FORAERONAUTICS

TECHNICAL

EXPERIMENTAL

L-

NOTE 3491

RATIO,

INVESTIGATION OF ECCENTRICITY

FRICTION,AND OIL FLOW OF LONG AND

SHORT JOURNAL BEARINGS WITH

-,-- -NUMBER CHARXS

_uMD
By G. B. Du.Bois,
F.W. Ocvirk,
andR. L,Wehe
Cornell

Washington
September1955
., ,

-,.

,.

-.;<d-J~d

r..
.-.,,,

,.+ 0
. L..

..-

-----

. .

..

TECHLIBRARY
fCAFB.
NM

NATIONAL
ADVISORYcoMlmrrEE
FORAERONAUTICS

lIIMll!llll

,.

DOLL55J

TECHNICAL
NOTE3491
EXPRIMENTMINVESTIGATION
OF ECCENTRICITY
RATIO,
FRICTION,
ANDOILFLOWOF IONGAND
SHORTJOURNAL
BEARINGS
WITH
ILMD-NUMBER
CHART;
By G. B. -iS, F. W. Ocvirk,
and R. L. Wehe

loading
are
Theperformances
ofplainbebrings
understeadycentral
therangeof
by single-line
curvescovering
compared
andsummarized
length-diameter
ratiosbothaboveandbelow1. Experimental
dataon
eccentrici~
ratio,friction,
andoilflowforlength-diameter
ratios
of 1, 1.,and2 areshownforcomparison
withearnerdataforlengthdataprovidecharts
diameter
ratiosof 1/4,1/2,and1. Thecaibined
ofplain-bearing
performance
whichcoverthepractical
rangeof lengthdismeter
ratio.
l/Cn in theform p/pN
(c~d)2(d/Z)2, where Cn
Theloadnumiber
is thecapacity
number,p is thecentral
unitbearingload, p is
speed,cd is thedismetral
theoilviscosi~,N is the journal
bearingclearance,
d is thebesring
diameter,
and Z is thebearing
length,
is usedas the.
basicvariable
onwhicheccentrici~,
friction,
to showthatthe.
andoilflowratedepend.Chartsarepresented
eccentrici~-ratio
datafallnearlyon a single13_ne
forbothlongand
shortbearings
if the (d/Z)
2 termin theloadnumberis takenas uni~
forbearings
of Z/d greater
than1; in effect,
theloadnumberfor
longbearings
becomesthereciprocal
of theScmmerfeld
number 1/s.
Friction
andoilflowrate,however,
areshownto dependon theload
numberwithoutthismalif
ication,
forbothlongandshortbearings.
Plotsof friction
datain termsof friction
ratio F/F. arepresentedto fecusattention
on thecliff
erencebetweenjournal
friction
riction
datafromother
andbearing
friction.Experimental
journal-f
sources
sreincluded.

.. -- . ..

..,.

..

. .

NACATN 3491

Analytical
curves
by SOmtnerfeld,
Cameron,
andWoodandtheshortbearingapproximation
areshownforcomparison
withtheexperimental
dataforeccentrici~
ratioandfriction.Curvesresulting
fromthe
integration
of Sommerf
eldspressure-distribution
function
foran extent
of 18ooof thepressure
filmarealsoshown.
INTRODUCTION
Tn an earMerreport(ref.1), theanalytical
curvesof theshortbearing
approximation
werefoundtobe in usefulagreement
withexperimentaldataon eccentricity,
friction,
andoilflowforcentrally
loaded
plainbearings
withvaluesof Z/d of 1/4,1/2,and1. A fewearly
experiments
withbearings
of Z/d of 2 indicated
@at it wasnecessary,
withregsrdto eccentrici~,
to limittheusefulness
of thecurvesto
bearings
of Z/d notexceeding
1. Forthisreason,
a conclusion
was
drawnthatfurther
testing
onbearings
of 2/d between1 and2 wasnecessaryto covertherangeofbearingconfigurations
usedin practice.
.
Thenewexperimental
datain thepresent
reportwereobtained
from
testsconducted
in theMachine
I!esign
Laboratories
at Cornell
University,
Ithaca,
NewYork,as a partof a bearing-research
investigation
conducted ~
underthesponsorship
andwiththefinancial
aidof theNational
Advisory ,
Committee
forAeronautics.
Experimental
dataonbesrings
of 1~-inch
diameter
with Z/d ratiosof 3/4,1, 1$ and2 arereportid
hereinand
correlated
withetisting
dataforvaluesof Z/d of 1/4,1/2,and1.
Ih thelong-bearing
tests,loadsup to 1,700poundsor 8E poundsper
squareinchwereusedat speedsup to 7,0~ rpmwithME 10 oil andwith
clearances
ramging
from0.0018to 0.0038inch. Thedataarecoqibined
in
chartscovering
therangeof valuesof Z/d in commonuse. Theintegrations
forthe%mnerfeldz casearegivenin theappentiwhichwas
prepared
by Professor
Wehe.
Analytical
I?ackground
Analytical
solutions
by Csmeron
andWo&i(ref.2) gitingpoints
are
locating
curvesfor Z/d ratiosof 1/4,1/2,1, andinfinity
available.
Sommerfeld
ref.3) provides
an analytical
solution
for
bearings
of infiniteZ)d. Theshort-bearing
approximation
applies
best
to shortbearings
andis usefulin therangeof valuesof Z/d Up to 1.
Ihus,no analytical
solutions
for Z/d in theneighborhood
of 2 are
available.
Thisexperimental
investigation
was initiated
to provide
experimental
data,particularly
on eccentricity,
usingbearings
of Z/d= 1,
1+,and2. A shortbearingof Z/d= 3/4 wasalsoincluded
in orderthat

NACATN 3491

a parallel
investigation
on themisaligning
characteristics
ofbearings
Wing thesametestbearings
andjournals.
(ref.4) couldbe conductid
usedin theexperimentation
weresimilar
to
Theequipment
andtechnique
1 fortheshortbearings.
thosereported
in reference
3%hissolution
forbearings
of infinite
length,
Sommerfeld
showed
thatthebasicnondimensional
grouping
ofbearing
variables
onwhich
eccentricity
andfriction
dependis theSommerfeld
nuniber
S_dVd2
pc(d)

(1)

appro-tion yieldssingleForfhit-eshortbe=ings,theshort-bearing
linecurvesof eccentricity
ratioandfriction
usingthecapacity
number ~ as thebasicvariable:

%=

%$(U

(2)

Itmaybe seenthatthevariables
of equations
(1)and (2)arecliff
erent
by thefactor (Z/d)2.
bad NMber
Thereciprocal
of thecapaci~nuuiberj
or loadnuuiber
l/~, is
whichis thereciprocal
form
usedin figures
1 to 4. Theloadnumber,
of thebasicvariable,
is chosenin orderto expandtheinteresting
highloadregionandto showthevariation
of eccentricity,
friction,
and
oilflowratein directproportion
to thecentral
load.
Theexperimental.
dataon eccentricity,
friction,
andoilflowrate
in figures
1 to 4 showtheextentto whichtheSOmmerfeld
nuniber
and
loadnuniher
reduceda~ to nearlysingleMne chartsforbearings
of Z/d
of the (Z/d)
2 termin thisregardis
from1/4to 2. Theinfluence
datafromreference
1 forbearings
of
shown
. Chartsof experimental
Z/d= 1/4,1/2,and1 axe.
reproduced
in figures
l(b),2(b),andA(b)
forcomparison
withchartsof datareported
hereinin figures
l(a)to
A(a)onbearings
of Z/d= 1, % - 2
Friction
Ratio
Jnfigure2, theratioof thefriction
torquemeasured
on the
bearingto theanalytical
Petroff
friction
at no load,thefriction

...-.

._____

_-

._.

..

NACAm 3491

(ref.1). Thisisbelieved
to
ratio F/Fo,is usedas a basicvariable
%e moreconvenient
to use thanthecoefficient
of frictionf %asedon
loadd focuses
attention
to thequantities
onwhichfriction
&ectly
depends.A similar
ratio,thepower-loss
ratio j, isusedby Wilcock
andRosenblatt
(ref.5). Thefriction
ratiois shownanalytically
to
dependon theloadnumberby theshort-bearing
approximation.
It is generalJy
recognized
thatthefriction
torqueon thejournal
is greater
thanthefriction
torqueon thebearingbytheamountof the
couplecaused
by thelatempl
displacement
of thejournal
in thebeming.
The13ne-of
actionof theloadanditssupporting
reaction
areseparated
by thisdisplacement
to forma couple.Ihe
difference
in friction
is
considerable
undercertain
conditions
axlcanbe shownon a singlecurve
infi~e 3 by plotting
thecouplefriction
ratio AF/Fo(d/Z)2
against
theIcednuaiber.
OilFlowFactor
0

Theoilflowratefromthetestbearingis represented
in figure4
bya single-line
curveobtainedby
theuseof theoilflowfactorq,
whichis theratioof theexperimental
flowrateto theanalytical
end
leakage
rateof theconverging
wedgeportionof theoilfilm. Thisis
plotted
against
a modification
of theloadnumbercalledtheinlet-oilis subpressure
nuniber
1I$.> inwhichtheinletoilsupplypressure
stituted
fortheunitbearing
pressure
in theloadnumber.Thismethcd
ofplotting
oilflowrateis thesameas thatusedin theearlder
report
covering
valuesof Z/d lessthan1 (ref.1) exceptthatthereciprocal
formof thenumiber
is usedin thisreport.

SYMBOLS
I

Dimensional
quantities:

cd

diametral
bearingclearance,
in.

radialbearingclearance,
cd 2, in.

bear@g diameter,
in.

eccentiici@
forcentral
loading,
in.

lateral
displacement
of journal,
in.

,,

NACATN 3491

vertical
displacement
Qf journal,
in.

Fb

frictio~
forceon stationary
bearingsurface,
lb

Fj

friction
forceon rotating
journsl
surface,
lb

F.

