1

:is

f:^,^ tii

<.!

DEPAnTLibiT OF

imm

ENGn!EEu<;:
.:

Li6t!s:X.fJ urn hf: ..^..^....l^.

Return

ti'i:>1>c^o!<

to

....^

^:::^

Clipboard;..

Shelf:
.

Ail|>f<rk!-;

.'Tt.

to be signed for in the loan

vvlioii

bo^^y^l'.v. borrowed, r.nd

,

returned.
\7eei<,

ujuiu^rpust be returned witiiin One

permission
is

i!-;!jiC &pccii-i

given for a longer

"

WORKS OF PROFESSOR JOSEPH LIPKA

PUBLISHED by

JOHN WILEY & SONS,

Inc.

Giaphical and Mechanical Computation An aid in the solution of a large number of problems which the engineer, as well as the student of engineering, meets xiv + 259 pages, 6 by 9, 207 figures, in his work,
.

2 charts.

Cloth, net $4.00.

Also published in two parts. 119 pages, 6 by 9, Part I. Alignment Charts, xiv 130 figures, 2 charts. Cloth, net $2.00. 140 pages, 6 by 9. Part II. Experimental Data, xiv 77 figures. Cloth, net $2.00.

+ +

By HUDSON, LIPKA,
The Engineers' Manual

LUTHER and PEABODY

Howard B. Luther, S.B., Dipl. Ing., and Dean Peabody, SB., Massachusetts Institute of Technology, iv + 315 pages. 5 by 7M- 227 figures. Flexible " Fabrikoid
Jr.,

By Ralph G. Hudson.S.B., assisted by Joseph Lipka,Ph.D..

bincung, net $2.75. A consolidation of the more commonly used formulas of engineering, each arranged with a statement of its application.

By HUDSON and A Manual of Mathematics By Professors Ralph G. Hudson,

Ph.D.,
iii

LIPKA
S.B.,

and Joseph Lipka,

Flexible

132 pages.

by 7H-

95 figures.

"Fabrikoid" binding, net $1.50. A collection of mathematical tables and formulas covering the subjects most generally used by engineers and by students of mathematics, and arranged for quick reference.

Table

of Integrals

Ph.D.

Professors Ralph G. Hudson, S.B., and Joseph Lipka, Contains a Table of Derivatives, Table of Integrals,Natural Logarithms, Trigonometric and Hyperbolic Functions. 24 pages. 5 by 7%. Paper, net 20 cents.

By

GRAPHICAL AND MECHANICAL

COMPUTATION
PART
I.

ALIGNMENT CHARTS

BY

JOSEPH LIPKA,
Institute of

Ph.D.

Assistant Professor of Mathematics in the Massachusetts

Technology

NEW YORK

JOHN WILEY &

London:
1

SONS,

Inc.

CHAPMAN & HALL,

92

Limited

COPYEIGHT, I918,

BY

JOSEPH LIPKA

ISI

L7

192/
pi.
I

If.

H.GILSON COMPANT BOSTON, U.S.A.

PREFACE
This book embodies a course given by the writer for a number of
years in the Mathematical Laboratory of the Massachusetts Institute It is designed as an aid in the solution of a large numof Technology.

ber of problems which the engineer, as well as the student of engineering,

meets

in his

work.

In the opening chapter, the construction of scales naturally leads to discussion of the principles upon which the construction of various a The second chapter develops the principles of a slide rules is based.

network of

of coordinate paper

showing their application to the use of various kinds and to the charting of equations in three variables. Engineers have recognized for a long time the value of graphical
scales,

charts in lessening the labor of computation.

Among

the charts devised

none are so rapidly constructed nor so easily read as the charts of the alignment or nomographic type a type which has been most fully developed by Professor M. d'Ocagne of Paris. Chapters III, IV, and V aim to give a systematic development of the construction of alignment charts; the methods are fully illustrated by charts for a large number It is the writer's hope that the of well-known engineering formulas. simple mathematical treatment employed in these chapters will serve to make the engineering profession more widely acquainted with this time

and labor saving

device.

formulas in the engineering sciences are empirical, and the value of many scientific and technical investigations is enhanced by the discovery pi the laws connecting the results. Chapter VI is concerned

Many

with the

fitting of

equations to empirical data.

Chapter VII considers

the case where the data are periodic, as in alternating currents and voltages, sound waves, etc., and gives numerical, graphical, and mechanical

methods

for determining the constants in the equation.

When

empirical formulas cannot be fitted to the experimental data,

these data

may

and integration, by the numerical, graphical, and mechanical methods developed in the last two chapters. Numerous illustrative examples are worked throughout the text, and a large number of exercises for the student is given at the end of each chapter. The additional charts at the back of the book will serve
computation,
differentiation,

interpolation,

still

be

efficiently

handled for purposes of further

IV

PREFACE
Bibliographical

as an aid in the construction of alignment charts.

references will be found in the footnotes.

the

The writer wishes members of the

to express his indebtedness for valuable data to

Institute of Technology,

engineering departments of the Massachusetts and to various mathematical and engineering

the idea of a

publications.

He owes

Mathematical Laboratory

to

Edinburgh. He is especially indebted to Capt. H. M. Brayton, U. S. A., a former student, for his valuable suggestions and for his untiring efforts in designing a large number of the alignment charts. Above all he is most grateful to his wife for her assistance in the revision of the manuscript and the reading of the proof, and for her constant encouragement which has greatly lightened the labor of writing the book.
Professor E. T. Whittaker of the University of

JOSEPH LIPKA.
Cambridge, Mass.,
Oct. 13, 1918.

CONTENTS.
CHAPTER
Art.
1
.

I.

SCALES AND THE SLIDE RULE.

Page
Definition of a scale
i i

2.

Representation of a function by a scale

Variation of the scale modulus

3. 4.
5. 6.
7.

2 5
-.

Stationary scales
Sliding scales

8.

The The The

logarithmic slide rule

solution of algebraic equations
log-log slide rule

on the logarithmic

slide rule

11

13

9.

Various other straight slide rules

10.

Curved

slide rules

15 16 18

Exercises

CHAPTER

n.

NETWORK OF
11.

SCALES.

CHARTS FOR EQUATIONS IN TWO AND THREE VARIABLES.

20
21

Representation of a relation between two variables by means of perpendicular scales

12. 13. 14. 15.

Some

illustrations of perpendicular scales

16.

Logarithmic coordinate paper Semilogarithmic coordinate paper the solution of algebraic equations of the Rectangular coordinate paper 2nd, 3rd, and 4th degrees Representation of a relation between three variables by means of perpen-

22

24

26
28-

"

dicular scales

Charts for multiplication and division Representing curves are straight lines 18. Three- variable charts. 19. Rectangular chart for the solution of cubic equations Representing curves are not straight lines 20. Three- variable charts.
1 7.

30
32

35 37

21

Use

of three indices.

Hexagonal charts

40
42

Exercises

CHAPTER
22.

III.

NOMOGRAPHIC OR ALIGNMENT CHARTS.

Fundamental
(I)

principle

44

Equation of form /i() -H/sW

parallel scales

h{w)

or fi{u)

fi^v)

=/s(w) Three
45-54

Chart 24. Chart

23.

for equation (I) for multiplication

and division
V

45 47

VI
Art.
25.
26. 27.

CONTENTS
Pack

Combination chart

for various formulas

Grasshoff 's formula for the weight of dry saturated steam

Tension in belts and horsepower of belting

(II)

48 50 52

Equation of form

hiu)

Four or more

+Mv) +Mw)+'

=/,(/)

or /i() -f,iv)

-Mw)

=Mt)
55-63
55 56

parallel scales

Chart for equation (II) 29. Chezy formula for the velocity of flow of water in open channels 30. Hazen-Williams formula for the velocity of flow of water in pipes 31. Indicated horsepower of a steam engine
28.

57
61

Exercises

64

CHAPTER

IV.

NOMOGRAPHIC OR ALIGNMENT CHARTS

(m) Equation
32.

(Continued).

of form/i()

= fiiv) Mw)

or },{u)

= /2(t;)A^"
<

Z chart.

65-67
65 66

Chart for equation (IH) 33. Tension on bolts with U.

(IV)

S.

standard threads

Equation of form j-rr

fTT

Two intersecting index

lines

68-75
68 6g 70
71

34. 35.

Chart for equation (IV) Prony brake or electric dynamometer formula 36. Deflection of beam fixed at ends and loaded at center 37. Deflection of beams under various methods of loading and supporting 38. Specific speed of turbine and water wheel
(V) Equation of form /i()

73
inter-

= fi(v) 'Mw) M^)

'

Two or more

secting index lines

76-87
76
in pipes

40.

Chart for equation (V) Twisting moment in a cylindrical shaft 41. D'Arcy's formula for the flow of steam 42. Distributed load on a wooden beam
39.

77 79 80

43. 44.

Combination chart

for six

beam

deflection formulas

General considerations
(VI)

84 87
87-91

Equation of form "^j-^

= "^y-^ Parallel or perpendicular index lines

45.
46. 47. 48.

Chart for equation (VI)

Weight of gas flowing through an orifice Armature or field winding from tests

87 89 90
91
:

Lame

formula for thick hollow cylinders subjected to internal pressure

of

(Vn) Equation = Ml") Mq)

'

form

M^) - Mv) = Mi) - Mq)

or

M^) Mv)
91-95
91

Parallel or perpendicular index lines

49. Chart for equation (VH) 50. Friction loss in flow of water

Exercises

94 95

CONTENTS

VU

CHAPTER
Art.

V.
(Continued).

NOMOGRAPHIC OR ALIGNMENT CHARTS

(Vin) Equation of form /i(m) +f2iv)
index lines
51. 52. 53.
54.

Page

777

Parallel

or perpendicular

97-104
97 99
loi lOl

Chart for equation (VIII)

Moment

of inertia of cylinder

Bazin formula for velocity of flow in open channels

Resistance of riveted steel plate
(IX) Equation of form

fi(u) '^f2{v)

Mw)

'

Mu) '^Mv) '^Mw)

scales

"^

'

'

/4(g)

Three or more concurrent

55. 56.

104-106
104 106

Chart

for equation (IX)

Focal length of a lens.

(X) Equation of form /i(m) +/2(i') '/s^w)
scales

= Mw) Straight and

curved
106-1 13

Chart for equation (X) Storm water run-off' formula 59. Francis formula for a contracted weir
57. 58.

106
107

60.

The

solution of cubic

and quadratic equations

no no
114-117
ii4

(XI) Additional forms of equations.

61.
62. 63. 64.
^

Combined methods

Chart for equation of form Chart Chart

,-1

-re

/i()
for equation of

tt^

+ fiiw) -rr\ U{v)

I

fz{w)

<

for equation of form/i(M) +/2(f) 'fziw)

= /4(g)

114

form fi{u)

/4(g)

+U{v)
r ,

fsiw)
<

14

Chart for equation of form "tt^ Chart for equation of form

7-7^ + U{v) = Mw) /i()

115 1^5

65. 66.

t^^ Mu)

+ vtx = Mv)

67.

Chart for equation of form/i(M) /2(g) +M^) ' M"^) = M"^) Chart for equation of form/i (u) /2(g) +Mi^) /(') = Mi) +M''^)
Exercises

116
1

17 117

Miscellaneous exercises for Chapters III, IV,

118

CHAPTER

VI.

EMPIRICAL FORMULAS NON-PERIODIC CURVES.

68.

Experimental data
(I)

120

The

straight line

122-127
bx a 122

69. 70.

The The

straight line, y
straight line, y

= =

+ bx

125

Vlll

CONTENTS
(II)

AxT.

Formulas involving two constants

Pagb 128-139
128
131
is

Simple parabolic and hyperbolic curves, y = c** Simple exponential curves, y = ai'' hx^ (where n 73. Parabolic or hyperbolic curve, >" = a % X or - = a bx 74. Hyperbolic curve, y =
71.

72.

known)

135 137

(III)

Formulas involving three constants

ax^

140-152

75. 76.
77.

The parabolic or hyperbolic curve, y = The exponential curve, y = a^^ -\- c The parabola, y = a \- hx \- cx^ The hyperbola,
y =
X -7
Y
c

+c

140 14J
145
149

78.
79.

The

logarithmic or exponential curve, log y

-\r

hx

-\r ex''

or y

= a^"^'

151

(IV) Equations involving four or

80.

more constants

152-164
152
153 156

81.
82.

83. 84.

The additional terms ce^'^ and cx*^ The equation y == a + bx + ce^' The equation y ae^^ -\- ce^^ The polynomial y = a + to + cx^

+ dx^ +

159
161

Two

or

more equations

Exercises

164

CHAPTER
EMPIRICAL FORMULAS
85. 86.

VII.

PERIODIC

CURVES.
170 170
173

Representation of periodic phenomena the harmonics of a trigonometric series 87. Determination of the constants when the function is known 88. Determination of the constants when the function is unknown

The fundamental and

Numerical evaluation of the coefficients. Even and odd harmonics Numerical evaluation of the coefficients. Odd harmonics only Averaging selected ordinates ... 91. Numerical evaluation of the coefficients. Averaging selected ordinates. 92. Numerical evaluation of the coefficients. Odd harmonics only
89. 90.
.

174 179 186

192

198

93.

Graphical evaluation of the coefficients 94. Mechanical evaluation of the coefficients.

Exercises

200

Harmonic analyzers

203 207

CHAPTER

VIII.

INTERPOLATION.
95.
96. 97.

Graphical interpolation
Successive differences and the construction of tables

209 210

Newton's interpolation formula Lagrange's formula of interpolation 99. Inverse interpolation

98.

Exercises

214 218 210 272

CONTENTS

IX

CHAPTER
Art.
loo.

IX.

APPROXIMATE INTEGRATION AND DIFFERENTIATION.

The
loi. 102.
103. 104.

Page

methods Rectangular, trapezoidal, Simpson's, and Durand's

necessity for approximate

rules.

Applications of approximate rules

224 224 227

231

General formula for approximate integration Numerical differentiation 105. Graphical integration 106. Graphical differentiation The planimeter 107. Mechanical integration.
108.

Integrators

109. 110.

The integraph
Mechanical differentiation.
Exercises

<

234 237 244 246 250 252 255 256

The

differentiator

Graphical and Mechanical Computation.

CHAPTER I. SCALES AND THE SLIDE RULE. Definition of a scale. A graphical scale is a curve

(l)

or axis on

which are marked a series of points or strokes corresponding in order to a set of numbers arranged in order of magnitude. If the distances between successive strokes are equal, the scale is said to be uniform (Figs, la, ib). If the distances between sticcessive strokes are unequal, the scale is said to be non-uniform (Fig. (lc). The strokes are drawn as fine as possible, perpendicular to the axis which
carries the scale.
I

T
ic.

Fig.

I a.

Fig. i&.

Fig.

u^ of a variable u.

(2^ Representation of a function by a scale. Form the table

Consider the function

10

^=0123456 789 = 49
u^

16

25

36

49

64

81

100

from the origin 0, lengths equal to ic = 0.04 u^ inches (Fig. 2a) mark at the strokes indicating the end of each segment the corresponding value of u. Thus, a stroke marked w is at a distance of 0.04 u^ inches from the origin. Fig. 2a is said to represent the function u"^ by a scale. The length 0.04 inches is chosen arbitrarily in this case to represent the unit segment used in laying off the values of u^ on the axis. This unit segment is called the scale modulus.

and on an

axis,

OX lay
;

ofif

or 2

rrn

-i

;:^

J 9

t,

fo

(u)
Fig. 20.

In general, any function f{u) of a variable u such that each value of the variable determines a single value of the function, may be represented

by a

scale.

Form
U = f{u) =

the table
Ui

U2
/(2)

. .

/(i)

f{Uz)

SCALES

AND THE

SLIDE RULE

Chap.

and on an axis OX lay off from the origin lengths equal to ic = mj{u) inches, and mark with the corresponding values of u the strokes indicating the end of each segment. Fig. 2h is said to represent the function /(m) by a scale. The length m inches is chosen arbitrarily to represent the unit segment used in laying off the values of /(w), and it is called the scale
modulus.

The equation x = mf{u)

is

called the equation of the scale.

(u)
Fig. 2b.

The uniform
/()
o
I

scale
2c,

is

a special case of this representation when

=
'1
1
1

u.

In Fig.
2
1 I

X
3

= mu, where
4
1

m=
6
1

0.5 inches.*

/
)
1

1
1

1 1

1 1

I
I

1 1

5
1
1
1

7
1 1 1 1 1

a
1 1 1
1

9
1 1 1 1

to
1

1 1

1 1

Fig. 2c.

The
j{u)

logarithmic scale
log u.

is

a special case of this representation

when

f
I

In Fig. 2d, X

= m
I
I I

log u,

where _w
a
5
IIII
I

12.5

cm.*
7

2
1 1

1 I

I
I

3
I I I I I I I I I I

e
1

a
I
I

9
[

'TO
j

l|

IHH

1 1

Fig. 2d.

The uniform and logarithmic scales.are the most important scales for our work. After we have constructed a scale for /() from a table of values of and /(m) we may wish to estimate the value of u corresponding t o a stro ke intermediate between two strokes of the scale, or to estimate on the scale the position of a stroke corresponding to a value of u intermediate between two values of u in the table. This process of interpolating on the scale is of course very much easier for uniform scales than for non-uniform scales. The accuracy of such interpolation evidently depends upon the interval between two successive strokes. Experience has shown that
,

in. (very rarely this interval should not be less than i mm. or about need it be as small as this) this may always be done by the choice of a proper scale modulus.
;

(^
Xi

Variation of

the scale

modulus.
is

_m with which a

scale for /(w)

varying the modulus constructed, we^et a series of scales

. . .
,

By

mif(u), X2

W2/(m), X3

msfiu),

all

representing the

same

* The values of given in the text are those which were employed originally in the construction of the scales; these values do not however refer to t he cuts which, in most cases, are reductions of the original drawings.

Art, 3

VARIATION OF THE SCALE MODULUS

. . . ,

Only one of these respectively. ntz, mz, need be constructed by means of a table of values of /(w), and the others may be derived graphically from this. In Fig. 3ai let OiXi carry the scale Xi = mif(u) we wish to construct Let be any conthe scale X2 = tn2f{u).
function with moduli mi,
scales
;

OOi and on this line choose O2 such that OO2/OO1 = W2/W1; through O2 draw O2X2 parallel to OiXi. If A is a point on OiXi marked ', then OiA = Wi/(m')> and OA will cut O2X2 in a point B ^^ such that O2B/O1A = 002/OOi. or O2B = Wi
venient
point;
join

X,

SCALES
transversal through
will

AND THE SLIDE RULE

be marked with the same value of
its

Chap.
u,

and

the scale on 02^2 will have for

equation x^
-

niifiji).

need not be drawn, but simply their If the transversals through are drawn, then we may get a scale of any required modulus by merely drawing a parallel to 0\Xi dividing the segment OOi in the required ratio; thus a line midway between O and Oi will carry a scale with modulus
transversals through

The

points of intersection with O2X2 indicated.

Fig.

3C.'*

W1/2, a line of the 2 W1/5, etc.

way from

to Oi will carry a scale with

modulus

This principle

is

illustrated in Figs. 3&

If

and 3c

for

uniform and loga.1, .2, .3,

. . .

rithmic scales respectively.

we mark

a uniform scale

.9, on the base line beginning at 0, then the lines through these points parallel to the left-hand scale with modulus will cut the transversals

m
,

in scales

w, .9 w, respectively. best to make the charts in these figures almost square, and to take w
.1

whose moduli are

m,

.2

It is

10

Art. 4
in. for

STATIONARY SCALES
the uniform scale and

'

m =

25 cm. for the logarithmic scale.

The

chart of uniform

scales will then

or architect's hexagonal scale,

be an amplification of the engineer's and the chart of logarithmic scales, an

amplification of the logarithmic slide rule.

necessary the scales in either chart may be extended. Note, however, that in the case of the logarithmic scales, the segment representing
If

m = 10 is of the same length as the segment representing the interval from w = 10 to m = 100, or, in general, the segment representing the interval from u = 10" to m = 10"+^. It is convenient to draw Figs. 36 and 3c on durable paper. Only the primary scale with 'modulus m, the base line and the transversals need be drawn. The paper may then be creased along any parallel to the primary scale to give a scale of the required modulus. Charts of this nature have been used to assist in constructing a large number of the scales and charts that follow, and much time and energy have been saved thereby. (Such charts will be found in the back of this book.) A relation between two variables u and v ('4^ Stationary scales. of the form v = /(w) may be represented graphically by constructing the two scales x = mv and x = mf{u) on opposite sides of the same axis or on adjacent or parallel axes with the same jnodulus and from the same origin or with origins in a line perpendicular to the axes. If C represents degrees Centigrade and F represents degrees FahrenWe construct the uniform scales x = m (F 32) heit, then F 2,2 = 1.8 C. and X m (i.S C) on opposite sides of the same axis, and from the same In Fig.^a,\ origin, i.e., the points marked C = o and F = 32 coincide.
the interval from m

to

f
C -20

!0

20 30

40

SO
10

60

70

80 90
30

WO
40

770

720 130 740 ISO 760 770 180

790 20O Z70

-10

20

50

60

70

BO

90

lOO

Fig. \a.

m=
If

0.02

in.,

so that the equations of our scales are x

C.

0.02 (F

and X = 0.036
pressure

By means

of such a figure,

we may immediately
in.,

32) con-

vert degrees Centigrade and Fahrenheit into each other.

is

expressed as pounds per sq.

P, and feet of water,

jc

W, then

P =
'
i

0.434

^-

Draw

the uniform scales

8

= mP and
10
'

w
'1

9
'

//
'

'

I'i

'

' 1

nM M MM
I
1'

'
1

'

1'

I'l'

'1

'i

i'i'
'
'

i'

\. \ \
' '

i'i'i
'

'

h'
'I

'
i

1' ['
I I'

I'

I'

r
I'

I'

I'

'

'

'

I'

r'

I'

'

I
I

'I

S
Fig.
ifh.

as
'I
I

'I

'

'

'I

'I

'

I I

'

10

m (0.434 W)
so that

from the same

origin.

In Fig/^J', the scale modulus

is i in.,

0.434 ^- We another scale for pressure expressed in inches of mercury,

we have

the scales x

= P and

may add M; thus

SCALES
0.492

AND THE SLIDE RULE

Chap.

scale has for equation x = 0.492 M. By (drawn with the aid of chartf36 or a pair of dividers) pressure may be read immediately in pounds per'sq. in., feet of water, or inches of mercury. If the relation between the two variables is of the form v = jogju, we construct the adjacent scales x mv and x = mlogu. If we take w = 25 cm., the logarithmic scale will be the same length as that of the logarithmic slide rule, and if the uniform scale is divided into 500 parts, we can use such a figure to read easily the values of the mantissas of the logarithms of all numbers to three decimals, and conversely to read the

P =

M, and

the

means of such a

figure

numbers corresponding
at
V |iM
/
i

to given mantissas (Fig. 4c).

0.4
i,i
|

The
0.$

slide

rule

0.2
iii,Ml,i
,
i

0.3

0.5
ipl,i
)|
i,,
,

0.6

0.7
i,^
i

0.8
ii|i
i i

1.0

ii

4ilVM

|ii,M;i|n,i|i^i^iJi, i|

i^||||)|||

l^li|n

(||

,|

3
Fig. 4c.

10

contains two such adjacent scales.

for the

The

chart, Fig./3?; could be used

same purpose

if

we

construct a uniform scale adjacent to the

primary logarithmic scale. If the relation between the two variables is of the form v u^, we may write this as log v = % log u. Here we construct the adjacent scales X = m log v and x = m (f log u), i.e., two logarithmic scales with moduli m and 3 m/5 respectively. We use chart 3c and get Fig. 4</, from
/
</
|
'
,

2
't,.v
i

3
i

4
.

56789/0
l
,
,i

20
i ,

30
in il^
,

40 SO 60 708090100
I

,.i
,

^n

i,i

i|

l^,.,

|, ,,

lp|,i
,

,||,

||,

|| |,

| ,

4
Fig.

56789
|

^H

n|4M

^1

,1
^

1,

,1

i(.
|

,^

10

20

4(f.

which we can read

when u
i

lies
-20.

within the limits

]

to lOO, or read

u when
u^,

V lies within the limits

to

We may similarly construct two adjacent scales for the relation p =

The chart Fig.f3c may conveniently where is any positive number. be used for this purpose. We may write the relation as \ogv = p log m, and we pick out on the chart the scales x = m log v and x = {ptn) log u, Since the axes carrying these scales are i.e., with moduli m and pm. parallel with origins in the same perpendicular, any perpendicular to the axes will cut out corresponding values of u and v. If /> < i we may use lip > i we write the relathe primary logarithmic scale for the v scale, tion in the form u = v^/^ and proceed similarly. If in the relation v W, ^ is a negative number, say, g, then = 10 m~, we merely shift the position of the deciIf we write V = vr". mal point in the value of v; then logf = log 10 glogw. Construct
,

t;

Art. 5

SLIDING SCALES

the adj^spent scales x

= m

log v
;

and x

= m

(log lo
is
a.s

q log u)

m
and

(qm) log u from the same origin

X

the latter scale

merely the scale

starting point

{qm) log u constructed from the point x

= m

^liiili s
1

'

^-OttOt^^ M>

<.

L_L Till

i'i' i|

iV 'iVii'n'
i |

ii

iiili i|'iMi

ti i|

'r'('i
I

i]

I'll
Ok

2
Fig. 4e.

proceeding to the

left.

Fig. 4e represents the relation v

u~^ con-

structed with the help of chart 3c.

(5.

Sliding scales.

