Aircraft Propeller Design

AIRCRAFT PROPELLER DESIGN
AIECRAFT PROPELLER
DESIGN
FRED E.yWEICK, B.S.

Aeronautical Engineer, Langley Memorial Aeronautical Laboratory,
National Advisory Committee for Aeronautics; Member,
Society of Automotive Engineers
First Edition
Seventh Impression
McGRAW-HILL BOOK COMPANY, Inc.

NEW YORK AND LONDON
1930
106
Copyright, 1930, by the
McGraw-Hill Book Company, Inc.
PRINTED IN THE UNITED STATES OP AMERICA
All rights reserved. This book, or

parts thereof, may not be reproduced
. in any form without permission of
the publishers.
THE MAPLE PRESS COMPANY, YORK, PA.

PREFACE
Great strides have been made in propeller design during the
past decade, and it is the object of this book to collect in one
volume the important experimental, theoretical, and practical
developments and to present them in a simple, usable manner.
The various factors have been considered not only from the point
of view of propeller performance and efficiency but also directly
in terms of the effect on airplane performance.
It is hoped that the work brings out clearly the dominating
influence which the airplane and engine designers have on the
propulsive efficiency obtained, due to their control of the con
ditions under which the propeller must operate. In this con
nection it is thought that certain chapters, such as that on the
desirability of gearing, should be of at least as great interest to
airplane designers as to propeller designers.
The writer has spent the past several years in experimenting
with and designing propellers, which is his justification for
attempting this work. His first work of this nature was as
propeller designer for the Bureau of Aeronautics, Navy Depart
ment. He was then transferred to the staff of the National
Advisory Committee for Aeronautics, where he was in charge of
the Twenty Foot Propeller Research Tunnel. Following this,
as chief engineer of the Hamilton Aero Manufacturing Company,
he had an opportunity to apply practically the experimental
results and to check them by means of flight tests. He has been
particularly fortunate in regard to this work in that, being
engaged in research work on the staff of the National Advisory
Committee for Aeronautics, experimental results were available
to him as soon as the tests were completed. This has given
him an opportunity to include the latest developments, and in
this connection the friendly attitude of the N.A.C.A. officials
has been of great help.
Valuable suggestions in regard to the material to be included
were received from Mr. Edward P. Warner, now editor of Avia
tion, Prof. William G. Brown of Massachusetts Institute of
Technology, Prof. F. W. Pawlowski of the University of Michi
v
vi PREFACE
gan, and Prof. E. G. Reid of Stanford University. Acknowledg

ment is also due to Dr. Clark B. Millikan and Dr. A. L. Klein of
California Institute of Technology and to Mr. Donald H. Wood
and Mr. John Stack, the author's associates at the Langley
Memorial Aeronautical Laboratory, all of whom checked and
made helpful criticisms of portions of the material. The author
is also indebted to his wife for revision of the manuscript.
Fred E. Weick.
Langley Field, Va.,
September, 1930.
CONTENTS
Page
Preface v
CHAPTER I
Introduction 1
Propeller terminology.
CHAPTER II
The Momentum Theory 5
The simple momentum or Froude theoryIdeal efficiency.
CHAPTER III
The Aiufoil 11
Fluid resistance and dynamical similarityThe aerodynamic
forces on airfoilsCoefficients and methods of plottingEffect of
variations in the shape of airfoilsElementary airfoil theory
Airfoil sections suitable for propellers.
CHAPTER IV
The Simple Blade-element Theory 37
The air forces on a blade elementThe efficiency of an element
Limitations of the simple theoryExample of analysis with blade-
element theory.
CHAPTER V
Modifications of the Blade-element Theory 51
The combined or inflow theoryThe blade-element theory with
multiplane interference correctionsExample with multiplane
interference correctionsThe induction or vortex theoryExample
with vortex theoryResume^ and comparison of various theories.
CHAPTER VI
Aerodynamic Tests on Propellers 82
Flight testsWind tunnel testsCoefficients and methods of
plotting results-The flow of air through a propeller.
CHAPTER VII
The Effect of Blade Shape on Propeller Characteristics . . 105
The effect of pitch, pitch distribution, blade width, number of
blades, blade thickness, airfoil section, and plan form.
vii
viii CONTENTS
Page
CHAPTER VIII
The Effect of Tip Speed on Propeller Performance 119
Tests on model airfoils at high speedTests on model propellers at
high tip speedsTests on full-scale propellers at high tip speeds
Practical data for finding efficiency at high tip speeds.
CHAPTER IX
Body and Propeller Interference 136
Expressions for efficiency with a bodyEffect of body size, shape,
and dragComparison of tractor and pusher arrangements
Practical use of body interference data.
CHAPTER X
The Effect of Propeller Characteristics on Airplane Perform
ance 159
Effect of propeller characteristics on speed, climb, take-off, cruising
performance, range, and enduranceEffect on performance at high
altitudes with ordinary and with supercharged engines.
CHAPTER XI
The Variable-pitch Propeller 180
Effect on speed, climb, take-off, and cruising performance of air
planes with unsupercharged enginesPerformance with variable-
pitch propellers and supercharged engines.
CHAPTER XII
The Gearing of Propellers 190
Example in which gearing greatly improves the performance
Example in which the performance is poorer with gearing than with
direct driveMethod of determining whether or not gearing is
desirable.
CHAPTER XIII
Tandem Propellers 201
Calculation of operating conditions and air forces on rear propeller
Model tests with tandem propellersMethod of dealing with
tandem propellers in which each propeller is treated as a complete
unit, with examples.
CHAPTER XIV
Materials and Forms of Construction 213
Propellers constructed of wood, micarta, steel, and aluminum alloy
The construction of variable-pitch propellers.
CONTENTS ix
Page
CHAPTER XV
The Strength op Propellers 229
Calculation of stressesApproximate method of stress analysis
Example showing approximate stress analysisEffect of deflection
on the stressesDestructive whirling test.
CHAPTER XVI
Summary of Factors to Be Considered in the Design op a Pro
peller 249
The selection of propellers to give the best high speed, climb, take
off, angle of climb, cruising, or all-around performanceLimita
tions imposed by airplane and engineChoice of the number of
bladesResume' of factors affecting propulsive efficiency.
CHAPTER XVII
Design Procedure, with Charts and Examples 257
Information furnished propeller designerA simple system for
selecting propellers of standard form, with examplesApproximate
method for propellers of any form, with examplesDesign by
means of the blade-element theory.
Symbols 285
Index 289
LIST OF SYMBOLS
A Propeller disc area.
A Area of section.
A Cross-sectional area of wind tunnel air stream.
a Slope of lift curve = dCh/da, a being in radians.
a Inflow velocity factor.
a' Rotational interference velocity factor.
B Number of blades in a propeller.
6 Blade width.
b Final slipstream or wake-velocity factor.
c Chord length of airfoil.
c Velocity of sound in air.
CF Centrifugal force.
CP. Center of pressure.
Co German lift coefficient = 100CL.
C German drag coefficient = IOOCd.
Cl Absolute lift coefficient = L/qS.
Cd Absolute drag coefficient = D/qS.
Cd{ Induced drag coefficient.
Cd Profile drag coefficient = Cd Cd-.
Cm Pitching-moment coefficient = M/qcS.
Cp Power coefficient = P/pn*D6.
Cq Torque coefficient = Q/pn2Ds.
Cs Coefficient of resultant force = R/qS.
Cs Speed-power coefficient =
CT Thrust coefficient = T/Pn*D*.

CT' Thrust coefficient = T/qA.
D Propeller diameter.
d Body diameter.
D Drag or air resistance.
Dc Drag coefficient = kD = D/pSV*.
Di Induced drag.
>o Profile drag.
EHP Excess horsepower available over that required for horizontal flight.
g Standard acceleration due to gravity = 32.17 ft./sec.2
H Total pressure head.
h Maximum thickness of an airfoil section.
HP Power, in horsepower units.
HPr Horsepower required.
hp Horsepower.
xi
xii LIST OF SYMBOLS
I
Moment of inertia.
J
Slip function = V/nD.
kL
British lift coefficient = L/pSV2.
fcc
British drag coefficient = D/pSV.
kt
British thrust coefficient = T/pn2D*.
feu
British torque coefficient = Q/pn2D6.
Kr
Engineering lift coefficient = L/SV2.
Kx
Engineering drag coefficient = D/SV2.
L
Representative linear dimension.
L
Lift.
Lc
Lift coefficient = kL = L/PSV2.
MPitching moment.
MResultant bending moment.
MaTwisting moment due to air loading.
Ma'Twisting moment due to centrifugal loading.
MqComponent of air bending moment due to torque.
Mq'Component of centrifugal bending moment in direction of torque.
MtComponent of air bending moment due to thrust.
Mt Component of centrifugal bending moment in direction of thrust.
MPH Velocity, in miles per hour.
m.p.h.
Miles per hour.
N
Revolution speed, in revolutions per minute.
iV\
Equivalent revolution speed of rear propeller of tandem series.
n
Revolutions per second.
P
Power.
p
Pressure.
p
Pitch, geometrical.
p'
Pitch at 0.75ft.
Q
Torque.
Qc
Torque coefficient = Q/pV2D3.
QHPTorque or brake horsepower.
q
Dynamic (or impact) pressure = }4pV*.
R
Resistance.
R
Resultant force.
R
Aspect ratio = span squared divided by area.
R
Tip radius of propeller blade,
r
Radius.
RPM Revolution speed in revolutions per minute.
r.p.m.
Revolutions per minute.
S
Area, wing area.
S
Nominal slip, where 1 S = V/np.
S
Solidity factor = 2wr/Bb.
S
Stress.
Sc
Compressive stress.
St
Tensile stress.
T
Thrust.
Thrust coefficient = T/pV2D2.
THP Thrust horsepower.
LIST OF SYMBOLS xiii
TR Thickness ratio of whole propeller = the thickness ratio of the
section at 0.75R.
TWR Total width ratio = WR X B.
V Velocity, usually considered velocity of advance or forward velocity.
Vr Resultant velocity of advance of a blade section.
Fj Axial component of full slipstream velocity = V(l + 2a).
W Weight.
w Specific weight, lb./cu. in.
w Tangential velocity.
WR Width ratio of a propeller blade = blade width at 0.75ft divided
by D.
a Angle of attack.
ao Angle of attack for infinite aspect ratio.
at Induced angle = e/2.
0 Blade angle.
& Blade angle at 0.75ft.
y Angle between lift and resultant force vectors of an airfoil, where
tan y = D/L.
c Angle of downwash.
J7 Efficiency.
0 Interference angle.
v Kinematic viscosity.
p Air density, mass per unit volume.
po Mass density of standard air at sea level = 0.002378 slugs per
cu. ft.
4> Angle between direction of motion of a blade element and the
plane of rotation, where tan <p = V/2irrn.
w Angular velocity.
AIRCRAFT PROPELLER
DESIGN
CHAPTER I
INTRODUCTION
A propeller is a device for providing a force or thrust, at

the expense of power generated by a motor, for driving a craft
of some type through a fluid medium, such as air or water.
In order to provide this thrust, the propeller must set a mass
of the fluid in motion in a direction opposite to that of the craft
being propelled.
While many types of propellers, including paddle wheels and
various feathering blade devices, have been proposed and in
some cases used, the screw propeller has been universally adopted
for aircraft propulsion. The air screw propeller, as its name
implies, screws or twists its way through the air, advancing as
it turns, and moving the aircraft with it. /The force which
drives the aircraft is the reaction obtained because the propeller
has pushed a certain mass of air backward. The backward-
moving air, called the slipstream, has kinetic energy due to its
motion, and this energy represents a loss. There are also other
losses, one being that due to the friction of the air on the blades.
The useful or thrust power developed by the propeller is therefore
less than the power delivered by the engine to the propeller, and
one of the most important aims of the propeller designer is to
obtain a high ratio of useful power to engine power or, in other
words, a high propulsive efficiency.
Aircraft propellers are nearly always driven by internal-
combustion engines, and each design of engine develops its
power most satisfactorily at some definite speed of revolution.
A propeller, therefore, should not only be efficient but should
also be designed to absorb the power of the engine at the correct
revolution speed.
1
2 AIRCRAFT PROPELLER DESIGN
In addition to its aerodynamic qualities, a propeller must

have sufficient strength to be safe and reliable. The attainment
of sufficient strength and resistance to fatigue for safety is a
stumbling block in the designing of propellers having radical
or unusual shapes or constructions but is relatively simple if
standard proved forms are adhered to. ^
Propeller Terminology.1The term propeller is used almost
universally in this country, whereas the word airscrew (or its
Fig. 1.Typical wood propeller and some terms which apply to propellers.
Germany, and France. From the propulsion point of view,
propeller seems the more fundamental term, but on the other
hand it may be said that the propeller operating in air is just
one form of airscrew, other types being the fan, the windmill,
and the anemometer. Either term is satisfactory, but the Ameri
can practice of using the term propeller is adhered to in this
book.
The diameter D of a propeller is the diameter of the circle
swept by the blade tips, or the distance between the tips (Fig. 1).
The pitch of any screw is the advance per revolution. A
propeller of fixed geometrical form, since it operates in a fluid,
1 A complete nomenclature for aeronautics including propellers is given in
Technical Report 240 of the National Advisory Committee for Aeronautics
(hereafter abbreviated to N.A.C.AT.B., also the Reports and Memoranda
of the British Advisory Committee for Aeronautics and its successor, the
Aeronautical Research Committee, are abbreviated to British R. and M.),
1927, and many of the definitions in this chapter are taken directly
therefrom.
INTRODUCTION 3
may have a variety of forward speeds at the same revolution

speed, so that the pitch, in the usual sense, is not fixed. The
advance per revolution, however, is of fundamental importance
in propeller operation and is called the effective pitch (Fig. 2).
There are two other kinds of pitch which are generally used
in connection with propellers: the geometrical pitch, and the
experimental pitch. These forms of pitch are not necessarily
constant along the radius of a propeller but may vary from
section to section. The geometrical pitch p of an element of a
propeller is the distance which the element would advance along
the axis of rotation in one revolution, if it were moving along a
helix having an angle equal to its blade angle (Fig. 2.). The
nomigaL-er standard geometrical pitch of a wholepropeller is_the
Fig. 2.Illustration of effective and geometrical pitch.
pitch of the section at two-thirds of the radius. If all of the

elements of a propeller have the same geometrical pitch, the
propeller is said to have uniform geometrical pitch.
The experimental pitch of an element of a propeller is the
distance the element would have to advance in one revolution
in order that there might be no thrust. The mean experimental
pitch, sometimes called the zero thrust pitch, of a whole propeller
is the distance the propeller would have to advance in one revolu
tion in order that there might be no thrust.
The boss of a propeller is the central portion in which the hub
is formed or mounted (Fig. 1).
The hub is the metal fitting inserted or incorporated in or with
the propeller for the purpose of mounting it on the propeller
or engine shaft.
The root of the blade is the portion near the boss or hub
(Fig. 1).
The aspect ratio of a propeller blade is the tip radius R which
is half the diameter, divided by the maximum blade width.
The width ratio WR of a propeller blade is the blade width at

radius 0.75R divided by the diameter.
The total width ratio TWR of a whole propeller is the width
ratio of one blade multiplied by the number of blades.
The thickness ratio of a section of a propeller blade is the ratio
of the thickness at that section to the blade width.
The thickness ratio TR of a whole propeller is taken as the
thickness ratio of the section at 0.75/2.
The blade angle is the acute angle between the chord of a
propeller section and a plane perpendicular to the axis of rotation
of the propeller.
The rake or tilt of a propeller blade is defined as the mean angle
which the line joining the centers of area of the sections makes
with a plane perpendicular to the axis of rotation (Fig. 1).
A tractor propeller is one which works in front of the body
or engine, relative to the direction of flight, while a pusher
is one which is placed behind the body or engine.
A propeller is termed right-hand, if, when viewed from a
position to the rear (in the slipstream), it rotates in a clockwise
direction. Conversely, if from the same position the rotation
is counterclockwise, the propeller is left-hand.
An engine is called right- or left-hand according to its rotation
when fitted with a tractor propeller. A left-hand engine is
therefore required by a right-hand pusher propeller.
CHAPTER II
THE MOMENTUM THEORY
As stated in the introduction, the propeller, in order to provide

a thrust, must give motion to a mass of air in a direction opposite
to the thrust. The simple momentum theory, developed by
Rankine and R. E. Froude, is based on a consideration of the
momentum and kinetic energy imparted to this mass of air.
The Froude and Rankine theories are essentially the same,
the main difference being that the Froude theory considers the
propeller disc as a_ whole, while the Rankine divides it into
elementary annular rings and deals with one ring at a time,
afterward summing up the effects of each to obtain the effect
of the whole.
In the Froude theory the propeller is assumed to be an advanc
ing disc producing a uniform thrust T, the air pressure being
therefore different in front of and in back of the disc by a constant
amount over its area. This hypothetical disc, sometimes called
an actuator disc, can be imagined as a propeller having an infinite
number of blades. It is also assumed that the flow of air is
streamline in character on either side of the disc and continuous
through the propeller, so that the axial velocity is the same
immediately in front of and immediately in back of the disc.
There is no torque on the disc, and no rotation or twist is imparted
to the air going through it.1 It is further assumed that the air
is a perfect fluid, being incompressible and having no viscosity.
In practice the propeller actually advances through the
air but it is more convenient for the purpose of analysis, and is
relatively the same, to consider the propeller disc as stationary
in a uniformly moving stream of air. The general conception
of the flow is then as shown in Fig. 3. The air stream, where
it is not influenced by the propeller, has the velocity V and
pressure p. The pressure is reduced to p' at the front of the
1 A. Betz has developed an extension of the momentum theory including

rotation, entitled Eine Erweiterung der Schraubenstrahl-Theorie, published
in Z.F.M., p. 105, 1920.
5
disc and receives an increment Ap in passing through it. The

velocity is given an increment aV on approaching the disc and
has the constant value V + aV in passing through it (the disc
is assumed to have no thickness). In the final slipstream or
wake the velocity has been given a further increase to V + bV,
where bV is the added velocity of the slipstream, and the pressure
has fallen to its original value p.
Actuator Disc,
.V VfUa) H VO+b)
P P'+Ap P
.V V
P P
Fig. 3.
The flow being regarded as potential except in passing through

the actuator disc, Bernoulli's equation1 may be applied separately
to the air in front of the disc and the air in back of the disc.
The total head in front of the disc is
h = p + y2Pv* = p' + y2Pv\l + ay
where p is the mass of air per unit volume; and the total head
in back of the disc is
H1 = p' + Ap + V2pV\l + a)2 = p + y2Pv\l + b)\
The difference in pressure in front of and behind the disc is
therefore
Ap = #1 - H
= [p + V2pV2(i + by] -[P + y2Pv*]
= PV>b(l + 1)
and if A is the area of the disc the thrust is

T = AAp
= APv*b(i + iy
1 Bernoulli's equation states that along a stream tube the total head of the
fluid H = p + MpV2 = constant. Its proof can be found in any good
textbook on hydrodynamics.
THE MOMENTUM THEORY 7
Also, since the thrust is equal to the change of axial momentum

in unit time,
T mass per unit time X velocity imparted
= AV(1 + a)pbV
= APV2b(l + a).
Equating the two expressions for thrust,
APV*b(l + a) = ApV*b(l + |)
and
a = y2b.
Thus, according to the momentum theory, half of the velocity
imparted to the slipstream occurs in front of the propeller disc,
and half behind.
Ideal Efficiency.The work done on the fluid per unit time,
which is the same as the rate of increase of kinetic energy of the
fluid, is
Energy = y2pAV{l + a)[V2(l + 6)2 - 72]
= PAbV\l + a)2.
In order to find the useful work done by the thrust, it is con
venient to go back to the state in which the propeller is advancing
with velocity V through fluid at rest, the thrust and velocity still
being in the same relation to each other as before. The work
done by the thrust on the propeller and aircraft in unit time is
now TV.
The efficiency, which is the ratio of the useful work to the total
work done on the air, is therefore
= TV
71 PAbV\l + a)2
= PAbV3(l + a)
PAbV3(l + a)2
= 1
~ 1 + a
The efficiency can also be obtained by considering the useful
work and the energy lost in the slipstream. When the propeller
is moving through a fluid which is at rest, the work done by the
thrust on the fluid is equal to the increase of kinetic energy of the
fluid in unit time, or
Energy = Yi (mass per unit time) (velocity imparted)2
= V2pAV{l + a){bVY
= MpAV*b(l + a)bV
= ViTbV
= TVa.
This energy is not regained from the air, and it represents a loss.
The efficiency is then
_ useful work
' useful work + energy lost in slipstream
= TV
TV + TVa
= 1
1 + a
This expression represents the ideal or limiting value of the
efficiency of a perfect propeller. It is never reached with actual
propellers because of the following additional losses :
1. The energy of rotation of the slipstream due to the torque.
2. The profile drag or friction of the propeller blades moving
through the air.
3. The loss due to the fact that the thrust is not actually
uniform over the disc area, but at best falls off near the perim
eter due to tip losses and at the center due to hub losses.
4. The loss due to the finite number of blades and the conse
quent variation of thrust at any one point with time.
Usefulness of the Ideal Efficiency.While the ideal efficiency
is never attained in practice, it represents a mark at which the
propeller designer can aim. The actual propeller efficiencies
obtained throughout the working range are usually from 80 to
88 per cent of the ideal efficiency, and so the ideal is useful in
determining approximately the actual efficiency which can be
obtained with a propeller under given conditions. The ideal
efficiency can be taken as a measure of how favorable the condi
tions for the operation of the propeller are, and the ratio of the
actual to the ideal can be taken as a measure of how well the
propeller works under the given conditions.
In examining the expression for the ideal efficiency it is interest
ing to note that the efficiency depends only on the ratio of the
velocity imparted to the air by the propeller to the forward
velocity of the propeller itself. It is instructive to examine this
expression to see how it is affected by variations in the forward
speed, the thrust, and the diameter of the propeller. Consider
the equation already obtained for the thrust,
T = APV%{l + a).
THE MOMENTUM THEORY 9
ti = zr and a = ]^b, the equation can be written

1 + a
2PAV2(l ~ v)
T =
?
or
2PAV*
1 - V T
It is convenient here to introduce a thrust coefficient
where q is the dynamic pressure of the fluid and is equal to

YipV*. The above expression involving efficiency then becomes
1 - v C'T
and the variation of efficiency with C't is shown in Fig. 4. It
will be observed that as the thrust coefficient, which represents
the thrust per unit area per unit dynamic pressure, increases,
1.00
0.20
0 I I I I I I I
0 5 10 15 20 25 10
Fig. 4.-Ideal efficiency vs. thrust coefficient

the efficiency drops rapidly at first and then more gradually.
By varying only one factor of C't at a time, Fig. 4 illustrates the
following points:
1. The ideal efficiency decreases with increase of thrust.
2. The ideal efficiency increases with increase of forward
velocity.
3. The ideal efficiency increases with increase of density of

the fluid.
4. The ideal efficiency increases with increase of disc area or
propeller diameter.
While considerations of the ideal efficiency indicate that the
diameter should be as large as possible, the loss due to the friction
of the blades increases as the diameter becomes greater, so that
it does not increase the efficiency of actual propellers to increase
the diameter beyond a certain limit.
CHAPTER III
THE AIRFOIL
In the blade-element theory of the propeller, the propeller

blade is considered an airfoil varying in cross-section and angle
from hub to tip. It is necessary, therefore, to have a general
understanding of airfoils and the aerodynamic reactions on
them before taking up the blade-element theory. The aim of
this chapter is merely to cover briefly the characteristics of
airfoils which are of importance in propeller design. More
comprehensive treatments can be found in some of the modern
books on aerodynamics.1
Fluid Resistance.If an object such as a flat plate is held
in a wind or air current, it is at once apparent that a force
is acting upon it. If the plate is held perpendicular to the
direction of the air flow, the only force will be resistance or drag
in the direction of the flow. If, however, the plate is inclined
at some angle, the force will also be inclined and its lateral or
cross-wind component may become much greater than the drag
in the direction of flow. An airfoil is a body so shaped that the
cross-force, called the lift, is high compared with the drag.
The forces acting on a solid body moving in a perfect fluid,
that is, a fluid which has no viscosity and is incompressible,
would vary as the square of the velocity V and would be pro
portional to the area on which the fluid particles acted. If <S
is an area denoting the size of the object, the resistance R would
be expressed by the equation,
R = KSV2,
where K is a constant or resistance coefficient determined
by experiment. Air, however, is not a perfect fluid and it
has been found by experiment that the resistance does not
always vary exactly as the square of the velocity and that the
coefficient K is not a true constant.
'"Airplane DesignAerodynamics," by Edward P. Warner, McGraw-
Hill Book Company, Inc., 1927, is recommended.
11
Dynamical Similarity.If K of the above formula were

always a true constant, it would be possible to make aerodynamic
tests on models of any size at any speed and apply the results
with perfect accuracy to geometrically similar bodies having
any other size and speed. The principle of dynamic similarity,1
as developed by Riabouchinski and Buckingham, states that
in the case of the motion of objects in a real fluid medium such
as air, having viscosity and being compressible, the resistance
varies according to the following equation :
where L is a convenient representative linear dimension of

the object, v is the kinematic viscosity of the fluid, and c is the
velocity of sound or the rate of travel of any compressive wave
in the fluid. From an examination of the above equation it
is at once apparent that if all factors materially affected the
results, aerodynamic tests would be representative only of the
actual conditions of speed and size at which they were made,
and model tests would be of no value for predicting full-size
performance.
Fortunately, however, all of the factors are not of great
importance, and the expression,
R = pSV2 X a constant.
is a fairly good first approximation. The factor f2 (Lg/V2),
while important in marine work, can be considered a constant
for air and eliminated from the expression. The factor /i (V/c),
representing the effect of compressibility, is usually eliminated
as unimportant for general airfoil work, since the velocities
ordinarily dealt with are far below the velocity of sound, but it
is of great importance in connection with propellers, for the tips
of the blades often travel at speeds higher than that of sound
in air, and the variation of R with V/c is great at these high
speeds. The expression LV/v is called Reynolds number and
is of considerable importance in airplane work, for there is
usually a great difference in size between the model and the
actual airplane.
1 A clear statement of the principles of dynamic similarity and dimensional
homogeneity is given in Some Aspects of the Comparison of Model and
Full-scale Tests, by D. W. Taylor, N.A.G.A.T.R. 219, 1926.
THE AIRFOIL 13
The exact variation of the aerodynamic characteristics with

Reynolds number and V/c is not known, and the results of
tests are usually assumed to vary in accordance with the first
approximation
R = pSV2 X a constant,
and attempts are made to correct for differences in Reynolds
number or "scale effect."
The Aerodynamic Forces on Airfoils.As stated at the
beginning of this chapter, if a flat plate is placed in a smooth
stream of air, either parallel or perpendicular to the flow, there
is a resistance in the direction of flow, but obviously no resultant
cross-force on the plate, since the air flow and pressures are
equal but opposite on opposite sides and therefore balance each
0 and 90 deg., there will be both a resistance and a lateral force

on the plate, and the respective proportion of these will vary with
every different angle. In many aeronautical problems it is
desirable to obtain a large cross-force or lift with a small resistance
or drag, and bodies called airfoils have been developed which
are much more effective in this respect than the flat plate.
Figure 5 shows a cross-section of an airfoil and the resultant
force acting upon it when it is inclined at a small angle in a
stream of air. The resultant force is usually resolved into
components perpendicular to and parallel to the relative airflow,
and these components represent the lift and drag of the airfoil.
The point where the resultant force vector intersects the chord
or base line of the airfoil section is called the center of pressure.
Force Coefficients.The aerodynamic forces on airfoils are
usually given for convenience in terms of dimensionless coeffi
cients of lift and drag. The coefficients used by the N.A.C.A.

are adhered to throughout this book, since they are considered the
most fundamental, having a physical significance, and are also
quite satisfactory for purposes of propeller design. These are
Lift coefficient = CL
qS }iPV2S
and
D _ D
Drag coefficient = CD
: qS MpV*S'
where L is the lift, D is the drag, q = }4pV2 is the dynamic
pressure due to the velocity of the air, and S is the projected
area of the airfoil in plan. The symbols CL and CD are used to
represent these particular coefficients only.
The Germans were the first to use the dynamic or impact
pressure in their lift and drag coefficients Ca and Cw. These
coefficients are the same as those used by the N.A.C.A.
excepting that they are made one hundred times as large in
order that whole numbers rather than small decimals may be
dealt with.
The British do not use the dynamic pressure in their coefficients
but use the density and velocity as such. Their coefficients are
therefore just one-half as large as the American NA.C.A.
coefficients and are given by the expressions
. _ L
L ~ oSV2
and
I. D
Kd ~ PSV2'
The symbols Lc and Dc are often used in place of kL and kD in

these expressions.
All of the above coefficients are dimensionless (that is, they
are pure or abstract numbers) and may be used with any con
sistent system of units such as the foot-pound-second or the
kilogram-meter-second systems. The English, or foot-pound-
second, system is used throughout this book.
Airplane designers have another set of coefficients called
engineering coefficients in which the velocity is in terms of miles
per hour, the area in square feet, and the mass density p is
included in the constant. The coefficients are therefore always
THE AIRFOIL 15
corrected to standard air for which p = 0.002378 lb.-ft.-4-sec.2

They are
Lift coefficient, K, =
and
Drag coefficient, Kx =
It is unfortunate that so many different coefficients exist for

practically the same purpose. Each one is related to any other
by a simple constant, however, and coefficients may be converted
from one type to another by the use of the following relations:
CL = 2(L or kL),
CL = S92K
cL = o.oic,
Le or kL = 0.5CL,
La or kL = imKv,
Le or kL = 0.005Ca,
Ky = 0.00255Ci;
Ky = 0.0051 (Lc or kL),
Ky = 0.0000255C.
The various drag coefficients have the same relations with each
other as the corresponding lift coefficients.
The ratio of lift to drag, L/D, which is the same as the ratio
of the lift coefficient to the drag coefficient, is used as a measure
of the effectiveness of an airfoil in producing lift at the cost of
drag. The L/D is especially important in connection with
propeller sections, for it affects the efficiency of the propeller.
The center of pressure CP. is usually given in the form of the
ratio AB/AC, shown in Fig. 5, which is the ratio of the distance
of the resultant force from the leading edge, on the chord line, to
the chord of the airfoil. Sometimes the moment about the lead
ing edge or that about the quarter chord point is used instead of
the CP. The moment coefficient used by N.A.CA. is
where c is the chord of the airfoil.

Methods of Plotting Coefficients.The lift and drag forces
on an airfoil are different for each different angle of attack.
The variations of lift, drag, and angle of attack are most con
veniently presented in graphical form, and many systems have

been devised for this purpose. In the one which until recently
has been in most common use, the coefficients of lift and drag,
and the ratio of lift to drag, are plotted as three separate curves
against the angle of attack. This method of plotting airfoil
characteristics is shown in Fig. 6. It is useful in connection with
the designing of propellers, because the angle of attack has an
important bearing on the pitch of a propeller.
.28 |4 T-a
24 1.2
20 1.0
16 0.8
L
D CL
12 0.6
8 0.4
4 0.2
0 0
-4 0 4 8 12 16" 20
Angle of Attack, cc
Fig. 6.Airfoil characteristics plotted against angle of attack.
The curves in Fig. 6 are for a typical propeller section. It will
be noticed that the lift coefficient increases with increased angle
of attack up to a certain maximum value, after which it falls off
as the angle becomes still greater. The drag coefficient has a
minimum value near zero degrees and increases with change of
angle of attack in either direction, becoming very great for the
high angles beyond maximum lift. It will also be noticed that
the lift coefficient is not zero at zero angle of attack but at an
angle of attack considerably less than zero. The attitude of a
typical airfoil when producing no lift is shown in Fig. 7.
THE AIRFOIL 17
In comparing the aerodynamic qualities of various airfoils,

the angle of attack is not of great importance, and a simpler
arrangement in which only one curve is used in place of the
above three is now in wide use. In this device, called a polar
Direction
ofAir Flow >.
Fig. 7.
diagram, the lift coefficients as ordinates are plotted directly

against the drag coefficients as abscissas, and the angles of attack
are indicated on the curve, as shown in Fig. 8. Since the drag
coefficients are relatively small, the scale for the drag coefficients
1.4
1.2
18
r
1.0
0.8

0.6 /
r
0.4
0
0.2
\'3
0
\
-0.2 _ . . ,
0 0.05 0.10 0.15 0.20 0.25
Co
Fig. 8.-Polar curve for an airfoil.
is usually made four or five times as large as that for the lift
coefficients. If the lift and drag coefficient scales are made
equal, the length of a vector from zero to any point on the curve
represents the coefficient of resultant force (hence the name polar
diagram), and the slope of the vector is the ratio of lift to drag.
The CP. is not of the same interest in connection with pro

pellers as with airplanes. It is of importance in propeller design
only because it affects the twisting of the blades which in turn
affects the pitch of the propeller. The CP. coefficient is usually
plotted against angle of attack and is shown for a typical pro
peller section in Fig. 9. It will be noted that as the angle of
attack is increased the CP. moves forward, tending to increase
the angle still further.
0|| 1 1 1 1 ..
L.
I 80
iool_L I 1 I I I
-4 0 4 8 12 16
Angle of Attack, cc
Fig. 9.
Pressure Distribution over Airfoils.The coefficients dealt

with up to this point apply to the position and the components
of the resultant force on an airfoil. This force is due to differ
ences in pressure over the upper and lower surfaces of the airfoil,
and it is sometimes of interest to know how the intensity of
pressure varies from point to point. Figure 10 shows the pressure
distribution over the midspan section of an airfoil at low and
high angles of attack.1 At the low angle, which is very nearly
that for zero lift, there is a positive pressure on the nose and there
are negative pressures on both the upper and lower surfaces.
(Positive and negative are relative to the static atmospheric
pressure. There is of course no absolute negative value of
pressure.) At the high angle, on the other hand, there is a
positive pressure on the lower surface and a negative pressure
1 The data for Fig. 10 were obtained from Pressure Distribution over Thick
AirfoilsModel Tests, by F. H. Norton and D. L. Bacon, N.A.C.A.r.ft.
150, 1922.
THE AIRFOIL 19
on the upper, both contributing to the lift. It is interesting

to note that at the high angle, the negative pressures on the
upper surface attain values much greater than the positive
pressures on the lower side, and the maximum pressures occur
near the nose, which explains the forward position of the CP.
at high angles. The total lift and drag forces can be obtained
by an integration of the pressures, but the drag obtained in this
manner is less than that obtained from force tests by the amount
of skin friction.
Fig. 10.-Pressure distribution at low and high angles of attack.
Effect of Variations in the Shape of Airfoils.From the

pressure-distribution diagrams in Fig. 10 it is evident that the
characteristics of an airfoil are mainly dependent on the upper
surface, so it is natural that changes in the upper surface have
considerably greater effect on the airfoil characteristics than
changes in the lower. Of the changes which have been made
in the upper surface, the variation of the amount of upper camber
is one of the most important, and many researches have been
made on series of airfoils having the same lower surfaces and
varying amounts of upper camber.1 The results of one of
these researches (N.A.C.A.T.R. 259) made on a series of flat-
bottomed propeller sections are shown in Fig. 11. In general,
1 Characteristics of Propeller Sections Tested in the Variable Density
Wind Tunnel, by Eastman N. Jacobs, N.A.C.A.T./2. 259, 1927. The Aero
dynamic Properties of Thick Airfoils, II, by F. H. Norton and D. L. Bacon,
N.A.C.A.T.R. 152, 1922. Airfoils for Airscrew Design, by W. L. Cowley
and H. Levy, British A.C.A.B. and M. 362, 1917-1918. Determination
of the Lift and Drift of Airfoils having a Plain Lower Surface and Variable
Camber, the Upper Surfaces Being Obtained by Varying the Ordinates in
a Constant Ratio, British R. and M. 60, 1911-1912
the lift coefficients are higher for the thicker sections at any
particular angle of attack, although the slopes of the lift curves
are lower for the thickest sections and the maximum lift coeffi
cients are lower than those for the sections of medium thickness.
The angle of attack for zero lift becomes lower as the upper
Maximum
Upper Camber
-- - 0.04
^- 0.08
Angle of Attack, cc
Fig. 11.-Variation of airfoil characteristics with upper camber.
camber is increased, reaching comparatively large negative

values for the thickest sections. Most important of all from
the point of view of propeller efficiency, the maximum effective
ness or L/D of the sections becomes lower as the upper camber
is increased.
A few tests have been made to determine the effect of varying
the position of the maximum ordinate of the upper surface of an
THE AIRFOIL 21
airfoil.1 The general conclusion from these tests is that for

most airfoils the best position of the maximum ordinate is about
one-third of the chord back from the leading edge. Moving it
forward slightly helps the maximum lift, while moving it back
ward slightly helps the maximum L/D. Thick sections seem
to be best when the maximum ordinate is a little back, and
thin sections when it is somewhat forward of the one-third
position.
Variations in the amount of lower camber have less effect
on the characteristics of airfoils than variations in the upper
camber, but they are nevertheless of some importance. Many
tests have also been made to determine the effect of varying
the amount of lower camber, and Fig. 12 shows the results found
from one investigation2 in which the models were tested at the
relatively high speed of 164 ft. per sec. Five airfoils were
tested in which the maximum lower camber varied from +0.05
(concave lower surface) to 0.07 (convex lower surface),
although the variations were not made in uniform steps. The
upper surfaces of all the sections were the same, having a maxi
mum ordinate of approximately 0.16. As can be seen from
Fig. 12, concave lower camber increases the lift and convex
lower camber decreases it. Convex lower camber seems to
have very little effect on the L/D, but judging from these tests
the L/D is reduced by concave lower camber. Other tests
(R. and M. 60) 2 indicate that concave lower camber also has
very little effect on L/D.
The effects of a multitude of other changes in the shapes of
airfoil sections have been tested, but a discussion of them all
does not lie within the scope of this chapter. The effect of
plan form is, however, of considerable importance. A series of
five airfoils having tips, as shown in Fig. 13, ranging from a
1 One of the best is The Determination of the Lift and Drift of Aerofoils
Having a Plain Lower Surface and Camber 0.10 of the Upper Surface, the
Position of the Maximum Ordinate Being Varied, British R. and M. 72,
1912-1913.
1 The Aerodynamic Properties of Thick Airfoils, II, by F. H. Norton and
D. L. Bacon, N.A.C.A.r.B. 152, 1922. Other tests on the effect of lower
camber are: Experiments on Thick Wing Sections, Expt. Dept., Airplane
Eng. Div., U. S. Army Air Service Bull, for December, 1918. Experiments
on a Series of Airfoils Having the Same Upper Surface and Lower Surfaces
of Different Camber, by L. Bairstow and B. M. Jones, British R. and M.
60, 1911-1912.
rectangle to a very slender ellipse, have been tested and the

results indicate that there is a minor improvement in L/D due
to rounding the tips but that there is no improvement to be had
Section No Maximum
Lower Camber
I + 0.05
2 0
-0.04
4 -0.05
5 -0.07
5.i. ,-2 1.6
L
14 D >tTO ?f~
3 4 1.4
i' w\
'!
>
I \( \ \ /
/hi i >
L j
&\
0 10 I0
i \
\
h \
nH t! y \
0.6
i
?i' A \\ \
'/ Is
A 0.6
i
u
7/ 1
/. //
/ OA
1 , '//
0" 6 12 20 24"
Angle of Attack, cc
Fig. 12.-Variation of airfoil characteristics with lower camber.
Fig. 13.
by making them narrower than the medium one, No. 3. 1 Another

series of tests was made on the characteristics of four airfoils
1 British A. C. A. Rept. T-477, February, 1915.
THE AIRFOIL 23
with various amounts of taper in plan, as shown in Fig. 14. 1

Those having a medium amount of taper had higher values of
L/D than either the one with no taper or the one tapered to
a point, and No. 3 was very slightly superior to No. 2. These
two sets of tests seem to indicate that in order to obtain the
highest possible efficiency, a wing or
propeller blade should be somewhat l
tapered and rounded at the tip,
although of course it is not certain "
that airfoil data of this kind can be ' >
applied directly to propeller blades.
The change in plan form having
the greatest effect on airfoil charac
teristics is change of aspect ratio,
which is defined as the span divided
by the mean chord, or the square of
the span divided by the wing area.
The effect of aspect ratio on the Fig. 14.
characteristics of an airfoil can be
satisfactorily explained and calculated theoretically and will be
considered in the next section.
Elementary Airfoil Theory.A theory of the airfoil, called the
induced-drag theory or the Lanchester-Prandtl wing theory,
was suggested by Lanchester in Great Britain in 1907 and worked
out separately by Prandtl and some of his associates in Germany.2
A brief physical concept and the results of the theory in the form
of final simple equations are all that will be given here.
Just as a propeller must give a backward motion to a mass
of air in order to develop a forward thrust (as was brought out
in the momentum theory), a wing must give a mass of air a down
ward motion in order to obtain an upward lift. If, then, a
horizontal stream of air is flowing past an airfoil which is giving
lift, the direction of some of the air is changed downward from
the horizontal by a certain angle called the downwash angle
1 The Aerodynamic Properties of Thick Aerofoils, II, by F. H. Norton
and D. L. Bacon, N.A.C.A.T.R. 152, 1922.
2 Prandtl has made a brief statement of his theory in English in Applica
tions of Modern Hydrodynamics to Aeronautics, by L. Prandtl, N.A.C.A.-
T.R. 116, 1925. See also Aerofoil Theory, by H. Glauert, British R. and M.
723, 1920-1921, which is a survey of the work of Prandtl, Betz, and others;
and "Aerofoil and Airscrew Theory," by H. Glauert, Cambridge University
Press, 1926.
(see Fig. 15). The mea/i direction of airflow past the wing is
indicated by the dotted arrow in Fig. 15 and has an angle equal
to half the downwash angle, or ^ ; and according to the induced-
drag theory, the angle of attack of the airfoil with respect to
this line of mean flow direction is the angle of attack which the air
foil would have if its span were infinite. The flow would then be
two dimensional and there would be no downwash, for the wing
would be acting upon an infinite mass of air. The flow about a
finite airfoil relative to the mean direction is assumed to be the
same as that for infinite span, and the components of the resultant
Fig. 15.
force referred to the mean direction of flow are called profile

lift and profile drag because they depend only on the cross-
sectional shape or profile of the airfoil.
The profile lift L0 and profile drag D0 are shown in Fig. 15
along with the ordinary lift L and drag D, which are referred to
the direction of the airflow before it is disturbed by the airfoil.
The two lifts L and L0 may be considered as having the same
magnitude, for they differ only by an insignificant component
of the small profile drag. The component of the profile lift,
however, has considerable magnitude in the direction of the
ordinary drag, and this component is called the induced drag,
for it is the drag which occurs because lift is obtained with a
wing of finite span.
The induced drag has the minimum value for an airfoil having
what is known as elliptical lift distribution, which means that
the lift grading across the span has the form of a semiellipse.
THE AIRFOIL 25
For this type of span loading the induced-drag coefficient is given

by the following expression :
where R is the aspect ratio, or the span squared divided by

the area. According to this equation the induced drag depends
only on the square of the lift coefficient and the aspect ratio
and is entirely independent of the airfoil section. The induced
angle, which is half the downwash angle is,
6 CL
where a< is in radians.

The lift coefficient for infinite aspect ratio is then considered
the same as that for a finite airfoil, and the drag (profile) and
angle of attack for infinite aspect ratio are
rt n L
and
CL
ao = a s>
irti
all of the angles being in radians.
Most model airfoil tests are made with airfoils having a rec
tangular plan form, for which the lift distribution across the span
is not elliptical. The above equations give values of the induced
drag which are as much as 5 per cent too low and values of
the induced angle as much as 25 per cent too low for wings of
rectangular plan form and aspect ratio 6, and since the angles are
important in dealing with propeller sections, the equations are
not satisfactory. More accurate expressions are used which were
developed by Betz for wings of rectangular plan form and
simplified to the following form by Glauert:1
a
ao = a j
a
1 The Calculation of the Characteristics of Tapered Wings, by H. Glauert,
British B. and M. 767, 1921-1922.
where a is the slope of the straight portion of the lift curve, or

dCL/da, and the angles are measured in radians. The values of
5 and are functions of R/a, where R is the aspect ratio, and
these functions are given in Fig. 16. If the angles are expressed
in degrees, the slope becomes
180 dCL
a - j >
7r act
0-6i 1 1n 1 1 1 1 1 1 1 1 1 1 1 1
0.5
0.4
t,tmiS
0.3
0.2
0.1
0
0 ' R 2 3
a
Fig. 16.
and the angle of attack for infinite aspect ratio becomes

180 CLl
ao = a
x a
The use of these transformation formulas can be made clear
by means of an example. Consider the airfoil of aspect ratio 6
and camber 0.10 (Fig. 20). At an angle of attack of 3.2,
CL = 0.543 and CD = 0.0300. The slope of the lift curve is
a = 4.08, and R/a = 6/4.08 = 1.47. Then, from Fig. 16,
= 0.256 and 5 = 0.228, and the induced drag is
CD. = ^ = 0.0559 CL2 = 0.0164,

' a
and the profile drag is
CDo = CD - CDi = 0.0136.
THE AIRFOIL 27
In like manner the induced angle is

= 180C* = 3 = 20o
ira
, and the angle for infinite aspect ratio is
ao = a c*i = 1.2.
In recent years there has been a tendency in wind tunnel
airfoil tests to use larger models with respect to the size of
the tunnel than formerly, the span now often being half the
diameter of the air stream or greater. In this way a larger
scale or Reynolds number is obtained but the effect of the
tunnel wall interference on the airfoil characteristics is important,
and the airfoil characteristics should be corrected to free air
conditions before being applied to propellers. Prandtl has
developed theoretical expressions for calculating the tunnel
wall effect,1 and as in the case of the Betz rectangular wing
formulas, Glauert has put them in slightly more usable form.2
As in the transformation to infinite aspect ratio, corrections are
made for the drag and the angle of attack at a constant lift.
The correction factors for wind tunnels of circular cross-section
are
ACD =
8A
SCL
Act =
8A '
where A is the cross-sectional area of the air stream and S is
the area of the airfoil. The corrections are to be added to the
drag coefficient and angle of attack as measured in a closed
tunnel and subtracted from the values measured in an open
jet tunnel. Since the corrections depend on the lift and area
of the airfoil and are independent of the aspect ratio or plan
form, they apply to any wing system whether of one or more
planes.
The theoretical corrections for tunnel wall interference and
the transformation equations to infinite aspect ratio have been
well checked experimentally and are sufficiently accurate for
practical use.
1 Applications of Modern Hydrodynamics to Aeronautics, by L. Prandtl,
N.A.C.A.T.R. 116, p. 49, 1925.
* Aerofoil and Airscrew Theory," by H. Glauert, Cambridge University
Press, 1926.
Effects of Scale and Compressibility.The principle of

dynamic similarity indicates that the lift and drag coefficients
of airfoils are not constants but vary with Reynolds number and
the compressibility factor V/c. A number of tests have been
made at values of Reynolds number ranging from those commonly
used in model tests to full scale for airplane wings and propeller
blades. These tests have been made in the variable-density
wind tunnel of the N.A.C.A., in which the Reynolds number
is varied by compressing the air and , thereby changing the
kinematic viscosity v. The variation of maximum L/D with
Reynolds number is shown for the U. S. A. 35-B airfoil in Fig.
17. 1 The lift and drag of different airfoils vary in different
24 1 1 1 1 1 1 1 1
0 1,000,000 2,000,000 3,000,000

Reynolds Nurnber
Fig. 17.
ways with Reynolds number, and no general rule can be given
at this stage. The greatest variation, however, occurs at the
low values of Reynolds number, and the high-scale tests are more
consistent and reliable.
No tests of real quantitative value have been made to date
on airfoils at velocities approaching that of sound in air, but
two series of tests have been made by Briggs and Dryden2 which
are of interest in showing the general trend of the variation
of airfoil characteristics with V/c. The variation of the maxi
mum value of L/D for a propeller section of thickness ratio
1 The data for Fig. 17 were obtained from The Comparison of Well Known
and New Wing Sections Tested in the Variable Density Tunnel, by George J.
Higgins, N.AC.A.T.iV. 219, May, 1925.
2 Aerodynamic Characteristics of Airfoils at High Speeds, by L. J. Briggs,
G. F. Hull, and H. L. Dryden, N.A.C.A.T.B. 207, 1925. Pressure
Distribution over Airfoils at High Speeds, by L. J. Briggs, and H. L. Dryden,
N.A.C.A.T./2. 255, 1927. Aerodynamic Characteristics of Twenty-four
Airfoils at High Speeds, by L. J. Briggs, and H. L. Dryden, N.A.C.A.T.R.
319, 1929.
THE AIRFOIL 29
0.12 is shown in Fig. 18. It may be said that in general there

is a decrease in lift and an increase in drag with increasing
V/c and a consequent reduction in effectiveness or L/D as the
speed approaches that of sound. Also, the variation of the
characteristics of thin sections is considerably less than for thick
sections, with both V/c and Reynolds number.
Airfoil Sections Suitable for Propellers.Because of strength
considerations, aircraft propeller blades are usually tapered in
thickness and plan form, so that the chord and camber of the
16,
Max.-.
Fig. 18.
airfoil sections are different at different radii along the blade.
It is necessary, of course, in order to analyze the aerodynamic
performance of a propeller, section by section, to have some
knowledge of the airfoil characteristics of the sections along the
blade. Since it is not feasible to test a series of airfoils for every
propeller designed, it is customary either to adhere to a standard
form of blade, representative airfoils of which have been tested,
or to use data from a series of airfoils of varying thickness ratio
and interpolate for the actual thickness ratio at each section
of the propeller if necessary. The latter method allows the
propeller designer greater freedom, for it can be used with any
plan form and variation of thickness along the radius, and it is at
present in general use in the United States.
The series of propeller sections which has been most widely

employed and tested is based upon the R.A.F.-6 airfoil modified
so as to have a flat lower surface. (The practice of making the
lower surface or face of propeller sections flat over the working
or outer portion of the blades is almost universal because it
facilitates construction, and it is entirely justified, since the best
flat-bottomed airfoils are quite as effective aerodynamically as
any others.) In the series based on the modified R.A.F.-6
section, the upper camber is varied by scaling all of the ordinates
up or down in the same proportion as the maximum ordinate.
The ordinates for these sections are given in terms of the maxi
mum ordinate in Fig. 19. If the blade width or chord and the
thickness of a section are known, all of the ordinates of the section
-Maximum Ordinate = Unify
V
) V
0.025 0.05 0.10 0.ZO 0.J0 0.40 0.50 0.60 0.70 0 80 0.90 1.00
Fraction L.E. T.E.

of Chord Rod 0.025 0.05 0.10 0.20 0.30 0.40 0.5o|o.6O 0.70 0.80 0.90 Had.
Ordinate 0.10 0.41 0.59 0.79 0.95 0.998 0.99 0.95 1 0.87 0.74 0.56 0.15 0.08
Fig. 19.-Ordinates of standard propeller section based on R.A.F.-6.
can be calculated by means of the relations given in Fig. 19.

