even moderate speeds

also very incomplete.
The
following values of/ must, in view of the incomplete infor-
mation, and also because of variations which come with slight
changes of conditions, be looked on as approximate values only.

Unless
it is
positively
known
that the surfaces will be kept free

oily substances, these values
from even
slight
contamination by
must be used with judgment.

* See "Lubrication and Lubricants": Archbutt
of these points.
&
Deely, for a
full
discussion

Coefficients of
99
F vict ion
(J) for Dry or Slightly Lubricated Surfaces.

3 to .5
2 to .6
Wood Wood
on
Wood Static
" " "
on Metals
or very low velocity " " " "
Leather on Metals
Leather on
"
"
" "
"
"
"
3 to .6 3 to .5
Wood
"
"
"
"
feet per
Metal on Metal
Cast Iron on Steel " " "
"
" " "
"
"
(average)
.3
velocity
"
"
440
2640 5280
minute.
"
"
.
.32
.2
= =
"
" "
"
.06
in the
There are no experimental data giving the decrease

value of/ at high speeds, for combinations such as
wood
or leather
on metals.
The data
a rough guide to
on steel will, however, serve as what may be expected to occur. It is to be parfor cast iron
ticularly noted that, in designing
brake shoes or other friction

allowance must
machinery where great

be
velocities are involved,
made
29.
for the decrease in the value of the coefficient.
Dry Rolling
Friction.
it
When
a curved
body
rolls
upon
than
the
a plane or curved surface,

has been found that the so-called
due
to the rolling action is
much
less
that
due
to sliding, for the
same
load.
If P =
the load;
F =
horizontal force required at the axis of a circular
body
to pro-
duce and sustain uniform motion; and

it
=
is
radius of rolling body,

a coefficient to be deis
has been found that
P F = -
k
r
where k
termined experimentally.
to
If r
be expressed in inches k
found
have a value of about
.02 for iron or steel rolling
on iron or
steel.
Neither the coefficient k nor the exact theory of rolling friction

is
at
present very accurately known.

is,
The most important

is
use
of rolling friction
in
as far as the present discussion
concerned
connection with roller bearings for shafting, and a fuller dislater.
cussion of these will be given
30. Friction of Lubricated Surfaces.
When
a lubricant
is
interposed between
resistance
is
pair
of
rubbing surfaces, the frictional
materially reduced because the surfaces are wholly or

MACHINE DESIGN
separated from each other by
the
lubricant.
of ways.
IOO
partially
The
If the
lubricant
may be
fed to the surfaces in a
number
is
motion
oil
is
intermittent,
and other conditions

oil
will allow,
a simple
If the
hole leading to the rubbing surfaces

is
often used.
motion
tinuous
continuous, some form of
cup which
(a)
will give a con-
supply
is
better.
Fig.
14
shows a
cup
of
oil
the
simpler type where a wick of cotton or wool draws up the

capillary attraction
is
and feeds
siphon
it
slowly into the

Fig.
oil hole.
by This
so-
sometimes
called
feed.
14 (b)
shows a
called sight feed
cup where the

it
oil falling
by gravity from the

also used to
cup can be seen as

lated
passes the hole e and the flow can be regu-
by the screw
oil
d.
Centrifugal action
is
some
is
extent to feed
to rotating parts.
Sometimes an opening
pad saturated with lubricant can be kept pressed up against the moving surface, thus lubricating the whole length of the journal continuously. For heavy lubricants, such as greases, where very heavy pressures are carried on the rubbing surfaces, so-called compression cups are often used and are constructed so as to force the lubricant in between the surfaces. Fig. 14 (c) shows a "ring oiled" bearing. The ring r running loose on the shaft s dips into the pocket below the shaft. The friction of the ring on the shaft causes it to rotate and draw up oil from the pocket. Sometimes chains are used instead of solid rings. For the most efficient lubrication the journal itself
in the bearing so that a
made
runs in a bath of
pressure.
oil (Fig. 15)
or
is
flooded with
oil
supplied under
of sup-
The
relative merits of these various
methods
plying the lubricant will be more apparent after a discussion of

the general laws of lubrication.
The
effect of friction,
and the
efficiency of lubrication of so-
called lubricated
surfaces,
may
conveniently be treated under
three heads:
(a)
Static Friction of
Lubricated Surfaces.
(b) Friction of
(c)
Imperfectly Lubricated Surfaces.
Friction of Perfectly Lubricated Surfaces.
31. Static
Friction
and Lubrication.
When
a pair of lu-
bricated surfaces are pressed together by a load, the pressure
tends to slowly expel the lubricant from between the surfaces.

Experiments and experience show that
lubricant.
If
it
IOI
is
very difficult even
with limited areas and heavy pressures completely to expel the
ordinary machinery, however,
is
allowed to stand
at rest for a short period of time, this action is sufficient to expel
so
much
of the lubricant that
may have been between
the sur-
faces
while running as to allow the metallic surfaces to

less in contact.
is
come
more or
cated surfaces
The static coefficient of friction very hence much higher than that of
for
it
of lubri-
surfaces
which move even very slowly;

at
will
be seen presently that even

their
low
velocities the surfaces tend to

It is
draw in the lubricant by
motion.
a well-known fact that heavy machinery always
offers a great resistance to starting after lying idle a short time
and often the rubbing surfaces,
if
not oiled before starting, will
Longitudinal Section
Fig. 15.
abrade each other before the lubricating action due

begins to take
effect.
to
running
The
and
oil
materials therefore for the rubbing
surfaces of heavy machinery should be carefully chosen for their

antifriction qualities,
grooves should be carefully provided
so that lubricant
can be applied as near the point of greatest
pressure as possible before motion begins.
The
coefficient of static friction for lubricated surfaces is not
very accurately
known and
it
varies
somewhat with the
pres-
sure and character of the lubricant.
fair
average value for

lbs.
metal surfaces and pressures ranging from 75 to 500

in. is .15.*
per sq.
See
Thurston's " Friction and Lost Work," page-, 316-317.
102
32. Imperfect
MACHINE DESIGN
Lubrication.
When
one
lubricated
surface
slides over another, the moving surface, even at low velocities,
tends to carry the lubricant,

surfaces.
if
properly applied, in between the
Thus
the layer of oil

it
which touches the surface
of a
journal adheres to
and
is
carried along under the bearing.
This layer
joins
it,
in turn tends to carry along the layer
which next ad-
because the viscosity of the lubricant opposes the shearof the journal.
ing action which results between layers on account of the action

of the
moving surface
is
it
In plane sliding surfaces
the lubricant
generally applied to the stationary surface and

in spite of the tendency of the slider to rub
tends to cling to
it off.
The
is
action of the sliding surfaces in drawing in the lu-
bricant
less
similar to that of the rotating journal, but in a
much
marked degree
as
ture of the case.
If
would naturally be expected from the nathe velocity of rubbing be very low, or the
is
pressure very high, or the supply of lubricant limited, the quantity

of lubricant that is carried in
very small and the surfaces in
contact are very slightly lubricated and

metallic contact.
may
even be in actual
The
materials, therefore, for the rubbing sur-
faces of slow-moving machinery should also be carefully chosen

for their antifriction qualities, as
even after the machinery has

metallic contact
been
successfully set in
motion
may
occur
between them.
If the velocity of
rubbing and the supply of lubricant be
in-
creased, the load remaining the same,

is
more and more lubricant

till,
thrust between the surfaces
by the action noted above
at
a point depending on the pressure, velocity of rubbing, and
vis-
cosity of the lubricant, the metallic surfaces are completely sep-
arated and the friction becomes only that due to the fluid
tion of the lubricant
itself.
fric-
This
last state is
known
as perfect
lubrication.
The
is
formation of this separating film with in-
creasing speed
probably gradual
and the character of the
contact most probably passes through a gradual change, from
contact which
tially
is
nearly metallic through successive stages of par-
fluid
contact to complete fluid separation.
The
exact
point at which perfect lubrication occurs for any given load,

velocity,
and lubricant
is
not accurately known, but what data are
103
available will be given in connection with the discussion of perfect lubrication fect lubrication
which follows.
It is
known, however, that perin
cannot be obtained without a plentiful supply of
the lubricant, as in the case
where a journal runs

It is
an
oil
bath,
or
is
is
supplied by so-called forced lubrication where the lubricant
delivered under pressure.
impossible or inconvenient,
however,
to lubricate the greater part of the
rubbing surfaces of
machines
in this
manner and,
therefore, all surfaces lubricated

holes,
is
by such means as simple oil where the supply of lubricant
oil
cups, oily pads,

restricted
etc.,
in
any way
must be
between
considered as imperfectly lubricated.
As already noted, the exact condition which

the supply
will exist
such surfaces depends on the pressure, the velocity of rubbing,
and character of the lubricant, and the temperature of

oil.
the bearing as affecting the viscosity of the

so
Naturally where
many
variables exist, experimental results are very discord-
ant,
and while an immense amount of work has been done, the

emphasize the great variation
It is
results only serve to
in conditions
with change of these variables.

if
evident, for instance, that
and pressure remain constant, almost any condition may be produced from metallic contact to perfect lubrication simply by varying the supply of lubricant. The law of variavelocity
tion of the coefficient of friction, with either varying pressure
is
or velocity,
oil is
also found to be modified
by the rate
at
which
at
supplied.
The
generally accepted theories for imperfectly
lubricated bearings running
under average conditions,

oil
i.e.,
normal temperature, and with good

are as follows *
:
supply from cups or pads,
(a)
Starting from rest with constant load, the coefficient of
friction first increases slightly
with increasing velocity and then

feet
decreases, until at a velocity

ute
somewhere below 200

oil
per minvalue
is
(and depending upon the

(see Fig.
16). f
supply) a
minimum
reached
With
further increase of velocity the

and Thurston's "Friction and Lost
the-
See Archbutt and Deeley, page Work," pages 296-312.

fit
is
58,
to
be noted that this discussion and
coefficients
given refer to
circular bearings
and
friction of rotation.
io4
coefficient increases
MACHINE DESIGN
till
the temperature affects the viscosity of
the lubricant to such an extent that abrasion and failure occur.

(b)
With constant
velocity
is
and very
light loads (see Fig. 17)
the coefficient of friction

the coefficient
As the load is increased, decreases very rapidly at first, and then more
very high.
coefficient again slowly increases.
slowly
till
pressures of about 100 to 200 lbs. per square inch are
obtained
(c)
when the The law of
variation of friction with temperature
is
very
complex and not well defined.

ever,
Its general characteristics,
how-
may
and
be expressed as follows: every combination of presvelocity requires a lubricant of a certain viscosity for
sure
best results.
At high speeds and
light loads, a light, thin oil
600
500
500
3400
a
/
/
c m
u
i
-300
ft
&300
m
-C
8 200
g200
100
\\^
,.005
Ah .100
/
.005
.01
.01
.015
.02
.025
.03
.015
.02
.025
Coefficient of Friction
Coefficient of Friction
Fig.
i 6.
Fig. 17.
will
tion,
be readily drawn in between the bearings, and
its fluid fric-
which constitutes the greater part

be
less
of the resistance in such

oil.
cases, will
than that of a heavier

its
Increasing the
temperature of a lubricant decreases
viscosity and, in the

friction.
above case therefore, would cause a decrease in
In
the case of the heavier loads and lower velocities, usually
met with
in machines, an increase of temperature decreases the viscosity
and may, owing

in friction.
to the expulsion of the lubricant, give
an increase
Care should therefore be used

case in
to
obtain an
oil
suited to the
hand,
for
sometimes a change
of lubricant is suffi-

cicnt to cause great trouble or,
105
to
on the other hand,

is
reduce the
failure
temperature
imperfectly
of
bearing
that
heating.
The
results
of
lubricated
bearings
generally
from
the
lowering of the viscosity by increased temperature, so that the

oil
film
is
no longer maintained and metallic contact and abrait is
sion ensue.
From
Figs.
the foregoing
evident that the coefficient of friction
for imperfect lubrication will necessarily
be a variable quantity.
//
16 and 17
show
the variation of
for varying velocities
and pressures.
With good lubrication and moderate
velocity
it
may
be as low as .005, and again with low velocity and poor

it
lubrication
may
rise to
.05
or more.
When
which
lbs.,
the velocity
is
exceedingly low, the coefficient approaches that of static friction

of lubricated surfaces, the average value of
is .15.
fair
average range for pressures from 50 to 500
and
velocities
from 50
is
to
500
ft.
per minute,
is
from
.02 to .008 and, for pur-
poses of design of ordinary machinery,

to
may be
is
taken at .015.
It
be noted that with imperfectly lubricated surfaces and low

the coefficient
of
friction
less
velocities
dependent on the
character of the lubricant, and more dependent on the character

of the
rubbing surfaces.
posite curves
sults.
The curves taken from a number
Figs.
16 and 17 are com-
of
actual experimental re-
They
/'.,
are not to be taken as giving exact values of the
coefficient
but serve to show graphically the general laws by

In interpreting such curves as Fig.
is
which
it
varies.
17
it
must
be kept in mind that, while the coefficient

creasing, the actual frictional resistance
like
decreasing or in-
may
is
not be changing in
the product of the
manner.
The
frictional
resistance
If,
load and the coefficient of friction.
for instance, the coeffi-
cient decreases as fast as the load increases, the frictional resist-
ance
will
remain constant.
The
curves show, however, where
best results
may
be expected when designing new machinery,

light
and throw some

of the relation
on proposed changes
in
running speed of and the
machinery already
installed.
exists
They
it
also indicate the complexity

velocity, pressure,
which
between
is
coefficient of friction.
When
considered that the temperait
ture also greatly affects these relations,
is
evident that a state-
106
MACHINE DESIGN
of these relations for imperfect lubrication, in the
is
ment
all
form of
a general law or mathematical expression,
impracticable,
and
such expressions are misleading.

33. Perfect
shown in the last article that any rotating journal will, by means of the molecular attraction between it and the lubricant, combined with the viscosity of the lubricant, draw more or less of the lubricant in between the journal and bearing, the amount so drawn in depending on the velocity and pressure. If the journal be allowed
Lubrication.
It
has been
to
oil,
run in an
oil
bath, or
is
otherwise plentifully supplied with
and the
velocity be high
enough
for the pressure carried,
it
is
found that
this action is so
marked
that the rubbing surfaces
are completely separated by a thin film of lubricant

friction
itself.
and the
becomes only that due
to the fluid friction of the lubricant
Mr. Beaucamp Tower experimenting with journal

that with a journal
friction (see
Proceedings of Institution of Mechanical Engineers, 1883) found
and bearing
to
above action was so marked as

sure such that the load
form a film
arranged as in Fig. 15, the of oil under presfluid borne.
was completely
The
distri-
bution of the pressure in this film was found to be as indicated
by the diagrams above the cross-sections, rising to a maximum at the middle and falling to zero at the edges of the bearing. Mr. Tower succeeded in this way in carrying a load of 625 pounds per square inch of projected area at a velocity of 471 ft. With a load of about 330 lbs. per sq. inch, ^,nd a per minute.
velocity of about 150
ft.
per minute, a
maximum
oil
pressure of
was found near the middle point of the bearing. It has been proved mathematically, and verified experimentally, that the conditions which exist in a bearing running under these
625
lbs.
conditions are as follows: the journal,
being
slightly
smaller
than the bore of the bearing, tends to be crowded back from the
side
where the lubricant

in
is
is
carried in, as
shown
in
an exag-
gerated manner
pressure
half
the figure, giving a wedging effect.

little
The
consequently greatest at a point a
more than
way beyond
the centre of loading where the distance be-
tween surfaces
is least.

The
is
it
IO'
exact relation which must exist between velocity and

is not known nor Enough is known,
pressure, to allow this pressure film to form,

likely that exact limits
can ever be
set.
however,
to serve as a general
guide for average conditions.

for circular journals the
Professor H. F. Moore found that
minimum
will just
limiting values of pressure
and
velocity
where the film
form may be approximately expressed by the expression

\
7.47
y,* where
w is in
pounds per square

to
inch,
and v in
feet
per minute.
Fig. 18,
The
values given by this expression arc plotted in

1,
Curve No.
and seem
check
fairly well with con-
siderable other data.
Curve
(2),
Fig. 18,
olive oil,
values obtained by
Tower with
shows the simultaneous where frictional resistat least well
ance was a minimum, indicating that the film was

300
No. 2
No.
r -.v.'
vzl
^100
my/*~^
JJOji.
^~
!
SO
100
150
200
250
300
350
,400
450
500
Feet perTVTinute
Fig. 18.
formed.
Curve
is
(3)
shows similar values
for mineral
oil.
The
values obtained by
Moore
are on the safe side judged by
Tower's
work, which
accepted as accurate, and probably do not indicate
the very lowest point at which a film will form.

that a film
Tower found
Curve
ture
(2).
would form considerably below the values given in In Moore's experiments, as in Tower's, the temperaat oo.
was constant
Moore's experiments were on mineral
Tower's experiments are very concordant oils. and conclusive, and show that the laws of friction for perfectly
results of
The
lubricated surfaces, for ordinary speeds

definite,
and pressures, are quite

pressure, very nearly.
16,
the coefficient of friction varying as the square root

velocity
of
the
and
inversely
as
the
American Machinist, Sept.
1903.
io8
MACHINE DESIGN
for olive oil the relation is expressed
Thus
P
very closely
by
y/ v
.2
It follows
from
/*
this,
that for any fixed velocity
and
is,
temperature the product of

the frictional resistance
load, for any velocity.
is
and
w will
be a constant.
That
practically constant with

to
change of
This was actually found
be the case in
the experiments, a variation of pressure per square inch from ioo

to
500 not appreciably affecting the resistance. Table VIII will serve to show the remarkable regularity of the results, and the low
values of the coefficient as compared with imperfectly lubricated
surfaces.
Much
lower values have since been attained in
oil-
testing machines,
under more ideal conditions, but such low
values must not be considered as attainable under ordinary practical
working conditions, while there
is
no good reason why such
coefficients as given
below cannot be obtained in well-constructed
machinery.
TABLE
VIII
OIL
BATH OF RAPESEED
Coefficients of Friction for Speeds as Below.
Load
in
lbs. per sq.
Inch.
105 ft. per Min.
157 ft. per Min.
262 ft. 209 ft. 314 ft. per Min. per Min. per Min.
366
ft.
per Min.
47i ft. 419 ft. per Min. per Min.
573 520 4i5 363 258 i53 IOO
.00102
.00095 .00093 .00084 .00139 .00200 0O357
.00107 .00162 .00277
.00108 .00105 .00107 .00960 .00162 .00239 .00423
.00118 .00115 .00119 .OOIIO .00178 .00267 .00503
.00126 .00125 .OOI30 .00122 OOI95 .00300 00576
.00132 OOI33 .00140 .00134 .00213 00334 .00619
00139 .00142 .00149 .00147 .00227 00367 .00663
.00148 .00158 00155 .00243 .00396 .00714
Tower's experiments have been amply
verified
and may be
accepted as reliable for the range which they cover.
The
ex-
periments of Stribeck and Lasche (see Chap. X) have extended

the range of knowledge on this point to velocities over 2,000
ft.
per minute.
Their experiments show that for velocities between

ft.
500 and 2,000
per minute the coefficient of friction, for a given
load and temperature, varies as the 5th root of the velocity; and
beyond 2,000
ft. is
independent of the velocity.
This point
is

discussed
still
IO9
further in Chap.
its
in connection with the design

is
of bearings, where
34.
principal application
found.
Summary.
statements
(a)
From may be made:
the foregoing discussion the following
The
friction of imperfectly lubricated surfaces
depends
partly
on the character of the surfaces themselves, and in a greater
degree on the character and amount of the lubricant supplied.

(b)
The load
that can be successfully carried
on an imperfectly
of lubricant
is
lubricated surface will vary greatly with the
amount
supplied, and must be kept very low where this supply

stricted.
c
re-
The
friction of perfectly lubricated surfaces
depends very
little
on the character
of
the
rubbing surfaces, but depends
mainly on the character of the lubricant.

(d)
is,
The
frictional resistance of perfectly lubricated surfaces
within the ordinary limits, independent of the intensity of

velocity'.
pressure and dependent only on the

(e)
The
coefficient of friction of perfectly lubricated surfaces,
for
any given pressure and temperature, varies very nearly as the

ft.
square root of the velocity for velocities up to 500
per minute;
approximately as the
fifth
root of the velocity for velocities beis
tween 500 and 2,000
ft.
per minute; and
practically independent
of the velocity for values

35- Efficiency.
It
above 2,000
ft.
per minute.
all
has been pointed out that

is
the energy
supplied to a machine
that
not transformed into useful work, but

lost in
overcoming frictional resistances There are many ways in which energy losses may occur in machines, and a careful distinction must be made between certain of these ways in order to get a clear
of
it is
some
always
and doing useless work.
definition
of the
term
efficiency.
Thus
the steam engine re-
ceives
its
supply of heat in the form of steam under pressure.

is
considerable portion of the heat so received
lost
by con-
densation of steam on the cooler cylinder walls, and some escapes
by radiation without doing any work whatever on the piston.
Of
the energy actually applied to the piston, part
is
transformed
lost in over-
into useful
work
at the driving belt,
and part
is
IIO
MACHINE DESIGN
frictional resistances just discussed at the various
coming the
constraining surfaces.
The
gas engine
is
subject to similar losses; a large part of the
heat of combustion escaping to the jacket water or to the at-
mosphere by radiation, and doing no work on the piston; while

only a part of the energy actually applied to the piston reappears
as useful work.
Hydraulic and
electric
machinery have similar

might be
as losses
elements of
loss.
The
first
class of these energy losses
called leakage losses, as they are of the
same character
is
by actual leakage
energy.
losses
of the
medium which
machine
itself
used to transmit the
The
all
losses in the
are
known
as frictional
and are common

of
it
to all
machines;
and no machine can
transform
the energy supplied into useful work, but must lose
some
in friction or other wasteful resistances.
Efficiency has been defined (Art. 2) as the ratio of useful
work
ence
and from the above it appears that a machine may have two efficiencies depending on whether referto energy supplied;
is
had
to total
energy supplied, or to that portion only of
the total energy which the machine transforms into useful and
useless work.
These
is
efficiencies are respectively
known
as the
Absolute
if
Efficiency
and
the
Mechanical
Efficiency.
Thus,
a gas engine
supplied with 1,000 thermal units, and trans-
forms 200 units into useful work, and 50 units into the useless
work
of friction,
its
absolute efficiency
is
200
1,000
.20,
and the
mechanical efficiency
lute efficiency is
is
200
.80.
The
consideration of abso-
250
of
beyond the scope of this work; for the design but the it does not need to be considered mechanical efficiency can seldom be neglected, since, in general, the amount of work to be done is fixed, and the source of energy must supply enough more energy than this to compensate for the
many machines
frictional losses of the
machine.
The mechanical
efficiency of
any
train of
all
mechanism
is
the
continued product of the efficiencies* of

* It
the several pairs of

used in a number of ways
may be
noted in passing that the term efficiency
is
other than as the ratio of
work done
to energy
expended.
Thus
the strength of a

constraining
occur.
III
losses
let
surfaces
in
the
train
at
which
frictional
Let any machine have n pairs of such surfaces, and
their respective efficiencies
be
e,
e v e 2, e 3 , e 4,
Let
let
E
be
be the mechanical efficiency of the whole machine, and

the total
amount
of energy available for transformation into
either useful or useless work.
Then, the amount
of energy
which
is
the
first
pair of constraining surfaces delivers to the second
X e, and the amount which the second delivers to the third is Ke X e and so on, till the amount of. energy delivered by the last element (or the work done) is K (e X e X e ej.
ly
x
But the mechanical
efficiency of the train is
work done
energy
supplied
K
X
et
(e
el
e2
e n)
K
X
e2
{e
ej.
machine may consist
of several trains of
mechanism.
If
these several trains are arranged in series so that the energy
passes from one to another consecutively, the efficiency of the
whole machine, by reasoning similar

is
to that in the last
paragraph,
the continued product of the efficiencies of the several trains
of
mechanism.
If,
however, the trains are arranged in parallel

is
so that the total energy several trains of
transmitted simultaneously through

train transmitting only a por-
mechanism, each
tion of the energy, the
above reasoning for the efficiency of the
whole machine does not hold.

to
each train
is
known, the
amount of energy supplied amount of work which it will deliver

If the
can be computed as above.
The sum
of all the work, delivered
by
all the trains,
divided by the total energy supplied, will be the
efficiency of the
If,
whole machine.
efficiency of the
therefore, the efficiencies of the several constraining sur-
faces of a
machine are known, the mechanical

compared
to the strength of the original
riveted joint,
unpunched
plate,
is
called
the efficiency of the joint,

in
when what
really
is
meant
is
is its
relative strength.
Again,
an
air
compn
->or. the ratio of the air actually
discharged per stroke, to the wholecalled the volumetric efficiency.
amount
It is
raised to the required pressure per stroke,
evident that such efficiencies are of a different character from those discussed
above and do not enter into the calculations of the efficiency of the machine, as a
whole, in the
manner indicated above.
112
MACHINE DESIGN
The mechanical
efficiency of
whole machine can be calculated.
any machine element
is,
however, a variable quantity;
for the
will
coefficient of friction of
any pair of constraining surfaces

its
vary with the lubricant and
method
of application, the tem-
perature, the alignment of the surfaces, the velocity of rubbing,
tions are constant, the
and the bearing pressure. Furthermore, when all other condisame pair of constraining surfaces will
have an entirely different efficiency for the same amount of
is
power transmitted, depending on the manner in which the load Thus, consider a simple wheel and axle driven by a applied. belt on the periphery of the wheel. With a given diameter of transmission wheel, the of a given amount of power will bring a certain definite frictional load on the bearings. If, however,
is
the diameter of the wheel

in a like ratio,
doubled, the belt speed
is
increased
and the
belt tension will, for the
same power
transmitted, be one-half of the former value;
and, as a conse-
quence, the frictional resistance at the bearings will be reduced

to one-half the original value,
stant.
the revolutions remaining con-
In general, therefore,
it
is
impossible to calculate precisely
from the analysis

will
of a design
if
be,
particularly
the
what the mechanical efficiency mechanism is at all complicated,

is
though a reasonable approximation

a similar type have been built,
the design of
in existence.
it
possible.
is
far
new ones on
For
all
efficiency tests
If machines of more accurate to base made on those already

tests
standard machines such

results
have been
made, and the recorded

sign of
chine
to
new machines of a new type is
of like
to
form a valuable basis for the decharacteristics. But when a matests are
ef-
be designed, and no recorded
be had that
will give
any information as
to the
probable
ficiency,
an estimate must often be made and the efficiency
cal-
culated as outlined above.
In general, a close approximation

is
can be made, and the making of such estimates

only with experience.
ficiencies of various
a great aid to
the development of that judgment in such matters, which
comes
ef-
In such cases a knowledge of the
machine elements becomes necessary.
If
the coefficient of friction for
any constraining surface could be

accurately determined,
efficiency with
it
113
calculate
its
would be possible
of certainty.
to
some degree
But, as before noted,

in
the quantity varies with the velocity of rubbing, with changes
bearing pressures,
necessarily
etc., and such methods of computation are cumbersome and to be attempted only where a very
is
close estimate
required.
The
dent
is
following are rough average values of the efficiencies of
the most
common
elements.
For more accurate values the
stu-
referred to the respective discussions of these various
elements which follow
Common Common
Bearing, singly Bearing, long lines of shafting
96-98
95
Roller Bearing
Ball Bearings
98
99
93
Spur Gear Cast Teeth, including bearings Spur Gear Cut Teeth, including bearings
Bevel Gear Cast Teeth, including bearings
96
92
95
Bevel Gear Cut Teeth, including bearings
Worm
Gear, varies with thread angle, see Art. 54
Belting
96-98
_ 95 97
Pin-connected Chains, as used on bicycles
High Grade Transmission Chains

8
97~99
CHAPTER V
SPRINGS
36. Distinguishing
Characteristic
of
Springs.
Springs
are
characterized by a considerable distortion under a moderate load.
Every machine member

is
is,
in a sense, a spring, for
no material and
distri-
absolutely rigid and the application of a load always produces
stress
and accompanying
it is
strain.
By proper
selection
bution of material
possible to control (within wide limits)
the degree of distortion under a given load.
An
absolutely rigid material would be practically unfit for the
construction of any
member
subject to other than a perfectly
shown in Art. 24) the stress due to a suddenly applied load would be infinite if the corresponding distorquiescent load; for (as
tion of the
member were
it
zero.
While
of
is
usually
desirable
to
to
restrict
the
distortions
most machine parts
very small magnitudes,
there are
many
load
the
cases in which considerable distortion under

is
moderate
desirable or essential.
is
To
meet
this last
requirement
member
37.
often given
some one
of the
forms commonly
Springs are in
called springs.
The
Principal Applications of Springs.
common
I.
use:
forces;
For weighing
as in spring balances,
dynamomemechanism
ters, etc.
II.
For controlling the motions
of
members
of a
;
which would otherwise be incompletely constrained for example, This in maintaining contact between a cam and its follower.
constitutes
III.
what Reuleaux has called "force closure." For absorbing energy due to the sudden application
;
of a
force (shock)
as in the springs of railway cars, etc.
IV. As a means of storing energy, or as a secondary source of

energy; as in clocks,
etc.
114
SPRINGS
"5
in
An important
to
class of
a
mechanisms
which springs arc used
weigh forces
is
common
type of governor for regulating the
speed of engines or other motors.
In those governors which use
springs to oppose the centrifugal, or other inertia actions, the springs automatically weigh forces which are functions of speed,
1
or of change of speed.
The
links, or other connections,
which
forces,
move
relative to the shaft with
any variation of the above
correspond to the indicating mechanism of ordinary weighing

devices.
The
forces
first
is
of the above-mentioned applications
the weighing of
the load
action.
usually the most exacting as to the relation between and the distortion of the spring throughout the range of In the second and third classes of application, it is fre-
quently only required that the

shall lie within certain limits,
cisely defined.
maximum
load and distortion
which often need not be very preis
The
use of springs for storing energy (as the
term spring
light
is
ordinarily understood)
pieces of
almost wholly confined to

little
mechanisms or
to operate
apparatus requiring but
power
them.
of
38. Materials
Springs.
Springs
are
usually
of
metal;
although other solid substances, as wood, are sometimes used.
A high grade of steel, designated as spring steel, is the most common material for heavy springs, but brass (or some other
alloy) is often used for lighter ones.
A
in
confined quantity of
air,
or other compressible fluid,
is
used
many important applications to perform the office of a spring. The air-chamber of a pump with its inclosed air is a familiar example of what may be called a fluid spring used to reduce shock ("water hammer"). The characteristic distortion of the solid
springs
is
a change in
form rather than
of volume;
while the
inci-
fluid springs are characterized
by a change of volume with
dental change of form.
Soft-rubber cushions, or buffers, are not infrequently employed

as springs,
action to
irv, in
and these are in some respects intermediate in their It is usually not necthe two classes mentioned above.
these simple buffers, or cushions, to secure a very exact
relation between the loads
and the distortions under such loads.
n6
The
will
MACHINE DESIGN
discussion of the confined gases (fluid springs)
is
not within
the scope of the present work;
hence the following treatment

Springs
be limited to solid springs.
39.
Forms
of Solid Springs.
may
be subjected to
actions which extend, shorten, twist, or
bend them, producing
'<^^^^>:
*
j^ A
>
<&*e
stresses in the material, the character of
the
form
I.
of the spring
which depends both upon and upon the manner of applying the load.
Flat Springs are essentially beams, either cantilevers, or
with more than one support.

flexure
These springs are subjected
to
and the resultant stresses are tension in certain portions of the material, and compression in others, with a transverse shear, as in all beams; the shear may
the load
is
when
applied,
SPRINGS
usually
117
be
neglected
in
computations.
The
ordinary
flat
beam
formula? for strength and rigidity
may be
applied to
springs,
with constants appropriate to the particular material and form of
beam
used.
Flat springs
may be
simple prismatic
it
strips, of
uniform cross-
section (Fig. 19 or 22), although
is
preferable that the form
of such springs approximate those' of the
"uniform strength"
beams
(Figs. 20 or 21;
23 or 24).
It is often desirable or practically
necessary to build up these
springs of several layers, leaves, or plates, producing a laminated

spring.
It will
appear from the discussion of these laminated
Fig. 30.
Fig. 30 (a).
springs that they
may
be properly treated as a modification of one
form
of
"uniform strength" beam.

as a spring
The
neutral surface of the
beam used
is
may
be
initially
curved, either to clear
other bodies, or to give the spring an advantageous form
when
it
under normal load.
See Fig. 27.

as in the " ellipti-
Two
or
more springs may be compounded,

In such cases, each spring
total deflection is the
set.
cal" springs or in the platform springs frequently used under

carriages.
rately,
and the
may be computed sepasum of the deflections of the

commonly used
to resist
separate springs of the

II.
Helical, or Coil Springs are most
actions which extend, shorten, or twist the spring relatively to

its
longitudinal axis.
These are sometimes improperly
called
spiral springs.
Il8
MACHINE DESIGN
The stress in the wire (or rod) of which a helical spring is made is somewhat complex, consisting of torsion combined with
tension or compression, or both.
is
In a "pull spring," one which

a torsion, and there
extended longitudinally under the load, the predominating

is
stress (with ordinary proportions)
is
secondary tensile stress in the wire.
In a "push spring," one

is tor-
which
sion,
is
shortened by the load, the predominating stress

stress.
with a secondary compressive

is
When
that
the helical
spring
subjected to an action which twists the spring (as a

is
whole) the principal stress in the wire

(tension
and compression
in opposite fibres)
due to flexure and the secondary
stress is torsion.
Helical springs are sometimes arranged in "nests," springs of
smaller diameter being placed within those of larger diameter,

(Fig. 30).
In these cases, the different springs of a
set are
com-
puted separately.
car trucks.
This
last
arrangement
is
common
practice in
form of These are best adapted for a twist relative to the axis of the spiral, and are usually employed when a very large angle of torsion between the two connections is necessary. In this form of spring, the stress in the material is that due to flexure or tensile and compressive stress on opposite
III. Spiral Springs, properly so called are those of the
the familiar clock spring.
sides of the neutral axis.
IV. Helico-Spiral Springs.
by the common upholstery spring which has been elongated, and given a permanent
the direction of
its axis;
The form of spring represented spring may be looked upon as a spiral

set, in
it
or
may be
considered as a modified
helical spring in
which the
radii of the successive coils are not
equal.
It is
thus intermediate between the two preceding classes.

is
This
last
form
not usual in machine construction;
though
it
has the advantage over the

siderable
lateral
common
and
it
helical
spring of con-
resistance,
it
may be
undesirable
employed
otherwise
is
to
to
advantage where
is
difficult
or
constrain the spring against buckling.
This spring
used only
springs
as a push spring, to resist a compressive action.
The
used on the ordinary disc valves of
pumps
are often of this
SPRINGS
II 9
form, as they will close up flat between the valve and guard. Car springs are sometimes made of a flat strip or ribbon of steel
wound
in this general form, with the edges of the strip parallel
to the axis of the spring.
V. Occasionally straight rods, usually of circular or rectangular

cross-sections,
are
employed
to
resist
torsion relative to their
longitudinal axis.
the
stress
is,
of course,
These are comparatively stiff springs, and torsional. Every line of shafting is
gives the ordinary forms of solid
necessarily a spring, in this sense.
The
springs;
following
summary
the kinds of loading to which they are subjected;
and
the predominating stresses resulting
from the
different loads.
GENERAL SUMMARY OF SPRINGS

Form
of Spring.
Load Action.
Predominating
Stress.
Flat Spring. Helical Spring.

II
CI <<
11
Spiral
"
Flexure or Bending. Extension, Pull. Compression, Push. Torsion, Twist. Torsion, Twist.
Tension Torsion Torsion Tension Tension
and Compression.
(plus).
(minus).
and Compression. and Compression.
41.
Computations of Simple Flat Springs.

flat
The
following
notation will be used in treating of

cross-sections.
springs with rectangular
P = load
/
applied to the spring.
= free length of the spring. p = intensity of stress in outer fibres. / = moment of inertia of most strained section. h = dimension of this section in plane of flexure. b = dimension of this section perpendicular to plane E = modulus of elasticity of material. o = deflection of the spring.
The
These
six
of flexure.
to 24, are the will
forms of rectangular section beams, shown by Figs, 19 most important of those usedassimple flat springs.
be designated Type
I,
II,
etc.,
as in the following
120
table,
MACHINE DESIGN
which gives the constants
to
be substituted in the general
formulae for computations relating to each type.
TABLE IX
Coefficients.
Type.
A
I
/3
B
1
K
1
c
3 2
As per
"
" " "
"
Fig. 19
2 3 2 2 3
1
4"8
1 "32" 1
4
~8
II
III
" "
" " "
20
21
3
1
T
1
"
" " "
24
1
3"
2"
1
6
IV
22
23
IT
1
4 6
8
3
I
V
VI
2
1
"6
6
6
"
24
theory of strength against flexure (equation J and tables and 2) gives: For rectangular section beams supported at the
I, II,
The
ends and loaded at the middle (Types

2
III).
2
. .
-PI = \pbh .\Pl = -pbh

4
6^
(1)
For the rectangular section
cantilevers, with load at free end,

1
PI
Or
beams may be
pbh'
(2)
the general formula for the strength of rectangular section
written
PI = A pbh2
In which the coefficient
(3)
has the values given in the Table.
The
theory of elasticity of
beams
gives
PI
or for rectangular cross-sections
EI PP
(4)
(5)
In which
/?
and
are as given in the Table, for the types under
consideration.
The
last
equation
(5)
may be
used for
all
computations as to
is
rigidity of flat springs (beams),
provided the elastic limit
not

SPRINGS
exceeded.
expression
121
The
is
only constant for the material which enters this
the
ratio of stress to strain

elastic
modulus of elasticity which holds up

Equation
(E);
to,
this is
simply the
but not beyond, the

this
limit;
hence any computation made by

(3)
formula
should be checked for safety.

purpose.
may be
used for this
To
illustrate,
I)
matic spring (Type
assume that a rectangular section prishas a length between supports of 1 = 30";
the load at the middle
load
lbs.; the deflection under this and the spring is made of a single Required the breadth (b) of the strip of steel H inch thick (h). spring, assuming the modulus of elasticity, = 30,000,000.
is
is
P 1,000
to
be
= 1.5
inches;
From
b
eq. (5)
= B
PI
-=rrr3
holt
1 1,000 X = ~ X ~
27,000
30,000,000
1.5
X S 12 = X 27
M+
n
2
inches.
This gives the width

deflection to load;
of spring for the required relation of the

is,
it
that
gives a spring of the required
stiff-
ness, provided the stress

limit.
It is
produced does not exceed the
elastic
if
necessary to check the spring as found above, for

is
the elastic stress

set,
passed, the spring not only takes a permanent
but the required ratio of the load to the deflection will not be
secured.
On
the other hand,
it is
often important for

is
economy
of material to use as light a spring as

or, in
consistent with safety;
other words,
it is
important not to have too low a working

load.
stress
under the
eq. (3)
:
maximum
3
From
P
PI = Ablt
stress is
it
1000
30
64
"
2x2.84X9
beyond the
II2
'
5 Ibs -
P r
1-
mch
This
steel,
elastic limit of
any ordinary grade of
hence
is
probable that some different form of spring should

in the thickness of the
be used.
plate,
change could be assumed, as
and new computations made with the new data. A thinner plate would reduce the stress, but it would demand a wider spring
for the required stiffness.
more general method
will
now be
com-
given,
by which
it is
possible to determine the proper spring for

trial
given requirements without the necessity of successive

putations.
MACHINE DESIGN
122
From
eq. (3)
PI
Plh
Ap
From
eq. (5)
Ap
BPP
BPl
pP
bh3 =
^r
pP
From
eqs. (6)
and
(7)
Plh
A
p
From
eq. (3)
PI
(9)
-TJre- J
and
the span
(/),
:
r c
Pl
(9)
The two
signing a
equations
spring
(8)
are in convenient form for dedeflection

(#),
flat
when
load (P),
and the material are given.

section prismatic
flat
Example
is
The span
of a rectangular
spring (Type
I) is
30 inches; and a load of
1,000 lbs. applied at the middle

inches.
If
to cause a deflection of 1.5
the
modulus
of elasticity be 30,000,000
and the
safe maxi-
mum working stress be

From
eq. (8)
:
taken
at
50,000
lbs.
b.
per sq. in.,* required
the dimensions of the cross-section, h and
I 50,000 X 900 rr PF = ~ mch. = -- X -2~2 =K ^ Ed 6 6 30,000,000 X 1.5 Taking h = ^ inch, to use a regular size of stock, p
will
be
somewhat
less
than 50,000, or
50,000 \\f% :-;
1
.
\ p
47,000.
From
b
eq. (9)
=C
~Pl
9 pnr is
1000X30X1024 3 = = - X 2 47,000 x 25
to
3Q.2 inches.
* If the spring
provided with stops

to
prevent deflection beyond a certain
amount, the
the material.
stress
due
such deflection
may
be nearly equal to the elastic limit of

that
is
very small factor of safety
is all
necessary.
SPRINGS
If this
123
width
is
inadmissible, a laminated or plate spring

article.
(8)
may
be used.
It will
See next
be noted that equation
does not directly involve

b.
either the load

a
or the breadth of spring

of given span
(/),
It is
evident that
(/?),
if
beam
(flat
spring)
and thickness
is
caused
to deflect a
given
amount
is
(),
the outer fibres will undergo
a definite strain
which
not dependent upon the width of the

to
beam
(b),
nor upon the force required
produce
this
change
in
relative positions of the molecules.
As
h,
the unit strain multiplied

it
by the modulus of
this stress
elasticity equals the unit stress,

/,
follows that
may
be computed from
and
<?
(which determine
the strain), in connection with E.

(b)
is
If the
breadth of the
beam
increased, the force (P) required to produce the given de(<?)
fection
will
be proportionately increased, but the intensity of
stress is not affected
by these changes alone.

rela-
This same conclusion may be reached from the following

tion,* in
which
/>
= the
radius of curvature due to load.

-v-
EI __ = ~TT = EI M
'
-- =
pi
2
Eh
.
.
(10) K
'
P~
is
()
simply proportional
(/>),
It
appears from eq. (n) that the stress

(h)
to the thickness
and the radius
of
curvature
for
any
given value of E.
so that eq. (10) or
(10)
The span /, and the deflection 8, determine p, (n) may take the place of eq. (8). Equations
and
(11)
are important in connection with the theory of
laminated springs.
42. Laminated, or Plate, Springs.
It
was shown
in the precedflat
ing article that the

fixed
maximum
thickness of a simple
spring
is
when
the
span, deflection, and modulus of elasticity are

intensity or
known, and the

[See eq.(8).]
it
working
stress
has been assigned.

(h)
will
With the value of the thickness frequently happen that a simple spring
breadth
(b)
thus limited
require exit
will
cessive
to
sustain the given load, and
is
often
necessary to use a spring built up of several plates or

*
leai
See Church's " Mechanii
3,"
page 250.
124
MACHINE DESIGN
Example:
in.;
= 60,000 lbs. per sq. P = 1,000 lbs.; / = 3o"; = 2", and = 30,000,000. A simple prismatic spring of
/>
rectangular section, with load at the middle of the span (Type

to
I)
meet the above requirements would have:
K -=rr =tX 6 h
Tr
pP
I 5
60,000
'
900
30,000,000
=
2
.15 inch.
PI ^ 7= 3 = =G X 2
p h
2
00
60,000
X 3 = x .0225
,
"*
SS/i inches.
This spring, consisting of a plate

inches wide, with a span of 30 inches,
ble one for
.15 inch thick

is
and
^)4
evidently an impracticathis plate
any ordinary
case.
Suppose
be
split into six
strips of equal width, each 33.3-^-6 = 5.5" wide, and that these
strips are piled
upon each other
as in Fig. 25;
then, except for
friction
between the various
strips, the
spring would be exactly
equivalent* as to stiffness and intensity of stress, to the simple
spring computed above.
While the form
of
laminated spring
cases, an-
which has
just
been developed might answer in some
other form, based

is
upon
the " uniform strength"

It
much
better for the ordinary conditions.
beam (Type II), may be developed

example except
as follows, taking the
same data
as the preceding
II, Fig. 20.
that the spring
is
to
be of Type
In the simple spring, Type
II,
Table IX
J rr P? = ~ X = Kp7T ho
60,000
000
30,000,000
1,000
^_ [__ 60,000
.22
e;
inches.
c A
PJ_ 2 TT-2 p h
;X x 2
X 30 w _^ = x .0506
14.8 inches.
laminated spring for the case under consideration
may be
in-
derived from this simple spring by imagining the lozenge-shaped

plate to be cut into strips
which are piled one upon another as

size of stock,
dicated in Fig. 26.
The
thickness of .225 inches does not corre-
spond
stock.
to a regular
commercial
however, and
it
will
usually be better to modify the spring to permit using standard

If
a thickness of
%"
be assumed for the leaves or plates,

(8) of
the stress, as found from eq.
the preceding article becomes:
SPRINGS
I2 5
p r
hEo = A r
/-
.25
^0,000,000
66,700.
900
If this stress is
considered too great, steel f\" thick might be
used,
4 when p =
^0,000,000
16
^0,000.
900
With
//
= TV, and
p
3
2
50,000,
= c
"
/>
/r
1,000x30x256 = 50,000 X 9
If
this spring,
middle, be replaced by
30" span, f\" thick, and 25.6" wide at the 11" 5 equivalent strips, each 25.6^5 = 5.
wide (nearly
l see Fig. 26, a laminated spring of good form A"),
and
practical dimensions will result.
In cases where the maxifixed,
mum
plates
allowable width of spring
is
a larger
number
if
of
may be
each
necessary.
Thus,
in the preceding
it
problem,
the
spring width must be kept within 4>2">

plates,
is
necessary to use 6
25.6-^-6 = 4. 27"
is
wide.
Fig.
In actual springs, the usual

27,
construction
that
shown by
in
which the several plates

tri-
have the ends cut square across instead of terminating in

angles.
These springs approximate uniform strength beams, and may be computed by equations (8) and (9) of Art. 41, remembering that b
;/
is
the breadth of the equivalent simple spring.

of plates
b
{
Or,
if
is
the
number
and
the breadth of each plate in the
laminated spring, n
= b.
not strictly applicable
The
when
last of these formula?, eq. (9), is
the ends of the plates are cut square across;
but
it
may
It is
generally be used with sufficient accuracy, provided the successive plates are regularly
shortened by uniform amounts.
quite
common
practice to have two or
more
of the plates extend
the full length of the spring.
This construction makes the spring

prismatic types (Type II
IV, depending upon whether
Is
a combination of the triangular and
and Type
the spring
I, is
or
Type V and Type

at the
supported
in
ends, or
a cantilever).
Mr. G.
R. Henderson
discussing the cantilever form (Trans. A. S.

126
MACHINE DESIGN
E., Vol.
XVI), says: "For a spring with length we would have (see eq. 5)
M.
all
the plates full
API'
Enb.h
so for one-fourth of the leaves full length, the deflection
would be
decreased approximately one-fourth of the difference between
6P1
Enb
By
h3
and
4PI
Enbh
3
- or 3
5.5P/ 3 "
Enbh
'
similar reasoning, for a spring loaded at the middle
and
supported at the ends, with one-fourth the plates extending the
whole length of the spring,
11
PI 3
32
En b
h3
This
may be
the
otherwise stated as follows

of full-length leaves is one-fourth the total
When
number
instead of
number
of leaves in the spring, use
^B
(8)
instead of
and \
in equations (5)
and
of the preceding article;
the values of
and
K being those given for the triangular forms,
Type V, as the case may be. shown in Fig. 27 is initially curved (when free), which is common practice. The best results are obtained by having the plates straight when the spring is under its normal full Type II The
or
spring
load
(if
this is practicable)
because the sliding of the plates upon

is
each other, with the vibrations,
then reduced to a minimum.
The
several plates of a laminated spring are usually secured
by
a band shrunk around them at the middle of the span.

stiffens the
This band
spring at the middle, and one-half the length of the

Fig. 27)
band (%
not
if
I,
may be deducted from

I
the full span to give

It is
the effective span to be used as
in the
above formulae.
uncommon
to
is
make
the longest plate thicker than the others,

full
but one plate
given the
length of the spring.

all
This cannot
of the plates
be looked upon as desirable practice, however, as

are subjected to the
same change
in radius of curvature; hence the
thicker plate
is
subjected to the greater stress.
See equation (11).
The
following formulas (derived from the preceding)

flat
may be
used in computing
springs; but
it
must be remembered that
SPRINGS
there
is
127
always
liability of
considerable variation in the modulus

to
of elasticity, hence such
computations can only be expected
give approximations to the deflections

tests of actual springs.
which
will
be observed by
These computations
will
be sufficiently
to
exact for
many
purposes; but
when
it is
important accurately
determine the scale of the spring

actual tests must be made.
(ratio of deflection to load),
In using these formulae the follow-
ing rules should be observed.

I.
When
the several plates are secured by a
band shrunk,
is
or
forced, over them, one-half the length of the band
to
be sub-
tracted
from the length
of the spring to get the effective length
of the spring.
II.
When
If
the plates
have different thicknesses, the

having the
stress
should be computed
III.
for the plate
maximum
thickness.
more than one
plate has the full length of the spring,
an appropriate modification of the values of the coefficients
and
K
B
should be made.
Thus, when one-fourth of the
total
number
of plates are full length,
\\
and \^
K should be used instead of

I, II,
and
(Type II or V)
in equations
III,
and IV, below.
EQUATIONS.
O
= B _ E
PP
.
n bx h
.3 3
(I) v J
*- b -^p
<)
'-"-45
>-*-
<
<-v>
-,
V -
<v
=
i^C
(V!)
P=
bt
fL
n
(VII)
= -r-^-p =
A
p h
^1 p h
n
(VIII)
"
128
MACHINE DESIGN
elastic limit
Experience shows that thin plates have a higher

than thick plates of similar grade of material.
In the practice of
a prominent eastern railway company, the values allowed for the
maximum
intensity of stress in flat steel springs are, for:
Plates "
\ inch T\
iV
"
thick,
^ = 90,000
lbs. sq. in.
"
" "
" "
p= 84,000 p= 80,000 p- 77,000 p= 75,000
" "
"
The above
values are satisfied by the equation p = 60,000 +

is
-^
*7
COO
in
which h
the thickness of plate in inches.

for the greatest stress to
is
These values are

the stops.
which the material
can be subjected, as when the spring
deflected
down
against
value
The modulus of elasticity, E, may vary considerably; but its may be assumed at about 30,000,000 in the absence of more
In designing a
definite data.
new
spring, the value of

is
is to
be found from
equation (III); then b t

deflection
found by equation (VIII).
The
other
formulae are useful in checking springs already constructed, for
due
to a given load, or the reverse;

If
for safety, etc.
43.
Helical Springs.
a rod or wire be
wound
into a flat
ring with the ends bent in to the centre, Fig. 28, and two equal
and opposite
will
forces,
+ P and
P, be applied to these ends
(perpendicular to the plane of the ring) as indicated, the rod
be subjected to torsion.
rod be
If a longer
wound into
forces,
a helix, with the two ends turned

is
in radially to the axis, the typical helical spring
produced.
If
two equal and opposite
+ P and
lie
P, act on these ends,
along the axis of the helix, they induce a similar stress (torsion)
in the rod, but as the coils
do not
is
in planes perpendicular to of direct stress along
the line of the forces, there the rod.
component
This direct
stress increases as the pitch of the coils in-
creases relative to their diameter; but with ordinary proportions
SPRINGS
of springs, the torsion alone
ternal forces
lie
129
need be considered, when the ex-
along the axis of the helix.
The
following notation will be used in treating of helical
springs of circular wire, subjected to an axial load:
P = the
r
(/
force acting along the axis.
= the
radius of the
coils, to
center of wire.
= the diameter of wire. /> = thc maximum intensity of stress in wire (torsion). 7 p = the polar moment of inertia of wire. = the transverse modulus of elasticity. = the "deflection" (elongation or shortening) of spring. = the number of coils in the spring.
s
'>
;z
= the
length of wire in the helix

helical spring
=2
n (approximately).
to
Suppose a
to
under an axial load
be cut across
of this section
the wire at any section,
and the portion on one side

that the
be considered as a free body, Fig. 29.

equilibrium
Neglecting the direct
stress,
demands
force shall
equal the stress
moment (Pr) of the external couple, or moment of resistance
m6
3 p d for circular section
'
If this free
portion of the helix

lines in Fig. 29,
it
is
straightened out, as indicated

its
by the broken
till
direction
is
perpendicular
still
to the radial end,
will
appear that the moment Pr
equals
the
moment
same
of resistance,
p d 16
3
.
Since the stress and strain are

it
the
in this helix
and the straight rod,

is
appears that the

in
energy expended against the resilience

(the length of wire affected
the
same
both cases
remaining constant).
Or, as the force
(P)
and the arm

are equal.
(r)
are the
same
in
both conditions, the distances

stress
through which
(p)
this force acts to

If
produce a given torsional

/
a straight rod of length
is
subjected to a
torsional
moment
Pr, the angle of twist being"- (in -measure), then
Pr =
[See Church's
9
-^
236].
"Mechanics," page
130
MACHINE DESIGN
the rod
is
The energy expended on

tiplied
is
the
mean force applied mulIf the
by the distance through which this force acts.

is
load
gradually applied, this energy
y Pr
2
In the case of the
corresponding helical spring, the mean force (J4P) acts through a distance equal to the "deflection" of the spring (<*), or the energy
expended
in the
is
j4
P8
is
>
As pointed out above, the energy expended
two cases
the same, or
y
2
Pro.
=
b
Pb
Ttd 4
.'. a
= r
r.Pr =
o-I
R V^ = - X X
E = dd
4
s
s =-L
64
(1)
Equation
to
(1)
may be used
for finding the load corresponding
an assigned deflection in a given spring.
The
equation can
to
be put in the following form for finding the deflection due

given load
S
^^
d4
{) (2 )
which
Or
the equation
may be employed
r,
for designing a spring in
the load and deflection are given, by assuming any two of the
three quantities,
latter
d and
n.
The most
convenient form for this
purpose
is
usually,
n
These equations
^T?
good only within the
is
(3)
for rigidity hold
elastic
limit of the material, as strain within this limit.
simply a ratio between stress and
It therefore
any of the above indicated computations

high for safety, or too low for economy.
becomes necessary to check for strength, and it will

is
often be found, after thus checking, that the stress
either too
The formula
under torsion
is
for the strength of a solid circular-section
rod
SPRINGS
131
~
Pr =
pd r 16
-pd?
.'.
' '
pd
3
'
16 r
(4)
= -^p'>
*=^
i6Pr
(4) is for safe strength,
It is to
be remembered that as equation
the
can
load (P) should be the maximum load to which the spring be subjected; but equation (3) may be used with any load
and the corresponding deflection. Example: The load on a helical spring

corresponding deflection
elasticity of material
is
is
1600
lbs.,
and the
to
be
4".
Transverse modulus of
= 11,000,000, and the maximum intensity of safe torsional stress = 60,000 lbs., wire of circular section. To design the spring, assume d=y%", and r = \y 2 "\ from eq. (3),
n
625
X
64
11,000,000
4,096
1,600
last
1.5
X
512
19.4.
27
(4),
Checking
for the stress
by the
equation in group
16
1,600
X
be
*X
found
it
=
is
So> 2 o lbs.
125
safe,
This
stress is
to
but
considerably below the
limit assigned,
and
may be
desirable to
work up
to a
somewhat
can
higher stress.
Another computation can be made (with a smaller
d or larger
r),
and by a
series of trials, the desired spring
be found.
The
following order of procedure avoids this element
of uncertainty.
The
load being given, assume a diameter of wire
and value
of safe stress, then solve in eq. (4) for the radius of coil.
some convenient dimensions (not exceeding that computed if the assumed stress is considered the maximum safe value). Next substitute these values of d and r (with those Thus, given for P, d and E) in eq. (3) to find the number of coils.
this radius
Make
with the data of the preceding example, assuming

r
d=W
f
;
=
is
p d?
16
-7
n P
X 60,000 X = ~2 H 16 X 1,600 X
-
12s
"
512
1-79
If the yi"
rod
wound on an arbor
3" diameter, the radius to the
centre of coils will be about 1.81";
stress
I32
MACHINE DESIGN
lbs.
would be 60,500
this value is
per square inch.

it
This
is
so slightly in ex-
cess of the assigned value that
may be
permitted, especially as
steel.
a moderate one for spring
Substituting in
eq. (3);
1%
$ d*
64
It
E = Pr
s
11,000,000
64
fix
1,600
5-93
X 62^ = X 4,096
T T T
may
be desirable to
upon the radius
of coil, rather than the
diameter of wire, in the
first
computation, in designing a spring.

3
From
eq. (4)
^
In other cases,
it
l6Pr
Ttp
1 ^
\P~r =I -72\
v
,
.
(5)
may
be desirable to assume the ratio of the

(4)
radius of coil to the diameter of wire, then from eq.
In either of the preceding conditions, a standard

should be chosen.
In checking a given spring,
it
size of wire
may
If
be required to determine
If the
either the safe load, or the safe deflection.

case, eq. (4)
former
is
the
may
is
be used
directly.
it is
required to find the

eq. (4) in eq. (2)
safe deflection, substitute the value of
from
and the
result
12.57
n^P
(7)
Ed
The weight
material
is
of a spring is a matter of
some importance,
as the
expensive.
The
following discussion shows that the
weight varies directly as the product of the load and the deflection, inversely as the
square of the intensity of stress in the wire,
and
directly as the transverse
modulus
of elasticity.
Hence
for
a given load and deflection, economy calls for a high working

stress
and a low modulus
of elasticity.
From
eq. (4)
P =
p
p 10
7T
d3
r
also for a
member under
torsion,
X ^j1
[Church's " Mechanics," p. 235].
SPRINGS
133
(/
X
r
d$E
E
s
2 - r
rs
ii
4 -
11
4 ~
rnp dE.
4
(8)
5-^ .\P*--^
s
(9)
But the volume
of the spring
is
v-%*#l = yi*&Tn
(10)
.;P*=fK,.:v = 1fPt
The
weight
is
fr
E,
......
<)
directly proportional to the
volume; hence for
given values of
of the load
and
p, the weight varies simply as the product
and the
if
deflection.
All possible helical springs (of
same weight for a given load same material and worked to the same stress. It can be shown that a helical spring of square wire must have 50 per cent greater volume than one of round wire, the stress and modulus of elasticity being the same in both. The
similar section of wire) have the
and
deflection,
of the
round section
is
generally admitted to be best for helical springs
under ordinary conditions.
A
limit
in
small wire of any given steel usually has a higher elastic
than a larger one, while there
is
not a corresponding change
the
modulus
of
elasticity
with change in diameter.

is
This
suggests the use of as light a wire as
consistent
with other
requirements.
An
Adams,
extensive set of tests of springs, conducted
by Mr. E. T.
to stress
"
in the Sibley College Laboratories, indicates that the steel

is
such as
used in governor springs

lbs.
may
be subjected
varying from about 60,000
per square inch with
y
A
wire to
80,000
lbs.
per square inch (or more) in wire yi" diameter.
The
following expression
springs
may
be used to find the safe stress in such
40,000
-
1 5
,000
(12)
Mr.
J.
Springs before the
W. Cloud presented Am. Society
a most valuable paper on Helical

of
Mechanical Engineers (Trans.,
134
MACHINE DESIGN
Vol. V, page 173), in which he shows that for rods used in rail-
way
springs
lbs.
(W
to i T g" diam.), the stress
may
be as high as
80,000
per square inch, and that the transverse modulus of
elasticity is
about 12,600,000.
being subjected to the same
Two or more helical springs are often used in a concentric nest

(the smaller inside the larger);
all
deflection.
This
is
common
practice in railway trucks,
where
they
the springs are under compression when loaded.
If these springs
if
have the same "free" height (when not loaded), and

are of equal height
when
closed
down
"solid,"
Mr. Cloud shows
that the length of wire should be the

set for
is
same
in each spring of the
equal intensity of
n,
stress.
The
is
1
"solid" height of a spring
H=d
and the length of wire
=2
~ r n\
hence the num-
bers of coils of the separate springs of the set are inversely as the
diameters of the wire and inversely as the radii of the

the ratio of
r to
coils;
or
same in each spring of the nest. This conclusion may be somewhat modified when it is remembered
d
is
the
that the wire of smaller diameter
may
usually be subjected to
somewhat higher working

helices;
stress
than the larger wire of the outer
and
also that the wire of these compression springs is
commonly
the seats.
flattened at the
end
to secure a better bearing against
See Fig. 30.
SUMMARY OF HELICAL SPRING FORMULA.
Pr = -7pd
16
r
3
.
(I)
E.d
3 64 r n
(VII)
r:pd3
16
(II)
(VIII)
>>*'<!
(HI)
_ ~
64
Pt%
d4
d
(ix)
'
'--*>/&)
i6Pr
d*E
2
(IV)
3 64 Pr
(X)
' '
P
(V)
2A
(XI)
*P#
(VI)
SPRINGS
35
Two common
in Fig.
30
(a).
fit
One end
methods of attaching "pull" springs are shown of the spring shows a plug with a screw
This plug
is
thread to
slightly,
the wire of the spring.
usually tapered
and the
it
coils of the spring are
somewhat enlarged by
screwing
in.
The
is
other end of the spring shows the wire bent

lies
inward
to a
hook which
(I)
along the axis of the helix.
The
(VIII)
former method
usually preferable for heavy springs.

to
Formulae
to (VII), inclusive, relate
strength;
to (X), inclusive, relate to rigidity, or elasticity.
In the absence of more exact information as to the properties

of the material of
which a
steel helical spring is
made, the
fol-
lowing values
may
be taken:
E
a
12,000,000,
p
44. Spiral
40,000
Springs
15
,000
or
Helical
in
Torsion.
The
following
formulae for either true spiral or helical springs subjected to torsion are derived from
leaux.
<p 9
"The
Constructor," by Professor Reu-
PRl EI
PR
'
In which
P = load applied to rotate axle, R = lever arm of this load, = angle through which axle turns,
<P
= length of effective coils, = modulus of elasticity (direct), / = moment of inertia of the section.
/
CHAPTER
45. General Considerations.
VI
RIVETED FASTENINGS
The
is
simplest form of fasten(Fig. 31), a straight
ing
b,
is
the rivet.
It consists of
c,
a head a
shank
and a second head
which
formed while hot and known
as a point.
gether,
is
When
it
is
desired to rivet two pieces

drilled as
MN
to-
mating holes are punched or

a
larger in diameter.
shown, the rivet

is
heated white hot and pushed into the hole which

little
purposely
made
The head
is
held up firmly
by a heavy bar or sledge and the point may be formed with a hand hammer, or with the aid of a forming tool or set. In riveting on a large
against the plate
scale
this
operation
is
per-
formed by hydraulic or pneu-
^p
matic machines.
The
relative
merits of the two methods will
be more apparent after further

discussion.
The
rivet is a per-
be
Fig. 31.
manent fastening and cannot removed without the destruction of either head or point.
It is largely
used in structures
such as bridges, the framing of buildings, ship work, boilers,

tanks, etc.
Fig. 32
shows various forms

at
of rivet
heads and points.

for small rivets
The
form shown
T
5
g-
is
most commonly used
in.
diameter, which are driven without heating, for
up to such work
The form at C is much as light tank and smokestack work. used in ship work, or wherever smooth exterior surfaces are
desired.
In machine work, where great accuracy
is
required,
the holes are reamed, and

pletely to
fill
the rivet carefully fitted so as
com-
the hole;
both heads in such cases are usually
countersunk and formed cold.
n6
RIVETED FASTENINGS
137
When
sile
a rivet
is
"driven" hot
it
shrinks in cooling, drawing

cold
it
the riveted parts firmly together.

stress
When
and
is
under a
tenit
due
little
to this shrinking,
for the
same reason
it
is
always a
smaller than the hole which

hot.
originally
com-
pletely filled
effect
when
rivet,
The
tensile stress
it
due
to this cooling
cannot be accurately determined as
depends on the temit is
perature of the
and the manner
in
which
driven.
are
Rivets
are, for this reason, unreliable as tension
members and
(Fig. 31)
is
seldom
so used.
In most cases the parts
M and N M
have the load
applied as shown, and the tendency

relative sliding
to shear off the rivet
and produce
load, P',
of
to
between
and N.
The normal
due
to the tensile stress in the rivet, holding the surfaces
M and N firmly in contact, sets up a frictional resistance equal

!'.P'
which opposes the action of P.
From experiments made

,
Conical
Sutton auiion
Conoidal
Countersunk ante Co
VPan
j
5fe
J
Fig.
3:
by Stoney *
at
it
appears that
lbs.
this frictional resistance
may
be taken
about 10,000
per square inch of rivet area.
Experiments
it
by Bach, and
evident that
stress
if
others,
show a much higher
resistance, but
is
the
normal pressure
of the rivet is such that a

is
equal to or greater than the elastic limit
induced, the
permancy of the resistance cannot be relied upon. In some French and German practice the design of the joint is based entirely upon the frictional resistance, but in England and America it is neglected, and the design based upon the tensile and shearing strength of the plates and rivets. Riveted joints are of many forms de46. Forms of Joints. pending on the character of the work to which they are applied.
In structural work, such as bridges, they are used simply to
*
"Strength of Riveted Joints," page
75.
I38
resist direct loads;
MACHINE DESIGN
but in boiler construction, and similar work,
they must not only resist direct loading but must also be tight
against fluid pressure.
This
last
requirement materially affects

of joints for
the proportions of the joint,
and makes the design
withstanding fluid pressure most important.

divided into two general forms.
(a)
Riveted joints are
Lap
Joints,
where the sheets
to
be joined are lapped on
each other and riveted as shown in Fig. 33 (a). (b) Butt Joints, where the edges of the sheets to be joined abut against each other, and have auxiliary butt straps or cover
plates riveted to the edge of each, as
lap joint
and these
Fig. 33
shown in Fig. 33 (e) and (f). more rows or seams of rivets, rows way be arranged in the form of " chain" rivet-
may have one

,
or
ing, Fig. 33
( c ).
(b)
or in the form of zigzag or staggered riveting,
butt joint
side of the
may have one or more seams of rivets on each joint, and these may also be arranged in either chain
shown
that
in Fig. 33 (f)
or staggered form, as
and
(g).
are very numerand the student is referred to any treatise on boiler work for fuller information on this point. The distance between rivets along the seam is called the pitch
The combinations
may
be thus
made up
ous,
or spacing,
tion
and
will
be denoted by
joint
it
s,
Fig. 33 (b)
An
examina-
of
any riveted
pattern as
will
show
that the arrangement of
rivets, or
may be
called, continually repeats itself as

joint, the repetition
the
"seam" extends
along the
occurring with
is
the greatest pitch, where the pitch of the various seams

as in Fig. 33 (h).
unequal
not
unit strip
is
equal in width to the pitch, the

the pitch of
is
maximum
the same.
pitch being taken
when
all
seams
is
The
transverse pitch
the distance between the
centre lines of adjacent seams Fig. 33 (b) and will be denoted The diagonal pitch is the distance between the centre of by s
. t
a rivet and that of the one nearest to
it
diagonally, in the next
row, and will be denoted by s d Fig. 33 (d). The margin is the distance from the edge of the plate to the center line of the nearest
row
of rivets, as e Fig. 33 (d).
It is
sometimes defined as the
distance from
the edge of the plate to the edge of the rivet hole.
RIVETED FASTENINGS
47. Stresses in
139
that exist in
Riveted Joints.
The
stresses
the various
members
of riveted joints are complex,
and do not
admit of refined calculation.

to the
Not only
are the plates subjected
apparent direct stresses of tension and compression, and

J
L
^
00
,
!(D
^V-^V
U-7
WR^
ête-cfr
AA
(e )
A
(0
\ip^jrxp-tp7
*m^
Fig. 33.
(a, b, c, d, e,
f,
g, h.)
140
the rivets to shear
MACHINE DESIGN
and compression, but often there are also difficult to analyze and provide for
lap joint, as that
bending actions which are

mathematically.
33
(a),
Thus a simple
shown
in Fig.
when
subjected to a load, tends to take the form shown
in Fig. 34.
The
force applied tends to
draw the
plates into the
same
plane, putting a bending action on the plate
greater tensile stress on the rivet head,
and rivet, a and a concentrated crushing

frictional resistance is
load on the corners of the sheets.

entirely destroyed
The
when
the conditions illustrated in Fig. 34 exist.
The above
defects are
more marked
in the lap joint than in
the double strapped butt
joint, as in the latter the plates are ini-
tially in line
and the condition shown
in Fig.
fill
34 cannot occur.
even here the rivets do not completely
the holes
when
cold,
But and
hence some bending of the rivet and concentrated crushing on the

plate
must
is
result.
Again,
while the quality of the material
forming the
joint
may
be well
known
or determined, the work-
manship
not so easily controlled and
may be
very defective
and
yet not
tests
show on the
exterior;
and while there have been
many
made
to find the ultimate strength of riveted joints,

of Mechanical
* Proceedings
Institute
Engineers,
1881,
1882,
1885,
1888.
Watertown Arsenal Reports, 1885, 1886,
1887, 1891, 1895, 1896.
RIVETED FASTENINGS
such
joint
tests
141
show only the stress at which a certain element of the failed, and do not throw any light on the distribution and
progress of the stresses in the various individual

ing the
test.
members dur-
Such
tests
have usually been performed on joints
made
of straight plates while in practice these are often curved.

therefore, while giving the only data available
These experiments,
in designing.
relative to the ultimate strength, should be used with
judgment
For these reasons the theoretical formulae deduced
for the design of riveted joints, as a rule, take cognizance only
of the apparent simple stresses
and provide
for the
unknown by
one of the
means
It
of a factor of safety.
has been found that riveted joints
may
fail in
following ways
(a)
Shearing of the rivet as in Fig. 35
(a).
(b)
(c)
Rupturing of the plate by tension as

Tearing of the margin as
Crushing the
Shearing of the margin as in Fig. 35
in Fig. 35 (b).
in Fig. 35 (c).
(e).
(d)
(e)
(f)
plate, or rivet as in Fig. 35 (d).

rivet holes
Rupturing of the plate diagonally between
by
tension, in staggered riveting.
Where
due
on
to
the joint
is
complex
in form, ultimate failure
one or more of the above causes.

Figures
may be The Watertown Arsenal

I,
reports include cuts of ruptured joints which are very instructive

this point.
(a)
and
(b),
on Plate
are reproduced
from these
failure
reports,
and show very
clearly all the

(c)
ways
in
which
has
may
occur in the plate.
Fig.
shows a
rivet that
(d)
been tested
to destruction in single shear, while Fig.
shows
one that has been similarly tested

It is
in
double shear.
obvious that no riveted joint can be as strong as the un-
perforated plate, since the very fact of
making
holes in
it
reduces
ratio
the cross-sectional area in the line of the rivet holes.
The
of the strength of the weakest element of the joint, to the strength of the imperforated plate,
ficiency of the joint.
is
called the relative strength or effirst
The
this
expression
It
is
is
more suggestive
if,
and
will
be used
in
work.
desirable to reduce the
strength of the plates as
little
as possible
by perforation; and
Ca;
142
Plate
I.
RIVETED FASTENINGS
therefore, the correct relation
43
between the
size of rivet
is
and cross
is
section of perforated plate, for equal strength,
established, an
excess of strength in other directions, as marginal distance,
not a defect but good design, as
it
insures that the full strength
of the perforated plate will be in service before rupture
can occur.
well-designed joint should hence
along the line of the rivet holes, at

destroy the rivets;
joint should
by tearing of the sheet about the same load as will

fail
and the
relative strength of a well-designed
be the ratio of the cross section of the perforated to
that of the unperforated plate, the shearing
and crushing
resist-
ance of the rivets being equal to the former.

not
exist,
If this equality
does
the relative strength of the joint can be
made
greater
by strengthening the weaker of these elements

the stronger.
48. Marginal Strength.
at the
expense of
The width
of
margin
is
independ-
ent of the proportions of the other elements, and hence can be
made
and
sufficient to
It
(e).
prevent tearing or shearing, as in Fig. 35 (c) has been found that, with the usual proportions, if
the margin be the rivet,

it
made equal
to
one and one half the diameter of

of the
will
be safe against both shearing and tearing from
rivet pressure.
Committee
Master Steam Boiler Makers
Association recently recommended, as a result of experiments,

that the distance
plate be
from the center
of the rivet to the edge of the
made
twice the diameter of the rivet, in order to insure
excess strength enough to either shear the rivets or rupture the

plate
by
tension.
It is
important, however, that the margin be not

as this
excessive in boiler
work
makes
it
more
difficult to
steam-tight joint by calking the edge of the plate.
make a The con-
sideration of the marginal strength can hence be omitted as far

as
its
it
influence on the relative strength of the joint

to
is
concerned,
as
can be made
depend on the diameter

and
Diagonal
Pitch.
of the rivet.
It
49. Transverse
has
been
deter-
mined, in a similar way as above, that in chain riveting the

transverse pitch should not be less than twice the diameter of the
rivet
or 2d, where d
It
is
the diameter of the rivet,
and
2.5
is
better.
has
also
been
demonstrated
mathematically
(see
Cathcart's " Machine Design," page 148) that in staggered riveting

144
MACHINE DESIGN
the transverse pitch should not be less than 0.4 times the pitch
along the seam, in order to avoid rupture along the diagonal

pitch,
and a greater distance

2
is
recommended
for safety.
Unwin
d,
(page 123) gives

riveting,
d as the
minimum
diagonal pitch in staggered

1.7
which would make the transverse pitch

that
and
recommends
strength.
somewhat greater distances be used
for
added
An
examination of the practice of several boiler-makappears from the above that the trans-
ing and insurance concerns shows that these values check fairly
well with practice.
It
verse pitch can also be

rivet,
made
to
depend on the diameter
of the
though
it is
not a direct function of the rivet diameter.

Since the margin
50. Theoretical Strength of Riveted Joints.
and transverse
pitch can be assigned from the diameter of the
rivet, three of the
ways
in
which a
joint
may
fail,
namely
c,
d and
/, page 141, can be omitted from the theoretical discussion of the strength of riveted joints, leaving a, b and e to be considered;
the problem being so to proportion the rivet and the pitch along
the seam as to give equal strength against failure in any of these
three ways.
Let,
rivet in inches.
d = diameter of
5 = pitch of rivets in inches. p = tensile strength of plates in pounds per sq. inch. p = crushing strength of plates or rivets in pounds
t c
per sq.
inch,
if
rivets are in single shear.
p\ = crushing strength of plates or rivets inch, where 'rivets are in double shear.
in
pounds per
sq.
= shearing
when
strength
of
rivets
in
pounds per
inch,
in single shear.
p\= shearing
It is
strength of rivets in pounds per
sq.
inch,
when in double shear. known that the unit shearing

is less
resistance of a rivet
is
greater in single shear than in double shear,
while the unit
crushing resistance
in single shear than in double shear.

a).
Consider
first
a simple lap joint (see Fig. 33

is
The
tensile
strength of the unperforated strip
P = stp
(1)
RIVETED FASTENINGS
145
along the seam of
The
tensile strength of the perforated strip
rivets is
T=
its
(s-d)tp
is
(2)
In the simple lap joint there

shearing strength
is,
but one rivet per unit strip and
5 =
The
it
T^
(3)
resistance to crushing of the rivet or the plate against

is,
which
bears
C = dtp*
relative strength,
(4)
For uniform strength against rupture and hence for greatest
T = S
Equating
(3)
=C
c
and
(4)
--p
4
.
= dtp
1.2*1
\ d
P, 1-
Equating
(2)
and
(3)
(sd)tp =
t
.
-d2
4
p
pt
=
(s
a
s
_.
The
relative strength
T P
d)
x
d
s
stp
In
Double-Riveted
Lap
Joints.
similar
joint.
may
be developed for any other form of

P
way equations Thus for double-
riveted lap joints
= =
1.27
and
*
[Pt
It is
known from experiments on

little
it is
indentation that the resistance to indentation

its
depends very
area.
on the form of the indenting body but mainly on
projected
Hence
10
customary
to take the resistance of rivets to
crushing as propor-
tional to their projected area.
146
MACHINE DESIGN
relative strength
The
d and
will
be greater than in the case
of the single-riveted lap

to d.
joint since s is greater in proportion
Single-Riveted Butt Joints with double cover plates.
Here,
'
d=.64 tf
a
= =
i.5 7 d p'
71
s
d
s
and the
relative strength
Double-Riveted Butt Joints with either chain or staggered

riveting
and double cover
plate.
Here,
*t*
Note that the

ative values of s
relative strength
d
,
and compare the
rel-
and d
used
in this case with those in single riveting.

It is evident
Thickness of Cover Plates.
that where only
one cover plate

of the
is
its
thickness should not be less than that
and in practice, single cover plates are made a little thicker than the main plate to insure an excess of strength. A butt joint with a single cover plate is shown at Fig. 33 (e). A
main
plate;
joint of this kind is really equivalent to
two lap
joints.
They
are used where a smooth surface

the longitudinal
is
desired or in such places as
sufficient strength.
than half
boilers where a lap joint has Double butt straps should not be made less the thickness of the main plate, and, for the same
seams of steam
reason as above,
it is
not unusual to increase their thickness to
about y$
where the cover plates are the same width.

is
Where
as in
the outer cover plate
narrower than the inner
plate,
RIVETED FASTENINGS
Fig. 33 (h), the outer cover plate
is
147
often of the
-X. to
same thickness
/.
as
the
main
plate
and the inner one from

Equations
for
T\
51. General
Riveted
Joints.
The
funda-
mental equations for riveted joints

form.
the
all
may be
put in a more general
The unit strip, as before, maximum pitch being taken

kt
is
of width equal to the pitch;
for
such width of unit strip
if
rows do not have the same

Let
pitch.
=p
in
-5-
Pa
k2
unit
strip
single
shear
n = number of rivets per = and m number of rivets per unit

p\-?-p\\
strip of joint.
strip in
double shear per unit
The
strip is
general expression for the net tensile strength of the unit
T=(s-d)l Pl *
The
in the unit strip
is
(1)
general expression for resistance to shearing of the rivets
m* n* S- A+ f ....
d?
2 d?
,
(2)
The
general expression for resistance to crushing of the rivets
in the unit strip is
C =
The
nd
k l pa
+ m
k 2 p' s
(3)
tensile resistance of the solid strip is
P = stp
Equating 5 and C,
eqs. (2)
(4)
and
s
(3)
and solving
2
for
d
(5)
= - X
nk,p + np
a
m k pj -7+ 2 m pi
....
which gives the proper diameter of

plate,
in
rivets for a given thickness of
when
the
number
of rivets in single shear
and the number
double shear and the corresponding shearing and crushing
resistances are
*
known.
do not
all
Where
the rows of rivets
have the same pitch, as

the sheet yield at
in
some forms
it
of
butt joints, the outer row or that farthest from the edge of the sheet has the greatest
pitch (see Fig. 33 h).
yield along this outer
It is
evident that
rivets;
it
if it
all
by tearing,
will
row of
for
cannot tear along an inner row without
shearing the outer row of rivets, and

ing
all,
cannot shear one row of rivets without shearyield
in
which case the
joint
would
by shearing of the
rivets
and not by
tearing.
148
Equating
MACHINE DESIGN
and
S, eqs. (1)
and
(2),
and solving
for s
'-^
Equating S and C,
eqs. (2)
/A
and
d
s
+ ''
'
(6)
(3)
-d
4
{n
m p') =
=
;
;
(n k t
+ m
f
k 2 p'
#*
Tz(P(np 4(nk
1
+ 2 m p + mk
;
s
p:)
w) KU
)
Equating
T
t
and
S, eqs. (1)
2
and
(2)
and solving
.
stp =
If the joint is
*d r (np
4
+2mp;)+dtp =P.
t
(8)
designed for
maximum
relative strength,
S = C, hence any one

form, or dividing
(2)
of these three quantities divided
by
T= P
gives the relative strength () of the ideal joint, for any given
by
(8)
np
+
.
2rn p
2
E=
i= P
~d
4
(w
p
t
mp
')
s
-\-
Substituting the value of d
as given
by
eq. (7)
2
and dividing
-d2
numerator and denominator by
4
(n
mp
),
kx
+ m
1
k2
If all rivets are in single shear,
(90
t
n
If all the rivets are in
k,
double shear
E=
1
-,
(9")
mk
p\
RIVETED FASTENINGS
Equation
(9)
49
is
applies to any form of riveted joint.
It
useful in finding the limiting relative strength of joint for any
form and materials; the actual proportions adopted may give a

lower relative strength, but can never give higher relative strength.
These general equations were

for the
originally
due
to
Professor
William N. Barnard, who also suggested the above expressions
maximum
in
relative strength in the general case.
The forms
due
which
eqs. (9), (9')
and
(9)" are
now
given are
to Professor
H. F. Moore.
in practice for boiler
The
following are rough average values of the relative strength
of joints as
made
work:
55
Single riveted lap joints
Double riveted lap
joints
70
65
>
Single riveted butt joints
Double riveted butt
joints
75
Triple riveted butt joints
80
85
Quadruple riveted butt
joints
52. Practical Considerations Affecting Proportions of Riveted
Joints.
It
is
to
be especially noted that the proportions of
riveted
joints
as given
on equal strength of
also that
rivet
by the foregoing equations are based and plate, and that any variation
It is
therefrom will destroy this theoretical equality.
apparent
any variation
in the strength of the material
used would
affect the proportions as given
table of rivet
by these equations, and that a diameters and pitches would have to be very exIt is
tensive to cover the entire range of practice.

in practice,
an advantage
has also been
however, to adopt regular diameters and pitches for a

It
given thickness of plate and form of joint.
found that as the thickness of the plate increases, the corresponding theoretical diameter of the rivet sometimes becomes too large
to
be easily driven, especially in the case of simple lap
joints.
In the case of boilers, or wherever fluid pressure must be withstood, the theoretical spacing
must sometimes be modified
in
order that
it
may
not be so great as to prevent the making of a
steam-tight joint.
expressions for T,
strength and
Wherever such variations are made the general S and C can always be used to check the
in
show
what direction the
joint
may
be strength-
150
MACHINE DESIGN
making
it
ened, with the fundamental object of
strong enough
in all other directions to insure full service out of the plate itself.
It
may
be noted that the bearing resistance of a rivet varies with

therefore, the rivet chosen
the diameter, while the shearing resistance varies with the square
of the diameter.
If,
be smaller in
the diameter
diameter than would be given by the theoretical equations, the

shearing resistance alone need be regarded
;
while
if
of the rivet be greater than the theoretical diameter, the bearing
pressure only need be considered.