Consider two functions f{u)

inf{u)

and

F{v)

and con-

struct their scales

x =

and x =

mF{v).

If these scales are placed

adjacent with their origins coinciding or in the same perpendicular to

the axes (we shall call this the stationary position), then for

any

pair of

values u and v opposite each other,

hence, mf{u)

we have OA = O'B
F{v).

(Fig. 5a),

and

mF (v),

or
f(u)

(in the stationary position)

(I)

This relation was illustrated in the examples of Art. 4. If now one of the scales is slid along the other scale through any distance d, then for any pair of values of u and v opposite each other, we have OA - O'B = d (Fig. 56), or mf(u) - mF{v) = d, or/(w) - F{v) =

SCALES

AND THE SLIDE RULE

Chap.

As an example,
log f or

consider the scales x

after sliding,

= m

log

u and x

=
by

m log v.
(I),

If

these are placed adjacent, then, in the stationary position,

log

V,

and

by

(II),

log

log w

constant, or

u log -

constant, or -

constant for

any pair of values of u and v opposite

pairs of values of

each other, and each other.

=
if

for

any two

u and

opposite

Thus

any three

of the last four quantities are given, the

Ui=5.2
/

8
I

9
I

10
I

7
v,-2.i

3/0.

1^-40

Fig. 5^.

fourth quantity
the scales until
in Fig. 5^, ^

may be

found at once; thus

Wi,

if

Ui, Vi, U2

are given, slide

is

Vi is

opposite

and read p opposite U2.

This

illustrated

'

where we read ^-^

2.5

^^^^-

scale

We may perform the same X = m log u sliding along an

4.00 operation

by means of a single logarithmic unmarked axis (Fig. $e)

.

tf3.7

u- 5.92
-I

7
u-3.7

to

n
V-2.S
Fig, 5e.
v.-

7
4.0

I I

tu

1st position:

place the scale x

= m
=

log

axis

and mark on the

slide

latter the values Ui

2d position:
the

the scale x

u adjacent to the unmarked and thlog u until Vi of this scale falls

opposite Ml of the

unmarked

axis; then read V2 of the scale opposite W2 of

unmarked
It is

axis.

evident that simple multiplication and division are special cases of the above, for if wi = i or 10, then ^2 = W2 ^i or 10 U2 Vi, and if
V2

or 10, then W2

Ui/vi or lo Ui/vi.

Art. 6
6.

THE LOGARITHLIIC SLIDE RULE

The
logarithmic slide rule.*

This instrument consists of several

and one uniform scale, some on the stock of the rule and others on the slide. Any two of these scales may be placed adjacent by means of a glass runner which has a fine hair line scratched on its under side and which is adjusted so that the hair line is always perparallel logarithmic scales

pendicular to the axes (Fig. 6).

All logarithmic slide rules

The

following

is

do not carry the. same number of scales. a description of the scales and their equations on the

4 S 6 7891
I. I I
1

4 S 6 7 691
I

,1

1
I

CI

^1

7
1

IT ^
6

CI rrnnr
I

J
7 8 9
T

a 9i^/
I

I'

'I

Fig. 6.

modern Mannheim standard

"the lo-inch rule."
cm. and we

rule (polyphase or duplex),

commonly

called
is

The
.

length of the graduated part of the rule

25

shall designate this length

by m.

The scales are distinguished

by the

letters
c,
.

A, B, C,
,

We

shall use the corresponding small

letters a, b,

to represent numbers on these scales.

= = =

ml.

:x
:

in log c.

D
CI
s).

= m log d.
10 = m log
1

X X

log a.
~ log k.
3

B
S

log b.
2
(100 sin

X X

'

Ci

^ l>ive{

'.X

= m log

= m log

(10 tan

/).

graduated so that we can easily read three Rules for the position of the decimal point may be given, but in computing it is best to disregard all decimal points and to estimate the position of the decimal point in the final result. The following are some of the relations which arise through the application of the principles of stationary and sliding scales to this type of rule. (Other illustrations will be found in the manuals issued by the manuscales are
figures in

The C and

any part

of these scales.

facturers.)

We
it

shall designate the stationary or initial position

by

(I)

Gunter invented the logarithmic scale and used compasses to Oughtred invented the straight logarithmic slide rule, consisting of two rulers each bearing a logarithmic scale, which were slid along each other by hand (1630). Rules in which the slide worked between parts of a fixed stock were known in England in 1654. Robertson constructed the first runner in 1775. Mannheim designed the modern standard slide rule (1850). Roget invented the log-log scale
calculate with
(1620).

* Historical Note.

in 1815.

See F. Cajori's History of the Logarithmic Slide Rule.

16

SCALES

AND THE SLIDE

RXJLE

Chap. I

and the position after sliding by (II). designated by the same subscript.
(i)

Numbers

opposite each other are

LandD:
=
log

(I) /

d and d

antilog
:

/.

(I is

only the mantissa.)

(2)

C and D (or A and B) c = log c = log d, (II) log c logd = const.,

(I)
.'.

d.
.'.

-j

const.

Art. 7

SOLUTION OF ALGEBRAIC EQUATIONS ON THE SLIDE RULE

II

.456', .ooo'456' are said to

contain three figures, the left-hand groups

in 45'.6, .045'6, .ooo'o45'6 are said to contain two figures, while the

left-hand groups in 4'.56, .004'56 are said to contain one figure.

We

read k in the first, left-hand group contains one, two or three figures.
(7)

middle or

last third of the

K scale according as the

CandX:
log c

(I)

^ log k,

.'.

\/k and k
.\

(?.

(II)

log c

^ log k
(or

const.,

-^ =
v^,

const.,

:.

-j~ =

-^'

^^

(8)

A
A

an^-X
I log a

and K)
.*.

(I)

I log ^,

Va =

or

k^

and k=a^.

(9)

and

S'.

(I) ^

\ log a

\ log (loosing),

.*.

loosing and

sin~*

100
^
^

(II)

loga

I log (loosing)

const.,

:. .

sin 5

const, or

sinji

sin 52

The

last relation

may
s,

be used
set
sin

in the solution of oblique triangles.

To To

find

y ='a
y

sin

we

a 90

y
sin 5

find

sin 5

we

set
sin 5

sin

90
s).
*

We
(10)
(I)

also note that cos s

sin (90

Dand
log d

T:

log (10 tan

t),

.'.

10 tan

and

tan~^

(II)

\ogd
find

log (10 tan =

t)

const.,

.'.

=
tan
t

const., or

tan

ti

tan

t^

To
To

y
y

=
=

d tan
-,

d
t,

set

tan 45
set
;

y tan

find

tan/
/

tan/'

tan 45

We
7.

also note that cot

tan (90

/)

The

solution of algebraic equations on the logarithmic slide rule.

The

between the D dnd CI scales expressed in Art. 6 (3), viz.: product of d and ci is the same for all such pairs of numbers opposite each other, may be used to assist in the solution of algebraic equations of the second and third degrees. Thus if we set ci = I over d = q, then over any number y on the D scale we shall find
relation
that, after sliding, the

- on the CI

scale (since

-j and

on the

scale; this Us

12

SCALES

AND THE SLIDE RULE

Chap.

illustrated in the

accompanying diagram:

CI

(Fig. 7).

We

may

use

ci

lo instead of ci

if

necessary, but care

must be taken

in reading the position of the

decimal point.

tst
fl/tf

6 7 8 9
I I I I

1
I

J
4
1

S
I

6 7
I I

891
I I I

a
Bl

19 6
I I
I

7
I

6 5
.1
I

""I

f"

n U

Fig.

7.

(i)

If

we

slide the

runner until the readings on the

- ox y^

and CI

scales

are the same, then y

=
>

q^

and

3/

db

Vg.

Thus,

if 3/^

5,

or

we have

CI

Art. 8

THE LOG-LOG SLIDE RULE

13

The nature of

the roots of this cubic equation are determined as follows:

I

only

real root;

if

is

+,

root

is

-+
4

-i 27<

+ +

if

is

root

is

3 real roots of which 2 are equal. 3 real and unequal roots; i root is
root? is

and

2 roots are

or

To

find

and 2 roots are +. the negative roots, we replace y by ji, and the positive

roots of the resulting equation are the negative roots of the original

equation.

We also
=
g
is zero.

note that the

sum of the three roots of the equation 'f -{- py The complete cubic equation z^ -\- az^ -\- bz -\- c = o must
y^
-\-

first

be reduced to the form

py = qhy the substitution

= y

To
the

facilitate the

the slide so that the

comparison of the A and CI scales, it is well to invert C scale is transformed into a CI scale and slides along

scale.
if

Thus,
since

y+

3'

=
I

5 or

y+

^, there

is

only one positive root

is

y
is

+
4

27

5 4

>

o and g

positive.

This positive root

found

A
by the
setting

1-33

CI

Again,

if

D y4
64 ^ 27

4-33 I-I53
3^

I-I53-

2 or

3/^

=y

there are three real roots, since

is

^
4

p^

The

positive

root

found by the setting

'

27

14
the required scales
of

SCALES AND THE SLIDE RULE

Chap.

we must
expressed

write this equation in the form (II) or (III)

in

the

principle

Art.
^

5.

Taking logarithms, we get

= -log wi, or 2__1 log log til - log Ci = log log W2 log 2

= _&

and taking logarithms again, we get

C2,

log

or log log

log c

const.

The
log
c.

equations of our scales are therefore x

m log log n
our scale

and x

=m

The

initial

point of the

w log log 10 = (since x = m log log 10^ = m log 10

The range from n
that
it

would be marked n = 10 (since x = wlog 1=0), and the end point would be marked n = lo^"
scale
==''w) if
is

to be

cm, long.

10 to w

10^" is

not a convenient range for n, so

what.

has been found best to modify the equation of this scale someAn instrument called the log-log rule (Fig. 8) has been constructed
l.0

lot

uo

LLi

LL3

'

'

'

'-n 4

r^
5

SCO

1000

1000

logo 10000 Kou

Fig.

8.

which the equation of the n scale is x = m log (100 In U) (where = loge // and e = 2.71828 the base of Napierian or natural logarithms). The scale is broken into three parallel scales of length m = 25 cm.
in
In//
. . . ,

the

first,

marked LLi with a range from

//

eO.oi,^^

qj approx.)

to

//

e^\

the second,
.

the

third,

marked LL2 with a range from // = e-^ to II = e,''^ marked LL3 with a range from II = e' to 11 =
approx.).

e*

(= 22,000

By

sliding the adjacent scales

=
or

in log

(100 In

//)

and x

= w log c, we
VTl
-^

have log (100 In//)

or
If

logc

const.,

const., or

const.,

v^ =
we
set

V^,
Ci

or

II2

Ih"';

hence we have the setting

Ih,

we have II2 = lli^ or lU = lli^. Of and C2/10 have the same position, but on the LL scales the decimal poiht must be left in its original position. It is easy to see on which of the three LL scales the result is to be read;
I

or 10 opposite

course on the scale x

= m

log

c, c^

thus

2*-^

gives the setting

LL2 LL3

22.6
4-5
I

.*.

22.6

C
and
y

2-^ gives

the setting

LL2 LL3

1.366 2

y~

1.366.

0.45

Am.

VARIOUS OTHER STRAIGHT SLIDE RULES

note above that the smallest value of
at once be replaced
// is

15

We
0.99

i.oi.

Values of

//

may

by

their reciprocals,

and the

reciprocals of

the final result taken,

smce

jj-

=(77-)

The LL and

the

C scales

in their initial position

may

also

be used to
(100 In
//)

find directly the natural logarithm of

a number, for

we have log

log c or In ^
It
is

//

100

with the log-log rule.

r

evident that compound interest problems are very easily solved Thus, the amount, A of the sum of $1 .00 placed i 1. 00 at
,

per cent interest for n years and

compounded

ar, q times a year,

is

given b by

(r
I
-\

\"'
I
;

the required setting

Many
9.

other illustrations

may be

C nq I or 10 found in the manuals published by

is

then

LL

(+

the manufacturers.

Various other straight slide rules. -^ As another illustration of the use of sliding scales, let us construct a slide rule for the expression

-4--.
u
V

If

we choose

for our scales

= m(-] and
\uj
have

= wfio
\
1

v)

),

then for any position of our scales

or

we
1

shall

[10

const.

= U
I

const.,

or
Ui

=
Vi

U2

V2

If

we choose
be 10

the modulus, m,

to be

I in.,

and the

total length of our scales to

in.,

then the range

o
' '

'

T'

I'I'iiii'

'I

II

Fig.

9.

of

u is from u = cc (x = from y = 0.1 (x = o) to

opposite 2

6) to y

=
(jc

o.i (x

10),
(Fig.
,

and the range

9.)

of v

is

00

10).
1

Now

if

we
read

set
off

^2=00

u,

we

shall

have
Vi,

- and we

may

any one
rule

of the three quantities Wi,

jf

the other two are known.

This

may be used

to solve the equation

-5-

Kx

+ -5- = K2

J\.

d where R

is

the com-

bined electrical resistance of two parallel resistances Ri and R2, or to soive

the equation
7-

/:

+h =
7-

-j,

where /

is

the principal focal distance of a lens

and/i and/2 are conjugate focal distances.

i6

SCALES

AND THE SLIDE RULE

Chap. I

A large number of slide rules have been constructed for solving various
special equations in engineering practice.

Among these may be mentioned: stadia slide rules for measuring the horizontal distance and ver-

tical height when the rod reading and the elevation of telescope are known; Nordell's sewer rule for solving Kutter's formula for circular sewers; Hudson's horse power computing scale for obtaining the indicated H.Ps of an engine (this rule has two slides); Hazen-Williams

hydraulic rule for finding the velocity of the flow of water through pipes
(see chart

on

p. 6i).
/

CIRCULAR SLIDE RULE

Fig. 10.

10.

Curved
TT

point, viz., 2

the angular magnitude about a radians, into looo equal parts by straight rays drawn
slide

rules.

Divide
;

through the point. with the number i

Choose one
(for

of these rays as initial ray

and mark

it

log i)

mark

the ray at the end of 301 parts

Art

lo

CURVED SLIDE RULES

with the number 2 (for 0.301 = log 2), and the ray at the end of 477 parts with the number 3 (for 0.477= log 3), etc. Then the angle will be divided logarithmically (Fig. 10). The circumference of any circle drawn with the point as center will likewise be divided logarithmically, the
points on the circumference carrying the

same numbers
,

as the rays

through them. Designating such a circumference by A the numbers on its radius by rg inches, the equation of the scale on ^ is x =
i.e.,

it

by

a,

and

(2 irra) log a,

the point marked ai

point, a

is

at a distance of (2 xra) log ai inches from the ini-

tial

I,

measured along the circumference.

We now draw a con=

{2 irn) log b

centric circumference, B, of radius n, carrying a scale x

and so constructed that the plane of the 5-circumference can rotate about the center. If in the initial position of the scales, i.e., when the numbers 0=1 and 6=1 are on the same ray, a ray cuts out the numbers ai and hi, then we have
(2 irra) log ai

)
If,

(2 irrb) log bi

=-

/-a
,

.-.

lo^ ai

log

I,

bi

J and

ai

U
&i.

Tb

after rotation of the 5-scale

through any angle, two rays cut the scales

in ai, bi
)
(

and
,

^2, 62,

then

(2 Trra) log

fli '

(2

Tn) log

bi

(2 irra) log a2
-)
, (2 Trrb) log 62

Ta =-.
?'6

logai

-loga2
1

log&i-log62.

I.

J,

fli

02
62
is

61

Hence the

ratio of

two numbers on the same ray

is

constant.

This prin-

ciple of rotating circular scales

therefore similar to the principle of sliding

straight scales.

Thus
3;

for
& or

Multiplication:
r^.
.
.

=a =T b
a

- = ^ we
,

set
^

one factor
dividend
divisor

product
other factor

Dimsion:

^^

i;

ay ~
b

1'

>

^^

set

A B

quotient

One advantage
difficulty of

of such
off

a circular rule

lies in

the fact that

we avoid

the

running

the rulC) as often happens in setting with the

straight slide rule.

In the instrument called "Sexton's

Omn imetre," which

Fig. 10 repro-

duces in part, the .4- and 5-circles have the same radius, about 3 inches, A so that it is approximately equivalent to a straight rule of 18 inches.
ray drawn on a strip of celluloid capable of revolving about the center aids in the setting of the scales. The plane of the 5-circle also contains the
scale X

(4

7r?'c)

log c\

and

for

two numbers, b and

c,

on the same

ray,

we

have, in the initial position,

^4.Tr,)Jogc
. ;
,

(2

irrb)

log b

re

-,
rb

2logc
or
-j

log b

^ =

I,

.*.

c^

and

= ^/r Vb.

l8

SCALES

AND THE SLIDE RULE

Chap.

It is evident that if c is to vary from i to lo, the C scale must consist ot two concentric circumferences, on one of which c varies from c = i (or X = o) to c = Vio (or X = 2 Trrc) and on the other from c = Vio to The C and B scales thus serve for finding squares c = lo (or X = 4 TT^c). and square roots. We may also combine the C and A scales, and after The instrument also contains three rotation we have a\/c^ = ch./C'^. concentric circumferences for the scale x = (6 -wrk) log k and a combination with the J5-scale, in the initial position, gives b = k^ or k = v^. The instrument further contains scales for sines, tangents, and versines, and a scale of equal parts. There are other forms of curved rules. "Lilly's Improved Spiral
;

and a
of

Rule," a disk 13 inches in diameter, consists of a spiral logarithmic scale circular scale of equal parts, and is equivalent to a straight rule

about 30 feet long; it gives results correct to 4 figures. "Thacher's Rule" consists of two logarithmic scales one on a cylinder and the other on a set of 20 parallel bars external to the cylinder. This is really an amplification of the straight slide rule, involving the same principle in its use; the rule gives four figures correctly and a fifth may be estimated.*

EXERCISES.
(Note.

For the constructions

in Exs.

4-10 use charts of uniform and logarithmic

(

scales. Art. 3.)

1.

Construct scales for the function

Vm

o to

= =
in.

100)

if

w is

i in.,

0.5 in.,

0.2 in., respectively.

2.
^^ in.,

Construct scales for the function log m (

respectively.

to

10)

if

w=5

in.,

10

in.,

3. 4.

Construct a scale for the function log

if

w =

10

Construct adjacent scales for converting inches

/.

(7) to

centimeters (C)

we have

C=

2.54

5. Construct adjacent scales for converting cu. ft. per sec. (C) to million gallons per hour (G) we have G = 0.0269 C. 6. Construct three adjacent scales for converting foot-lbs. per sec. (F) into horsepower = 1.356 X lO"' F. (H.P.) and kilowatts (K)] we have H.P. = 1.818 X lo'^ F, and
;

7.

Construct adjacent logarithmic scales for the following:

(c) V

=-

(ft)

= =

(c)

= <^; =
'.

(d)v=^^;
8.

ie)v

u-i;

(/) r

Construct adjacent logarithmic scales for

A =

where

D = diameter

144

in inches
*

and

A =

area of circle in sq.

ft.

For descriptions and

illustrations of this rule

and other

rules, see

"Methods

of

Calculation, a

Handbook
Bell

of the Exhibition at the Napier Tercentenary Celebration,"

published

by G.

&

Sons, London.

EXERCISES
9.

19

Construct adjacent logarithmic scales for h

= -^
02,5

where h

pressure head

in

ft.

and

P =
and

pressure in

lbs.

per sq.

in. for

the flow of water.

10.

Construct adjacent logarithmic scales for A

ft.

= j

where h

velocity

head in
11.

velocity in

ft.

per sec. for the flow of water.

li^

Show how

to construct a slide rule for the relation V^

k (cos

cos

/).

12.

Solve by means of the logarithmic slide rule the following equations:

(a)y
(c)

y'

+ 3y-7=o;
'f

(b)

y^
y^

6y

+y+5
-\-f

o;
1

-\- 1

0;
ie)

(d)

+y = o;

y^

- 3^ +

o.

CHAPTER

II.

NETWORK OF SCALES. CHARTS FOR EQUATIONS IN TWO AND THREE VARIABLES.

Representation of a relation between two variables by means of. A relation {u, v) = o between two variables u and V may be represented by means of two perpendicular scales instead Construct the scales x = mif{u) and y = m2F{v) of two adjacent scales.
II.

perpendicular scales.

where f{u) and F{v) are any functions of u and v, on two perpendicular axes OX and OY, and through thepoints marked on these scales draw Any pair of values of u and v perpendiculars to the axes (Fig.(lio). that satisfy the equation (w, v) o will determine a point, viz., the
intersection of the corresponding perpendiculars to the axes; thus the

pair of values

Ui, Vi will

perpendicular to
V

OX through u =
all

correspond to the point of intersection of the F through Ui and the perpendicular to

is

vi.

The

locus of
(m, v)

such points

a curve which

is

said to represent

the relation

=
x

o.

The

rectangular or Cartesian equation of this

curve referred to the ax es

tions of the scales,

OX and OY may be found by solvingjthe equa,

X and

y,

mif(u) .a nd y = m2 F(v) for u and v in terms of and substituting these values in (w, v) = o.

Art. 12

SOME ILLUSTRATIONS OF PERPENDICULAR SCALES

Vk

21

curve in the point P, and read

to

at the foot of the perpendicular

from

P P

OF.
\g(\io]

On a

transparent sheet, draw two perpendicular index

lines,

and moves along the curve, keeping the index

/
1^7

intersecting in a point P.

Slide the sheet so that the point

lines parallel to the axes; then,

Vh-

when

lu cuts

OX

in Uk,

will cut

OF in

(It is

a simple mechanical

matter to keep the index lines parallel to the axes.)

Fignic)
strip.

Draw

the scale x

W]/(m) with axis O'X' on a transparent

^ITde the strip perpendicular to

Vk-

F until
Vk

the point Uk

falls

on the

curve; then at 0' read

In any case, the interpolation of Uk and

on the u and

v scales is easily

done by
12.
(i)

sight.

<

Some

illustrations of perpendicular scales.

v

Consider the relation

u^.

If

we

construct two uniform scales

y = mv on OX and F respectively, and draw perpendiculars to the axes through the points marked on the scales, we shall have the (Fig. ,i2a^ rulings of an ordinary piece of rectangular coordinate paper. Here, v = u^ will be represented by the locus \yhose Cartesian equation We plot this curve from a table is y/m = x^/m'^ or x'^ = my, a parabola.

= mu and

Note that we could with different moduli have constructed the scales x y Wi and W2, but the corresponding Cartesian equation m^x^ = mi^y still represents a parabola.
of values of

u and

v satisfying the

equation

= miu and

= u^. m^v

22

NETWORK OF SCALES
set
niiu^ are

Chap. II

one such scale X =

and the slope, 2, of the Hne. Note that the points on the marked with + and values of u.
In the former, the straight line
it is is

It is evident that the representation in Fig. 12b is

that in Fig. 12a.

much simpler than much more easily

constructed and every point on

definitely determined, while, in the

latter, the curve between plotted points is only approximated. Of course, it is easier to interpolate on the uniform M-scale in Fig. 12a than on the non-uniform w-scale in Fig. 126. Furthermore, in Fig. 126, we can project the point in which the representing line cuts f = 10 vertically on = o and thus draw a second section of the line parallel to the first seci;

tion, for

which

ranges from 10 to 20; this process

may
is

again be used to

get further sections of the line.

A
(3)

third representation of the equation v

u^

given in the next

this In

article.

&%2
X

Consider the relation v If -f In a (In t; = logz;).

ae~^"'.

We

can write

i;

we

construct the perpendicular scales

will
\-

miu^ and y

= nhln v,
is

our relation

be represented by the straight

This
line
is

line

whose equation

y = nh

mi u

\na.

easily con-r

structed

by means

of the points

o, v

Fig.(i2c^we have taken a

of natural logarithms
13.

6, b

\, nii

I a = a and = t v = -. In e = 0.2 in., W2 = 2 in. A table

was used

to construct the scale

Logarithmic coordinate paper.

p, q,

Consider the relation u^v^ =

We
plogu

on OY.

a,
-\-

where
q log V

and a are any numbers.

and y

= log a. If we = m log v and draw

can write this construct the perpendicular scales x

the perpendiculars to the axes,

= m
we

log

shall

Art. 13

LOGARITHMIC COORDINATE PAPER

23

have the ruHngs of a sheet of logarithmic coordinate paper. In Fig/iO m = 25 cm. and u and y vary from i to 10. (Logarithmic paper can be constructed for larger ranges of the variables and with various Our relation will be represented by the straight line whose moduli.)
Cartesian equation ^
of
is

+ mm

:!c

log a,

which

may be

plotted

by means

two pairs
10

of corresponding values of

u and

v or

by means

of one pair

24

NETWORK OF SCALES
will

Chap. II

then represent the relation v = u^, where u varies from i to lo and V from i to lOO, and hence for all values of u and v since 4:he scale log u, for example, where u varies from lo^ to 10^+^ (p = integer) X = log u, where u varies from can be made to coincide with the scale x =

M'A,

must be determined independently in each case. In finding u when v is given, divide v into groups of two figures each beginning at the decimal point, as in arithmetic (see Art. 6 (4)) if the left-hand group contains only one significant figure, use
I

to 10.

The

position of the decimal point

OM, if it contains two significant figures, use the section M'A = 0.64, read u = 0.8, but when v = 0.064, read u = 0.253. (2) For the relation v = I iru^, the volume of a sphere 4n terms of its diameter, the representing straight line passes through the point u = 2, V = ^T and has a slope equal to 3; this gives the section BC in Fig.!a3^ We continue this line by projecting C into C on OX and drawing C^
the section
;

thus

when

parallel to B'C, then projecting

to

CD, and

complete this
parallel to
is

last section

D into D' on OX and drawing D'E parallel by projecting E into E' on OF and

drawing E'F

Our relation In values u and v.