Petroff
friction
forceat no loadandzeroeccentricity,
ti~N2d(d/cd),
lb

bearinglength,
in.

journal
speed,rpm

journal
speed,rps

film
P

localfluidfilmpressure,
lb/sqin.
applied
central
unitbearingloadonprojected
area,lb/sqin.

PC

capsule
pressure
of load,lb/sqin.

PO

inletoilpressure,
lb/sqin.

P~

pressure
at e = 0 and I-Cin Somerfeldsolution,
lb/sqin.

applied
central
bearingload,lb

p@y

components
of central
loadparallel
andnormalto lineof
bearing
andjournal
centers,
lb

experimental
totalrateof oilflowfrombearing,
cu in./sec

bearing
radius,
in.

%4>tg

temperature
at stations
4 and9, respectively,
OF

surface
speedof journal,
in./sec

centipoise
oilviscosity,
or reyns
6.9x 106

attitude
angle,anglebetweenloadlineandlineof centers
ofbearing
andjournal,
deg

-iab~

gitinglocalfilmpressure,
deg

...

NACATN 3491

6
Nondimensional
quantities:
&d

clearance
ratio

Cn

%0

pNt d~
inlet-pressure
capaciwnumber,p
o ()()
cd d

l/&

had muiberforvaluesof 2/d lessthan1

l/~.

inlet-oil-pressure
number

coefficient
of friction,F/P

F/F.

friction
ratio

b/Fo

bearingfriction
ratio

Fj/Fo

journal
friction
ratio

M/F.

cou@e friction
ratio

Z/d

length-diameter
ratio

eccentricity
ratioor attitude
forcentral
loading,e/Cr

horizontal
component
of eccentrici~
ratio,
%/+
.

nv

vertical
component
of eccentricity
ratio, ev/Cr

oilflowfactory
*

@d
Sommerfe3d
nuuiber,
,P()
cd

1/s

than1
loadnumberforvaluesof 2/d greater

XJY

coordinates

viscosity,
centipoise

.. .

-----

NACATN 3491

APPARATUS
Thebearing-testing
machine
usedin theseexperiments
is illustrated
in figures
~ to 8 andis thesameas ~t describ~d
in reference
1. The
mannerinwhichthetestelements
were
supported
andlosdedis shownin
figures
5 and6. lHgure7 showsthemechanical
systemformeasuring
journal
displacements,
andfigure8 givesthelocations
of thethermocouples
Wed to determine
bearingtemperatures.
TestBearing
andJournals
A single
bronzebearing
andfive.
steelshaftsof 1~-inchnominal
diameter
wereusedin theconfiguration
shownin figure~. Eachof
theshaftsrepresented
a givenlength-diameter
ratioanda givenclearanteas listed
below:

Average
diametral Clearance
shaft Z/d clearance, ratio,
c~d,
%
in.
/in.
in.
6A

6B
6C
6D
6E

2
0.00252 0.ool&j
2
.00376
.oo2~
11
~
.00196
.oolk2
1
.00183
.CX)133
;
.00258
.00187

Thesetestelements
werealsousedin themisaMnement
exmriments
as reportid
in reference
4. SAE10 oilwasusedas thetestl-tiricant
andwasfedto thetestbearing
at a pressure
of40 to 16opoundsper
squareinchthrough
a l/8-inch-diameter
oilholelocated
opposite
the
central
load. Theoilwaspreheated
to 140F nearthepump,butthe
temperature
of theoilwasmeasured
as it entered
thebearing
because
of smallheatlossesin theoillines.
Thetestshafts
weredrivenby a high-s~ed,
direct-current,
variable-speed
aircraft
motorhavinga speedrangeof 1,000to 10,(NOrpm.

---

NACATN 3491
Loading
Appsxatus
As shownin figure6, thecentral
loadwasapplied
hydraulically
by a pressure
capsule
andwastransmitted
to thehearingthrough
an
oil-pressurized
spherical
seat. Theoilflowthrough
thespherical
seatfloatsthebearinggivingit freedom
to rotateanddisplace
on the
application
of central
load. Thelineof actionof theloadpasses
accurately
through
thecenterof thespherical
seat,whichis located
at thecenterof thebesring.
Displacement
Measurements
Thecoordinate
displacements
of thejournal
endsrelative
to the
bearing
weremeasuredly
themechanical
arrangement
illustrated
in figure7 inwhichhorizontal.
andvertical
motions
aretransmitted
by bronze
riderson leversthrough
vertical
ralsto fourO.0001-inch
dialindicators.Thisis thessmesystemreported
in reference
1 exceptthat
thelevers
weremcd.ified
to givea 2:1magnification
of thedisplacementsto improve
theaccuracy
in reading
thedials.A duplicate
setof
rodsis usedfortemperature
compensation.

l&iction
Torquemeter
Friction
torquewasmeasuredlythehydraulic
torquemeter
shownin
figure6. Thetorquemeter
consists
of twol/2-inch
pistons
whichare
connected
to themastextending
fromthetestbearingso thatthepiston
forcesopposeeachother.A flowof lightoilis forcedintothespace
at theheadof eachpistonandis discharged
through
portspartia12y
covered
by thepistons.Thefriction
torqueof thebearing
applies
a
forceto thepistons
to causea slightchangein thedischarge
portarea,
resulting
in a pressure
differential
on thetwopistons
to balance
the
applied
force. Thepistons
arecentered
in theirboresby apressurizi.ng
systemofre~evedareas,thuseliminating
frictional
contact.
Thedifference
in pressure
on thetwopistonsis indicatedby
a
mercury
manometer
andis proportional
to thetorqueon thetestbearing.
Manometer
readings
wereconverted
to friction
torqueby direct
caMbration
underrunning
conditimsby app@inglmuwnincrements
of moment
to thebear~ hubby theweightsshownin figure6 andrecording
the
corresponding
changes
in themamxneter
readings.
OilFlow~asurements
AS shownin figure6, a drainholein thebottomof themachine
housing
allowstheoilflowing
fromthetestbesring
tobe collected

.. .____

. ..

mcllTN 3491

in a panwhichmaybe removed
forweighing.&o shownaretheslinger
ringsamdbaffles
whichprevent
themidng of thetest-bearin~
oilwith
theoilfromthesupport
bearings
andthe-spherical
seat. -

-rature

Measurement

tion-constaxtan
thermocouples
wereusedto measure
bearing
temper.
aturesat 14 locations
in thebearing
hubwithin1/16inchof thebearing
surface
as shownin figure8. Thermocouple
9 gavethebearing-hub
temperature
at a point2 inchesradially
fromtheoilfilm,andthermocouple16wasusedto determine
thetemperature
of theincoming
oilat
theoilinletto thetestbearing.
TESTPROCEDURE
Forthefiveshafts,
twoseriesof experiments
wereconducted
at
a nmiberof constant
speedsandinletoilpressus. Displacemmt
experiments,
in whicheccentricity
wasmeasured,
weremadeon allshafts.
IRriction
andoilflowexperiments
weremadeseparately
fromthedisplacementtestsbecauseof thenecessity
of disconnecting
thebronzeriders
whichwerea sourceof unwanted
friction.Measurements
of friction
and
oilflowweremadeon thelongerbearings
of ~ = l; and2.

Displacement
Experiments
Variations
in displacement
wereobtained
by varying
thecentral
loadat constant
speed.Itwasfoundthatby operating
underconstant
speedthevariations
inbearingtemperature
weresmall.A nearly.
constanttemperature
of theapparatus
at thermal
equi~brium
assured
a
minimum
thermal
effectin themeasurements.
Beforethedataweretaken,
thetestelements
andthemeasuring
apparatus
werebroughtto equilibrium
temperature
by rumning
themat constant
speedandcentral
loadfora
periodof 20 to 30 minutes.
At theendof thewarmuppericxi,
thenetcentral
loadandtheinlet
oilpressure
werereduced
to zeroso thattheindicator
dialscouldbe
setto a datumposition
representing
zerodisplacement
of thejournal
wasthenapplled
and
relative
to thebearing.An inletoilpressure
dataas readfromthefourdialswerethen
heldconstant.Displacement
recorded
forvarious
increments
of central
load. Theloadwasincreased
in fourto eightincrements
andthendecreased
in thesameincrements
to permitaveraging
of data. Thefollowing
datawererecorded:Journal
displacements,
capsule
pressure
of central
load,bearingtemperalnmes
at

.-

-- -- .. ___

._

....._.__

... ..._

10

NACATN 3491

critical
locations,
inletoilpressure,
speed,direction
of rotation,
inletoiltemperature,
roantemperature,
andoiltemperature
afterthe
heater.Thesameprocedure
wasfollowed
fortheopposite
direction
of
rotation
so thatthedatacould%e averaged
forthetworotations.
As
shownin tableI, constant-speed
runsat approximately
500,1,200,2,500,
and5,0CK)
rpmwereconducted
foreachof thefivetestshaftsto determinetheeffects
of Z/d andclearance
on thebearingcharacteristics.
me ~
app~edcenttiloaci
in tietestswas1,713poundson s~t 6C
andthe~unit
loadonprojected
areawas 874psi on shaft6D.
Themaximum
bearing
temperature
recorded
was198F at 7,000rpmfor
s~t 6D.
Friction
andOilFlowExperiments
Friction
torqueandoilflowrateweremeasured
simultaneously,
withthedisplacement-measuring
apparatus
temporarily
disconnected
in
orderto eliminate
thefriction
of theridersandtheoilflowto them.
Theprocedure
ofwarming
up andapplying
increments
of loadat constant
speedwasthesameas in thedisplacement
experiments.
At eachload,
aftertheoilflowstabilized,
thefric$ion
torquemeter
wasread,and
theflowratewasdetermined
by weighing
theoilcollected
in periais
of 1 to 2 minutes.
Thefollowing
datawererecorded
foreachloadcondition:
l&lction
torqueas indicated
by themanometer,
weightof oilflowfromthetest
bearing,
timeof oilflow,capsule
pressure,
bearingtemperature
at
critical
locations,
inletoilpressure,
speed,direction
of rotation,
inletoiltemperature
, roomtemperature,
andoiltemperature
afterthe
heater.Thesamedataweretakenfortheopposite
direction
of shaft
rotation
so thatdatacouldbe averaged
forthetworotations.
As
shownin tableII,constant-speed
reins
at approximately
5~, l,2~j
and5,0~ rpmwereconducted
on thelongj~ruals6A,6B,and6c. The c.
msximumapplied
central
loadwas1,688poiihds
witha maximum
unitload
onprojected
areaof 59 poundsper squsreinchon shaft6c. me maximumbearingtemperature
recorded
was181F at 5,OOOrpn.