If the joint to
be made does not have to withstand

the
fluid or
gaseous pressure,
plate,
design
can,
for
ordinary thickness of
be made to conform closely

but when
to the theoretical proportions
for equal strength;
fluid or
gaseous pressure must be
withstood, as in boiler work, the spacing of the rivets for thick

plates
ness;
must be less than the theoretical spacing to insure tightand in all cases as the plates increase in thickness, the
rivet, as
diameter of the
already noted,
is
for practical reasons
reduced in diameter from that required for equal strength.
The
between diameter by average practice may be expressed by the equation,

relation
of rivet
and thickness
\/t
of plate as fixed
1.2
(10)
For plates above &" thick
this
equation will give rivets smaller

all
in diameter than required for equal strength in
directions.
As before pointed
out, the rivets in such cases

If the
need only be
to
checked for shearing strength.

determined by equation
(10),
diameter of the rivet be
and the pitch so chosen as
make
the tensile strength of the perforated plate equal to the shearing

strength of the rivet, the
maximum
relative strength of joint
possible with the rivet chosen will be obtained,
and
it
will
be
found that the
joints will
It is
be steam-tight.
Example.
the joint apart
plate
is to
required to design a riveted joint, as shown
in Fig. 33 (h), for a boiler shell in

is
which the force tending

shell.
to pull
6,000 pounds per inch of length of

steel of
The
be of
60,000
to
lbs.
tensile strength, the rivets
have a shearing strength of 49,000 lbs. per square inch in single shear, and 42,000 lbs. per square inch in double shear. The factor of safety is to be 5.
are to be of steel
and are
RIVETED FASTENINGS
151
The
the joint the
allowable stress per inch of length of the shell outside

is
60,000
12,000
the
lbs.
If the joint
were as strong as
plate
unperforated
plate
thickness
of
the
would be
6,000
12,000
V inch.
.80,
The
relative strength of the joint will not
be
less
than
and hence the thickness of the plate must be
X
.80
5"
The diameter
The
rivet
size oj the
ill of the rivets will
1.2 %/ _5_
^
nearly.
8
punched
hole
and
the
diameter of the driven
will be
|f
"
Equation
(6) of this
chapter gives the relafor
tion
between pitch and diameter of
rivet
equal strength
against shearing of the rivets
and tearing d
of the plate.
Th
= =
,
*jrr np.+
4
L_
2mpn
-J
x
tp
Here, n
and
m =
L
4.
Hence,
.'.
lMt\ *>f**> +
4
54
>
<4 x 42 ,ooo) -|
-I
60,000
16
8 inches nearly.
The
relative strength of the joint is

o
15 ^^-
= 88%,
hence the design
is safe.
If the pitch
found as above should be considered too great,

it
either
on account of very high steam pressure or because
is
desired to
make
the structure
stiffer,
a smaller pitch could be

less.
used, but the relative strength
would be
Where no
will
fluid pressure is to
be withstood the above methods
always give satisfactory results for joints in tension.
For
struc-
joints in compression the student is referred to treatises
on
tural work. 52.1.
The Making
of Riveted Joints.
It is
evident that the

first-class
following precautions must

riveted joints.
be observed in making
152
(a)
MACHINE DESIGN
The
plates
driven, to prevent a fin
must be in close contact before the rivet is from forming between them and thus
holes
making a
(b)
tight joint impossible.
The mating
must be "fair"; that
is,
they must be in
perfect alignment to insure full cross section of the rivet at the
junction of the plates.

(c)
(d)
The The
rivet
must completely
fill
the hole.
its
rivet should be carefully driven so that
strength,
or that of the plate, will not be weakened by poor workmanship.

(a)
In hand riveting the plates are drawn up together, before

is
is
the rivet the rivet
driven,
to
by a
bolt placed in a hole near that in
which
be driven.
With comparatively
if
thin plates this
method
well.
is
will
accomplish the result
the holes have been accu-
rately spaced,
and if the plates have been properly rolled and fit For heavy work, where for other reasons machine riveting necessary, the riveter is sometimes provided with a powertill
driven closing device which holds the plates up
the rivet nips
from
cooling.
(b)
drilled.
rivet.
The rivet holes in the plate may be either punched or They are generally made about tV inch larger than the
Generally speaking
it is
cheaper to punch the holes than In structural

however, some
is
to drill
them, and hence in the cheaper kinds of work, and with
thin plates,
punching
is
almost always resorted
to.
work
the holes are generally punched.
There
the
it
are,
serious
objections
to
punching.
When
punch
forced
through a plate the amount of metal which

of a
removes in the form

filled
is is
"plug" This hole.
is
is
not equal to the amount that originally
the
accounted for by the fact that punching
not a
pure shearing action, but that during the process there

of metal
a flow
from under the punch
to the walls of the hole, setting

if
up a
stress in the material, and,
the metal
is at all
hard, serithis
ously impairing the strength of the plate.

action
either
is
It is
found that
confined to a thin ring next to the hole, and that by
reaming out the hole about xV inch all around, or annealing the plate, this weakening effect disappears. The process of
punching
is
apt to
make
inaccurate
work and,
therefore,
fair.
when
the plates are brought together the mating holes are not
The
RIVETED FASTENINGS
old practice of driving a taper drift pin into such holes
1 53
and
If,
drawing them into line by force the injury thus done to the plate being often very
is
now
largely
prohibited,
serious.
however, the holes are punched a
little
small,
and reamed
ing
is
to the proper size, the difficulties
and put together due to punching
are largely overcome.

small, are
Thin plates, in which the effect of punchpunched and used without reaming or annealing.
>2 inch thick
Plates
more than
-
should always be either annealed
or have the holes reamed
after
punching.
always better
if
drilled,
and
all first-class all

if
Heavy plates are boiler work requires
the holes to be drilled in place and
burrs carefully removed.
This
last is
important, as the burr,
allowed to remain,
may
seriously impair the strength of the head.
small countersink,
on the other hand, materially contributes

rivet.
to the strength of the
When
;
plates are annealed the
work should be properly
done
for
if
the plate be overheated structural changes take place
that materially
weaken
too rapid nor the

holes
The heating should not be temperature above a medium cherry red. The
the metal.
made by punching
rivet holes
form, and where they are used as they
machine the
rivet has a
(c)
somewhat tapering in come from the punching should be punched so that they will come
are necessarily
stress.
together as in Fig. 36; for the rivet drives better and the tapering
tendency to relieve the head of part of the tensile

it is
Since
it
necessary to have the hole a
little
larger than
the rivet,
is
clear that the rivet
when
driven must be upset

it
throughout
the hole.
difficult
its
entire length in order that

rivets
may
completely
fill
Large
to
should therefore be machine driven, as

if
it is
upset heavy rivets, especially

fill
hand.
If
a rivet does not completely
centrated bearing, or shearing stress,
by undue conmay be brought on its

of great length,
the hole an
neighbor.
(d)
On
the other hand, care
must be exercised
and
in
machine
riveting that the pressure applied does not create such a flow in
the rivet as unduly to strain the plate,
also that the pres-
sure applied
is
not great enough to crush the plate directly.
Practice allows about 80 tons per square inch of rivet area.
Machine
riveting,
when
well done,
is
superior to
hand work,
154
the plates being held
MACHINE DESIGN
up
firmer,
if
and
also because the

is
impact
cold, is
from hand
care
riveting, especially
the rivet
worked too
liable to result in the
breaking
off of the
head.
In either case
rivet
should be exercised that the point

if
is
formed on the
concentrically;
the dies are not properly set eccentric rivetre-
ing will occur.
In machine riveting the pressure should be

till
tained on the rivet
cold enough to hold the plate firmly,

specifications.
This
is
sometimes recognized in writing

spacing of the rivet
is
If the
correct
and the
is
riveting well
done, the joint will be tight against ordinary pressures.

tight joints
is
Where
This
do not
result the
edge of the plate

Fig. 36,
" calked."
often done, as
shown
in
in
by means
of a sharp-nosed
tool
which tucks the sharp bevelled edge of the plate under-
neath, as
shown
an exaggerated manner.
at B, Fig. 36, is preferable.
There
is liability
of injuring the lower plate in using a sharp-nosed tool as T,
and
the
method shown
The
plates should
be bevelled before riveting, as the method of hand bevelling after

riveting, as often
done in
practice, is almost sure to result in
some
well
injury to the lower plate.

52.2. Strength of
Materials for Riveted Joints.

is
It is
known
shear.
that the strength of rivets
different in single
and double
The
following
may be
used as average values.

p s* 40,000
pet
60,000
72,000
Iron rivets single shear Iron

Steel Steel
"
double
single
"
39,ooo
"
"
"
"
49,000
42,000
80,000
100,000
double
Steel Plates for Boiler
Work
less
are generally specified to have

lbs.
a tensile strength of not
than 55,000
per square inch,
and not more than 65,000 lbs. per square inch, for if the tensile strength is too high and the metal is hard they are liable to crack while being worked. For structural steel construction the student is referred to handbooks on structural work. For iron plates an average value may be taken as 45,000 pounds per
* Master Boiler Makers' Association Rules, page 150. f Proceedings
sign," page 132.
Inst.
Mech. Engineers, 1885, and
Unwin's "Machine De-
RIVETED FASTENINGS
square inch.
It is
55
shown by experiment
this
that the metal between
the rivet holes has a higher apparent tensile strength than that
of
the
unperforated plate;
increase being sometimes

if
as
high as
20%.
It
is
questionable, however,
this
should be
taken account of in designing, especially where the holes are
punched, as the operation of punching

this peculiar increase.
may more
than
offset
52.3. Factor of Safety.
In boiler work the factor of safety
is
taken
at
about
which, of course, brings the working stress well

If the joint is to
below the
pressures,
elastic limit.
be subjected to hydraulic
to
where heavy shocks may have

It
it
be withstood,
this
factor should be increased.

52.4. Practical Rules.
has already been noted that practi-
cal considerations
make
necessary to modify the theoretical
equations for uniform strength.
There are many
sets of practical
rules for designing riveted joints a
number
of
which
will
be
found
in the references given below.
Rules of the Hartford Steam Boiler and Inspection Co.

Rules of the American Bureau of Shipping.
Rules of the Master Steam Boiler Makers' Association.

Rules of the U.
S.
Board
of Supervising Inspectors.
See also Cathcart's "Machine Design."
Proceedings of Inst, of Mech. Engineers, 1885.
Machine Design." Wm. M. Barr's " Boilers and Furnaces." " Steam Boiler Construction," W. S. Hutton.
Unwin's
"
CHAPTER
VII
SCREWS AND SCREW FASTENINGS

53.
Form
of Screws.
Screws, as used in machines,
may be
divided into two classes.

(a) (b)
Screw fastenings.
Screws for transmitting power.
of the thread
The form
Thus,
depends upon the service required.
for screw fastenings, the full
as
shown
in Fig. 37 (a), or
modified forms of
threads, as
shown
in Fig. 37 (b)
easily cut
and
(c),
are
most used because they are strong and

dies.
by machine
is
They
are inefficient for transmitting power, but this

it
desirable quality in fastenings, as
reduces the
liability of
un-
screwing. For transmitting power the square thread, Fig. 37 (d), is most used, since its efficiency is higher than that of any other
form.
It
cannot be cut with a

for
die,
however, and
of thread.
it is
difficult to
compensate
the half
wear with
this
form
For these reasons
thread, Fig. 37
is
(e), is
often used for transmitting
power
when wear
form
an important
factor.
is
standard thread of this form
In Fig. 37 (e) the Acme shown. The efficiency of this

it
of thread is a little less than the square thread but
can be
of a
cut with a die and wear can be compensated for by

longitudinally split nut;
this
means
it
compensation making
very de(f)
sirable for such service as lead screws of lathes, etc.

illustrates the buttress thread,
Fig. 37
which
is
often used to exert pressure

is
in
one direction only.

with this
The
pressure face
perpendicular to
the axis of the screw,

of 45
axis.
and the back face usually makes an angle This screw has, therefore, the efficiency and the strength
all
of the square thread
of the
thread.
The
underlying principles of
screws are the same, and before dis-
cussing the various forms and classes in detail the fundamental
equations relating to their action will be developed.

156

54. Friction
157
In
and Efficiency of Square Threaded Screws.
Fig. 38 (b), let
represent a nut
moving on a square thread,

at the
under the action of a tangential force P, acting

radius of the thread.
couple, so that there
mean
Let
is
this force
be applied by means of a
no
lateral pressure against the screw.

is
Let
W
or
represent the load under which the nut

it
moved, and
(Fig.
consider that
is
can move only in an axial direction, hence there
friction
between
N and W.
This
frictional force
38 b)
may
It
may
not act at the same radius as P, and the work due
to this frictional force will
vary with the radius at which
it
acts.
can be considered as forming a resisting
moment opposing
the
Full (a)
Sellers or U.S. Standard
(b)
Whitworth Standard (c) ^14.5
Square Thread
(d)
-b-W
Acme Standard Thread

(e)
Buttress Thread
(f)
Fig. 37.
turning
quired turning
moment due to P, moment of P,
hence,
is to
when
computing the
reIt
be added to that value.
can therefore, for simplicity, be omitted temporarily from the

discussion.
Let
= coefficient of friction between thread and nut. = coefficient of friction between load and collar. n d = nominal or external diameter of screw. r nominal or external radius of screw. d = diameter of screw at bottom of thread,
ft
l
l
r,
radius of screw at bottom of thread.
i58
r
MACHINE DESIGN
frictional radius of collar.
dm
r s
= mean = mean
diam. of thread
+
2
dj
radius of thread
rx
= =
pitch, or angular
advance of thread per turn.
angle
made by
thread with a plane perpendicular to
the axis of screw.
If
now
is
the thread be developed as in Fig. 38 (a),

is
it is
seen,
since the thread
a true
helix, that the action of the
thread
and nut
identical with that of a

,
body
N sliding up an inclined
s
plane of length nd m and vertical height

carrying a load
ting
tion
equal to the pitch, and
W which
is
free to
move
vertically only.
Omitthe fric-
F F
temporarily, the forces acting are the load
W,
normal reaction R.
,
between the thread and nut, the driving force P, and the It is required to determine for any angle
the value of
incline.
all
required to slide the body
N
2
(turn the nut)

i*R.
up the
resolving
The
frictional
resistance
F =
Hence,
forces parallel to ac
159
.
P cos
.'.R
Resolving
all
W sin a P cos a W sin =

a
ft
R =
(j)
a.
....
(2)
forces perpendicular to ac
R
Substituting in
(3)
P sin a = W cos
the value of
(3)
obtained in
(2)
p_
rsinoj^co^-l w l-COS an
sin a-J
_
,
Since (Fig. 38 a) cos
and
sin
equation
(4)
may be
written
p=
The
friction,
w[^lA,~]
L-7T
ft
s-l
( J/ 5)
F l ,=p 1 W, and if r be the radius at which F moment of F around the axis of the screw = p Wr and when this resistance is considered the total moment of P
c
x
acts, the
around the axis
is
Prm =
If the load is
Wrm P +
** dm~]
+ H Wr.
moment
...
F
x
(6)
being lowered, the directions of
and
are
reversed,
and
in this case the turning
that
must be
applied
is
In equation 6
the
first
term of the right-hand side of the

of
equation
the
-
is
the
moment
is
of the resistance at the thread, while
second term
the
moment
tan
the collar
friction.
If
dm
s,
that
is
if
/,
the
if
moment
there
is
of the re-
sistance
at
the thread will be zero,

if
it
and
no
collar
friction, or
is
very small,
as in the case of ball-bearing
thrusts, this will give a condition of equilibrium, the friction of
the thread alone just sustaining the load,
and
will
be equal to
l6o
zero.
MACHINE DESIGN
,
If the pitch s is made greater than p x dm the moment of the resistance at the thread becomes negative; and if increased
till its
numerical value
is
equal to the
moment
of collar friction,
the entire right-hand side of the equation will be equal to zero
and the load will just be sustained by the friction of the thread and collar combined. If the pitch is still further increased, the entire right-hand side of the equation becomes negative, and the moment P rm must be applied in the direction of raising
x
the load,
or.
the screw will " overhaul," the nut exerting a turnin the
ing
moment
downward
direction.
To
find the limiting value of
where the nut
will not over-
haul, equate the right-hand side of the equation

5
(6') to
zero and
solve for
whence
tan
m
a.
=
rm
is
(7)
If rc
tion,
=
o,
=
o, in
which case there
no moment due
to collar fric-
tan
P as before.
rc
Hf.-
and
ft
p1
tan a
=
I
2 p
2
' ' ' *
(8)
and
n a
if
l
be taken
as.
.1
(see
Art.
65)
tan
.2
whence
= tt ii To find
the efficiency of the screw, consider that the load has

to the pitch, that /*,=/*,
been raised a distance equal

then
e
and
rc
= rm
efficiency
=
,
since s
= * dm
e
tan
= W s or p rm *dm P and inserting the value of P from equation

work done
energy expended
tr =
W
s
2 n
(5)
=
dm
W -d
fs
tan a
TT + pWnd,
.
+
(1
arc
dm n
tan
p tan
+
2
fi
tan a
nearly J
(o)
If the collar friction is zero, or
very small, as in the case of
ball-bearing thrust collars

e
l6l
.
.
tan
a.
(i
ft
tan
it
*)
(10)
tan a 4If,
in equation (9),
ft
be taken as
.1,
and tan
as .2 as before,
e will
equal 50 per cent (nearly), and a brief reflection will show
that in
no case can the
efficiency of a self-sustaining
hoisting
screw exceed 50 per cent.

=
Suppose the load

repose.
relieve
If
(Fig. 38) to
be just
is
sustained by the frictional resistance of lowering, that

;jl
tan a
is
or
angle
of
to
now
force
apthe
plied,
just
sufficient
this
frictional
resistance,
load will be sustained
and the reaction R. If the frictional resistance of raising were zero, the slightest addition to P would move the body up to plane. But the fric-
by the
force
tional resistance of raising
sequently, before the
is equal to that of lowering, and conbody can be started up the plane, a force 2 P must be applied; which is twice the force required to balance the frictional resistance, and the efficiency would then
be 50 per cent.
similar reasoning will apply to other hoisting
devices which are barely self-sustaining on account of friction,
namely, that the force which must be applied to
start the
load
is
equal to the friction due to lowering plus the friction due to
Hence the maximum mechanisms is 50 per cent.

raising.
efficiency for
such self-sustaining
In designing screws for power transmission
it
is
desirable to
know
the pitch angle that will give

If
maximum
efficiency for the

(9)
conditions taken.
the
first
it is
differential
of equation
be
taken and equated to zero,
found that the
maximum
efficiency
when
collar friction is considered will occur
when
fi
tan
If
fect,
ft
= =
v
51
.
fi
.1
=
as in transmission screws
1.23
where lubrication
is
imper-
tan a
and
oil
In the case of
bath lubrication, as in
tan a for
worm
gearing,
ft
may
be as low as
a
.05
/
.
when
maximum
(10) for
1?
/'
efficiency
1.3 18 or
= 5 2-49
In a similar
way from equation

tan
maximum
efficiency
a=Vi +
62
MACHINE DESIGN
for n
.05, a
Whence
and
for
/i
= .1, a = 42 = 43 34'
on the
efficiency of the screw is
The
and
is
effect of the pitch angle
of great importance in designing screws for
power transmission,
Threaded Screws.
at
more
fully discussed in Arts.
64 and
65.
55. Friction
and Efficiency
of Triangular
With triangular threaded screws the normal pressure

threads
is
the
greater than with square threads;
hence the friction

In Fig. 38
(c)
at the threads is greater, other things
being equal.
the
normal pressure
for a square thread is indicated

is
by
while
<p,
the normal pressure for a triangular thread
R'
=R
sec
in
which
<p
half the angle
between the adjacent faces of a thread.
R" represents
screw, and
its
the radial crushing action on the thread of the
equal and opposite reaction tends to burst the nut.

thread, as in the Sellers system, or the
sec 30
With 6o angular
common
thread,
R =R
f
= 1.15^.
it is
The
friction increases directly
about 15 per cent greater in the 6o angular thread than in the square thread.
as the
If
normal pressure; or
inequations
(1)
and (s),R
sec
<p
be substituted
fori?,
then by
collar
a similar method of reasoning as in square threads,

friction is neglected
when
P =
d -^n w p+/' L d sec

7T
(p
....
moments around
(II)
If the collar friction is considered, the
the
axis
when
raising the load
may be
a p
7tdm
written
'
Prm =
If
+ *d Wr\-^ U^-^j d
\~s 's
sec
<p~~\ <p
{is sec <p-i <p-
-\+HWr.
l
(12)
Pr = Wrm m
The
+ i.i$ p* d R+i-iSM'O
's
B
Ltt
i.itfisJ 5^
efficiency of a triangular threaded screw, following the
same reasoning and /*! = /*, is
as
for
square threaded screws, taking rm = rc
W *dm tan
*dP
2*rP
163
W * dm tan a
'
W*dm Lr
rs
or e
^m
/^sec^J
nearly.
/*
findm sec9~]
/t
w W
7T rf
tan a (i tan a
ft
tan sec 9)
^ sec
(i3 fl )
For the thread on a one-inch bolt in the Sellers system tan a = .04, and takings = .1, e = 11%. The efficiency of the threads on standard bolts is hence seen to be very low and this, as
G3 CD
(d)
WQ
g(g)
|p(e)
EHj(f)
K(h)
Fig. 39.
has been pointed out,

to prevent
is
an advantage
in fastenings as
it
tends
them from unscrewing.

Screw fastenings are used
for hold-
56.
Screw Fastenings.
ing two machine parts together in permanent position, or for
adjusting one part relatively to another.

of screw fastenings but all
1.
There
is
a great variety
may be
3.
roughly classified as follows:
Through-Bolts;
2.
Studs;
Tap-Bolts
or
and
as
Cap-Screws;
4.
Machine Screws;
Through-Bolts.
5.
Set Screws.
through-bolt,
"bolt"
it
is
com-
monly
called, Fig. 39 (a), has a solid

It
is
head on one end and a nul on
the other.
the best form of screw fastening
and should
164
MACHINE DESIGN
always be used when the hole can be drilled completely through

the two pieces to be held together.
Studs.
Sometimes
it is
not possible or desirable to
drill
a hole
entirely through both pieces
which are
to
is
be held together, and

often used.
in such cases a stud-bolt, or "stud," Fig. 39 (b),

is
stud,
a circular bar having a thread cut on each end.
A
The
hole
is
tapped in the part that cannot be drilled clear through,

of the stud is screwed firmly into the hole.
and one end
part that can be drilled
through
is
slipped over the stud, and
a nut on the outer end clamps the two parts firmly together.
Where a
fastening.
through-bolt cannot be used the stud

It
is
the next best
should be a tight
fit
in the
tapped hole, and when

if
once screwed in should not be taken out, especially
the hole
is
tapped into cast
iron, as repeated
removal wears out the threads.
The
length of the tapped hole should be at least one and one half
times the diameter of the stud, in order to secure ample frictional

resistance against turning
when
the nut
is
unscrewed.
.(c),
Tap-bolts and Cap-Screws.
Tap-bolts, Fig. 39
and cap-
screws, Fig. 39 (d), have a solid head
on one end and a thread on
the other.
They
are used under exactly the
same circumstances
as the stud but are not as good a fastening, as they necessarily must be unscrewed from the tapped hole whenever they are removed. Where they have to be frequently unscrewed, and especially
if
the hole
is
tapped into cast iron, they should be avoided.

is
The
next
only difference between tap-bolts and cap-screws
in the size
head (see and the cap-screw for the same size of bolt having a smaller head slightly rounded on top. Tap-bolts are much used in such work as securing patches on boilers, where a large head is desirable. Cap-screws are a little more ornamental and are much used in cheaper grades of machinery. They are a standard article in the market and hence can be bought very cheaply. Machine Screws. Under the term "machine screws" are
of the head, the tap-bolt having a standard
,
and form
article)
included
slotted
many forms
of
of small screws usually provided with a
head so that they
most usual forms
may be set up with a screw-driver. The machine screws are shown in Fig. (39 e, f, g

and
h).
65
At
c is
shown an oval
fillister
head; at
/a
flat fillister
head; at g a
flat
countersunk head; and at h a round head.
Machine screws are designated, for convenience, by numbers, Thus the the larger numbers indicating the larger diameters. smallest size, as given in Brown and Sharp's catalogue, is number 000 the diameter of which is .03152. The difference in diameter between consecutive numbers is .01316. The diameter of a
number o screw
is
.0578, so that the diameter of
any number
by the formula d = .0131 + .0578; in inches, and the serial number of diameter where d is the is not to be confused with the number The number the screw. Machine screws larger than number 16, of threads per inch n.
larger than this is given
which
is
about
%"
in diameter, are not
much used
in
machine
work, another standard, to be discussed later, being used for

sizes
above that diameter.

far,
Manufacturers have, so
screw.
been unable
to agree
upon stand-
ard numbers of threads per inch, for a given diameter of machine
machine screw may have 20 or 24 threads per inch, so that these screws are usually specified by naming the size number first, followed by the number of threads per inch. Thus, an 18-20 machine screw means size 18 and 20 threads
Thus a number
12
per inch.
Because of the great confusion
now
existing regarding
this point, the
American Society of Mechanical Engineers appointed a committee to establish, if possible, a system of standards This committee has reported and their for machine screws.
of the Transactions. of screw-fastening fre-
recommendations can be found in Vol. 28

Set Screws.
Set screws are a
form
quently used to prevent relative rotation of two machine parts.
Thus
shaft b
in
Fig.
40 the hub a
set
is
prevented from revolving on the

of the set screw is square
(a),
by the
screw
c.
The head
while the point

,
may
be cup-shaped as in Fig. 40
round as
is
in
Fig. 40 (b) or conical as at c inFig. 40.
When
is
the set screw
made
it
in the
form shown
at Fig.
40
(a),
the point
hardened
to enable
If
to cut into the shaft, thus increasing its
holding power.
the
the
screw
is
made
of tool steel the hardening

if
may be done by
The
ordinary process of tempering;

result
made
of
wrought iron the same

objection to
may
be obtained by case hardening.
i66
the cup-shaped
MACHINE DESIGN
end
is
that
it
makes a burr on
is
the shaft which
sometimes greatly interferes with the removal of the hub.

obviate this a small conical depression
shaft with the
To
sometimes made
at Fig.
in the
end
of a drill
and the form shown
40 used.
Set screws are not reliable for heavy
work and should be used

U.
S.
only
when
the load
is light.
Standard Screw Threads.

fastenings larger than
Sellers or
System.
Screw
to
inch diameter are

order to
made according
some standard system
in
secure interchangeability.
The
first
system of
this
kind was that introduced into England
Fig. 40
is
by Sir Joseph Whitworth. The form of the Whitworth thread shown in Fig. 37 (c). The thread angle is 55 and the top and bottom of thread are rounded off as shown.
The
this
recognized standard screw thread in the United States

S.,
is
the Sellers, U.
or Franklin Institute thread.

in Fig.
is
The form
is
of
thread
is
shown
39
(b).
The
is
thread angle
6o;
the top of the thread

filled
cut off and the bottom of the thread
in
as shown.
This standard
full
not used exclusively in
this country,
however, but a
thread, as
is
shown
in Fig. 37 (a),
without the flattened tops and bottoms
also in
common
use.
The
angle of such
thread
is
generally 6o in machine bolts
and the number

ticular.
of threads per inch usually corresponds to those
of the Sellers system, but there are
many
is
variations in this parstrictly
Where
the Sellers standard
not
adhered
to

it
167 one manufacthan
is
advisable, therefore, to
buy machine
bolts of
turer only or so specify as to insure interchangeability.
The
Sellers screws
full
have much greater
tensile strength
screws with
threads of equal angles and pitch, because the

is
thread of the former

flattening at the tops
only three-fourths as deep owing to the
and bottoms.
TABLE x
SELLERS, U.
S.,
OR FRANKLIN INSTITUTE STANDARD BOLTS
Bolts and Threads.
Hex Nuts and Heads.

.
Sq.
Nuts and Heads.
>3
2
u
t/5
'0
"o
03
53
<"
5
.2-
<S2 ^
(LI
H
No.
Is
C
In ches.
W c
4>
_;
rt
ft
P
Inches.
0 <
Inches.
P2 0hC/2
3cS
II
H
IS
SI'S
Inches. [Inches.
Inches. Inches. Inches. Inches.
20 18
16 14
13
1
185
240
2Q 4
A
4
\
344 400
4 54
507
i
\
10
9 8
7 7
1
1
620
731 83 7
.027 045 .068 093 126 162 202 .302

. .
tl
25
ft ft ft
9
A
i
5
A
TU ft
TS
3
8
ft
TIT
& II
I
iA
i^t ift i4 ift 23V aft
ft
it
I iT
lA
I*
A
5
8 8
1%
4
T
1
8
ft ft itV
13
ft itV
it
420
ii ii if
A ^
i\
6 6 Si
5 5
1
1 1 1
*l
2
A
A
l
4i A\ 4 4
:
2
2
2
940 065 160 284 389 491 610 712 962 176 426 629
550 692 .890

.
*A t 5
ift
2
2T
ift It
2
ItV
It
i57
1.293 1.510 1.741 2.050 2.300 3-030 3-7I9 4. 620 5.428
*A
\
2.','
It It
:!
1 1
III
2
2^
2ft
3
3
lA
1
2#
2^
A
!,
2^
*A
2f 2ft 3* 34 3* 4i
4i
2f 2ft
14
1 1
S
s
:
ift 9 IT .
ift
3ft 3ft
4
*A
2{ aft 3$
3
7
3A
3tt 3t 4tV 4*
4
it
2
JJt
2|
*A
2V. 2ft
4ii 4ft
5ft
6
2*
2\ 3
.29
:;
3*
L'
5J
4t 4
6ft
The
area of a 1"
full
6o thread
is
.482 square inches, while

is
the area at the bottom of a Sellers thread
.55
square inches,
or 14 per cent greater.
The Whitworth system

the
of threads differs
from the
Sellers
in
shape of the thread, as noted above, and the number of

is
threads per inch
also different for
some diameters.
Thus
the
l68
Sellers
MACHINE DESIGN
system gives 13 threads per inch for
gives
12.
X"
bolts while the
Whitworth
The Whitworth system

size of bolt
flat
gives
somewhat
is
stronger screws as the diameter at the root of the thread

greater for the
root
is
same
is,
and the rounded shape

easier to
at the
stronger than the
root of the Sellers thread.
The
Sellers
thread
however,
much
produce than the
Whitworth.
by the SelAbove this size the standard is not adhered to rigidly, as the size and pitch of the screw becomes rather large for convenience. Thus a 6" bolt in the Sellers system will have 2}^ threads per inch. It is common,
gives the proportion of screws as fixed
lers
Table
standard for bolts up to
2^"
therefore, to
rare,
make
these larger sizes, which are comparatively
with 4 threads to the inch.
In Germany, France, and other European countries other sys-
tems are in use, some of which are based on metric

57. Pipe Threads.
units.
is
The
Briggs
system of
pipe threads
the established standard in the United States.
The numbers
of
threads per inch for the various sizes of pipe are given below
yi" pipe, 27 threads per inch.
%" and %" and

2^2
/;
yi" pipe, 18 threads per inch.
pipe, 14 threads per inch.
1" to 2" pipe,
n X threads per inch.

details as to Briggs
and
over, 8 threads per inch.
For form
of threads
and other
system,
see Trans. A. S.
M.
E., Vol. VIII,
page
29.
58. Straining Action in Bolts
due to External Load.
The
load applied to the bolts
is
generally one which tends to separate
the connected members, in the direction of the axis of the bolt, and
this action is resisted
by a
tensile stress in the bolts
but bolts are
sometimes used
pieces,
to prevent the relative translation of

is
two or more
when
a shearing stress
is
produced in the
bolts.
When
the action of the load
oblique to the axis, the stress in the bolt
may be combined
If
tension and shearing.
any screw
is
tightened up under load there
is
an
initial
direct stress (tension or compression)
and usually a
torsional stress

due
to friction
69
between the threads of the screw and the nut.

screwed up
hard, as in
to
With
bolts or studs
making a steam-
tight joint, the initial tension

in excess of that
due
screwing up
may be much
will
due
to the
working load.
This
be treated
more
bolt
fully later.
If the
load applied to the bolt produces a shearing action, the

fit
shank should accurately
the holes in the connected pieces,

if
at least for the portions
near the joint; and
P is the load per bolt,

stress,
d the diameter of the bolt (shank), and p s the shearing
In a bolt subjected to a load which produces tension, the mini-
mum
cross section sustains the greatest stress.
This smallest
cross section, in
common
if
bolts, is
through the bottoms of the

to
threads.
41,
Thus
only
a load
be applied
an eye-bolt, as in Fig.
will
the
stress that will
be induced in the bolt
be that
to the
due to the external load P.

the threads,
If
p be the
tensile stress
due
load P, and d x the diameter of the bolt at the bottom or root of
P = 1
4
t
d? P-'.p
= 4r* n a
i
(14)
Values of d are given in Table X, page 167, for the various

sizes of Sellers screws.
For a given diameter and pitch of screw
the area at the bottom of threads

full
would be considerably
less
with
threads.
59. Initial
Tension in Bolts due to Screwing up.
If the
bolt is
used simply to hold two machine parts together, as in

(a),
Fig.
39
and there
to
is
no external load tending
to separate the
parts, the stress in the bolt will
be the resultant of the tensile
due
screwing up the nut, and the torsional stress due
to the frictional resistance at the thread.
In the Sellers system the pitch angle of the thread
(a)
varies
varies
from about 3 from .054

r
in a
X"
screw, to i-5o' in a 3" screw;
or tan a
to .032 in this
same range.
rm
,
If,
therefore, in
.1,
equation (12),
.15,
it
be taken equal to \
p be taken at
l
and
/,
appears that
varies from .35
W with a /
2
"
screw to .32
170
MACHINE DESIGN
a 3" screw.
this, so it
W with
more than
The coefficients may be assumed
of friction will vary
much
that for the ordinary range
of screw fastenings
W approximately P or the tension in the screw W =

P =
.33
33
The
by the
turning moment,
Pr m due
,
to the
wrench
pull, is resisted
frictional
moment
body
of the nut or collar

frictional
and the
frictional
is
moment
at the thread.
This
moment
at the thread
transmitted to the
of the bolt, so that the bolt itself is sub-
jected to a twisting
moment
collar.
is
equal to Pr m minus the frictional
moment
at the
nut or
The
resultant stress, therefore,
under these circumstances,

direct stress,
stress as
that due to
combined twisting and

is
and
it
can be shown
(see Art. 67) that the resultant

(1),
determined by equation
page 49,
from 15
to 20
per cent greater than the direct stress alone.
Refined calculations regarding the resultant

to screwing
stress in bolts
due
up
are, in general, useless
and misleading, especially

Since a me-
in the case of fastenings less than J/&" in diameter.
chanic using a wrench of ordinary proportions can easily rupture
any of these small fastenings,
it
follows that the actual stress in
such bolts depends entirely on the judgment of the mechanic.
series of
experiments was
made
in the Sibley College
Labosizes
ratory, a few years ago, to determine directly the probable load
produced in standard bolts when making a

of bolts used
tight joint.
The
were
%"
}i", 1"
and \%".
One
set of
experiments
was made with rough nuts and washers, and another set with the nuts and their seats on the washers faced off. A bolt was placed in a testing machine, so that the axial force upon it could be weighed after it was screwed up. Each of twelve experienced mechanics was asked to select his own wrench and then to screw up the nut as if making a steam-tight joint, and the resulting load on the bolt was weighed. Each man repeated the test three times for every size of bolt, and each had a helper on the 1" and \%" The sizes of wrenches used were 10" or 12" on the 2 " bolts sizes, up to 18" and 22" on the \%" bolts. The results were rather dis-
171
cordant, as should be expected; the loads in the different tests
were rather more uniform, as well as higher, with the faced nuts
and washers.
The
general results indicate:
(a)
that the initial
load due to screwing
up
for a tight joint varies

is,
about as the
diameter of the bolt;
that
a mehanic will graduate the pull

(b)
on the wrench
in
about that
ratio,
lbs.
That
the load produced
may be
in
estimated at 16,000
per inch of diameter of bolt, or

(15)
W
which IT
is is
16,000 d
the initial load in
pounds due
to screwing up,
and
This
it
the nominal (outside) diameter of the screw thread.
value of
is
rather above the average for the tests;
but
is
considerably below the

cess of the load
tight joint.
If the initial
maximum, and it is probably not in exwhich may reasonably be expected in making a
load due to screwing up be divided by the cross-
sectional area of the bolt at the

intensity of the tensile stress
bottom of the threads, the

is
initial
obtained.
The above
experi-
ments indicate that

inversely as the
this intensity of stress varies, approximately,

(d)
nominal diameter
of the bolt;
and that
it
may
frequently equal or exceed
30,000 -
lbs.
per sq. in
is,
(16)
In addition to this tensile stress there considerable twisting action on the bolt.
as before stated, a
(16)
Equation
would
bolt;
give a stress of 60,000 lbs. per square inch
on a K-inch
It also
and
size
this result is substantiated
by the
fact that steel bolts of this
were broken
in the course of the experiments.
agrees
with
as
common
experience which forbids the use of screws as small
K-inch
for cases requiring the nuts to
be screwed up hard.
In these experiments, the average effective lever
arm
of the
from 15 times the diameter, or 30 times the radius, of the screw; hence, if it be assumed as in the previous
far
wrench was not
paragraph that the turning force acting

is
at the radius of the

is,
rf
screw
P=.$$W
the force applied at the
wrench
16,000
in
pounds, about
p =
1
=
3
"
33lF
30
-33
172
MACHINE DESIGN
discussion indicates that the factor of safety should
The above
be increased as the size of the screw decreases, and of course this factor should be varied with the conditions of the case, as in
some applications the nuts

hard than in others.
are
much more
apt to be screwed
up
A set of experiments was made by Mr.

A.
S.
James McBride (Trans.
M.
safety,
page 781), which show that the factor of as bolts are frequently used, is very low, even with a very
E., Vol. XII,
moderate external load.

cates that the stress
One
to
due
McBride indiscrewing up a 3^-inch bolt was

case cited by Mr.
nearly one-half the ultimate strength, or probably very near the

elastic limit.
His direct determinations of the efficiency of a
standard 2-inch screw bolt shows an average of only 10.19 per

It is probably this low efficiency which saves many screws cent. from being broken, as the frictional loss reduces the tension produced in the bolt by screwing up. The excessive friction makes
the screw bolt a useful fastening, as
it
reduces the tendency to
"overhaul" or unscrew.
60. Resultant Stress
on Bolts due
It
to
Combined
in
Initial
Tension
bolts
and External Load.
was shown,
Art.
59,
that
may be subjected to a high tensile stress by screwing up the nuts. The question often arises as to the effect of the combined
and the external, or useful, load. by some that the resultant load on the bolt is simply Others contend the sum of the initial and the external loads. that the application of the external load does not change the
action of this initial tension
It is stated
stress in the bolt, unless this external load exceeds the initial load
due
to screwing up;
that
is,
that the resultant load
is
equal to
the initial load alone, or to the external load alone, whichever is the greater.
Neither of these views

tained in practice.
is
entirely correct for conditions at-
They
lies
represent the extreme limiting cases
and every actual case

If the bolt itself
between them.
could be absolutely rigid while the members

it
forced together in screwing
up yielded under
pressure, the total
load on the bolt would be equal to the

the external load.
If,
sum
of the initial load
and
however, the members pressed together

were absolutely
rigid,
73
only the bolt yielding, the total (resultant)

initial
load on the bolt would be the load alone, whichever

is
load alone, or the external
the greater.
is
The
first
of the
in
above conditions
approached by the arrange-
ment shown
of 2,000
is
Fig. 42.
Screwing up the nut compresses the
spring interposed between
A and
is 2
B.
Assume
1
that an axial force
pounds
till
will
compress
this spring
inch;
then
its
if
the nut
screwed up
the spring
inches shorter than
free length,
the load on the bolt, due to screwing up,
must equal the reaction
Fig. 47.
Fig. 46.
Fig. 48.
Fig. 43.
of the spring, or 4,000
lbs.
Assume,
also, that the
extension of
the bolt under this screwing-up action, or under the initial load
of 4,000 lbs.,
is
.02 inch.
Now,
if
an external axial load of say
2,000 lbs. be applied to the eye at the bottom of B, this added load
would tend further

inch;
to increase the length of the bolt
by about
.01
but this further extension of the bolt would reduce the
compression on the spring by a corresponding amount and thus

slightly
diminish the spring reaction.
With such great

of the connected
differ-
ence between the rigidity of the bolt and

bers, the load
mem-
on the bolt becomes practically the sum of the
174
initial
MACHINE DESIGN
and the external loads, but the resultant load is necessarily somewhat less than this sum in any possible case. The arrangement shown in Fig. 43 is one which approaches
the other limiting case mentioned above.
Suppose the bolt

lbs. in
to
be
a spring which
is
subjected to an axial load of 4,000
screw-
ing the nut up two inches, and that the corresponding yielding of
the
member B
is .02
inch.
is
The
4,000
initial
lbs.,
is
load on the bolt (which
is
the spring in this case) the contact surfaces of

axial load
and the pressure between

it.
and
equal to
If
an external
be
now
applied to the eye in B, the pressure between

is
the contact surfaces

external load.
reduced by an amount nearly equal to
this
But, unless the external load exceeds the
initial
enough to separate these contact and entirely remove the pressure between them, because the load on the bolt (spring) cannot change without changing the length of the bolt, and with the above data the bolt would have to stretch an additional .02 inch (equal to the initial yielding of the connected members) before the contact surfaces would be entirely
load, the bolt will not elongate
surfaces
relieved of pressure.
It therefore
appears that the addition of an
external load in this case does not materially affect the resultant
tension on the bolt as long as this external load does not exceed
the initial load.
If the external
lbs.),
load
is
greater than the initial
load (say 6,000

inches)
;
the elongation of the bolt increases (to 3 the resultant load on the bolt will be simply the external
is sufficient
load alone, because the latter
entirely to relieve the
pressure produced between the contact surfaces in screwing up.
In
all
ordinary practical cases the difference in rigidity between

is much less than in the exThe resultant load on a bolt
the bolt
and the connected members
treme conditions considered above.
may
be anything between the
sum
of the initial
and
the external
loads as a
maximum, and
the greater of these two loads alone as a

limit
minimum. This resultant load approaches the maximum when the bolts are rigid relative to the connected members
Fig. 44
;
as in
and
this resultant
approaches the
minimum
limit
when
the bolts are relatively yielding, as in Fig. 45. case the designer can
tell
In any particular
which
limit is the
more nearly ap-
proached, and he should be governed accordingly.

The Locomotive
(Nov.,
1897) contains an
excellent
is
75
article
on the resultant load on
bolts,
and a
relation
derived from
which the following method of treatment has been developed:
The
application of this
method depends simply upon the
ratio of
It
the yield of the connected

will usually this ratio
members
to the yield of the bolts.
not be
difficult to assign
a sufficiently close value to
even when the actual magnitudes of yielding are un-
known;
in fact, only a
rough approximation
this ratio
to the value of this
ratio is necessary.
call the initial
Let
be called y and
y
let
x;
load on the bolt due to screwing up
',
the exter-
nal (useful) load
and the
total (resultant) load
W.
Then
it
can be shown that
W
If the yield ratio (y) is
+ x
2.
known, the value

y.
of
is at
once found by
the
above relation of x and

is
If the yield of the
connected
mem-
bers
is
between
and
times that of the bolt, the resultant load
equal to the
initial
load added to from 0.5 to 0.8, the external
load.
If a tight joint is
made
with short rigid bolts or studs, con-
necting flanges which are separated by an elastic packing, or with

a metal contact at
some distance from the center

is
line of the bolts,
as indicated in Fig. 44, the applied load

ation since the value of y
the external load
is
an important considerIn some other cases
relatively great.
may be
a minor consideration as affecting the
strength of the bolt.
When
the conditions are such that the nut

is
is
not apt to be
screwed up hard, that
when
the initial load
may be
safely
neglected, design for the external load alone.
The
following suggestions
may
serve as a guide in practical

initial
problems involving the resultant load on bolts when the

load due to screwing up
(a)
is
apt to be considerable.
If a bolt is
manifestly very
much more
the greater.
yielding than the

initial
connected members, design the bolt simply for the

for the external load,
(b)
load or
whichever
is
If
the probable yield of the bolt
that
of
the
connected
is from one-half to once members, consider the resultant load
176
MACHINE DESIGN
from one-fourth
to one-half the external
as the initial load plus
load.
(c)
If the yield of the
connected members
is
probably four or
initial
five
times that of the bolts, take the resultant load as the
load plus about three-fourths the external load.

(d)
In case of extreme relative yielding of the connected
memof the
bers, the resultant load

initial
may be assumed
at nearly the
sum
and external
it is
loads.
61. Allowable Stress in
Screw Fastenings.
From
the fore-
going
seen that small screw fastenings are very liable to be
heavily overstrained by the initial load due to screwing

nut.
up
the
While the body of the bolt

is
is
well designed to resist heavy
loads a source of weakness
found in the threaded portion.
The reduced
stress,
area, due to cutting the thread, localizes the greatest and cracks are very liable to start from the roots of the threads, especially where the thread is of the full V form.
For these reasons the ordinary apparent

in most
ings.
fibre stresses allowed
machine members cannot be permitted
in screw fastenis
For ordinary purposes, where overstraining

allowed, for steel.
not likely
to occur, or for large bolts, 8,000 to 10,000 lbs. per square inch
For such work as steam and hydraulic joints, where the initial stress may be large, from 6,000 to 8,000 lbs. per square inch should be allowed, depending on the conditions
may be
and quality
1
of material employed,
and
if
shocks are liable
to occur, stresses as low as 3,000 to 4,000 are often preferable.
Example
through
The
cylinder of the steam engine

is
is 1 2
inches in
diameter, and the cylinder head

bolts.
held in place by 10 steel

pressure
is
The maximum steam
100 pounds per
square inch.
are
If the contact surfaces of the

is
head and cylinder
ground together so that no packing

of the bolts so that the
necessary,
what must
the bolt
lbs.
be the diameter
per square inch
maximum stress in
necessary to insure a steam-tight joint will not exceed 7,000

?
it is
In
this case
evident that the bolts are
much more
yielding
than the parts which they hold together and the conditions are
those of case a in the previous paragraph.
the initial load on the bolt
It is also clear that
must be greater than the external
77
load due to the steam pressure in order to insure a steam-tight

joint.
fair
If this initial
load be taken at twice the external load a

is
margin of safety
secured.
If
W\
be the
initial
load and
the external load per bolt then
TFo
"
=
4
2
-12 2
10
100
1,130 lbs.
whence
W
if
=
x
2,260 lbs.
Whence
d be the diameter at the root of the thread
X
4
7,000
2,260.
Therefore d
= .64
inch (at root of thread)

It is to
which corresponds
closely to a $4" screw.
be noted that
is
while a total stress of 7,000 lbs. per square inch of section

sufficient to insure a tight joint,
much
greater stress
nut,
if
may be
is
in-
duced by the workman

or inexperienced.
in screwing
up the
he
careless
Example
If in the
above example
of the studs
steel studs are
used and
rubber packing y" thick be placed between the contact surfaces,
what must be the diameter
Here the parts held together are more elastic than the studs and the conditions may be taken as corresponding to those of
case
c.
As
before, the initial load
W\
may
be taken at twice the
external load.
Then
_.
Tk
9
"
and
X 12' X 100 = 1,130 4 X 10 W = 2 W = 2,260.

l
From
(c)
paragraph 60 the
total load
w
=
\\
= w, +
%w
=
2,260
+ (X X
=
1,130)
3,107 lbs.
- d{
hence
4
X
and d
x
7,000
3,107
.76 inch (at root of thread)

to a
J/&
which corresponds closely
inch screw.
178
MACHINE DESIGN
stress
The maximum
which the workman may, perhaps,

is
in-
duce in the stud by screwing up the nut
p
which
= =
30,000
-,
30,000
y%
34,000
lbs.
approximately,
will
be increased a
little
it
by the external
load.
This
is
close to the elastic limit;
but
may
be noted that even should

not so
the elastic limit be slightly exceeded the efficiency of the fast-
ening
is
not impaired, since here permanency of form
is
essential as in
machine parts which transmit motion.

bolts,
62. Resilience of Bolts with Impulsive Load.
In bridge work and other cases requiring long
it is
very
common
may
or
to
make
the cross-section through the
body
of the bolt
about equal to the section at the bottom of the threads.

be done by upsetting the ends where the thread
is to
This
be cut,
by welding on ends made from stock somewhat larger than that used for the main length of the bolt. The most apparent result of this practice is to economize material without sacrifice of strength (as the shank still has an area of cross-section equal to the threaded portion), and if the weld (when the ends are welded) is perfect, the strength of the bolt
is
not reduced.
It
seems probable that

rods.
this
reason
it
is
responsible
for the original adoption of this practice, since
has been most
generally used in long

to shock, there is
tie
However,
in case of bolts liable
an even more important reason for such concan be shown that the reduced section not only
strength under static load, but
it
struction
since
it
maintains the
full
very greatly
last fact
increases the capacity of the bolt to resist shock.
This
has not been very generally recognized, as appears from the com-
mon
It
application of such reduced shank bolts only to structures,
rather than to machines.
has been seen that the resistance of a tension

is
member under
a static load
in a
determined solely by
its
weakest section; while,
member
subjected to shock, impact, or impulsive load, the

total
resistance
depends upon the

in Art. 24, the
extent of distortion of the
member due
As shown
to a given intensity of stress.
maximum stress with impulsive load is
79
W
P
(h
k d
+ d) A
For a
stress within the elastic limit
2W
( h
+ 8\
2W ( h
This shows clearly that

or with impact, the stress
for a given load,
produced
in a
W, applied suddenly member of sectional area,

Hence,
if
is
greater as
becomes
less relative to h.
is in-
creased, the stress produced

action;
becomes
less for a given

is
impulsive
or the resistance to such action
greater for a given
value of the
If
maximum
stress.
is
an ordinary bolt
subjected to shock in a direction to pro-
duce tension, the

the
tion will
sections,
stress will
be a
maximum
at the sections
through
bottom of the threads; the bolt

hence the
stress will
will elongate,
but the elonga-
be confined largely to the very short reduced (threaded)
be
much
less in the larger portion
of the bolt.
In a Sellers bolt of one inch diameter the area

is
of
the shank
.78 sq. inches, while the area
at
the bottom of
threads
1U lbs.
is
only .55 sq. inches.
Therefore a stress on A' of 30,000

21,000 on the
.,
r 11
per sq.
in.
,SO,000 X = - - -SS - =
.78
full sections,
bup-
pose the elongation per inch of length at a stress of 30,000 (taken

as the elastic limit)
1
f nnnr"* Each inch of section A will elongate while each inch of full section A ( = .78 sq. in.) will have
is
o\o ",
a stress of only 21,000
lbs.,
with a corresponding elongation of
Assume the thread to be 1" long, and the To X two remainder of the bolt to be 5" long. It will appear that the mean
=-0007".
stress
on the threaded portion
(1") is
about the
mean
of 30,000
and 21,000, or say 25,500 lbs. per square inch; as the mean secHence the tion is an average of .55 and .78 square inches. elongation for this threaded 1 inch, when the stress on ^' = 30,000,
is
.00085", while the other 5" (of area
A)
will elongate
under
this load 5
X
4-
.0007 .0035
.00085
= =
.0035".
The
total elongation will
then be
.00435 inches.
l8o
MACHINE DESIGN
ith = Ty,
w = A'p
8250
o
X^
55
2
30,000
^- =
.00435
.10435
.0416
^11 =
344
lbs.
Now, suppose
to
the 5" shank of this bolt were reduced in section
an area
^4'
= .55.
Then
5
the elongation of this portion under
the above load would be

the total elongation
.'.
.001
.005", instead of .0035"
and
would be d
.00085
.
.005
.00585.
.55
30,000 *
-^ =
,ooc;8s
8250
0.553
457
lbs.
This
The
bolt
to A'.
latter load is 33 per cent greater than the preceding. preceding example shows that the elastic resilience of the
was increased 33 per cent by reducing the body of the bolt Of course the gain would be still greater with a longer It may be well to remember that the "long specimen" is bolt. more apt to contain a weak section than is a short specimen; but, on the other hand, the sharp notching of the threads is quite
liable to start a fracture at their roots.
If the bolt is strained
beyond the
elastic limit, the portion thus
strained yields at a
that given above.
much
in. in
greater rate, relative to the stress, than

stress of
With a load which would produce a

30,000
)
30,000
lbs.
per sq.
the larger portion (area A), the stress in
the reduced portion (area
.785
will
be
43,000
lbs.
per sq. inch.
Hence, the
is
effect of a
long section in resisting
shock without rupture

elastic
much
greater even than that
shown
for
deformation only.
section of the shank of the bolt may be reduced as in Fig. by turning down the body of the bolt to about the diameter at the bottoms of the threads. The collars a and a' may be left to form a fit in the hole. This form is easy to make, but does not fit the hole throughout its length, and it is weak in torsion. Fig. 47 is somewhat more expensive, but fits the hole better, and is somewhat stronger in torsion. Fig. 48 is the form which gives the best fit, and is also the strongest in torsion. If very long it is difficult to make; otherwise it is perhaps the best.
46,
The

These high
l8l
resilience bolts only increase the resistance to im-
pulsive load, not to dead load.
They
are good forms to use in
such cases as the so-called "marine type" of connecting rod,
where the bolts are subjected to considerable shock. For cylinder head bolts, and other cases where a tight
the
joint
is
main consideration,
this
form
of bolt
tests,
may be
entirely unsuited.
Professor Sweet prepared, for
some
bolts such as are used
in the connecting rod of the Straight
Line Engine; of these, half

half
were
solid (ordinary form) bolts,
and the other
were of the
form shown
in Fig. 48.
Tests of a pair of these bolts, one of each kind, showed an elongation at rupture of .25" for the solid bolt, which broke in the
thread;
while the drilled bolt elongated 2.25", or 9 times as
much, and it broke through the shank, the net section of which was a trifle less than that at the bottom of the threads. Drop tests showed similar results. These tests indicate the superior ultimate resilience of the reduced shank bolts. It was shown in Art. 24, page 77, that where a machine member must absorb considerable shock, a rather weak yielding material might be safer than one which is stronger and stiffer, because of the greater elastic resilience of the weaker and more
ductile material.
This principle
is
of
importance in designing
fastenings which are subjected to shock where they

sarily
must neces-
work under high

of
stress.
63. Location
Fastenings.
As
it
previously
stated, screw
fastenings are generally intended to be tension
and from the foregoing discussion

in this
appears that
members only, even when used

be used should
all
manner alone they
are subjected to very high stresses.

is
The
upon
conditions under which a fastening
to
therefore be carefully considered in order that
forces acting
of
may fastening may

it
be provided
or
for.
Further,
the location
the
may not be advantageous, thus greatly affecting its required size. Thus in Fig. 49, if the bolts alone are depended upon to resist the downward force P, they must be carefully fitted, to insure that
each bolt receives
its full
share of this
shearing load.
Through
bolts only can be used in such a case

If the
as studs or tap bolts cannot be accurately fitted.
down-
182
MACHINE DESIGN
force
is
ward
resisted
by a projecting
ledge, as at A,
fit
which
"is
the preferable way, the bolts need not

either studs or tap bolts can
to rotate
the holes closely
and
around
of the
be used. The bracket now tends and the moment of the load P I must equal
of the bolts
the
is
sum
moments
round the same
point.
It
evident that the lower bolt must be considerably larger than
the upper bolt, to be equally effective.
In small work
it is
con-
make all bolts the same size, the sum of their resisting moments being made equal to P I. In large work the bolt at C is often made large enough to exert a moment equal to P I, and the bolts near A serve only to insure correct location. The upper bolt should, in any case, be located as far away from A as
venient to
possible.
4^
Fig. 50.
Fig.
Fig. 4q
In
it
many machine
parts, such as flywheel rims
and brake bands,
often occurs that the bolts cannot be placed directly in line
with the applied force but must be at a distance / (Fig. 50) from The bolts in such cases may be subjected to its line of action.
both flexure and direct
stress.
Thus
in
Fig.
50,
if
the bolts
fit
the holes in the lugs tightly such a combination of stresses will
be induced.
in Fig.
51.
In such parts as brake bands the connecting bolts
are often used as a

If
means
of adjustment against wear, as
shown
in such a case the lugs be
weak and
is
yielding, the
threaded portion of the bolt will be subjected to both flexure and

direct stress.
The threaded
portion of the bolt
particularly
weak
against flexural stress because cracks are easily started at
the root of the threads,
and where screws are used
in this
manner
they should be designed with a large factor of safety.

64. Screws for
83
the
Transmission of Power.
It
has been
for
pointed out, in
Art. 53, that the square thread is
most used
transmitting power because
of
its
higher efficiency,
and that
is
when wear must be compensated

serviceable (see Figs. 37 d
half
for the half
thread
most
and
).
Where
the thread angle of the
thread
is
small, as in the
Acme
thread, the general equa-
tions which have been deduced for the square thread
may
be
used without great error.
Equation
(6),
Art. 54, expresses the relation
which
exists be-
tween the turning moment which must be applied to the screw,
and the moments due

thrust collar.
to the load, friction at the thread
and
at the
An examination
of this equation
shows that
for a
given applied force P,
the load
which can be overcome desince s

s, and inadded to the
creases with an increase in the value of the pitch

creases with a decrease in the value of
s,
is
numerator
plied
of the fraction
and subtracted from the denominator.
This can be seen
in another
way by
considering the energy sup-
the force P be applied through a complete revolution, or a distance of - d m the load will be raised a distance equal to the pitch s. Evidently, if 5 is
and the work performed.
If
decreased, a greater load can be raised by a given force P; since

(neglecting friction), the force applied, multiplied
by the space
through which
it
moves, must be equal to the load multiplied by

it
the space through which
is
raised.
In other words, the me-
chanical advantage of the screw can be varied by reducing the

pitch angle;
and
it
is
evident that by reducing the pitch angle a
small force applied at a long radius

load.
may be made
to raise a great
In order that the thread on the screw

strong, with similar materials, the thread
arc
and nut may be equally and space on the screw
made equal
As
to
each other and therefore equal to half the

is
pitch.
the pitch
increased the axial width of both thread

if
it
and space are necessarily increased, and

very heavy thread
is
desired to keep
the section of the thread square in form, this soon results in a
when compared
to the cylinder
is
on which
it
is
formed.
If
the depth of the space
reduced, to avoid reducing
the diameter of the cylinder, the bearing surface of the screw
184
MACHINE DESIGN
is
and nut
reduced, which
is
not desirable.
It is
customary in
axis of the
such cases to divide the axial width of the thread and space into
several equal parts, arranged alternately
round the
screw, thus forming several parallel threads and spaces.
The
means, be greatly reduced and surface be provided. wearing ample Such screws are called multiple threaded screws and may have two, three, or more
this
depth of the space can, by
parallel threads.
tical
The
theory of such screws
is
evidently iden-
with that of the single threaded screw.
Power Transand (10), while expressing the general relations which exist between efficiency and the pitch angle, do not show clearly the effect upon the efficiency due to varying this
65. Friction
of
and Efficiency
(9)
Screws for
mission.
Equations
angle.
In Fig.
52
fi
t
these equations have been plotted for various
constant values of
and these curves show graphically the
effect
of varying the pitch angle.
An
examination of
this figure
shows
that for the value of n chosen the efficiency increases rapidly, as
the angle increases,
and 50 angle, becoming zero again near 90 It is to be noted that between 20 and 6o the efficiency does not vary materially with change of angle, and that when the
tween 40
,
.
up to 15 or 20 attaining a maximum beand then decreasing with an increase in the

,
efficiency of the
screw alone
is
considered, steep pitched threads,
from 30 to 50 pitch angle, give maximum efficiency and hence a more durable thread. It is seldom feasible to use such
as
pitches in practice, for reasons that will be presently discussed.
The
curves in Fig. 52
will
be
found useful in making
trial
assumptions for the efficiency of screws.
Screws for transmitting power are usually

freely, hence, in general, their
difficult to lubricate
rubbing surfaces are imperfectly

coefficient
lubricated (see Art. 28).
The
of friction for screws
working under pressures ranging from 3,000 to 10,000 lbs. per square inch, and at low velocities, has been experimentally determined * by Professor Albert Kingsbury.
From
his experifj.
ments
it
appears that, for these conditions, the value of
may be
* Transactions of American Society of Mechanical Engineers, Vol. 17.
Efficiency
1 86
MACHINE DESIGN
For pressures lower than 3,000 lbs. per square and velocities above 50 ft. per minute, the value of /i may.15.
taken at
inch,
be assumed
at .1,
if
fair lubrication is
maintained.
The may be
bearing pressure per unit area of thread surface that

carried on a screw thread, will vary greatly with the
If
conditions of service.
the velocity
is
low,
and wear not an

important,
important factor, as in the case of jack screws, very heavy pressures
may
be carried; but where accuracy of form

ft.
is
and
where the velocity exceeds 50
per minute, the pressure
per unit area should not exceed 200
lbs.,
and
is
for
such service as
it
lead screws, where maintenance of form

as low as possible.
If
essential,
should be
PT = load carried.
p = intensity
of pressure per unit of projected area of thread. of threads per inch.
n = number
/
= length
x
of nut in inches.
d = outside, diameter of thread.
= inside
diameter of thread.
Then
= pnl-[d
4
d>]
the
is
(17)
The
load per unit of projected area
is
same
as the load per
unit of true area, since the projected area
equal to the true
area multiplied by cos
and the
axial or projected pressure is
equal to the normal or true pressure multiplied by the same

function.-
66. Stresses
in Art.
in
Transmission Screws.
It
has been shown

thread
is,
59 that the resisting

(6),
moment
at the
from
equation
equal to
r$
\
-TT
+ d
ft
- d ~i ln
or to the total turning

5
/i-l
moment applied, minus This moment induces a

direct action of the load
* This statement applies
is
the frictional
moment
at
the collar.
torsional stress* in the

is
screw.
The
to
induce a tensile or compressive
collar
most usual case only, where the thrust power is applied. If the collar is not located at the end to which the power is applied, the total torque, Pr m of equation (6) is transmitted through the body of the screw.
strictly to the
located at that end of the screw to which the

the screw equal to
87
stress in
(where
is
is
the area at the
root of the
six
if
thread)
if
the screw
short.
If the
screw
is
over
times as long as
is
its least
diameter d lf the compressive
stress,
the screw
in compression, will
be that due to W, considering

(1)
the screw as a long column.

III,
Equations
and
(2) of
Chapter
page
48,
and
of
their discussion in Art. 18, are therefore, ap-
plicable to the design of such screws.

67. Design
Screws
for
Power
Transmission.
An
in-
spection
of
Fig.
52 shows that screws of small pitch have very

it
low efficiency, and
would seem desirable
for that reason to

it
keep the pitch as great as possible.

increases as the pitch decreases.
On
was
the other hand,
was
54
pointed out in Art. 64 that the mechanical advantage of a screw

It
also
shown
in
Art.
that a self-sustaining screw could not have an efficiency of over
50 per cent, and for perfect safety against overhauling
it
should
be
much
less
than this value.
The
if
best pitch, for a given set of
conditions,
flicting
may
therefore be a
requirements.
is
Thus
the turning
compromise between these conmoment which can be

with hand
applied to a screw
limited
(as is often the case
power) a low pitch must be selected in order to attain mechanical

advantage.
In such a case
it is
obvious that care should be used
in selecting the materials of the screw
and nut, so as
to obtain as
low a
coefficient of friction as possible.

steel
Thus
a bronze nut will

(See Art.
run well on a
28.)
screw with imperfect lubrication.

a more efficient pitch
Again
in
such cases as the screws in certain machine
tools, as plate planers,
may
be taken.
If
there
is
no tendency for the screw to overhaul, and the necessary

applied, the pitch of
moment can be
selected.*
maximum
efficiency
can be
Example.
The
force required to
is
open or close a certain sublbs.

It
is re-
merged sliding water gate
estimated at 6,000
quired to design a single-threaded steel screw such that one

* In certain forms of saw-mill carriages these conditions exist,
for
man
netting the log over to the
and the screws saw may be and are made with a very efficient
pitch.
1 88
MACHINE DESIGN
hand wheel 40
is
exerting a pull of 60 lbs. at the periphery of a
inches in diameter, attached directly to the screw, can operate

the
gate.
to
The
be 4
ft.
greatest
unsupported length of the screw

friction
found
Let the coefficient of
= .15,
the crush-
ing strength of the material
= 30,000
lbs.
per square inch, the

lbs.
maximum working
Since the screw
tensile or
compressive stress = 10,000
per
square inch, and the coefficient of elasticity = 30,000,000.

is
in compression in closing the gate

at
be designed as a long column square

if
both ends (case
must 4), and

it
so designed
it
will
have surplus strength when in tension.

is
The
effect of the
thread in stiffening the screw

total
small and will

is
be neglected. be 10,000
lbs.
The
maximum
unit stress in the screw
to
and it is evident that the maximum compressive stress p will form the larger part of the total stress; p may therefore be taken at 9,000 lbs. per square inch, and the mean compressive stress p assumed at 6,000 lbs. per square inch whence
f
:
the
trial
area
of
the
screw
at
the
root
of
the
thread
P = =
;
6,000
6,000
i sq.
inch, or a diameter of
1 Ms
inches.
Check-
ing this assumption by formula
N page 94
/ = 1T
I+
P.
2
(l\*
77
~^^
9,000
2
mn E\p/
closely
^8V =
stress
15
5>3olbs.
4 Xtt
X 30,000,000V. 28/
assumed.
per cent.
which checks
enough with the mean
Assume
the efficiency of the screw and collar at
The
energy which the operator can supply in one complete revo-
lution of the wheel
ttX 40
60
7,600
in. lbs.
Hence the
dis-
energy delivered at the nut will be 7,600
.15
1,140 inch lbs.
But during one revolution

tance equal to the pitch
s,
of the screw the gate
must move a
against a force of 6,000 lbs.
Hence
.
X
-r
6,000
1,140
*.
.19" or say
0,000
.2" so
that the thread
may
be easily cut in a
lathe.
is
Since the thread
square the outer diameter of the screw will

be dA
89
/i"
the outer diameter

ter of the
screw
+ .2 = 1.325, or in order to use a standard tap may be taken as \]4" The corrected diameat the root of the thread will be 1.5 .2 = 1.3"
,
and the corrected mean diameter

these
a
is
will
be 1.5 .1
1.4*.
For
to
values
tan
X
5
.2
=
1.4
.046 which
corresponds
= 2 40'.
From
curve
(Fig. 52) the efficiency of the
screw
about 13 per cent, and the original assumption of efficiency
is sufficiently close.
The
in.
."
twisting
lbs.
II
'
moment applied to the screw = 60X20 = 1,200 The frictional moment at the collar is approximately
.15
rm
6,000
X X
.7
630
in.
lbs.
Hence the
torsional
moment
at the
nut*
T =
The
page
91,
60
20
630
570
in. lbs.
is
torsional stress
due
to this
moment
by equation E,
p = ^
e
16
T
rr 3
16
*
X
:
570
r=
1,320 lbs.
(1.3)
.*.
by equation
(I),
page
2 ]
(48),
the
maximum
\/9,ooo
2
direct stress, i,32o)

2 ]
An
A [p
l
f + 4 A = K [o>oo +
is
+ (4X
9,600
is
lbs.,
which
less
than
the
assigned limit and the
design
therefore correct.
The
ment
that a
is
increase in the
maximum
stress
due
to the torsional
mo-
here only about 6 per cent.

greater
Where
the screw
is
short, so
much
is
mean
direct stress
can be carried, or where

If the
the screw
only in tension and hence admits of a high tensile
stress, this increase

is
may be from
it
15 to 20 per cent.
screw
made
of cast iron
should also be checked for shearing by
equation
* It
is
(2) of Art. 16.

at the
assumed that the
collar
is
upper end where the power
is
applied.
CHAPTER
VIII
KEYS, COTTERS, AND FORCE FITS

68.
Forms
of steel,
of Keys.
Keys are wedge-shaped

etc.,
pieces, generally
made
tion
which are used primarily
to prevent relative rota-
between shafts and the pulleys, gears,
which they carry.

keyed
to
On
account of the frictional resistance which they induce between
the surface of the shaft
and the member which
is
it,
they
also often prevent relative sliding of the parts.
Keys are most

in
usually rectangular in cross-section;
but occasionally they are

is
made
of
circular form.
saddle key
shown
to
Fig.
53.
its
This form of key does not require the shaft

holding power
is
be cut;
but
so small that
it is
used only for light work.

the
For
(c)
small loads, or as a safeguard

pin as
when
hub
is
shrunk on, a round
shown
in Fig. 54 (a) is often used. of applying
Figs. 54 (b)
and 54
show two other methods

stitute for
round taper pins as a subin Fig.

55.
keys.
flat
key
is
shown
This form
its
requires a small portion of the shaft to be cut away, and
hold-
ing power
is
much
is
greater than that of the saddle key.
The sunk
and
is
key, Fig. 56,
the most secure form of key fastening,

It is so called It
more
used than any other.
because
it
is
-sunk into a
keyway
or groove cut in the shaft.
thus requires more metal

flat
to be cut
away from
the shaft than the
key,
and
this
is
taken into account in designing shafting since the metal
must be removed
is
from the outer

applied loads.
fibre
where
it
is
most serviceable
for resisting
The keyway
is
cut in a shaft for a sunk key
made
of the
parallel to the axis of the shaft; but the
keyway
in the
hub
pulley or gear which

in Fig. 56 (b).
to
be
made
fast is cut tapering as
shown
driven
The
sides of the key are parallel, as
shown, and
is
should
fit
well in both shaft
and hub.
190
When
the key
KEYS, COTTERS,
in,
AND FORCE
FITS
TQ1
the shaft and
the shaft
set
hub arc drawn tightly together on the side of opposite to the key, and the frictional resistance thus
to prevent relative sliding of the parts lengthwise
If the
up helps
of the shaft.
bore of the hub
is
tapering, or
if
the key
fits
Fig. 53.
Fig. 54 (a).
Fig. 54 (b).
Fig. 54
(c).
Fig. 55.
more
tightly at
one end than
at the other, the part

its
keyed on may
required, as in
be thrown out of alignment so that

to the axis of the shaft.
plane
is
not perpendicular
is
Where
great accuracy
flanged couplings on shafting, owing to this tendency, the faces

of the flange or part secured to the shaft after the
should be faced in place

to
key
is
driven.
If the part
keyed on does not have

a tight or press
fit
be removed often, the hub

should such occur.
may be made
on
the shaft, thereby preventing largely the tilting action of the key
In the Woodruff system (Fig. 57), the key
Fig. 57.
is
a circular segment and the
keyway may be
will
cut with a milling
cutter.
This allows the key

in the
to adjust itself to the taper of the
keyway
hub, hence
it
not
throw the keyed part out

this
of perpendicular
alignment.
With
system, the hub must
192
MACHINE DESIGN
These keys are used
largely in
be forced on over the key.

chine tools.
ma-
In general, the part to be secured on the shaft

position
is
placed in
and the key driven in. This makes it necessary to exkeyway along the shaft at least the length of the key, the tend (except when the hub is at the end of the shaft) unless the diameter of the shaft is enlarged under the hub, sufficiently to allow the
shaft proper.
ally, it is
keyway to be Where it is desirable

called a
cut without cutting into the

to
withdraw the key occasion-
shown in Fig. 56 (b), in draw key.* Sometimes, however, it is not desirable to extend the keyway beyond the hub, in which case the keyway in the shaft is made the same length as the key, and the hub is driven over the key into its correct position. Much more force is necessary to drive the hub into place in this manner than to drive the key, on account of the friction between When the hub is a sliding or an easy fit the shaft and the hub.
often provided with a head, as
it is
which case
on the shaft, and only one key is used, there is a tendency to throw the hub eccentric to the shaft. Under these circumstances there is a tendency for the hub to rock and work loose on the
shaft, especially
if
the direction of motion be reversed.
In such
cases two keys
set
90
apart
make
much more
secure fastening
If
as this gives three lines of contact and prevents rocking.

of these keys
is
one
a saddle key, as
shown
is
in Fig. 53, the fitting is
greatly facilitated
and the fastening
almost as secure as with
two sunk keys.

69. Stresses in
relative rotation,
Sunk Keys.
it
Since keys are designed to prevent
is
evident that every key must transmit a
certain
torsional moment or torque. This torsional moment may be equal to the total torque transmitted by the shaft, or the key may be required to transmit only a part of it. This would
indicate that keys of different sizes should be used with any given
diameter of shaft, depending on the load which the key must

transmit.
For practical reasons, however, such as standardiza-
Where
draw key cannot be used

it
the point of the key
is
sometimes case-
hardened so that
will not upset so readily in
being driven out.
KEYS, COTTERS,
tion
AND FORCE
FITS
1 93
of the shaft
and interchangcability, it is desirable that the dimensions and key should bear a fixed relation to each other. All practical systems of keys, therefore, give a fixed size of key for each diameter of shaft, the dimensions of the key, presumably,
being such that
its
strength
is
equal to the torsional strength
of the shaft.
Shafts are usually designed for torsional stiffness
rather than torsional strength, which results in a shaft consider-
ably larger than necessary as far as strength
is
concerned.
If,
under these circumstances, the key

it
is
designed as indicated above

the shaft
is
will also
have excess strength.
Where
short
and
is
designed for strength alone, the key should be more carefully
considered.
Keys
resisting a torsional
moment
manner
are subjected to
simple
crushing, or to crushing
and shearing, depending on the manof

fitting.
ner of their application and
The
ordinary
to the
sunk key
(Fig.
56
a), is
subjected to a force,
lt
due
,
and to a resisting force, F 2 due to The effect of the reaction from the hub which it secures. these two forces is to set up a shearing stress along the middle
pressure from the
shaft,
section of the key at the outer surface of the shaft.
They
also
form
a couple
which tends
to rotate the
key in the keyway.

results, resisted
This tend-
ency to rotate should be, for best

of the
hub and
is
shaft against the top

fit
the key
not a tight
pressures,
and
if
will
by the pressure and bottom of the key. If on the top and bottom these resisting be concentrated near the corners. This
be sufficient to crush the key
at these
concentrated pressure
points,
may
and allow
it
to roll in the
keyway, deforming both the

to
keyway and key and subjecting the key

action rather than simple shear.
If
a severe crushing
the conditions of service
require a continual reversal of motion a state similar to that
shown
and
in Fig.
56
(c)
is
induced, where the resisting
forces
have been moved inward and their moment arm made so
short that their magnitude

position.
It is
must be very great

which
to hold the
stress
key in
This may bring a severe bursting

fit
on the hub.
well
evident, therefore, that keys
sidewise only, cannot

fit
be depended on to carry as great a load as those which
on the top and bottom.

13
Where
great accuracy
is
required, as
194
in
MACHINE DESIGN
machine tool construction, the hub is often made a force fit on the shaft and the key fitted only on the sides, so that it cannot throw the parts out of relative alignment by radial pressure.
Referring to Fig. 56
(a)
Let
"
the length of the key or

" thickness of the key
hub
t=
b
"
" "
T= p=
" breadth of the key " torsional moment applied to the shaft " force acting at the radius of the shaft so that
P-=T
2
Then for shearing stress p

and
the
since the torsional
p=
moment
n ........
(1)
moment
d
applied to the shaft must equal
of the crushing load applied to the side of the
key
T = P- = F,a
2
orP- =
2
p rc
l-(- -) V2
2
4/
....
(2) J v
If
be considered
reduces to
to act at the radius of the shaft
(which can
be done without serious error for keys as ordinarily proportioned)

equation
(2)
r = PJ L 2
Equations
(1)
......
is
(3)
and
(3)
may be
used to compute the stresses in any

to equal the crushing
sunk key.
If the
shearing resistance of the key
resistance, then
from
&
(1)
and
c
(3)
p lb
.
.
= p l~
or
2 P*
t
=
'
(4) V4;
If
p t = b, and the key is square for equal resistance to shearing and crushing. For machinery steel, such as is generally pc
2
s
,
used in keys,
p =
a
.8,
and hence from
(4) for
equal strength in
'
KEYS, COTTERS,
shearing and compression
t
AND FORCE
If,
FITS
195
= i.6b.
key
in addition, the
moment
of the shearing resistance of the

resisting
is
to
be equal
to the torsional
moment
of the shaft, then
T=
Pjb-.=ra
(5)
where p\ is the shearing stress in the outer fibre of the shaft. For steel shafts and keys, which are most common, p = p\ whence
from
(5)
lb-=~r 10 2
The minimum
is
*d3
length of hub (7),as determined by practice, which
necessary to give a good grip on the shaft, should not be less
than
Substituting this value of

3
in equation (6)
bd _ *jF
2
4
.
.
d d = - nearly 12
4
16 -
(7)
The above would,

or thickness
t
therefore, give keys of breadth b
d =
4
depth
j.6 b
.4 d,
and minimum length -
d.
Keys
as
used in practice conform closely to these rules as far as length
and breadth are concerned; but,

of the shaft, the thickness
is
to avoid cutting
away
so
much
usually
much
less
than that given
above.
An
average value of the thickness
may
it is
be taken at
b.
8
This gives a key considerably thinner than

it
wide and makes
weakest in crushing.
The
crushing resistance can, however,
be increased by lengthening the key or by using a hard grade of

steeL
Keys designed
since the friction
as above usually have
an excess of strength,
between the shaft and the hub materially decreases the load actually brought upon the key. In addition, as
has been pointed out, shafts are most usually designed for
stiff-
ness or angular distortion, and therefore are greater in diameter
196
MACHINE DESIGN
If the
it
than would be required for strength alone.

proportional to the shaft diameter as above,
key
is
made
fail
must, therefore,
have excess strength against rupture; and such keys seldom

unless subjected to severe shock or extraordinary loads.
There are no fixed standards various machine builders having

following table
for the
their
dimensions of keys,
standards.*
own
2
The
may be
taken as representing average practice

is
when
the length of the key

less
it
not less than
d.
If the
length
must be
puted, as
than
this value, the
crushing stress should be com-
may
be necessary to use two keys.
TABLE
XI
DIMENSIONS OF FLAT KEYS IN INCHES
Diam.
of Shaft d.
1*
I*
If
2i
5
31
1
8
10
Breadth of Key
5
.
16
3
8
T6
9
a
4
1*
if
X4
a
ii
Thickness of Key
/.
5
.
3 2
T6
32
T6
16
T6
13
ii
The
drives
taper of sunk keys
is
usually about
]/%"
per foot of length.
is shown in Fig. 58. This key by compression or as a strut. The keyways are more difficult to cut, the keys more difficult to fit, and the shaft is cut
Another form of sunk key
deeper than for the
common
form.
It
has been used with great

desirable to have
success on very heavy work.