The aerodynamic characteristics are given for this series of
sections in Figs. 20 and 21 for aspect ratio 6. The values have
been corrected for tunnel wall interference to free air conditions
and have been faired and cross-faired to eliminate minor irregu
larities due to experimental error. The airfoils were tested in
the variable-density wind tunnel of the N.A.C.A., both at 1
atmosphere and at 20 atmospheres pressure, corresponding to
ordinary model test-scale and full-scale Reynolds numbers,
respectively.1 The sections, excepting for the very thickest
ones, have very little scale effect with Reynolds number, and the
characteristics given in Figs. 20 and 21, which are taken from
the high-scale tests, are considered satisfactory as far as scale is
concerned for propeller sections moving at velocities up to 300
or 400 ft. per sec. Model propeller tests usually fall within this
1 Characteristics of Propeller Sections Tested in the Variable Density
Wind Tunnel, by Eastman N. Jacobs, N.A.C.A.T.#. 259, 1927.
THE AIRFOIL 31
range of tip speed but full-scale propellers do not, so that for

full-scale propellers a scale correction of some kind is ordinarily
applied.
It will be noticed that in Fig. 21 the effectiveness of the
airfoils is plotted in terms of D/L and that a scale for 7,i which
is the angle between the lift and the resultant force vectors, is
oc
Fig. 20.-Standard propeller sections based on R.A.F.-6. Aspect ratio = 6.
also given. This is for convenience in working with the blade-

element theory, in which 7 or its tangent D/L is used directly.
The sections near the hub of a propeller are usually given
a convex shape or negative lower camber. If the series of
propeller sections based on the modified R.A.F.-6 is used, the
convex sections are obtained by putting two regular flat-faced
sections face to face and rounding the leading and trailing edges
as shown in Fig. 22. There are no entirely satisfactory data

regarding the effect on the aerodynamic characteristics of adding
0.34
Fig. 21.-Standard propeller sections based on R.A.F.-6. Aspect ratio = 6.
Fig. 22.
various amounts of lower camber in this manner to airfoil sections
of various thickness. The sections having convex lower camber,
however, being near the hub, are relatively unimportant, and it
THE AIRFOIL 33
is sufficiently accurate for ordinary purposes to reduce the lift

of the upper flat-faced section alone by an amount depending
on the added lower camber as given in Fig. 23. The L/D is but
very slightly affected by a moderate addition of lower camber
and may be taken as being the same as that for the flat-faced
upper section alone.
The characteristics for the standard series of flat-faced pro
peller sections based on the modified R.A.F.-6 are also given for
infinite aspect ratio, or two-dimensional flow, in Figs. 24 and 25.
04 1iiiiiiiiiii|
ACL 0.2
0.04- 0.06 008 0.12

Convex Lower Camber
NOTE: For a Section with LenverCamber,CL-Cl.(fiatface)~a^.
Fig. 23.-Correction to lift coefficient for convex lower camber.
These propeller sections have become more or less standard,

mainly because of the considerable amount of test data available
regarding them,1 although lately other sections have been
developed similar to the Clark-Y and the U. S. A. 35-B which
are somewhat better aerodynamically and are being used in
propellers where very high efficiency is desired. The Clark-Y
and U. S. A. 35-B sections have approximately the same thickness
as the standard propeller section of thickness ratio 0.12, and the
1 In addition to the reference in the previous footnote, test data on these
propeller sections may be obtained from the following sources: "The Air
plane Propeller," prepared by the Propeller Section, Eng. Div., U. S. Army
Air Service, 1920, and published by the Government Printing Office.
Wind Tunnel Studies in Aerodynamic Phenomena at High Speed, by
F. W. Caldwell, and E. N. Fales, N.A.C.A.r.B. 83, 1920. Aerodynamic
Characteristics of Airfoils at High Speeds, by L. J. Briggs, G. F. Hull,
and H. L. Dryden, N.A.C.A.T.K. 207, 1925. Pressure Distribution over
Airfoils at High Speeds, by L. J. Briggs, and H. L. Dryden, N.A.C.A.-
T.R. 255, 1927.
CL and L/D curves obtained from high-scale tests in the variable-

density tunnel are plotted for these three sections in Fig. 26.
The lift curves for all three airfoils are practically the same
Fig. 24.-Standard propeller sections based on R.A.F.-6. Infinite aspect ratio.
for ordinary angles of attack, so that in the absence of test

data on a series of propeller sections based on the Clark-Y or
the U. S. A. 35-B, the lift curves for the standard sections may
THE AIRFOIL 35
be used directly for the new-type sections. The values of L/D

for the newer sections range from 10 to 30 per cent higher than
those for the standard propeller sections.
0.32
Fig. 25.-Standard propeller sections based on R.A.F.-6. Infinite aspect ratio.
Several other series of sections suitable for propellers have

been tested, 1 mainly for the purpose of checking various forms
1 Airfoils for Airscrew Design, by W. L. Cowley and H. Levy, British R.
and M. 362, 1917-1918. Aerodynamic Characteristics of Aerofoils, II, N.A.-
C.A.T.R. 124, 1921, Fage and Collins airfoils 1 to 6. Comparison of Model
Propeller Tests with Airfoil Theory, by W. F. Durand, and E. P. Lesley,
N.A.C.A.T.B. 196, 1924. A Consideration of Airscrew Theory in the
of blade-element analysis against model propeller tests. In

some of these, the airfoils were tested at the same values of
Reynolds number as the corresponding sections in the model
Standard Propeller Section,
Thickness Ratio, 0.12
d Clark-Y
U.S. A -35B
1
/ r\ \P.S.T.R. 0.12
11 1.4
// \
\
v \ . Oar k-Y 1 \
// \ \
20 1.S./ -A 1.2
35-B \ \
\ \\ 1.S.A
f Clc irk- f \ IS-B
\ \ /
r 1.0
\ N
^ \ v/
p.s:'
Thicknes 5 \
}i// Hat o 0.1 - / >\ 0.8
\\ \\
14 \ 0.6
\
J
Ji \ 0.4
f
<
10 0.2
\
/o * o 0
Angle of Attack, cc
Fig. 26.
propellers, and since the tip speeds were very low compared
with the speed of sound in air, this practically eliminated errors
iue to scale.
Light of Data Derived from an Experimental Investigation of the

Distribution of Pressure over the Entire Surface of an Airscrew Blade, and
also over Aerofoils of Appropriate Shapes, by A. Fage, and R. G. Howard,
British R. and M. 681, 1921-1922. Experiments with a Family of
Airscrews, Part III. Analysis of the Family of Airscrews by Means of
the Vortex Theory and Measurements of Total Head, by C. N. H. Lock,
and H. Bateman, British R. and M. 892, 1923-1924.
CHAPTER IV
THE SIMPLE BLADE-ELEMENT THEORY
While the momentum theory is useful for determining ideal

efficiency, it gives a very incomplete account of the action
of screw propellers, neglecting among other things the torque.
In order to investigate propeller action in greater detail, the blades
are considered as made up of a number of small elements, and
the air forces on each element are calculated. Thus, while the
momentum theory deals with the flow of the air, the blade-
element theory deals primarily with the forces on the propeller
blades. The idea of analyzing the forces on elementary strips
of propeller blades was first published by William Froude in
1878. 1 It was also worked out independently by Drzewiecki
and given in a book on mechanical flight published in Russia
seven years later, in 1885. 2 Again, in 1907, Lanchester pub
lished a somewhat more advanced form of the blade-element
theory without knowledge of previous work on the subject.
The simple blade-element theory is usually referred to, however,
as the Drzewiecki theory, for it was Drzewiecki who put it into
practical form and brought it into general use. Also, he was
the first to sum up the forces on the blade elements to obtain
the thrust and torque for a whole propeller and the first to intro
duce the idea of using airfoil data to find the forces on the blade
elements.
In the Drzewiecki blade-element theory the propeller is
considered a warped or twisted airfoil, each segment of which
1 The Elementary Relation between Pitch, Slip, and Propulsive
Efficiency, by William Froude, Trans. Inst. Naval Architects, 1878.
2 This fact, which is not generally known in English-speaking countries,
was called to the author's attention by Prof. F. W. Pawlowski of the Univer
sity of Michigan. Drzewiecki's first French paper on his theory was
published in 1892. He wrote in all seven papers on aircraft propulsion
which were presented to l'Academie des Sciences, l'Association Technique
Maritime, and Le Congres International d'Architecture et de Construction
Navale, held on July 15, 1900. He finally wrote a book summing up all of
his work called "Thcbrie Generale de 1'Helice Propulsive," published
in 1920 by Gauthier-Villars in Paris.
37
follows a helical path and is treated as a segment of an ordinary

wing. It is usually assumed in the simple theory that airfoil
coefficients obtained from wind tunnel tests of model wings
(ordinarily tested with an aspect ratio of 6) apply directly to
Fig. 27.
propeller blade elements of the same cross-sectional shape.1
The air flow around each element is considered two-dimensional
and therefore unaffected by the adjacent parts of the blade.
The independence of the blade elements at any given radius
-j 1 X,
.Ziirn
Fig. 28.
with respect to the neighboring elements has been established
theoretically2 and has also been shown to be substantially true
for the working sections of the blade by special experiments3
1 In his book, Drzewiecki suggested that the airfoil characteristics could be
obtained from tests on special model propellers.
2 "Aerofoil and Airscrew Theory," by H. Glauert, Cambridge University
Press, 1926.
3 Experiments to Verify the Independence of the Elements of an
Airscrew Blade, by C. N. H. Lock, H. Bateman, and H. C. H. Townend,
British R. and M. 953, 1924.
THE SIMPLE BLADE-ELEMENT THEORY 39
made for the purpose. It is also assumed that the air passes
through the propeller with no radial flow (i.e., there is no contrac
tion of the slipstream in passing through the propeller disc) and
that there is no blade interference.
The Aerodynamic Forces on a Blade Element.Consider the
element at radius r, shown in Fig. 27, which has the infinitesimal
length dr and the width b. The motion of the element in an
aircraft propeller in flight is along a helical path determined
by the forward velocity V of the aircraft and the tangential
velocity 2wrn of the element in the plane of the propeller disc,
where n represents the revolutions per unit time. The velocity
of the element with respect to the air Vr is then the resultant
of the forward and tangential velocities, as shown in Fig. 28.
Call the angle between the direction of motion of the element
and the plane of rotation </>, and the blade angle /?. The angle
of attack a of the element relative to the air is then a = /3 <t>.
Applying ordinary airfoil coefficients, the lift force on the
element is ~7 , J
Let t be the angle between the lift component and the resultant
force, or 7 = arc tan D/L. Then the total resultant air force
on the element is
y4pVr2CLbdr
dR =
cos 7
The thrust of the element is the component of the resultant
force in the direction of the propeller axis (Fig. 28), or
dT = dR cos (4> + 7)
lApVSCjbdr cos Q + 7)
cos 7
V
and since Vr =
J
For convenience let
CLb
K
sin2 (t> cos 7
and
K cos O + 7).
Then
dT y2PV*T4r,
and the total thrust for the propeller (of B blades) is

T = MPV*Bf*T4r.
Referring again to Fig. 28, the tangential or torque force is
dF = dR sin (<t> + y),
and the torque on the element is
dQ = rdR sin (</> + 7),
which, if Qc = Kr sin (<t> + y), can bewritten
P dQ ^y2pV2Qcdr^
The expression for the torque of the whole propeller is therefore
The horsepower absorbed by the propeller, or the torque horse

power, is
QHP ~ "550 '
and the efficiency is
THP TV
' QHP 2tuQ
The Efficiency of an Element.Because of the variation
of the blade width, angle, and airfoil section along the blade,
it is not possible to obtain a simple expression for the thrust,
torque, and efficiency of propellers in general. A single element
at about two-thirds or three-fourths of the tip radius is, however,
fairly representative of the whole propeller, and it is therefore
interesting to examine the expression for the efficiency of a
single element. The efficiency of an element is the ratio of the
useful power to the power absorbed, or
= dTV
v ~ dQ2im
_ dR cos (<t> + y)V
dR sin (<j) + y)2irnr
tan <t>
tan (4> + y)
Now tan <t> is the ratio of the forward to the tangential velocity,
and tan y = D/L. According to the simple blade-element
theory, therefore, the efficiency of an element of a propeller
depends only on the ratio of the forward to the tangential velocity

and on the L/D of the airfoil section.
The value of 4> which gives the maximum efficiency for an
element, as found by differentiating the efficiency with respect
to <t> and equating the result to zero, is
* = 450 - i
The variation of efficiency with <t> is shown in Fig. 29 for
two extreme values of 7. The efficiency rises to a maximum
at 45 t/2 and then falls to zero again at 90 y. With
an L/D of 28.6 the maximum possible efficiency of an element
according to the simple theory is 0.932, while with an L/D of
1.20
1.00
^0.80
o
.1o 0.60
uj 0.40
0.2b
0
0 10" Z0 30' 40 50 60 70 8Q 90
Fig. 29.
9.5 it is only 0.812. At the values of <t> at which the most
important elements of the majority of propellers work (10 to
15) the effect of L/D on efficiency is still greater. Within
the range of 10 to 15, the curves in Fig. 29 indicate that it is
advantageous to have both the L/D of the airfoil sections and
the angle <t> (or the advance per revolution, and consequently
the pitch) as high as possible.
The Limitations of the Simple Blade-element Theory.
According to the momentum theory a velocity is imparted to the
air passing through the propeller, and half of this velocity is
given the air by the time it reaches the propeller plane. This
increase of velocity of the air as it passes into the propeller
disc is called the inflow velocity. It is always found where
there is pressure discontinuity in a fluid. In the case of a
wing moving horizontally, the air is given a downward velocity,
as shown in Fig. 15 of Chap. Ill, and theoretically half of this

velocity is imparted in front of and above the wing, and the other
half below and behind. This induced downflow is present in
the model wing tests from which the airfoil coefficients used '
in the blade-element theory are obtained; the inflow indicated
by the momentum theory is therefore automatically taken into
account in the simple blade-element theory. As shown in Chap.
Ill, however, the induced downflow is widely different for
different aspect ratios, being zero for infinite aspect ratio.
Most model airfoil tests are made with rectangular wings having
an arbitrarily chosen aspect ratio of 6, and there is no reason to
Fig. 30.
suppose that the downflow in such a test corresponds to the

inflow for each element of a propeller blade. In fact, the general
conclusion drawn from an exhaustive series of tests,1 in which
the pressure distribution was measured over 12 sections of a
model propeller running in a wind tunnel, is that the lift coefficient
of the propeller blade element differs considerably from that
measured at the same angle of attack on an airfoil of aspect
ratio 6. This is one of the greatest weaknesses of the simple
blade-element theory.
Another weakness is that the interference between the propeller
blades is not considered. The elements of the blades at any
1 A Consideration of Airscrew Theory in the Light of Data Derived from
an Experimental Investigation of the Distribution of Pressure over the
Entire Surface of an Airscrew Blade, and also over Airfoils of Appropriate
Shapes, by A. Fage, and R. G. Howard, British R. and M. 681, 1921.
particular radius form a cascade similar to a multiplane with

negative stagger, as shown in Fig. 30. Near the tips where
the gap is large the interference is very small, but in toward the
blade roots it is quite large.
In actual propellers there is a tip loss which the blade-element
theory does not take into consideration. The thrust and torque
forces as computed by means of the theory are therefore greater
for the elements near the tip than those found by experiment.1
In order to eliminate scale effect, the wind tunnel tests on
model wings should be run at the same value of Reynolds number
(scale) as the corresponding elements in the propeller blades.
Airfoil characteristics measured at such a low scale as, for example,
an air velocity of 30 m.p.h. with a 3-in. chord airfoil, show peculi
arities not found when the tests are run at a scale comparable
with that of propeller elements. The standard propeller section
characteristics given in Figs. 20, 21, 24, and 25 (Chap. Ill)
were obtained from high Reynolds-number tests in the Variable
Density Tunnel of the N.A.C.A., and, fortunately, for all except
ing the thickest of these sections there is very little difference in
characteristics at high and low Reynolds numbers. These values
may be used with reasonable accuracy as to scale for propellers
operating at tip speeds well below the speed of sound in air, and
therefore relatively free from any effects of compressibility.
The poor accuracy of the simple blade-element theory is very
well shown in a report by Durand and Lesley,2 in which they
have computed the performance of a large number of model
propellers (80) and compared the computed values with the
actual performances obtained from tests on the model propellers
themselves. In the words of the authors:
The divergencies between the two sets of results, while showing
certain elements of consistency, are on the whole too large and too
capriciously distributed to justify the use of the theory in this simplest
form for other than approximate estimates or for comparative purposes.
The airfoils were tested in two different wind tunnels and in
one of the tunnels at two different air velocities, and the propeller
characteristics computed from the three sets of airfoil data
1 An Analysis of the Family of Airscrews by Means of the Vortex
Theory and Measurements of Total Head, by C. N. H. Lock, and H.
Bateman, British R. and M. 892, 1923.
2 Comparison of Model Propeller Tests with Airfoil Theory, by William
F. Durand, and E. P. Lesley, N.A.C.A.r.R. 196, 1924.
differ by as much as 28 per cent, illustrating quite forcibly the

necessity for having the airfoil tests made at the correct scale.
In spite of all its inaccuracies the simple blade-element theory
has been a useful tool in the hands of experienced propeller
designers. With it a skilful designer having a knowledge of
suitable empirical factors can design propellers which usually fit
the main conditions imposed upon them fairly well in that they
absorb the engine power at very nearly the proper revolution
speed. They are not, however, necessarily the most efficient
propellers for their purpose, for the simple theory is not suffi
ciently accurate to show slight differences in efficiency due to
changes in pitch distribution, plan forms, etc.
Example of Analysis with the Simple Blade -element Theory.
In choosing a propeller to analyze, it is desirable that its aero
dynamic characteristics be known so that the accuracy of the
calculated results can be checked. It is also desirable that
the analysis be made of a propeller operating at a relatively
low tip speed in order to be free from any effects of compressibility
and that it be running free from body interference. The only
propeller tests which satisfy all of these conditions are tests of
model propellers in a wind tunnel. We shall therefore take for
our example the central or master propeller of a series of model
wood propellers of standard Navy form, tested by Dr. W. F. Dur-
and at Stanford University.1 This is a two-bladed propeller
3 ft. in diameter, with a uniform geometrical pitch of 2.1 ft.
(or a pitch-diameter ratio of 0.7). The blades have standard
propeller sections based on the R.A.F-6 airfoil (Fig. 19), and
the blade widths, thicknesses, and angles are as given in the first
part of Table I. In our analysis we shall consider the propeller
as advancing with a velocity of 40 m.p.h. and turning at the rate
of 1,800 r.p.m.
For the section at 75 per cent of the tip radius, the radius is
1.125 ft., the blade width is 0.198 ft., the thickness ratio is
0.107, the lower camber is zero, and the blade angle |3 is 16.6.
The forward velocity V = 40 m.p.h.
= 40 X 88
60
= 58.65 ft. per sec,
1 Tests on Thirteen Navy Type Model Propellers, by W. F. Durand,
N.A.C.A.r./2. 237, 1926, propeller model C.
and
1,800
n =
60
= 30 r.p.s.
V
The path angle <t> = arc tan
2irrn
58.65
= arc tan
2tt X 1.125 X 30
= 15.5.
The angle of attack is therefore
a = 0 - <fi
= 16.6 - 15.5
= 1.1.
From Fig. 21, for a flat-faced section of thickness ratio 0.107
at an angle of attack of 1.1, y = 3.0, ' and, from Fig. 20,
CL = 0.425. (For sections having lower camber, CL should be
corrected in accordance with the relation given in Fig. 23, and y
is given the same value as that for a flat-faced section having the
upper camber only.)
Then
sin* 4> cos y

= 0.425 X 0.198
0.26722 X 0.999
= 1.180,
and
Tc = K cos O + 7)
= 1.180 X cos 18.5
= 1.119.
Also,
Qc = Kr sin (0 + 7)
= 1.180 X 1.125 X sin 18.5
= 0.421.
The computations of Tc and Qc for six representative elements
of the propeller are given in convenient tabular form in Table I,
and the values of Tc and Qc are plotted against radius in Fig. 31.
The curves drawn through these points are sometimes referred
to as the thrust and torque grading curves. The areas under the
curves represent
Tcdr
and
these being the expressions for the total thrust and torque
per blade per unit of dynamic pressure due to the velocity of
advance. The areas may be found by means of a planimeter,
0.6 0.3
0.4 0.2
0.2 0.1
0.6 0.8 1.0 1.2 1.4.

Radius in Ft.
Fig. 31.Thrust and torque grading curves.
proper consideration, of course, being given to the scales of

values, or the integration may be performed approximately
(but with satisfactory accuracy) by means of Simpson's rule.
In using Simpson's rule the radius is divided into an even number
of equal parts, such as ten. The ordinate at each division can
then be found from the grading curve. If the original blade
elements divide the blade into an even number of equal parts it is
not necessary to plot the grading curves, but the curves are
advantageous in that they show graphically the distribution of
thrust and torque along the blade. They also provide a check
upon the computations, for incorrect points will not usually form
a fair curve.
Table I.Computations fob Pbopelleb Analysis with Simple Blade-
element Theoby
D = 3.0 ft. Forward velocity = 40 m.p.h. = 58.65 ft. /sec.
p = 2.1 ft. Rotational; velocity = 1,800 r.p.m. = 30 r.p.s.
r/R 0.15 0.30 0.45 0.60 0.75 0.90

r(ft.) 0.225 0.450 0.675 0.900 1.125 1.350
b (ft.) 0.225 0.236 0.250 0.236 0.198 0.135
hv/b 0.190 0.200 0.167 0.133 0.107 0.090
w& 0.180 0.058 0.007 000 000 000
0(deg.) 56.1 36.6 26.4 20.4 16.6 13.9
2wn 42.3 84.7 127.1 169.6 212.0 254.0
1.389 0.693 0.461 0.346 0.277 0.231
0 (deg.) 54.2 34.7 24.7 19.1 15.5 13.0
a = 0 - <t> (deg.) 1.9 1.9 1.7 1.3 1.1 0.9
7 (deg.) 3.9 4.1 3.6 3.3 3.0 3.0
0.998 0.997 0.998 0.998 0.999 0.999
Cl 0.084 0.445 0.588 0.514 0.425 0.356
sin 0 0.8111 0.5693 0.4179 0.3272 0.2672 0.2250
A'- cos 7('' sin' </> 0.0288 0.325 0.843 1.135 1.180 0.949
0 + 7 (deg.) 58.1 38.8 28.3 22.4 18.5 16.0
cob (<t> + y) 0.5280 0.7793 0.8805 0 . 9245 0.9483 0.9613
Tc = Kcos (* +7).... 0.0152 0.253 0.742 1.050 1.119 0.912
sin (0 + 7) 0.8490 0.6266 0.4741 0.3811 0.3173 0.2756
Qc = Kr sin (<*> + 7) . . . 0.0055 0.0916 0.270 0.389 0.421 0.353
If the abscissas are denoted by r and the ordinates at the

various divisions by y\, y2, . . . yn, according to Simpson's rule
the area with ten equal divisions will be
R
^ F(r)dr = -jiVi + 2(y* + V + Vi + V*) + ^ + Vi+ye +
Jo
y% + 2/10) + yu\.
The area under the thrust-grading curve of our example is
therefore
RTcdr = ^So + 2(0.038 + 0.600 + 1.050 + 1.091) +
4(0 + 0.253 + 0.863 + 1.120 + 0.912) + 0]

= 0.9075,
and in like manner
f* Qcdr = 0.340.
The above integrations have also been made by means of a

planimeter, and the average results from five trials agree with
those obtained by means of Simpson's rule within one-fourth
of one per cent.
The thrust of the propeller in standard air is
T = IipV*Bf*TJr
= MX 0.002378 X 58.652 X 2 X 0.9075
= 7.42 lb.,
and the torque is
Q = y2PV*B$*Qcdr
= HX 0.002378 X 58.652 X 2 X 0.340
= 2.78 lb.-ft.
The power absorbed by the propeller is
P = 2mQ
= 2 X x X 30 X 2.78
= 524 ft.-lb. per sec.
or
524
HP =
550
0.953,
and the efficiency is
TV
' 2-rnQ
= 7.42 X 58.65
524
= 0.830.
The above-calculated performance compares with that meas
ured in the wind tunnel as follows:
Calculated Model test
Power absorbed, horsepower 0.953 1.073

7.42 7.77
0.830 0.771
The power as calculated by the simple blade-element theory

is in this case over 11 per cent too low, the thrust is about 5
per cent low, and the efficiency is about 8 per cent high. Of
course, a differently calculated performance would have been
obtained if propeller-section characteristics from tests on the
same series of airfoils in a different wind tunnel had been used,
but the variable-density tunnel tests are probably the most
reliable of all.
Some light may be thrown upon the discrepancy between the
calculated and observed performance by referring again to the
pressure distribution tests on a model propeller.1 In these
AERODYNAM1C DATA OFAEROFO1L SECT1ON C.
Directmeasurement of forces on an aerofoil ofaspectratio 6 with square ends
o Calculated from the pressure distribution over the median section ofthe aerofoil
of aspect ratio 6
Calculated from the pressure distribution over the section C ofan aerofoilshaped
as an airscrew blade but without twist
1.6,
.
1.4
Curvt s of ( /
12
a 1.0
/ -
08
-
Cu rves c
> 0.6
-
0.4
-
0.2
-
-2 0 2 4 6 8 10 12 14 18 20
Angle erf Attack in Degrees
Fig. 32.{From R. and M. 681.)
tests the pressure distribution over several sections of a propeller

blade was measured while the propeller was running in a wind
tunnel, and the three following sets of tests were made on
corresponding airfoils:
a. Standard force tests on airfoils of aspect ratio 6.
b. Tests of the pressure distribution on the median section
of the above airfoils of aspect ratio 6.
1 A Consideration of Airscrew Theory in the Light of Data Derived from
an Experimental Investigation of the Distribution of Pressure over the
Entire Surface of an Airscrew Blade, and also over Airfoils of Appropriate
Shapes, by A. Fage and R. G. Howard, British R. and M. 681, 1921.
c. Tests of the pressure distribution over a special airfoil

made in the form of one blade of the propeller, but without
twist, the pressure being measured at the same sections as in the
propeller blade.
The results of these three sets of airfoil tests are shown for
the section at three-fourths of the tip radius in Fig. 32, which
has been taken from the report. It will be noticed that the
coefficients of resultant force CR agree quite well for the median
section of the airfoil of aspect ratio 6 and the corresponding
section of the special propeller-blade airfoil but that the resultant
force coefficient for the entire airfoil of aspect ratio 6 is con
siderably lower. It is natural, then, that the calculated thrust
and power of a propeller should be too low when based on airfoil
characteristics for aspect ratio 6.
CHAPTER V
MODIFICATIONS OF THE BLADE-ELEMENT THEORY
Many modifications to the simple blade-element theory

have been suggested in order to make it more complete and to
improve its accuracy. Most of these modified theories attempt
to take into account the blade interference, and, in some of
them, attempts are also made to eliminate the inaccuracy due
to the use of airfoil data from tests on wings having a finite
aspect ratio, such as 6. The first modification to be made
was in the nature of a combination of the simple Drzewiecki
theory with the Froude momentum theory.
The Combined Momentum and Blade-element Theory.
According to the momentum theory, half of the increase in
velocity in the slipstream occurs in front of the propeller. The
forward velocity of the propeller with respect to the air passing
through the propeller disc is therefore greater than the aircraft
velocity V. In the combined theory the inflow is calculated
by means of the momentum theory, and the sum of the inflow
velocity and the aircraft velocity V is used in place of V in
calculating the angles of attack and the forces on the blade
elements.
When an airfoil of aspect ratio 6 (or any other finite aspect
ratio) is tested in an airstream, there is a certain downwash,
and the portion of the downwash which occurs in front of the
airfoil corresponds to the inflow to a propeller blade. Thus,
part of the inflow, and usually the larger part of it, is inherently
accounted for in the airfoil tests, although not correctly, for
the conditions in the propeller do not necessarily correspond
to those for aspect ratio 6. Now since a certain portion of
the inflow is present in the model wing tests as downwash, the
value of one-half the wake velocity does not apply, and the
portion of the wake velocity representing the inflow correction
is, in practice, nothing more than an empirical factor, ranging
in value from one-third to two-thirds, to make the calculated
performance fit the actual. Obviously any value greater than
one-half cannot be theoretically correct.
51
Calculation of Inflow.The inflow is calculated by equating

the thrust to the momentum imparted to the air per unit time.
For an elemental annulus of a propeller the relation for axial
inflow is
dT = mass per unit time X velocity imparted
= 2m-drV(l + xb)pbV
= 2wpV2b{l + xb)rdr,
where x is an empirical factor representing the ratio of the inflow
to the total increase of velocity in the wake bV.
The air in passing through a screw propeller having both
thrust and torque receives not only an increase in axial velocity
but a rotational velocity or twist as well. This rotational
velocity in the wake may be calculated by equating the torque
to the rate of change of angular momentum. For an elemental
annulus the relation is
dQ = mass per unit time X tangential velocity X radius
= 2irrdrV(l + xb)pwr
= 2ttPVw(1 + xb)r2dr,
where w is the tangential velocity in the wake and is in the
direction of the torque.
According to the Prandtl airfoil theory,1 there can be no
rotational inflow in front of the propeller, and this has been
shown by experiment to be substantially true.2 There is, how
ever, an interference flow in the plane of the blades themselves,
the air being carried around a certain amount in the direction
of the torque before it has passed entirely through the propeller.
This rotational interference flow is ordinarily very small and
is usually neglected in the application of the combined theory.
As in the case of the axial inflow, the greater portion of it is
taken into account in the use of airfoil data for aspect ratio 6,
although not, of course, accurately.
In the routine application of the combined theory it is custom
ary not only to neglect rotation but also to calculate only the
average axial inflow for the entire propeller instead of that for
each element. Sometimes the central portion of the propeller
Press, p. 209, 1926.
2 The Measurement of Airflow round an Airscrew, by C. N. H. Lock and
H. Bateman, British R. and M. 955, 1924. An Investigation into the
Nature of the Flow in the Neighborhood of an Airscrew by J. R. Pannell
and R. Jones, British R. and M. 371, 1917.
MODIFICATIONS OF THE BLADE-ELEMENT THEORY 53
is considered ineffective and the effective disc area A' is taken as

0.7 or 0.8 of the total disc area A. In this case the inflow is
determined by the expression
T = A'PV*b (1 + xb).
If the thrust, as is usually the case, is not known, the efficiency
can be estimated and the approximate thrust calculated from the
relation
TV_ 550 X v X HP
V = F or T =
55OHP
Expressions for Thrust, Torque, and Efficiency of a Blade
Element.Considering the combined Froude-Drzewiecki theory
and neglecting rotational interference flow, the velocity diagram

for a blade element is given in Fig. 33. The velocity of advance
of the element relative to the air is taken as
u = V + xbV
and the angle of attack is
' . - -
where
u (l + xb)V
tan <t>' = = (1 + xb) tan <t>.
1-wrn 2wrn
Substituting <t>' for <t> and u for V, and following the same
procedure as for the simple blade-element theory, the expressions
for the thrust and torque on an element become
%pu2CLbdr cos (</>' + 7)
dT =
sin2 <t> cos 7
dg = }jpu2CLbrdr sin (4>' + 7)
sin2 <t>' cos 7
These can be written

dT Yzpu* 77 dr,
dQ Mpu* Qc' dr,
where
77 K' cos + 7), '
K'rain (<*>' + y),
and
CLb
K'
sin2 <' cos 7
The efficiency of an element according to the combined theory
is then
dTV
v = dQ 2irn
tan <t>'
(1 + xb) tan (0' + 7)
The total thrust and torque for the whole propeller are found
in the. same manner as with the simple blade-element theory.
In fact, the entire procedure is the same except that u <t>', and a'
are substituted respectively for V, 4>, and a.
Limitations of the Combined Theory.The combined momentum
and blade-element theory is subject to the same limitations as the
simple blade-element theory regarding the use of airfoil data from
tests of model wings having a finite aspect ratio and with regard to
tip losses and scale effect. Theoretically the only possible point
of improvement over the simple theory is that it may be said that
the empirical inflow factor takes into account, in a rough manner,
the blade interference. It does this very inaccurately, however,
for the factor x, representing the portion of the added velocity
of the slipstream which occurs in front of the propeller, should not
actually be a constant for all propellers but should vary consider
ably even when only one set of airfoil data is used. From the
practical point of view, it seems to be a fact that, in general, anal
yses made by means of the simple theory come closer to the
measured propeller performance than those made with the com
bined or inflow theory.1 This is the case with our example in
the last chapter, where the power and thrust calculated by means
'See "The Design of Screw Propellers for Aircraft," by H. C. Watts,
Appendix to Chap. VII, Longmans, Green and Co., London.
of the simple theory were low. Since any amount of inflow would
reduce the effective angle of attack of the blade elements, the
power and thrust calculated by means of the combined theory
would be even lower.
It seems that the combined Froude-Drzewiecki theory, although
it has been widely used in the past, is gradually being abandoned
for more accurate methods of analysis and will probably soon
pass out of use entirely.
The Blade-element Theory with Multiplane Interference Cor
rections.As stated in Chap. IV and illustrated in Fig. 30, the
elements of the blades of a propeller at any given radius may be
considered as forming an infinite multiplane having negative
stagger. The interference effect on an element caused by the
corresponding elements of the other blades, and also by the posi
tion of the original element itself on previous and following
revolutions, is analogous to the downwash and other interference
effects on one plane of an infinite multiplane system.
A series of wind tunnel tests has been made by R. McKinnon
Wood and associates, in which the changes in lift and drag on an
airfoil, due to the interference of a cascade or multiplane series
of airfoils, were measured.1 From the multiplane interference
data obtained from these tests, factors have been derived with
which propeller elements may be corrected for blade interference.
The experiments show that the interference depends on the lift
of the airfoils but is practically independent of the drag and
therefore of the airfoil section. The interference also depends
on the spacing of the airfoils and the angle <t> between the airflow
and the line joining the cascade (Fig. 30). The results of the
experiments are given as a correction SCL to the lift coefficient
and another correction e to the angle of attack. The angle is
in the nature of a downwash angle and is defined by the relation
CV (CD - SCD\ = /CA
+ tan e,
cv \CL-5CL)a \CL),at
CV and CL' being values with interference.

1 Multiplane Interference Applied to Airscrew Theory, by R. McK.
Wood, F. B. Bradfield, and M. Barker, British R. and M. 639, 1919. Also
Preliminary Investigation of Multiplane Interference Applied to Propeller
Theory, by R. McK. Wood and H. Glauert, British R. and M. 620, 1918.
Also, Experimental Investigation of a Wing Profile in a Lattice Arrange
ment, by K. Christiani, L.F.F., vol. 2, No. 4.
Since e was found in the experiments to be practically inde

pendent of the drag, it has been assumed that the interference
corrections can be expressed as functions of CL, <, and S, where
which is the circumference at any radius divided by the total

blade width at that radius and represents, in a way, the solidity.
The interference corrections SCL and e are plotted in terms of
CL, 4>, and S, in Figs. 35 and 36.
Fig. 34.
Calculation of the Forces on a Blade Element.The forces

on a blade element are calculated in the same manner as with
the simple blade-element theory, excepting that the airfoil
coefficients are corrected for blade (or multiplane) interference.
Referring to Fig. 34, the lift component on an element is now
dU = HpVr2CL'bdr,
where
Cl = Cl SCL,
and the resultant air force on the element is
dR = y2PVr*CL'bdr
cos (y' e)
The thrust component is then
dT = dR cos (<t> + y')
_ }4pVR2Ci'bdr cos (</> + 7')
COS (7' e)
&nd since the corrected forces still refer to the original velocity
and angle of attack,
V
Vr =
sin <t>
and
%pV*CL'bdr cos (0 + 7')
dT =
sin2 4> cos (7' e)
Let
CL'b
K =
sin2 <t> cos (7' t)
and
Tc = K cos O + y').
Then
dT = y2PV*Tcdr,
and the total thrust for the propeller (of B blades) is
T = y2PV*Bf*Tcdr.
'i he component of the resultant force on the element contributing
to the torque (Fig. 34) is
dF = dR sin(</> + 7'),
the torque on the element is
dQ = rdR sin(< + 7'),
and if
Qc = Kr sin(0 + 7O,
dQ = Y2PV*Qcdr.
Then the torque for the whole propeller is
Q = y2PV*BfQBQcdr.
The efficiency of an element with interference corrections is

dTV
v dQ2wn
= dR cos + f,)7
dR sin (</> + y')2irrn
_ tan <^>
~ tan (<t> + 7')'
T/ie Limitations of the Blade-element Theory with Multiplane
Interference Corrections.This modification is subject to the
same limitations as the simple blade-element theory in regard
to the unsuitability of airfoil data obtained with aspect ratio 6,
and there is also some doubt as to the aspect ratio effect in the
multiplane tests, in which airfoils of aspect ratio 6 were mounted

with one end against a large plate. Check tests were run on a
series of airfoils of aspect ratio 12 without a plate, however,
and although the lift and drag coefficients were different, the
correction factors 5CL and e agreed quite well with those found
with the plate.
The theory with interference corrections necessitates the same
apparently justifiable assumption as the simple theory regarding
the independence of the neighboring blade elements and has the
same limitations concerning tip losses and scale and compressi
bility effects.
In practice, it has been found that the blade interference
is apparently well represented by the multiplane corrections
but that with airfoil characteristics obtained from tests in any
one particular wind tunnel on model wings of aspect ratio 6,
the calculations have a consistent error when compared with
model propeller tests. This fact led to the idea of calculating
the airfoil characteristics backward from model propeller tests,
so that when these calculated airfoil characteristics were used
in the blade-element analysis along with multiplane interference
corrections, the propeller characteristics could be computed
more accurately. This was done, using data from tests on a
series of 13 Navy-type model propellers as a basis. 1 The method
used would require too much space to describe here, but the
details may be found in three Technical Notes published by the
N.A.C.A.2 The propeller-section characteristics as calculated
from the model propeller test data are given in Fig. 37.
Analyses by means of the blade-element theory with multi
plane interference corrections, when made using the calculated
airfoil characteristics of Fig. 37, agree with the measured per
formance of the 13 model propellers throughout the range of
the tests within two or three per cent, or within approximately
the limits of experimental error of the model propeller tests.
Since the family of 13 propellers covers the ordinary range of
'Tests on Thirteen Navy Type Model Propellers, by W. F. Durand,
N.A.C.A.T.. 237, 1926.
2 Propeller DesignPractical Application of the Blade-element Theory,
by Fred E. Weick, N.A.C.A. T.N. 235, 1926. Propeller DesignExtension
of Test Data on a Family of Model Propellers by Means of the Modified
Blade-element Theory, by Fred E. Weick, N.A.C.A. T.N. 236, 1926. Navy
Propeller Section Characteristics as Used in Propeller Design, by Fred E.
Weick, N.A.C.A.T.N. 244, 1926.
pitch ratios, blade widths, and blade thicknesses likely to be

found in practice, apparently the blade interference is sufficiently
well represented by the multiplane interference corrections for
ordinary design purposes. In fact, this method with the cal
culated~propeller-section characteristics is used by the writer
Fig. 35.-Lift correction SCl for multiplane interference.

in preference to all other forms of the blade-element theory
because in his experience it has proved to be the most accurate.
Example of Analysis by Means of the Blade-element Theory
with Multiplane Interference Corrections, Using Calculated
Airfoil Characteristics.For purposes of comparison, we shall
calculate the performance of the same propeller which was
(j, <(> 4,
Fig. 36.-Downwash correction for multiplane interference.
analyzed by means of the simple blade-element theory in Chap.

IV. The dimensions of the propeller are given again in the first
part of Table II. The propeller is advancing with a velocity
Angle o-f Attack

Fig. 37.Propeller airfoil characteristics calculated from model propeller tests.
Standard propeller sections based on R.A.F.-6.
of 40 m.p.h. and turning at the rate of 1,800 r.p.m. We shall
run through the analysis, as before, for the section at 75 per cent
of the tip radius. For this section, as before,
V
d> -- arc tan
= 15.5,
and the angle of attack is

a = fi - <t>
= 1.1.
From Fig. 37, for a flat-faced section of thickness ratio 0.107
at an angle of attack of 1.1, CL = 0.490. (For sections having
a convex face, CL is corrected in accordance with the relation
given in Fig. 23.)
Now, in order to find the multiplane interference corrections
it is necessary to calculate the value of S:
2-kt
S
Bb
2 X ir X 1.125
2 X .198
17.84.
From Fig. 35, for CL = 0.490 and <f = 15.5, SCL = 0.050
for S = 15 and SCL = 0.030 for 8 = 20. Interpolating, for
S = 17.84, SCL = 0.035. (This interpolation can be made
quite accurately, if desired, by finding SCL for three or four
values of S, plotting a curve of SCL against S, and finding bCL
for the desired value of S from the curve.)
Then
Cl = Cl SCl
= 0.490 - 0.035
= 0.455.
From Fig. 36, with proper interpolation, the downwash
correction e is found to be 0.5.
Then
a' = a e
= 0.6,
and from Fig. 37, the value of D/L for an angle of attack of 0.6
is 0.0650 for this section.
Since by definition,
it follows that
y' = arc tan (0.0650 + tan 0.5)
= 4.2.
Then
R _ Crib
sin2 4> cos (7' e)
= 0.455 X 0.198
0.26722 X 0.998
= 1.264,
and
Te = K cos (<*> + 7')
= 1.264 X cos 19.7
= 1.190,
Table II.Computations fob Propeller Analysis with Multiplane

Interference Corrections and Calculated Airfoil Data
D = 3.0 ft. Forward velocity = 40 m.p.h. = 58.65 ft. /sec.
p = 2.1 ft. Rotational velocity = 1,800 r.p.m. = 30 r.p.s.
r/R 0 15 0 30 0 45 0 60 0 75 0.90
r(ft.) 0 225 0 450 0 675 0 900 1 125 1.350
&(ft.) 0 225 0 236 0 250 0 236 0 198 0.135
hu/b 0 190 0 200 0 167 0 133 0 107 0.090
hL/b 0 180 0 058 0 007 000 000 000
0 (deg.) 56 1 36 6 26.4 20 4 16 6 13.9
2irrn 42 3 84 7 127.1 169.6 212.0 254.0
tan <t> = V/2hrrn. . 1 389 0 693 0 461 0 346 0 277 0.231
<t> (deg.) 54 2 34 7 24 7 19 1 15 5 13.0
a = 0 - <f (deg.) . 1 9 1 9 1 7 1 3 1 1 0.9
Cl , 0 175 0 535 0 680 0 600 0 490 0.410
S = 2irr/Bb 3 140 5 99 8 49 11 99 17 84 31.42
SCL.. 0 030 0 070 0 075 0 055 0 035 0.020
CjJ = Cjj bCjj. 0 145 0 465 0 605 0 545 0 455 0.390
(deg.) 0 9 1 7 1 4 0 9 0 5 0.2
a' = a e (deg.) . 1 0 0 2 0 3 0 4 0 6 0.7
(D/L)* 0 0730 0 0792 0 7060 0 0652 0 0650 0.0662
tan e 0 0157 0 0297 0 0244 0 0157 0 0087 0.0035
tan 7' = (D/L)* 4 0 0887 0 1089 0 0950 0 0809 0 0737 0.0697
7' (deg.) 5 1 6 2 5 4 4 6 4 2 4.0
<(> + 7' (deg.) 59 3 40 9 30 1 23 7 19 7 17.0
sin <t> 0 8111 0 5693 0 4179 0 3272 0 2672 0.2250
sin2 <t> cos (7' e) 0 0496 0 339 0 868 1 203 1 264 1.041
COS (0 + 7') 0 5105 0 7558 0 8652 0 9157 0 9415 0.9563
Tc = Kcob (0 +y').. 0 0253 0 256 0.751 1 101 1 190 0.996
sin (0 + 7') 0 8598 0 6547 0.5015 0 4020 0 3371 0.2924
Qc = Kr sin (0 + 7') 0 0096 0 0998 0.294 0 436 0 479 0.411
and
Qc = Kr sin (<t> + 7')
= 1.264 X 1.125 X sin 19.7
= 0.479.
The computations of Tc and Qc for the six elements of the
propeller are given in Table II.
The procedure of plotting Tc and Qc against the radius, and
integrating for the total thrust and torque, is exactly the same
as with the simple blade-element theory and gives the following
values for thrust, power, and efficiency:

Thrust, pounds 7.84 7.77
Efficiency 0.776 0.771
For this analysis it will be noted that the calculated values are
within one per cent of the model test results, but the fact that
they are closer than two or three per cent is merely fortuitous.
Modern Wing Theory Applied to the Blade-element Analysis ;
the Induction or Vortex Theory of the Propeller.The Lanches-
ter-Prandtl wing theory has been applied to propeller analysis by
Betz and Prandtl1 but not in a practical form suitable for the
designing of propellers. Glauert, with the aid of a simplifying
assumption, was the first to combine the wing theory with the
blade-element analysis in a form similar to that of the older
blade-element theories familiar to propeller designers,2 and
his method is followed in all essentials here. The method of
procedure is changed, however, to facilitate its use.
It will be recalled that, according to the wing theory, a finite
wing giving lift is accompanied by a change of direction or down-
1 Schraubenpropeller mit geringstem Energieverlust (The Screw Propeller
Having the Least Energy Loss) , by A. Betz, with an addendum by L. Prandtl.
Math.-phys. class paper, Gottingen, 1919. Also published in " Vier Abhand-
lungen zur Hydrodynamik und Aerodynamik," pp. 68-92, Gottingen, 1927.
Applications of Modern Hydrodynamics to Aeronautics, by L. Prandtl,
N.A.C.A.r./2. 116, pp. 53-59, 1925.
Press, 1926. An Aerodynamic Theory of the Airscrew, by H. Glauert,
British R. and M. 786, 1922. Notes on the Vortex Theory of Airscrews,
by H. Glauert, British R. and M. 869, 1922.
wash of the airflow. The amount of downwash depends on the

lift of the wing and the span, and it becomes zero for a wing
having infinite span. In the case of infinite span, the flow past
the wing depends only on the shape or profile of the airfoil
section.
Now the flow about a finite wing may be divided into two
parts: that due to the profile and that due to the downwash.
The part due to the profile, which is the local flow at the airfoil,
is the same as that for the infinite span wing. This profile or
two-dimensional flow is influenced by the downwash only in
that the angle of attack is affected to the extent of half the down-
wash angle. Thus the downwash, or induced downflow, repre
sents an interference flow within which the two-dimensional
profile flow takes place.
In the application of the airfoil theory to propeller blade-
element analysis, the two-dimensional or profile characteristics
are used for the airfoil elements, and the interference flow is
calculated from the momentum generated, after the manner
of Froude and Rankine. Thus, since only the profile lift and
profile drag are used, the questionable aspect ratio effect is
eliminated. Also, since the interference flow is determined
by the number of blades, their spacing, and the forces on each,
the blade interference is accounted for.
In this form of the blade-element theory, as in the others,
it is assumed that there is no radial flow through the propeller.
The interference flow is distinguished from the inflow and
outflow as follows:
The inflow is the flow immediately in front of the propeller.
The outflow is the flow immediately in back of the propeller.
The interference flow is the average flow through the propeller
(described by Glauert as the velocity field of the system of trailing
vortices which acts as an interference on the blade elements).
The wake is the final flow in the slipstream far behind the
propeller.
Calculation of the Interference Flow.The interference flow
at any one point in the propeller plane is periodic, varying
as each blade passes. The exact calculation of this periodic
flow is difficult and complicated, so that Glauert replaces the
actual flow by its average value. This is the simplification
previously referred to and is the same as assuming that, for
the calculation of the interference flow, the propeller has an
infinite number of blades and therefore constant thrust and

torque throughout each annulus.
Considering only the axial component of the airflow through
the propeller, the interference flow, inflow, and outflow velocities
all have the same value, which, according to the momentum
theory, is half the wake velocity. Calling the axial interference
velocity aV and neglecting the rotation in the outflow, the
momentum theory yields the following value for the thrust
on an elemental ring or annulus:
= 2irrdrV(l + a)p X 2aV
= 47rp72(l + a)ardr.
The rotational velocity or twist given to the air flowing
through the propeller may be calculated by equating the torque
to the rate of change of angular momentum. The rotational
interference velocity is equal to half the rotational velocity
in the wake, although there can be no rotational inflow.1 The
flow may be visualized by considering the propeller disc as having
a certain finite thickness. There is no rotation in front of
the disc and full rotation immediately back of it. The rotational
interference velocity, or the average value of the rotational
velocity passing through the disc, is half the rotational outflow,
and the rotation in the outflow is the same as that in the wake.
(The rotational velocity can be imparted suddenly because it
does not involve a large change in diameter of the slipstream.
The increase of axial velocity does involve a considerable change
in diameter, which, since the static pressure outside the slipstream
is constant all over it, requires a correspondingly large distance
in which the change of diameter takes place.) If w is the
tangential interference velocity, the torque for an annulus may
be written
= 2wrdrV(l + a)p X 2w X r,
and calling the rotational interference factor a', where
w = a' X 2irrn,
1 The fact that there can be no rotation before the air reaches the propeller
is clearly shown by Glauert in his book "Aerofoil and Airscrew Theory."
He states in effect that the circulation around the blade and the trailing
vortices, which have equal strength but opposite rotation, nullify each other
in front of the propeller blades.
the expression for torque becomes

dQ = 4*rp7(l + a)(2irrn)a'r2dr.
Thus if the total thrust and torque for an elemental annulus
are known, the axial and rotational interference factors can be
calculated by means of the above equations.
The Aerodynamic Forces on a Blade Element.The forces on
a blade element are calculated in much the same manner as with
the other types of the blade-element theory, excepting that
airfoil characteristics for two-dimensional flow are used (the
dRt'-K tea-
subscript zero denoting two-dimensional flow or infinite aspect

ratio characteristics), and the angle of attack is based on the
interference flow. Referring to Fig. 38, the lift on an element is
dL0 = WrJCMr,
and the resultant air force on the element is
JJ3 HPVr<?CLbdr
dR =
cos 70
The thrust component is then
dT = dR cos (<t>0 + to)
}4pVro2CLb dr cos (<t>0 + 70)
cos 70
and since
_ 7(1 + a)
ro sin <*> '
y2PV\l + a)2 CLbdr cos (0o + 70)
dT =
sin2 <t>o cos 70
Let
R = CL&(1 + a)2
sin2 </>o cos ya
and
Tc = K cos (4>0 + To).
Then
dT = Hp72rcdr,
and the total thrust for the propeller is
The component of the resultant force on the element contributing

to the torque (Fig. 38) is
dF = dR sin (0o + y0).
The torque on the element is
dQ = rdE sin (#0 + To),
and if
Qc = Kr sin (<t>0 + y0),
then
dQ = }ipV2Qcdr,
and the torque for the whole propeller is
Q = y2PV*BfQRQcdr.
The efficiency of an element is now

dTV
v dQ2*n
_ dR cos (0o + yo)V
dR sin (<t>0 + 7o)2xrn
_ tan 0
tan (</>o + To)'
and since from Fig. 38,
tan <b = : tan 0o,
1 +a
the expression for efficiency may be written
_ (1 a') tan <t>o
' ~ (1 + a) tan (0o + To)
This shows that the efficiency is reduced by both the rotational
and axial interference factors and also by the profile drag of the
airfoil section.
Procedure in Computing the Performance of a Propeller.