REsuErs
Theexperimental
dataon eccentricity
ratio,friction,
andoil
flowrateareshownin figures
1 to 4 in a formintended
tobe tieful
in thedesignof plainbesrings
withvaluesof Z/d as highas 2. The
nondimensional
coordinates
shownarethosedetermined
fromtheanalyticalsolution
of reference
1. I!&
coorctha.te
pointvaluesshownwere
determined
fromthefollowing
expressions
by usingtheWmwn quantities
fromtheexperimen&l
tests.

NACATN 3491
IOadnuniber:

Eccentricity
ratio:
n = e/~ =

&v)2
+(%)2

nv = +cr

(4)

nh = e~Cr
Bearing
friction
ratio:
b
~=

(5)

Couplefriction
ratio:
L@
()()

(6)

Oilflowfactor:

q=

(7)

YCd2C@@)Il

Ihlet-oil-pressure
nuniber:
1/%0 = (l/@(pO/p)

. .. .. ..

. .. .-.

._._.

._.

.. ._ _

____

TtACA
TN 3491

12

LoadNunber
Theexperimental
quantities
in theloadnuuiher
(eq. (3)) arethe
load,viscosi~,speed,diametral
clearance,
andbearingdiameter
and
length.Theunitload p on a projected
areawasdetermined
from
by thepistonwas calcuP P/td andthetotalnetload P applied
la=tid
from P = 5.0(pc- 12.5) where pc is thecapsule
pressure
acting
on thepistonareaof 5.0squareinches.Thetareweightof thebearing
andattached
appsratus
is represented
in thesetestsby a pressure
of
12.5poundspersquareinch.
Viscosi@valuesfortheSAIZ
10 oilweretakenfromthecurveof
.
figure9 andconverted
to reynsby divl~ thecentipoise
valuesby
6.9x 106. Oilfilmviscosityp wasdetermined
fromthetemperature
of thermocouple
k located
1/16inchfromthebearingsurface.
Dismetral
clearance
at roomtemperature
wasdetermined
frommeasurementsofbearingandjournal
withelectrolhit
gages.Forrunning
conditions,
clearance
wasdetermined
by subtracting
thechangein clearance
givenin figure10 fromroom-temperature
clearance.
Theclearance
in
figure10 represents
thecalculated
clearance
chsmgecaused
by differ1 as a function
of the
entialthermal
expansion
ofbearingandjourna
temperature
gradient
in thebearing
housing.Thermocouples
k and9 were
1/16inchand2 iJ2Chf33
, respectively,
fromthebearingsurfaces,
giving
temperature
differences
(* - @) inMcativeof thegradient.Theslope
of thecurvein figure10 is appro-tely one-third
of theslopeof
thatusedin reference
1. Additional
studyindicated
thattheear~er
expansion
problem
wasreevaluslopeshould
be smaller.Thedifferential
atedanalytically
by themethodof mshenko (ref.6).
Eccentiici@
Ratio
Calctitions
of experimental
valuesof eccentricity
ratio n in
figurel(a)weredetermined
fromthecoordinate
vertical
andhorizonti
displacements
of theshaftmeasured
at theriderslocated
beyondthe
endsof thebearingl= inchfromthecenterMne of thebearing(cf.
16
fig.5). Eightobservatima
of a coordinate
displacement
wereaveraged
to determine
themeanvalueat a givenload. Readings
of twodials,
oneat eachend,as theloading
wasbeingincreased
andthendecreased,
Fourmoreobservations
wereavailable
forthe
gave.f
ourobservations.
opposite
direction
of rotation.
Ih orderto correct
thevaluesof theeccentricity
ratio,the
slightdisplacements
at zeroloadandzeroinletoilpressure,
at which
it wasassumed
thatthejournal
andbearingaxeswerecoincident,
were

13

NACATN 3491

averaged
andsubtracted
fromtheaveraged
valuesof displacement
under
load. Ml displacements
werehalvedbecauseof the2:1multiplication
of themeasuring
system.
Anothercorrection
wasmadeto account
fortheeffectof shaft
deflection
on displacement
measurements.
Calculations
of shaftbendi&
deflections
weremadeassmning
a uniformly
distributed
loadoverthe
lengthof thebearingandvar@ngmmen%s of inertia
M-the sMghtly
stepped
shaft:Mflectioncorrections
weremade,on
measured
vertical
displacements
only.,Because
thejournal
withinthebearinglengthis
deflected
to a fourth-degree
parabola,
thecorrections
weremadeto
simulate
thevertical
displacement
of an ideaUystraight
journal
located
one-fifth
of theheightof theparabola
below~ a~x. Thevertical
andhorizontal
components
ofeccentrici~
ratio + and ~ werecalculated
by dividing
thecorrected
coordinate
displacements
by %heradial.
clearance
(eqs.(4)), andtheeccentricity
ratio n was calculate&
as
thesqme rootof thesumof thesquares
of theccmiponents.
Theexperimental
dataof eccentricity
ratioshownin figurel(b)
forshortbearings
arefromreference
1 andsreshownforcomparison
withfigurel(a)forlongbearings.
Bearing
IZriction
Ratio
Experimental
valuesofbearingfriction
ratiO l?b/l?oin figure2(a)
werecalculated
fromeqwtion(5)inwhichthefriction
force Fb was
determined
fromthemercury
manometer
readings
of thetorquemeter.
Multiplication
of themanometer
valuesby thecalibration
factorFc
gavethefriction
torque.Theaverqge
valueof Fc was0.475poundinchof torqueperinchdifference
ofmercury.Dividing
thefriction
torqueby thebearing
r@ius gavethefriction
force ~.
Theexperimental
friction
datdforshortbearings
fromreference
1
areshownin figure2(b)forcomparison
withthelong-bearing
datain
figure
2(a).
Couplel?riction
Ratio
datashuwnin figure2 represent
measured
values
lhe
friction-ratio
of friction
actingon thebearingof thejournal-bearing
combination.
As discussed
in another
section.
of thisreport,
thefriction
ratioof
thejournalFj/Fo maybe obtained
analytically
by addingthecouple
friction
ratio Fb/Fo
. Thecouple
friction
ratio AF/Fo to thebearing
friction
ratiois of interest,
sincewhenmultiplied
by (d/Z)
2 it is

NACATN3491

14

a single-line
function
of theloadnumberas shownin figure3. The
partof thefriction
on therotating
meniber
whichis dueto thecouple
formedby thecentral
loadandthelateral
displacement
of therotating
memberrelative
to thestationery
meniber
is represented
nondimensionally,
by @F..
Theexperimental
datashownin figure3 werereduced
fromdata
givenby McKeeandMcKee(ref.7) andPE&u13mhandBlair(ref.8).
Bothinvestigators
present
friction
datafromtestsin thefour-bearing
data
machine
whichmeasures
journal
friction.McKeeandMcKeepresent
in experimental
curvesof f againstZN/p with cd/d and Z/d ratios
given,andPalmlichandBlairgivetabular
experimental
dataof f(d/cd)
2 withthe Z/d ratiogiven.Thesedataareconagaitit@/p(d/cd)
expressions:
vertedto Fj/Fo and l/~ valuesby thefollowing

f = f#P

(9)
.

(lo)

F. = 2&lNzd(d/Cd)
(Petroff
friction)

TheMcKeesmdMcKeeconversions
are
6.9x 106x 60

(n)

i = (~/p)(d/Cd)
*(Z/d)
2

fp

2x2pNf2d(d/cd)
f(P/Zd)
21-c2pN
(d/Cd)

P-(]

cd (%/d)(2/d)2

@ F

(c#d)(Z/d)*
(12)

15

NACATN 3491
ThePalsulich
andBbir conversions
we

(13)

Fj
=
F.

(fd/cd)
z2 1

(14)

(!()
21f2-~

from
Valuesof M/Fo(d/Z)
2 shownin figure3 weredetermined
equation
(6)usingtheexperimental
valuesof Fj/Fo andtheanalytical
valuesof Fb/Fo fromtheanalytical
curveof figure2.
OilFlowFactor
Theoilflowfactorq in figure4(a)was calcu&ed from
equation
(7)as in reference
1. Theoilflowrate Q in cubicinches
persecondwas converted
fromthemeasured
flowin poundsperminute
Because
usingthespecific
gravi~of theoilat bearingtemperature.
displacement
measurements
werenotsimultaneous
withmeasurements
of
(7)weretaken
oilflow,valuesof eccentricity
ratio n in equation
fromtheexperimental
dataof figurel(a)forcorresponding
testconnumber 1ICpoj~fist which q iS
ditions
. Theinlet-oil-pressure
plotted
in figure4, wasdetermined
fromtheloadnumberandtheratio
(8).
pressurep as in equation
of inletoilpressurep. to unitbearing
Exper-ntaloilflowfactors
fortheshortbearings
of reference
1
areshownin figure4(b)forcomparison
withthelo~-bearing
datain
figure4(a).
Experimental
datafromMcKee(ref.9) forabearingwlthasingle
oilholearealsoshownin figures
4(a)and4(b)andcompare
wellwith
theexperimental
dataof thisinvestigation.
Additional
McKeedata
fromreference
9 enablea curvetobe drawnfora bearing
havingan
axialoil~oove whichis somewhat
shorter
thanthelengthof thebearing.
ANALYSIS
ANDDISCUSSION
Thedependence
of eccentrici~
ratio,friction,
andoilflowon
theloadnumberl/~ is shownin figures
1 to4. The (Z/d)2term

. .