70. Feathers or Splines.
Sometimes
it
is
the
hub
free to slide axially along the shaft,

it.
but constrained to
is
rotate with
In such cases a feather or spline
used.
The
sides of the spline are parallel

rigidly to the shaft or
it
and
it
may be
either fastened
may move
with the hub.
Small splines
are frequently dovetailed into the shaft (or hub), as shown in

Fig. 59
(a),
while larger ones are often held in place by means

(Fig.
of countersunk screws
59 b), or
rivets.
it
A common way
of securing the feather so that
will
move
* See Kent's "Mechanical Engineer's
Handbook," page
977.
KEYS, COTTERS,
with the hub
is
AND FORCE
FITS
197
shown
in Fig. 60.
Splines are subjected to a
shearing stress across the mid-section at the radius of the shaft,
and
any
to a crushing stress
fitted
Being
on the sides in the same way as sunk keys. loosely on the top and bottom, they do not produce
offer
friction
less
between the hub and the shaft and, therefore,

resistance than
much
sunk keys
to
the rolling action imto
posed upon them (see Art. 69).
This rolling action tends
bring a concentrated crushing force at a and b (Fig. 59
a), if the
Fig. 58.
Fig. 59.
feather
is
not rigidly secured to either the hub or the shaft.
For
this reason,
and
in order also to provide
ample wearing
surfaces,
feathers are usually given a greater radial depth than
sunk keys,
and from
their general
It
is
proportions are often
distinguished as
is
square keys.
evident that the holding power of splines
not equal to that of sunk keys.
with
The following table common practice:
gives dimensions of feathers
which agree
TABLE
XII
DIMENSIONS OF FEATHER KEYS IN INCHES
Diam.
of Shaft d.
li
li
if
4
t
3h
10
Breadth of Feather
b.
T*
1*
if
ij
if
ai
Thickness of Feather/
A
is,
i*
if
if
if
a*
a|
The
length of feather keys
in general, greater
than that of
sunk keys, for the same
size of shaft, in
order to reduce the bear-
ing pressure and increase the wearing surface on the sides.
198
71. Cotters.
sliding
MACHINE DESIGN
A cotter is
a form of key used to prevent relative

Fig.
between two members.
61
shows a method
of a cotter.
of
this
securing a piston rod to a piston by

case the connection
is
means
In
permanent
in character, the cotter being

is
removed only when the piston or piston rod

newed.
rod
is
repaired or re-
In other forms of cotter ed joints of
this character the
not tapered, but

of a shoulder or
means
rectangular in section,
is prevented from sliding into the boss by by the cotter alone. The cotter is usually but sometimes the edges are rounded so
as to avoid sharp corners in the opening cut through the rod or

to facilitate
machining.
In light work a taper pin of circular

Fig. 62
section
is
often
used as a cotter.
shows an arrange-
Fig. 60.
Fig. 61.
Fig. 62.
ment
of a gib
is
and
cotter
(commonly known
is
as a gib
and key),
such as
used on the ends of the connecting-rod of steam engines.

to prevent spreading of the strap.
The
function of the gib
This arrangement permits a small amount of adjustment between

the strap and the connecting-rod for taking
up wear on the pin

form shown
in Fig.
and
61
brasses.
72. Stresses in Cotters.

is
A
at
cotter of the
beam supported
the ends.
The
exact distribution
of the loading is indeterminate, as the

to concentrate the load
bending of the cotter tends

It is suffi-
near the points of support.
ciently accurate, however, to consider the load as uniformly distributed.
The
area of the surface of the cotter where

also
it
bears
on the rod, and
on the hub, should be
sufficiently great to
prevent crushing of the material.
This indicates that the diam-
AND FORCE
eter of the
FITS
1 99
hub should,
is
for similar materials,
be twice that of
the rod,
which
the usual proportion.
The
section of the cotter
at the point of
support should be great enough to prevent shearing,

it is
and
in
many
cases
sufficient to
compute the
is
section for shear
alone, neglecting the
bending action.
When
is
a cottered joint of this character

tight
made, the
out.
cotter
must be driven in
especially true
enough
to prevent its
is
backing
This
when
the load
a reversed one as in the case

initial stress in the
of the steam-engine piston.

cotter
This induces an
and rod, over and above that due
to the load P.
The
to the
it
conditions, in fact, are
somewhat
similar to those
which
exist in
screwed fastenings (see Art. 60).
The
all
initial stress
due
driving of the cotter cannot be accurately computed, though
may be
very great.
For
this
reason
calculations of dimensions
based on the
maximum
if
applied load should be modified to suit
the conditions of service
and the materials
of
which the
joint is
made.
as
is
Thus
the rod be of brass
common
cast iron.
in
pump
and the hub or boss of steel, work, the proportions would be different
the materials were of steel or of steel
from those employed
if all
and
= the diameter of the rod where the cotter passes through. = thickness of cotter. b = breadth of cotter. Then, in order that the net cross-section of the rod may be as
Let d
"
'
strong in tension as the cotter and rod, where
they bear upon
each other, are in crushing,
For a
steel
rod and
steel cotter
where p
= .& p
= .^d
(2) (3)
Good
practice gives b =4/ i.y6d

of cotters, as
The taper
that there
is
shown
in Fig. 61,
should be so small
no danger of backing out and should not exceed
inch per foot of length.
An
auxiliary locking device
is
often
200
MACHINE DESIGN
used in arrangements such as shown in Fig. 62, in which case

the taper
may be
as great as
in 8.
62, the
stress
In the form of cotter shown in Fig.

driving the key
due
to
may be
disregarded, and the design based on the
maximum
applied load.
The
student
is
referred to treatises on
this form.
steam-engine design for relative proportions of

It is often
necessary to allow a rather high bearing pressure
on the
cotter to avoid large
and clumsy proportions.
An
ex-
amination of successful practice shows an allowable pressure of

15,000 pounds per square inch as computed from the applied
load.
FORCE AND SHRINKAGE FITS

Crank discs, the hubs of heavy and work in general which is to be subjected to shock or vibration, must be fastened to the shaft more securely than can be accomplished with a key, when In such cases the bore of the hub is a sliding fit on the shaft. the hub is made slightly smaller than the diameter of the shaft, and the shaft is forced cold into the hub or the hub is expanded by heating till the bore is slightly larger than the shaft, then The first slipped over the shaft and allowed to cool in place. method is known as a force or pressure fit, and the second as a
fly-wheels, impulse water wheels,
shrinkage
fit.
The
degree of tightness or "grip" required beservice.
tween shaft and hub depends largely on the

shafts
Thus, with
between
up
to three or four inches in diameter, a difference
the diameter of the shaft and the bore such that the parts
may
Such
be driven together with a hand sledge

a
fit is
is
often satisfactory.
called a driving
fit,
diameter and the bore

ture spiders
is
very small.
and the difference between the shaft With such work as armafor the press
fit
and fly-wheel hubs, the allowance

facilities for erection.
depends largely on the
If the parts
can be
forced together in the shop, where adequate means, in the form

of a powerful hydraulic press
is to
be had, an allowance requiring
a pressure of one hundred tons or more

the parts
to
may be made. But if must be erected in the field, this allowance may have be reduced on account of the difficulties of erection. It is
KEYS, COTTERS,
AND FORCE
FITS
201
usually possible in the case of armature spiders, fly-wheel hubs,

etc., to
obtain a sufficiently tight grip on the shaft by

fit
a press
without inducing undue stress in the parts.
means of Dependmeasure,
ence for preventing relative rotation
may
be, in a large
placed upon the key in all such cases. In such work as crank shafts when
parts,
it
built
is is
often necessary to insure as strong a grip

possible without inducing
up from separate upon the
shaft as
undue
stress.
greater
difference between
the shaft diameter

fits
then allowed than in forced

together by shrinking.
are of importance
74. Stresses
of the hub is and the parts are usually put
and the bore
In the latter cases the stresses induced
and should be carefully considered.

to
r,
Due
Force
Fits.
If
x be
2
the elongation or con-
traction of
any radius
then 2 -
is
the corresponding elongation

r.
or contraction of the circumference
The
elongation or
2 ~
x
.
contraction of the circumference per unit of length
is
- r
If
p be the
then
stress
which would induce
this
change of length of
circumference, and
E be
the coefficient of elasticity of the material,
E _ -L-
or
x . ii
(I)
In Fig. 63,
let
represent a hollow shaft on which has been

of the contact surface
forced or shrunk a
hub or boss B, the radius

and the inner radius
being
r2
Before the operation of pressing, the outer radius of
the shaft
was
is,
r2
+e
2,
of the
hub was
r2
2.
The hub B
therefore, in the condition of a thick cylinder sub-
jected to an internal pressure,
and the shaft
is
in the condition
of a thick cylinder subjected to an external pressure.

est tensile stress will
The
great-
be found at the inside surface of the hub,

stress at the inside surface of the
and the greatest compressive

shaft.
If,
therefore, e be the difference
between the outer radius
of the shaft
and the inner radius of the hub, before pressing,

2
.
then
= e + e'
2
202
MACHINE DESIGN
.
Let p be the unit tensile stress in the hub at a radius r 2 Let p c be the unit compressive stress in the shaft at a radius
t
r2
Let Let
be the unit radial pressure between
and B.
rt
be the internal radius of the
shaft.
Let r3 be the external radius of the hub.
Then from
(i)
= r 2
+-r
or
2)
(2)
The
is,
general equation for the stress in thick cylinders of this kind

2
2 ft
*
W,
2 f2
W +
2
/
4.
^*
y2
(W x
-W
r2
O
2\
...
Applying
\o/
Where
r x is the inner radius of the cylinder, r 2 the outer radius,
w
p
the internal unit pressure,
the external unit pressure

r.
and
this
the tensile or compressive stress at any radius
equation to the shaft,
w = o, r = r
1
2,
whence the compressive stress

2
at the surface of the shaft is
(2 r 2
r, )
2^
(4)
In a similar way substituting in the general equation

r 2 for r 3 ,
r x for r 2 ,
for
inner surface of
and w 2 2 the hub is
for
w
+ '
the unit tensile stress on the
^ ~2
A =
Dividing
(4)
\" r 2
'
/..
('.*
^" O = ^2
'
2 3 r3 / ) O
(5)
by
(5)
Pc
a
(6)
A
From
(2)
and
(6)
^
and p Fc =
*
_EeL_
r
(
+
+
w;
p)
r2
(a
.......
/?)
(8)
J^
Merriman's " Mechanics
of
The
following
treatment
is
from
Professor
Materials," 1906 edition, page 396.

that adopted in this work.
The
notation has been changed to agree with
KEYS, COTTERS,
AND FORCE
FITS
203
When
the shaft
is solid,
r x in the
above equation becomes zero
and the equations are much simplified. Example. A hollow steel shaft 10 inches outside diameter and 2 inches inside diameter is to have a steel crank shrunk upon
its
end.
The hub
of the crank
is
18 inches in diameter.
What
must be the difference between the diameter of the shaft and the bore of the crank so that the tensile stress at the inner surface
of the
will
hub
shall not exceed 20,000 lbs. per square inch ?
What
be the corresponding compressive stresses at the outer and
inner surfaces of the shaft?

Here/',
Take = 30,000,000.
r3
1,
r2
=
2
5,
9
(2
and
p 20,000
t
\\
hence
=
2 r2
2
+
rj
2
4 r?
^7
3(>-2
O
_
=
(2
X
2
3(5i)
2N
5 +
)
(4
i )
3 = -
4
2
'
and
_
.
4 rj
)
2
(4
2
X
)
''"3W-0^
(7)
3(9
-5
_ ~~
)
23
Then from
p% r2
(a
Efi
20,000X5(^+2.23)
fi
30,000,000
.0044
2.23
From
e
""
30,000,000
5
r/(
+1)
X .0044 x X 2.98
.75
6,700
From
14)
-A
6 7
'
8,900
lbs.
and substituting
18,500
It is
this value in (3),
making
= r, and w =o,
l
it is
found that the compressive stress at the inner surface of the shaft
is
lbs.
per square inch.

if
evident that
be assumed, which
is
usually the case,

It
the resulting pressure and stresses can be computed.

K
should
be noted that p must be well within the elastic limit to prevent the hub yielding and relieving the pressure. It appears, as
pointed out by Professor Merriman, that the allowances
practice for force
fits,
made
in
induce stresses which should be considered
204
if
MACHINE DESIGN
Thus, in the example
2
other stresses are to act on the members.
given, the total allowance or difference
between the diameter of
the shaft
and the bore
of the
hub would be
would be
X. 0044 = .0088;
.0088
and the allowance per inch

which
is
of diameter
.00088",
10
close to average practice lor force
is
fits,
where .001 " per
inch of diameter
often allowed.
A somewhat greater allowance
Fig. 64.
Fig. 65.
is
generally
made
for shrinkage
fits,
as here the difficulty of forcing
on the hub does not occur.

75. Practical Considerations in
Force and Shrink
Fits.
The
allow-
foregoing equations, while giving the probable stresses and radial

pressure resulting from a force or shrink
fit
made with an
ance
force
e,
are limited in their application to the practical
fits.
fits,
There
is
is,
generally speaking, no difficulty
making of in making
shrink
with any practical allowance, as far as getting the
parts together
concerned; although greater

fits
skill is
required in
fits,
handling shrink
ever, the
is
than force
fits.
In making force
how-
amount
of pressure that can be applied to the parts
often a controlling factor.
tween the shaft and hub (w 2)
The probable radial pressure bemay be found as above, but little is

and
it is
known
of the coefficient of friction in such work,
evident
that this quantity will vary greatly with the character of the
material, the finish of the surface,
and the lubricant applied.

hence
it is
Experimental data are lacking on
this point,
almost im-
KEYS, COTTERS,
AND FORCE
FITS
205
possible to estimate the resistance to slipping offered
by
force or
fits
shrink
fits.
In general, shrink
fits
are
superior to force
since their surfaces are very dry and unlubricated, while those
of a
force
fit
are lubricated.
fit itself,
Total dependence
but a key
is
is,
therefore,
seldom placed on the

will vary for
(a)
also used for safety.
Experience shows that the pressure required to

fit
make
a force
any given diameter.
Directly as the length of the

Directly as the allowance e
hub
(b)
(c)
(d)
As some function of the radial thickness of hub With the character of the materials and the finish
the surfaces.
of
It is
evident that a mathematical expression accurately expressing
these relations would not be practicable, and recourse
must be
had
to successful practice.
An
allowance of .001 " per inch of diameter will represent
average practice in this country for such work as crank shafts,
crank pins, and in general where a
tight
fit
is
required.
is
For
often
armature spiders, or fly-wheels, one-half this allowance

sufficient.
For shrink
fits
a greater allowance
is
often made,
although the foregoing discussion indicates that this should not
be
much exceeded
considering the stresses induced.

is
For further information and practical data the student

referred to the following:
Transactions of the American Society of Mechanical Engineers, Vol.
XXIV.
Machine Design" by Forrest R. Jones. "Machine Design" by W. L. Cathcart. Machinery, Vol. Ill, No. 9, May, 1897.
76.
"
Thin Bands or Hoops.
If the ring or
or shrunk on to a
member be
thin,
band which is forced radially, compared to its

ring.
diameter, the assumption can be made, without appreciable error,

that the stress
is
uniform throughout the cross-section of the

in the
is
The change
due
to
of
form
member on which
the
band
it
is
placed
compression
so small in such cases that
may
due
be
to
neglected, and the stress in the
band may be taken

This
as that
is
stretching
it
over an incompressible body.
practically
206
MACHINE DESIGN
applicable to any ordinary shape of band, but rigidly true for

circular shapes only.
Thin bands band
of this character are usually
shrunk into
position.
Example.
thin steel
is to
be shrunk on
is to
to
a casting
is
whose external
48 inches.
linear
dimension where the band
be placed
What must be
the length of the inside face of the
band
so that the stress per unit area
due
to shrinking will be 30,000 lbs. ?
What
will
be the area of the cross-section of the band in order
that the total stress in the
band may be 60,000
lbs. ?
= the length of band before shrinking. Then 48 / = total amount of elongation of band. ,48^ = unit elongation of1,band. and
Let
I
Whence,
then,
,
if
E, the coefficient of elasticity, be taken as 30,000,000,

30,000,000
'
E =
unit stress
-
'
unit strain
30,000 r = 48
or
47.0s ^' yo ins.
T~
The
total area of the cross-section of the
band
will
be
A =
60,000
30,000
square inches which
may
be distributed in any convenient be shrunk
proportions.
If the part
on which the band

is
is
to
is
circular in
form, the band
in the condition of a thin cylinder subjected to
an internal pressure w per unit area, where w is the radial pressure between the band and the part on which it is shrunk.
Therefore by Art. 78,
per unit width of the band
wd =
P, where
.
P is
the total stress
orw =
Thus
in the
above prob-
lem and
let
the
band be shrunk upon a
circular
hub
x
of diameter
4".
per
let
the cross-section of the
band be
%"
Then P =
lbs.
60,000
15,000,
and
w =
-
a
2P
15,000
4 square inch.
4&
tt
1,962
The
steel tires of
locomotive driving wheels are usually shrunk

of .001" per inch of diameter
)
on with an allowance for shrinkage which gives .001 inches elongation (
per inch of circumference.
KEYS, COTTERS,
AND FORCE
FITS
tire
207
Taking
= 30,000,000,
is
and considering the
a thin band, the
unit stress in the tire
=-.
E A =
Forms
30,000,000
.001
30,000
lbs.
77. Other
of Shrink Fits.
Many machine
steel links or
parts such
as fly-wheel rims are held together

into place.
by
bands shrunk
is
The
theory outlined in the preceding article
clearly
applicable to these members, and their dimensions should be

carefully calculated so that they will not be overstrained
by the
shrinking alone.
If
such members are so designed that they will

elastic limit
be stressed up to the
liable to
from shrinkage alone, they are

elastic limit,
is
be strained beyond the
when an
external
load greater than the total shrinkage stress
applied to the parts
which they hold together, and the link, taking a permanent set, becomes ineffective. In computing the dimensions of such links
allowance must sometimes be
made
for the
is
compression of the
small and
parts held together, but ordinarily this neglected.
may
be
Occasionally a bolt or link
is
used to reinforce a cast-iron

in
member
shown.
place.
against tensile stress.

is
Thus
open frames,
Fig. 65, a
large bolt
sometimes placed on each side of the throat as

bolts are usually put in hot
These
and allowed
to cool in
As
ordinarily applied, the benefit derived from

If
them
is
questionable.
they are designed and fitted so as to put the

at
frame in compression
by the working load
strained
A, an amount equal to the tension induced
P at this same point, without being themselves

elastic limit
beyond the
when
the load
is
applied, then
If,
no
stress
can come upon the frame
itself
from the force P.
however, the bolts and frame are each to carry part of the load,
care should be exercised that the stress induced in the bolts by
the initial load
stress
due
to shrinking is so
low that the additional

initial stress
due
to the external load
does not raise this
beyond the
elastic limit, thus giving the bolts
a permanent set
and destroying
cross-section
bolt B.
their usefulness.
Let A, Fig. 64, represent a cast-iron
member
of
uniform
which
is to
first,
be reinforced against tensile stress by the

that the nut
is
Suppose,
screwed up
till
it
just
208
MACHINE DESIGN
If
bears firmly on the casting.
now an
external tensile load
1S
applied to the casting, the bolt and casting will be elongate

the
iron
same amount A.
is
But the
coefficient of elasticity of ca
steel.
st
only about one-half that of
Hence, since
p=Et>
the stress per unit area in the casting will only be one-half th;
in the steel.
If
2,000 lbs.
is
the allowable unit stress in th e
casting, 4,000 lbs. per unit area is all that
can be thus obtainec

a total compres"
-
in the bolt.
This would lead
to unnecessarily large bolts.

is set
Suppose, however, that the nut

sive load
up
till
is
applied to the cast iron.

casting compressed, the
The
bolt will be elongat

of elongation
r
ed* and the

members.
also
If
amount
o;
compression depending on the cross-section of the respective'
The
unit stress induced in the bolt
and casting
bolt,
will"
be proportional
to the area of their respective cross-sections.
now an
external tensile load
is
applied to the
the'
to
is
tendency
is to
relieve the compressive stress in the casting
and
increase the tensile stress in the bolt.

sufficient to elongate the bolt as
ally
When
the load applied
much
as the casting
was
originIf the
itself,
compressed, the casting will be relieved of
external load
it is
all stress.
is
applied to the bolt through the casting

is
evident that practically the same result
obtained; and after
the compressive stress in the casting
addition to
is fully
relieved
any further
still
induces a tensile stress in the casting and
fur-
ther increases the tension in the bolt.
Usually the cross-sectional
area of the casting
is
very
much
greater than that of the bolt.
Furthermore the compressive

initial
stress
induced in the casting by the
load on the bolt
is
usually very small
compared
can
to the
tensile stress
induced by the working load.

deformation
in
For these reasons

usually
the compressive
the
casting
be
neglected without appreciable error;
and the
bolt
may be
(See
Art.
designed on the basis of the external load alone.

60,
Case
a.)
Example.
load
In Fig. 65
is
/.
let
the section
AB
be stressed by the
whose arm
Let
It is
be the location of the gravity
axis of the section
AB.
desired to keep the stress at
not
* See Art. 59.
KEYS, COTTERS,
AND FORCE
FLTS
material
209
is to
greater than 3,000 lbs. per square inch.

cast iron.
The
be
Let
" "
P = 60,000.
I
= moment of ^ = 10 inches.
inertia of section
= 4,500.
Then from Ple\ + ~T~)
Also
(If),
let
the area of the section be 200 square inches.

91, the tensile stress at
page
A
)
due
to
is
p=
(P
I
=
to
( 60,000
I
60,000
30
io\
J
4,300
lbs.,
and
...,., desired
it is
200
4>5
lbs.
reduce this to 3,000
by reinforcing
bolts.
These
rein-
forcing bolts serve the double purpose of increasing the factor
by reducing the fibre stress, and also of decreasing the frame at the point where the work is done. Let Then the compressive stress these bolts be located 8" from O. induced at A by P' is
of safety
deflection of the
,
AT
\A
P'l'e \
(F_
^200
P'X
io \
'
'
4>5o
But p
p must equal 3,000; therefore
4,300
(
I
P'
V200
P'
8\
)
450
3,000. J
'
Whence P' = 57,000. This is when the full working load P

at the root of the thread
1
the total tensile load on both bolts,

is
applied.
If the
lbs.,
maximum
stress
be taken at 15,000
1
then the area of
each bolt at the root of the thread
is
57)
2
x 15P00
The
i.o sq. m.,
which corresponds closely
to
\%"
bolt.
area of
the
body
is
of a
i^"
bolt,
where most of the stretching takes place,
2.4 square inches.

is
Hence the working
stress in the
body
of the
bolt
2.4
sq. 11,880 lbs. per r *i
in.
That r portion
is
of the
boss which immediately adjoins the throat
subjected to an aver-
age tensile stress nearly equal to the fibre stress at the surface
of the throat or 3,000 lbs. per square inch.
The upper and
lower
portions of the boss have
little
or
no
tensile stress
induced in
them, as a consideration of a section such as

axis
is
XX
whose gravity
at O', will
14
show.
It will
be reasonable to estimate that
2IO
the stress in the boss
MACHINE DESIGN
is
equivalent to the full stress of 3,000 lbs.

its
per square inch through 14 inches of
length, the total length
being 19".
Neglecting the compressive deformation of the boss
due
bolt
to the initial load
from the
bolts, the stress

lbs.,
induced in the
will
when the stress in the boss is 3,000 X 2X14/19 = 4,500 lbs. (remembering that
of
steel
is
be 3,000
of
the
coefficient
elasticity
initial
twice
that
of cast
iron).
Whence
the
stress
in
the bolt will be
11,880
4,500 = 7,380.
The
be
allowance for shrinkage necessary to give this
initial stress will
a -
= ^9A!^
h
30,000,000
0046
The number of
threads per inch on a 1^4" bolt
is 5.
Hence after
the nut has been set
up snugly
.0046
it
should be given
5
.023
of a turn, or should
be turned through 360
.023=8.2 degrees.
This
is
most
easily
done in the case
of large bolts
set
if
the nut with reference to the bolt
when
it
and then heating the body

It is to
of the bolt,
by first marking up snug in a cold state, necessary, and rotating
the nut the desired amount, allowing
to cool in position.
be especially noted that a very small shrinkage allowto
ance
is
needed
is
induce a great stress in the
bolt.
If too great
an allowance
limit,
is
made, the bolts may be stressed beyond the

set the first
elastic
and take permanent
time the external load

applied, a force
applied.
When the external load is again
much
the
will
if
smaller
than the total load
will
strain the
casting
to
point where the bolt becomes effective.

strain the casting further than
it
The
total load
did originally, and even
the stresses induced are not sufficient to rupture the casting, the
stiffness of the
frame
is
materially decreased.
CHAPTER
IX
TUBES, PIPES, CYLINDERS, FLUES, AND THIN PLATES

78. Resistance of
Thin Cylinders
to Internal Pressure.
If a
hollow circular cylinder, whose walls are very thin compared to

its
diameter,
is
subjected to an internal bursting pressure, a
tensile stress is
induced in the walls.
This
tensile stress is
reduced
near the ends by the action of the ends themselves which tend to
hold the walls together.
Let Fig. 66 represent one-half of a por-
tion of a thin cylinder so far

effect
removed from the ends that
their
may be
Let
"
neglected.
unit internal pressure
w = the
r
/
d = the diameter of the cylinder
"
" "
"
/>
= the radius of the cylinder = the thickness of the cylinder walls p = the unit tensile stress in the longitudinal section = the unit tensile stress in the transverse section
t
"
=the length
of the part considered

all
Consider the half of the cylinder as a free body, and resolve

forces perpendicular to the cutting plane.
The normal
is
pressure
on a longitudinal
strip of length
and width rdo

normal
sin
r dd.
The
is
component
of this force perpendicular to the cutting plane

0.
wlr do
/
sin
The
do sin
this
total pressure
to this
plane
I
is
= wIrJ
do
= w
d.
For equilibrium
in the
two sides of the cylinder.

2
normal force must equal the Hence
resisting
stress
1 1
= w
d
(1);
orp =
w = ~j-
pt
( 2 );
or
w
Yp
(3)
In other words, the unit longitudinal stress in the walls of a thin
212
cylinder
is
MACHINE DESIGN
equal to the product of the diameter into the unit inter-
nal pressure, divided by twice the thickness of the cylinder walls,
and
is
independent of the length of the cylinder.
Fig. 66.
Fig. 67.
If a transverse section of the cylinder (Fig. 67)

it
will
be seen that the
total pressure
be considered, on the head, which tends

.
~d2 w
-,
to
this
cause rupture along
transverse
section,
is
and
must be equal to the intensity of the transverse duced multiplied by the area of the metal in such a
stress pro-
section, or,
= "dtp,
-(Pw
wd A = 4*
to
(4)
w = 4p -Tt
t
.
(5)
or
wd 4p
t
(6)
A comparison of
reason
it
(1)
and
(4)
shows the
practice to
stress in transverse sections
be only one-half of that in longitudinal

is
sections.
For
this
very
common
make
the circumferential
seams
of a boiler shell with a
single
riveted joint,
when
the
longitudinal seams are double or triple riveted.
* In this discussion the mutual interaction of the longitudinal and transverse

stresses is neglected.
laterally as
if
If
a tensile stress p t
7ip t
is
induced in a body, the body contracts
acted upon by a stress

is
of pt where a
edition,
Poisson's ratio.
acting at right angles to the line of action (See Merriman's " Mechanics of Materials," 1906
pi in the
page 359.)
Therefore the true longitudinal stress

is
above case
(since A equals 3 for steel)

P\
wd
(1)
wd
This gives a lower value than equation

safetv.
and hence the
latter is
on the
side of

TUBES, PIPES, CYLINDERS, FLUES,
79.
AND THIN PLATES
213
Thin Spheres.
Since
all
the meridian sections of a sphere
are the
same
it
as the transverse section of a cylinder of equal
diameter, given by
shell,
is
evident that the stress in the walls of a sphere

If spherical
is
(4).
heads, of the
is
same thickness
to
as the
are placed on a cylinder which
withstand internal
stress
pressure, they will be subjected to a
maximum
equal to the
transverse stress in the shell.

80. Resistance of Non-Circular
Thin Cylinders
to
Internal
Pressure.
Suppose a cylinder
to
have a cross-section made up of

the upper half as a free
circular arcs as in Fig. 68. (section along the
Take
axis).
body
major
Let the resultants of the com-
ponents of pressure which are normal to the plane of the section
be
TFj,
W W
2i
3,
for the portion
marked
I,
II, III, respectively.
Then
these resultant forces per unit of length of the cylinder are
as follows:
wr //y
sui
<f
<f
u> r
cos
cos ?')
= wm
Wn -
wR
r w
sin S
dO = wRicosO' +
cos 6")
=wm
IV
- w W* iv 3
J
1
sin o
f/
<pd*p
= wr (
3
cos -
+
2
cos
,f
<p
7YI-,
Therefore
W
if
= w (m +
l
m +
to
nt 3 )
= wA
axis,
In a similar way,
the section
is
taken along the minor

is
the resultant force normal to this axis
found
be wB.
is
In like
manner
by the
force
the resultant force
normal
to
any section
(per unit of
length of cylinder) equal to the intensity of pressure multiplied

axis of that section.
is
As
is less
than A, the resultant
wB
less
than
uA;
or the force tending to elongate the
minor axis
axis.
is
greater than the force tending to elongate the
major
If the
tube were perfectly
flexible, its
form of cross-section
all
would become, under pressure, one

or circular.
in
which
axes are equal,
A
it
rigid material offers resistance to

is
such change of
to the direct
torm, and a flexural stress

tension, but
produced in addition
to the circular
approaches nearer
sure increases.
circular cylinder
The
existence of this flexural
form as the presstress in a nonof Figs. 69
becomes apparent from a comparison
214
and
all
MACHINE DESIGN
70.
In Fig. 69 (circular section) the lines of normal pressure pass through a single point (the centre of the circle) the re;
and P 2) also passes through this same point, hence these forces form a concurrent system, and they are in equilibrium. In Fig. 70, however, the pressures do not in themselves form a concurrent, nor parallel, system of forces,
sultant (P r ) of the tensions (P 1
hence they cannot be balanced by a single force
(as the resultant
but there must be a moment, or moments, of stress for equi-
librium.
similar course of reasoning could be applied to a

for such a section
cylinder of any non-circular cross-section;

(Fig. 71)
could be considered as
made up
of circular arcs, each of
which could be treated
(like the special case of Fig. 68)
by
inte-
FlG.
Fig. 71.
grating between proper limits.
A direct inspection
will also
show
that in any non-circular section cylinder, subjected to internal

pressure, the pressure tends to reduce the cylinder to a circular
cross-section.
Suppose the original cylinder

its
(Fig. 71) to
flat
be cut
along the greatest axis of
cross-section,
and that a
it,
bottom
lower
total
coinciding with this section-plane be secured to

portion of the cylinder being
entirely
the
removed.
The
pressure on this bottom evidently balances the components of

the pressure on the curved surface which
lie
normally to
this flat
bottom; hence, the resultant of these normal components of pressure equals
(a
a)
=w A,
per unit of length of cylinder.
In a similar way, the resultant of components of pressure acting
normally to any other section
(as
b,
Fig. 71)
equals
AND THIN PLATES
215
(b
b)
=w B < w A
This direct method might have

re-
been used in the preceding cases (Figs. 66 and 68) without

course to the calculus.
It is
apparent, then, that any cylinder under internal pressure
tends to assume a circular cross-section.

circular section, but departing
cylinder of nominal
to
from the true form
some
extent,
tends to correct this departure under internal pressure;

circular cylinder ternal force,
it
or
if
under internal pressure

its
is
deformed by any ex-
tends to resume
circular shape.
is in
Thus
a circular
If
cylinder under internal pressure

the section
is
"stable equilibrium."
is
other than a true circle there
a flexural
stress, as
well as tension,
when under
pressure.
.
RESISTANCE OF THIN CYLINDERS TO EXTERNAL PRESSURE

81. Theoretical Considerations.
If a thin
hollow cylinder of
it is
circular section
is
subjected to an external pressure,
obvious
that a course of reasoning similar to that in Art. 78 will that a compressive stress
is
show
induced in the walls of the cylinder,
the value of which will be given
by formula
(1) Art.
78 or
p
If the cylinder
wd
where p
is
a compressive stress.
were perfectly cylindrical, of uniform thickness,

material, there seems to be
and
of
homogeneous
no reason why
failure should
occur until the compressive stress reaches the yield
point of the material.
form, uniform in
But tubes are never absolutely circular in thickness, or homogeneous in character; and
is
hence failure occurs long before the compressive yield point

reached.
collapse,
tube which
fails
under external pressure
is
said to
and the forms of collapsed tubes are very characteristic.
Fig. 72
shows the form of cross-section

that the
of collapsed tubes,
and
Unwin* has shown

the
number
of lobes
depends on the
to the
ratio of length to diameter, the smaller this ratio the greater being
number
of lobes.
This peculiarity
is
undoubtedly due
influence of the heads placed in the ends.
For values of
See
" Elements of Machine Design," page 101, 1901 edition.
2l6
MACHINE DESIGN
6,
greater than about 4 to

c
only the forms of collapse shown at
and
d, Fig. 72,
appear.
If the non-circular cylinders of either Fig.
68 or 71 be con-
sidered as subjected to external pressure, the force tending to

increase the major axis will be seen to be greater than that tend-
ing to increase the minor axis;
hence the external pressure

In a cylinder of nominal
will
is
cause
collapse,
unless the flexural rigidity of the material
sufficient to prevent this action.
cir-
cular section any departure from the ideal section will be increased
by the external pressure. Or, if a cylinder of true circular section is deformed in any way while under external pressure,
this pressure will
tend
still
further to increase the deformation.
Fig. 72.
Fig. 73.
In other words, a cylinder under external pressure

equilibrium."
is
in " unstable
As
perfectly true circular sections
and homogene-
ous materials are not attainable in practice the danger of collapse
must be taken into consideration

to
in designing pipes, tubes, or flues
withstand external fluid pressure.

Since the wall of an ideal thin tube
is
subjected to a uniform
compressive
stress, it
may be
considered as being in the same
condition as a long column; and theoretical equations expressing

the relation between the external pressure, the stress,
and the
In
dimensions of the tube have been developed on

view of the fact that the theory of long columns
satisfactory,
it is
this basis.
is itself
most un-
not surprising that such equations do not accord
with
actual results,
and they may be

pressure
is
safely disregarded, but the
analogy between long compression

jected
to
members and tubes subOther deductions

while throwing some light
external
instructive.
based upon the theory of
elasticity,
AND THIN PLATES
21
on the form of rational equations expressing these

not as yet applicable to practical problems.
82.
relations, are
Long Tubes,
Pipes, etc.
Until
very recently
the only
experimental results on the collapse of thin tubes were those due

to Sir
William Fairbairn, who, in 1858, made a
series of careful
experiments on short tubes and deduced therefrom the following

formula:
W = 9,675,600
where
(i)
is
the unit external collapsing pressure in

/, /,
square inch and
pounds per and d are the thickness, length, and outside

Fairbairn himself modified this
diameter respectively in inches.
equation, for simplicity, to the form
.w = 9,675,600
(2)
Many other equations

Thus Professor Unwin
have been deduced from the experiments of
Fairbairn, usually of the
same form but with
different exponents.
gives the following as the result of a careful
resume of Fairbairn's work For tubes with a longitudinal lap-joint

2.21
w =
7,363,000
j-^-^
(3)
For tubes with a longitudinal butt-joint
w =
9,614,000 7n jii6
like
(4)
For tubes with longitudinal and cross joints

boiler flue
2-35
an ordinary
w =
15,547,000
^^
is
(5)
Other writers have deduced similar equations from the same data. Fairbairn's experiments were conducted with tubes whose
lengths were small compared to their diameters.
the effect of the supporting action of the head
his equations
In such tubes
noticeable;
hence
make
the allowable pressure vary inversely as
some
if
function of the length.
Now
it
is
reasonable to suppose that
2l8
MACHINE DESIGN
the tube were long enough the head would have no effect, except
near the ends, and the collapsing pressure would be independent

of the length.
In a similar
way
if
the tube were very short, the

intensity of
walls should theoretically yield
by crushing, and the
the compressive stress
would be given by formula
(i) or,
iv
In
1906 Professor A. P.
experiments
Carman published*
at the
the results of a
set of
made
of
Engineering
Experiment Staconclusively
tion of
the University
Illinois,
which
prove
that Fairbairn's equations
hold
only for tubes whose lengths
are from four to six times their diameters; and that beyond that ratio the collapsing pressure is independent of
the
length.
He
found
that
the
results
of
his
experi-
ments could not be well expressed by a

devised two equations to cover the range
pressing the relation which exists between
;
single
equation, but
these equations ex-
and
a
Thus
for
values of
-V025, and
t\
length
greater
than 4 to 6 times the
diameter, he gives
w=

(t
where k and
c are constants to
be determined experimentally and depending upon the material.
For brass tubes
w=
93.3 6 5
steel
~ ~
d
A74
(6)
For seamless drawn cold
w=
For lap- welded
steel
95,52095>520
2,090
.....
(7)
w =
83,270-
1,025
(8)
Professor R. T. Stewart,f in an elaborate set of experiments

* See Bulletin of the University of Illinois Engineering Experiment Station
Vol. Ill, No. 17, June, 1906.
f See Transactions of American Society of Mechanical Engineers, Vol.

1906.
XXVII,

on lap-welded
steel boiler
AND THIN PLATES

for the National
2ig
tubes
made
Tube
Company, found
that for values of
- \
.023
the results of his
work could be expressed by the following:
w =
86,670-
1,386
(9)
which corresponds closely with (8) of Professor Carman's work, showing the accuracy of the experimental work.
For values
results of his
of
- <.o2 5
Professor
Carman found
that
the
work could be expressed by an equation
of the
form
where k' as before is an experimentally determined constant, whose value for thin brass tubes is 25,150,000, and for thin colddrawn seamless steel tubes 50,200,000.
Professor Stewart found that for values of
practically the
below
.023,
or
same
limit as above, his results
were expressed by
.
. .
w =
The
value of
1,000
(1-/1-1,600^)
.023
is
(11)
for
about 600
lbs.,
which corresponds
closely with the
upper limiting value of

less
obtained from (10).
For values of -
than .023, the corresponding values of

(10)
as found by either
or
(11),
do not
differ
materially.
in
Furthermore, tubes in
gineering
which- ^.02
are not
much used
for
en-
work under external
pressure,
and
convenience
therefore equation (10) will be adopted.

83.
Stewart and
Summary of Equations for Long Tubes. The works of Carman deal entirely with tubes which are so long
is
that the supporting effect of the heads
negligible, or in
which the
length
is
at least four
times the diameter.
Their experiments, while
220
MACHINE DESIGN
conducted separately, supplement and corroborate each other.
As given above, the equations are not in the most convenient form for use by the designer, since usually /, d and w are known and t is required. Transposing these equations, therefore, they
may be
written as follows:
of
les For values
^ .023 and
pressures less than 600 equation
10 becomes
3
Vt
'
and 50,200,000
steel tubes.
<">
for thin
where
= 25,150,000
for thin brass tubes,
cold-drawn seamless tubes or lap-welded
For values
of
.023
and pressures greater than 600
lbs.,
equation 6 becomes
1
where
"
for brass tubes
....
. . . .
r~
d(w +
c)
(I 3>
= 93,365 and
= 2,474
" seamless cold-
drawn
"
steel
" lap-welded steel
= 95,520 and c = 2,090 = 83,270 and c = 1,025
The
following approximate formula, which covers practically
the whole range of values of
is
suggested by Professor Car-
man
as useful in
making rough
calculations.
2
I
w =
where k" = 1,000,000
for lap-welded
steel
T //
"(V
this,
<*3a)
for
cold-drawn seamless tubes and 1,250,000
tubes.
From
d \wc
the following usually
more convenient formula can be derived:

3b)
tube 4 inches outside
Example.
lap-welded
steel boiler
diameter and 10
feet long, is subjected to
an external pressure of
the thickness be in
?
300 pounds per square inch.
What must
order to have a factor of safety of at least 6
AND THIN PLATES

is
221
Here the assumed collapsing pressure

300 X 6 = 1 ,800
lbs.
per square inch.
Applying equation
/
(13)
d (w
c)
= 4(1,800+1,025) =
83,270
.14 inch.
Here the
ratio
- =
a
.14
-35>
an d hence equation
(13) applies.
4
less
In case this ratio should be
than .023, which will seldom
occur, a second solution, using equation 12, should be made.

84. Short Cylinders, Flues, etc.
When
the cylinder or flue
is
short,
i.e.,
~;\4
to 6, the effect of the
heads should not be neg
lected as in
Carman's and Stewart's work, and Fairbairn's apis
proximate equation
applicable, or
t
2
w =
or transposing
9,675,600^
(14)
t= \_wl_
Xl
9,675,600
If
a cylinder under external pressure could be depended
upon
to
fail
only by actual crushing, instead of
through
collapse
in
(buckling),
then the formula
w=
pt
2
would apply, as
CL
remembering that under external pressure If this equation gives a lower working pressure than (14) the flue designed by it will be safe against The rules of the Lloyd's Marine Register allow the collapse.
internal
pressure;
is
the stress p
compression.
following
pressure
in
boiler
flues:
w=
1,07^,200
=-=
1/
2
t
.
This
is
CL
Fairbairn's equation with a factor of safety of
9.
The
2pt
British
flues
Board
of
Trade
rules allow a
working
stress in furnaces
and
.
of about 4,000
when computed by
the equation
w=
This
222
is
MACHINE DESIGN
little less
than that allowed by the U.
S.
Board
of Supervising
Inspectors.
Hence
for the
same allowable pressure under these two

2
rules
w=
or
If,
I
p
d
8,000
1,075,200
d
t
Id
134.4
t
therefore,
<
134.4
equations
(1), (2),
and
(3)
may be
safely
used.
It will
be observed that
this relation limits the use of these

flues.
equations to comparatively short
Thus
a flue
%"
thick
could only be 34" long to have these equations applicable.
In
practice long flues of large diameter are reinforced at short intervals
rings, thus
by heavy rings of rolled or other section, known as collapse making the flue consist virtually of a series of short flues, to which equations (1), (2), and (3) may be applied. Various Insurance and Government inspection departments These rules give rules for proportioning flues and furnaces. change from time to time, and if the boiler is to be insured in any company the specific rules prescribed by it should be consulted.
Thus Lloyd's
Register for 1906-7 gives

1,075,200
T
2
t
w =
.
w X
;
when
/)
I.
>i2ot
and
w =
50(300/
a
when
<
120/
where
/, t
and d are
all
in inches.
Various other authorities give similar equations with practically the
same
coefficients.
85. Corrugated
Furnace Flues.
stiffer
Flues corrugated as in Fig.
73 are very
flues,
much
against collapse than plain cylindrical
made
and with proper dimensions of corrugations may be safely of any desired length. Their peculiar shape also permits
of expansion
and contraction under the influence

1
of heat.
When
the corrugations are not less than
inches deep, and not more
than 8 inches from centre
to centre of corrugations,
and plain
AND THIN PLATES
223
S. I.
portions at the ends do not exceed 9 inches, the U. S. B.
allows a working pressure of

:
"
14,000
~^r~
Board
of
This
is
also the formula of the British
Trade.
Lloyd's Register for 1907-8 gives a

ing various types of flues.
number
of rules for design-
The
tion
following references contain valuable practical informathis subject
on
Lloyd's Register of British and Foreign Shipping.

"
Steam
Boilers,"
by Peabody and
S.
Miller.
Rules and Regulations of U.

spectors.
Board
of
Supervising In-
Rules and Regulations of the American Bureau of Shipping.
Rules of the British Board of Trade.

Rules of the Bureau Veritas.
Seaton and Rounthwaite's Pocket Book.
THICK CYLINDERS
86.
When
the wall
of
cylinder,
is
which
thick
is
subjected to
to
internal or
external fluid
it
pressure,
relatively
the
internal diameter,
in
it
can no longer be assumed that the stress
the wall
is
is
uniformly distributed over the cross-section, but
greater at the inner surface
and decreases
is
is tensile,
to a
minimum
at the outer surface
whether the pressure

is
internal or external.
When
the pressure
is
internal the stress

is
and when the
pressure
external the stress
compressive.
Many
of
formulae have been deduced to express the relations
between pressure,
stress, and cylinder thickness. Of these, that Lame, deduced in 1833, s perhaps best known. Clavarino's* modification of Lame's formula, which was published in 1880, is
*
now much
used and will be adopted in this work.

is
Ordinarily the cylinder

internal pressure alone;
subjected to either external or
but in a gun tube, for example, which
The
student
is
advised to read the discussion of thick cylinders given in Merri-
man's "Mechanics of Materials," edition of 1906.
224
MACHINE DESIGN
it,
has a hoop shrunk upon

the cylinder
is
the
more general case occurs
in
which
subjected to both internal and external pressure.
Let
"
w = the internal unit pressure. " w = the external "

1
" r 1
" r 2
"
"
= the internal radius of the cylinder. = the external " p = the unit stress at the inner surface. " " " " outer p 2 = the
1
Then by
Clavarino's equation the unit stress at any radius r
is
(w l
- w )~]
2
.
(16)
If the external pressure
w
1
be zero, which
is
the most usual
case, the greatest tensile stress is at the inner surface,
and
is
''"TLtf^J
or fi
w r r* +
4r2
-\
(I7)
,t
rî-i*
L3P1
4î J
....
(l8)
Example.
is
cast-iron cylinder 20 inches in internal diameter,
to withstand
How
an internal pressure of 1,000 lbs. per square inch. thick must the wall be in order that the stress at the inner
surface
may
r
not exceed 4,000
lbs.
per square inch
Here
in (18)
=10,
w = 1,000
1
and p x = 4,000.
Hence
substituting
rjLAJ^q i
L3
A ~4 îJ
it is
IO
3x4,000+1,000 1 = i2g L 3 X 4,000 4 X 1, 000 J
or the cylinder walls must be 2.8" thick.
From
(16)
found that p 2 the

,
stress in the cylinder walls
at the outer fibre, is 2,620 lbs.
PRACTICAL CONSIDERATIONS
87. Cast-iron pipes are widely used for
underground water
on account
or
for
pipes and to
of
their
some extent
also for gas pipes, largely
durability
against
corrosion.
For
steam,
high pressures generally, cast-iron pipes are
now seldom used

because
pipes
of
their
AND THIN PLATES

all
225
unreliability.
For
made
of
wrought iron or
steel are
ordinary purposes most used, although
in special cases,
such as marine work, copper and brass are
preferred.
Wrought-iron or
while steel
steel
pipes
may be
either lap-welded or butt-
welded, the latter being
which case
commonly used for the smaller diameters, piping may be "drawn" so that there is no seam, in it is known as "seamless drawn tubing."
its
Standard Piping is designated by
nominal internal diameter.

of 1.3 15 inches.
Thus standard
of
1
i-inch gas pipe has a nominal internal diameter
inch,
and an external diameter
So-called
standard wrought-iron piping

100
lbs.
may
be used for pressures up to
with safety.
For
still
higher pressures, such as are found
in high-class
steam plants, thicker pipes, known as extra strong,

to several
are used.
For hydraulic work, where pressures up

lbs.
thousand
piping,
per square inch must be withstood,

strong,
is
still
thicker
are
known as double extra made by decreasing the
used.
These heavy pipes
internal diameter of the standard
pipe, thus keeping the outside diameter and hence the screw
threads for the flanges to one standard.*

i-inch pipe (nominal size)
.95 inches,
Thus an extra strong would have an internal diameter of

of the
and a double extra strong

inches in
for
same nominal
size
would have an internal diameter

diameter remaining
1.3 15
of .587
inches,
the external
all cases.
For large cylinders both

cast iron is
yet,
still
steam and hydraulic

will
service,
much used and probably
be for some time
on account of the ease with which complicated iron castings

In the case of steam-engine cylinis
can be made and machined.

der.^ the
thickness of the walls
fixed
by considerations other
than those of strength, such as stiffness and securing good casting.
The
proportions of steam cylinders as fixed by practice
are the best guide.
An examination
of current practice
shows
the average thickness of low-speed engines to be given
by the
The
Sttldenl
is
referred to Kent'-, " Engineer's Pocket

sizes of pipes,
Book," or similar works,

Sec
also
for full
tables of standard
flanges, etc.
current trade
catalogue-.
5
226
following,
in inches,
sq. inch.
t
MACHINE DESIGN
o$d + .3 inch,* where
= thickness and
= diameter
when
the steam pressure does not exceed 125 lbs. per
Kent's " Mechanical Engineer's Pocket Book" gives the following as representing current practice,
= .0004^ + .3,
where d =
diameter in inches and p = pressure in pounds per square inch. If p be taken as 125 pounds this equation reduces to that given
by Barr.
Cast iron
is
also
much used
for the cylinders of hydraulic
machines, although
steel castings are better in general.
In such
with
all
cases equations (16) to (18) developed above, in
common
equations based on the theory of

caution
elasticity,
should be used with

Further-
when
cast iron
is
selected for the cylinder.
more it must be borne in mind that the thicker the cylinder walls, the more liable are they to be porous in the interior, where made
of castings.
It is safer, therefore, as
a rule, to carry a high work-
ing stress, within safe limits, and insure sound castings, than to
design thick walls which are open to suspicion, in order to get a

theoretically lower stress.
3-inch wall, for instance, with a

is
working
stress of 5,000
pounds per square inch
preferable to a
4-inch wall with a working stress of 3,000 pounds per square inch.
Care should also be exercised in cylinders made of castings to

avoid excessive thickness of metal at any point, thus insuring
Thick castings of any metal are very liable to give trouble by leaking on account of porosity, if subjected to high pressures, and cast-iron cylinders are often lined with brass or
sound
castings.
bronze liners to obviate
this difficulty.
etc.
87.1. Pipe Couplings, Flanges,
Methods
for
securing
the ends of pipes together have
become
of greater importance
as higher steam pressures have been employed.
The most
usual
method
for
accomplishing this purpose has been to thread the

(see Art. 57)
ends of the pipes

flanges.
and secure them together with
either a cylindrical pipe coupling, a pipe union, or a pair of pipe
All of these are in very
up
to 100
common use. For pressures pounds per square inch and pipes not over 12 inches in
J.
* See "Current Practice in Engine Proportions," by

tions A. S.
H. Barr; Transac-
M.
E., Vol.
XVIII.

diameter these
AND THIN PLATES
227
may
be used with success, but for higher pressures

satisfactory.
and larger diameters they are not so is used on small pipes only.
In the ordinary screwed
cut the thread accurately,
The union
difficult to
fitting of large size
it is
and
to
screw
the fitting on
tight
leakage at A, Fig. 73 (a). This can be remedied to some extent by making the threaded portion of the pipe
enough
to prevent
long enough to project through the flange slightly, and then

facing off pipe
and flange
at B, Fig.
so as to
make
Even
a smooth surface, and
permitting the packing or gasket (P) to cover up the screwed

joint, as
shown
if
73 (a).
is
this joint,
if
however,
is
liable to leak
the
workmanship
poor or
the flanges
do not
align properly.
To
obviate the difficulties of the screwed joint on pipe of
Fig. 73 (a).
Fig. 73 (b).
Fig. 73
(c).
larger diameter, the flanges are

in
sometimes shrunk on as shown

In order to insure tightness,
flange, the
Fig. 73 (b)
(see also Art. 73).
and secure a firmer grip on the usually expanded into the flange,
flange.
end of the pipe
is
as
shown
in Fig. 73 (b).
The
gasket usually covers up the joint between the pipe and the
This form of coupling

if
is
not well suited for high-pressure

is
work, however, especially

side
the pipe
not machined on the out-
before the
flange
is
shrunk on.
to
When
subjected to the
heavy straining action incident
expansion and contraction,

is
as in heavy steam mains, the pipe
and leaking
a recess,
will ensue.
(Fig. 73 b),
sure to work in the flange These flanges are sometimes fitted with into which a strip of soft metal, such as
228
MACHINE DESIGN
but this can hardly
copper, can be caulked to check small leaks;
be considered
satisfactory in high-grade
work.
A
by
somewhat
rolling the
better grip of the flange is sometimes obtained
pipe into a groove in the flange, as shown at E, Fig. 73 (b) In the so-called Van Stone joint (Fig. 73 c), the ends of
the pipes themselves are flanged over the flanges so formed.
and the
joint
made between
The
flanges F,
holding the flanges proper together,
become clamps for and may be loose on the pipe.

it
This
last feature is
a very useful one, as

joint
greatly facilitates
erection.
This form of
has been used with success in high-
pressure work.
For the highest grade

welded
to the pipe,
of
work, wrought-steel flanges are
making
the pipe
is
and
its
flanges one piece.
This construction, while expensive,

pipe and the highest pressures.
almost essential for large
To
(b).
prevent the packing from blowing out, the flanges are

fitted
sometimes
with a recess and tongue as shown at

is
H,
Fig. 73
This construction
essential for very high pressures, as

if
in hydraulic work, but should be avoided

lines, as
it
possible in steam
makes
it
difficult to
renew the packing.*

to establish
Although several
in this country.
efforts
have been made
standard
use
dimensions for pipe flanges, several systems are in
common
The most important
of these systems are the
standards adopted by the Flange Standardization Committee,

of the A. S.
M. E., and that known
as the Manufacturers' Standard.
The The
first is
and the second

student
is
used for pressures up to 125 pounds per square inch for pressures up to 250 pounds per square inch.
referred to standard handbooks,
and the catalogues
of various manufacturers for details of these several systems.
See also Transactions A.
S.
M.
E., Vol.
XXI.
88. General Theory.
THIN PLATES The theory of
the stresses induced in

is
thin plates,
when
subjected to load pressures,
one of the most
uncertain portions of the mechanics of materials, due in part to

* For a fuller description of
Van
2,
Stone joints see articles by

1909.
W.
F. Fischer,
Power, Feb.
23, 1909,
and March

the complexity of the problem,
AND THIN PLATES
229
and
in part to the scarcity of
corroborative experimental data.
The
subject has been investi-
gated mathematically by Grashof, Bach, Unwin, Merriman, and

others,
verities
and the experimental work of Bach, Benjamin, and others in a measure some of their conclusions. The mathematical by various
to
results obtained
efficients, the
authorities differ mainly in the co-
general form of the equations being in most cases
similar.
Those due
Merriman*
will
be used in this treatise

safe, generally,
as they are simple, easy to apply,

as any others.
and give values as
Let
the thickness of the plate in inches;

/
of circular plate in inches;
angular plate
per unit area
r = the radius = length and b = breadth of rectin inches; p = maximum tensile stress; w = load in lbs. and P = concentrated load in lbs. Then
;
For
flat
circular plates supported but not fixed at the edges,
carrying a distributed load
for \ >
p
*
\w
wrought iron or
steel
(1)
and
For
r \|
9 w for cast iron o
(2)
flat circular plates fixed at
the edges (encastre), carrying
a distributed load
t
2
r \l
w
for
wrought iron or
steel
....
(3)
and
For
w
for cast iron
(4)
Mi'
flat
circular plates supported but not fixed at the edges,

is
carrying a concentrated load P, which
applied to the centre
of the plate over a small circle of radius ro so that
P=T
.
r
{
=
=
1
1
r v'
r r
-1
for steel or
wrought iron
(5)
and
- + - log
"
w for cast iron ...

-1
(6)
See
Merriman's
Mechanics
of Materials," 1907 edition,
page 409.
230
For
MACHINE DESIGN
flat circular plates
is
.
encastre carrying a concentrated load
P, which
applied to the centre of the plate over a small circle
of radius r o
t
= =
ro
- + -log
O
for
o
steel or
wrought iron
I O
(7)
and
r -J
T- + - /of "I
(5), (6), (7)
(8)
for cast iron
...
.
(8)
In equations
and
is
the pressure per unit
area on the small circle whose radius
is r
i.e.,
P w =
o
7Z
The
value of
should not exceed the elastic strength of the material.
Example.
a load of 4,000
A
is
circular cast-iron plate 20" in diameter supports
lbs. at its centre,
the load being applied by a bolt
whose head
2" in diameter.
How
thick must the plate be,

it
if
simply supported, in order that the tensile strength in exceed 6,000

lbs.
shall not
per square inch?
Here
P =
.
4,000; r
10; ro
=
X
and p
;
6,000.
W =
n
P
n
r
7, 2
4,000
I,26S ^ lbs.
and
10 =
2.3
whence from equation

-1
(6)
-V[l
Rectangular Plates.
1,240-
4
If 2Z
6,000
and 2b are the length and breadth

supported but not
per square inch
fixed,
of a rectangular plate, then for plates
and
for a uniformly distributed load
**\
and
for fixed edges
/i.
,~m,.
lb^-
W 2(?Tb
r
3
)P

If
/
AND THIN PLATES
231
equation
(9)
reduces to
'
w
/
\lf 8 ^ p
and equation
(10) reduces to
'-'^J
The above
implicitly, but will serve as
(I2)
equations, as before stated, are not to be relied on
approximate guides only.
This
is
is
particularly true in cast materials
where heavy ribbing

of the
used
and where trained judgment

89. Flat Stayed Surfaces.
is
perhaps the best guide.
One
most important cases

flat
of flat plates occurs in boiler work,
where large
areas are held
against pressure
by stays
at regular intervals over the surface.
These stays are usually screwed into the plate and the projecting end is slightly riveted over to insure steam tightness. The various
Inspection
Bureaus and Insurance Companies give practical
formulae for the design of such plates, and these can be safely
Thus the U. S. Board of Supervising Inspectors* and the used. American Boiler Makers' Association rules give for steel plates
w = -7Where
Cxf
t
3)
w = pressure
in lbs. per sq. in.,
thickness
of plate in
sixteenths of an inch, s
= greatest
thick
pitch of stays in inches,
and
C=
a constant as below given
C = 1 1 2 for C = i2o " C = 140 " C=i6o "
plates
y
"
and under.
"
over T thick. with stays having a nut inside and outside.

" "
"
washers
.5
.5
as thick as
the plate and of a diameter at least

* See General Rules
the greatest pitch.
and Regulations of U.
S.
Supervising Inspectors; also
Rules of American Bureau of Shipping.
CHAPTER X
CONSTRAINING SURFACES
machine must move with

retained
in
definite relative motion, they
As the various members of a must be
by constraining surfaces. Thus a shaft is held in position by bearings which locate its axis The of rotation, and by collars which prevent motion endwise.
correct
position
relative
motion of a pair of constrained members
may be
its
that of
sliding, as in the case of an engine crosshead and
guide;
rotation, as in the case of a shaft journal
and
its
bearing; rolling,
as in roller
and
ball bearings;
or a combination of
some
of these
as illustrated in certain forms of cams, where both sliding

rolling exist.
and
Dry
metallic surfaces, under
any appreciable load,
even when smoothly machined, will not slide over each other
without abrasion.
surfaces separated
It
is
therefore necessary to keep rubbing
by a
thin film of
some kind
of lubricant,
is
and
the whole subject of the design of constraining surfaces
closely
connected with the theory of lubrication.*

It
has been pointed out in Chapter IV, that when bath or

is
forced lubrication
surfaces
is
maintained, the friction between two rubbing
independent of the character of the material of which
the surfaces are composed; but
when
shown
on
the surfaces are "imper-
fectly" lubricated the frictional resistance
depends somewhat on
the metals used. Experience has
that like metals usually
do
not rub together well.

steel
Thus
steel
steel (except
when hardened),
poor combina-
on wrought
iron, or cast iron
on cast
iron, are
light. If two rubbing surfaces of cast iron can be run together for some time without cutting, they take on hard glazed surfaces which will run
tions except
where the velocity is low and the pressure
well together. This
is
well illustrated in slide valves
and pistons of
* See Chapter IV.
232
steam engines.
lubricated
will
233
when
Care must be exercised that the surfaces are well Soft steel and wrought iron first put in service.
steel,
both run well on hardened

steel at
and hardened
Steel
steel
mayif
be run on hardened
very high pressures and velocities,
the surfaces are ground true, and polished.

iron will run very well
tin, zinc,
and wrought
on brass or bronze.
etc.,
antimony, lead,
The alloys of copper, commonly known as anti-friction

steel
or babbitt metals, run extremely well with iron

the
or
wroughtare
of
journals.*
Innumerable
alloys
of
this
kind
market under different names.
They can be made
upon any
degree of hardness, depending largely upon the proportion of
antimony used.
Very hard
alloys of this kind are
sometimes
known
is
as white brass.
In using babbitt metal for heavy pres-
sures, care should
be exercised that the particular alloy selected

to flow
hard enough so as not
under the applied pressure.

influence
Other materials, such as wood, are sometimes used for rubbing

surfaces.
The
conditions
which
the
selection
of
materials for rubbing surfaces, and the practical considerations
governing their application, will be more fully discussed in connection with the several forms of constraining surfaces.
The most common forms

translation
of
motion in machines are rectilinear
and
rotation;
therefore the most important forms of
constraining surfaces are

(a)
Sliding
surfaces,
for
the
constrainment of rectilinear
motion.
(b)
Journals and bearings, for the constrainment of motion

of rotation.
SLIDING SURFACES
91.
Forms
of Sliding Pairs.
The
stationary
member
of a pair
of surfaces,
which have
relative sliding motion, is usually called
the guide, while the
moving part has various names depending
on the
service, as the
ram
will
of a shaping machine, the table of a
planing machine, or the crosshead of an engine.
The
general
term sliding
* Sec Kent's
member
be used here to denote the moving
" Mechanical Engineer's Pocket Book" for detailed analysis and

of the best
properties of
some
known
alloys.
234
member.
(a)
MACHINE DESIGN
Sliding pairs
may be
classified
by the degree
of lateral
constrainment afforded the slider by the guides, and this

Partial lateral constrainment.
may be
(b)
Complete
lateral constrainment.
In either case the rubbing surfaces of the guide and sliding
member may be
either square, angular, or circular.
Thus
Fig.
74 (a) chines while Fig. 74 (b) shows a set of square guides for a similar purpose. In each case the lateral constrainment is only partial,
the tendency of the platen to raise being resisted
Fig. 75 (a)
shows a form of angular guide
much used on
planing ma-
by
gravity.
shows the crosshead of a steam engine with an angular

is complete. Fig. 75 (b) is crosshead with circular guiding surfaces.
guide.
Here, lateral constrainment
also a steam-engine
This form of surface may be considered as a special form of the
Z5~
Fig. 74
(a).
Fig. 74
(b).
angular type.
If the circular
is
;
guiding surfaces have a
common
centre at O, the crosshead
prevented from rotating around
only by the connecting-rod
and as long as
is
it is
it is
so held
from
rotating the lateral constrainment
complete.
If the surfaces
have different centres as

take place.
Figs. 76 (a)
2,
obvious that rotation cannot
and 76
is
(b)
show square and angular
guides where constrainment
complete.
The
characteristic
is
which distinguishes the square guide from

for wear, while in the case of the
the angular one
that in the square guide two sets of adjustments
must be made
to
compensate
angular guide one set only is needed.
Thus
in Fig. 76 (a), vertical
by lowering the piece A, while lateral wear is taken up by the set screws C which press against the wearing strip or gib B. In Fig. 76 (b) lateral and vertical wear are both compensated for by the set screws C which press upon the gib D. Sometimes D is made tapering and provided with a
for
wear must be compensated
screw adjustment so that
pensating for wear.
it
35
can be moved endwise, thus comset
In such cases the
screws
are omitted.
As
faces,
to the relative merits of

it
square and angular guiding sursquare surfaces are easier to
may be
fit
said, in general, that
machine and
angular guide
than the angular ones.
There are many
places,
however, such as the cross slides of lathe carriages, where the
much more convenient. In places such as lathe commonly used have the advantage of autoV matically taking up lost motion, no matter how badly they are
is
beds the
worn.
guides
But, as a rule, the bearing surfaces of such
guides are
very small and wear soon begins to be apparent, especially as the
wear from the carriage
is
usually concentrated on a short por-
C
1>
T7^X-D
rm
/
(b)
C
Ê
ll
P-1M
(a)
Fig. 75.
Fig. 76.
tion of the bed.
There
is
a tendency
flat
among manufacturers
to
discard the
guide in favor of
is
surfaces.
English practice,
in this
especially in large tools,

particular.
in
advance of American practice
combination of
Principles.
V
a
and
flat
guides
is
also often
used.
92. General
If
short
block,
Fig.
77,
slides
backward and forward upon another member B, carrying a fixed load P, it is evident that, if the material in A and B were homogeneous and the velocity were uniform throughout the stroke,
the frictional resistance
and consequent wear would be practically

in practice.
uniform over the whole surface of B.

difficult to attain
and seldom occur
These conditions are Since A must

it
be stopped and started at each end of the stroke,

the velocity cannot be uniform; although in
as plate planers this condition
is
follows that
some machines such Usually, howapproximated.
236
MACHINE DESIGN
from zero
at the
ever, tfie velocity varies
beginning to a
maximum
somewhere near the middle

greatly.
of the stroke, as in the case of engine
crossheads, shaping machines, etc.
Again, the load
Thus
in the
steam engine, the normal pressure

its
P may vary P be-
tween the crosshead and
guide
is
zero at each end of the stroke
and a
maximum
near mid-stroke.
at
The
velocity of the crosshead

to a
also varies
from zero
each end of the stroke
maximum
near mid-stroke.
In the ordinary case the greatest frictional
resistance* and wear will occur near mid-stroke, because both

velocity
and normal pressure between the bearing surfaces are

If the
greatest at this position.
crosshead could be
made
the
same length
of the stroke,
as the guide, the unit bearing pressure, at the
middle
would be practically uniform over the whole surface, and would be small compared to the unit normal pressure attained when the crosshead is short. For positions of the crosshead near
mid-stroke the wear would be approximately equal over the whole
surface,
and much
less
than when the crosshead
is
very short,
but
still
theoretically greater than at the
end positions, when both

It
has been found by and guide are made the same length, the wear, even under varying load and velocity, is very small, and more uniform over the entire contact
velocity
and normal pressure are
zero.
experience that
when
the
sliding block
surfaces.
It is
seldom possible, however,

as the guide.
to
make
the sliding
member
rams and the tables of planing machines, the sliding member is, in some machines, shorter than the guide, and in other machines longer. In most cases of this kind the wear is liable to be greater on one part of the guide, or sliding member, than on another. Thus in the case of a shaping machine the ram seldom operates at full stroke, and the wear on the back end
the
same length
Thus
in lathe carriages, the
of shaping machines,
of the
ram
is
very small, the result being that

of the
when appreciable
back end
all,
wear takes place on the forward end
ram, and the guides

of the
are readjusted to compensate for the same, the
ram
will
not
pass
through
the
guides
at
hence
the
* See Art. 32.
adjustment
2 37
must
be
somewhat
slack,
and
accurate
work
cannot be done.
has corrected this
built,
In other machines the excessive local wear

result occurs.
comes on the guide, and a similar
Professor Sweet*
difficulty, in certain
machines which he has
by reducing the wearing surface on that portion of the sliding member or guide, as the case may be, where the tendency to
the following convenient method on the surface of a sliding member. Fig. 78 shows a sliding surface such as is found on the ram of a shaping machine, where little wear occurs on the back or right-
wear
is least.
He has suggested
of laying out the wearing strips
hand end
as here shown.
The shaded
portions represent the
parts of the surface which have been relieved, leaving the wearing
p
\
I
^-^F
Fig. 77.
a^
S// Sl
sT^
Fig.
7!
-4
(a)
Fig. 79.
draw the diagonal From a draw the line ac, a, b across the surface to be relieved. making any convenient angle with the horizontal. Lay off ce equal to the width of the face x. Draw de parallel to ac and take
strips S,
1
S and S
2,
etc.
To
lay off the surface,
the vertical distance above the point of intersection of ab and de

for the first gap,

first
vertical distance
strip,
below
the point of intersection for the
wearing
repeating this
operation to the end of the surface.
Similar wearing strips should

if
it
be cut in the opposite direction on the other member,

comparatively long;
guide, the guide only need be relieved.
93. Bearing Pressures
in
is
but where a short block slides in a long
on Sliding Surfaces.
It is
pointed out
Article 32 that the tendency of a loaded flat surface to expel

is
the lubricant
resisted to a certain degree
by the
viscosity of
* Professor
Sweet has embodied some of

that are Usually
his experience in this line in a little

will well
book
called
"Things
Wrong," which
repay reading.
238
the lubricant,
MACHINE DESIGN
This
in
and its power to adhere to the stationary member. power is much less marked in sliding surfaces than It is rotating surfaces, as here the motion is intermittent.
resisting
difficult therefore to lubricate sliding surfaces as efficiently as
rotating surfaces, and, in general, they must be considered as " imperfectly " lubricated surfaces. The unit bearing pressure
that can be sustained
by
sliding surfaces
is,
therefore,
it is
much
less
than can be borne by rotating journals.
Further,
difficult to
obtain initially true sliding surfaces and, as pointed out above, very difficult to maintain their accuracy under service.
sliding part,
The
be
and
also the guides themselves, should, therefore,
designed for rigidity; in fact considerations of strength seldom
need
to
be considered, but the guides should be so

It is
stiff
that
localized pressure will not occur.
not surprising, in view
of these considerations, that the allowable bearing pressures as
fixed
by practice vary
greatly,
even with similar classes of work.
Owing
it
to the difficulties of lubrication
and compensation
if
for wear,
may be
stated, as a general principle, that the bearing pressure
must be kept
so low that
wear
is
inappreciable,
accurate sur-
faces are to be maintained.
The
different
following are average values of bearing pressures for
forms of sliding surfaces, as fixed by practice

lbs. to
Crossheads,* stationary slow-speed engines 30

high
"
"
"
50
lbs.
marine engines
10 " 50 "
" 30 " " 75 "
94. Lubrication of Sliding Surfaces.

difficult to lubricate efficiently
Sliding surfaces are very
on account of the " wiping" action
member. In high-speed engines, bath lubrication commonly obtained by enclosing the running parts, and allowing them to run in what practically amounts to an oil bath. Where this cannot be done, care must be exercised in the manner in which the lubricant is supplied. If possible, when
of the sliding
is
the guide
is
horizontal, the lubricant should be supplied near
the middle of the guide.

also
The oil
grooves in the moving part should
be given careful consideration.

* See Trans. A. S.
From
the theory of lubrica753.
M.
E., Vol.
XVIII, page

tion
it is
239
evident that the
oil
channels on
all
constraining surfaces
should be at right angles to the direction of motion, wherever the

velocity
If
is
great
enough
to
draw lubricant between

is to
the surfaces.
to
made
otherwise their effect
relieve
any tendency
form
a pressure film.
The
grooves in crossheads, and other sliding
members, should,
in 79 (b).
therefore, be made as in Fig. 79 (a) and not as In either case the grooves should be stopped some
distance from the edge of the surface so as not to facilitate the
escape of the
tion
oil.
When
the load
is
so
heavy that forced lubricaFig. 79 (b)

also
oil
is
correct; the oil being forced in at O.
must be used, the system of grooves shown in Care should
be taken
grooves,
If
that the outer edges of the slider,
and the edges of the
are chamfered so as to assist the entrance of the lubricant.
the edges are square
and sharp
their scraping effect
may
seriously
impair the lubrication.
Where
the guiding surfaces are very
long, as in planing machines, oiling devices such as rollers dipping

in
an
oil
pocket, placed at intervals along the guides, are very
effective.
JOURNALS AND BEARINGS

BEARINGS
Forms of Bearings. The part other member, which constrains a as a shaft, is known as a bearing.
95.
machine frame, or rotating member, such That portion of the

of a
is
rotating journal.
member which engages
with the bearing

all
known
as a
Journals are necessarily circular in
cross-sections,
but their profile
may be
cylindrical, conical, spherical, or
even
(See
more complex
Art. 105.)
in form, as in the case of thrust bearings.
One
(a)
or
more
of the following considerations affect the design
of the bearing proper:
Rigidity, in order that the alignment

affected
may
not be seriously
by
to
deflection.
(b)
(c)
Strength, to resist rupture under the greatest loads.
Adjustment,
compensate
for wear.
oil film,
(d)
(ej
Formation and maintenance of an

Automatic adjustment,
to insure alignment.
240
(a
its
MACHINE DESIGN
and b)
.
The inside diameter, or
bore of the bearing, and also
length are fixed by the dimensions of the journal which engages

it;
with
and the required strength and
rigidity
may be
secured
by a proper
distribution of metal in accordance with the general
principles discussed in Chapter III,

of bearings, as far as strength
and which apply
to all
forms
and
stiffness are
concerned.
Usually the question of strength does not enter into the design
of the
main part
of the bearing.
If,
however, the cap A, Fig. 80,

is
should be called upon to carry the load, as
often the case,
its di-
mensions should, in general, be checked

should be such that
that the cap
stiffness is secured.
is
for strength,
and
its
design
The
exact distribution of
the pressure over a bearing

is
not known;
but the assumption
beam loaded
and
at the centre
and
of a length equal
to the distance
between the cap bolts

strength
will give
dimensions on the
greatest
safe
side
for
deflection.
The
bending
moment and
Table
1.
deflection for such
beams
are given in Case IX,
It is
is
impossible to adjust the cap bolts so as to be sure
that the load
uniformly distributed
among them, and
the un-
certainty of the initial stress due to screwing
up the nuts makes

load.
If
the problem
more
difficult.
For
this
reason the cap bolts should

only two
be designed
load;
to carry
more than the apparent
bolts are used each should be designed for two-thirds of the total
if
four are used each should be able to carry one-third of
the load with an apparent stress of not
more than 6,000
lbs.
per
square inch.
c, d and e, affect the form of the bearing. and d. It is evident that the metal of the bearing will wear away most rapidly in the line of greatest pressure, hence adjustment for wear should also be along this line. It follows
The
last three items,

first c
Consider
also that the bearing should
be parted
if
at right angles to the line
of greatest pressure.
Thus,
the load on the shaft be a simple
vertical load P, as in Fig. 8c,
wear
If this
will take place only
on the
bottom half
of the bearing.
wear
is
so small as not to
if
interfere with the alignment of the shaft, or
all
the bearings
on the shaft wear uniformly, adjustment may be made by lowering the cap A. If the shaft must occupy a fixed position relative to the frame of the machine, alignment must be maintained by
raising the lower bearing surface.
24I
lower wearing surface
is
usually
Where this is desirable the made separate from the pillow-
block, as in Fig. 82, thus allowing the bearing to remain fixed

in position, while the
wearing part
may be
raised to compensate
for the wear.

all
If the
load P, Fig. 80, be in an
upward
direction,
necessary adjustment
It
may be made by means
of the cap.
will auto-
was shown
in Articles 32
and 33 that a journal
matically tend to form a film of lubricant between

bearing.
If the conditions
fluid
itself
is
and the
supplied
under which the lubricant
are correct,
pressure
may
thus be created between the

is
journal and bearing provided the surface of the bearing
con-
tinuous for some distance on each side of the line of action of the
Fig. 80.
load.
It is
The
greatest pressure will be found near this line of action.
evident that the bearing
shown
in Fig.
80
fulfills
both these
requirements for vertical load either upward or downward; but

is
unsuited for lateral pressure from the standpoints both of ad-
justment for wear and lubrication.

Suppose, however, that the journal carries a heavy vertical
load
(Fig. 81),
and
is
subjected at the same time to a heavy
horizontal belt pull P,.

the arrangement of parts
The
resultant of these forces
is
2,
and
shown
will
in Fig. 81 is correct for

If
motion
of rotation in either direction.
P\
:i
be reversed in direction
and the arrangement is not correct for adjustment against wear,' and very defective as far as lubrication is concerned, as the surface is broken near the point
the resultant of
x
and
be
if,
242
MACHINE DESIGN
where the greatest film pressure should exist. Bearings of this form are often used in steam-engine work, and in such cases
the
force
due
to
the
in
steam pressure
on the piston,
adjustment
is
continually
reversed
is
direction.
in Fig. 82.
Another
for
a similar case
shown
Here the shoe or bottom

liners,
"brass" can be raised up by introducing thin "shims," or

underneath
it;
while lateral wear can be taken up by setting out
the "cheek pieces" B,
by means
of the
wedges D.
Provision
is
thus made, by this arrangement, for taking up wear in

tions
all direc-
and keeping the
shaft accurately aligned
and located.
For
horizontal pressures in either direction the resultant

close to the point at
passes
which the bearing
is
parted;
best conditions for lubrication
do not
exist.
and hence the Pressure films more
Fig. 8-
or less perfect, depending on the
oil
supply, will form on the lower
shoe, but the continual reversing of the lateral pressure P, hardly allows time for the formation of pressure films on the cheeks.
These
carried
reversals in pressure, however, allow the lubricant to be
by the
shaft, first
under one cheek, and then under the

effectively.
other, thus lubricating
them
Sometimes a bearing consists
of a conical
bushing
split at
some convenient place, as shown in Fig. 83.
nut A, and screwing up on B, the the frame C, thus closing the bore of the bushing
By releasing the bushing may be forced into

slightly
and
compensating
bushing
its
is
for wear.
It is
obvious that once the bore of the

rectify
worn
eccentric,
no amount of taking up can

in this
shape; in fact taking up wear

fit
manner tends
to destroy
the
of the journal in the bearing.
Occasionally the journal
itself is
2 43
is
made
conical,
and adjustment
for
wear
made by moving
is
the shaft endwise.
The
application of such bearings
limited
to short shafts, such as machine-tool spindles.
Machine bearings are made in many forms, depending on and service. The bearings are sometimes split into three pieces, and various other means of compensating for wear are used, but the fundamental principles outlined above regarding the point where the bearing should be parted apply
the location
to all forms.
Consider the
lines of shafting,
is
last
item
(e,
automatic adjustment).
In long
it
which tend rapidly
to get out of adjustment,
desirable that the bearing be so constructed as to adjust itself
automatically to the changing

in
position
of
the
shaft,
order to avoid localized
pressure,
sult
which would
heating.
refastalso,
in
In
running
such
as
machinery,
countershafts, dy-
namos, and
motors,
is
where
neces-
perfect alignment
sary, self-adjusting bearings
have
been
found
almost
Fig. 84.
essential.
Fig.
84 shows a
bearing of this kind as used in
dynamo and motor
bearings.
The
sleeve
has a spherical surface turned upon the outside,
the centre of the surface being at O.
This surface engages with
a similar surface bored in the outer casing B.
The
sleeve
may
swivel in any direction, but the centre line of the shaft must
always pass through O.

this kind
it
When
a shaft has only two bearings of
is
evident that perfect alignment can be secured,

Similar devices are
within the range of motion of the sleeves.
used in the case of long shafting, where

used.
It is
many
bearings must be
obvious that the fundamental principles regarding

for
adjustment
wear and maintenance of the

also.
oil film,
apply
to all
bearings of this form

96.
Practical
Construction of Bearings.
It
was shown
in
244
MACHINE DESIGN
and the white
It is to
Article 88 that metal such as brass, bronze,
alloys
make excellent
on account of
bearing surfaces for wrought-iron or steel journals,

their anti-friction qualities.
be noted that
even in the case of perfect lubrication, where the character of the
rubbing surfaces
care
is less
important once the
oil
film
is established,
must be exercised
is
in the selection of the material for the
bearing surface, in order that abrasion

film
may
film.
not occur before the
formed, or in case of failure of the
There
is
a further
advantage in having the bearing surface softer than the journal,

in that
it is
very desirable to have the journal maintain

it is
against wear, which
a soft surface than
it
its form more likely to do when rubbing against would against one harder than itself. The
bearing
itself
should be
rigid, so as to insure
proper alignment of
the shaft.
Rigidity, against even

if
moderate pressure, could not
ordinarily be attained
of the white alloys,
the entire bearing
member were made

of brass
and economy prohibits the use

It
is
and
bronze for the entire bearing.
customary, therefore, to
main body of the bearing of cast iron (or sometimes a and to fit into it wearing surfaces of the softer These wearing surfaces may be either rigidly attached metals. In Fig. 80 is shown a to the main castings or may be removable.
the
steel casting),
make
bearing of the type

babbitt-metal lining
commonly used
is
for
heavy shafts when the
rigidly
attached by means of dovetail

is
shaped recesses, into which the babbitt

state.
poured
in a
molten
The
necessary shrinkage due to cooling, which would

is
leave the lining loose in the recesses,
usually overcome by
hammering the babbitt, when cold, till it again fills the recesses, and then boring the babbitt to size. For cheap work the lining is often cast to size on a metal mandrel and no further work put upon it, but for all good work the bore of the lining is cast small enough to allow of hammering or peening, and then boring to a
smooth
turning
surface.
Fig. 81
shows removable linings of brass or
bronze which are circular in section, and are prevented from
This parting by the parting piece B. piece, or also permits taking lip wear by reducing its thickness as occasion requires. Fig. 82 shows an arrangement of wearing surfaces common on horizontal steamin place
when
" liner,"
engine bearings.
245
The cap C
is
is
babbitted with some form of

it,
cheap metal since there

either
no wear upon
all
the pressure being
downward
or sidewise.
The "quarter boxes"

is
B, and the
lower box or shoe A,

lined with babbitt.
may
be of brass or bronze, or of cast iron

there
Where
danger of the boxes breaking,

it
through pounding by the shaft, and where

babbitt metal, they
is
desired to use a
and babbittand compensalined. When tion for wear is important, as in the case of machine tools, it is customary to make the wearing surfaces removable as indicated
of brass or bronze
may be made
cast-iron wearing surfaces are used,
work the bearing surface is part of machined to the required size. Hardened steel bearing surfaces are obtained by making circular shells or bushings," of the required internal diameter, and of sufficient
in Fig.. 81.
For
less
accurate
the
main casting
itself,
'
thickness to insure strength.
These bushings are forced
into
openings in the main casting and no provision for taking up wear

is
made.
If the forcing
operation closes the bore of the bushing,

T
it is
"lapped" out with emery and oil to the required size. W here the bearing must work under water, as in the case of a propeller
lignum
vitae
shaft or the lower bearing of a vertical turbine water wheel, a

lining of
or other hard
wood
is
often used.
The
of
surrounding water furnishes the only lubricant necessary in such

cases.
detailed
is
description
of
the
many arrangements
bearing surfaces
beyond the scope of this treatise. When the bearing must work under trying conditions, as on shipboard or in a heated room, and there is some question as to
whether the heat of
friction will
be dissipated by radiation, the
bearing
is
cast hollow so that water
may be
circulated around
it,
thus carrying off the heat and maintaining the lubrication.
In
an emergency, water
of the bearing,
may
be allowed to run over the outside
accomplishing the same purpose.
High-grade
marine work, and large stationary-engine

important bearings.
installations, are often
equipped with a complete system of water circulation on the most
JOURNALS
97.
Theoretical
Design
of
Journals.
The
considerations
affecting the design of
any journal are one or more of the following:
246
(a) (b)
(c)
MACHINE DESIGN
Strength to resist rupture.
Rigidity, or stiffness, to prevent undue Maintenance of form against wear.
yielding.
(d)
(e)
Maintenance of lubrication. Radiation of the heat due to

first
frictional resistance.
The
are
two considerations,
covered by the general principles laid
more
fully considered in
and rigidity, are Chapter III, and Chapter XI, where the special probstrength
down
in
lems in connection with shafts are discussed.

material dictates that
Economy
of
theminimum diameter
of shaft
be consistent
with the applied bending and twisting moments.
The
third consideration (c) particularly affects such journals
as those on the spindles of grinding machines and machine tools

generally,
where the accuracy of the product depends on the accuUsually, in such cases, the wearing surface
racy of the journals.
must be so great, in order to reduce the wear to an inappreciable amount, that the consideration of strength does not enter into the
computations.
The considerations, (d) and (e), are closely correlated. It was shown in Articles 32 and 33 that if the unit bearing pressure on the journal is not too great, the lubricant, because of its viscosity, may be drawn in between the journal and the bearing, thereby reducing the frictional resistance. This frictional resistance can never be reduced to zero even with perfect lubrication.
The energy
thus absorbed appears as heat, and
is
radiated
to the surrounding air
by the metallic surfaces

till
of the bearing,
the temperature of which rises

that at which heat
is
the rate of radiation equals
being generated.
In well-designed ma-
chinery the temperature of the bearing should not exceed 150 F.
The
raising of the temperature of the bearing has a tendency to
lower the viscosity of the lubricant, and

too hot, the lubricant
it
if the bearing becomes becomes so thin that the pressure squeezes out completely, and failure of the bearing by abrasion occurs.
It is evident, therefore, that
a journal of given dimensions
may
in
carry a given load very satisfactorily under certain conditions, and

fail
absolutely under others, the
same lubricant being used
each case.
The
consideration of the proper radiation of the heat
generated
is,
t
247
herefore, most important. It
may be assumed, without

is
serious error, that the rate of radiation of heat
proportional to
of heat units
the projected area of the bearing.
The number
which
ing air,
will
be radiated from a unit of surface, at any given
difference in temperature
is
between the bearing and the surround-
a fixed quantity for
any
is
set of conditions;
and
if
the
heat of friction per unit area

at the desired
greater than can be radiated
bearing temperature, the temperature of the bearing

equilibrium
is
must
for
rise
till
obtained.
It
follows therefore that

friction per unit
any desired bearing temperature the work of
of projected area of bearing
must not exceed the

jr
rate of radiation
per unit of projected area, or
fiwV=K
where
a*
or
wV =
H
(1)
is
the coefficient of friction,
the load in pounds per
unit of projected area,
V
to
the velocity of rubbing in feet per
minute, and
foot
the rate of radiation per unit of projected area in
pounds per minute,
be determined experimentally.
if
f*
It is to
be considered as constant,
increasing the diameter of a journal (the
number
of revolutions
affect
and the
bing
total load
remaining constant) does not materially
the development or dissipation of heat, since the velocity of rubis
increased in the
If.
same
ratio as radiating surface is in-
creased.
however, the bearing be lengthened, the radiating
surface
is
increased and the
work
of friction
remains unchanged,
statement, while
with the same total load as before.
This
is
last
true for imperfectly lubricated surfaces,

true
for
only approximately
as
will
bearings
with
perfect
lubrication
be
seen
presently.
The amount
shown
in Fig.
of heat
which
will
be radiated from a bearing
has been experimentally determined by Lasche.*
The
curves
84
(a)
are those
shown
in his Fig. 57,
transformed
into English units,
and with the
scale of radiation further modified
so as to read in foot
pounds per square inch
of projected area
per second, instead of per square inch of actual bearing surface.