D'E; F will project into the initial point B on completely represented by these sections for all finding u when v is given, divide v into groups of three figures each beginning at the decirnal point as in arithmetic (see Art. 6 (6)) according as the left-hand group contains one, two, or three significant

OX.

figures,
(3)

use the

first,

second, or third section, respectively.

For the relation u

v^-'^^

10,

where u
;

is

the pressure and v

is

the

volume

of a perfect gas, our first representing section,

HK,

passes through
is

the point u

10, v

with slope

1/1.41

the second section, K'L,

easily constructed

and these two sections

will serve for the variation of v

from

to ID.

If

the sections are continued, later sections will overlap

the preceding ones.

14.

Semilogarithmic
.

coordinate
are

paper.

Consider
We

the

relation

= p q^^ where p and q \ogv = ulogq -\- \ogp. If we and y = nh log v, and draw

can write this construct the perpendicular scales x = niiu

the perpendiculars to the axes,

any numbers.

we

shall

have the rulings of a sheet of semilogarithmic coordinate paper. In Fig. T4',; wi = W2 = 25 cm. and u varies from o to i while v varies from 0.1 to I. (Semilogarithmic paper can be constructed for larger ranges Our relation will be repreof the variables and with various moduli.)
sented
log p,
of

by the

straight line

whose Cartesian equation

is

= log 2 +

which can be plotted by means of two pairs of corresponding values The folv, or by means of one pair and the slope W2 log q/mi. lowing examples will serve to illustrate the use of semilogarithmic paper, -\(i) The relation v = o.i e^-^" can be written logy = (2.1 log e) u logo.i, where e is the base of natural logarithms, and is represented

u and

Akt. 14

SEMILOGARITHMIC COORDINATE PAPER

straight line (section

25

by the

OA

in Fig. 14),
i,

which passes through the

points u

o, y

0.1

and u =

o.Sij.

To

extend the range of

our variables
tally into

we need not extend

the chart but merely project

horizon-

A' on OY, draw A'F' parallel to OA, project F' vertically into F" on OX, and draw F"E" parallel to OA ; then for OA, u varies from o to I and V from o.i to 0.817, while ior A'F' and ^"E", u varies from i to 2 and V from 0.817 to 6.668.
/.o

o.s

0.6

o.s

0.4

(V)
0.3

0.2

26
'

NETWORK OF SCALES

'

Chap. II

tion V = p q", and run our eyes along this line until we find the point where u and v are equal this is the required value of v. Thus to find the solution oi v = 0.6 e~-^^^ ", we draw the line BC representing the equation v = o.6e~9i" and run our eyes along this line (watching the u and v scales) to the point D where we read u = v = 0.46
;

(Fig. 14)-

draw the
equation

Again, to find the solution of the equation log line OA representing the equation log i'
is

t;

0.912 u

0.912

i,
i

we

[this

equivalent to the equation considered in

(i), since 2.1 log e

0.912 and log 0.1

run our eyes along this line until we read u = V = 0.132. We can find another value of v satisfying the equation by running our eyes along the section F"E" until we read u = v = 1.17. 15. Rectangular coordinate paper the solution of algebraic equations of the 2d, 3d,

= i] and

and 4th degrees.

We may use the rulings of a sheet

= u and
circle

of rectangular coordinate paper to solve graphically algebraic equations of the 4th, 3d,

and 2d degrees.

Let the scales he x

y^

=
{x

v
^

where
4*

the modulus
(i)
If

is i

we draw
r^

the parabola

2 x

and the

hY

with center at {h, k) and radius r, the ordinates of their points of intersection are found algebraically by eliminating x between From the these two equations and solving the resulting equation for y. first equation v/e have x = y^/2, and substituting this in the second equaiy

ky =

tion

we

get I-

hj

(y

ky =

r^,

or

y + 4 (i If

h) y^

S ky

-{-

4{h^

+
t

k^
is

r"^)

o.

we
get

divide this last equation

by

t^,

where

an arbitrary number,

we

or

z^

-{-

az^

-\-

bz

-}-

=
,

o,

where

-, a

=
j^

= -

Conversely, the real roots of the equation ^ -\- az^ -\r bz -\- c = o are found by measuring the ordinates of the points of intersection of the parabola y^ = 2 x and the circle with center at

4-.. ._ -5- and radius r = h^ k^ ..__.........^,,,._= y

,

fe

-\-

where

The

an arbitrary number, and dividing these ordinates by /. t allows us to throw the center of the circle to a Y, or as near to or as far convenient point always to the right of
t

is

introduction of

Akt. is

THE SOLUTION OF ALGEBRAIC EQUATIONS

27

from the vertex of the parabola as is convenient such that the circle will not cut the parabola at an angle too acute for accurate reading of Note that the one parabola, y^ = 2 x, will serve for findthe ordinates. ing the real roots of all equations of the fourth degree (Fig. 15), To /w^ g^^ r& + 5 = o, we first substisolve the complete equation t/^

-2

28

NETWORK OF SCALES
and the other points
2^ -\-

Chap. II of intersection will

origin or vertex of the parabola,

Hence, the real roots az -\- b = o are found by measuring the ordinates of the equation ^ of the points of intersection of the parabola y^ = 2 x and the circle with
give the real roots of the cubic
az
-\-

o.

center at h

4.

at'^
'

bt^

^^"^''^
"x"

^" arbitrary number) and

passing through the vertex of the parabola, and dividing these ordinates by /. Note that the one parabola y^ = 2 x will serve for finding the real
roots of
all

fourth and third degree equations (Fig. 15).

v^

To

solve the

complete equation

pv^

-\-

qv

-{-

o,

we
-\-

first

substitute v

/)

and

this

equation takes the form

2.

z^ -\-

az

o,

and then proceed

z^

a."

above.

Example

Let us find the

of the circle
1.88,
is

real roots of the

equation

33
-'if

=0.

Here the center

at h

=
4

^^

=-=
8

i.

We

read z

y
3.

1.53,

0.35 (Fig. 15).

Let us find the real roots of the equation v^ t^v"^ -\- v = 0. Let z = V -\- i; then the equation becomes 2^ 2 2 5 = o. Here 4 the center of the circle (Fig. 15) is at h = i -\- ^ f = 2,, k = ^ fi = 5, We read 3' = 4.18; hence 2 = y/t = 2.09, and v = z i = 1.09. if / = 2. If we draw the parabola y^ = 2 x and the straight line y = mx k (3)

Example

of slope

and

3'-intercept k, the ordinates of their points of intersection

are found from the equation y^ ^

my

22k =0.
-\

If

we
f

divide this
-^

by

^ (where
z2

is

an arbitrary number), we get (yj

o,

yj H

o or
the

+ as + 6 =

where

=
2^

real roots of the

equation

= + a2 + 6
'V

2
:

7.

2 ^= ^i^

Conversely,

ordinates of the points of intersection of the parabola

straight line of slope

o are found by measuring the y = 2 x and the

and

^'-intercept ^

bt

(where

is

an arbitrary number), and dividing these ordinates by t. Note that the one parabola y^ = 2 x will serve for finding the real roots of all fourth, third, and second degree equations (Fig. 15). Example 4. Let us find the real roots of the equation :? 1.452
5.6 ^

o.

Here the slope of the ^

line is

m=
i.

22
=
read 2

. ,

145^

145

and

its ;y-in-

tercept \sk
3.20.
16.

145^

^^ =

3.86,

if ^

We

3/

= - 175 and

of perpendicular scales.

Representation of a relation between three variables by means An equation in three variables of the form

Art. i6

RELATION BETWEEN THREE VARIABLES

29

by generalizing the method an equation in two variables. Thus, if we assign to wa value, say Wi, we shall have (t){u,v,Wi) = o, an equation in two variables u and v, which can be represented by the method of perpendicular scales, x = mif{u), y = miF{v), as a curve; By assigning to w a succession this curve is marked with the number "Wi.
(f){u,

v,w)

=0 can be represented

graphically

employed in Art. 11 for the representation of

of values, Wi, Wi, Ws,

we

get a series of representing curves, each

marked with
ables
is

its

corresponding value of w.

The equation

in three vari-

network of curves. It is evident that the same equation (f){u, v,w) =0 can similarly be represented by a network of curves found by assigning a succession of values to u (or v) and marking each curve with its corresponding value of u (or v).
said to be represented
this

by

(W

30
until

NETWORK OF SCALES

Chap. II

OX
4,

passes through v

^,

and at

its

intersection with the curve

w =
best,

we

read u

3.7.

The

task of

drawing

the w-curves is often very great,

to

and

it is

therefore

whenever convenient,

choose the scales for

u and

v so that the repre-

senting w-curves are straight lines.

This will not only lessen the labor of construction but will evidently increase the accuracy of our charts.
17.

Charts

for

multiplication

and

division.

This

example

will

illustrate

how

the choice of scales determines the nature of the represent-

ing curves.
(i)

The equation uv = w can be

for

represented

= mv

our two perpendicular

scales,

by taking x = mu and and drawing the corresponding

4.^
Fig. 170.

^.'^

network. The equations of our representing w-curves are of the form xy = mhv, a set of rectangular hyperbolas (Fig. 17^) which are quite difficult to draw and for which the interpolation is very inaccurate. (2) The equation uv = w can be represented by choosing x = mu and y = mw/ ioioronr two perpendicular scales. The equations of our representing z/-curves are of the form y = y^r/io, a set of radiating straight lines of slope v/io. Fig.' 17^ illustrates this chart, where u and v vary from I to 10. The valuesljf v are placed at the end of the representing lines. For saving of space, the values of w are placed at the points

Art. 17

CHARTS FOR MULTIPLICATION AND DIVISION

31

where the horizontals cut the Une f = 10 and the rulings above this Of course, the position of the decimal point diagonal need not be drawn. in the value of a variable may be changed with a corresponding change The great disadvantage of this chart is that the z;-lines in the result.

32

NETWORK OF SCALES

Chap. II

over a range from o to lOO or, in general, from o to lo*, similarly for the range of v, with corresponding change in the range for w. (3) The equation uv = w can be represented by writing it in the log v = log w and choosing x = m log u, y = m log v for form log u our two perpendicular scales, i.e., we have the rulings of a sheet of log-

arithmic paper.

The equations

of our representing w-curves are of the

i

form X

-^

= m\og

w, a set of parallel straight lines (Fig.

line

yd)

These

parallel lines are

very easily drawn, for the

z;

w =

k,

for example, cuts

(or the 3'-axis) at

hence the line

the point on IQP to I0P+^ with corresponding readings in the range for w.

= ^, and cuts v = i (or the x-axis) at u = k; w = k is a line joining the point on OX marked u = k with F marked v = k. The ranges for u and v may be read from
illustrated in (2)

The methods
tion of

and

(3)

may be

extended to any equa-

the form f{u) F{v) = 4>{w). If we choose x = mif{u) and mz<i>{'w) for our perpendicular scales, then the equations of our repre-

senting y-curves have the form y

F{v)
=

x,
nh.

a set of radiating

lines.

But

if

we choose x = mi
'V -^

log f{u)

and y

log F{v) for our perpen-

dicular scales, then the equations of our representing w-curves have the

form

X Wi

f-

m^

log <^(w), a set of parallel lines.

The following examples

tions in three variables,

18.

Three-variable charts.

Representing curves are straight

lines.

illustrate the construction of charts for

equa-

where the perpendicular

lines.

scales are so chosen

that the representing curves are straight

(i) The equation w = V^a/S^ can be written 5 log w = log a + 4 log ^, and if we choose x = Wi log a, y = W2 log /3 for our perpendicular scales and let Wi = W2 = 10, the equations of our representing w-curves are of the form X + 4 >' = 50 log w, a set of parallel straight lines. These lines have the slope \. They are most easily constructed by noting that when a = /3 = ^, we have w = k also. We, therefore, draw a system of parallel lines through the points a = k, ^ = k with slope j and mark

w = k (Fig. 1 8a). = c, for adiabatic expansion of certain gases where /> = pressure and v = volume, can be written log p -f 1.41 log = log c. If we choose x = 10 log v, y = 10 log p for our perpendicular
these with the corresponding value
(2)

The equation

pv^-*^

scales, the

equations of the representing c-curves have the form

3'

1.41

=
V

10 log

c,

set of parallel straight lines.

structed

I,

by noting p = k there

that the slope

is

1.41,
=
,

These lines are easily conand that through the point

passes the line c

k (Fig. i8a).
for the elastic limit of rivet steel,

(3)

Consider the equation /

-!^

where

P is the actual load in pounds at the elastic limit, D is the diameter

Art.

i8

REPRESENTING CURVES ARE STRAIGHT LINES

33

of the bar,

and/ is the

fiber stress in

niiP,

= W2/ for our

pounds per square inch. If we choose perpendicular scales, the equations of the repre-

senting D-curves have the form y

lines.

^^

x,

a set of radiating straight

For Fig.

18b,

Wi

2 W2, and the lines were constructed

by means

/o

34

NETWORK OF SCALES

Chap. II

SSjooa

Art. 19

CHART FOR THE SOLUTION OF CUBIC EQUATIONS

Consider the equation

35
the dis-

(4)

h-

= where u and
,

v are

tances of an object and its image from a lens and w is the focal length of the lens, or where w is the combined resistance of two resistances u and v
in parallel.
If

we choose
;

= m/u,

= m/v

for our perpendicular scales,

36
z

NETWORK OF SCALES
0.4

Chap,
{p, q) falls

n
on

and

=
of,

0.5.

According as the point

outside

of,

the boundary

or within the triangular shaped region on the

left,

we

can read one, two, or three values of z, and the corresponding cubic equation has one real root only, 3 real roots two of which are equal, or 3

Art. 20

REPRESENTING CURVES ARE NOT STRAIGHT LINES

37
2^ -\- z

original equation are then 2

kz'.

Thus

to solve the equation

2'^

4.

o,

let z
ife

=
2,

kz',

and the equation becomes

get
2'^

+ +

-r^

2'

o;
z'

if

we

choose
If

=
2

we

0.25

z'

0.5

o, for

which we read
c

0.69,

and hence

1.38.

the complete cubic equation u^

first

+ clu^ +
2^

6w

is

given, this

must

stitution

be transformed into the equation a

+ ^2 + g

o by the sub-

3 In a similar way we may build a rectangular chart for the solution ^z -|- g = o, or for any trinomial equation of the quadratic equation 2^

z^

pz"

+g

o.

Representing curves not straight lines. chimney draft. Extensive researches have been carried (i) Chart for out by the Mechanical Engineering Department of the Massachusetts Institute of Technology to determine an equation expressing the draftof a chimney in terms of its height and the temperature of the flue gases. No simple relation between these quantities has been found. From the experiments performed, it was found that if Ti is the absolute temperature in degrees Fahrenheit of the flue gases measured 3 feet above the center of the flue (the lowest temperature point recorded), H2 is the height of the chimney in feet, and T2 is the absolute temperature in degrees Fahrenheit of the flue gases at the top of the chimney, then
20.

Three-variable charts.

n= 0.32 I: (ila

3) L\

J f)"-]
("-

Now

if

is

the draft in inches of water, with the outside air at a tem-

perature of 70 F., then

D=

0.192 ^0.075

^j {H2 -

3).

If the value of 7*2 from the first of these equations is substituted in the second equation, we shall have an equation in three variables, D, Ti,

and H2.
where
75

In Fig. 20a, our two perpendicular scales are x = niiTi and y nti = 200 m^, and the representing i?-curves are drawn for

W2Z),

H = 50,

300 ft. Thus, for a chimney 150 ft. high and for an absolute temperature of 1139.5, we read that the draft is 0.955 iri. of water. (2) Experimental data involving three variables are often plotted by

means
of

of a network of curves, and such a chart takes the place of a table double entry. Fig. 20& gives a chart useful in heat flow problems where the temperature difference is an important factor. The chart gives the difference between the temperature of pure water under various gage pressures and the temperature under various vacuums. (Corrections must be applied for solutions.) Let P denote the gage pressure in

38
lbs.

NETWORK OF SCALES
per sq.
in.,

Chap. II

T, the temperature difference in degrees Fahrenheit,

and

vacuum in inches. We first construct the perpendicular scales X = miP and y = nh.T (in Fig. 206, Wi = 2 W2) then the F-curves are constructed by means of a table, part of which is as follows:
F, the
;

Art. 20

REPRESENTING CURVES ARE NOT STRAIGHT LINES

is

39

Such a table
Thus, a
or 4.92

constructed with the aid of Peabody's

is

vacuum

of 25 in.

0.4912

Steam Tables. equivalent to a barometric pressure of 4.92 in. 2.42 lbs. per sq, in., and this gives a temperature of
is

I33.2; a gage pressure of 5 lbs.

(adding the atmospheric pressure of 14.7

equivalent to total pressure of 19.7 lbs. lbs.), and this gives a tempera-

Vacuum
28 Ji'
210
28'
27'/3'

PRESSURE

POUNDS

GAGE.

Fig. 2oh.

ture of 227. I

93-9-

we thus have a temperature difference of 227.! I33.2 In this way, we subtract the temperature for a 25-in. vacuum
; .
.

from the temperatures at gage pressures of o, 5, 10, lbs., and these corresponding values of P and T are plotted giving the curve marked V = 2^ in. The table is completed for various values of F = 26,
.

26^,

...

in.

and the corresponding curves are drawn.

The curve

40

NETWORK OF SCALES

Chap.

II

marked "atmospheric"
oration.
21.

gives the temperature difference for open evap-

OX

In Fig. 21a let Use of three indices. Hexagonal charts. and OF be perpendicular axes and let OZ be the bisector of tbeir

Fig. 2 1 a.

Fig. 2ib.

Fig, 21C.

angle.

From any

point

draw P/i, P/2,

Ph

perpendicular to
in C.

OX,

OY

and
cut

OZ respectively, cutting OZ in D, and draw AE perpendicular

OX\nA,OY\nB, OZ
to

OD, cutting OZ

Let Ph also Then in E.

OC = 0-|-Z)+PC = 0^cos45**+^Z>cos45 + PPcos45=^^;^^-

Art. 21

HEXAGONAL CHARTS
if

4I

Thus,
z

the axes

OX, OY, OZ carry

the scales x

mf{u), y

mF(v),

V2

p^ (t){w)

respectively, then the three perpendiculars

from any point

to these axes will cut

<^(w)

them

so that
/-

V2
thus,
indices
/i,

V2

or

f(u)

F(v)

<f>(w}

any equation
h,
I3

of this type may be represented by three scales. The may be drawn on a transparent sheet and this sheet is

Fig. 2id.

moved over

the paper keeping the indices perpendicular to the axes.

- + - = - by this method. U V w We can choose the modulus on OZ the same as the moduli for OX and OF by following the construction illustrated in Fig. 216. Here OX and OF cut at an angle of 120 and OZ bisects this angle. Join OP and
Fig. 21C charts the equation
let

angle

COP =

a.

Then

42

NETWORK OF SCALES

Chap. II

OA = OP COS (60 + a) = OP (cos 60 cos a - sin 60 sin a). OB = OP cos (60 - a) = OP (cos 60 cos a + sin 60 sin a). OA -\-0B = 2OP cos 60 cos a = OP cos a = OC. .: Thus our three scales are x = w/(w), y = mF(v), z = m<i>{w). Fig. 2id charts the equation uv = w or log u + log = log w by this method.
z;

EXERCISES.
1.

2.

Represent the equation v = u^ by a straight line, using natural Represent the equations (a) 2 u' 2,v^ = 6, (b) m^ 3 z;2 = j^

scales.

by

straight lines, using natural scales,

and

(c) m" _|_ ^,2 ^ find graphically the simultaneous solutions of

2

the three equations taken in pairs. 3. Find graphically the simultaneous solutions of the equations v
i

= 6e~^ and

10 e
4.

u2

Solve graphically the equation u = 6 e ^^. 5. Construct a sheet of logarithmic coordinate paper and draw on lines representing the relations:
(a)
V.

it

the straight

';

(b) V

u^;

(c)

-;
tr
(e)
.4
;

(d)
(f)

C = vD (circumference of circle); pi^-*^ = 2 (adiabatic expansion of a

h

=
y

D"^

(area of circle);

gas)

(g)

= =

^2

^2

TTT: (^ D4-4 2 g of water).

velocity head in

ft.,

velocity in

ft.

per sec. for flow

6.

Construct a sheet of semilogarithmic coordinate paper and draw on

it

the straight

lines representing the relations v

7.

0.2 e'-^"

and

0.85 (i.5)~3".

Solve graphically the equation u

0.2 e'-^".
In {pv'')

8.

Show how

to solve for

p the equation

-\

-^

pv

c,

where

a, b, c, k,

and

R are known

various values are assigned to v. 9. Construct charts for the relations V = wr^h and S = 2 irrh (volume and lateral surface of a cylinder) using parallel straight lines only. ID. Plot the equation y = 2 x" for various values of n, positive and negative, (a) as a set of curves, (b) as a set of straight lines. n. Plot the equation y = e"^ for various values of n, positive and negative, (a) as a
set of curves, (b) as a set of straight lines.
12. Plot the following experimental data for the relative humidity obtained by a dew-point apparatus, using the wet bulb temperature, degrees Fahrenheit, as abscissas and the dry bulb temperature, degrees Fahrenheit, as ordinates.

constants,

when

WET BULB TEMPERATURE

Dry Bulb

(DEG.

P.).

EXERCISES

43

13. Construct a chart similar to that of Ex. 12, using the difference between wet and dry bulb temperatures as abscissas and the dry bulb temperature as ordinates. Which is the better representation, that oi Ex. 12 or that of Ex. 13? 14. From the chart of Ex. 13, with the aid of graphical interpolation, form a table

giving the relative humidity for dry bulb temperatures of 40, 50, ference between wet and dry bulb temperatures of 0, 5, 10
set of curves for the difference of

110,

45,

and difand draw the

temperatures using the dry bulb temperature as ordib.

nates and the relative humidity as abscissas. 15. Plot the curves y = a sin bx for various values of a and
16.

The Brake Horsepower

of

an engine (B.H.P.) with n cylinders

of d in. diameter,
is

according to the rating of the Association of Automobile Manufacturers,

given by

B.H.P.

<Pn
.

Plot the representative curves for

in., (a)

2, 4, 6, 8, 12, letting

d range

from if
17.
lbs.

in.

to 5

using rectangular coordinate paper, (b) using logarithmic coordiof

nate paper.

The volume,
in.

per sq.

V, of one pound of superheated steam which has a pressure and a temperature of T degrees, is given by (Tumlirz's formula)

V-= 0.596 p
18.

0.256 cu.

ft.

Plot representative lines (a) radiating, (b) parallel.

Solve the following equations

(o)
(c) z3

by means

of parabola
(ft)

and
12 z

circle.

(Art.

isO

+ 3z -I-

z - 3 2^ I (e)2*+2-i=o;

= o; = o;
by means

3' z*

+z + 5 =
-32^

0;

(<f)

(/) z*

+ 7 = o; + 3 = o.

19.

Solve the following equation

(c) z3
(c)

of the rectangular chart of Art. 19.

(&)
2^

- 2^ -

+ 3z -

o;
I

62

+ =

o;

(d) 2^

+ z + 5 = o; + + 2 - = o.
z2
I

CHAPTER

III.

NOMOGRAPHIC OR ALIGNMENT CHARTS.

22.

Fundamental

principle.

The

methods employed

in

the pre-

ceding chapter for charting equations are very useful in a large

of problems in computation, but they

number
(i)

have certain disadvantages:

is

the labor involved in their construction

great, especially

when

the rep-

resenting curves are not straight lines; (2) the interpolation must largely be made between curves rather than along a scale, and thus accuracy is
sacrificed; (3)

the final charts appear very complex, especially

if

the

methods are extended to equations involving more than three variables. The methods to be explained in this and the following chapters are applicable to a large number of equations or formulas and possess certain distinct advantages over the previous method: (i) the chart uses very few lines and is thus easily read; (2) interpolation is made along a scale rather than between curves, with a corresponding gain in accuracy; (3) the labor of construction is very small, thus saving time and energy;
(4) the

chart allows us to note instantly

the change in one of the variables due

to changes in the other variables.

The fundamental
ment charts consists
of
/(m,
V,

principle involved iiT

the constnifction of nomographic or alignin the representation

an equation connecting three variables,

w)

o,

by means

of three scales
lines) in

along three curves (or straight

such a manner that a straight

the three scales in values of u,
satisfying the equation.
Fig. 22.

line cuts
v,

and

The transversal

We shall

now make a study of some

vis called an isopleth or index line (Fig. 22)^ of the equations which can be rep-

resented in this way, and of the nature and relations of the scales representing the variables involved.*
nomographic or aJignment charts developed by M. D'Ocagne in his "Traite de Nomographie." Further references may be given to "The Construction of Graphical Charts," by J. B.
have been most
fully
*

The

principles underlying the construction of

Peddle; "Nomographic Solutions for Formulas of Various Types," by R. C. Strachan (Transactions of the American Society of Civil Engineers, Vol. LXXVIII, p. 1359), and
to various smaller articles that have appeared from time to time in Engineering Journals.

44

AlKT.

23

CHART FOR EQUATION

(I)

45

(I)

EQUATION OF THE FORM /i(u) +/2(v) =Mw) or /i(u) /2(r) =/3M. THREE PARALLEL SCALES.

[The second form of equation (I) can Chart for equation (1). immediately into the first form by taking logarithms of both be brought
23.

members;
Let
base
line.

thus, log/i(w)

log/2(t')

log/sCw).]