Glauert suggests calculating the characteristics of a propeller
by first assuming a number of angles of attack for each of several
blade elements for which r, b, B, and /3 are known. CL and CDo
are then found from two-dimensional airfoil data for each angle
of attackj>nd-ealculations~are made for a, a', the rate of advance,
and-the thrust and torque per blade element. Curves of thrust
and torque are then plotted against rate of advance, and from
these curves the thrust and torque for each element at any one
desired rate of advance are found, plotted against radius to form
grading curves, and integrated to give the thrust and torque for
the whole propeller. This is an indirect method of solution which
is very cumbersome for analyzing or designing a propeller for a
given rate of advance.
A form of procedure suggested by Miller and Harpoothian,1 in
which the interference angle (0, Fig. 38) is calculated and the
succeeding steps taken in the same manner as with other forms of
the blade-element theory, has been found more satisfactory by
the writer.
In order to obtain a solution for 0, we may equate the two
expressions for dT, one of which was obtained from the equation
for momentum and the other from the forces on the blade ele
ments (the elements of all the blades at a given radius). After
simplifying, we get
a _ CLBb cos (</>o + 70)
1 + a 8wr sin2 4>o cos 70
It is convenient here to introduce the quantity
2xr
S = Bb'
which is the ratio of the circumference of an annulus to the totai

blade width. This is the same quantity as is used in the multi
plane interference corrections. Then
a CL cos (0o + 70)
1 +a 4S sin2 <t>0 cos 70
By equating the two expressions for dQ, we get in like manner
a' _ CL tan 4> sin (<fto + 70)
1 + a, 4<S sin2 <t>o cos 70
1 Applications of the Aerodynamic Theory of the Propeller, by William
H. Miller and Edward Harpoothian, Curtiss Aeroplane and Motor Com
pany, Rept. 3190, 1928.
and from the above two equations it follows that

a' = a tan 0 tan (0o + 70).
From Fig. 38 it is apparent that
tan 0 = ^7-^1 tan <t>o>

(1 + a)
from which
tan <t>
a' = 1 - (1 + a)
Now equating the two expressions for a', we get, after simpli
fication,
a _ cot 0o (tan 0o tan 0)
1 + a 1 + tan 0 tan (0o + 70)'
and, in turn, equating the two expressions for ^ x - we get
cos (0U + 70) + tan 0 sin (0o + 70)

CL 4 sin 0o cos 0o(tan 0o tan 0) cos 70'
which simplifies to
8 _ 1 tan 70 tan (0o 0)
CL 4 sin 0o tan (0o 0) '
and since <t>0 <t> = 6, this may be written,
S _ 1 tan 70 tan 6
<TL ~ 4 sin (0 + 0) tan 6
The interference angle 0 is in this equation an implicit function
of <S, C/,, 0, and 70, which are, except for the almost negligible
factor 70, the same quantities which govern the multiplane
interference corrections.
It is interesting to note that the interference angle 6 is the
same factor in the flow past a propeller element that half the
downwash angle e/2 is in the flow past an element of a wing.
Since d is an implicit function, it can not be solved directly
from the above equation. It can be solved quite easily, however,
by means of the chart in Fig. 39. In finding 6, since 6 affects
the angle of attack a0 and therefore CL and 70, the correct values
of CL and 70 are obtained at the same time. The use of the
chart is shown clearly in the example in the next section.
In order to solve the equations for the thrust and torque

on a blade element, it is necessary to know only one more factor
(1 + a). We already have the equation
a cot 0o(tan <t>0 tan <t>)
1 + a 1 + tan <t> tan (0o + To)'
y v 5
e(whenyo=0)
Fig. 39.
and solving for 1 + a we get
j , a _ tan <t>o[l + tan <t> tan (</>0 + To)]
tan <I>[1 + tan <t>0 tan (<0 + To)]
ZL
cemg eth noitulso fo sith uoiatque si lheati 'usoiedt st osjb

dhesqdraocctJ jCb sauera fo n i-iahc 'Sxj) '(Qf hej, tcejje fo hei
a\5tuv 70 si tou abliaceitou no het %ivvjo puv so 70 sha uebe d.ettiiuo
Of
suoipyyymi'j fo 9yf uotjanpuI j,o X3}X0j^ fLumu, fo J9jpdorP .
hcLi/Vi eth uoijacqdda J iiofare Xaohet erlledoad ûemepe-adbl
analysis, the aspect ratio question is settled in a very satisfactory

manner. The airfoil characteristics are still obtained from wind
tunnel tests of model wings, however, and the calculated propeller
performance is entirely dependent on the accuracy of the model
wing tests, especially with regard to the angle of attack for zero
lift and the profile drag. It also depends on the accuracy of the
wing theory, which is based on many assumptions and is not
entirely unquestionable, particularly where the angle of attack is
of importance.
The blade interference is disposed of very neatly by the
calculation of the interference flow, based upon the assumption
that the thrust and torque are constant with time throughout
any one annulus. It is known, however, that the air flow
through a propeller is highly periodic in nature; and the assump
tion that, for the purpose of calculating the interference flow,
it may be considered constant may lead to comparatively large
errors for highly loaded propellers. In fact, from several
comparisons between calculated and test performances made
in England1 it seems possible that the calculated interference
flow may be noticeably in error, for in every case the calculated
thrust is too high when the blades are working at high angles
of attack and slightly low at very low angles of attack. Of
course, this may be due to slight errors in the airfoil tests or the
model propeller tests, either of which seems quite likely, for just
the opposite has been found in some check calculations made
on a family of Navy-type model propellers by the propeller
research section of the N.A.C.A. In these checks the calculated
thrusts were too high at the low angles of attack of the sections
and the agreement was better for the higher-pitch propellers
(in the British tests it was better for the low). The calculated
efficiency for the Navy-type propellers was in every case at least
5 per cent too high.
In general, however, the agreement between calculated and
measured performance is fair, being very much better than
1 Experiments with a Family of Airscrews, Part III. Analysis of the
Family of Airscrews by Means of the Vortex Theory and Measurements
of Total Head, by C. N. H. Lock and H. Bateman, British R. and M. 892,
1923. The Accuracy of the Vortex Theory of Airscrews in the Light of
Recent Experimental Work, and Its Application to Airscrew Design, by
H. Glauert and C. N. H. Lock, British R. and M. 1040, 1926. An Analy
sis of the Pressure Distribution on a Model Airscrew by Means of the Vortex
Theory, by A. Fage, British R. and M. 940, 1924.
with either the simple blade-element theory or the combined

momentum and blade-element (inflow) theory. It can be used to
bring out small relative differences in performance, such as
those due to slight changes in pitch distribution along the
radius, for which the simple and inflow theories are entirely
inadequate. Apparently, however, it is not superior to the
multiplane interference method in this respect, and it is con
siderably less accurate for calculating the performance of a
propeller than the multiplane method with computed airfoil
characteristics. On the other hand, it is of course possible that
with airfoil characteristics for two-dimensional flow computed
from model propeller test results, the accuracy of the induction
or vortex theory might be equal to or better than that of the
multiplane theory.
Example of Propeller Analysis.For our example of analysis
by means of the induction or vortex theory we shall again compute
the performance of the propeller which was used with the other
forms of the blade-element theory, and, as before, we shall
run through the analysis in detail for the section at 75 per cent
of the tip radius. With a velocity of advance of 40 m.p.h. and a
rotational speed of 1,800 r.p.m., we have, as before,
V
0 = arc tan
2irrn
= 15.5,
and
2irr
S =
Bb
17.84.
The gross angle of attack
a0 + e = fi - <t>
= 16.6 - 15.5
= 1.1.
We shall now find 6 from Fig. 39, and also get CL and 70.
The solution is simple if the steps given here are followed, and
after a little practice it can be made in a moment or two. The
procedure is as follows:
Assuming a value for 6 of, say, 1.0,
a0 approx. = 1.1 - 1.0 = 0.1.
Then for ao = 0.1 and h/b = 0.107, CL = 0.470 (from Fig. 24),
and S/Cl = 17.84/0.470 = 38.
From Fig. 39, for S/CL = 38 and <t> = 15.5, 0 (when y0 =

0) = 1.3. (The value of 0 when 70 = 0 is found at the bottom
of the chart.) Now with our first approximation of 6 = 1.3,
the above steps are repeated, giving for the second approximation,
a0 = -0.2.
CL = 0.438,
1 = 40-8;
0, (70 = 0) = 1.2,
and repeating again for the third approximation,
a0 = -0.1,
CL = 0.445,
Si = 4-1'
0, (70 = o) = 1.2.
This is the same as our second approximation, and in general
it may be said that the second approximation usually gives
the final result within the limits of accuracy of our computations,
i.e., one-tenth of 1. We have, however, still to consider 70,
which affects 0 very slightly. From Fig. 25, for a0 = 0.1
and hu/b = 0.107, 70 = 1.9. Referring again to Fig. 39,
the value of 6 when 70 = 0, in our case 1.2, is projected vertically
upward to the sloping lines of 70, and then from the value of 70,
in our case 1.9, a horizontal line is projected to the right-hand
vertical scale for the final value of 6. In our case 0 is still
1.2, and in fact for nearly all practical cases the effect of 70
is negligible. Our final values of 0, a0, CL, and 70 are therefore
6 = 1.2,
a0 = -0.1,
CL = 0.445,
70 = 1.9.
Then, from Fig. 40, for <t> = 15.5 and 6 = 1.2, 1 + a = 1.076.
Next,
K = CLb(l + a)2
sin2 #0 cos 70
= 0.445 X 0.198 X (1.076)2
sin2 16.7 cos 1.9
= 1.234,
Tc = K cos (4>0 + To)

= 1.234 cos 18.6
= 1.170,
and
Qc = Kr sin (<t>0 + y0)
= 1.234 X 1.125 sin 18.6
= 0.443.
The computations for Tc and Qc for the six elements of the
propeller are given in Table III.
Table III.Computations for Propeller Analysis by Means of the
Induction or Vortex Theory
D = 3.0 ft. Forward velocity 40 m.p.h. = 58.65 ft. /sec.
p = 2.1 ft. Rotational velocity = 1,800 r.p.m. = 30 r.p.s.
r/R. 0 15 0.30 0.45 0.60 0.75 0.90

r (ft.) 0 225 0.450 0.675 0.900 1.125 1.350
b (ft.) 0 225 0.236 0.250 0.236 0.198 0.135
hv/b 0 190 0.200 0.167 0.133 0.107 0.090
hL/b 0 180 0.058 0.007 000 000 000
0 (deg.) 56 1 36.6 26.4 20.4 16.6 13.9
42 3 84.7 127.1 169.6 212.0 254.0
tan <t> = V/2rrn 1 389 0.693 0.461 0.346 0.277 0.231
<t> (deg.) 54 2 34.7 24.7 19.1 15.5 13.0
S = 2wr/Bb 3 14 5.99 8.49 11.99 17.84 31.42
(ce0 + 6) = 0 - </> (deg.) 1 9 1.9 1.7 1.3 1.1 0.9
6, from chart (deg.) .... 0 9 1.9 2.2 1.7 1.2 0.7
o (deg.) 1 0 0 -0.5 -0.4 -0.1 0.2
Cl 0 172 0.465 0.580 0.520 0.445 0.396
To (deg.) 2 0 2.2 1.9 2.0 1.9 2.0
1 +a 1 01 1.043 1.080 1.083 1.076 1.053
0 = 0 + 6 (deg.) 55 1 36.6 26.9 20.8 16.7 13.7
0 8201 0.5962 0.4524 0.3551 0.2874 0.2368
rr CLb (1 + a)2
K . 9 0 0581 0.336 0.827 1.144 1.234 1.060
sin' <t>o cos 70
<t>0 + To (deg.) 57 1 38.8 28.8 22.8 18.6 15.7
COS (<<>o + To) 0 5432 0.7793 0.8763 0.9219 0.9478 0.9627
Tc = K COS (<t>o + To) . . 0 0315 0.262 0.725 1.055 1.170 1.020
0 8396 0.6266 0.4817 0.3875 0.3190 0.2706
Qc = Kr sin (<t>u + to) . . 0 0110 0.0947 0.269 0.400 0.443 0.387
The procedure of plotting Tc and Qc against the radius and

integrating for total thrust and torque is the same as with the
other blade-element analyses and gives the following values for
power, thrust, and efficiency:

Thrust, pounds 7.71 7.77
0.829 0.771
The calculated thrust, it will be noted, is within 1 per cent

of the test value, but the calculated power absorbed is 7.5 per
cent low, and the calculated efficiency is therefore 7.5 per cent
too high.
Resume, and Comparison of the Various Theories in Their
Representation of the Flow through a Propeller.All of the
forms of the blade-element theory discussed neglect the con-
High Speed Climb Zero Advance

Fig. 41.Airflow through a propeller.
traction of the slipstream and the radial component of the

flow through the propeller. The flow of air through a conven
tional propeller has been accurately measured1 for three different
rates of advance corresponding to flying at high speed, climbing,
and standing still on the ground. The streamlines in a longi
tudinal plane are shown in Fig. 41. In all three cases there is a
1 The Measurement of Airflow round an Airscrew, by C. N. H. Lock and
H. Bateman, British R. and M. 955, 1924.
slight radial flow near the propeller axis due to the position of the
small guard body, but since this is in the region of the hub it is
unimportant. There is very little radial flow even through the
tip sections for the high-speed condition and also very little
contraction of the slipstream. For the high-speed condition,
therefore, which is usually near the point of maximum efficiency
of the propeller, it may be concluded that the neglect of radial
flow and contraction of the slipstream will not cause appreciable
errors in the computation of propeller performance. This is
not quite so true for the climbing condition and apparently
not true at all for the condition of zero advance, for which
calculations with any of the theories may be expected to be
seriously in error.
Fig. 42.-Typical calculated and measured thrust grading curves.
All forms of the blade-element theory also neglect tip losses.

Figure 42 shows a typical comparison between the calculated
thrust grading curve for a propeller and that measured from the
difference in total head in front of and behind the propeller. It
will be noticed that the calculated thrust is too high near the tip
of the blade. This is true for torque, also.
Figure 42 brings out another point: that of the drag, or nega
tive thrust at the center of the propeller. Although none
of the forms of the blade-element theory shows a negative
thrust at the center of the propeller, actually, due to the hub
and a body, which, although it may be small, must be near the
center of the propeller, the thrust grading curve usually passes
through zero near the hub, and there is a region of negative
thrust at the center. The amount of this negative thrust is

very difficult to determine accurately. Sometimes, especially in
Great Britain, an amount equal to the drag of a disc the
thickness and diameter of the hub boss is subtracted from
total thrust. This is, however, a gross overestimate of the
actual negative thrust, because the presence of even a very small
guard body in back of the propeller tends to reduce the drag of
the boss in two ways : It forms, in a sense, a continuation of the
boss, relieving the region of low pressure behind the boss which
is the cause of most of the drag; and, also, the body tends to
slow the air ahead of it in which the hub is located, still further
reducing the drag. It is probably more accurate to consider
the hub a part of the body (as it actually becomes when a spinner
or revolving nose fairing is used) and to neglect the small negative
thrust in the propeller computations than to make any such
correction as that mentioned above. It is of interest in this
connection to note that for our examples the thrust is already low,
and deducting a hub correction would make it still lower. The
torque, of course, never becomes less than zero excepting at
extremely high rates of advance entirely beyond the working
range of the propeller.
The independence of the neighboring blade elements is another
assumption common to all forms of the blade-element theory, and
this seems well substantiated in theory and experiment excepting
at the hub and tip, where radial flow and tip losses complicate
the flow.
All forms of the blade-element theory are also limited by
the accuracy of the airfoil tests upon which the forces are based
(excepting when calculated airfoil characteristics are used)
and by the effect of scale and the compressibility of the air.
As has been stated, the flow through a propeller is highly
periodic in nature, and it is interesting to compare the various
theories in regard to the manner in which this periodic flow is
accounted for. The variation in velocity at one point behind a
propeller has been measured by Drzewiecki by means of a
strobometer. The strobometer is a stroboscopic pitot tube
connected to a manometer by means of a valve which allows it to
operate for only one particular instant during each revolution of
the propeller. In his experiments Drzewiecki used a propeller
having only one blade, the centrifugal force being balanced by
a suitable counterweight, and the instantaneous velocity of the
air was measured at 20 points on a circle behind the blade. The

variation in velocity found during one revolution is shown in
Fig. 43.
The free air velocity with respect to the propeller was 16
meters per second, and it will be noticed that as the blade passed
there was a rapid increase in velocity to over 21 meters per
second, then a sudden decrease to about 18 meters per second,
25, , , . ,
90 180 270 360

Position of Blade with respect to Pitot
Fig. 43.
followed by a more gradual decrease to the original free air

velocity. The interference inflow is evidently in this case zero,
for the blade is always attacking air having the free air velocity.
It is also evident, however, that if the propeller had been
equipped with two or more blades, the increase in velocity due
to one blade would not have died out in 180 deg. or less, and the
remaining increase over the free air velocity would have repre
sented an interference velocity. There would also have been an
interference velocity remaining after 360 deg. if Drzewiecki's
single blade had been wider and consequently more powerful.
The accurate computation of this highly complicated periodic

flow is not at the present time possible. The periodic flow is
largely present, however, in the model wing tests. Thus in the
simple blade-element theory the periodic nature of the flow is
inherently accounted for empirically in the airfoil characteristics
used, but incorrectly because of the aspect ratio factor, and also
because interference is neglected.
The multiplane interference method takes account of both the
periodic nature of the flow and the blade interference empirically,
and does it more satisfactorily than any other system, but has the
disadvantage of the questionable aspect ratio effect.
In the induction or vortex theory the periodic nature of the
flow is neglected in so far as the interference calculations are
concerned, but that portion of the local periodic velocity which
is present in two-dimensional flow is represented empirically in
the airfoil characteristics for infinite aspect ratio. In this
theory, therefore, the periodic nature of the flow is not so well
represented as in the multiplane interference theory, but the
aspect ratio factor is satisfactorily cared for. 1
1 For an interesting comparison of various theories of the propeller, the
reader is referred to a paper entitled Review of Airscrew Theories by Maj.
A. R. Low, published in the Journal of the Royal Aeronautical Society for
February, 1923. The discussion by McKinnon Wood, Fage, Bairstow,
Glauert, Lanchester, and others is of particular interest.
CHAPTER VI
AERODYNAMIC TESTS ON PROPELLERS
In order to determine the performance characteristics of

propellers, and also in order to provide a check for the theoretical
computations, aerodynamic tests are made either in flight or
in aerodynamic laboratories called wind tunnels. To determine
the complete characteristics it is necessary to measure thrust,
torque, revolutions, velocity of advance, and the temperature
and pressure of the air. The density is calculated from the
temperature and pressure. The results are usually given in terms
of non-dimensional coefficients involving these quantities, in
order to simplify the comparison of propellers with one another.
Flight Tests.It is, of course, desirable that propellers be
tested under the conditions in which they are used, and the only
tests which fulfil these conditions perfectly are actual flight
tests. In flight tests, however, the thrust and torque have not
been successfully measured directly1 but are obtained indirectly
from the performance of the airplane and dynamometer tests
on the engine.
The lift and drag characteristics of the airplane are first
found from glide tests at various angles of attack, and then the
airplane is flown full throttle at various air speeds and rates
of climb. Measurements are taken of the flight-path angle,
attitude of the airplane, air speed, revolutions, temperature,
and atmospheric pressure. The thrust credited to the propeller
is obtained partially from the drag of the airplane at the same
speed and angle of attack, as found from the glide tests, and
partially from the portion of the thrust power which produced
the climb. This latter may be positive, zero, or negative,
depending on the flight path.
1 Hub dynamometers, or special propeller hubs, which measure the torque

or thrust (or both) by means of springs or hydraulic pressure, have been
built and tested but have either given inaccurate results or have been imprac
tical to operate. A good, light hub dynamometer would be extremely
useful for checking propeller characteristics and engine power in flight.
82
AERODYNAMIC TESTS ON PROPELLERS 83
The full-throttle power of the engine is usually measured for

various revolution speeds on a dynamometer before and after
the flight tests, and the engine is credited with the same power
at the same revolutions in flight, corrections being applied for
temperature and pressure. This procedure introduces an error,
for the power of an engine varies considerably with changes in
installation, cooling, carbureter altitude control setting, etc.
It is also fairly well established that, in general, aircraft engines
deliver somewhat less power in the air than under the more ideal
conditions of a dynamometer test.
Since neither the thrust nor the power of the propeller can be
determined accurately, flight tests are useful mainly as a rough
check but will not bring out small differences in propeller charac
teristics such as those due to changes in pitch distribution, plan
form, and blade section.
The flight tests require a great deal of time and labor and
the use of several special instruments to measure the angle of
flight path, angle of attack, air velocity, etc.
Rough information regarding the power absorbed at one rate of
advance of a propeller can also be obtained from ordinary airplane
performance tests, in which the high speed is obtained over a
measured course in full-throttle level flight. In this type of test
the power is usually obtained from a type test on another
engine of the same design, providing an even greater chance
for error.
Wind Tunnel Tests on Model Propellers.A wind tunnel is an
aerodynamic laboratory in which the object to be tested is placed
in a moving stream of air. Thus while the relative motion of the
body and the air is the same as in steady flight, actually the body
does not advance. This makes it possible to support it on
balances or scales with which the air forces on the body
can be accurately measured. When the body is a propeller,
suitable apparatus must be provided to rotate it and measure
its torque and thrust. This apparatus is more complicated
than that required for ordinary airfoil and drag tests, and prob
ably for that reason nearly all tests on model propellers have
been confined to a very few tunnels, notably the Stanford
University tunnel in this country and one of the tunnels of the
National Physical Laboratory in England.
The Stanford tunnel is of the open-jet, or Eiffel, type, having
a large experiment chamber with a stream of air passing through
the center. Figure 44 is a diagram of the tunnel showing the

cones used for the entrance and exit of the air. The model
propeller being tested is mounted on a long spindle which amounts
to little more than a shaft for over a diameter back of the pro
peller. The body interference may therefore be said to be nil
for all practical purposes. Means are incorporated in the driving
mechanism to measure the torque, thrust, and revolutions.
The British N.P.L. type of tunnel is similar to that at Stanford
University except that the airstream is enclosed throughout the
entire tunnel, including the portion passing through the experi
ment chamber, and the cross-section of the airstream is square
instead of circular. This closed type has the disadvantage that
it is necessary to apply a correction for wall interference, whereas
none is required for the open-jet- type for propeller tests.1
Fig. 44.-Diagram of wind tunnel at Stanford University.
In the British tests the model propellers are driven by an

electric motor enclosed in a small guard body, the nose of which is
shown in Fig. 41. The whole installation is supported by wires in
such a manner that the thrust and torque can be measured.
The model propellers tested in wind tunnels usually have a
diameter of 3 ft., although sometimes models as small as 2 ft. in
diameter are used. The tests without a body or fuselage are
useful for checking the theoretical computation . of propeller
performance and for finding the effect on propeller characteristics
of variation of the number of blades, blade width, and other
geometrical proportions. The model test results cannot be
applied directly to full-scale aircraft propellers, however, on
account of the effect of scale, compressibility, and distortion of
the propeller blades in operation. Even when the model
propellers are run at the same tip speed as the full-scale propellers,
the results are not the same, due apparently to a difference in
size and therefore Reynolds number. The body interference can
1 See The Advantages of an Open-jet Type Wind Tunnel for Airscrew
Tests, by H. Glauert and C. N. H. Lock, British R. and M. 1033, 1926.
be taken care of in the model propeller tests by providing a model

body, and this has been done in a few cases.
Full-scale Wind Tunnel Propeller TestsIn 1927, the
N.A.C.A. completed a wind tunnel with a 20-ft. airstream
capable of velocities up to 110 m.p.h. in which full-size pro
pellers are tested under conditions that are, for all practical
purposes, those of actual flight. The whole airplane, or the
fuselage only if desired, is mounted in the experiment chamber
and the propeller driven by an actual aircraft engine. This
tunnel is of the open-jet, double-return passage type, as shown
diagrammatically in Fig. 45. A view of the experiment chamber
Return Passage
Experiment Chamber
'Fan
Exit Cone ------ = =Z
Guide Vanes--
Fig. 45.-Diagram of N.A.C.A. 20-ft. Propeller Research Tunnel.

with a stub-winged cabin monoplane mounted on the balance is
given in Fig. 46. The power of the engine is measured by means
of a special dynamometer designed to fit inside the fuselage.
The thrust, drag, lift, rolling moment, and pitching moment are
measured on scales on the floor below. A special electric starter
is used to start the aircraft engine and is then swung clear of the
airstream during the tests. The deflection and twist of the
propeller blades in operation are also measured, by means of a
device in which a telescope is sighted on the leading and trailing
edges.
Incidentally, the propeller fan shown in the exit cone in
Fig. 46 afforded an interesting design problem. It is 28 ft. in
diameter, weighs 7,000 lb., and is driven at 375 r.p.m. by a
2,000-hp. Diesel power plant. It was desired to make the
blades adjustable in pitch because the exact requirements of
the tunnel could not be known in advance. Wooden blades

could not be satisfactorily held in adjustable blade roots without
a great deal of development work, and duralumin forgings
could not be made in so large a size, and so finally the blades
were made of heat-treated cast-aluminum alloy. They have
been operating satisfactorily since May, 1927.
Coefficients and Methods of Plotting Results of Propeller
Tests.As with airfoils, non-dimensional coefficients are usually
used for comparing the characteristics of propellers. In this
Fig. 46.-View of the 20-ft. Propeller Research Tunnel showing the experiment
chamber, balances, exit cone, and propeller fan.
way the results of tests on a propeller can be applied directly
to any geometrically similar propeller regardless of size or exact
revolutions or velocity of advance, due consideration being given
to scale and compressibility effects.
One of the most fundamental characteristic magnitudes
affecting propeller performance is the advance per revolution V/n.
This is made non-dimensional by giving it in terms of the diam
eter, in which case it is often called J and becomes
_P
J =
nD'
where V is the velocity of advance, n the rate of rotation, and
D the diameter, all in consistent units (such as foot-pound-
second or kilogram-meter-second).
Plotting propeller characteristics on a basis of V/nD is similar

to plotting airfoil characteristics on a basis of the angle of attack.
The V/nD is closely related to the angle of advance 4> and therefore
the angle of attack of each section of a given propeller. The
relation is
V V
tan <t> = ~ = =; X a constant.
2rrn nD
According to the blade-element theory, the efficiency of an
element is given by the expression,
_ tan <t>
v ~ tan (*+ y)'
The efficiency therefore depends on the angle 4>, and, conse
quently, V/nD.
The V/nD is sometimes termed the slip function, because it is
intimately connected with the slip of a given propeller. The
nominal slip S is defined by the relation,
i _ s = v/nD = X.
p/D np
This slip is merely nominal, because it is based on the pitch
of the chord of the airfoil section, which has no particular aero
dynamic significance when different airfoil sections are used.
Its use is sometimes convenient, however.
Coefficients of propeller thrust and torque can be derived from
the expressions for the thrust and torque on an element in the
simple blade-element theory. The thrust for an element was
found to be
dT = MPV*CLbdrC0\{+ + 7)-
sin2 <t> cos 7
Since b and r vary directly with the diameter, and since <t> is a
function of V/nD, the expression for thrust may be written
dT cc ovwf^y
and for the whole propeller,

T = TcPV2D2,
where the non-dimensional thrust coefficient Tc is a function
of V/nD.
A similar but slightly different form of the coefficient is
sometimes used (particularly in Germany) in which the dynamic
pressure q = }4f>V2 is used in place of pV2, and the disc area

A = 7rD2/4 in place of simply D2. The expression then becomes
T = thrust coefficient X qA.
The torque on an element of a propeller blade, according to
the simple blade-element theory, is
dQ = &VH!J**&M,
sin2 <j) cos 7
from which it follows that
dQ oc Pv>D>f(j~y
and for the whole propeller,

Q = QcpV*D\
P.c. Tc
or 11 p7c
0.12 1.2 1
1
1
0.10 1.0
0.08 0.8 l\
\\
\\l\
0.06 0.6 \\
\\
\\
0.04 0.4 \\
0.02 .0.2
V
0 0
0 0.2 0. 1- w 0. 5 0. 5 1.0
'iTU
Fig. 47.
Both Tc and Qc are usually plotted against V/nD as shown

for a typical propeller in Fig. 47. The coefficients are applicable
to propellers of any size as long as dimensional homogeneity is
preserved as to both the propeller and the airflow. In other
words, neglecting scale and compressibility effects, the coeffi
cients apply to all geometrically similar propellers operating

at the same value of V/nD.
The coefficients can be obtained by means of the principle of
dimensional homogeneity as well as from the blade-element
theory. The coefficients Tc and Qc have more or less passed
out of use and given way to more convenient ones with scales
covering a smaller range of numbers so that more nearly uniform
accuracy is maintained over all portions of the curves. Other
non-dimensional coefficients can be obtained from Tc and Qc
by merely multiplying or dividing them by some power of the
quantity V/nD. The power can be so chosen as to cancel out
. 0.008 0.08
0.006 0.06
0.004 0.04
0.002 0.02
0 0.2 0.4- 0.6 0.8 1.0
Fig.; 48.
of the expression the terms for either the velocity or the diameter.
The following coefficients are in terms of the revolution speed
instead of the velocity :
CT = Tc X (^j
= T
pn2D*'
Q
pn2D*
The coefficients are shown plotted against V/nD in Fig. 48.
They are extensively used in England, where they are called
kT and kQ.
In practice it is usually desirable to use coefficients of power

and efficiency instead of thrust and torque. If P represents
the power absorbed by the propeller,
2irnQ
and
Qcz--
n
Substituting P/n for Q in the last equation for the torque coeffi
cient, we get as a power coefficient,
P
CP =
pn3D6
where CV = 2wCQ, or, to relate the American power coefficient
to the English torque coefficient, CP = 2irkQ.
0.4 0.02
0.2 0.01
Fiq. 49.
The efficiency may then be found from the thrust and power
coefficients as follows:
TV
v = P
= CTpn2D*V
CP A nD
The curves of power coefficient and efficiency, when plotted

against V/nD (Fig. 49), completely determine the aerodynamic
characteristics of a propeller.
Speed-power Coefficient.All of the above coefficients are

useful for comparing the thrust, torque, and power of different
propellers. They are not particularly helpful, however, in
selecting or designing propellers to fit certain specific requirements
for driving some form of aircraft. Furthermore, V/nD is not a
fair basis upon which to compare the efficiencies of various
propellers which might fit the same aircraft at the same values
of V and n but have different diameters and therefore different
values of V/nD.
Frequently tests are made on series of propellers which are
geometrically similar except that the pitch is varied in uniform
steps. A propeller may be selected from such a series to fit any
combination of conditions (P, n, V, and i?) required for a particular
aircraft. In order to select conveniently the best possible
propeller, it is desirable to have a coefficient involving the above
required factors but not involving the diameter or any dimension
of the propeller. Such a coefficient involving only the known
or required factors (P, n, and V) is also a fair basis on which to
compare the efficiency of different propellers which might be
used to drive the same aircraft.
A coefficient involving P, n, and V without D may be obtained
from CP and V/nD in the following nlanner:
Coefficient = (ZZ^j
CP
nW X P
= ell
Pn2'
This is sometimes called a speed-power coefficient because it
contains the rotational and forward speeds and the power
absorbed. The reciprocal of the coefficient as given above was
used by Drzewiecki. In this form, however, it covers a very wide
range of values, making its use impractical. Admiral Taylor,
in dealing with marine propellers,1 has made use of the square
root of the reciprocal of the above coefficient, but even in that
case it is necessary to use logarithmic or semilogarithmic scales
for plotting the data in order to obtain about the same degree of
accuracy throughout the entire range of the curves. For aero-
1 "The Speed and Power of Ships," by D. W. Taylor, John Wiley & Sons,
Inc., 1916.
5000 -f-30
1000- r-4.0
900 - 4500 4
800-
t-28
700 - 4000
600-
3500 -^26
E-3.5
500-
400 -t 3000 -+
--24
300 -i 2500 -F
HP. t3.0 HP '/5 R P M f-22 RPM^
200 -F 2000
1900
1800
150 -E 1700
1600
1500
1400
100 -2.5 -18
90 1300
80 Scales of Hp'5 1200 -17
70 and R P M z/sfor 1100
60 Solution of 1000- -16
50 pV5
900
V Pn 2 -15
40 800
. Q.638xM PH
_E-2.0 ~ HP'/sxR P M2/s -14
30 700
for Standard
Atmosphere 600 -13
20 oitSea Level
500- -12
400 i-11
-t-1.5 -10
Fio. 53.
To obtain the highest possible speed, the highest possible

propeller efficiency is desired at the high-speed condition of
operation, i.e., value of C. For the point at which the line of
V/nD for maximum possible efficiency crosses C, = 1.38, we find
that V/nD = 0.75 and the pitch is given by the blade angle
22 deg. at 0.7 5R (Fig. 51 or Fig. 52).
The diameter is then
D = 88 X MPH
RPM (V/nD)
= 88 X 125
1,800 X 0.75
= 8.15 ft.
The efficiency of this propeller is 0.83. If with this efficiency
the thrust horsepower is more or less than the amount required
to drive the airplane at 125 m.p.h., the new speed may necessitate
the selection of a slightly different propeller.
For a second example let us assume that a propeller of our
series having a diameter of 9.0 ft. is on hand and that the blade-
angle setting which will make it fit the above airplane is desired.
In this case C, remains the same at 1.38, but the V/nD becomes
V = 88 X 125
nD 1,800 X 9
= 0.68.
Now from either Fig. 51 or Fig. 52, for C = 1.38 and V/nD =
0.68, the blade-angle setting is found to be 17.5 deg. at 0.75.R,
and the efficiency is 0.81. 1
Eiffel Logarithmic Diagrams.Another method for choosing a
propeller, which has had some use, is that of the logarithmic
diagram devised by Eiffel.2 In this type of chart, CP is plotted
against V/nD in a manner similar to that of Fig. 49, except that
logarithmic scales are used, so that actually log CP is plotted
against log V/nD, (Fig. 54).
1 Larger charts having more pitch settings represented than Figs. 51 or 52
are necessary for accurate use. Charts of this nature are given in Figs.
165 to 175.
2 Only a brief description of one type of logarithmic diagram is given here.
For a more complete treatment of the subject see "Les Recherches surles
Helices Aeriennes," by G. Eiffel (French text) 1919, "Nouvelles Recherches
sur la Resistance de l'Air et l'Aviation," by G. Eiffel (French text), 1914,
and Experimental Research on Air Propellers, by William F. Durand,
N.A.C.A.r./E. 14, 1917.
The logarithmic scales enable us to deal directly with the

quantities P, V, n, and D. Consider a point on the curve of
log CP; the abscissa is log V/nD which is equal to log V log
n log D. This can be thought of as three individual segments
placed end to end. In like manner, since CP = P/pn3Ds, the
ordinate can be formed of the three segments, log P, 3 log
n, and 5 log D, placed end to end. Now the two segments
involving n, one ordinate and one abscissa, can be combined to
give an oblique segment the length of which is proportional
Log Cp
to n. In the same manner an oblique segment can be used

instead of the two segments involving D. One can now pass
from the origin to the point on the curve of log CP by the path
OABCD (Fig. 54), and the segments measure respectively
V, n, D and P.
The slope of the n segment will always be 3/1, and that of the
D segment 5/1. For practical use the diagram is usually con
structed with the scale for the abscissas twice that of the ordinates
so that the slopes of the n and D segments become actually 3/2
and 5/2, making possible more accurate plotting. Complete
diagrams are given in Figs. 55 and 56. Scales of V, n, D, and P
0.70
SO 40
60 70 80 90 100 lib
Velocity in Miles per Hour
' Fig. 55.-Logarithmic diagram for propeller 4412 set at 15.0 at 0.75R, with
efficiencies marked on curve of log Cp.
| , i 1 Iii H ) ii^ iiiii | i i'i I |'ih fi'i Wi'i[ liii)liiîi1li'iiil).Uj.l).V.iir.|).^jiU^"

JO 40 50 60 TO 80 90 100 110 120 130 140150160
Velocity in Miles per Hour
Fig. 56.Logarithmic diagram with lines of constant efficiency, for the same
series of propellers as in Figs. 51 and 52.
are added to the diagrams in units of m.p.h., r.p.m., feet, and

horsepower, the scales for n and D being on lines of the proper
slope placed at any convenient position on the diagram. Origins
for the n and D scales are assumed at, say, 1,200 r.p.m. and 10
ft., respectively, and these determine the location of the scales of
V with respect to V/nD and HP with respect to CP. The scales
are usually calculated for standard sea level atmosphere (15C.
and 760 mm. Hg). The efficiencies can be marked on the curves
of log Cp, as has been done in Fig. 55, or separate curves of
thrust-power coefficient can be drawn, in which case the distance
between the curves of input and thrust power is a measure of
the efficiency.
In using the logarithmic diagram, the segments of V, n, D, and
P are drawn in from the origin to a point on the curve of log CP,
each segment being parallel to its respective scale. The lengths
of the V and HP segments are measured from the origin, and the
n and D segments from the circles (assumed origins) on the scales.
In drawing the various segments, the direction of each should
be in the sense that it is from the origin or assumed origin (up if
above, down if below). The logarithmic diagram can be used
for the graphical solution of the same problems as the charts
of V/nD vs. C,. If any three of the quantities P, n, V, and D
are known or assumed, the fourth and also the pitch and efficiency
can be determined. The method of procedure can be best shown
by means of examples. Assume that we have a propeller geo
metrically similar to the one whose logarithmic diagram is given
in Fig. 55 and that we wish to find the diameter required to
absorb 80 hp. at 1,500 r.p.m. and 70 m.p.h. The velocity
segment OA is the horizontal scale from the origin to 70 m.p.h.
Obviously the segments can be drawn in any order as long as
they are added vectorially, and so in order to keep within the
diagram, the horsepower segment AB is put in next, parallel
to the HP scale. Its length is equal to the distance from the
origin to 80 hp. on the scale. Then the revolution segment
BC is drawn parallel to the r.p.m. scale, its length being equal
to the distance between the origin (1,200 r.p.m.) and 1,500
r.p.m. Finally the diameter segment is drawn parallel to the
diameter scale from C to the propeller curve, and its length, when
measured from the origin (10 ft.) on the diameter scale, gives a
diameter of 8.35 ft. The efficiency is 0.74.
In Fig. 56 the same series of propellers is given as in Figs. 51 and

52, and curves of constant efficiency are drawn in the same
manner as in Fig. 52, as well as the curves for maximum possible
and peak efficiency. We shall take as an example the same
problem as before; i.e., we shall select the propeller giving the
highest speed for an airplane with an engine developing 200
hp. at 1,800 r.p.m. and design for an airspeed of 125 m.p.h. at
sea level. The velocity, power, and revolution segments OA,
AB, and BC are drawn in as before, and the diameter segment
is drawn from C to where it intersects the line for maximum
possible efficiency, at D. The length CD, measured on the
diameter scale, gives the diameter at 8.15 ft., the same as before.
Also, the blade angle is again 22 deg. at 0.75R, and the efficiency
is 0.83.
If we wish to find the blade angle of a 9-ft. diameter propeller
to fulfil the above conditions, we go through the same process
as before, but instead of drawing the diameter segment to
intersect another line, it is terminated at the point E, corre
sponding to the 9-ft. diameter. For the point E we then see that
the blade angle is 17.5 deg. at 0.75/2 and the efficiency is 0.81.
Measuring the Flow of Air through Propellers.The momen
tum theory shpws that in order to produce a thrust, a propeller
must give the air through which it passes a backward motion.
Similarly, due to the torque, the air is given a rotational motion.
The result is a column of air, or slipstream, which follows a
twisting or spiral path. The diameter of this column, according
to Bernoulli's equation, will vary with the velocity, so that the
greater the velocity the smaller is the cross-sectional area and
consequently the diameter of the column. Figure 41 of Chap. V
shows the lines of average flow in a longitudinal section of the
slipstream of a propeller for three different rates of advance.
It will be noted that the diameter of the slipstream changes
more at the lower rates of advance where the thrust, and, there
fore, the additional velocity, is greater.
The most complete measurements of the flow through a
propeller have been made in England with a four-bladed 16-in.
model propeller.1 Using the data from these tests, Fig. 57
was drawn, showing the increase in velocity in a longitudinal
section of the slipstream for a rate of advance corresponding
1 An Investigation into the Nature of the Flow in the Neighborhood of an
Airscrew, by J. R. Pannell and R. Jone3, British R. and M. 371, 1917.
to climbing flight. The axial or fore-and-aft component of the

added velocity in the slipstream is indicated by arrows whose
ends are connected by dotted lines, in planes 2 and 4 in. ahead
of the propeller plane and 1, 3, and 8 in. behind it. It will be
seen that there is an appreciable increase of velocity, or inflow,
4 in. or one-fourth of a diameter ahead of the propeller. The
contraction of the slipstream is evident from the radii at which
the velocity becomes zero. Just behind the propeller the flow
at the outer edge of the slipstream reverses slightly, and the edge
...2. -1 -A
Propeller
Fig. 57.Longitudinal section of slipstream. Increase in axial velocity caused
by propeller is shown by arrows. Four-blade propeller, p/D =0.9,V/nD =0.63.
of the slipstream in this region is quite sharply defined. At the
center the drag of the hub causes the velocity to become negative
also.
In Fig. 58 the tangential or rotational component of the slip
stream velocity is shown for each of the above positions in front
of and in back of the propeller plane. While some rotational
velocity was measured in front of the propeller, it was very small
and in opposite directions at various radii, averaging about
zero. It seems, therefore, that the assumption of no rotation
in the inflow, as made in the vortex or induction theory of the
propeller, is true, at any rate for practical purposes. The
rotational component of the velocity is seen to be greater near

the center, and it also has the tendency to pass through zero and
attain small negative values at the edge of the slipstream.
4"Ahead of 2"Ahead of ("Behind 3"Behind 8"8ehind

Propeller Propeller Propeller Propeller Propeller
Fig. 58.Tangential velocity of slipstream. Arrows indicate magnitude to
same scale as in Fig. 57.
Figure 59 shows the change of both the axial and rotational

components in passing through the propeller. Only the ele
mental annulus at three-fourths of the tip radius is considered,
f '75l I I I I I I I I
8
01 2J45678
Distance Back of Propeller in Diameters
Tig. 59.Variation of axial and tangential components of slipstream velocity
with distance from propeller.
the values being obtained from the same data as Figs. 57 and
58. It is seen that for this propeller and rate of advance, the
increase of axial velocity begins about three-fourths of a diameter
forward of the propeller and reaches its maximum value at about
half of a diameter back of it, after which it gradually decreases,

although it still has over half its maximum value at eight diam
eters back. The diameter of the slipstream is smallest at the
point of maximum velocity. In a perfect frictionless fluid it
would continue indefinitely with this diameter and velocity, but
in air, which has considerable viscosity, the energy of the slip
stream is gradually dissipated and the velocity becomes less and
the diameter greater with increased distance. At the propeller
plane, it will be noticed, the increase of axial velocity has reached
just about half of its final maximum value, which is in agreement
with the momentum and induction theories.
The rotational or tangential component is practically zero right
up to the front of the propeller. It attains its full maximum value
immediately on passing through the propeller and then gradually
falls off, but less rapidly than the axial component. It appears
reasonable also that the rotational interference velocity at the
plane of rotation (halfway through the propeller) is half the final
maximum value.
It should be remembered that all of these velocities and com
ponents of velocities represent average values. The velocity is
not, of course, constant with time but is highly periodic in nature
(see Fig. 43, Chap. V).
Also in these tests there was no body present. With a body
of average size behind the propeller, the slipstream does not
usually neck down but is spread so that its diameter is at no point
smaller than that of the propeller.1
1 See report on VE-7 flight test, Investigation of Slipstream Velocity, by
J. W. Crowley, Jr., N.A.C.A.r.B. 194, 1924.
CHAPTER VII
THE EFFECT OF BLADE SHAPE ON PROPELLER

CHARACTERISTICS
In this chapter the results of wind tunnel tests are used to show
how variations of the blade shape effect the aerodynamic per
formance. Only propellers operating without body interference
and at comparatively low tip speeds are considered, the effects
of tip speed and body interference being taken up in separate
chapters. Most of the effects could also be shown by means of
an advanced form of the blade-element theory, but not that of
tip shape or of pure size.
The Effect of Pitch on Propeller Characteristics.In Fig. 49
are given curves of efficiency and power coefficient as a function
of effective pitch V/nD. When V/nD is zero, the efficiency is
also zero, which results from the definition of efficiency for a
moving vehicle; for when the forward velocity is zero, no useful
work is being done even though the thrust may be quite large.
The efficiency for a given propeller increases with increase of V/nD
up to a certain maximum point and then falls off with increasing
rapidity as the V/nD is increased further, until it finally reaches
zero again at the value of V/nD for zero thrust. The thrust and
efficiency, of course, become zero at the same value of V/nD.
When the propeller is operating at a V/nD between zero and
that for zero thrust, it is said to be in the propeller state. When
it is between the values of V/nD for zero thrust and zero torque
or power, it is in the brake state, for the thrust is negative but
power is required to produce it. At values of V/nD above that
for zero torque, the propeller is in the windmill state, for it then
receives energy from the air and helps to turn the engine. This
condition may have serious consequences in the case of an
airplane in a steep dive where the speed of advance becomes
very great and the engine is rotated more rapidly than is safe.
Curves of efficiency vs. V/nD for a series of model propellers
varying in pitch are given in Fig. 60. 1
N.A.C.A.r.fl. 237, 1926.
105
The models were of the standard Navy wood form having

uniform pitch; i.e., the geometrical pitch based on the flat face
of the sections was the same at all radii. The pitch-diameter
ratios varied from 0.5 to 1.1. The maximum efficiency attained
by the different propellers increases from 0.69 for p/D = 0.5 to
0.83 for p/D = 1.1. If the pitches were carried higher, the
maximum efficiencies would continue to increase at a lessening
'0 I 1 1 r~-1 1 1 1 1 1 1 1 1 1 1
0 0.1 0.4 0.6 0.8 1.0 11 \A

iTU"
Fig. 60.Efficiency curves for a series of model wood propellers. AR = 6,
2 blades, WR = 0.066, TR = 0.107.
rate up to a p/D of about 1.6 or 1.7, after which they would

decrease again. This is to be expected from the blade-element
equation for efficiency,
_ tan 0
" " tan (<*> + y)'
for V/nD is a function of <t> (Chap. IV, Fig. 29).
Most airplanes have propellers with pitch ratios of 0.6 to 0.8,
and higher efficiencies could be obtained if the pitch could be
increased. Since the propeller must fit particular values of P, n,
and V, however, it is usually possible to increase the V/nD, and
thereby the p/D, only by gearing the propeller to operate at a
lower revolution speed.
In order to compare the efficiencies of propellers having
various pitch ratios on the basis of the performance required
THE EFFECT OF BLADE SHAPE 107
on a given airplane, i.e., at the same required values of P, n, and

V, the efficiencies can be plotted against the speed-power coeffi
cient C as has been done for a series of metal propellers in Fig.
51 (Chap. VI).
It is interesting to compare the efficiencies actually obtained
with the ideal or limiting efficiency as determined by the momen
tum theory. In Chap. II it was shown that
_ 2p X -j- X T-
In order to express this in terms of power instead of thrust, we

may substitute Pij/V for T, which gives
ttD2 V3
= T ( 7V
2C\nD)
0.2 0.4 0.6 OA 1.0 1.2 U
Via. 61.Comparison of actual and ideal efficiencies.
The ideal efficiencies as given by this expression are plotted in

Fig. 61 along with the actual efficiencies for the two extreme
propellers of the series of Navy wood models. The maximum
efficiency of the high-pitch propeller is 88 per cent of the ideal,
while that of the low-pitch propeller is only 78 per cent of the
ideal. At the lower values of V/nD, however, the low-pitch
propeller approaches the ideal more closely than the high-pitch
propeller.
The variation of the thrust and power coefficients of this

series of propellers is shown in Figs. 62 and 63, respectively.
0.16 ,
0 02 0.4 0.6 0.8 1.0 l.Z U

V
nD
Fig. 62.-Thrust coefficient curves.
Both the thrust and power coefficients increase with the pitch
ratio throughout the ordinary range.
0liiii ' iiiiiiiii
0 0.2 OA 0.6 0.8 1.0 1.2 1.*

nD
Fig. 63.-Power coefficient curves.
Pitch Distribution.The effect of even wide variations of the

distribution of the pitch along the radius is comparatively
small, the average pitch for the whole propeller being the impor
tant thing. The above Navy wood models were designed with
uniform or constant pitch at all sections, for this has been found
to give very nearly the best efficiency for all cases without body
interference. A slightly greater efficiency can be obtained,
however, by having the pitch increase with the radius so that
it is somewhat higher than the average at the tip and lower at the
hub.
This point is well illustrated by some British tests1 on a family
of model propellers. These propellers had detachable blades
which could be turned in the hubs. The blades of one propeller
0.10
-n\
0.60
//
A
\
0.60 w
0.40 II
J*
0 0.4 0.8 v 1.2 1.6 2.0

nD
Fig. 64.Effect of pitch distribution on efficiency. WR - 0.072. TR =0.103.
which were designed to have a uniform pitch of 0.7 of the diameter

were turned up 16 deg. so that the average pitch was about 1.5
times the diameter, and comparative tests were made with a
uniform-pitch propeller of p/D = 1.5. The efficiency curves
of the two propellers are given in Fig. 64, and it is seen that the
efficiency of the reset propeller, which had a much greater pitch
at the tip than at the hub, was slightly higher throughout the
entire working range than that of the uniform-pitch propeller.
The effect of increasing the pitch with the radius is smaller
on propellers of lower pitch. From an extensive series of tests
on model metal propellers with various pitch distributions,
1 Experiments with a Family of Airscrews, by A. Fage, C.N.H. Lock,
R. G. Howard, and H. Bateman, British R. and M. 829. 192?.
the curve in Fig. 65 is given as approximately the best pitch

distribution for thin-bladed propellers of all pitches, without
body interference. Fortunately the best conditions require a
proportionately greater increase of pitch toward the tip for
high-pitch than for low-pitch propellers, making it possible to
design an adjustable-pitch propeller with a practically uniform
pitch at a low pitch-diameter ratio, which can be set at any
desired average pitch with the assurance that it will have very
nearly the best pitch distribution at any setting. Thus one
blade form does very well for propellers of any pitch within
the ordinary limits of practice.
i 1 1 i 1 i
0.8 I I I I I I
0 0.2 0.4 0.6 0.8 1.0
Radius
Fig. 65.-Approximately best pitch distribution for thin metal propellers
without body interference.
With the propeller operating near a body such as a fuselage

or nacelle, the above pitch distributions are no longer the best,
for the body reduces the velocity of the air through the propeller,
and this reduction is usually much more pronounced near the
center than near the tips. Designers, therefore, sometimes
reduce the pitch still more near the hub, and occasionally they
reduce it or "wash it out" near the tip also, in the hope of
reducing the tip losses.
Blade Width. The blade width is sometimes expressed as the
mean width of the working portion (the outer two-thirds or
three-fourths) of the blade in terms of the diameter. Also,
sometimes the aspect ratio, which is nominally considered the
tip radius divided by the maximum blade width, is used as a
measure of the blade width. Neither of these is satisfactory if
propellers outside of a single family or series are to be compared,
for neither is a criterion of the mean effective blade width.
The blade is more effective near the tip because of its greater
speed there relative to the air. The radius which best represents
the average of the thrust and torque grading curves is usually
at from 70 to 75 per cent of the tip radius. We shall therefore
take as a measure of the mean effective blade width the blade
width at 0.75R. This is expressed non-dimensionally in a ratio
called the width ratio WR, which is denned by the relation
WR = ^gLR.
In making changes in blade width alone, it is assumed that the

blade sections remain of the same form, i.e., that the thickness
is varied in proportion to the width.
0.05
0.04
0.03
0.02
0.01
0
WR 0 0.02 0.04 0.06 0.08 0.10 0.12 0.14
TOR 0 0.04 0.08 0.12 0.16 0.20 0.24 0.28
Fig. 66.Variation of power and efficiency with blade width.
The variation of power and efficiency with blade width is

shown in Fig. 66 for a series of propellers1 all having a pitch-
diameter ratio of 0.5 and operating at the value of V/nD for
maximum efficiency (V/nD = 0.45). The power coefficient
increases with the blade width, but the increase in power is less
than the increase in blade width. The efficiency drops off with
increasing blade width. Both of these effects are due largely
to the increased slipstream velocity with the wider blades and
incidentally, could be found by means of the advanced forms of
the blade-element theory, but not with the simple theory.
'Experiments with a Eamily of Airscrews, by A. Fage, C.N.H. Lock,
R. G. Howard, and H. Bateman, British R. and M. 829, 1922.
Since high-pitch propellers act on more air per revolution, the

effects are less pronounced than with propellers of low pitch.
Number of Blades.According to the blade-element theory,
the torque and thrust for all of the blade elements in an elemen
tary annulus depend on a solidity factor S, where
2wr
S =
Bb
This factor contains the product of the blade width and the num
ber of blades, representing the total blade width for the annulus.
As long as the product Bb is the same, the individual values of
B and b are immaterial, and therefore according to the theory,
a two-bladed propeller should have the same thrust, torque, and
efficiency as a propeller having four blades, each being half as
wide. The British model tests previously referred to1 show that
this is essentially true and that the curves of Fig. 66 also hold for
propellers having any number of blades if the total width ratio
is used, where
TWR =
The model tests deviate slightly in the matter of efficiency,

however, the efficiency of the multibladed models (three, four,
and six blades) being on the average from one to two per cent
lower than that of the two-bladed propeller having the same
total blade width. It is possible that this is due to scale effect,
for the tip speeds obtained in the tests were low and the maximum
blade widths of some of the propellers were less than an inch and
a half.
Blade Thickness.Applying the same reasoning as in the case
of the blade width, the most satisfactory measure of the thick
ness of a propeller blade is the thickness ratio of the section at
0.75R, or
In Dr. Durand's tests on a family of Navy-type model wood

propellers,2 four propellers of p/D = 0.7 formed a series varying
1 Experiments with a Family of Airscrews, A. Fage, C. N. H. Lock, R. G.
Howard, and H. Bateman, British R. and M. 829, 1922.
N.A.C.A.r.R. 237, 1926.
only in thicknees ratio. The power coefficient and efficiency

curves for these four propellers are given in Fig. 67. It will be
noticed that the V/nD for zero efficiency (or zero thrust) is
greater for the thicker propellers. This is to be expected from
the fact that the thicker airfoils reach zero lift at lower angles of
attack (Chap. III). Also, just as the thicker airfoil sections
have a greater lift and drag at the same angle of attack, the
thicker propellers have greater power and thrust coefficients at
0.02 0.2
Fig. 67.Efficiency curves for wood propellers of various thicknesses. WB