Traction and Transmission, January, 1903, page 52.
248
Curve
i
MACHINE DESIGN
represents actual experimental results, with bearings of
still air.
the usual proportions, in
Curve
2 is for
bearings which
are connected to large iron masses, or which are ventilated by air

currents.
Curve 3 was calculated from theoretical considera-
6
IN FT. LBS.
10
12
IN.
14
16
RADIATION
PER SEC PER SQ.
PROJ. AREA.
Fig. 84 (a).
from a very thin bearing or sleeve and indicates that radiation is more effective as the bearing becomes thicker, as might be expected; for metal is a better contions.
It.
gives the radiation
ductor of heat than
air,
and hence the thick bearing more
easily
carries the heat
249
away
to a greater radiating surface.
The
for
values
obtained from these curves equation

(1).
may
therefore be used
in
Lasche points out that though these experiments
represent only a limited variety of conditions, they are probably
on the
If,
safe side
and
will serve at least as a very useful
check
in
designing.
in
designing a journal, the value of

(1)
fi
can be deter-
mined,
equation
and
Fig. 84 (a)
give the relations
which
must
and
exist
between the velocity and pressure in order that the
safe bearing temperature
may
not be exceeded; or
if
the pressure
velocity are fixed
by other circumstances,
air.
Fig. 84 (a) indicates
whether radiation must be assisted by

water circulation or currents of
artificial
means, such as
98. Imperfectly Lubricated Journals.

Articles 32
surfaces,
is
It
has been shown in
and 33 that the value
of
p-,
for imperfectly lubricated
a very variable quantity, even for the
neous values of velocity and pressure.

velocity,
oil
same simultaNot only does it vary with
and temperature, but the regularity of the supply (over which the designer has little control) affects it
pressure,
much more seriously. Further, bearings running under the same nominal load and velocity give widely different values of frictional resistance and temperature rise, depending on whether
the load
is
constant or intermittent, or whether the motion

etc.
is
steady or vibratory,
Notwithstanding
this,
equation
(1)
may
be
made
to serve as a useful
<>.
check in doubtful cases by assuma constant
ing a safe value of
The assumption
is
sometimes made that

jr
fi
is
and
formulae of the form
wV = =
C, where
is
a constant that
has been determined from practice, are

minute, Mr. Fred
much
used.
Thus
if
w be
V
in
expressed in pounds per square inch of projected area and

feet per
W. Taylor*
is
gives for mill
work
C=
24,000; and says that
C = 12,000
not safe for cast-iron bearings
with ordinary lubrication. If the rise of temperature in the bearing
be taken as 75
and
be taken as .015, which

S.
is
ordinarily a safe
Transactions A.
M.
E., Vol.
XXVII.
250
value, then
MACHINE DESIGN
from curve
1,
Fig. 84 (a),
K=
222,
whence
C = =
222
.015
1=5,000.
From
curve
2,
K=
384 whence for ventilated
bearings
C =
25,600.
These values agree with Mr.
Taylor's limits better than would be expected.

All formulae of this empirical
form must be considered,
as
far as imperfectly lubricated journals are concerned, as applying
only to the conditions and range for which they have been found
true,
and
for
which
,u
is
apparently constant.
This
is
more
evident when the wide variation of the value of such constants

as determined
by practice
is
considered.
gives, as the practice of the
General Electric
Thus Mr. H. G. Reist Company on generator
bearings, a limiting value of
C = 50,000 for bearing
30 to 80 pounds per square inch.
pressures from Mr. H. P. Been gives the
practice of one of the largest Corliss engine builders as
C=
140
60,000 to 78,000 for bearing pressures not higher than
pounds per square inch. Unwin, page 249, gives values
of a similar constant,
/?,
which
corresponds to the following values of C:

Locomotive Crank Pins Locomotive Axles Marine Engine Crank Pins
Stationary Engine
250,000 to 375,000 200,000

50,000 to
15,000 to
75,000
Crank Pins
50,000
20,000
Railway Carriage Axles
75,ooo to 100,000
7>5 to
Crank Shaft Bearings
The
great variation in these values of
C is no
more than might
be expected in view of the foregoing, and also in view of the

difference in lubrication
and
in radiating capacities of bearings,
due
to
material,
form,
and
location.
While, therefore, these
coefficients
may form
a guide, and while doubtful cases

(i),
may
be
checked for heating by equation

the bearing pressure
of
is
care should be exercised that
kept within the limits which will admit
good lubrication. The allowable bearing pressures as fixed by practice for various classes of machines are given in the following table, and it may be noted that these are more accurately
. .
known than
radiation K.
the values of
,,
25J
or the values
of
the coefficient
of
Economy
in the use of material
and the importance of mini-
mizing the work of friction suggest that the diameter of the

journal shall be as small as
ness.
is
consistent with strength
and
stiff-
With
the diameter of the journal determined
by these
TABLE
XIII
BEARING PRESSURES FOR VARIOUS CLASSES OF BEARINGS

Allowable Bearing Pressure lbs. per Square Inch.
Class of Bearing and Condition of Operation.
Bearings for very slow speed as in turntables in bridge work Bearings for slow speed and intermittent load as in
7000 to 0000
punch presses
Locomotive Wrist Pins Locomotive Crank Pins Locomotive Driving Journals Railway Car Axles
Marine Engine Main Bearings

{
ggS^SSto
(high speed)
{
[
Marine Engine Crank Pins

Stationary Engine
3000 to 4000 3000 to 4000 1500 to 1700 190 to 220 300 to 325 275 to 400 400 to 500 400 to 500
60 to ISO tO
120 2sO
Main Bearings
for for
dead load* steam load

.
(high speed)
I
overhung crank. centre crank

[
Stationary'
Engine Wrist Pins (high speed)

(slow speed)
for
900 400 1000

80 200 800 1000 500 1500 1500 100
1500 600 to 1800

to to
to to
Stationary Engine
Main Bearings
for
dead load.* steam load
140
Stationary Engine Crank Pins (slow speed) Stationary Engine Wrist Pins (slow speed) Gas Engines, Main Bearings Gas Engines, Crank Pins Gas Engines, Wrist Pins Heavy Line Shaft Brass or Babbitt Lining Light Line Shaft Cast Iron Bearing Surfaces
to
Generator and
Dynamo
Bearings
to to to to to 15 to 30 to
400 1300 1500 700 1800 2000 150

25
80
considerations,
it
is
evident that the length of the journal must

is
be such that the bearing pressure

It
within the allowable limit.
may
be, however, that the length of the journal thus deterwill
mined
be so great that localized pressure
may
result;
or
it
may be
that the type of
machine
will not allow
space enough for
Weight
of shaft, flywheels, etc.
252
MACHINE DESIGN
such a length of bearing.
made
In such cases the diameter must be and the length may be correspondingly decreased. While practice shows wide variations, it is found that the ratio
larger
its
of the length of the journal to
diameter
is
fairly well
defined for any given class of machinery.

fore, that,
It often occurs, there-
when
journals are designed with the ratio as fixed by
practice, they
have an excess of strength while barely satisfying
the conditions as to bearing pressure.
The
practice
following are average values of
as
found
in
good
TABLE XIV
TYPE OF BEARING
Marine Engine Main Bearings Marine Engine Crank Pins Stationary Engine Main Journals Stationary Engine Crank Pins Stationary Engine Crosshead Pins
Ordinary Heavy Shafting with Fixed Bearings. Ordinary Shafting with Self-adjusting Bearings Generator Bearings
.
. .
Values of -y
a
to
to
I
1
1
ifcto 2.5
1
to
1
.
to 3 to 4
3
From the foregoing the following statements 99. Summary. may be made regarding imperfectly lubricated journals (a) The minimum diameter of a journal is fixed by the con:
siderations of strength
(b)
and
stiffness
under the loads applied.

of friction
The
smaller the diameter of the journal for a given co-
efficient of friction, the less is the

liability to heating.
(c)
is
work
and consequent
The tendency
is
of the bearing to heat, other things equal,
not materially affected by changing the diameter of the jour-
nal,
but
(d)
reduced by increasing the length.

projected area of the journal must be such that the
The
bearing pressure will be kept within the allowable limits for the
particular conditions;
and the
ratio of length to diameter
must
is
not be so great that severe localization of bearing pressure

result.
These considerations may require a larger liable to bearing than the previous requirements alone would demand.
(e)
253
The work
of friction, per unit area,
must not exceed the
rate of radiation, per unit area, for the allowable bearing temperature.
100. Perfectly Lubricated Journals.
It
was shown
in Article
33 that if a journal is supplied with sufficient lubricant, of proper viscosity, the journal itself may draw in the lubricant till
a film
is
formed under such a pressure that the load
will
be en-
tirely fluid-borne.
With any given
set of conditions,
therefore,
and perfect
lubrication, a definite journal velocity will permit
the carrying of a definite load per unit area
upon the journal,
and once the
relation
is
established between the load, velocity,

it
and
coefficient of friction,
is
constant,
and not unstable, as
in the case of imperfectly lubricated surfaces.

It
was further shown that the following statements are true
regarding perfectly lubricated surfaces.

(a)
The
friction of perfectly lubricated surfaces for a given

little
velocity
depends very
on the materials which form the
rubbing surfaces, but does depend largely on the character of

the lubricant.
(b)
The
film
ft
frictional resistance of perfectly lubricated journals

is,
for
any given velocity

oil
is,
within the limits of pressure under which
the
may be
maintained, independent of the pressure;
(that
a constant)
(c)
The
coefficient of friction of perfectly lubricated surfaces,
for
any given pressure, varies very nearly as the square root of
velocity, for velocities
up
to
500
ft.
per minute;
approximately
as the
fifth
root of the velocity for velocities between 500

is
and
2.000
ft.
per minute; and
practically independent of the velocity
for values
above 2,000
ft.
per minute.
The The
first
of these statements has

is
been abundantly verified
by experiment and
discussed more
is
fully in Article 90.
following table, which
one of the several given in
Tower's report, shows clearly the truth of the second statement,

for the frictional resistance is seen to
remain practically constant
with
is
all
loads at any fixed velocity.
The
also seen to vary very nearly as the square root of the velocity.
$$,
Table VIII, Article
which was deduced from Table XV, shows
2 54
MACHINE DESIGN
the coefficients of friction for the range of pressures and velocities

given, the latter not exceeding 500
ft.
per minute.
TABLE xv
BATH LUBRICATION
Frictional resistance in pounds per square inch of projected area of bearing surface = j. w, for velocities in feet per minute as below. Temperature = 90 F.
n.So
3b 1
a
u
1 05 ft.
157
ft.
209
ft.
262
ft.
314
ft.
366
ft.
419
ft.
471
ft-
573 520 4i5 363 258 i53 100
583
.496 .386 .306
.62
.678
597 495 .401
.721
546 445
35
.648
539 444 503 459 576
758 .691 .582 .488

547 .510 .619
794
Seized
.771 655 .561 .626
.605 .714
735
.619
532 583 .561 .663
277 248 277
357 .306
357
.416 3 6 4
423
459 .408
53
Tower's experiments
ties,
at different
temperatures show that the

veloci-
coefficient of friction, for the
above range of pressure and
decreases as the temperature increases.
His principal experi-
ments, from which Table
XV
is
taken, were conducted at 90 F.
and without
artificial
means
of cooling the bearing.
The
differ-
ence between the coefficients of friction obtained at 90 F. and those obtained at temperatures as high as are usually allowed in
practice,
can be neglected, as far as designing
is
concerned,
especially since those at 90 are
on the
safe side.
Tables VIII and
XV may,
therefore,
be taken as representing
fairly well the rela-
tion existing
this range,
between pressure, velocity, and

It is to
frictional resistance for
which fortunately covers the most usual conditions

be noted that
at the greatest pressure
in practice.
and
highest velocity, the bearing seized, indicating that with such

velocity a lower pressure
is to
must be assigned,
if
a perfect
oil film
be maintained, or that with
this greatest load
a lower velocity
must be assigned, if the bearing is to radiate the heat of friction. The work of friction at 471 ft. per minute with a load of 520 lbs.
is
seen to be
.771X471=365
ft.
lbs.
per minute, or about 6

it
ft.
lbs.
per second.
From
curve
I,
Fig. 84 (a),
appears that to
radiate this
amount
of energy, the bearing
must
attain a tempera-
ture of over
255
no F. above
that of the surrounding atmosphere, or
a total temperature of at least i8oF.

It is to
be especially noted that within the limits of pressure

oil film will
is
where a perfect
given velocity,
pressure.
form, the frictional resistance, for a
practically constant
and independent
of the
(See Table XV.)
The
and
coefficient of friction, for bearft.
ings running at velocities of over 2,000

lubrication, have
per minute with perfect
been quite
fully
determined by
Wm. O. Lasche.*
The experimental work was

clusive,
very extensive, the results very con-
and should be carefully read by designers of high-speed
machinery.
discussion of these experiments
is
beyond the
scope of this treatise,
but a few of the most important results will

at these
be considered.
Lasche found that
high velocities the
coefficient of friction
was
practically independent of the velocity,
but varied inversely with the pressure as in the
Tower
experi-
ment^ and
that
if
also varied inversely with the temperature.
w be the bearing pressure in

fiw
or
(t
He found pounds per square inch, and t

=
51.2
(2)
the temperature of the bearing in Fahrenheit degrees, then
32)
~
/'.
w =
('-32)
ft.
(v 6
per minute the co-
For
velocities
between 500 and 2,000
efficient of friction varies about as the fifth root of the velocity, as
shown by
the experiments of Stribeck.
As
far as designing is con-
cerned, the difference between the coefficients for this range and
those found by Lasche for the higher velocities, may be neglected, and Lasche's equation may be applied, without serious error, to
all velocities
above 500
(3)
ft.
per minute.
Equation
may be
written
" V
where V
is
is
51
V
(4)
7Z7^
ft.
the rubbing velocity in
per minute.
Since
,"
wV
the frictional loss per unit of projected area in foot
minute, equation (4)
may be
used to
pounds per which a heat compute the

1903.
* See Traction and Transmission, January,
256
MACHINE DESIGN
perfectly-lubricated high-speed journal can radiate per square
inch of projected area, and not rise above a temperature
t.
The
limiting values of the pressure under
which a perfect
oil
film can be maintained, at these high velocities, have not been

fully determined.
In Lasche's experiments a load of 213 pounds In Kingsbury's* experiments loads from

per minute.
the
oil
per square inch of projected area was carried at a velocity of

1,968
ft.
per minute.
80 to 86 pounds per square inch were repeatedly carried at velocities
up
to
1,990
ft.
In both Kingsbury's and

or
Lasche's work
self,
either
circulation,
the
bearing
it-
was
artificially cooled,
thus materially assisting the radia-
tion.
The
values.
values given by these experiments were obtained on
experimental machines and

ings gives velocities ranging
may be
looked upon as limiting
Successful practice in the design of steam turbine bear-
from 1,800
to 3,000
ft.
per minute,
with pressures inversely as the velocity ranging from 80 to 50 pounds per square inch. Where the pressure is as high as 90
pounds per square kept below 1,800
inch,
ft.f
it
is
found that the velocity must be
per minute.
The
empirical equation
wV =
=
150,000
is
much used
for this class of

It is
work, and gives

evident that with
values agreeing with those just quoted.

these high velocites the radiation
must be
assisted.
Thus
t
let
2,000 and w =
75 in accordance with the empirical rule just

at
given,
and
let it
be required to keep the temperature
150 F.
or a temperature of say 75 F. above the atmosphere. Then by equation (4) the frictional work is,
51 wV=
11
= 867 (150 32)

.
2,000
ft.
lbs.
per minute or 14.

if
ft.
lbs.
per second, whereas the bearing alone,

iron frame, will,
connected to a heavy
ft.
from Curve
2,
Fig. 84(a), radiate only 6.4
lbs.
per second.
Since the specific heat of both water and
oil
are
known
the supply of either necessary to carry off the excess
heat of friction can be calculated.

* Transactions A. S. M. E., Vol. XXVII, page 425. f See " Steam Turbines," by Frank Foster, page 181.
It is to
257
ft
be noted that with perfect lubrication the product

is
for
any velocity
a constant quantity.
It follows therefore that
for
any given
total load
the unit bearing pressure should be

it
kept as high as possible provided
does not exceed the max-
imum
allowable value for the given conditions.

is
For
if
the unit
bearing pressure
decreased, either by increasing the diam-
eter or length of the bearing, the coefficient of friction is cor-
respondingly increased
is
hence the
total frictional resistance
also increased.
Care should, of course, be exercised in any

is
case that the heat of friction
properly carried away.
10 1. Examples of Journal Design.
Journals generally form
an integral part of a shaft or spindle, and the determination of

the stresses acting
upon them
I
is
a part of the solution of the
xl.,,
]*
M
>TFig. 85. Fig. 86.
stresses in the shaft itself.
It is desirable,
however, to point out
some
of the special features of journal design.
The
actual distribution of pressure over a journal, in the

is
direction of the axis,
not known; but there

is fairly
is
every reason to
believe that the distribution
uniform.
Thus
bearings,
as a rule, wear quite uniformly over their entire length, where

fair
alignment
is
maintained.
It is
customary, in the absence

to strength
of exact data, to
rigidity
assume
for
computations as
journal
is
is
and
and
that
its
the load on the

length.
concentrated at the
middle of
will
This assumption
on the
safe side,
sometimes give shaft diameters excessively large as

is
far as
strength
concerned.
(a, b
The
]
following examples
cases of journal design.

7
It is
and c) show the most important assumed in each case that the bear-
258
MACHINE DESIGN
is
ings are imperfectly lubricated, which

dition,
the most
common
con-
but the application of the theory to perfectly lubricated

is
journals
obvious.
(a)
.
Example
This case
is
is
illustrated in Fig. 85
tre of the bearing
fixed at O,
Here the cenby the construction of the machine.

.
The
centre line of the pulley

belt
is
also fixed at
XX, by
the location
which the
must occupy, so
a.
that the pulley overhangs the

of the pulley d
is
bearing by the distance
The diameter
side
is
40
inches, a = 10 inches, the pull on the tight side of the belt

lbs.,
is
500
and the
pull
on the slack
300
lbs.
It is
required to
determine the dimensions of the journal.
The
stresses
induced in the journal are, torsional
stress
due
to the twisting
moment (T T
1
d
2)
flexural stress
due to the bend.
moment (7\ + T 2) a, and shear due to the direct pull 7\ + T 2 The last is small and is usually neglected (see Art. 26), and the journal may be considered as subjected to a combined bending
ing and twisting moment.

fore, applies.
Formula
K
2)
or
(page 49), there-
The bending moment

The
twisting
M = (T
(7\
,
+T
2)
a =(500 + 300) 10 = 8,000

(500
p,
moment T =
2
T - =
taking
,
.
8
n
.
300) 20
.
4,000
Hence
M = 8,000 = I
4,000
2
= x and
=
p
it is
found from
Figure 9 that equation

.'.
applies
=-[x + Vx + 2
i]T = -[2 +
2
V2 +
2
i]4,ooo
3
.
8,480
From
equation J, page 94,
M
3
pTld pi =
or
.'.
rfi
_ 3i^. .
71
X 8,480 _ x 10,000
2%"
XIV)
tl
2}i inches (nearly), or say

7
2
If the length of the
bearing be taken at
will
M1
.
inches (see Table
the bearing 5
pressure F
be
7\
+ T = *% X 7
500
300
50
lbs.,
15-75
which
is
a safe value.
If the
259
/'.
number
of revolutions be
300 per minute, and
be
taken as .015, the work of friction per unit of projected area will
be 50
ft.
.015
300
'-
133
ft.
lbs.
per min., or 2.2
lbs.
per sec.
From Curve
1,
Fig. 84 (a),
it is
seen that to radiate
this
amount
of energy the temperature of the bearing will rise

air.
about 50
the design
above the surrounding

is satisfactory.
This
is
a safe value and
Example
crank pin
(b).
Let the
line of action of the
load pass through
the centre line of the journal, as in the case of the steam-engine

in Fig. 86.
Let the length of the crank be 18 inches,

pressure on the
and the
pounds.
total
maximum
crank pin be 25,00c
What
should be the dimensions of the crank pin in
order to be safe against rupture and overheating?
Referring to Table XIV,

ter are short
it is
seen that journals of this charac-
compared
stiff
to their diameter,
and hence are usually
strong enough and
enough
if
designed for a bearing pressure
low enough to prevent overheating.
Let
be
taken as 1.25.
From Table XIII

safely carried
it is
seen that 900 lbs. per square inch

If
may
be
on
2
.
this type of pin.
d be the diameter of the

is
pin and
1.25
the length, then the projected area of the pin
lxd =
dXd = i.2$d
Whence 1.25^X900 = 25,000

or d 2
" . .
= 22.2
5
^ = 4.7 or say
1
inches,
inches.
and
= 5X1.25 =6.25
to
The
this
pin
it
may now
is
be checked for strength.
In a short pin of
dis-
kind
more accurate
assume the load uniformly

assume
it
tributed along the pin, than to
as concentrated at the
middle.
The
pin may, therefore, be considered as a cantilever
uniformly loaded with a load
W = 25,000.
3,
Whence from Table
I,
case
the
maximum bending moment
Wl =
25,000
6.2 5
78,125 inch pounds
260
MACHINE DESIGN
from equation
7,
page
94,
pi
or p
Me M - =
32
is
-=-
or
^>
6,400
pounds
nearly,
which
safe
value.
In a similar way the pin

desired,
may
be checked for deflection,

I.
if
by means
(c)
.
of case 3,
Table
Example
Sometimes the location of the bearing is dependis
ent on the diameter of the shaft, which
unknown, and
in
such
case a tentative method must be adopted.
Thus
is
in Fig.
86 neither
the length of the bearing B, nor the thickness of the crank
hub
t,
can be definitely decided upon

diameter of the journal.
till
something
known about
the
The diameter must
therefore be asto the
sumed, and then checked by the equations which apply

case.
Usually a close estimate can be
made from
is
existing
ma-
chines of similar type.

for example,
it is
In the case of the steam-engine shaft,

that the
known
main journal
frequently about
one-half the diameter of the cylinder.

(b)
The
data taken in example
correspond to a cylinder diameter of about 18 inches, and the
journal diameter
may
therefore be
assumed as 9
inches.
From
Table XIV, the length
of the journal
may be
taken as 20 inches.
The
length of the hub should be at least 8 inches, for this di-
ameter.
The
boss under the pin

pin,
may
is
be taken as yi" in height
and since the
from case
(b),
6.25 inches long, the total
distance from the centre of the crank pin to the centre of the shaft
journal
may be assumed as 21 % inches. The projected area of the is 9X20 = 180 square inches, which gives a bearing
pressure of
= 140 pounds per square

25,000 -
inch; and from Table
IoO
XIII
this
it is
seen that this
is
a safe value as far as the load due to steam a heavy fly-wheel
pressure
is
concerned.
If the shaft also carries
must be taken
stresses
(a).
into account (see next chapter).
The
induced in the journal are of the same character
as in case
Taking the length
of
crank
= i8
inches,
and the
pressure on the pin
= 25,000 M = 25,000 X 21 }i = 537,500
as before, then the bending
inch pounds, the twisting
moment moment
26
T =
x,
25,000
18
= =
450,000 inch pounds, whence
=
to
1.19
and taking
.8,
equation
is
found by Fig. 9
apply to
the case.
.'.
[.v
vV
1]
r =
[1
19
+ Vi.K) +
2
1]
45> 000
616,500.
From J
which
is
t/
(as in
example
ix a) p
32
TZ
ê =
-=-
32
X
7T
616,-soo
d
is
8,600,
a safe value and the design
satisfactory.
102. Lubrication of Journals.
The
point of application of
the lubricant
is
of
utmost importance, and the method of supplying

affects
the lubricant to the journal sometimes materially
the
design of the bearing.
The most common methods
of feeding
lubricants to rubbing surfaces as given in Article 30 apply fully to
journals and
may be
classified as follows:
'
Common oil hole. Common wick or siphon feed cup. Common drop sight feed cup.
Oily pad against journal. Ring or chain feed.
Centrifugal
oiler.
Imperfect Lubrication
Compression grease cup.

[
Perfect Lubrication
Bath lubrication. Flooded lubrication.

Forced lubrication.
In flooded lubrication (sometimes erroneously called forced

lubrication), the oil
is
supplied to the bearing under a low pressure

is
which insures that the journal

application, as
in
always flooded at the point of

it
bath lubrication, but
does not force the

oil is
lubricant between the surfaces.
In forced lubrication the
supplied at a pressure in excess of the film pressure at the point

of application,
and
is
thus forced in between the surfaces, no
reliance being placed
on the tendency of the journal
to
draw
in
the
lubricant.
The compression
grease cup, while supplying
the lubricant under slight pressure, gives only imperfect lubrica-
262
MACHINE DESIGN
is
tion as the supply of lubricant
not copious as in the case of
forced lubrication.
In applying any of these methods of lubrication, therefore,

except the compression grease cup and forced lubrication, care
should be exercised that the point of application
is at,
or near, a
point of lowest pressure and at the place where the journal will
naturally
sure
is
draw
in the lubricant.
Thus,
in Fig. 80,
if
the pres-
always downward, lubricant can be supplied at

in either direction.
If the
for
motion
oil
pressure were upward, an
hole at
would not only be
useless for supplying lubricant
but would be fatal to good lubrication, as any tendency for a

pressure film to form would be destroyed by relief of the pressure
at the hole.
In such a case the lubricant should be supplied

if
from underneath, or
wise an
oil
the direction of rotation were anti-clockat
hole as
shown
/ would be good design.
In forced
lubrication the point of application should be the point of greatest
bearing pressure, and the hydraulic pressure under which the

is
oil
supplied should be greater than the
maximum bearing pressure.

to
While the decreased

evident,
it
friction
due
effort
perfect lubrication
is
does not follow that an
should be
made
to design
every bearing so as to secure this advantage.
In some places a
simple
oil
hole
is sufficient, in
others a constant supply from a
wick feed
oiling
lations,
will suffice,
is
while again, with greater speeds, a ring
device
necessary.
In
many modern power
instal-
with
either
steam turbines or reciprocating engines,

the
very complete apparatus for supplying flooded lubrication will
be found.
oil,
The
bearings are constructed so as to catch
all
as
it
leaves the journal,
and pipes convey
it
to a central
receiver.
A pump
in-
continually circulates the
oil to
the various
bearings, and
the best installations the
oil is filtered
during the
lubrication
circuit.
The same
results are obtained
and cooled by flooded

is
as
with bath
lubrication.
Forced
lubrication
resorted to only where the bearing pressures are excessive and
beyond those which can be supported by the natural action the film formed by rotation of the journal. (See Art. 33.)
of
The
location and character of the
oil
grooves deserve special
attention.
If the velocity of the journal is so
low as
to
draw
in
little
263
lubricant the
oil
grooves should be so cut as to allow the
lubricant to flow in near the points of greatest pressure.

or scores
Grooves,
on the journal
heavy.
itself,
have been found helpful in drawing
in the lubricant
under such circumstances; especially where the
lubricant
is
But where the velocity
minute
(see Fig. 16),

oil
and
for
above 25 feet per ordinary pressures, care should be

is
used that no
grooves are cut that will tend to prevent the

If the lubricant is delivered at
oil
formation of the pressure film.
H,
Fig. 80,
kind running from
and the pressure is downward, which will distribute the
grooves of any
over the surface
oil
of the journal, are allowable so long as they terminate at a little
distance from the edge of the bearing.

/,
If the oil is
delivered at
and the pressure
is
either
downward
or upward, the grooves
should be cut
at right angles to the direction
of motion, so as to
If cut
distribute the oil along the entire length of the bearing.
diagonally they will extend under the journal toward the point of
greatest oil pressure, thus relieving
of a pressure film
any tendency
will
to the
formation
it
and the lubrication
not be as good as
would be
if
no grooves were present.

of all oil grooves should be carefully
oil
The sharp edges

to facilitate the
removed
scraped
passage of the
under the journal.
The sharp
edges of the bearings themselves should also be
filed or
away
for the
same reason. some

due
little
Where one bearing
surface encircles
nearly one-half of the shaft, as in Fig. 80, the surfaces should be

relieved for
distance from the parting line
to
help
the
wedging action
of the oil
to
and
to insure the journal against
side pressure
springing of the bearing under the load.
bearing which binds sidewise will not lubricate properly.
THRUST BEARINGS
a
When
a shaft
is
subjected to
heavy end thrust, either from the weight of the parts carried
or on account of the
are used to
suffice,
power transmitted, the simple
collars
which
prevent end thrust in ordinary shafting will not
and bearings of special form, known as thrust bearings,

is
it
must be provided. If the bearing is taken on the end of the shaft
designed so that the thrust

is
called a step-bearing or
264
footstep-bearing.
MACHINE DESIGN
If the thrust
bearing must be placed at some

it is
distance from the end of the shaft

104. Step-Bearings.
as
is
called a collar bearing.

is
If the
motion of rotation
very slow,
the case in swinging cranes and similar work, a simple cast-
iron step, as
if
shown
is
in Fig. 87, will

If,
meet the requirements, even

is
the pressure
heavy.
however, the velocity
high, this
simple arrangement will not give good results, even

pressure per unit area
is
when
the
low.
It
may be assumed,
The
without great
fitted
error, that the unit pressure
between the faces of a newly
step-bearing
is
uniform
at all points.
velocity of rubbing,
it is
however,
is
maximum at the
outer edge, and, theoretically,
zero at the geometric centre of the pivot.
Since the wear

it
is
pro-
portional to the product of pressure and velocity,
follows that
the surface will wear unevenly,
the greater wear taking place

at the outer edge.
This
will
bring a concentrated pressure

at other points,
and heating and

It is
cutting
may result.
always
advisable in heavy work, for this

reason, to
Fig. 87.
Fig. 88.
remove the wearing
surface near the center, where
the motion is slowest, and where eventually the greatest concentration of pressure is likely to be produced (see Fig. 87). Decreasing the bearing pressure by increasing the surface, is effective within limits, since the
area increases as the square of the diameter while the velocity

of
rubbing increases directly as the diameter.

however, increases the average
Increasing the
of
radius,
moment arm
the
It is
and hence increases the
lost energy.
often better, therefore, to carry a higher bearing pressure,
and
thus keep the diameter of the pivot small.

If a
number
of discs are placed

(Fig. 88),
between the
step,
or pivot,
and the bearing

relative velocity
they have the effect of reducing the
between adjacent surfaces;

is
if
and
if
the rotative
velocity of the pivot
high, they are very useful as a safeguard
against cutting;
for
abrasion should begin between any pair
of
discs,
265
the lubrication
motion
will
cease at that point
till
These washers are usually made alternately of steel and brass, or some other metal, and the upper and lower washers are fastened to the shaft and bearing respectively. An oil hole passes through the centre of the washers, and radial grooves cut across the faces permit a flow of oil between the
became
effective again.
surfaces, centrifugal action assisting the lubrication.
If the
top
of the bearing
is
connected to the bottom by an
oil
passage, as
shown
at
(Fig. 88), the centrifugal action will set

oil,
up a con-
tinuous circulation of the
making
is
the lubrication effective.
The
shaft
unit pressure
between washers
the
same
as
between the
combination
and the
is
first
washer, but the relative motion between the
surfaces
decreased and the wear thus reduced.
Fig. 89.
Fig. 90.
Fig. 91.
of
hardened and ground
steel
washers, alternating with brass or
bronze washers, makes an effective bearing.

order to allow the shaft automatically to adjust
Sometimes the
its
washers are made lenticular in shape, as shown in Fig. 89, in

alignment.
of
For very
light
work the
shaft
sometimes
rests
on a pair
hardened
steel buttons, or a
hardened steel ball which runs between hardened

In the submerged step-bearings of
is
steel surfaces is introduced.
water turbines the shaft, which
often capped with bronze, rests

is
on a lignum
ing water.
vitae
step
and lubrication
effected
by the surround-
If the outline of
a step-bearing be
made
that of a tractrix*
(Fig. 90),
it is
found that the tendency to wear in an axial direction

* See Church's " Mechanics," page 181.
266
is
MACHINE DESIGN
at all points;
uniform
in fact
if
two homogeneous
flat
surfaces
wear into the form of a tractrix by experiment. This is, therefore, the correct shape, theoretically, for all step-bearings; but on account of the difficulty and expense of machining the surfaces, it is seldom The tractrix has been called Schiele's Anti-friction Curve used. This is a misnomer, after the discoverer of the above property.
are rotated together they tend to
as has been proven
however, for the friction of a tractrix-shaped step
is
much
higher
than that of a plain pivot.

It is evident that the
rubbing surfaces of
all
the step-bearings
which have been discussed can be
submerged
in
an
oil
bath.
is
The
lu-
brication thus obtained
not to be
confused with that obtained on horizontal

rotating
bearings
discussed
force
to
formerly.
While
oil
centrifugal
does drive the
from the centre

action
the outside, there
is little
on the
to
part of the surfaces themselves tending, on account of

its
Viscosity,
draw
in
the lubricant between them, as
horizontal bearings.
Such
lubri-
cation cannot therefore be looked on
THRUST BEARING OF CURTISS VERTICAL STEAM TURBINE

Fig. 92.
as perfect lubrication
although giv-
ing
excellent
of
ments
The experiBeaucamp Tower* on a

results.
steel foot step, three inches in

eter, gives
diamIt
considerable
information on this subject.

oil
was
found that a single diametral
groove was better than more,
and pressures up
fully
to
160 pounds per square inch were success-
carried
at
128 revolutions per minute.
The
foot
step
was
and rested directly on the bearing, no washers being interposed. At 240 pounds per square inch the
freely lubricated,
bearing seized.
If
under heavy loads the maintenance of lubrication
is
im-
* Transactions of the Institute of Mechanical Engineers, 1891, page in.
267
portant, the lubricant should be supplied at the centre of the step-
bearing under a pressure such that the metallic surfaces are

forced apart and the load
is
fluid-borne.
Fig. 92
shows a recent
form of the step-bearing used on the Curtiss steam turbine.

vertical shaft
The
A, which supports the heavy rotating parts of both

is
turbine and generator,

it.
carried on the disc
which rotates with
The lower
is
disc
is
can be adjusted
vertically,
the screw E,
and
prevented from rocking on
by means of by the screws F.

load
is
Oil
forced between the discs through the central pipe Enforcing
the discs apart and escaping into the cavity G.
The
is
thus
completely fluid-borne and perfect

,Oil
lubrication
maintained.
Holes
Fig. 93.
The oil atH.
passes from
upward through the guide bearing escaping
105. Collar Thrust Bearings.
When
the thrust bearing must

of the shaft, the shaft
be placed
is
at
some distance from the end

shown
in Fig.
provided with collars integral with
itself,
which bear against
the resisting surfaces as
93,
which
illustrates a
thrust bearing as used for marine work.
In cheap work, or
where the load

the collar
is
small, a single collar
is
sometimes used.
Occa-
sionally a series of washers, as in Fig. 88, are interposed
between
and the bearing
ring.
is
The
objection to the single-
collar bearing for

to
heavy loads
that the large diameter necessary
obtain a practical bearing
pressure
increases the
work
of
friction,
due
to the increased velocity,
and the difference between and

at its outer
the rubbing velocities of the ring at the shaft
diameter results in unequal wear.
The
outer diameter of the
ring, or collar, is usually, therefore, not
more than one and one-
268
MACHINE DESIGN
which
limits the
half times the diameter of the shaft,
width of
face of the collar even in large shafts to a few inches;
and the
necessary area
is
obtained by using a number of rings.
In small or cheap work, the bearing surfaces of the thrust block

are sometimes
made
integral with the bearing proper; but usually
they are
made
detachable.
Thus
the
main casting
of the block
may be
of cast iron
and the bearing rings
of brass are inserted
held in place by radial grooves cut in the block.
and These rings must
be scraped
its
until
each collar on the shaft bears properly against

uniformly distributed.
is to
mating
ring, so that the thrust is
The
most modern practice
in
marine work
make
the bearing rings
horseshoe-shaped, as in Fig. 91, so that each ring can be with-
drawn without disturbing any other portion

shaft.
of the bearing or
Occasionally the horseshoe collars are adjustable along
more easily brought to a proper bearing. In work each horseshoe has its own independent water circulation, so that local heating may be prevented, and the lower
the shaft so as to be
first-class
part of the bearing constitutes an

dip.
oil
bath into which the collars

oil.
This
oil
bath also has a water circulation for cooling the
106. Friction
total load
and Efficiency of Thrust Bearings.

flat
If
p.
be the
on a
circular pivot of radius r t
and
be the co-
efficient of friction,
then the frictional
moment resisting rotation is

*
1
M = -hPy
3
If r x
(1)
be in inches and
is
be in pounds then the energy
lost
per
minute in foot pounds
E = -fiP
3
ri
271
= .^qp-P
12
rx
(2)
where
is
the
number
of revolutions per minute.

if
In a similar manner
side radius r lf
the thrust be taken on a collar of outr 2,
and inside radius
then
Y
**\
(Y ^
-/j.P
3
^h) ....
3
(3)
andfi-. 3 49MPiV-(jjji=^)
* Church's " Mechanics," page 180.
(4)
269
The
vertical
efficiency of a thrust bearing
cannot always be expressed
as a function of the
power transmitted.
a
little to
Thus
in the case of a
shaft carrying
heavy load of gears,
the
frictional
resistance of the step has
do with the power transmitted.
In the case of the thrust bearing of a steamship the frictional
moment and energy

force P.
loss are directly proportional to the driving
In either case, however, the frictional
energy loss must be added to the turning

supplied, as the case
moment or the moment or the energy
may
be.
The
following coefficients of friction are taken from Tower's
experiments
TABLE XVI
Minute as given below.
Pressures in lbs. per Unit Area.

50 R.P.M.
128 R.P.M.
194 R.P.M.
290 R.P.M.
353 R.P.M.
20
40 80 120 140
.0196 .0147 .0181 .0221
.0080 .0054
.0063.
.0102 .0061 .0045 .0052 .0062
.0083 .0093
.0178 .0107 .0064 .0048 .0046
.0167 .0096 .0063 53 .0054
At 50 and 128 R.P.M., the

the other velocites the bearing
coefficient
oil
supply was restricted, but at

flooded.
was
In
all
cases the
increased at revolutions
to
below 40
the
centre).
R.P.M., which
force
was probably due

to
ity,
the
decrease of
centrifugal
(the bearing being oiled
from the
This would seem

rubbing velocpivot
indicate
that
devices for reducing
relative
such as multiple washers (Art. 102),
may
be carried to an
flat
extreme, causing more friction than a plain

centrifugal action
is effective.
where
In the case of thrust collars, such

in
as
shown
in Fig. 91,
running
an
oil
bath, the surfaces themsimilar to that of the
selves tend to
draw
in lubricant in a
way
ordinary journal.
The
coefficients of friction for this class of
thrust should therefore be as low at least as those given above.

107. Bearing
Pressures
is
on Thrust
Bearings.
is
Where
the
velocity of rubbing
very low and wear
not important, as in
the case of swinging cranes, very heavy unit loads
may be
put
270
MACHINE DESIGN
bearings, especially
velocity
is
if
upon pivot Where the

portant,
they rotate in
an
oil
bath.
is
high, or even moderate,
and wear
im-
much
lower pressures must be carried with imperfect
lubrication, than
velocity.
on ordinary bearings running at the same With forced lubrication, as in the step-bearing shown
it
in Fig. 92,
tained.
is evident that very heavy pressure may be mainon the other hand, too many collars are used on a collar thrust bearing, in an effort to keep the bearing pressure down to a low value, there is danger that all of the collars will
If,
not bear simultaneously.
The
following are average values of
bearing pressures, for thrust bearings, as found in practice:
TABLE XVII
Mean
Velocity in
ft.
per Min.
Character of Lubrication.
Bearing Pressure in lbs. per Square Inch.
Very slow as
50 125 200 500
in to to to to
hand cranes
50 125 200 500 800
ft.
Up to
Bath Bath Bath Bath Bath
as as as as as
in in in in in
Fig. 87 Fig. 88 Fig. 88
2,000 to 3,000
Fig. 88 Fig. 88
200 i5 100
5.
75 to 50
Thrust Bearing and Bath

Lubrication as in Fig. 93.
Example.
bearing.
Speed in knots
Design the thrust journal for a steamship having
the following data, and estimate the frictional loss in the thrust
(1
knot= 6,080
ft.
per hour)
15
Indicated horse-power of one engine

Inside diameter of thrust collars
5>oo
14"
Outside diameter of thrust collars

Allowable pressure per
sq. in. of surface
21"
40
120
lbs.
Revolutions of the shaft per minute
Owing
to
frictional losses in the engine,
propeller,
is
and
shaft
only about two-thirds of the indicated power

thrust block.
delivered to the
The
pressure against the thrust block multiplied
by the distance through which -the ship moves per minute must
equal the energy delivered to the block per minute; or
thrust,
if
P be the
the speed of the ship in knots per hour,

2
/.
and the indicated
horse-power be denoted by
H.P., then -Xl.H.P. X 33,000
27
60
__
_ PxSx
60
6,080
QJ-
n _ 2 X I.H.P. X ^
3
33,000
6,080
X
2
I.H.P.
217
i
X 5 X
5
1
*7
Hence
in the
above example
r OOO X P =
72,300.
The
area of each thrust collar
= 4
(21
14
192 sq.
in.
There
fore the total allowable pressure
on each
collar
=192 X 40 =
and the 7,680 '

If the
number
of collars
oil
72,300
7 ,680
o.z J 10. y D or say
bearing runs in an
.01
bath, the coefficient of friction
will not
be more than
under the worst ordinary conditions.

.349
Therefore from
(4)
E =
f*
PN
.01
%"
r
\j
= .349 X
72,300
120
rio.5 3 -7 3 -| U10.5 a 7 I
1
405,000
ft.
lbs.
per min. or 12.3 H.P.
ROLLER AND BALL BEARINGS

108. General Consideration of Rolling.
in Article 29 that the resistance
less
It
due
to rolling friction
was pointed out was much
than that due to sliding
friction,
for a given load.
The
application of this principle to very heavy loads

as in the case of
and low speeds,

is
moving heavy bodies en
rollers,
of great
antiquity;
but only in recent years have mechanics been able to
produce surfaces of such a character as could carry even very

light loads at high speeds
on
either roller or ball bearings.
At
present, however, bearings of this character can be obtained

will
which
run well under very severe conditions.
upon any other surface with which it theoretically makes line or point contact, the two surfaces tend mutually to deform each other, the amount of deformation depending on the character and hardness of the materials
a curved surface rolls
When
forming the surfaces, and the intensity of the load sustained.

the surfaces of both
light, the
If
members
is
are very hard,
and the load
is
very
deformation
negligible
and true
rolling
can be practi-
272
cally attained.
MACHINE DESIGN
When, however, any appreciable load
and the load
is is to
be
carried
the mutual deformation of the
surfaces destroys the
theoretical line or point contact

surface.
borne on a small
a combination
is
This occurs even when the surfaces are very hard, and
is
the action instead of being that of pure rolling,

of rolling
and
sliding.*
The
It
true theory of this action, which

is
very complex, has not been fully demonstrated and
beyond the
it
scope of this
treatise.
can readily be seen that
is
closely
connected with the

character, which
elastic properties of materials,
on which much
of this
research work has been done.

is
Undoubtedly the work
of
most value
in the design of roller or ball
bearings,
is
that of Professor Stribeck
whose masterly report has

to this
been translated into English by Mr. Henry Hess,| and

translation reference will be
made
hereafter.
elastic limit of
If the intensity of pressure
be such that the
the materials
is
exceeded, permanent deformation will occur.
In the case of
roller or ball bearings this
may
result in the destruc-
tion of the surfaces either
by
flaking off locally, or
by simply
this
crushing out of shape.

character
is
In either case continued action of
destructive to the bearing.
Experiments on either
balls or rollers to determine the ultimate crushing load are, therefore,

is
misleading and useless as far as the design of such bearings

It
concerned.
appears from experiment and experience that

carry fairly
bearings of this character can be constructed to
heavy loads
at
high speeds for a long period of time provided

is
the intensity of pressure
not too great.
It
is
obvious from
the foregoing that the materials used in such bearings must be
homogeneous, and of uniform hardness.
modern
ball
and
roller
bearing has been
The made
success of the
possible
by im-
proved materials and workmanship rather than by new

Referring to Fig. 94,
rollers or balls,
it
theories.
is
evident that
when two adjacent
A and
B, touch each other, the directions of
motion of the
common
points of contact are in opposite directions.
It is often stated that this results in *
considerable frictional loss;

by
rolling a
The
student
may
demonstrate
this action
round lead pencil on
a piece of soft rubber under pressure.

f See Transactions A. S.
M.
E., Vol.
XXVIII.
and sometimes small intermediate
balls, or rollers, are
loss.
273
used as
inter-
shown
at
C, Fig. 94, to obviate the supposed
Such
mediate balls or rollers must be kept in place by a cage such as

, Fig. 94, and this cage will give rise to a greater frictional loss
than that which
it is
expected to remedy.
show
that very
little
pressure can possibly exist between
A and
B.
The
only pressures that can be exerted by the guiding surfaces

balls or rollers are in a radial direction or
upon the
surfaces,
normal
to the
and these have no component tending
to force the ad-
jacent rollers or balls together.
Sometimes the
,
rollers or balls are
separated by a guiding cage (see Fig. 95) and if any appreciable pressure could exist between adjacent rollers or balls the same
would necessarily
This theory
tional loss
is
exist
between them and
this
guiding cage.
not borne out by experience, as these cages, in
well-built roller bearings,
do not wear appreciably.

undoubtedly very small.
The
is
fric-
from
this source is
The
small,
friction of roller
this is a very
and
ball bearings while at rest
very
and
important point in the design of heavy,
slow-moving machinery where, with ordinary sliding bearings,

it
often takes a
much
greater effort to start the machinery

full
from
rest
than to maintain motion at
speed.
ROLLER BEARINGS
109.
Forms
of Bearings.
Roller bearings, in
common
with
the ordinary bearing, are classified as radial or thrust bearings,
according to the manner in which the load

typical
is
is
sustained.
form of construction of
in Fig. 95.
,
roller
bearings for radial loading

steel,
shown
shell of
hardened
it.
B, surrounds
against
the shaft
this shell
and
is
secured firmly to
The rollers C bear

is
B, and against an outer shell D, which
secured to the
bearing proper, E.
Both
rollers
and
shells are usually
made
of
high carbon steel hardened and ground, or of mild steel case-
hardened.
The
rollers are held parallel
with the axis of the shaft
by means of a cage F which is made of brass or other soft material. Some form of cage is necessary in all roller bearings on account
of the tendency of the rollers to twist out of line with the shaft,
thus replacing the theoretical line contact with point contact, and
18
274
also causing
MACHINE DESIGN
an end pressure and cramping on the rollers. This end thrust is sometimes provided for by putting a
If
tendency
to
small ball at each end of the roller to act as a thrust bearing.

the axis of the roller
is
not parallel to that of the shaft,

it
it
cannot
make
line contact
with the shaft unless
assumes a
spiral form.
Fig. 94.
If the surfaces
which confine the

very small, as
it
roller are accurately
made, and
the clearance
is
should be, the roller cannot get
out of parallelism with the shaft without being bent into a spiral
form.
If the rollers are
if
hardened
this
may
result in fracturing
them, especially
they are relatively long.
To
obviate this
Fig. 96.
Fig. 97.
trouble the rollers are sometimes

at
H,
in Fig. 95, or the roller is

roller
shown in Fig. 96. steel strip spirally upon a mandrel, thus making a hollow flexible roller. It is to be especially noted that neither of these methods For continuous will compensate for inaccurate workmanship.
Hyatt
made in short lengths, as shown made flexible as illustrated by the This roller is made by winding
line contact the outer
2 75
and inner
great accuracy, placed in
must be machined with very accurate alignment, and the rollers

shells
must be guided
so as to
remain perfectly parallel

to
to the shaft.
These conditions are

service.
difficult
obtain
initially,
and almost
impossible to maintain with great accuracy under continuous
The
97.
rollers
in
bearings for radial
loading
may be
adjustroller is
cylindrical or they
in Fig.
may
be conical as in the Grant bearing shown
The
construction here
is difficult
shown permits
of
ment
for
wear, which
to obtain
where the
is axial,
cylindrical.
If the direction of the load to
be carried
roller thrust
bearings of the form shown in Fig. 98 are often used.
The
shaft
PLAN OF CAGE
Fig. 98.
Fig. 99.
carries a thrust collar
B and
the thrust
is
taken on the frame of
the
ring
machine by a corresponding collar C. A hardened steel D is attached to B and rotates with it, while a similar ring
fastened to the stationary part C.
rings, carrying with
is
The
conical rollers
them the cage F. A thrust ring prevents the rollers from moving radially outward. The apex angle of the roller should not exceed 15 and in most cases is kept down to 6 or to prevent serious end pressure against 7 this retaining ring. It is evident that where the roller is conical
roll
between these
in
form, the apex
oj the cone
lying in
its
tlie
centre line of the shajt,

is
the velocity of any point in
periphery
proportional to
its
distance from
will
the axis o! the shaft and, theoretically, true rolling
be obtained.
276
MACHINE DESIGN
Bearings of this character with conical rollers are expensive

to
make
in
an accurate manner, and a simpler form, as shown in
Fig. 99, is
sometimes used. Here the rollers are cylindrical in form and are made in short lengths so as to reduce relative
slipping.
The
outer rollers rotate faster than the inner rollers,

of the rollers are such that
and the lengths and arrangement ridges are not worn in the seat.
roller
Space does not permit of discussion of the

are the same,
many forms
of
bearings on the market; but their fundamental principles
and the student
is
referred to current trade cata-
logues for variations in methods of construction.
no. Allowable Bearing
Pressures.
It
is
evident that the
bearing pressure in roller bearings must not be great enough to

stress the material of either roller or
elastic limit,
bearing surface beyond the

little
but theoretical considerations are of
service in
the actual designing of such bearings.
The most
reliable ex-
perimental data bearing on the subject are the results of Stribeck's
work.
In
roller
bearings under radial pressure the load
is
con-
sidered as carried on one-fifth of the total
number
of rollers;
and
the quantity equivalent to the projected area of the ordinary
bearing, as far as carrying capacity

as the product of length
is
concerned,
is
considered
and diameter
of a single roller, multirollers in
plied
by
one-fifth the total
number
of
the bearing.
if
Thus, according
to Stribeck, for cylindrical bearings,
N= W=
w =
d
total
number
of rollers.
in lbs.
lbs.
total load
on bearing
load on one roller in
diameter of roller in inches (mean diameter for conical

rollers)
= length of roller in inches. k = a constant to be determined Then w = k d

I
I
experimentally.
(1)
(2)
and
N d
for
From Stribeck's* experiments

rollers
k has a value of 550 for unhardened
and bearing surfaces and 1,000

* See Transactions A. S.
hardened surfaces.
444.
M.
E., Vol.
XXVII, page
In the case of thrust bearings the load
distributed over the total
277
considered as
may be
number
of rollers.
Bearings of the
type shown in Fig. 99 have been constructed to carry a load of
156,000 pounds at 250 revolutions per minute.
BALL BEARINGS
in. Theoretical Considerations.

roll
Let the ball A, Fig. 100
(b),
along the circular path * B, with pure rolling motion, making
point contact with the path.
Let the path

plane.
be parallel
it rolls
to the
plane
CD, and suppose

from
this
also that the ball as
remains a
if
fixed distance
same
Then
it
it is
evident that
rolls
with pure rolling motion along B,
will rotate
around
P
jL
/A
/
6
a
ci
e-
Nj^\
/
fe
-b
>
Fig. 100
(b).
Fig. 100 (a).
some one
will
of its diameters, at right angles to
B, as an
axis,
and
make
contact with
it
along the edges of such a disc as would
be cut from
by a plane passing through the point of contact
b perpendicular to the diameter around which the ball rotates.
Thus
rotate
the ball
may
rotate
i.
around Ok as an
It is clear,
axis,
and
roll
along
the edges of the disc b
however, that the ball can
around only one diameter
at a time,
and preserve true
has two concentric paths of whose points of contact with the ball are b and e (Fig. 100 a) respectively, then it must roll along two discs b-i and e-l, and these discs must have a common axis of rotation O k perpendicular to their planes and passing through Further, the discs must be so placed that the centre of the ball.
rolling contact with B.
If the ball
(b)
contact as
and E, Fig. 100
The
guiding surfaces of ball bearings are almost invariably circular in form.
278
the lines
i I
MACHINE DESIGN
and
b e intersect
on the
line
m, passing through the
common
centre of
and E;
for then
Yl
rb
YV
may be
or the circumferences of the rolling discs are proportional to the
circumferences of the paths of contact, and true rolling

attained.
It is
not possible to have more than two points of

its
contact between the ball and one of
guiding surfaces, with
pure
rolling, as the proportionality given
above
is
not true for
any other points on the
line
b except those given.
The above
with
principles are fundamental

circular guiding surfaces.
and apply
to all ball bearings
112. Spinning.
Usually one of the guiding members
is
fixed
and the other rotates, the friction between the moving member and the ball causing the latter to roll. If the load carried is so small that no distortion of the surfaces takes place, and true point
contact exists, this frictional force will act tangent to the outer
circumference of the disc of contact and be parallel to
its
plane.
Such
load
theoretical conditions never exist in practice, as the sur-
faces of contact are deformed, even under light loads,

is
and the
fric-
carried on a small area instead of a point.

the
ball
is,
The
tional force rotating
hence, indeterminate and in
general has components which tend to rotate the ball
about
other axes than the one which will give pure rolling motion.
It is clear that
inaccurate workmanship will give the
same
result.
This action
is
known
as spinning
and
is
necessarily
accompanied
by
friction.
113.
Forms
of Bearings.
Ball bearings are divided into three
types, according to the character of the load
and the way
it
is
sustained by the bearing

(a)
Radial bearings, for loads acting at right angles to the
shaft.
(b)
Thrust bearings,
the shaft.
for loads acting parallel to the axis of
(c)
Angular bearings,
for
taking loads both perpendicular
and
parallel to the axis of the shaft.
Each of
by the
these types
279
maybe
either a two-point, three-point or four-
point bearing, depending on the

ball
number
of points of contact
made
on the guiding
race
surfaces.
114. Radial Bearings.
Figure 1 01
(a)
shows a two-point radial
oearing.
is
The
is
secured to the shaft A, while the race
secured to the other
member
C.
Either
or
C may
be the
at E>
If the
rotating part.
In order to place the balls in the raceway an
opening
race
is
often cut in the side of one of the races, as

filling
shown
and the opening then closed with a
piece as shown.
is
stationary this filling piece can be located on the unit.
loaded side and no wear brought upon
If
is
stationary the
in locating the
opening must be cut in
it,
and the same care used
Fig. ioi
(a).
Fig. ioi (b).
filling
piece with reference to the load.
If
both the shaft
A and
load
is
hub
rotate this cannot be accomplished,

filling piece,
and the
full
brought upon this
thus decreasing the capacity of the
bearing to sustain a load, on account of the break in the surface

of the race.
If the velocity of the rotating
is
member
is
high this
break in continuity of the race

If
fill
destructive to the bearing.

of balls necessary completely to of
about half the

is
total
number
is
the race
used, each race
may be made
filling in
one
solid piece.
In such cases the bearing
assembled by moving the inner race

the balls
over eccentrically to the outer race,

distributing them.
in
and then
Separators of elastic material are then pushed

to
between the balls
maintain correct spacing.
These separat-
ors, also, often act as reservoirs for lubricant.
They may be
of
2 8o
MACHINE DESIGN
or such soft material or
felt
may be made
for
in the
form
of a helical
spring.
This construction
of balls is
is
showninFig. 101
(b).
The
lessened
number
compensated
by using
balls of larger
diameter and hence greater carrying capacity.
The
carrying capacity of radial ball bearings, according to

is
Stribeck's experiments,
not affected materially by velocity,

is
within reasonable limits, so long as the velocity of rotation
uniform; but sharp variations of velocity at high speed reduce

the capacity.
115. Thrust Bearings.
Fig. 102 illustrates a four-point thrust
difficulty in filling in the balls
<j>
bearing.
Here there
is
no
when
the
races are solid.
In Fig. 102 the angles

it is
this is not necessary as
and $' are equal, but evident that any line drawn through
Fig. 102.
which
and intersecting the ball circle will locate a pair of rolling discs will roll on B, without interfering with the pair shown which may roll on A.
The
parallel.
initially,
surfaces
and
are sometimes
made both
flat
and
It is difficult,
however, to obtain absolute parallelism,
between C,
D and the ball races, and much more difficult

An
error
to
maintain this parallelism under running conditions.
in alignment, either
from poor workmanship or deflection under on one

to
side.
If possible, therefore,
load, of less than one thousandth of an inch will cause concen-
trated loading of the balls
such bearings should be seated on spherical surfaces, as shown

at
D, thus allowing the races
adjust themselves correctly.
Mr. Henry Hess
states that speed is
28l
in
an important factor
such
bearings and he gives 1,500 revolutions per minute as a max-
imum.
simple form of ball thrust bearing

balls
is
shown
in Fig. 103.
Here the
run against
flat
hardened surfaces,
A and
B, and
some soft alloy. The cage may be made to retain the ball loosely by drilling the openings for the balls almost through as shown in Fig. 103 (b), inserting the ball and then closing down the upper edge a little with a set as shown at e, Fig. 103 (b).
are kept in position
by a cage
C made
of
116. Angular Bearings.
If possible, radial
loads should be
supported by radial bearings, axial loads by thrust bearings,
and angular bearings should be avoided.
Radial bearings should
not sustain heavy axial loads and thrust bearings should not be
loaded axially.
For
light loads the
angular bearing will sustain
pressure in either of these directions.
forms of angular bearings.
Fig. 104 (a), (b), (c),
There are innumerable and (d) may be

and are often adputs the
taken as typical of two-, three-, and four-point angular bearings.
The
races can be
made continuous
is
in all cases,
justable.
This
to
last feature,
while sometimes necessary and often

really a detriment as
it
claimed
be an advantage,
bearing at the mercy of an unskilled person.

ball bearings
Properly designed
if
do not wear appreciably, and
wear does take
place
it
will
occur on the loaded side only; and adjustment cannot

for this,
compensate
It
is
but only hastens the failure of the bearing.

all
evident that
the arrangements
shown
in
Fig.
104
282
fulfill
MACHINE DESIGN
the requirements for pure rolling contact as outlined in
Art. 109.
The path
of the ball
is
not so definitely determined
at a, Fig. 104, as in the
other forms.
For
this
reason the radius

of the ball,
ball.
of the ball races should, in order to prevent
wedging
not be greater than three-quarters the diameter of the

the
For
same reason the angle

2
.
<j>
in Fig. 104 (b) should not be less than
about
cally,
In Fig. 104
(b)
and 104
it
(c)
the point a may, theoretiroll
be anywhere, as long as
It
lies
between the discs which
on the outer raceway.

117. Allowable Load.
should be so placed, however, as

c.
nearly to equalize the loads at b and
The
allowable load which
may be
races.
put upon a ball bearing will depend on the following

(a) (b)
(c)
The character of the materials forming the The shape of the raceways. The diameter of the balls.
Ball bearings
balls.
fail
balls
and
by over stressing the material of the induced is far beyond the elastic limit, and often repeated, the surfaces will flake off and failure will occur. Experiments on the crushing strength of balls or races are useless and misleading as the life of the bearing depends on the elastic and not the crushing strength. Evidently none but hard materials can be used for appreciable loads and these must be homogeneous in texture. Case-hardened materials are of doubtful value for severe service. For most trying circumstances special steels and alloys will no doubt be much used. (b) Theoretically, a ball supports the load on a point, but
(a)
raceways or
If the stress
practically the
unavoidable distortion of the material increases

It
the point to a small surface.

matically,
can be demonstrated mathe-
and
is
evident on reflection, that a greater bearing
surface will be formed for a given distortion of ball and ball race
the
more
closely the cross-section of the ball race corresponds to
the cross-section of the ball.

rect
On
the other hand, and as a diit is
consequence of
this increase of surface,
found that the

approaches the
friction increases as the cross-section of the races
cross-section of the ball, a result to be expected.

It is
almost impossible to machine and adjust ball bearings
of three- or four-point contact so that the load is uniformly dis-
tributed at the various points of contact.
It
is
283
borne out by
experiment and
it is
well
known
that two-point bearings can carry
heavier loads, than any other form for a given diameter of ball.
(c)
The
allowable load which a ball can carry varies with the
square of the diameter.
These
Stribeck,
statements
have
been
proven
experimentally
by
who found
that the carrying capacity of a ball could be
expressed by
w = kd
where
2
.
(1)
w =
k
greatest load on one ball in pounds.
= =
a constant
depending on the material and shape of
ball races.
diameter of ball in inches.

is
Stribeck assumes that the total load

total
carried on one-fifth of the
number
of balls.
pounds on one row
of balls,
W be the total N the total number of N N = kd? W = w

If,
therefore,
load in
and
balls,
(2)
For hardened
k
steel races to
made
of
good quality of
steel
450
750
for
for flat or conical races, three- or four-
point contact.
k
=1,500
two-point contact
and raceways whose

d.
2 radius of curvature equals /i
With more
perfect materials Stribeck states that these values

fifty
may be
increased
per cent.
It
is
118. Practical Considerations.
clear that in order to
insure an even distribution of load, initially, the
workmanship
in order to
on both balls and races must be very accurate; and
maintain this distribution the material must be uniform in quality
and hardness throughout.
It is also
found
that, for best results,
the surfaces must be highly polished
and
free
from scratches.
The bearing must be kept free from acid and rust and provision made for excluding dust and grit and for retaining a supply of
lubricant, the function of the lubricant being largely to prevent
rusting.
284
MACHINE DESIGN
stated,
it
As before
one row of
of
has been found better to carry the load on
balls, if possible.
Where
this
cannot be done special
provision should be
balls carry its
made
to insure that
each of the several rows
proportionate load.
This usually leads

is fixed,
to
some form
by the load
ard
sizes.
of equalizing-device
which complicates the design.

approximately,
in stand-
The minimum
diameter of the shaft
carried,
and
balls are
made, commercially,
full
In designing a bearing for a
tentative computation
must generally
of the balls.
number and diameter

diameter of the
circle
number of balls a be made to fix the proper Knowing these, the exact
passing through the centre of the balls can
be determined as follows
(a),
any two
balls in contact as
i,
H and O
as shown.
draw a line connecting the centres of G and H, and draw the radii O G, Also let r be the radius of the ball and R
the radius of the ball circle.
Then
and
= RsinO
180
R
.
180
This value must be increased

clearance.
sufficiently to allow for the necessary
The methods
varied, that
of applying ball bearings are so
no attempt can be made here

is
to illustrate
the student
this point
referred to the following sources of
numerous and them, and information on
Transactions of A.
S.
M.
E., Vol.
XXVII and
Vol.
XXVIII.
Trade publications
generally.
CHAPTER XI
AXLES, SHAFTS, AND SHAFT COUPLINGS
The terms axle, shaft, and spindle are applied 119. General. somewhat indiscriminately to machine members which are so constrained by journals and bearings as to admit of motion of These rotating members may be subjected to simple rotation. torsion or bending, or to combinations of torsion and bending.
Shear, also, usually exists as
in
the
case
of
loaded beams.
Rotating members
to the
may
be
classified
roughly as follows, according

26), or to the
predominating
stress (see Art.
particular
purpose for which they are intended.

(a)
Axles, loaded transversely
and subjected principally
to
bending.
(b)
Shafts,
subjected to torsion or combined torsion and
bending.
which directly carry a tool for actually doing work, and which as a consequence must have
(c)
Spindles, or short shafts
accurate motion.
The
(a);
axles of railway freight cars are
good examples of case
transmission shafting in factories, or the shafts of steam

(b)
;
engines are good examples of

spindles illustrate
(c).
while lathe and milling-machine
Considerations of strength seldom enter into the design of

spindles.
In these
members
is
torsional stiffness
and accuracy of
form
tions.
in the bearings are, usually, the
most important considera-
When
in
the spindle
there
is
designed with these latter require-
ments
view,
usually an excess of strength against
rupture.
The
discussions given in Art. 12 apply in this case,
and
it
will
not be considered further here.
120. Axles.
Let
(Fig.
is
105) be an axle
which carries the
loads
Pn P
and
3,
but
not subjected to any torsional stress
286
MACHINE DESIGN
Suppose the axle
except that due to negligible bearing friction.

to
be supported by the bearings

is
N and N.
The
distribution of
the bearing reactions
and the assumption is This assumption is the middle of the bearings, as indicated. on the safe side, so far as the strength of the shaft is concerned, as
the slightest deflection of the shaft tends to concentrate the re-
indeterminate, as explained in Art. 95, usually made that they are concentrated at
action at the inner edge of the bearing.
The
axle can, therefore,
be treated as a simple
beam
(Art. 14).
If the load
were zero,
and the loads

(which
is
and
were equal and symmetrically placed
the most usual condition, as in car axles), the case
space diagram
Fig. 105.
would be
structive,
identical with
Case
XIV
of
Table
I.
It will
be in-
however, to
make
a solution of the general case given
above.
The
bending.
principal stress to which the axle
is
subjected
is
simple
Shear also
exists in every section;

it is
but from the general
theory of
beams
(Art. 14)
known
that, usually, this latter

If,
may
is
be neglected in the body of the

short,
shaft.
however, the shaft
and consequently need not be large to withstand the applied bending moment, the section of the bearing at XX should be
checked for shearing
will
it is
stress.
be where the bending
The dangerous section of the shaft moment is a maximum, and hence
necessary to determine this
maximum moment, which
also
AXLES, SHAFTS,
involves
reactions
the
AND SHAFT COUPLINGS
287
determination of the unknown reactions. The may be determined mathematically by taking moments
.
around
Then,
RJ = P
. .
+ PJ + P,l P ! + P l + PJ *t x
and
R =P +P +
2
x
P.,
- R
The bending moment at any section is the algebraic sum of all Thus the the moments at either side of the plane considered. = = (I l P (l l and R bending moment at P 2)
2
2)
this value
may be
used in equation J of Table VI
to
determine the stress for a given cross section, or to determine
the cross section for an
assumed
stress.
A
at
graphical solution
once where the
is much more convenient, as it shows maximum bending moment is located. In
Fig. 105, denote the forces P,,
2,
etc., thus, ab, be, cd, etc.,
and
draw
x ,
the corresponding force

2,
diagram as shown, making

to
AB =
P BC = P
and
CD = P
3,
any convenient
scale.
It is to
be noted that these forces are drawn consecutively downward,

since they act in that direction,
and
their
sum,
AD, must
equal
the
sum
of the reactions, or vertical forces.
pole, as O,
ab, in the
and draw OA, OB,

in the force
OC
Take any convenient and OD. From any point on

ob, parallel
space diagram, draw oa and
respectively to
OA
and
and
be
OB
diagram.
From
the intersection of ob
draw
oc, parallel to
OC, and
in similar
manner draw^ od.

parallel to oe,
Join the intersection of oa and ea with the intersection of od and

de,
thus locating the closing string
oe.
Draw OE
locating E.
to the
Then
in the force
diagram
DE = R
2,
and
EA =
P,
assumed
scale of the force diagram.
The
to the
vertical ordinates of the space
polygon are proportional
bending moments
at the points considered.

is
The numerical
value of any bending
moment
the continued product of the
length of the ordinate, the perpendicular distance of

the reciprocal of the scale of the space diagram,
from AD,
reciprocal
and the
288
MACHINE DESIGN
Thus
if
of the scale of the force diagram.
the ordinate at
some
point be 2" long, the pole distance be 2X", the space scale be
1%"
2]/ 2
to
1 ft.,
or yi
size,
and
1"
5,000 lbs. on the force diagram;
then the bending
moment
at the point considered is
M=
5,000
200,000
inch lbs.; and from this
moment
the diameter of the shaft
may be computed.
The fundais
121. Shafts Subjected Principally to Torsion.
mental relations existing in a shaft which
subjected to torsion
only have been fully discussed in Article 12, and for such cases
or where other stresses, such as those due to bending, are negligible, Article 12 is applicable.
Shafts subjected to pure torsion

is
rarely occur in practice, as bending
almost always present

to the
due
to the weight of the shaft itself,

it
and
weight of pulleys
which
supports, as well as to belt pull, etc.
There are many

predominant, and
cases, however,
where the torsional
stress is
where the secondary bending

in long factory shafting,
at
effect is difficult to
is
compute.
Thus
where the power
supplied to the shaft
one point, and

all
is
given off in small increments at short interto^
vals
is
along the shaft, the bending due
the pull of the belts
small.
This
is
especially true
if
care is exercised to place the
pulleys as close to the bearings as possible.

If the shaft is of considerable length, the
is
angular distortion
of importance,
and
it
may
often occur that a shaft having
sufficient torsional strength will

If the
not have proper torsional
stiffness.
power
is
applied at one end of the shaft, and taken off at
the other end, computations for both strength

easily
ever,
and
stiffness are
made and may be
of service.
In nearly
all cases,
how-
power is delivered in varying quantities all along the shaft, and such computations are not only difficult to make but would
its
indicate that the diameter of the shaft should vary at different
parts of
length.
This would be undesirable, as

etc.,
it is
important
that shafting, hangers,
should, as far as possible, be uniform
and interchangeable
the driving point
for
convenience and economy;

it
and the
practice of reducing the diameter of the shaft as

is
extends from
confined to larger shafting (say over 3" in
diameter).
The
design of shafts subjected principally to torsion,
therefore, is usually based
on the formula
for torsional strength,

AXLES, SHAFTS,
AND SHAFT COUPLINGS
289
modified by practical coefficients which experience has shown

will
provide for stiffness against torsion and bending.

Referring to equation E, Article 12,
'-^-'-J?*
If
w
and
be the equivalent force applied
at the periphery of the shaft,
so that
if
T =
Pr, where
r is
the radius of the shaft in inches;
be the number of revolutions of the shaft per minute; then
the horse
power transmitted
will
be
H.P. =
2PmN
33,000
2TnN
33,000
12
12
or
T =
33,000
X 12 X H.P. 27CN
63,024
H.P.
Substituting this value of
in (1) above,
A
where
is
X
p
s
N
If shear-
a constant
depending on the
stress assigned.
ing stress alone

as 9,000 lbs.
were to be considered, ps might be taken as high per square inch, for steel shafting. In order to
secure stiffness, and to provide for the indeterminate bending in

line shafts,
it
is
customary
to
assume a lower
stress (or higher
factor of safety),
for
depending on the material used, and the service

is
which the shaft
intended.
The
larger
and more important

stress, as the failure
the shaft, the lower should be the

of a
working
is
head shaft or shaft of a prime mover
accompanied by great
inconvenience and expense.
The
following factors of safety are
indicated by successful practice:
For head shafts

"
line shafts carrying pulleys
. .
15
.10
etc.
7
"
small short shafts, countershafts,
For
steel shafting, the
allowable stress for the above factors would
be about
'9
4,000, 6,000,
and 8,500 respectively, whence
290
For head
shafts,
MACHINE DESIGN
For
line shafting carrying pulleys,
<Z
3-75^^
3-3^?Lp
(4)
For small short
shafts, countershafts, etc.,
(5)
It
must, however, be borne in mind that a universal rule

class of shafting;
cannot be laid down for any
and cases
v/ill
always arise which need further consideration than given by the
above equations.
the
For example,
span
in the
span of shafting where
power
is
applied by a large belt the bending action

this particular
excessive,
eter to
and
than the remainder of the
may be may have to be of larger diamshaft. The student is referred

on the
size of
handbooks*
for tabulated data
transmission
shafts for various purposes.
It is to
be especially noted that a
shaft carrying a transverse load, which applies a bending
moment
to the shaft, is subjected to a reversed stress as the shaft rotates.