AX, BY, CZ be

three parallel axes with

ABC any

transversal or

(Figs. 23a, 236.)

Draw any

index line cutting the axes in

^^
^EQ.:><>;

A^-

,^

_-Fig. 23a.

Fig. 236.

the points u,

'v,

respectively, so that

Au =

x,

Bv =

y,

Cw =

z.

How

are x, y, z related?
li

AC CB =
:
:

Wi

m^,

to

AB, then
'.Dv

the triangles
or

Ew

= AC CB
m^x

if through v and w we draw lines parallel uEw and wDv are similar, and Eu Dw = x z z y = mi m2.

and

+ m\y =

(mi

+ W2) z

or

X mi

mi

W1W2 mi + W2

Now
mi

if

AX, BY, CZ

carry the scales x

mifi(u), y

mzfiiv), z

m^'

-fz{w), respectively,

the last equation becomes /i(w) -{ f^iv)

fz{w),

and any index

line will

cut the axes in three points whose correequation.

sponding values

u, v,-w satisfy this

We
X

also note that for the equation fi{u)

fi{v) =

fz{w), the scales

mifi{u)

and y

m^fiiv) are constructed in opposite directions, as

Uiv)

in Fig. 236.

Hence
(i)

to

chart equation (I) /i(w)

parallel lines (x-

-\-

fsiw), proceed as follows:

Draw two

and

3'-axes)

on these construct the scales x

mifi{u) and y

any distance apart, and m^fiiv), where Wi and

46

NOMOGRAPHIC OR ALIGNMENT CHARTS

of the u-

Chap. Ill
y-scales

W2 are arbitrary moduli. The graduations start at any points on the axes.
(2)

and

may

Draw a

third line (z-axis) parallel to the x:

and

3'-axes,

such that

(distance from x-axis to z-axis)

(3)

(distance from z-axis to >'-axis)

= mi

m^.

Determine a starting point

for the graduations of the w-scale.

This

B A

{y

may be the point C (z = o) cut out by the line from A {x = o) to = o). If the range of the variables u and v is such that the points

do not appear on the scales, a starting point for the w-graduanevertheless be found by noting that three values of u, v, w satisfying equation (1) must be on a straight line; thus, assign values to u and V, say Wo and Vo, and compute the corresponding value of w, say wo, from equation (I) mark the point in which the line joining u = Uq and V = Vo cuts the z-axis with the value w = Wo and use this last point as a
tions

and

may

starting point for the w-graduations.

(4)

From

the starting point for the w-graduations, construct the scale

z

...
mzfz{w)

W1W2 -7 /3(w). mi + m2
r r

General remarks.

In practice the index

lines

need not be drawn; a

straight edge or a transparent sheet of celluloid with a straight line

scratched on
i.e.,

under side or a thread can serve for reading the chart, one of the variables when two of them are given. The distance between the outside scales and the moduli for these scales should, in general, be so chosen that the complete chart is almost square. Then any index line will cut the scales at an angle n ot less than 45, and its points of intersection with the axes is more easily noted and It is the corresponding interpolation on the scales is more accurate. rarely necessary to choose the moduli so that the length of the longest
its

for finding the value of

scale greatly exceeds 10 inches.

Charts of logarithmic and uniform scales similar to those described have been used in laying off the scales needed in the construction Much time and energy have been of most of the charts which follow. saved thereby. For greater convenience, the modulus of the primary or left-hand scale was taken to be 10 in. instead of 25 cm. In laying off the w-scale with the help of these charts, the following procedure will increase the accuracy of the construction. Assign two or three sets of values to u and v, and compute the corresponding values Draw the of w\ let these be (mq, z'o, wq), (mi, I'l, Wi), and {u^, V2, Wi). index lines (mq, Vq), (mi, Vi), and (w2, ^2), and mark the points in which these Fold the chart lines cut the z-axis with the corresponding values of w. along the scale with modulus ms, and slide this scale along the z-axis until the points of the scale numbered Wq, Wi, W2 practically coincide with the like-numbered points on the axis. This procedure is especially important when the modulus, mz, is quite small.
in Art. 3

The

cuts in the text are reductions of the original drawings.

Aet. 24

CHART FOR MULTIPLICATION AND DIVISION

47

Chart for multiplication and division. The equation v = w. log v = log w, we have an equation If we write this equation as log u Let u and v range /o~, of the form (I). from I to 10; then w ranges from i
24.

to 100.

Construct (Fig. 24a), 10

in.

apart, the parallel scales x

and y =^m2 log W2 = i i, the 2-axis is mid\^ay between the x- and y-axes. The line joining w = i and v = i must cut = i, and using this last the z-axis
10 log u
:

i*

= nti log u = 10 log v.

Since Wi

mw
=

point as a starting point,

the scale z

construct
5 log w.

W1W2
(wi

+ W2)

w =

I-

Fig 240.

The index
V

line in the

3.

w =

21.

completed chart (Fig. 246) gives the reading m Since the u- and y-scales are logarithmic scales, we
100-1

= 7, may

JO

9
6

908070 60-

7
6
5-

50-

40-

k
30-.
20-.

4-.

-4

ft/}

fWJ
3-.

(V^

ro-

9: 87-

6-

5
i-

-2

2-

/J

7-J

CHART

FnOR

MULTIPUCATION AND D/VISION UV^W

Fig. 246.

48

NOMOGRAPHIC OR ALIGNMENT CHARTS

Chap. Ill

read these scales as ranging from lo^ to iqp+i where p is any integer, with a corresponding change in the position of the decimal point in the value of w.
25.

Combination chart for various formulas.

(I),

As a further
fi{v)

illustra-

tion of product formulas of the type

/i(w)

>

Ji{w),

we

shall

now

represent several such formulas in the

same

chart, with the

same

outer scales but varying inner scale.

Our chart (Fig. 25) represents six formulas and undoubtedly others could be added. In all cases, if Wi and m2 are the moduli of the scales on the x- and 3'-axes, then the modulus of
the scale on the 2-axis
2-axis is
is

ma

mirrh/inii
:

m^, and the

position of the

determined by the ratio mi nh. (i) U' V = w for multiplication and division. This has already been charted in Art. 24. The equations of the scales are
a:

=
i
:

10 log w,
i.

10 log

v,

5 log w,

and mi

mz

The

index line gives the reading w

3, u

5,

w =

15.

(2) v'uTij* = w occurs in the McMath "run-off" formula. equation can be written log 4 log w = 5 log w and hence

The

X
Let mi

= mi log u,
10

y = m2 (4 log v), and m2 = 10/4, then ms =

2
2,

= ma

(5 log
:

w)

and mi

m2

i.

The

equations of our scales are

10 log w,

lologi',

10 log w.

A
V

starting point for the w-scale

then

w =
=

I,

is found by noting that when u = i and and by aligning these three points. The index line

gives the reading

(3) pi^'*^

3, u

5,

w =

4.5.

c gives the pressure-volume relation of certain gases under adiabatic expansion. The equation can be written \ogp 1.41 log V = log c, hence

x
If

mi\ogp,

1. 41

we choose mi = 10 I. The equations

:

= m2 (1.41 logz;), 2 = m3logc. and m2 = 10/1.41, then ma = 4.15 and nii nh =

y
:

of our scales are

x=io\ogp,

=
is

lologv,

4.15 log

c.

A
V

starting point for the c-scale

found by noting that when p

and

then

gives the reading

(4)

V=
mi

and by u = 2>, v = ^, w = 29. 0.785 D^H, the volume of a circular

I

aligning these three points.

The

index line

cylinder.

The equation

can be written 2 log

Z)

+
y

log

(log

log 0.785), hence

X
If

(2 log D),

we

choose mi

= mi log H, z = nis (log V log 0.785). and m2 = 10, then ma = 3.33 and mi m2 = i
:

2.

r
/Ut.
25

COMBINATION CHART FOR VARIOUS FORMULAS

49

a0)

y^

A-

H JO S3
lA

VA

N
il
I I I
I

g
I '
I

0,

00
1 1
I

N
1

<o
1.
I
I I

III

I I

II

TT
I I I I

*An/

cO PZSO'A.
I 1
I

l\
L
I

3a3HdS JO si/vniOA
II
I I

11

I,

...

V<?

'?

"?

y^

i
i

Hi

|i

II

1
1 I

o=,^.,Acf

s\^o

n jo^umoA
I i

3iinss3tid

"^

I'

'

I"'

'
I

"
I

'

<

'

'

'

'

1'

'

'

'

'

'

'

I'

' I

I I

I'

Miiiiiii

4\= A X ny
\

JO j.ono
'

^
"^^

>

^j HI

1'l

'II

J'

' '

'i

'
' '
I

'll

>

'

1 '

1 1

'

'

'

nn

'HfO

SB0-A-V3aNllAC\ UVlOOy/O JO 3MimOA

\
'

"

'

'

MM

1 1

\i

^Hzaf9-0 03SM3J
\

-J-jnO NI AJ.llNVnb

\
I"" "
I

1
1

1 I 1

'

' ' I '

|i""" ii|

""

I I

"
1

I 1

1 '

I 1

' I

I '

1 I

1 '

I 1

'

'

'

'

'
I

'

r-

a-d-n-a JO ssmtTA

50

NOMOGRAPHIC OR ALIGNMENT CHARTS

equations of our scales are

Chap.

Ill

The

iologZ>,

lologiY,

3.33 log V,

where we have discarded the expression 3.33 log 0.785 in the value of z; this may be done since this expression merely helps to determine a starting point for the F-scale; thus the point
2

F =

is

at a vertical distance

= 3.33

log 0.785

from the base

line

AB

of Fig.

23a or 236.

We shall

however determine a starting point for the F-scale by noting that when D = I and = i, then F = 0.785, and by aligning these three points.

D = H = 5, V = t,^. ^ = 0.524 D^ the volume of a sphere. The equation can be written F = 0.524 D'D^or\ogD-\-2 log D = log V - log 0.524, hence X =^ mi log D, z = m^ (log V log 0.524). y = nh (2 log D),
The index
(5)

line gives the reading

2,,

If we choose mi = 10 and m2 = The equations of our scales are

5,

then mz

3.33

and mi

m^

i.

10 log

Z),

10 log D,

3-33 log ^.
in the value of
2.
i,

where we have discarded the expression 3.33 log 0.524

We
V=

find a starting point for the

scale

by noting that when

i,

D =

0.524 and

we
D2

align the three points

D =

D=

i,

and

F=

0.524.

(6)

Q =

6.3

Vh
If

second which

flows

under a head
(log

H feet.

Q log

6.3).

gives the quantity of water, Q, in cu. ft. per through a pipe having a diameter D ft. when = The equation can be written 2 log D + I log we choose Wi = 5 and W2 = 20, then W3 = 4 and

200^
f54h-t

Wi
^^

W2

4.

The equations

of

our scales are

y
'006-^
,

x=io\ogD, = lologfl", 2 = 4 log Q.

Again we discard the expression log 6.3 in the value of 2, and

find

a starting point for the ^-scale by = i, noting that when D = i and

then

Q =

6.3

and by aligning these

three points.

The
3,

index line gives

the reading

D =

H=

5,

^04
Aiff-*

j-> -20

26. Grashoflf's

formula
D^Pi^-^''

M?

Q = 127. = 0.0165

for the steam in pounds per second flowing from a reservoir at pressure Pi pounds per sq.
CLOOi-

^Pi0 97

0.01296

Fig. 26a.

weight,

n;,

of dry saturated

through a standard converging orifice of A diameter D in. to a pressure of P2 pounds per

in.

sq. in. or circular orifice of

sq. in.,

if

Pi

0.6 P2.

ZX)0-{
ISJO;

iso-

iO'O3.0 84) 7S> 6.0

r300

o
250'

5U)-

4.0-

3,01000.90-

-200
2.0<o

0.80t.O

-ISO

^0.70^

0.9 0.8 0.7

0.6

0.S
0.4- -\

^0.50-i
It]

03
0.2

100
-90

^0.40^
0.1

-80
-70

k
0.30-

0.09 0.08 0.07 0.06

60

0.0S
Si

0.040.03-

-SO

a02
o.zo-\

40

o
k
0.010J>090.0060.007-

o
30

0./S

0.006
0.005-

0.0040.003-

a/<H
O.OOZ-'

20

GRASHOFF-S FORMULA W0.0/6S A

Fig. 266.

0.97
P^

51

52
If

NOMOGRAPHIC OR ALIGNMENT CHARTS

we
write the equation as

Chap.

Ill

2 log Z)

+ 0.97 log Pi =
(I).

(log

w
=

log 0.01296)

we have an
X

equation of the form

The
z

scales are
nis (log

= mi{2 logD),

= m2

(0.97 log Pi),

w-

log 0.01296).

Let

vary from

o.i in. to 2.0 in.,

then
if

to log 2

0.301, a range of 1.301;

:)c

\ogD varies from logo.i = i we choose Wi = 5, the equation of

= 10 log D and the scale will be about 13 in. long. the D-scale will be Let Pi vary from 20 pounds to 300 pounds, then log Pi varies from log 20 = 1.301 to log 300 = 2.477, a range of 1.176; if we choose m^ = 10/0.97, the equation of the P-scale will be 3' = 10 log P and the scale will be about 12 in. long. Then W3 = minh/inii -f W2) = 3.37. The equations of our scales are now
X

10 log

Z),

10 log P,
3'-axes

3.37 log w.

Construct (Fig. 26a) the x- and

divide this distance in the ratio mi
:

in.
:

apart; the z-axis will

We have therefore from the x-axis and 4.04 in. from the 3'-axis. A starting point for the w-scale is found by aligning D = i, Pi = 100, and w = 1. 13. The completed chart is given in Fig. 26b.
m^
4.85
10.

drawn the

z-axis at a distance of 1.96 in.

27.

Tension in
.

belts,

T ~ ii

=
first

e""^^^^^^",

and horsepower
is

of

belting,

H.P.

-^
stress in

33)000

In the

of these formulas. Pi

the allowable

working

pounds per

in. of

width, or the tension in the tight side

of the belt; the value of this

may be

obtained either from the manu-

facturer of the belt or

by breaking a

piece in a tension machine; a suitable

Pi may vary from about 50 to 75 for and from 100 to 150 for double belts, a is the arc of contact in degrees of belt and pulley and may vary from 100 to 300. / is the coefficient of friction and is assumed (in this chart) to have the value 0.30 for leather belts on cast-iron pulleys. P2 is the tension in the loose side This formula may be written of the belt in pounds per in. of width.
factor of safety should be added.
single belts

log P2

log Pi

=0.01745 fa log e
(I).

or

log Pi

0.002274 a

log T^

which

is

in the

form
Pi,

The
y

scales are

m\ log

= W2

(0.002274 a),

nis

log P2.

from log 50 = 1.6990 to log 150 = 2.1 761, a range of 0.4771; if we choose Wi = 10, the equation of the Pi-scale will he x = 10 log Pi and the scale will be about 5 in. long. Again, a has a range of

Now

log Pi varies

200;

if

we

choose

ntz

-,

the equation of the a-scale will be

40 (0.002274)

Art. 27

TENSION IN BELTS
the scale will be 5
in. long.

S3

= 5V and
5.24.

Then m^ = mim^/iini
z
in.

+ nh)

The equations
X

of our scales are

lologri,

= iV
W2

5.24 log r2.

Construct (Fig. 27a) the x- and ^'-axes 3.75

divide this distance in the ratio Wi
:

apart; the z-axis

I
/

must
:

10

002274)

^^

11.
x-

We
axis

have therefore drawn the and 1.96 in. from the

z-axis at a distance of 1.79 in.

y-axis.
is

from the

90^

T-IOO

starting point for the T2 scale

and T2
is

found by aligning Ti = 80, a = 150.. = 36.5. The completed chart

given in Fig. 27^, and indicates the

reading Ti

80 pounds, a = 150, = 36.5 pounds. ^2 In the second of the formulas, S

^^150
36.5

806--

is

the distance traveled

by the

belt

in feet per minute,

and may vary

Ti

from 300 to 6000;

Ti

is

the

difference in the tensions

1.96'IJ9"vary from 10 to 200; horsepower which a belt of one inch width will transmit; then, knowing 10 the horsepower which we wish to Fig. 270. transmit we merely divide to get the width of the belt desired. The equation can be written

and may H.P. is the

\300

log (ri

T^)

+ log 5
y

log

H.P.

+ log 33,000,
z

which has the form

X

(I).

The
T2),

scales are

Wi log {Ti

nh log

S,

nis

log

H.P.

Now
of

log

(T*!
if

7*2)

varies from log 10

5,

to log 200

2.3010, a range

1. 3010;

we choose Wi =

the equation of the

scale will

(T'l

7'2)-scale will

be about 6.5 in. long. Log S varies from log 300 = 2.4771 to log 6000 = 3.7781, a range of 1.3010; if we choose nh = 5, the equation of the vS-scale wilt hey = $ log 5 and the scale will be about 6.5 in. long. Then W3 = WiW2/(wi W2) = 2.5. The equations of our scales are
be X
5 log (Ti
T2)

and the

5 log (Ti

-T2),

3;

=
i
.

5 log 5,

2.5 log

H.P.
this

Construct the x- and 3'-axes 4 distance in the ratio Wi W2 = i

:
:

in.

apart; the z-axis

must divide

A starting point for the H.P.

scale

is found by aligning Ti T2 = 10, 5 = 300, and H.P. = 0.091. The completed chart is given in Fig. 27c and indicates the reading Ti Ti =

100,

6"

1000,

H.P.

3.0.

54

NOMOGRAPHIC OR ALIGNMENT CHARTS

Chap. Ill

NIlAl

y3d 133J Ni (S)1133 JO a33d9

I
I I I . I I 1 1 I
I I I

I I

/
/
!?

/
I

I i

"" ii "

/
i 1 1 1 1

III

/l

i""i""i

nil

II

^<,
ui!?
*)

C,
<0

/ H2a/M JO HOMI
/
inii
I
I I

iJ3d 'dH

^^

I I

ji

II

I'lMimM. M

S33y930 N/
1
1 1
1

(-o)

lOVlNOO JO
1

OifV

|
1 1 1

II

\
II
I I I
I I

\ i\m
I

iiniiii

II

|i

i|

^ O

HiamHONI d3d 'Sff7(^J.)3di^ 39001 NO NO/SA/31

"

\
(

ll
|

ll

l[1lll|lll

I I

II|

I|" HIIII
I

|I

I I

si

<

>

^HJ.a/M

HONI dJd S&KltJa/S IHOIl NO N0ISN31

Art. 28

CHART FOR EQUATION

/i(u)
.

(II)

55

(II)

EQUATION OF FORM
or

fi{u)'f2{v)'Mw)

'

= Ut) + f2{v) + Mw) + FOUR OR MORE PAR=Ut).

ALLEL SCALES.

[The second form of equation (II) 28. Chart for equation (II). can be brought immediately into the first form by taking logarithms of = log/4(/).] both members; thus log/i(w) logfsiw) ^ogfiiv) Equation (II) is merely an extension of equation (I) and the method of charting the former is an extension of the method employed in charting

the latter.

For
tion
in

definiteness, let us Consider the case of four variables

the form fi(u)

is

+ fiiv) + fsiw) = U{t).

form
(I)

This equation

in the

and the equaLet fi{u) -\- hiv) = q. and can therefore be charted by means

tt^f
<>~-(6Ly/
~<iv.

6w,

(U. V)

^w.

<5>

Fig. 28c.

Fig. 286.

of three parallel scales,

28a.)

but the g-scale need not be graduated. (Fig. = fi{t), which is also in the form (I) -\- fsiw) and can therefore be charted by means of three parallel scales one of which is the g-scale already constructed. The graduations of the u-, v-, and w-scales may start anywhere along their axes, but a starting point for the graduations of the /-scale must be determined by a set of values u = uo, V = Vo, w = Wq, t = to satisfying equation (II) thus, join Uo and Vo by a straight line and mark its point of intersection with the g-axis; join this point with Wo cuttfng the /-scale in a point which must be marked

We

then have q

/o;

this last point is

then used as a starting point for constructing the

t-

one joining points on the u- and y-scales, the other joining points on the wand /-scales, intersecting the g-axis in the same point. Fig. 28a illusscale.

To

read the completed chart

we

thus use two index

lines,

trates the position of the scales. It is thus easy to find the value of any one of the four variables when the other three are known. The extension of this njethod to equations of the form (II) containing more than four variables is obvious.

56

NOMOGRAPHIC OR ALIGNMENT CHARTS

Chap. Ill

Considering again the case of four variables, /i(w) fi{t), we can write this in the form fi(u) +/2(^) = fiit)
chart each of the equations /i(w)
-\-

+ fziv) + fsiw) =
fz{w) = q and fz{w) = q hy

fiiv)

q and

fi^t)

means of three parallel scales, with the g-scale (v/hich is not graduated) in common. Again, to read the cl art, we use two index lines, one joining points on the u- and w-scales and the other joining points on the w- and
/-scales, intersecting

the g-axis in the Scime point.

Fig. 286 illustrates

the position of the scales in this case.

29.

Chezy formula

c y/rs.

Here,
c is

for the velocity of flow of water in

open channels,
is

v is the velocity of flow in ft.

,

per sec, r

the hydraulic

radius in
surface,

ft.

(area divided

and

(See Art. 53 for

by wetted perimeter) 5 is the slope of the water a coefficient depending on the condition of the channel. the construction of a chart computing c by the Bazin
r

formula.)
to 20

ft.

Let our variables range as follows: s from 0.00005 to o.oi, Writing the equation ft., c from 10 to 250.
^ log 5

from

o.i

^ log r
(II).

+ log c

log

t;

we have an equation of q, we can write

(i)

the form

Introducing an auxiliary quantity,

ihgs
= Wi

-\-^\ogr

and
first

(2) g

+ log c
z

=
=

log

i>.

We now construct a chart for

are

the

of these equations.
nf^q.

The

scales

Now log 5 varies

is
:x:

(^ log s), y = mi{\ log r), from log 0.00005 = 5.6990 10

if

to log o.oi

8.0

10,

a range of 2.3010; and

we choose

Wi =

10, the equation of the 5-scale

in.

5 log 5

and the

scale will

be about 11.5

long.

Again, log

r varies

i. 3010, a range of 2.3010; and if we choose from log o.i to log 20 = 10, the equation of the r-scale is y = 5 log r and the scale will be W2 about 1 1.5 in. long. Then W3 = miW2/(wi -{- m2) = 5. The equations

= I

of our scales are

5 log

5,

5^ogr,

5q.

Construct (Fig. 29a) the x- and y-axes at any convenient distance, say 8 in. apart; the z-axis must divide this distance in the ratio Wi W2 = 1:1, and hence the z-axis is drawn midway between the x- and y-axes. The g-scale need not be graduated. We continue the construction by charting the second equation. The
:

scales are

ntaq,

m/^

log

c,

We

log 10

use the same g-scale as above so that = I to log 250 = 2.3979, a range of 1.3979; ^^^

= nit log v. W3 = 5. Log c

b
if

varies from

choose mt
-\-

5,

the length of the scale will be about 7 in. The equations of our scales are 2.5.
z

Then
&

nts

= mzmj{m3
V.

m^ =

2.

5 log

c,

2.5 log

Art. 30

HAZEN-WILLIAMS FORMULA

57

Construct (Fig. 29a) the a-axis at any convenient distance, say 10 in. from the 2-axis. The graduations of the c-scale may start anywhere
along the a-axis for symmetry, we shall place the scale opposite the middle The 6-axis must divide the distance of the scales already constructed. and it is therefore i the z- and a-axes in the ratio ^3:^4=1 between get a starting point for the y-scale midway between them. drawn
;
: ,

We

by making a
o.ooi, r

single

com5

0.01 qqi^

-.

putation; thus,

when
c

^20
r-^SO

=
v

and

=
r

100,

we have
join s

3.16; hence,

O.OOI

and

=
in

i,

II

15

cL

^^100

cutting
point,

the

g-axis

and then join and c = 100, point

this

cut-

ting the &-axis in a point

which must be marked

V

S"
i_/^

3.16.

Starting at this
4'
o.otxx?s

last point and proceeding

^-0.02
*-0.f

along the axis, the u-scale

is

graduated from v *

0.02

to v

^ Fig.

29a.

100.

The completed chart is given in Fig. 2gb. To read the chart we need merely remember that the {s, r) and (c, v) index lines must intersect on The index lines drawn in Fig. 296 show that when 5 = o.ooi, the g-axis. = I ft. and c = 100, then v = 3.16 ft. per sec. r Hazen- Williams formula for the velocity of flow of water in 30. The quantity of water discharged, pipes, V = CR^-^^S"^-^ (0.001)-'^. = 4 t:RW. The first of these formulas has been derived experimen^ tally by Hazen and Williams, who have also constructed a slide rule

is

for its solution.

V is
C

the velocity of discharge in

full;

ft.

per sec. from circular

ft.

pipes or channels flowing

cross-section divided
to length of pipe;

the hydraulic radius in

;

(area of

by wetted perimeter)
is

is

the slope or ratio of rise

condition of the inner surface of the pipe.

following table of values for C:
Brass, block tin, lead, glass

a coefficient depending on the material and the Williams and Hazen give the

Cast
*'

iron,

" "

very smooth new, good condition.

old,

"

"

"

140-150 140-145 125-135 ...100-125

. .

Cast iron, old, bad condition

Steel pipe, riveted, new " " " old

60-100
105-115 90-105 1 10-135

Masonry conduits

Replacing (o.oi)~^ by its value 1.3 18 and expressing R in inches instead of in feet, the formula becomes V = 0.2755 CR-^S-^, and this can be written as
0.63 log

R + 0.54 log 5 +

log

C+

log 0.2755

log

58

NOMOGRAPHIC OR ALIGNMENT CHARTS

Chap.