NACATN3491

16

in theloadnumberservesto reduceexperimental
dataforvarious
values
of 2/d to nearlya singlecurve.A singleexception
is thecurvefor
eccentricity
ratioin figurel(a)in whichtheexperimental
datacluster
neara singleHne onlyif the (2/d)2termin theloadnunberis taken
as unityforthelongbearings
forwhich Z/d is weaterthan1. For
thebearingfriction
ratio
in figure2,
thecouplefriction
ratioin
figure3, andtheoilflowfactorin figurek, the (Z/d)2termappears
tobe basica22y
applicable
forbesrings
withvaluesof Z/d from1/4
to 2.

.
&

Figures
1 to 4 areintended
to serveas designchartsforbearings
havingloadnumbers
up to 100covered
by theexperimental
datafrom
thesetests.As shownin figurel(a),eccentricity
ratiosaveraging
n= 0.93 wereattained
experimentally
at loadnumbersof approximately
90.
Abovethis,thevalueof eccentiici~
ratioapparently
becomesasymptotic
to n= 1.00 as theloadnmiberbecomesWinite.
Eccentricity
Ratio

Twosetsof experimental
dataon eccentricity
ratioappearin figurel(a)andaxeshownseparately
in figures
l(b)andl(c). Thosein
figurel(b)areforthelongbearings
of Z/d= 3 to 2 fromthe
experiments
usingshafts6A to 6E. tifigwe l(cf aredatafromreference1 forshortbearings
of Z/d= 1/4to 1. Itmaybe seenthatthe
makes
deviceof assuming
valuesof Z/d of unityforthelongbesrings
thelong-sadshort-bearti
eccentricity
d&tanearlycoincident.
The
in theloadnmiber,
unitbearing
pressure
on projected
area p appears
ratherthanthe.total
load,andtheabovedevicecanbe interpreted
to
meanthattheafiowable
unitbearingloadis practically
independent
of
Z/d forvaluesof Z/d abovelbut thatforvaluesof Z/d lessthan1
theallowable
unitbearingloadf~
as thesquareof Z/d.
Theexperimr@xal
titsof figures
l(b)andl(c)arealsoshownin
figures
n(a) andn(b) usingthecapacity
number& as theabscissa
l/~. In figure11,the
ratherthanitsreciprocal,
theloadnumiber
usefulloadregionis compressed
in therangeof abscissas
fromO to 0.1,
whereasin figures
l(b)andl(c),whichusetheloadnuniber
l/b, this
regionextends
from10 to infinity
in proportion
to theunit.load.
The
load-nuniber
methcdofplotting
avoidsdevoting
90 percent
of thegraph
to theL@ht-loadregionbelowa loadnumberof 10.
Theexperimental
valuesof eccentrici~
ratioarematerially
larger
thanthoseindicatedby
theanalytical
curveof theshort-bearing
approximation.Someof thedifference
betweenthecurvesmaypossibly
be due
to thesensitivity
of theexperimental
datato smallerrorsin determina:
tionof eitherthecoldclearance
or theclearance
at running
temperature.

NACATN 3491

17

Usualmeasurements
of thebearingclearance
tendto givereadings
at
thetopof thewaviness
of thesurfaces,
ratherthantheeffective
disrm+r. An analytical
correction
forthechangeof clearance
at
running
temperature
wasusedin calculating
theexperimental
dataand
of
shouldbe included
in makinguseof thecurves.An appreciation
themagnitude
of thechangein eccentrici~
corresponding
to a difference
in n valueof 0.90to 0.95is givenby considering
thatthisrepresents
a changein displacement
of aboutO.~O@O inchin thesetests.
Perhapsthemostimportant
factorrelated
to thedifference
between
theanalytical
curvegivenby theshort-bearing
approximation
andthe
e~rimentaldatais theomission
in theanalytical
solution
of the
effectof thecircumferential
pressure
cliff
erential
on theleakage
in
solution
thecircumferential
direction
of theoilfilm. Theanalytical
includes
thesideleakage
andthecircumferential
flowdueto rotation,
buttheomission
of partof thecircumferential
leakage
mentioned
would
theoretically
causetheexperimental
valueof n to be abovetheanalyticalcurveby a small.
unknown
amount.It is alsopossible
thatthe
difference
betweentheanalytical
andexperimental
curvesis related
to
theuseof a temperature
nearthebearingsurface
to obtaintiscosity,
ratherthanthetemperature
of theoilfilm. A constant
viscosity
was
alsoused,ratherthanone&angingfrompointto pointwiththeoil
filmtemperature
andpressure.
Allof thesefactors
arerepresentative
ofmethodsthatwould
commonly
be usedin predicting
bearing
performance
fromthecurves.
Theexperimental
curveis thusrepresentative
oftheeffecttobe
e~ectedwhenthesemethods
areused.
Considering
thesefactors,
a singleexperimental
Me is shownin
figurel(a)whichis theaverage
of au of thedatain bothcurvesl(b)
andl(c).Whilefigures
l(b)andl(c)differslightly
in thelight-load
region,
thesingleexperimental
lineof figure1(a)is considered
to be
a practical
interpretation
of thedata. Forpractical
useandtobe on
theconservative
side,it is recommended
thattheexperimental
curvein
figurel(a)be used,ratherthantheanalytical
curve.
b figurel(a)theexperimental
valuesareabout15 percent
greater
thantheanalytical
onesforloadnunibers
between10 and100. Forload
nunbers
aboveabout60 it is an advantage
to havean approximation
that
becomesasymptotic
to a valueof n oflinthesam waythat,the
than100
analytical
curvedoes. Forvaluesof theloadnur&r greater
theexperimental
valueevidently
liesbetween
theanalytical
value
Largerthan0.93as itbecomes
asymptotic
and1. Thiseffectcanbe
valueof n.
obtained
by adding0.5(1- n) to theanalytical
the
Blending
of (1/d)2transition.
- Thedeviceof considering
(Z/d)
2 termin theloadnuniber
as havinga valueof 1 forvaluesof Z/d

18

NACATM 3491

greater
than1 canbe expressed
mathematically
by considering
the
Z/d termto havean unlmown
exponent
of x. Thevalueof theexponentx
changes
from2 forvalues of Z/d lessthan1 to zerofor .Z/dgreater
thBxl
1. Thetransition
of theexponent
froma valueof 2 to zerocan
thenbe considered
to occurin a smoothmanner,
andtheshapeof the
transition
curveandtherangeof Z/d inwhichthe transition OCCllrS
remainto be determined
experimentally.
Thedifference
betweena smoothblending
of thechangein the
by thefactthatthe
exponentx andan abruptchangeis minimized
transition
occursin theneighborhood
of a valueof Z/d of 1, where
thevalueof (Z/d)xis 1 regardless
of thevalueof x. Theabrupt
changemaybe sufficiently
accurate
forpractical
purposes,
as indicated
by thefollowing
arbitrary
exsmple.Assumethatthevalueof x should
be 2 at Z/d= 1 andthendecreases
~early to zeroat Z/d= 1.2
where (1.2)0=1. Themaximumchangeoccursat a valueof Z/d of
about1.1where (Z/d)x
= (1.1)
1,withtheeffectdecreasing
in either
direction
from1.1. Thusthemaximum
effectof thismethcil
ofblending
is a changein loadnumberof about10percent
at midrange,
if therange
inwhichthetransition
occursis 0.2pointsof Z/d.
An~ical curves
.-Analy-ticalJy,
information
on eccentricity
ratio
against
loadnuniber
forshortbearings
is available
fromthesolutions
by theshort-bearing
approxhnation
(ref.1) andby themethodof Cameron
andWood(ref.2). Theseareshownas theuppergroupof curvesin
figures
12 and13. Thecurvefromtheshort-bearing
approximation
is
a singlecurvetheoretically
applicable
to shortbearings.TheCsmeron
andWoodexactsolution
of Reynolds
equation
givespointsindicating
of theseupper
separate
curvesfor Z/d= 1/4,1/2,and1. A comparison
curves
withtheexperimental
curvewhichis repeated
in figure13 shows
thattheplotting
of thedataas a function
of theloadnumberusing
Theuseof theload
the (Z/d)2termfor Z/d< 1 is justifiable.
nuniber
in thismannerpermits
directcomparison
of analytical
and
experimental
dataforau valuesof Z/d on onecurve.Thelocations
of theCsmeron
andWocxi
pointsarealsofoundto be in relatively
close
agreement
withtheexperimental
curvesadsubstantiate
theuseof the
curvein figurel(a)forpractical
purposes
whenusingtheloadnuniber
aa a parameter.
Ih figmes 12 and13,thelowergroupme analytical
curvesfor
infinite
valuesof Z/d fromtwosolutions
by Scmmerf
eld (ref.3) and
valuesof
onefromCsmeron
andWood(ref.2). Thesecurvesforinfinite
Z/d of necessity
referto theabscissa
scalelabeled1/S sinceno
It canalsobe saidthatthe
finitevaluesof Z/d areavailable.
curvesareplotted
in accordance
withthedevicetiatinfinitely
long
bearings
areconsidered
tobe of Z/d= 1, suchthattheSommerfeld
that
numberapplies
ratherthanthecapaci~number.It is interesting
theinfinite-hearing
curvesagreewiththeshort-bearing
curvesat the

.--

*
\.,

NACATN 3491

19

highloadnumbersalthough
theyaregreatly
divergent
at lowloadnumbers.
solution
for
It isperhaps
forthisreasonthatuseof theSomnerfeld
bearings
with Z/d= 1 hasbeensuccessful.
However,
forbearings
solution
wouldbe lessconshorter
than Z/d= 1, useof theSommerfeld
frcxn
1.
servative
as Z/d decreases
Thedifferences
in thethreecurvesshownforinfinite
values
of Z/d area resultof different
assmnptions
as to thecircuml?erential
extentofpressurized
oilfilm. TheSommerfeld
solution
basedon a %t
or 360etientof thefilmassumes
thatnegative
pressures
contribute
to
theloadcapaci~of a bearing.Ih theCameron
andWoodsolution,
the
uponassumptions
filmextentis somewhat
greater
than1800jdepending
regarding
boundary
conditions
of thecircumferential
pressure
gradient.
feldspressure-distribution
Thethirdcurveshownisbasedon Sommer
function
forsm infinite
bearing,
buttheintegrations
to determine
film
loadcapaci~as a function
of n arebasedon a pressurized
of m or 180extent,
thuseliminating
theeffectof negative
pressures.
Equations
(15)and (16)belowshowthedifference
in theanalytical
functions
of theSommerfeld
typeon thebasisof extentof theoilfilm.
Forthe % film
1/2
d
2=(2
+n2)(l-n2)
=PN
s
(15)
p cd
X%2n

()

Forthe m film
pNfd\2=
S.yy

(2+n21-n2
)(
-

- #n2)+ kn~~

(16)

of equation
(16)areshownin
Theintegrations
forthe
determination
theappendix.