If,
in addition, the twisting
moment
high,
varies in magnitude, the

stress
factor of safety,
(see Arts. 25
owing
26),
to
complete or partial reversal of
and
must be
and
this
accounts for the
low
stresses allowable with
such shafts.
122. Shafts Subjected to Torsion

shafts,
and Bending.
In engine
head shafts driven by heavy belts, and many others, the torsional stress is not predominant and may, in fact, be less than
that
due
to bending.
full
discussion of the relations which

it
exist in this case has been given in Article 16 and
remains
to
show the application
of this discussion to actual cases of design.
From
that
if
Article
16 (equations K,
and
Fig. 9),
it
appears
for
the bending and twisting
moments can be determined
any
section, the theoretical diameter of the shaft at that section

* See Kent's "Mechanical Engineer's Pocket Book," page 869.
AXLES, SHAFTS,
can be found.
without
AND SHAFT COUPLINGS
29
Usually the twisting

but the bending
difficulty,
moment can be determined moment is often difficult to

is
determine, and sometimes the designer must be content with an
approximation.
One
of the greatest sources of uncertainty
the location of the reactions at the bearings.
Usually, as already
pointed out, the safe assumption
is
made
that these reactions are
concentrated at the centre line of the bearing.

is
When
the shaft
is
of appreciable length (15 or 20 diameters), the error
small;
but in
such cases as the crank shafts of multiple-cylinder engines,

is
where the distance between the centres of bearings

or five diameters, or
less,
it
only four
is
is
evident that the assumption
in the direction of excessive safety.
In line shafting, particularly with the usual swivel bearings,

the error from this source
of such shafting
is
small,
to
and
at first sight the conditions
would appear
approximate those of a con-
tinuous beam.
While such an assumption might be safely made

it
when
the shafting has been put in perfect alignment,
would not
soon
be safe as a general principle, as perfect alignment, even under

best conditions,
is
of short duration,
and bending
It
stresses
appear as a result of lack thereof.

in this case, the safest
appears, therefore, that,
procedure would be to treat each span
as
if
disconnected at the bearing,
when computing bending

is
moments.
A
shaft
typical case of
combined twisting and bending

106
Art.
(a),
5.
the engine
shown
in
in Fig.
(c),
the data taken being those of the
example
Case
Here the
shaft
is
supported by the
carries a
bearings at the points
X and X',
downward
lbs.
as indicated,
and
heavy
of
generator spider at Y.
the shaft
itself,
The weight
a
of this spider,
and that
with the probable magnetic pull which

little,
may
occur
when
lbs.
the shaft wears
is
estimated at 22,000
The maximum
is
is
pressure (P) on the crank pin, due to the
steam pressure,
the crank
25,000
This force
is
maximum when
it
about
vertical, and, at that position,
exerts a twist-
ing
moment on
is
the shaft
from the crank
to the point
F*
where
in a
power
delivered,
and
also a
bending moment on the shaft

hub
of the
The
reinforcing effect of the
.s;>i<lrr is
neglected.
292
horizontal direction.
MACHINE DESIGN
The weight of the moment in a downward
generator,
direction,
etc.,
exerts
a simple bending
and
at right
angles to that induced by P.

the direction
Fig. 106 (b) shows, isometrically,
and point
of application of the various forces
and

AXLES, SHAFTS,
reactions,
AND SHAFT COUPLINGS

maximum
how
293
equivalent
and
it
is
required to find the

shaft.
(c),
bending moment on the

It
was shown
in
Example
for the
Article 99,
a tentative solu-
tion could be
made
diameter and length of the main jour-
nal, thus fixing the distance of its centre line
from the centre
line
of the crank pin at 21
K"-
Other data
fix
the distance between
bearings as
7'
9".
is
Graphical Analysis
here very convenient, and the order of
procedure
(a) (b)
(c)
will
be as follows:
to the
Find the bending moment due Find the bending moment due
steam pressure P.
to the
dead load W.
Combine
these bending
moments
to find the
maximum
all
resultant bending
(a)
moment.
the
Consider, for convenience, that the force P, and

it,
reactions due to
so that
have been rotated into the plane of the paper
is
represented as acting vertically.
Draw
the force *
to
it,
diagram, O' B' A'
C for
force P,
and the reactions due
to
a convenient scale, here taken as 8,000 lbs. per inch, taking
on a horizontal
of the space
later
line
through A', thus making the closing string

also
diagram
horizontal,
work.
Draw
1 ft.
the space diagram,

it,
MNP
which
is
convenient for
P, and
for force
the reactions due to
as shown.
The
scale of the space dia-
gram
for
is
Y^" to
or yV size.
(b)
In a similar manner construct the force diagram

r
f
BC O,
force
for the
T J I, W, and the corresponding space diagram force W, making the pole distance = A O taken here
\
as 3 ".t
(c)
To combine
S
the bending
moments
distance
at
any
it
section, as Z,
take the intercept

the
to
T, on P.
I J, and lay
T'
off
is
as S'
on
diagram
the
MN
The
proportional
combined bending moments and may be used as an ordinate VV in the diagram of combined bending moments
DGFE.
It
often occurs
that the shaft carries a heavy flywheel at Y,

* See also Article 120. f
Reduced
to one-half size in cut.
294
MACHINE DESIGN
may
also
instead of a generator, and a heavy belt
run on the
weight
wheel.
It is evident that the resultant force
due
to the
of the wheel
and the
pull of the belt, can be determined, both in
magnitude and
will not
direction.
In general, the direction of
this force
be
vertical,
but will
make an
angle,
<p,
less
than 90 with
the direction of the force P.
In such a case the moments
may be
combined by the
angle
<.
triangle of forces taking into consideration the
The
numerical value of any
moment
it,
is
the continued product

re-
of the ordinate
which .represents
the pole distance, the
ciprocal of the scale of the space diagram,
and the reciprocal
of
the scale of the force diagram.
Thus
the
maximum
bending
moment, which occurs
at
Y =
The
value
is,
iT
yX3X
V6 X
is
485,400 inch pounds.
twisting
moment
seen by inspection to be uniform

it
over the whole length of the shaft which

as before, 25,000
affects.
Its
lbs.;
numerical
18
=
to
450,000 inch
and these
two moments may be combined,

of the
cle 16.
determine the safe diameter
main part of the shaft according to the methods of ArtiA graphical method will be given later, which somewhat facilitates the numerical work of this computation. The methods outlined above are clearly applicable to any shaft which has not more than two points of support since in such
cases the reactions can be readily found.
A
to
to
convenient diagram
is
shown
in Fig.
107 for determining
the diameter of a shaft, of solid circular cross-section, subjected
any moment, and with any intensity of

15,000
lbs.
fibre stress
from zero
per sq. inch.
This diagram can be used for

for
either simple
bending or twisting moments, or

Its use in
combined
first.
bending and twisting actions.
connection with probwill
lems involving simple twisting moments

If
be discussed
is
the twisting
moment, d the diameter

intensity of stress in the
of the solid
circular shaft,
and p the
3
.
most strained
fibres,
T =
pd
Therefore, for a given diameter of shaft,
296
is
MACHINE DESIGN
Thus,
if
directly proportional to p.
4",
64,
and
the
T =
inch
scale
.196
lbs.
X 64P =
12. 57^.
if
If
p be taken as 10,000,
T = 125,700
(to
if
In Fig. 107
ordinates represent
to each division);
moments
and
"A"
of 500 inch lbs.
abscissas
lbs.
represent intensity of stress (to the upper scale, "5," of 1,000
per sq. inch to each division), the point a corresponds to

125,700, p
T =
10,000,
4".
As
the
moment
varies directly
of shaft, the
as the intensity of stress, for

relations
shaft)
a,
any given diameter
between corresponding values of

be represented by the straight
origin O.
and p
(for a 4"
will
line
through the point
and the
In a similar manner straight lines through

for other shaft diameters.
the origin are
drawn
To
inch
determine the diameter of shaft for a

with a fibre stress of 12,000
lbs.
moment
(or
of 90,000
lbs.,
per sq. inch, pass along
the horizontal through the point
marked "9"
T =
90,000)
on scale "A,"
(or
tal
to the vertical line
through the point marked " 12"
p = 12,000) on and vertical (b)
scale "J3."
lies
The
intersection of this horizon-
at its
below the diagonal marked 3.4 outer end; or the shaft should be about 3.37" (or 3H")
a
little
diameter to give a stress of 12,000
lbs.
per sq. inch.
The
oblique line nearest to the point located in the last

figures, viz.:
example bears three
".732
1.58
3.4,"
and the
signifi-
other diagonals each bear three separate figures.
The
cance of these designations will be explained by further

tions.
If
illustra-
T = TV T = 16
7:
of
90,000,
or 9,000 inch lbs.,
and p
12,000,
f/fm
,
j d
16
a| y 12,000
a
*
90,000 v
n >
aJ
3-37
Vio =
1.56
10
since d varies as the cube root of T,

3-37"if T = 900, = .726". Vô" To use the diagram when T =
and when
T =
90,000, d
=
-f-
In a similar way,
or T
of 90,000, d
3.37
900,
and p
12,000, consider
scale
A"
as representing the
moment
in 100 inch lbs.;
pass
along the horizontal through 9 on this scale to the vertical, through u b" and take the first 12 of scale "B," as before, to the point
AXLES, SHAFTS,
figure borne
AND SHAFT COUPLINGS

(.732)
297
by the nearest diagonal

or,
as the approximate
diameter of the shaft;

.726".
If
by interpolation, find the diameter

consider scale
T =
9,000,
12,000;
"A"
as representfigure
ing the
moment
in 1,000 inch lbs.,
and read the middle

is
on
be
the nearest diagonal (1.58) as the required approximate diameter

of the shaft;
1.56".
If the
or,
by
is
interpolation, the diameter
found
to
moment
12,000,
greater than 130,000, the
diagram
if
is
quite
as applicable as for smaller moments.
Thus
as
T=
"5"
7.
900,000
the
and
scale u
at
consider
scale
"A"
representing
moment
b
,}
in
100,000 inch
lbs.
The
horizontal
through 9 of
intersect
"A"
and the
vertical
through 12 of scale
is
as before.
The
required diameter
about
26";
be-
cause the diameter was found to be about .726 for a

of 900,
moment
must be 10 times as great for a moment of io 3 X 900 = 900,000. For p = 12,000 with a moment of 9,000,000 3 inch lbs. '(= io X 9,000), the diameter is 10 X 1.57 = 15.7",
and
it
etc.
It
thus appears that the diagram covers

it
all
moments, withif it
out being of such impracticable size as
would be
were not
for the peculiar designation of the oblique lines
and the method
"A." The diagram can also be used for simple bending moments. The expression for the bending moment in
of using scale
a shaft of solid circular section
is
= pd";
32
while the expression for a twisting
moment
is,
as given above,
Therefore, with a given diameter and numerically equal fibre

stress,
is
numerically equal to 2M.
To
determine d for given

the equivalent T,
values of p and M, multiply
M by
2 to get
and
with this value of T, proceed as in the former examples.
For finding the diameter appropriate to a combined bending and twisting moment, the equivalent twisting moment,
T =
VM
298
is to
;
machine design
be determined see Art.
1 6,
equation
ing
moment
is
readily determined
This equivalent twist4 from the diagram by the use of

.
scale
"C"
at the
bottom
of Fig. 107
and a pair
of dividers,
the simple bending

(T) are given.
moment (M) and

Suppose
Consider scales
the simple twisting
when moment
40,000;
Example:
M
M
on
2
30,000;
U
T =
to
and p
13,000.
"A" and
at 3
scale
C"
measure
with
moments
in 10,000 inch lbs.
Take
on
scale
"i"
one point of the dividers, and
at 4
"C"
with the other
point of the dividers; then the distance between 3 on scale
"A"
and 4 on
scale
"C"
about the point

o
at 3
represents \/ U
+ T
2
.
Swing the dividers
on scale
A"
;
as a centre until the other
point reaches scale
1,+^
e,
8
in this
"A" (at point 8) then o. .8 on scale "A " = = M + VM + T = T With the value of
2
2
.
found
way, proceed as in case of a simple twisting
moment. The intersection of the horizontal through 8 (T e ) and the vertical through 13 (p) is at point "c." Since the moments correspond to units of 10,000 inch lbs. on scale "A"
the largest figures of the diagonals are to be read in determining
the diameter.
The
point
between 3.0" and

as 3.15".
3.2''';
"c" therefore indicates a diameter of by interpolation the diameter is taken

the diameter
is
By computation
found
to
be 3.14".
shaft 3 T diameter would be proper for this case. of the diagram in connection with equations and
The
3
use
of
Table
VI
is
obvious from the above.

of Fig.
The diagram
or the
stress.
107
is
equally convenient for finding
the intensity of stress in a given shaft under a
known moment;
of 1,000,000
moment on
Thus,
if
a given shaft corresponding to any intensity of
7^"
shaft
is
subjected to
moment
inch
lbs.,
consider the
moment
units as 100,000 inch lbs., pass

to a point slightly
horizontally from 10 on scale
"A"
below the
diagonal marked .776 (7.76" diameter), and then vertically up-
ward
If
to scale "2?,"
where the
stress is
read as about 11,000
lbs.
per sq. inch.

it
is
required to find the twisting
to
an intensity of
point
slightly
stress of 9,000 lbs. per sq. inch
diameter;
pass vertically
moment corresponding on a shaft i%" downward from "9" on scale U B" to

diagonal
above the
marked "1.49";
then
AXLES, SHAFTS,
horizontally to 5.9 on scale
AND SHAFT COUPLINGS
299
"A"
As
1.49
is
the middle
lbs.;
number
on the diagonal, the moment units are 1,000 inch
therefore
T =
5.9
1,000
5,900 inch lbs.
123. Torsional Stiffness and Deflection of Shafting.
When
shaft has considerable length, the matter of torsional stiffness

is
important.
rule,
common
in practice, is to limit the twist
in
the shaft to one degree for every 20 diameters in length.

to 0.075 degree for every foot
Another rule limits the twisting

in length.
The
lateral deflection of the shaft should not exceed
y-JV per foot of length, to insure proper contact at the bearings. Theoretical considerations, however, do not enter so largely into
the spacing of bearings of line shafting, as does the construction
of the
framework
to
which the bearings are fastened.
Care
should be exercised in laying out such structures, that provision

is
made
for fastening the hangers close
enough together
to avoid
excessive deflection.
For the average range
of velocities found
in practice the following
formulae* can be used for ordinary
small shafting.
L =
\/
d
d
for shaft without pulleys
(1)
. .
L =
where L
If
for shaft carrying pulleys
(2)
distance between hangers in feet and d
diameter
of shaft in inches.
T be
the twisting
moment
in foot lbs. applied to a shaft,
then the power transmitted at
N revolutions per minute \s2T-N;

moment,
for a given
from which
the smaller
of
it
appears that the greater the velocity of the shaft,

the required turning
is
amount
power transmitted.
If
a slightly deflected shaft
is
rotated, centrifugal force, acting
on the eccentric mass
of the shaft, tends to equalize the forces in
which hold the shaft deflected

as a of centrifugal force is small,
one plane and
to whirl the shaft
whole around the axis of rotation.

plane.
At low speeds the action

will
and the deflecting force
hold the
shaft deflected in
its
As
the effect of centrifugal force
increases with the velocity, while the effect of the deflecting force
is
constant,
it
is
clear that as the speed
is
increased the centri-
* See also Kent's " Mechanical Engineer's Pocket Book," page 869.
300
fugal force will, at
flecting force,
MACHINE DESIGN
some speed, balance the effect of the deand the shaft will become unstable. Beyond this For a given speed the shaft will whirl about the central axis. diameter of shaft there is one definite speed within which it will
maintain a stable condition with a given deflection.
If
L =
distance between bearings in
feet,
diameter of
shaft in inches,
critical
and
N = the revolutions per minute, then for the
speed *
^-75^1 .......
This equation
the
refers to the bare shaft only
(3)
and
it
determines
maximum
safe span.
Where
pulleys are carried at
some
of the
distance from the bearings, the span, L, must be less than the
value given by equation

pulleys,
(3)
on account
of the
added mass
and the great
liability of the latter to

is,
be unbalanced.
The
speed of shafting in practice

critical speed.
almost always, considerably
below the
Shafting 124. Practical Considerations, Hollow Shafting, etc. up to 3" in diameter is, in this country, made of cold-rolled steel. Such shafting is true and straight and needs no turning whatever. If keyways are cut the shaft must, in general, be carefully straight-
ened afterward, as the cutting
relieves, locally, the skin tension
due
to the cold -rolling f thus causing the shaft to
warp.
Larger
sizes of shafting are forged
and machined.
The
use of hollow shafts not only reduces the weight for a
given strength, but the removal of the metal from the core of a
steel shaft (or of the ingot
from which
is cast,
it is
made) very greatly

stress.
increases
its reliability
under repeated application of
Shortly after a steel ingot
the exterior solidifies

is
and
still
becomes comparatively cool while the internal portion

fluid.
The subsequent
contraction, during complete cooling,
is
much
mass.
less
less in the exterior walls
than
it
is
in the hotter interior

this period,
it
Unless the interior
is
"fed" during
will
be
dense than the outer portions and shrinkage cavities are apt to
* See Rankine's "Milhvork," page 549. t See Article 12.
AXLES, SHAFTS,
AND SHAFT COUPLINGS
301
be present near the centre of the ingot. have been adopted

to
reduce
this evil,
Numerous expedients among which is "fluid

overcome
compression," or subjecting the ingot to heavy pressure immediately after

it
is
poured.
The
difficulty is not entirely
by such means, however, as the walls of large ingots become too

rigid to yield to the pressure before the interior is entirely solidified.
The external walls "freeze," after which the internal shrinkage is made up by metal flowing from the upper portion toward the
bottom as long as any
of
it
remains
fluid.
This leaves a shrinkage

liberated during cooling
cavity at the upper end of the ingot.

collects in this cavity also.
to
Gas
The
result of these
two actions
cut off
is
form what
is
called the "pipe,"
which frequently extends

is
to a
considerable depth.
The
top end of the ingot

all
and and
not
remelted, but this does not insure removal of

it
of the pipe,
is
also involves
much
all
expense.
If
the portion cut off
sufficient to
remove
of the pipe, a piece rolled or forged

is
from
the ingot contains a flaw near the centre which

into a long crack
rolling
if
drawn out
the ingot
is
worked
into a long piece.
The
work
and forging may squeeze the

it
sides of the cavity together
so that
is
is
not easily detected at any section, but as this

a temperature
is
done
at
much below
its
that corresponding to
is
welding, the defect
not removed.
This flaw
is
more
or less
irregular or ragged; hence

fracture,
far
form
stress,
favorable to starting
finally
under variations of
to
which may
effectually
extend
enough
cause rupture.
If the ingot is
bored out, the pipe

is
is
removed, and
It will
the metal remaining
superior to that of a solid shaft.

is
if
be evident that casting a hollow ingot

boring out one which was cast solid; for
not the equivalent of

the ingot
is
cast hollow
the outer and inner walls cool before the intermediate
mass does,
In
fact,
and the shrinkage

shaft
effect
takes place in the latter.

is
made from
a hollow ingot
worse than the solid
shaft, in
the respect that the former has the defective material nearer the
outer fibres where the stress
i.-
'jreater.
3 2
MACHINE DESIGN
COUPLINGS AND CLUTCHES

125. General Description.
Couplings are machine members

of
which fasten together the ends

of
is
two
shafts, so that rotary
motion
one causes rotary motion of the other.

to
be broken only at rare intervals, as in
Where the connection making of repairs, the

Such couplings are
desired to disengage
couplings are generally constructed so that they must be partially

or wholly dismantled to separate the shafts.
known
is
as
permanent couplings.
is
When
The
it is
the shafts at will, the coupling
of a different construction
and
generally
known
as a clutch.*
use of clutches
is
not,
how-
Fig. 108.
Fig. 109.
ever, confined to securing together the ends of shafting, but they
are
much
used for engaging and disengaging pulleys at
will, in
connection with the shafts on which they are placed.

service clutches
For
this this
making use
is
of friction are
much
used,
and
particular type
discussed in Chapter XIII.

to bring
Couplings should be placed near a bearing, so as
the joint in the shaft near a supported point, and should be placed
on the
side of the bearing farthest

is
away from
is
the point where
power
applied,
is
so that
when
the shaft
disconnected the
running part
126.
to
supported near the end.
Permanent Couplings.
in Figures 108, 109,
Where
the axes of the two shafts
be connected are parallel and coincident, couplings such as are
shown
no, and
in
are used.
Fig.
108
illustrates a type of
coupling
known
as a split-muff coupling.
and discussion of
M.
E., 1908, for a full description
various forms of clutches.
AXLES, SHAFTS,
AND SHAFT COUPLINGS
0^0
The
fore,
parts
A and B
is
are separated
be clamped
to the shaft
by the
by a small space and can, thereFor heavy work a bolts C.
key as shown
provided, but in lighter shafting friction alone
may
suffice to
prevent relative rotation.

Sellers
Fig.
no
shows the
Muff Coupling.
Here the
circular
The tapered wedges B, B, are drawn inward by the bolts C. drawn are they tighter the hence wedges are split as shown at D, inward the more firmly they clasp the shaft. For light work no
key
is
necessary, but for the full capacity of the shaft keys are
advisable.
Couplings such as shown in Figures 10S and
no
are regularly
Fig. iio.
Fig.
hi.
manufactured
cities of
in
standard
sizes,
and the student
is
referred to
the trade catalogues of manufacturers for dimensions and capa-
such couplings.
Coupling, Fig. 109,
effective
is
The Flange
and
also
one of the most
common
one of the most
forms of permanent couplings.
The
general proportions are usually designed empirically, but the
bolts should be designed so that their
combined resistance
and the
to a
torsional
moment, around the

fitted
axis of the shaft, will be at least as
great as the torsional strength of the shaft itself;
bolts
should be accurately
so as to distribute the load evenly
among them.
Let
D=
= n = r = p =
d
%
diameter of the shaft in inches

diameter of the bolt in inches
the
number
of bolts
radius of bolt circle in inches
allowable shearing stress per square inch, for
steel.
34
MACHINE DESIGN
Then
-
nD
16
tz
<P
n
4
whence d
w
--
(i)
Good
practice gives
but this number

etc.
may be
be
modified for convenience in spacing,
The
bolts should
carefully fitted to insure that each one carries

load.
its full
share of the
it
The
projecting outer flange
is
an important feature as
covers the revolving bolt heads, thus protecting
workmen from
becoming entangled. For best results the flanges should be pressed on to the shaft and the faces trued up in place, thus
Fig. ii2.
insuring greater accuracy of alignment.

in all
This should be done

marine
good work.
great strength
is
When
Fig.
and
reliability are desired, as in
work, the flange
sometimes forged
solid with the shaft, as in
in. Here the bolt holes are sometimes bored tapering, and reamed after the flanges are placed together, thus insuring a perfect fit for the bolts, and also facilitating their withdrawal.
When
cident, or
the axes of the two shafts are parallel, but not coin-
when
there
is
danger of parallel and coincident axes

is
wearing out of coincidence, Oldham's Coupling, Fig. 112,

used.
to
It consists of
its own
often
fast
two heavy flanges (A and B), each keyed
respective shaft,
and an intermediate
disc C.
The
disc
has a tongue running diametrically across each face, these tongues
AXLES, SHAFTS,
AND SHAFT COUPLINGS
305
being placed at right angles to each other and fitting into grooves
cut in the flanges.
With
this
coupling the rate of rotation of
the driven shaft
is
identical with that of the driver, or, in other

is
words, the angular velocity
the same.
The
coupling
is
often
used on the propeller shafts of small power boats.

If the
axes of the two shafts,

angle
of a
A and
B, Fig. 113, intersect and
make an
by means
called.
is
with each other they
may be
coupled together
it is
Hook's Coupling or Universal Joint, as

is fitted
often
In this coupling each shaft
with a jaw
F.
which
pin-connected to an intermediate
intermediate
member
The
holes in
this
member
for receiving the pins

this
are at right
angles to each other.
With
arrangement the angular velo-
city of the driven shaft is not the
same
at all points of the revolu-
tion as that of the driver.*
The
construction
shown
in Fig. 113
Fig. 113.
is
very
common, but
is
the difference between the angular velocity

of the driven shaft
is less
of the driver struction
and that
when
the con-
such that the axes of the pins
intersect.
The
is
construction required to
usually
in
make
the axes of the pins intersect

113,
more complicated than that shown in Fig. rough work the simpler design is adopted.
and hence
If another shaft C be coupled to B so that A and C make the same angle with B; if also the pins G, G in B are parallel to each other and all three shafts lie in the same plane; then the angular
velocity of
will
be identical with that of
A and
vice versa.
Empirical practice makes the diameter of the pin

one-half the diameter of the shaft.
128. Positive Clutches.
equal to
Positive clutches are
much
198.
used for
* Sec " Kinematics of Machinery," J.
H. Barr, page
306
starting
MACHINE DESIGN
and stopping such machines as punch presses which
intermittently.
must work
work.
They
are
made
in so
many forms
given in the
that a description of
them would be beyond

S.
the scope of this

is
very
full
description of
many forms
Transactions of the A.
M.
E., Vol.
XXX,
to
has already been made.
Fig. 114 illustrates
which reference the most common
form
of disengaging coupling for

fast to the shaft to
heavy work.
The
part
is
is
made
be driven, while part A, which
com-
pelled to rotate
by the feather F, can be moved
axially along
the driving shaft.
ring R, fitting the groove
loosely in a
radial direction, is connected
by the pins
to
an operating lever
Fig. 114.
Fig. 115.
which
is
not shown.
When
the part
is
moved forward
till
the
jaws / engage,
will drive
positively in either direction.
In
order to facilitate the engaging of the jaws they are often
made
as in Fig. 115, but in this case the driving can be in one direction
only.
The
total cross-sectional area of the
jaws must be such
that they will not shear off under the load,

faces
and the area of the jaw

is
must be
sufficient to prevent crushing.
Frequently, for light work, only one feather
used, but two
feathers are, in general, better, both on account of the driving

effort
and
for ease of operation.
129. Flexible
Couplings.
flexibility in
Where
it
is
desirable to have a
small amount of
is
a shaft, a flexible coupling, such as
shown
in Fig. 116, is employed.
These members are much

steam engine,
In the con-
used for connecting rapidly revolving machines to prime movers,

as in the case of a
dynamo
directly coupled to a
the object being to prevent
undue
stress,
or bearing pressure,
shafts.
from lack of accurate alignment of the two
AXLES, SHAFTS,
struction shown, the shafts
,4
AND SHAFT COUPLINGS

and
3 7
arc fitted with heavy flanges,

elastic material
F, which carry pins, P.
Links of leather or other
connect pins on one flange with pins on the other, there being as
many
links as there are pins in each flange.
This arrangement
Fig. 116.
allows for a slight angle between the axes of the shafts, or for a
small lack of coincidence in the axes.
The
pins in one disc are
sometimes placed on a smaller diameter than those on the other,

so that in case of failure of the links the pins will not strike
and
cause breakage.
CHAPTER
BELT, ROPE,
XII
AND CHAIN TRANSMISSION

When power
is
to
be transshafts
mitted from one shaft to another, especially

are not far apart, in such a
when such
two must be constant,

employed.
ratio
manner that the velocity ratio of the some form of toothed gearing is usually
it is
When, however,
not necessary that the velocity
remain constant,
flexible elastic
connectors are
is
much
used.
is
When
the distance through which
power
to be transmitted
comparatively short (50 feet or

or manila, are
less), flat belts,
or ropes of cotton
most common; while
for longer distances steel
ropes have certain advantages.
For small amounts of power,

used.
round
lately
belts of leather are
much
Chain
drives,
which are
virtually flexible connectors
running on toothed wheels, have
come
into extended use for transmitting
power over commaintain
paratively short distances.

positive velocity ratio
They
are very efficient,

shafts,
between the two

is
and can be used
when
the distance between shafts
too
great for convenient
use of gears.
Leather belts are made by cementing, sewing, or riveting

together strips of leather cut from oak-tanned ox-hides.
Where
only one thickness
is
used they are
known
as single leather belts;

to obtain a
where two,
belt,
three, or four thicknesses are
needed
they are
known
respectively as double, triple,
heavy and quadruple
belts.
Cotton belts are made either by weaving in a loom, or are
built
up
of several layers of canvas,

fold.
sewed together, with a special
composition between each

this country.
Rubber
belts are
They are very little used in made of several layers of canvas,

Belts of raw-
held together with, and completely covered by a rubber composition.
They
are very effective in wet places.
hide are also used to some extent.

308
BELT, ROPE,

make them
309
The ends
either
of all belts are joined, to
continuous,
by lacing or sewing, or by some kind of special fastening of which there are many on the market, or by making a permanent The latter method is much joint by cementing and riveting.
preferable where
it
can be applied, as
as strong as the rest of the belt,
makes the joint practically and gives a smooth surface which

it
runs better than any
joint.
Other kinds of
joints
reduce the
strength of the belt from 60 to 75 per cent, but inasmuch as the lacing can be replaced and the belt itself has its life prolonged by
reduced load,
ful as
it
this initial loss of efficient strength is
not as waste-
at first appears.
131. Theoretical Consideration of Belts

117, let
and Ropes.
is
In Fig.
is
represent a pulley whose centre
at
O, and which
Fig. 117.
connected by a belt as shown to the pulley B, whose centre
is at
Or
When no
turning
moment
is
applied to the driving pulley A,
the tensions in the two parts of the belt are the same, except
possibly for friction of the bearings, tension with which the belt
total initial
is
and
is
that due to the initial
placed upon the pulleys.
Let
this
tension on each side of the belt be called
It is
evident that this initial tension will cause the belt to
exert a pressure
upon the
If
pulley,
and
this pressure will
induce
a frictional resistance opposing relative sliding between the belt
and the
pulley.
now
to
will
a turning
B,
moment
is
applied to A, and
belt
a resisting
moment
the pull
upon the
in
due to
part.
this
increase the tension in the lower part
of the belt,
and decrease the tension

tensions be called
the
upper
Let
It
is
these
new
total
and
respectively.
evident that the tendency of the belt to slip around the pulley,
3io
owing
is
MACHINE DESIGN
to the difference in tension
on the two parts of the
belt,
resisted
by the
between the belt and pulley.

)r
The difference in tensions when the turning moment resisting moment applied to
If the difference
tends to rotate the pulley B, and

(7\
T
1
becomes equal
which
to the
B, rotation will take place.
between
and
is
is
necessary to
to
overcome the
frictional
resisting
resistance
compared between the pulley and belt, no
moment,
will
small
the
slipping
of the belt in
practice,
on the pulley
occur.
To
obtain this result
would necessitate the use of very large belts, power transmitted. It has been found to be better practice to use smaller belts and allow the belt to slip somewhat.
relatively, for the
In addition to the slipping action noted above,

subjected to what
is
all
belts are
known
as creep.
Referring again to Fig. 117
consider a piece of the belt of unit length moving on to the pulley
under a tension
As this piece of belt, of unit length, moves to N, the tension to which it is around with the pulley from subjected decreases from 7\ to T 2 and the piece, owing to its
Tv
elasticity,
shrinks in length accordingly.
The
is
pulley A, there-
fore, continually receives a greater length of belt
than
it
delivers,
and the
velocity of the surface of the pulley
faster than that
of the belt
which moves over
it.
In a similar way the pulley

it
B
its
receives a lesser length of belt than

velocity
surface.
is
delivers,
and
its
surface
slower than that of the belt which
moves over
results in
of
is,
This creeping of the belt, as it moves over the pulley, some loss of power, and is unavoidable. The total loss speed due to both slip and creep should not exceed 3%; that
the surface speed of the driving pulley should not exceed that
of the driven pulley

this value to
by more than 3%.
Good
practice limits
about 2%.
When
the total slip approaches

off of
is
20%,
there
is
danger of the belt sliding
the pulley entirely.
Since the pulling power of a belt
proportional to the differ-
ence between
and
2,
it
is
necessary to
know
the relation
which
Let
t
exists
between these quantities.

pulley.
the tension per square inch of belt section at any point
on the
BELT, ROPE,
/,
311
=
=
the tension per square inch of belt section on the tight

side in pounds.
/_,
the tension per square inch of belt section on the slack

side in pounds.
the
in
maximum
allowable tension per square inch of belt
pounds.
/= = w = q =
v
effective pull of belt per square inch of cross-section
(/,
/ 2 ),
in
pounds.
the velocity of the belt in feet per second.

the weight of one cubic inch of belt in pounds. the
reaction
of
pulley
against one linear inch of

belt
of
the
width
con-
sidered, in pounds.
c
the
centrifugal
force
of
one cubic inch of belt in
pounds
ft
at the
given speed.
friction
=
=
the coefficient of
between
r
belt
and
pulley.
the
radius of
the pulley
in inches.
= =
the angle of
in degrees.
belt
contact
the angle of belt contact in radians

of
=
of
.0175 a
belt
will
The
c
centrifugal force
one cubic
inch
be
12
wir gr
1
hence the centrifugal force of one linear inch of

1
belt
having
ww
square inch of cross section will be
Let the cross-sectional area of the belt be one square inch and
consider an elemental portion of
It is
its
length as
is
shown
in Fig. 118.
held in equilibrium,
following forces:
(a)
(b)
(c)
when
slipping
impending, by the
(d)
The centrifugal force = c ds The radial reaction of the pulley The frictional force = jt q ds. The tensions / and t -f dt.
against the belt
q ds.
312
Resolving
all
MACHINE DESIGN
forces vertically
q 2 ds
ds
do
t
sin
+
2
(t
dt) sin
do
2
(i)
Here d
d
is
so small that sin

2
may be
taken as equal to
of d
do
in radians, without appreciable error,
and the product
and sin
do may be
(i)
neglected.
Hence
may be
written
qds
but
cds
=
,
tdO
7
(2)
12
wv*
g
r
and ds
2
do
* .
cds
12
WV
gr
ds
=
t
12
WV
g
= =
(t
for convenience.
Hence from
(2)
q ds
dO
z d
z)
dO
(3)
From
equality of
moments around
t
O
fi
dt
q ds
(4)
(3)
.'.dt
p.qds
Substituting in (4) the value of q ds obtained from
p.
(t
z)
d
d / u r do J
,
rh h
J
/
t
dt
t
z
j
L
log
ti
<*
H.
f-
(5)
and common
S
0.434
/*
=
k
0.439
11
'l
=10
=10
0.0076
ju.
=10
for convenience
(6)
BELT, ROPE,

tx
313
value of
Now
t2
t2
*
.
t2
f and
substituting
this
in (6)
and reducing
'
,_
io
-i
=
I
h r where C
~C 1
' '
(7)
10
and
[tl
-z]
f^]
v
=[t -z]C
(8)
If (8)
be multiplied through by
it
will express the horse-
power
(h.p.)
which a
belt of
one square inch cross-sectional area
will transmit or,
132. Practical
/ 1)
=h.p. =
[t
-z]
the
i)
....
equations
(9)
Coefficients.
,
In
above
the
must be known or assumed before /* a solution for t or / can be made. The angle of contact, a, can be taken from the drawing of the drive in question, and some
following quantities a
x
and
z,
allowance should be
if
made
for the conditions of operation.
Thus
the belt
is
to
run in a horizontal position, with the slack side
on
top, the full theoretical value of a
may be
taken.
If,
however,
the slack side
must be on the bottom (an arrangement which

if
should be avoided
position,
possible) or
if
the belt
is to
be run in a vertical
the theoretical
also applies to
some reduction must often be made in value of a to allow for sagging of the belt. This belts running at high speed, where centrifugal
lessen the arc of contact.
force tends to
The
tity,
^ is an exceedingly variable quanchanging with the character and the condition of the surfaces
of contact, the initial tension of the belt,
and the
rate of slip.
It
has been found by experiment that, within reasonable

coefficient increases with the slip
limits, the
and
that, as before stated,
maximum
3 per cent is
rate of slip, including creep, not in excess of about
good practice. Experiments made by Professor Died er/<
richs in the laboratories of Sibley College gave the values of
314
MACHINE DESIGN
first
shown in the
column
of the following table.
Allowing for the
difference between conditions in the laboratory

in practice, the value
shown
in
and those found the second column may be used
in designing leather belts.
For pulleys made of pulp, For pulleys made of wood, For pulleys made of cast iron,
// fi
ft
= = =
o 29
.
o 20
.
0.31
o .46
0.22
o .30
Values considerably above these were found for paper pulleys of

special construction.
The
used),
quantity
z is
proportional to the weight of the belt per

is
cubic inch.
For ordinary leather (which
most commonly
w may
be taken from 0.03 to 0.04, an average value being
0.035 pounds.
Table XVIII has been calculated with a value of w = 0.035, XIX is abbreviated from "Transmission of Power by Belting"* by Wilfred Lewis.
while Table
TABLE
Values of
XVIII
for
ft.
per sec, or
V=
0
ft
per minute,
w =
80
o35-
40
60
70
90
100
no
120
130
140
V
z
1,800 2,400 3,000 3,600 4,200 4,800 5>4o 6,coo 6,600 7,200 7,800 8,400
n-75 20.9
3 2 -5
47.0
64.2
83 -4 io 5-5 130
-5
1576
187.6 220.2 255-5
Example.
capacity,
fi
Design a belt to operate a dynamo of 15 H.P.

the belt velocity
is
when
2,400
lbs.
ft.
per minute.
Assume
0.30, a
180
and
(9)
tt
200
From
ditions:
h. p.
l
equation
the
horse-power transmitted by a belt

is
having a cross-sectional area of one square inch
for these con-
u \t
z\
CV =
55
r
1
200
.61
40
20
qJ
=7-9
55
.'.
the cross-section required
1.9 sq. in.
which
is
equivalent to a belt
* Transactions A. S.
-fa"
thick and 8" wide.

E., Vol. VII,
M.
page 579.
BELT, ROPE,
315
be
The
1.9
will
total tension
(7\) in the tight side of the belt will
200
380
lbs.
The
total tension
(T 2 )
to
in the slack side

pull,
be
this
value minus the required effective
P, which
is
found by dividing the foot pounds of work

velocity of the belt or,
be done by the
15
X 33P =
2,400
206.
Hence
T =
9
T P =
x
380
206
174 pounds.
TABLE XIX
Values of
(Nagle)
Degrees of Contact
/"
90
100
10
120
130
140
ISO
160
170
180
.210
.270
2 3
.270
319
381
288
3 64
307
325
.408 .480 544
342
359
376
.467
.20
295
342
407
.386
.428
53
567
448
524 590 .646
695
25
325
376
;
354
.408
45
7
432
457
.520
575
544
.610
.667
438
489
.467
494
548
597
35
423
467
.520
567
.600
.649
.692
.624
673
715 785
40
45
.502
536
579
.652
.624
.667
715
757
.822
507
578
544
.617
.610
640
7i3
737
.805
55
.684
739
763
Equations
(7)
and
(8)
involve the relations which exist between
and
for a given set of conditions,
but they do not indicate

tension
the relation between
them and the
initial
x
3.
It
was
formerly supposed that the
sum
of
and
equal to iT.^
calculations.
and
this relation
may
still
was constant and be used for very rough

2
Mr. Wilfred Lewis* has shown, experimentally,
that this
is
not true.
The
ratio of stress to strain in leather
and
rubber increases with the strain instead of being proportional to

it
as in ductile metals.
When
is
a belt transmits power the tension

side
till
is
increased on the tight side
and decreased on the slack
the difference in tension
equal to the required driving force.

M.
E., Vol. VII,
page 566.
31
MACHINE DESIGN
is
This
accomplished by what virtually amounts to shortening
the belt on the tight side, a given amount,
by transferring
this
amount
between
to
the slack side.
Because, however, of the relation
stress
and
strain noted above, the increase of tension
on
the tight side, due to this
amount
of shortening, is greater than
the decrease of tension on the slack side due to an equal

of lengthening, and, as a consequence, the
amount
Sugges-
sum
of the
two ten-
sions
tion:
is
increased* as the effective pull
is
increased.
Place a rubber band over the fingers of the two hands

stretch
it
and
moderately
then twist one of the hands in either

of force tending to bring the
direction
and the increase
hands
together will be apparent.
In the case of a long horizontal belt the increase in the

of the tensions
is
still
sum
further
augmented
in
driving, because
the tension on the slack side (with a proper initial tension in the
due to the sag of the belt from its own weight; and thus the tension on the slack side tends to remain nearly constant, while the tension on the tight side increases with the power transmitted, at a given speed. It is found that the sum of the tensions on the two sides, when driving, may exceed the sum
belt) is largely
and in by the strength of the belt. In addition to the causes discussed, the tension on both parts of the belt are increased by the centrifugal action due to the mass of that portion of the belt which is rotating round This latter cause increases the stresses on both the pulley axis. the tight and slack sides of the belt, and decreases adhesion between the belt and the pulley, but does not increase the loads on the shafts which produce pressure at the bearings and flexure of
of the initial tensions in vertical belts,
by about 33 per cent
horizontal belts the increase
may be
limited only
the shafts.
Large
tension.
belts should therefore
be put on with care, as to

is left
initial
Ordinarily, the initial tension

it
to trained judg-
would seem that the more advanced practice of splicing the belt under a known initial tension will add to the life of large and important belts.
ment, but
M.
E., Vol. VII,
page 569.
And chain transmission

133. Strength
of
317
of
Belting.
The
ultimate
strength
good
leather belting will vary

inch.
from 3,500 to 6,000 pounds per square Professor Benjamin * gives the strength of cotton belting as
about the same as good leather.
He
also
found that four-ply
rubber belting had a tensile strength of from 840 to 930 pounds The ultimate strength of belting seldom per inch of width.
enters as a factor in belt design, as the real strength of the belt
is
in the joint.
Where
the ends of the belt are laced together, a
maximum working
found
to
stress of
be good practice;
gether, thus
making
to
it
200 to 300 pounds per sq. inch is and where the belt is cemented to"endless," a working stress of 400 pounds
per square inch

varies
may
be used.
The
thickness of leather belting
from T\
-^ inch for single leather, and from
to ]/ 2
inch for double leather.
Hence
for single leather,

for laced belts.
belts.
p = 50 p
to 75
pounds per inch of width

belts
100 pounds per inch of width for cemented
For double leather
p may be taken
life
at twice these values.
Lower
stresses
than these are often advocated, and undoubtedly

of the belt.
lower stresses increase the
134. Velocity of Belting.
In equation
(8)
when
belt
t 1}
f=
and the
If
belt will exert
no turning
force, the centrifugal force
relieving all frictional resistance

/j
between the
and
pulley.
be taken as high as 400 pounds,

2
and w =
.035 this will
occur
when
400 or when
6
400 whence v
175
it.
per second or 10,500 feet per minute.

If
equation
it
(8)
be multiplied through by
v,
the velocity of
is
the belt,
will express the rate at
which energy
being de-
livered, or
fv = v [t - z] C = v\t t
x
If
now
.3,
w =
v[t 1
.035, a
=
2
]
180,
which are average conditions,
the equation
becomes
jv =
.013 v
0.6
0.6
.0078 v*
* See
"Machine Design," by Benjamin, page
186.
3l8
MACHINE DESIGN
and equating
Differentiating the right-hand side with respect to v

to zero
0.6^ .0234 v 2 = o
or v
=
tx
5.1
\/
(10)
which gives the relation between v and
for
maximum
power.
When
=
t
400, v
x
102 feet per second or 6,120 feet per minute
and when
275 pounds, v
85 feet per second, or 5,1*00 feet
per minute.
It is
often necessary to run belts at

it is
much
lower
speeds than these; but
not economical to exceed these limits.
speed of a mile per minute
may be
taken as about the economithis is also
cal
maximum
limit;
and
it
so
happens that
about the
limit of safety for ordinary cast-iron pulley rims.
For durability
feet
combined with
minute
efficiency,
speed of 3,000 to 4,000
per
may be
taken as a
fair value,
though practical limitations

fix belt
such as speed of shafting and diameter of pulleys often

velocities at
much
lower values.
135. Efficiency of Belting.
The
slip
losses of
power
in belt transto
mission consist of the loss due to
and creep, that due
bend-
ing the belt over the pulley, and the frictional losses at the shaft
bearings, due to belt pull.
The
2
first
two, slip and creep, should

is better.
not exceed 3 per cent, and
per cent
The
loss
due
to
bending the belt

life
is,
usually, negligible although the effect
on the
of thick belting running
on small pulleys
considerable
is
if
important.
the belt
The
losses at the bearings
may be
laced on under great initial tension in order to

this condition
must be carry the load, and

absolutely nec-
should be avoided except where

belt.
it is
essary to use a short
well-designed belt transmission

it
should have an efficiency at least as high as 95 per cent, and may be as high as 97 per cent including bearing losses.
136. Other Equations,
Common
as 305
Rules.
If in
equation
(9),
w be
taken as 0.032 and

h- P-
tx
pounds the equation reduces

2
]
to
-55
0.0000216 v
h.
vC
(11)
where
C =
as before
If the
it
and
horse-power per square
inch of belt area.
equation be multiplied by
will express the total
the area of
the belt cross-section,

mitted, or H. P.
horse-power trans-
.55
0.0000216 v 2 ]
vC A
....
(12)
BELT, ROPE,

is
319
iden-
Professor Diederichs has pointed out that equation (12)

tical
with that reported by Mr. Nagle to the A.
S.
M. E.* and
commonly known by his name. given in Table XIX.
Values of
have already been
In the transactions of the American Society of Mechanical

Engineers, January, 1909, Mr. Carl Barth presents a
more
ex-
tended mathematical treatment of the driving capacity of
belts.
He
also presents scientific
methods
for
measuring the tension
in belting.
Many
other formulae of a strictly empirical charauthorities
acter
are
given by different
and
last
some
of
them
speeds
that a
are
very convenient.
In general
these
formulae neglect
to belt
centrifugal action
and are hence applicable only

per minute.
below 2,500
single
feet
Thus
wide
Kent's
common
rule
is
leather
will
belt
one inch
1
traveling
"
1,000
feet
per
minute
Pocket
transmit
H.P.
Mechanical Engineer's
Book," page 877, gives a number of these so-called

Considerations. the
F.
literature
practical rules.
137. Practical
One
the
of
the
most valuable
is
contributions
to
of
subject
" Notes
on
Mr.
Belting," by Mr.
tions of the
M.
Taylor, in Vol.
of
XV
of
the Transac-
American Society
Mechanical Engineers.
Taylor kept an accurate record of measurements and observations
on
belts in use at the
Midvale
facts
Steel Co.'s works, for nine
years,
and gives many valuable
and practical suggestions

it
in his paper.
satisfactory abstract of
is
not possible here.
belts.f
Mr. Taylor advocates thick narrow belts rather than thin wide He sums up his investigation in 36 "Conclusions," among which are:
"A
of 35
double leather belt having an arc of 180
will give
an
belt
effective pull
on the face of the pulley per inch of width of

for
oak-tanned and fulled leather, or 30 pounds for other types of leather belts and 6- to 7-ply rubber belts."
pounds
"The number
* Vol. II, page 91.
t
is
of lineal feet of double belting,
inch wide,
While
in general this
conclusion
is
justifiable, care
life
should be taken that
it
not carried to the extreme where the
of the belt
may be
shortened by ex-
cessive bending.
320
MACHINE DESIGN
passing around a pulley per minute, required to transmit one
horse-power
is
950
feet for
oak-tanned and fulled leather
belt,
and 1,100
feet for other types of leather belts,
and
6- to 7-ply
rubber belts."
"The most economical

is
average total load for double belting,
square inch of section.
65 to 73 pounds per inch of width, i.e., 200 to 225 pounds per This corresponds to an effective pulling
of 30
power
effect
pounds per inch
of width."
little
"The
on
speed at which belting runs has comparatively

its life, till it
passes 2,500 or 3,000 feet per minute."
"The
It
belt speed for
maximum economy
should be from 4,000
to 4,500 feet per minute."
should be especially noted that Mr. Taylor advocates a

belt tension of about one-half that ordinarily used.
first
maximum
materially.
This would, of course, increase the

size of belt required to
cost of the installation
His values, however, are not based on the
minimum
simply transmit a given horse-power, but

will transmit that
on the
size of belt
which
horse-power for a given
time with
minimum wear and
loss of time
due
to
breakage or
taking up to restore tension.

or not,
it
Whether
his practice is followed
indicates the true aspect of the problem,
and
is
a step
in advance.
In laying out belt drives, care should be taken to keep the

diameters of pulleys reasonably large.
action to which the belt
is
is
The
it
constant bending
subjected as
runs around the pulley

is
a great source of wear, and where the pulley

to the thickness of the belt, this
it is
very small,
compared
For
this
may
be excessive.
reason also
probably better to run the hair side of

is
the belt next to the face of the pulley as this side
cracked by bending, than the
flesh side,
it
which
is
more easily more soft
and
pliable.
Mr. Taylor says
is
safe to
run double leather
belts
on pulleys 12 inches in diameter.

total length of the belt or distance
The
when
will
between shaft centres
also deserves attention.
belt being elastic, acts like a spring

it.
tension
is
applied to
The
longer the belt the greater
be the total stretch for a given load.

produce
less
Suddenly applied loads,
therefore,
stress in
long belts than in short ones
BELT, ROPE,
(see Art.
24).
If,

is
32
too
however, the distance between centres
great,
compared
to the size of the belt, the belt is liable to flap
and run unevenly on the
pulleys.
is
For small, narrow
belts a
maximum
distance of 15 feet
good
practice, while for heavier
belts 25 feet is
found
satisfactory.
A number
Engineers.
Society by:
of important
to
investigations of belt transmission
have been reported
the
American Society
of
Mechanical
Professor G.
See the following papers in the transactions of the
Mr. A.
F. Nagle, Vol. II, page 91;
Lanza, Vol. VII, page 347; Mr. Wilfred Lewis, Vol. VII, page 549; Mr. F. W. Taylor, Vol. XV, page 204; Professor W. S.
Abstracts of these as well as other valuable data are given in Kent's " Mechanical Engineer's Pocket
Aldrich, Vol.
136.
XX, page
Book," pages 876
to 887.
138. General
to
FIBROUS ROPE DRIVES Considerations. When the amount

is
of
power
be transmitted
large,
the width of belt required

is
may be
run wide
excessive, even
when
the belt
made
very thick.
To
belts successfully, the shafting
must be kept
in perfect parallel
alignment, and the distance between shaft centres must not be

too great.
satisfactory
large,
For these reasons rope drives have been found very where the amount of power
to
be transmitted
is
and the distance
of transmission relatively great.
They
are also particularly serviceable for connecting shafts
which are
not parallel, as in the case of "quarter-turn" drives, especially
where a
belt
would have
to be of considerable
width and would,
as a consequence, run badly.
In all fibrous rope drives the surfaces of the pulleys or "sheaves"

are provided with wedge-shaped grooves to receive the rope
and
thereby give the rope a better grip on the sheave.

of
For drives
moderate length, 40
to 150 feet, fibrous ropes of cotton,
hemp
al-
or manila fibre are chiefly employed.
For transmitting power

is
comparatively great distances, wire rope
more common,
though fibrous ropes are also used for comparatively long transmissions.
In
all
long-distance transmission the rope
must be
supported
at intervals
by
idler pulleys.
322
Fig.
MACHINE DESIGN
119* shows a typical rope drive where the
is
line shafting
of each floor of a mill
driven by
its
own
rope drive from the
main
shaft of the engine.
139. Materials for Fibrous Ropes.
or rawhide, are used to a limited extent,
Round ropes of leather, when the amount of

is
power
in
to
be transmitted
is
small.
it
Rawhide
is
especially useful
damp
places, but since

its
costs about six times as
much
as
vegetable fibre rope,
application
very limited.
Leather
belts or ropes of square f or
wedge-shaped section have also been
Fig. 119.
used to a limited extent.
In certain
is
localities in
Great Britain,
but
hemp, which
is
a local product,
quite extensively used;
cotton and manila fibre are by far the most
common
for transis
missions of any considerable
size.
In this country manila fibre
used almost exclusively, while in England and on the Continent

cotton rope
It is
is
also
much employed.
obvious that as a twisted rope of any fibrous material
bends while passing over the sheave, there must be a certain

Reproduced by permission from " The Blue Book of Rope Transmission." For a fuller discussion of such ropes see " Machine Design," by H. J. Spooner.
BELT, ROPE,

The
is
323
is
amount
of internal friction.
result of this action
very
noticeable in any old manila rope which has been used without
lubrication.
filled
When
fibre,
such a rope
fibre,
broken open
it is
found
to
be
with powdered
due
is
to the internal chafing.
For
this
reason manila
cated,
which
twisted
naturally rough,
rope,
is
usually lubriparaffine,
while being
into
with
tallow,
soapstone, graphite, or
some such
lubricant.
Cotton
rise
fibres,
on the other hand, are smoother and hence give
to less internal friction.
They
from
are, therefore,
usually laid
up dry
into rope, a dressing or lubricant being applied to the

rising
exterior to prevent small fibres

starting the rope to fraying.
on the
outside, thus
This dressing also excludes moisoils
ture
and retains the natural
in the interior fibres.
Cotton
fibre is not as strong as manila.
makes the following comparison between cotton and manila rope: " As compared with manila, then, the advantages of cotton ropes of the same diameter are: Greater flexibility, greater elasticity, less internal wear and loss of power due to bending of the fibres, and the use of smaller pulleys for a
Professor Flather*
given diameter of rope.
cost, lesser strength,
Its
disadvantages are:
loss of
Greater
first
and possibly a greater
power due
any case
to
pulling the ungreased rope out of the groove

is
in
this
usually small with speeds over 2,000 feet per minute."

140. Theoretical
Considerations.
131,
if
The
general
equations
(7), (8),
and
(9), of Art.
which were deduced
for flat belts
hold also for round ropes
the proper notation be substituted.
In these equations the unit mass of belt was taken as one cubic
inch.
With ropes
in length
it
is
more convenient
to take a piece of
rope
one inch
and one inch
in diameter.
With
the following
exceptions, therefore, the notation used here will be the

that used in Art. 131.
same
as
Let
1
it/
the weight of a piece of rope
inch in diameter and
inch long.
Let
z'
w' v 1
where w' has the value above.

g_
* "
Rope
Driving,'' by J. J. Flather, page 81.
324
MACHINE DESIGN
Let t\
tight side.
the tension in a rope of
inch diameter on the
Let
C =
f
new
(8)
coefficient
= C
modified on
account
of
wedging
effect of groove.
Then
equations
and
(9)
become
(13)
f = [t\-z']C>
mdh.p. =
In equations
pulley
(8)
[t\
Cv
'
z'\
55o
(14)
and
(9)
the frictional force between the

p.
and the
belt for a flat belt is taken as
radial pressure between the pulley
and the
belt.
where q is the In a grooved

is
pulley the pressure between the pulley and the rope
greater
than the radial pressure in the ratio of cosec

is
to unity,
where
the angle between the sides of the groove.
The
p.
frictional
resistance between the rope
and sheave
is
therefore
q cosec
If
pi
cosec
be substituted for
fi
in the quantity
(equations
8 and 9) the result

(13)
C may
f
be used as indicated in equations
and
(14)
for rope drives.
The
value of
//
for rope sheaves
has not been determined with any degree of accuracy.
Professor
is
Flather* after reviewing what experimental data there

subject, concludes that 0.12
is
on the
a fair value and computes the
following values of
</>
/j.
cosec
0.12 cosec
(p
TABLE XX coefficient of friction =

3
35
0.12 cosec
Angle of groove.
40
45
5o
55
6o
.46
.40
35
3 1
.28
.26
.24
It is
obvious that
if
<j>
be used instead of
/*
in
Table XIX, the
corresponding values of
* "
C in Table XIX will be

Driving,
'
the
new
constant
Rope
page 112.
BELT, ROPE,
AND CHAIN TRANSMISSION =

.31.
325
C.
Thus HO =
and
45
(/>
If also a
180
from Table
XIX =
than 45
.61 about.
is
satisfactory
,
The angle 45 has been found to be the most If the angle 6 be less most commonly used.
is in-
the wedging action, hence the pulling capacity
creased, but the

it
power
is
loss
and wear
of rope
due
to
drawing
out of the grooves
greater.
For such sheaves, with
= 45
(15)
and a
180
h.p^.6i[f 1
As before
stated, reliable data
-z'}~
....
for ropes are scarce,
on the coefficient of friction and designing engineers have approached

to the subject is that of
the problem of rope drives without regard to this coefficient.
One
of the
most important contributions
Mr. C. W. Hunt (see Transactions A. S. M. E., Vol. XII). The notation of Mr. Hunt's article has been changed somewhat to
correspond with that used in this
text.
Let d
L
11/
W
T T T
x
= diameter of the rope in inches. = sag of rope in inches. = distance between pulleys in feet. = weight of one inch of rope oj one-inch diameter. = weight of one foot of rope of diameter d. = total tension in rope on tight side. = total tension in rope on slack side. = tension necessary to give the rope adhesion. = the total tension applied to each side of the rope due
to centrifugal force.
effective turning force
P =
7\
Then 7\ = T +
and
T = T +
2
K K
+ P
in a groove
Mr. Hunt says that "when a rope runs

sides are inclined
sufficient
whose
toward each other

-j-
at
2.
adhesion when 7\
T =
2
an angle of 45 there is However, he assumes a

his equation, for
somewhat
different ratio in the
development of
which he assumes "that the tension on the slack side necessary

for giving adhesion
is
equal to one-half the force doing useful
work on the driving
side of the rope."
326
MACHINE DESIGN
Or T = - and T t = T
2
+K + P = -+K
2
+ P =
^-P+K
2
and
P T = T + K = - +
2
o
by assumption.
..P = ^[T -K]

1
(16)
If
equation (16) be multiplied through by
it
will express the total
horse-power transmitted or
H.P.-l^-^ ....
The
180
tical
I7 )
tension
on each
side of the rope for

1
an arc of contact of
w' v 2
,
and a rope
of one inch diameter
is
which
is
iden-
with the constant
z'
in equation (14).
Mr. Hunt's formula
therefore
may be
written
M_ip W] j_^w]
where
h.p. is the
(l8)
horse-power transmitted by a rope one inch in

is
diameter.
(15)
This
differs
identical in
it
form with the theoretical equation
and
from
only by a negligible
amount
in the value
of the coefficient.
It
would seem therefore
that
Mr. Hunt's assumptions give by using the value 0.12

z
results very close to those obtained
for
as
recommended by
It is to
Professor Flather.
be noted that the values of

z'
.
given in Table
XVIII
may be
used in computing values of

for the weight w'.
The
quantities are the
same except
In Table XVIII,
.035.
w=
the weight
of one cubic inch oj leather
In equation
(18),
w'
the
weight oj one inch of rope oj one inch diameter
.028 for manila
rope and .022 for cotton rope.
If,
therefore, the values given in
Table XVIII are multiplied by - they are

ropes,
applicable
to
manila
and
if
multiplied by
they
may be
used for cotton ropes.
BELT, ROPE,
Example.
327
What diameter of manila rope is necessary to when running 4,000 feet per minute, in grooves Take t\ = 200 pounds, and w' = .028. having an angle of 45 \ z' = From Table XVIII, z, for the given velocity = 64 nearly.
transmit 25 H.P.
. .
64
X
5
51.
From
equation (18) the horse-power which a rope

is
one inch in diameter will deliver under these conditions

h.p.
[/',
- z'] =
[200
51]^
=
12.
h.p.
.*.
the cross- section required
12.
twice the area of a one-
inch rope which corresponds to a rope \yi" in diameter.

Fig.
120* shows curves based on equation
(17), giving the total
horse-power transmitted by ropes of various sizes for 7\
20od 2
and
will
be found convenient for making calculations.
141. Strength of Fibrous Ropes.
The
ultimate strength of
manila transmission ropes

for cotton
may be
taken as about j,oood2 and
rope as about 4,600^* where d

stress
diameter of rope in
less
inches.
The working
must be taken very much

life
than
these values or otherwise the
of the rope is
much
shortened.
For manila rope Mr. Hunt recommends that the working tension
(T
be not over 200 d 2

dr
The same
factor of safety
would give
but
130
as the allowable
working tension
for cotton ropes;
since cotton ropes are
somewhat
less affected
by internal chafing
the working tension may, perhaps, be safely taken at a rather
higher value.
142. Velocity
of
Fibrous
Ropes.
The
centrifugal
z'
force
produces a tension in a rope of one inch diameter of

or in a rope of diameter d the centrifugal force
=
d
1)
12
v1
2 IV
The
g
allowable stress in the rope
is
200 d 2
The
12
centrifugal force will
w'
d?
equal the allowable tensile stress
when
the
v2
200 d 2 or
From "The Blue Book
of
Rope Transmission," by
American Mfg. Co.
328
MACHINE DESIGN
feet per second, at
when v = 140
becomes zero
If
which speed the

stress.
effective pull
for this allowable
working
equation (18) be differentiated and the differential be equated
to zero as in Art. 134, the resultant equation will give the value
55
2" Roge
\
50
50
\
45
\
1? i"R ope
45
\
\
40
\
35
1 i
(A
30
1
/
IVfi"B )pe
\ \
\ \
L
35
I
B OQ g
OS
25
Ph
i
1
/ /
1*1 [~R ope
\ \
u '""R
:
\ \
\
I
O
ft
25
g O
20 Pm
15
/1 // 1
/' z "/ /' '/
'/>
"Re>pe
\\ VA
\\ \
20
15
10
1
I/,
^^
* S*B ape
3 /<i"
Rope
k& ?> i m
^
]
?eet per
MiE ute
\\\ \0 ^ ^
X^N
cm
00
Fig. 120.
of the velocity
where the work done

This
is
is
maximum,
to
for a
rope
feet
of one inch in diameter.
found
be about 4,900
per minute.
stress,
Since the centrifugal force, and the total working

this limiting velocity
both vary as the area of the rope
BELT, ROPE,
applies to
all sizes

which
is
329
of ropes, a conclusion
borne out by
the curves of Fig. 120.

It
has been found, in practice, that the most economical speed
for ropes is
from 4,000
to
5,000 feet per minute.

is
If
speeds
greater than this are used, the wear on the rope
excessive.
is
For a fixed value of
T =
x
200
<P the first cost of
a rope
a mini-
mum
at
about 4,900
if
feet as above,
is
and
this first cost is greater
by 10 per cent
the velocity
increased to 6,000, or decreased to
3,700 feet per minute.
when
the velocity
vivcr
is
The first cost is increased 50 per cent reduced to 2,400 feet per minute with
Fig. 121.
T =
x
200 (P but the reduction in speed increases the
life
of the
rope.
143. Systems of Rope-Driving.
There are two methods of

In the Multiple, or English
placing fibrous ropes on the sheaves.
system, several separate ropes run side by side, each rope forming
a closed circuit in exactly the
same manner
as a flat belt,
and
running constantly in
its
own
particular groove
on each
is
pulley.
In the Continuous or American system one rope only
used, the
till
rope being carried continuously from one pulley to the other

all
the grooves are filled,

it
and
it is
then spliced; so that the rope

is
as
to
[eaves the last groove of the driven sheave

first
returned
the
groove of the driver, or driving pulley, by means
of an idler, or guiding sheave.

so that through
(see Fig. 121).
it
This
idler is
usually arranged
a suitable tension
may
be put upon the rope
330
MACHINE DESIGN
it
Regarding the merits of the two systems

the multiple system
is
may
be said that
the simpler,
considerable security against the loss

as
it is
and that it also provides of time due to breakdowns,

the broken rope
not likely that more than one rope will break at a time.
failure of a rope does occur,
at a
When
may be
removed and repaired
more convenient opportunity, allowing

Occasionally,
the other ropes to carry the load temporarily.
however, the breaking of a rope in the multiple system

great delay, on account of the broken rope
in one of
may
cause
becoming entangled the rope sheaves and winding up upon it before the
machinery can be stopped. In this system the individual ropes must be respliced occasionally to take up the sag in the rope due
to stretching.
The
velocity ratio transmitted
by a new rope
will
be different from that transmitted by an old one which has worn

smaller,
and hence
its
fits
d~wn
radius.
farther into the grooves, thereby
changing
tendency
effective
The
velocity
ratio
of
the
two
sheaves can, however, have but one value, and, therefore, the
will
be for either the old or the new ropes to carry the
whoie load.
result in a
When
is
the
driving sheave
is
the larger, this will
tendency to throw more load on the old ropes; when

the smaller the tendency
is
the driving sheave
to
throw more
of the
load on the larger and
new
ropes.
The unequal speed
ropes, of course, leads to unequal stress;
and slipping and conthan
sequent wear are sure to occur.
The continuous system

due
is
more
flexible in its application
the multiple system; for, owing to the limited sag in the ropes
to the action of the weighted idler, the rope
safely at
any angle.
This form of drive

of
is,
therefore,
may be run much

principal
used for vertical and quarter-turn drives, and, generally, where

the
transmission
is
a complicated nature.
The
objections to the system are the danger of loss of time due to a
breakdown, and the unequal straining
of the various spans of
the rope particularly with a varying load or inequality of grooves.
When
a load
is
suddenly applied
to the
continuous system
all
the spans on the slack side
become
is
slacker except that which
runs over the idler and which

greater load
is
kept at a fixed tension.
A much
hence brought on the driving span of rope next to
BELT, ROPE,
the idler
331
and some time must elapse before this load can be equalized over all the spans. Mr. T. Spencer Miller* has pointed
out that the general tendency to unequal straining
may
be some-
what obviated, where the sheaves are of
different diameters,
by
making
the angle of the groove in the
small sheave somewhat
sharper than that in the larger, so that the product of the arc of
contact and the cosecant of half the groove angle
are equal;
thus making the tendency to slip equal.
The above
two systems.
are the principal points of difference between the
The
particular conditions of the installation
must
be considered in making a choice between them.

144. Sheaves for Fibrous Ropes.
The
sheaves over which
ropes are to run deserve special attention.
Care should be taken
Fig. 122
(a).
Fig. 122 (b).
Fig. 122
(c).
that the
form of the grooves, and, the
effective diameters are the
same
for all grooves of the
same sheave and
the surfaces should
be accurately finished and well polished, as any roughness or
unevenness seriously
of
affects the life of the rope.
much experimentation two forms

(a)
of grooves as
As the result shown in Fig.

In Fig. 122
(b)
122
and 122
(b)
have become most common.
the sides of the groove are straight while in 122 (a) the sides are
curved.
This curving of the sides makes the angle of the groove

flatter at the
somewhat
reduced
.slip
bottom and hence when the rope has been

it lies
in
diameter from wear
lower in the groove and will
little
more
This
readily than
is
when
it is
new and occupies a higher

article.
position.
of
importance in relieving the old rope of a

also said to assist the rope to roll in the
tendency
to pull
harder as indicated in the preceding

is
The curved
outline
* Transactions A. S.
M.
E.
Vol. XII, page 243.
332
MACHINE DESIGN
it
groove, a very desirable feature since

the rope.
distributes the
wear on
in the
The curved
it
groove
is
therefore
much used
groove.
is
multiple system.
rotates as
In the continuous system the rope necessarily

first
passes round the idler to the
The The grooves

the stretch
angle of the groove, as before stated,
usually 45
of idler pulleys for simply supporting the rope
when
is
great are not
made v-shaped but
as
shown
in Fig.
122
(c).
The wear
internal
of fibrous ropes is
both internal and external, the

to
wear being due largely

as
chafing of the fibres on
each other in bending the rope over the sheaves.

son sheaves should be
large as
possible,
For
this rea-
and, in general,
should not have a diameter less than forty diameters of the rope.
145. Deflection or Sag.
is
Where
the span between the pulleys
considerable the
amount
of deflection is
sometimes of import-
ance.
pulleys,
Since the deflection varies with the distance between

the size
and speed
it
of the rope
and the
difference in
elevation of the pulleys,

existing
is
impossible to express the relation
between them in a
single formula.
For the simple case
of the horizontal drive the

side
approximate deflection on the driving

for the
may be determined both
continuous and multiple
systems and also the deflection of the slack side of the continuous
system, where uniform tension
is
maintained by a tension weight.
In the multiple system, however, ample allowance must be
made
on
the slack side, as
new
ropes stretch very rapidly, and the
deflection
formed.
for
excessive before resplicing can be perMr. Hunt gives the following equation (transformed), computing the deflection in horizontal drives:
may become
A=.-zi-j^r^. A
2W
4
(I9 )
Where
tion,
T is
the total tension on either the slack or tight side deit is
pending on the side for which
desired to compute the deflec-
W the
to
weight of rope per
foot,
the span in feet
and A the
is
deflection in feet.
Where
will
the
2
,
tension on the driving side
assumed
be equal to 200 d
regardless of speed, the deflection
on the driving side
be constant for a given span.
As the
BELT, ROPE,
333
tension in the rope due to centrifugal action increases as the square

total tension T 2 on the slack and hence the deflection on the slack side decreases with the velocity, the span remaining constant. The value of T, may be computed and substituted in equation
of the velocity, there
is
an increasing
side for a fixed value of 7\;
(19) to find the deflection.
Mr.
Frederick
Green *
gives
the following
approximate
formula for computing the deflection
A =
Where
that
W XD ~sV~
as in equation (19),
(2o)
the symbols are the
same
and from
which he has calculated the following table on the assumption
T
x
200
(P.
TABLE XXI
Sag on Slack Side.
Distance
Sag on
Driving Side, All Speeds,
Feet.
between
Pulleys, Feet.
Velocity, Feet per Minute.
3.000
4,000
4,Soo
5,000
5.500
3 40 5 60 70 80 90 100 120 140 160
.19 34
45
80
I I
39 .69
1 .1 1
36 .64
.0
1 1
33
59 92
3
7
3
53 .84
1 .2
S3
.76
1
8 4
.0
i-7 .2.1
1.4 i-7
2 .1
3 2 4
5
3
41
5-4
9 9 12 9
2.9 3-5 4-3 6.2 8-5

11
.
1.4 1.9 2-5 3-2 4.0 5-7 7:8 10.2
1.6
2
.
3 3 7 5 3
7 2
2.7 3-3
9 5
4.8 6.6 8.6
WIRE-ROPE TRANSMISSION
146. General.
Ropes made
of iron or steel wire have been
used to a considerable extent for transmitting power over comparatively great distances.
The
introduction of electrical transfield as far as

still
mission has, however, greatly curtailed the

transmission
* See
is
power
concerned;
although wire ropes are
much
"The Blue Book
of Elope Transmission,"
by American Mfg. Co.
334
MACHINE DESIGN
etc.,
used for conveying materials such as coal, rock,

of buckets attached at intervals along the rope.
by means
rope in
The
such installations moves at very low velocities and constitutes a

different
problem from that of power transmission.
Wire ropes
are also
much used
for hoisting
work such
as elevator
and mine
work and for carrying static loads masts and suspension bridges. 147. Materials for Wire Ropes.
of
as in supporting smokestacks,
wrought
used.
iron,
open hearth
Wire ropes are usually made For very steel, or crucible steel.
severe
is
work especially strong crucible steel known as plough steel For a few special cases, copper and bronze are employed.
A. Roebling's Sons Co. publications give the
tensile strength of various
fol-
The John
Swedish Iron
lowing values for the
kinds of wire.
45, 000 to 100,000 lbs. per sq. in.
Open Hearth
Crucible Steel
Steel
50,000 to 130,000
130,000 to 190,000 190,000 to 350,000
lbs. lbs.
per sq.
in.
per sq. per sq.
in. in.
Plough
Steel
lbs.
They also
state that
a testing machine,
of all the wires.
it is difficult to obtain from a sample of rope in more than 90 per cent of the aggregate strength This is due to the difficulty of getting a perfect
grip
on the rope so that
all
the wires will carry their full share of
the load; and also because the inner wires of a strand are shorter
than the outer wires and are therefore more quickly overloaded.
The
wires,
on account of the twisted construction, also tend to

under heavy loads.
mutually cut into each other, thus rendering them more liable
to fracture
On
account of this latter action
ropes
made with
a short twist break at a lower percentage of
their full strength
than those of a longer
twist.
Wire ropes for power transmission are usually made of iron or soft steel and are laid up with a soft core of hemp in order to give greater flexibility. They cannot be run on metallic surfaces and the sheaves must be lined at the bottom with soft rubber or similar yielding material.
148.
Power Transmission by Wire Rope.
Great care must be taken that the rope does not chafe and, unlike the
sheaves for fibrous ropes, the grooves in sheaves used
for wire rope are so
formed that the sides
of the groove
do not
at the
compress the ropes.
In wire-rope sheaves, the radius
BELT, ROPE,
335
bottom
itself
of the groove
is
always greater than that of the rope

flat
so that wire-ropes drive, like
belts,
simply through
the friction on the bottom of the groove, due to the tension of

the rope.
or
The
lining of the
bottom of the groove
(leather,
wood,
some other comparatively soft material) gives increased friction as well as less wear of the rope. The sheaves should be as large
possible
to
as
minimize the bending
effect
on the rope; one
hundred rope diameters being often taken as the minimum

diameter of the sheave.
The
is
general theory and equations developed for fibrous rope

It
hold also for wire rope, proper constants being substituted.
evident from this discussion that wire ropes can safely trans-
mit a greater amount of power than fibrous ropes of the same

diameter, because of the
much
higher allowable tensile stress.
is taken from a circular John A. Roebling's Sons Co., shows the power that may be transmitted by iron ropes of various sizes with sheaves of These values are for a different diameters and rotative speeds. rope made with six strands around a hemp core, each strand
The
table
on the following page, which
of the
consisting of seven wires.

for the
This table does not make allowances

to the
change of
stress
it
due
change of centrifugal force
at
various speeds;
but
does consider the influence of the sheave

stress.
diameter on the bending
For example
a y%" rope on an
eight-foot sheave running 100 r.p.m., transmits
only 32
H.P.;
while the same rope transmits 64 H.P.,

foot sheave at
80 r.p.m. or at the
120
it
when running on a tensame linear velocity. By rea manila rope

of
ferring
to
Fig.
is
seen that
at the
1%"
diameter transmits only 30 H.P.,
most economical
velocity,
or at about twice the velocity in the above instance.
For hoisting and for transmission, must be much smaller than those given
if
the sheave diameters
in the
preceding table, a
strands around a
more
flexible
rope
is
used.
This consists of
is
six
hemp
core, but each strand
made up
of 19 wires,
which
are, of
course, of smaller diameter than those used for corresponding sizes

of seven-wire strands.
in
The
lining of the bottoms of the grooves

If
it
the sheaves should be maintained in good repair.

irregular,
be-
comes
through wear, the rope
may
be bent
at
a sharp
33&
MACHINE DESIGN
TABLE
XXII.
TABLE OF TRANSMISSION OF POWER BY WIRE ROPES *

u
<u
to
M-l
Optn fe.S
u
S.fl
8'g
V-i
s
-^
O
Ph
s
<<
<uP4
.
'o
1
Ph
4)
<up4
3
3 3 3 3
o H
8o
IOO 120 I40
23
C O
IS
Q
i to
1h
5*
3
3
3
7
w
35
I40
20
19 19
A
f f
1
23 23 23
23
1 1 I
3*
4
8 8 8
8
9
80
IOO I20
26
32
4*
4
5
19
19
t
39
45
f
I
80
IOO 120
I40
4 4 4
5
23
80
IOO 120 140
j
I
20
i
(10
20
19
23 23
22
22
1
1
_7_
6
7
9 9
9
140
20
19
}
A A A
9
f
I
47 48 58 60 69
73 82
f
1
5
80
IOO 120
j
"i
20
19
!
16
9
11
16
5
5
T%
10
10 10 10
12 12 12 12
80
IOO 120 140
ii9
1
18
f
5
xi
1
22 22
21
TV
16
*3
1S
140
ji9 1i8 |i9

i
8
\
f.
11 16
84 64 68 80
85
i
1
18
H
T6"
!
j
Q6
102 112 119
6 6 6 6
7
80
IOO
1 1
1
14
17
ji9
(
18
i7
21
21
80
IOO 120
(18
1
\
1 1
120 140
20
23
(18
(
21
1
9
17 18 i7
T6
11
16
\
4 1
93 99 116 124 140 149

173 141
80
IOO 120
20
16
20
25
120
16
1
I
f
20 20
*
_9_
14 14
80
IOO
16
30
P {?
I*
1
I*
1
148 176
185
Taken from a
J.
publication of the John A. Roebling's Sons
Company,
of
Trenton, N.
The above table gives the power transmitted by Patent Rubber-lined Wheels and Wire Belt Ropes, at various speeds. Horse -powers given in this table are calculated with a liberal margin for any
temporary increase of work.
BELT, ROPE,
337
angle in passing over the high spots of the lining, with a resultant
increase in the stress of the wires.
This
last action,
however,
is
not equivalent, so far as the
life
of the rope is concerned, to run-
ning over a correspondingly smaller sheave, for every portion of
each wire
is
bent around each sheave once during every circuit

while
it
of the rope;
is
not likely that the same portion of the
rope will frequently come in contact with any single irregularity

in the lining.
ROPES AND CABLES FOR HOISTING

149. Fibrous
Ropes
for Hoisting.
In power transmission
it is
usually possible to install sheaves large enough to prevent the
bending action from seriously affecting the

in hoisting
size
life
of the rope;
but
work
this is
not always possible, on account of the

the resulting tackle.
if
and clumsiness
1
of
Thus, a manila
used for
rope of
inch diameter,
used for power transmission, should
run over a sheave

hoisting
it
at least
40 inches in diameter but

to
if
might be required
run over a block sheave 12 inches
or even 8 inches in diameter.
The
internal friction
life
and external
of the rope,
chafing are, in such cases, very great and the
even when working
at a
lower
stress, is greatly
shortened;
but
in hoisting tackle, the
frequency with which any portion of the

is
rope passes over the sheaves
much
less
than
is
ordinarily the
case in power transmission, on account of lower speed.
Theoretical considerations are of

installations,
little
or
no help
in hoisting
and recourse must be had
to successful practice
on
which, fortunately, there are considerable data.

table,
The
following
from a paper presented by Mr. C. W. Hunt, before the
A.
S.
M.
E., gives the results of a
long series of observations, and

It
indicates the most economical size of rope for a given load.
has been found, by experience, that ropes larger or smaller than

those
recommended
in the table are shorter-lived
under the load
indicated.
The
speeds indicated in the table are defined as
follows
"Slow"
Derrick, crane,
and quarry work;
50 to 100 feet
per minute.
33%
MACHINE DESIGN
"Medium" Wharf and

minute.
cargo work;
150 to 300 feet per
"Rapid"
400 to 800
C.
feet per
minute.
TABLE XXIII WORKING LOAD FOR MANILA ROPE

A.
B.
D.
E.
F.
G.
H.
Diameter of
Ulti-
Working Load
Rapid.
in Pounds.
Minimum Diameter of Sheaves

Medium.
mate
Medium.
Slow.
Inches.
Rope, Strength, Pounds.
Rapid.
Slow.
I* I* 1* I* if if
7,IOO 9,000 11,000 13,400 15,800 18,800 2I,8oo
200 250 300 380 45 530 620
400 500 600 750 goo

1,100 1,250
1,000 1,250 1,500 1,900 2,200 2,600 3,000
40
45
12 13
5
55
14
15
9 10
11 12
60
65 70
16
17
18
13 14
150.
elevators
Wire Hoisting Ropes. On overhead travelling cranes, and mine work, iron or steel cables are used almost
it
exclusively, as here
is
usually possible to install sheaves or
drums
of large diameter.
is
For rough
service,
deep mine work or
wherever great strength
necessary, these ropes are sometimes

steel.
it
made
of crucible or
plough
Great care should be exercised

is well,
in installing such ropes
and
if
in general, to obtain the
advice of the manufacturers before selecting any rope

crucible steel, especially
made
and
is
of
great safety
is
desired.
factor of
for
safety of at least 5 should be used in ordinary work,

elevator, or similar work, a factor as high as 10 or 15
some-
Table XXIV, taken from a publication of the John A. Roebling's Sons Co., gives data on standard hoisting ropes. For open-hearth steel the strength as given for iron rope
times desirable.
may be
increased 25 per cent.
It
will
be noticed that these

5.
tables are based
on a factor of safety of
151.
(a) (b)
(c)
CHAINS AND CHAIN TRANSMISSION Chains may be conveniently divided into three
Chains for raising and supporting loads. Chains for conveying purposes.
Chains for power-transmission purposes.
classes:
BELT, ROPE,
Chains for Hoists.

first
339
In the
class are
such chains as are

this
used on cranes and hoisting appliances.

are
Chains of
character
made
with elliptical-shaped links and should be manufactured
of the best
is
wrought iron
to insure perfect
welding where the link
joined.
The
links themselves should be as small as possible

to the pull of
to
minimize the collapsing action or bending due
the adjacent links, and also that due to winding the chain
circular
close, or
upon a
drum.
Such chains are sometimes called
short-link,
crane chains.
TABLE XXIV
STRENGTH OF IRON AND STEEL HOISTING ROPES
Swedish Iron.
Crucible Steel.
v
.5
fe
DO
a
ooC
60
t-
p;
S
.2,23
u
<
ii .95
to
P O
J)
91J
.g
<U
> D
il
5
114
95 78 62
<D
3*
<
s
2f 2* ai
2
If If I* If I*
9-8 5 8.oo 6.30 4.85
22.8 18.9 15.60

12 .40 9 .60
16
15 13 12
2f 2i 2|
2
n-95
9-8 5 8.00 6.30 4-8 5
415
3-55 3.00 2.45
2 .00
1
48 42 36
3i
25 21
17
10
8.40
7 .20
8* 7i
7
If If
415
3-55 3.00 2-45
2
1
228 190 156 124 96 84

72 62
45 6
10
37-9 31.2 24.8

19 .2
9i 8i
8
7i
6 .20
Ii If
16.8 14.4 12.4

10 .0
5.00
4 .20
3 40
2 .60
1 1
6i
6
^
ii
1
6* 5! 54
5
50
42
ii
i
.00 .58 .20
A A
1 i
i J f
1
-58 .20
.62
13
5i 4i
4
0.89
.50
9-7
6 8
5 5
.94
36
.
10
39
4 4
3 4
2
1
1
0.30
.
22
5
7 .2
0.88 0.68 0.50
3i 2| i
2
t 1
0.89 0.62
.50
34 26 19.4
13 6
11
8.4 6.8 5-2 3.88

2 .72
2
4*
4
3i
3
tV
.20 -36 .00
ii
1
0.15
.
034
.24
A A
1
o-39 0.30
.22
8 8 6 8
5
1.76
1
1
2i If Ii
Ii
I
0.15
.
10
10
3 4 2 4
0.68 0.48
The
strength of a chain link in tension
is less
than twice the

is
strength of a bar of the iron from which the chain
made on
account of the bending action due to the manner in which the

load
is
applied,
and
also
on account of the weld.
If
the
breaking load in pounds, and d
the diameter in inches of the
340
MACHINE DESIGN
is
bar from which the link

equation
made, then the following empirical
may be
used for iron crane chains.

2
W
In
54,000 d
(1)
The working
value or
W=
load
(W)
2
should not be more than one-third this

(2)
18,000 d
many
cases a lower stress than indicated

the load
is
by
(2)
should be
adopted.
Whenever
not a direct pull, but severe
bending
stresses are also induced, as in chain "slings" for
handling
heavy iron castings, the chain should have great excess of strength.
Chains should be carefully inspected and tested or " proved" before using. The " proof " usually applied is one-half the ulti-
Where chains are used for hoisting work, they are become badly strained. Annealing by heating allows a readjustment of the structure of the iron, and this should be done periodically with all such chains, particularly chains used
mate
load.
likely to
for slings.
This also affords an opportunity
to
thoroughly inspect
chains which are greased in operation.
The
uncertainty regardit
ing the exact condition of a chain in service, and the fact that
gives no warning of weakness, but may break at a load below the normal working load, have caused them to be largely replaced, on
such appliances as overhead cranes, by

of the strength of the latter
tion.
is
steel rope.
The
state
more
easily determined
by inspecshape
in
Weldless*
steel
chain rolled from a bar of
special
is
has lately come into use to some extent.
The
chain
made
lengths of from 60 to 90 feet, and the lengths are joined together
by a link made
of special welding steel.
They
are said to be
much
stronger than iron chains.
152. Chain
Drums and
Sheaves.
Drums on which
crane
chains are to wind should be carefully grooved so that alternate

links
lie flat
on the surface
of the
drum; and should have

the drum.
sufficient
capacity to receive the chain in one layer, as overwinding brings

severe stresses on the parts
eter of the
wound upon
The diamis better.
drum
should in no case be less than twenty times the

this
diameter of the chain used, and thirty times

* See
diameter
"Machine Design," by H.
J.
Spooner, page 452.
BELT, ROPE.
341
Where
it
is
not possible to have the chain wind upon a drum,
pocket chain wheels are often used.
These wheels are made with

fit.
pockets around the periphery into which the links

links are prevented
The
the wheel;
from coming out by a guide over a portion of and hence cannot slip on the sheave. Anchor chains,
of certain forms of chain blocks for raising weights,
and the chains
run over such sheaves.

153. Hoisting-Hooks.
weights deserve special attention.
The hooks used for raising heavy They are usually made of
Fig. 123 (a).
Fig. 123 (b).
steel or iron
some
employed to hook can be kept low the use of steel castings may be justified; but where the load is great and the fibre stress in the hook necessarily high, to avoid clumsy proportions, the hook should be forged from ductile material. Let the hook in Fig. 123 (a) be subjected to a vertical load P;
forgings although steel castings are
If the stress in the
extent.
then
XY,
the most dangerous section,
is
apparently acted upon
by a
direct stress p'

flexural stress p"
P
due
(where
to the
is
the area of the section)
and
by a
moment Pa;
the stress p" being
342
tensile at
MACHINE DESIGN
and compressive
at F.
The
theory of article 19,
therefore applies, apparently,
and equation
M (Table
VI)
may
be used
to design the section, or
p=
made
of materials
p'
P"
P Pae + A I
-r
this
Experience shows that members of
kind,
even when
whose
tensile
and compressive
usually
elastic limits
are about the same, almost invariably yield to rupture on the
tension side;
and the section
is
made
as shown,
the
gravity axis being located nearer the load, thus decreasing the
tensile stress
and increasing the compressive
stress, as
computed
by the above equation.

in a
Recent investigations* have shown that

in this
curved
beam loaded
manner, the neutral axis does

is
not coincide with the gravity axis, as in straight beams, but

located nearer the tension side, and the above theory
defective, as the true tensile stress
is
is
therefore
greater than that given
by
equation
M.
That
this is true is
fail in
borne out by the fact men-
tioned above, that hooks
tension
when designed with an

stress,
is
apparent compressive stress considerably above the tensile

although the elastic limit for either stress
about the same.
The
however, somemay be what complicated and it is believed that equation safely used if due care is taken in assigning the limits of stress. Hooks for small cranes and hoists are much more likely to be
applicationf of the
is,
more accurate theory
loaded frequently to their

large loads;
full
full
capacity than hooks for raising
thus a hook on a five-ton crane
may
be loaded to
its
capacity several times every day, while the hook of a twenty-ton
crane would be thus loaded at rare intervals.
The
stresses in
small hooks must therefore be kept low, and fortunately this can
be done without making the hook clumsy.
As
the size of the

in-
hook
increases,
however, the stresses must necessarily be
creased to avoid clumsiness, but the larger the hook the less
frequently will
it
be
fully
loaded and a working stress as high as
* See "Strength of Materials," by Slocum and Hancock.