0.066.
the same V/nD, and due to the higher drag the efficiency is
slightly less. The range between the thinnest and thickest wood
propellers is not great, and the thickest propeller of Fig. 67 is
only a little over one per cent less efficient than the thinnest.
Aluminum-alloy propeller blades can be made much thinner,
and the gain in efficiency over wood blades is appreciable but not
great at ordinary tip speeds (up to about 900 ft. per sec). In
Fig. 68 a curve of the maximum efficiencies obtainable with a
series of full-scale standard wood propellers is given along with
a similar curve for a series of standard thin-bladed metal pro
pellers, both being plotted against C, so that they are on a fair
basis as regards similar operating conditions. Both series of

propellers were tested on a Vought VE-7 airplane with a 180-hp.
Wright E-2 engine in the N.A.C.A. 20-ft. Propeller Research
Tunnel.1 The thinner metal propellers are from 4 to 7 per cent,
more efficient than the wooden ones under the same operating
conditions, i.e., for the same values of P, n, and V.
'0 i 1 1 i 1 1 1 1 1 1 i 1 1
Fiq. 68.Maximum possible efficiencies of ordinary wood and metal propellers.
Blade Section.If, as is nearly always the case, the airfoil

sections of propellers are made with a flat lower surface, or face,
the shape of the profile is fairly well limited by the thickness
ratio. With these factors fixed, any reasonably good airfoil
section will give very nearly the same propeller characteristics
as the best possible section.
From the simple blade-element theory we have the relation
_ tan 4>
* ~ tan U + T)'
where <t> is the helix angle of the path of the element and depends
on the effective pitch and y is the angle whose tangent is D/L and
represents the effectiveness of the airfoil. Considering this
relation, it would be expected that any particular change in airfoil
1 Full-scale Tests of Wood Propellers on a VE-7 Airplane in the Propeller
Research Tunnel, by Fred E. Weick, N.A.C.A. T.R. 301, 1928. Full-scale
Wind Tunnel Tests of a Series of Metal Propellers on a VE-7 Airplane,
by Fred E. Weick, N.A.C.A. T..R. 306, 1928.
effectiveness y would have a greater importance at low than at

high pitches.
This is verified by full-scale wind tunnel tests1 on three pairs of
metal propellers which were exactly the same except for the air
foil sections. In one pair the sections throughout the outer
half of the blades all had a thickness ratio of 0.06, in one pair
it was 0.08, and in the third, 0.10. One propeller of each pair
had sections based on the R.A.F-6 and the other, on the Clark-Y.
The efficiency curves for the thinnest pair set at low, medium, and
high pitches are given in Fig. 69. Wind tunnel tests on Clark-Y
i.o ,
RAF-6
Clark-Y
0.8
\N v \
\\ \\ \ \\
S// / \\
0.6 / \
/ A
/ /' A<
/ A / /y
0.4 ff// ' / / \\ \
\|
/ /' ''/'
//
0.2 1 /' //
P MS" " 19\
0.2 0.4 0.6 0.8 1.0 1.2 1.4

nO
Fig. 69.Comparative efficiency of propellers with R.A.F.-6 and Clark-Y type
airfoil sections.
and R.A.F-6 type airfoils of the same thickness have shown

that the Clark-Y has lower drag at low lift coefficients, and a
higher maximum L/D. In Fig. 69, the Clark-Y propellers
have higher efficiencies at the high range of V/nD values near zero
thrust and efficiency. Also, the peak efficiency of the low-pitch
propellers is noticeably higher with the Clark-Y, but there is no
appreciable difference at the high pitch. The differences are
somewhat more pronounced with the thicker propellers.
It is of interest to note that the same envelope or "maximum
efficiency for V/nD" curve is touched by all of the propellers
regardless of airfoil section, indicating that if any pitch setting
may be selected, the advantage of the Clark-Y is more apparent
Comparison of Full-scale Propellers Having R.A.F.-6 and Clark-Y Air
foil Sections, by Hugh B. Freeman, N.A.C.A.T.B. 375, 1930.
than real. It does have an advantage, however, for low-pitch

propellers operating at or beyond their peak efficiency.
Although the Clark-Y has some advantage at the high values
of V/nD, the curves show that with the higher-pitch propellers
at the lower values of V/nD corresponding to the take-off and
even the climbing range, the propellers with the R.A.F-6 type
sections are decidedly superior.
Propeller No. 5
1.Q
0.8
Y
0.6
0.4
0.2
0.2 OA 0.6 0.8 .1.0

V
nO
Fig. 70.-Effect of plan form on efficiency.
Plan Form.The effect of plan form upon propeller perform

ance is not great, although there is some aerodynamic advan
tage in a blade tapering toward and rounded at the tip. The
tapering blade is also advantageous from the point of view of
strength.
In Fig. 70 are plotted the efficiency curves for two Durand
propellers1 which are similar except for the plan form. One had a
blade of constant width with slightly rounded corners at the tip,
1 Experimental Research on Air Propellers V, by W. F. Durand and E.
P. Lesley, N.A.C.A.T.fi. 141, 1922; propellers 5 and 7.
and the other had a tapering blade with rounded tip. With this
extreme difference in plan form the difference in maximum
efficiency was only 3 per cent.
The two plan forms most commonly used with detachable-
blade aluminum-alloy propellers are shown in Fig. 71. They
are known as the wide-tip and narrow-tip plan forms. Tests
on full-scale propellers in the N.A.C.A. Propeller Research
Tunnel have been made at moderate tip speeds with propellers
of both plan forms, the actual blade thickness in each case
being the same so that the tip sections of the wider propeller
had lower values of TR. The efficiency was found to be about
1 per cent greater with the wide-tip plan form.
1.0 r
1 |
^ Wide Tip
| 0.8
9
jC -T3V 0.6 Narr
a
CO
BV | 0.4
J3
CO I
a 0.2
2
0.4 R
0.7 R 0.6 R aeR
Radius
Fig. 71.-Plan forms of wide and narrow tip aluminum alloy propellers.
Sweepback and rake of the blades have no appreciable effect on

the aerodynamic characteristics of a propeller, except that they
affect the twist of the blades while operating, and therefore the
pitch. The blades of wood propellers are often swept back
(see tapered plan form in Fig. 70) in order to obtain smooth
running qualities and to eliminate flutter.
The Effect of Size.There appears to be very little scale
effect with change in pure size from the ordinary 3-ft. wind tunnel
models to the 6- to 12-ft. propellers used on aircraft. There
is one difference, however, between model and full-scale wind
tunnel tests which is noticeable even when they are run at
nearly the same tip speeds. This is illustrated by a typical
comparison of model1 and full-scale2 wind tunnel tests, the
1 For tW model tests see Comparison of Tests on Air Propellers in Flight
with Wind Tunnel Model Tests on Similar Forms, by W. F. Durand and
E. P. Lesley, N.A.C.A.r.B. 220, 1926.
Research Tunnel, by Fred E. Weick, N.A.C.A.r.B. 301, 1928.
efficiency curves of which are given in Fig. 72. Both propellers

were tested in the presence of the same shaped bodies, and under
as nearly as possible the same conditions. The full-size propeller
has a higher V/nD for zero thrust and its maximum efficiency
also occurs at a higher V/nD. In other words, the "dynamic"
or zero thrust pitch of the full-size propeller is larger than that
10 1 1 1 1 1 1 1 1 1 1 1 1 1
Full Scale
Fig. 72.-Effect of size on efficiency.
of the model, and the maximum efficiency is therefore slightly

higher and occurs at a greater V/nD. This fact has also been
indicated in flight tests but their lack of accuracy makes the
results of dubious value. It is apparent, therefore, that if
propellers are to be accurately designed to operate at a particular
position on the efficiency curve, say at maximum efficiency,
accurate full-scale test data must be available.
CHAPTER VIII
THE EFFECT OF TIP SPEED ON PROPELLER

PERFORMANCE
As was shown in Chap. Ill, the aerodynamic force on an

airfoil is a function of V/c, where c is the velocity of sound in air,
or the rate of travel of any compression wave. While this com
pressibility function has been found to be unimportant in con
nection with airplane wings, which even in the fastest cases travel
at velocities far below that of sound, the tip sections of propellers
on modern aircraft engines often approach and in some cases
exceed the velocity of sound.1 In these cases in which the veloc
ity approaches or exceeds that of sound, the effect of V/c is of
vital importance, especially in regard to reduction of efficiency.
Three forms of experimental work have been carried on to
gain information on the effect of tip speed on propeller perform
ance coefficients. These are (1) tests on model airfoils at
velocities up to and beyond that of sound; (2) wind tunnel tests
on small model propellers run at tip speeds up to and beyond the
velocity of sound; and (3) wind tunnel tests on full-scale pro
pellers run throughout the same range of tip speeds.
The model airfoil and propeller tests have thrown a consider
able amount of light on the subject, showing a great variation
of coefficients with V/c, especially at high velocities. They are,
however, limited by low Reynolds number, and in the airfoil
tests by a questionable aspect-ratio effect, so that with our
present limited knowledge the results have no quantitative
value, and the more recent full-scale propeller tests are the only
ones giving information which can be used for practical
computations.
Tests on Model Airfoils at High Speeds.The first model
airfoil tests at velocities approaching that of sound were made in
an open jet of air emerging from a standpipe with a nozzle
1 The velocity of sound in air is independent of pressure but varies approxi
mately as the square root of the absolute temperature. It is therefore
independent of altitude except for a secondary temperature change. It
varies from 1,050 ft. per sec. at 0F. to 1, 160 ft. per sec. at 100F.
119
12J4 in- m diameter, an air stream having any velocity from 550
to 1,000 ft. per sec. being supplied by means of a large centrifugal
compressor.1 Model airfoils having a chord of 3 in. and extend
ing entirely across the airstream were used. The aspect ratio
effect is therefore questionable, especially at high angles of
attack. Six airfoils were tested, all having standard propeller
sections based on the R.A.F.-6 (see Fig. 19) but varying in thick
ness ratio from 0.10 to 0.20.
An extension of the above investigation was made in which the
pressure distribution over the same six airfoil sections was meas
ured on models of 1-in. chord extending entirely across a 2-in.
air jet with velocities up to 1,250 ft. per sec.2
A still further extension was finally made in which force tests
were made on 24 1-in. chord model airfoils in a 2-in. air jet.3
The 24 airfoils included the original series of standard propeller
sections with the addition of three thinner sections, the thinnest
having a thickness ratio of 0.04. There were also five sections
based on the Clark-Y, the ordinates all having been scaled up or
down to give thickness ratios varying in even steps from 0.04
to 0.20, and a few other miscellaneous sections.
The results obtained for the same section in the three different
ways do not agree in value, indicating that the aspect ratio effects
and probably also the effect of Reynolds number at high speeds
are very large. As pointed out by the authors of the reports,
the results cannot be considered as having quantitative value,
at any rate not until more is known regarding the effect of
Reynolds number and regarding corrections for the aspect ratio
condition occurring with a 1-in. chord airfoil extending entirely
through a 2-in. jet of air. The tests no doubt have some qualita
tive value, however, for they all indicated the following effects :
1. The lift coefficient decreases with increase of velocity, very
slightly for thin sections but rapidly for thick ones.
1 Aerodynamic Characteristics of Airfoils at High Speeds, by L. J. Briggs,
G. F. Hull, and H. L. Dryden, N.A.C.A.r.ft. 207, 1925. An earlier investi
gation was made in a wind tunnel capable of air speeds up to about 600 ft.
per sec, which is reported in Wind Tunnel Studies in Aerodynamic Phenom
ena at High Speed, by F. W. Caldwell and E. N. Fales, N.A.C.A.T.R. 83,
1920.
2 Pressure Distribution over Airfoils at High Speeds, by L. J. Briggs
and H. L. Dryden, N.A.C.A.r.K. 255, 1927.
3 Aerodynamic Characteristics of Twenty-four Airfoils at High Speeds,
by L. J. Briggs and H. L. Dryden, N.A.C.A.T.ft. 319, 1929.
THE EFFECT OF TIP SPEED 121
2. The drag coefficient increases with increase of velocity, the

rate rising abruptly at a critical speed well below the speed of
sound. The increase at the higher speeds is very great, especially
for the thicker sections.
3. The center of pressure moves back with an increase of
velocity.
4. The effects of high speed are less for thin than for thick
sections.
5. The angle of zero lift decreases with increase of velocity up
to a critical speed, after which it approaches zerd degrees.
6. The last series of tests indicated that except for thin
sections at high lift coefficients the Clark-Y series are somewhat
better than the R.A.F.-6 series. Also, one section having a
flat undersurface and a circular arc upper, with a thickness ratio
of 0.08, was found to be better than any other section of the same
thickness ratio at the higher speeds.
Tests on Model Propellers at High Tip Speeds.Tests have
been made in a 7-ft. wind tunnel at the Royal Aircraft Establish
ment in England on several 2-ft. model propellers run at tip
speeds ranging from about half to somewhat over the speed of
sound.1 Very complete investigations were made including
not only the usual force measurements of thrust and torque but
also a measurement of the airflow through the propeller by means
of total head pitot tubes and special yawmeters. The airflow
measurements provided a basis for the computation of the thrust
and torque grading curves, and, what is even more interesting,
the airfoil lift and drag coefficients were calculated by means
of the vortex blade-element theory. Thus the effect of high
speeds on the computed lift and drag coefficients for the model
propeller blade sections can be compared directly with the results
of the model airfoil tests made in this country. In general they
show the same tendencies, the effects shown by the computed
propeller section characteristics being as follows:
1. The lift coefficient increases slightly with velocity up to
about 0.7c, after which it falls off rapidly. (The model airfoil
tests showed a decrease throughout the entire range.)
1 The Effects of Tip Speed on Airscrew Performances, by G. P. Douglas
and R. McKinnon Wood, British R. and M. 884, 1923. Wind Tunnel Tests
with High Tip-speed Airscrews, by G. P. Douglas and W. G. A. Perring,
British R. and M. 1086, 1091, 1123, 1124, 1134, and 1198, published in 1927
and 1928.
2. The drag coefficient increases with increase of velocity, the

rate rising abruptly at a critical speed of about 0.6c.
3. The center of pressure moves back with an increase of
velocity.
4. The effects of high speed are less for thin than for thick
sections.
0.08
0.07
.Calculated Curve
Tip Spet ds
0.06 s-l.lt e.
V..l.06c.~
/.0.97c.
/.-0.93c.
./r-OMc.
// r- 0.75c.
0-05 '0.57c.
vV
\\ \\. x \v
0.04[ . \\
\ \\ ^1
0.03
\\
\\\\
O02
0.01
0.1 02 03 _V OA 0.5 OS
nD
Fig. 73.Thrust. Coarse pitch R.A.F.-31o airscrew, Curves of kt measured
at various tip speeds.
The results of the model propeller tests indicated that Reynolds

number as well as V/c apparently had a large effect at high
tip speeds. This was then proved by special model tests1 and
also by full-scale flight tests.2 The same conclusion is indicated
by a comparison of the model propeller test results with the
results of wind tunnel tests on full-scale propellers. All of these
1 Wind Tunnel Tests on High-speed Airscrews, by G. P. Douglas and
W. -G. A. Perring, British R. and M. 1174, 1928.
1 Full-scale Determination of the Effect of High Tip Speeds on the
Performance of an Airscrew, W. G. Jennings, British R. and M. 1173, 1928.
comparisons show that the effect of high speeds on the coefficients

is less pronounced at high values of Reynolds number than at
0.007 r
Tip Sp\eds
0.006 ' .-.1.07c.
-.'...-0.98c
/ .--094c
. r--o.esc
. :..-. 0.75c
0.005 ' 0.57c
0.004 \\
\\
0.001 >\
Calcufa+ed <\
0.00EI
0.1 0.3 0.4 0.5 OA
nD
Fig. 74.Torque. Coarse pitch R.A.F.-31a airscrew. Curves of fcg (measured)
at various tip speeds.
Fig. 75.Efficiency. Coarse pitch R.A.F.-31a airscrew.

low. Unfortunately this complicates matters and makes full-
scale tests necessary for obtaining quantitative results applicable
to actual propeller problems.
In all of the model propeller tests the efficiency fell off as the
tip speed was increased, throughout the entire range of the tests.
The drop in efficiency was small at the lower tip speeds but
became greater and of fairly constant rate at tip speeds above
about 0.75c or 0.80c.
Typical curves showing the thrust, torque, and efficiency

coefficients at the different tip speeds are shown for one of the
model propellers in Figs. 73, 74, and 75. (This propeller had
R.A.F.-31a sections with a thickness ratio of 0.126 over the
entire working portion of the blade.)
The thrust coefficient falls off and the torque coefficient
becomes greater as the tip speed is increased. The dotted lines
represent coefficients calculated by means of the vortex theory

from low-speed model airfoil tests. It is interesting to note
that while the measured thrust coefficients for a tip speed of
0.57c check the calculated values rather well, the measured
torque coefficients for all tip speeds tested are substantially higher
than the calculated.
-2i 1 r-rn mii 1n
Thrust and torque grading curves as obtained for the same

model propeller from the measured air flow through it are given
for the various tip speeds in Figs. 76 and 77. These figures
show very graphically the manner in which the thrust of the
various sections decreases and the torque increases with increase
of speed through the air. The tip sections are the first to be
influenced, the effect working in as the speed goes up. The

values over the inner half of the blade are practically the same
for all tip speeds, and they agree reasonably well with the
calculated values.
Wind Tunnel Tests on Full-scale Propellers at High Tip
Speeds.The full-scale propeller tests were run on thin-bladed
aluminum-alloy propellers in the N.A.C.A. 20-ft. Propeller
Research Tunnel, at tip speeds from about 400 to 1,300 ft. per
sec. (0.35c to 1.15c).1
Ten propellers in all were tested, four being of forms in common
use and having diameters of 9 ft. and 9 ft. 6 in., and six also
having diameters of 9 ft. 6 in. but being especially designed to
afford a comparison of different airfoil sections and thickness
i.o
0.8 9 Ft. 61n. Full Sizt Propeller
0.6
0.4
0.2
05 0.6 0.7 0.B0.9 1.0 1.1
V
E
Fig. 78.-Comparison of tip-speed effect on typical model and full-scale propellers.
ratios. Of the six special propellers, three had standard propeller
sections based on the R.A.F.-6 and the other three had sections
based on the Clark-Y. In each case the three propellers made
a series of different thicknesses, all of the sections in the outer
half of the blade having thickness ratios of 0.06 in one propeller,
0.08 in another, and 0.10 in the third.
The full-scale tests gave results somewhat different from those
of the model tests, although in both cases there was a great
reduction of efficiency at the highest tip speeds. In the full-scale
tests the efficiency was constant throughout the entire usual
range of tip speeds, from the lowest at which the tests could
be run, up to a critical speed of about 0.9c. Above this critical
1 Full-scale Tests on a Thin Metal Propeller at Various Tip Speeds, by
Fred E. Weick, N.A.C.A.T.R. 302, 1928. Full-scale Tests of Metal Pro
pellers at High Tip Speeds, by Donald H. Wood, N.A.C.A.r.B. 378, 1930.
tip speed the efficiency fell off rapidly at a constant rate of

decrease. This is shown in Fig. 78, where a typical curve
showing the maximum or peak efficiency of a full-scale propeller
is plotted as a solid line against tip speed. The line, it will be
noticed, has two straight portions connected by a short curve at
the critical speed. The dashed line on Fig. 78 gives similar
values from tests on a 2-ft. model. The efficiency falls off
throughout the entire range. The critical speed is much lower
and less definite than for the full-scale propeller. Beyond the
critical speed the model efficiency line also becomes fairly
straight, but its slope is not so great as that for the full-scale
propeller.
0.03
8 1200-*,
"2 1100-y
1000^
!22*
7: fl00->
0.02 1= 100T,
g 600
8 500'
o.
0.01 R
0.10 0.20 0.30 040 0.50 0.60

FTd
Fig. 79..Power coefficient curves for a full-scale metal propeller at various
tip speeds.
Figures 79 and 80 show curves of the power coefficients and

efficiencies for the above full-scale metal propeller at various tip
speeds. This propeller had a pitch-diameter ratio of 0.4 and
Clark-Y type sections with a constant thickness ratio of 0.08
throughout the outer half of the blades.
The efficiency curves, it will be noticed, all lie within the
experimental error of a single curve for the tip speeds under the
critical, but above 1,000 ft. per sec. they drop off rapidly. The
power coefficients also vary but slightly up to the critical speed,
particularly at values of V/nD near maximum efficiency, as
shown in Fig. 81. The variation which does occur below the
critical speed can be accounted for by the deflection or twisting
of the blades in operation, which was measured during the
tests. This is a fortunate thing, for it means that the results
of the regular full-scale wind tunnel propeller tests can be applied
directly to all ordinary cases of propeller operation where the tip

speed is below the critical. At higher tip speeds, however, the
power coefficient increases rapidly with increase in tip speed, and
at a higher rate for the lower rate of advance and higher angles of
attack of the blade sections.
0.80, , 1 1 1 1 H
v
TTT)
Fig. 80.Efficiency curves for a full-scale metal propeller at various tip speeds.
Figure 82 shows the effect of section-thickness ratio on both
the critical speed and the rate of decrease of efficiency beyond the
critical speed. The ratio of the maximum efficiency at any tip
speed to that at low tip speeds is plotted against tip speed for
the series of three 9 ft. 6-in. propellers having Clark-Y type
sections and thickness ratios of 0.06, 0.08, and 0.10. The
critical tip speed for the 0.06 thickness ratio is 90 ft. per sec.
higher than that for the 0.10 thickness ratio, and above the
critical speed the maximum efficiency decreases at the rate
of about 11 per cent per 100 ft. per sec. increase for the 0.06
thickness ratio as compared with 13 per cent for the 0.10. These
results are in agreement with the indications of the model airfoil
and model propeller tests which all showed a greater high speed
or compressibility effect on thick than on thin airfoil sections.
"400 600 800 1000 1200 140O

Tip Speed in Ft. per Sec.
Fig. 81.-Variation of power coefficient with tip speed.
Since in the full-scale tests the critical speed varies only from
about 0.85c to 0.95c, and since the velocity of the air passing
just over the upper forward part of the airfoil is somewhat
greater than the velocity of the section with respect to the air
i.i
"5 i.o
Secti on 7/ ickn aiio
.S 0.7
Tip Speed in Ft. per Sec

Fig. 82.Effect of thickness ratio on loss of efficiency at high tip speeds.
in general, it seems quite likely that the critical speed is that

at which the air flowing over the nose reaches the velocity of
sound. This is substantiated to some degree by the fact that the
critical tip speed is lower for the thicker sections which have
greater nose curvature and therefore greater air velocity over

the nose portions.
As the tip speed is increased beyond the critical value, more
and more of the blade runs above the critical speed and is affected
by the compressibility factor. This effect is shown in Figs. 76
and 77 for a model propeller, but the model tests were apparently
influenced by Reynolds number to a much greater extent than
the full-scale tests, for they did not have such definite critical
speeds.1
The propellers having Clark-Y type sections had slightly
better efficiency below the critical tip speeds than the propellers
of the same thickness ratio with R.A.F.-6 type sections, but the
loss at the higher tip speeds was slightly greater. (This agrees
with the results of the model airfoil tests.) It seems that since
the reduction in efficiency at the higher speeds is accompanied
by a more turbulent state of airflow, the better airfoils with
smoother flow at low speeds suffer more at high speeds. The
Clark-Y sections, however, are very slightly superior throughout
the entire range covered in the tests.
Practical Data for Correcting Efficiency for High Tip Speed.
In both the full-scale and model propeller tests, in order to obtain
high tip speeds with the wind tunnel and motive power equipment
available, it was necessary to use propellers of rather low pitch.
The tip-speed effects were therefore greater than for propellers
of more ordinary pitch, and the results are not directly applicable
to ordinary propellers of any pitch. As it affects efficiency,
however, the tip-speed effect may be taken as a change of the
L/D of the blade sections, and on this basis the results may be
calculated for any pitch.
In the simple blade-element theory, the efficiency of a section
is
_ tan 4>
v ~ tan (0 + T) '
where tan y = D/L and tan <t> is proportional to the effective
pitch. By calculating the values of y for various tip speeds from
the propeller tests, the results can be applied by means of the
1 The air flow through the propellers was measured in the full-scale
propeller tests, and the thrust and torque grading curves as well as the
section airfoil characteristics will be available when the computations
are completed. Unfortunately they are still unfinished as this book goes
to press.
above relation to propellers operating at any effective pitch,

or V/nD.
Figure 83 gives the correction factor to the maximum efficiency
of ordinary thin-bladed metal propellers as computed by the
above method from the average results of tests on four standard
propellers. One is that shown in Fig. 176 for which test data are
given in Figs. 51, 52, and 165 to 175. One of the others was geo
metrically similar to it but 9 ft 6 in. in diameter, and the other
two were of the same general characteristics except for the plan
form which was of the narrow-tip design shown in Fig. 71. Two
of the propellers were tested at two different pitch settings, and
the results give a satisfactory check on the method of computing
the tip-speed correction factors for the other pitches.
Figure 83 gives the tip-speed correction factor to the efficiency
of ordinary aluminum-alloy propellers with sufficient accuracy
for performance computations. Fortunately the correction
factor is approximately the same for a given propeller operating
at various values of V/nD, so that while the values in Fig. 83
apply strictly to the region near maximum efficiency only, they
can be used with fair accuracy for any flight condition.
It will be noticed that the tip speed in Fig. 83 is taken as
, irnD
tip speed = >
cos <j)t
where the subscript t refers to the blade tip, and
tan 0( = ^D'
This expression for the tip speed takes into account not only
the peripheral speed in the plane of rotation but the speed of
advance as well. Propellers operating at higher values of
V/nD reach the critical tip speed at lower revolutions, assuming
constant diameter. For convenience in computing the tip
speed, Fig. 84 gives a curve from which the value of cos <j)t may be
obtained for a propeller operating at any V/nD. Although the
value of cos <j)t has practically no effect on the tip speed at low
values of V/nD, its effect reaches a value of about 10 per cent or
100 ft. per sec. at high values of V/nD.
As an example showing the use of the tip-speed correction chart
(Fig. 83), let us assume a metal propeller having a diameter of
10 ft. 6 in. operating at 2,100 r.p.m. on an airplane traveling at
110 m.p.h. (These figures represent an actual case, the Navy

SC-1 three-purpose airplane.) For this combination
_F = 88 X 110
nD ~ 2,100 X 10.5
= 0.44
and from Fig. 84, cos <t>t = 0.99. Then the actual tip speed is
wtlD
T.S. =
COS (t>t
= t X 35 X 10.5
0.99
= 1,167 ft. per sec.
Fig. 84.Curve for finding cos 0<.
From Fig. 83, for a tip speed of 1,167 ft. per sec. and a V/nD
of 0.44, the correction factor for efficiency is 0.86. The efficiency
of our propeller operating at the above V/nD but with a tip
speed under the critical would be, we shall say, 0.70. The
efficiency at a tip speed of 1,167 ft. per sec. is then
v = 0.70 X 0.86
= 0.60.
The high tip speed in this case causes a loss of 14 per cent in
the useful or thrust power available.
As shown in Fig. 83, thin-bladed aluminum-alloy propellers of
average form can be run at tip speeds up to 1,000 ft. per sec.
without an efficiency loss due to high tip speed. Figure 85
shows directly the maximum propeller diameter which can be
used at any revolutions and speed of advance to keep within the
limit of 1,000 ft. per sec. and thereby avoid tip-speed loss.
In applying the tip-speed correction to airplane performance

calculations, it should be kept in mind that with ordinary engines
and propellers the maximum revolutions occur only at high speed,
and the revolutions, and consequently the tip speeds also, are
appreciably lower in taking off, in climbing, and in cruising
flight. It is therefore quite possible to have a fairly large tip-
speed loss at maximum horizontal speed, but if the tip speeds in
climb and cruising are below the critical it is at the same time
\
Airpijne Speed
,100 M PH
M.P.H.
^,220 M.P. H
^JOOM.P.H
5L
1000 1200 1400 1600 1800
2000 2200 2400 2600 2800 3000
R.P. M.
Fig. 85.Maximum diameters for average metal propellers to operate without
tip-speed loss.
possible to escape tip-speed loss in all performance excepting

maximum speed. This, except in the case of racing and some
military airplanes, is of little practical importance.
Effect of Tip Speed on Noise.It has been noticed in all of the
model and full-scale propeller tests mentioned in this chapter, as
well as in electric whirling tests of propellers for strength deter
mination, that the sound emitted increases with tip speed. As
the tip speed is raised above 1,000 to 1,100 ft. per sec. or through
the neighborhood of the speed of sound, the propeller noise
suddenly increases in intensity, especially in or near the plane of
rotation. During the full-scale tests in the N.A.C.A. Propeller
Research Tunnel, at tip speeds above the critical a pulsation

was set up in the air which caused the hands and faces of the
observers to tremble, indicating that violent compression waves
were passing through the air.
Obviously, one way to reduce propeller noise is to reduce the
tip speed.
CHAPTER IX
BODY AND PROPELLER INTERFERENCE
All propellers used on aircraft operate near other bodies,

such as power plants, fuselages, or wings. In practically all
cases, the propeller is attached directly to the power plant and
is located within a few inches of it. Since both the propeller
and body affect the flow of the air through which they are passing,
but affect it in entirely different ways, each influences the aero
dynamic characteristics of the other.
0.12
0.10
0.08
c,
and 0.06
0.04
0.02
0
0 02 0.4 0.6 0.8 1.0
V
nD
Fig. 86.Comparison of thrust and power coefficient curves for a propeller
alone and with a VE-7 airplane.
Effect of Body on Propeller Characteristics.A body moving
through the air has a resistance, and a certain amount of air is
dragged or moved along with it. Also, even though the body
were moving through a perfect non-viscous fluid, the velocity
would be reduced near the nose. If a propeller is operating
in the air affected by the body it will in effect be working at a
lower rate of advance than if it were in free air. Thus, with
136
BODY AND PROPELLER INTERFERENCE 137
respect to the local air in which it is working, it has in effect

a lower V/nD and the blade elements work at higher angles of
attack. Both the thrust and the torque are therefore greater
than if the body were not present. This is shown in Fig. 86, in
which curves of the propeller thrust and power coefficients are
given for a model propeller alone, and also mounted in front of a
model VE-7 airplane.1
The body does not have the effect of a change of velocity
which is uniform throughout the entire plane of the propeller
disc. With bodies of ordinary shape, having their longitudinal
0 4 8 12 16 20 24.
Radius in Inches
Fig. 87.-Survey of velocity in propeller plane. VE-7 model.
center lines in the neighborhood of the propeller shaft axis, the

velocity of the air with respect to the body is reduced very greatly
at the center but practically none out as far as the propeller tips.
The velocity distribution in the propeller plane, but with the
propeller removed, is shown in Fig. 87 for the VE-7 model air
plane.2 It will be noticed that due to the landing gear, the
outer half of the propeller blades must work at lower effective
1 Comparison of Tests on Air Propellers in Flight with Wind Tunnel Model
Tests on Similar Forms, by W. F. Durand and E. P. Lesley, N.A.C.A.T..R.
220, 1926, model propeller B'.
s Tests of Five Metal Model Propellers with Various Pitch Distributions
in a Free Wind Stream and in Combination with a Model VE-7 Fuselage,
by E. P. Lesley and E. G. Reid, N.A.C.A.T.ft. 326, 1929.
velocities of advance when the blades are down than when

they are up. The average velocity at the propeller tip radius
(18 in.) is the same as the free air velocity, but at the radius
corresponding to the edge of the propeller hub the velocity has
been reduced to one-half the free air value.
It is interesting to compare this velocity distribution in the
propeller plane produced by the body alone with the velocity
distribution caused by the propeller alone as shown in Figs.
57 and 58. Near the hub the velocity is reduced by both the
body and the propeller. This means that they both set up
resistance to the forward motion in the hub section, but fortu
nately when placed one close behind the other they cause a total
resistance which is smaller than the sum of the two separate
values when each is operating alone. Throughout the rest
of the blade, or what is commonly called the working portion,
the velocity is increased by the action of the propeller but
decreased by the drag of the body. Also, due to its torque
reaction the propeller gives the air back of it a twisting or rota
tional motion.
Unfortunately the mutual interference is too complicated for
the individual effects of the propeller and body alone to be simply
added together vectorially and the resultant velocity to be used
as a basis for the computation of the propeller characteristics
by means of the blade-element theory. Computations of this
nature have been made by the author for a series of model
propellers run in combination with the model VE-7 fuselage,
but the results did not check the test results. Several methods of
blade-element analysis including the effect of body interference
have been published,1 but none at the present time satisfactorily
covers the problem.
Effect of Propeller on Body Drag.If the body behind (or
in front of) a propeller is not of extremely small size, it lies in
the region of air which has been given an increase of axial velocity
by the propeller, and the drag of the body is increased. This
slipstream effect on body drag has been found from a large
1 The Airflow round a Body as Affecting Airscrew Performance, by C. N.

H. Lock, H. Bateman, and H. C. H. Townend, British R. and M. 956,
1925. Analysis of Experiments on an Airscrew in Various Positions within
the Nose of a Tractor Body, by C. N. H. Lock, British R. and M. 1120, 1927.
Influence of Fuselage on Propeller Design, by Theodor Trailer, N.A.C.A.-
T. M. 492.
number of tests to vary directly with the propeller thrust coeffi

cient Tc in such a manner that
- = a + bTc
no
. IT
~ a ~t~ PV2D2'
where R is the resistance of the body when in the propeller

slipstream, RB is the resistance of the body alone at the free air
velocity, T is the thrust of the propeller (or tension in the pro
peller shaft) when the body is present, and a and b are constants.
FlQ. 88.-Increase of body drag with propeller thrust coefficient.

Theoretically a is equal to 1, and bTc represents the ratio
of the increase in drag due to the influence of the propeller.
Actually, model tests show that for ordinary airplane fuselages
the values of a are very close to 1. In one series of tests,1
however, in which smooth streamline bodies having very blunt
noses were used, the values of a ranged from 1 to nearly 2. This
was very likely due to the propeller increasing the drag very
considerably even at zero thrust, because of the disturbance of
the air flow over the smooth body caused by the relatively poor
1 Experiments with a Family of Airscrews Including Effect of Tractor and
Pusher Bodies, Part II, by A. Page, C. N. H. Lock, H. Bateman, and D. H.
Williams, British R. and M. 830, 1922.
hub sections of the propeller. Values of b vary from about 1 to

about 13, showing that the increased drag due to the slipstream
varies greatly with different bodies. Curves of R/R0 plotted
against values of Tc are given for various bodies in Fig. 88. It
will be noticed that the increase in drag due to the propeller is
much greater for good aerodynamic bodies of smooth form than
for ordinary fuselage shapes. In fact, from the tests made to
date it may be said that the greater the drag coefficient of the
body the lower will be the value of b, or the relative increase in
drag.
Various Expressions for Efficiency.If the efficiency is found
for a propeller acting near a body in the same manner as for the
propeller alone, the higher thrust and torque of the propeller
will be considered, but not the increased drag of the body. Thus
the apparent efficiency is found by means of the same expression
as the efficiency of a propeller in free air, or
TV
Apparent efficiency = =
ZirniJ
where T and Q are measured for the propeller while running in the
presence of the body. The thrust is increased more than
the torque by the presence of the body, and the apparent
efficiency is therefore always higher than the efficiency of the
propeller alone at the same V/nD. Values of apparent efficiency
of over 140 per cent have been measured in model propeller tests
with blunt-nosed bodies, showing that this is not a true measure
of the efficiency. It is therefore misleading and should not be
used.
In another manner of computing the efficiency of a propeller
with a body, the combination of propeller and body is looked upon
as a complete propulsive unit. The net thrust is then considered
the resultant horizontal force, or the propeller thrust minus the
body drag, each measured in the presence of the other. The
expression for net efficiency is then
Net efficiency = ^^p^-
The net efficiency is always lower than the free propeller efficiency
or the apparent efficiency, fairly typical curves of each being
shown in Fig. 89. The net efficiency is useful for showing the
effectiveness of engine-propeller units which are rather isolated
from the rest of the aircraft, such as the power nacelles used on
dirigible airships. It is not useful, however, for cases where the

propeller is near some essential part of the aircraft, such as a
thick wing or a fuselage, which is more than merely a power-plant
housing.
If the body in the presence of the propeller is considered a
necessary part of the aircraft, such as a place in which to carry
load, or a sustaining wing, then the drag R0 which it would
have if the propeller were not near it should not be deducted from
1.0 r
0 0.1 0.2 0.3 0.4

0.5 0.6 0.7 O.B 0.9 1.0
V.
nD
Fig. 89.Various forma of efficiency for tractor body B, B. and M. 830.
the propeller thrust. The body under the influence of the

propeller, however, has a higher drag R than it would have in
free air. The thrust effective for propulsion is therefore the
actual propeller thrust T in the presence of the body, minus the
increase in body drag, or (R R0), which is caused by the pro
peller. The efficiency based on this effective thrust T (R R0)
is called propulsive efficiency and is given by the equation
T - (R - B,),
Propulsive efficiency = v =
2wnQ
The symbol j? is used to represent both propulsive efficiency
and the efficiency of an isolated propeller, the context being
sufficient to make clear which is meant.
The propulsive efficiency is a reasonable measure of the pro

portion of useful work done by a propeller in the presence of a
body, and it is the only type of efficiency used for that purpose
in this work.1 If there is no body, present, both R and R0
become zero and the expression becomes identical with the one
given for the propeller alone. The propulsive efficiency curve is
given in Fig. 89 along with the curves for the other forms of
efficiency. It will be noticed that the propulsive efficiency is
closer to the efficiency of the propeller alone than are the apparent
or net efficiencies. As a matter of fact, with ordinary bodies the
propulsive efficiency is usually only slightly below the correspond
ing isolated propeller efficiency, and under certain conditions it
can even become slightly greater.
Effect of Body Size.Model tests have been made in England2
with three different-sized but similar bodies behind an average
two-bladed propeller having a pitch-diameter ratio of 0.7.
The bodies were circular in cross-section and had smooth low
drag shapes but blunt nose portions. The three ratios of maxi
mum body diameter d to the propeller diameter were 0.40,
0.60, and 0.75. The smallest 0.40 is about the average value
found with most open cockpit fuselages on small airplanes, while
0.75 is larger than is ordinarily used in practice. In Fig. 90
the propulsive efficiency curves are shown for the propeller
with each of the bodies along with the efficiency curve of the
propeller alone. The maximum propulsive efficiency with the
smallest body is only about 2 per cent less than the maximum
efficiency of the propeller alone. This happens to be fairly typical
for airplane fuselages of about th same size with respect to the
propeller, although the shape of the body is not representative.
The maximum efficiency is about 5 per cent less with the medium
body and 7 per cent less with the largest body than the efficiency
of the propeller alone.3 These percentage differences in effi-
1 In airplane performance computations the use of propulsive efficiency,
instead of isolated propeller efficiency, eliminates the necessity for and the
inaccuracies due to the correction for the extra drag of the parts in the
slipstream.
Pusher Bodies, Part II, by A. Page, C. N. H. Lock, H. Bateman, and D. H.
3 The percentage difference is based upon the ratios of the actual values of
efficiency, not on the direct difference. Thus an efficiency of 0.75 is 6.25
per cent lower than an efficiency of 0.80.
ciency remain approximately the same throughout the whole

flight range.
The fact that the propeller operates in slower air with the
larger bodies is shown by the larger values of V/nD for both zero
thrust and maximum efficiency with the larger bodies.
The above tests give a good idea of the manner in which the
size of body and propeller affects the propulsive efficiency through
a fairly wide range, but unfortunately the bodies used in the tests
v
nO
Fig. 90.Propulsive efficiency of an average propeller with various sized
bodies.
are quite different from ordinary aircraft, body forms. Full-

scale tests have been made, however, covering a smaller range of
relative sizes but with an actual 200-hp. nine-cylinder air-cooled
Wright J-5 engine in the nose of an average open-cockpit fuselage
with an undercarriage. The tests were made in the 20-ft.
Propeller Research Tunnel of the N.A.C.A.1 The engine had
normal cowling leaving the cylinder heads exposed, as shown in
Fig. 91. Four metal propellers were tested, varying in diameter
1 Full Scale Tests with a Series of Propellers of Different Diameters on a
Single Fuselage, by Fred E. Weick, N.A.C.A.r.B. 339, 1930.
from 8 ft. 11 in. to 10 ft. 5 in. All were geometrically similar

to the propeller shown in Fig. 176 and were set at a blade angle
of 15.5 deg. at 0.75R. The maximum cross-sectional area of the
fuselage was 11 sq. ft., and assuming a circular cross-section
having the same area, the tests were run at values of d/D from
0.36 to 0.42. The overall diameter of the engine was 45 in.
The results showed little difference in the characteristics of the
four propellers tested, the only one of importance being an
increase of propulsive efficiency of the order of 1 per cent for a
5 per cent increase of propeller diameter, or a 5 per cent decrease
Fig. 91.Open cockpit fuselage with Wright J-5 engine.
in the ratio d/D (see Fig. 92). This is the same general tendency
as, but to a greater degree than, the results of the tests with the
blunt-nosed model bodies which showed a variation of about 1
per cent in efficiency for a 10 per cent change of d/D. It is
thought that the full-scale tests give good results for practical
use within the limited range of the tests but that a mean value of
1 per cent change in efficiency with an 8 per cent change in
diameter is probably a more accurate figure to use if the range is
considerably larger.
Effect of Body Shape and Drag.With the smooth blunt-
nosed bodies used in the model tests shown in Fig. 90, the pro
pulsive efficiency is lower with any of the bodies than the efficiency
of the propeller alone. This is not necessarily the case with all
bodies, for tests have been made at Stanford University1 in

which with certain bodies the propulsive efficiency was found
to be higher than the efficiency of the propeller alone. Figure
93 tabulates the results of some of the tests, in which a 3-ft.
propeller having a pitch-diameter ratio of 0.9 was tested in front
of a series of bodies all 1 ft. in diameter but with various amounts
Diam. 10 -S
-Diam. 9-11 \ s
Diam. 9'-S -,\
Diam.
0.8
0.2
Diam e'-n"
Diam 9'-5"a nd9lh 311
Diarr JO'S
0
0 0.2 0.4 v 0.6 0.8
7TD
Fig. 92.Propulsive efficiencies with different sized propellers on an average
fuselage.
of taper at the propeller end. It will be noticed that with all of
the bodies except that having the greatest taper, the propulsive
efficiency is higher than the efficiency of the propeller alone.
Also, for all of the bodies with a cylindrical portion the propulsive
efficiency is greater for the bodies with greater resistance. This is
partly because the bodies are small enough to be almost entirely
in the central portion of the propeller where the poor propeller
sections have the effect of decreasing rather than increasing the
velocity of the air over the body, and partly because the poorer
bodies close to the central portion of the propeller partially
blank off or shield this poor portion of the propeller which
1 The Effect of Slipstream Obstructions on Air Propellers, by E. P. Lesley
and B. M. Woods, N.A.C.A.r.ft. 177, 1924.
represents a loss rather than a gain when the propeller is operating

alone.
It may be said that in general the propulsive efficiency is
higher with bodies of poor aerodynamic shape than with those
having low drag. To have a helpful influence on the propulsive
Percentage
Difference r
Maximum fromEff. 0
Propulsive of Propeller Rp
Efficiency Alone ~ qA
Propeller Alone
0.608 0%
0.809 +0.1% 1.101
0.835 +J.4% 0.695
,P 0.824. +2.0 % 0.140
XZ I Dtf 0.818 +1.2% 0.122
0.797 -1.2% 0.092
Direction of
Air Flow
Fig. 93.-Some bodies with which the propulsive efficiency is higher than the
efficiency of the propeller alone.
efficiency, however, the high drag must be due to a poor nose

where the effect is close to the propeller. This is shown by some
English tests in which a propeller was run with its poor hub
portion enclosed in a spinner forming the nose of a good streamline
body, and excrescences in the form of annular rings to increase
the drag were added to the body at various distances from the
nose.1 All of these annular rings had the same amount of area.
Pusher Bodies, Part IV, by H. Bateman, H. C. H. Townend, and T. A.
Kirkup, British R. and M. 1030, 1926.
The propulsive efficiency is less with the excrescences than with

the smooth body, as tabulated in Fig. 94. Also, as would be
expected, it is less with the smooth body than for the propeller
alone. Considering only the cases in which the excrescences are
present, however, the propulsive efficiency is higher if they are
located nearer the nose, even though the drag also is higher.
FlG. 94.-Effect of body excrescences on propulsive efficiency.

In fact, even the net efficiency is higher for the forward position
of the annulus, indicating that if the body must be poor, say due
to a radiator, it will improve the general effectiveness to put
the poor portion at the nose and close to the propeller rather
than back away from it.
The model tests referred to have been made with bodies which
were either of nearly perfect streamline form or of some simple
geometrical shape. Neither of these types is actually used on
airplanes. A few tests have been made which show the effect
of actual airplane shapes on propulsive efficiency. One of these
was on a model of the VE-7 airplane tested with several standard-

type wood propellers.1 The airplane is shown in Fig. 95. The
propulsive efficiency with a propeller of pitch-diameter ratio
Fiq. 95.The VE-7 airplane.
0.7 is compared with the efficiency of the propeller alone in Fig.

96. The thrust coefficient CT is here based on the effective
thrust T {R R0), which may be called the tension in the
propeller shaft minus the increase in body drag due to the
0.16 0.08 r
>7
0.14 0.07 .\ 0.7
\\
\\
0.12 0.06 . s 0.6
CP y\\
"a 0.05 as
' s > \ \\ 1
//"
I1 a. 0.08 l> - 0.04-| '/// \ 0.4-
I \
0.06 0.03 J\ 0.3
0.04 0.02 \V 02
\ 1
Propeller alone \ I
0.02 0.01 *Ji 0.1
0 0
0.2 OA
v 0.6 as 1.0
rTD"
Fiq. 96.-Comparison of propulsive efficiency with and without VE-7 airplane.
propeller. The maximum propulsive efficiency is about 2 per

cent lower than the maximum efficiency of the propeller alone,
and this is within one or two per cent of the reduction obtained
Tests on Similar Forms, by W. F. Durand and E. P. Lesley, N.A.C.A.T.B.
220, 1926.
with most ordinary airplanes having either air-cooled or water-

cooled engines in the nose of the fuselage. This is shown by
many full-scale tests in the N.A.C.A. Propeller Research Tunnel,
in which the same propellers have been tested on the VE-7
airplane and on both open-cockpit and cabin fuselages with
both radial air-cooled and V-type water-cooled engines.1 The
propulsive efficiency is very nearly the same with all fuselages,
it being highest for one with a radial air-cooled engine having com
plete or N.A.C.A.-type cowling, and lowest for a V-type water-
cooled engine (Curtiss D-12) entirely enclosed in a smoothly
rounded cowling. No radiator was in the airstream in the latter
case, and the body had the lowest drag coefficient of any tested.
The maximum propulsive efficiency ranged about 2 per cent less
with this body than with any of the others, and about 4 per cent
less than the one with N.A.C.A. cowling.
Effect of Position of Propeller with Respect to Nose of Body.
The English tests referred to, in which an annular ring was
located at various distances from the nose of a streamline body,2
also included different positions of the propeller in the nose
portion. Two bodies were used, one having a pointed nose and
the other having a rounded nose, both being of good streamline
shape, and both being the same back of the nose portion. The
nose portion of the body formed a propeller spinner in each case,
and three different locations of the propeller were selected with
each body so that 0.25D, 0.39D, and 0.46D of the propeller were
enclosed. The maximum propulsive efficiency and the reduction
from the efficiency of the propeller alone are tabulated in Fig.
97.
With both shapes of bodies the efficiency is reduced as the
propeller is moved back into the body, even though more of the
inefficient hub section is enclosed. With the forward position
the efficiency with the pointed nose is slightly higher than that
1 Drag and Cooling with Various Forms of Cowling for a "Whirlwind"
Radial Air-cooled Engine, Parts I and II, by Fred E. Weick, N.A.C.A.T.R.
313 and 314, 1929. Full Scale Wind Tunnel Tests of a Series of Metal
Propellers on a VE-7 Airplane, by Fred E. Weick, N.A.C.A.r.B. 306, 1929.
Working Charts for the Selection of Aluminum Alloy Propellers of a Standard
Form to Operate with Various Aircraft Engines and Bodies, by Fred E.
Weick, N.A.C.A.T.R. 350, 1930.
2 Experiments with a Family of Airscrews, Including Effect of Tractor and
Pusher Bodies, Part IV, by H. Bateman, H. C. H. Townend, and T. A.
Kirkup, British R. and M. 1030, 1926.
with the blunter rounded nose, probably because with the pointed
nose the propeller is located farther forward of the main portion
of the body. For the two rear positions the efficiency is greater
with the rounded nose.
The same tendency of the propulsive efficiency to become
greater as the propeller is moved forward with respect to the
body has been shown to be true for actual fuselage shapes, by
Percentage
Difference
Maximum from Eff. Ro
Propulsive ofPropeller
Efficiency Alone ~qA
0.699 -4.07. 0.066
c
0.660 -9.27. 0.066

3(>
0.6S5 10.07. 0.066

3>
0.690 -5.07. 0.066
0390--? 0.678 -6.87. 0.066
0.668 -8.27. 0.066

Direction of
Air Flow
Fig-. 97.Effect of position in nose on propulsive efficiency.
both model and full-scale tests. The model tests were made at
Stanford University with a 3-ft. propeller and a model DH-4
fuselage.1 The DH-4 airplane had a water-cooled Liberty
engine with a large flat-nosed radiator and no spinner. Moving
the propeller from % to 4 in. ahead of the radiator increased
the propulsive efficiency 4 per cent with a propeller having a
pitch-diameter ratio of 0.7, and 5 per cent with one of 0.9.
1 Interaction between Air Propellers and Airplane Structures, by W. F.
Durand, N.A.C.A.r.ft. 235, 1926.
The full-scale tests were made in the N.A.C.A. Propeller

Research Tunnel with an open-cockpit fuselage having a radial
air-cooled Wright J-5 engine with ordinary cowling as shown in
Fig. 91.1 The same 10 ft. 5 in. metal propeller was tested on
the fuselage with a direct-drive engine and also with a geared
engine of the same type, the propeller being located 7.5 in.
farther forward in the latter case. The propulsive efficiency was
found to be 4.4 per cent greater with the propeller in the forward
position. This is probably an extreme case due to the high drag
but small body interfering with the central portion of a large
propeller.
Comparison of Tractor and Pusher Arrangements.In the
British tests with the smooth blunt bodies previously referred
to,2 three pusher arrangements were tested as well as the three
tractor. The same propellers were used with all of the bodies,
and each of the pusher bodies had the same diameter as one of
the tractor bodies. The pusher bodies had the bluff portion
at the rear and consequently had higher drag coefficients.
The maximum propulsive efficiencies are given in Fig. 98 for
both the pusher and tractor bodies. Within the limits of experi
mental error there is no difference between the propulsive effi
ciencies obtained with the pusher and tractor bodies having the
same diameter. In both cases it is lower with the larger bodies.
With the particular bodies tested the net efficiencies are higher
for the tractor than for the pusher arrangements, due to the
greater drag of the pusher bodies. This is likely to be the case
in practice also because of the desirability of keeping the propeller
close to the engine, but it is not necessarily so for there have been
a few cases in which the propeller was mounted some distance
from the engine on an extended shaft.
Mutual Interference between Propellers and Wings.Very
little experimental work has as yet been done on the mutual
interference between propellers and wings,3 although some model
1 The Effect of Reduction Gearing on Propeller-body Interference, as
Shown by Full-scale Wind Tunnel Tests, by Fred E. Weick, N.A.C.A.T.jB.
338, 1930.
Pusher Bodies, Part II, by A. Fage, C. N. H. Lock, H. Bateman, and D. H.
3 A rather complete series of tests with a 4-ft. propeller mounted with a
nacelle in a large number of positions with respect to a wing having a span
of 16 ft. is now underway in the N.A.C.A. 20-ft. Propeller Research Tunnel.
tests showing the general effect of a propeller on a wing have been

made at the Gottingen Aerodynamics Laboratory in Germany.1
These tests indicate that if the propeller is above the wing, it
increases the natural circulation and therefore also the lift.
Conversely, if the propeller operates below the wing, the lift is
reduced somewhat. The effect on lift is, however, ordinarily
quite small.
Maximum
Propulsive
Efficiency
Direction of
A ir Flow
Fig. 98.Comparison of propulsive efficiency with pusher and tractor bodies.
The wing also has an effect on the propeller characteristics,

for if the propeller is above the wing it is in a region of faster-
moving air, and its thrust and power coefficients are less than
in free air at the same V/nD. Again, if the propeller is below the
wing the effect is the opposite. If the propeller axis is near the
chord line of the wing, the lift and drag of the wing increase due to
the increased velocity in the slipstream, and the propeller thrust
and torque increase because the wing drag reduces the air velocity.
1 Mutual Influence of Wings and Propeller, by L. Prandtl, N.A.C.A.T.N.
74, 1921.
With most present-day airplanes, and particularly the single-

engined ones, the effect of the wings on the propeller character
istics is small. Full-scale tests have been made in the N.A.C.A.
Propeller Research Tunnel in which five different single-engined
airplanes having monoplane and biplane wings were tested both
with and without the wings.1 The propulsive efficiency was
1.0 9 Ft. Diameter Propeller Set 1S a t 42 "Rad 0.10