Comparing
thelowergroupof curvesin figure13whichreferto
bearings
of infinite
valuesof Z/d withexperimental
dataof Z/d
of 2 or lessseemsto indicate
thattheassumption
of Z/d of infini~
givestoolowa valueof eccentricity
ratioat lightloads.The

assumption
of infinite
valuesof Z/d el.imina$es
thkeffectof end
leakage
on thepressure
distribution
in theoilfilm. Thespreadof
of Z/d of 1.5and2 is closer
theGta in figurel(b)forbearings
to theuppergroupof curvesat ldghtloadsthanto thecurvesfor
infinitk
valuesof Z/a.

..

...-

20

NACATN3491
I!earing
Friction

Curvesofbearingfriction
ratio ~/F. forlongbearings
and
shortbearings
areshown,respectimly,
in figures
2(a)and2(b). Both
curvesshowtheexperimental
dataplotted
as a function
of theload
nuniber
l/~ inwhichthe (Z/d)2termis included.Itmaybe seen
thattheshort-andlong-bearing
experimental
data,whencompsred
with
theanalytical
curve,arenearlycoincident
without
distinguishing
betweenshortor longbesrings
in theloadnuniber
as is donein the
curvesof eccentricity
ratio.Thecurveobtained
by meansof thedevice
of using 1/S forthelongbearings
in which Z/d is takenas unitv
is shownh figure14(a)forthesameexperimental
dataof figure2(a).
It maybe seenthatthesamedatasxeinbetteragreement
withtheshortbearinganalytical
curvein figure2(a)thanin figure14(a).It is
interesting,
however,
thattheexperimental
dataforlongbearings
in
figure14(a)arein fairlyreasonable
agreement
withtheanalytical
curvesforinfinite
valuesof Z/d by Cameron
andWood(ref.2) and
by theSommerfeld
solution
for1800filmextentof thisreport.
Anal#sical
curvesforbearing
friction
ratiofromthesolutions
of
Cameron
andWod (ref.2) andtheshort-besring
approximation
(ref.1)
agreewellwhenplotted
a@nst loadnumberl/Cn forshortbearings.
Theseareshownin figure14(b)for Z/d= 1/4,1/2,and1. Originally,
theCsmeron
andWoodcurvesforfriction
wereplotted
on coordinates
of f(r/cr)against
theSommerfeld
nwiber.A repottingof theircurves
ofbearing
friction
as in figure14(b)showstheinfluence
of (Z/d)2
in drawing
thecurvesfor Z/d= 1/4,1/2,and1 together.

Journal
l?riction
ratio F3/Fo against
TheCameron
andWoodcurvesof joun&lfriction
loadnumberin figurel~(a)showa greater
divergence
with 2/d thanthe
bearing
friction
curves.
in figure14(b).Theyalsoshowthatthedifferencebetweenjournal
friction
andbearing
friction
maybe greatforhigh
leadnunibers
andlargevaluesof t/d. Thessmecurvesplottedon coordinates
of f(r/c&)againstS alsoshowa difference
betweenjournal
andbearing
friction,
butthisdifference
is difficult
to distinguish
friction
sincethehighloadvaluesplotneartheorigin.By plotting
ratioagainst
loadnuniber,
thedifference
betweenjournal
andbearing
friction
is shownmoreclearly.
Analytically,
theshort-bearing
approximation
doesnotindicate
a
difference
betweenjournal
andbearing
friction
becauseof theassumptionttitthecircumferential
veloci~profile
in theoilfilmis linear.
Ih thisrespect,
theagreement
of theCameron
andWoodandshort-bearing
curvesforbearing
friction
is of considerable
interest.Fromequilibrium

21

NACATN 3491

conditions,
thebearing
friction
torqueis theoretically
lessthanthe
journal
friction
torqueby thecouplePe sin~. Curvesforjournal
friction
canbe determined
fromrelationships
of theshort-bearing
approximation
by addingthiscoupleto thefriction
torqueof thebearing.
!lhe
curvesof journal
friction
in figure15(a),labeled
short-bearing
approximation,
arebasedon thisaddition
in accordance
withthefollowing
analysis.
Thedifference
betweenjournal
andbearing
friction
torqueAF(d/2)
is equalto thecoupleof theload P andthelateral
displacement
of
thejournal~=esin~:

A+=

Pe sin@

(17)

lh termsof friction
ratio:

g=

%&-E)

(18)

Substitution
of P = ~N(d/cd)2(t/d)2(l/~)Zd,
e = n(cd/2),
and
thePetroff
frictionF. = 2#pNZd(d/cd)
gives

& _ n sin@ (Z/d)2

F.

2f12

(r

AFd
F. 1

(19)

Cn

.nsin@l

(20)

2fi2 ~

..-

Since sin@ andtheloadnumberl/Cn arefunctions

of n as
givenin reference
1, thecurveof (Al?/Fo)
(d/t)
2 against
loadnumber
fromequation
(20)is a singlelineas shownin figure3.
ratiowhichis addedto
Theratio Al?/Fois thecouplefriction
thebearing
friction
ratioto obtainthe~ournal
friction
ratio
(21)

.
.

.....

._

..

.- ..

NACATN 3491

22

As shownin figure15(a),theplotof journal

friction
ratio
against
loadnumbergivesa familyof curves
whichtiein closeagreementwiththecurvesfromtheCameron
andWoodsolution.
~rimental dataon journal
friction.
- Sinceexperimental
dataof
journal
friction
axenotavaila%le
fromthisinvestigation,
datafrom
McKeeandMcKee(ref.7) andPalsulich
andBlair(ref.8) areshownin
figures
15b) to l~(f)forcomparison
withtheanalytical
curvesfrom
equations
19)and
(21)
.
Bat-a
fromthese
investigators
wereobtained
[
fromthefour-bearing
machine
whichmeasures
thefriction
torqueof the
journal.
McKeeandMcKeepresent
datafora 1# -inch-diameter
%earing
with
a rangeof clearance
ratiosfromO.00~ t; O.O&O for Z/d= 1/4,1/2,
3/4,1, ad 2.8. Comparatively
lightloadswereused,buthighload
nunibers
of theorderof 5~ wereobtained
at lowrotative
speeds.These
dataareshownin reference
7 as experimental
curves,
ratherthanpoints,
plottedon coordinates
of f (friction
coefficient
= F/P)againstZN/p.
Sincethe Z/d andcd/d information
aregivenwiththecurves,
conis possible
as
version
of thedatato Fj/Fo and l/Cn coordinates
shownin figures
15(b)to 15(e).

Figures
15(b)andla(c)compare
theMcKeeandMcKeedatawith
equation
(21)fortheshortbearings
of Z/d= 1/4and1/2 andfigures15(d)and15(e)compere
themfor Z/d= 1 and2.8. Thesecurves
showthatagreement
of theanalytical
andexperimental
curvesisbette(r
for Z/d= 1/2,1, and2.8 thanfortheveryshortbearingof 2/d= 1/4
wheretheloadswereverysmall.-Since
onesetof datafor 2/d= 1 is
givenin tabular
formin reference
7, it is possible
to showpointdata
in figures
15(d)and15(e)forthefriction
risebeyondthehookpoint
in theregionofboundary
lubrication.
Thesequence
of datapoints
aroundthehookis shownby the ZN/p curveinsetin figure15(d).
Thepointsabovethehookrapidly
leavethescaleof the Fj~Fo curve.
of slope 0
Thehookpointis indicated
on the Fj/Fo curveby an increase
of thelinewhichtendsto risealmostvertically
to pointsoffthe
vertical
scale.Themorerounded
thehookon the ZN/p curve,the
moregradual
thechangeof slopein the Fj/Fo c~e.
Highloadnumbers
areshownin theexperimental
curvesof figures15(b)and15(c)for Z/d= 1/4and1/2 becauseof theinfluence
of the (t/d)2 termin theloadntier, andthenwdmumloadnunibers
for Z/d= 1 and2.8 sremuchsmaller.
PalstichandBlairpresent
experimental
datafroma four-bearing
machinesimilar
inprinciple
to theMcKeeandMcKkemachine
butmodified
to operate
at higherloadsandgreater
speeds.Dataarepresented
for
b