" t See
On
Theory
of the Stresses in
Crane and Coupling Hooks.
With Ex-
perimental Comparisons with Existing Theory," by Professor Karl Pearson and
Mr. E.
S.
Andrews.
Messrs. Dulan
&
Co.
BELT, ROPE,
343
15,000 pounds per square inch, or more, is as safe in a fifty-ton hook as 10,000 pounds per square inch would be in a ten-ton
hook.
(See Art. 25.)
The most valuable data on crane hooks is that given by Mr. Henry R. Towne in his ''Treatise on Cranes," as a result of both Fig. 123 (a) and the mathematical and experimental work. following formulae give the most important dimensions of a hook according to this work, and these proportions have been much used with uniform success. The basis for each size is a commercial size of round iron or dimension A.
In the following
in tons of 2,000
formula A
pounds.
to provide
is
the nominal capacity of the
hook
The dimension
ample room
is
assumed
arbitrarily but
so
as
for the slings.
The
following measure-
ments are then expressed
in inches:
D =
G=
The
.5
a +
1.25
H
I
.75
= 1.0SA = 1.33 A
= 1.13A L = 1.05 A
made from
following gives the capacity of the hooks
various sizes of bar stock:
TABLE XXV
Capacity of hook
Size of bar
in tons
i
1
I*
1
IT
4
2
IO
in
inches
T 1 1
I*
If
if
2*
2h
2|
3l
It is to
be noticed that the stresses allowed by Mr. Towne's
proportions are very low.
Thus
in a ten-ton
hook the dimension
is
3^4"
or,
allowing for finishing, the dimension

tensile stress in the
It
B may be
taken
as 3",
which would give a
pounds per square inch.
shank of only 3,000 should be borne in mind, however,

the designer has no
that hooks are subjected to
much abuse and
assurance that they will always be loaded with a true axial load,
for
improper arrangement of the sling often throws the load more

is
toward the point of the hook and the member

carry a bending
called
upon
to
moment
greatly in excess of that for
which
it is
intended.
When, however, hooks larger then those covered by Mr. Towne's work are to be designed his proportions lead to clumsy dimensions. Thus a twenty-ton hook would require a shank
344
MACHINE DESIGN
diameter and a fifty-ton shank would be 6}4" in diameter.
4X" in
Fig. 123 (b)
shows a twenty-ton hook of Norway iron which has

practice.
been successfully used in

$yi" in diameter
is
The threaded shank
being
pounds per square inch but yet is only as large as the shank of a ten-ton hook as given by Mr. Towne's dimensions. Examination of current practice and measurements taken from a number of large hooks
therefore stressed to about 6,000
in successful service indicate
an allowable
tensile stress at
X,
as
computed by equation
M of Table
VI, ranging from 10,000
lbs.
per square inch in ten-ton hooks, to 15,000 lbs. per square
inch in fifty-ton hooks.

154.
Conveyor Chains.
Chains for conveying and elevating
materials, such as grain, coal, ashes, etc., are usually
made
of
Fig. 124 (a).
Fig. 124 (b).
Fig. 124 (c).
malleable iron, the links hooking together in some manner.

style of
This
forms,
is
chain
is
known
as link belt.
On
account of the diverse
purposes to which they are applied, they are
made
in
many
and the
usually
selection of the particular
form
for a given
problem
made
in conference with the
manufacturer or taken from
trade catalogues giving the desired information.
This form of
agri-
chain
is
also extensively used in
rough machinery, such as
cultural implements, for the transmission of power.
In such
cases the chains
must be run at low speeds, as they become noisy and unreliable even at moderate velocities. The chains heretofore 155. Chains for Power Transmission.
discussed move, necessarily, at low velocities, but of late a
demand
has arisen for chains which
may be run
at high speeds for the
purpose of transmitting power.
Such chains are used when a
BELT, ROPE,
positive velocity ratio
shafts,
345
must be maintained between the connected and where the distance between shafts is so great as to
make
and
tooth gearing inconvenient.
Of
this
class there
are at
present three principal types, namely, roller chains, block chains,

so-called silent chains.
roller
Fig. 124 (a) illustrates the simplest
form of
links,
chain in which the pin
is
riveted fast in the outer

roller
and
rotates in the inner links.
The
and
lessens the
friction against the tooth.
In this form of chain the wear between

is
the pin and the inner link
excessive,
for this reason
it is
now
little
used for power transmission.

In this form
It is
sometimes made
is
without the roller and with several inside links and

as stud chain.
it is
then
known
used for very low velocities only.
Fig. 125.
Fig. 126.
The form shown

ing
in Fig. 124 (b) is
most common.
Here the bushis
is
pressed into the inner links, and the pin, which
riveted
fast to the
outer links, bears over the whole length of the bushing.
The
124
roller
(c),
rotates
on the bushing.
In the block chain, Fig.
the pin also bears over the whole thickness of the block
roller is necessarily omitted, there is
D, but since the

to
more
friction
against the tooth.
Roller chains
may
be used for velocities up
about 800
feet
per minute, and block chains up to about 500
feet
per minute.
The
defect in the operation of the roller or block chain
may
is
be seen by referring to Figs. 125 and 126.

new, and has the same pitch as
as
the wheel,
it fits
When
the chain
down on
the wheel
shown
in Fig. 125,
but in a very short time the chain stretches
346
slightly,
MACHINE DESIGN
due
to
wear
of the joints, thus increasing the pitch of
the links.
The
wheel, on the other hand,
not change the pitch.
may wear but The operation of the chain is then

of the tooth as they
is
this
does
as
shown
in Fig. 126, the increased pitch causing the rollers to ride higher
and higher on the back
move
refund the
The roller A coming down to its seat.

sprocket.
shown
Before
fully seated while
is
just
can become
fully seated
must
rise,
and this action takes place
when A and B
are carrying
full load.
As a consequence the chain does not run quietly and smoothly and the wear is excessive, thus limiting the speed at which the chain may be run. This difficulty is sometimes met
by the arrangement shown
in Fig. 127.
Here the pitch
of the
Fig. 127.
chain
when new
is
made
a
is
little less
than the pitch of the driving L; the pitch
sprocket,
and clearance
allowed between the roller and the
tooth, so that the driving

of the chain being
is
done by the
last tooth
such that the incoming
roller
M just clears the

the root as at
is
back of the
first
tooth and seats
itself close to it at
N.
As the chain
stretches, the rollers

till
move backward toward

reached,
the faces of the teeth,
a condition like that in Fig. 125
and riding commences.

is
The
127.
pitch of the driven sprocket wheel
is
and the condition when new As the chain stretches, the rollers move gradually backward away from the driving faces of the
to that of the chain,
made equal
that
shown
in Fig.
tooth, the driving being
done on the
last tooth
P.
It is
evident
BELT, ROPE,
347
that this construction extends the time preceding the condition
shown
in Fig. 126.
When
this construction is used, the
form of
it
the tooth must be slightly modified.
is
obvious that
if
the outline of the tooth
M be
an arc of a
circle
struck from the centre of the roller (2), this roller will swing from
its
position
i'
rolling
on the face
of the tooth,
and
this is the
(2)
usual outline.
is
But before
it
roller (3)
can take the load, which
about to give up,
must be
fully rooted against the next tooth;
Therefore as
(3) to
whereas (from Fig. 128), a small distance now separates the two. (2) rolls up the curve of the tooth it should allow
slowly settle back in place.
The
tooth outline
little
is
therefore
struck (as
shown on
if),
from a point a
inside the pitch
polygon so as to give a curve tangent to the
first
and
last positions
Fig. 128.
of the roller.
This outline
is
also necessary for the
back of the
than
tooth in order to allow the incoming roller to swing in without

striking.
The
velocity of the chain
is,
therefore, a
little less
the theoretical velocity on account of this continual slipping back-
ward.
Brief reflection will
driven sprocket
rooted in place;
will creep as
it
may be
is
show that the tooth outlines of the struck from the centre of the roller when and that when the chain is stretched a little it wound upon the driven sprocket.
is
When
I\
roller
the roller (4), Fig. 127,

(5)
is
about to
roll
up the face of
not in contact with

will
(wear
having begun);
contact with
hence the chain
move ahead
till
(5) is in full
is
M.
ex-
The
load
is
greatest defect in this construction
the fact that the
carried entirely on one tooth and hence the
wear
is
348
cessive.
MACHINE DESIGN
This
may be
so great that the chain creeps forward on
the driven wheel so as to cause the incoming roller to strike the
tooth S, Fig. 127.
The above
chains.
is
difficulties
are overcome in the so-called silent
In these chains the inevitable stretching of the links

for in a peculiar
is
compensated
manner.
The
true theory of the

is
action of these chains

as follows:
to take
very complex; but the general action
as
the chain stretches, the links continually tend
up a position farther and farther away from the centre of

thus increasing the length of the sides of the pitch
suit the elongation of the link.
its
the sprocket;
polygon to
Each
link therefore
remains in constant contact with
own
tooth,
from the time of
Fig. 129.
engagement
till
release takes place.
The
is
links seat themselves
without sliding action and the operation
nearly noiseless.
In the Renold chain of this type, Fig. 129, the links move
relative to
each other on a round pin P, the shouldered ends of

fitted to the links
which are riveted into a washer W, thus holding the chain together.
In a later form half bushings of bronze are so
that the pin has a bearing over
its full
length;
but the relative In the

of
motion of the pin to the bush
is still
a sliding motion.
Morse chain
and B, are
links
this sliding is eliminated
by an ingenious form
steel parts,
rocker joint shown in Fig. 130.
The hardened
fitted respectively to the sets of links,
D and C.
While
thus
keeping contact along a fixed line they rock on each other as the
and
move
relatively to
each other, and sliding
is
BELT, ROPE,
eliminated.

simple
tension
349
When
transmitting
between the
surfaces as
sprockets, the parts
and
are in contact on
flat
shown
little
at E.
This construction has the advantage of requiring
or no lubrication, hence the chain
maybe run
at
higher speeds
than others requiring lubrication, the speeds of which are limited
by the
cant.
velocity at
which centrifugal action throws
off the lubri-
The Morse chains also work well in dusty places. The efficiency of both of these chains is very high, the makers
Morse chain claiming an
efficiency of nearly 99 per cent.
of the
Fig. 130.
Such chains are particularly useful

too far apart for gearing,
places where
positive
for connecting shafts far

is is
which are
and not
It
enough
for a belt,
and
in
connection
desirable,
as
in
motors
driving heavy machine tools.

this
to

tension
form
of
transmission requires no definite

to
on the
slack side of the chain

the tight side;
produce a certain driving force on

is
and hence the pressure on the bearings
much
reduced, for a given effective pull on the wheel rim.
CHAPTER
XIII
APPLICATIONS OF FRICTION
Friction Wheels for
Power Transmission
When
and
it
is
required to drive
driving
is
a rotating
member
intermittently,
the rate of
not necessarily positive, friction wheels have been found very

useful.
They
is
are particularly applicable
where the amount
of
power
comparatively small, as in feed mechanisms, but they
may
also be used for
heavy work when properly constructed.

of friction wheels
For continuous driving the transverse sections
\ S^^S
-
m
Fig. 132. Fig. 133.
-IM
used in
must be
practice.
circular in cross-section,
and
this form, only, is
Figures 131 and 132 show
common
forms of
friction wheels.
In Fig. 131
let
be the driving wheel which rotates continuously,

is
and
let
be the driven wheel which
required to be driven
intermittently.
The
it
shaft of
is
so mounted that, by
means
of a
lever attached to the bearing,
A may
be pressed up against
B
no
with a force P, or
contact exists.
If
can be moved
slightly
away from
until
now
the force
is
applied to the bearing

is set
(which should be close to A), an equal and opposite force
up
in the bearing of B,
and the wheels are pressed together

35
at the
line of contact. will
}
35*
The
ft
resistance to slipping at the line of contact
be fiP where
is
the coefficient of friction of the materials

if
t
of
which the wheels are made; and

Theoretically,
uP
is
equal
to,
or greater
than, the resisting force at the surface of B,

rotate.
will
cause
to
A and B
will roll together with
pure rolling
motion, but practically this cannot be attained, as even with very
hard materials the wheels

(See Art. 108.)
flatten slightly at the line of contact.
Fig. 132 illustrates the application of friction wheels to shafts
which are not
parallel to
each other, the wheels here having the

is
form of
tion
rolling cones.
Obviously the principle
of wide applicaFig.
and many combinations

illustrates
of friction wheels are used.
133
a friction wheel arranged so that the driver
Fig. 134
Fig. 135 (b).
Fig. 135 (a).
can rotate the driven wheel
in either direction,
depending on
for
whether
it
is
pressed against the surface
or the surface n.
Fig. 134
shows a form
of friction
mechanism much used
imparting variable speed to the driven shaft.
The
at
driver
may
be moved along the shaft
at will.
is
When
A' the angular
velocity of
velocity of
B B
is
minimum.
As A
is
increases.
When A
it
moved inward, the rotative moved across the centre of B
to the other side, the direction of rotation of
is
reversed.
If
were
infinitely thin,
would, theoretically,
it
roll
upon
with
pure rolling motion.
Since, however,
must have an appreciable
width of face, and since the velocity of

it is
varies with the radius,
evident that there must be

this
some
sliding at the line of contact.
For
reason the thickness of
must, for best results, be kept
small compared to the radius of B.
352
MACHINE DESIGN
157. Materials for Friction Wheels.
The
driven wheels of
friction devices should
always be
made
of a harder material than

is likely at
the driver, for the reason that the driven wheel
any
time to be held stationary by the load, while the driving wheel

revolves against
it
under pressure.
This action, while severe

it
on the
driver, does not tend to

flat
wear
out locally, while
it
does
are,
rapidly wear
spots on the driven wheel.
Driven wheels
therefore, almost universally
made
of iron,
of
wood, leather, paper, rubber or of

the most
and driving wheels some composition of

various forms of
these;
common
being leather and
paper.
158. Practical Coefficients.
The
tangential force F, exerted
by
A upon
B, Fig. 131,
\.
is
dependent on the pressure

necessary to
and the
the
It is, therefore,
know
allowable pressure per unit of length along the contact elements
and
also the value of
/jl
for the particular materials used.

is
most comprehensive investigation of these relations
that
The made
by Professor Goss,* whose experiments cover a variety of materials. He recommends the following pressures, which are about one-fifth
of the ultimate crushing strength
of
the respective materials.
Safe Working Pressures per Inch of Contact.

Material.
Pressure.
'.
Straw
fibre
150
Leather fibre Tarred fibre
240
240
150
100 to 150
all
Leather
Woodf
Professor Goss found that the coefficient of friction for
the
wheels tested approached a
maximum
2
value
when
the slip be-
tween the two wheels was about

limits,
per cent, and, within narrow
was
practically independent of the pressure of contact.

to
He
found these values

values up to .515.
range for different combinations from low
In these experiments the friction due to the
bearings was neglected.

roller type and, probably,
The
bearings,
less
however, were of the
absorbed
M.
E.
power than the ordinary

XXIX.
* See Transactions A. S. f
Vol.
The
value for
wood
is
not from Professor Goss's paper.
bearing.
353
between
Professor
of
//
Making due allowance

those
for
the
in
difference
practice,
laboratory conditions and
found
Goss recommends the following approximate values

the various combinations.
that allowance
coefficient
* for
In this connection
it
is
to
be noted
must be made
the
for a decrease in the value of this
when
the linear velocity of

driver
is
the
driver
is
great,
in
the
case where
starting the driven
wheel under
load (see Art. 28).
Working Values of Coefficient of

Materials.
Friction.
Coefficient of Friction.
and cast iron and aluminum Leather fibre and cast iron Leather fibre and aluminum Tarred fibre and cast iron Tarred fibre and aluminum Leather and cast iron Leather and aluminum Leather and typemetal Wood and metal
Straw
fibre
o 26
.
Straw
fibre
0.27
0.31 o 30
.
0.15 0.18
0.14 0.22
025 025
If
159.
Power Transmitted by Friction Wheels.
be the
velocity of the surface of the friction wheels in feet per minute,
the total normal pressure in pounds,
the resulting tangential
force,
and
a the coefficient
is
of friction;
then since
F =
/iP,
the
is
rate at
fiP V,
which power
transmitted in foot pounds per minute

is
and the horse-power
HP
of
if
nPV
(1)
/
33,000
d be the diameter of the driver in inches,
the length of face
in inches,
the allowable load per inch of face,

is
and
N the number
.
of revolutions per minute, the horse-power
P.
fiwl 12
nd
N
-
x 33P
0.000008
uwldN.
face,
(2)
Example.
How many
horse- power can be transmitted
by a
run-
straw-fibre friction pulley of 8" diameter
and 6"
when
ning at 500 r.p.m., the driven wheel
:o
be of cast iron?
*The
23
coefficient for
wood
is
not from Professor Goss' paper.
354
Here d
'
. .
MACHINE DESIGN
8",
6",
N
X
500,
fx
=
X
0.26,
w =
8
150
H. P.
.000008
0.26
150
500
7.5
It
may be
a
little
noted that the horse-power per inch of width of
face
is
more than
unity, for a surface speed of 1,000 feet,
as in the above example.
This corresponds closely
to the empiri-
cal rule given for belts in Art. 136;
and corroborates the empirical

face
is
rule often used that the

friction
same width of
necessary for a
wheel as for a
belt, to
transmit a given horse-power at
the given speed.
In the case of bevel wheels (see Fig. 132) the component

of the applied force
presses the wheels together
and
R =
The
velocity of the
mean circumference
of the driver
may be
taken as the velocity of transmission.
In face friction driving as in Fig. 134, the width of the driving

wheels should be kept as narrow as possible for best
If
results.
is
the
velocity
of the outer edge
of the driving
wheel
not
more than 4 per cent

above
drive
coefficients
at
greater than that of

used.
the inner edge, the
may be
Where
the driver must, at times,
a short distance from the centre, lower values of the
must be taken. and are then known as wedgeof

this
The
faces of a pair of metal friction wheels are sometimes

in Fig. 135 (a),
formed as shown
faced friction wheels.
The
object
construction
is
to
secure a greater resistance to slipping, with a given radial pressure.

It is to
be noted that the number of wedges does not

it
affect this
ratio,
but decreases the wear by distributing
over several sur-
faces.
This
last
item
is
important, as
it
is
easily seen that the
contact surfaces of the driver and the driven wheel can have the
same
velocity at
one point only, and that
at all
other
result.*
points
slipping or a grinding action occurs
and wear must
The
teeth therefore should not be very long.
In Fig. 135
surface,
(b), let
be the radial force applied
to the
wedged
the tangential force transmitted,
r>
the reaction on
10
* See "Kinematics of Machinery," by John
H. Barr, p^ge
each face and 26 the angle of the wedge; then the wedge
in equilibrium
is
355
held
by the force P, the reactions

due
to the
and the
frictional
resistances a
forces
wedging
action.
Equating
vertical
P =
(R
\
s i
n 6
'-
-f
cos
11
0) or since
/
F =
ff l
2
(
R\
)
or
R =
.
F p-
'
P =
or
sin
I'-
+F
ll
cos 6
(1)
F =
p -r sin + a
1
(2)
cos 6
less
To
avoid sticking the angle
should not be
than 30
FRICTION BRAKES
160. Friction brakes
are used for controlling
machinery by absorbing energy through
and stopping from
some moving
part,
and dissipating
it
as heat.
Brakes used in
heavy work, and as dynamometers for measuring energy, must

often be fitted with water circulation to carry
away
the heat.
for
The
student
is
referred to treatises on
power measurement
is
a discussion of dynamometers.
161.
Block Brakes.
The
simplest form of brake
the block
brake as shown in Fig. 136.

lever A, presses the block
Here the force P, acting on the

Let the reaction
if
C against the wheel B.

Then
will
between the wheel and the block be R.

tangential frictional resistance
be rotating, a
its
aR =
oppose
motion.
With the arrangement shown

passes through
as a free
in Fig. 136, the line of action of
F
A
the centre of the fulcrum of A.
Considering
body and taking moments around O, then
for rotation
in either direction
(a
b)
= Rb
or since
R =
Fb
35^
MACHINE DESIGN
In Fig. 137 the line of action of

effect
does not pass through
O
A
and
its
therefore in writing the equation for the equilibrium of
must be considered, whence
Fb p =
a
r~i
b
en
u]
(2)
be used for rotation in a clockwise direction, for the arrangement shown, and the plus sign for rotasign
is
The minus
to
Fig. 137.
Fig. 138.
tion
in the opposite direction.
It is to
for clockwise rotation
when =
fi
or
when
c,
P =
o;
that
is,
the brake
is
self-acting
and
it
if
put in contact the
moment
of the frictional force will apply
with ever-increasing pressure.
Obviously such proportions should be avoided.

In a similar manner for Fig. 138
P =
Û+*]
R
R
r
(3)
the plus sign referring to clockwise rotation, for the arrangement
shown, and the minus sign to rotation in the opposite direction.

In
this class of
is
brakes the pressure of the brake

at the
against
the wheel
wheel.
opposed by an equal force
bearing near the

effect
is,
In the calculations above, the braking
due
to
friction of the journal is neglected, as its lever
arm
usually,
small.
this
It
cannot be neglected in designing the bearing, and for

is
reason this form of brake
not well adapted to heavy work.
357
Fig. 139 shows an arrangement of brake beams for heavy work such as is used in mining machinery. The force W, which may
be applied by a steam cylinder, acting on the system of levers,

causes the brake beams
of the wheel,
B and B to press equally on opposite sides and causes no pressure on the bearings of the drum.
in
If
T be the tension
2
each of the rods A and A. the frictional force

is
is
exerted on the wheel

that
F =
ti
T.
it
If the pin
is
so located
when
the load
applied
moves
to O',
and the centre
Fig. 139.
line of the
effect is
rod
passes through the centre of the pin P, a toggle
obtained and the tension intherods^. and^4may be

in fact with
made
any desirable value;
such an arrangement care must
be exercised in adjusting the brake that such pressures are not
brought on the pins as

operating force
will
cause failure by shearing.

is
When O
and the
a
moves flown to O' the brake

valuable quality in a brake.
ally lined with
"locked"
This
in position
may be removed.
wooden blocks
131),
last feature is often
Brakes of
as shown.
this character arc gener-
162. Strap Brakes.

lected
(see
If the effect of centrifugal
force
is
neg
Art.
and the
total
tensions in the
band (7\
358
and
MACHINE DESIGN
2)
be taken instead of the tensions per inch of width, equa-
tion (6) of that article reduces to
irrr 10
Where
also,
(I)
k
is
0.0076
fi
a,
a being the arc of contact in degrees.
If,
the total frictional force exerted
by the band upon the
wheel,
F = T -T,
i
(2)
It is
obvious that these equations are applicable to the discussion

Figs. 140, 141,
of
band brakes.
arrangement of band brakes.
and 142 show the most usual In Fig. 140 the end of the strap
Fig. 140.
Fig. 141.
Fig. 142.
which
is
subjected to the greatest tension
is
anchored, for con-
venience, at the pin which serves as a fulcrum for the operating

lever L;
it
could be anchored to any other convenient part of the
frame.
From
(1)
and
(2),
T =
2
F
10
Taking moments around
Pa = T
or
Fb
2
10
P =
Fb
a(io
K
-i)
(3)
'
which expresses the relation between the applied force

frictional resistance applied to the wheel.
and the
In Fig. 141 the end under greatest tension
lever
is
359
attached to the
and the end of
least tension is
anchored, hence for this case

io
k
P =
Fb r
a
Lio
-1
4
1
In Fig. 142 the end under greatest tension

at a shorter
is
anchored
to the lever
radius than the end of least tension; so that the force
which
it
exerts assists the operating force P.
This
is
known
as
a differential brake.
For
this case in a similar
manner
as above
K^]
o'S,
It is to
bl
if
io
b2
P =
is
o,
and
the
band
will
brake automatically; that brake

will
is if
any force
itself
applied
to the lever, the
continue to set
increasing force
of
till
motion ceases or rupture occurs.

its
up with ever This form

tendency to
brake
is
exceedingly dangerous on account of

if
jjl
"grab," especially
is
materially increased through a change
in the character of the friction surfaces.
Strap brakes are usually

light
made
of
wrought iron or
steel.
In
be
lined
work they may engage with a
cast-iron surface or
may
lined with leather; but in very heavy
work they should be
with wood.
FRICTION CLUTCHES AND FRICTION PULLEYS

163. Friction clutches
though made
in a great variety of
forms
well-
can, in a large measure, be classified under four principal types,
namely, Conical, Radially Expanding, Disc, and Band.

designed clutch should start
its
load quickly but without shock,

It
and should disengage quickly.

that
is,
should be "self-sustained,"
when
the clutch
is
driving,
no external force should be

In addition,
it
necessary to hold the contact surfaces together.

is
often necessary that the clutch should "lock" in place, after
the
manner
of the
brake
in Fig. 139.
164. Conical
Clutches.
is
Fig.
143 shows the elements of a
conical clutch which

shaft
self-sustained.
it.
The cone F
is
fast to the
S and
rotates with
The
pulley
rotates
upon
and
3 6
carries with
it
MACHINE DESIGN
the levers E.
When
the thimble
is
forced under
the rollers C, the levers
force the cone surfaces in contact.
Heavy
the
springs at
is
(not
shown) throw the surfaces apart when
thimble
withdrawn.
The
is
relation
between the transas that of the
mitted frictional force
and the force

the
applied to the cone, in a
direction parallel to the axis,
same
wedge
gearing in Art. 159, or
F = -^-ir stn $ +
The
is
11
P
fi
cos n
.....
(6)
angle
should not be
less
than
io, unless
some mechanism
provided for separating the cone surfaces, positively, when
Fig. 143.
Fig. 144.
desired.
For clutches that do not operate frequently, metal

is
is
surfaces are often used; but where the operation of clutching

frequent, one surface
165. Radially
usually lined with wood, cork, or leather.

Fig. 144
Expanding Clutches.
shows the
ele-
ments of a radially expanding, self-sustained clutch. The clutch body A is keyed to the shaft, while the pulley C rotates loosely upon the shaft. The circular segment B, which fits the inner surface of C, can be moved radially upon A. The loose ring G is operated axially by a forked lever fitting on the pins P. When the sleeve E is forced inward by the ring G, the links D force the segments B outward against C. In the arrangement shown the links have a toggle effect and can exert enormous pressure against
B,
36
hence adjustment must be carefully performed.
This
is
usually accomplished by
able,
making
the length of the link

devices,
adjust-
by means
of turn-buckles or similar
which also
provide a means of compensating for wear.
Usually the sleeve
has motion enough to carry the inner end of the link slightly past
the centre position shown, thus locking the clutch in place.
166. Disc Clutches.
disc clutch as
Fig. 145
in
shows the elements of amultipleautomobile work for connecting
sometimes used
the engine to the transmission shaft,

shaft
being fast to the engine
and
to the transmission shaft.
The
by
part
carries a
number
of discs, C,
which
fit
loosely in a radial direction but are
prevented from rotating relatively to

hold L, the cover of the case, in place. placed alternately between the discs
bolts
which also
second
set of discs
D,
are carried on the part
and compelled
spring
to rotate
with
it
by the keys G.
heavy
as
helical
(sometimes
made
of rectangular
section
shown)
presses the two sets of discs together with a

the clutch
T
,
is in
and the
shafts connected.
known load P, when The sleeve B while
compelled by the feather
to rotate
with the transmission shaft
can be moved axially by means of the grooved collar / and the

I being
ring J;
made
fast to
but built separately from

is
it
for
constructive purposes only.
When B
moved
to
the right the
spring
is
compressed and the pressure on the discs relieved.
The
It
is
discs often run in an oil bath to prevent " grabbing."
readily seen that while the force, P,
which presses each pair of

of pairs of contact
contact surfaces together
is
the same, the total frictional force
transmitted
is
proportional to the
number
surfaces n or
F = /inP
If the
(7)
mean
friction radius of the discs

is
be
r,
the frictional
mo-
ment
transmitted
Fr =
is
jinPr.
In Fig. 145, n
7.
The
known as the Weston clutch. Obviously any number of pairs of discs may be used. For large work the discs are sometimes made of iron and wood (or wood-faced).
above form of clutch
For small work, alternate discs of
steel
and brass are employed. and the discs run

in
Many
pairs of contact surfaces are then used
3 62
oil to
MACHINE DESIGN
prevent " grabbing."
The width
of the wearing faces of
the discs should be
made
small to prevent undue wear toward the

It is
outer edges of the discs D, as in a thrust block (Art. 105).

better to use
face.
more
discs of a smaller diameter than a few of great
167.
Band
Clutches.
Fig.
146 illustrates the elements of a
band clutch.
shaft)
The
clutch wheel
(which
carries the
wood-lined band C.
is
may When
be
the
fast to
one
thimble
(which slides on the shaft)
forced under the lever E, the iron
band
is
tightened and clutches the rim of the driven wheel B.
Fig. 145.
Fig. 146.
Obviously the principles involved are identical with those of the

strap brake, Fig. 140 of Art. 162.
For
light
work the band may
be lined with leather, but in heavy work, such as mine hoisting,

blocks of bass wood, or other soft wood, are used.
lining
is
The wood
usually
made
fast to the strap,
though occasionally on
self-
very large diameters they are attached to the wheel so that they
may
be turned true in place.
These clutches are made
locking by arranging for a toggle effect in some one of the operating levers.
Occasionally the band

the wheel to be driven.
is
made
to
expand inside
is
of the
rim of
It is to
be noted that
this case is not
the
same
as the one just discussed, but
a special case of a

radially
363
expanding clutch. The outward force exerted by the band may be computed by the theory of Art. 78, considering the band as a thin cylinder under compression, the compressive stress at any section being that due to the pressure applied by the
operating lever.
168. Magnetic Clutches.
A number
of clutches have recently
appeared which are operated magnetically.

generally of the disc type.
These are most

In
Evidently the general principles above,
regarding transmissive power, apply also to these clutches.
magnetic brakes, the load
is
usually applied by a spring, or
weight, and released by magnetic action, thus insuring safety

against accident should the electric service
Practical Coefficients for Brakes
fail.
and Clutches.
The most
In the
usual combinations of friction surfaces for brakes and clutches

are wood, leather, or cork with iron;
and iron with
iron.
multiple-disc type, brass or bronze on iron or steel are sometimes

used.
Mr. C. W. Hunt gives the following values
of
jj.
as the
result of considerable experience in designing clutches,
namely:
iron 0.2.
cork on iron, 0.35; leather on iron, 0.3; and for
wood on
It
For iron on iron
>i
may
be taken as 0.25 to
0.3.
should be
remembered
that
if
the friction surfaces are to be engaged at
high velocity, lower values must be assumed than for lower
speeds (see Art. 28).
The
pressure per unit area of surface
is
also
if
an important
this is
feature in the design of friction machinery, for
taken
too high, excessive wear will result.
Thus
in disc clutches the
pressure
is
usually taken at not
more than
25 to 30
pounds per
surfaces
square inch and lower values are desirable.
Wooden
should not be loaded beyond 20 to 25 pounds per square inch. If the clutch or brake is to operate frequently, ample surface
mu-t be provided
References:
to properly radiate the heat generated.
Transactions A.
Transactions
S.
M.
E., Vol.
XXX,
1908.
Inst.
Mechanical Engineers, July, 1903.
CHAPTER XIV
TOOTHED GEARING
169. General Principles.
When
it is
necessary that rotation
of one
it is
shaft shall produce definite
and
positive rotation of another,
evident that friction wheels, as discussed in the preceding
chapter, will not suffice where any considerable

is
amount
of
power
to
be transmitted.
In such cases the peripheral surfaces of

evident that any pair of surfaces
the transmission wheels are provided with teeth, so that the
motion
shall be positive.
It is
which
will roll together with
pure rolling motion, so as to give the

serve as a basis for the design of a
treat fully of
required velocity ratio,

pair of toothed gears;
may
and works on mechanism
the
methods
of
drawing the sections
of such surfaces for various
conditions and velocity ratios.

surface thus outlined shall
the angle which the shafts
friction wheels,
Whether the elements of the be parallel or otherwise will depend on

with each other, as in the case of
classified
make
and tooth gearing may be
* according to
the character of the pitch surfaces,
and the
relation of the axes,
thus:
Kind.
Relation of Axes.
Pitch Surfaces.
Spur
Bevel
Parallel Intersecting
Cylinders
Cones
Cylinders
Screw
Skew
Twisted Face
Not in one plane Not in one plane
Any Any
of these are spur, bevel,
Hyperboloids Any of the above
None,
strictly
The most important
and a few
special
forms of twisted and screw gears.
The motion
is
transmitted by a
pair of properly designed toothed gears
identical with that

If r x
of the base curves or surfaces rolling together.
and
r2
be
* See "Kinematics of Machinery," by John H. Barr, page no.

3 64
TOOTHED GEARING
365
the instantaneous radii of such a pair of surfaces at the point of
contact,
and
w,
and
<"
2
be their instantaneous angular velocities,
then
of both shafts
is
In the most
common
case the angular velocity
constant and hence r x and r 2 are constant, and
the rolling surfaces are circular in cross-section.
Thus
by the
Fig. 147
shows a portion of two gears whose
rolling surfaces are a pair of

circles
circular cylinders, represented in cross-section
C
on
and D.
D.
It
If the teeth are
properly proportioned the motion trans-
mitted will be identical with that produced by the rolling of
can be shown that the condition which such tooth outlines
Fig. 147.
Fig. 148.
must
that
fulfil in
order that the velocity ratio
may
be constant,
is
tlie
common normal
to the tooth outlines at the
point of contact
circles.
must always pass through the point of langency of the rolling There are many curves which can be used for tooth
outlines,
and which would
fulfil
the condition, but in practice the involute
only two are

cycloid.
commonly employed, namely,
and the
Fig.
147 illustrates a portion of two gears with involute teeth. wheel,
The upper
tact a
M,
a,
is
the driver.
Contact between two teeth

to the point of con-
has just begun at
and the
common normal
h passes
through the pitch point O.
As the wheels
rotate
the [joint of contact will
move along
the
the line
aOb
till
contact
to the
ceases at
b.
Hence
in
involute system the
normal
366
point of contact
the pitch circles.
Fig. 148
MACHINE DESIGN
makes a
fixed angle with the
common
tangent to
shows a portion
of
two gears with cycloidal
teeth.
Contact
is just beginning at a,
and as the gears rotate the point
of
contact will
b.
move along
the curved path a
b,
contact ceasing at
The normal
to the first point of contact is
drawn, and
it is
clear that the inclination of the

of the pitch circles, is a
normal
to the
common
tangent
maximum
at this point,
and continually
varies in direction though always passing through the point O.

It
can be shown that in the involute system the angular velocity remain constant, within the
limits of action,
ratio will
whether
the pitch circles are tangent or not;
but for the transmission of
constant velocity ratio with cycloidal gearing the pitch circles
must remain tangent.
The involute
it
gear, therefore, has a decided
advantage for general use and

cycloidal for most work.
has practically superseded the
A
is
fuller
treatment of the theory of

this
gear-tooth outlines, which
beyond the scope of
work,
will
be found in treatises on mechanism.*

170. Interchangeable Systems of Gearing:
It is desirable in practical
Standard Forms.
work
that any gear of a given pitch

pitch.
shall
run properly with any other gear of the same
In
order that this
may be
so,
certain limitations
of the tooth.
must be placed
In the cycloidal
upon
the
form and dimensions
system interchangeability
may be
all
accomplished, as far as the

the
tooth outlines are concerned, by keeping the diameter of
describing circle constant for
gears of the series.
Any
involute tooth outline will run properly with any other
similar outline;
and any gear with involute
teeth will
run with
any other gear having similar

will allow
teeth, as far as the length of the
involute outlines will permit, providing the thickness of teeth
them
to
mesh.
In order to obtain involute outlines of

series of gears with fixed
sufficient length,
circles,
and a
9,
nominal pitch
the angle
Fig. 147,
made by
the line of action with the
common
* See
tangent to the pitch circles must have a proper value,

for all gears of the series.
J.
and be constant
Design," part
In the systems in
also, "
"Kinematics of Machinery," by i, by F. R. Jones.
H. Barr, page in;
Machine
TOOTHED GEARING
though there most common use this angle is 14K dency in modern work toward a greater angle.
,
367
is
ten-
It is
found undesirable
in practice to
make
is
gears with less than
twelve teeth;
and
in
some
cycloidal systems the radius of a
twelve-tooth gear of the required pitch

of the describing
circle.
taken as the diameter

this will re-
For a twelve-tooth gear
sult in radial lines for the tooth outlines below the pitch circle,
i.e.,
the tooth will have radial flanks.
In the practice of the

circle
Brown
is
&
Sharpe Mfg. Co., the diameter of the describing
the radius of the fifteen-tooth gear of the series.
This gives
spaces
between the flanks of the teeth on the twelve-tooth, or
smallest gear, so nearly parallel that they
may
be cut with a
rotary cutter.
It is
evident from Figs. 147 and 148 that the tooth outlines of
any system
till
may be
extended both above and below the pitch line

the longer the teeth the earlier
will
they meet.
It is also clear that
will they
engage with each other, the greater

will
be the arc of
contact, contact.
of teeth
and the greater
be the number of teeth continually in
The
is
distribution of the load over a
number
of pairs
in itself
conducive to smooth running; but on the other
hand, extending the arc of contact away from the pitch point,
increases the sliding between teeth, and also the velocity with
which the teeth approach each other.
The
tooth also
becomes
weaker as
it
is
lengthened, the thickness remaining the same,
and
for these reasons a practical limit is placed
on the length of
is,
teeth.
The
length of tooth adopted in practice
therefore, a
compromise between conflicting conditions, which experience has shown will give good results. The distance along the pitch line from any point on a tooth to
a corresponding point on the next tooth,
pitch
;
is
called the circular
and
will
be noted by
s.
The
by
/,
thickness of the tooth along
the pitch line will be denoted

gears,
Fig. 151.
In the case of cut

/
where no clearance
of têars,
is
allowed between teeth,
-.
In
some form>
such as shown in Fig. 150, where a metal
pinion engages with a gear having
wooden
teeth, the pitch
may
not be equally divided, but the metal tooth
may be
thinner than
368
the
MACHINE DESIGN
wooden
tooth.
If
be the number of teeth and

If the
the
diameter, then evidently
Ns = nD.
number
will
of teeth
be divided by the diameter, the quotient, or

of diameter,
S.
is
the teeth per inch
called the diametral pitch
and
s
be denoted by
.
Since
S =
The
and
-,
S X
71
.'
S = - and
s
71
diametral pitch
is,
ordinarily, the
most convenient
for use,
and in this country practically all interchangeable systems are based upon the diametral pitch. Thus a gear 24" in diameter and 3 diametral pitch would have 24 X 3 = 72 teeth,
and the
circular pitch
would be
1.05 inches.
In the sys-
tem
of teeth
is
adopted by the Brown
&
Sharpe Mfg. Co., and
which
Let
" " "
used very extensively in America, the following proSee Fig. 151.

the outside diameter of the gear.
portions are given for cut teeth.
" " "
= D = D = S = = 5 a = = c
Z>]
2
the pitch diameter of the gear. the diameter of a circle through bottom of space.
the diametral pitch. the circular pitch.
the
addendum =
when
height of tooth above pitch line.
the clearance between top of tooth
and bottom
of
space
gears are in mesh.

total
"
= = = =
the
dedendum, or
depth of space below pitch
line.
"
the thickness of tooth on pitch line
width of
space on pitch line in cut teeth.
"
N
h
the
number
of teeth in gear.
"
the total height of tooth.
Then
= DS =
5
2
7T
71
5
20
10
TOOTHED GEARING
I
369
d h
x
= s = a + c = 2a +
B =
^--^ and
D =D2
2 (a
c)
In the case of rough gear teeth, cast from a wooden pattern, the
thickness of the tooth must be less than the width of the space,*
and the clearance

in cut teeth. If
at the
bottom
of the space
must be greater than

For
of
the gears are machine-moulded, the difference
need not be quite so great as in pattern-moulded gears.

pattern-moulded gears good practice gives
gears, to 0.47 5 for small gears,
/
0.45 5 for large
and the corresponding width
the space
/
0.465 to 0.48
5.
would be 0.55 s to 0.53 s. For machine-moulded gears 5 and the corresponding space would be 0.54 5
to 0.52
Table
in
XXVI gives
of
dimensions of gear teeth for cut spur gears,
accordance with the standards of the

171.
Brown
is
&
Sharpe Mfg. Co.
Methods
Making Gear Teeth.

solid
Metallic gear wheels

cast or forged solid,
are either cast
from a pattern, or the rim
and the teeth are cut from the

cating cutters.
metal by rotary or reciproit is
Where
the gear teeth are cast,
very important
that the pattern itself be very accurately

greatest care in moulding,
it is
made;
for even with the
impossible to obtain true spacing,
on account of shrinkage and displacement due to " rapping" the For this reason, and on account of the pattern in the sand.
difficulty of obtaining
smooth
surfaces, greater clearance
must be
allowed in cast gears than in cut gears, as already noted.
Wooden
patterns are very unreliable for such work, on account of their
tendency to warp and shrink, and permanent patterns should be
made
of metal.
If the
pattern for a spur gear
is
withdrawn from
the sand with a
movement
parallel to the length of the tooth, the
tooth pattern must have draft, or be slightly tapering to facilitate
drawing, and consequently the cast tooth must also be tapering.
Care should be taken
in
assembling such gears, that the tapers
in
i-
* The difference between the ommonlv ailed " backlash."

(
thickness of the tooth and the width of the space
24
37
the
MACHINE DESIGN
two gears are reversed
sets of teeth
to avoid
having the thick ends of
both
at
one end.
come together, thus concentrating the pressure Rough cast gears, of the kind described above, are
used only for rough or large work, and not for high speed.
particular defect of spur gears
The
due
to draft does not exist in bevel
gearing.
In gear-moulding machines the pattern consists of a segment

of
the
gear pattern,
carrying several teeth.
The
pattern
is
TABLE XXVI
PROPORTIONS OF GEAR TEETH
Diametral
Pitch.
Circ :ular Pi tch.
Thick ness of
Addendum
Dep
Space
Pitch
To oth.
t
5=a
*&
+
c
Depth **\'ooth.
2
S
z
I* I* If
2
a* a* 2|
3
i
3i 4
5
1416 5133 0944 795 2 S7o8 3963 2566 1424 0472 8976 7854 6283
5 236
1 1
5708 2566 0472 8976 7854 6981 6283 5712 523 6 4488 3927
3 J 42
0000
1571
2
I
.8000 .6666
5714
.5000 4444 .4000 3636
3333
2857 .2500 .2000 .1666 .1429 .1250
.
6
7
4488
39 2 7 349 3 J 42
2618 2244
1963 J 745 J57 1 1428 130Q 1208 1122 1047 0982
9257 7714 66l2 5785 5*43 4628 4208 3857 33 6 2893 2314 1928
l6 53
I
I I
1571 7257 438i 2326 0785 9587 8628 7844 7190 6163
8
9 IO
ii 12
mi
1446 1286
5393 43 x 4 3595 3082 2696

2397 2i57 1961 1798 1659 i54i !437 1348 1269 1 198
11 35
13 14 15
2856 2618 2417
2244 2094
1963 1848 1745 1653 1571
16
17 l8 19
20
0924 0873 0827 0785
.1000 .0909 0833 .0769 .0714 .0666 .0625 .0588 o555 .0526 .0500
"57
1052
0964 0890 0826 0771 0723 0681 0643 0609 579
1079
mounted on an
"indexed."
axis in such
manner
that
it
can be rotated
accurately through any portion of a complete revolution, or
In forming the mould the segmental pattern

is
is
placed in position and sand

is
rammed around
it.
The
pattern
then withdrawn radially and rotated to the next succeeding
TOOTHED GEARING
position
(the
371
indexing
device insuring accurate spacing), the

till
operation being repeated
the whole circumference

is
is
moulded.
The mould for

this last
the
hub and arms
then completed, in large work

If
being often accomplished by means of cores.

is
machine
at
moulding
well
done the
results are far superior to those obtained
by pattern moulding, and gears may be made that can be run

moderately high speeds.
Obviously, however,
all cast
gears are
much more inaccurate than

Metallic gearing, even
inclined to be very noisy
cut gears, and the latter are preferable
where high speeds and smoothness of action are required.
when accurately cut and aligned, is when run at a peripheral speed of more
if
than 1,200
lash"
feet
per minute, especially
any appreciable " backare unavoidof
exists.
Relieving the points of the teeth, slightly, reduces
the tendency to produce noise.
Where high speeds

is
able the teeth of one of the mating gears
sometimes made
are
wood
or rawhide.
Wheels with wooden

are not as
teeth
known
as
mortise wheels.
They
much
used as formerly, because
modern methods
of gear-cutting produce metallic gears of such
accurate form that they
may be run
in places
where mortise gears

In
were formerly considered
indispensable.
making mortise
is fitted
wheels the wooden teeth are roughed out and the shank
into openings cast in the rim of the wheel, as
shown
in Figs.
149
and
150.
The
teeth are held in place
by the
keys, K, or pins, P,
Fig. 149.
Fig. 150.
as shown.
The
teeth proper are dressed to correct
form with
hand
tools or
by special machines using a
fine circular
saw
for a
cutter.
Usually the large gear, only,

teeth, the pinion
is
made with wooden

This
is its
or " mortise"
rational since
being
made
of metal.
the pinion, on account of the shape of
teeth, is the
weaker of
372
the two,
MACHINE DESIGN
and
also because the teeth of the pinion
come
into contact
more
frequently,
and hence
suffer greater wear.
In such com-
binations,
less
the
metal gear frequently has teeth of thickness
than - and the wooden gear teeth of thickness greater than

to equalize strength.
-,
See Fig. 150.
In recent years gears
made of rawhide have been The blanks for rawhide gears

prepared
tallic
rawhide discs
much used for high speeds. made by cementing specially together under great pressure. Meare
rivets
discs,
on each side of the blank, held together by
passing through the blank, assist the rawhide teeth in retaining

their form.
The
teeth are cut in the blank in the

cut.
same manner
gearing
the
that
metallic
is
teeth are
In
of
using rawhide
pinion
almost always
made
rawhide and the larger gear of
Such a combination may be run at a very high rate of speed, 3,000 feet per minute being a not unusual velocity. Rawhide gears are almost noiseless in operation but care must be
cast iron or brass.
used that they are not subjected to extreme moisture nor run in
too dry an atmosphere.
Formerly
it
was cheaper
to cast gear teeth,
but the development
of gear-cutting
machinery has changed the situation where a large

small teeth are to be made.
number
methods
speeds,
of gears with of
Modern
gear-cutting produce teeth of great accuracy, and
have also so greatly reduced the cost of production that for high
and where smoothness
of action
is
necessary, cut gears
have largely superseded cast gears even in large work.
There are many methods

operation, the most
of cutting gear teeth in practical

of cutting spur gears being
common method
by the use
with the
of a rotating cutter.*
The
outlines of gear teeth vary
number
of teeth in the gear, the pitch or thickness of

is
tooth remaining constant, and, theoretically, a different cutter
required for every different diameter of gear in a series of the
same
pitch.
To meet
same
this
requirement would lead to an excessive

It is
number
of cutters for each pitch.
found in practice, how-
ever, that the
cutter can be used, without serious error, for
* See " Gear-Cutting Machinery," by
Ralph E.Flanders.
TOOTHED GEARING
several sizes of gears of a given pitch.
In the system adopted by
the
Brown & Sharpe Mfg.
Co., only 24 cutters are used for each
pitch in the cycloidal system,
and only 8 cutters
for
each pitch in
the involute system, as given below.

in the first
The
letters
and numbers
column
are the manufacturer's designations, for pur-
poses of ordering cutters.
TABLE XXVII
CUTTERS FOR CYCLOIDAL TEETH
itter
A
B C
" " " " u " " "
D
E F
CUtS 12 teeth. " " *3 " " 14 a " r5 " "
Cutter " "

U " " " " " "
N O
p
16
G H
I
" " " " " " "
it
17 18 19
R
S
Q
T U
cuts " " u U " " " " "

" "
27
to
29
teeth.
20
a " "
3 34 38 43 5 60
75
" " "
33 37 42
" " "
K
"
21 to 22 teeth. " " 23 24 " " 2^ 26
100 150 2 so
49 59 74 " 99 " 149 " 249
or
more
Rack.
TABLE XXVIII
CUTTERS FOR INVOLUTE TEETH
Cutter No.
i
cuts
" "
" " " " "
from 134 teeth

" "
to rack.
4
5
6
7
" " " " "
55 to 134 teeth. " 35 " 54 " 26 " 34 " 21 " 25 " 20 17 " " 16 14 " " 12 " 13
When
the teeth
gear-cutting
is
carefully done, very accurate
work may
all
be accomplished.
It is to
be noted, however, that the form of
when
cut with a set of cutters, as above, are not
theoretically correct;*
and even
itself,
in best practice the error in the
gear-cutting machine
cutters
coupled with that due to dullness of

hardness in the
and deviation due
to different degrees of
metal,
may
be considerable.
172. Forces Acting
on Spur Gears.
In Fig. 151
let
the gear
circle of
let
A
be
drive the gear B.
Let
be the velocity of the pitch

circle of
A;
and
let
V
b
be the velocity of the pitch
B.
Also
&
the equivalent driving force acting at the pitch circle of A, and
W
*
be the equivalent resisting force acting
at the pitch circle

forms
There are gear-cutting machines which,
theoretically, generate correct
of teeth for all gears of a series.
374
of B.
If
MACHINE DESIGN
now
the tooth outlines are properly constructed, the line
of action of the actual driving force
will
always pass through
the pitch point and the angular velocity ratio of
to
will
be
constant.
rolled
The
.
action of the pitch circles will be as though they

their linear velocity will
upon each other and

h
be the same
or
V = V
&
From
the principle of
work
W
The
is,
=W
h b
h b
Therefore
h b
tangential driving force exerted by one gear
upon another
any correct
therefore, independent of the angle of pressure, in
Fig. i^
system of gearing, and the action

if
is,
in this respect, the
same
as
a pair of teeth were continually in action at the pitch point.
The
the shaft
distribution of the reaction at the bearings due to the
pressure between teeth
(W
lf
Fig. 151),
and
its
bending
effect
on
which supports B,
will
depend upon the
relative positions
of the gear
directly proportional
action
CD
and bearings; but the latter will, in any case, be to As the obliquity of the line of is increased, the angle 6 (Fig. 151) is increased and
hence sec
is
also increased.
Therefore, since
1
W, =
1
sec
0, '
TOOTHED GEARING
the pressure
375
on the bearings
is
increased with an increase on the
obliquity of the line of action; but the torque on the driven shaft
remains unchanged.
In cycloidal gearing the obliquity varies from a
the beginning of the contact to zero
in the line of centres;
maximum
it
at
when
the contact point lies increases
and, during the arc of recess,
to a
maximum
0,
at the
end
of contact.
,
The maximum
value of
the angle
teeth.
Fig. 148,
is
about 22
,
with usual forms of cycloidal

equals 1.08, or the
When
equals 22
is
sec
maximum
normal pressure
rotative force.
about 8 per cent greater than the tangential
The
obliquity
is
constant throughout the arc of action in

0,
involute gears, and the angle
Fig. 147,
is
usually
14X
is
or 15
When
15
sec
1.035, or the
normal pressure
3^
per
cent greater than the tangential force.

the influence of friction has
In the above discussion
been neglected.
During the arc

deflects
of
approach the
action of
frictional
in
force
(Fig.
151)
the line of
such a way as to increase the effective obliquity.

of recess
it
During the arc
acts in the opposite direction
and deforce
is
creases the obliquity.
The
for
influence of this frictional

its
small and may, usually, be neglected, but

to a
action accounts,
that
certain
degree,
the
well-known
fact
gears
run
more smoothly during

It is
recess than during approach.
more than one pair of teeth shall be owing to the unavoidable inaccuracy of form and spacing previously noted, it is not safe to depend upon a distribution of the load between two or more teeth of a gear. It
usually intended that
in action at all times, but,
is
safest to provide sufficient strength for carrying the entire load

single- tooth.
on a
In the rougher classes of work, this load
may
be concentrated
and
if
all
one end of the tooth, as indicated in Fig. 152, such gears should be carefully inspected and corrected,
at
intended
to
carry
heavy and important loads.
With
it
well
is
supported bearings, and machine-moulded or cut gears,
not
unreasonable
to
consider the load as fairly well distributed across

if
the face of the gear,
the face does not exceed in width about

(see Fig.
three times the circular pitch
153).
The
obliquity of the line of pressure gives rise to a crushing
376
MACHINE DESIGN
component of the normal which results from the
component,
with
the
action on the teeth (due to the radial
force), in addition to the flexural stress
tangential
component.
This
crushing
ordinary proportions of teeth, does not exceed 10 per cent of the
normal pressure.
flexure,
is
Its effect is to
reduce the tensile stress due to

Since cast iron
this
and
to increase the compressive stress.
far
stronger in compression than in tension,
may be
makes
often
neglected in gears
made
of that metal, while in the case of steel,
or composition gears, the margin of safety assumed usually

it
unnecessary to consider
173. Strength of Spur
this
component.
Gear Teeth.
The assumption
made
that the teeth of spur gears can be considered as rectangular
cantilevers, in determining their strength, is not satisfactory, es-
pecially when treating pinions having a small
number
of teeth. Fig.
at the
154 shows four gear teeth which have the same thickness
Fig. 154
pitch line
and the same

(b)
height.
is
The
tooth
marked
(a) is
one of
an involute rack;
teeth; * (c) is
one of an involute pinion having 12

(d) is
one of a cycloidal gear having 30 teeth;

of
one
of a cycloidal pinion of 12 teeth.
Mr. Wilfred Lewis,

the
to
first to
Wm.
Sellers
& Co., seems to

gearing.
have been
investigate the strength of gear teeth with
due regard
the actual forms used in
modern
His work was
published originally in the Proceedings of the Engineers' Club of

Philadelphia, in January, 1893, an<^ n ^ s
method
of investigation
was
as follows
Accurate drawings of gear teeth were

line of action of the
large scale,
and the
normal
force,
made on a when acting
on the point
*
of a tooth
was drawn
may have
in;
its
see Fig. 155.

teeth
From
the
The
12-tooth involute pinion

it is
weakened by a correction
for interference; but
usually better to correct the points of the mating gear.
TOOTHED GEARING
377
intersection of this line of action with the centre line of the tooth,
a parabola was
locating a
drawn tangent
of
to the sides of the tooth,
thus
beam
uniform strength equal to the effective strength
of the tooth (see Article 15).
The
points of tangency a, a, locate
the weakest section of the tooth,

to this section is
W
e
I.
and the bending moment applied Then from equation J, page 94.
hr
6
y bph
3
bps(y -)
3
orir =
bfs(y z ) =
=
bps(y)
in inches,
(x)
Where
stress,
the breadth of the tooth

s
=
is
the tensile
and
the circular pitch.
The
factor y
a variable,
depending on the shape of the tooth.
Mr.
W1
Lewis
are
found
that
its
value
is
practically
s,
l>
independent of the pitch

proportional
to
(since
h and
|S
1/
the pitch), but depend-
\ 1 //
\ '>
ent
gear.
mainly on the number of teeth in the
Tabulated values of
this coefficient
may
ak.
be found in Kent's " Mechanical Engineers'
r
\
Pocketbook," page 901.

lated "values,
From
these
tabufollow-
Fig. 155-
Mr. Lewis deduced the
ing equations in which
the
number
of teeth in the gear.
For the 15
tooth pinion,
involute system and
the cycloidal system using

of the 12-
a generating circle
whose diameter equals the radius

/
W=bpS
o.684\
[0.124
r ) N
(2)
For the 20 involute system,
(3)
Mr. Lewis' investigations on cycloidal gears were made on a

system using the radius of the 12-tooth pinion as the diameter
of the describing circle.
Modern
practice sometimes
makes
the
radius of the 15-tooth pinion the diameter of the describing
circle,
which gives somewhat weaker teeth than the
first
system.
The
378
difference
is
MACHINE DESIGN
small, however,
compared
to the variation in the

little
assumed
for small
stress, p,
and since cycloidal teeth are now
used
and moderate-sized
is
gears, equation (2) will be
adopted
or p,
in this
work
for standard gears.
The
Lewis' formula
convenient for determining

is
W, b,
s,
where the number of teeth (N)

problem in design
known.
is to is
known; but a very common

(s),
determine the pitch
when
the pitch
is
diameter of the gear
given and the number of the teeth

to this last stated
un-
The formula may be adapted
problem
as follows.*
To
accord with modern practice, circular pitch will
also be transformed to diametral pitch.
Let
"
"
D =
w =
S =
the pitch diameter.

the load per inch of face.
the diametral pitch
N=DXS Therefore W= bsp (.124

Then
sb
or s
- '-^) =
X ^ X
^(.124-^)
(4)
or since
w =
W
=
-f-
and therefore S
*-(. 194
V
J \
.g
_^1^\
p
(6)
...
(6)
'
The
pitch can be found from equation

p,
for
any values of w,
D, and
shafts
when
the face of the gear

is
common problem
and
as
follows:
their velocity ratio is
A is known or assumed. The distance between two known required the pitch of spur
;
gears to connect these shafts for a given load and working stress
on the
teeth.
The
centre distance of the shafts, and the velocity
ratio fix the diameter of the gears.
be governed by the space available, or

designer
The face of the gears may it may be assumed by the
upon other considerations.
To
M.
illustrate;
suppose
* See a discussion by John H. Barr, Trans. A. S.
E., Vol.
XVIII, page 766.
TOOTHED GEARING
379
15,000
lbs.,
is
smaller gear
to
p = 8,000 lbs. per square inch, and that the be 40 inches diameter. Assume also that the
be taken as 6 inches.
6
face of the gear

face
is
may
-f-
The
load per inch of
w =
15,000
2,500
lbs.,
hence,
2,sOo\
/
^ 5 =
;>,ooo/
(.194
2,500 V
\1 ~
,o?8
2.1c;
J 8,000x40
)=
Li
or
say
diametral pitch.
The diagrams shown

equation
(5).
diametral pitch
pitch.
in Figs. 156 and 157 are plotted from That in Fig. 156 covers the range from 12 to 6 and Fig. 157 covers the range from 5 to 1 diametral
The
abscissas (Scale A) represent pitch diameters of gears
in inches,
and the ordinates (Scale B) the load in pounds per inch Any of width offace, for a stress of 6,000 pounds per square inch.
any other
is
other stress could have been taken for plotting the diagrams, and
may
be used in solving problems by them.

to illustrate, let
(5)
curve
6,000.
drawn
for each pitch;
S =
1.5, let
p =
Substituting these values in
W =
or
.389
6,000
V 1.5
2.15 D X 2.2s)
w
3.7,
hence when
D=
w=
o;
D=
10
w =
983;
D =
20,
w =
1,270, etc.
Plotting these corresponding values of
and
as abscissas
and ordinates, respectively, the curve

is
for a diametral pitch
= 1%
drawn through
the points.
If,
The
other curves are constructed

is
in a similar
manner.
then, the diameter of the gear
known,
pounds
the allowable load per inch of face for a stress of 6,000
per square inch
may
be found by passing vertically upward from
the given diameter on scale A, to the curve corresponding to the

pitch,
and then moving horizontally

at the top of the
is
to
the left-hand scale B,

face.
which gives the required load per inch of
Scale
is
re-
produced
marked 6,000
drawn from
diagram, as scale C, and a 45 diagonal the lower right-hand corner of the
3 8o
MACHINE DESIGN
to scale C.
If,
diagram
then, instead of
moving
the
horizontally
from the pitch curve
to scale
B
Scale
on the
left,
movement be
horizontally to the right (or left) to the diagonal
and then
vertically
upward
to the scale C, the
marked 6,ooo same reading will

?
be obtained on
as originally found
on B.
'
TOOTHED GEARING
Furthermore,
if
381
other diagonals be drawn, as shown, from
various points on scale C, they
may
B
be used to read loads per inch
Scale
^N
b*L
\
w ^L
!
*/" 7*
\
Û
f
i
1
/^
7<
/ 4 A /4>
/
-^
s u \vr\^%
\
l\\\
\i
DO
\\\\
I
\ \Y
-^
///
/
/
/
\
/
X/
'
/#
A?
/
>
X
\
!
V /
/
/
/
/ /
,*1/
TO
7
/
A
1\
/
/
7
/
A /
/
4 4
5
V
/
4>
r
> g
0.
cr
/
\
/
/
/I
W /// K
//
K
7^"
/If
/i
j-
/ /
/
^K /
/
'
/
f / p^ II
l
=
-ti^ // /|
,
//JK /
y
/
,r
1
/
/
|a
1 s
/E^// / /
_77z // -?$/-Y/ /
/
/ // r
/
/
j
77
1
/
<-v
&
F^
--
te
/
// IV-/-/
/fe
////,
8
L /////
V/// /// //
/ y
,
/ /
71
Ty
</
^\
/////
'/////
ay ^^
/
}
^r
"
8
Si
oov ^cT
S
Dfaunetnl Pitch
Circular
Tit.
S
-1
CO
t*
$
H
00
M
OS
of face for stresses corresponding to these respective points

this
on
scale,
since
from equation
(4)
it
appears that the
stress
varies directly with the load.
These
stresses are indicated along
382
the several diagonals.
MACHINE DESIGN
Thus
to find the pitch
when
w=
2,500,
8,000,
D =
40,
from 2,500 on
scale (C), Fig. 157, pass verti-
cally
downward
to the diagonal
marked p =
is
8,000; then horizon-
tally to
a point on the vertical rising from
40",
on scale A.
which
The
pitch curve nearest this point

pitch.
for 1 diametral pitch
would be the required

If it is
required to find the load per inch of face for a gear of

stress,
given diameter and pitch, with an assigned
start at the
point on scale (A), corresponding to the diameter; pass

the given pitch curve
;
upward
to
thence horizontally (right or

thence
left
as the case
to scale
may
(C),
be), to the proper stress diagonal;
upward
where the unit load may be read
off.
(See Fig. 156 or 157.)
and unit load are known quantities, on scale A, to the proper pitch curve; thence horizontally to a point under the unit load on scale C, and the stress may be found by interpolating between
If the diameter, pitch,
pass
upward from
the diameter reading
the adjoining diagonals.

It
may
be noticed that the different pitch curves in either Fig.
156 or 157 have a
common
tangent through the origin O.
The
i.e.,
points of tangency correspond to the diameters of gear at which

cycloidal teeth have radial flanks for their respective pitches;
a 12-tooth gear.
The
various curves have not been extended far
beyond
this point as in that case
they intersect, and
mar
the
clearness of the diagram.
Since this intersection occurs after

is
the diameter of the 12-tooth gear
reached,
it is
evident that the
remainder of the curve
is
of
no practical importance.
In
fact,
increase of pitch, or a less
number
of teeth for a given diameter,
beyond
this point will
not give additional strength, because the
length of tooth will be increased, and the flanks will be undercut to an extent which
more than compensates
for the
added
thickness of tooth.
The data may be such

157, will
lie to
that a point corresponding to a, Fig.

all
the left
and above
the pitch curves,
i.e.,
above
same data were substituted This means in equation (4) an imaginary quantity would result. that the unit load taken cannot be carried with the stress and diameter assumed, by any possible pitch. The only recourse for
the
If the
common tangent through O.
TOOTHED GEARING
a gear, of the given diameter and total load
face,
333
W,
is to
increase the
and thus reduce
or to use a material which permits a
higher intensity of stress.

It
should be noted that the teeth of the smaller gear of a mating
pair are
is
weaker
in
form than those of the
larger.
greater
on the
teeth of the smaller gear since
The wear, they come in
also,
conis
tact
more
frequently.
Hence, in general,
if
the small gear
properly designed the larger gear will have sufficient strength.
Fig. 158.
This does not apply

teeth
teeth
to certain
forms of reinforced or " shrouded"
discussed later, nor, necessarily,
where the thickness of

of
and spaces are unequal, nor where the mating gears are
Gear Teeth.
If
different material.
174. Strength of Bevel

teeth, Fig. 158,
a pair of bevel gear

at a, Fig.
have just come into contact as shown

is
147, then the driving force
applied to the point of the driven
tooth by the root of the driver.

will
The
tooth of the driven wheel
be deflected a certain amount, while the deflection of the drivSince the deflection of the driven tooth
ing tooth will be negligible.

is
caused by the rotative
effort of the driving gear, the
magnitude
384
of this deflection at
MACHINE DESIGN
any point on the
line of contact of the
two
teeth will be proportional to the
movement
its
of the corresponding
contact point of the driver or to
distance from the axis of
rotation of the driver; hence, from similar triangles, will also be
proportional to the distance from
its
own
axis of rotation.
all
Now
points,
the cross-sectional outlines of the tooth are similar at
and
it
can be ^hown that in the case of simple cantilevers of

deflection.
It
form the load applied is proportional to has just been shown, however, that the deflection
similar
of the tooth at
any point
is
proportional to the distance of that point from the
axis of rotation.
Hence
the load
also be proportional to the distance
on the tooth at any point must from the axis, being least at
the small end,
and greatest
at the large end, the
mean value
be-
ing at the middle of the tooth.

the
Therefore a spur gear which has
same width of face, and teeth of the same form and pitch mean section, will have, theoretically, the same strength as It can also be shown that in simple cantilevers the bevel gear. of equal breadth and similar outline, the stresses induced at coras the
responding points on
applied
is
the
cantilevers
are equal,
if
the load
the
proportional to the linear dimensions.

stresses are the
Hence
maximum
It is
same
at all sections of the tooth.
evident that the relation established above between the
mean section of a bevel gear and a spur gear with similar teeth, may be (and often is) used as a means of designing bevel gears. It is much more convenient, however, to deal with the teeth at
the outer or large end.
of the
If, also,
the pitch radii are used instead
addendum
radii, the error will
not be great.
Let r 1
"
"
11
"
11
= r2 = r = b = w = w=
2
the pitch radius at the small end of the tooth.

the pitch radius at the large end of the tooth.
the
mean
pitch radius of the tooth.
the width of the face of the tooth along the elements.

the load per inch of face at the radius
r.
.
the load per inch of face at the radius r2
" "
at
W=
w=
e
the resultant load on the tooth
= w
b.
the equivalent load per inch of face which,

r2,
if
acting
a radius
would produce the same
rotative effect as the
actual load.
TOOTHED GEARING
Since the load on the tooth varies as the radii, the total
sul tii nt load will act at a radius
385
re-
= - - 3 ('V >V)
2
(r
3
^-.
and the torsional
moment due
to the resultant force,
W,
will
be
(r2
--''i
Xow
by
definition
2
b r2
(r 9
3
=
r
W
3
)
i?.
Therefore,
w = - w 7=
3
^ (V >YV2
(7)
Also, since the load varies with the radius,

71:
w
r
r2
w
+
rx r2
w
+
>\
r2
2wr r + r
2
(8)
And from
(7)
2 -'
and
'2
"
-.
>'l
(8)
2
=
U>*
t
r2
W -j~i
will
W
(r
3
3 )
; '2
:r
3 (V 2
rx )
r2
;
'l
''I')
(V
will
^-
(o)
The
actual load,
w2
always be greater than

therefore,
in the ratio
shown above.
carrying the
bevel gear,
be more heavily however,

(5)
loaded at the large end than a spur gear of the same diameter, and
same
torque, in the ratio
shown above.
r, is
If,
is
known,
can be computed, and used in equations

Usually
and
(6) of Art.
173 instead of w.
-
made
not less than
-r 2
3
When r,= -r 2
3
1.4,
and
this value
can be used
in
It
computing
unless the face of the gear
is
excessively long.
should be especially noted that in solving problems in bevel
gearing, either
by the diagrams
that
f
of
Figs.
156 and 157 or by
(([nations '^) and (6) of Art. 173, the diameter
D, which must
as the
be substituted therein,
circle,
is
corresponding to the formative

sec
0,
whose radius the tooth is fixed by
is
R =
r2
Fig.
158,
form
.
of
this radius
and not by the radius r2
The
* See also Mr. Lewis's article,

Jan., 1893.
Proceedings Engineers' Club of Philadelphia,
25
386
MACHINE DESIGN
computations should be made for the smaller of the two gears, as

in the case of spur gears.
Example.
the
with a velocity ratio of 3 to 2
Design a pair of bevel gears to transmit 50 H.-P. the gears to be of cast iron, and
;
maximum
fibre stress to
be 4,000 pounds per square inch.
The
revolutions per minute of the shafts are to be 300 and 200
respectively.
Lay
to 2.
off the
that corresponding radii of
Then it is
and OT, Fig. 158, and draw OU so and are in the proportion of 3 found that = 34 and 0' = 56 Assume tentaaxes
OV
tively that r 2 for the large gear
15" and r 2 for the small gear

4".
10",
and take a
trial
width of face of
The
feet
velocity at the radius r 2
7T
IO
300
1,575
per minute.
Hence
the equivalent total load at a radius r 2

1,050,
W
Tir
50
X
2
1.2
33,000
=
k
or
w =
=
i,575
=
1,050
262 pounds.
4
e
Therefore
face.
w = w X
X = =
262
1.4
Since sec 6
sec 34
1.2, the
367 pounds per inch of diameter of the formative

Fig. 157, or
circle is
20
24"
and from the diagram,

that for
from
24", 4,000 pounds, (6), it is found and w = 367 pounds, the diametral pitch is very nearly 3X? which may therefore be selected. This would give 70 teeth for The width of face is a the small gear and 105 for the large gear. little more than three times the circular pitch and is therefore in accordance with good practice. Gear teeth of all kinds are likely 175. Stresses in Gear Teeth.
equation
D=
p =
to
be subjected
to shock, unless
running
at a very
low
velocity,
and the danger from shock increases as the velocity increases; hence the allowable stress must be reduced as the velocity is increased. Reliable experimental data on the allowable stress
in gear teeth are lacking, although
be found in
treatises
on the
subject.
many empirical rules are to The values given by Mr.

The
following
Lewis, in the paper already quoted, are probably as reliable as

any, for teeth that bear along their entire length.
equations have been deduced from Lewis's work:
TOOTHED GEARING
387
\
(600 7?)
;
600
+ V)
(10)
for steel r p
20,000 f-
V600
+ V)
(11)
where
V =
the velocity at pitch line in feet per minute.

stress
It
would
be probably safe to take the

for
bronze as r p
=
is
12,000
(
\6oo
+ V)
(12)
since the resilience of bronze
greater than that of cast iron.

is
An
old empirical rule for rough cast teeth
W
where W, as before,
is
200
(13)
the total load, 5 the circular pitch,
and
the width of face of tooth.
made
cast
of
beech or
The strength of wooden mortise teeth, maple, may be taken as about one-half that of
and the strength
of
iron,
under the same circumstances;
good rawhide gears
may
be taken as equal to that of similar gears be noted that a rawhide gear

will en-
made
of cast iron.
It is to
dure considerable more shock than one
made
of cast iron.
While rough cast teeth are more

only,
likely to bear
on one corner
they
are
stronger
in a
than
cut
teeth of
the
same
pitch,
which compensates
there
is,
measure
for this defect;
furthermore,
usually, an excess of strength, to allow for wear, in all
new
gears,
and the subsequent wear tends
to correct the initial
unequal bearing, along the elements.
On
there
safely
April,
is
account of the increased
liability to shock,
with increase
of speed,
and
also because of the noise of operation at high 'speeds,
a limit to the speed at which any form of gear
may be
which
and conveniently operated.
Mr. A. Fowler,
in Engineering,
at
1889, gives the following as
maximum
values
gearing
may
be successfully operated:
Ft. per
Min.
Ordinary cast-iron gears

Helical cast-iron gears
1,800
2,400
Mortise wheel and
<
ast-iron pinion
2,400 2,400
3, 000
Ordinary
cast-steel gears
Helical cast-steel gears
Special cast-iron machine-cut gears
3,000
388
Although higher
MACHINE DESIGN
velocities are occasionally
found in practice,
these are undoubtedly
maximum
average values and, in general,
more than two-thirds the values given Rawhide gearing, which operates almost noiselessly, may be run satisfactorily up to 3,000
the velocity should not be
above, on account of noise and wear.
feet per minute.
176.
Width
of
Face of Gears.
Equation
the load per inch of face that
may be
of the
(5) of Art. 173 gives applied to a tooth of the
given form and pitch, the total load depending on the width of
face as
shown by equation
with increase
of
(4)
same
article.
The
of
dur-
ability of
a tooth for a given load

face.
is,
therefore,
difficulty
theoretically
increased
The
is
securing
in-
uniform distribution of the load along the contact element

creases, however, as the
width of face
increased,
and
this imposes
a practical limit to the width of the face.

intensity of the load on the tooth
is
On the other hand, if the

too great, excessive wear
may
stress
result.
The
equations given above do not take wear into
account, the allowable load being fixed with reference to the

alone.
On
this basis
a large tooth
may
The
carry a
much
higher load per inch of face, but the wear will be proportionally greater,
the velocity being the same.
empirical rule
given in equation (13) of Art. 175 assigns a load of 200 pounds per inch of face, per inch of circular pitch. For a tooth of 1"
circular pitch this load will give,
of only 2,000
by the Lewis equation, a stress pounds per square inch, for moderate -sized gears. This is a very low stress, for ordinary speeds, so that this rule would give more durable teeth than the Lewis equation, as
ordinarily applied.
Experimental data on the durability of teeth are lacking.

is
It
evident, however, that the allowable load will
depend
largely
on the character of the service, velocity of rubbing, lubrication, and the material used. Thus, for ordinary cut cast-iron teeth under constant service, the value given above (200 lbs.) is probably conservative; while with teeth of high-grade steel
greater loads
of over 2,000
much
may be
carried.
Cases are on record where loads

of face
pounds per inch
were successfully carried,
with a peripheral velocity of over 2,000 feet per minute, the
TOOTHED GEARING
389
pinion being of forged steel and the gear a steel casting, 4.92"
circular pitch.
to
Well-made gears
of
rawhide
may be
loaded up
150 pounds per inch of face, per inch of circular pitch; but in
no case should the load exceed 250 pounds per inch of face.* In the case of machines such as punching-machines which
work
intermittently,
and whose operation extends over a short

is
space of time, the element of wear
not so important in the de-
sign of the teeth; but in such gears as those connecting street-
railway or automobile motors with the driving axles, where the
work
is
both continuous and severe, wearing qualities
may be
fully as
important as strength; and gears
made
of steel or other
hard materials
may have
makes
;
to
be used solely on
this account.
Good
practice
the face of the tooth about three times
the circular pitch
but in fixing the pitch and width of face, in

points
extreme cases,
sidered.
177. Other
to the
the
discussed
above
should
be
con-
Forms
of
Gear Teeth.
Gear teeth made according

dis-
Brown
is
&
Sharpe standard, on which the foregoing
cussion
based, have been found very satisfactory for average
conditions,
and are
in
most common use

it
in this country.
For
extreme conditions, however,
has been found necessary to
reinforce such teeth, or to use teeth of a different form.
very
common way
of reinforcing teeth of cast gears is
by
shrouding, which consists in casting an annular ring of metal on
one or both ends of the

the gear tooth
teeth, as
shown
in Fig. 159.
cast as an integral part of the gear casting,
This ring is and hence strengthens

are shrouded to the
their outline,
by
practically twice the shearing strength of the
cross-section of the tooth,

top.
when both ends

from
The
teeth of the pinion are,
always
is
weaker than those of the gear, and the wear on them

greatest.
also
The shrouding
should, therefore, be put on the pinion;
and
if
carried to the top of the tooth
on both ends
it
will give
them an excess
widths of
in
of strength over those of the gear, with usual

If the gears to
face.
be reinforced do not
differ greatly
diameter, the teeth of both
may be shrouded
the
half
way
up.
Private communication from
New
Process
Raw Hide
Co.
390
Shrouding
is
MACHINE DESIGN
used mostly on rough cast gears, the shroud practi-
cally prohibiting the cutting of the teeth

If the gears are to
by the usual methods.
run in one direction only, and where very

buttress tooth,
is
heavy pressures are to be withstood, a form of tooth as shown in

Fig.
1 60,
and known as a
may
be, but
seldom
is,
employed.
The
driving face, A,
made
of correct theoretical
outline, while the
give the required strength,
back face B may be of any outline* that will and clear the teeth of the mating gear.
form
The
front face should be of standard cycloidal or involute
obliquity of generator than
and the backs are preferably involute forms, with a much greater would be permissible in driving.
For some time past there has been a marked tendency f on the part of the designers of gearing for extremely trying service
to depart
from the Brown
&
Sharpe standard, and
to use teeth
by that standard. In some instances the same angle of pressure has been retained, while in Mr. C. W. Hunt reported others this angle has been increased. to the A. S. M. E. in 1897 (Vol. XVIII) the results of the adoption of such a system and gives full information for their design.
shorter than those given
somewhat
few other manufacturers have adopted similar systems.
The
need of a small gear of great strength, in automobile work, has

increased the
demand for a stronger form of tooth, and it would seem that the old standard must be modified or a second standard adopted for extreme service. The most prominent form of these
so-called
" stub teeth," at present,
is
that
advocated by the
Fellows Gear
Shaper Co.
In
is
this
system an involute tooth
with a pressure angle of 20

c.8 as high as that of the
used, the
addendum being about

Sharpe standard.
Brown
&
gives a tooth nearly twice as strong as the old standard.
This Some-
times these stub teeth are given the height of a standard tooth of
smaller pitch;
thus a 6-pitch stub tooth
may have
the length of
is
a standard 8-pitch tooth, in which case the gear

described as a 6-8 gear.
sometimes
Notwithstanding the fact that the arc

is,
of contact in stub-tooth gears
generally speaking, less than in
* See " Kinematics of Machinery," by John H. Barr, page 131.
f See a paper by R. E. Flanders, Trans. A. other systems.
S.
M.
E., Vol.
XXX,
resume of
TOOTHED GEARING
the old standard, they run well and will undoubtedly be
391
more
used in the future.