1 1 1
0.9 With Wma 0.09
mhojtWn
0.8 0.08
0.7 0.07
0.6 |0.06
C
0.5 \ 0.05
//
0.4 \\ 0.04.
0.1 0.01
\
0.2 / 0.02
s 0.01
0.1
0 1 \\ 1 . 0
V
nD
Fig. 99.Effect of wing on propulsive efficiency with a single-engined airplane.
reduced from 1 to 3 per cent by the presence of the wings, the loss
being about the same for the biplane and monoplane wings
tested. The loss was slightly greater with high-pitch than with
low-pitch propellers.
Typical curves of propulsive efficiency and power coefficient
are given in Fig. 99 for a cabin monoplane with and without the
wing. The power coefficient CP is slightly greater with the wing
1 The Effect of the Wings of Single Engine Airplanes on Propulsive
Efficiency as Shown by Full-scale Wind Tunnel Tests, by Fred E. Weick
nnd Donald H. Wood, N.A.C.A.T.AT. 322, 1929. Full-scale Tests of Wood
Propellers on a VE-7 Airplane in the Propeller Research Tunnel, by Fred E.
Weick, N.A.C.A.T./2. 301, 1929.
in place, indicating that the wing reduces the velocity through

the propeller plane even though the leading edge is more than
half the propeller diameter to the rear.
The VE-7 airplane was tested both with and without the tail
surfaces, and there was no measurable difference in the propeller
characteristics.
The effect of an exceptionally large wing on propulsive effi
ciency has been tested at Stanford University.1 The wing chord
was 2.37 times the propeller diameter and the thickness was
Fig. 100.Effect of large wing on propulsive efficiency.
0.42D. The tests were made with two propellers having pitch-
diameter ratios of 0.7 and 0.9, and with each propeller located at
three different distances in front of the wing. Figure 100 shows
the percentage loss in propulsive efficiency compared with the
efficiencies of the propellers alone. It will be noticed that while
the loss is great if the propeller is very close to the wing, there
should even in this extreme case be little loss if the propeller were
located at a distance of D/5 or more ahead of the leading edge.
As in the case of the smaller wings and most bodies, the loss in
1 Interaction between Air Propellers and Airplane Structures, by W. F.
Durand, N.A.C.A.r.ft. 235, 1926.
propulsive efficiency is slightly greater with the propeller having

the higher pitch.
The above tests represented the case of a wing engine installa
tion for a proposed Army night bomber, the central fuselage
and nose engine arrangement of which was also tested. As shown
in Fig. 101, due to the great size of the body and wing, and the
machine gun turret in the nose, this probably represents the most
extreme case of body and propeller interference tested to date.
The projected area of the body covers almost the entire propeller
disc area. The loss in propulsive efficiency was 25 per cent for
both the 0.7 and 0.9 pitch-ratio propellers close to the body, but
.13-
Fig. 101.-Central section of proposed night bomber with extreme propeller
interference.
if the results are extrapolated it seems that a very small loss

might be expected if the propeller were placed at least D/2
ahead of the body.
Effect of Propeller Pitch Distribution.As shown by Fig. 87,
the effect of an ordinary type of body located near the center
of the propeller disc is to reduce the velocity of the air passing
through the propeller, this reduction being great at the hub and
little, if anything, at the blade tips. Tests have been made with
five model propellers having different pitch distributions, operat
ing both alone and with a model VE-7 airplane.1 The model
propellers were of metal and, except for the blade angles, were all
geometrically similar to the typical metal propeller shown in Fig.
177. Model A had a uniform pitch over the entire blade with a
pitch-diameter ratio of 0.7. The others all had the same average
pitch, but the pitch of B increased from 0.6J? to the tip, that of C
decreased over the same range, while that of D increased and E
1 Tests of Five Metal Model Propellers with Various Pitch Distributions
in a Free Wind Stream and in Combination with a Model VE-7 Fuselage,
by E. P. Lesley and E. G. Reid, N.A.C.A.r.B. 326, 1929.
156 AIRCRAFT PROPELLER DESTGN
decreased from 0.6R to the hub. Curves showing the distribu

tion of pitch along the radius for the various models are given in
Fig. 102, along with values of the maximum efficiencies of the
propellers alone and the maximum propulsive efficiencies with
the model airplane.
Maximum
Max.Eff. Propulsive Percentage
Propeller Eff-witt) Losswith
i or P/o=0.7fc_ Alone VE-7 Body
" '
fife 5|- Prop. A 0.811 0.762 6.0%

o
o as no
7r
1.0
P/d05 - P/d' Prop.B 0301 0.767 4.2% '
J 1
0 0.5 1.0
'.r fP/D
Prop.C 0.800 0.750 6.2%
Prop.D 0.811 0.750 7.5%
*p]D Prop.E 0.811 0.790 2.5 7o
Fig. 102.--Effect of pitch distribution on propulsive efficiency with VE-7 body.
The maximum efficiencies of the propellers alone are all sur

prisingly near the same value. Models A, D, and E all had
maximum efficiencies of 0.811. These models all had a uniform
or nearly uniform pitch over the outer portions of the blades,
and the results indicate that without a body present the sections
near the hub are relatively unimportant. Propellers B and C
had efficiencies about one per cent lower.
With the model airplane, the propellers had propulsive effi

ciencies in the order in which their pitch distributions fitted the
flow conditions set up by the body (as shown in Fig. 87). Pro
peller E had a propulsive efficiency of 0.79, which was noticeably
better than any of the others. This propeller had a pitch
distribution of such a nature that the angles of attack at maxi
mum efficiency, as calculated by the simple blade-element theory
on the basis of the average velocity given in Fig. 87, were prac
tically the same over the entire blade. The pitch distribution
therefore fitted the flow conditions caused by the body remark
ably well. Propeller B, which also had the pitch increasing from
hub to tip but did not fit the airflow so well, was next with a
propulsive efficiency of 0.767. Model D, in which the pitch
increased toward the hub, had as high an efficiency as any when
tested alone but as low an efficiency as any when tested with the
airplane, giving it the greatest loss. This indicates that the
sections near the hub are more important when the propeller
is working with a body.
It is also interesting that the loss in efficiency with the uniform-
pitch propeller A in front of the VE-7 model was 6 per cent, while
the average loss with five uniform-pitch model wood propellers
in front of the same model airplane1 was only 4 per cent. The
smaller loss with the wood propellers is probably due to their
thicker and therefore less efficient sections near the hub, for the
poorer sections would have less to lose by being forced to work at
poor angles by the body interference.
Practical Use of Body Interference Data.The simplest and
most accurate method of taking body interference into account
in designing a propeller is to base the design on wind tunnel
tests made with the particular design of body and propeller
desired. This is of course not ordinarily possible, but a selection
can be made from the results of full-scale tests made in the
N.A.C.A. Propeller Research Tunnel with series of typical metal
propellers tested at various pitch settings and with several
typical forms of bodies. Since there is fortunately little differ
ence in the interference effects on either propulsive efficiency
or power absorbed with moderate changes in body form, the
results of the above tests can be used directly for selecting a
Tests on Similar Forms, by W. F. Durand and E. P. Lesley, N.A.C.A.-
T.R. 220, 1926.
suitable propeller of the type tested to operate with any ordinary

form of single-engined tractor airplane.1 Corrections for
different relative propeller and body sizes can be based on the
information given in this chapter.
For bodies which do not fall within the range of the above tests,
or for propellers of different forms, the results of the various
tests which have been made can be used indirectly to show the
approximate effect to be expected. For example, although
few tests have been made with pusher installations, the results
of the few which have been made show that the effect on pro
pulsive efficiency and power absorbed is about the same as for
tractor installations having about the same size and resistance
of the body. A pusher propeller can therefore often be satis
factorily obtained by selecting it to fit an equivalent tractor
installation.
1 The test results are given in convenient form for this purpose in Working
Charts for the Selection of Aluminum Alloy Propellers of a Standard Form
to Operate with Various Aircraft Engines and Bodies, by Fred E. Weick,
N.A.C.A.r.ft. 350, 1930. The charts are also given in Chap. XVII.
CHAPTER X
THE EFFECT OF PROPELLER CHARACTERISTICS

ON AIRPLANE PERFORMANCE
In calculating the performance of an airplane, it is usual to

compute first the power required for horizontal flight at various
air speeds. The performance at any speed is then found from
the relation between the power required and that available,
son 2000
1800 z
1600
1400
0 20 40 60 SO 100 120 140

Air Velocity, M.P.H.
Fig. 103.-Curves of power required, power available, and revolutions; 8.17-ft.
metal propeller.
The curves of power required vs. air speed are of more or less
similar form for all airplanes.1 For showing the effect of pro
peller characteristics pn airplane performance it is convenient
to take as an example a typical airplane having a normal per
formance, such as the Vought VE-7. The curve of power
required2 vs. airspeed for the VE-7 is given in Fig. 103.
1 The manner of obtaining the power required may be found in books on
airplane design, such as "Airplane DesignAerodynamics," by E. P. War
ner, or "Engineering Aerodynamics," by Walter S. Diehl.
* The curve was obtained from flight tests reported in Characteristics
of Five Propellers in Flight, by J. W. Crowley, Jr., and R. E. Mixson,
N.A.C.A.T..R. 292, 1928.
159
Power Available.The power available under various con

ditions depends upon the characteristics of the power plant
and propeller. With the gasoline engine, as used in airplanes
today, the full-throttle torque is very nearly constant throughout
the working range of revolution speeds, although it drops off
slightly with increased revolutions.
A typical curve of full-throttle horsepower at sea level vs.
revolutions per minute is given in Fig. 104 for an engine rated
at 200 hp. at 1,800 r.p.m. In ordinary practice, if 1,800 r.p.m.
220
140 1 I I I I I I I I I I I I I I
1100 1400 1500 1600 1700 1800 1900 ZOOO
R.P.M.
Fig. 104.Full-throttle horsepower vs. r.p.m.
is specified by the engine manufacturer as the maximum revolu
tion speed for service use, the propeller is designed so as to absorb
the full engine power (200 hp.) at the maximum horizontal
sea-level speed and 1,800 r.p.m. These are often called the
maximum horsepower and revolution speed, although they are
actually merely the maximum specified for safety.
The useful or thrust horsepower available at maximum speed
at sea level is the product of the maximum brake horsepower
and the propulsive efficiency of the propeller.
EFFECT ON AIRPLANE PERFORMANCE 161
As an example, let us select from the series of metal propellers

whose aerodynamic characteristics are given in Figs. 51, 52, and
56 the one which will give the highest possible sea-level speed.
This will, of course, be the propeller having the highest efficiency
under the high-speed conditions. We find, either by means
of the method of the speed-power coefficient C or with the Eiffel
logarithmic diagram, that the highest speed is obtained with
the propeller having a diameter of 8.17 ft. and a blade-angle
setting of 22 deg. at 0.7572. (p'/D = 0.95, where p' = p at
0.75E).1
Neglecting the effect of the relative size of propeller and body,
this propeller gives a propulsive efficiency of 0.833 at maximum
speed. That efficiency was obtained, however, from tests with
a 9-ft. propeller on the VE-7 airplane, and as shown in Chap.
IX, our 8.17-ft. propeller will give a slightly lower efficiency
due to the effect of body interference. Taking the effect on the
efficiency to be 1 per cent for an 8 per cent difference in propeller
diameter, the actual propulsive efficiency with our 8.17-ft.
propeller will be 0.833 X 0.989 or 0.824, giving a thrust horse
power of 164.8 and a maximum speed of 128.5 m.p.h.
It is now desirable to find the useful power available at lower
speeds so that we can draw a curve of power available vs. airspeed.
Since the torque coefficient of the propeller increases as the
advance per revolution becomes less (see Figs. 47 and 48), but
the engine torque remains essentially constant, a decrease in
airplane speed will cause the engine to slow down to the point
where the torque of the engine and that of the propeller are equal.
To find the full-throttle r.p.m. corresponding to any other
than the maximum speed, the curves of V/nD vs. C, (Fig. 51 or
52) can be used. We shall proceed to find the r.p.m. at which
our 8.17-ft. metal propeller will turn at an air speed of 80 m.p.h.
First, we shall assume three revolution speeds which are likely
to cover the range desired, say 1,700, 1,600, and 1,500 r.p.m., and
find the value of C for each at 80 m.p.h. and the reduced power
corresponding to each revolution speed. (The latter is obtained
from Fig. 104.) This gives us three points through which we
1 To find this it is first necessary to assume a trial value for the propulsive
efficiency, from which the approximate thrust horsepower is calculated and
an approximate value of the maximum speed obtained from the power-
required curve of Fig. 103. This speed is then used in selecting the propeller.
If the efficiency does not check that assumed, the process is repeated.
can draw a curve of r.p.m. vs. C, (curve 1, Fig. 105). Now, for
each of the above three values of C,, we can find from Fig. 51
or 52 the value of V/nD for our propeller, which has a blade-angle
setting of 22 deg. at 0.75/2, and then, knowing V/nD, V, and D,
we can solve for the revolutions. This gives us another set of
values of r.p.m. vs. C,, which when plotted give curve 2, Fig.
105. At the intersection of the two curves, r.p.m. = 1,618
and C, = 0.939, which are the values for our propeller at full
throttle and 80 m.p.h.1 At 1,618 r.p.m. the engine develops
1700
2
ol 1600
1500,0.90
Fig. 105.
179.9 hp. (Fig. 104), and for C, = 0.919 the propulsive efficiency
is 0.715 X 0.989 = 0.707 (Fig. 51 or 52). The useful or thrust
power available at 80 m.p.h. is therefore
179.9 X 0.707 = 127.2 hp.
Following the same procedure, the revolutions and the power
available have been found for several air speeds and the curves
drawn in Fig. 103. The maximum horizontal speed is shown
by the intersection of the curves of power required and
power available. The vertical distance between the power-
required and power-available curves shows the excess of power
available above that required for horizontal flight, and this excess
can be used for climbing. At speeds higher than the maximum
horizontal speed (higher speeds can be attained by diving) the
1 The tests on which these figures are based were made with the propeller
axis parallel to the direction of airflow, whereas in actual flight the propeller
axis assumes angles as high as 10 or 12 deg. with respect to the flight path.
Full-scale tests in the 20-ft. Propeller Research Tunnel of the N.A.C.A.
have shown, however, that the propeller characteristics are practically
unaffected within that range. Other tests confirming this conclusion are
reported in Experimental Investigation of Aircraft Propellers Exposed to
Oblique Air Currents, by O. Plachsbart and G. Krober, N.A.C.A. T.M.
562, igso.
full-throttle revolutions will go above the maximum permissible,

and the engine should be throttled.
For the purpose of showing the effect on the performance of
the VE-7 of using other propellers of our metal series, the power-
available curves have also been computed for propellers having
blade-angle settings at 0.75R of 20, 17.5, 15, and 12.5 deg. (The
corresponding values of p'/D are 0.86, 0.74, 0.63, and 0.52.)
The propeller with a setting of 17.5 deg. operates at the peak
--8.11 , \ .
160 6.53i, 20
fiA1 nt ~V*
IP
140 ' 10 ISi 12'5'-'
Powt r
Availab le
u 120
OID-
QJ
/
Ig 100 /
t
jE 80
60
40
VelocHj.M.P.H
Fig. 106.Curves of power available with metal propellers of various diameters
and pitches.
of its efficiency curve at the maximum horizontal speed. The

power-available curves for these propellers, along with that for
our original "high-speed" propeller having a setting of 22 deg.,
are given in Fig. 106.
Effect of Propeller Characteristics on Maximum Speed.
Since the engine revolutions and power are specified for the high
speed condition of flight, the propeller must conform to the engine
requirements. In our example this means that it must absorb
200 hp. at 1,800 r.p.m. at the maximum speed of the airplane,
and since the brake horsepower is determined, the propulsive
efficiency is the only factor in the propeller design affecting the
thrust horsepower at maximum speed, and therefore the speed

itself.
The high-speed performance with the five metal propellers is
tabulated below:
Blade angle at 0.75K, deg 22.0 20.0 17.5 15.0 12.5

P'/D 0.95 0.86 0.74 0.63 0.52
Diameter, ft 8.17 8.53 9.04 9.61 10.18
Propulsive efficiency 0.833 0.828 0.810 0.780 0.710
Body correction factor 0.989 0.994 1.001 1.010 1.017
0.824 0.822 0.811 0.788 0.721
Max. speed, m.p.h 128.5 128.4 127.8 126.6 122.7
It is noteworthy that even wide extremes of pitch ratio and

diameter do not have a great effect on the high speed. The
propeller operating at its peak efficiency (17.5 deg. at 0.75R,
9.04-ft. diameter) is the one which in the past has been generally
160
140
fc* 120
o
ui
o 100
L.
f 80
60
40
60 TO 80 90 100 110 120 130
Velocity , M.P.H.
Fig. 107.-Curves of power available with wood propellers.
used as a propeller for all-around service on airplanes of this
power and speed. It gives a maximum speed only 0.7 m.p.h.
lower than that with the best high-speed propeller.
The propeller having the lowest pitch setting, which setting
is considerably lower than would ordinarily be used in practice,
also has the lowest efficiency at high speed. The maximum

speed with this propeller is 122.7 m.p.h., which is 5.8 m.p.h. less
than the best obtained. It is interesting to note that a reduction
in propulsive efficiency from 0.824 to 0.721 or 12.5 per cent
reduces the maximum speed of the airplane only 4.5 per cent.
The curves of power available for three typical wood propellers,
having pitch ratios of 0.6, 0.7, and 0.8, are shown in Fig. 107.
The data for these propellers, as well as for the metal propellers,
have been obtained from full-scale tests in the N.A.C.A. 20-ft.
Propeller Research Tunnel at Langley Field.1 The high-speed
performance with the wood propellers is as follows:
P/D 0 8 0.7 0.6

8 0 8.43 8.78
0 768 0.758 0.710
Max. speed, m.p.h 125 5 124.9 122.1^
1 These propellers had approximately the same diameter calculated, and so the body
interference factor is automatically taken into account.
The propeller with a pitch ratio of 0.8 has approximately the
best pitch ratio for high speed. It is interesting to note that the
diameter is nearly the same as that of the metal propeller giving
the highest speed and, also, that the maximum speed obtained
with the best metal propeller is 3 m.p.h. or 2.4 per cent higher
than that with the best wood propeller. This difference would of
course be greater if the propellers were operating at high tip
speeds.
A simple rule for finding the effect on the maximum speed of
ordinary differences in propulsive efficiency is that the percentage
variation in velocity is one-third that in efficiency, although this
usually underestimates the velocity variation by about 10 per
cent and it is somewhat more accurate to use 0.37 instead of one-
third. This rule, while of course of a very approximate nature,
is quite useful for making preliminary estimates and is usually
sufficiently accurate for designing propellers to fit their conditions.
Effect of Propeller Characteristics on Rate of Climb.When
the power available is greater than the power required, as is the
case for an airplane flying between its minimum and maximum
Research Tunnel, by Fred E. Weick, N.A.C.A. T.B, 301, 1928. Full-scale
Wind Tunnel Tests of a Series of Metal Propellers on a VE-7 Airplane, by
Fred E. Weick, N.A.C.A. T.R. 306, 1928.
horizontal speeds, the excess power available may be used for

climbing (see Figs. 106 and 107). The rate of climb in feet per
minute is given by the relation
D . . .. . EHP X 33,000
Rate of climb = ^ >
where EHP is the excess horsepower available over that required

for horizontal flight and W is the weight of the airplane in
pounds. Obviously the greatest rate of climb for a given air
plane is obtained at the speed at which the vertical distance
between the power-available and power-required curves is the
greatest. This speed is usually in the neighborhood of 0.6 of
the maximum speed with modern airplanes.
Because its torque coefficient becomes greater as the advance
per revolution decreases, the propeller holds the engine to a lower
revolution speed than that at the maximum velocity of the air
plane, even though the throttle is left wide open. As a result
the thrust or useful power at climbing speed is less than that at
the maximum horizontal speed, both because the engine delivers
less power at the lower revolutions and because the propulsive
efficiency is less at the lower rate of advance per revolution at
climbing speed.
With a given airplane and engine, the propeller which will
give the highest thrust horsepower at climbing speeds will give
the highest rates of climb. Referring to the power-available
curves for the three wood propellers in Fig. 107, it is seen that
the curves cross, and that while the highest-pitch propeller is
the best for maximum speed, it is also the poorest for climb,
and the one having the lowest pitch gives the greatest power
available for climbing.
In Fig. 108 are curves of the rate of climb vs. air speed for our
VE-7 at sea level, the total weight of the airplane being 2,200 lb.
Each curve shows the climb for one of the wooden propellers.
The one having a pitch ratio of 0.6 gives a maximum rate of
climb of 1,095 ft. per min., the one with a pitch ratio of 0.7
gives 1,062, and the best speed propeller with a pitch ratio of
0.8 gives only 1,020.
It will be noticed that on Fig. 106 the curves of power available
for the thin-bladed metal propellers do not cross above the speed
for maximum climb, so that the propeller giving the highest hori
zontal speed also gives the greatest thrust horsepower available at
climbing speeds and consequently the best climb! The curves of

rate of climb vs. air speed for our series of five metal propellers
1200
p/d -0.6. p/d 0.5
1
1000
c8
2 800
| 600
400
ZOO
v
60 70 80
90 100 110 IZ0 110
Velocity, M.P.H.
Fig. 108.Rates of climb vs. air speed with wood propellers.
1200
60 70 80
90 100 110 120 130
Velocity, M. P. H.
Fig. 109.-Rates of climb vs. air speed with metal propellers.
on the VE-7 are shown in Fig. 109. These curves show that for
these propellers the lower the pitch the lower is the maximum
climb, which is the reverse of the case with the wood propellers.
Another interesting point is that there is a difference of only 30 ft.
per min. between the best and the poorest maximum rates of climb.
The reason for this difference in climb between the performance
of the wood propellers and the metal propellers is made clear by
a study of the curves in Fig. 1 10. The top set of curves shows the
1ZOOr
0.95
' 0.90 0.7 Q'>-
g
<- _4 >- - 2
0.85 r-- ~Z
7zs
12.5 /.
0.80
o
'.4-O 0.75
0.80
0.75 12.56
l / .5
, 2r '
0.70
1
Cu 0.65 16 c .7
*- 0.60 0. 8
0.5 0.6 0.7 , 0.8 a9 1.0
Fig. 110.
maximum rates of climb obtained with both the wood and metal
propellers with the various pitch ratios. Unfortunately, the low
est-pitch wood propeller tested was not low enough to give the
maximum possible climb with that type of propeller, and so the
point for p/D = 0.5 has been calculated from tests on model
propellers of similar form. The wood propeller giving the maxi
mum climb would from the top set of curves of Fig. 110 have a
pitch ratio of about 0.56 and would give a rate of climb of 1,100
ft. per min. The metal propeller giving both the highest speed
and the best climb gives a rate of climb of 1,150 ft. per min.,
which is about 5 per cent better than that with the wood pro
peller giving the best climb and about 13 per cent better than with
the best high-speed wood propeller.
The middle set of curves on Fig. 110 shows the reduction of
engine revolutions at a speed of 75 m.p.h., which is about the
average speed for maximum climb with all of the propellers being
investigated. The ratio of the r.p.m. at 75 m.p.h. to the r.p.m.
at maximum horizontal speed is plotted against pitch ratio.
Since the engine torque is approximately constant, the ratio
of the revolutions is also an indication of the fraction of the
maximum permissible brake horsepower which is being used at
climbing speed. With both the wood and the metal propellers
the reduction in revolutions is less for the propellers of high than
low pitch, but the variation with pitch is much more pronounced
with the metal propellers. The highest-pitch metal propeller
makes use of a larger portion of the maximum horsepower
than any of the wood propellers, and the lower-pitch metal
propellers make use of considerably less than any of the wood
propellers. Thus the high-pitch metal propeller has a great advan
tage over the lower-pitch metal propellers in the amount of engine
power used in climb, while the high-pitch wood propeller has but
a very slight advantage over the lower-pitch wood propellers.
The bottom set of curves in Fig. 110 shows the variation of
propulsive efficiency with pitch ratio at a climbing speed of 75
m.p.h. The metal propellers are decidedly more efficient in
climb than the wood propellers. From the curves in Fig. 110,
it is seen that in climbing flight the drop in engine revolutions is
less for the higher-pitch propellers but that the efficiency is
greater with the lower-pitch propellers. Since with the wood
propellers the change in efficiency with pitch ratio is greater
than the variation in revolutions, the low-pitch wood propellers
are better in climb. Conversely, since with the metal propellers
the variation of revolutions is greater than that of efficiency, the
higher-pitch metal propellers are better in climb.
A simple and useful formula for quickly finding the effect of
changes in propeller characteristics on the rate of climb is as
follows:
A one-point change (i.e., from 0.71 to 0.72) in either the pro
pulsive efficiency or the fraction of the maximum revolutions
results in a change in the rate of climb, in feet per minute, of
330 X HPm. r 330
W ' power loading
This shows that in the case of a single-seater fighter type of

airplane, in which the weight per horsepower is very low, the
propeller can influence the absolute rate of climb more than with
a heavily laden transport airplane. Due to the much greater
climb of the fighter type, however, variations in the propeller
characteristics will have a larger relative effect on the climb
of the transport type. As an example of this difference suppose
that we have a typical 600-hp. fighter airplane weighing 3,000
lb. and having a maximum rate of climb of 3,000 ft. per min.,
and also a 600-hp. transport weighing 9,000 lb. and having a
rate of climb of 400 ft. per min. A total difference of five points
in efficiency and fraction of maximum revolutions, which as can
be seen from Fig. 110 is easily obtainable, will cause a difference
in the rate of climb of the fighter of 330 ft. per min. and a differ
ence in that of the transport of only 110 ft. per min. The
climb of the fighter, however, is changed only 11 per cent,
while that of the transport is changed 27.5 per cent.
In special cases where the climb of an airplane far outweighs in
importance the other performance characteristics such as the
high-speed and cruising performance, the propeller can be
designed to absorb the full permissible power of the engine at
climbing speed. The pilot is then relied upon to throttle the
engine at all higher air speeds.
The Effect of Propeller Characteristics on Take-off.The
take-off distance, or the length of the run of a land plane between
the time of starting and the time of leaving the ground, depends
on the accelerating force, which is the excess of thrust over the
resistance. The thrust becomes less as the speed increases,
and the resistance, which is partly due to rolling friction, varies
widely with different fields, making it practically impossible
to calculate the take-off distance accurately. An approximate
idea of the relative take-off qualities with different propellers
can be obtained, however, by comparing the average thrusts
over the take-off period. The average thrust can be found
with sufficient accuracy by taking the average of the thrusts
at zero advance1 (called static thrust), at the take-off speed or
1 The static thrust can be conveniently found by means of the equation
QCT 2ttQCt
T = or T = - >
DCq DCP
where Q is the full-throttle engine torque, which is substantially constant.
slightly higher, and at a third speed halfway between the other

two.
The various thrusts obtained with our metal and wood pro
pellers on the VE-7 airplane are tabulated below for speeds of
0, 30, and 60 m.p.h.
Blade angle Thrust Thrust Aver Approxi

Diam Static
Mate at 0.75ft thrust, at 30 at 60 age mate
rial eter, m.p.h., m.p.h., thrust take-off
or lb.
ft. pitch ratio lb. lb. lb. run, ft.
8.17 22.0 deg. 731 713 656 700 364

8.53 20.0 deg. 860 755 664 759 336
Metal 9.04 17.5 deg. 962 803 660 806 316
9.61 15.0 deg. 1,050 809 656 836 305
10.18 12.5 deg. 1,080 815 656 850 300
8.05 0.8 930 761 616 769 332
Wood 8.43 0.7 950 780 628 786 325
, 8.78 0.6 970 801 650 807 316
In the last column approximate take-off distances are given

for an average field and no. wind. These distances were obtained
from an average value by varying the distances for the various
propellers in inverse proportion to their average thrusts. This
is not, of course, strictly correct, but since the average resistance
is very small compared with the average thrust, the distances
are not greatly in error relative to one another.1
It will be noted that the best wood propeller in climb, i.e., that
having a pitch ratio of 0.6, is also the best of the wood propellers
in take-off. Also, the lowest-pitch metal propeller gives the
best take-off of all, and the highest-pitch metal propeller, which
1 Flight tests have been made by the author which give an interesting
check on the performances calculated for the VE-7 with the five metal
propellers. A Stearman biplane with a 200-hp. Wright J-5 engine was
tested with four metal propellers of different diameters, all being geometri
cally similar to the ones used for the VE-7 calculations. In each case the
pitch setting was adjusted to give 1,900 r.p.m. at maximum speed. The
performances measured were as follows:
Diameter, ft 8 8.5 9 9.5

Max. speed, m.p.h 114.5 114.5 114.0 112.0
Max. climb, ft./min 850 850 850 850
Take-off time, sec 10.5 8.3 8.2 8.0
gave the highest speed and the best climb, gives the poorest
take-off. The shortest distance, however, is only 18 per cent
less than the longest, which in this case makes a difference of
only 64 ft. This is practically negligible with ordinary airplanes
operating from average fields but might, of course, become
important in the special cases of small fields or heavily laden
airplanes.
The problem of a seaplane taking off from the water is some
what different from that of a land plane, for the resistance of the
water rises to a very large value at a speed called the "hump"
speed which is usually in the neighborhood of one-half the take-off
speed. It is often difficult to get sufficient propeller thrust
to push over the hump speed with heavily loaded seaplanes, and
sometimes it is necessary to design the propeller to absorb the
maximum permissible power of the engine at the hump speed.
In that case, as in the case where the full power is used in climb,
the pilot is relied upon to throttle the engine at higher air speeds.
Effect of Propellers on Cruising Performance.After an
airplane has taken off and climbed to a reasonable altitude it is
usually flown at reduced throttle, both because it is easier on the
engine, which has a limited life, and because the fuel consumption
is lower and the range and endurance are greater.
A reasonable cross-country cruising speed for our VE-7, which,
with the 8.17-ft. propeller has a maximum speed of 128.5 m.p.h.
at 1,800 r.p.m., would be about 100 m.p.h. The power required
for horizontal flight at 100 m.p.h. is found from Fig. 103 to be
83.2 hp. Assuming that the propulsive efficiency is the same at
cruising speed as at high speed (the reason for this assumption
will appear later), the brake or engine power required is
83.2/0.824 = approximately 100 hp.
The revolution speed to which the engine is throttled to give
100 hp. at 100 m.p.h. can be found in the same manner as for
full-throttle flight at reduced speeds, i.e., by (1) assuming three
revolution speeds which cover the likely range, say 1,300, 1,400,
and 1,500; (2) computing the value of C, for each; (3) from the
curves of C, vs. Vn/D finding the value of V/nD corresponding
to each value of C,; (4) calculating the r.p.m. corresponding to
each V/nD; (5) plotting curves similar to those of Fig. 105 of
both the original and final values of r.p.m. vs. C; and (6)
finding the actual values of r.p.m. and C, at the intersection of
the two curves. The values for cruising at 100 m.p.h. with
100 hp. are r.p.m. = 1,420 and C, = 1.39. The efficiency for
C = 1.39 is 0.821, which checks the assumption of 0.824 closely
enough so that another trial would still give 1,420 r.p.m.
It will be noticed that the C, for cruising, 1.39, is nearly the
same as that for high speed, 1.42. This is also true of the values
of V/nD at high speed and cruising, for the drop in revolutions
is very nearly proportional to the drop in air speed. The most
efficient propeller at high speed is therefore also very nearly
the most efficient at cruising speeds. Actually, the most efficient
of our series of metal propellers for cruising at 100 m.p.h. with
the engine throttled to 100 hp. at 1,420 r.p.m. would have a
diameter of 8.20 instead of 8.17 ft., a blade angle at 0.7522 of
21.9 instead of 22 deg., and an efficiency about one-tenth per
cent greater, which is negligible.
At a cruising speed of 80 m.p.h., the 8.17-ft. propeller would be
throttled to 1,193 r.p.m., and its propulsive efficiency would be
0.818 as compared with 0.820 for the most efficient propeller
under the same conditions.
The revolutions of all of our five metal and three wood pro
pellers would be nearly enough the same at any cruising speed so
that the wear on the engine can be considered the same for all.
The more efficient ones would actually turn a little more slowly,
for they would require less engine power to give the same thrust
power. The fuel consumption is also less in about the proportion
that the efficiency is greater.
The VE-7 cruising at 100 m.p.h. requires approximately 100
hp. from the engine, and except for the limitations imposed by
the propeller, it would be possible to obtain this at any r.p.m.
above that at which the engine is held down to delivering 100 hp.
at full throttle, or in this case from 900 to 1,800 r.p.m., the latter
being the upper limit for safety. A curve of the specific fuel
consumption in pounds per horsepower per hour for our 200-
horsepower engine developing 100 hp. at various revolutions is
given in Fig. 111.1 The specific consumption with the 8.17-ft.
metal propeller throttled to 1,420 r.p.m. is 0.524 lb. per hp. per hr.
With a propeller which would absorb the 100 hp. at 1,050
r.p.m. or below, the fuel consumption would be reduced to 0.495
lb. per hp. per hr. There would be no advantage in going lower
1 This figure was drawn from data given in Cruising Performance of Air
planes, by E. E. Wilson and B. G. Leighton, Navy Dept. Bur. Aeronautics
Tech. Note 133, 1924.
than 1,050 r.p.m., for the fuel consumption would not be any less
and it would be harder on the engine.
The metal propeller of our family having the highest efficiency
while absorbing 100 hp. at 1,050 r.p.m. and a cruising speed of
100 m.p.h. would have a diameter of 9.65 ft., a blade angle at
0.75ft of 24 deg., and a propulsive efficiency on the VE-7 of
0.851 X 1.010 = 0.859. The actual fuel consumption at 100
m.p.h. with this propeller, considering its higher efficiency as well
as the lower specific consumption at 1,050 r.p.m., is 9.5 per cent
better than with the best cruising propeller throttled to 1,420
r.p.m.
OM
.2 x
tfc 0.5Z
E 9-
m 3= 0.51
gcd
- S 0.50
900 1000 1100 1200 1300 WOO 1500

R.P.M.
Fiq. 111.Fuel consumption of engine rated at 200 hp. at 1,800 r.p.m., when
developing 100 hp. at various revolutions.
The large propeller would, however, entail great disadvantages

which would practically prohibit its use. The high-speed full-
throttle r.p.m. would be only 1,150, so that only 131 hp. instead
of the full 200 would be taken from the engine, and the high speed
would consequently be only 111 m.p.h. instead of 128.5. For the
same reason the maximum rate of climb would be reduced to less
than half of its normal value. Also, it is probably harder on the
engine to develop 100 hp. at 1,050 than at 1,420 r.p.m., due to
the higher explosion pressures and temperatures.
Range and Endurance.Both the range and the endurance are
directly proportional to the propulsive efficiency and inversely
proportional to the specific fuel consumption. It is therefore
advantageous, of course, to have a high propulsive efficiency at
cruising speeds and a low fuel consumption, but the importance
of these may be outweighed by other factors where the maximum
range or endurance are desired without regard to the other per
formance. In these cases it is also advantageous to have a
propeller which will enable the airplane to take off with the
largest possible fuel load, and the best propeller is a compromise

which must be worked out for each individual case.
Performance at Altitude.The density of the air decreases
with altitude, as shown for standard atmosphere in Fig. 112,
where p0 is the density at sea level, or 0.002378 slug per cu. ft.
8,000 12,000 16,000 20,000 24,000 28,000

Altitude Above Sen Level in Ft.
Fig. 112.Variation of air density with altitude.
With an airplane flying horizontally at any angle of attack, the
following equation shows the relation of lift, density, and velocity:
L = MpVCJS,
= pV2 X constant.
Since the lift is equal to the weight, which remains constant
(fuel loss with time being neglected), the airplane must fly faster
in the lighter air encountered at altitudes above sea level, in
order to maintain level flight at the same angle of attack. Using
the subscript zero to denote sea level, the necessary speed at
altitude is given by the equation
y -
If the angle of attack does not change, the L/D remains the
same regardless of altitude, and since L remains constant, D does
also. The thrust horsepower required at the new altitude and
speed is then
DV
HPT =
550
D7o
where HPr^ is the thrust horsepower required at sea level.

The power required curve for our VE-7 at an altitude of 10,000

ft. has been calculated by means of the above relations and is
given in Fig. 113 along with the curve for sea level.
j.
r
i
/
nF
boi
.1
V
U
V
V -4
VyC. W*
TO SIS
Velocity, M.P.H.
Fig. 113.Power available and power required curves for sea level and 10,000
ft. with 8.17-ft. metal propeller.
1.0 1
'0 4,000 8,000 12,000 16,000 20,000 24,000 28,000

Altitude Above Sea Level in Ft.
Fig. 114.Variation of engine power with altitude, at constant r.p.m
The power available depends on the effect of altitude on both

the engine power and the propeller characteristics. As shown in
Chap. VI and also by the blade-element theory, the thrust,
torque, and power of a propeller vary directly with the density

of the air, but the efficiency is independent of the density, remain
ing the same for the same advance per revolution (V/nD) regard
less of altitude.
The power output of the engine also falls off with altitude, but
the decrease is greater than the decrease in density. Figure 1141
shows the decrease with altitude of the ratio HP/HP0, where HP
is the horsepower at any altitude and HP0 is that developed at
sea level at the same revolutions.
The curve of power available at 10,000 ft. with our 8.17-ft.
metal propeller is shown in Fig. 113 along with that for sea-level
and the power-required curves. The power-available curve at
altitude is found in the same manner as that for sea level except
that the lower density and power must be considered. The
formula for C, is
r ~
C' - \Pn2
"F~6-
It will be noticed that P and p are both in the expression to the

one-fifth power. The simplified formula for sea-level density
and in terms of engineering units was given in Chap. VI as
r = 0.638 X MPH
' ~ HPX BPM^
For altitudes this may be changed to
c = 0.638 X MPH
\P/Po/
where HP is the actual horsepower developed at altitude and
p/po is obtained from Fig. 112. The one-fifth power of the whole
HP
factor r- may be found from the HP scale in Fig. 53 (Chap. VI).
P/Po
The performance of the VE-7 with the 8.17-ft. propeller is
tabulated below for an altitude of 10,000 ft. and also for sea
level:
1 Figure 114 is based on. data from The Variation in Engine Power with
Altitude Determined from Measurements in Flight with a Hub Dynamom
eter, by W. D. Gove, N.A.C.A.T.B. 295, 1928; and The Determination of
the Horsepower Height Factor of Engines from the Results of Type Trials of
Aircraft, by J. D. Coales and A. L. Lingard, British R. and M. 1141, 1927.
10,000 ft. Sea level
Max. horizontal speed, m.p.h 122.8 128.5

R.p.m. at max. speed 1,748 1,800
Brake hp. at max. speed 137.4 200
1.40 1.42
0.822 0.824
Max. rate of climb, ft./min 613 1,150
Air speed for max. climb, m.p.h .... 78 75
The value of the C, for high speed is very nearly the same for
10,000 ft. as for sea level, from which it follows that the propeller
giving the greatest efficiency at the maximum sea-level speed
will also give approximately the highest efficiency at maximum
horizontal speed at 10,000 ft. Also, the maximum revolutions
at 10,000 ft. are only 3 per cent lower than at sea level, so that
very nearly all of the permissible engine power is used at 10,000 ft.
Our example therefore leads to the conclusion, which is fortu
nately true in practically all cases, that with an unsupercharged
engine the best propeller for a given purpose (high speed or
climb) at sea level is also the best for that purpose at altitudes.
The ceiling, or the altitude at which the rate of climb is zero, is
about 21,000 ft. for our VE-7 with the 8.17-ft. metal propeller.
Ordinarily, the propeller giving the greatest rate of climb at
sea level will give the best climb at all altitudes and also the
highest ceiling.
Performance at Altitude with Supercharged Engine.In
order to overcome the loss of power at high altitudes, engines
are sometimes fitted with superchargers, or blowers which supply
air to the carbureter at sea-level pressure, up to a certain "criti
cal" altitude. Above the critical altitude the power falls off in
about the usual manner.
We shall now consider the performance of the VE-7 on the
assumption that the engine is fitted with a supercharger having
a critical altitude of 20,000 ft. (p/p0 = 0.5327). We shall assume
for convenience that the sea-level power curve (Fig. 104) also
applies at 20,000 ft., although this is not strictly true, for the
actual sea-level conditions are not reproduced. At altitudes
lower than the critical, it is usually necessary to reduce the
amount of supercharging to avoid excessive cylinder pressures
and temperatures, so that approximately the same power is
obtained from sea level to the critical altitude. The effective

power is reduced somewhat by the power absorbed by the
supercharger, but we shall neglect this factor in our calculations.
With these assumptions, our 8.17-ft. metal propeller still gives
the best air performance at sea level. At 20,000 ft., however,
with full supercharging and full throttle, it would allow the
engine to turn up 2,350 r.p.m., and at this revolution speed the
engine would develop about 240 hp. If the motor could be
run under these conditions, which are of course far beyond the
specified safe limits, the airplane would attain a maximum hori
zontal speed of 168 m.p.h. at 20,000 ft. Actually, however,
the engine should be throttled to the specified safe limit of 1,800
r.p.m., and if this is done the performance is about the same
as if no supercharger were used. It is evident, therefore, that
our 8.17-ft. propeller is not suitable for use with a supercharged
engine at high altitudes.
The best high-speed propeller selected from our metal series
to absorb 200 hp. at 1,800 r.p.m. at 20,000-ft. altitude has a
diameter of 9.15 ft. and a blade angle at 0.75.R of 23.6 deg.
The propulsive efficiency and maximum horizontal speed at
20,000 ft. are 0.849 and 159.5 m.p.h., respectively. With this
propeller a fine performance is obtained at altitudes in the
neighborhood of 20,000 ft. and above, and a ceiling of 30,000 to
40,000 ft., depending somewhat on the properties of the super
charger and engine. At sea level, however, even assuming
that the engine develops a full 200 effective hp. at 1,800 r.p.m.,
the propeller holds the engine down to 1,320 r.p.m., at maximum
speed, the engine develops only 117 brake hp., and the maximum
horizontal speed is only 116.5 m.p.h. as compared with 128.5
with the unsupercharged engine. Also, the maximum rate
of climb at sea level with this propeller, neglecting the additional
weight due to the supercharger installation, is only 690 as
compared with 1,167 ft. per min. without a supercharger.
These examples serve to show that the same propeller is not
suitable for both high and low altitudes when a supercharged
engine is used. If a fixed propeller is used, the performance
must be sacrificed at one end or the other, or a compromise
of some kind must be made. Usually when a supercharger is
used the main interest is in the high-altitude performance, and
the performance at low altitude is sacrificed.
CHAPTER XI
THE VARIABLE-PITCH PROPELLER
Chapter X brought out the fact that no one propeller of

fixed shape and size gives the best results for all conditions of
flight with an airplane. If the propeller is designed for one
condition, say maximum speed at sea level, the full permissible
engine power and best propulsive efficiency are not obtained at the
other conditions such as take-off and climb. With a super
charged engine, the loss of performance at low altitudes is very
serious if the propeller is designed for a high altitude, and vice
versa.
Performance with Variable Propellers.Curve 1 of Fig. 115
shows the power available for our VE-7 at sea level, based on the
assumption that at each speed the full 200 hp. is absorbed at 1,800
r.p.m. by the most efficient metal propeller of our series for the
speed, power, and revolutions. This would require a propeller
whose size and pitch could be independently varied in flight.
No such propeller is available at the present time, but propellers
have been made and successfully tested1 in which the pitch can
be varied in flight at the will of the pilot or, in one case,
automatically. In all of these propellers the pitch is varied, as
in the case of the adjustable-blade metal propellers, by turning
the entire blade in the hub.
If the pitch of our 8.17-ft. metal propeller were varied in such
a manner that it would absorb the full 200 hp. at 1,800 r.p.m. at
all air speeds, the power available would be as shown by curve
2 of Fig. 115. Curve 3 shows the power available with the
8.17-ft. propeller when the blade angle is fixed at 22 deg. at 0.7572.
The difference between curves 2 and 3 shows that the power
available when the full allowable power is used throughout the
1 Although several designs of variable-pitch propellers have been success
fully flight tested, none has been used after the tests. Apparently the
extra weight and complications, including in most cases an added control
for the pilot to attend to, have overshadowed the advantages of the variable-
pitch feature in the propellers so far designed. The construction of a few
of these propellers is explained in Chap. XIV.
180
THE VARIABLE-PITCH PROPELLER 181
entire speed range is materially greater than the power available

with a fixed propeller except at maximum speed. A variable-
pitch propeller cannot, of course, increase the maximum speed
over that obtained with the best fixed high-speed propeller.
Above 80 m.p.h. it makes practically no difference whether the
pitch alone is varied or whether both the best pitch and diameter
180
160
p.Av iihb e
140 > &t //
a-
<n
Ct
120
V 10.0
fe ioo 9.5 c
* - V
-- 9.0 'is
o 80 Bes+nu 8.5 |
r
5.0 S
60 22"
PA .r-
.' " tl 20 *
B ,. \is; ade
40 "Ar sr/e 18" 5
V- 16. %>CD
to 14
B 12 St
10
60 TO 9060 100 110 120 110
Velocity, M.P.H.
Fm. 115.Power available with variable pitch and diameter propellers.
are used at each speed to give the highest efficiency obtainable

with the particular series of propellers used. At maximum speed
the best diameter is 8.17 ft., and there is no difference between
curves 1 and 2. Below 80 m.p.h., however, it is markedly
advantageous to have a diameter greater than 8.17 ft.
Curves 4 and 6 of Fig. 115 show the variation in diameter and
blade angle required to give the highest efficiency at each speed
while absorbing the full 200 hp. at 1,800 r.p.m., resulting in
power-available curve 1. Curve 5 shows the variation in blade
angle of the 8.17-ft. variable-pitch propeller to absorb 200 hp.

at 1,800 r.p.m. at each speed and give power-available curve 2.
Curves of the rate of climb at sea level vs. airspeed are given
for all three propellers in Fig. 116. The climb with the 8.17-ft.
variable-pitch propeller is practically as high as that with the
hypothetical variable-pitch and variable-diameter propeller
except at low speeds. It is decidedly better than that with the
fixed propeller at all speeds below maximum, and the maximum
rate of climb is 193 ft. per min. or 16.5 per cent better.
1600
1400
1200
.| 1000
a;i.
a.
*!.
Sl. 800
E
G15 600
400
200
0
60 TO 80 90 100 110 120 110
Velocity, M.P.H.
Fig. 116.Rates of climb with variable pitch and diameter propellers.
This difference is appreciable with the VE-7, but with the

heavily loaded transport airplane mentioned in Chap. X having
a rate of climb of 400 ft. per min. with 600 hp. and a total weight
of 9,000 lb., the same relative increase in power available with a
variable-pitch propeller would result in an increase of 145 ft.
per min. or 36.2 per cent.
As shown by Fig. 116, a propeller of larger diameter than that
which is best for high speed would increase the rate of climb still
more, and, what is also important, the maximum rate of climb

would occur at a lower forward speed, resulting in greater angle
of climb.
The propulsive efficiency with the three propellers (1) the
hypothetical variable-pitch and variable-diameter propeller;
(2) the 8.17-ft. variable-pitch propeller; and (3) the fixed 8.17-ft.
propelleris shown as solid lines on Fig. 117. It is noteworthy
that there is very little difference between the efficiency curves
of the three propellers, except at very low speeds where the
variable-pitch and -diameter propeller is noticeably better than
the other two.
^. - -
Ideal Efficient
3
5>
<s'Z>
Actual P -opulsive
Efflci ?ncy
Noi'AT1 ON
1. Var able Pitch and Variat le Diamete r
2. Var 'able Pitd 1
3. Fix 'J Pitch
60 70 80 90 100 no 120 130
Velocity, M. P. H.
Fig. 117. Comparison of actual and ideal efficiencies.
It is interesting, also, that from the speed for maximum climb

(75 m.p.h.) to the maximum speed the fixed propeller has the
highest efficiency of the three.
One of the advantages often claimed for variable-pitch pro
pellers is that at climbing speeds, where the V/nD is low, the
blade angles of the variable-pitch propeller are reduced and the
sections therefore work at more nearly their best L/D than those
of fixed -pitch propellers which have necessarily been designed
with the maximum speed in mind also. Since the propeller
sections operate at more nearly the .angle of attack for maximum
L/D, it is concluded that the propeller efficiency is higher. The
same conclusion may easily be drawn from an incautious study
of a series of efficiency curves plotted on a basis of V/nD for a
number of propellers varying in pitch (see Fig. 60), for at the
lower values of V/nD, the low-pitch are more efficient than the
high-pitch propellers. Ordinarily, however, as in the case of our
example (illustrated by curves 2 and 3 in Fig. 117), the opposite
is truethe efficiency of the variable-pitch propeller is lower than
that of the fixed-pitch propeller throughout practically the entire
flying range. In our case the efficiency of the fixed propeller is
even higher than that of the variable-pitch and -diameter pro
peller, between the maximum speed and the speed for maximum
climb.
Two useful conclusions may be drawn from the above dis
cussion and a study of Fig. 117:
1. Any gain in climbing performance at sea level with a
variable-pitch propeller will usually be entirely due to the use
of greater engine power than can be obtained with a fixed pro
peller designed to hold the engine down to the specified maximum
revolutions at high speed; also, the variable-pitch propeller will
usually be slightly handicapped by a slightly lower efficiency;
2. While propeller characteristics plotted on a basis of V/nD
are satisfactory for showing the comparative values of the
characteristics themselves, the V/nD basis is unsatisfactory
and in many cases misleading when used for showing the relative
performances of propellers on aircraft.
It is interesting to compare the ideal efficiencies of the three
propellers according to the momentum theory. This may be
obtained from the expression (explained in Chaps. II and
VII)
w(V/nD)3 _ ttC.5
2CP 2{V/nDY
The curves of ideal efficiency for the three propellers are shown
as dotted lines in Fig. 117. It will be noted that the ideal effi
ciency of the fixed propeller, curve 3, is also greater than that
of the variable-pitch propeller, curve 2. This is due to the
lower V/nD and the greater power input with the variable-
pitch propeller. With the best possible propeller of the metal
series, however, in which the diameter is varied as well as the
pitch, the ideal efficiency is higher than for either of the other
two.
Effect on Take-off.In order to see the effect of the variable-
pitch propellers on the take-off run, we shall find the best of our
metal propellers for producing static thrust while absorbing the
full 200 hp. at 1,800 r.p.m. To do this it is convenient to make

use of the relation
TD = 2tCt
Q ' C, '
which is plotted for the various propellers of our series at zero
advance in Fig. 118. The diameter corresponding to each
angle setting or pitch, as obtained from the relation,
P 200 X 550
CP =
pn'D6 0.002378 X (30)3Z>6
T
1400
1
.
.T-i faiic Thn D
12 1200
\ 10 1000
D''
8 800
CO
6 600
4 400
2 200
Blade Angle ai 0.75 R
Fig. US.Static thrust for various blade-angle settings.
is also given in Fig. 118. The torque of our 200-hp. engine at
1,800 r.p.m. is 584 lb.-ft. Then, since for any blade angle D, Q,
and TD/Q are known, it is easy to solve for T. The static
thrust obtained in this manner is also plotted in Fig. 118. The
maximum static thrust is obtained with a 9.5-ft. propeller set
to 10.4 deg. at 0.75^.' This maximum static thrust is 1,275
lb. as compared with 950 lb. for the 8.17-ft. variable-pitch
propeller set to 19.4 deg. (from Fig. 118) and 731 lb. for the
1 The fact that the static thrust decreases with diameters greater than
9.5 ft. and blade settings lower than 10.4 deg. is partly due to the poor pitch
distribution of these propellers at very low pitches. With propellers having
uniform pitc'h the diameter for the best static thrust would be somewhat
larger.
8.17-ft. propeller with its pitch angle fixed at 22 deg. at 0.75R.