Ill

10-

9a-_

7-

6s-

.001-

0009-

oooa.0OO7-

9-e.7

6.59

J-

.0001.00009.ooooa.O0007-

.00006'
/tofxa'

AxT. 30

HAZEN-WILLIAMS FORMULA
of the
in.,

59

which
in.

is

form

(II).

to 20

5 from

o.oooi to 0.05,

Let our variables range as follows: C from 25 to 200.

R from o.i
q.

We

first

construct a chart for 0.63 log

R+

0.54 log

S =
ntzq.

The

scales are

= mi

(0.63 log R),

Now

logi? varies from logo.i

= m2 = i

(0.54 log 5),

to log 20

1.3010, a range of

5/0.63, the equation of the i?-scale will be 2.3010; if we choose = 5 log i? and the scale will be about 11.5 in. long. Again, log 5' varies X

mi =

from log 0.0001

0.05

10 to log

10, a range of 8.6990 2.6990; if we choose W2 = 5/0.54, the equation of the 5-scale will

he y = $ log 5 and the scale will be about 13.5 in. long. Then W3

WiW2/(wi
X

+ W2)
= = =

4.27.

The

equations of our scales are

5 log R, 5\og S,

y
z

4.27

q.
:!C-

Construct (Fig. 30a) the

and y-axes 11.7 in. apart; the 3-axis must divide this distance
in the ratio Wii
:

"
/,,_r
1

^-o.aoa/

mz

= -4~
0.63

Fig. 30a.

=
in.

5.4

6.3.

We

0.54 have therefore

in.

drawn the
from the

z-axis at

from the x-axis-iand 6.3

i

y-axis.

The

a distance of 5.4 q-scale need not

be,

graduated.

We now
The

continue the construction

2

by charting
log V.

+ log C + log 0.2755 =

=
nti (log

scales are
2

mzq,

C+

log 0.2755),

b
4.27.

= m^ log
Log C

V.

We use
log 25
nii

the same g-scale as above so that W3

varies from

1.3979 to log 200 = 2.3010, a range of 0.9031; if 4.27, the equation of the C-scale will he a = 4.27 log

we choose

C and
2.14.

the

be about 4 in. long. equations of our scales are

scale will
2

Then W5 = mzmJimz

W4)

The

4.27

q,

4.27 log C,

2.14 log V.

Construct (Fig. 30a) the o-axis at a distance of 4.8 in. from the 2-axis. Although the graduations of the C-scale may start anywhere along the o-axis, the C-scale is only about 4 in. long and we shall get a more symmetrical chart

by

placing the scale opposite the middle of the 5-scale.

6o

NOMOGRAPHIC OR ALIGNMENT CHARTS

Z>-axis

Chap. Ill

must divide the distance between the z-axis and a-axis in the W4 = i i, and is therefore midway between them. We get a starting point for the F-scale by making a single computation, thus, when R = 2, S = o.ooi and C = 100, we have V = 1.02 hence, join R = 2 and S = O.OOI, cutting the g-axis in a point, and then join this point and C = 1.00, cutting the 6-axis in a point which must be marked V = 1.02.
ratio

The

W3

OS

h04

st>n

/s

2O00

roon

100010

eo60.0/

9
676-

200JOO

.009

ooa 007
10

S
4-\

d Cj
:

.006

i^

.00s

.^90
BO
70

700^.
-

.6

6
.4

60SO 40-

-.O02 <0

I
.00s .006 .004
./ -

-oor

i
,

k.

-.0009

fS

I'"
^o.sA
2tOJO-A

08-

'-j0007^\
-.0006

-.aw5>i to r0004

210'.\
to'

/?-5
/fey;
/?

C
V
^.OOOk

o.'-'

HAIENAND WILLIAMS FORMULA FLOW OF WATER IN PIPES

V'CR''"S''"'O0O/'^

Q-AV'47TR^V.

Fig. 306.

by adding a scale for For circular pipes, we have Q = 4 ttRW, where V is the velocity of discharge in ft. and R is the hydraulic radius (one-fourth of the diameter of the pipe) in ft., ov Q = 0.0873 RW, where R is expressed in inches. We write this equation in the form
enlarge the usefulness of our chart
ft.

We shall

now

Q, the quantity discharged in cu.

per sec.

2 log

i?

-I-

log

V=

(log

(2

log 0.0873),

Art. 31

INDICATED HORSEPOWER OF A STEAM ENGINE

61

and our scales are

X

= Wi

(2 log

R),

m2 log V,
x

= W3

(log

Q
we

log 0.0873).

As we want to use the and the equation of the

b

i?-scale already constructed,

take Wi

=;

5/2,
to

scale

is

5 log R, as above.

We also want
= 2.14 and = 1.15, and

use the F-scale already constructed and hence take W2

get

2.14 log V, as above.

is c

Then W3 = m\mi/{mi

W2)

the

equation of the (2-scale

are 7.8 in. apart.

1.15 log Q.

In Fig. 30a, the x- and 6-axes

this distance in the ratio

The

c-axis

must divide

Wi

W2

2.5

tance of 4.2

in.

accomplished by drawing the c-axis at a disfrom the x-axis. We get a starting point for the Q-scale
2.14; this is
2,

by

aligning

R =
2

V=

1.02,

and Q =

0.36.

Fig. 306 gives the completec| chart.

The

index lines drawn indicate

that

when

R =

in.,
ft.

5 =

o.ooi,

and C

100, then

V=

1.02

ft.

per sec.

and Q
31.

0.36 cu.

per sec.
engine,

Indicated horsepower of a steam

is

H.P.

=
33>ooo

Here,

the

mean

effective pressure in
ft.,

pounds per

sq. in.,
,

is

the

length of the stroke in

is

the area of the piston in sq. in

and

is

the speed in revolutions per minute.

This formula

is

used extensively in steam engine testing practice.

obtained from the indicator card and is equal to its length. In double-acting steam engines air com-

The

pressure, P,

is

area divided

by

its

water pumps we have the fluid acting on both sides of the piston alternately. Here we must apply the formula to each end and add the results in order to get the total power output. For purposes of illustration we shall here use the diameter, D, instead
pressors, air engines, or

of the area,

and write
H.P-

(33,000) (4) (12)

""^^^'w

0.000001983

PLDW

where

L
(i)

is

expressed in inches, as

is

more common.

We
(i)

shall divide the charting of this equation into three parts:

PL = PL = q
X

g,

(2)

D^N =

t,

(3)

can be written log

scales are

-f-

= 0.0000019833/. log L = log g, which has

H.P.
2
i

the

form

(I),

and our

m\ log P,
10 to 200, log

= nh log L,

W3 log g.
long.

If
if

P varies from

P varies from

to 2.3010, a range of I.3010;

in.

we choose Wi =

10, the P-scale will

be about 13

If

varies
if

from 2 to 40, log L varies from 0.3010 to 1.6020, a range of 1.3010; we choose W2 = 10, the L scale will be about 13 in. long. Then W3 mifrh/{'mi -f- nh) = 5, and the equations of our scales are

3c=iologP,

3'=iologL,

5log2.

62

NOMOGRAPHIC OR ALIGNMENT CHARTS

Construct (Fig. 31a) the x- and
^'-axes 10 in. apart.
:

Chap. Ill

The

z-axis

divide this distance in the ratio mi

nh

and

is

therefore

must drawn

midway between them, but

(2)

the g-scale need not be graduated.

V^N =
a

form

(1),

can be written 2 log and our scales are

t

D+
log

log

=
c

log

/,

which has the

= W4 (2 log D),

nts

N,

= m^ log

t.

to 40, log D varies from 0.3010 to 1.6020, a range of and if we choose W4 = 5, the equation of the Z)-scale will be a = If iV varies from 50 to 10 log D and the scale will be about 13 in. long. 1000, log iV varies from 1.6990 to 3.0000, a range of 1.3010; and if we choose ms = 10, the iV-scale will be about 13 in. long. Then wg = W4W6/(w4 + W5) = 3.33, and the equations of our scales are
If

D varies from 2

1. 3010;

10 log D,

10 log N,

3.33 log L

Construct (Fig. 31a) the a- and 6-axes 10

have the same range and

their scales

Since D and L in. apart. have the same modulus, ior y = 10 log L and a = 10 log D, we find g^
it

convenient to

make

the a-axis

coincide with the y-axis, and to

ilni-

The

use one scale for both L and D. c-axis must divide the distance between the a- and &-axes
in the ratio

m^

m^

2,

or

the c-axis
'

is

at a distance of

3.33

in.

F^G. 31a.

i-scale

from the a-axis, but the need not be graduated.

(3)

H.P.

log 0.000001983),
z

0.000001983 qt can be written log g -f log / = (log H.P. which has the form (I). Our scales are
c

= W3 log g = 5 log q, where W7 = mzmj{mz -f

c-axes are 8.33 in. apart
:

=
=

niQ
2,

log

3.33 log

t,

ntj

log H.P.,

and hence </ = 2 log H.P. The 2- and must divide this distance in the ratio W3 We = 5 3.33; thus the J-axis must be at a distance of 5 in. from the 2-axis and must coincide with the y- and a-axes. Thus, one side of this triple axis carries the scale for L and D, and the other side, the scale for H.P. To get a starting point for the H.P. scale we make a single comws)

and the

rf-axis

putation:
I>

line joining

P = 50, L = 14, P = 8, iV = 300 give i7.P. = 26.7; hence the P = 50 and L = 14 cuts the g-axis in a point, the line joining

= 8 and N = 300 cuts the /-axis in a point, and the line joining these two points cuts the rf-axis in a point which must be marked H.P. = 26.7. Fig. 31& gives the completed chart, and the index lines indicate that = 300 when P = 50 pounds per sq. in., L = 14 in,, D = 8 in., and revolutions per min., then H.P. = 26.7.

"""- "

'

I.. ...

....
^

, .

11
I 1

A/m yjd 'A3y a// (A/J OJ3d9

i
I

\
^

(N-a)

!\
^
i

o M

S3H3N/ N/ (V 3>i0yA0S/V{O)yJ23iVV/O AfOlS/c/

uui
1

iiiii

1,1,1

t.

'j

i,i,j ,ii

'

?T
,

<

'ii Vi

'

' i I

i'

' i

n 'i

'

' I I

'

H.a

'

a3d(^13A3a(dHJy3AA0d 3Si/0f/

w
I
\
\
\

(I'd J

5S

//:^A//

'0Sy3dS97 N/(d)3daSS3dd 3A/103dJ3

\
1

A/t^J/V
1

CTT-T-rX-Trl

-1

M M

J ".'M"M
'

I.

03

64

NOMOGRAPHIC OR ALIGNMENT CHARTS

EXERCISES

Chap.

Ill

Construct charts for the following formulas. The numbers in parenthesis suggest limiting values for the variables. These limits may be extended if necessary. Additional exercises will be found at the end of Chapter V.
I-

W'

AP = -

D^Pu
It

Rarikine's formula, for the weight, w, in

pounds per

steam flowing from a reservoir at pressure Pi pounds (20 to 300) per sq. in. through an orifice of A sq. in. or of diameter D in. (o.i to 2.0) to a pressure of Pi pounds per sq. in., if Pi = 0.6 Pi.
sec. of

Francis' formula for the discharge, Q, in cu. ft. per sec. over a 2. Q = 3-33 btfirectangular weir h ft. (2 to 15) in width due to a head of ft. (0.5 to 1.5) over the crest.

3.

Z*

2 In

+ 0.5. Self-inductance,
each
r

L, in abhenries per cm. length of one of

two

parallel straight cylindrical wires

cm.

(o.i to 0.25) in radius, their

axes d cm.

(2.5 to 144) apart,

and conducting the same current

in opposite directions [distance

small compared with length of wires].

^.

P=

50,210,000 ^25)
sq. in. of

Stewart's
steel

formula for the collapsing pressure, P, in

t

pounds per
(l

Bessemer

tubing

in.

(0.02 to 0.13) in thickness

and

in,

to 6) in external diametei.
5.

sec.
(0.1

Q = % ^2 gcbfP. Hamilton Smith formula for the discharge, Q, in cu. ft. per over a contracted or suppressed weir b ft. (2 to 20) in width due to a head of ft. to 1.6) over the crest, if the coefficient of discharge is c (0.580 to 0.660). [g = 32.2.]

6.

P
d

=
is

0.196

Load,
/.

P, in pounds supported by a helical compression

spring;

B. S. gage), r

the diameter of the wire in inches (0.102 to 0.460 or No. 10 to No. 0000, is the mean radius of the coil in inches (0.5 to 2.0), / is the fiber stress in

pounds (30,000 to 80,000). Conveyor-belt calculations; p is the correct number 7. p = kgW {L + 10 H). of plies (i to 15) is the width of the belt in inches (10 to 60), g is the weight of material handled in pounds per cu. ft. (30 to 125), L is the length of the belt in ft. and is the difference in elevation between the head and tail pulleys in ft. (Z, -f 10 fl" 100 r to 1500), ^ is a constant depending on the type of drive {k 1/250,000 for a simple

H
k

drive with bare pulley, k

= 1/300,000 for a sirnple drive with rubber-lagged pulleys, 1/375,000 for a tandem drive with bare pulleys, k = 1/455,000 for a tandem drive with rubber-lagged pulleys). [Charted in Metallurgical and Chemical Ertgineering, =
XIV,
8.

Vol.

Jan.

i,

1916.]

H^

= V

15 ird^VD.

Flow of
(i

steam through pipes;

is
ft.

is

the weight of steam

is

passing in pounds per min.

(f to 36),
is

to 30,000), d

the inside diameter of the pipe in inches

the velocity of flow in

per sec. (15 to 250),

the density of the

steam at the mean pressure (use a steam table and plot D for values of the absolute pressure from i to 215 pounds per sq. in.) [Charted in Electrical World, Vol. 68, Dec.
9, 1916.]
9.
is

p = V^DK,

where

= ^^ ^i

^V Flow

of

steam through pipes; p

the pressure drop between the ends of the pipe in pounds per sq. in. per 100 ft. of pipe (o.oi to 20), and V, d, and D are defined in Ex. 8. [Charted in Electrical World,
Vol. 68, Dec. 9, 1916.)
10.
/**

ft.

14.7*

0.0007

ft'

-fy

Blast-pressure furnace; H

is

the height of the

furnace in
ft.

is the number of cu. (50 to ioo);D is the bosh diameter in ft. (10 to 25), of air at 70 F. per minute (5000 to 80,000), p is the blast-pressure in pounds gage (2 to

25).

[Charted in Metallurgical and Chemical Engineering, Vol. XIV, Mar. 15, 1916.]

CHAPTER IV. NOMOGRAPHIC OR ALIGNMENT CHARTS

(HI)

(Continued). or

EQUATION OF FORM
/i(u)

= MvY'^^'

f,(u)

=f2{v)'Mw) CHART.

Chart for equation (III). [The second form of equation (HI) can be brought immediately into the first form by taking logarithms of both members.] The first form of equation (IH) is the same as the second form of equation (I), but in Art. 23 we used three parallel logarithmic
32.
-

scales,

while here

we

shall use three natural scales,

two

parallel

and a
axis

third oblique to them.

In Fig. 32a, let

AX

and

BY

he two

parallel axes

and

AZ

any

oblique to these and cutting these in

and

respectively.

Draw any

Fig. 32a.

Fig. 326.

index line cutting the axes in the points u, v, w so that Au Aw = z; note that Au and Bv are oppositely directed.
y,

= x, Bv = y, How are x,

and 2 related? Let AB = k.

:

Then
:

in the similar triangles

Auw
or

and Bvw,
x

Au Bv = Aw wB,

or

z,

-r

y.

Now

if

^Z and B Y carry the scales x = mjiiu) and y =

=

equation becomes /i(m)

nhfiiv), the last

rr^

r f^iv),

and

if

AZ

carries a scale for

w such

that

W22

Wi

{k

z)

= hiw)

or
65

k
mifz{w) -\rnh\

66

NOMOGRAPHIC OR ALIGNMENT CHARTS

Chap. IV
line will cut the

the equation becomes /i(w)

tion.

f2(f)

'

fz{w),

and any index

v,

axes in three points corresponding to values u,

satisfying this equa-

Hence,

to

chart equation (III) /i(w)

fsiw) /2(y) proceed as follows:

Draw

three axes

AX, BY, and AZ, where

AX
z

and

BY are

parallel

oppositely directed, and

AB
y

is

any convenient

length, k.

With

and and

as origins, construct on these axes the scales

mifi{u),

nhf2{v),

!^/-^^.

mijaiw)
it is

+ W2
,

Note that

for the construction of the w-scale,

necessary to com-

pute the value of z for every value of w which is to appear on the chart. To avoid this computation, proceed as follows: On B Y, choose a fixed point F at any convenient distance, I, from B
(Fig. 326),

and on

AX
Then

construct the scale

AC =
AZ.

x'

fsiw).

From
in w,

F as
and

center, project the points

let

C on

the axis

Let

FC

cut

AZ

Aw =
z

z.

in the similar triangles

:l

ACw
k

and BFw,
-^

:k

or

j.

I -\-

x'

niifsiw)

+ nh
first

Hence
x'

to construct the scale z

-

mifi{w)
fsiw)

,\

+ rrh
,

construct

the scale

on

AX, and

then project this scale from the fixed point

on

BY (where BF

/)

to the axis

AZ marking corresponding points with

the same value of w.

This type of chart

33.

is

illustrated in the following example.

Tension on bolts with U. S. standard threads,

D= =

1.24 '-^t \/

0.088.

Here D
we

j v| ^ ft
is

-{-

is

the outside diameter of the bolt in inches,

is

the

load on the bolt in pounds, and ft

sq. in.
Tf If
.' r V the write .u equation as L

the tension fiber stress in pounds per

AD = ft x
if
,

0.088)2
r^

we have an equation

of the

form (HI).
X

The
r

scales are

m,L.

nnft,

,mi(D - 0.088)2 = l-^-^^^.

o.oooi, the equation

Let

L vary up to

100,000 pounds;

we choose Wi =

0.0001 L and its length will be 10 in. Let/( he x vary up to 100,000 pounds; if we choose tth = 0.0001, the equation of the If we choose /rscale will he y = 0.0001 ft and its length will be 10 in. the fixed point or center of projection, F, on the y-axis so that / = 8.3 in., then the equations of our scales are
of the L-scale will

0.0001 L,

0.0001 ft,

x'

5.4

(D

0.088)2.

Art. 33
If

TENSION ON BOLTS WITH

varies from \ in. to 4
off
in.,

U.S.

STANDARD THREADS

67

we compute

the corresponding values

of x'

and lay

the scale on the x-axis.

We

then project this scale from

the point

to the
'

oblique axis, marki

9-

^.<^in-nn<,a,70r

ng

corresponding

points with the

same

value of Z> (Fig. 33a).

The
point

final

chart,

showing neither the

F nor

the prois

jecting lines,

given
100.000

m Fig. 330.
side of

On one

the oblique

axis the threads per inch corresponding to the various diameters

have been
ft

given.

The

index line indicates that

when L = 20,000 pounds and

37,000 pounds per sq. in., then D^= i in. and there are 8 threads to the inch. Similar charts can be built up for various other threads.

20.000

3Q000-.

tension bolts with

u.s. st'd.

threiads.

Fig. 336.

68

NOMOGRAPHIC OR ALIGNMENT CHARTS

Chap. IV

(IV)

EQUATION OF FORM fi(u) _ fsjw)

SECTING INDEX LINES.

TWO
of

INTER-

34.

Chart for equation (IV).

large

number

volving four variables can be written in the form (IV)

/i(w) -fiiv)

such equations
in

equations

in-

as

-Mw) =

fi(t)

or

/i(w) -fiiv)

fi{w) 'ji{t), etc.

(II),

(IV)

is

included in the second form of equation

but

Equation Art. 28 we used

logarithmic scales whereas here

we

shall use natural scales.

Let

AX,

coincide with

BY and AZ, BTho: two pairs of parallel axes, where AZ may AX (Fig. 34a) or AZ may make any convenient angle with

Fig. 34a.

Fig. 346.

AX (Fig. 34&), and where AB

point

is

common
2,

transversal.

Through any
v,

P on AB
related?

dfaw two index

lines cutting the axes in the points w,

w,
z,

and and

q so that
t

Au =

x,

Bv =

y,

Aw =

and Bq

t.

How

are x, y,

From
X
:

the similar triangles in these figures,

we have
:.

AP PB
:

and

= AP PB,
:

t.

Now

ii

AX, BY, AZ, BT carry the X = Wi/i(w), y = nhfiiv),

: :

scales
z

msfsiw),
/i(m)
:

ntifiiq),
:

where Wi rui = ms mi, the relation becomes and two index lines intersecting in a point on u, V, w, and q satisfying equation (IV).
Hence,
to

f-i.{v)

= fziw)

fiiq)

AB
:

will cut

out values of
as follows'.

chart equation (IV) /i()

/2(f)

=fz{w)

/4(g) proceed

Through the ends

parallel axes

AX

segment AB oi any convenient length, draw the and BY and the parallel axes AZ and BT, where AZ
of a

Art. 35

PRONY BRAKE OR ELECTRIC DYNAMOMETER FORMULA

make any convenient
y
angle with

69

may

coincide or

AX.
i

On

these axes

construct the scales

niifiiu),

mifiiv),

= W3/3M,

=
:

mifi{q)

where the moduli are arbitrary except for the relation Wi

wg

= W3

W4.

To
and
ing

read the chart, use two

lines,

index
V,

one joining u

and the other joing,

w and
The
35.

and

intersect-

ing in a point on

AB.

following examples

illustrate this

type of chart:
Fig. 35a.

Prony brake or

electric

dynamometer

formula, H.P.
for

^'^'^'^
.

33>ooo

The sketch

in Fig.

35a gives the method

measuring the power of a rotating shaft. Either the prony brake or the electric dynamometer may be used. With such an arrangement the power is given by the above formula, where L is the length of brake arm in feet, is the speed of the

shaft in revolutions per minute,

and

is

the load on scale in

pounds.
If

we

write the equation as

:z

H.P.

N
.T

W = 5260/L
of the

77-

we have an equa^
(IV),

tion

form

and our

scales are

X
400

miH.P.,

y
/

=
=

ntiN,

msW,

W4-

5260

O
Fig. 356.

60

The

following

table

exhibits

the limits of

the variables, the

choice of moduli, and the equations and approximate lengths of the scales.
Scale

Limits

Modulus
X

Equation

Length

H.P.

wi = 0.15 o to 5000 nh = 0.00225 o to 400 THs = 0.03 W2W3 = 0.00045 0.5 to 5 W4 =

o
to 80

= y = z =
^

0.15 H.P.

0.00225
0.03

12" in III
-4

12'

2.37

1"

mj

are laid off on opposite sides of the

In Fig. 35&, the x- and z-axes coincide, so that the H.P.- and 1^-scales common axis, and starting from the

70

NOMOGRAPHIC OR ALIGNMENT CHARTS

origin; similarly for the

Chap. IV

same

N- and

L-scales.

The

index

lines,

one

joining

L and

W and the other joining N and H.P. intersect on the transW

versal joining the zero points of the scales.

The completed chart is given in Fig. 35c, and the index lines show that = 50 pounds, and N = 2000 r.p.m., then H.P. = 38. when L = 2 it.,

31

36.

Deflection of
.

beam

fixed at ends

and loaded

at center.

A =

Here, A is the deflection of beam in inches, v^ 192 EI load on beam in pounds, L is the length of beam in feet,

is

the total

E is the modulus

Art. 37

DEFLECTION OF BEAMS
and /
is

71

of elasticity of material in inch units,

the

moment

of inertia in

inch units.

We
scales

shall take

30,000,000 for
,

steel, so

that the equation

(IV),

may be

written as

W
ji
3.333.000 /

which has the form

and gives the

= miA,

= nhU,

ntsW,

= W4

(3,333.000) 7.

The

following table exhibits the choice of moduli

and the equations

of the scales.

Scale

Limits
to 1.5

Modulus
Wi W2
iris

Equation
8

Length

up up Up

10 to 35 to 300,000
to

3000

nii

= 8 X = = 0.000,224 y = = 0.000,04 z = = W2W3 = 0.000,000,001,12 =

,

0.000,224 L^
0.000,04

^
T

12" 10" 12"

II"

0.003,735 I

72

NOMOGRAPHIC OR ALIGNMENT CHARTS

These
six cases

Chap. IV

may

be represented by

six charts similar to those dis-

cussed in Arts. 35 and 36, for the equation can be written 7^

A =

LENGTH OF BEAM
I I I I

(L)

IN

FEET

I-

U-1

soo-

I.OOi

1,500-

2,000-

2.S00-

SjOOO^

r-T-T

I I I

'

'

'

'

'

I
I

DEFLECT/ON OF BEAM

(A)

fN /NCHES.

BEAM FIXED AT

E/yOS -LOADED AT CENTER.

Fig. 366.

which has the form (IV) when a value

scales are

is

assigned to

a.

In

all cases,

the

wiA,

m<JJ,

= msW,

= W4

(3,333.ooo) ^-

and

(i) to (4), the x- and z-axes are perpendicular, and in cases (5) the x- and z-axes coincide. The scales are arranged so that there (6), In all is only one common transversal joining the zeros of all the scales. and / and the index line joining A and L cases the index line joining

In cases

must

intersect

on the common transversal.