178. Strength of
Gear Rims and Arms.
The rim
of the gear
wheel must not only be strong enough to
resist the forces
brought
upon
teeth
it,
but
stiff
enough
also to prevent improper, action of the
due
to springing of the rim.
section of
rim between two
arms may be considered
as a
beam
fixed at the ends

is
and carrying
0,
a load at the middle, the value of which
sin
Fig. 151. 1.25

t,
Good
where
practice
/
makes
the thickness of the
rim
at least
is
the thickness of the tooth on the pitch line.
For small
In
gears this proportion gives ample stiffness, but for very large
gears stiffening ribs are also sometimes necessary.
many
cases the thickness should be sufficient to allow of dovetail-
Shroud
Fig. 159.
Fig. 160.
Fig. 161.
ing a tooth into the rim, in case of accidental breakage of one

or
more
teeth.
Gear wheels are seldom run

induce
at
peripheral
velocities
which
dangerous centrifugal
stresses.
The
principles governing the design of such wheels are discussed,
however, in Chap.
XV.
wheels
The arms
ing that each
of gear
may be
treated as cantilevers, assum-
arm
carries a load
W n
where n
is
the
number
of
arms, and
of either
the tangential load. Computations for strength arms or rims must, however, be considered as giving minimum dimensions, stiffness being the prime requirement, and due regard must be paid to proportions of rim, arms, and hub,
to minimize shrinkage stresses
due
to cooling.
392
MACHINE DESIGN
179. Efficiency of Spur Gearing.
The
experimental data on
the efficiency of spur gearing are very meagre.

available data are those obtained
of
Probably the best

for details
by Mr. Wilfred Lewis,

VII.
of
12
which see Trans. A.
S.
M.
E., Vol.
His investigation
was made with a cut spur pinion

of 39 teeth.
teeth
meshing with a gear
The
circular pitch
was
\%
inches and the face
3^
inches.
tooth,
The
load varied from 430 pounds to 2,500 pounds per
and the peripheral speed ranged from 3 feet to 200 feet The measurements included the friction at the teeth, and the friction at the bearings. The efficiency, as observed, varied from 90 per cent at a velocity of 3 feet per minute
per minute.
to over 98 per cent at
200
feet per
minute.
It
appears that the
friction at the teeth is a small part of the loss with
good cut gears,
the greater portion of the loss being at the journals.
The
effi-
ciency of bevel gears
is
somewhat
less
than that of spur gears,

friction
on account of the axial thrust, which induces

the
between
hub
of the gear
and the
collar at the supporting bearing.
HELICAL OR TWISTED GEARING

180. General Principles.
Suppose a spur gear to be cut into
n small
sections
If
by a
series of planes perpendicular to the axis of
rotation.
each section be then placed a proper distance ahead

(a), it is
is
or behind the adjacent section, Fig. 161
evident that they

just
may be
of
it is
so arranged that
its
some one
section
th
coming into
contact with
mating section when the n
section in advance
in contact at the pitch point.
some
section will always be in contact near the pitch point,
With such an arrangement and
n points of contact with the mating gear between the pitch point and the point which marks
there will always be approximately the beginning of tooth action.
Since the action of gear teeth
is
smoothest when contact

of gearing runs
is
near the pitch point, this arrangement
more quietly and smoothly than ordinary spur gearing, and it was at one time used in machine tool and similar work where smooth action is very desirable. As the number of sections is increased, the total width of
the gear remaining the same, the spacing of these sections being
TOOTHED GEARING
393
kept uniform as before, the form of the stepped tooth approaches

that
the teeth
shown in Fig. 161 (c). When the number becomes infinite become helical in form, and contact is continuous along
is
that portion of the face which

evident,
within the arc of contact.
It is
however,
is
that
since
the relative position of adjacent
laminae
arbitrary,
and may follow any desired law, the outline of

is
the tooth in an axial direction
not necessarily helical, but
may
This
have any desired shape;
although these teeth are most usually

to
cut.
made
form
of
helical,
this
form being more practical
of gearing is also
known
as twisted gearing,
is
for
an ob-
vious reason.
The
action of such gears
identical with that
common
spur gearing, and should not be confused with that
of screw gearing,
also twisted
though certain limiting forms of the
latter are
gears.
screw gear must have regular or unia twisted gear does not necessarily
form helical
have
teeth, while
this limitation.
must be normal which tends to move the gear in an axial direction, causing end thrust on the This can be obviated by making two sets of helical shaft collars.
Since the pressure,
teeth
to the surface, there is a
W, between mating
component, Fig. 161
(c),
teeth
on each
gear,
one right-hand and one left-hand, as shown in

desired to use cut teeth the wheel
is
Fig. 162.
When
it is
some-
made in two parts and fastened together, or the wheel may be made in one piece and the two sets of teeth staggered so as to
times
allow
them
to
be cut; but in both of these constructions there
is
some
loss of strength
due
to the
absence of the reinforcing action
of teeth cast solid as in Fig. 162.
Gears of
this type are also
called herring-bone gears.
With the arrangement shown
in Fig.
162, care
accurate, or end play

of
must be used that the alignment in an axial direction is must be provided so that the middle plane both gears coincide; otherwise the full load will be thrown
181. Strength of Twisted Gears.
If the effective
its
on one-half the gear and the object of the double gear defeated.
load which
(c),
one tooth of a twisted gear transmits to

the length of the tooth be
mate be W, Fig. 161
then the total load normal to the face
is
cosec
6.
If
denoted by
by
b,
then
b cosec
0.
/, and the breadth of the gear Hence the load per inch of face on
394
.
MACHINE DESIGN
,
,
a twisted tooth
WcosccO W, = =
/
TT
or the
b
same
as
in
b cosec
spur gear of face
b.
This would be
strictly true if all points in
the liae of contact were at the

tation as in a spur gear.
of contact
same distance from the

never
so, in
axis of ro-
This
is
twisted gears, the line
always extending; diagonally across the tooth face.
The
and
however, is small, and on the side of safety, assumed that the load per inch of face in twisted gears is the same as that of a spur gear of equal width and equally loaded. This diagonal distribution of the load across
error
it
due
to this,
may
be
the
tooth
to
face,
decreases
the lever
tends
break the tooth; the
arm of the force which amount of decrease depending
Fig. 162.
Fig. 16-
Fig. 164.
on the amount of twist in the tooth. If the twist is so great that when the end in advance is going out of contact the other end is just coming into contact, the line of contact will run diagonally across the tooth from point to flank, and the average arm of the
driving force will be about one-half the height of the tooth.
the twist be
If
made equal
It is clear,
to the pitch, tooth action is continuous
at every point of the arc of action
and
this proportion is the
one
most used.
pitch
is
however, that the assumption often
made
same
that twisted teeth are twice as strong as spur teeth of the
not true for teeth of usual proportions, a difference of 25
per cent being, perhaps, as
much
as can safely be assumed.
On
account of continuous tooth action and consequent smoother

operation in twisted gears, the effect of shock
is
lessened some-
TOOTHED GEARING
what.
395
Twisted gears have been used with success on heavy windrough cast and
ing and hoisting engines, the teeth being often
both gear and pinion half shrouded, making a very strong tooth.
SCREW GEARING
182.
Forms
of
Screw Gears.
When
it is
the axis of
two shafts are
not parallel and do not intersect,

surfaces on which gear teeth
line contact.
possible to lay out contact
may
be constructed which will give
Gears of
this
kind are
known
as skew-bevel gears.
If
They
are difficult to construct,
and are very rarely used.
the load can be carried on point contact, pitch cylinders
may be
described on the axes, Fig. 163, and on these surfaces helical

teeth
may
be constructed which will transmit the desired motion.

spiral * gears, the latter
Such gears are known as screw or

being really a misnomer.
name
While the teeth of such gears resemble
those of helical twisted gears, their theory and action are quite
different;
for,
in addition to the conjugate rolling
action, as in spur gears, there is also a sliding
and sliding component along

action of screw
at
the elements between contact surfaces.
The
gearing
is
very smooth.
The
special case
where the axes are

helical teeth, is
right angles,
and where a large wheel having many
helical teeth
meshes with a small one having a very few

that
an im-
portant one on account of the great reduction in velocity ratio
may thus be obtained. This last arrangement is commonly known as a worm and worm-wheel. Fig. 165 illustrates such a worm and worm-wheel, the teeth on the worm wheel being
truly helical in
form and cut

result is
at
an angle
to suit the
worm
thread
or helix.
The same
sometimes obtained by using a plain
spur gear, and setting the axis of the

with the plane of the gear.f
contact, and on the
line cut
worm
is
at the
proper angle
is
The
contact in these cases
point
worm wheel
tooth
confined to points in a
from the working surface of the tooth by a plane passing
spiral gears, see a "Practical Treatise
methods of laying out and producing so-called on Gearing," by Brown & Sharpe Mfg. Co., and also "Worm and Spiral Gearing," by F. A. Halsey. t A highly successful form of this arrangement is the worm-and-rack drives on planing machines, first used by W'm. Sellers & Co.
* For a full discussion of the
396
MACHINE DESIGN
through the axis of the
worm
at right angles to the axis of the
worm
area.
wheel.
In practice the point of contact becomes a limited

of this
The advantage
is
form
of
worm
wheel, like
all
spur
gears,
that the teeth can be cut with a rotary cutter,
and patterns
for
rough cast teeth are comparatively easy

It is possible,
to construct.
however, to construct a
worm
wheel in such a
Referring
manner
shown
is
as to secure line contact, as in spur gearing.

it
to Fig. 164,
can be seen that when the single-threaded

,
worm
moved
to a
rotated through 360
any median section as
is
forward an amount equal to the pitch of the

position B;
worm wheel
and that rotation
of the
to a translation of these sections
worm, in general, is equivalent backward or forward. The

with
action
is
equivalent to translating a rack of similar proportions,

if
and, in fact,
the
the teeth of the
worm itself is moved axially it will engage worm wheel in the same manner as a rack
does
with a gear.
In the involute system of gear teeth the rack has
straight sides,*
in
and this property is usually taken advantage of making worm gearing, since a worm thread of such a crossis
section
easily
machined.
The
sides of the involute rack face
are at right angles to the line of contact,
aO
b,
Fig. 147,
the inclination of the sides to each other the standard system

.
is 2 0,
Fig. 147,
and hence and in
= 29 If other planes such as 2 N be worm and worm wheel parallel to the median plane X X, Fig. 164, it will cut a trapezoid from the worm somewhat different from that cut by the median plane. The rackpassed through the
like action of these trapezoids
would, however, be similar to those

clear that the shape of the
on the median plane, and

wheel tooth in the plane
rectly
it is
M N may be
so
made
as to
wormmesh cor-
with this new trapezoidal section. It is evident that if enough such sections be taken, a complete tooth outline may be formed that will give line contact with a worm across its full face. It is evident also that any other form cf worm thread may be
similarly treated.
The
preceding
discussion
demonstrates the possibility of
line contact in
screw gearing, and suggests a method by which
* See " Kinematics of Machinery," by John H. Barr, page 125.
TOOTHED GEARING
the teeth of such gearing could be drawn,
397
There
is
no practical value
in
and hence constructed. actually making such drawings; but

process.
teeth having this property of line contact are automatically pro-
duced by what
of tool steel is
is known as the hobbing made of the exact form of
A worm
worm.
wheel
the desired
This
worm
Fig.
1
is
made
it is
into a cutter
is
by cutting
as a hob;
flutes across the face as in
68.
This
known
tempered
used as a milling cutter.
and when hardened and The wheel blank, which

is
has been turned to correspond to the outside of the teeth,
mounted
in a gear cutter, or a special
hobbing machine, and the
Fig. 165.
Fig. 166.
hob
is
also
mounted
in correct relation to the wheel,

little
but with the
axes of the wheels a

final distance.
greater distance apart than the required
toward the
is
The hob is then rotated and at the same time fed worm wheel till the proper distance between the axes reached, thus cutting the teeth in the worm wheel in a very
Sometimes the wheel
is
accurate manner.
caused to rotate simply

results are obtained
if
by the action of the hob, but much better

it
is
driven
positively, with
the proper velocity ratio, from the
cutter spindle
by means
teeth of the wheel are
of positive gearing. In heavy work the roughed out or " gashed" before hobbing.

398
Fig. 166
MACHINE DESIGN
shows a
has been hobbed, and its form of wheel occasionally mating worm. Fig. 167* shows a used where the wheel is sometimes rotated by hand or when the
projecting teeth are undesirable.
worm wheel which
Such wheels may be hobbed,
but are usually cut by the approximate method shown in Fig. 169, where a cutter is fed radially inward toward the axis of the worm
wheel, producing what
is
known
as a drop-cut wheel.
In the
Hindley
worm
the pitch line of the
worm
is
curved to coincide
with the pitch line of the wheel, thus obtaining contact on several
teeth at the
same time.t
183. Velocity Ratio of
Worm
is
Gearing.
per turn of the
worm
thread
called the lead.
The axial advance Thus in Fig. 164
the lead of the single-threaded worm

to the axis,
shown
is
the distance, parallel
from any point on the tooth section A, to a corresponding point on the section B, and is equal to the circumferential If the worm were double-threaded the pitch of the worm wheel. lead would be twice this amount, or equal to the distance between
corresponding points on A and C, and would then be twice the pitch
of the
worm wheel.
The lead
pitch,
be three times the
makes one
revolution, a
worm would and so on. If a single-threaded worm tooth of the worm wheel is moved a
of the triple-threaded
distance equal to the pitch.
In the case of a double-threaded

twice the pitch; and in general
worm
if
the tooth would be
moved
N be the number of teeth in the worm wheel, and n the number angular velocity of worm N = of threads on the worm, then, angular 01 worm-wheel n
.
,
.
vei.
Evidently a very great velocity ratio

tively small
is
possible with a
compara-
worm-wheel.
is
It is to
be
especially noted that the
angular velocity ratio
independent of the diameter of the worm.

wheel, which must be decided upon by
The
pitch of the
worm
a given
consideration of the strength of the teeth, fixes the radius of the
worm wheel for worm may then
number
of teeth;
but the radius of the
be varied to
suit other conditions.
* Figs. 165, 166, and 167 are reproduced from
Browne
&
Sharpe's "Treatise
on Gearing."
t See
"Worm
and
Spiral Gearing,"
by F. A. Halsey.
TOOTHED GEARING
184. Efficiency of
399
general expressions
Worm
Gearing.
The
for the efficiency of screws,
deduced
Since the
in Art.
54 of Chap. VII,
thread
is,
apply also to worm-gearing.
worm
(a) of
usually,
a so-called angular thread, equation 13

strictly
that Article
would
apply.
is
However, the inclination

and
(10) of that article,
is
of the face of
worm
threads
so small that the error introduced in using the simpler

(9)
which were deduced from These equations show that the efficiency of all screw gears is a function of the angle which the thread makes with a plane perpendicular to the axis, and of
equations
the square thread,
small.
=-,d
Fig. 167.
Fig. 168.
Fig. 169.
the coefficient of friction, assuming that the coefficient of friction at the thrust collar is the
same
as at the tooth.
to this subject is the
One
of the
most valuable contributions
experimental work of Mr. Wilfred Lewis.*

Fig. 170 have
The
the
full lines in
been plotted from the diagram on which he has

results.
summarized
also
of
his
They show
clearly
increase
of
efficiency with increase of
thread angle at
all velocities.
They
(2) of
show a remarkable agreement with the
theoretical equations
Art. 54.
The
its
dotted curve
is
reproduced from curve

is
Fig. 52,
and
close
agreement with Mr. Lewis' curves

for
to
be
noted.
This dotted curve was plotted

* Trans. A. S.
a value of
//
0.05.
M.
E., Vol. VII,
page 297.

MACHINE DESIGN
of
this coefficient for
400
Mr. Lewis' calculated average value

velocity of 20 feet per
minute
(5)
is
0.059 an(^ ^ or IO ^ eet P er
m hiute
0.074.
Curves
(4)
and
in Fig.
52 may, therefore, be taken
as supplementary to those in Fig. 170, and may be used, as they were intended, for designing slow-moving and poorly lubricated A theoretical curve plotted from equation (9), Art. 54, screws.
with a value of
\i
0.014
(which would be obtained only at

1,
high speeds), will coincide very closely with curve

aoo
onft
Fig. 170.
_3tt|5o
90
^-~~ 1^--- '

/-
j-- -
M'^tZr"
'
'
"
-7^-^^^-----^^---4W^^-J---'-^^ =:S5 i-Z .^

J1
'
^
<s^
^^
^^
'
***~'z5r \c\ ~~z~-"~'~'
-/.
->
'
^
.^
Jtf
^.^
-
^.""'^" "**"
"''
**'*'
^"^Cg=
~\j
-'"'*
""'^
- *~"
-"~~~
-70
,7 ^ y
-^ "\T-
^ ^ ^"i^/
^^ 7
J!
I
60
4r
50
^.^ Z ^^ Z ^^ ^ ^ / -2 v ^ ,z
^^
^^ ^ ^"^
s7
-''^-^^
M-/ ?
6
y
^ 3 /
# ^Z J ? y
40
- f
30 10 15
20
25
30
35
40
45
Lead Angle in Degrees

Fig. 170.
This coincidence is closer than might be expected from the nature of the problem and the assumptions on which equation
(9) is
based.
Mr. Lewis' value of
p.
for these velocities *
(200
feet per
minute) ranged from 0.026 to 0.015, his average value
being 0.02.
Mr. Halsey \ has examined the design

* Velocity here
of a
number
of success*
between worm
means velocity
of rubbing at the point of contact
and worm wheel.

f See
"Worm and
Spiral Gearing," page 38.
TOOTHED GEARING
ful
401
and unsuccessful worms used
for
transmitting power and
found that every worm among those examined whose lead angle was greater than 12 30' was successful, and every worm whose lead angle was less than o was unsuccessful, and quotes Mr. James Christie, who has had considerable experience with this form of
gearing, as giving
1
15' as the
lower limit for successful design,

It is
which
still
further corroborates the general theory given.
to be noted,
on the other hand, that there

,
is little to
be gained in
using a pitch angle above 30 not to be understood that
the increase in efficiency being

is
very small, while the side thrust on the wheel

is
it is
increased.
It
never proper to design a

for there are
worm
low,
with a lead angle less
than 9
many
cases, not
is
primarily for power transmission, and where the velocity

in
which worms of
less pitch are not only effective but neces-
sary.
oil,
In Mr. Lewis' experiments the

the end of the
worms ran
shaft
in a
bath of
and the
efficiencies given include journal friction, the thrust

at
being taken
worm
by a loose brass
steel
washer running between two hardened and ground

(see Art.
washers
104).
effect
The
and
of
the velocity of rubbing on the
coefficient of
friction of imperfectly lubricated surfaces,
was noted
in Art. 32,
Fig. 17 of that article indicates, in a general

all
way, what
may
per
be expected with sliding surfaces:
experimental results going to

at
show
that the lowest coefficient
was obtained
about 200
fixes
feet
minute.
Mr. Lewis, as the

in perfect
result of his
work,
200 feet per

gearing,
minute as the point of
maximum
efficiency of
worm
which
is
accord with the general theory of lubrication.

gearing, although running in an oil bath,
The
surfaces of
worm
must, from the nature of the contact, be classified as imperfectly

lubricated surfaces.
limit,
An
increase of velocity may,

it is
up
to a certain
decrease the coefficient of friction, but
not possible at
any speed, with the small amount of surface contact obtainable

in
screw gearing, to create a true
oil
film so that the load
would
It
be fluid-borne (Art. 33). 185. Limiting Pressures and Velocities in
Worm
is
Gearing.
was
stated in the last two articles that the best results are obtained
from
worm
26
gearing
when
the rubbing velocity
about 200
feet
402
MACHINE DESIGN
per minute and the lead angle not less than 12
30'.
It is
not
always possible, however, to keep the design within these
limits.
it
Thus
in order to obtain mechanical advantage (see Art. 64)
may
be necessary
kinematic
worm requirements may

to use a
with a very small lead angle, and

necessitate a
much
higher velocity
than 200 feet
at the pitch line.
The
allowable axial load that
may be
applied to a
worm under
experi-
varying velocities has not been very accurately determined, the
law undoubtedly being complex

varies,
(see Art. 32).
Enough
mental work has been done, however,
to
show
that the pressure
approximately, inversely with the velocity;
or the law
may be
roughly expressed as
W V = K, where W = the axial load

worm
on the worm,
V =
the velocity of rubbing in feet per minute, and
K=
98)
.
a constant to be determined by experiment (see also Art.
In Lewis' experiments, made on cast-iron worms and

oil
was found that the limiting value of K, i.e., where cutting began, was about 1,500,000. Smith and Marx* quote corresponding pressures and velocities, attributed to Stribeck, obtained with hardened steel worm and bronze worm wheel running in an oil bath, which give an average allowBach and Roser experimenting with able value of 690,000 for K.
wheels, running in an
bath,
it
;
soft-steel
worms and bronze worm
wheels, succeeded in carrying
a pressure of 800 pounds at a velocity of 1,700 feet per minute,
which gives
1,360,000.
It
would appear,
therefore, that for
average conditions and bath lubrication of the

safe, for velocities
worm
it
will
be
up
to 1,500 feet per
minute, to take
. . . .
WV
The above
shown
that
if
750,000
(14)
discussion has reference to
worms
as ordinarily con-
structed with straight-sided threads.
Mr. Robert Bruce f has

are
the sides of the

carried.
worm
made
concave &
of this
much
form
due,
greater load
may be
With improved threads

6,720,000.
he has succeeded in carrying 25 tons at a velocity of 120 feet per

minute, corresponding to
K=
This great gain
is
* " Machine Design" page 301.

t Proceedings of Institution of the year 1906.
Mechanical Engineers
(British),
page 57 of
TOOTHED GEARING
403
without doubt, to the improved lubrication obtained by what

practically
amounts
to surface contact,
between the mating convex
and concave surfaces of the teeth. 186. Design of Worm Gearing.

the
In general, the strength of
worm
exceeds the strength of the teeth in the worm-wheel;

is is
and where the worm

usual case, the wear
is
made
of a harder material, which
is
the
It
greatest
on the worm-wheel
teeth.
usually sufficient, therefore, to design the wheel teeth alone,
considering
them
as simple spur gear teeth as in Art. 173.
In
the case of rough-cast,

that the entire load
is
or drop-cut teeth,
gearing
it
is safe to
two, or even three,
carried by a single assume that the load is distributed between teeth, depending on the number of teeth in the
must be assumed tooth; but in hobbed

it
wheel.
Example.
to be 20 to
is
Design a
worm
gear to connect two shafts which
are 11 inches apart,

is
1,
and
to transmit
7^
H.-P.
The
velocity ratio
the
worm
shaft
,
angle
not to be less than 15
is to make 320 R.P.M., the lead and the worm wheel is to be cut
with a hob.
The
be
will
solution of problems in
worm
gearing must, generally,

to
1,
tentative.
If the velocity ratio is to
be 20
the
worm-wheel
have
20, 40, or
60 teeth, depending on whether the

It is difficult to
worm
is
single-, double-, or triple-threaded.
obtain a high
lead angle with a single-threaded
worm
without making a very
large thread, therefore a trial assumption will be
triple-threaded
worm, and 60
trial
teeth in the wheel.
made with a Twenty inches

trial pitch
may be
taken as a
diameter for the wheel, and the
circumference will therefore be 63 inches approximately.

circumferential pitch be taken as one inch, the lead of the
If the
worm
thread will be three inches, and can therefore be easily cut in a

lathe.
The
corrected circumference of the wheel will then be
60",
corresponding to a pitch diameter of 19.11".
The
pitch
diameter of the worm, with the given distance between centres,

will
be 2.9"; hence the tangent of the lead angle
=
n
2.9
0.33, or the lead angle is 18
15',
which
is
an
efficient angle.
404
MACHINE DESIGN
of revolutions per
The number
be
320
minute of the
worm
wheel
will
=16.
Hence the
velocity of the
worm wheel
at the pitch
line
=
12
=
worm
will
80
feet per
minute.
The
total
axial
thrust
on the
city of
be
equals
00
3,100 pounds.
The
velo-
rubbing
the length of one turn of the
worm
thread multiplied by the
number
of revolutions per minute, or
V=
71
(cos 18'
2.0X320 X ~ . 15") x
of velocity
=
12
7T
X 2.0X320 =25$ JJ 0.95x12
ft.
per minute. r
The product
3,100
and
axial pressure
is
on the worm
790,000 which by equation (14)

high.
255 X a safe value, although
somewhat
The
load
may be
will
considered as distributed between two teeth,

at the root at least
and each tooth
have a face or length
equal
to the pitch of the
worm
(see Fig.
164), or say 2.75".
Hence
the load per inch of face of
tooth
X2
=
.
560 pounds.
75
From
the diagram, Fig. 158,
it is
seen that this load corresponds

1
to a fibre stress of
about 5,000 pounds per square inch with
inch circular pitch.

it is
equation (10) of Art. 175, however, seen that for the velocity, 80 feet, the allowable stress is 7,000
From
pounds, hence the tooth has an excess of strength to provide

against the wear, which falls heaviest on the
worm
wheel.
From curve
(1), Fig. 170, it is
found that the
efficiency is
about
90 per cent; hence the horse-power which must be supplied to

furnish j.k J H.-P. at the worm-wheel shaft will be
'
8.4H.-P.
0.90
= 277,200
tion of the
foot-pounds per minute, or 866 foot-pounds per revolu-
worm.
The
torque T, which must be applied to the
worm-wheel
shaft, will
be
T =
XXV,
1,650 inch-pounds.
The depth below

circular pitch
is,
the pitch line of a standard tooth of one inch
from Table
0.3857 inches;
therefore the
TOOTHED GEARING
diameter of the worm
at the root of the
405
thread
2.9-
(2
2.13*,
and from equation E, page

1,650
rj-
94, the torsional stress

.
X 0.3857) p =
s
.
16 X 16 T =
-=
850 pounds per square inch, which
is
very
if
low.
the
to
The design may, therefore, be considered satisfactory worm is to be cut integral with the shaft. If, however, it
fitted
is
be bored out and
over the shaft, further calculation as to
the strength of the shaft which

187. Thrust Bearings for
loss in
may be
fitted is necessary.
Worms.
An
important frictional
worm
gearing occurs in the thrust bearing, which therefore
deserves special attention.

applies in this case.
The
general discussion in Art. 104
The
type of bearing
much
used, and of late ball bearings have
shown in Fig. 88 is met with considerable
success in such places.
CHAPTER XV
FLYWHEELS AND PULLEYS
1 88.
Capacity of Flywheels.
There
are two distinct types of
flywheels; namely, those
redistribute energy, as noted in Articles
which
also act
absorb and and 6.2, and those as a pulley or band wheel and transmit power
sole function is to
2, 4,
whose
continuously.
When
a flywheel
is
attached to a train of mechanit
ism in which the supply of energy varies

excess energy,
tends to absorb any
thus having
is
its
velocity
increased.
When
the
work
to
be done
in excess of the energy supply, the wheel tends
to furnish the deficiency at the expense of its kinetic energy,
with a resulting reduction of velocity.

effective
Flywheels, therefore, to be must vary in velocity; the allowable amount of variation depending on the conditions of the case. Thus in engines driving electric generators, the variation from normal speed
may be
limited to one-half of one per cent, or
less,
while in such
machines as punching machines, the variation

as twenty per cent.
If
may
be as great
W be the weight in pounds of a body moving with a velocity

is
of v feet per second, then the kinetic energy in foot-pounds which the
body possesses
Wv
x
where g
,
32.2.
If the velocity
of the
is
body be changed from v to v 2 the change in kinetic energy work which the body will do, or the energy it will absorb, depending on whether its velocity is decreased or increased. If, then, the work to be done or the energy to be absorbed with a given change in velocity is known, the necessary weight of the body may be found; for if K x be the kinetic energy of the body
the
when moving with a

,
velocity v t ,
and
be the kinetic energy at a
velocity v 2 then the energy delivered or absorbed during a
change
of velocity
is
406
407
*)
E = K -K,=
X
Wv? "- Wv? 2
2g
W (v*-v
2
(1)
If the
body
is
rotating around a fixed axis, the velocities of differ-
ent points in the

axis.
body vary
as the distance of these points
from the
For
,o
this case
the kinetic energy of the
body
is
W
2
o>
where
is
the radius of gyration,
and w the angular
velocity.
Hence
for rotating bodies equation (1)
may
be written
E=K -K X 2
p -^
(>*
./)
( 2)
or since
W
o
2
/,
the
moment
t
of inertia * of the body,
E=
In
all
- K, = 1
(<
2
t
22 )
....
(3)
cases of flywheel design the effect of the
neglected,
and
in nearly all cases the effect of -the
hub may be arms is so small
as to be negligible,
and the rim only need be considered.

sufficiently accurate to take the
When
radius
such
is
the case
it is
mean
(2)
of the rim
as the radius of gyration,
and equation
becomes
identical with equation (1) since, in general,
>
v.
In the case
of wheels with
it is
many heavy
arms, or heavy disc wheels, and where

inertia effect of the
desirable to
compute the
wheel
closely,
as in direct driving of electric generators, equation (2) or equation

(3) is
applicable.
In the case of a wheel with arms whose sides

it is
are parallel, or nearly so,
to
be noted that the square of the

2
radius of gyration of the arms or p
is
very nearly equal to
Hence
total
for this case,
if
W be
the weight of the rim
and
%R W the
2
.
weight of the arms
EExample
K - K = j- (W + K WJ o
t
K
to
- 2 2)
(4)
(1).
is
A
to
punching machine
is
make 30
strokes
per minute and

*
solid
punch holes
inertia of
yi" in diameter in steel plate
y
2
"
The
student should distinguish clearly between the
moment
of inertia of a
body and the moment of
an ami.
See Church's "Mechanics," Art.
86,
page 91.
In the case of a circular disc /
WR
2
.
408
thick.
MACHINE DESIGN
Since the machine
may be
used for shearing
also, it
should
be capable of punching a
of
hole, or of
doing the equivalent amount

feet per minute, and,
work
in shearing, at every stroke of the punch, continuously.
The
belt speed is to
be about 600
type,
from
existing
will not
machines of the same

exceed 85 per cent.
it is
known
that the efficiency
It is
required to find the cross-
section of the flywheel rim.
Let Fig. 171 represent the machine under discussion.
The
mechanism in the head, A, is a slotted crosshead;* so that the punch moves with harmonic motion. Let the diagram, Fig.
171
(a),
represent the path of the pin F, in the crosshead.
When
Fig. 171 (a).
Fig. 171.
the pin
is at b
the
punch
enters the plate,
lower face of the plate

the
when
the pin
is at c.
and emerges from the When the pin is at d
punch is at its lowest position, and has entered the die A " at e the punch is withdrawn from the plate and at /is at its highest position. The pin, therefore, moves through an angle of 30 while the work of punching is being performed. The preliminary layout also shows that the diameter of the driving pulley should not exceed 18", and the mean diameter
of
the
flywheel
rim should not exceed
42".
preliminary
estimate also fixes the ratio of the diameter of the pinion

the diameter of the gear
to
as
to 6 ;
hence the driving shaft
will
make 30 X
180 R.P.M.
The
circumference of the driving
* See "Kinematics of Machinery," by John H. Barr, page 184.

pulley
409
belt speed
= -3 =
IoO
600
180
3.34 feet and, therefore, the diam-
eter of the driving pulley will
be 13* which
is
well within the

(see
limit
set.
The machine makes an energy

is
cycle
Art.
4)
every stroke of the punch, or every six revolutions of the driving

shaft.
While the hole
being punched the flywheel
is
giving
up
energy to
assist the belt,
and during the remainder of the cycle

sheet,
the belt withdraws the
punch from the
and
restores the
wheel to normal speed.
The
will
greatest pressure
which the punch must exert on the plate
be at the beginning of the punching operation, and will be
equal to the area of the metal in shear multiplied by the shearing

resistance;
belt
orP =
60,000* = 70,800 -X-X~X 4

2
it
lbs.
If the
had
to exert this effort unaided,
would have
to
be of double
As the punch passes through the plate the shearing resistance decreases, The average until it becomes zero as the punch passes out. pressure may therefore be taken as half the maximum, and the
leather 15 inches wide;
hence the need of a flywheel.
total
work performed
in
punching
70,800
is
X - X
11 =
loss.
1480
is
ft.-lbs.
small and
The work of withdrawing the punch from may be considered as part of the frictional
machine
is
the sheet
Since
the efficiency of the
85 per cent the belt must supply
=
0.85
1740
ft.-lbs.
every energy cycle.
The energy
delivered
by the
belt
belt per cycle is the

1
product of the difference of the belt

Since the speed of the belt
is
tensions (T
2)
multiplied by the distance through which the

is
moves
(see Art. 131).
600
feet
per minute, and the time of the cycle
-^ of a minute, the belt
moves a distance
1740
or
= Afip- = 20 feet T - T = 87 lbs.

x
per cycle.
Hence
of
(7
2)
20
The
effect
centrifugal
(8),
ji
force
Art.
*
may
131,
be neglected at
0.6/ for
this belt speed,
hence equation
gives/ =
an arc of contact of 180

is
and
The
shearing strength of a plate in punching
about equal
to its tensile
strength.
410
0.3,
MACHINE DESIGN
where/ =
the effective pull per inch of width of belt
and
tension per inch of width of belt on the tight side.

of machinery where excessive slipping of the belt
/ is
In
this class
sure to occur,
should be taken at not more than 40
lbs.
per inch of width;

of the belt
whence
7=0.6X40 =
inches.
24
lbs.,
and the width
= 3K 24
moves through 30 or -^ of a revolution, during the operation of punching, and since the punch makes 30 strokes per minute, the time consumed during
Since the pin, Fig. 171
(a),
,
the operation of punching will be
12
= ^ of 30 360
ft.
a minute.
the
The
150
belt,
therefore,
moves 600
operation and supplies 87
1.7
X ^-q = 1.7 = 150 ft.-lbs.,
during
leaving 1750
1600
ft.-lbs. to
be supplied by the flywheel.

or 3
The
driving
shaft
makes 180 R.P.M.,

since
revolutions per
second
and,
therefore,
the
2
mean
7T
radius of the
wheel
=21
inches,
vt
21
33 feet per second,.
The
allowable variation in velocity
may be
taken as 10 per cent,
hence v 2 = 33 X 0.90 30 feet per second. Therefore, neglecting the effect of the hub and arms, the weight of the rim is from
equation
(1)
W = -^ < ~ v? =
2
gE
32.2
2
1,600
33
~ 3 i
lbs.,
545 ^s.
One
cubic inch of cast iron weighs 0.26
hence the number of
cubic inches in the rim will be
0.26
2,100.
The mean
cir-
cumference of the rim

cross-section
42
=
.
132 inches;
.
therefore the
,
2,100
132
=16
may
square inches, or a section 3-H inches
wide by
inches deep
be selected.
quite easy to
In the above example
it is
compute
the
amount
from the conditions of the problem. In the more general case this cannot be done so readily, and methods such as those outlined in Articles 5 and 6 of Chapter
of energy to be redistributed,

II
4II
done
in
must be employed, where the diagram representing work to be is superimposed upon that representing the energy supplied,
such a manner that the excess and deficiency
may
be measured.
Care should also be exercised that the solution covers a complete

energy cycle in order that the solution
excess or deficiency.
may
be based on the greatest
Thus
in Fig. 2 (d) the flywheel redistributes
energy on both strokes, but the

represented by
maximum
Again
in
excess of effort
is
and not by
engines
internal-combustion
v with
such machines as
governors,
"hit-and-miss"
giving a very variable energy supply, the design of the flywheel
may have
to
be based on a hypothetical performance of the
engine covering a
number
of successive strokes, or in other cases
may be based on
perience.
empirical constants which are the result of ex-
In some machinery, such as steam engines,

to limit the variation in velocity to a definite
it
is
desirable
and below the mean velocity. If v be the v 2 the maximum and minimum velocities respectively, then it is sufficiently accurate for most work to take v = ]/ (v^ + v 2 ), 2 but the true relation between these quantities depends on the manner in which the velocity changes. For very exact work, as
t
amount both above mean velocity, v and
in parallel operation of alternating generators, to take this into account.*
it
may
be necessary
Example
(2).
Find the weight of a cast-iron flywheel neces-
sary to limit the speed of the engine discussed in Art. 5 to a total

variation of not
more than
is
0.01,
i.e.,
0.005 above or below the
mean
the
speed.
The wheel
to
also to act as a belt wheel,
and the
sides of
belt speed is to
be about 4,000 feet per minute. be parallel and their

effect is to
The
arms are
be considered.
Since the engine

the wheel
makes 160 R.P.M.,

2$ J
feet.
the circumference of
4,000 = 160
;
Therefore the diameter of the
wheel
8 feet, which
of the
may
be taken without great error as a
mean diameter
will
wheel rim, since the thickness of the rim

face of the wheel should be at least 16
not be great.
* See a paper by
The
I.
J.
Astrom, Trans. A.
S.
M.
E., Vol.
XXII, page
972.
412
MACHINE DESIGN
preliminary
inches wide to give the necessary width of belt.

layout gives 6 arms, each weighing 170 pounds, or
1,020.
&
170
X6=
Let n
the revolutions per second

5,
-- and 60
R = the
Then
>
radius
=
n
4 feet; from Art.
E =
4,620 foot-pounds.
71
=
a>
160 -7
16.8 radians.
60
Now,
problem,
0.995 w
=
2
x
and hence from the conditions

1.005^
of
o>
the
<d
1.005
X
2
I0^8
=
2
16.88
and
I ^-7-
Therefore
(e^ 2
o>
2 2 )
(16. 88
16.
6.1
From Equation
or
Ty
(4),
(W +
&
X^
a)
= &(l*-a,*\
1,020
3
to
2gE
#
2
K
,
-"2sr
2 )
Therefore
JFa
= 2X32.2X4,620 7 7 16x6.1 3 = 2,720 + 1,020 = 3,740 lbs.,

is
2,720. ?/
which must
be added the weight of the hub

wheel.
to obtain the total
weight of the
Since the rim

is to
to
be 16 inches wide and the mean

will be
diameter
be 06 inches the thickness
2 720 0.26x^X96x16
2.2 inches or say 2 X A inches.
189. Practical Coefficients.
It is
evident that the ratio of
the energy to be absorbed, to the total energy supplied per
energy cycle, will vary in different machines, and also in the
same machine under
different conditions.
Thus
in the
punch-
ing machine the flywheel absorbs and redistributes nearly the

entire energy supply per cycle, while in the steam-engine
example
the
amount absorbed
is
about one-third the
total.
It is readily
seen that in the steam engine this ratio will vary with the point
of cut-off,
steam pressure, weight of reciprocating

It is to
parts, etc.,
and
therefore, in general, tabulated values of this ratio are deceptive
unless they refer to specific conditions. the weight of the flywheel

is
be noted that
directly proportional to the energy to
be absorbed and inversely proportional to (v*
v 2 2 ).
The latter

is
413
is
usually a small quantity and, therefore,
if
large the weight
of the flywheel
of the cost,
may
be excessive,
which
is
undesirable because
and
also because
heavy wheels bring great loads on

losses.
the bearings, causing
frictional
For
this
reason
it
is
always desirable so
be absorbed.
to
arrange the sequence of events in the energy
supply and work to be done as to minimize the excess energy to
This
is
illustrated in Art. 6, Fig. 5,

(or increased)
K may be greatly decreased

positions of the crank pins.
where the area by changing the relative

is
This procedure
closely restricted.
is
of great import-
ance to avoid wheels of great weight in large steam engines when

variation in velocity
must be
The
allowable variation in velocity
fixed with reference
evident that some work require much more constant velocity than others, and experience has shown what the limits in variation of
to the character of the
work
to
be done.
It is
classes of
velocity
may be
for successful operation.
The
2
following limiting
represent average
values of the proportionate variation

practice.
The
particular case of direct driving of alternating
generators in parallel must, in general, be treated with reference

to
the allowable variation per pole,
and when,
therefore,
the
number
of poles is great the total allowable variation is corre-
spondingly small.
TABLE XXIX
Values of
v
0.10
to
For For For For
punching machines and similar machines

engines driving stamps, crushers, etc engines driving pumps, saw mills engines driving machine tools, weaving and paper
mills
0.15
o .05
0.20
0.03
to
For engines driving spinning mills for coarse thread .... For engines driving spinning mills for fine thread For engines driving single dynamos
For engines driving alternators in parallel
0.025 0.016
0.01
o .007
to to
0.03
0.025
to
0.02
o .003 to o .0003
190. Stresses in Flywheels.
The
velocity of the rims of
all
flywheels
If the
is,
from the nature of their requirements, very high.

is
wheel
to act as a
band wheel, the
desirability of obtaining
414
MACHINE DESIGN
high belt speed (Art. 134) brings the peripheral velocity up to It has been shown that the 4,000 or 5,000 feet per minute.
capacity of a given wheel
velocity and, therefore,
alone,
is
proportional to the square of

the wheel
is
its
when
to act as a flywheel
economy
in the use of material, or the limiting of the
external dimensions,
makes high speed very
desirable.
Great
care should be used in the design of such wheels, for a flywheel
which breaks
limb,
at normal speed is exceedingly dangerous to life and and when such wheels "explode" or break from overspeed-
ing, the results are usually very disastrous.
Unfortunately, mathematical analysis of the stresses in
fly-
wheels and pulleys
is
not satisfactory or conclusive.
In the case
stresses
of small wheels cast in one piece,

of great
unknown shrinkage
magnitude .may
exist,
which render useless any refined

elastic theory of
calculations.
In large wheels built up of sections, the presence
of joints vitiates
any calculations based on the
and when the parts are of cast materials and of large sectional area, there is no assurance that the character of the material is uniform throughout. It is importhe strength of materials;
tant,
however, to understand fully the general character of the
stresses even
their
when no accurate computations can be made

is
as to
magnitude.
free to
Consider that the rim of the wheel in Fig. 172

radially, the
tion.
If the
is
expand
arms exerting no restraining force

wheel be rotated on
its
in a radial direc-
axis the action of centrifugal
force
such as to cause an outward pressure on every part of the
rim, in exactly the
same manner
(see Art. 78)
as in a boiler shell acted

;
on by
an internal pressure
tensile stress
the rim expanding until the
induced in any section
A, Fig. 172, balances the

If, on the hub and rim,
tendency of the wheel to separate along that section.

other hand, the arms are rigidly attached to both
and are so
negligible,
inelastic that their stretch,
under the action of the

of the rim,
is
centrifugal pull
it
due
to their
is
clear that
own mass and that they may be placed so

and practically no
close together
that the rim cannot exuand,
stress will exist
in the rim, the centrifugal action being balanced
by the
stress in
the arms.
415
Flywheels approximating both of these conditions are sometimes built, but in the most usual case the arms stretch a certain
amount and are not placed

sults similar to that
close together, so that a condition re-
shown, in an exaggerated manner, in Fig.

stretching
173.
Here,
the
arms, though
somewhat, do not
and, therefore,
stretch
enough
to allow the
is
rim to expand
less
freely,
The hoop tension section of rim between each pair of arms is so long that it becomes a beam fixed at the ends and loaded uniformly by the unbalanced centrifugal action; the greatest bending moment being at the arm,
the
somewhat
than that in the free ring.
and a bending moment
of half the
maximum occurring at
the centre
pîg.
172.
Fig. 173.
of the span.
The maximum
tensile stress will
be the
sum
of the
hoop tension and the
tensile stress
due
to the
bending action.
stress will,
The
If
relative values of the
hoop tension and bending
evidently,
depend upon the amount which the arms
stretch.
they should stretch enough owing to their
own
centrifugal force,
will occur;
so that the rim
expands
will
freely,
no bending action
while
if
they are so inelastic as completely to restrain the rim,
no hoop tension
will
be induced, but the
full
centrifugal force
be applied
to
of stretch of the
bend the rim. With any intermediate amount arms the rim will be held in equilibrium, partly
by the hoop ten-ion and partly by the restraining action of the
MACHINE DESIGN
41
arms, the latter being a measure of the unbalanced centrifugal

force of the rim,
and
of the
is
bending
stress
caused thereby. Since
the expansion of the rim

the
directly proportional to the stretch of
arms
to
it is
clear that the

If,
hoop tension
for instance,
is
also directly propor-
tional
the stretch.
the
arms
stretch one
quarter the amount necessary for free expansion, the hoop tension
will
be one-quarter that due to free expansion, and the bending

be proportional to three-quarters of
the centrifugal
exists
stress will
force of the rim.
The mathematical relation which
between
these stresses
size
is
complex, and will of course vary with the relative
and shape of the rim and the arms. If the rim is of a wide thin section, and the arms are few, the bending stress may be very Professor Lanza* has shown that, with the proportions serious.
ordinarily used,
the arm, theoretically,
stretches
about threeIt is also to
quarters the
amount necessary
if
for free expansion.
be noted that
the wheel
is
to act as a
wide thin rim, the bending action

still
band wheel, and has a on the arms as at B, Fig. 173,
further distorts the rim

side.
and increases the bending on the
forward
D = R = Let / = Let v Let w = = Let

Let Let
I
the
the
mean diameter of the rim in feet. mean radius of the rim in feet.
the thickness of the rim in inches. the velocity of the rim in feet per second.
the weight of the material per cubic inch.

the length of the rim between
arms
in inches.
Consider a section of the rim one inch wide on the face.

centrifugal force per unit of length
this section is c
(1"),
The
circumferentially, of
w ti? = Rg
and, therefore, by Art. 78, the total load '
which tends
to separate
such a ring along a diameter

if
is
wtv
X
is
2D, and the unit
stress in the section,
no bending
exists,
therefore,
=
x
12
ztRg
=;
D =
12
wv
g
. ,
10'
nearly, for iron wheels y

'
k) '
* Trans. A. S.
M.
E., Vol.
XVI, page
208.

The maximum bending moment
arms,
in
417
the rim occurs at the
and
its
value
is
cl2 *
12
considering the rim as a
straight
beam.
The
stress
due
to this
bending moment when no
hoop
tension exists is therefore,
*where
If
e is
Me ~S7
cPe
If
the
(6)
the distance to the outer fibre, and
moment
of
inertia of the cross-section of the elementary ring.
now
the stretch of the
arms be taken as three-quarters
that necessary for free expansion of the rim, the total unit tension in
the rim will be
#-** + *#, -(^ + ^7)

if
(iv
cPe\
12
(7)
n be the
number
of
arms
in the wheel,
*
nD
,
and
if
the
cross-section of the rim be rectangular

(7)
-5
whence equation
reduces to
For
16
= =
&&,
to
D =
ft.,
and n
6.
Professor
Lanza
finds the stress
due
hoop tension
575 and the stress due to
bending
5,060 or p the total stress

(8)
1,
5,635.
2
For the same

4,600, or a total
data equation
stress
gives j^Pi
581 and
%p =
5,81
which agrees quite
closely.
The above
placed upon
article,
it
equation
may be
used for checking, roughly, the
allowable stress in flywheel rims, but implicit faith must not be

for the reasons given in the first
paragraph of
similar
this
and
all
results obtained
from
this or
formula?
should be checked by successful practice wherever a doubt arises.
The
equation does, however, show clearly that in wheels having

I,
* Sec Table
case 17.
I is
tit should be noted that this

the entire cross-section.
for a
unit (1") of width of rim
and not
for
27
41
MACHINE DESIGN
is
thin rims, or few arms, the bending stress
much
greater than
ac-
that
due
is
to
hoop
tension,
and care should be exercised

at high speed.
cordingly
(5)
when such wheels must run
Equation
often taken as a basis for the design of flywheels, using,

If
therewith, a large factor of safety to cover uncertainties.

in equation (5)
be taken as 1,000
(a factor of safety
is
20),
then
6,000 feet per minute, and this
found
to
be a safe peripheral
be noted, however,
speed for ordinary cast iron wheels.

that this speed
so,
is
It is to
safe only because experience has

is
shown
it
to
be
and
not, as will be seen, because the stress
necessarily as
low as 1,000 pounds. Example. Compute the
stress in the
rim of the cast-iron

assuming that the
flywheel discussed in example
(2) of Art. 188,
arms
stretch three-quarters the
amount necessary for free expansion

2.2,
of the rim.
Here n
=
v2
6,
D=
and v
= 60
4,000
66.6.
.-.
from
(8),
( +
2
0.025)
\ J
(8
The
stress, if
+
2
0.025
1,668 lbs. per sq. in.
based on equation
(5) ,
would be 444 pounds per

rest, or
square inch.
When
the
a.
flywheel
is
is
being accelerated from

off,
when
the
energy supply
suddenly cut
called
as
it
may be
in a
steam engine,
arms may be
fixed at
upon
to carry the full torque load.

stiff
Each
car7).
arm of a wheel with a very
rim approximates a cantilever

at the other,
I,
beam
If the
one end, free but guided
and
rying a concentrated load at the free end (see Table
case
rim
is.
thin
and
flexible,
the arms approximate a simple
cantilever loaded at the free end.
In addition, the arm
is
sub-
jected to a tensile stress
due
to the centrifugal action of its

it
own
is
weight, and that part of the rim which

ently
supports, so that appar-
equation
to
(Table VI)
applies.
The
direct
stress
difficult
compute, however, and since the bending

is
stress in
the simple cantilever
twice that of a cantilever with the free end

guided,
it
419
is
considered sufficiently accurate to compute the
arm
as a simple cantilever
Let Let
P = a =
=
and neglect the direct stress. the greatest force due to the belt pull
arms.
at the rim.
the length of the arm.
the number of Then from J (Table VI)
Let n
Pae
^~nf
from which the
determined.
stress p, or the
stress
( 9)
moment
of inertia I,
may be
The
allowed should not exceed 2,000 pounds
per square inch, for cast iron, on account of the uncertainties

of the case,
and a lower value is sometimes desirable. The statement sometimes made that the arms should be as strong against bending stress as the shaft is against torsional stress, is
misleading
as,
in general, shafts are designed for stiffness,
and
not for torsional strength.
The
shaft of a
steam engine may
have
to
be very large to avoid excessive deflection and, as a
consequence,
may have
great excess of torsional strength.
191. Construction of Wheels.

for velocities
Flywheels and band wheels,

minute, are usually
below 5,000
of
feet per
made
of
cast iron
on account
low
cost.
For higher
velocities
steel
castings are used,

plates, or
and
in
extreme cases wheels made of
steel
wire-wound wheels have been constructed.

written v
Equation
unit
tensile
(5)
may be
1.64 a
A w
The
allowable
is,
strength divided by
the weight per cubic unit
therefore, a
measure of the value of the material for

cast-iron wheels
this purpose.
For
this
reason some woods are superior to cast iron for wheel rims, and
which have burst have been successfully
re-
placed with wheels having rims

Difficulties
made
of
wood.*
the diameter of
in transportation
limit
wheels
cast in one piece to about ten feet, cast in
two parts
to
and the diameter of wheels about twenty feet. Wheels from about sixare usually
teen feet in diameter
upward
made
in several sections.
Small flywheels and band wheels are usually cast in one piece, or
* Trans. A.
S.
M.
E., Vol. XIII,
page 618.
420
MACHINE DESIGN
in
made
two parts
for convenience in erecting.

stresses will
In either of the
exist.
latter cases
unknown shrinkage
most probably
These shrinkage stresses are sometimes relieved by casting the hub in several pieces, each piece being cast integral with one or more arms. The openings between the parts are afterward filled with lead, and rings are shrunk upon the hub to hold the parts Experience shows that solid cast-iron wheels, when in place. properly proportioned, are safe up to 6,000 feet per minute which, fortunately, is also about the limit of efficient belt speed. If, however, the
wheel has a very wide thin rim

if
it
cannot be considered
safe at this speed, particularly
balance weights are attached to
the rim between the arms, thus increasing the centrifugal bending
Fig. 174.
Fig. 175.
force.
If joints exist in the rim, their relative
strength must be
considered.
Band wheels
to
of wrought-steel construction
feet in
can
now
be obtained up
strong,
about 4
diameter;
they are light and
and are rapidly coming
into favor.
Where speeds above 6,000 feet per minute are necessary, wheels such as shown in Fig. 174 are sometimes built. Here the rim and hub are of cast iron, each cast in one piece, and the spokes are of steel. The spokes are placed in the mould, and
the metal poured around them, so that on cooling they are gripped
very firmly.
is practically
The
spokes are placed close together so that there
no bending of the rim, and the rim is also prevented from expanding freely. Wheels of this construction are used for large band saws at velocities above 10,000 feet per minute, under heavy service, with perfect success.
42
In Fig. 175 the rim is cast separately in one or more pieces. The arms do not constrain the rim radially, but leave it free to
expand.
The
stresses in the
is
rim when cast in several pieces so

cantilevers.
that shrinkage
only,
ter
not a factor are those due to centrifugal force
and the arms are simple

illustrate
Wheels
of this charac-
have been used with success in rolling-mill work.*
Figs. 174
and 175
lies
wheels which correspond closely to the limiting
types discussed in Art. 190.
The
construction of most wheels
band wheel with arms and hub cast in one piece and the rim in sections. The joints in the rim are simple flange joints, placed midway between This is the most dangerous location possible, on the arms. account of the added bending effect due to the centrifugal force
between these types.
Fig. 176 illustrates a
the
Fig. 176.
Fig. 177.
of the flanges
which add
to the
strength.
The
best location
is at
mass without contributing to the the arm, and many wheels are
well.
built thus, the
arm being
bolted to each segment, and the seg-
ments themselves bolted together as

placed between the arms,
length of the span
it
Where
the joint
is
should be about one-quarter the

the arm, as at A, Fig. 176, where,
away from
by the theory of elasticity, the bending moment is zero. Fig. 177 shows a heavy flywheel with an arm and a segment of the rim cast together. The arms are secured in the hub by means on one side may be movable axially so as more firmly to clamp the arms. The segments are held together at the rim by means of links of rectangular cross-section shrunk in place. This
of fitted bolts.
The hub may be
solid or the flange
construction
is
very
common.
M.
Occasionally links
are
also
* Trans. A. S.
E., Vol.
XX, page
944.
422
MACHINE DESIGN
In Fig. 178
are joined
bolts.
shrunk into recesses on the outer face of the wheel.

the segments
are
called "prisoners,"
at
held together by T-headed links, sometimes
shrunk in place.
are fastened to
The segments
is
the
arms,
which
is
them by through
hub
is
This construction
that in Fig. 177.
simple and the machining
easier than with
flanged connections.
The
construction of the
similar to
Fig. 178.
Fig. 179.
It is evident that the
wheels
is
most important.
Wheels with thin, flanges as shown in
manner of joining segments in built-up Wheels seldom fail at the hub. wide sections are almost always joined by
Fig. 180.
When
such joints are used they
should be well ribbed for
stiffness, as indicated,
and the
bolts
should be placed as near the rim as possible, so that the lever

arm. a shall be as great as possible compared to the
arm
b (see
C=
Hib'
m
Fig. 180.
Fig. 181.
Fig. 182.
Art.
63).
A much
is
better arrangement
is
shown
It
in Fig.
181,
where an arm
placed on each side of the joint.
This
is
par-
ticularly applicable to wheels cast in
two
parts.
may be
noted
that thin rims are often stiffened

trie
by
light circumferential ribs at
outer edges.
Mr. A. K. Mansfield has pointed out (Trans.
A.
S.
M.
E., Vol.
XX)
that these ribs
maybe
a source of weakness.

The
are
423
greatest bending moment is near the arm where these ribs on the tension side of the beam. A rim having such ribs is not necessarily as strong against bending in this direction, as one and when of rectangular cross-section having the same area
;
ribs are used the section
The
prisoner link
modulus should be calculated. shown in Fig. 178 has certain advantages over
It is
the link
shown
in Fig. 177.
is
evident that the depth of the
recesses in Fig. 177
limited, while in Fig. 178 the slot
can extend
entirely across the section
and the link can be made as wide as

it is
the rim
itself.
Furthermore,
possible to
machine both wheel

to
and link
in Fig. 178 accurately,
which
is difficult
do with the
construction in Fig. 177.
This permits of greater accuracy in

induced in the link by shrinking
77.
it
computing the
in place, the
initial stress
importance of which has been noted in Art.
If
the rim be
made
I-shaped,* as in Fig. 182, the links can be so pro-
portioned that the joint will be as strong as the rim proper or
even stronger.
While,
evidently,
the
relative
strength
of
the
joint
com-
pared to the solid rim will vary with the exact proportions selected,
average practice gives the following apparent values:
Flanged
Flanged
joint, bolted,
midway between arms

end of arms
25
joint, bolted, at
Linked Linked
joint as in Fig. 177
50 60
65
1 1
joint as in Fig. 178 Fig. 182
Linked joint as in
Solid rim
.00
.
00
It
must
is
not
be
inferred
from
as,
the
above
that
a
thin
solid
rim
with
necessarily
the
best;
obviously,
a wide
rim
unknown shrinkage
strain
may
not be as safe as a narrow

if
deep rim of the same sectional area

joint.
held together
by a good
or
For extreme
wheels with rims
of steel castings,
latter type
velocities,
wheels built up of
steel plates,
Fig. 179 shows a flywheel of the used in rolling-mill work (see Power, Feb. 4, 1908).
made of are now
plates fastened to a central spider
made
used.
The rim
is
made
of laminations held to the spider
by
dovetails,
* See Trans.
S.
M.
E., Vol.
XX, page
944.
424
as
MACHINE DESIGN
shown, the laminations being assembled with overlapping
joints.
Heavy
outside plates
clamp the whole structure together

In the particular case noted above, 250
feet
by means
of a
of through bolts.
the velocity of the wheel rim
is
per second.
to
Descriptions
number
of
examples of such wheels are

S.
be found in the
Transactions of the A.
M.
E., the
magazine Power, and other
periodicals. Wheels for great speed have also been constructed by winding the rim with many turns of steel wire. The rotors of some forms of electric generators, steam turbine rotors, and similar revolving members are often loaded as shown at W, Fig. 176. Such loads add to the centrifugal force acting on the rim, but do not add to the strength of the rim. Due al-
lowance should be made in such cases; particularly

or loads are placed near a joint as
if
the load
shown
and
in Fig.
if
176.
is
The
large,
teeth of gear wheels constitute such a load,
the wheel
and the peripheral speed high,

especially
this
should be considered.
Balance
weights, placed between the arms, should be carefully considered,
when
the rim
is
thin
and the
velocity high.
192. Experiments
on Flywheels.
The
best experimental data
upon the strength
of
flywheels are from tests conducted

S.
by
Professor Benjamin and reported to the A.

these experiments were
M. E.*
While
made
to
determine the bursting speed of
small cast-iron wheels only, and throw no light on the increase of

stress
with an increase of speed, they are very valuable as indicat-
ing the
manner
in
which various types of wheels
fail.
Being
conducted on small wheels, due allowance must be made for the

difference in quality between the metal of small
in estimating probable bursting stresses.
to
and large castings
These experiments go show that solid cast wheels will burst at a peripheral velocity somewhere near 400 feet per second, and such wheels are safe only Rim joints at a velocity of not more than 100 feet per second. flange joints, common the arms, particularly the midway between
were found to reduce the strength materially.

of various joints
The
strength
was found
to
be about as tabulated in Art. 191.

See also "Machine Design," by C. H.
*See Trans.,
Benjamin.
Vols.
XX and
XXIII.
425
Extra loads, such as balance weights located between the arms,
were found
effect.
to
be very dangerous, on account of the added bending

If the radial
193. Rotating Discs.
depth of a wheel rim be

equations deduced in the
great
compared
to its axial width, the
preceding articles do not apply, the difference being analogous

to that existing
between thick and thin cylinders.
Mathematical
with those
analysis of the stresses in a rotating disc, in

existing in thick cylinders
common
under internal pressure, are complicated

Experimental data, corroborat-
and not altogether

of these stresses is
satisfactory.
ing the theories, are also lacking.
full
mathematical treatment
beyond the scope
of this treatise,
and only
enough
will
be inserted to show the general character of the

rapidly rotated on
problem.
When
a radial
a disc of uniform thickness
is
its
axis, the principal stresses

stress, at
induced are a tangential tension, and
every point in the disc.
Let
"
"
" "
"
"
"
= the outer radius of the disc in inches. = the inner radius of the disc in inches. r = the radius at any point. = Poisson's ratio = ]4, for steel and Y for cast * N = revolutions per minute. w weight of one cubic inch of the material. p = the tangential stress at any radius r. p = the radial stress at any radius r.
r2
r,
'
s
iron.
Then
it
can be shown* that for a
flat
disc of
uniform thickness,
having a hole at the centre of radius r lt

/>
= o.ooooo 3 5 5 T^Y [(3 + A)(r

2
Hr
2
1
+ ^)-(i+3/-)r](ii)
2
1
and
p'
= 0.00000355
solid
vN
[(3
+ /) (r + r
2 2
-^-L -r ) J
2
(12)
For a
disc
0.000003 55
p'
wN
2
[
(3 4-
/)
r2 (3
(1 X)
(
+
r2
2
;-)
'
2
l
(J J
3)
4)
and
* Sec
= 0.00000355
wN2
y2
)
1
"Theory
of the
Steam Turbine," by A. Jude, pages 192 and 204.
The
notation and units have been changed to correspond with those used in this text.
426
It is to
MACHINE DESIGN
be noted that the radial
stress is less at
any point than

of
at
the
corresponding tangential
stress;
and an examination
is
equation (n) shows that this tangential stress
maximum
and a minimum At the surface of the bore or where r =

the surface of the bore
at the outer periphery.

r^
the stress
2 )
0.00000355
rx
oj
2
[
(3
A)
2 r
+
2
(1
3
it
A)
2
]
If
now
be taken so small that r

is
is
negligible,
appears
that the tangential stress
0.00000355 w
2
[
(3
A)
2
\
which
(13).
is
just twice that obtained

effect of
is,
by making
r '
in equation
The
even a very small hole at the centre of a
rotating disc
therefore, to increase the stresses greatly.
Example.
circular steel
saw
inch in thickness and 80
inches in diameter has a hole 4 inches in diameter in the centre
stress at
and runs at the rim and

Here
rate of 500
R.P.M.
A
Determine the tangential

r2
also at the hole.
N=
in (11)
500,
w=
r
0.28,
= %,
40,
and
rx
2.
Whence
rim
is
making
r2
=
500
40 the tangential stress at the
0.00000355
(1
X +
0.28
1)
X
2
]
2
[
(3
+ K)
(40
2 )
40
=
=
535
2.
2
[
lbs.
per sq.
in.
and
p
at the hole
making
rx
0.00000355
(1
0.28
2
]
500
(3
-
+ K)
SCL-
(4o
40
2 )
1) 2
2,643 lb s
P er
The
foregoing equations, (11) to (14), are deduced on the

is
hypothesis that the material

It is clear that
perfectly elastic
and homogeneous.
they cannot be intelligently applied to built-up
wheels of the disc type, and must also be applied with caution to
brittle materials.
They
are of great value, however, in showing
the general character of these stresses and the location of the

greatest stress, thus indicating the shape
for
which discs should have
uniform strength; for a
show
that such
discs
must be thickened
at the centre to
reduce the stress at that

point.
427
shapes, reference
turbine.
It is
For complete mathematical analysis of discs of different may be made to the various works on the steam
evident that great care should be used in selecting
for high-speed discs.
and working the material

are not
Rolled sheets
good
for
is
very high speeds on account of their
seamy
Discs
structure,
terials of
which
conducive to incipient cracks, and cast ma-
brittle structure
must be of
first-class quality.
forged
down from much
thicker ingots give the safest construction.
References
of the Steam Turbine," by A. Jude. "Steam Turbines," by L. French. "The Steam Turbine," by Dr. A. Stodola.
"The Theory
CHAPTER XVI
MACHINE FRAMES AND ATTACHMENTS
194. Stresses
in
Machine Frames.
Since
machine frames
must, in general, receive the reactions from the forces applied

to the various
moving members by the energy transmitted,
it is
obvious that the stresses induced in frame

cases, very
members
are, in
most
complex and beyond mathematical
analysis.
If it is
essential that the
moving members be held in accurate alignment, as in the case of machine tools, the predominating requirement For these reasons for the frame is stiffness and not strength. the design of machine frames, in general, must be governed largely by judgment and experience, the cases where complete matheHowever, even in cases matical analysis is possible being rare. where judgment must be the guide, it is not only helpful, but
sometimes necessary
to check, as closely as possible, the stresses
in certain important sections,
by applying those fundamental

fit
formulae of Table VI, page 94, which apparently

stances.
the circum-
In
all
cases,
what may be termed a ^qualitative

is
analysis" of the frame
very desirable as a guide in properly
distributing the material,
and
in determining the
forms of the
various sections.
and line of action upon a given section are known, the stresses
If the character, value,
of every force acting
in the section can be

static
determined by applying the fundamental requirements for

equilibrium of the section, namely
(a)
:
The The
algebraic
sum
of
all
horizontal
component
forces
must
(b)
o.
algebraic
o.
sum
sum
of all vertical
component
forces
must
=
(c)
The
algebraic
of all the
428
moments must = o

The stress, sum of all the
in
429
any direction,
at
any point,
will
be the algebraic
stresses acting in that direction, at that point, as

(a),
{b),
found by applying
classification of
and
(c).
It is
impossible to
of
make
machine frames that would be
any
service,
but
the principles will be illustrated

It is to
be noted that
it
is
by applying them to typical cases. seldom possible to find the required
dimensions of a section, directly, by solving the particular equations from Table VI which apply; but, in general, the section must be assumed from the conditions given, and then checked
for strength or stiffness.
Fig.
183 illustrates a type of frame which

It is
is
quite
common
and known as an open frame.
one of the few types where
Fig. 183.
Fig. 184.
a mathematical analysis can be made with some degree of completeness. In the case of a punching-machine frame as illustrated
in Fig. 183, great stiffness
is
not essential and the design
may
to
be based on the strength required.

outlined as
Suppose the frame
be
shown
so that the dimensions of the cross-section at
any place maybe assigned.

at the sections
Evidently,
if
the stresses are checked
BC, DE, and FG,
the strength of the frame will be
fully
determined.
In the .section
BC, whose
gravity axis
is at
consider the
portion of the frame above
BC as a free body.
to
It is in
equilibrium
under the action of the exterior force P, due
punching, and the
43
internal forces exerted
MACHINE DESIGN
upon it by the lower half of the frame. There are no horizontal forces. The vertical force P must be balanced by an equal and opposite force at the section BC, which
induces a tensile stress uniformly distributed over the section,

the intensity of which
is
p1
P pounds per square

j
inch,
where
A
is
is
the area of the section.
The
only
moment
acting on the part
Pa, due to the action of P, which tends to rotate the upper part
of the
frame around
1}
the gravity axis of the section, causing a

resisting
resisting tension at B,
and a
compression at C.
is
The
maximum
intensity of these flexural stresses
given by the
fundamental equation for flexure in beams
(see J,
Table VI), or
p2
P ae where
is
is
the distance from
to
the outer fibre
and 7 X
axis Oi.
will
the
moment
of inertia of the cross-section
The
greatest tension will therefore be at
around the and its value
be p
p1
p2
is,
which
is
equation
of
Table VI.
direct stress,
This
therefore,
a case of combined flexure and
which
is fully
discussed in Art. 19, Chapter III.

is
Consider next the section DE, whose gravity axis
at
2,
and suppose the part of the frame at the left of body. There are no horizontal forces and the
DE
to
be a free
vertical force
must be balanced by a vertical tensile pull upon the upper part of the frame by the lower part. The resultant of this tensile pull, which is distributed uniformly over the whole section, may This acting at the centre of gravity. be represented by 2 K
force
may
be resolved into the components
HK
=P
perpen-
dicular to
BE, and producing a

and
is
tensile stress at right angles to
the section
P =
2
parallel
to
DE
and producing a
shearing stress along the section.

the section
The
due
only
that due, as before, to P,

stress
moment acting upon whose moment arm is a 2

.
The
moment, as determined by equation J, Table VI, may be combined with the direct stress Pj, as in the section BC, to find the maximum tensile or
tensile
and compression
to this
compression
stress.
The
shearing stress
is

pa
This
is
43
Po
-p where
A =
2
the area of the section
DE.
usually small and
may
FG.
be neglected except near the

(see Art. 14,
ends of the
beam
as in the section
FG
Chapter
III).
Consider
last the section
As
before, there are
no
hori-
zontal forces, but the vertical force

vertical resisting force
must be balanced by a
stress at
which induces a shearing

is
the
section.
The
intensity of this shearing stress
A,
where
is
A3
is
the area of the section.
Since the area of the section

it is
much
value.
smaller than at
DE
is
or
BC,
advisable to compute
its
The moment Pa 3
of the section
tion J, of
balanced as before by the resisting

stress
moment
and the resulting
may
be computed by equa-
Table VI.
Evidently these general principles
may be
applied to any section.

Fig.
riveter.
185 illustrates an open frame as applied to a power
The
of the
rivet
which
is to
be "driven"
is
is
placed between the
dies
D and D v
and pressure
applied to the movable die D, by
means
power cylinder R.
The
pressure which
is
applied
may
be very great (150 tons or more), and unless the jaws are
properly designed they

fail to
may
spring so
much
that the dies will
align properly,
and faulty work

is,
will result (see Art. 53).
Stiffness
and not strength

for
if
therefore, the essential factor in

stiff
the design;
general,
the parts are
enough they
will be,
in
strong enough.
is
The
yielding which most affects the
alignment
that
due
to the
bending of the frame B, and the
stake C, and that which
may
result
from the elongation
of the bolts
which hold these members together.
When
the riveting pressure
P
O,
is
applied, the
beams
and
tend to rotate around the point

bolts.
this
tendency being resisted by the tension in the
The
be
load which
may
.
be applied to the bolts by the force

If the
will
P
t
nuts on the bolts are set up so that a
combined
total initial tension
somewhat
greater than
is
induced
in the bolts, the stretching of the bolts,
and the consequent openwill
ing up between
the
frame and the stake,
be negligible.
43 2
(See Art. 60
MACHINE DESIGN
and
Fig. 43
and
also Art.
77.)
The
intensity of
stress in the bolts should not
exceed 6,000 pounds per square inch.
The upper
II.)
part of the frame, B, approximates a cantilever of

a.
and Case 1 of Table which occurs at D may, therefore, be computed and the maximum stress which occurs at E F may be checked by Equation J of Table VI. The stake, C, approxiuniform strength of length
(See Art. 15
The maximum
deflection
mates a cantilever
of
uniform
cross-section,
and may therefore be
Fig. 186.
treated in a similar manner.

J,
(See Case
1,
Table
I,
and Equation
Table VI.)
Fig.
186 illustrates a closed frame as applied to a vertical
steam engine.
guide
is
is
The back column, B, which
carries the crosshead

It
of cast iron, while the front columns, C, are of steel.
required to check the stresses in these columns
when
the piston
is
ascending and also when
it is
descending, the rotation of the
engine to be taken in a clockwise direction as indicated.
When
draw
resisted
the piston
is
ascending, the steam pressure tends to
and bed closer together. This tendency is by P', the combined thrust on all the columns, the vertical component of which must equal P, the total steam pressure on the piston. It may be reasonably assumed that the back column
the cylinder

carries one-half of the total thrust,
433
and that each
of the front
columns
carries one-quarter.
The
P'
thrust of the
back column,
may
be resolved into components perpendicular and parallel to
the face of the foot.
The
vertical
component
will
equal
The
horizontal
component
tends to spread the foot of the column

stress in
it.
outward and induce a bending

therefore, be secured to the
The column
if
should,
bed by
fitted bolts, or,
the bolts
are loose in the holes, the foot should be well dowelled to the bed;
or, better
still,
the foot should
fit
against a ledge cast on the bed
plate.
will
then be balanced by an equal and opposite reaction
at the feet of the front
in the
columns, thus setting up a negligible tension bed and leaving a compressive force only on the column.
By
similar reasoning each front
column
is
subjected to a com-
pressive load
P' and the
total horizontal
component
R is balanced
by that of the back column through the bed.
The
tension or compression in the piston-rod
and connectingnormal
to
rod, either ascending or descending, have a resultant R'

to the guide,
which may have a large value where the connecting-
rod
is
short
compared
to the crank.
This resultant tends
bend B, and hence
also, in a left-hand direction, the

feet.
bending
being resisted by the fastenings at the

cylinder, however, constitute a very
The columns and

and except
stiff
structure,
where the frame

be neglected.
is
made up
of light construction this effect

',
may
This reaction, R' however, also bends the column
locally, that is as a
beam
If
encastre at
is
S and N,
the effect of
being greatest when the crosshead
near half stroke.
(See
Case
18,
Table
I.)
then
it
be desired to check the central

stress
section
of the
column, the long column
due
to
P'
must
be added
to
the flexural stress
due
to
R'.
The sum
of these
stresses should not, of course be greater than the allowable stress

for the material used.
as long
columns
(see equation
The columns, C, need N, Table VI).
only be checked
434
MACHINE DESIGN
the piston
is
When
descending the steam pressure tends to

cylinder.
separate the bed
and the
The
reactions at
M and N
R
f
are reversed in direction
and the columns are put
in tension, the
horizontal components inducing negligible compression in the
bed.
The most dangerous

stress will
section in this case will be under
and the
be that due to
plus the
P'
2
tensile stress
due
to
the bending effect of
f
.
The
fastenings of the columns to the
cylinder and to the bed plate must, of course, be sufficiently strong

in tension to resist the force tending to separate the cylinder
and
bed.
In the foregoing examples the lines of action of
all
forces
acting on the section considered, lay in a plane of symmetry of

the section,
axis at right angles to this plane.
and the section tended to rotate around a gravity While this is the most usual
Fig. 187
case, occasionally the force or forces acting are not in a plane of
symmetry.
Thus
may
represent the cross-section of the
column
of a radial drilling machine, in
which
to
it is
required to
in the
check the stresses when the force P, due

position shown.
If
drilling, is
be the centre of gravity of the section, the
tendency to rotate will be around the axis X' X' at right angles to
PC,
the
arm
of the force P,
and the
resistance of the section against

inertia of the
tensile
such
rotation will
be measured by the moment of
section with reference to this axis.
The maximum
and
the
compressive stresses will occur at the fibre farthest removed from
X' X' or
pressive
at
M
P
is
and N, the
stress at
M being
The
tensile
when
direction of
upward
to the plane of the paper,

is
and com-
when
its
direction
downward.
centre of gravity,
C,
may
be located readily, by finding the intersection of any pair

If the section
of gravity axes.
Fig. 187,
it is
has an axis of symmetry, as UV,
necessary only to find the axis at right angles to

as follows:
at
UV.
This
is
most readily done graphically
Divide the
Fig. 187.
section into small areas, as
From
be, cd,
the centre of
shown by dotted lines gravity of each area draw
xxin
UV.
parallel lines ab,
preferably at right angles to the

lay off
known
axis
In Fig.
187
(a),
AB, BC,
etc.,
proportional to the respective areas
435
whose gravity axes are ab, be, etc. Take any pole O and draw AO, BO, etc. From any point on ab, draw ao indefinitely, parallel From the same point draw ob parallel to OB. From the to AO. intersection of ob and be draw co parallel to CO, and from its The intersection of intersection with cd, draw od parallel to OD.
ao and od locates the gravity axis
evident that this
XX (see
also Art. 120).
It is
method may be applied when both axes are unof inertia of the section
known.
The moment
around
X X
r
may be
lines,
most readily found by transforming the area of the figure into

an equivalent figure with
RP
as a base, as follows:
Draw
Fig. 187.
Fig. 188 (b).
X' X and plot the intercepts made by it on the given section, on each side of CB as ordinates of an equivalent section, shown in Fig. 187 by the dotted line L. The accuracy of the work may be checked with a planimeter, as it is
as
X" X"
parallel to
evident that the area of the transformed section will be equal to

that of the original.
Divide
lines
this
equivalent figure into approxiparallel to
mate
rectangles,
by
drawn
X' X'
as
shown
at
Then the moment of inertia of r around the axis X' X' will r. be its moment of inertia round its own gravity axis parallel to
X'
f
,
plus
its
area into the square of the distance between these

of the
axes.
will
The sum
be the required
moments of inertia of all such small moment of inertia of the section.
areas
436
MACHINE DESIGN
195. Distribution of Metal in Frames.
Machine frames
are
usually
made
of castings,
on account of
their
complicated shapes,
cast iron being the material

idly
most used, while
steel castings are rap-
coming
into use for severe work.
In addition
to the stresses
induced in the frame by the energy transmitted by the machine,

it
may
also be subjected to severe accidental stresses
due
to
such
causes as shrinkage, or the settling of a part of the foundation.
Both these classes

erally
of stresses are, in general, very
complex and genfre-
beyond mathematical
left to
analysis,
and the problem must
quently be
is
the
judgment
of the designer, especially if stiffness
a large factor.
its
Economy
in the use of metal, however,

shall
demands
that
distribution throughout the frame
be in accord
with the best analysis possible, and, therefore, the general prim
ciples governing the
forms of sections must be kept in mind.
The most
trying stresses to which a frame
may
It
be subjected
are torsion, flexure, or a combination of these.

in Art. 12 that the hollow section (Fig. 7) is
resisting torsion, and,
if
has been noted

effective for
most
this
7,
be the predominating
or modified sections as
stress sections
such as are shown in Fig.