The thrusts at 0, 30, and 60 m.p.h. and the corresponding take-off
runs are tabulated below for the three propellers:
Thrust Thrust Aver Approxi

Static
at 30 at 60 age mate
thrust, m.p.h., m.p.h.. thrust,
lb. take-off
lb. lb. lb. run, ft.
Prop, having best diameter &

1,275 1,000 779 1,018 250
Variable-pitch prop., 8.17-ft.
dia 950 840 738 842 302
Fixed pitch prop., 8.17-ft. dia. . 731 713 656 700 364
The take-off run with the 8.17-ft. variable-pitch propeller

is about the same as with the best fixed-pitch propeller for
take-off (Chap. X). The fixed-pitch propeller, however, has a
very poor performance in climb, high speed, and cruising.
With our hypothetical variable-pitch and -diameter propeller,
for which the diameter is 9.5 ft. for 0 and 30 m.p.h. and 9.29 ft.
for 60 m.p.h., the take-off run is reduced to approximately 250 ft.
Effect on Cruising Economy.We found in the section on
cruising in Chap. X that the most economical propeller for
cruising at 100 m.p.h. would hold the engine down to 1,050
r.p.m. while absorbing approximately 100 hp. The diameter
of this propeller is 9.65 ft. and the blade angle at 0.75.R is 24 deg.
If the pitch of the 8.17-ft. propeller is increased to absorb 100 hp.
at 1,050 r.p.m., a blade angle of about 33 deg. is required, an
efficiency of only 0.77 is obtained as compared with 0.851 with
the 9.65-ft. propeller, and the fuel consumption is actually
greater than with the 8.17-ft. fixed-pitch propeller with the
engine throttled to 1,420 r.p.m.
The best fuel consumption with the 8.17-ft. propeller is
obtained at 1,270 r.p.m. (blade setting at 0.75/2 = 26 deg., rj =
0.829, specific fuel consumption = 0.507), in which case the fuel
consumption is 5 per cent higher than with the best propeller
at 1,050 r.p.m. and 3.3 per cent lower than with the fixed pro
peller at 1,420 r.p.m.
Summary of the Effect of a Variable-pitch Propeller on Per
formance with an Unsupercharged Engine.With an unsuper-
charged engine, the increase in performance at altitudes with a
variable-pitch propeller is very much the same as that at sea

level. A slight increase in high speed over that with a fixed
propeller is also obtained at altitudes, however, due to the fact
that the power corresponding to the maximum permissible
revolutions can be taken from the engine.
The final performances obtained with our three propellers are
compared in the first three columns of the following table :
8.17-ft. 8.17-ft. Variable Best 9.00-ft.

fixed variable pitch & diam variable
pitch pitch diameter eter, ft. pitch
Max. speed, m.p.h 128.5 128.5 128.5 8.17 127.8

Max. rate of climb, ft./min. . . 1,167 1,360 1,387 9.11 1,380
Take-off run, ft 364 302 250 9.5 255
Relative effectiveness, cruising
at 100 m.p.h., based on best . 0.92 0.95 1.00 9.65 0.99
It is evident that even with a variable-pitch propeller the best

possible performance in all conditions of flight cannot be obtained
with a single propeller of constant diameter, and the propeller
must at best be a compromise. If it is desired to obtain a high
maximum speed, the 8.17-ft. propeller is the best. If any of
the other items is more important, a propeller of larger diameter
is required for the best performance.
If a compromise propeller 9 ft. in diameter is used, the per
formance is given by the figures in the last column of the table
above. The high speed, climb, and economy in cruising at
100 m.p.h. with this 9-ft. variable-pitch propeller are all within
1 per cent of the maximum obtainable with a variable-pitch and
-diameter propeller, and the take-off run is only 2 per cent
longer.
The performance with this 9-ft. variable-pitch propeller
(and 9 ft. may be taken as about the best compromise diameter)
is appreciably better than the performance with a fixed propeller
in climb, take-off, and cruising. It is not, however, sufficiently
better to warrant a large increase in weight (which has been
neglected in our computations) and in cost and complication.
Performance with Variable-pitch Propeller and Supercharged
Engine.It was shown in Chap. X that a fixed propeller was not
suitable for both high and low altitudes when used with a super
charged engine. If reasonable advantage of the supercharger
is to be obtained it is necessary to use a propeller having change
able characteristics.
We found in Chap. X that the best high-speed propeller for the
supercharged VE-7 at the critical altitude of 20,000 ft. had a
diameter of 9.15 ft. and a blade angle at 0.75R of 23.6 deg. This
diameter is nearly the same as that of our 9-ft. compromise
variable-pitch propeller for all-around sea-level performance.
As a matter of fact, a variable-pitch propeller having a diameter
of either 9 or 9.15 ft. would make a satisfactory all-around pro
peller for all altitudes. If the climb at 20,000 ft. and above were
important, it would be advisable to use a slightly larger diameter,
say 9.5 ft., sacrificing about 2 m.p.h. more in maximum speed at
sea level.
The supercharger is very limited in usefulness with a fixed
propeller, for the performance is either mediocre at all altitudes
or very poor at high or low altitudes, but the variable-pitch
propeller and the supercharger make a happy combination for
high-altitude flying without sacrificing low-altitude performance.
The supercharger enables the engine to develop full power at
altitudes, and the variable-pitch propeller makes it possible to
use the full power under all flying conditions. It must not be
forgotten, however, that they add both weight and complication
and also extra duties for the pilot. Since the pilot already has
much to absorb his attention, it is highly desirable that the con
trol of the variable-pitch propeller be made automatic but subject
to control by the pilot if necessary or desired.
Other Uses for Variable-pitch Propellers.Some variable-
pitch propellers are constructed in such a manner that the pitch
can be reversed for the purpose of giving a braking action or
backward thrust. This is particularly advantageous in the
maneuvering of lighter-than-air craft such as dirigibles while
landing or taking off.
The effectiveness of the reversible-pitch propeller as an air
brake to shorten the landing distance of an airplane has been experi
mented with by both the Army and the Navy in this country.
There is no doubt that the reverse thrust substantially reduces
both the gliding angle and the distance run on the ground before
stopping. Difficulty was experienced, however, with the effect
on the controllability, both in the air and on the ground, where it
was very difficult to hold the airplane straight and keep it from
ground looping. These tests were made several years ago, and
1 it is likely that the ground looping could be overcome by the use
of modern individually operating wheel brakes.
With reversible propellers, it is advisable to have the engine
throttle connected with the pitch-changing mechanism in such
a manner that the throttle is closed while the propeller blades are
passing through zero pitch. If this is not done and the throttle
is accidentally opened when the pitch is at or near zero, the engine
will race and may destroy itself.
Often dirigible airships, and sometimes multi-engined air
planes, cruise with one or more engines cut out. Sometimes the
engine and propeller stop, and sometimes the propeller acts as a
windmill with sufficient torque to keep the engine rotating. In
either case, and particularly in the latter, the drag is high and
could be reduced appreciably if the propeller blades could be
turned fore and aft so that their average pitch was infinite.
With multi-engined airplanes having fixed propellers, if one
engine fails the speed of the airplane is reduced and the other
propellers do not allow their engines to turn at full revolutions and
deliver full power. The condition is similar to that of a normal
airplane in climb, and variable-pitch propellers would improve
the performance in the same way as in normal climb.
CHAPTER XII
THE GEARING OF PROPELLERS
Due to the fact that the power output of gasoline engines is

directly proportional to the revolutions as well as the mean
effective pressure, one obvious way to cut down the weight per
horsepower of the engine is to increase the revolution speed.
Unfortunately this reduces the value of C, at which the propeller
operates, and the propeller has a lower pitch and a lower maxi
mum efficiency. The knowledge of this fact has been influential
in more or less limiting the maximum revolutions of aircraft
engines, but in spite of it there has been a gradual tendency
toward higher maximum engine revolutions.
There is also a tendency toward engines delivering more and
more power, and this too decreases the value of C, and therefore
the maximum possible propeller efficiency. In addition, the
combined high powers and revolutions are accompanied by high
tip speeds and the consequent efficiency losses which may be
very great. Due to the losses accompanying one or more of these
items it has in some cases been found advisable to mount the
propeller on a separate shaft which is geared to turn more slowly
than the engine crankshaft. This, of course, increases the weight,
the complication, and the cost.
Factors Affected by Reduction Gearing.The question of the
desirability of reduction gearing is an involved one because of
the many factors affected. The advantages and disadvantages
which must be balanced against each other may be summarized
briefly as follows:
Advantages:
1. Increase in propulsive efficiency and consequently in airplane per
formance, due to
o. Greater propeller pitch ratio.
6. Larger size of propeller with respect to body.
c. Elimination of tip-speed loss, if any.
2. Less propeller noise due to lower tip speed.
Disadvantages :
1. Loss of power in gears.
190
THE GEARING OF PROPELLERS 191
2. Extra weight of propeller, gearing, and structure.
3. Possible effect of greater propeller diameter on landing gear height,
engine location, etc.
4. Greater number of moving parts requiring lubrication and service.
5. Greater cost.
The problem is further complicated by the effect of the gain in
propulsive efficiency on the various items of airplane performance,
such as high speed, cruising speed, take-off, climb, and cruising
economy. Due to the number of factors involved, no simple
general rule can be made to cover all cases, but each case should
be worked out individually and in detail. Two examples follow
which show the nature of the results to be expected in rather
extreme cases, as well as a method in which the problem can be
attacked for any given case by airplane, engine, or propeller
designers.
Example 1.For our first example we shall take the case of an
express or mail transport airplane powered with a single engine
delivering 800 hp. at 2,000 r.p.m.1 We shall assume that our
engine is water cooled with a nose radiator and that with a direct-
drive propeller the mutual propeller-body interference is the same
as with the VE-7. The gross weight will be 12,000 lb., with a
wing loading of 15 lb. per sq. ft. and a power loading of 15 lb.
per hp. With a direct-drive propeller the maximum speed at
sea level will be 130 m.p.h. and the stalling speed 65 m.p.h.
The power-required curve is shown in Fig. 119.
The propeller characteristics and the resultant power-available
curves are calculated for the direct-drive condition and also
for gear ratios of 5:4, 5:3, and 5:2, the corresponding maximum
propeller revolutions being 2,000, 1,600, 1,200, and 800. The
computations are made in the same manner as those in Chap. X,
based on the full-scale metal-propeller test data given in Figs.
51, 52, and 167. The form of these propellers is shown in Figs.
176 and 177. Owing to the high power and revolutions, the
direct-drive propeller will have an efficiency loss due to the
high tip speed which is inevitable at full throttle and maximum
speed. The propeller diameter for the direct-drive installation
is therefore selected as that which just escapes a tip-speed loss
when cruising at 1,750 r.p.m. It then also escapes this loss in
1 There is now one water-cooled engine, the Packard 1 A-2500, which is
manufactured to this particular rating, and it is likely that there will be both
air- and liquid-cooled engines of about the same power in the near future.
the take-off condition and in climbing at very low air speeds,

all of which are more important in this type of airplane than the
maximum speed. A compromise diameter of 11 ft. gives about
the best results.
The three geared propellers are selected, considering take-off,
climb, and high cruising speed, so as to have diameters halfway
between those for the best high-speed propeller and the propeller
operating at the peak of its efficiency curve at maximum speed.
800i 1 1 1 1 1 1 1 1 1
100
0 II I I
60 70 60 90 100 110 120 130 140 150
Veloci+^M.P.H.
Fig. 119.-Power available with various gear ratios. Transport airplane.
A 2 per cent loss in power in the gears is assumed for all ratios.
This value is not the result of accurate measurements, which are
difficult to make, but seems reasonable from data on gears
and also from the comparative measured power outputs of direct-
drive and geared engines of the same design.
In regard to body interference, it is assumed that the data
from the VE--7 tests are correct for the direct-drive propeller
and that for the geared propellers the propulsive efficiency is
1 per cent higher for each 8 per cent increase in propeller diameter
(see Chap. IX).
The main characteristics of the four propellers are tabulated

below, and the power-available curves obtained with them,
assuming that the full-throttle engine torque is constant within
the range of revolutions required, are given in Fig. 119.
Gear ratio Direct 5:4 5:3 5:2

Max. r.p.m 2,000 1,600 1,200 800
C 1.040 1.183 1.408 1.670
V/nD 0.520 0.589 0.728 0.922
Diameter, ft 11.00 12.68 14.42 17.31
Po.76, deg 15.9 16.2 19.7 24.7
ij, with VE-7 0.756 0.784 0.828 0.855
Tip speed, ft./sec 1,168 1,080 930 770
0.873 0.954
Body interference factor 1.016 1.030 1.045
0.660 0.760 0.853 0.893
Thrust hp. at max. speed . . 528 596 669 700
Max. speed, m.p.h 130.0 136.3 142.5 145.0
The effect of gearing on the maximum speed is surprisingly

great, it being 15 m.p.h. higher with the 5 : 2 gear ratio than with
direct drive. Nearly half of this effect is due to the large tip-
speed loss with the direct-drive propeller, and it will be noticed
that there is a substantial tip-speed loss even with the 5 : 4 reduc
tion gear.
The variation of propeller diameter and weight with gear
ratio is shown in Fig. 120. Regarding the limitations on pro
peller diameter with our transport airplane, the landing gear
would ordinarily be sufficiently high to give the required 9 in.
of ground clearance in flying position1 with a 14-ft. propeller
diameter, this landing gear height being necessary to give the
wings the proper angle of attack in landing. If the regular
two-blade propellers of our calculations are used, it will be
necessary to increase the landing-gear height for gear ratios
greater than about 5:3. For the 5:2 ratio this increase would
be about 20 in. The drag and weight would also be increased
somewhat, but this factor we shall neglect. If the gearing
happened to be of the spur type with the propeller axis above the
crankshaft axis, a larger diameter, probably at least 15 ft.,
could be used without increasing the landing-gear height.
With our transport airplane the effect of the gearing on cruising
speed is of more interest than the effect on maximum speed
1 This is the present Department of Commerce requirement.
Assuming that the cruising is done at two-thirds full power,

the engine revolutions will be 1,750, and there will be no appreci
able tip-speed loss even with the direct-drive propeller. The
cruising speeds obtained with the engine turning 1,700 r.p.m.
are plotted along with the maximum speeds for the various
gear ratios in Fig. 121. Without the tip-speed losses in the
direct drive, the gain with reduction gearing is not so great in
cruising speed as in high speed, but it is nevertheless quite
Gear Ratio
. Direct 5J4 5^1 5:2
substantial, being nearly 8 m.p.h. with the 5:2 ratio. The

cruising economy and range are also better at the higher cruising
speed by the percentage increase in speed, and they are much
better at the same cruising speed.
The stalling speed is of course affected by the extra weight with
gearing and a larger propeller diameter. The propeller diameters
and weights with the various gear ratios are given in Fig. 120.
For our calculations of stalling speed, climb, and take-off it is
assumed that the extra engine weight due to the gearing is about
130 lb. for all ratios. With this and the added weight due to
the larger propeller, and also a reasonable increase in the weight
of the structure supporting the engine and propeller, the total
increased weight is about 200 lb. for the 5:4 ratio, 300 lb. for the
5:3 ratio, and 400 lb. for the 5: 2 ratio. The resultant stalling
speeds given in Fig. 121 show an increase of about 1 m.p.h.
with the 5 : 2 ratio.
Gear Ratio
150
|
140 136 Ma t/m 'ipeed
110 a
if3.9.
121 s
. 120 11 '.Bz Cnnsir gStjeec
x m0
al
* 110
.au
u 100
'^
90
60
70 Sfa lint Sp *ed

65 q 65 S i.. 65 S S
SO
Z000 1500 1000
Propeller R.P.M.
Fia. 121.-Variation of maximum speed, cruising speed, and stalling speed with
gear ratio. Transport airplane.
Gear Ratio
800 Direct 5:4 5:3 5:2
93
72
. 700 6
J3 6
E
5 600
a
2
500!
2000 1500 1000
Propeller R.P.M.
Fig. 122.-Variation of maximum rate of climb with gear ratio. Transport
airplane.
It is apparent from the power-available curves of Fig. 119

that the rate of climb is noticeably greater with the larger gear
reductions. The maximum rates of climb obtained with the
various reduction ratios are given in Fig. 122, the extra weight
with gearing being allowed for. The maximum rate of climb
with the 5 : 2 ratio is 180 ft. per min. or approximately 30 per cent
greater than that with the direct drive, and apparently with
greater gear reduction the climb would increase at an even
greater rate.
In order to calculate the relative take-off distance obtained
with the different gear ratios, the static thrust is obtained by-
means of the curve of TD/Q given in Fig. 118. As is apparent
from this curve, the static thrust is very low for the highest-pitch
Sear Ratio
900 B7C
81 '0
c 800 7<12--'
a 75 7
? 700
r*
600
2000 1500 1000
Propeller R.P.M.
Fig. 123.-Effect of various gear reductions on take-off run. Transport
airplane.
propellers, and this tends to limit the gear reductions which can
be satisfactorily used. Assuming a take-off run of 800 ft. with
the direct-drive propeller, the lengths of run required with the
various gear reductions are calculated in the manner shown in
Chap. X and are given in Fig. 123. The shortest take-off run
is obtained with a propeller having a maximum revolution speed
of about 1,500, and with less than 1,100 r.p.m. the take-off is
poorer than with direct drive.
An examination of Figs. 120 to 123 shows that for the case
of our example, gearing the propeller increases the performance
to an extent which would more than justify the additional cost
and weight. Just what gear ratio is best is difficult to say, but it
would probably be between 5:3 and 5:2. The 5:3 ratio escapes
all tip-speed loss and gives an increase in all of the performance
items considered excepting stalling speed. The 5:2 ratio gives
slightly larger gains in most items but gives a loss in take-off,
and with greater reductions this loss would become prohibitive.
Also, the propeller size has approached, if not exceeded, the
practical limit. Everything considered, the ratio of 2 : 1 giving

the propeller 1,000 r.p.m. is probably the maximum gear reduc
tion advisable in this case.
With a reduction of 2:1, the propeller diameter is 15.6 ft.,
and an increase in landing-gear height of about 9 in. is necessary.
Neglecting the effect of this,
a. The high speed is increased 14 m.p.h.
b. The cruising speed is increased 7 m.p.h.
c. The maximum rate of climb is increased 140 ft. per min. or
23 per cent.
Against these gains we have the very slight disadvantages that
a. The take-off distance is increased 2.5 per cent.
b. The stalling speed is 1 m.p.h. higher.
In case the diameter of the two-blade propeller is considered
too great, a three-blade propeller could be used. This would
have a diameter of about 14 ft. and the original landing gear
could be used unchanged, but the weight would be approximately
70 lb. more than with the 15.6-ft. two-blade propeller. The
propulsive efficiency would be only about 2 per cent less at
maximum speed, and the all-around performance only a trifle
less than with the two-blade propeller.
The three-blade feature has one distinct advantage where
very large propellers are used. It is noticeably smoother running
in turns and in yawing flight. This is due to the fact that in
turns the gyroscopic moments are balanced and uniform with
three-blade but not with two-blade propellers. Also, both in
turns and in yawing or slipping flight the air loads on the blades
vary during each revolution, and with two-blade propellers one
blade has its greatest loading at the same time that the other
has its lightest, while with three-blade propellers as each blade
gets its extreme loading it is balanced by two others with medium
loadings. Unless large, slow-turning two-blade propellers are
mounted on very rigid structures, the vibrations they cause in
sharp turns may be severe enough to make the use of three or
more blades advisable.
If the above example were a flying boat instead of a land
airplane, almost all of the performance items excepting take-off
would be affected in the same manner by gearing. It is likely
that there would be no difficulty in getting the seaplane above
the "hump" speed with any of our gear ratios. Assuming
this to be true, the best and most certain take-off would be
given by the combination having the greatest thrust at the take

off speed, or about 65 to 70 m.p.h. As is evident from Fig. 119,
this is the propeller having the largest diameter and the greatest
gear reduction.
Gear Ratio
13
1800 1500 1200 900

Propeller R.P.M.
Fig. 124.- -Variation of propeller diameter with gear ratio. VE-7 example.
Example 2.The foregoing example represents a rather
extreme case where gearing is not only advisable but practically
necessary. For another example we shall take a smaller and
lighter airplane, our VE-7 airplane dealt with in Chap. X, having
Gear Ra+io
140
130 W ax'in Sae( d Aj .5

123-4 129.1 1U .4
Q-
Z 120
.o" 1t -*-\~116.1 117.1 //7.9
'. Cruisinq Spe ed ati
w 110
u 11 001znqi ne 1 4.
<
100
90
Propeller R.P.M.
Fig. 125.-Effect of various gear ratios on maximum speed and on cruising
speed. VE-7 example.
an engine delivering 200 hp. at 1,800 r.p.m. We shall take the
8.53-ft. direct-drive propeller of Chap. X, having a blade-angle
setting of 20.0 deg. at 0.75/2, and select the geared propellers
to operate at the same relative portion of their efficiency curves.
The propeller diameters with various gear reductions are shown

in Fig. 124. The original VE-7 will have proper clearance for
only a 9-ft. propeller, and the landing-gear height must be
increased if any reasonable gear ratio is to be used.
Gear Ratio
1300 Direct 6:5 3^2 2:i
12 a-
. 1200
Hi 11
E
U 1100
Ka
1000
1800 1500 1200 900
Propeller R.P. M.
Fig. 126.Effect of gear reductions on maximum climb. VE-7 example.
The high speed, cruising speed, climb, and take-off distance,

calculated in the same manner as for the previous example, are
shown for various gear ratios in Figs. 125, 126, and 127. With
the VE-7, which with direct drive has no high tip-speed loss but
has a reasonably high value of C (Ca = 1.42 as compared with
Gear Ratio
.600, Direct 6:5 3.12 211
,.
t
400 98
6
1800 1500 1200 900
Propeller R.P.M.
Fig. 127.Effect of gear reductions on take-off distance. VE-7 example.
1.04 for the transport example), the gains in both maximum speed
and cruising speed are very slight even with large gear reductions.
They would be even less than shown if the necessary increased
landing-gear height were considered.
The maximum climb is less with small gear reductions than
with direct drive, due to the power loss in the gears and the extra
weight. With this example there is no appreciable gain in climb

with any reasonable gear reduction.
The take-off run is longer with any gear reduction than with
direct drive, and it becomes worse as the reduction ratio is
increased. This is due to the fact that even the direct-drive
propeller has a reasonably high pitch and the higher-pitched
geared propellers have greatly reduced static thrust.
Obviously, with no appreciable increase in any performance,
and with a definite loss in take-off and possibly in climb, gear
reduction of any amount is not desirable for the VE-7 example,
even neglecting such items as cost.
General.As shown by these examples, the problem of whether
or not gearing is desirable, and if so what reduction ratio will
give the best results, is a complicated one because of the many
factors involved. Obviously in cases where the maximum speed
is important, such as in racing airplanes, it is important that there
be no appreciable loss due to too high tip speed, and gearing can be
used to eliminate this loss if it exists with direct drive. Also,
if the efficiency with direct drive is low, due to a combination of
high power and high revolutions, gearing will probably help.
Since increasing either the horsepower or the revolutions
decreases the value of C and since the value of C, is directly
connected with the maximum efficiency obtainable, a value of
C, may be used as a criterion for the desirability of gearing.
In general it may be said that if with direct drive the value of
C, is higher than about 1.3, and if the tip speed is under the
critical, gearing is undesirable. If, however, the value of C,
is lower than 1.3 or if there is a tip-speed loss, gearing may sub
stantially improve the performance. In order to find the best
gear ratio and whether or not the increase in performance is
worth the cost, performance computations considering all of
the factors affected should be made, after the manner of the
examples given.
CHAPTER XIII
TANDEM PROPELLERS
In some multiple-engined airplanes, two engines, each driving

its own propeller, are placed in tandem. The rear propeller,
which is a pusher, operates in the slipstream of the front pro
peller, which is a tractor.
Since the propellers are usually spaced about one diameter
apart, the front propeller is not influenced to an appreciable
extent by the action of the rear one, and it may be considered
as if it were working alone. The rear propeller, however, works
in a twisting slipstream which has been given both an additional
axial velocity and a rotational velocity by the front propeller.
If both propellers absorb the same power and turn in opposite
directions, the rear propeller will produce a rotation which is
equal and opposite to that produced by the front propeller,
and there will be no general rotation in the final slipstream back
of the rear propeller. This condition is obtained with the usual
tandem arrangement in which two similar engines are used end
to end, one pointing forward and one to the rear. This tandem
arrangement, therefore, does not suffer the usual loss of efficiency
due to the rotation in the slipstream.
Calculation of Slipstream Velocities in which Rear Propeller
Operates.The velocity in the slipstream of the front propeller
(neglecting the rear propeller) can be found by applying the
momentum theory to one annulus at a time. Thus:
= 2irrdrV(l + a)P2aV.
The symbols are the same as in the momentum and blade-element
theories (Chaps. II and V), and from any of the forms of the
blade-element theory, dT for an annulus also has the value
dT = MpV'BT. dr,
where Tc is calculated from the air forces on a blade element.
Equating both expressions for dT,
2rrdrV{l + a)P2aV = MPV2BTC dr
201
from which, if Vi is the axial component of the full slipstream

velocity,
and
The rotational component of the slipstream velocity is found

in like manner from the equation for angular momentum,
= 2irrdrV(l + a)P2wr,
where w = a'(2wrn) is the interference tangential velocity at the
plane of the front propeller, as in the induction or vortex theory.
From the blade-element theory dQ also has the value,
dQ = HpV2BQc dr,
and equating both expressions for dQ,
2irrdrV(l + a)P2wr = xApV2BQc dr
from which the tangential velocity in the final slipstream is
2w = 2a'(2irrn) = . ,
_ VBQC
Calculation of the Air Forces on the Elements of the Rear

Propeller.The regular blade-element theory equations for the
thrust, torque, and efficiency of the elements can be applied to
the rear propeller of the tandem arrangement by substituting
Vx= 7(1 + 2a) in place of V and (2wrn)i = (2vrn) + 2a'(2xrn)
in place of the ordinary tangential velocity 2irrn.
It is interesting to examine the difference in the expressions
for the efficiency of an element of the front and rear propellers
according to the vortex blade-element theory. The efficiency of
an element of the front propeller, which is the same as for an
isolated propeller, is
= dTV
v dQ(2irrn)
_ (1 a') tan <t>0
(1 + a) tan Oo + To)
TANDEM PROPELLERS 203
The useful work performed per second by an element of the

rear propeller is the thrust times the velocity of advance of the
airplane, the latter being less than the velocity of the slipstream
in which the rear propeller works. With the subscript 1 pertain
ing to the rear propeller, the efficiency of an element of the rear
propeller, when the propellers are turning in opposite directions, is
= dTlV
v dQl(2xm)
= (1 + 200(1 - a'i) tan 0ol
(1 + 2a) (1 + aO tan (<*>o, + 70l)'
Thus the efficiency of the element of the rear propeller is increased
by the rotational velocity factor 2a' and decreased by the axial
velocity factor 2a. Ordinarily, and especially in take-off or
climbing flight, the value of a is much larger than that of a', and
the net result is a loss in the efficiency of the rear propeller.
Since the area of the slipstream is not constant but becomes
smaller back of the propeller, there is some radial flow and the
air passing the front propeller at a given radius may (and usually
does) pass the rear propeller at a different radius. This com
plicates the use of the blade-element theory for tandem pro
pellers, but with an ordinary body housing the two engines
between the propellers, the slipstream is spread so that it is not
far wrong for the main working portions of the blades to assume
that the same air passes both propellers at the same radius.
Model Tests with Tandem Propellers.Only a few isolated
tests1 have been made with tandem propellers, and the only
systematic investigation of sufficient size to throw much light on
the subject was made by Eiffel in 1919. 2 In his investigation,
Eiffel tested two main series of tandem propellers, one in which
both propellers were of the same diameter and form, but of
opposite hand, and one in which the rear propellers were 12.5
per cent smaller in diameter. Both series were made up of
propellers of several pitch ratios, . and the blades of the rear
propeller were adjustable as to pitch. The pitch of the rear
propeller was in every case adjusted so as to make it absorb the
1 Some Experiments with Tandem Combinations of Airscrews, by A.
Fage and H. E. Collins, British R. and M. 421, 1918. Some Further
Experiments on Tandem Airscrews, by A. Fage and H. E. Collins, British
R. and M. 605, 1918.
2 Tandem Propellers, by G. Eiffel, published in the French magazine
VASrophile, November, 1919.
same power at the same revolutions as the front propeller, at the

speed for the maximum efficiency of the front propeller.
The pitch required by the rear propeller of like diameter to
absorb the same power as the front propeller is shown for various
pitch ratios by the dotted line in Fig. 128. It is interesting that
with propellers of the form tested having pitch ratios higher than
1.0, the rear propeller requires less pitch than the front one, show
ing that at high rates of advance per revolution the rotational
1.4
s
0.50 1.2
E jo 045
3S O
040 of/ 1.0 L
* o
.8 S 0.35 lb
/
/ /
0.30] 08'
f/
= </3 0.25 /
. J 0.20 0.6 o
Ol C- // +-
5 "5 0.15 f/ DCQ
0.10 0.4 1
if ^ 0.05 V,
:3-g
0 0.2
-0.05
-0.10
0.4 0.6 0.8 1.0 1.2 1.4-
Pitch Ratio of Front Propeller
Fig. 128.
component of the slipstream has a greater influence on the pitch

of the rear propeller than the axial component. The solid line
in Fig. 128 shows the loss of efficiency of the rear propeller as
compared to its efficiency at the same airplane speed and revolu
tions when isolated. The loss is large for the low pitches, but
it is smaller for the higher pitches and actually becomes a gain
for pitch ratios higher than 1.2. The tandem sets with the
smaller rear propellers show the same general tendencies.
While Eiffel's tests could have been more complete, especially-
regarding the operating regimes corresponding to climb and take
off, they, along with the other tests referred to, tend to illustrate
the following points:
1. With propellers of normal proportions having a spacing of

about one diameter, the front propeller is not appreciably affected
by the rear propeller with regard to power, but with high power
coefficients and low pitches the efficiency of the front propeller
may be reduced as much as 2 or 3 per cent.
2. The efficiency of the rear propeller is greater if the propellers
rotate in opposite rather than like directions.
3. It is advantageous as regards efficiency for tandem pro
pellers to have high pitch ratios.
Since the tandem arrangement is ordinarily used on highly
powered but rather slow airplanes, propellers of high pitch ratio
can usually be obtained only by gearing the propellers to run at
lower revolution speeds than the engines. Gearing is also
advantageous in that the larger diameters of geared propellers
enable them to work on a larger mass of air with consequent
higher ideal efficiencies. This is particularly desirable with
tandem propellers, for their greatest disadvantage is their low
efficiency at take-off and climbing speeds due to the fact that
both propellers work on the same air.
The Diameter of the Rear Propeller.It has been the practice
of some designers, especially in England, to make the rear pro
peller markedly smaller (in the neighborhood of 20 per cent) in
diameter than the front one, mainly to make certain that the
rear propeller always operates entirely within the slipstream of
the front one. In climbing flight the slipstream does not follow
the propeller axis, and the rear propeller is consequently not in
the center of the slipstream of the front propeller, so it is neces
sary to make the diameter of the rear propeller quite small if it
is always to work entirely within the slipstream of the front one.
In order to obtain a reasonable pitch with the small diameter, it
has usually been found necessary to use four blades on the rear
propeller. The small rear propeller has the disadvantage that
if the front engine fails in flight the rear propeller has a com
paratively low efficiency at the low speeds possible with one
engine.
The diameter of the rear propeller is naturally somewhat
smaller than that of the front if both propellers are of the same
general form and are selected to give the highest efficiency, for
the rear propeller works at higher effective velocity and revolu
tions. When it is remembered that (1) due to the body between
the two propellers, the slipstream of the front propeller is spread
so that it retains approximately the full diameter of the front

propeller;1 (2) the transition from the full slipstream velocity to
the velocity of the outside air is not sudden ; and (3) the velocity
is far from constant within the slipstream, varying greatly with
radius at climbing speeds and also being of a pulsating nature,
there seems to be no good reason why the rear propeller should
not have so large a diameter as the front propeller as far as
operating conditions and smooth running qualities are concerned.
There is a variation of the air load on any propeller in climb due
to the angle between the propeller axis and the flight path and
the consequent greater angles of attack of the blade elements
going down than those going up. For a rear tandem propeller
in climb, to this load variation is added that due to the operation
in the pulsating and varying velocity of the slipstream of the front
propeller, and the conditions are apparently as good for a pro
peller the size of the front one as for a smaller one. There is
therefore no apparent advantage in using a rear propeller which is
smaller than that giving the best efficiency.
A Method of Dealing with Tandem Propellers in Which Each
Propeller Is Treated as a Complete Unit.It is usually more
convenient to deal with a propeller as a whole than to deal with
each section separately and then sum them all up. For this
purpose the average axial and rotational velocities in the slip
stream of the entire front propeller are used, the rotational
velocity being considered as an increment in the revolutions of
the rear propeller. (It is an increment only if the propellers
rotate in opposite directions.) The revolutions of the rear
propeller are then in effect (relative to the air in which it works)
higher than the engine revolutions, just as the forward velocity
is in effect higher than that of the airplane, due to the axial
component of the slipstream from the front propeller. If V\
and Ni are the velocity and revolutions (in m.p.h. and r.p.m.) of
the rear propeller with respect to the air in the slipstream of the
front propeller, and V and N are the airplane velocity and the
engine revolutions, the relations between them are given by
the following empirical equations-
1 Investigation of Slipstream Velocity, by J. W. Crowley, Jr.,

N.A.C.A.r.B. 194, 1924.
or
7, _ (l^V/nDYr, + Q v + 0.50)
F V c-'
and
+ 1
JV 5(7/nD)
or
iV, = (7/nP)'
JV 5C.6 f 1.
The coefficients CV, Cp, n, C and V/nD apply, of course, to the
entire front propeller.
If rii is the efficiency of the rear propeller with respect to 7i
and Ni, the actual efficiency of the rear propeller is
_ 7 tfi
Examples. Let us take as an example an airplane having a

maximum sea-level speed of approximately 120 m.p.h. with two
engines in tandem, each developing 500 hp. at 1,800 r.p.m.
From the front propeller, which is treated as if alone, C, =
1.10; and from Fig. 51 or 52 (we shall choose our propellers
from the metal series tested on the VE-7 airplane and assume
that the nose of our engine nacelle is similar to that of the VE-7),
the V/nD for the propeller having the highest efficiency for the
above C is found to be V/nD = 0.573. The diameter is then
10.23 ft., the blade angle at 0.75R is 18.0 deg., and the propulsive
efficiency is 0.778.
The ratios of the velocity and revolutions in the slipstream
to those of the airplane and engine are
7, = (IWV/nDYr, + 0My+ Q 5Q
= 1.170
and
Ni = (V/nD)*
+ 1
N 5C6
= 1.0135.
The velocity and revolutions of the rear propeller with respect
to the air in which it works are therefore
7i = 1.170 X 120
= 140.4 m.p.h.
and
Ni = 1.0135 X 1,800
= 1,824 r.p.m.
Then for the rear propeller C>i = 1.28, and assuming that the
curves of Figs. 51 and 52 apply also to pushers, which is true
within practical limits, we find that (V/nD)i = 0.687, D = 9.86
ft., the blade angle at U75E is 20.5 deg., and 171 = 0.816. The
actual efficiency of the rear propeller is then
1 = X 0.816
= 0.707,
and the average efficiency of the front and rear propellers is
0.742. This is a trifle higher than the best efficiency which can
be obtained with a single propeller of the same family absorbing
the whole 1,000 hp. at 1,800 r.p.m. and 120 m.p.h.
Since the efficiency of the rear propeller is greatly reduced
by the axial velocity added by the front propeller, it might
reasonably be thought that the overall efficiency could be
increased by using a front propeller of greater diameter, thereby
reducing the slipstream velocity. A small gain would result if the
diameter could be increased by the use of narrower blades. If,
however, the larger diameter is obtained by the use of lower
pitch, as would normally be done, the efficiency of the front
propeller drops more than the ratio Vi/V, so that the overall
efficiency is reduced slightly.
The thrust horsepower available at various air speeds has been
computed for the above tandem propellers, on the assumption
that the full-throttle engine torque is constant, and plotted in
Fig. 129. The front propeller yields from 10 per cent more
useful power at 120 m.p.h. to 22 per cent more at 60 m.p.h.
It is interesting to compare the tandem arrangement with a
side-by-side arrangement having two engine-propeller units
similar to the front one. The thrust horsepower available with
two propellers similar to the front one is shown by the top curve
in Fig. 129. This does not make a fair comparison with the
tandem arrangement, however, for the side-by-side arrangement
would necessitate the use of two engine nacelles instead of one.
Taking the difference between the drag of two nacelles and that
of one long nacelle housing both engines as being 150 lb. at 100
m.p.h., which appears to be a reasonable value for this sized
installation, another curve has been plotted on Fig. 129 in which

allowance has been made for this difference in drag. If this
curve is compared with that for the tandem installation, it is
seen that the tandem propellers give a higher thrust power at
speeds above 100 m.p.h. and therefore are better for high speed
800
100
0I I I I I I I
60 70 80 90 100 110 120
Air Speed, M.P.H.
Fig. 129.Thrust horsepower available with direct-drive tandem propellers.
and cruising, but the side-by-side propellers give greater thrust

power below 100 m.p.h. and are therefore better for take-off
and climb. The differences, however, are not large.
It is sometimes argued that one disadvantage of the tandem
arrangement is that if the front engine fails, the rear propeller,
which is designed to operate in the relatively high velocity of
the slipstream, will hold the engine down to a low revolution
speed and consequently give a very low thrust horsepower.
As a matter of fact, the rotation in the slipstream ordinarily
affects the engine revolutions slightly more than the added
axial velocity, so that with propellers of ordinary form, the full-
throttle revolutions of the rear propeller are very slightly greater
with the front propeller stopped than with it running at full

power. The efficiency is slightly lower than that of the front
propeller, due to the higher pitch and the low speeds obtainable
with one engine, and so the thrust horsepower available with
the rear propeller running alone is very slightly lower than that
for the front propeller alone, as shown by the lower dotted curve
in Fig. 129.
With side-by-side engines, if one stops, the thrust power of the
other is the same as for the front engine of the tandem set, but
the thrust is to one side of the center of resistance of the airplane,
which increases the difficulty of control and also the drag.
All things considered, the tandem arrangement seems slightly
superior to the side-by-side in case either engine stops.
Eiffel's model test indicated that the loss in efficiency of the
rear propeller is less for propellers of high than low pitch. The
pitch of the propellers of our example could be increased by
gearing the propellers to turn more slowly than the engines.
Larger diameters would then be necessary, which would also be
advantageous especially at take-off and climbing speeds.
Assuming that our two 500-hp. engines are geared so that the
propellers turn at half the crankshaft speed, and assuming a
gear loss of 2 per cent so that the actual brake horsepower
delivered to the propellers is 490 hp. at 900 r.p.m., the best front
propeller for 120 m.p.h. has a value of C, of 1.46, a diameter of
14.7 ft., a blade angle at 0.75.R of 22.7 deg., and an efficiency of
0.838 X 1.038 = 0.869.1 Then Vi/V = 1.092 and Ni/N =
1.0122, and the best rear propeller has a diameter of 14.42 ft.,
a blade angle at 0.75.R of 24.0 deg., and a value of 171 of 0.852 X
1.038 = 0.884. The actual efficiency of the rear propeller is
1.0122
v = 1.092 X 0.884 = 0.819,
and the average efficiency for the two geared propellers neglecting
gear losses is 0.844 as compared with 0.742 for the direct-drive
installation.
Curves of thrust horsepower available, similar to those for
the direct-drive propellers, are given for the geared propellers
in Fig. 130. A 2 per cent gear loss has been allowed for in these
curves. It will be noticed that the loss for the rear propeller
1 This value for the efficiency is obtained by considering that due to body
interference the propulsive efficiency increases 1 per cent with an 8 per cent
reduction of d/D.
900
Annl I I I I I I I
60 70 80 90 100 110 120 130
Air Speed, M. P. H.
Fig. 131.Comparison of tandem and side-by-side, geared and direct-drive
propellers.
is less for the geared installation. Also, at high speed the geared
tandem arrangement has more advantage over the side-by-side
than is the case with the direct drives, and the geared side-by-side
arrangement has less advantage over the geared tandem in climb
and take-off.
In Fig. 131 the geared tandem and side-by-side arrangements
are compared with the direct-drive tandem and side-by-side
arrangements. Both geared sets are better than either of the
direct-drive ones throughout the entire range.
CHAPTER XIV
MATERIALS AND FORMS OF CONSTRUCTION
Propellers are subject to two main types of loads, those due to

centrifugal forces and those due to air forces. Since the centri
fugal force is greater for heavy than light materials, a low weight
per unit volume is advantageous. The strength should of
course be high, particularly the fatigue strength, which is the
stress which may be repeated indefinitely without causing failure.
The air loads fluctuate somewhat with time due to gustiness,
unsymmetrical conditions, and the elasticity of the propeller
blades, and the fluctuating tends to set up vibrations in the
blades. Hysteresis, or internal friction, is therefore advan
tageous in a material for damping vibrations.
A good propeller material should also hold its shape in various
climates and under varying weather conditions and should be
able to withstand impact with seaplane spray, raindrops, and
small pebbles without appreciable damage.
Wood Propellers.In some ways wood is almost an ideal
material for aircraft propellers, but in others it is not well suited.
Its best qualities are a high strength-weight ratio, a high fatigue
strength, and great internal friction. Tensile stresses up to 5,000
or 6,000 lb. per sq. in. and compressive stresses up to 3,000
lb. per sq. in. are permissible with most propeller woods, the
fatigue strength being almost as high as the fiber stress at the
elastic limit. Thus the ratio of tensile fatigue strength to weight
is higher for wood than for any other successfully used propeller
material. The disadvantages of wood, however, in most cases
outweigh the advantages, and although in 1922 it was universally
used, at the present time it has been replaced by other materials
for all propellers subject to severe service.
Wood propellers warp with varying climatic conditions,
especially in tropical climates. In order to minimize this warping
and twisting, wood propellers are built up of several layers or
laminations, all glued together. The laminations are from
to 1 in. thick, the usual value being % in., and all of the lamina
213
tions of the same propeller having the same thickness, except

possibly the two outer ones. The laminations are placed parallel
to the plane of rotation so that a section through the propeller
~_
A fter Glueing ^_
After Finishing
Fig. 132.Section through blade of laminated wood propeller.
blade appears as shown in Fig. 132. 1 Wood does not satis

factorily withstand the impact of pebbles, rain, or seaplane spray.
In fact, one-half hour of running in hard rain is sufficient to
ruin the leading edge entirely.
For this reason wood propellers
are usually sheathed . with
metal at the tip and part way
along the leading edge. Three
forms of metal tipping are
illustrated in Fig. 133. The
Navy tip, which is for pro
pellers subjected to severe sea
plane spray, is a complete tip of
brass 0.022 in. thick, fastened
on with brass wood screws
- Army Scalloped Tip where the blade is over ^
Fig. 133. in. thick and with copper
rivets where it is thinner.
In England a similar tip is used excepting that saw cuts are put
in the metal as shown. This provides a certain amount of
freedom which keeps the tipping from cracking at other places,
1 For a description of the process of manufacture of wood propellers, see
"The Airplane Propeller," published by the War Department in 1920.
MATERIALS AND FORMS OF CONSTRUCTION 215
as is common with the Navy tip, especially on flexible propellers.

Some American manufacturers use a somewhat similar arrange
ment in which the tip is made up of a number of overlapping
sections. The Army uses a scalloped tipping of terne plate
which is expected to crack between scallops.
Fig. 134.Typical fabric-sheathed wood propeller.
The best propeller woods are walnut, birch, oak, and Honduras
mahogany. Walnut is probably the best all-around wood and
has the special advantage of holding its shape well. Birch is the
toughest and is best used in experimental propellers of doubtful
Fig. 135.Wood propeller with Navy tips.
strength. It is a better propeller wood than walnut except for

a greater tendency to warp. Oak is used by the Navy because
of its ability to withstand seaplane spray better than the others.
It wears away without splitting. Mahogany is lighter than the
others, and it holds its shape well.
While wood is strong in tension it is weak in longitudinal

torsion and splits easily along the grain. For this reason the
blades of wood propellers are often sheathed with fabric. The
fabric is glued on and increases both the torsional strength and
stiffness. A typical fabric-sheathed wood propeller of Navy
form is shown in Fig. 134. A similar propeller with a Navy-type
metal tip is shown in Fig. 135, and a typical four-bladed propeller
in Fig. 136. The latter two pictures show the metal-flanged hubs
with through bolts commonly
used with wood propellers.
Micarta Propellers. A
material called micarta has
been developed by the West-
inghouse Electric and Manu
facturing Company and has
been applied by the same com
pany, in cooperation with the
Army, to aircraft propellers.1
As used in propellers, micarta
is a composition formed of
layers of cotton-duck fabric
impregnated with a synthetic
fig. ise.-Four-biaded wood propeller. resin binder and baked under
pressure. The propeller is
formed of laminations of cotton-duck micarta and is finished as
it comes from the mold, except for slight trimming at the parting
line and curing for about 20 hr. at a temperature of 100C.
The hub is fitted with a steel bushing having large splines
extending into the micarta, the bushing fitting directly on the
engine shaft. The propeller is balanced by locating the bushing
at the center of gravity. A micarta propeller with hub bushing
is shown on a balancing stand in Fig. 137. Micarta propellers
are also made with detachable blades held in a steel hub in much
the same manner as the aluminum-alloy blades described later.
Micarta is not affected to an appreciable extent by climatic
changes, and it withstands rain, cinders, and pebbles somewhat
better than wood. It does not withstand them with entire
satisfaction, however, and the latest micarta propellers are
1 For a detailed account of the materials, methods of construction, and
design see Micarta Propellers, Parts I to IV, by F. W. Caldwell and N. S.
Clay, N.A.C.A.T.iV. 198, 199, 200, 201, 1924.
being tipped with metal in much the same manner as wood

propellers.
The properties of micarta depend largely on the fabric used
for the laminations. With an ordinary good grade of cotton
duck, as usually used in propellers, an ultimate tensile strength
of from 9,000 to 10,000 lb. per sq. in. and a longitudinal shearing
strength of about 4,000 lb. per sq. in. are obtained. With a
special fabric having 90 per cent of the total weight of the cloth
in the warp, which is placed lengthwise of the propeller blade, a
tensile strength of 26,000 lb. per sq. in. has been obtained.
Ordinary propeller micarta weighs approximately 0.05 lb. percu.in.
Fig. 137.Micarta propeller for Liberty Engine. (Photograph from N.A.C.A.