NACATN 3491

23

a bearing
approximately
2 inchesin diameter
and11 incheslongwitha
F
diametral
clearance
of 0.004inch. Speedsas highas 5,000rpmandunit
loadsto 4,000poundsper squareinchwereused. Theexperimental
data
srepresented
in reference
8.usingcoordinates
of f(r/c)and S,where
c is theradialbearingclearance.
se y be Convertid
b %/Fo
and l/Cn valuesas shownin figure15(f).An exactvalueof Z/d is
notgiven,buta inspection
of thetabular
dataindicates
thatthebearing
was2* by Ii inchestith Z/d= 0.607.Figure15(f)showsthecomparison
of experhental
datawithequation
(21).Wad numbers
in excessof 1,OCO
wereobtained
as shown. Theexper~ental
dataagreewellwiththeanalytical
curvebelowa loadnumberof 1~ butarebetweentheanalytical
curvesof I/d= 0.607and1 at thehighloadnumbers.In reference
8,
theplot
of thefriction
dataon f(r/c)and S coordinates
didnot
exhibit
a hookpointevenat theveryhighloadnumbers.
An important
feature
of theanalytical
curvesin figure15(f)is
thaton thecoordinates
showntheyindicate
Journal
friction
to be
several
hundred
percentof thebearing
friction
at highloadnumbers
andhighvaluesof Z/d.
BoththeMcKeeandMcKeeandPalsuMch-Blair
datawereusedto
determine
thecouplefriction
ratio AF/Fo forplottingAl?/Fo(d/t)2
dataof couplefriction
ratioweredeteras in figure3. Experimental
minedby subtracting
theshort-bearing
analytical
valueofbearing
friction
ratiofromtheexperimental
journal
friction
ratio.As shown
in figure3, mostof theexperimental
dataareabovetheanalfiical
beyondthe
curve.TheMcKeeandMcKeecurveswhichdidnotcontinue
hookarein relatively
goodagreement
withtheanalytical
line,except
forthedataat 2/d= 1/4,whichareoffthescale.Themethodof
plotting
thedifference
of journal
andbearingfriction
as in figure3
magnifies
thedifferences
shownin figures
15(b)to 15(e)andisworthy
of wideracceptance
as a toolto magnify
thespreadof friction
data
at highloadswhicharehiddenneartheoriginof the ZN/p curves.
valuesof 2/d.-Analytical
curves
icalcurves:infinite
Analyt
ofbearing
andjournal
friction
forbearings
of infinite
lengthare
shownin figure16 on coordinates
of friction
ratioand 1/S. The
threesetsof curvesshownarederived
fromtheanalytical
solutions
of Cameron
andWood(ref.2) andSommerfeld.
Twosetsof theanalyticalcurvesarebasedon theSommerfeld
pressure-distribution
function,
onesetas givenby Sommerfeld
(ref.3) fora ~ or 360pressurized
filmandonesetfora m or 1800pressurized
filmas determined
in
theanalysis
of theappendix.In allthreesolutions,
thefriction
force F is determined
fromintegrations
of filmshearing
stresses
theeffectof filmpressure
on theshearing
stresses
over360. However,
depends
upontheangular
extentof thepressurized
film. As shownin

...

24

NACATN 3491

theappendix,
thefriction-ratio
curvesfortheSomerfeldm casemay
be determined
fromthesumof twointegrations,
onefromtheSommerfeld
shearing
stressfunction
forthepressurized
filmof m extent,
andone
fromtheshort-%earing
approximation
function
forthenonpressurized
film
of n extentwherea linearcircumferential
velocity
profile
is asswmd.
Equations
(22)and (23)belowshowthedifference
in formof thefrictionratiofunctions
of the 2JCand m casesof theSomnerfeld
solution.
Forthe 2YCfilm
21- ~2)

Fb
=
F.

(2+ n2)(l- n2)1/2

(22)

9=
F.

21+2n2
2 1/2
(2 -1n2)(l- n )

Forthe z film
Fb 1
=F. 2

(4 - n2)
2+n21-n
)(
(

2 1/2
)

(23)

AE shownin figure16jtheSOmmerfeld
m curveandtheCameron
andWoticurveforbearing
friction
arealmostidentical
in theloadnumberrangeshownandarein fairagreement
withtheexperimental
data
forlongbearings
in figure14(a).Theanalytical
curvesforjournal
friction
aresomewhat
divergent,
theScmmerfeld
% curvegivingthe
largest
friction
values.
TheSommerfeld
% solution
showsthegreatest
difference
between
bearingandJournal
friction
as shownin figure16. lhis
is a direct
resultof theassumption
in the ~ solution
thatnegative
pressureq
contribute
to theload-carrying
capacity
of a bearing
andcausedisplacements
alwaysnormalto theloadlineas shownin theclearance
circlediagram
of figure17. Themomentarmof thefriction
coupleis
alwaysgreatir
therefore
in the % solution
thanthemomentarmsof
the x solution
or theCameron
andWocdsolution
forthesamevalue
of n.

.- .

:U

NACATN 3491

25

IfinfiniteZ/d analytical
solutions
areto be usedin longitwouldseemthattheCameron
andWocil
finite-be@ngappl.icat~ons,
andSommerfeld
x solutions
arethemostrealistic
sincemostexperimenters
havereported
displacement
curvesof thesemicircular
form
shownin figure17.
.,

Thedisplacement
curvefortheSommerfeld
z caseshownin figure17 isbasedon equation
(24)deliermined
fromtheanalysis
in the
appendix:
(24)

oilFlow
figure4(b)
Theexperimental
dataforoilflowfactorq shown.in
arefromreference
1 butarereplotted
in orderto showoilflowrate
as a reciprocal
function.Theinlet-oil-pressure
number l/~. containstheloadnumberl/en andtheratioof oilsupplypressurep.
in reference
1 is
to theunitbearing
pressure.Thecurvepresented
shownforcomparison
in figure18. Bothcurvesshowtheexperimental
dataforshortbearings
havinga singleoilfeedholelocated
opposite
theload. Alsoshownaredatafromtestsby McKee(ref.9) for%earings
witha singleoilholeandforbesrings
withan axialoilgroove.The
curvesof figure4(b)applyto bearings
of Z/d up to 2, although
only
theshort-bearing
dataareshownbecause
thelong-bearing
dataplotso
neartheorigin.An enlarged
plotwithscalesfivetimessizeis shown
separately
in figure4(a)withdatafor ~ = 1, 1A
*Y and2. Thesolid
d
linedrawnthrough
theexperimental
dataforlongbearings
witha single
oilholein figure4(a)is thesamelineshownin figure4(b).

As in~catedin reference
1, theoilfluwfactorq (eq.(7))is
theratioof totaloilflowrate Q issuing
fromtheendsof thebearing
discMgh.g
frOIII
to theanal@icaUydetermined
flowrate fidZcd(N~/2)n
flowrate,based
thepressurized
180of thebearing.Theanalytical
on theshort-bearing
approximation,
maybe viewedas a volumetric
displacement
ratein cubicinchespersecond.It is equalto thedifference
filmthiclaless
betweentheflowrateentering
at thepointof maximum
filmthiclmess,
andthatleaving
thebearing
at thepointof minimum
undertheeffects
of rotition
andviscosity,
at zerocircumferential
experipressure
gradient.Thetotalflowrate Q wasdeterminedly
ment.al
measurement.
By empirical
methodsofplotting
theexperimental
data,itwasfoundthattheoilflowfactorq appears
to be a function
of theinlet-pressure
capacity
numberho:

.z

....

...

NACATN 3491

26

thereciprocal
formo? thecur&s of figures
k(a)andk(b)
However,
seemsto give--even
bettercorrelation
as a function
of theinlet-oilI
as
given
in
equation
(
8).
~resswember 1I%0 (l/cn)(Po/P)
IE@ures
k(a)andk(b)areintendedto
serveas a meansforthe
determination
of oilflowrate Q fromknownconditions
involving
the
variables
in theloadnuniber
andthelumwnoilfeedpressurepo. It
shouldbe noticed
thatthedeticeof taking Z/d as 1 forvaluesof
Z/d greater
than1 is notusedfortheoilflowcurves,
exceptindirectly
whenthevalueof n so obtained
entersintothecalculation
of the
analytical
partof theoilflowrate.
.
Figure4 is notintended
to be usedas a deviceto predict
the
oilsupplypressurep-o.No criterion,
otherthanoperating
temperature,
valueof oilflowfactorq or
is available
whichspecifies
theminimum
performing
successfully
on thick
oilflowrate Q. Casesofbearings
filmsunderstarved
conditions
of oilflowareknown,butthedegreeto
whichstarvation
canbe tolerated
is notkn6wn.lHgurek(a)indicates
thatbearings
hatig Z/d greater
thsn1 areaptto havean oilflow
factorq lessthan1, whichmaybe an indication
thatoilgrooves
are
desirable
forlongbe~gs to ficrease
oil fluwandloweroperating

-rature as Wcatedby figure37 in reference

1.
USEOF CURVES
l?Lgures
1 to 4 maybe usedto predicteccentricity,
minimum
film
thiclmess,
bearing
andjournal
friction,
journal
powerloss,andoil
thevariables
contained
flowrateforbearings
to I/d= 2. However,
in theloadnumberandinlet-oil-pressure
numbermustbe known,including
therunning
temperature
of thebearing
whichdetermines
filmviscosity.
Inletoilpressure
mustalsobe lumrn.Forbearings
of I/d of 1 or
less,theloadnumbermaybe calculated
including
tie (l/d)*termto
determine
theabove-named
performance
characteristics.
However,
for
bearings
of Zjd ~eaterthan1, twoloadnumbers
arerequired.The
loadnumber1/S withoutthe (z/d)*termis usedto determine
the
eccentric@ratio n fromfigurel(a)whichisusedin thecalculation
of oilflowrate Q fromfigure4. Ioadnumberl/Cn including(Z/d)2

-.