Art. 38

SPECIFIC SPEED OF TURBINE

AND WATER WHEEL

73

The completed
that there
oSS!5!2!S!i5

chart, Fig. 37, clearly distinguishes the six cases so

it.

is

no

difficulty in reading
5 X

LENGTH OF BEAM IN FEET.

5SS5
Si
!!i

DEFLECTION IN INCHES-BEAM FIXED AT ENDS -LOADED AT CENTEP.

comb/nat/on chart -deflect/ on ofbeams.

Fig. 37.

tion will

Another and more compact method be given in Art. 43.

38.
Specific speed of

of charting this

composite equa-

turbine and water wheel.

N,

N VH.P. t

The formula

gives

the specific speed of a hydraulic reaction turbine

74

NOMOGRAPHIC OR ALIGNMENT CHARTS

also of a tangential water wheel.

Chap. IV

and
is

Here,

iVg is

the specific speed, H.P.

the horsepower,

is

the

number

of revolutions per minute,

and

is

the head of water on turbine or

wheel in

tensively

and

in

The formula is exused in Hydraulics water power engineering

ft.

work; the reaction turbine is used when the head is low and the quantity of water available
is

relatively large, the value of

varying from lo to lOO, while wheel is used when the head is great and,
iVg

the tangential water

yo.oosN
Fig. 38a.

as is usual in such cases, the water limited, the value of iV varying from 2 to 6. Because

of this difference in the range

is

of iVa for the

If

two

cases, it

best to construct separate charts

TV
-v/rr

we

write the equation as -^

^, we have an equation of the

= mz VH.P.,
t

form

(IV),

and our

scales are

= miN

= nhN,

= m^^.
of the

The

following tables exhibit the choice of moduli

and the equations

scales.

Reaction Turbine
Scale

Limits
10 to 100

Modulus

Equation

Length

N,

N
H.P.

up to 2000
" "
" 1000

H
Scale

" 200

= = = W4 =
mi W2 ms

0.1

0.005
0.28

= y = z =
X
t

0.1

Ng

9"
10''

0.005
0.28

^
H^

VTrF.
10"

^' = 0.014

0.014

Tangential Water Wheel

Limits
2 to 6

Modulus
mi W2 W3
mi

Equation

Length

N,

N
H.P.

up
"

2000 " " 1000

to

= = =

1.7

0.00476
0.28
^

X y
z

= =
=

1.7

Ns

7"

0.0047 6 0.28

N
H^

VKpI

9b"
10'

" 2000

= -^ =

0.000784

0.000784

Fig. 38a shows the position of the scales. The x- and z-axes coincide so that the N,- and iY.P.-scales are constructed on opposite sides of the

Art. 38

SPECIFIC SPEED OF TURBINE

axis; similarly for the

AND WATER WHEEL

The

75

common
shown

N- and

H-scales.

charts for the re-

action turbine
in the

and

the tangential water wheel have been combined as

i.e.,

diagram,

the axes for the former have been placed per-

pendicular to the axes of the latter, and both charts use the same transSPECIFIC SPEED
(Ns)

OF REACTION TURBINE.

SI

J!

\horse power

(R. T.)!

200-

100

300- 200

400- -300

I
1000

500eoo-

400

-900

JSO-soo
BOO

^ BOO
-JOO

700

BOO

*
^
5;

600

^
y:^

k
<.

900
1000-

500

-mo
1100-

^00

Uj

%
-^1200-

ItOO

Iq

o
300-^

\l

/200

^'^'^IJOOi.

7
100

'Kl400-

1400
1500

Ig
Itl

>

1500-leoo

Key:

1600-

-1700
.1700-

BO

i
I I
I
I

-leoo

1900
\

^ leoo-

2000

ai

HEAD (H)

IN I
I

YT u
I

IR.T.)

^ %

.l
.

^
...',
|
.

^ J M ^

,|
.

i..l.,

I
,..l|

M
..,l

,..i i

<

,, ,. .T

.... n' .

"T^

SPEED

IN) IN REV.

PER MIN.

(R. T.)

SPECIFIC

SPEED OP TURBINE AND WATER WHEEL.

N
VTTp.

Fig. 386.

versal

on which the index lines intersect, one Index and the other joining H.P. and H.
Fig. 386 gives the
lines

line joining iV

and JV

show that when

iV,

then

when

= =

completed chart; for the reaction turbine, the index = 1000 r.p.m., H" = 70 ft., and B..P. = 201, 70; for the tangential water wheel, the index lines show that = 700, and H.P. = 275, then N, = 4.6. 1000 r.p.m.,

76
(V).

NOMOGRAPHIC OR ALIGNMENT CHARTS

-

Chap. IV

EQUATION OF FORM f,{u) = f2{v) fs{w) Ut) .... TWO OR MORE INTERSECTING INDEX LINES.

Charts for equation (V). A large number of equations involving three or more variables can be written in the form (V), which is similar to the second form of equation (II), but in Art. 28 we used log39.

we shall use natural scales. Equation (IV) a special case of equation (V) when there are four variables present. We shall here consider the cases where equation (V) contains three, five, or six variables. The method of charting to be employed is an amplifiarithmic scales, whereas here
is

cation of the

method described
Three variables.
/i(w)
:

in Art. 34.

Case

(i).

fi{u)
:

f2{v)

'

Mw).
is

be written as
scales are
oc

f2{v)

fz{w)

i,

which
z

of the

This equation can form (IV); the

t

=
:

mifi(u),
:

niifiiv),

nisfsiiv),

nii,

where Wi W2 = W3 W4. Here the g-scale is replaced by a fixed point, P, on the y-axis and at a distance W4 from B. The first index line joins u and V, the second index line joins w and the fixed point P; the two lines must intersect in a point on AB. (Figs. 34a, 346.) The fixed point, P, may be used as a center of projection from which the w-scale may be projected on the transversal AB. We shall then have

two

parallel scales

and a third

scale oblique to these,

v,

and a

single index

line will cut the scales in values of u,

and

satisfying the equation.

This method was employed

bolts in Art. 33.

in charting the

formula for the tension on

in Art. 40.

An example
Case
{2).

illustrating case (i)

is

worked out

Six variables.

h{u)'fM)'Si{i')=h{v)'h{'^)'U\s).

This

equation can be written as

h{u)

h{v)

=Mw)

and

p :Uq) = Mr)

Ms).

Each of these equations has the form (IV) and can therefore be charted by the method described in Art. 34. In Fig. 39a, the p-, v-, and r-scales
lie

common axis, but the ^-scale need not be graduated. To read the chart we need two pairs of index lines; the index lines (u, v) and (w, p) intersect in a point on AB, and the index lines (p, q) and {r, s) inalong a
tersect in a point

on BC.
is

An example
Case{^).
tion can

illustrating case (2)

M''^)
'

worked out
'

in Art. 41.

Five variables.

Mo)
and

Mi')

= Mv) M"^)
:

This equa-

be written as

Mu) Mv)
:

=M'iv)

Mo) ^ Mr)

and can be considered as a special form of case (2), where the 5 scale (Fig. 39a) is replaced by a fixed point through which the fourth index line must pass. An illustrative example is worked out in Art. 42.

Art. 40

TWISTING

MOMENT

IN A CYLINDRICAL SHAFT
:

77

We may also
and

chart the equation /i(m)

f^iv)

described in Art. 34, and the equation described in Art. 32. The arrangement of the scales

p =

= fsiw) p by the method /4(g) Mr) by the first method

:

is

shown

in Fig. 396,

this

arrangement

is

more compact than that

of Fig. 39a,

and em-

ploys only three index lines instead of four.

Fig. 39a.

Fig. 396.

scales are carried

In Fig. 396, the r-scale lies along the transversal on the same axis the index lines (u,
;

AB and
v)

the q- and w(w, p) inter-

and

on the transversal AB, and the third index line aligns p, r, and q. An illustrative example will be found in Art. 43. = 0.196 FD^. Here 40. Twisting moment in a cylindrical shaft, F is the maximum fiber stress in pounds per sq. in., D is the diameter of the shaft in inches, and is the
sect

twisting
If

moment
:

in inch

pounds.

we

write the equation as

5.1

M:
Our

D^

= F

we have an equation
(V), case (i).

of the

form

scales are

X
z

= miM, = msF,
of

y
t

= nhD^, = mi (5.1).
table exhibits

The
the

following

choice

moduli and

the

fixed

equations of the scales:

Scale

Fig. 40a.

Limits

Modulus
Wi W2 W3

Equation

Length
9-5" 6.4"

M
D
F

up up

to 190,000

i" to 4"
to 16,000

= = =

0.00005
0.1

X y
2

0.0005

= 0.00005 ^ = o.iD^ = 0.0005 P

8"

78

NOMOGRAPHIC OR ALIGNMENT CHARTS

Chap. IV

'2.00020.000

4000- -40,000

6.000ao.ooo

<0

8,000-60.000

10,000-

-100,000

k k
^12.600^120.000

k o
14.000140.000

16.000- -160,000

i
-180.000

TWiST/NG

MOMENT IN
Fig.
402>.

CYLINDRICAL SHAFTS

M- 0.196 F D^

Art. 41

D'ARCY'S

FORMULA FOR THE FLOW OF STEAM IN PIPES

79

hence / = 5.1, and we have a fixed point on the from the origin. We construct the M- and Fscales on the same axis from a common origin, and the Z)-scale and the (Fig. 40a.) The two index lines, one joinfixed point on a parallel axis.
minis/ m\
i,

Now nti =

y-axis at a distance 5.1 in.

ing

M and D and the other joining F and the fixed point, must intersect
common
transversal joining the zeros of the scales.
lines
2>

on the

J^
show that
in.

The completed chart is given in Fig. 406, and the index when F = 12,000 pounds per sq. in. and D in., then
pounds.
41.

M = 63,500
P=
fwd^

D'Arcy's formula for the flow of steam in pipes,

Here, P is the drop in pressure in pounds per sq. in., that is, the difference between the pressure, pi, at the entrance to the pipe and the pressure, p2, at the exit of the pipe; B is the weight of steam flowing in pounds per minute; L is the length of the pipe in feet; c is a quantity which varies

with the nature of the inner surface of the pipe w is the mean density of i.e., the average of the density at the entrance and the density at the exit of the pipe; d is the diameter of the pipe in inches. This formula is extensively used in engineering practice. We usually desire the pres;

steam,

sure drop between two points.

solve for

The

chart to be constructed will however

of the

We

any one of the six variables involved. have an equation involving six variables
in Art. 39,

and as suggested

we

shall separate

it

into

involving four variables, and build

the square root of both

up a Z chart for

form (V), case (2), two equations each each of these. Taking

members

of the equation,

we

write

it

VLB = VPcVwVd^,

or

VP

r=r-

=
i/c

and equating both members to an auxiliary quantity, Q, we write

VP^^B

Vl
We

<2'

^^^ ^"

Q Vd^

^Vw
i/c'

now construct a Z chart for each of these equations, the two charts having the ^-axis in common.
For the
Scale
first of

these equations

we have

the following table:

8o

NOMOGRAPHIC OR ALIGNMENT CHARTS

Chap. IV

The P- and 5-scales (Fig. 410) are placed on the same axis and starting from the same origin, and the L- and (J-scales on a parallel axis, but the
Q-scale
is

not graduated.

For the second equation we have the following table:

Scale

Limits

Modulus

Equation

Length

Q
d

o to 10 o to 10 30 to 70

= Wb = We =
fHi

0.002
0.06
r

6 180
nti

s
t

= =
=

0.06

Vd^

19'

Vw
(a)

m^

= W5W6 =

180

The

iv-

scales are

and ^-scales are placed on the same axis hence the w- and Lon the common Q-axis. The d and c scales are placed on a
;

parallel axis (Fig. 410).

We
(c,

use four index

{d,

lines.

The

w) and

Q) lines must inter-

sect

of the corresponding scales,

on the common transversal and the {Q, B) and (L, P) lines must intersect on the common transIt is thus

versal of the corresponding scales

a simple matter to find

the value of any one of the six

variables

when

the other five are

known.
of
Fig. 41a.

Thus, to find the value

P when c, w, d, B, and L are known, proceed as follows (Fig.

:

41a)
tion of
(c,

join the point of intersec-

w) and the

common

transversal (a) with d, cutting the (2-axis

in a point, Q; join the point of intersection of {Q,

B) and the common

with L, cutting out the required value of P. Fig. 41 & gives the completed chart, and the index lines show that when c = 40, w = 2, d = y in., B = 300 pounds per minute, and
transversal
(/3)

L =

800

feet,

then

P=

1.34

pounds per

sq. in.

42. Distributed load

on a wooden beam.
pounds per
sq.

F=
in.;

WL
njii
-

Here, F
B
is

is

the

maximum
beam
of the

fiber stress in

is

the length of the the width

in inches.

in ft.;

is

the total load on the

beam

in

pounds;

beam

in inches;

and

is

the height of the

beam

In

construction work, the total load on the

beam (depending on

the load

which the floor must support), the allowable fiber stress (depending upon the kind and quality of the wood), and the length of the beam, are usually known; and the width and height of the beam are to be determined.

Art. 41

D'ARCY'S

FORMULA FOR THE FLOW OF STEAM IN PIPES

81

D'ARC Y EQUA r/OA/-FL OtV OF STEAM /Af PJPES.

Fig. 416.

82

NOMOGRAPHIC OR ALIGNMENT CHARTS

Chap. IV

we have two unknown quantities we can of course get various combinations of these to satisfy the equation. By means of the chart to be
Since

constructed
tions of
fixed

these combinawidth and height are

readily seen,

and any desired

combinations
chosen.

may

then

be

We have an equation involving five variables of the

form (V), case (3), and introducing an auxiliary quantity,
Q,

we

shall

separate

it

into

two equations; thus,

F Y=
Fig. 42a.

W
-7^

and

Q IP ^=

We now
= mzW,

construct a
in

chart

for each of these equations, the

two charts having the Q-axis

common.

For the
X

first

of these equations our scales are

ruxF,

= m2

(L),

nii (Q),

and we have the following

Scale

table:

Limits

Modulus
nil

Equation

F L

up up up

to

2000

to 40

to 15,000

W2 W3

= = =

0.003
0.3

0.0008

Art. 42

DISTRIBUTED LOAD ON A WOODEN BEAM

83

15.000

14.000-

13.000-

12000-

1.000-

40-'

DISTRIBUTED LOAD ON A WOODEN BEAM- F-^^wr

Fig. 426.

3WL

84

NOMOGRAPHIC OR ALIGNMENT CHARTS

^-scales are placed on the

Chap. IV
;

The H- and
L- and
parallel axis,

same

axis (Fig. 42a)

jy'-scales

are on the

common
is

Q-axis.

The

^-scale

on which there

also a fixed point, /,

hence the placed on a at a distance of 9.0 in.

is

from the

origin.

We use four index lines. The (L, F) and {Q, W) lines must intersect on the common transversal of the corresponding scales, and the {B, Q) and (/, H) lines must intersect on the common transversal of the corresponding scales. It is thus a simple matter to find the value of any one of the five variables when the other four are known. Thus to find the when F, L, W, and B are known, proceed as follows: (Fig. value of 42a) join the point of intersection of (L, 7^ and the transversal (jS) with

W,

and the transversal value of H.

If

cutting the Q-axis in a point, Q; join the point of intersection of (B, Q) (a) with the fixed point, /, cutting out the required

we wish we can

project the if-scale on the transversal (a) using

the fixed point,

fixed point, /,
will

/, as a center of projection.

We

can then discard the

and the index

line

through

it,

for the index line (B, Q)

then cut the transversal (a) in the required value of H.

Given

then F, L, and
the index line

W, we determine the point Q as above, and by rotating through Q we can cut out any desired combination of B

and H.
that

The completed chart is given in Fig. 426, and the index lines show = 10,000 pounds, L = 15 ft., F = 1,200 pounds per sq. in., when

and

B=

in.,

then

H
is

12 in.
six

43.

Combination chart for

beam

deflection

formulas.

A =

192

EIP beam in

=^777".

Here, W
feet,

the total load in pounds,

is

the length of the

is

the

moment

of inertia in inch units,

is

the deflection

the modulus of elasticity (30,000,000 for steel), and factor which determines the method of loading and supporting.
in inches,
is

is

Thus

when

the

beam

is

(i)

(2) fixed at
(3) (4)

(5) (6)

both ends and uniformly loaded, both ends and loaded in center, supported at both ends and uniformly loaded, supported at both ends and loaded in center, fixed at one end and uniformly loaded, fixed at one end and loaded at the other,
fixed at

P P P P P P

= = = = =

Pi P2 P3 P4 Ft P^

= 2; = 1 =f = i = j\ =
;

^^:i'

The equation

thus involves five variables and

is

of the

form (V), case

introduce an auxiliary quantity, Q, and separate the equation (3). into two equations; thus,

We

^3

2^

and

-3

= AP.

Art. 43

COMBINATION CHART FOR DEFLECTION OF BEAMS

first of

85
if

The
write

these equations has already been charted in Art. 36,

we

for

indeed

is

the deflection of a

beam

fixed at

both ends and

loaded in center, i.e., for P = i. We shall here use the same method of charting and the same scales employed in Art. 36. The scales are

miQ,

= m^D,

= mzW,

m^.

(3,333.ooo) J,

and the following table exhibits the choice

Scale

of moduli:

Limits

Modulus
Wi = 8 m2 = 0.000,224 W3 = 0.000,04
3^

Equation

Length
;

x=8Q

W
I

10 to 35 up to 200,000

= 0.000,224 L^ 2 = 0.000,04 W

10"

8"

up

to 2000

W2W3 lUi = W4

= 0.000,000,00112
in
J 200,000^

The
43a);

scales are arranged

the form

of

a rectangle

(Fig.

the L- and /-scales start

from one vertex, B, and the Wand (2-scales start from the opposite vertex. A, but the Q-scale The two inis not graduated. and / dex lines, one joining and the other joining L and Q must intersect on the transver-

sal

AB.
chart the equation
the

3000-'

Ax,ts-Q

We now Q = AP by

Fig. 430.

method described

in Art. 32.
s

miQ,

= W2A,
parallel

The scales WiP WiP -f W2

are

where the
the 5-axis

xis

and r-axes must be

and extend

in opposite directions,

the transversal through the origins of these axes,

and k

is

the

distance between the origins.

These conditions are met

if

in Fig.

43a

(where the

ic-axis is

already constructed)

we make

the r-axis coincide

with the 3'-axis,'and the ^-axis with the transversal from A to B. We have drawn AB 13" long, and we choose nh = 8, hence the equations
of our scales are (Fig. 43a)

= 8Q,

8 A,

13

P+i

The A- and L-scales are carried on opposite sides of their common axis. The six points Pi, P2, Pe of the P scale are easily constructed by means of the table
. . . ,

P.
s:

5^

^\

8.67"

6.5"

3.7"

2.6"

0.52"

0.2"

86

NOMOGRAPHIC OR ALIGNMENT CHARTS

find the value of

Chap. IV
fol-

To

A when

J,

W,

L, and
(/,

are known, proceed as

lows: join the point of intersection of

W) and

the transversal .45 with

v>

>^

i?

VJ

MOM Erf T
^f-

OF INTERTIA' (I) /N tNCH UNITS.

Art. 45

CHART OF EQUATION

(VI)

87

the

= 100,000 lbs., L = 25 ft., and P = Pz, i.e., when / = 1000 inch units, beam is supported at both ends and uniformly loaded, then A = 1,17 in.

All the equations charted thus far 44. General considerations. can be brought under the general forms

fi{u)+Mv)+Mw)+
Most

'

=m

and

Mu)'Mv)'Mw)

=Mt).

of the formulas of engineering can be written in

one of these forms.

We have used various methods of charting these equations, employing In the case of three variables, the underlogarithmic and natural scales.
lying principle has been that one index line will cut the scales in three

values satisfying the equation.

lying principle has been that

In the case of four variables, the underlines intersecting

two index

on an auxiliary
;

axis will cut the scales in four values satisfying the equation

this

method

has been extended to equations involving more than four variables.

In the remainder of this chapter, we shall chart various forms of the above equations by methods requiring the use of parallel or perpendicular index lines. In Chapter V, we shall consider some equations which cannot be brought under either of the above forms, but which may be charted by methods requiring the use of parallel or perpendicular index lines or by methods involving the construction of curved axes. We shall end Chapter V with a brief discussion of various combined methods.

(VI)

EQUATION OF FORM Uu) Mv) =fs(w) Uq). PARALLEL OR PERPENDICULAR INDEX LINES.
:

45.

Chart of equation (VI).

Consider
that

two pairs

of intersecting

axes

AX, A Y and BZ,

y.r

BT so constructed

BZ

is

either parallel to or

u
Fig. 450.

X
Fig. 456.

coincides with
(Figs. 45a, b).

AX

and

BT

is

either parallel to or coincides with

AY,

Draw two

parallel index lines,

one meeting

AX and A Y,

and the other meeting

BZ

and

BT m

u, v, w,

and q

respectively, so that

88

NOMOGRAPHIC OR ALIGNMENT CHARTS

X,

Chap. IV

Au =

Av =

y,

Bw =

z,
:

and Bq =
y

t.

Then,
if

in the similar triangles

uAv and wBt, we have x

the scales
,

t.

Hence

AX, A Y, BZ,
t

BT

carry

Wi/i(m),

y
nti
:

nhfiiv),

mzfziw),

niifiiq),

respectively,

where

nti

niz

mi, then
/i(w)
:

X :y

z :t

becomes

fziv)

fsiw) :fi(q),

which

equation (VI), and a pair of parallel index lines, (u, v) and {w, q) A pair of will cut out values of u, v, w, and q satisfying this equation. celluloid triangles will aid in reading the chart.
is

Consider again two pairs of intersecting axes AX, A Y and BZ, BT so constructed that BZ is perpendicular to -4-X' and 57" is perpendicular Draw two perpendicular index lines, one meeting to .4 F (Figs. 45c, d).

y /"
/^

;ff

\
\
A

K^ r
,-

Fig. 45c.

Fig. 45d.

AX
X :y

and

AY and
:

the other meeting

spectively, so that

Au =

X,

Av

=^ y,

BZ and BT in u, v, w, and q reThen again Bw = 2, and Bq =

t.

our axes carry the scales described above, a pair of perpendicular index lines, {u, v) and {w, q), will cut out values of u, v, w, and q satisfying equation (VI). A sheet of celluloid with two perpenz
t,

and

if

dicular lines scratched on

If

its

under side
:

will aid in

reading the chart.

the equation involves only three variables, i.e.,fi{u) = fiiv) 'fz{w), i; here the g-scale is the equation can be written /i(w) fziv) = Mw) replaced by a fixed point through which the second index line must always
:

pass.
It is evident that there are other positions for the axes

illustrated in Figs. 45a, b,

c,

d that

will satisfy the conditions

than those imposed by

the problem.

Art. 46

WEIGHT OF GAS FLOWING THROUGH AN ORIFICE

Weight
of gas flowing through

89

46.

an

orifice,

576V

Here,

the weight of gas in pounds flowing per second, d is the diameter of is the velocity of the gas in ft. per sec, and V is the specific volume in cu. ft. of the gas in the orifice.
is

the orifice in inches, v

^^^

vo

CO

DISCHARGE OF GAS THROUGH AN ORIFICE. Wz Av _ ^d^ V

I4./I.

576 V

Fig. 46.
If

we

write the equation

d^

183.5 V,

we have an equation

of

the form (VI).

Fig. 456.

We shall
scales are

build

up a chart
z

similar to that represented

by

The

miw,

m^d^,

m^v,
of

nti

(183.5 F),

and the following table exhibits the choice

Scale

moduli
Equation
Length

Limits

Modulus

w
d
V

o o O

to 0.4
to

0.2 to 2

3000

V
The
Fig.

to 200

= 10 = I = 0.00135 = W2W3 = 0.000135 W4

Wi W2 W3
Wi

X y
z
/

= low = d^ = 0.00135 ^
=
0.02475

4" 4" 4"

{w, d) and {v, V) index lines must be parallel. The chart is given in 46 and the index lines drawn show that when v = 1500 ft. per sec, = 80 cu. ft., and d = 1.6 in., then w = 0.26 pounds per second.

^
47.

NOMOGRAPHIC OR ALIGNMENT CHARTS

Armature
or
field

Chap. IV

^2 234.5 + t2 Here Ri and Rz are resistances in ohms and k and h are the initial and final temperatures Centigrade in an armature or field winding.

winding

from

tests.

^ = ^MiLlLli

__

3\

4
-JR,)

,60

Resista nee

at Initial Temperature
C. E. Lansil

TEMPERATURES IN AN ARMATURE WINDING FROM TESTS

Ri

R,_ 234. S 234. 5

* t2

* t,

Fig. 47.

We
X

have an equation

of the

form (VI) and we

shall build

up a chart

similar to that represented

by
z

Fig. 45c.

The

scales are
^1),
/

miRi,

nhRi,

= mz

(234.5^4-

W4 (234.5

+ h),

and the following table exhibits the choice of moduli:

Scale

Art. 48

LAMfi

FORMULA FOR THICK HOLLOW CYLINDERS

ti

91

are 23.45 in- from the point and /-axes, which are respectively perpenBut it is a simple matter to arrange the dicular to the X- and 3'-axes. axes so that the h- and./2-scales are within close range of the Ri- and R2The {Ri, R2) and (^i, k) index lines must be perpendicular. The scales. chart is given in Fig. 47, and the index lines drawn show that when = 2.71 ohms. ti = 65, ^2 = 33, and Ri = 3.04 chms, then R2

We

note that the points

o and

^2

of intersection, B, of the 2-

48.