T.N. 199.)
Like wood, micarta has a high degree of internal friction, and

the fatigue strength is apparently fairly high. Unlike wood,
however, micarta has no tendency to split, and micarta propellers
often run satisfactorily under conditions which have proved too
severe for wood propellers and certain types of metal propellers,
conditions such as that of a wing propeller operating partially
in the slipstream of a nose propeller, with a consequent highly
fluctuating air load on the blades. Due to the non-splitting
properties of micarta, micarta blades can be made from 20 to
40 per cent thinner than wood blades, resulting in slightly greater
efficiency. At the present time the cost of micarta propellers
is high and they are used but little.
Steel Propellers.Many steel propellers have been designed
and built, in a large variety of both solid and hollow forms of
construction, but only one type, the Leitner-Watts, has proved

sufficiently satisfactory to be used in service. This type is made
with detachable blades which fit in a two-piece split-steel hub.
The blades are hollow and are built up of thin laminated sheets
of mild steel pressed to form. The number of laminations
decreases from hub to tip, giving the material in effect a tapering
cross-sectional area. A section showing the construction of the
blade root is shown in Fig. 138.
The shapes of the blades and the
laminations, which are terminated
in a sort of tapered fork so that
the change in cross-sectional area
is not abrupt, are shown in Fig.
139. The blades are pressed in
-Seam
- Ends of .
\ Laminations
A solid mass is formed here [

by welding betweenplates
Fig. 138.-Section showing con Fig. 139.Leit

struction of blade root of Leitner- ner-Watts laminat
Watts propeller. {From catalogue ed steel blade.
of Metal Airscrew Company, Ltd.)
halves, and the two halves are fastened together by welding
along the edges. In order to stiffen the flat face a seam is put
in it and internal spacers are riveted in between the two faces.
Under fluctuating loads the friction between the laminations
tends to damp out vibrations, which, in the opinion of the writer,
is the reason for the success of the Leitner-Watts propellers when
all other steel propellers have failed. Steel has very little internal
friction if stressed below the elastic limit, and steel propellers,
even when apparently stressed well below the fatigue strength,
have repeatedly failed, probably because the vibrations build

up to the extent that they cause stresses above the fatigue limit.
The Leitner-Watts propellers were the first to be made with a
limited number of standard-sized blades and with two-, three-,
and four-bladed hubs to take the blades, making it possible to
take blades and hub from stock to fit any ordinary aircraft
engine. The hubs are made with removable bushings which fit
the various engine shafts.
The Leitner-Watts propellers are made no thinner than wooden
propellers, and they are consequently no more efficient aero-
dynamically. They are not being widely used at this time, for
they are more expensive and less efficient than the thinner solid
aluminum-alloy propellers which have been developed during
the last few years. They still have a field, however, in very
large-sized propellers, especially for airships, where they can be
made lighter than the present solid aluminum-alloy propellers.
Aluminum -alloy Propellers.Aluminum alloy has now been
used for propellers for several years and its use is rapidly
increasing. It has already displaced wood in the high-power
high-revolution range, and it is gradually displacing it in the
low-power range.
The composition and physical properties of all aluminum
alloys used for propellers are practically the same, although
they are made by various manufacturers and bear different trade
names.1 There are slight variations in the alloys and the
methods of heat treating, but all are made up of approximately
95 per cent aluminum, 4 per cent copper, and 1 per cent other
ingredients and impurities. In propellers, the alloys are used in
the rolled or forged state, and all of the alloys, after heat treat
ment, have approximately the following physical properties:
Ultimate strength 55,000 to 60,000 lb. per sq. in.
Fatigue strength 10,000 to 15,000 lb. per sq. in.
Modulus of elasticity 10,000,000 to 11,000,000 lb. per sq. in.
Weight per cubic inch 0.100 to 0.103 lb.
The aluminum alloys have no definite yield point or elastic limit,
although the yield point is often arbitrarily considered as from
30,000 to 40,000 lb. per sq. in. The material does not have a
straight-line stress-strain curve, the deflection not being exactly
1 Aluminum Company of America, 25S; Baush, duralumin; Vickers
(England), duralumin; Schneider and Company (France), alferium; Metall-
bank and M. G. (Germany), Aeron.
proportional to the load applied even with comparatively low

loads. This indicates a hysteresis or internal friction in the
material which is helpful in damping out vibrations. Stresses
up to the fatigue limit, which is usually considered about 12,000
lb. per sq. in., can be repeated indefinitely without failure, and
it is the aim in designing propellers to keep the stresses, including
those due to vibrations and fluctuations, below this value.
Aluminum, although it has- a slight grain structure when
rolled or forged, has very nearly the same strength in all direc
tions. Due to this homogeneity, and also to the stiffness of the
aluminum alloys, propellers of this material can be made with
very thin sections which are advantageous as regards efficiency,
especially at high tip speeds.
Fig. 140.Reed twisted-slab type of aluminum-alloy propeller.
The first aluminum-alloy propellers to be used were of the

Reed twisted-slab type. Propellers of this type are made from
flat slabs of aluminum alloy, the blades being machined and hand
filed to the form of a flat tapered airfoil and then twisted to the
proper pitch angles. The center portion is fitted to the ordinary
flanged hub for wooden propellers by means of filler blocks, as
shown in Figs. 140 and 141. The twisted-slab type of propeller
was rather widely used a few years ago but during the last few
years its use has declined greatly.
The twisted-slab type of construction with filler blocks was
considered by the inventor a temporary expedient for adapting
metal propellers to wood propeller hubs, and it has been replaced
by a more satisfactory form called the Reed R type. The R-type
hub are squeezed tightly down around the blade shanks by means
of clamp rings, which are shown separately in Fig. 145 and on a
FlQ. 145.Parts of split-type hub for three-bladed propeller.
propeller in Fig. 143. The dimensions of the blade roots recom

mended by the Department of Commerce and standard for the
Army and Navy are given in Fig. 146.
tRound Corners i/^
STANDARD BLADE A B c D E F
END No. H
N0.O0-0T0 50HP. 2495 -tO.000
-0.003 <uuu -0.010 0 500
U-5UU +0 002
-0.000 04375 +0-O1C i-iio
l">'< +0.010
-0.010 3/32
-0.010
NO.O-50T0 250HP 3.000 1 , 25D 0 000 9C9C
2625 +0.010
-0.010 uwa -o ooo uw -0.010 1687 40010
056Z +0.010 -0.010 %
. +0.000
NO.1-25070500HP 1 flit +0.000 *M5 ,,7n<S +0.010
-0.003 .5-3 -0.010 riA-rc -0.000 0.750 1JS8 ,
-875 +0.002 +0.010
"'5 -0.010 S/32
a qnr +0.000
NO.2-500T0800HP a.soo 18885 ,9 1A7C +0X510
-0.003 3876 l9cn +0.0OZ
-0.010 123U -0000 1.000 iggfg x,a 3,25 +0.0iO
-0.010 %2
FlQ. 146.Department of Commerce standard ends for detachable blades for
aluminum-alloy propellers.
The blades of these propellers are forged solid, and when

a particular design is made in large quantities, they are drop
forged with dies which leave them within about ^6 m- of their

final size. They are then machined to the proper dimensions,
largely by means of hand grinding with flexible disc grinders
and hand filing. Usually each blade is balanced against a master
blade kept for that purpose
only.
Propellers made of aluminum
alloy are, including the steel
hub, of about the same weight
as micarta propellers for the
same purpose and are usually
(but not necessarily) slightly
heavier than wood propellers
for the same purpose. .
Aluminum alloy is not only
well fitted for use in propellers
but also has the additional
Fig. 147.Aluminum-alloy propeller advantage over any other
bent in an accident. material used to date that a
propeller damaged in an air
plane accident (Fig. 147) can be straightened and heat treated
by the manufacturer at nominal cost and used again.
Aluminum alloy resists the erosion of rain, pebbles, and sea
plane spray better than wood or micarta, but it will not satis
factorily withstand severe seaplane spray. In one extremely
severe taxiing test made by the Navy, in which the writer partici
pated, approximately one-fourth inch of the leading edge of an
aluminum-alloy propeller was eroded off in ten minutes. After
the smooth edge has been partially eaten into, the impact of the
spray and occasional solid waves tears the metal off rapidly in
irregular pieces. For ordinary seaplane service the conditions
are not so severe, and aluminum-alloy propellers usually give
reasonably satisfactory service if they are kept smooth by rub
bing with crocus cloth and, when the blades become slightly
rough or pitted, an occasional filing. If the propellers are
operated near salt water it is advisable to protect them from
corrosion by wiping them with oil after each flight.
The Construction of Variable -pitch Propellers.Although
variable- or controllable-pitch propellers have not yet progressed
beyond the experimental stage, it seems certain that in the near
future they will be used, at any rate on supercharged airplanes,
and it is therefore of interest to describe the construction of a

few of the most promising types.
One of the greatest problems with the early variable-pitch
propellers was to provide a suitable way of attaching the wooden
blades to the hub in such a manner that the full strength of the
wood was obtained. This was finally worked out in a fairly
Fig. 148.Sleeve for fastening wood blade in metal hub.
satisfactory way for low powers, by fitting the blade root in a

tapered-steel sleeve as shown in Fig. 148. With micarta, or
aluminum alloy, standard or nearly standard detachable blades
can be used, and the blade fastening is not a difficult problem.
In all of the variable-pitch propellers which have been flight
tested on airplanes, the centrifugal force on the blades is taken
on ball or roller thrust bearings in order to make the friction
Fig. 149.:Principle of operating mechanism of Hart variable-pitch propeller.
against turning as low as possible. In order to keep the weight

within reason it is usually necessary to use bearings which are
very highly loaded. It is therefore important that the blades
be made as fight as is practicable. The air loads, which put a
side force on the blades, are not so great and are often taken on
plain radial bearings.
The Hart variable-pitch propeller was one of the first to be

built and tested in this country. The pitch control is manually
operated, the blade angles being set to any desired value by means
of a lever in the pilot's cockpit. The operating mechanism,
which is quite simple, is diagrammed in Fig. 149. This shows a
cross-section of the root of one blade having an arm extending
out to one side and projecting through a slot in the hub casing.
This arm is connected by a link to a collar which rotates with the
propeller and can slide back and forth along the axis of the engine
shaft. Each position of the collar corresponds to a certain
propeller pitch angle. The collar has a groove fitted with a
yoke, the position of which is controlled by a lever in the cockpit.
Fig. 150.Hart variable-pitch propeller with counterweights and micarta

blades. (Photograph from N.A.C.A. T.N. 200.)
Due to the fact that the blades are not symmetrical about a
radial center line, the centrifugal force produces a twisting
moment which is much larger than and opposite to that produced
by the air force and which tends to turn the blades in the hub,
making it necessary to apply an undesirably large force to change
the pitch in one direction. To overcome this the Army, which
has undertaken the development of the Hart-type propeller, has
designed small counterweights which project forward from the
blade roots near the hub and which produce twisting moments
due to centrifugal force approximately balancing those on
the blades themselves. These are shown in Fig. 150, which is a
view of a Hart variable-pitch propeller with micarta blades, on a

JN-4H airplane with pitch control from both cockpits.
Another form of variable-pitch propeller is that invented
and developed by Spencer Heath, in which engine power is
used to change the pitch. The principle of operation is shown
diagrammatically in Fig. 151, which represents a brake drum con
nected through a gear train to each
blade of the propeller. If the brake
drum is allowed to revolve freely, it
will turn with the crankshaft, all gears
will remain stationary, and the pitch
will remain the same. If, however,
the brake drum is held stationary, the
gears will be put into action, and the
Spur Gears
pitch angle will change as long as the Brake Drum
braking force is applied. In order to
change the pitch in the opposite direc
tion, a second brake drum is connected
to the worm shaft through an idler
which reverses the direction of rota
tion. During normal flying none of
these gears is operating, and the
blades are locked in position by the
worm gears which are irreversible.
In actual construction the mechan
ism is complicated by the necessity
for a large gear reduction, entailing
several extra gear wheels, and also Fig. 151.Diagram showing
by the necessity for a pitch-angle operation of Heath variable-
indicator which entails many small pitch propeller. iFrom N.A.
C.A. T.M. 70.)
continuously operating gears. It has
the advantages, however, that the pitch can be changed by a
light pull or push on a button on the end of the brake-control
rod, that the angular range of the blades is unlimited, and that
there are no running gears or slip joints in the pitch-changing
mechanism.
Another promising variable-pitch propeller is that designed
by R. W. Turnbull of Canada and tested by the Royal Canadian
Air Force. In the Turnbull propeller the pitch is changed by a
small electric motor which is located in a spinner at the front of
the hub and is geared to the blades.
The most recent and perhaps the most promising variable-

pitch propeller is the hydraulically and automatically operated
Hele-Shaw Beacham. This propeller has a double-acting
hydraulic ram or piston which is located in a spinner in front of
the hub and operates a sliding collar similar to the one in Fig. 149,
which in turn changes the pitch angle of the blades through a
link and arm. The whole mechanism revolves with the propeller.
The piston is moved back and forth by oil pressure, the oil being
brought in through the engine shaft, which must be drilled and
fitted with a rotating oil-tight joint, from a variable-stroke
pump driven by the engine. The stroke of the pump is con
trolled by a governor, also driven by the engine, which can be
set by the pilot for any desired engine revolutions. If the revolu
tions fall below the set value, the governor acts on the pump in
such a manner as to increase the oil pressure on the side of the
piston which decreases the pitch, and vice versa. Thus the action
is entirely automatic but can be over controlled by the pilot
by merely changing the revolution setting, which is a highly
desirable feature, especially in high-altitude and acrobatic
flying. If the hydraulic gear should fail, a spring returns the
blades to a predetermined "normal" pitch, and the flight is
completed with what is in effect an ordinary fixed propeller.1
If variable-pitch propellers are used as brakes for landing
airplanes or for reversing an airship, the pitch angles must
pass through zero. If the throttle were kept open while the
blades passed through zero pitch, the engine would race and
very likely destroy itself. To eliminate this possibility, the
pitch-control mechanism is connected to the throttle, usually
by means of a cam of some kind, in such a manner that the
throttle is closed as the blades are moved through the zero pitch
position.
Although all of the variable-pitch propellers described have
apparently been successfully flight tested, none has been used
further in service. The Gloster Hele-Shaw Beacham propeller,
however, is new and is apparently being given service trials by
the British Government. All of the propellers are quite heavy,
but there is a possibility of reducing the weight materially by
refinement of design.
1 The Gloster Hele-Shaw Beacham variable-pitch propeller is thoroughly
described in a paper by the inventors, published in the July, 1928, issue of the
Journal of the Royal Aeronautical Society.
CHAPTER XV
THE STRENGTH OF PROPELLERS
The calculation of the stresses in propellers, which have both

a non-uniform shape and a varying load, is a complex problem,
and no strictly correct method of calculating the stresses has been
devised. The ordinary theory of bending is used as the best
practical method, but it is based upon assumptions which are
of very doubtful value when applied to a body with as complex a
shape as that of a propeller.1
The Loads on a Propeller Blade.There are three main types
of loadg which aft nn propeller blades in operation:
V^ThgJoad duejtoâirjorces^
2. Tê_load_due to centrifugaHorces?)
3. The load due to gyroscopic forces and to other inertia
forces due-to-unsteady motion:
In our "analysis of propeller stresses we shall limit the case to
steady rectilinear flight, so that only aerodynamic and centrifugal
loadings need be considered.2
The propeller blade may be looked upon as a cantilever beam,
in which case the stresses at any given cross-section depend on the
loads on the portion of the blade from the section in question to
the tip. It is not a cantilever beam in the ordinary sense, for
the centrifugal force tends to hold the blade straight out after
the manner of a tie.
A section of a propeller blade is shown in Fig. 152, where
the direction of rotation is given by OA, the direction of transla
tion by OB, and OC represents the radial axis through the center
of gravity of the section 0 perpendicular to the plane of the
section.
1 See Some Notes on the Calculation of the Working Stresses of an Air
screw, by A. Fage and H. E. Collins, British B. and M. 420, 1918. These
notes contain the basis of the method of stress analysis given in this chapter.
'Loads due to gyroscopic and other inertia forces are small compared
with the air and centrifugal loads, and a propeller which is strong enough
to endure the latter two types of loads for an indefinite period of time is also
strong enough to stand the slightly higher total loading during the relatively
short periods in which the gyroscopic and other inertia forces act.
229
The loading on the section due to the air forces on the portion
of the blade between the section and the tip may be represented
by:
1. A bending moment which may be represented by the two
components
a. MT, due to thrust, about the axis OA.
b. Mq, due to torque forces, about the axis OB.
2. A twisting moment MA about the axis OC, due to the fact
that the resultant air forces for the various sections along the
blade do not pass through the axis OC.
3. A shearing force across the section.
Fig. 152.
The centrifugal loading on the section for the general case

where the centers of gravity of all of the blade sections do not
lie on a straight radial line perpendicular to the axis of rotation
may be represented by
1. A bending moment which may be resolved into the two
components
a. Mt, about the axis OA.
b. Mq, about the axis OB.
2. A twisting moment MY about the axis OC, which is due
partially to the fact that the centers of gravity of the sections
are not all on a straight radial line and partly to the fact that the
sections are airfoils of unsymmetrical shape.
3. A tensile force CF acting radially outward along OC.
4. A longitudinal shear perpendicular to the plane of the
section.
The shearing forces are small and are ordinarily neglected
even in wood propellers, which are particularly weak in longi
tudinal shear. It should be kept in mind, however, that the
cross-shear due to the air load also results in a longitudinal
shear and that when the air load fluctuates, as it usually does,
THE STRENGTH OF PROPELLERS 231
the resultant varying longitudinal shear has a tendency to cause

working and deterioration of the glue joints.
The twisting moments also cause an unimportant shear in
the blades and are usually neglected. In wood propellers, large
twisting moments tend to split the blades lengthwise and should
be avoided. The most important effect of the twisting moments
is more aerodynamic than structural, for the accompanying
angular deflection affects the propeller pitch.
In computing stresses the loading at a section is therefore
considered as made up only of a tensile force CF, the centrifugal
bending moments MT' and MQ', and the air-load bending moments
MT and MQ. The combined resultant bending moment on the
section is then
M = V(Mr+ Mt'Y + Wq + MQ')\
Fig. 153.
Calculation of the Maximum Stresses at a Section.The

typical propeller section shown in Fig. 153 has the principal
axes OX and OF. The resultant bending moment M is about
the axis OD, which is at an angle X with the minor principal
axis OX.
The neutral axis ON is found as follows: The resultant bending
moment M is resolved into two components: M cos X about the
minor principal axis OX; and M sin X about the major principal
axis OF. Then if Ix and Iy are the moments of inertia about
the minor and major principal axes, respectively, at any point P
(Fig. 153), which is a distance x from OF and a distance y from
OX, the stress according to the ordinary theory of bending is
yM cos X xM sin X
6 = *x Jjy
The stress at all points on the neutral axis is zero, and so by

equating the above expression to zero the equation of the neutral
axis is found to be
y = x ~ tan X.
The neutral axis (Fig. 153) is inclined to the minor principal

axis OX by an angle 5, where
v I
tan S = - = tan X.
X ly
The bending moment about the neutral axis is then
M cos (X S),
and the moment of inertia about the neutral axis is
J = Ix cos2 S + Iv sin2 5.
The maximum tensile and compressive stresses in the section
occur at the points farthest from the neutral axis (yt and yr,
Fig. 153). The maximum stresses due to bending are then
r, M cos (X S)yt
bt = I
M cos (X 8)yc
c f
If A is the area of the section, the uniform tensile stress across the
section due to the centrifugal load is
CF
bt - -x-
The total maximum tensile and compressive stresses at the

section are then
Total St = M C0S (7X ~ 5)y< + f
and
Total S6 = Mcos
/ A
There is at present no way of knowing the accuracy of the
stress calculated in this way, for there is no strictly rigorous
method to compare with.
Calculation of Loads.The air forces are calculated by means
of the blade-element theory, the unit resultant force for each
section being obtained from the thrust and torque per unit radius.
The centrifugal force acting on any given particle in the pro

peller is represented by the equation
W
CF = *r,
9
where W is the weight of the particle, g is the acceleration due to
gravity, is the angular velocity in radians per second, and r
is the perpendicular distance between the particle and the axis
of rotation. In engineering units the angular velocity u becomes
2t X RPM/60 and the equation takes the form
Wr(2w X RPMV
CF - T\ 60 ) '
The propeller blade is usually divided into elements spaced

either 6 in. or 0.15 R apart along the radius. The weight W
of an element per inch of radius becomes
W = Aw,
where A is the cross-sectional area of the section in square inches
and w is the specific weight of the material in pounds per cubic
inch. The centrifugal force on an element is given by
d(CF)
(CF) =
_ . w/2ir
w/2tt X RPMV
dr g\ 60 )
Variation of Stresses with Tip Speed in Geometrically Similar
Propellers.As can be seen from the above equation the centri
fugal force is proportional to n2D4, and the blade-element theory
shows that for any one value of V/nD the air forces also are
proportional to n2Z)4. Both the centrifugal and air-load bending
moments are therefore proportional to n2Db, and
Bending stress ~
Since I/y is proportional to D3, the above expression may be

written,
n2Db
Bending stress ~
~ n2D2.
Also, since the uniform tensile stress due to centrifugal force is
given by the expression CF/A, and since A~D2,
n2D*
Centrifugal stress ~ ^r
~ n2D2.
Thus at any one value of V/nD, and neglecting scale effect, all
of the stresses in geometrically similar propellers of like material
vary as n2D2 and therefore as the square of the tip speed.
An Approximate Method of Stress Analysis.The above
method of stress analysis is very laborious as well as of doubtful
accuracy. An approximate method, which for all practical
purposes is as accurate as the more complete method, will now
be given, based on the following simplifications:
1. The resultant air force is assumed to be perpendicular to
the chord at all sections. This is essentially true and causes
negligible errors.
2. The chords of all sections are assumed to lie in a plane, so
that the resultant air forces also lie in a plane. This neglects
the change of blade angle along the radius and results in the cal
Assumed
cNeutral Axis-,.
I Assumed Resultant
Air Force
Fig. 154.
culated stress being slightly too high, an error on the safe side.
The error is negligible in propellers of ordinary and low pitch
and is never likely to be greater than 5 or 10 per cent even in the
worst case of the hub sections of a very high-pitch propeller.
3. The neutral axis is assumed to be parallel to the chord
(Fig. 154). This, as shown by calculations in R. and M. 420,
causes no error in the maximum compressive stresses, but the
computed maximum tensile stresses due to bending are 25 to
30 per cent low for the ordinary range of angles between the
actual neutral axis and the assumed one parallel to the chord.
The tensile stresses due to bending as found by this approximate
method are therefore multiplied by the factor 1.30.
4. Twisting forces are neglected.
Example of Approximate Stress Analysis.We shall make an
approximate analysis of the stresses in the standard aluminum-
alloy propeller shown in Fig. 176 (Navy Bureau of Aeronautics
Drawing 4412, set at 15 deg. at the 42-in. radius, giving a medium
pitch). The blades have standard propeller sections based on
the R.A.F.-6 airfoil (Fig. 19). The blade widths, thicknesses,
and angles are given in the first part of Table IV. The centers
of gravity of all of the blade sections lie on a straight radial
line perpendicular to the axis of rotation, and so, neglecting
deflection in operation, there are no bending moments due to
centrifugal force. The areas and the locations of the centers of
gravity of the standard (R.A.F.-6) propeller sections may be
found from Fig. 155.
We shall analyze the stresses for the case where the propeller
is absorbing 170 brake hp. at 1,700 r.p.m. and 82 m.p.h.
>~dJ!ii6h .
k 0A46b
Single Camber
C.G.Tata/ Section. C.S. Upper

^T- 0.4/6 hu .
^ h ? h
< 0.446b > ^C.G.i-Ower
-fa ' >
Double Camber
Single Camber
Area = 0.74 bh
I *0.047Zbh3
Double Camber
x.OA16(hu*hL)
Area^0J4b(hu+hL)
I-0.0472b(hj> + V) + 0.112bhuhJhu+hLi)
Fig. 155.Centers of gravity, areas, and moments of inertia of standard
propeller sections based on R.A.F.-6 (see Fig. 19).
For the section at 0.45/2 the cross-sectional area (Fig. 155) is

A = 0.74 bhu
= 0.74 X 7.25 X 0.86
= 4.62 sq. in.
The centrifugal force per foot radius is
d(CF) _ l2rAw /2ir X RPMV
dr ~ g \ 60 )
= 12 X 2.01 X 4.62 X 0.101 /2t X 1,700V
32.2 \ 60 )
= 11,100 lb. per ft.
The centrifugal force per foot radius is found in like manner

for all of the sections and is plotted against radius in Fig. 156.
The area under the curve from any section to the tip, proper
consideration being given the scales, represents the total centrif
ugal force acting on the section. This area for the 0A5R or
12,000
10,000
6000
6000
4000
2000 i
OR 0.1 5R 03 )R \
0 1JR 02OR OAi5R 0.(i
1 Z J 4 5
Radius in F+.
Fig. 156.Centrifugal-force loading curve.
2.01-ft. section, as found by means of a planimeter, is 8.86

squares, and the centrifugal force acting on the section is
CF = 2^00 X 8.86
= 17,700 lb.
The uniform tensile stress due to the centrifugal force is then
= 17,700
4.62
= 3,830 lb. per sq. in.
The thrust and torque grading curves for one blade of the
propeller are shown in Fig. 157, having been computed by means
of the blade-element theory and modified slightly to agree with
an actual full-scale test on the propeller operating under the

given conditions in the N.A.C.A. Propeller Research Tunnel at
Langley Field. From the thrust and torque, the resultant unit
160
Total Resulfani
120
//
//
// rque . \ 11
> fit h
80 // . \\
r / V
/ \
/ \
t ,
/'<
-L e. I
40 A' y 'Force in Direction of Torque
4'A-j-n
0 O.JOR 0.45R 0.60R 0.751 R 0.901 R
1 E.J
Radius in F+.
Fig. 157.-Air-loading curves.
air load is found at each section, and this is also shown in Fig.
157.
The shear at each section is then found from the area under
the curve of the resultant air loading. For the section at 0A5R
400 r
this area is 7.18 squares and the shear is 40 X 7.18 = 287 lb.
The shear for each section is then plotted against radius (Fig.
158), and the integration of this curve gives the bending moment.
Thus for the 0.45jB section, the area under the shear curve
between 0.45/2 and R is 3.58 squares, and the bending moment

at 0.45.R due to the air load is
M = 100 X 3.58
= 358 ft.-lb.
- 4,290 in.-lb.
From Fig. 155,
yt = 0.416/iu
= 0.416 X 0.86
= 0.36 in.
and
yc = 0.50 in.
Also, from Fig. 155.
I = 0.0472 bhu3
= 0.0472 X 7.25 X 0.86s
= 0.217 in.4
The maximum stresses at the 0.4522 section due to the bending
moment caused by the air load are
1.3 My,
S, = j
_ 1.3 X 4,290 X 0.36
0.217
= 9,250 lb. per sq. in.
and
c _ My,'
oc j-
= 4,290 X 0.50
0.217
= 9,900 lb. per sq. in.
The total maximum tensile stress at the section, which occurs
on the face near the leading edge, is then
St = St (bending) + St (centrifugal)
= 9,250 + 3,830
= 13,080 lb. per sq. in.,
and the total maximum compressive stress at the section is
Sc = Sc (bending) St (centrifugal)
= 9,900 - 3,830
= 6,070 lb. per sq. in.
The maximum tensile stress at each section is plotted as curve

1 in Fig. 159, showing that the maximum for the whole blade
occurs at the 0.45/2 section.
,16,000
, No AHowan ce for
\Deflec
S; 12,000
,Def ecfion Consi'der

/J
| 4000
EH
a
xR
Fig. 159.-Maximum tensile stress at each section.
The Effect of Deflection on the Stresses.The centers of

gravity of the sections of this propeller all lie on a straight radial
line perpendicular to the axis of rotation, so that if the deflect
ion of the blades under load is neglected, and this has been the
accepted practice and was done in our example above, there
are in this case no bending moments due to centrifugal force.
1 2 3
Radius in Ft.
Fig. 160. Deflection of propeller in operation.
With the propeller operating under the above conditions (i.e.,

170 hp., 1,700 r.p.m., and 82 m.p.h.) in the Propeller Research
Tunnel, however, both a forward deflection and a twist were
measured, as shown in Fig. 160. The change of blade angles
affects the pitch (this was allowed for in the thrust and torque
computations) and the forward deflection brings in bending
moments due to centrifugal force which are opposed to the air
Table IV. Computations for Stress Analysis of Standard Aluminum-
alloy Propeller, Neglecting Deflection
Bu-Aero Propeller 4412
D = 8 ft. 11 in. Hp. = 170 R.p.m. = 1,700 M.p.h. = 82
r/R 0.133 0.30 0.45 0.60 0.75 0.90

r (ft.) 0.594 1.34 2.01 2.68 3.34 4.02
b (in.) 3.87 7.00 7.25 6.88 5.71 3.90
hu (in.) Circular 1.22 0.86 0.64 0.44 0.23
section
0 0 0 0 0
Centrifugal Force Loads

<E1
dr - l2A"(l\'
" ff\ 60 / -ArX
Ar X 12X0^Pi
32.2 V/2OO.700V
60 )
= 1,195 Ar
A = 0.746^ 11.78 6.32 4.62 3.26 1.86 0.66

Ar 7.00 8.46 9.29 8.74 6.21 2.65
d{CF)/dr 8,350 10,100 11,100 10,420 7,430 3,160
C. F. (lb.) 32,200 25,100 17,700 10,500 4,500 780
Air-load Bending Moments
Thrust per ft. (lb.) 0 34.0 75.5 113.7 128.8 124.8

Torque per ft. (ft.-lb.). . 0 26 53 83 99 96
Torque force, per ft.,
(lb.) 0 19 26 31 30 24
Resultant air load, per
ft., (lb.) 0 39 80 118 131 127
Shear (lb.) 340 327 287 220 135 44
M (in.-lb.) 9,840 6,830 4,290 2,270 828 108
Stresses
V, (in.) 1.94 0.51 0.36 0.27 0.18 0.10

Vc (in.) 1.94 0.71 0.50 0.37 0.26 0.13
/ (in.') 11.08 0.600 0.217 0.085 0.023 0.0022
S,, bending 1,720 7,540 9,250 9,370 8,400 6,400
Sc, bending 1,720 8,100 9,900 9,900 9,300 6,400
St, direct centrifugal . . . 2,730 3,970 3,830 3,220 2,420 1,180
4,450 11,510 13,080 12,590 10,820 7,580
Sc, total 4,130 6,070 6,680 6,880 5,220
force bending moments; thus the stresses are actually lower than
the values found by assuming no deflection.
The unit bending moment at the 0.45/2 section due to the
centrifugal loading at the 0.60/2 section is the loading at the
latter section (Table IV) multiplied by the difference in forward
" * 2000
0.45 V 0.6 >R 0.7 >R 0. 0R\

0 V
Radius in Ft.
Fig. 161.Bending moment at 0.45ft due to centrifugal force.
deflection at the two sections (it being assumed that the for
ward deflections are the same as those in the assumed plane of
the resultant air loading), or
~ = (deflection at 0.60/2 - deflection at 0.45/2)

= 10,420 (0.20 - 0.07 in.)
= 1,355 in.-lb. per ft. radius.
This is plotted in Fig. 161 along with the corresponding values
for the 0.75/2 and 0.90/2 sections, and the area under the curve
represents the total bending moment at the 0.45/2 section due to
centrifugal force. The area is 3.55 squares, and the bending
moment at the 0.45/2 section due to centrifugal force is
M' = -1,000 X 3.55

= -3,550 in.-lb.
The value is negative because it is opposite to the air-load bend
ing moment, which was taken as being positive. The bending
moments due to centrifugal force have been found in like manner
for the other sections and are given in Table V, along with the
moments due to the air load.
Table V.Stresses Considering Deflection
r/R 0 133 0.30 0.45 0.60 0.75 0.90

9,840 6,830 4,290 2,270 828 108
M', centrifugal -5,050 -5,050 -3,550 -1,610 -590 -78
M+W 4,790 1,780 740 660 238 30
St, bending 840 1,960 1 ,600 2,720 2,420 1,780
840 2,110 1 ,710 2,880 2,670 1,780
Si, direct centrifugal 2,730 3,970 3,830 3,220 2,420 1,180
Si, total 3,570 5,930 5,430 5,940 4,840 2,960
250 600
1
The maximum tensile stress at 0.45.R due to bending is now
found to be
c 1.3(M + M')y,
8, = j
1.3(4,290 - 3,550) X 0.36
0.217
= 1,600 lb. per sq. in.
as compared with 9,250 lb. per sq. in. when the deflection was
neglected. Thus it is obvious that neglecting the deflection
gives an entirely erroneous idea of the bending stress, and in this
case a value of the tensile stress which is several hundred per
cent too great, and an apparent compressive stress along the
whole blade with a value as high as 6,880 lb. per sq. in., while
when the deflection is considered there is no compressive stress
except for a negligible quantity near the tip (Table V). Curve
2 of Fig. 159 shows the total maximum tensile stress found along
the blade when the deflection is taken into account. It will be
noted that the maximum is about 6,000 lb. per sq. in., less than
half of that found when the deflection was neglected. In the
propeller under consideration, therefore, and the same applies to
most propellers, it is hopeless to get even a reasonably accurate
approximation to the actual stresses without a fairly accurate
knowledge of the deflections under load.
The fact that the forward deflection of the blades in operation
reduces the bending stresses suggests the possibility of eliminating
the bending stresses entirely by so inclining the blade to start
with that the bending moments due to centrifugal force just
balance those due to the air load. This can be done for one set
of operating conditions (i.e., one value of V/nD) only, but if
average conditions are taken the stresses in the propeller can be
substantially reduced, and clever designers have for years taken

advantage of this fact. It is more important with thick rigid
propellers than with relatively thin and flexible propellers which
tend to take the proper inclination when operating.
These stresses were found for the propeller turning 1,700 r.p.m.
which corresponds to a tip speed of 795 ft. per sec. Since the
stresses vary as the square of the tip speed (at constant V/nD),
and since the fatigue limit of the aluminum alloy used is in the
neighborhood of 12,000 to 15,000 lb. per sq. in., it should be safe
according to the stress analysis to run geometrically similar
propellers at the same V/nD, at tip speeds up to about
which corresponds to about 2,400 r.p.m. with our particular 8

ft. 11-in. propeller. Actually, due partly to the fact that the
air load fluctuates, partly to the fact that vibrations may increase
the stresses, and partly to the questionable accuracy of the
stress analysis, the safe limit for indefinitely long operation is
probably somewhat less than the tip speed of 1,125 ft. per sec.
indicated.
The stress analysis of our example was made under the condi
tion of full-throttle climbing flight. At higher full throttle air
speeds the revolutions, and consequently the centrifugal force,
would be higher, but the air load would be lower, while at lower
speeds the reverse is true. The maximum calculated stresses
are therefore not ordinarily greatly different at the various full-
throttle speeds, although the case of full power at zero advance
seems the most severe in practice.
Flutter and Weave of Propeller Blades.Flutter is a form of
severe vibration of the propeller blades. It is usually considered
a torsional vibration in which the blade angles increase and
decrease rapidly, but it may be connected with bending, for
fore-and-aft deflections are ordinarily accompanied by a change
of angle also. Flutter is accompanied by an unpleasant increase
in noise which is sometimes of a rapid hammering nature. It
sometimes sets up uncomfortable pulsations in the air for some
little distance from the propeller. Both tests and practical
experience show that severe flutter is accompanied by a suffi
ciently great stress, or variation of stress, to cause failure if
continued long enough.
Weave is a form of comparatively slow waving backward and

forward of the propeller blades while turning. It is much less
severe than flutter but is also accompanied by a stress variation
of some magnitude. This is apparent from our stress-analysis
example, in which a forward deflection of about % in. at the tip
changed the maximum fiber stress from 13,080 to 6,000 lb. per
sq. in. A weaving motion of only ^-in. total amplitude would
therefore be accompanied by a change in stress in the neighbor
hood of 1,500 to 2,000 lb. per sq. in. in this propeller.
There are also vibrations other than those associated with
flutter and weave. Propellers, like all elastic bodies, have
natural periods of vibration, and the vibrations may build up to
alarming proportions if the natural frequency corresponds to
some frequency of impulse of the engine or airplane, such as that
of the torque impulses. The natural frequency of geometrically
similar propellers has been found by both theory and experiment1
to vary inversely as the diameter. If the frequency of the
propeller and that of the torque vibrations coincide, a propeller
of smaller diameter might be used, in which case if the diameter
were small enough the vibration frequency would be above the
range of the torque impulses, thus preventing resonance.
Vibrations are also caused by (1) unbalance of the centrifugal
and air loads which is present if one blade is heavier than the
other and the propeller does not balance statically, or if angles and
shape of both blades are not the same so that one blade carries
more air load than the other or others; (2) periodically varying
air loads which occur because of unsymmetrical air flow about
the axis of rotation due to the proximity of unsymmetrical
fuselages, wings, and landing gears;2 (3) periodically varying
air loads due to the inclination of the axis of rotation to the flight
1 The investigation of vibration periods was made by the Army Air Corps
at McCook Field, Dayton, Ohio.
2 One particularly bad case of periodic fluctuation of the air load has been
found in large airplanes having three side-by-side engines, the center engine
being located somewhat ahead of the two wing engines, and the wing-engine
propellers operating partly in the slipstream of the center engine. Even if
the overlapping is not large, severe pulsations are set up in the side propellers
when they get into phase with the center propeller, and few propellers last
long under such conditions, micarta being about the best because of the
combination of high fatigue limit with homogeneity. Other cures for the
situation have been to use a three-bladed center propeller and two-bladed
side propellers so that the propellers will always be out of phase, or to move
the side engines farther out.
path, particularly in climbing flight, for as a blade goes up it

has a smaller angle of attack than normal and as it goes down it
has a larger one; (4) irregular variations of the air, such as gusts.
Since the propeller axis is seldom exactly parallel with the
direction of flight, and also since the air is seldom free from gusts
and currents, propellers practically always operate under more
or less fluctuating loads, and therefore the fatigue limit of the
material should not be exceeded even by the peak stresses if a
propeller is to run indefinitely without failure. Unfortunately
with present methods it is not possible to calculate the stress
fluctuations, propellers usually being designed so that the calcu
lated stresses are well under the fatigue limit. The only final
criterion of strength is actual operation over an extended period
of time, although a fairly satisfactory substitute is an overload
whirling test for a period of several hours. After propellers of a
certain form have been thoroughly tested and their real strength
and endurance are known, it is fairly safe to apply the ordinary
stress analysis to new propellers of generally similar form, but it
cannot be applied with any assurance to new forms.
Fatigue Failures.Most of the few failures of conventional
forms of aluminum-alloy propellers are fatigue failures which
started at some flaw or irregularity in the surface. It is well
known that fatigue failures due to fluctuating stresses are likely
to start in cracks or irregularities, such as sharp internal corners,
even though the actual stresses are very low. Figure 162 shows
an example of fatigue failure in the blade root of an aluminum-
alloy propeller caused by a sharp internal corner just outside the
steel hub. The blade had been hand filed to fit a hub to which
it did not belong, and a small but sharp internal corner was left
in it. The failure started as a crack (at top of the photograph)
which gradually worked its way in toward the middle until the
effective area was reduced to a point where the centrifugal force
pulled the blade off. It will be noticed that at the top of the
picture the metal is bright, due to rubbing. The hole in the
center was filled with lead for purposes of balance.
Even lettering stamped on a propeller blade has caused com
plete failure due to the notch effect in the surface. The surface
of all propeller blades should be smooth and fair with no sharp
curvatures along the blades and no irregularities of any kind.
Since fatigue failures start as small cracks which gradually
develop in size, it is common practice to etch aluminum-alloy
propellers after every 100 hr. or so of service and to examine them

for cracks. The blades are swabbed with, or preferably immersed
in, a 10 to 20 per cent solution of commercial caustic soda of the
same temperature as the propeller, for from two to six minutes.
They are then rinsed in water and swabbed with or immersed in
a 5 per cent solution of nitric acid until bright, after which they
are thoroughly washed with water and dried. The surface is
carefully inspected for cracks, preferably with a low-power
Fig. 162.-Example of fatigue failure in blade root of aluminum-alloy propeller.
magnifying glass. At any cracks or flaws found the surface is

sanded off a little and given a deeper local etch. Small longi
tudinal seams which are rolled into the material are not impor
tant, but cracks across the blade which penetrate under the
surface are usually a sign of impending fatigue failure and are
cause for rejection.
Failures of Wood Propellers.Good propeller woods have
great toughness, and the fatigue limit is very near to the ultimate
strength. Practically no wood propellers of reasonably normal
form fail in direct tension or compression, and of those few that
do, the failure can usually be traced to a poor grade of material.

Wood is not so uniform in its properties as metal, and it is not so
easily tested, so that occasionally even with rigorous inspection
some poor wood will creep in.
Usually wood propellers fail by the blade splitting longitudi
nally near the tip. Because of their grain structure, wood pro
pellers are weak in torsion, and so large twisting moments should
be avoided. Flutter in wood propellers will nearly always cause
longitudinal splitting if continued long enough. A slight amount
of sweepback has been found to make for smooth running without
nutter and is usually used with wood propellers.
If wood propellers are thick and heavy enough to be reliable
and hold their shape under various weather and operating con
ditions, and also to withstand normal handling, they are usually
sufficiently strong for ordinary requirements with tip speeds up
to about 900 ft. per sec.
Destructive Whirling Tests.Due to the effect of flutter and
fluctuating air loads, as well as to the complicated nature of the
stresses in propellers and the inability to calculate them with a
reasonable assurance of accuracy, the only real criterion of pro
peller strength and endurance is actual operation and use. This
not only takes considerable time under ordinary operating
conditions, but also it may be highly dangerous since a serious
propeller failure in flight, in which either one whole blade or a sub
stantial part of a blade breaks off, always results in the remaining
unbalanced force tearing the whole or a large part of the engine
completely out of the aircraft.
It has therefore been the practice to test the strength and
endurance of propellers by whirling them by means of electric
motors in armored stands or on an oversized gasoline engine.
The stands for the electric motors are usually completely enclosed
to the ground, making the air flow past the propeller unsym-
metrical, the propeller consequently being subjected to a fluctuat
ing air load. Instead of testing a propeller for several hundred
hours at the maximum power for which it is designed, it is whirled
at an overload for a shorter period of time. The usual test is
ten hours without failure of any kind at 100 per cent power over
load for metal propellers and at 50 per cent power overload
for wood propellers. If the propeller develops a violent
flutter at any speed, it is run for an additional ten hours at
that speed.
Nearly all of the electric whirling tests made in this country

have been made by the Army Air Corps at Dayton, Ohio. New
whirling equipment is now being completed at Wright Field
near Dayton, on which propellers up to 40 ft. in diameter can be
tested while absorbing up to 6,000 hp., and ordinary-sized
propellers can be turned as high as 5,000 r.p.m.
If propellers are whirled with an oversized airplane engine,
the requirement is usually ten hours at 75 per cent power over
load for metal propellers and 40 per cent for wood propellers.
CHAPTER XVI
SUMMARY OF FACTORS TO BE CONSIDERED IN THE

DESIGN OF A PROPELLER
Before undertaking the actual design of a propeller to give a

certain performance, it is well to summarize the pertinent factors
which have been considered in detail in the previous chapters.
Particular Airplane Performance Desired.As shown in
Chap. X, an airplane will have somewhat different performance
characteristics if equipped with different propellers of the same
geometric form, the diameter, and pitch onlybeing varied in such
a manner that the revolutions at maximum speed remain the
same.. AT certain combination of diameter and pitch will give
the highest maximum speed, and certain others will give the best
cHmbângle of climb, take-off, cruising economy, etc.
High Speed.^Ôrdinarily the design of a propeller for a racing
airplane represents the simplest and most straightforward case,
for~the~only ~requirement is that the propeller should be that
giving the highest possible propulsive efficiency at the maximum
speed condition,, In most cases the performance in take-off and
cTfinb will be reasonably satisfactory. There have been a few
cases-, however, in which racing seaplanes have been unable to
leave the water with propellers designed to give the maximum
possible speed. The condition is difficult, for with engines
delivering over 1,000 hp., the propeller must turn at a reason
ably low revolution speed to avoid a high tip-speed loss, and
with airspeeds of over 300 m.p.h., a very high-pitch propeller is
required, resulting in a very low static thrust and poor take-off
qualities. The best solution of the propeller phase of the prob
lem, assuming that the revolutions are fixed and that a fixed-
pitch propeller is to be used, is probably to increase the diameter
and blade-width ratio and decrease the pitch as much as is neces
sary to get a satisfactory take-off.
Rate of Climb.The propeller giving the highest rate of climb,,^
if it is one of a family of given form, can be selected by calculating!
the performance for a series of pitches and diameters, each hold- ;
ing the engine to the same full-throttle revolutions at maximum

speed, as was done in Chap. X for a series of metal and a series of
wooden propellers. With the relatively thin-bladed metal
propellers the pitch and diameter giving the maximum rate of
climb are approximately the same as those giving the greatest
high speed. All combinations of pitch and diameter throughout
quite a large range will, however, give very nearly the same rates
of climb.
With the thicker wood propellers the best climbing propeller
is not the one which also gives the greatest speed but is one
having a considerably larger diameter and smaller pitch. In
our example of Chap. X, the wooden propeller giving the greatest
rate of climb would have had a pitch-diameter ratio of about
0.56 and a value of V/nD at high speed of about 0.65, from which
_ <? = YlnD J 0 65
P/D 0.56
= 1.16,
and the nominal slip S would have a negative value of 16 per cent.
A negative slip of 10 per cent would give practically the same
climb, and this value may be taken as an approximate one to
which good climbing wooden propellers of ordinary pitch ratios
may be designed.
Take-off and Angle of Climb.Good take-off characteristics
and a steep angle of climb are of great importance, and in practi
cally all commercial and the majority of military airplanes, these
characteristics overshadow the rate of climb. In fact, when a
"good climbing propeller" is ordered, what is almost always
desired is a propeller which will give a short take-off run and a
steep angle of climb and not a high rate of climb. It happens
that with both wood and metal propellers, the ones having the
largest diameters and the lowest pitches which will give a reason
able high speed give both the quickest take-off and the greatest
angle of climb.
Cruising Speed.The propeller which will give the highest
cruising speed is within all practical limits the same as that
giving the highest full-throttle speed in level flight. This is due
to the fact that as the motor is throttled in level flight, the air
speed is reduced in very nearly the same proportion as the revo
lutions. In fact, if the maximum speed and revolutions of an
airplane are known, the cruising speed at any revolution speed is
FACTORS IN THE DESIGN OF A PROPELLER 251
easily and quite accurately obtained from the relation that a 10

per cent decrease in revolutions results in an 11 per cent decrease
in air speed, a 20 per cent decrease in revolutions resulting in a
22 per cent decrease in air speed. The propeller in cruising
flight therefore works at about the same values of C, and V/nD
as it does at maximum speed, and with the revolutions at maxi
mum speed fixed, the propeller giving the highest full-throttle
speed will also give the highest cruising speed at any particular
revolutions.
All-around Performance.If, as is ordinarily the case, good
take-off and angle of climb are considered as having about equal
importance with high maximum and cruising speeds, a compro
mise or general service propeller is selected. This will of course
have a larger diameter and smaller pitch than the best high-speed
propeller but a smaller diameter and greater pitch than the
propeller giving the shortest take-off and the steepest climb.
Fortunately a propeller about halfway between the two extremes
will give a speed very nearly as high as the best high-speed pro
peller and at the same time give a take-off nearly as short as the
best take-off propeller. The compromise or general service
propeller, whether of wood or metal, may be taken as being the
one which operates at its maximum or peak efficiency at the
high-speed condition of flight. The efficiency here referred to is
the propulsive efficiency of the full-scale propeller operating with
body interference. With wood propellers this peak efficiency
usually occurs at a nominal slip value between zero and negative
5 per cent, while with metal propellers it comes between zero and
positive 10 per cent.
Special Performances.For special performances not men
tioned above, such as high-altitude flying with a supercharged
engine, no simple general rules are available and it is usually
preferable to calculate the performance with a series of at least
three propellers, as was done in Chap. X, and then choose the
best.
The above discussion has applied only to airplane propellers.
The problem of designing propellers to give maximum perform
ance on lighter-than-air craft is simpler, for only the high-speed
condition need be considered. In this case, where there is no
induced drag due to dynamic lift to be considered, the forward
speed varies directly with the revolutions when the engines are
throttled, so that the propeller works at exactly the same values
of C, and V/nD at all steady speeds. The best high-speed

propeller is therefore the best at all throttled speeds. If thin
metal propellers are used the same propeller is also the best,
even though some of the engines on an airship are shut down
and the ship is cruised with the remaining ones running at or
near full throttle, as is sometimes done to improve the fuel
economy. (__
Limitations Imposed by Airplane and Engine.f-The diameter
of a propeller is often limited by insufficient ground clearance,
interference with a part of the structure such as a seaplane float,
or interference with another propeller or its slipstream on a multi
engined airplane. The present Department of Commerce
requirement for ground clearance on a landplane is 9 in. in flying
position, and on a seaplane the propeller is required to have 18
in. clearance from the surface of the water. One inch is given as
the minimum allowable distance between the propeller tips and
a body such as a fuselage, boat hull, or float, and more is desirable.
In cases where propellers are operating side by side in the same
plane with no body between them, they should clear each other
by a minimum of about 6 in.
Where side propellers are located somewhat to the rear of a
center propeller, as is usually the case with airplanes having three
engines, care should be taken that they have sufficient clearance
to be essentially free of the slipstream of the front propeller.
If there is an appreciable amount of overlap the blades of the
side propellers are subjected to a severe periodic fluctuation of
air load which is almost certain to cause failure sooner or later.
In this connection it should be remembered that although the
slipstream of an isolated propeller necks down behind the plane
of rotation, the center propeller is likely to be working on the
nose of a rather large body which actually spreads the slipstream
somewhat. It is therefore desirable that the projected propeller
discs have a few inches' clearance between them.
In the cases where the diameter is limited, there are several
courses which may be followed. If take-off is not of great impor
tance, a small high-pitch high-speed propeller may be satis
factory. If take-off is important it may be advisable to use a
propeller with extra-wide blades or, in extreme cases, with three
or four blades.
Choice of the Number of Blades.Two-bladed propellers are
used in all ordinary cases, for the fewer the blades the lighter,
cheaper, simpler, and more efficient will be the propeller; and two
is the smallest number of blades with which proper balance of
mass and air forces can be obtained. Three or four blades are
used if the diameter is limited to a size where a two-bladed pro
peller is unsuitable, or in cases where the operation is not smooth
with a two-bladed propeller due to unsymmetrical body inter
ference or the slipstream of another propeller. If there is an
unsymmetrical airflow through a two-bladed propeller, ordinarily
one blade is at its highest angle of attack and load at the same
time that the other is at its lowest, and this is likely to cause
undesirable vibrations. With three or more blades the uneven
load distribution is spread more evenly around the propeller disc
and the vibrations are greatly reduced. Vibrations are also set
up in two-bladed propellers when the airplane is turning, due to
the varying gyroscopic moment of the two-bladed arrangement
and, when the airplane is sideslipping, due to the uneven air
loading. These effects are too small to be noticed in ordinary
installations, but with very large and slowly turning geared
propellers the pulsations may become so great as to necessitate
the use of three or more blades, which will remedy the situation.
Vibration difficulties considered, it is usually advisable to have
three or more blades in the rear propellers of tandem series, in
large geared propellers, in propellers operating very close to a
body which is far from symmetrical about the axis of rotation,
and in propellers which operate partially in the slipstream of
other propellers.
Metal propellers are seldom made with more than three blades,
but wooden propellers are usually made with either two or four,
due to the ease of construction.
Resume of Factors Affecting Propulsive Efficiency.In order
to collect the various factors, such as the geometric proportions
of propellers, which are directly connected with the propulsive
efficiency, the important ones are given briefly here.
Pitch Ratio.For a series of propellers differing only in pitch,
the maximum or peak efficiencies become greater as the pitch
is increased, within the range of present-day aircraft practice.
Since the propeller must operate reasonably near its peak effi
ciency range at the maximum speed of the airplane, the pitch
depends largely on the relative values of the maximum air
speed, the revolutions, and the power absorbed by the propeller,
all of which are conveniently combined in the coefficient C.
The higher the value of C again within the range of present-

day practice, the greater is the maximum possible efficiency
obtainable.
After the speed, power, and revolutions are fixed, and these
are usually fixed by the airplane and engine designers, the pro
peller designer still may select from a fairly large pitch range to
stress the particular kind of airplane performance desired, as
outlined in the foregoing sections on performance.
Pitch Distribution.If the pitch of a propeller operating on a
body is smaller near the hub than at the tip, a somewhat higher
propulsive efficiency is obtained than if the pitch is uniform over
the entire blade. The maximum difference is of the order of
three or four per cent with average metal propellers.
Blade Width and Number of Blades.The total width of all the
blades of a propeller is the critical factor affecting the perform
ance, regardless of the number of blades or the width of each.
Thus a four-bladed propeller will have the same performance as
an otherwise similar two-bladed propeller if the blades of the
two-bladed one are twice as wide, making the total width the
same. With blades of average width (WR = 0.053), the maxi
mum efficiency of a two-bladed propeller is ordinarily about 3
per cent greater than that of a three-bladed propeller of the same
diameter and about 7 per cent greater than that of a four-bladed
one. A two-bladed propeller having extremely narrow blades
(WR = 0.02) is about 2 per cent more efficient than the average
one. These differences are somewhat less for high-pitch pro
pellers and greater for low-pitch propellers. Also, they are
slightly greater at the climbing range than at the high-speed or
maximum efficiency range.
The above values are for propellers of the same diameter and
pitch. Actually if a four-bladed propeller replaces a two-bladed
propeller on the same airplane and engine, it will have about the
same pitch but a smaller diameter and consequently a higher
pitch ratio. This higher pitch ratio is accompanied by an
increase in maximum efficiency which partly offsets the decrease
due to the greater width ratio.
Thickness Ratio.Ordinary aluminum-alloy propeller blades
have thickness ratios about one-half as great as those of ordinary
wood propellers, and if the R.A.F-6 type sections are used, the
thinner metal propellers are from 4 per cent (with high pitches)
to 7 per cent (with low pitches) more efficient than the wooden
ones. If Clark-Y type sections are used the effect of thickness is

less pronounced.
Blade Section.Of the two airfoil sections most commonly used
in propellers, the Clark-Y type gives somewhat better effi
ciencies at low angles of attack while the R.A.F-6 type is superior
at high angles of attack. This gives the Clark-Y type an advan
tage in low-pitch propellers in which high angles of attack are
not reached but gives the R.A.F-6 type an advantage in high-
pitch propellers with which it has a much better take-off and
climbing efficiency. Also, in the high-pitch propellers the max
imum efficiencies are practically the same with either type of
section. All-around performance being considered, it may be
said that the Clark-Y type sections are better with low-pitch
propellers while the R.A.F-6 type are better with high-pitch
propellers.
Plan Form.If the blades are given some taper and the tip is
rounded, the exact shape or degree of taper has no appreciable
affect on the propulsive efficiency. The same is true of sweep-
back and rake or forward tilt.
Tip Speed.Ordinary thin-bladed aluminum-alloy propellers
can be operated at tip speeds up to about 1,000 ft. per sec. with
out loss of propulsive efficiency, but as the tip speed exceeds this
critical value the efficiency drops off very rapidly. With the
thicker wooden propellers the critical speed is lower and the loss
above it greater.
Body Interference.The propulsive efficiency, which includes
the increase of body drag due to the propeller as well as the effect
of the body on the propeller, ranges for ordinary bodies from the
same as to about 7 per cent less than the efficiency of the isolated
propeller. For most bodies the value is about 2 or 3 per cent.
Roughly, the lower the drag of the body the worse will be the
effect on the propulsive efficiency. It is advantageous to have a
large propeller diameter with respect to the mean body diameter,
and for the average case it may be said that the propulsive
efficiency increases about 1 per cent for an 8 per cent increase of
the ratio of the propeller diameter to the body diameter.
Strength.One of the most important and difficult problems
in designing a propeller of a new shape or form of construction is
to establish thoroughly its ability to operate indefinitely without
failure. This usually requires careful stress analyses and over
load whirling tests. With geometrically similar propellers of
proved form, however, since the stresses vary as the square of the
tip speed regardless of the size, a limiting tip speed can be found
below which any propeller of the family can be operated safely.
This may even be applied to different pitch settings with reason
able accuracy if the actual tip speed at maximum air speed is
used, and not merely the tip speed in the plane of rotation. For
the commonly used family of aluminum-alloy propellers shown
in Figs. 176 and 177, the limiting tip speed below which the
calculated stresses will not exceed the fatigue strength is about
1,100 ft. per sec. For ordinary, routine design procedure it is
more satisfactory to use the more conservative value of 1,000 ft.
per sec. Thus if a propeller of this form operates under the
critical tip speed for efficiency loss at the maximum speed of the
aircraft, it should be safe from the strength point of view. It
should also, however, operate smoothly, without flutter or exces
sive vibration.
CHAPTER XVII
DESIGN PROCEDURE, WITH CHARTS AND EXAMPLES
In this chapter a simple method of design procedure is given

along with accurate working charts which may be used for the
selection of metal propellers of a standard form. An approximate
method is also given for using the same charts for propellers
of different form. The use of the methods to select propellers
for various performances is shown by means of examples.
Information Furnished Propeller Designer.The primary
data required for the design of a propeller are of course the correct
brake horsepower, the revolutions at which that power is
obtained, and the maximum air speed. One of the most difficult
problems of the propeller designer has been to obtain accurate
values of the power and air speed for new engines and airplanes,
due partially to a lack of accurate tests on the new types and
partially to the optimistic results usually expected by the makers.
If the engine actually delivers less power than the propeller is
designed to absorb, the propeller will hold the revolutions down
to a lower value than desired, which in turn cuts down the horse
power output and consequently the airplane performance. The
lower high speed obtained causes a further reduction in revolu
tions which tends further to reduce the speed. The same con
ditions occur if the airplane fails to make the speed designed for,
even though the engine power is up to standard. It is therefore
necessary to get accurate power and speed data if the propeller
is to perform satisfactorily.
It is better to design the propeller from the power-required
curve of the airplane than merely the anticipated maximum
speed, for with it and the propulsive efficiency of the particular
propeller designed, the power available and power required can
be balanced and the high speed determined more accurately.
An accurate power-required curve is not, however, ordinarily
available.
One method, and usually the most satisfactory one, of finding
accurate values of the power and speed is to run high-speed tests
257
of the airplane in question, with a propeller of known charac

teristics. The speed is then accurately determined over a timed
course and the engine power can be calculated from the propeller
characteristics.
The propeller designer should also know the kind of perform
ance desired from the airplane, the form of the body and engine
and the relative location of the propeller, and any limitations,
such as the maximum permissible diameter.
A Simple System for Selecting Propellers of a Standard
Form.-Due to th limitations imposed by the material used,
including such factors as weight, strength, and satisfactory
smooth operation, all propellers made of one particular material
will naturally fall -within a fairly small range of blade width and
thickness ratios, for any ordinary performance required. Since
small changes in the width and thickness ratios have no important
effect on the propulsive efficiency, it is quite reasonable to give
the same blade form to all ordinary propellers of a given material.
Thus both the designing and manufacturing operations are
greatly simplified.
A convenient method of selecting propellers of a given family
to operate under any particular performance conditions is the
one based on the coefficient C,,1 which is outlined in Chap. VI.
The coefficient C, involves only the conditions under which
the propeller must operatei.e., the power, revolutions, air
speed, and air densitybut has no connection with the dimen
sions of the propeller itself. The value of C can, therefore,
be computed from the desired performance factors alone. In
engineering- units and for standard sea-level density its value
may be easily found from the equation
r = 0.638 X MPII
' HP* X RPMX"
the values of HP^ and RPM$~ being obtained from scales given
in Fig. 163 (also Fig. 53).
Working charts, in which curves of V/nD and propulsive
efficiency are plotted against C, for even blade-angle settings,
are given for a standard form of aluminum-alloy propeller
operating with six different engine and fuselage combinations in
1 The Eiffel logarithmic diagram is also convenient for this purpose but
requires the use of simple drawing equipment.
DESIGN PROCEDURE, WITH CHARTS AND EXAMPLES 259
5000^30
r^4.0
1000-F
900- r 4500-
800- i
25
700-L 4000 -:
600-
3500-^zs
50CH:E-3.5 it
40 0-. 3000-=
4-24
30CHF 2500-
E-3.0 2'i
200-F
* HPi/5 2000
HP RPM 1900 R P M 2/S
1800 ir 20
150-F 1700
1600
1500
2.5 1400
100- 18
90- 1300
80- 1200 17
70- pV5
1100
Pn 2 )
60- 16
1000
and in Engineering
50-
Units for Standard 900 15
40- Atmosphere at 800
Seoi Level
E-2.0 14
30Ê 0.638 x M P H 700 -i
HPVsxR PM2/s 600 -.r 13
20^
500 -+- 12
400 -Hi
-t-1.5 300 t-
Fig. 163.Scales for finding HPH and RPMH.
Figs. 164 to 175. These charts are drawn from full-scale wind
tunnel test data published by the N.A.C.A.1
A drawing of the blade of the 9-ft. propeller used in all of the
tests is given in Fig. 176, and the blade form is given in non-
dimensional ratios which can be applied to any diameter, in Fig.
177. Airfoil sections based on the R.A.F-6 are used. The
various pitches were obtained by merely turning the whole blade
in the hub, the pitch being approximately uniform for a pitch-
diameter ratio of 0.5. As shown in Chaps. VII and IX, this
Fig. 164.Open-cockpit fuselage with 400 hp. Curtiss D-12 engine. No

radiator (corresponds to case with wing radiators). Smoothly faired nose.
Maximum cross-sectional area, 11.6 sq. ft. Data in Fig. 165.
combination gives very close to the best pitch distribution

for operation with an ordinary body over a large range of pitches.
In fact, propellers of this standard form will give efficiencies which
are within ordinary practical limits as high as can be obtained
with any form capable of filling the same strength and per
formance requirements.
Torsional Deflection of Blades.As was brought out in Chap.
XV, propellers deflect and twist under load, so that the pitch of
an operating propeller is often quite different from the pitch of
1 Working Charts for the Selection of Aluminum-alloy Propellers of a
Standard Form to Operate with Various Aircraft Engines and Bodies, by
Fred E. Weick, N.A.C.A.T.fl. 350, 1930. For a method of selecting wood
propellers of a standard form, see Propeller DesignIII: A Simple System
based on Model Propeller Test Data, by Fred E. Weick, N.A.C.A. T.N. 237,
1926. Also, Simplified Propeller Design for Low-powered Airplanes, by
Fred E. Weick, N.A.C.A.T.N. 212, 1925.
Fig. 165.Open-coekpit fuselage with D- 12 engine. Propeller 4412.

the same propeller in the static condition where there is no load.