187, are correct.
It
shown
in Fig.
was
also noted in Art. 19
(Fig. 10) that in
the case of cast iron, or other metal
whose
tensile strength is
much
is
less
than
its
compressive strength, a great saving of material
effected
Fig.
by massing the metal on the tension side as shown in 188 (a) thus making the tensile and compressive stresses more
;
in proportion to the strength of the material.
If
(in
then the pre-
dominating
stress in
a frame
is
simple flexure
is
a given plane),
if,
a section like that
shown
in Fig. 188 (a)
allowable, but
if
in
addition, torsional strength

flexure
is
must be withstood, or
the plane of
may
change, a section similar to that shown in Fig. 188 (b)

it
better design, since

(a).
combines the merits of both Figs. 187 and

better to
188
Sometimes
it is
make
the section so large that
the flexural stress
can be safely withstood by a wall of uniform

is
thickness, as in Fig. 187, as the construction of the pattern
simpler and the shrinkage stresses less serious than in such sections as
shown
in Fig. 188.
The metal
in the walls will be
much
sounder, also, as the thick sections of Fig. 188 are very likely to
have a porous
interior,
due
to shrinkage.
Cast-iron parts more

than four or
this
five
437
inches thick are almost sure to be defective in

walls of such sections as
manner.
The
shown
in Fig.
188
should taper uniformly from the thick part to the thin parts, and
all
corners should be well rounded, and

as
filleted, to
minimize as
far
possible the concentration of shrinkage stresses.
Thin
This
wide flanges or webs should not be cast integral with thick heavy
parts, as
is
unequal shrinkage and porosity are sure

edge of the rib
to result.
especially true of thin ribs cast
on the tension
side of large
sections, as the
is liable to
crack through shrinkage,
thus starting rupture across the entire section.
Small brackets
or other attachments of thin sections should never be cast

large frame, as they seldom cast well.
on a
section of
moderate
re-
thickness
is
often stronger than a thicker one, since the greatest

is
strength of cast iron
in the outer skin.
It
should also be
stiff
membered
to
do the
to
when a frame is both required work at low speeds,

that even
strong and
it
enough
may
not have mass
enough
rapidly.
dictated
set up when running more more metal in the frame than is Openings for supporting or by other requirements.
absorb the vibrations

call
This may
for
removing cores should be placed near the gravity axis so as

to
reduce the strength as

196.
little
as possible (see Fig. 191).
Attachments and Supports.

is
The
general appearance of
more by the outline of the main frame than by that of any other member. This outline should, therefore, be clearly shown, and not obliterated at places by the various attachments which restrain the moving parts or support the frame. In Fig. 183 is shown the outline of a frame in which the various
a machine
affected
sections
have been proportioned in accordance with the loads
brought upon them, and the various bosses
and the support S main member. Fig. 184 illustrates the same machine with the attachments merged into the main member, thereby destroying the character of the design, and also making it more difficult to judge of the relative strength of various
appear as attachments
to the
sections of the frame.
The form
the service
is
it
of an attachment will, of course, be governed

is
by
it
required
to
render and the manner in which

If the outline of
loaded and supported.
the attachment
is
438
MACHINE DESIGN
based on theoretical considerations, care should be exercised

that
if
all
the modifying influences are duly considered.
Thus
of
parabolic outlines are given to an attachment, such as the
housings
the
H for supporting the tool in Fig.

by the bending
192, the
upper end
housing must be modified from the theoretical parabolic

effect of the force
outline indicated
P, so as to
is
provide for the shearing effect at the upper end, which
fre-
quently neglected.
If the
(See also Article 15.)
frame
rests directly
on the
floor its outlines
should be
carried
down
to the floor in
such a manner as will give an appear-
Fig. 189.
Fig. 190.
ance of
stability.
Thus
Fig.
189 shows such a machine frame
on which the
Fig.
vertical outline of the
back
of the
frame
is
undercut.
carried
190
shows the same machine with the

is
outline
straight to the floor
as stability
and the improvement in appearance, so far concerned, is obvious. Fig. 191 shows the outline
.the upright,
of a planing
machine in which
form
of a leg.
U,
is
carried to the
is
floor at V, in the
This construction
designed to
not correct,
as
is
an attachment
to the bed, to
all
resist the force of

itself
the cut
and
to
transfer
it
the bed,
which should
be
stiff
it.
enough
withstand
such stress
thus brought
upon
Any
settling of
the foundation might affect
the alignment
of
439
and hence the arrangement shown in Fig. 192
is
more nearly
correct.
In large machines the frame usually rests directly on the

foundation, and should have sufficient stiffness to resist distortion
due
to
the settling of the foundation, since the latter
is
very
difficult to avoid.
In smaller machines the frame
is
carried on
(a) cabinet or box shown in Fig. 193. The choice of support will, of course, depend on the type and In any case the number of points of size of the machines. support should be as few as possible. If the machine can be
supports, which
pillar
may
be of two general types,

(b)
supports (Fig. 192), and
legs as
supported on three points

affected
it is
evident that the frame cannot be

It is difficult, in general,
by
settling of the foundation.
Fig. 192.
to obtain three-point support, but
it is
seldom necessary
is
to place
supports as close together as in Fig. 191 (which

actual design), where the frame
is
taken from an
Fig.
carried on eight points.
192 shows the same frame properly carried on box supports,

the supports themselves being so
stiff
as materially to assist the

to so-called
frame and practically reducing the support

support.
box-pillar, the
two-point
Small machines can often be supported on a single
overhanging parts of the frame having a parabolic

in
outline
as
suggested
height
if
Fig.
sides
189.
If
the
box
pillar
is
of
considerable
the
top;
for
the
should
taper
slightly
toward
at
made
parallel the pillar will

It
is
appear wider
the top than at the bottom.

of support throughout,
i.e.,
preferable to use one form

pillars or
all
all
box
legs,
and not
one or more of each.
"
440
MACHINE DESIGN
the frame
must be supported on legs, as in Fig. 195, these should not curve outward as in Fig. 193, unless it is absolutely essential in order to obtain stability.
When
Spreading the legs
as in Fig. 193 lengthens the distance between the reactions, R, R,
and, therefore, increases the bending effect on the bed and legs
as a whole.
The
leg
shown
in profile in Fig. 194 is better

legs should
and
much
outline
easier to
make.
The
be so placed that the
forms a continuation of the principal vertical outline
of the frame, as
shown
in Fig. 194.
the end view of the legs as
The same remarks apply to shown in Figs. 195 and 196. The
of Fig. 195 are not only use-
complex curves and ornate features
Fig. 193.
Fig. 194.
Fig. 195.
Fig. 196.
less
but expensive.
It is
not always possible or desirable to
make
machine frames and supports with simple straight-line outlines; but where curves are necessary they should be as simple as possible and in general the best results can be obtained by using arcs of circles or parabolas. Ornamentation of a fanciful nature is not permissible anywhere, as it really detracts from the appearHarance of the machine, and adds to the cost of production.
;
mony
of
of design can be attained
by making the various members
correct
proportions to withstand the loads brought upon
them, and by using the simplest and most direct design with
smooth
It is
transition curves between straight lines which intersect. a proverb in design that " what is right looks right.
INDEX
Absolute
efficiency,
no
Belts, slip of,
310
Accumulator, hydraulic, 29 Air compressor, 26

Air reservoir, 29
Anti-friction metals, 233
velocity of, 317
Bending moment, equivalent or ideal, 49

Bevel gears, ^8^
Block brakes, 355

89
Boiler plate, strength
rivets,
of,
Apparent factor of
Axles, 285
safety,
154
strength
of,
154
Bolts, allowable stress in, 176
Babbitt metal, 23$

Ball bearings, 277
efficiency of, 172
experiments on the strength

for reinforcing castings, 207
initial
of,
170
allowable load on, 282
Bands,
thin, 205
tension in, 169
Barnard, Prof.
ings), 149
W. N.
(riveted fasten-
location of, 181
Professor Sweet's experiments on,

of,
Beams, general theory
40
181
resilience of, 178
of,
of uniform strength, 42
Bearing pressures on journals, table

251
resultant stress in, 172
straining action in, 168, 169, 172

stud, 163, 164
tap, 163, 164
on
sliding surfaces, 238
Bearing, step, 264

Bearings, allowable pressure on, 251
ball, 271,
through, 163, 164
277
of,
Brakes, block, 355

coefficient of friction for,
collar,
264 243
363
construction
differential,
friction,
359
forms
metals
of,
239
355
for,
2^
strap, 357
perfectly lubricated, 253
Briggs' system of pipe threads, 168
radiation of heat from, 247
Butt joints in plates, 138, 146
273 table of proportions

thrust, 263
Belt,
roller, 271,
of,
252
Cap
screws, 163
Prof.
Carman,
of,
A. P., experiments on
314 transmission, theory of, 308

weight
of,
example of design
tubes, 218
Carrying strength, 89
Belting, efficiency of, 318
314
Chain drums and sheaves, 340 Renold, Morse, 348

313
roller,
Belts, coefficient of friction of,
block, stud, 345
construction
of,
308
Chains, 338
conveyor, 344 for power transmission, 344
creep
of,
310
practical consideration of, 319
practical rules for, 318
proof
test of,
340
44i
442
Chains,
silent,
INDEX
345
Efficiency of belting, 318
of bolts, 172
strength
of,
340
weldless, 340
of riveted fastenings, 141
Clavarino's formula, 223

Clutches, allowable pressure on, 363
of screws, 157
of square- threaded screws, 157 of triangular-threaded screws, 162
Efficiencies of
band, 362
coefficient of friction for,
363
machine elements, 113
conical, 359
disc,
Elastic limit, 34
resilience, 77
361
359 magnetic, 363 radially expanding, 360

shaft, 301,
friction,
Elasticity, coefficient of,
34
Energy
cycle, 6
in air compressor, 26 in steam engine, 16
305
Coefficient of elasticity, 34
of friction for screws, 184
Coefficients of friction for brakes
clutches, 363
Energy problems, 6
redistribution of, 29
and
Euler's formula for columns, 62
of friction for friction wheels, 353 of friction of pivots, 269
Factor of safety, 35, 88 on boiler work, 155

Factors of safety, table
of,
Collar bearings, 264, 267
91
Columns, eccentric loading

or long struts, 61
of,
73
Fairbairn,
flues,
Sir
Wm., experiments on
217
Compression and torsion, combined, 57 in machine elements, 36 Conservation of energy, 3-6

Constraining surfaces, materials
of,
Fatigue of materials, 82
Feather keys, 196

Feathers, table of dimensions
of,
197
232
Flanges, pipe, 226

Flather, Prof.,
Continuous system of rope-driving, 329

Cotters, stresses in, 198
on rope
drives, 323,
324
Flexure and direct
Coupling, flange shaft, 303 Hook's, 304
58 torsion combined, 43
stress,
in
machine elements, 40
joints,
Oldham, 304
Couplings, flexible shaft, 306
shaft,
Flues, 211
Flywheel rim
422 412
301
Flywheels, 406
coefficients of fluctuation,
Crank-effort diagram, 20
Cycloidal gear teeth, 365 Cylinders, thick, 223

thin, 211,
construction
of,
419
of,
experiments on the strength

general theory
stresses in,
of,
424
215
406
413
Deflection of ropes, 332

table of, 333 Deformation, work
Force
fits,
200
practical considerations in, 204

of,
77
stresses
due
to,
201
6, 9,
Differential brake, 359
Forces acting on machines,

Friction, applications of, clutches, 359
31
Discs, rotating, 425
350
Efficiency, absolute,
definition of, 6
no
109
coefficient of, 97, 99, 104, 105
general theory
of,
of,
96
general theory
laws
of,
98
mechanical,
no
of circular surfaces, 97
INDEX
Friction of dry surfaces, 98
of
flat
443
surfaces, 97
Gears, rawhide, allowable load on, 389 Gordon's formula for columns, 67
of lubricated surfaces, 99 of triangular threads, 162
Helical
of screws, 157 of rolling, 99

static,
gearing, 392 Hindley worm, 398 Hobs and hobbing, 397
100
of general laws of, 109
Hoisting mechanism,
9,
29
summary
wheels, allowable pressures on, 352

coefficients of friction for,
Hook's coupling, 305 Hooks, hoisting, 341

strength
of,
353
341
forms
of,
350
for,
table of proportions of, 343
materials
352
by, 353
power transmitted
wedge-faced, 354
Hoops, 205 Hunt, C. W., on rope driving, 325 system of gear teeth, 390
Impact, shock, 78
Imperfect lubrication, 102
work
Furnace
of,
97
corrugated, 222
flues,
Gear
teeth, allowable stresses in, cut,
386
Inertia effects in general, 29
369
redistribution of, 29
Inertia forces in steam engines, 19
cycloidal, 365
Fellows system of stub, 39c
Involute gear teeth, 365
Hunt
system, 390
involute, 365
Johnson's,
66
J. B.,
formula for columns,
machine moulded, 369 methods of making, 369

proportions
of,
T. H., formula for columns, 65

Journals, bearing pressure on, 251
368, 370
shrouding
strength
stub,
of,
389
design
of,
376
390
of,
of, 257 imperfectly lubricated, 249
of, 245, 257 examples of design
wear on, 388

width of face
perfectly lubricated, 253
388
wheels, forces acting on, 373 mortise, 371
Keys, draw, 192

190 forms of, 190
flat,
rawhide, 372
strength of rims
and arms, 391

364
saddle, 190
stresses in, 192
Gearing, efficiency of spur, 392

general principles
helical or twisted,
of,
sunk, 190
table of dimensions of sunk, 196
392
of,
herring-bone, 393 interchangeable systems

screw, 395 skew-bevel, 395
spiral,
Woodruff, 191
366
Kinematics, 6
Lame's formula, 223
395 standard form-
Lap
of,
joints in plates, 138, 145

of, 247,
366
Lasche, experiments
255
strength of twisted, 393
worm, 395 Gears, allowable speed

bevel, 383
Launhart's formula, 85 Lewis', Wilfred, formula for gear

Live load,
82
teeth,
of,
387
effect of,
444
INDEX
Resilience, 76
elastic,
Load, steady, dead, suddenly applied, 31

Lubrication, imperfect, 102
77
methods
of, 100,
261
of,
of bolts, 178
of journals,
methods
261
Ritter's
formula for columns, 68

joints, 138,
of sliding surfaces, 238

perfect,
Riveted fastenings, butt
146
106
chain riveting, 138

efficiency of, 141
Machine
attachments, 437
factor of safety in, 155

failure of, 141
design, definition of, 1
frames, 428
distribution of metal in, 436
stresses in,
428
432
forms of joints, 137 general considerations, 136 general equations for, 147
lap joints, 138, 144
stresses in closed, stresses in open,
429
making
of,
151
of,
screws, 163, 164
marginal strength
143
supports, 437
McBride, James, experiments on

ciency of bolts, 172
effi-
practical consideration of, 149 practical rules for, 155

relative strength of, 14*-
Mechanical advantage, 28
efficiency,
no
staggered riveting, 138 strength of materials foi, 154

stresses in, 139
Mechanism, definition of, 2 Micro flaws, theory of, 83 Moore, Prof. H. F., experiments of, 107 on riveted fastenings, 149
theoretical strength of, 144
Riveting, machine, 153

Rivets, diagonal pitch of, 138, 143
Morse
chain, 348
pitch of, 138
Multiple system of rope-driving, 329
Oil
film,
106
in perfect lubrication, 106
transverse pitch of, 138, 143 Roller bearings, 271, 273 allowable load on, 276
grooves, 262
Rope-driving, sheaves for fibrous, 331 systems of fibrous, 329
Oldham
Perfect
coupling, 304
lubrication, 106
Rope
transmission, theory
of,
309, 323
(by wire), 338
Ropes, cotton, 322

Ropes, deflection of fibrous, 332
fibrous hoisting, ^^y
Pipe couplings and flanges, 226

threads, 168
Pipes, 211, 217
hemp,
of,
leather, etc.,
322
Piping, practical considerations
224
Manila, 322
materials for fibrous, 322 materials for wire, 334 strength of fibrous, 327
Pivots, coefficient of friction of, 269
Planing machine, 30 Plates, thin, 228
Power, definition
Pulleys,
of,
strength of fibrous hoisting, 338
406 Punching machine,
10,
60
strength of wire hoisting, 339 velocity of fibrous, 327
wire hoisting ropes, 338
Rankine's equation
141
for columns, 67
Rotating
discs,
425
Relative strength of riveted fastenings,
Screw
fastenings, 163
Renold chain, 348
gearing, 395
INDEX
Screw and screw fastenings, 156
Screws, bearing pressure on, 186
cap, 163, 164
coefficient of friction of, 184
445
flat,
Springs,
design
of,
119
forms
of,
116
128
helical, 117
design
of,
design
of,
for
power transmission,
springs in torsion, 135

spiral,
187
efficiency of, 157
118
of,
laminated or plate, design

materials
spiral,
of,
123
for
for
power transmission, 183 power transmission, efficiency

184
of,
115
118
teeth, strength of,
of,
Spur gear
156
376
forms
gearing, efficiency of, 392

gears, allowable speed of, 387
friction of, 157
machine, 163, 164 mechanical advantage

multiple-threaded, 184
allowable stress
of,
183
in, 386 machine-moulded, 369 width of face of, 388
stresses in transmission,
186
U. S. or Sellers standard, 166
Stayed surfaces, 231 Steam engine, energy distribution

Step bearing, 264
Stewart,
Prof.
in,
16
Whit worth standard, 166

Sellers shaft coupling,
303 standard screws, 166
R. T., experiments on
tubes, 218
Set screws, 163, 165
Storage battery, 29
Strain, definition of, 32
Shaft clutches, 301

coupling, flange, 303
couplings, 301
Shafts, allowable deflection of, 299
Straining action, nature
of,
32
table of formulae for, 94
allowable span
of,
299
Strap brakes, 357 Strength of materials, table

Stress,
of,
93
factors of safety for, 289
compound,
33,
40
hollow, 300
definition of, 32
subjected to torsion, 288
predominating or primary, 40
strain diagram, 33
subjected to torsion and bending,
290
torsional stiffness of, 298
working, 35
Stribeck, Prof., experiments of, 255, 272
of,
whirling
299
Stub gear Stud
teeth,
390
164
Shaping machine, energy distribution

in,
bolts, 163,
10
Sweet, Prof., method of relieving sliding

surfaces, 237
Shear in machine elements, 36
Shock
Shrink
in
machine members, 78
200, 207
fits,
Tap
389
bolts,
164
practical considerations in, 204
Shrouding of gear
teeth,
Taylor, F. M., rules for belting, 319 Temperature, coefficient of expansion,

75
stresses
Sliding surfaces, 233
bearing pressures on, 238

lubrication of, 238
due
to,
75
Spheres, 213
Splines, k>6
Tension in machine elements, 35 Thrust bearing for worms, 405 Thrust bearings, 263
allowable pressures on, 270
efficiency of, 268
Springs, applications of, 114

characteristics of, 114
flat,
116
Toothed gearing, angular velocity of, 365
446
Toothed gearing,
366
Torsion
57
classification of,
INDEX
364
of,
Van Stone
pipe flanges, 227

formula, 86
interchangeable
systems
Weyrauch's
and compression, combined,
Whitworth standard screws, 166 Wire rope transmission, ^^^

theory
of,
and flexure, combined, 43 in machine elements, 36 Tower, Beaucamp, experiments

Tower's experiments, 106
334
ropes, materials for, 334

of,
254
power transmitted by, 336

Wohler's experiments
of,
84
Towne, H.
R., experiments
on hooks,
of,
Work
of deformation, 76
343 Triangular threads, efficiency

friction of, 162
162
Working stress, 35 Worm and worm wheel, 395

gearing, design of, 403
efficiency of,
Tubes, 211, 217 Twisted gears, 392
399
limiting pressures on, 401

limiting velocities of, 401
Ultimate
US.
strength, definition of, 34
velocity ratio of, 398
Unions, pipe, 226
Hindley, 398
of,
standard screws, table
167
thrust bearing, 405
Short-title Catalogue
OF THE
PUBLICATIONS
or
JOHN WILEY & SONS

New York
London:
CHAPMAN & HALL,
Limited
ARRANGED UNDER SUBJECTS

Descriptive circulars sent on application. Books marked with an asterisk (*) are All books are bound in cloth unless otherwise stated. sold at net prices only.
AGRICULTUREHORTICULTUREFORESTRY.
Armsby's Principles of Animal Nutrition Budd and Hansen's American Horticultural Manual: Part I. Propagation, Culture, and Improvement
8vo, $4 00
Part II. Systematic Pomology Elliott's Engineering for Land Drainage (Second Edition, Rewritten) Practical Farm Drainage. 8vo, Graves's Forest Mensuration 12mo, Green's Principles of American Forestry (Woll.) 12mo, Grotenfelt's Principles of Modern Dairy Practice. * Herrick's Denatured or Industrial Alcohol 8vo, Waugh's Landscape Gardening. (New Edition, Rewritten. In Kemp and Preparation) * McKay and Larsen's Principles and Practice of Butter-making 8vo, Maynard's Landscape Gardening as Applied to Home Decoration 12mo, 12mo, Sanderson's Insects Injurious to Staple Crops Sanderson and Headlee's Insects Injurious to Garden Crops. (In Preparation)
.
12mo, 12mo, 12mo, 12mo,
50 50 1 50 1 50 4 00 1 50 2 00 4 00
1
1 1
50 50 50
* Schwarz's Longleaf Pine in Virgin Forests Stockbridge's Rocks and Soils
12mo,
8vo, 8vo,
25
Winton's Microscopy of Vegetable Foods Woll's Handbook for Farmers and Dairymen
2 50 7 50
1
16mo,
50
ARCHITECTURE.
Baldwin's Steam Heating for Buildings Berg's Buildings and Structures of American Railroads Birkmire's Architectural Iron and Steel Compound Riveted Girders as Applied in Buildings Planning and Construction of American Theatres Planning and Construction of High Office Buildings Skeleton Construction in Buildings
1
12mo,
4to,
8vo, 8vo, 8vo, 8vo, 8vo,
2 50 5 00 3 50 2 00 3 00 3 50 3 00
Briggs's Modern American School Buildings Byrne's Inspection of Materials and Wormanship Employed
8vo, $4
in Construction.
00
16mo,
Carpenter's Heating and Ventilating of Buildings * Corthell's Allowable Pressure on Deep Foundations Fieitag's Architectural Engineering Fireproofing of Steel Buildings (Fourth Edition, Gerhard'r Guide to Sanitary Inspections. vised and Enlarged) * Modern Baths and Bath Houses
Sanitation of Public Buildings Theatre Fires and Panics Johnson's Statics by Algebraic and Graphic Methods
8vo,
00 4 00
3
1
12mo,
8vo, 8vo,
25
3 50 2 50
1
1 1
Entirely
Re12mo,
8vo,
12mo',
50
50 50 00
3 00
12mo,
8vo, 8vo, 16mo, mor., 8vo,
4to,
Kellaway's How to Lay Out Suburban Home Grounds Kidder's Architects' and Builders' Pocket-book Merrill's Stones for Building and Decoration Monckton's Stair-building Patton's Practical Treatise on Foundations Peabody's Naval Architecture Rice's Concrete-block Manufacture. Richey's Handbook for Superintendents of Construction Building Foreman's Pocket Book and Ready Reference. * Building Mechanics' Ready Reference Series: * Carpenters' and Woodworkers' Edition * Cement Workers' and Plasterers' Edition * Plumbers', Steam-Fitters', and Tinners' Edition. * Stone- and Brick-masons' Edition Sabin's House Painting Siebert and Biggin's Modern Stone-cutting and Masonry Snow's Principal Species of Wood
8vo, 8vo, 8vo,

.
16mo, mor. 16mo, mor.
2 2 5 5 4 5 7 2 4 5
1 1
00 00 00 00 00 50 00 00 00
50 50 50 50
16mo, mor. 16mo, mor. 16mo, mor. 16mo, mor. 12mo,

8vo, 8vo,
1
1 1 1
Towne's Locks and Builders' Hardware Wait's Engineering and Architectural Jurisprudence
18mo, mor.
8vo, Sheep,
Law of Contracts Law of Operations

tecture
8 vo, Preliminary to Construction in Engineering and Archi-
3 3 6 6 3 00
00 50 50 00 00 50
8vo, 5 00 Sheep, 5 50 Conditioning Wilson's Air 12mo, 1 50 Worcester and Atkinson's Small Hospitals, Establishment and Maintenance, Suggestions for Hospital Architecture, with Plans for a Small Hospital.
12mo,
25
ARMY AND NAVY.

Bernadou's Smokeless Powder, Nitro-cellulose, and the Theory of the Cellulose Molecule 12mo, 2 50 Chase's Art of Pattern Making 12mo, 2 50 Screw Propellers and Marine Propulsion 8vo, 3 00
* Cloke's Enlisted Specialists'
Examiner
Gunner's Examiner
Craig's
8vo, 8vo,
4to,
Azimuth Crehore and Squier's Polarizing Photo-chronograph

* Davis's
8vo,
Elements of Law 8vo, * Treatise on the Military Law of United States 8vo, 24mo, mor. DeBrack's Cavalry Outpost Duties. (Carr.) * Dudley's Military Law and the Procedure of Courts-martial. .Large 12mo, 8vo, Durand's Resistance and Propulsion of Ships * Dyer's Handbook of Light Artillery 12mo, Explosives High 8vo, Modern Eissler's * Fiebeger's Text-book on Field Fortification Large 12mo, 18mo, Hamilton and Bond's The Gunner's Catechism * Hoff 's Elementary Naval Tactics 8vo, 8vo, Ingalls's Handbook of Problems in Direct Fire * Lissak's Ordnance and Gunnery 8vo,
.
2 00 1 50 3 50 3 00 2 50 7 00 2 00 2 50 5 00 3 00 4 00 2 00
1 1
00 50
4 00 6 00
* Ludlow's Logarithmic and Trigonometric Tables * Lyons's Treatise on Electromagnetic Phenomena. Vols. * Mahan's Permanent Fortifications. (Mercur.)
I.
and
II..8vo,each, 8vo. half mor.
8vo, $1 00 6 00 7 50
1
Manual for Courts-martial * Mercur's Attack of Fortified Places * Elements of the Art of War
16mo,mor. 12mo,
8vo,
50
2 00
4 00
1
Nixon's Adjutants' Manual 24mo, 8vo, Peabody's Naval Architecture * Phelps's Practical Marine Surveying 8vo, 8vo, Putnam's Nautical Charts 8vo, Rust's Ex-meridian Altitude, Azimuth and Star- Finding Tables Sharpe's Art of Subsisting Armies in War 18mo, mor, * Tupes and Poole's Manual of Bayonet Exercises and Musketry Fencing.
7 2 2 5
1
00 50 50 00 00 50
24mo,
* Weaver's Military Explosives Woodhull's Notes on Military Hygiene
leather,
8vo,
16mo,
50 3 00 1 50
ASSAYING.
Betts's Lead Refining by Electrolysis 8vo, Fletcher's Practical Instructions in Quantitative Assaying with the Blowpipe.
4 00
1
Pardoe's Manual of Practical Assaying. (Sixth vised and Enlarged) 8vo, Lodge's Notes on Assaying and Metallurgical Laboratory Experiments.. 8vo, Low's Technical Methods of Ore Analysis 8vo, Miller's Cyanide Process 12mo,
Furman and
16mo, mor. Edition. Re-
50
3 00 3 00 3 00
1
1
Manual
of
Assaying
12mo,
its
Minet's Production of
Aluminum and
Industrial Use.
Treatment of Gold Ores Ricketts and Miller's Notes on Assaying Robine and Lenglen's Cyanide Industry. (Le Clerc.) Ulke's Modern Electrolytic Copper Refining Wilson's Chlorination Process Cyanide Processes
O'Driscoll's Notes on the
(Waldo.). .12mo, 8vo, 8vo, 8vo, 8vo,

.
2 2 3 4 3
1
12mo, 12mo,
00 00 50 00 00 00 00 50 50
ASTRONOMY.
Comstock's Field Astronomy for Engineers
Craig's Azimuth Crandall's Text-book on Geodesy and Least Squares Doolittle's Treatise on Pracical Astronomy Hayford's Text-book of Geodetic Astronomy
8vo,
4to,
8vo, 8vo, 8vo,
Hosmer's Azimuth 16mo, mor. Merriman's Elements of Precise Surveying and Geodesy 8vo, * Michie and Harlow's Practical Astronomy 8vo, 8vo, Rust's Ex-meridian Altitude, Azimuth and Star-Finding Tables * White's Elements of Theoretical and Descriptive Astronomy 12mo,
50 50 00 00 00 1 00 2 50 3 00 5 00 2 00
2 3 3 4 3
CHEMISTRY.
Physiological Chemistry in Thirty Lectures. (Hall and Defren) 8vo, * Abegg's Theory of Electrolytic Dissociation, (von Ende.) 12mo, (Matthews.) 8vo, Alexeyeff's General Principles of Organic Syntheses. 8vo, Allen's Tables for Iron Analysis 8vo, Armsby's Principles of Animal Nutrition (Mandel.) Large 12mo, Arnold's Compendium of Chemistry. Association of State and National Food and Dairy Departments, Hartford 8vo, Meeting, 1906 8vo, Jamestown Meeting, 1907
* Abderhalden's
5 00
1
25
3 00 3 00
4 00 3 50
3 00 3 00
Austen's Notes for Chemical Students 12mo, $1 50 (In Preparation;. Baskerville's Chemical Elements. Bernadou's Smokeless Powder. Nitro-cellulose, and Theory of the Cellulose Molecule 12mo, 2 50 Biltz's Introduction to Inorganic Chemistry. (Hall and Phelan). (In Press). Laboratory Methods of Inorganic Chemistry. (Hall and Blanchard). 8vo, 3 00 * Blanchard's Synthetic Inorganic Chemistry 12mo, 1 00 * Browning's Introduction to the Rarer Elements 8vo, 1 50 * Claassen's Beet-sugar Manufacture. (Hall and Rolfe.) 8vo, 3 00 Classen's Quantitative Chemical Analysis by Electrolysis. (Boltwood.).8vo, 3 00 Cohn's Indicators and Test-papers 12mo, 2 00 Tests and Reagents 8vo, 3 00 * Danneel's Electrochemistry. (Merriam.) 12mo, 1 25 Dannerth's Methods of Textile Chemistry 12mo, 2 00 Duhem's Thermodynamics and Chemistry. (Burgess.) 8vo, 4 00 Effront's Enzymes and their Applications. (Prescott.) 8vo, 3 00 Eissler's Modern High Explosives 8vo, 4 00 Erdmann's Introduction to Chemical Preparations. (Dunlap.) 12mo, 1 25 * Fischer's Physiology of Alimentation Large 12mo, 2 00 Fletcher's Practical Instructions in Quantitative Assaying with the Blowpipe. 12mo, mor. 1 50 Fowler's Sewage Works Analyses 12mo, 2 00 Fresenius's Manual of Qualitative Chemical Analysis. (Wells.) 8vo, 5 00 Manual of Qualitative Chemical Analysis. Part I. Descriptive. (Wells. )8vo, 3 00 Quantitative Chemical Analysis. (Cohn.) 2 vols 8vo, 12 50 When Sold Separately, Vol. I, $6. Vol. II, $8. Fuertes's Water and Public Health 12mo, 1 50 Furman and Pardoe's Manual of Practical Assaying. (Sixth Edition, Revised and Enlarged.) 8vo, 3 00 * Getman's Exercises in Physical Chemistry 12mo, 2 00 Gill's Gas and Fuel Analysis for Engineers 12mo, 1 25 * Gooch and Browning's Outlines of Qualitative Chemical Analysis.
Grotenfelt's Principles of Modern Dairy Practice. (Woll.) Groth's Introduction to Chemical Crystallography (Marshall) Hammarsten's Text-book of Physiological Chemistry. (Mandel.) Hanausek's Microscopy of Technical Products. (Winton.)
Large 12mo, 12mo, 12mo,

8vo, 8vo,
25
2 00
1 25 4 00 5 00 2 00 2 50 4 00 3 00 1 00
* Haskins and Macleod's Organic Chemistry 12mo, Hering's Ready Reference Tables (Conversion Factors) 16mo, mor. * Herrick's Denatured or Industrial Alcohol 8vo, Hinds's Inorganic Chemistry 8vo, * Laboratory Manual for Students 12mo, * Holleman's Laboratory Manual of Organic Chemistry for Beginners. (Walker.) 12mo, Text-book of Inorganic Chemistry. (Cooper.) 8vo, Text-book of Organic Chemistry. (Walker and Mott.) 8vo, * Holley's Lead and Zinc Pigments Large 12mo, Holley and Ladd's Analysis of Mixed Paints, Color Pigments, and Varnishes.
00
2 50 2 50 3 00
2 50 3 00
1
Large 12mo, 8vo, Hopkins's Oil-chemists' Handbook Work Physiological Chemistry. Laboratory in .8vo, Directions for Jackson's Johnson's Rapid Methods for the Chemical Analysis of Special Steels, SteelLarge 12mo, making Alloys and Graphite 8vo, (Tingle.) Landauer's Spectrum Analysis. * Langworthy and Austen's Occurrence of Aluminum in Vegetable Prod8vo, ucts, Animal Products, and Natural Waters Lassar-Cohn's Application of Some General Reactions to Investigations in (Tingle.) 12mo, Organic Chemistry. Leach's Inspection and Analysis of Food with Special Reference to State Control 8vo, (Lorenz.) 8vo, Lob's Electrochemistry of Organic Compounds. Lodge's Notes on Assaying and Metallurgical Laboratory Experiments.. 8vo, 8vo, Low's Technical Method of Ore Analysis 12mo, Lunge's Techno-chemical Analysis. (Cohn.) * McKay and Larsen's Principles and Practice of Butter-making 8vo, 12mo, Maire's Modern Pigments and their Vehicles
25
3 00 3 00 2 00
1
00
50 00 00 00 1 00 1 50 2 00
7 3 3 3
Mandel's Handbook for Bio-chemical Laboratory
12mo, $1 50
60 25
* Martin's Laboratory Guide to Qualitative Analysis with the Blowpipe
12mo, (Chemical and Bacteriological.) 12mo, Mason's Examination of Water. (Considered Principally from a Sanitary Standpoint.) Water-supply.
* Mathewson's First Principles of Chemical Theory Matthews's Laboratory Manual of Dyeing and Textile Chemistry 2d Edition, Rewritten Textile Fibres. * Meyer's Determination of Radicles in Carbon Compounds.
8vo, 8vo, 8vo, 8vo,

(Tingle.)
4 00
1
00
3 50
4 00
1 1
Third Edition 12mo, 12mo, Miller's Cyanide Process 12mo, Manual of Assaying (Waldo.). .12mo, Minet's Production of Aluminum and its Industrial Use. 12mo, Mixter's Elementary Text-book of Chemistry 12mo, Morgan's Elements of Physical Chemistry 12mo, Outline of the Theory of Solutions and its Results * Physical Chemistry for Electrical Engineers 12mo, 16mo, mor. Morse's Calculations used in Cane-sugar Factories * Muir's History of Chemical Theories and Laws 8vo, Mulliken's General Method for the Identification of Pure Organic Compounds. Vol. I. Compounds of Carbon with Hydrogen and Oxygen. Large 8vo, Nitrogenous Compounds. (In Preparation). Vol. II.
.
2
1
3
1
1
1
25 00 00 50 50 00 00 50 50 00
5 00
The Commercial Dyestuffs. (In Press). Vol. III. O'Driscoll's Notes on the Treatment of Gold Ores
Ostwald's Conversations on Chemistry.
8vo,
2 00
1
Part One. (Ramsey.) 12mo, Part Two. (Turnbull.) 12mo, Owen and Standage's Dyeing and Cleaning of Textile Fabrics 12mo, * Palmer's Practical Test Book of Chemistry 12mo, * Pauli's Physical Chemistry in the Service of Medicine. (Fischer.) .12mo, Penfield's Tables of Minerals, Including the Use of Minerals and Statistics of Domestic Production 8vo, (Biddle.) Pictet's Alkaloids and their Chemical Constitution. 8vo, Poole's Calorific Power of Fuels 8vo, Prescott and Winslow's Elements of Water Bacteriology, with Special Reference to Sanitary Water Analysis 12mo, * Reisig's Guide to Piece-Dyeing 8vo, Richards and Woodman's Air, Water, and Food from a Sanitary Standpoint 8vo. Ricketts and Miller's Notes on Assaying 8vo, Rideal's Disinfection and the Preservation of Food 8vo, Sewage and the Bacterial Purification of Sewage 8vo, Rigg's Elementary Manual for the Chemical Laboratory 8vo, Robine and Lenglen's Cyanide Industry. (Le Clerc.) 8vo,
.
50
2 00 2 00
1
00 25
00
5 00 3 00
1
15
50 00
2 00 3 00
Ruddiman's Incompatibilities in Prescriptions Whys in Pharmacy Ruer's Elements of Metallography. (Mathewson). (In Press.) Sabin's Industrial and Artistic Technology of Paint and Varnish Salkowski's Physiological and Pathological Chemistry. (Orndorff.) Schimpf 's Essentials of Volumetric Analysis Manual of Volumetric Analysis. (Fifth Edition, Rewritten)
8vo,
12mo,
8vo, 8vo,
4 00 4 00 1 25 4 00 2 00 1 00
3 00 2 50
1
12mo,
8vo, 8vo, 8vo,
25 25 50 00 00 50 50 00 50 00 00 00 50 00
5 00
1
* Qualitative Chemical Analysis Smith's Lecture Notes on Chemistry for Dental Students Spencer's Handbook for Cane Sugar Manufacturers 16mo, mor. Handbook for Chemists of Beet-sugar Houses 16mo, mor. Stockbridge's Rocks and Soils 8vo, Stone's Practical Testing of Gas and Gas Meters 8vo, * Tillman's Descriptive General Chemistry 8vo, * Elementary Lessons in Heat 8vo, Treadwell's Qualitative Analysis. (Hall.) 8vo, Quantitative Analysis. (Hall.) 8vo, Turneaure and Russell's Public Water-supplies 8vo, Van Deventer's Physical Chemistry for Beginners. (Boltwood.) 12mo, Venable's Methods and Devices for Bacterial Treatment of Sewage 8vo, Ward and Whipple's Freshwater Biology. (In Press.) Ware's Beet-sugar Manufacture and Refining. Vol. 1 8vo, Vol.11 8vo,
2 3 3 2 3 3
1
4 5
1
4 00 5 00
Washington's Manual of the Chemical Analysis of Rocks * Weaver's Military Explosives Wells's Laboratory Guide in Qualitative Chemical Analysis
8vo, $2 00 g vo 3 00 8vo! 1 50
'
Short Course in Inorganic Qualitative Chemical Analysis for Engineering Students 12mo, Text-book of Chemical Arithmetic 12mo, Whipple's Microscopy of Drinking-water 8vo, Wilson's Chlorination Process 12mo,
1
1
50 25 50 50
3 50
1 1
Cyanide Processes Winton's Microscopy of Vegetables Food Zsigmondy's Colloids and the Ultramicroscope.
12mo,
8vo,
(Alexander)... Large 12mo,
7 50 3 00
CIVIL ENGINEERING.
BRIDGES AND ROOFS.

ING.
HYDRAULICS.
MATERIALS OF ENGINEER-
RAILWAY ENGINEERING.
3 00
Baker's Engineers' Surveying Instruments 12mo, Bixby's Graphical Computing Table Paper 19i X24J inches. Breed and Hosmer's Principles and Practice of Surveying. Vol. I. Elementary Surveying 8vo, Higher Surveying Vol. II. 8vo, * Burr's Ancient and Modern Engineering and the Isthmian Canal 8vo, Comstock's Field Astronomy for Engineers 8vo, * Corthell's Allowable Pressure on Deep Foundations 12mo, Crandall's Text-book on Geodesy and Least Squares 8vo, Davis's Elevation and Stadia Tables 8vo, Elliott's Engineering for Land Drainage 12mo, Practical Farm Drainage. (Second Edition Rewritten.) 12mo, * Fiebeger's Treatise on Civil Engineering 8vo, Flemer's Photographic Methods and Instruments 8vo, Folwell's Sewerage. (Designing and Maintenance.) 8vo, Freitag's Architectural Engineering 8vo, Goodhue's Municipal Improvements 12mo, * Hauch and Rice's Tables of Quantities for Preliminary Estimates. 12mo, 8vo, Hayford's Text-book of Geodetic Astronomy 16mo, mor. Hering's Ready Reference Tables (Conversion Factors) 16mo, mor. Hosmer's Azimuth 12mo, Howe' Retaining Walls for Earth * Ives's Adjustments of the Engineer's Transit and Level 16mo, bds. Large 12mo, Johnson's (J. B.) Theory and Practice of Surveying 8vo, Johnson's (L. J.) Statics by Algebraic and Graphic Methods (In Preparation). Kinnicutt, Winslow and Pratt's Purification of Sewage. * Mahan's Descriptive Geometry 8vo,
.
25
3 2 3 2
1
3
1 1 1
5 5 3 3
1 1
3 2
1 1
00 50 50 50 25 00 00 50 50 00 00 00 50 50 25 00 50 00
25 25 4 00 2 00
1
50
Merriman's Elements of Precise Surveying and Geodesy Merriman and Brooks's Handbook for Surveyors Nugent's Plane Surveying Ogden's Sewer Construction Sewer Design Parsons's Disposal of Municipal Refuse Patton's Treatise on Civil Engineering Reed's Topographical Drawing and Sketching
Rideal's
8vo,
16mo, mor.
8vo, 8vo,
2 50 2 00
3 50
12mo,
8vo, 8vo, half leather,
4to, 8vo, Peele.).8vo, 8vo, 8vo,
3 2 2 7 5
Sewage and the Bacterial Purification
of
Sewage
4
3
1
Riemer's Shaft-sinking under Difficult Conditions. (Corning and Siebert and Biggin's Modern Stone-cutting and Masonry (McMillan.) Smith's Manual of Topographical Drawing. Soper's Air .and Ventilation of Subways * Tracy's Exercises in Surveying Tracy's Plane Surveying * Trautwine's Civil Engineer's Pocket-book Venable's Garbage Crematories in America Methods and Devices for Bacterial Treatment of Sewage
12mo, 12mo, mor. 16mo, mor. 16mo, mor.

8vo, 8vo,
2 2
1
3 5 2 3 00
00 00 00 50 00 00 00 50 50 50 00 00 00 00
Wait's Engineering and Architectural Jurisprudence
Law Law
8vo, of Contracts of Operations Preliminary to Construction in Engineering and
8vo, $6 00 Sheep, 6 50 3 00
Architecture
8vo, Warren's Stereotomy Problems in Stone-cutting * Waterbury's Vest-Pocket Hand-book of Mathematics for Engineers. 2|X5| inches, mor. Webb's Problem's in the Use and Adjustment of Engineering Instruments. 16mo, mor. 8vo, Wilson's Topographic Surveying
8vo, Sheep,
5 00 5 50 2 50
1
1
00
25
3 50
BRIDGES AND ROOFS.

Boiler's Practical Treatise on the Construction of Iron * Thames River Bridge
Highway
Bridges.. 8vo,
Oblong paper,
8vo, 8vo,
Burr and Falk's Design and Construction of Metallic Bridges Influence Lines for Bridge and Roof Computations
Du Bois's Mechanics of Engineering. Vol. II Small 4to, Foster's Treatise on Wooden Trestle Bridges 4to, Fowler's Ordinary Fopndations 8vo, Greene's Arches in Wood, Iron, and Stone 8vo, Bridge Trusses 8vo, Roof Trusses 8vo, Grimm's Secondary Stresses in Bridge Trusses 8vo, Heller's Stresses in Structures and the Accompanying Deformations.. ,8vo, Howe's Design of Simple Roof- trusses Wood and Steel 8vo.
.
Symmetrical Masonry Arches Treatise on Arches Johnson, Bryan and Turneaure's Theory and Practice
8vo, 8vo,
in the
2 00 5 00 5 00 3 00 10 00 5 00 3 50 2 50 2 50 1 25 2 50 3 00 2 00 2 50 4 00
Modern Framed Structures Merriman and Jacoby's Text-book on Roofs and Bridges:
Designing of Small 4to, 10 00
Part I. Stresses in Simple Trusses 8vo, 2 50 Part II. Graphic Statics 8vo, 2 50 Part III. Bridge Design 8vo, 2 50 Part IV. Higher Structures 8vo, 2 50 Morison's Memphis Bridge Oblong 4to, 10 00 Sondericker's Graphic Statics, with Applications to Trusses, Beams, and Arches 8vo, 2 00 Waddell's De Pontibus, Pocket-book for Bridge Engineers 16mo, mor. 2 00 * Specifications for Steel Bridges 12mo, 50 Waddell and Harringtoon's Bridge Engineering. (In Preparation.) Draw-spans. parts Two in one volume Wright's Designing of 8vo, 3 50
HYDRAULICS.
Barnes's Ice Formation Bazin's Experiments upon the Contraction of the Liquid Vein Issuing an Orifice. (Trautwine. ) Bovey 's Treatise on Hydraulics Church's Diagrams of Mean Velocity of Water in Open Channels.
8vo,
3 00
2 00 5 00
1
from
8vo, 8vo,
Oblong
Hydraulic Motors Coffin's Graphical Solution of Hydraulic Problems
Flather's Dynamometers, and the Measurement of Folwell's Water-supply Engineering
Frizell's
4to, paper,
50
8vo,
2 00
Power
16mo, mor. 12mo,

8vo, 8vo,
Water-power
Fuertes's Water and Public Health Water-filtration Works Ganguillet and Kutter's General Formula for the Uniform Flow of Rivers and Other Channels. (Hering and Trautwine.)
12mo, 12mo,
2 50 3 00 4 00 5 00 1 50 2 50 4 00
Water
in
8vo,
Large 12mo, $1 50 8vo, 3 00 8vo, 2 50 the Carrying Capacity of Large, Riveted, Experiments on Metal Herschel's 115 8vo, Conduits 2 00 8vo, 2 00 Hoyt and Grover's River Discharge Hubbard and Kiersted's Water-works Management and Maintenance. 8vo, 4 00 * Lyndon's Development and Electrical Distribution of Water Power. 8vo, 3 00 Mason's Water-supply. (Considered Principally from a Sanitary Standpoint.) 8vo, 4 00 8vo, Merriman's Treatise on Hydraulics 5 00 * Molitor's Hydraulics of Rivers, Weirs and Sluices 8vo, 2 00 * Richards's Laboratory Notes on Industrial Water Analysis 8vo, 50 Schuyler's Reservoirs for Irrigation, Water-power, and Domestic Watersupply. Second Edition, Revised and Enlarged Large 8vo, 6 00 * Thomas and Watt's Improvement of Rivers 4to, 6 00 Turneaure and Russell's Public Water-supplies 8vo, 5 00 Wegmann's Design and Construction of Dams. 5th Ed., enlarged 4to, 6 00 Water-Supply of the City of New York from 1658 to 1895 4to, 10 00 Whipple's Value of Pure Water Large 12mo, 1 00 Williams and Hazen's Hydraulic Tables 8vo, 1 50 Wilson's Irrigation Engineering 8vo, 4 00 Wood's Turbines 8vo, 2 50
Hazen's Clean Water and
How to Get It Filtration of Public Water-supplies Hazelhurst's Towers and Tanks for Water-works
MATERIALS OF ENGINEERING.
Baker's Roads and Pavements 8vo, Treatise on Masonry Construction 8vo, Black's United States Public Works Oblong 4to, (In Press.) Blanchard's Bituminous Roads. Bleininger's Manufacture of Hydraulic Cement. (In Preparation.) * Bovey 's Strength of Materials and Theory of Structures 8vo, Burr's Elasticity and Resistance of the Materials of Engineering 8vo, Construction Highway Byrne's 8vo, Inspection of the Materials and Workmanship Employed in Construction.
5 00 5 00 5 00
7 50 7 50 5 00 3 00 6 00
16mo,
Church's Mechanics of Engineering Du Bois's Mechanics of Engineering. Vol. I. Kinematics, Statics, Kinetics
Vol. II.
8vo,
Small 4to,
7 50
The
Theory Small 4to, 10 00 * Eckel's Cements, Limes, and Plasters 8vo, 6 00 Stone and Clay Products used in Engineering. (In Preparation.) Fowler's Ordinary Foundations 8vo, 3 50 * Greene's Structural Mechanics 8vo, 2 50 * Holley's Lead and Zinc Pigments Large 12mo, 3 00 Holley and Ladd's Analysis of Mixed Paints, Color Pigments and Varnishes. Large 12mo, 2 50 Johnson's (C. M.) Rapid Methods for the Chemical Analysis of Special Steels, Steel-making Alloys and Graphite Large 12mo, 3 00 Johnson's (J. B.) Materials of Construction Large 8vo, 6 00 Keep's Cast Iron 8vo, 2 50 Lanza's Applied Mechanics 8vo, 7 50 Maire's Modern Pigments and their Vehicles 12mo, 2 00 Martens's Handbook on Testing Materials. (Henning.) 2 vols. ...... 8vo, 7 50 Maurer's Technical Mechanics 8vo, 4 00 Merrill's Stones for Building and Decoration 8vo, 5 00 Merriman's Mechanics of Materials 8vo, 5 00 * Strength of Materials 12mo, 1 00 Metcalf 's Steel. A Manual for Steel-users 12mo, 2 00 Morrison's Highway Engineering 8vo, 2 50 Patton's Practical Treatise on Foundations 8vo, 5 00 Rice's Concrete Block Manufacture 8vo, 2 00
Stresses in of Flexures
Framed
Structures, Strength of Materials
and
8vo, $3 00 Richardson's Modern Asphalt Pavements 5 00 Richey's Building Foreman's Pocket Book and Ready Reference. 16mo,mor. * Cement Workers' and Plasterers' Edition (Building Mechanics' Ready Reference Series) 16mo, mor. 1 50 Handbook for Superintendents of Construction 16mo, mor. 4 00 * Stone and Brick Masons' Edition (Building Mechanics' Ready Reference Series) 16mo, mor. 1 50 * Ries's Clays Their Occurrence, Properties, and Uses 8vo, 5 00 * Ries and Leighton's History of the Clay-working Industry of the United States 8vo. 2 50 8vo, Sabin's Industrial and Artistic Technology of Paint and Varnish 3 00 Smith's Strength of Material 12mo, 8vo, Snow's Principal Species of Wood 3 50 12mo, 2 00 Spalding's Hydraulic Cement Text-book on Roads and Pavements 12mo, 2 00 8vo, Taylor and Thompson's Treatise on Concrete, Plain and Reinforced 5 00 In Three Parts. Thurston's Materials of Engineering. 8vo, 8 00 Part I. Non-metallic Materials of Engineering and Metallurgy.. .8vo, 2 00 Part II. Iron and Steel 8vo, 3 50 Part III. A Treatise on Brasses, Bronzes, and Other Alloys and their Constituents 8vo, 2 50 8vo, Tillson's Street Pavements and Paving Materials 4 00 Turneaure and Maurer's Principles of Reinforced Concrete Construction. Second Edition, Revised and Enlarged 8vo, 3 50 Waterbury's Cement Laboratory Manual 12mo, 1 00 Wood's (De V.) Treatise on the Resistance of Materials, and an Appendix on the Preservation of Timber 8vo, 2 00 Wood's (M. P.) Rustless Coatings: Corrosion and Electrolysis of Iron and Steel 8vo, 4 00
:
RAILWAY ENGINEERING.
Andrews's Handbook for Street Railway Engineers
Berg's Buildings and Structures of American Railroads Brooks's Handbook of Street Railroad Location Butts's Civil Engineer's Field-book Crandall's Railway and Other Earthwork Tables Transition Curve * Crockett's Methods for Earthwork Computations Dredge's History of the Pennsylvania Railroad. (1879)
Fisher's Table of Cubic Yards
3X5
inches, mor. 4to,
25
16mo, mor. 16mo, mor.

8vo,
16mo, mor.
8vo,
Papei Cardboard, Godwin's Railroad Engineers' Field-book and Explorers' Guide. 16mo, mor. Hudson's Tables for Calculating the Cubic Contents of Excavations and Em.
5 00 1 50 2 50 1 50 1 50 * 50 5 00 25 2 50
1
1
bankments
Ives and Hilts's Problems in Surveying, Railroad Surveying and Geodesy
8vo,
00
Molitor and Beard's Manual for Resident Engineers Nagle's Field Manual for Railroad Engineers * Orrock's Railroad Structures and Estimates 8vo, Philbrick's Field Manual for Engineers 16mo, mor. Raymond's Railroad Engineering. 3 volumes. Vol. (In Preparation.) I. Railroad Field Geometry. Vol. II. Elements of Railroad Engineering 8vo, (In Preparation.) Vol. III. Railroad Engineer's Field Book. Searles's Field Engineering 16mo, mor. Railroad Spiral 16mo, mor. Taylor's Prismoidal Formulse and Earthwork 8vo, * Trautwine's Field Practice of Laying Out Circular Curves for Railroads.
16mo, mor. 16mo, 16mo, mor.
50 00 3 00 3 00 3 00
1
3 50
3 00
1
50 50
12mo, mor.
2 SO
2 00 2 50 5 00 5 00 2 00
Method of Calculating the Cubic Contents bankments by the Aid of Diagrams Webb's Economics of Railroad Construction
of Excavations
and Em8vo,
Railroad Construction Wellington's Economic Theory of the Location of Railways Wilson's Elements of Railroad-Track and Construction
Large 12mo, 16mo, mor. Large 12mo, 12mo,
DRAWING.
Barr's Kinematics of Machinery * Bartl&tt's Mechanical Drawing
8vo, $2 50 8vo, 3 00 8vo, 1 50 8vo, paper, 1 00
2 50
*
Coolidge's
'
"
"
Abridged
Ed
Manual
of
Drawing
Coolidge and Freeman's Elements of General Drafting for Mechanical Engineers Oblong 4to, Durley's Kinematics of Machines 8vo, Emch's Introduction to Projective Geometry and its Application 8vo, French and Ives' Stereotomy 8vo, Hill's Text-book on Shades and Shadows, and Perspective 8vo, Jamison's Advanced Mechanical Drawing Svo, Elements of Mechanical Drawing 8vo, Jones's Machine Design: Part I. Kinematics of Machinery 8vo, Part II. Form, Strength, and Proportions of Parts 8vo,
4 2 2 2 2
00 50 50 00 00 2 50
1
50
3 00
Kimball and Barr's Machine Design. (In Press.) MacCord's Elements of Descritpive Geometry
Kinematics; or. Practical Mechanism Mechanical Drawing
Velocity Diagrams McLeod's Descriptive Geometry * Mahan's Descriptive Geometry and Stone-cutting
Industrial Drawing.
8vo, 8vo,
4to,
3 00 5 00
8vo,
Large 12mo,
8vo, 8vo, Svo,
(Thompson.)
Moyer's Descriptive Geometry Reed's Topographical Drawing and Sketching 4to, Reid's Course in Mechanical Drawing 8vo, Text-book of Mechanical Drawing and Elementary Machine Design.. 8vo, Robinson's Principles of Mechanism 8vo,
Schwamb and
Smith
Merrill's
Elements of Mechanism
8vo,
(A. W.) and Marx's Machine Design 8vo, Smith's (R. S.) Manual of Topographical Drawing. (McMillan) 8vo, * Titsworth's Elements of Mechanical Drawing Oblong Svo, Warren's Drafting Instruments and Operations 12mo, Elements of Descriptive Geometry, Shadows, and Perspective 8vo, Elements of Machine Construction and Drawing 8vo, Elements of Plane and Solid Free-hand Geometrical Drawing. 12mo, General Problems of Shades and Shadows 8vo, Manual of Elementary Problems in the Linear Perspective of Forms and Shadow 12mo,
.
4 00 1 50 1 50 1 50 3 50 2 00 5 00 2 00 3 00 3 00 3 00 3 00 2 50 1 25 1 25 3 50 7 50 1 00 3 00
Elementary Projection Drawing 12mo, Plane ProbVems in Elementary Geometry 12mo, 8vo, Problems, Theorems, and Examples in Descriptive Geometry (Hermann and Weisbach's Kinematics and Power of Transmission.
Manual
of
00 50 1 25 2 50
1
1
Klein. )
Wilson's (H. M.) Topographic Surveying * Wilson's (V. T.) Descriptive Geometry Free-hand Lettering Free-hand Perspective Woolf's Elementary Course in Descriptive Geometry
8vo, 8vo, 8vo, 8vo, 8vo, Large 8vo,
5 00 3 50 1 50 1 00 2 50 3 00
ELECTRICITY AND PHYSICS.

* Abegg's Theory of Electrolytic Dissociation, (von Ende.) 12mo, Andrews's Hand-book for Street Railway Engineering 3X5 inches, mor. Physics. (Magie.) .Large ... 12mo, Anthony and Brackett's Text-book of Anthony and Ball's Lecture-notes on the Theory of Electrical Measure12mo, ments 8vo, Benjamin's History of Electricity
Voltaic Cell
25 25 3 00
1 1 1
00
8vo,
3 00 3 00
10
Betts's Lead Refining and Electrolysis Classen's Quantitative Chemical Analysis by Electrolysis. * Collins's Manual of Wireless Telegraphy and Telephony
8vo, $4 00 (Boltwood.).8vo, 3 00
12mo,
* Mor.
50
Crehore and Squier's Polarizing Photo-chronograph * Danneel's Electrochemistry. (Merriam.) Dawson's "Engineering" and Electric Traction Pocket-book. Dolezalek's Theory of the Lead Accumulator (Storage Battery),
.
8vo,
.
.
2 00 3 00
1
12mo, 16mo, mor. (von Ende.) 12mo,

8vo,
25
5 00
2 50 4 00 3 00
Flather's
Duhem's Thermodynamics and Chemistry. (Burgess.) Dynamometers, and the Measurement of Power
Getman's Introduction
to Physical Science (Mottelay) Gilbert's De Magnete. * Hanchett's Alternating Currents
12mo, 12mo,
8vo,
2 50
1
Hering's
*
Ready Reference Tables (Conversion
Factors)
12mo, 16mo, mor.
00
8vo, Hobart and Ellis's High-speed Dynamo Electric Machinery 8vo, Holman's Precision of Measurements Telescopic Mirror-scale Method, Adjustments, and Tests.. .Large 8vo, * Karapetoff 's Experimental Electrical Engineering 8vo, 6 8vo, Kinzbrunner's Testing of Continuous-current Machines 2 (Tingle.) Landauer's Spectrum Analysis. 8vo, 3 Le Chatelier's High-temperature Measurements. (Boudouard Burgess. )12mo, 3 (Lorenz) 8vo, 3 Lob's Electrochemistry of Organic Compounds. * Lyndon's Development and Electrical Distribution of Water Power. .8vo, 3 * Lyons's Treatise on Electromagnetic Phenomena. Vols, I .and II. 8vo, each, 6 * Michie's Elements of Wave Motion Relating to Sound and Light 8vo, 4 12mo, 1 Morgan's Outline of the Theory of Solution and its Results * Physical Chemistry for Electrical Engineers 12mo, 1 * Norris's Introduction to the Study of Electrical Engineering 8vo, 2 Norris and Denmson's Course of Problems on the Electrical Characteristics of (In Press.) Circuits and Machines. * Parshall and Hobart's Electric Machine Design 4to, half mor, 12 Reagan's Locomotives: Simple, Compound, and Electric. New Edition. Large 12mo, 3 * Rosenberg's Electrical Engineering. (Haldane Gee Kinzbrunner.) .8vo, 2 8vo, Ryan, Norris, and Hoxie's Electrical Machinery. Vol. I 2 12mo, Schapper's Laboratory Guide for Students in Physical Chemistry 1 * Tillman's Elementary Lessons in Heat 8vo, 1 Large 12mo, 2 Tory and Pitcher's Manual of Laboratory Physics Ulke's Modern Electrolytic Copper Refining 8vo, 3
.
2 50 6 00 2 00
75 00 00 00 00 00 00 00 00 00 50 50
50
50 00 50 00 50 00 00
LAW.
* Brennan's
Hand-book
of Useful Legal Information for Business Men.
16mo, mor.
* Davis's Elements of Law ., 8vo, * Treatise on the Military Law of United States 8vo, * Dudley's Military Law and the Procedure of Courts-martial. Large 12mo, Courts-martial 16mo, mor. Manual for 8vo, Wait's Engineering and Architectural Jurisprudence Sheep, 8vo Law of Contracts Law of Operations Preliminary to Construction in Engineering and Architecture 8vo, Skeep,
.
5 00 2 50
7 00
2 50
1
50
6 00 6 50 3 00
5 00 5 50
MATHEMATICS.
Baker's Elliptic Functions (Bocher) Briggs's Elements of Plane Analytic Geometry. * Buchanan's Plane and Spherical Trigonometry
8vo,
1
12mo,
8vo,
1 1
50 00 00
11
Byerley's Harmonic Functions Chandler's Elements of the Infinitesimal Calculus * Coffin's Vector Analysis
gvo, $1 00
Compton's Manual
of Logarithmic
Computations
* Dickson's College Algebra * Introduction to the Theory of Algebraic Equations Emch's Introduction to Projective Geometry and its Application Fiske's Functions of a Complex Variable Halsted's Elementary Synthetic Geometry
12mo, 12mo, 12mo, Large 12mo, Large 12mo,

8vo, 8vo, 8vo, 8vo,
2 00 2 50
Elements
of
Geometry
* Rational Geometry
12mo,
:
Hyde's Grassmann's Space Analysis Johnson's (J. B.) Three-place Logarithmic Tables
*
Vest-pocket
size,
8vo, paper,
inches,
50 50 1 25 2 50 1 00 1 50 1 75 1 50 1 00
1
100 copies,
Mounted on heavy
cardboard,. 8
X 10
10 copies, Johnson's (W. W.) Abridged Editions of Differential and Integral Calculus. Large 12mo, 1 vol. Curve Tracing in Cartesian Co-ordinates 12mo, Differential Equations 8vo, Elementary Treatise on Differential Calculus Large 12mo, Elementary Treatise on the Integral Calculus Large 12mo, * Theoretical Mechanics 12mo, Theory of Errors and the Method of Least Squares 12mo, Treatise on Differential Calculus Large 12mo, Treatise on the Integral Calculus Large 12mo, Treatise on Ordinary and Partial Differential Equations. .Large 12mo, Karapetoff's Engineering Applications of Higher Mathematics.
.
15 5 00 25 2 00
2 50
1
1 1
1
3
1
3 3 3
00 00 50 50 00 50 00 00 50
(In Preparation.) Laplace's Philosophical Essay on Probabilities. (Truscott and Emory.) 12mo, * Ludlow and Bass's Elements of Trigonometry and Logarithmic and Other Tables 8vo, * Trigonometry and Tables published separately Each, * Ludlow's Logarithmic and Trigonometric Tables 8vo, Macfarlane's Vector Analysis and Quaternions 8vo, McMahon's Hyperbolic Functions 8vo, Manning's Irrational Numbers and their Representation by Sequences and
.
2 00
3 00 2 00
1 1
1 1 1
00 00 00
Mathematical Monographs.
S.
12mo, Edited by Mansfield Merriman and Robert Octavo, each No. 1. History of Modern Mathematics, by David Eugene Smith. No. 2. Synthetic Projective Geometry, by George Bruce Halsted. No. 3. Determinants, by Laenas Gifford Weld. No. 4. Hyperbolic Functions, by James McMahon. No. 5. Harmonic Functions, by William E. Byerly. No. 6. Grassmann's Space Analysis, Edward Hyde. Probability by W. No. 7. and Theory of Errors, by Robert S. Woodward. No. 8. Vector Analysis and Quaternions, by Alexander Macfarlane. No. 9. Differential Equations, by William Woolsey Johnson. No. 10. The Solution of Equations, by Mansfield Merriman. No. 11. Functions of a Complex Variable,
Series
25 00
Woodward
by Thomas
S.
Fiske.
Maurer's Technical Mechanics Merriman's Method of Least Squares Solution of Equations Rice and Johnson's Differential and Integral Calculus.
8vo, 8vo, 8vo,
4 00
2 00 1 00
1
1
vols, in one.
Large 12mo,
50
Large 12mo, Elementary Treatise on the Differential Calculus 8vo, Smith's History of Modern Mathematics * Veblen and Lennes's Introduction to the Real Infinitesimal Analysis of One Variable 8vo, * Waterbury's Vest Pocket Hand-book of Mathematics for Engineers. inches, mor. 2JX5| 8vo, Weld's Determinants 8vo, Wood's Elements of Co-ordinate Geometry 8vo. Woodward's Probability and Theory of Errors
3 00
00
2 00
00 00 2 00 1 00
1
1
12
MECHANICAL ENGINEERING.
MATERIALS OF ENGINEERING, STEAM-ENGINES AND BOILERS.
Bacon's Forge Practice Baldwin's Steam Heating for Buildings Barr's Kinematics of Machinery * Bartlett's Mechanical Drawing " " " * Abridged Ed * Burr's Ancient and Modern Engineering and the Isthmian Canal Experimental Engineering Carpenter's Heating and Ventilating Buildings (New edition in press.) Clerk's Gas and Oil Engine. Compton's First Lessons in Metal Working Compton and De Groodt's Speed Lathe
12mo, $1 50 12mo, 2 50 8vo, 2 50 8vo, 3 00

8vo, 8vo, 8vo, 8vo,
150
3 50 6 00 4 00
1 1 1
12mo, 12mo,
8vo, paper, Coolidge's Manual of Drawing Coolidge and Freeman's Elements of Geenral Drafting for Mechanical EnOblong 4to, gineers 12mo, Cromwell's Treatise on Belts and Pulleys 12mo, Treatise on Toothed Gearing 12mo, Dingey's Machinery Pattern Making Durley's Kinematics of Machines 8vo, Gear-cutting Machinery Large Flanders's 12mo, Flather's Dynamometers and the Measurement of Power 12mo,
50 50 00
2 50
1
1
50 50
2 00
4 00
3 00 3 00 2 00
1 25 2 00
Rope Driving
Gas and Fuel Analysis Goss's Locomotive Sparks
Gill's
for Engineers
12mo, 12mo,
8vo,
(In Preparation.) Greene's Pumping Machinery. Hering's Ready Reference Tables (Conversion Factors) * Hobart and Ellis's High Speed Dynamo Electric Machinery
16mo, mor.
8vo, 8vo, 8vo, 8vo, 8vo,
Hutton's Gas Engine

Jamison's Advanced Mechanical Drawing Elements of Mechanical Drawing
Jones's Gas Engine
50 00 00 00 50 4 00
2 6 5 2 2
Machine Design: Kinematics of Machinery Part I. 8vo, 1 50 Part II. Form, Strength, and Proportions of Parts 8vo, 3 00 16mo, mor. 5 00 Kent's Mechanical Engineer's Pocket-Book 8vo, Kerr's Power and Power Transmission 2 00 Kimball and Barr's Machine Design. (In Press.) (In Press.) Levin's Gas Engine. 8vo, Leonard's Machine Shop Tools and Methods. 8vo, 4 00 * Lorenz's Modern Refrigerating Machinery. (Pope, Haven, and Dean).. 8vo, 4 00 MacCord's Kinematics; or, Practical Mechanism 8vo, 5 00 Mechanical Drawing 4to, 4 00 Velocity Diagrams 8vo, 1 50 MacFarland's Standard Reduction Factors for Gases 8vo, 1 50 Mahan's Industrial Drawing. (Thompson.) 8vo, 3 50 Mehrtens's Gas Engine Theory and Design Large 12mo, 2 50 Oberg's Handbook of Small Tools Large 12mo, 3 00 * Parshall and Hobart's Electric Machine Design. Small 4to, half leather, 12 50 Peele's Compressed Air Plant for Mines 8vo, 3 00 Poole's Calorific Power of Fuels 8vo, 3 00 * Porter's Engineering Reminiscences, 1855 to 1882 8vo, 3 00 Reid's Course in Mechanical Drawing 8vo, 2 00 Text-book of Mechanical Drawing and Elementary Machine Design. 8vo, 3 00 Richards's Compressed Air 12mo, 1 50 Robinson's Principles of Mechanism 8vo, 3 00 Schwamb and Merrill's Elements of Mechanism 8vo, 3 00 Smith (A. W.) and Marx's Machine Design 8vo, 3 00 Smith's (O.) Press-working of Metals 8vo, 3 00 (Woodward and Sorel's Carbureting and Combustion in Alcohol Engines. Large 12mo, 3 00 Preston.) 8vo, 3 50 Stone's Practical Testing of Gas and Gas Meters 13
.
Thurston's Animal as a Machine and Prime Motor, and the Laws of Energetics.
Treatise on Friction and Lost Work in Machinery * Tillson's Complete Automobile Instructor * Titsworth's Elements of Mechanical Drawing
and
Mill
Work.
Warren's Elements of Machine Construction and Drawing * Waterbury's Vest Pocket Hand-book of Mathematics for Engineers.
12mo, $1 00 .8vo, 3 00 16mo, 1 50 Oblong 8vo, 1 25 8vo, 7 50

.
2JX5|
Weisbach's Kinematics and the Power of Transmission.
Klein.) Machinery of Transmission
inches, mor.
00
(Herrmann
and Governors.
(Hermann
Wood's Turbines
8vo, Klein.). .8vo, 8vo,
5 00 5 00 2 50
MATERIALS OF ENGINEERING.
Bovey's Strength of Materials and Theory of Structures Burr's Elasticity and Resistance of the Materials of Engineering
*
Church's Mechanics of Engineering * Greene's Structural Mechanics * Holley's Lead and Zinc Pigments Large 12mo Holley and Ladd's Analysis of Mixed Paints, Color Pigments, and Varnishes. Large 12mo, Johnson's (C. M.) Rapid Methods for the Chemical Analysis of Special Steels, Steel-Making Alloys and Graphite Large 12mo, Johnson's (J. B.) Materials of Construction 8vo, Keep's Cast Iron 8vo, Lanza's Applied Mechanics 8vo, Maire's Modern Pigments and their Vehicles 12mo, (Henning.) Martens's Handbook on Testing Materials. 8vo Maurer's Techincal Mechanics 8vo, Merriman's Mechanics of Materials 8vo, * Strength of Materials 12mo, A Manual for Steel-users Metcalf's Steel. 12mo, 8vo, Sabin's Industrial and Artistic Technology of Paint and Varnish Smith's ((A. W.) Materials of Machines 12mo, Smith's (H. E.) Strength of Material 12mo. Thurston's Materials of Engineering 3 vols., 8vo, Non-metallic Materials of Engineering 8vo, Part I. 8vo, Part II. Iron and Steel Part III. A Treatise on Brasses, Bronzes, and Other Alloys and their 8vo, Constituents 8vo, Wood's (De V.) Elements of Analytical Mechanics Treatise on the Resistance of Materials and an Appendix on the Preservation of Timber 8vo, Wood's (M. P.) Rustless Coatings: Corrosion and Electrolysis of Iron and
;
8vo, 8vo, 8vo, 8vo,
50 50 6 00 2 50 3 00
7
7
2 50
3 00 6 00 2 50
7 50 2 00 7 50 4 00 5 00
1
00
2 00 3 00
1
00
8 00 2 00 3 50
2 50 3 00
2 00
Steel
8vo,
4 00
STEAM-ENGINES AND BOILERS.

12mo, Berry's Temperature-entropy Diagram (Thurston.) 12mo, Carnot's Reflections on the Motive Power of Heat. 12mo, Chase's Art of Pattern Making Creighton's Steam-engine and other Heat Motors 8vo, Dawson's "Engineering" and Electric Traction Pocket-book. .. 16mo, mor. 18mo, Ford's Boiler Making for Boiler Makers * Gebhardt's Steam Power Plant Engineering 8vo, 8vo, Goss's Locomotive Performance 12mo, Hemenway's Indicator Practice and Steam-engine Economy 8vo, Hutton's Heat and Heat-engines .8vo, Mechanical Engineering of Power Plants 8vo, Kent's Steam boiler Economy
.
2 00 1 50 2 50 5 00 5 00 1 00 6 00 5 00 2 00 5 00 5 00 4 00
14
iss's Practice and Theory of the Injector 8vo* MacCord's Slide-valves g vo Meyer's Modern Locomotive Construction 4to, Moyer's Steam Turbine 8vo, Peabody's Manual of the Steam-engine Indicator 12mo, Tables of the Properties of Steam and Other Vapors and TemperatureEntropy Table 8vo, Thermodynamics of the Steam-engine and Other Heat-engines. 8vo, Valve-gears for Steam-engines 8vo, Peabody and Miller's Steam-boilers 8vo, Pupin's Thermodynamics of Reversible Cycles in Gases and Saturated Vapors. (Osterberg.) 12mo. Reagan's Locomotives: Simple, Compound, and Electric. New Edition. Large 12mo, Sinclair's Locomotive Engine Running and Management 12mo, Smart's Handbook of Engineering Laboratory Practice 12mo, Snow's Steam-boiler Practice 8vo, Spangler's Notes on Thermodynamics 12mo,
. .
.
$1 50 2 00 10 00
4 00 1 50
1
00
5 00 2 50 4 00
1
25
Valve-gears Spangler, Greene, and Marshall's Elements of Steam-engineering
8vo,
8vo
Thomas's Steam-turbines
Thurston's
Handbook
Tables
of
cator and the
Engine and Boiler Prony Brake
Trials,
and the Use
Handy
Manual Manual
8vo, of the Indi8vo, 8vo,

2vols., 8vo.
50 00 50 00 1 00 2 50 3 00 4 00
3 2 2 3
5 00
1
of Steam-boilers, their Designs, Construction, of the Steam-engine
and Operation 8vo,
(Patterson). 8vo. Weisbach's Heat, Steam, and Steam-engines. (Du Bois.) 8vo, Whitham's Steam-engine Design 8vo, Wood's Thermodynamics, Heat Motors, and Refrigerating Machines. . 8vo,
.
Part I. History, Structure, and Theory Part II. Design, Construction, and Operation Steam-boiler Explosions in Theory and in Practice Wehrenfenning's Analysis and Softening of Boiler Feed-water.
5 10 8vo, 6 8vo, 6
1
12mo,
50 00 00 00 00 50
4 00 5 00 5 00 4 00
MECHANICS PURE AND APPLIED.

Church's Mechanics of Engineering 8vo. Notes and Examples in Mechanics 8vo. Dana's Text-book of Elementary Mechanics for Colleges and Schools .12mo, Du Bois's Elementary Principles of Mechanics: Vol. I. Kinematics 8vo.
Vol. II.
Statics.
6 00 2 00 1 50
Mechanics of Engineering.
Vol. I Vol. II
* Greene's Structural Mechanics James's Kinematics of a Point and the Rational Mechanics of a Particle.
* Johnson's (W. W.) Theoretical Mechanics
3 50 8vo, 4 00 Small 4to, 7 50 Small 4to! 10 00 8vo, 2 50

2 00 3 00 7 50
1
1
Large 12mo, 12mo,

8vo,
Lanza's Applied Mechanics * Martin's Text Book on Mechanics. Voi.
I, Statics * Vol. II, Kinematics
12mo,
and Kinetics. 12mo.

8vo,
25 50
Maurer's Technical Mechanics * Merriman's Elements of Mechanics Mechanics of Materials * Michie's Elements of Analytical Mechanics Robinson's Principles of Mechanism Sanborn's Mechanics Problems
4 00
1 00 5 00 4 00 3 00 1 50 3 00 3 00 1 25
12mo,
8vo, 8vo, 8vo,
Schwamb and
Merrill's
Wood's Elements
Principles of
Elements of Mechanism of Analytical Mechanics
Large 12mo, 8vo

8vo,
Elementary Mechanics
12mo,
15
MEDICAL.
(Hall and in Thirty Lectures. 8vo, $5 00 Defren.) 12mo, 1 00 von Behring's Suppression of Tuberculosis. (Bolduan.) 12mo, 1 50 Bolduan's Immune Sera 8vo, (In Press.) (Gay). Bordet's Studies in Immunity. Davenport's Statistical Methods with Special Reference to Biological Varia16mo, mor. 1 50 tions 8vo, 6 00 (Bolduan.) Ehrlich's Collected Studies on Immunity. * Fischer's Physiology of Alimentation Large 12mo, 2 00 2 50 (Rosanoff and Collins.).. .Large 12mo, de Fursac's Manual of Psychiatry. 8vo, 4 00 Hammarsten's Text-book on Physiological Chemistry. (Mandel.) 1 25 Jackson's Directions for Laboratory Work in Physiological Chemistry .8vo, 12mo, 1 00 (Lorenz.) Lassar-Cohn's Practical Urinary Analysis. 12mo, 1 50 Mandel's Hand-book for the Bio-Chemical Laboratory * Pauli's Physical Chemistry in the Service of Medicine. (Fischer.) ..12mo, 1 25 * Pozzi-Escot's Toxins and Venoms and their Antibodies. (Cohn.). 12mo, 1 00 12mo, 1 00 Rostoski's Serum Diagnosis. (Bolduan.) 8vo, 2 00 Ruddiman's Incompatibilities in Prescriptions
* Abderhalden's Physiological Chemistry
Whys
in
Pharmacy
12mo,
.
00
Salkowski's Physiological and Pathological Chemistry. (Orndorff.) .. .8vo, * Satter lee's Outlines of Human Embryology 12mo, Smith's Lecture Notes on Chemistry for Dental Students 8vo, * Whipple's Tyhpoid Fever Large 12mo, Woodhull's Notes on Military Hygiene 16mo, * Personal Hygiene 12mo, Worcester and Atkinson's Small Hospitals Establishment and Maintenance, and Suggestions for Hospital Architecture, with Plans for a Small Hospital 12mo,
2 50
1
25
2 50 3 00
1
50 00
25
METALLURGY.
Betts's Lead Refining by Electrolysis. Bolland's Encyclopedia of Founding and Dictionary of in the Practice of Moulding
8vo,
4 00
3 00 2 50 2 50
1
Foundry Terms used

12mo, 12mo, 12mo, 12mo, 16mo, mor. 12mo,
Iron Founder
Supplement
Goesel's Minerals and Metals: * Iles's Lead-smelting
Douglas's Untechnical Addresses on Technical Subjects A Reference Book
00
3 00 2 50 3 00 2 50 3 00 2 00 2 50
1
Johnson's Rapid Methods for the Chemical Analysis of Special Steels, Steel-making Alloys and Graphite Large 12mo, Keep's Cast Iron 8vo, Le Chatelier's High-temperature Measurements. (Boudouard Burgess.) 12mo, Metcalf 's Steel. A Manual for Steel-users 12mo. Minet's Production of Aluminum and its Industrial Use. (Waldo.). 12mo, Ruer's Elements of Metallography. 8vo. (Mathewson) Smith's Materials of Machines 12mo,
00
Tate and Stone's Foundry Practice
12mo,
Thurston's Materials of Engineering. In Three Parts 8vo, Part I. Non-metallic Materials of Engineering, see Civil Engineering, page 9. Part II. Iron and Steel 8vo, Part III. A Treatise on Brasses, Bronzes, and Other Alloys and their Constituents 8vo, Ulke's Modern Electrolytic Copper Refining 8vo, West's American Foundry Practice 12mo, Moulders' Text Book 12mo,
2 00 8 00
3 50
2 50 3 00
2 50 2 50
16
MINERALOGY.
(In Preparation.). Baskerville's Chemical Elements. Pocket-book form. $2 00 Boyd's Map of Southwest Virginia * Browning's Introduction to the Rarer Elements 8vo, 1 50 Mineralogy. (Penfield.) 8vo, 4 00 Brush's Manual of Determinative 16mo, mor. 3 00 Butler's Pocket Hand-book of Minerals 8vo, paper, 1 00 Chester's Catalogue of Minerals Cloth, 1 25 * Crane's Gold and Silver 8vo, 5 00 Dana's First Appendix to Dana's New " System of Mineralogy" Large 8vo, 1 00 Dana's Second Appendix to Dana's New "System of Mineralogy." Large 8vo, 12mo, 2 00 Manual of Mineralogy and Petrography
. .
How to Study Them Large 8vo, System of Mineralogy Text-book of Mineralogy Technical Subjects Addresses on Untechnical Douglas's
Minerals and
12mo,
8vo,
50
half leather, 12 50
12mo,
8vo, Eakle's Mineral Tables (In Preparation). Eckel's Stone and Clay Products Used in Engineering. 16mo, mor. Goesel's Minerals and Metals: A Reference Book 12mo, Groth's Introduction to Chemical Crystallography (Marshall). * Hayes's Handbook for Field Geologists 16mo, mor. 8vo, Iddings's Igneous Rocks 8vo, Rock Minerals of Rock-forming Minerals in Thin Determination Sections. 8vo, Johannsen's
4 00 1 00 1 25
3 00
1
1
25 50
5 00 5 00
With Thumb Index

* Martin's Laboratory
Merrill's
5 00
Guide
to
Qualitative
Analysis
with
the
'.
pipe Non-metallic Minerals: Their Occurrence and Uses 8vo, Stones for Building and Decoration 8vo, * Penfield's Notes on Determinative Mineralogy and Record of Mineral Tests.
.
Blow12mo,
60 4 00 5 00 50
8vo, paper,
* * * *
Tables of Minerals, Including the Use of Minerals and Statistics of Domestic Production 8vo, Pirsson's Rocks and Rock Minerals 12mo, Richards's Synopsis of Mineral Characters 12mo, mor. Ries's Clays: Their Occurrence, Properties and Uses 8vo, Ries and Leighton's History of the Clay-working Industry of the United
8vo, 8vo, 8vo,
00 2 50 1 25 5 00
1
States * Tillman's Text-book of Important Minerals and Rocks Washington's Manual of the Chemical Analysis of Rocks.
,,,,,,
2 50 2 00 2 oo
MINING.
* Beard's Mine Gases and Explosions Boyd's Map of Southwest Virginia * Crane's Gold and Silver * Index of Mining Engineering Literature
.Large 12mo, Pocket-book form,

8vo, 8vo,
* 8vo, mor.
3 2 5 4 5
1
Douglas's Untechnical Addresses on Technical Subjects Modern High Explosives Goesel's Minerals and Metals: A Reference Book
Eissler's
12mo,
8vo,
16mo, mor.
8vo,
Ihlseng's Manual of Mining * Iles's Lead Smelting

Peele's
4 3 5
2 3 3 3 2
1
12mo,
for Mines Difficult Conditions.
8vo, Riemer's Shaft Sinking Under (Corning and Peele).8vo, * Weaver's Military Explosives 8vo, 2d edition, rewritten Wilson's Hydraulic and Placer Mining. 12mo, Treatise on Practical and Theoretical Mine Ventilation 12mo,
Compressed Air Plant
00 00 00 00 00 00 00 00 00 50 00 00 00 50 25
17
SANITARY SCIENCE.
Association of State and National Food and Dairy Departments, Hartford 8vo, $3 00 Meeting, 1906 8vo, 3 00 Jamestown Meeting, 1907 * Bashore's Outlines of Practical Sanitation 12mo, 1 25 12mo, 1 00 Sanitation of a Country House Sanitation of Recreation Camps and Parks 12mo, 1 00 (Designing, Construction, and Maintenance.) 8vo, Folwell's Sewerage. 3 00 8vo, 4 00 Water-supply Engineering Fowler's Sewage Works Analyses 12mo, 2 00 Fuertes's Water-filtration Works 12mo, 2 50 Water and Public Health 12mo, 1 50 Gerhard's Guide to Sanitary Inspections 12mo, 1 50 * Modern Baths and Bath Houses 8vo, 3 00 Sanitation of Public Buildings 12mo, 1 50 Large 12mo, 1 50 Hazen's Clean Water and How to Get It Filtration of Public Water-supplies 8vo. 3 00 (In Preparation.) Kinnicut, Winslow and Pratt's Purification of Sewage. Leach's Inspection and Analysis of Food with Special Reference to State Control 8vo, 7 50 12mo, 1 25 Mason's Examination of Water. (Chemical and Bacteriological) Water-supply. (Considered principally from a Sanitary Standpoint). 8vo, 4 00 * Merriman's Elements of Sanitary Enigneering 8vo, 2 00 Ogden's Sewer Design 12mo, 2 00 Parsons's Disposal of Municipal Refuse 8vo, 2 00 Prescott and Winslow 's Elements of Water Bacteriology, with Special Reference to Sanitary Water Analysis 12mo, 1 50 * Price's Handbook on Sanitation 12mo, 1 50 Richards's Cost of Cleanness 12mo, 1 00 Cost of Food. A Study in Dietaries 12mo, 1 00 Cost of Living as Modified by Sanitary Science 12mo, 1 00 Cost of Shelter 12mo, 1 00 * Richards and Williams's Dietary Computer 8vo, 1 50 Richards and Woodman's Air, Water, and Food from a Sanitary Standpoint 8vo, 2 00 * Richey's Plumbers', Steam-fitters', and Tinners' Edition (Building Mechanics' Ready Reference Series) 16mo, mor. 1 50 Rideal's Disinfection and the Preservation of Food 4 00 8vo, Sewage and Bacterial Purification of Sewage 8vo, 4 00 Soper's Air and Ventilation of Subways 12mo, 2 50 Turneaure and Russell's Public Water-supplies 8vo, 5 00 Venable's Garbage Crematories in America 8vo, 2 00 Method and Devices for Bacterial Treatment of Sewage 8vo, 3 00 Ward and Whipple's Freshwater Biology. (In Press.) Whipple's Microscopy of Drinking-water 8vo, 3 50 * Typhoid Fever Large 12mo, 3 00
Value of Pure Water Winslow's Systematic Relationship of the Coccaceae
Large 12mo, Large 12mo,
00
2 50
MISCELLANEOUS.
Emmons's
Geological Guide-book of the Rocky International Congress of Geologists Ferrel's Popular Treatise on the Winds Fitzgerald's Boston Machinist Gannett's Statistical Abstract of the World
Mountain Excursion of the Large 8vo.

8vo,
Haines's American Railway

*
Management
(Winton)
18mo, 24mo, 12mo,

8vo,
Hanusek's The Microscopy of Technical Products.
1 50 4 00 1 00 75 2 50 5 00
18
A Collection of Published Papers on OrJacobs's Betterment Briefs. ganized Industrial Efficiency 8vo, $3 50 Metcalfe's Cost of Manufactures, and the Administration of Workshops.. 8vo, 5 00 Putnam's Nautical Charts 8vo, 2 00 Ricketts's History of Rensselaer Polytechnic Institute 1824-1894. Large 12mo, 3 00 Rotherham's Emphasised New Testament Large 8vo, 2 00 8vo, Rust's Ex-Meridian Altitude, Azimuth and Star-finding Tables 5 00 Standage's Decoration of Wood, Glass, Metal, etc 12mo, 2 00 Thome's Structural and Physiological Botany. (Bennett) 16mo, 2 25 (Schneider) Westermaier's Compendium of General Botany. 8vo, 2 00
Winslow's Elements of Applied Microscopy
12mo,
50
HEBREW AND CHALDEE

Gesenius's
TEXT-BOOOKS.
Old Testament Scriptures.
Small 4to, half mor, 12mo, 5 00 1 25
Hebrew and Chaldee Lexicon
to the
(Tregelles.)
Green's Elementary
Hebrew Grammar
19
\V
"+
<*
<>
V>
P-
->
^-
d?
^,<<
\<^
^
rv
LIBRARY OF CONGRESS
H9MB

Elements of Machin

Hochgeladen von

Dokumentinformationen

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Elements of Machin

Hochgeladen von

Copyright:

Verfügbare Formate

HHH

hHHHHBI Hi HH bHrKBH HGH

H. BARR, M.S., M.M.E.

Manager, Smith Premier Works.

FIRST EDITION FIRST THOUSAND

JOHN WILEY & SONS

Electrotyped and Printed by Publishers Printine Co..

the outgrowth of the experience of the authors

in teaching Machine Design to engineering students in Sibley

for convenience, treated separately

portion of the subject which treats of the proportioning of machine

before proportioning the various

and convenient to outline the mechanism members.

forces resulting therefrom.

Proportioning the various parts to withstand these forces.

and the fundamental principles underseldom made clear

and such information

energy transformation in machines

in general, that relating

thorough understanding of these

however, in most cases, essential to success-

a consideration of the machine as a whole

discussion of typical energy and force problems

clearer to the beginner.

While the treatment presented presupposes a knowledge of

brief discussion of the

given in Chapter III, partly to

cation of the various formulae to engineering problems

modify logical theory.

hoped that the

numerical examples which are intro-

duced throughout the work may, in conjunction with the analy-

given, suggest proper treatment of practical prob-

which could be conve-

one college year and yet present the salient

interest to the practising designer.

has been the endeavor in

the preparation of the

book not only

to develop rational analytical

treatment, with due regard to constructive considerations

other practical limitations, but to reduce the analysis to such

forms and terms that definite numerical results can be obtained

Considerable of the matter contained in the book has already

selves freely of "the

given in the text.

authors are especially indebted to Professor G. F. Blessing

Swarthmore College, Professors W. N. Barnard, L. A. Darling,

have given instruction in the course

be very grateful for further suggestions

or criticisms which will

improve the book.

Ithaca, N. Y., June, 1909.

Definitions and Fundamental Principles

Straining Actions in Machine Elements.

Formulas for Strength and Stiffness,

Tubes, Pipes, Flues, and Thin Plates

Belt, Rope, and Chain Transmission,

Friction Wheels, Friction Brakes,

transform energy obtained

from natural sources into useful work

Useful work involves both motion and force, hence

definite force, for

parts are so put together as to give definite constrained