Lame formula
-77

for thick hollow cylinders subjected to internal

pressure. "

d^

^ f~p
is

Here, D
sq. in.,
is

is

the exterior diameter of the cylin-

der in inches, d
fiber stress in

the interior diameter of the cylinder in inches,

is

the

pounds per

and p

is

the internal pressure in pounds

per sq.

in.

The formula
cylinders.

extensively used in the design of thick

pump

and press

used in ordnance work on big guns, to determine what is known as the elastic resistance curve of the steel at various sections of the gun from breech to muzzle. We have an equation of the form (VI) and we shall build up a chart
It is also

similar to that represented

by

Fig. 4$d.
z

The

scales are
t

ntid^,

niiD^,

nisif

p),

= mi{f -\-

p),

and the following table

Scale

exhibits the choice of moduli

92

Art. 49

CHART FOR EQUATION

(VII)

93
inches apart, and an-

Consider a pair of parallel axes

other pair of parallel axes
pair;

AX

and BY,
(Fig.

ki

CZ and DT,

ki inches apart, parallel to

the

first

AB

and

CD

are also parallel.

49a.)

Draw two

parallel

index

lines,

one intersecting

AX

and

BY and

the other intersecting

CZ

<r*

Fig. 49a.

Fig. 496.

and

Dq =

and q respectively, so that Au = x, Bv = y, Cw = z, L Draw vE and qF parallel to AB and CD respectively. Then in the similar triangles vEu and qFw, we have x y ki = z t k^. Hence if AX, BY, CZ, DT carry the scales
u, v, w,
:
:

DT xn

X
z

= =
:

Wi/i(m),

y
t

nhfsiw),

= =

mjiiv),
^2/4(2),

where Wi

^1

= W2

^2,

then
z

ki

k2

becomes
fi{u)

-Mv) =Mw)

-fiiq),

and a pair

of parallel index lines, (m, v)

and

(w, q), will cut

out values of u,
are

v,

w, and q

satisfying this equation.

If

CZ and

DT

instead of parallel to
is

Ax and

drawn perpendicular By, and CD

Fig. 49c.

perpendicular to (Fig. 496), then a pair of perpendicular index lines, (u, v) and
{w, q), will cut out values of u,
v,

AB

w,

To
scales
in the

represent the equation /i(m)

f2{v) = /^{w) + fi(q),

and q satisfying the equation. the w- and q-

must be laid off in opposite directions. If the axes are arranged form of a square, or if the second pair of axes coincide with the first pair (Fig. 49c) then ^1 = ^2; hence, Wi = W2 and all four scales have the

94

NOMOGRAPHIC OR ALIGNMENT CHARTS

Chap. IV

same modulus. Because of the restriction on the choice of moduli,

of chart
is

not a very useful one.

We shall

this type only give a single illustration.

20

-1

FR/CT/ON LOSS /N P/PES LOST HEAD = 2dg

Fig. 50.

20-^

50.

Friction loss in flow of water.

ft.,

Sir? H = ^^-7-. Here 2dg

is

/ is

the length

of pipe in

v
is

is

the velocity in

ft. ft.

per sec, d

the internal diameter of

pipe in

ft.,

the lost head in

due to

friction,

is

the friction factor,

and g

32.2.

Art. so
If

FRICTION LOSS IN FLOW OF WATER

replace g

95

in

/ by 0.02 inches, our formula becomes

we

by

32.2,

(for clean cast-iron pipes)

and express

_
or

0.02

UP'

(12)
'

268.33
/

H _v^
d'

2 (32.2) d
(log B. of the

+ log 268.33)
.

log

2 log v

log

an equation

form (VII)

We

shall

arrange the axes as in Fig. 49c.

The
X

scales are
(log

= mi

H + log 268.33),

3'

= Wi log

I,

= Wi

(2 log v),

nti

log

d.

The

following table exhibits the limits of the. variables

and the equations

of the scales:

Scale

96
4.
<

NOMOGRAPHIC OR ALIGNMENT CHARTS

=

Chap. IV

pd

1. Thickness,
to 15,000).
.

t,

in inches o

a pipe of d inches internal diameter

(o to 60)

to withstand a pressure of

p pounds per

sq. in. (o to 100)

with a fiber stress of/ pounds

per sq.
5.

in. (o

p =

Rb
-77

Approximate

formula for flange rivets in a plate girder;

is

the

depth of the girder (20 to no), V is the vertical shear in pounds (50,000 to 275,000), R is the rivet value in pounds (1000 to 20,000), p is the pitch of the rivets in
effective

inches
6.

(i

to 9).
ft

-rrj-^

M Intensity of stress, /, in pounds per sq. in.

.

(750 to 1300) in the outer

fiber of

a rectangular beam, h inches in depth

inch-pounds.
2
it/i*
.

due to a bending moment of

7.

^ij

Field

M ..'...
intensity,

(3 to 20)

and

b inches in breadth (2 to 16)

H,

in lines per sq.

cm. at a point on a

line

through the center and normal to the plane of a circular turn of wire of negligible section conducting a current of / abamperes (o to 1000), the radius of the circular turn being r cm. (4 to 12) and the distance of the point from the wire being d cm. (4 V2 to 12 V2).
8.

C =
ft.

'.

Centrifugal
(i
-

force;
is

is

the weight in pounds

ft.

(i

to 150), v

is

the

velocity in

per sec.

to 50), r

the radius of the path in

(o.i

to 10), g

32.2,

C is

the centrifugal force in pounds.

<|>\^ -4- rT = wh (I sin )

ft.

Resistance to earth compression; w

h
is

is

the weight of the

<l>

earth in pounds per cu.

(o to 35.000).
10.

(o to 130),

the depth in

ft.

(o to 15),

is

the angle of
sq. ft.

repose of the earth (15 to 60),

P is the ultimate load on the earth in

some

pounds per

Apply the methods Apply the methods

of this chapter to charting

of the formulas of the

combination chart. Art. 25.

11.

of this chapter to charting the formulas in Exercises 7, 8,

and

9,

at the end of Chapter III.

CHAPTER

V.
{Continued),

NOMOGRAPHIC OR ALIGNMENT CHARTS

(Vm) EQUATION OF

FORM Uu) +Uv) = fj^-

PAR.

ALLEL OR PERPENDICULAR INDEX LINES.

51.

Chart for equation (VIII).

in opposite directions,

Consider two
AT m
u, v, w,

parallel axes,

AX and
AB.

BY, drawn
where
(Fig. 51a.)

and two intersecting

axes,

AZ and ^J",

AZ coincides with AX and A T coincides with

Draw two
parallel index lines,

the transversal

one intersecting

AX and BY

and the other intersecting

AZ

and

and g

respectively, so

(ife^

c ,

Fig. 51a.

Fig. 51&.

that

Au =
lines,

X,

Bv =

index

and let and Awq, we have

:

y, Aw = z, Aq = AB = k inches. Then
t.

Draw

BC
=

parallel

to these

in the similar triangles

A CB

Now

if

AC AB = Aw Aq AX,BY,AZ, and ^T carry

:

or

-{

t.

the scales
z

tnifiiu),
:

y
:

mif2{v),

mzfziw),

mrji{g),

where Wi

fe

= ms
'.k

W4, then
/i() -\-h{v)

x-\-y

= z:t becomes
97

= fz{w) :Mq),

98

NOMOGRAPHIC OR ALIGNMENT CHARTS

of parallel index lines, (w, v)

Chap. V

and any pair

values of u,

and

v,

w, q satisfying this equation.

{w, q), will cut the axes in This type of chart is illus-

trated in Art. 52.

In Fig. 516,

AC =
when

and are drawn in the same direction, and hence so that this arrangement serves to represent equation (Vm) fi{u) and f2iv) are opposite in sign, or an equation of the form
X

AX

BF

y,

fi(u)-f2iv)=Mw):Uq).
In the construction of the chart for equation (VUI), we note the folThe U-, W-, and g-scales are all laid oflf from the same origin, although we could have constructed parallel to and parallel
lowing: (i)

AZ

AX

AT

-^r

Fig. 51C.

to

AB

without affecting the relations of the

scales.

(2)

The

ti-

and

v-

scales are constructed in opposite directions or in the

same

direction -ac-

cording as /i(m) and

scales

fiiv) have like or unlike signs. (3) The u- and yhave the same modulus, mi, and the moduli and the length of
k,

the transversal,
(w, v)

are connected

by the

relation k

miniA/nia.

(4)

The

always parallel. If the equation (VIII) has the form /i(m) fiiv) = fi{u) -.fiiq), containing only three variables, it can be charted in a similar manner. Here the w-scale coincides with the w-scale, so that the (w, q) index line coincides with the (m, v) index line; hence a single index line cuts the scales This type is illustrateci in values of w, v, and q satisfying the equation.
(w, q) index lines are

and

in Art. 53.

and drawn in the same Consider again two parallel axes, and two intersecting axes DZ and DT, where DZ is perpendicular to is perpendicular to the transversal AB (Fig. 51c). and Draw two perpendicular index lines, one intersecting and and
directions,

AX

BY

AX

DT

AX

BY

Art, 52

MOMENT OF

INERTIA OF CYLINDER

99

the other intersecting

DZ
z,

and

DT m
L

u, v, w,

and q

respectively, so that
to the first of these

Au =

X,

Bv =

y,

Dw =

Dq =

Draw BC parallel

index lines. Then the triangles ACB and Hence sides are mutually perpendicular).

Dwq

are similar (since their

AC AB = Dw
:

Dq,

or

Now

if

AX, BY, DZ, and

DT carry

the scales
z

wi/i(w),
:

y
:

= miMv),
then

msfziw),

niifiiq),

where mi
X

= mz
y :k

nii,

z :t

becomes

/i(w)

fiiv)
v)

= fz{w)

:fi{q)

and any

pair of perpendicular index lines, (u,

v,

and

(w, q), will cut the

axes in values of u,
trated in Art. 54.

w, q satisfying this equation.

This type

is illus-

In Fig. ^id,

AX and BY are drawn

=/3M
inertia
:/4(2).

in opposite directions,

and hence

AC =
52.

-\-

y,

so

that this arrangement serves to represent equation

(Vni)/i(M) +/2(t')

Moment
is

of

of

cylinder.

(3 r^

fi^).

Here,

is

the total weight of a right circular cylinder in pounds, r the height in inches, and /
is

is

the radius

in inches, h

the

moment
its

of inertia in inch

units of the cylinder about an axis through

center of gravity and

perpendicular to the axis of the cylinder.

of

Writing the equation as 3 r? + ^^ = 12 / form (VIII), and we shall follow Fig. 51a chart. Here

W, we have an equation

in the construction of the

= Wi =

(3 r^),

mji^,

niz (12./),

= m^W.

Choosing k
Scale
r

15",

we have

the following table:

Limits

Modulus
X

Equation

Length

= y = z =
t

0.024

^^

0.008 h^

15" 13"

0.000,002,4 I 14"

0.000,375

^"

The
r

(r,

h)

and

(7,

W)
and

index lines must always be parallel.

Fig. 52
10* inch

gives the completed chart,

and the index

lines

drawn show that when

10

in.,

30

in.,

W=

20,000 pounds, then J

20

units.

r2S

40-i

MOMENTOF fNERTlA OF RIGHT CIRCULAR CYLINDER.

12

Fig. 52.

Art. S3

BAZIN'S

HYDRAULIC FORMULA
v

lOI

53.

Bazin formula for velocity of flow in open channels,

c Vrs,
is

where c

=
0.552

Here, m

is

the coefficient of roughness, r

the

+ -7^
Vr
ft.

hydraulic radius in
of the

(area divided by wetted perimeter), 5 is the slope water surface, and v is the velocity of flow in ft. per sec. We shall first build a chart for the coefficient, c. The equation can
/-

be written
0.552

V?

+m
,

87 -^-,

Vr

or

m r V.+^^^
,
.

^^,
87
'
,

Vr

which

is

^-scale,

the special form of (VIII) where the w-scale coincides with the and hence only one index line is required. Hence

r = miVr,

m
^^

\
)>

ntii

r = miVr,

= mA^-^

/o.552 c\
!

Choosing k = 15
Scale
r

in.,

we have

the following table:

Limits
0.2 to

Modulus
Wi = 2.7 m\ = 2.7 W4 = ^ = 15
out values of
r,

Equation

Length 12" 10"

c

20

m
c

0.06 to 2

10 to 155

= y = t =

2.7

Vr

4.89

0.0952

15"

One index
that

line cuts

m, and
c

c satisfying

the equation.

Fig. 53 gives the chart for this formula,

and the index

line

drawn shows

when

ft.

and

m=

i.i,

then

78.5.

can consider the equation v = c y/Ts as an equation of the form and we can build up a logarithmic chart accordingly. We have already constructed such a chart in Art. 29 and Fig. 296 may therefore be used to supplement Fig. 53 for a complete solution of our problem. In Fig. 2()h, the (r, s) and (c, v) index lines must intersect on the g-axis. Thus, when r = 4 ft., 5 = o.ooi, and c = 78.5, we read v = 4.96 ft. per sec. R = (p D) tft. Here, 54. Resistance of riveted steel plate.

We

(n), Art. 28,

R
in

is

the resistance of riveted steel plate to tearing between rivet holes

is

rivet hole in inches,

is the diameter of the the thickness of the plate in inches, and ft is the fiber stress of steel in tension and equals 55,000 pounds per sq. in. The
t

pounds, p

the pitch of the rivet in inches,

is

formula

is

used extensively

in boiler design

and

in structural
,

work.

The equation can be

form (Vin), and we We have
X

written p

-D =

55:^^

an equation of the

shall follow Fig. 51c in the construction of the chart.

mip,

= -miD,

m-iR/ $^,000,

t'

= m^.

aO

,aO

r-2
-J.9
,*P

-l.B
-1.7

o
,iP

-1.6

-2.

-7O

-1.5
to <0

-A

-1.4
-1.3

-1.2

,90

tr

-l.I

k
^oO

-10

k
-.9

A^O

k
,1^^

\-.7

k
6
,1^0

-.5

-.4
,aO
.3

-.2
io-^

.00
-./

COEFFICIENT /N BAZI N'S HYDRAULIC FORM ULA

0.552 *lPr
102
Fig. 53.

Art. 54

RESISTANCE OF RIVETED STEEL PLATE

iP3

'/2

\-l

\^

-l'/2

RESISTAN\CE TO TEARING (R)/N THOUSANDS OF LBS.

RESISTANCE OF RIVETED STEEL PLATE. TEARING BETWEEN RIVET HOLES.

(P-D)
t ft =

ft=

55000

Fig. 54.

104

NOMOGRAPHIC OR ALIGNMENT CHARTS

Chap.

Choosing ^
Scale

lo

in.,

we have

the following table:

Limits

Modulus
Wi mi Ws

D R

up up up
up

to 6
to 2 to 200,000 to
I

= = =

2 2
1.65
ruzk

mi

mi

8.25

Art. 55

CHART FOR EQUATION

(IX)

lOS

Now

if

AX

carries the scales

mifi{u), xi

mifz{w) and

^F carries

the scale y

m^f^iij),

then

=
X y
line cuts

xi

becomes

77-^fi{u)
u, v,

77-r + f2{v)

t-t-t-

fsiw)

and any index

lowing: (i)

out values of
3'-axes

satisfying this equation.

In the construction of the chart for equation (IX)

we

note the

fol-

The

:v-

and

other and they carry the

2-axis divides the angle

may make any convenient angle scales x = mifi(u) and y = niif^iv).

x-

with each
(2)

The

between the
:

and
:

3'-axes into

two angles whose

(3)

sines are in the ratio W]

mi,

i.e.,

AD Dw =
and

Wi

m^.

The

ac-axis
z-

also carries the scale Xi axis

m\fz{w),

this scale

is

projected on the

by lines parallel to the ^'-axis, the points and their projections being marked with the same value of w. If Wi = W2, then AZ bisects the angle XAY = a (Fig. 556). Then

Aw AD =
:

sin (180

a)

sin -, or

= Aw
.

= sm a a sm .

2
Xi

a sm .

22
2

cos -

a
Xi

a sm -

= Wi

2 cos - 1 fsiw).

In this case the w-scale

may
y

be constructed on AZ, and the scales are

mifi{v), s

mifi(u),

= Wi f 2

cos

^j f%{w).

Finally,

if

we take a =
X

120, our scales are

simply

mifiiu),

mifiiv),

mifsiw).
is

The method
extension of the

of charting the second

form of equation (IX)

method employed

for charting the first form.

merely an Consider

the case of four variables,

fM +
By
/,

f2iv)

+ Mw)
'

Mq)

introducing an auxiliary variable,

write
^""^

we can
I
I

fM'^Mv) We chart each

t

L. = i

1j I _ I rfz{w)~U{qy
by
Fig. 55c.

of these equations

means
a

of three concurrent scales with

/-scale

common

lines are necessary,

and

g-scales.

which need not I5e graduated. (Fig. 55c.) Two index one cutting the u- and y-scales and the other the w^ The (, v) and {w, q) index lines must intersect on the /-axis.

io6

NOMOGRAPHIC OR ALIGNMENT CHARTS

Equations of the form (IX) are not very We shall only give one illustration.
Focal lengths of a lens.

Chap.

common

in engineering prac-

tice.

56.

7+ r= ~

Here,

is

the focal

distance of the object,

is

the focal distance of the image, and

is

the

principal focal length of the lens.

We shall
are

take our x- and ^'-axes at an angle of 120, and the 2-axis as

the bisector of this angle.

Let Wi

0.5,

then the equations of our scales

z

o.sf,

0.5 F,

0.5 p.
line

that

The completed chart is given by Fig. 56. when / = 6 and F = 9, then p = 3.6.

The index

drawn shows

^<i'
Jo

^:
f<^'

Fig. 56.

Another formula which

may be
i?2

charted in the same

way

is

i?l

Rz

R
2?j,

where R
Rz,
. . .

is

the circuit resistance of a circuit containing resistances Ri,

connected in parallel.

(X)

EQUATION OF THE FORM /i(u) +/2(v) STRAIGHT AND CURVED SCALES

-/sCtr) =/4(ii>).

in

(We note that the variable w occurs 57. Chart for equation (X). and both members of the equation.) Consider two parallel axes BY and a curved axis CZ (Fig. 57). *Draw any index line cutting these

AX

axes in u,
in

v, and w respectively. Draw wD parallel to AX, cutting AB D, and draw wE and vF parallel to AB. The triangles uEw and wFv

Akt. 58

STORM WATER RUN-OFF FORMULA

107
if

are similar, hence

Bv =

y,

AD
X

Eu Fw = Ew: Fv = AD DB. = 2i, Dw = z and AB = k, we have

:

Therefore,

Au =

x,

z:z

=
X

zi: k

zx

Zi

or

{k

Zi)

-\-

Ziy

kz,

or

+k

Zi

=
k

nh

Zi

and
X

if

Wi/i(w),

Zi

kz
zi
\

mifiiv),

=
Zi

mifi{w),

this relation

becomes
niik

/i(w)

+ Mv) Mw) = f,{w).

Solving for

Zi

and

we

get

mifziw)

+ W2
/4(w).

_ ~
Hence
x

W1W2
mifsiw)

W2

to chart equation (X) pro:

ceed as follows

Construct the scales

par-

= oti/i(m),
axes

= m2f2iv) on two
and

allel

AX

BY

extending

,,^aV

in the

same

direction.

U AB =
AB,

inches, construct -the points of the

curved scale
to w,

CZ by

assigning values
Zt

and

layings off along

=
Fig. 57.

AD = niifziw)
allel

m\k
-\-

mi

fz{w),

and parW1W2

^f-

to

AX,

z = D'w

mifziw)

+ nh fiiw),
v,

and marking the point thus

found with the corresponding value of w.

the three scales in values of u,

Then any index

line will cut

and

satisfying equation (X).

-

To
scales

chart the equation fi{u)

fiiv)

fz{w)

fi{w),

we

construct the

mjiiu) and y = mifiiv) in opposite directions. This equation 58. Storm water run-off formula, q -f Nq^ = P. arises in the storm water run-off formula given by C. B, Buerger, in the Trans. Am. Soc. C. E.. Vol. LXXVIII, p. 1139, where and P are quantities which depend upon the sewer run, the area, and the rainfall, and q
X

is

the run-off in cu.

If

ft.

per sec. per acre.

we

write the equation

P
Zi

Nq^

q,

we have an equation
z

of the

form (X), with the scales

X

miP,

ntiN,

niik

Wi5^ -f

W2

=
i

W1W2
Wi5*
-\-

W2

Let P, N, and q vary from o to

10,

and take Wi

= m2 =

and k

14".

io8

NOMOGRAPHIC OR ALIGNMENT CHARTS

Then our
it

Chap.

scales are

= P,

y= -N,

zi

2^

=
z*

"^/v

The
(Fig.

axes

AX

directions,

and BY are drawn in opposite and the length of AB is 14 in.

58a).

We

assign
Zi,

values to

q,

and on

AB
we

we

lay off

AD =
z

and

parallel to

AX

lay off

DD' =

and mark the point D'

with the value assigned to q. We join the points D' by a smooth curve, thus giving a curved scale for the variable q. Any index
line will

The completed chart is given by Fig. shows that when P = 6 and N = 5, then

then cut out values of P, N, and q satisfying the equation. 586, and the index line drawn
q

cu.

ft.

per sec. per acre.

EXP/fSS/ON IN

BUERGER

RCf/V-OFF

FORMULA.

Fig. 58&.

Art. 59

FRANCIS FORMULA FOR A CONTRACTED WEIR

109

-IS

2S

--f

-*S

FRANCIS FORM ULA.

CONTRACTED WE/R.
Q "3.33 (B-0.2H)
Fig. 59.

^S

no

NOMOGRAPHIC OR ALIGNMENT CHARTS

Chap.

Here, Q
is

Francis formu a for a contracted weir. (? = 3.33 (B 0.2 H) hK is the quantity of water flowing over weir in cu. ft. per sec, B is the head over the crest of the the width of the weir in ft., and
59.

weir in
If

ft.

we

write the equation

3.33 i?2j5

equation of the form (X), with the scales x

niiQ,

-rfrom o to
k

= 0.666^2 we have an y = m^ (3.33 B),Zi =

-3

(0.666 i^J).
to 33.
If

Let

vary from o to
0.3,
,

5,

8,

and Q from o

we choose Wi =

nh

0.6,

and

12,

our scales are x

o.^Q, y

= 2B,Zi=

H^

+2

z= ^^
H^

The axes AX and

of

+2
2i,

B Y are drawn

in opposite directions,

and the length

We assign values to H, and on AB we lay ofi AD and parallel to 5 F we lay off DD' = z, and mark the point thus found with the corresponding value of H. We join the points by a smooth
is

AB

12 in.

curve, thus giving a curved scale for the variable

will

H. Any index line then cut the scales in values of B, H, and Q satisfying the equation. The completed chart is given in Fig. 59, and the index line drawn shows
60.

that

= 1.2 ft., then Q = 12. i cu. when B = $ ft., and The solution of cubic and quadratic equations.
o,
M7^
first

ft.

per sec.
^

tu^

+ pw + q =

+ pw + q
;*,

o,

w?'

rvtxp-

+ pw + 9 =

0.

Let us consider
the equation as g The scales are
'

the cubic equation v^

+ pw =
nhp,

we have an
km^

/>w g = o. Writing equation of the form (X).

X
If

iniQ,

Zi

miw

+
,

W2

w,

mymi
niiw

+ W2

wp.

we

and k

allow p and q to vary from = 10", our scales are

10

to

+10, and choose Wi = W2

z

q,

p,

Zi

= low
;

11^
;

In Fig. 60a, the p- and g-scales are constructed on

starting at

XX' and YY'

viz.,

and
. .

B
.

respectively.
,

Assigning positive values to w,

Zi

w=

o, o.i, 0.2,
z,

ID,

we compute

and

and lay

off

AD

Z\

and

DD' =

and mark the

points, D', thus found with the corresponding

values of w.

We draw a smooth
Then any index

curve through these points, getting the

line will cut the three scales in values of
line joining

curved axis AZ.

q, IV,

and p satisfying the equation, or an index

note that the line through

p and

q will

cut the curve in w, a root of the cubic equation.

We
YY'
u^
in

E (p =

+ pw +

A (q = o) and D' (w = Wo) will cut Wo^) since these values of q, w, and p satisfy the equation 2 = 0. This observation allows us to construct the points of

Art. 6o

THE SOLUTION OF CUBIC AND QUADRATIC EQUATIONS

III

the curved scale as follows:

OnAB, we

mark

the points with the corresponding value of w,

vertical

through these points. Let the point D through this point. Then the

Zi = lo w/(w+i), and draw the verticals be marked Wq and let DD' be the

layoff the scale

line

joining

{q

o)

with

E {p =

Wo^) will cut DD' in a point of the curve which must be marked with the value Wq. Thus the points of the scale
are rapidly constructed.

Interpolation on
either along

the w;-scale

may be made
by

the curve or,

in
Fig.
of

projection, along
for the cubic
is

AB.

The complete curve

means
it

drawn 606 (curve marked "C"). By

we can
i.

find the positive roots

of the equation.

Example
-\-

Solve the equation w^

joining
in

w 6 = o. The index line P =2 and q = 6 cuts the curve

2
1.46 (Fig. 60&).

w =
Fig: 60a.

The negative

roots of

the equation
w' and

can be gotten by letting

w =

finding the positive roots of the resulting equation.

beyond the
w'^
lie

limits

10 and +10,
o,

let

w =

and q lie aw'; the equation becomes

choose
a. so

Up

+^w' a

-\r-^= a
2.

or w'^

+ p'w' + g' =
i
-L

o;

p' that r and

<?'
u.

within the limits

10 and +10.
the
/3

Example
let

Solve
get

equation

la^

2\ow

9000
let

=
a

o.