It was found in the test with the 400-hp. D-12 engine installation
shown in Fig. 164 that the deflection was not sufficient to affect
the propeller characteristics to an appreciable extent for powers
up to 200 hp., but above that amount the power and thrust
coefficient become greater for a given V/nD, and the efficiency
drops off slightly. In order to make the results of the tests with
all engines and bodies comparable, the tests from which the work
ing-chart data were taken were run with the D-12 engine throttled
to 200 hp., which was approximately the power of the other
engines.
Fig. 166.Complete VE-7 airplane with wings and tail surfaces. Open-
cockpit fuselage with 180 hp. Wright E-2 water-cooled engine and nose radiator.
Maximum cross-sectional area of fuselage 9.6 sq. ft. Data in Fig. 167.
The change in propeller characteristics which occurred at the

higher powers can be accounted for by considering that the blades
twisted to increase their pitch. This increased pitch is substan
tiated by the fact that the thrust and efficiency coefficients
obtained with the high powers are about the same as those
obtained with lower powers but at slightly higher pitch settings.
Also, deflection measurements which were taken during the
tests show that the blade angles increased with increase of power,
but the measurements were unfortunately not sufficiently accurate
to use as a basis for showing the exact variation.
Considering the variation of the propeller coefficients with
power as being due entirely to deflection, the working charts can
be satisfactorily used for engines of all powers if only the deflection
Fiq. 167.VE-7 open-cockpit biplane. Propeller 4412.

in operation is known. It is necessary only to consider the blade

angles as those existing under operating instead of static
conditions.
Although accurate deflection data covering a large range
of powers, bodies, and propellers are not available, a useful
approximate rule for direct-drive propellers similar to the design
used in these tests has been based on the data obtained with the
D-12 engine. This rule is that the working charts may be used
without considering deflection in operation for powers up to 200
hp., but above 200 hp., the average blade angle increases at the
Fig. 168.-Open-cockpit fuselage with J-5 engine and conventional cowling.

Maximum cross-sectional area of fuselage, 11.0 sq. ft. Data in Fig. 169.
rate of % deg. for each increase of 100 hp. This makes an

increase of 0.5 deg. for an engine of 300 hp., 1.0 deg. for 400 hp.,
and 1.5 deg. for 500 hp.
This approximate rule has worked quite satisfactorily in select
ing hundreds of direct-drive propellers to fit commercial airplanes.
For propellers geared to run at one-half the engine revolutions
the angular deflection is in the neighborhood of half as large.
Use of Working Charts.-In order to find the diameter and
pitch setting of a propeller of this form for any particular set of
operating conditions, it is merely necessary to
1. Calculate the value of C, for the conditions under which the
propeller is to operate.
2. Choose the pitch setting for the propeller operating at the
desired portion of its efficiency curve (depending on the airplane
performance desired) and the above C,.
0.5 0.6 0.7 0.5 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 Z.O
Fig. 169.Open-cockpit fuselage with J-5 engine. Propeller 4412.
3. Find the V/nD for the above C, and pitch setting, from the
lower curves.
4. Knowing V/nD, n, and V, calculate D.
If the diameter of the propeller is fixed to start with, V/nD is
also fixed, and the pitch setting can be found directly from the
curves of V/nD vs. C,.
Examples.Several typical examples showing the use of this
system for the selection of propellers of this standard form are
given here.
Fig. 170.Cabin monoplane with J-5 engine. No cowling over cylinders or

crankcase. Maximum cross-sectional area of fuselage alone, 21.3 sq. ft. Data
in Fig. 171.
Example 1.For our first example we shall select a propeller

for a small open-cockpit sport airplane, with a conventionally
cowled radial air-cooled engine, delivering 150 hp. at 2,000 r.p.m.
The maximum speed expected is 115 m.p.h. High speed is of
greater importance than the other performance characteristics,
but a fair take-off is also desired.
First solving for C,, we find that
r = 0.638 X MPH
' HP* X RPM*
= 0.638 X 115
2.72 X 21.0
= 1.29,
the values of 150^ and 2,000% having been obtained from Fig.
163.
Fig. 171.Cabin monoplane with J-5 engine. Propeller 4412.

Then from Fig. 169, which is the working chart based on body
conditions similar to those specified, we find that the maximum
possible efficiency, and therefore the highest speed, will be
obtained for a C, of 1.29 with a propeller operating at a V/nD
of 0.705. This will, however, give a poor take-off, as was shown
in the example in Chap. X. A good average take-off would be
obtained with a propeller operating at the peak of its efficiency
curve at the high-speed condition, but this would mean a slight
sacrifice of speed. The V/nD for the propeller operating at
Fig. 172.-Cabin fuselage with J-5 engine. Large amount of conventional

cowling, leaving only the top portions of the cylinder heads and valve gear
exposed. Maximum cross-sectional area, 21.3 sq. ft. Data in Fig. 173.
the peak of its efficiency curve and at a value of C, of 1.29 would

be 0.645 (also from Fig. 169). Now it happens that with a
propeller about halfway between these two the take-off will be
very nearly as good as with the peak-efficiency propeller, while at
the same time the maximum speed will be within all practical
limits (about 0.1 m.p.h.) as high as with the best-speed propeller.
We shall therefore choose a propeller with a maximum speed
V/nD about halfway between 0.645 and 0.705, or 0.675. Then
from Fig. 169 the blade angle at 0.75R is 19.7 deg.
Knowing the V/nD, the revolutions, and the velocity,
T) = 88 X MPH
RPM X (V/nD)
88 X 115
2,000 X 0.675
= 7.50 ft.
.V 0.9,
nD
Fig. 173.-Cabin fuselage with J-5 engine. Propeller 4412.

Thus our propeller will have a diameter of 7 ft. 6 in., a

blade angle of 19.7 deg. at 75 per cent of the tip radius, and
from the upper curves of Fig. 169 the propulsive efficiency
is 0.818 for the airplane without wings. With ordinary biplane
wings this would be reduced, as shown in Chap. IX, to about
0.805.
Also, from Fig. 179 the weight with an average hub will be 35
lb., and from Fig. 178 the blade angle at the 42-in. radius, which
is used as a standard,1 is found to be 17.3 deg.
Fiq. 174.Cabin fuselage with J-5 radial air-cooled engine and N.A.C.A. type
complete cowling. Maximum cross-sectional area 21.3 sq. ft. Data in Fig. 175.
If a propeller giving the best possible take-off and angle of

climb were desired without a large sacrifice of high speed, a
propeller of greater diameter and lower pitch would be selected.
As shown by the example in Chap. X, a propeller with an angle
setting 2 or 3 deg. lower than that of the propeller operating
at the peak of its efficiency curve at maximum speed will give
just about as good a take-off as can be obtained with any, while
at the same time it represents about the limit in the low-pitch
direction if a substantial decrease in high speed is to be avoided.
Now from Fig. 169, the propeller operating at the peak of its
efficiency curve at C, = 1.29 has a blade angle at 0.755 of 17.4
deg. Our take-off propeller will therefore have an angle of, say,
1 For propellers under 8 ft. in diameter it is better to use a smaller radius,

preferably the even 6-in. station nearest 0.75ft.
Fig. 175.Cabin fuselage with completely cowled J-5 engine. Propeller 4412.
Section,
taken
looking
from Hub 2.S7S J^~ T*.." n'M
Curve ofMaximum Thickness
Fig. 176. Metal blade for 9-ft. diameter standard propeller 4412.
b h.
D b
0.075 0.30
0.070 0.28
0.065 0.26
0.060 0.24
0.055 0.22
0.050 0.23
0.045 0.13
0.040 0.16
0.035 0.14
0.030 0.12
0.025 0.10
0.020 0.08
0.015 0.06
0.010 0.04
0.005 0.02
0 0
Fig. 177.Blade-form curves for standard propeller.

15.0 deg. at 0.75R, the V/nD (from Fig. 169) is 0.61, and the
diameter is
88 X 115
2,000 X 0.610
= 8.30 ft.
The propulsive efficiency at maximum speed is about 5 per cent
less with this propeller than with the other, and if the maximum
9 10 11
Diame+er in Ft.
Fig. 178.Difference between the blade angle at a radius of 42 in. and
that at 75 per cent of the tip radius, for propellers of the form shown in Figs.
176 and 177.
speed of 115 m.p.h. is obtained with the first propeller, the speed
with the take-off propeller would be about 0.37 X 5 per cent =
1.9 per cent less (see Chap. X), or about 113 m.p.h. Using this
maximum speed the value of C, is 1.26, and for a blade angle of
15 deg. at 0.75/2 the V/nD is 0.600, and the diameter is
88 X 113
D =
2,000 X 0.600
= 8.30 ft.
Thus a difference of only 2 m.p.h. does not make an appreciable

difference in the propeller selected.
The tip speeds of these two propellers, as can be seen from Fig.
85, are both well below the critical speed of 1,000 ft. per sec,
and so they will have no tip-speed loss and should be satisfactory
in regard to strength.
300
200
5
too
0
Diame+er in Ft.
Fig. 179.Approximate weights of two-bladed aluminum-alloy propellers
of the form shown in Figs. 176 and 177; based on actual propellers with hub
and blade root sizes most likely to be used.
Example 2.A propeller is desired for a load-carrying transport

monoplane having a maximum speed of 135 m.p.h. with an
uncowled radial air-cooled engine delivering 575 hp. at 1,950
r.p.m. It is important that the airplane be given as short a take
off as possible as well as the highest possible cruising speed.
Solving for C we get
= 0.638 X 135
* 3.57 X 20.7
= 1.17.
Then from Fig. 171 we find that for the propeller operating at
its peak efficiency at maximum speed, and this is about the best
compromise propeller giving both a good take-off and a good

cruising speed, the V/nD is 0.557 and the blade angle at 0.75/2
is 14.2 deg. The diameter is
n =- 88 X 135
1,950 X 0.557
= 10.93 ft.
The tip speed of this propeller at maximum speed is 1,135 ft.
per sec, the efficiency factor due to tip-speed loss is found from
Fig. 83 to be 0.905, and the actual efficiency at maximum speed
is 0.759 X 0.905 = 0.687. Although there is a 9.5 per cent
tip-speed loss at high speed, at cruising speeds below about
119 m.p.h., the revolutions will be not over 1,720, the tip speed
not over 1,000 ft. per sec, and there will be no tip-speed loss.
The propulsive efficiency at cruising speeds will therefore be
about 0.759. Also due to the fact that the revolutions are under
1,720 in take-off and in climb, there will be no tip-speed loss in
those conditions. Since there is no tip-speed loss except at
high speed, which is unimportant in a transport airplane, the
propulsive efficiency is considered satisfactory.
The 575-hp. propeller will have a twist in operation increasing
the average blade angle about 1.9, so that when the propeller
is not operating the blade angle at 0.75R is 14.2 - 1.9 = 12.3.
Then, from Fig. 178, /342 in. /o.76s = 1-9, and the setting at
the 42-in. radius is 12.3 + 1.9 = 14.2.
The weight of this 10.93-ft. propeller, as shown by Fig. 179,
would be about 131 lb. Actually, due to the fact that the tip
speed is higher than the conservative limit of 1,000 and even
slightly higher than the maximum safe limit found for a similar
propeller in Chap. XV, it would undoubtedly be advisable to
make the propeller somewhat heavier and stronger. In fact, the
loading of this propeller is severe enough to warrant a careful
overload whirling test.
Example 3.In order to show the effect of altitude on the
design system, we shall take as our next example a propeller
for an airplane intended for high-altitude cross-country flying,
and powered with a radial air-cooled engine having a gear-
driven blower in the intake manifold which gives a certain
amount of supercharging. The engine thus equipped develops
400 hp. at 2,000 r.p.m. at a standard altitude of 8,000 ft. Above
this altitude the power falls off in about the usual manner, as
shown in Fig. 114, while below it the engine may have to be

partially throttled to avoid overstressing. We shall therefore,
design the propeller to absorb 400 hp. at 2,000 r.p.m. and 8,000 ft.
altitude, under which conditions the maximum horizontal speed
is expected to be 170 m.p.h.
As was shown in Chap. X, we can write the engineering unit
form of the equation for C, as
C = 0-638 X MPH
/HP\x ?*'
( , ) XRPM
yvpo/
where HP is the actual horsepower developed at the altitude
being considered, p is the air density at that altitude, and p0
is the density of standard air at sea level. The ratio p/p0 can
be obtained from Fig. 112, and for an altitude of 8,000 ft. we
find that the value is 0.79. Then
r = 0 638 X 170
* (400/0.79)* X (2,000)**
= 0.638 X 170
3.48 X 21.0
= 1.48.
We shall assume that this airplane has a large fuselage and
that the engine is fitted with complete N.A.C.A. cowling, making
the body conditions similar to those in Figs. 174 and 175.
Assuming further that high cruising speed and rate of climb are
the performance characteristics to be favored, we shall select
the propeller halfway between that giving maximum possible
speed and that working at the peak of its efficiency curve.
From Fig. 175, the V/nD for this condition is about 0.80, the
corresponding blade angle at 0.75R is 21.7, and the diameter is
n = 88 X 170
2,000 X 0.80
= 9.35 ft.
The blade angle with 400 hp. is 1.0 greater in operation than
for the static condition, and from Fig. 178 we find that /342 in.
/3o-7B = 0.1. The angle setting at the 42-in. radius is there
fore 21.7 - 1.0 - 0.1 = 20.6.
The tip speed at maximum air speed is 1,006 ft. per sec.,
which is just about the maximum for no appreciable tip-speed
loss and also for a conservative strength allowance.
Example 4.Propellers having three blades can be satis

factorily selected with the working charts for two-bladed pro
pellers by means of an approximate rule. A two-bladed propeller
will always absorb just about 70 per cent of the power that a
propeller having three of the same blades will absorb, regardless
of the pitch or rate of advance. If we wish to select a three-
bladed propeller to absorb a certain horsepower, we may therefore
simply select a two-bladed propeller to absorb 70 per cent of this
horsepower at the same speed and revolutions and then provide
the same blades and settings for the three-bladed propeller.
The efficiency of the three-bladed propeller will be about 3 per
cent less at the peak and 4 per cent less in the climbing range
than that of the hypothetical two-bladed propeller having the
same blades.
To illustrate this method we shall find the proper three-bladed
propeller for a seaplane having a water-cooled engine delivering
300 hp. at 1,800 r.p.m. and a maximum speed of 135 m.p.h.
The airplane has a nose radiator similar to that shown in Fig.
166, but there is not sufficient clearance for a two-bladed pro
peller giving a satisfactory take-off. A general service propeller
(one operating at the peak of its efficiency curve at maximum
speed) is desired.
We shall select a hypothetical two-bladed propeller to absorb
0.70 X 300 or 210 hp. The value of C, for this propeller is
r = 0.638 X 135
* 2.91 X 20.1
= 1.47.
Then from Fig. 167, for the propeller operating at the peak of its
efficiency curve at a value of C, of 1.47, we find the V/nD is
0.740, the blade angle at 0.75.R is 19.0 deg., and the diameter is
n = 88 X 135
1,800 X 0.74
= 8.92 ft.
Our three-bladed propeller will therefore have a diameter of
8.92 ft. The blades have the same loading as those of the two-
bladed propeller which absorbs about 200 hp., and so the deflec
tion in operation may be considered negligible. The actual
blade angle at the 42-in. radius will be, from Fig. 178, 19.0
0.6 = 18.4.
The efficiency at maximum speed will be about 3 per cent

less than that of the two-bladed propeller, or 0.97 X 0.825 =
0.80.
The maximum tip speed is well below 1,000 ft. per sec, so that
there is no tip-speed loss and the strength is satisfactory.
Method of Using Charts for Propellers of any Form.As
has been shown previously,1 the blade section at 0.75R best
represents the average of the thrust- and torque-grading curves
of most propellers. By considering that the section at 0.75/2
represents the whole propeller, and then making allowances
for the differences in width ratio, thickness ratio, and airfoil
section between the 0.75R section of the propeller in question
and that of the standard propeller used for the charts, a fairly
accurate approximate method is developed for using the charts
for the selection of propellers of any reasonably conventional
form. The plan form, taper in thickness, and pitch distribution
are all of secondary importance, and the section at 0.75.R gives a
good average for them all.
Any difference in blade-width ratio from that of our standard
propeller (0.053) may be handled in the same manner as
the problem of selecting a three-bladed propeller, which has been
shown in the previous section. For differences in blade width
up to about 20 per cent or so, it is permissible to assume that the
power absorbed varies directly with the width ratio. Then if we
wished to use the charts for a propeller having a width ratio of
0.06 to operate on an engine developing 200 hp., we should select
a standard propeller for the proportionate power which it would
absorb if it had the same diameter and pitch. The standard
propeller would therefore be selected to absorb 200 X ^q^qqq^
= 177 hp. If the difference in blade width is greater than about
20 per cent, the ratio of powers, and also the approximate effect
on the maximum efficiency, can be obtained from Fig. 66.
If the thickness ratio of the airfoil section at 0.75/2 is different
from the standard value of 0.076, approximately the same effect
may be produced by giving the standard propeller a different
blade angle of such an amount that the aerodynamic pitches are
the same. The thrust and power coefficients will then be very
nearly the same over the range near maximum efficiency. This
1 See also Propeller DesignI: Practical Application of the Blade-ele
ment Theory, by Fred E. Weick, N.A.C.A.T.AT. 235, 1926.
is easily understood when it is recalled that the slope of the lift-

coefficient curve vs. angle of attack is practically the same for all
airfoil sections regardless of thickness. If plotted on a basis of
the angle from zero lift, the lift curves will all lie in essentially
the same place for the low and medium lift coefficients. The
drag coefficients of course vary widely, but they are small and
have only a very slight effect on the power absorbed by a propeller.
4 l 1 1 1 1 1 1 1 1 1 1 1 1
-t\ I I I 1 I I I I I 1 I 1
0.04 0.06 0.08 0.10 0.12 0.14 0.16
TR = Thickness Ratio of Section at 0.75 R.
Fig. 180.-Correction to blade angle for section thickness ratio at 0.75.R.
Blade angle of standard propeller to be greater than that of special propeller
by above amount. Suitable for R.A.F.-6 or Clark-Y type sections.
The angle of attack for zero lift can be obtained from airfoil
tests made in wind tunnels, or they can be calculated with equal
accuracy by means of a very simple method devised by Dr.
Max M. Munk.1 The corrections of the blade angle due to
R.A.F.-6 or Clark-Y type sections of various thickness ratios may
be found from Fig. 180.
The effect of different airfoil sections may be taken into
account in the same manner as the different thickness ratios, i.e.,
by considering only the difference between their angles of zero
lift. Fortunately the two airfoils most commonly used in pro
pellers in this country, the R.A.F.-6 and the Clark-Y types,
happen to have the same angles of zero lift for sections of the
1 The Determination of the Angles of Attack of Zero Lift and of Zero
Moment, Based on Munk's Integrals, by Max M. Munk, N.A.C.A.T.iV. 122,
1923.
same thickness ratio. Either section may therefore be replaced

by the other without affecting the power absorbed.
All of the above factors apply to the power absorbed by the
propeller, but not to the efficiency. The efficiency can usually
be obtained with a fair degree of accuracy by using the efficiency
of the standard propeller as a basis, the secondary factors being
largely a matter of good judgment.
Example 5.As an example of the design procedure showing
the use of the charts with a propeller of different form, we shall
assume a form having a width ratio of 0.062, a thickness ratio
of 0.084, and Clark-Y type airfoils (for the propellers with which
the charts were obtained, WR = 0.053, TR = 0.076, and the
sections were based on the R.A.F.-6). A general performance
propeller is needed for a cabin monoplane having a maximum
speed of 140 m.p.h. with an uncowled radial air-cooled engine
developing 400 hp. at 1,900 r.p.m.
First considering the difference in blade widths, we shall select
one of our standard propellers to absorb 400 X q = "*42 hp.
at the above revolutions and air speed. For this the value of C,
is
r = 0.638 X 140
3.21 X 20.5
= 1.36.
From Fig. 171, for the propeller operating at the peak of its
efficiency curve the V/nD is 0.688, the blade angle at 0.75/2 is
17.9 deg. and the diameter is
/) = 88 X 140
1,900 X 0.688
= 9.43 ft.
Now from Fig. 180 we find that our special propeller having a
thickness ratio of 0.084 will have a blade-angle setting 0.3 deg.
lower than that of the standard propeller. No correction is
necessary for the Clark-Y type airfoil. Our correction for
twist in operation we shall take as being approximately the same
as with the standard propeller, or 1.0 for 400 hp. The pitch
distribution of the special form of propeller must be known in
order to get the angle at 42 in. Assuming it to be the same as
standard, we find from Fig. 178 that 42 in. /30.76 = 0. The
actual blade-angle setting of our 9.43-ft. special propeller will

then be 17.9 - 0.3 - 1.0 + 0 = 16.6.
The tip speed of this propeller is well below 1,000 ft. per sec,
and so there will be no danger of tip-speed loss. While this also
indicates that the strength will very likely be satisfactory,
the safety should be assured by comparison with tests of other
propellers of the same form or by special tests on the particular
design to be used.
Example 6.Let us now suppose we have on hand a 9-ft.
propeller with a width ratio of 0.062 and a thickness ratio of
0.084 and that we wish to find the suitability of this propeller
for the airplane in the above example. The value of C. for a
standard 9-ft. propeller will be the same as in the previous
example, or 1.36, but since the diameter is now fixed the V/nD
is also fixed at
V_ = 88 X 140
nD 1,900 X 9.00
= 0.721.
Now from Fig. 171, for C. = 1.36 and V/nD = 0.721 we find
that the blade angle at 0.75.R is 20.0 deg., and that the propeller
is about halfway between that giving the highest possible effi
ciency and that operating at the peak of its efficiency curve.
Thus the high speed, cruising speed, and rate of climb will be
slightly better than with the all-around propeller, but the take-off
and angle of climb a little poorer.
The blade thickness and twist corrections to the blade angle
are the same as before, or 0.3 and 1.0, respectively. For
our 9-ft. propeller, assuming the standard pitch distribution,
Pu in. 180.75B = 0.5. The blade angle at the 42-in. radius
is therefore 20.0 - .3 - 1.0 - .5 = 18.2.
Example 7. Analysis of Flight Tests.It is often desired to
analyze the results of flight tests made with one propeller in order
to find the horsepower developed by the engine and to use this
value for the design of another propeller. This is one of the best
methods of finding the horsepower to use in designing a propeller
to fit accurately a particular airplane, for any errors which may
be in the design system will cancel out. For our example we
shall find the power developed by a radial air-cooled engine in a
full-throttle flight test in which it turned its propeller at 2,050
r.p.m., with the airplane, which was a cabin monoplane, making
a maximum level speed of 165 m.p.h. The propeller had a

diameter of 9 ft. 6 in., a blade angle of 17.5 at 42 in., a blade
width ratio of 0.055, a thickness ratio of 0.071, R.A.F.-6 type
sections, and a special pitch distribution with which f}i2 ta.
0o.76B = 0.2.
First we shall find the equivalent blade angle at 0.75/2 for a
standard propeller. The engine has a rating of 480 hp., and as
this will probably be approximately correct, we shall assume
the corresponding twist of 1.4 in operation. Then with a
thickness correction of 0.2 (from Fig. 180) and the above
value of /342 m. Po- hr = 0.2, the equivalent blade angle
of the' standard propeller af 0.75R is 17.5 + 1.4 - 0.2 +
0.2 = 18.9.
Also, since we know the airspeed, revolutions, and diameter,
the V/nD is fixed at
= 88 X 165
nD 2,050X9.50
= 0.746.
Now, from Fig. 171, for /0.76b = 18.9 and V/nD = 0.746
we find that C, = 1.46. Then from our equation for C,
NpH _ 0.638 X MPH
C, X RPM*
= 0.638 X 165
1.46 X 21.2
= 3.40,
and from Fig. 163 the horsepower for the standard propeller is
450. Finally the power absorbed in the flight test by our special
0 055
propeller of WR = 0.055 was 450 X = 467 hp. This is
close enough to the rated 480 hp. that our angle of twist of 1.4
is satisfactory. The tip speed is slightly over 1,000 ft. per sec.
but not enough to affect the power appreciably.
Design by Means of the Blade-element Theory.The full-
scale wind tunnel propeller test data given in convenient chart
form provide an accurate and simple means for selecting pro
pellers of a certain blade shape. Even with propellers of some
what different shape, such as the ordinary run of aluminum-alloy
propellers used in this country at the present time, the approxi
mate method given above is surprisingly accurate.1 If, however,

a propeller of radically different shape is desired, it is usually-
advisable either to make special aerodynamic tests or to make an
analysis with the blade-element theory, section by section.
Designing a propeller by means of the blade-element theory
is not a direct procedure like that of the working charts but is a
cut-and-try process. The complete form of the propeller must
be first assumed, and the blade-element analysis is then used to
tell whether or not the assumed propeller will fill the require
ments. If not, a new assumption is made based on the result
of the first analysis, and the process is repeated.
Various means have been used to determine the approximate
form of the propeller to be analyzed, but the success of this step
has depended almost entirely upon the experience and ability
of the propeller designer. Probably as satisfactory a method
as any is to use the standard charts with the approximate system
for blades of other forms, as given in the previous section.
After the first approximate form of the propeller has been
determined, the blade-element analysis can be satisfactorily
made either with the multiplane interference correction method
or by means of the induction or vortex theory, both of which are
given in Chap. V. The use of the simpler forms of the blade-
element theory is not advisable, for their accuracy is no greater
than that which can be obtained with the simpler chart method
for propellers of any reasonable form.
1 The satisfactory accuracy of the approximate method of applying the
chart data to propellers of different forms has been proved by more than
130 flight tests made by the author.
INDEX
Airship propellers, 251

Alferium, 219
Adjustable-pitch propellers, 221 All-around performance propellers,
Aerodynamic tests on propellers, 82 251
Aeron, 219 Altitude, effect on airplane per
Air, variation of density with alti formance, 175
tude, 175 on design of propellers, 275
Air-brake, variable-pitch propeller, on engine power, 176
188 performance with supercharger,
Air-loading curves, 237 178
Airflow through propeller, 77 relation with density, 175
Airfoil, characteristics, compressi Aluminum-alloy propellers, erosion
bility effects, 28 of, 224
effects of scale, 28 etching of, 245
methods of plotting, 16 repair of, 224
from propeller tests, 58 weight of, 224
chart for, 61 Aluminum alloys, properties of, 219
plan form, effect of, 22 Analysis of flight test data, 281
propeller, area of, 235 Angle of climb, best propellers for,
center of gravity of, 235 250
moments of inertia of, 235 Apparent efficiency, 140
section, effect on efficiency, 114, Approximate stress analysis, 234
' 255 example of, 234
effect with different pitch, 115 Area of R.A.F-6 propeller sections,
sections, Clark-Y vs. R.A.F-6, 115 235
suitable for propellers, 29 Army tip, 214
shape, effect of, 19 Aspect ratio, 3
effect of lower camber, 21 effect, 65
position of maximum ordinate, Attachment of wood blades, 225
. 21 Automatic variable-pitch propeller,
upper camber, 20 228
tests at high speeds, 119 Axes, neutral, 231
theory, 23 principal, 231
Airfoils, effect of high speed, 119
forces on, 13 B
pressure distribution over, 18
for propellers, 11 Balance, 244
Airplane performance, 249 weights, variable-pitch propeller,
effect of propeller characteristics, 226
159 Balances, test, 83
Airscrew, 9 Balancing, against master blade, 224
285
Bearings, variable-pitch propeller, Body, drag, effect on propulsive

225 efficiency, 144
Bending, moments on propeller effect on airflow, 137
blades, 230 interference, 136
theory of, 229 data, practical use of, 157
Birch, 215 effect of body on propeller, 136
Blade, angle, relation with pitch effect on efficiency, 255
ratio, 94 effect of propeller on body, 138
deflection, 239 size, effect on efficiency, 142
detachable, 221 Boss, 3
elements, independence of, 79 Brake, horsepower, 160
interference, simple blade-element effect of variable pitch, 184
theory, 42 state, 105
section, effect of, 114 variable-pitch propeller, 188
on efficiency, 255 Brass tipping, 214
shape, effect of, 105 Bushings, engine shaft, 219
thickness, effect on efficiency, 254
twist, measurement of, 85 C
width, effect of, 111 Caustic soda, 246
on efficiency, 254 Center of pressure, 15
measure of, 110 chart of, 18
Blade-element theory, comparison Centers of gravity of R.A.F-6
of forms of, 77 sections, 235
independence of elements, 38 Centrifugal force, loads due to, 230
induction or vortex theory, 64 Charts, design type, 93, 99, 257
check with tests, 77 use of, 95
direct procedure, 69 working, 257
example, 74 examples, 266
Glauert's procedure, 69 use of, 264
limitations of, 72 Chrome-nickel steel, 221
tabular form, 76 Chrome-vanadium steel, 221
modifications of, 51 Clamp rings, hub, 223
with multiplane interference, 55 Clark-Y airfoil, 33
comparison with test data, 64 propeller efficiency with, 115
example, 59 Climb, best propeller for, 166, 249
limitations of, 57 calculation of rate of, 166
tabular form, 63 effect of pitch, 166
resume^ 77 of propeller, 165
simple, 37 of reduction gearing, 195, 199
accuracy of, 43 of variable pitch, 182
comparison with test data, 48 maximum, speed for, 166
example, 44 metal vs. wood propellers, 168
limitations of, 41 tandem vs. side-by-side propellers,
scale effect, 43 209
tabular form, 47 Coefficient, power, 90
tip loss, 43 speed-power, 91
use in design, 282 fifth root of, 92
Blades, choice of number, 252 use of in curves, 93
number of, 112 in design, 95
INDEX 287
Coefficient, thrust, 87 Detachable-blade propellers, 221
torque, 88 Diameter, front propeller of tandem
Coefficients, airfoil, conversion fac series, 208
tors for, 15 limitations, 252
basis of, 87 propeller, 2
lift and drag, 13 rear propeller of tandem series, 205
propeller, 86 Disc area, effective, 53
Combined theory, 51 Distribution of pitch, 108
application of, 52 best, 110
limitations of, 54 effect of body, 155
Compressibility effect, 12, 119 Downwash, 65
Consumption, effect of variable angle, 23
pitch, 186 multiplane interference, 55
fuel, 173 effect on inflow, 51
at cruising speed, 173 Drag, induced, 24
Controllable-pitch propellers, 180 equations for, 25
construction of, 224 profile, 24
Convex lower camber, 31 equations for, 25
effect on lift, 33 Drzewiecki theory, 37
Corrosion, protection against, 224 Duralumin, 219
Counterweights, variable-pitch pro Dynamical similarity, principle of,
peller, 226 12
Cracks, shown by etching, 246
surface, effect on fatigue failures, E
245
Critical tip speed, 126 Effective thrust, 141
chart for, 134 Efficiency, apparent, 140
Cross shear, 230 of blade element, maximum, 41
Cruising speed, best propellers for, simple blade-element theory, 40
250 comparison of actual with ideal,
effect of propeller on, 172 107
fuel consumption at, 173 correction for high tip speed, 130
effect of airfoil section, 255
D axial interference, 68
blade thickness, 112
Damaged propellers, repair of, 224 blade width, 254
Data, working, from full-scale tests, body interference, 140, 255
260 high tip speeds, 123, 126, 128
Deflection, blade, measured, 239 number of blades, 112, 254
effect on stresses, 241 pitch, 106
in twist, 260 distribution, 108, 254
approximate rule for, 264 ratio, 253
Design, with blade-element theory, plan form, 116, 255
282 profile drag, 68
charts, 93, 94, 99, 257 propeller-wing interference, 151
examples, 266 rotational interference velocity,
use of, 95, 264 68
simple method, 258 scale, 118
Destructive whirling tests, 247 size, 117
Efficiency, effect of thickness ratio, Flight tests, analysis of data, 281

254 on propellers, 82
tip speed, 255 Flow through propeller, 77, 101
variable pitch, 183 measurements of, 102
factors affecting, 253 periodic nature of, 79
ideal, 7 Fluctuating air loads, 244
computation of, 7 Fluid, perfect, 11
effect of diameter, thrust, and resistance, 11
forward velocity, 9 Flutter, 243
usefulness of, 8 Forging of blades, 223
induction or vortex theory, 68, 73 Form of a standard metal blade, 272
inflow theory, 54 Four-bladed propellers, when used,
metal vs. wood propellers, 113 253
methods of plotting, 93 Frequency, natural, effect of size,
multiplane interference theory, 57 244
net, 140 Froude momentum theory, 5
propeller, 90 I el consumption, 173
propulsive, 141 effect of variable pitch, 186
effect of body shape, 144 Full-scale propeller tests, 85
of body size, 142 Fuselage, effect on air flow, 137
of propeller location, 149
tandem propellers, 203, 207 G
various expressions for, 140
with body interference, effect of Geared propellers, effect on static
pitch distribution, 155 thrust, 196
Eiffel logarithmic diagram, 97 on weight, 193
, planation of, 98 example, transport airplane, 191
use of, 100 VE-7 airplane, 198
Element, blade, air forces on, 39 factors affected, 190
Elements, blade, independence of, 38 gyroscopic vibrations, 253
Endurance, flight, 174 when desirable, 200
Engine power, effect of altitude, 176 Gearing, with tandem propellers, 210
English coefficients, 89 Glauert theory, 64
Erosion, by seaplane spray, 224 Glide tests, 82
Etching of propellers, 245 Gloster Hele-Shaw Beacham vari
able-pitch propeller, 228
F Governor, for variable-pitch pro
Fabric sheathing, 216 peller, 228
Failures, fatigue, 245 Grading curves, integration of, 46,
photo of, 246 48
of wood propellers, 246 Grinding of blades, 224
Fan, wind tunnel, 85 Gyroscopic forces, 229
cast aluminum alloy, 86 moments, effect of three blades,
Fatigue failures, 245 253
photo of, 246
strength, 213 H
Filler blocks, for twisted-slab pro
pellers, 220 Hart variable-pitch propeller, 226
Finishing of blades, 224 Heath variable-pitch propeller, 227
INDEX 289
Hele-Shaw Beacham variable-pitch Inflow, rotational, 52
propeller, 228 simple blade-element theory, 42
High speed, effect of airfoil section, theory, 51
121, 130 expressions for thrust, torque,
effect on airfoil characteristics, 28, and efficiency, 53
120 limitations of, 54
on angle of zero lift, 121 Information furnished designer, 257
on drag coefficient, 121 Instantaneous velocity, 80
on lift coefficient, 120 Interference, angle, 69
propeller for, 97, 249 chart for solution of, 71
tandem vs. side-by-side, 209 relation to downwash, 70
section thickness, 121, 129 blade, downwash correction chart,
High tip speeds, 119 60
full-scale propeller tests, 126 lift correction chart, 59
model propeller tests, 121 body, 136
practical corrections for, 130 effect of body on propeller, 136
Horsepower, maximum, 160 of propeller on body, 138
Hub, 3 data, practical use of, 157
detachable-blade propeller, 221 effect on efficiency, 255
screw type, 222 factor, effect on efficiency, 68
split type, 223 rotational, 66
drag, 78 flow, 52
Hubs, for micarta propellers, 216 axial, 69
split, 219, 222 calculation of, 65
for wood propellers, 216 chart for solution of, 72
Hydraulic variable-pitch propeller, definition of, 65
228 rotational, 66
Hysteresis, 213 propeller and wing, 151
extreme example of, 155
I velocity, 80
Ideal efficiency, 7 L
computation of, 7
effect of diameter, thrust, and Laminations, metal propellers, 218
forward speed on, 9 wood propellers, 213
usefulness of, 8 Leitner-Watts propellers, 218
Inclined propeller axis, effect of, 162 Lift coefficient, engineering, 14
Independence of blade elements, 79 English, 14
Induced drag, 24 N.A.C.A., 14
calculation of for rectangular coefficients, conversion factors for,
plan form, 25 15
example, 26 Lighter-than-air craft propellers, 251
Induction theory, 64 Limitations of blade-element theo
Inertia, forces, 229 ries, 41, 54, 57, 72
moments of, 231 to diameter, 252
Inflow, 65 Loading curve, air, 237
axial, 52 centrifugal force, 236
calculation of, 52 Loads on propeller blades, 229
empirical factors for, 51 calculation of, 232
Logarithmic diagram, Eiffel, 97 Navy tip, 214

explanation of, 98 Net efficiency, 140
use of, 100 Neutral axis, 231
Longitudinal shear, 230 Nitric acid, 246
Lower camber, effect on airfoil Noise, effect of tip speed, 134
characteristics, 21 Nomenclature, 2
Number of blades, 112
M choice, of, 252
Machining of blades, 224
Mahogany, 215 0
Major principal axis, 231 Oak, 215
Master blades, 224 One-piece forged propellers, 221
Maximum airplane speed, effect Outflow, 65
of propeller, 163
of variable pitch, 181 P
horsepower, 160 Performance, airplane, 159, 249
revolutions, 160 all-round, propellers for, 251
Metal propellers, 217 effect of altitude, 175
efficiency of, vs. wood, 113 with variable pitch, 180
Micarta propellers, construction of, Periods of vibration, 244
216 Pitch, definition of, 2
strength of, 217 distribution, best, 110
thickness of, 217 effect, 108
Minor principal axis, 231 of body interference, 155
Miscellaneous propeller forms, use on efficiency, 254
of design charts for, 278 effect on efficiency, 105
Model propeller tests, pressure dis on ideal efficiency, 107
tribution, 49 on power, 108
propellers, 84 on thrust, 108
tests, correct scale for, 12 effective, 3
Moment coefficient, airfoil, 15 experimental, 3
of inertia of R.A.F-6 propeller geometrical, 3
sections, 235 ratio, effect on efficiency, 253
Moments, bending, 230 relation with blade angle, 94
of inertia, 231 relation to speed-power coefficient,
twisting, 230 107
Momentum theory, 5 variable, 180
Multiplane effect, 42 Plan form, effect of, 116
interference, 55 on efficiency, 255
chartfordownwash correction, 60 forms commonly used, 117
chart for lift correction, 59 Polar diagram, 17
effect of lift, 56 Power, accurate values for, 257
of profile, 55 available, computation of, 161
of solidity factor, 56 effect of altitude, 176
of spacing, 55 of pitch, 163
of reduction gearing, 192
N coefficient, 90
Narrow-tip plan form, 117 effect of tip speed, 127, 129
Natural frequency, effect of size, 244 effect of pitch, 108
INDEX 291
Power, excess, 166 Right-hand propeller, 4
required, 159 Rings, hub clamp, 223
effect of altitude, 175 Root, blade, 3
Prandtl theory, 23 standard for aluminum-alloy pro
Pressure distribution, over airfoils, 18 pellers, 223
on model propeller, 49 Rotation of slipstream, 202, 207
Principal axes, 231 Rotational velocity, computation
Profile lift and drag, 24 of, 66
Propeller gearing, 190 Run before take-off, effect of pro
sections, based on R.A.F-6, 29, 30 peller, 170
test data, 260
Propulsive efficiency, 141 S
effect of body size, 142
of body shape, 144 Safety, structural, 255
of propeller location, 149 Scale effect, 117
factors affecting, 253 terms of, 12
Pusher arrangement, 4 Seaplane service, 224
effect on efficiency, 151 Selection charts, types, 93, 94, 99
use of, 95
R of propellers, 257
R-type propeller, 220 Shape of blades, effect of, 105
R.A.F-6 airfoil, efficiency with, 115 Shear curve, 237
R.A.F-6-type propeller sections, air Shearing forces, 230
foil characteristics of, 31-35 Sheathing, wood propeller, 214
ordinates of, 30 Side-by-side propellers, comparison
Racing airplanes, propellers for, 249 with tandem, 208
Rain, erosion due to, 224 Simple blade-element theory, 37
Rake, 4 computation of efficiency, 40
effect on stresses, 242 of thrust, 39
Range, effect of propeller on, 174 of torque, 40
Rankine theory, 5 limitations of, 41
Rate of climb, best propellers for, Simple design method, 258
249 Simpson's rule, 46
Rear propellers of tandem series, 201 Size, effect on coefficients, 117
diameter of, 205 Slab, twisted propeller, 220
effect of pitch, 204 Slip, function, 87
efficiency of, 203 nominal, 87
Reed propellers, 220 Slipstream, dissipation of, 104
Repair of metal propellers, 224 effect on body drag, 139
Resistance, fluid, 11 measurement of, 102
Reversible-pitch propellers, use as size of, 102, 104
brakes, 188 velocities, calculation of, 201, 206
Revolutions, computation of for Solid forged propellers, 220
various air speeds, 161
maximum, 160 Solidity factor, multiplane inter
reduction of by gearing, 190 ference, 56
Reynolds number, 12 Sound, velocity of in air, 119
effect of, 28, 122 Special performances, propellers for,
at high tip speeds, 126 251
' Speed, accurate values of, 257 Sweepback of wood propeller blades,
as affected by gearing, 191, 195, 198 247
cruising, 172 Symbols, list of, xi
high, effect of on airfoil character
istics, 119 T
practical corrections for, 130 Take-off, best propellers for, 250
maximum airplane, effect of pro effect of propeller on, 170
peller, 163 of pitch and diameter on, 171
of pitch, 164 of reduction gearing on, 196,
propeller for, 249 '199
variation with propulsive effi of variable pitch on, 184
ciency, 165 of tandem vs. side-by-side ar
tandem vs. side-by-side arrange rangements on, 209
ments, 209 Tandem propellers, 201
Speed-power coefficient, 91 air forces on rear propeller, 202
with altitude term, 177 comparison with side-by-side, 208
use of, 91, 95 diameter of front propeller, 208
Split hubs, 219, 222 effect of gearing, 210
Splitting, of wood propellers, 247 efficiency of rear propeller, 203,
Spray, erosion due to, 224 207
Standard form, use of, 258 example calculations, 207
metal blade, 272 ' flow at rear propeller, 202, 206
Static balance, 244 model tests, 203
thrust, 170 simple method of design, 206
effect of variable pitch, 185 Taper of blade, effect of, 116
Steel propellers, 217 on efficiency, 255
failures of, 218 of plan form (airfoil), 23
Strength, 229, 256 Tensile force, 230
fatigue, 213 Terms, propeller, 2
of wood, 213 Terne-plate tipping, 215
Stress, analysis, accuracy of, 232 Test results, coefficients used, 86
approximate, 234 methods of plotting, 86
example, 234 Tests, model propeller, 83
on propellers, 82
effect of flutter, 243 flight tests, 82
of weaving, 244 whirling, 247
when safe to apply, 245 Theory, airfoil, 23
Stresses, calculation of, 231 Glauert, 64
effect of deflection, 241 induction, 64
of rake or tilt, 242 momentum, 5
variation with tip speed, 233 Prandtl, 23
Strobometer, 79 resume' of blade-element theories,
Supercharged engines, 178 77
propeller requirements, 179 vortex, 64
Thickness, blade, effect of, 112
use with variable-pitch propellers, ratio, 112
187 angle correction for, 278
Surface irregularities, effect on fa blade, 4
tigue failures, 245 effect on efficiency, 254
INDEX 293
Three-bladed propellers, use of de Tractor, comparison with pusher,

sign charts with, 277 151
when used, 253 Turnbull variable-pitchpropeller, 227
Thrust, coefficient, 87, 89 Twist, propeller blade, 239, 260
effect of high tip speed, 122, 124 approximate rule for, 264
of pitch, 108 measurement of, 85
effective, 141 Twisted-slab propellers, 220
grading curves, 46 Twisting moment due to centrifugal
comparison of calculated and force, 226
measured, 78 moments on propeller blades, 230
horsepower, computation of, 161 Two-dimensional airfoil character
induction or vortex theory, 67 istics, use of, 65
inflow theory, 53
multiplane interference theory, 56 U
simple blade-element theory, 40 Upper camber, effect on airfoil
static, 170 characteristics, 20
Tilt, 4
effect on stresses, 242 V
Tip, loss, 78 Variable diameter, 180
simple blade-element theory, 43 Variable-pitch propellers, 180
speed, computation of, 132 compromise diameter for, 187
corrections for efficiency, ex construction of, 224
ample, 132 Hart, 226
data, effect of pitch, 130 Heath, 227
effect of, 119 Hele-Shaw Beacham, 228
on efficiency, 255 Turnbull, 227
model vs. full size, 126 effect on climb, 182
on noise, 134 on cruising economy, 186
limiting value for allowable on efficiency, 183
stress, 243 on engine power available, 184
loss, practical data, 130 on speed, 181
relation to stresses, 233 on take-off, 184
Tip speeds, high, 119 use, as brake, 188
full-scale propeller tests, 126 with supercharger, 187
model propeller tests, 121 Velocity of sound, 119
Tipping of wood propellers, 214 Vibration, 244
Torque coefficient, 88, 89 Vortex blade-element theory, 64
effect of high tip speed, 123, 125 limitations of, 72
gasoline engine, 160
induction or vortex theory, 68 W
inflow theory, 53 Wake, 65
multiplane interference theory, Wall effect, wind tunnel, 27
57 Walnut, 215
simple blade-element theory, 41 Warping, of wood, 213
Torsional deflection, 239, 260 Weave, 243
approximate rule for, 264 Weight, standard metal propellers
vibration, flutter, 243 274
Total width ratio, TWR, 4, 112 Whirling tests, 247
Tractor, 4 Wide-tip plan form, 117
Width, blade, effect on efficiency, Wing-propeller interference, 151

110, 254 Wood blades, attachment to metal
mean effective, 111 hub, 225
ratio, WR, 4, 111 propellers, 213
allowance for odd widths, 278 failures of, 246
effect on efficiency, 254 laminations of, 214
on power, 111 manufacture of, 214
total, 112 strength of, 213
Wind tunnel, British N.P.L., 84 tipping of, 214
Eiffel type, 83 for propellers, 213
N.A.C.A. 20 ft. throat, 85 strength of, 213
Stanford University, 83 warping of, 213
tests, 82 Working charts, 257
wall effect, 27 use of, 264
Windmill state, 105 examples, 266
Wing theory, 23 Wright Field, whirling-test rig, 248
5 AA A 30
'.J

Aircraft Propeller Design

Uploaded by

Copyright:

Available Formats

Aircraft Propeller Design

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Aircraft Propeller Design

Uploaded by

Copyright:

Available Formats

AIRCRAFT PROPELLER DESIGN

FRED E.yWEICK, B.S.

McGRAW-HILL BOOK COMPANY, Inc.

All rights reserved. This book, or

THE MAPLE PRESS COMPANY, YORK, PA.

gan, and Prof. E. G. Reid of Stanford University. Acknowledg

CT Thrust coefficient = T/Pn*D*.

A propeller is a device for providing a force or thrust, at

In addition to its aerodynamic qualities, a propeller must

may have a variety of forward speeds at the same revolution

Fig. 2.Illustration of effective and geometrical pitch.

pitch of the section at two-thirds of the radius. If all of the

The width ratio WR of a propeller blade is the blade width at

THE MOMENTUM THEORY

As stated in the introduction, the propeller, in order to provide

1 A. Betz has developed an extension of the momentum theory including

disc and receives an increment Ap in passing through it. The

The flow being regarded as potential except in passing through

and if A is the area of the disc the thrust is

Also, since the thrust is equal to the change of axial momentum

ti = zr and a = ]^b, the equation can be written

where q is the dynamic pressure of the fluid and is equal to

Fig. 4.-Ideal efficiency vs. thrust coefficient

3. The ideal efficiency increases with increase of density of

In the blade-element theory of the propeller, the propeller

Dynamical Similarity.If K of the above formula were

where L is a convenient representative linear dimension of

The exact variation of the aerodynamic characteristics with

0 and 90 deg., there will be both a resistance and a lateral force

cients of lift and drag. The coefficients used by the N.A.C.A.

The symbols Lc and Dc are often used in place of kL and kD in

corrected to standard air for which p = 0.002378 lb.-ft.-4-sec.2

It is unfortunate that so many different coefficients exist for

where c is the chord of the airfoil.

veniently presented in graphical form, and many systems have

In comparing the aerodynamic qualities of various airfoils,

diagram, the lift coefficients as ordinates are plotted directly

The CP. is not of the same interest in connection with pro

Pressure Distribution over Airfoils.The coefficients dealt

on the upper, both contributing to the lift. It is interesting

Fig. 10.-Pressure distribution at low and high angles of attack.

Effect of Variations in the Shape of Airfoils.From the

camber is increased, reaching comparatively large negative

airfoil.1 The general conclusion from these tests is that for

rectangle to a very slender ellipse, have been tested and the

by making them narrower than the medium one, No. 3. 1 Another

with various amounts of taper in plan, as shown in Fig. 14. 1

force referred to the mean direction of flow are called profile

For this type of span loading the induced-drag coefficient is given

where R is the aspect ratio, or the span squared divided by

where a< is in radians.

where a is the slope of the straight portion of the lift curve, or

and the angle of attack for infinite aspect ratio becomes

CD. = ^ = 0.0559 CL2 = 0.0164,

In like manner the induced angle is

Effects of Scale and Compressibility.The principle of

0 1,000,000 2,000,000 3,000,000

0.12 is shown in Fig. 18. It may be said that in general there

The series of propeller sections which has been most widely

Fraction L.E. T.E.

CT Thrust coefficient = T/PnD.