_ -

NACATN 3491

27

is required
in thedetermination-of
besring
andjournal
kriction,
couple
friction
ratio,andinlet-oil-pressure
nuniber.
maximum
bearingtemperature
by theuse
A methcdof approximating
of a heatbalance
diagram
is described
in reference
1. lRtgures
2 and3
canbe usedto estimate
bearing
friction
andtheincrement
of couple
friction
to obtainthefriction
of therotating
element
whichgives.
thepowerlosssndheatgenerated.
Mgure 4 givesoilflowd.ata
for
usein estimating
theheatcarried
awayby theoilflow.
CONCLUSIONS

Thefollowing
conclusions
maybe drawnfromtheresultsof the
experimental
investigation
of eccentricity
ratio,friction,
andoil
flowof longandshortjournal
bearings
withloadnumbercharts:
1. The-performance
of plainbearings
understeadyloadis reduced
to single-klne
chartscovering
theusualrangeof length-to-dismeter
ratios Z/d from1/4to 2 by theuseof theloadnumberin &manner
described.
2. Inclusion
of the (Z/d)2termin theloadnumbercollects
the
experimental
datain a singlelinein allcasesexceptoneof eccentricity
ratiowhere Z/d valuesgreater
than1 ne takenas uni-&y
in
computing
theloadnumber.
in theloadnuniber
to determine
3. The (t/d)2termis.included
thefriction-ratio
andtheoilflowfactorfrcnn
thefigures.
4. Theaddition
of theincrement
of friction
torqueformedbythe
sideward
displacement
of theloadto thefriction
torqueon thestationary
element
to predict
thefriction
torqueoftherotattig
element,
heat
generated,
andpowerlossis supported
by analytical
curvesandexperimentaldataof otherinvestigators
whohavemeasured
journal
friction.
5. Themethodof plotting
friction
in twoparts,thebearing
frictionratio ~/F. andtheincrement
of theloadcoupleAF/Fo,permits
hydrodynamic
friction
tobe determined
fromsingle-line
curveswhich
havebeenshownto applyto theusualrangesof clearance
andlengthof friction
fromratiosrelated
to
to-diameter
ratio.Theprediction
thePetroff
friction
at no loadis morerational
thantheuseof a
friction
coefficient
basedon load.
to
6. Thereciprocal
formof theloadnumberl/Cn is preferable
~

Theloadnumberincreases
in directproportion
to
as a parameter.

28

NACATN3491

unitloadandis properly
weighted
to havemeaning
as a quantitative
measureof relative
loading.Theareaon thegraphdevoted
to the
13.ght-load
regionis minhized,andtheeffectofhighloadnunibers
is shqn moreclearly.
Cornell
Universi@,
Ithaca,
N. Y.,August20,l-.

29

NACATN 3491
APPENDIX
INI!Q3RATIONS
FORSOMMERI!ELD
Y( CASE

Thefo120wing
integrations
givingeccentxici~
ratio,attitude
angle,andfriction
ratioarebasedon theSomerfeldpressuredistribution
function
(ref.3) fora bearing
without
endleakage.However,a basicchsmgein assumptions
ismadeconcerning
thecircumferentialextentof thepressurized
film. Sommerfelds
integration
to determinetheresultant
forceor load P includesthenegative
pressure
regionof thediverging
wedgeof theoilfilmas we= as thepositive
pressure
regionof theconverging
wedge. Tn effect,
a working
film
of & or 360is assumed
suchthatthenegative
pressure
regioncontributes
to thesupport
of theload. Thefollowin$
analysis
isbased
on theassumption
thatsupport
of theloadby negative
pressures
is
negligible
andthattheworking
filmis in theconverging
wedgeandis
of YC or 180in circumferential
extent.Thesameassumption
ismade
in theshort-bearing
approximation
(ref.1).
Eccentrici@-Ratio
Function
Thefollowing
equation
is thepressure-distribution
fuhction
given
thedatumfilmpressure
at the
by Sommerfeld
in which PS represents
leading
andtrailing
endsof theconverging
wedgeof theoilfilm:
6VUr nsine(2+ncos
Pfilm~
2+n21+ncose2
)(
[(

e) +P

(27)

to be zerogagepressure
such
Ih thefollowing
stepS Ps is assumed
thatequation
(27)givesfilmgagepressures.
Theanvariablee
is measured
fromthelocation
ofmaximum
filmthickness.
The X and Y coordinates
arechosenparallel
andperpendicular,
respectively,
to thelinejoining
thejournal
andbearingcenters
which
components
of theresultantP
areat an eccentricity
e. Tbeorthogonal
determine
inte~ations
over m radians
are Px and py. T& following
ratio n:
therelationship
of theload P as a function
of eccentrici~

....-

30

NACATN 3491

P~=z

r cose de
J e=o~film

~2

d 2
=@12
2+n2 )(l-n2
cd
()
.

l-sine de
z *filmlr
o

J
.:!.

n sin2e(2
+ n cose) de
(2+ 112)(1
+ n .0se)2
d2-

()

@L ~

&m
~2 1/2
(2+n2 )(l-n )

(29)

,.

PX2+.PY2

>.

Substituting
P= pld and U

I-KIN

(2 + n2)(l- n2)

&M[Yt2(l
- n2)+ 4n~1/2
.

.-

-.

(31)

31

NACATN 3491

Equation
(31)is thesameas equation
(16)in thebodyof thisreport.
Theanalytical
curvesof eccentricity
ratioagainst
Sommerfeld
number
representing
equation
(31)areshownin figures
12 and13.
Attitude
Angle

Theangular
location
of theresultantP withrespect
to theX-tis,
or attitude
angle #, is determined
fromthefollowing
expression:
tan@=

Y
Px

=&&!!
2n

(32)

Theattitude
anglealsogivesthean*-po8itionJof theminimum
film
(32),
thiclmess
relative
to theloadlhe as shownin figure17. Equation
whichis thessmeas equation
(29),showsthedisplacement
pathin figure17 to be nearlysemicircular
ratherthans~aightas in theSOmmerfelil
% solution.
l?riction
Ratio
Equations
forfriction
forceson thebearing
andjournalare
givenby
%mnerfeld(ref.3) forthe Zh radians
or 360extentof thepressurized
film. To determine
thefriction.
forces~ and Fj Orlbearing and journal,
respectively,
forthe%mmxrfeld,
x case,witha pressurized
film
extending
1800or 1/2of thearc,haH of thevaluesgivenby Samerfeld
maybeusedto represent
theeffectof thepressurized
halfof the oilfilm
on friction:

()

Fb = pu2~d

23cl-n
(

2)

(33)

1/2
(2+ n2)(l. n2)

(*)
To equations
(33)~d (%) mustbe tidedthefriction
of thenonpressurized
filminwhichthevelocity
profile
is Mnear. Halfof the
valuesgivenby theequations
of theshort-bearing
approximation
may
tieusedto represent
thisfrictionF:

..

...

32

NACATN3491

Thetotalfriction
forcesFb and FI erethesumsof equations(33)and(35)andequations-(~)
and(35)
:
Fb+F

Fj+F
241

%)[(2

2n2)

+ n2)(l- ~2)/2+ (, . :2)/:

ti(k
+ 5n2)
@J2~
cd 2-1( n2)(l-n2)12

(37)

IMvidingFb and Fj by theload P in equations

(36)and(37)
results
in equs.tiuns
in termsof coefficient
of frictionf~ and fj.
SLibstituting
u = YCdl?
, P =pZd, and S= (@/P)(d/cd)2,
fb(d/Cd)
=S

~24-n2
(
)
(2+ n2)(l- n2)1/2
*2(4+ ~n2)

f3(d/cd)
=S
(2 + n2)(
1- n2)1/2

NACATN 3491

33

To obtainfriction
in termsof friction
ratio,equations
(36)and

(37)aredivided
by thePetroff
valueof frictionFo:
I-r(4
- n2J

I.LUt
(d/Cd)

Fb
=
F.

2 1/2
(2+ n2)(l- n )
=1
z

(4- n2)

,(40)

)
(2+n2 )(l-n2i2

(4+ ~n2)

q=~

TO

(41)

(2 + n2)(l- n~l2

(23)andare
Equations
(40)and (41)arethesameas equations
shownplottedin figure16 forcomparison
withtheSomnerfeld
21
solution.

-.

-. .. ...

..

~-

.-. .

34

NACATN 3491
REFERENCES

l&edW.: Analytical
1. D@ois, GeorgeB.,andOcvirk,
Derivation
and
Experimental
Evaluation
of Short-Eearing
Approximation
forIiU
NACATNS2808
Journal
Bear@S.
NACARep.11.57,
1953. (Supersedes
~d 2809.
)
2. Cmeron,A.,andWoml,W. L.: TheFullJournal
Bearing.Proc.
Institution
Wch. Eng.(Icmdon),
vol.161,W.E.P.Nos,.
47-54,
1949,pp.59-72.
Theoryof Lubrication
Friction.
A.: TheHydrodynamic
3. Sommerfeld,
Zs.Math.undphys.,
vol.-5O,
nos.1 and2, 19&, pp.97-155.
4.DuBois,
G. B.,Ocvirk,
F. W.,andWehe,R. L.: Experimental
~vestigationof Misallying
Couples
andEccentiici@
at theHs of MisalinedPlain
Bearings.NACATN3352,l=.
Murray:OilFlow,KeyIbctorin
D. F.,andRosenld.att,
5. Wilcock,
vol.74,no.5,
Sleeve-Bearing
Performance.
Trans.A.S.M.E.,
JtiY1952,PP, W9-866.

6. !l?imoshenko,
of Materials.
PartII - Advanced
Theory
S.: Strength
andproblems.Secondcd.,D. VanI?ostrand
Co.,Ihc.,1941,
pp.258-264.
of Journal
Bearings
as
7- McKee,S. A.,andMcKEe,T. R.:Fciction
Influencedby
Clearance
andLength.Trans.A.s.M.E.,
u-51-15-161,
vol.51,Aug.1929,PP. 593-595.
8. Palsulich,
J.,andBlair,R. W.: TestingofHighlyIoadedSleeve
Bearings.SAEJour.,vol.~, no.9, Sept.1946,pp.481-490.
Bearings.Trans.A.S.M.E.,
9. McKee,S. A.: OilFlowinPhin Journal
vol.74,no.5, J- 1952,pp.841-848.