If

we
and

w =

aw',

we

w _

210
a^

QOOO
a^

Now

10,

The index line joining p 2.1 and w'^ 2.1 w' g = o. = 9 cuts the curve in w' = 2.42; hence w = 24.2 (Fig. 606). g We may similarly build an alignment chart for the quadratic equaThe method of construction of the curved axis tion w^ + pw -\- q = o. we have
from the construction of the curved axis for the DD' = z = wi^/(w -\- i). Again we note that the points A (q = o), D' (w = Wq), and E {p = Wq) must lie on a straight line, since these values of q, w, and p satisfy the
for this equation differs

cubic only in that w^ replaces w^, so that

equation

w"^ -f-

pw

-{-

o.

We may

use this fact in constructing the

(curve
of the

The complete curve is drawn in Fig. 60& marked "Q"); by means of it we can find the positive roots
points of the curve.

equation.

Example
line joining

3.

Solve the equation

q

ii/^

5.15

-f-

2.42

o.

The

index

p = 5.15 and

2.42 cuts the curve in

w =

0.52 and

w =

4.65.

alignms^nzg^hart for solution of quadratic and cubic equations.

112
Fig. 606.

Art. 6o

THE SOLUTION OF CUBIC AND QUADRATIC EQUATIONS

-\-

II3

The complete cubic equation w^ + nw^ formed into the equation u^ + p'w' + g'
w'
,

pw -\- g o may be transo by the substitution w

by means
of

or

we may proceed

to solve

it

directly

an alignment

3
chart.

In the construction of the curved axis

we have
v^

AD = Zx =

low
let

and BD'
the curve
yl

w^ = -

+ ww^

7)+''(-^)'

In Fig. 60C,

and the curve

A Co
i'

to the

Q correspond to the quadratic uP' + pw + 2 = o = o (where n = o). Then cubic liP -\r pw -\q_

DQ=- w w^ +
Thus
tance
off

DC,= we simply

up

w+
off,

DD' = DCo-\-n' DQ.

i'

starting at Co,

lay

along DD', n times the fixed dis-

DQ

to arrive at the point C of the curve corresponding to the com-

plete cubic equation.

We

thus rapidly lay

a uniform from interval equal to DQ, and get scale with

Co in either direction

the points
C_3,
. .
.

Ci,

C2,

C3,

C_i,

C_2,

corresponding
3,
. . . ,

to

the curves

I, is

2,

-I, -2,

-3

This

done along the various verticals

and for the values of w for which n = o was plotted, and the curves for = 10 For intermediate to w = 10 are drawn. interpolate between two values of n we
curves.
Fig. 606 gives the complete chart for

the

equation

u^

nw^

pw

-{-

o.

Thus

to solve this equation,

we draw
q
lie

the index line joining

this cuts the

-curve in the required root.

To

find negative roots,

to find the roots

when p and

beyond the

limits

p and q; and 10 and +10, we

q

proceed as in the case of the simple cubic w^ Example 4. Solve the equation lu^ -\- w^
index line joining

pw

-\-

o.

2.1

3.2

o.

The
i

p = 2.1 and

= 3.2

cuts the curve

w =

in

w=
=
o

1.6.

Solve the equation ui^ -{- g6 w^ -\- 721.879 w 7,826.051 5. equation occurs in the problem of the equilibrium of arches). If we let w = aw' = 10 w', the equation becomes w'^ 7.22 w' 9.6 w'^ ~ 7-83 = O and the index line joining p = 7.22 and q = 7.83 cuts the curve n = 9.6 in w' = 0.59. Hence w = 5.9.

Example
(this

We
w^ pw

+
-\-

pw^
q

can similarly plot alignment charts for any trinomial equation niiP -f g = o and for the fourth degree equation /*

o.

114

NOMOGRAPHIC OR ALIGNMENT CHARTS

Chap.

(XI)

ADDITIONAL FORMS OF EQUATIONS. METHODS.*

Chart for equation of form
Jl(w)

COMBINED
form

6i.
is

ttt + ttT" = fsiw) This fttJ2{V)

a generalization of equation (IX).

and

AY and

a curved axis

CZ

(Fig. 6i).

Consider two intersecting axes Let any transversal cut the

AX

v, and w respectively, and draw wD parallel to A Y, so that Au = X, Bv = y, AD = x', and Dw = z. Then in the similar triangles, uDw and uAv, we have Dw Av = Du Au, or

axes in u,

=XX
if

/
:

X,

.'.

z/x' -; + ~y = X
,

Now

X
x'

= = =

Wi/i(w),

nhfziv),

mifi{w)

p = ^' /4(w)

or
tion

nhfsiw)

/4(w),

this

equa-

becomes
/i(w)^/2(t;)

Fig. 61.

/3(;)'

Thus, the w-scale

x'

is

tnifsiw)

and

a smooth curve through the points determined by ifhfaiw) f^iw) and marked with the corresponding

values of w.
62.

Any

index line will cut the scales in values of

u, v,

w satisfy-

ing the equation.

Chart for equation of form /i(u) +/2(v) 'faiw) =/4(q).

/

We

introduce an auxiliary variable

(l) fiiv)
'

and write

Mw) =

and

(2)

Mu) +

= Mq).

Equation (i) has the form (III) and may be plotted by the method of Equation (2) has the Art. 32, but the /-scale need not be graduated. the method of Art. 23. The position of form (I) and may be plotted by the scales is illustrated in Fig. 62. The {v, w) and (w, q) index lines must intersect on the /-axis. In63. Chart for equation of form/i(tt) /4(q) +Si{v) 'Sz{u)) = i. an auxiliary variable t, we write troducing

(i) /x(w)

Uiq)

and

(2)

+ Mv) Mw) =
-

i.

Art. 32.

(i) has the form (HI) and can be plotted by the method of Equation (2) is a special case of the form (X), where /4(w) = i, and can be plotted by the method of Art. 57. The /-scale is not gradu-

Equation

* These forms, involving three or four variables, occur rarely in engineering practice, and are therefore treated very briefly.

Art. 65

CHART FOR EQUATION OF FORM

The
position of the scales
is

(Xl)

"S
The
(w, g)

ated.
(v,

illustrated in Fig. 63.

/-axis.

and

w) index lines must intersect on the

^^

Fig. 62.

Fig. 63.

64.

Chart for equation of form "^77^

/i(")
/,

-f

'/2(f)

Uw)'

Introducing

an auxiliary variable
(i)

we

write
^

/4(g)

tH =

and

(2)

+ -=.
/2W

Mw)

Equation
Art. 32.

(i)

has the form (HI) and

(2) is of

may be

plotted

by

the

method

of

Equation
of Art. 55.

the form (IX) and

may

method

position of the scales.

The ^scale is not graduated. The {u, q) and (v, w) index

be plotted by the Fig. 64 illustrates the

lines

must

intersect

on the ^axis.

Fig. 64.

Fig. 65.

65.

Chart for equation of form "^7^

/i(u)
/,

+ ttt = l Introducing an Mv)

7
fz{i^) + ^-^^~ = Mv)
I.

auxiliary variable

we

write

(i)

/i() -ii)^

Mu)

and

(2)

ii6

NOMOGRAPHIC OR ALIGNMENT CHARTS

(i)

Chap.

Equation
Art. 32.

has the form

(2) is

Equation
/-scale

The common
is

(III) and can be plotted by the method of a special case of the form charted in Art. 61. need not be graduated. The position of the scales

illustrated in Fig. 65.

The

{u, q)

and

{v,

w) index

lines

must

intersect

on the
66.

^axis.

We introduce a new variable

(i)

Chart for equation of form /i(u) -f^iq)

t,

4-/3(1?) -f^iw)

=Mw).

and write
(2)

Mu)Uq) =

and

t-\-Mv)Mw) =h{w).

Equation
Art. 32.

(i)

Equation
of Art. 57.
is

has the form (HI) and ma,y be charted by the method of (2) has the form (X) and may be charted by the

method

The

^axis need not be graduated.

The

position

of the scales

illustrated in Fig. 63.

The

(w, g)

and

{v,

w) index lines

must

intersect

on the ^axis.

^9^

Fig. 66.

Fig. 67.
is

An
nels,

interesting application of this type

given

by D'Ocagne.*

He

considers Bazin's formulaf for the velocity of flow of water in open chan-

Vrs,

where

=
0.552

m
V?

We

can combine these equations into one equation,

V

SjVrs
0.552

m
Vr
<

or

87

VI _ w _ 0.552 ~ r V}

Here
* Traits t

?^ =
V

and

1-^ = ^. r
Vr

We have charted

de Nomographie, p. 233. this formula by means of two charts in Art. 51.

Art. 67

CHART FOR EQUATION OF FORM

illustrates the positions of the scales.
axis,

(XI)

II7

Fig.

66
r)

By

placing the
chart.

scales

on the same

we

get a very

compact

m- and 5The {v, s) and

index lines must intersect on the /-axis. 67. Chart for equation of form /i(tt) hiq) -\-SM f^{w) We introduce an auxiliary variable t, and write /6(it?).
(m,
-

= Mq)

(l) /i()

/2(g)

- = Mq)
t

and

(2)

Mv) Mw)

= Mw).

Both equations have the form (X) and can be plotted by the method of Fig. 67 Art. 57 with a common /-axis, which need not be graduated. the positions of the scales. The (w, q) and {v, w) index lines illustrates must intersect on the /-axis.

EXERCISES
Construct charts for the following formulas.
limiting values for the variables.
tional exercises will
1.

The numbers in parenthesis suggest These limits may be extended if necessary. Addibe found at the end of this chapter.
-\-

V=
ft. ft.

tzH

{\IP

d^).
is

Volume of a cask or buoy;

is

the diameter of the

ft.

base in

(o to 10),

the diameter of the middle section in

(0 to 10),

is

the

height in

ft.

the volume in cu. ft. (o to 800). 2. Q = 3.33 <& [(^ -f- A)* A^]. Francis' formula for the discharge, Q, in cu. per sec. over a rectangular, suppressed weir b ft. in width (2 to 15) due to a head of
(o to 10),
is

ft.

over the crest (0.5 to

yv

1.5),

considering the velocity head h

ft.

(o to o.i)

due to the

velocity of approach.
3.

Sp. gr.
is

_
.

Here,

is

the weight in pounds of the solid in air (o to

gr. is

100), w'

the weight in pounds of the solid in water (o to 95), sp.

the specific

gravity (o to 20).
4.

ST'

Gordon formula
is

for

columns with ends rounded and maxi-

9000 r^

mum

allowable compression stress of 20,000 pounds per sq.

in
ft.

in.;
(i

is

the_ length of

the

column

(10 to 50), r

the radius of gyration in inches

to 12), /is the allowable

stress in
5.

pounds per

sq. in.

(1000 to 20,000).

Equivalent resistance, R, of a parallel circuit the respecD=D'"l""D""t~Er* "2 JCi 'vj tive branches of which have resistances i?i, i?2, and R3 ohms (i to 10) and containing no e.m.f. 6. s = Vo< 5 gP. Distance, s, in feet (260 to +260) passed over by a body projected vertically upwards with an initial velocity of Vo ft. per sec. ( 260 to -|- 260)
rC

in

seconds (o to 17); g
7.

32.2.

V=

0.6490

j^.

Volume,
P

V, in cu.

ft.

of

one pound of superheated

steam which has a pressure of

pounds per

sq. in. (50 to 250)

and a temperature

of

T degrees

(280 to 650).

Il8

NOMOGRAPHIC OR ALIGNMENT CHARTS

MISCELLANEOUS EXERCISES FOR CHAPTERS
III, IV,

Chap.

V.

Construct charts for the following formulas. The numbers in parenthesis suggest limiting values for the variables. These limits may be extended if necessary.
1. P' = x^Pi^P., and P" = \^PiPi^. First intermediate pressure, P', in pounds per sq. in., and second intermediate pressure, P", in pounds per sq. in. of a three stage air compressor which compresses air from a pressure of Pi pounds per sq. in. (14.4 to 15.2) to a pressure of P2 pounds per sq. in. (500 to 3000).

2.

Wr^.

Moment
T"^'

about

its axis;

is

of inertia, 7, in inch units of a right circular cylinder the total weight in pounds (o to 25,000) and r is the radius in inches

(o to 25).
3-

Y
its

Radius

of gyration,

K,

of a right circular cylinder about

r

an
is

center of gravity and perpendicular to the axis of the cylinder; the radius in inches (o to 25), h is the height in inches (o to 40).
axis
4-

through

^=

3
361n -

Capacitance,
r

C, in microfarads of

two

parallel cylinders

10^

per cm. length; each cylinder

in practical cases).
5-

cm. in radius

(o.i

to 0.25), their centers separated

of dielectric constant ^ (^

by

a
i

distance of d cm. (2.5 to 144), arid immersed in a

medium
ft.

^ = 0.596

-p

0.256.

Volume,
P

V, in cu.

of one

pound

of superheated of

steam which has a pressure of

degrees (280 to 650).

pounds per

sq. in. (50 to 250)

and a temperature

6. B.H.P. = 3.33 {a 0.6 Va) vT?. Boiler horse-power, B.H.P. (o to 500) for chimney design for power houses; A is the internal area of chimney in sq. ft. (6 to 16), is the height of the chimney in ft. (50 to 150).

7. '

H.P.

321,000
in. (o.i

nsd^
.

Horse-power, H.P., transmitted by a solid circular shaft of J i.

diameter d

to 6) at n revolutions per min. (50 to 2500) with a fiber stress in shear of 5 pounds per sq. in. (o to 50,000).
8. AT = I 6c sin A. Area of a and included angle A (0 to 90).

9.

</

K, with = 0.013 VDJp. Piston-rod diameter, inches

triangle,
d, in

sides b (o to 10)

and

c (i to 10)

(i

to 6) of a steam engine;

the piston diameter in inches (12 to 24), / is the length of the stroke in inches (12 to 60), p is the maximum steam pressure in pounds per sq. in. (80 to 150).
is

10.

A=

593

/y-T-'

Sectional

area.

A, in circular

mils, of

a copper wire for

which the total annual cost of transmitting energy over a line conducting a constant current of / abamperes (o to 100) will be a minimum; c is the cost of generated energy in dollars per kilowatt hour (0.005 to 0.02), c' is the cost of the bare copper wire in dollars per pound (0.15 to 0.35), h is the number of hours per year that the line is in use (4 X 300 to 24 X 300), p is the annual percentage rate of interest on the capital invested in copper
(4 to 6).
11.

^
=B

(a*

-}.

b^),

Moment
T^ or Ti

of inertia, /, of a flat rectangular plate about

an
in

axis perpendicular to its plane

and passing through the center;

ft.

is

the total weight

ft. (2

pounds
12.

(o to 30,000),

is

the length in

(5 to 25), b
=

is

the breadth in

to 10).

Ti

+ VB^ +

is

B =

Bending moment,
is

B, in a circular

shaft for which the twisting

moment

T,

and

Tj

the twisting

moment which would

give the

same

effect as

B and T acting

together.

MISCELLANEOUS EXERCISES
13.

II9
sq. in.

p =

st
.

Allowable internal pressure,

r

p, in

pounds per

on a hollow
s is

cylinder of inner radius

inches;

is

the thickness of the cylinder in inches,

the

working strength of the material (20,000 pounds per sq. in. for steel). Resistance, R2, in ohms (o to 5) of a conductor 14. R2 = Ri[i + o. {ti ti)\. of k C. which has a resistance of i?i ohms (o to 5) at tC. (20 to 30) and is made of a material which has a resistance temperature coefficient of a at ^1 C. (a = 0.00393 when /i = 20 and may be taken as a constant for copper). Fteley and Stearns' formula for the discharge, 15. Q = 3.31 ftW^ + 0.007 6. Q, in cu. ft. per sec. over a suppressed weir b feet in width (5 to 20) due to a head of H feet over the crest (o.i to 1.6). Depth, D, in ft. (i to 55,000) to a stratum, where A is the tan A. 16. D =

dip in degrees
17.
(l

(i

,T

H sin A. Thickness,

to 89),

and

is

the horizontal distance in

T, in
ft. (i

ft.

(100 to 1,000).

to 1,000),

where

is

the dip in degrees

to 90),

and

is

the horizontal distance in

ft.

(lOO to 1,000).

C is the dip of the projected 18. tan C = tan A sin B. -^ Projection of dips. angle in degrees (o.i to 89), A is the dip of the bed in degrees (i to 89), B is the angle of projection in degrees (i to 90). = f R^KC. Explosion formula. is the number of half-pound blocks 19.

of T.N.T. (20 to 10,000),

R is

the radius of rupture in

ft.

(0.5 to 15.0),

is

a constant

for the material (o.io to 0.50),

20.
r

d^

=8rb 4lf.
of

a constant for tamping (0.1 to 5.0.) Diameter, d, of the base of a segment of a sphere of radius
is

and height
21.

segment
h.

h.

V = Trrb^ ^b^. Volume,

7=
COS^a

V, of a

segment of a sphere of radius

and

height of segment
22.

N .^N

is

the

number of

teeth

(i

to 100) in a spiral gear, o

is

the angle

{0 to 80) of teeth

with axis,

T is

the number
6; 21-26,

of teeth for

14,

No. 8; 14-17, No. 7; .17-21, No. 135, No. 2; 135 up. No. 1).

No.

5;

which to select cutter (1226-35, No. 4; 35-55, No. 3; 55-

INDEX.
Adiabatic expansion formula, 48 chart for, 33, 49 Alignment or nomographic charts
also

Chart, Grasshoff's formula, 51 Hazen-Williams formula, 60

{see

Charts,

alignment

or

nomo-

horsepower of belting, 54 indicated horsepower of a steam engine, 63

graphic)

fundamental principle of, 44 with curved scales, 106

with four or more parallel scales, 55 with parallel or perpendicular index
lines, 87, 91, 97 with three or more concurrent

Lame

formula for thick hollow cylin"run-off," formula, 49

of inertia of cylinder, 100

ders, 92

McMath
moment

multiplication
scales,

104

and prony brake, 70

division,

47

with three parallel scales, 45 with two intersecting index lines, 68 with two or more intersecting index
lines, 76 with two parallel scales and one inter-

resistance of riveted steel plate, 103

solution of quadratic
tions, 112
specific

and cubic equaand water

speed

of

turbipe

wheel, 75

secting scale, 65

storm water

run-oflf

formula, 108

Approximate differentiation, 224 Approximate integration, 224

Area,

tension in belts, 54 tension on bolts, 67

twisting

moment

in a cylindrical shaft,

by approximate integration rules, 227 by planimeter, 246 Armature or field winding formula, 90
chart
for,

78

90

49 49 Charts, hexagonal, 40 Chart with network "of scales, for

of sphere,

volume volume

of circular cylinder,

Bazin formula, loi chart for, 102, 116

adiabatic expansion, 33 chimney draft, 38

elastic limit of rivet steel,

34

Center of gravity, by approximate integration rules, 231

equations in three variables, 28

equations in two variables, 20
multiplication and division, 30, 31 solution of cubic equation, 36

Chart, alignment or nomographic, for

adiabatic expansion, 49 armature or field winding, 90 Bazin formula, 102, 116

temperature difference, 39

Chezy formula
chart
for,

for flow of water,

56

Chezy formula

for flow of water, 58

58

D'Arcy's formula for flow of steam, 81

deflection of beams, 72, 73, 86

Chimney

draft formula, 37

chart for, 38

discharge of gas through an

distributed load on a
focal length of a lens, 106

orifice,

89

Coefficients in trigonometric series evalu-

wooden beam, 83

ated,

Francis formula for a contracted weir,

109
friction loss in pipes,

94
XI

by six-ordinate scheme, 179 by twelve-ordinate scheme, 181 by twenty-four-ordinate scheme, 185 -for even and odd harmonics, 179

xu

INDEX
Focal length of a lens, chart for, 35, 40, 106
slide rule for, 15

Coefficients in trigonometric series evalu-

ated,
for for for

odd harmonics only, i86 odd harmonics up to the fifth, 187 odd harmonics up to the eleventh,

Fourier's series, 170

189
for

Francis formula for a contracted weir, chart for, 109

Friction loss in pipes, 94 chart for, 94

10

odd harmonics up to the seventeenth,

191
graphically, 200

Fundamental of trigonometric

series,

170

mechanically, 203 numerically, 179, 186, 192, 198 Constants in empirical formulas deter-

Gauss's interpolation formula, 219 Graphical differentiation, 244

mined by method of averages, 124, 126 method of least squares, 124, 127 method of selected points, 124, 125
Coordinate paper, logarithmic, 22
rectangular, 21

Graphical evaluation of coefficients, 200 Graphical integration, 237

Graphical interpolation, 209
Grasshoff's formula, 50 chart for, 51

semilogarithmic, 24

Harmonic analyzers, 203 Harmonics of trigonometric

Hazen-Williams formula, 57 chart for, 60 Hexagonal charts, 40 Horsepower of belting, 52

series,

170

D'Arcy's formula for flow of steam, 79

chart
for,

81

Deflection of beams, 70, 71, 84 chart for, 72, 73, 86

Differences, 210
Differentiation, approximate,

224

chart for, 54 Hyperbola, 149 Hyperbolic curves, 128, 135, 137, 140

graphical, 244

mechanical, 255 numerical, 234

Differentiator, 255 Discharge of gas through an
orifice,

Index

line,

44

Indicated horsepower of steam engine, 61

89

chart for, 63 Integraph, 252

Integration, approximate, 224

chart
chart

for,

89

Distributed load on a
for,

wooden beam, 80

applications

of,

83

Durand's

rule,

226

Elastic limit of rivet steel, 32

chart for, 34

by by by by by
in, 124,

Durand's
Simpson's

rule,

227 226 233

rectangular rule, 225

rule, 226,

trapezoidal rule, 225

Weddle's

rule,

233
.

Empirical formulas,

general formula for, 231

125,

determination of constants
173. 174

graphical, 237

mechanical, 246
Integrators, 250

for non-periodic curves, 120

for periodic curves, 170

Interpolation, 209

involving 2 constants, 128

Gauss's formula
graphical, 209

for,

219

involving 3 constants, 140 involving 4 or more constants, 152 Equations, solutions of {see Solutions of algebraic equations)

inverse, 219 Lagrange's formula for, 218 Newton's formula for, 214, 217

Experimental data, 120, 170 Exponential curves, 131, 142, 151, 153, 156

Isopleth, 44

Lagrange's interpolation formula, 2i8

INDEX
Lamd
formula for thick hollow cylinders,

XI II

127 Logarithmic coordinate paper, 22

of, 124,

91 chart for, 92 Least Squares,

Rectangular coordinate paper, 21 Rectangular rule, 225

Resistance of riveted steel plate, loi chart for, 103

method

Logarithmic curve, 151 Logarithmic scale, 2

definition of,

equation

of,

Maxima and minima by approximate

differentiation formulas, 236

logarithmic, 2

representation
Scale modulus, 2
Scales,

of, 1

McMath
chart

"run-off" formula, 48

for,

49

Mean

effective pressure

by approximate

network
sliding, 7

of,

20

integration rules, 228

perpendicular, 20
stationary, 5

Mechanical differentiation, 255 Mechanical integration, 246 Moment, by integrator, 250

Moment

of inertia,
rules,

Semilogarithmic coordinate paper, 24 Simpson's rule, 226, 233

by approximate integration by integrator, 250

230

Slide rule,
circular, 16

Moment
chart

of inertia of cylinder, 99
for,

for electrical resistances, 15 for focal length of lens, 15

Lilly's spiral, 18

100 Multiplication and division, charts

for,

30, 31, 41,

47

logarithmic, 9
log-log, 13

Newton's interpolation formula, 214, 217 Nomographic or alignment charts {see Alignment or nomographic charts) Numerical evaluation of coefficients, 179,
186, 192, 198

Sexton's omnimetre, 17 Thacher's cylindrical, 18

Solutions of algebraic equations,

Numerical differentiation, 234 Numerical integration, 224 Numerical interpolation, 215

Parabola, 145 Parabolic curves, 128, 135, 140 Periodic phenomena, representation 170 Planimeter,

by by by by on

means of parabola and circle, 26 means of rectangular chart, 35 means of alignment chart, 1 10 method of inverse interpolation, 22I
the logarithmic slide rule,
1

Specific speed of turbine

and water wheel,

73 chart
of,

for,

75
108

Storm water run-off formula, 107

chart
for,

Straight line, 122, 125

Amslef polar, 248 compensation, 249 linear, 249 principle of, 246 Polynomial, 159
Pressure and center of pressure,

Tables, construction

of,

213

Temperature
chart
for,

difference, 37

39
52

Tension

in belts,

by
231

chart for, 54

Tension on
chart
for,

bolts,

66

approximate integration Prony brake, 69

chart
for,

rules,

67

70

Rates of change, by approximate

entiation formulas, 235

differ-

Trapezoidal rule, 225 Trigonometric series, 170 determination of constants in, 173, 174 Twisting moment in a cylindrical shaft, 77 chart for, 78

XIV
by approximate integration 229 Volume, by approximate integration 229
Velocity,

INDEX
rules,

Volume
chart

of sphere, 50
for,

49
233
rules,

rules,

Weddle's
228

rule,

Volume

of

circular

cylinder,

48

Work, by approximate integration

chart for, 4Q

PLEASE

DO NOT REMOVE
FROM
THIS

CARDS OR

SLIPS

POCKET

UNIVERSITY

OF TORONTO

LIBRARY

1
TA 151 L7 1921 pt.l
Engin.

Lipka, Joseph

Graphical and mechanical computation

^Jf-^-mUmmmmtBi

J;