TA31E I.- EWHUWRTAL

COIiDITIONS
IN DI??PMCEMENTExPERIMmm

[& -in.-dam. steeljcmmal and bronzebemingj

8
SAE10
Ofl

Shal?t l/d

6A

cd>
in.
0.-52

Ofl SU@id

hole qqx)aitiC@ti=

-d,
N,
m

62

&z

1,438
1,538
1,653

3/4

&g
1~

170

1.20

223
266

5,(YW
.3,030

1,4T8

yil

132
156

1,438

3&l

938

329

134

80

31?a
938

137

32$)

1.28
132

o.oo142 1,2W

1%3

w
2
2-40
140

2,500

1,438

508

151

5,030

1,713

601

179

80
80
80
80

523
814

133
fig

5,m

988
1,538
1,610

Y,Coo

1,660

;$

80
m

128
139
158

::
40

o.m183 o.cxl133 1,203

O.o@8

80
M

1,q

yo

2,5~

6E

m
9

oil inlet
presmme,
Po) w4
~.

1,mo

1,200

kamm
UIllt WxLm!Jmbearing
lad, p,
*$,
q)
lb/6qin.

2,50il

1,2CN

6D

Icai]

0.cH3376 o.oo.?~

0.00M

1./8-in.
diem.

1,007
1,293

5,CCI0
2

MdJm.unlcaa,
P, lb

0.00183 5m,
l,m
5,m

63

thrcq@

o.(xl187 1,(X20
2,500
5,CW

489
838
m.

,%
343
g

TABLE II.- ExPERIMmmL

CONDI!IICONB
IN FRICTION M?O OIL FU)W EXP~S

fl-in.-diem. steel Journal W

[ 8

br~ze bearing; ME 10 oil mrpp~ed

+bX@il
8@-iL
-&W

Ofi

ihaftl/&

6A

cd)

k
&

sped,

~,

WsLimsm
lad,

0.OCE-52o.00IEFJ1,203

o.w~
6B 2 o.oc576

P, lb
1,283

hole

OppOBi&

COdHd

bad

M3dmurn bearing

=y

T4>

t~.>

lb/aq in.

OF

338
338
g;

136
135
133
130

248
458

132
tig

220

1,2(XI
1,2(XI
500
500

l,2a!3
1,288
938
938

5,000

1,663

5(XI

836
1,138
1,663
1,663

~8
@8

w
136
160
158

938

329

132

529
529

1-33
134
Ml
1~

;$%
5,0co
6C 1$ o.ool~o.cQ142 1,200
1,200
1,203
5,m
5,m

938
936
1,689
1,688

;%

a,l inlet oil flow,

pressure,
Q,
lb/ndn
Po) w.w ~.
o.M
R
120
80

.(2$8
.155
.m+

MO

.148
.285

80
:
120
40
1%
1%

6.332
.4M?
.648
.7%
O.*5
.do
.W
.132
.175

/.0

!2

>9
,$
.7
,6
5
.4(
,3
.Zc
.1

?!

.j

/0
Loa

D
FOR

(a) Data for both long

aM
frmu data in figures l(b) and l(c).
Figure l.- Eccentricity ratio aga~

load number.

2.

Data are cambined

(b) Data for long bearings with

Z/d .1,

~,

and 2

ti

.gh@ bearlnga with

table I for experherrtal conditions.

Figure 1.- Continued.

Z/d =

3/4. se,

?
h

(c)

/0

90

40

50

Datafor short bearinga with Z/d = 1/4,

1/2,
and 1.
reference 1.

60

?080

Experimental data are taken from

Figure 1.- Concluded.

w
w

I
I

(a) Data for long bearinge with

Figure 2.-

l/d = l% 9Jld
2.

See

table II for exprimerrtal condltione.

E
*
5

Bearing friction ratio againd load number.

~
F

?
,

Figure 2.- Concluded.

-F
m

~ O/4D

~UM-~y

/~N)

(.)2

(-)

F4
Figure 3.- Couple friction ratio agaimt load number.

[/v LITr-

(a)

Data

for

long

t7/L-p/?E5JoicG

bearinga. See table II

for

hUMB6E&,

experimmt.al

y~,

conditlona

ycn

of

SHE

6A,

6B,and6C.
G

Figure 4.-

Oil

flow

factor

againat

inlet-oil-pressure

number.

!,-

J---J-

\.. ,$
3

IW#

IT

70
(NL~T.

olk - PRESSUR

NfJM&!ER

%%

(b) Data for short bearinga.

Figure 4.- Concluded.

.gn$

Figure

5.-Configuration of temt bearing and aheft showing location of oil hole, thermocouples,
and dieplacemmt-mamring

riders.

Figure

6.-Schemat~c
tiagram
of appsratwforapplytng
central
1-, formeaauring
friction
torque on teat bearing, and for rceaauringoil flow through test bearing.

1
1
I

-%z/M3 Suff=twr

>

-qq:;-::

Figure7.-Mechanlcal system for maasuri~ displacement of Journal rehtive. to bearing.

51 m

r-h

/-45

+-.
---*,

OIL 70

BEARING
/47

q}

k7%-#-&AR--l=F=l
Figure 8.- Diagrem of thermocouple locations. All thermocouples except
I/16Inch frm bearing surface.

9 and16 arelocated
g
Q
y-

>

7U

NACAIIN
3491

49

/520

100

TEMPZEX?TURE,
.

ax?

300

Figure9.-Viscosity-temperature
characteristics
of S.AE
10 oilusedin
Measumentsof viscosity
weremadein a standard
experiments.
Saybolt
viscosimeter.

. _

.-__

. ...

20
/$

16

/.?

Jo

a
6

4
2
0

/234,5
CU4NG&

67
IN

DMII+KTKRL

@910//
f=Lg~K

IZ~S/+&

R~C~

~d.

Figure 10.- Changes in diemetral.clemce

as a function of temperature difference of two
points in teat-bearing wall. tk and tg are bearing temperatures at l/16 inch and
2 inches, res~ctively, from bearing surface. Running clearances of test bearing are
determMed by subtracting change in clearance frcnnrmm-temperatwe clearance.

E
b
a
g
P

(a)

Lcng

bearing,with

Z/d = 1, 1~, and 2

and short bar-

dataare sue aq In figure l(b); for

exprimmtd

with

Z/d m

3/4. ~e~ti~

conditions, see table I.

Figure11.- Eccentricity ratio against Smmerfeld number and capacity number.

vl

m
,

,6
,7
,6

..

Short beaings with

Z/d = 1/4, 1/2, and 1. _tiental

and are from reference 1.
Figure Il.- Concluded.

a!%ta are

same

as

In figure l(c)

1-

Figure K!.- Analytical curves of eccentricity ratio against lmd

number.

wl

-!=

,9

.8

c.4M6.eoM

SOMM6RFELD

.9

5HOR T. J9E14 R /NG

,A PP~OXIIWa
&.VD

TION
WOOD

.G

Nl

\
4
,3

,2

,
\

\ \

>
$

x.&

%
(y .
\

.*,C

mca

FigureI-3.
- AnEJ@Aml curvesof eccedmicity

ratio

Capaclty number.

against

Ekmnerfeld number and

(a) Coiuparimn

if

long-bearing data with analytical curves.

Figure 14,- Bearing friction ratio a@.mt

load number.

,,

LOa D

NUMBER]

%(592($)

(b) Analytical cwveE for short bea@ngs.

Figure

14.- Concluded.

i
t

(a) Analytical curves for long and short bearinga; Camron and Wood cum-m tdmn
reference 2.
Figure

15.- Curvesof Jcmrnalfriction

ratio ag*13t load number.

frm

58

NACATN 3491
.

(b)Comparison
of experimental
curvesofMcKeeandMcK&e(ref.7)with
-icdcmes
of equatio~(19)and (21);Z/d.1/4.

100
LORD

zoo

NUMB ERA

3C70

+%7

500

Y&O

(c)Comparison
of experimental
curvesofMc~e andMcKee(ref.7)with
analytical
curvesof equations
(19)and (21.);
Z/d= 1/2.
Figure15.
- Continued.

NACATN 3491

59

208

/00

~ 0.4D

~tiM8ER

3da

5U0

4.90

~ ~tn

(d)Comparison
of experimental
curvesofMcKeeandMcKke(ref.7) with
analytical
curvesof equations
(19)and (21);Z/d= 1.

qli

(e)
.

Comparison
of experimental curves of McKee and McKEe (ref. 7) with
analytical curves of e~tions
(19) and (21);Z/d= 2.8.
Figure15.- Continued.

. -.. ...z

..

__

. ..

.-

~
(f) Comparison of expdmmtal
data of Palsulich and Blair (ref. 8) with analytical curves of
equations (19)and (21); 10KW and short bearlnga.
Figure U.- Concluded.

ki
g
P

Figure 16.-

Analytical

curves

of

journal

friction ratio against

emd beax
values of 71 d.

1/S

for Imfinlte

62

NACATN 3491

SffORT

\ -.

6E4R

CA ME ROAJ

.-~o

ING
&UD

4PPzO%IM4710RV
WOOD

MM ERFKLD

GCCEAJTJZICITY

KG

T~O,

,/

.2

.3

.4
,5
.6
/

,7

Y2ii

(4

&
.9

Figure17. - Clearance
circle diagram showing paths of journal &kplace
ment relative to bearing under central loa~
in terms of eccentri
city ratio against altitude ar@e;

analytical curves.
.

. .

/?-8

PP
.20

,25

Lo30

.35

,40,

,, $-

Figure 18.- Oil flow factor against inlet-pressure capacity number for short bearings with
z/d = 1/4,1/2, 0.62, and 1. Exprhental
data are from references 1 and 9 andare also
shown in figure l(b).