https://ntrs.nasa.gov/search.jsp?

R=19900018747 2018-02-21T00:42:33+00:00Z

_rH

I''_ °

L_
m

'1
0_
1,-,,_. ii

©
 __ T

r. --v--_k

7_

<i

-£

_ £---- _.T£= _

L
NASA
Reference
Publication
1240

1990

Liquid Lubrication
"Space
in

Erwin V. Zaretsky
Lewis Research Center
Cleveland, Ohio

N/kSA
National Aeronautics and
Space Administration
Office of Management
Scientific and Technical
Information Division
Trade names or manufacturers' names are used in this report for identification
only. This usage does not constitute an official endorsement, either expressed or
implied, by the National Aeronautics and Space Administration.
 rapid evaporation of the liquid or semisolid grease lubricants
Summary
normally employed. Since lubrication ordinarily takes place
The requirement for long-term, reliable operation of aero- by means of a film entrained between sliding or rolling sur-
faces, the loss of this film due to evaporation and mechanical
space mechanisms has, with a few exceptions, pushed the state
of the art in tribology. Space mission life requirements in the working can result in failure of the mechanism.
early 1960's were generally 6 months to a year. The proposed With many metals the lubrication function is strongly
U.S. space station scheduled to be launched in the 1990's must influenced by the presence or absence of oxide films on these
be continuously usable for 10 to 20 years. Liquid lubrication metals. The surface oxides frequently act as protective films

systems are generally used for mission life requirements longer and, in some cases, contribute to the final surface films through
than a year. Although most spacecraft or satellites have reached either chemical reaction or chemisorption. At altitudes greater
their required lifetimes without a lubrication-related failure, than 89 km (55 miles) oxygen and nitrogen do not exist as

the application of liquid lubricants in the space environment the ordinary molecular species but rather in the atomic or ionic
state. The reaction rates between most metals and atomic
presents unique challenges. This report reviews the state of
the art of liquid lubrication in space as well as the problems oxygen are markedly different from those with molecular
and their solutions. oxygen. At altitudes greater than 1287 km (800 miles) atomic
hydrogen and helium are the principal species present.
Mechanisms in a closely sealed satellite are likely to operate
Introduction in a water vapor pressure of greater than 10 7 torr during
their first year in orbit. The atmosphere may include trace
Tribology problems have grown with aerospace advances
quantities of carbon monoxide, heavy hydrocarbons, and light
made over the past 30 years as shown in figure 1 (from Kannel
silicone polymers with an occasional burst of ammonia from
and Dufrane, 1986). In the beginning of the space age life
the decomposition products of the hydrazine motors. Some
requirements were minutes or hours. In the early 1960's they
components on the exterior of the satellite operate in a much
were generally 6 months to a year. Early deep-space probes cleaner environment, the gas pressure generally being less than
were notable exceptions, but these probes had relatively few
10- l0 torr, but with occasional bursts of ammonia (Robbins,
mechanical assemblies with high cyclic requirements. By the
1975). Because of the scarcity of oxygen, oxide films are
mid-1960's mission life requirements had increased to 3 to
formed at a rate inadequate for lubrication.
5 years. By the mid-1970's life requirements of 7 to 10 years
At extremely low pressure levels (where gas conduction and
were common (Ahlborn et al., 1975). The proposed U.S. space
convection are absent) the temperature levels will normally
station scheduled to be launched in the 1990's must be
be dictated by thermal radiation. Heat will be absorbed by
continuously usable for 10 to 20 years; that is, a 10-year design
radiation from any object that the mechanism "sees," and the
life requirement with a 20-year life goal (Dolan and
mechanism will, in turn, reject heat to outer space by radiation.
McMurtrey, 1985). Despite significant advances in lubrication
Various mechanisms will have different temperature levels
and mechanical component technology, the demands of these
depending upon their relative rates of heat gain and loss.
aerospace systems appear to grow faster than the technology. Robbins (1975) reports that the temperature is controlled by
Lubrication problems in space include
using carefully selected thermal blankets to balance the heat
(1) Very low ambient pressure absorbed from the 6000 K (10 341 °F) radiation of the Sun,
(2) Presence of atomic species other than the normally
the heat radiated by the satellite at 300 K (81 °F), and the
encountered molecular species
internal heat generation. He states that in order to provide an
(3) Thermal radiation
equable temperature for electronic components most satellite
(4) Absence of a gravitational field
mechanisms are designed to operate within a narrow
The absolute pressure outside the Earth's atmosphere (e.g.,
temperature band, usually in the range 280 to 320 K (45 to
above 1609 km (1000 miles) altitude) is approximately 10 - 13
117 OF). Lubricant evaporation from surfaces is a function
torr; the absolute pressure in interstellar space is approximately
of temperature. Hence, a mechanism's temperature is
l0-16 torr. Figure 2 shows pressure as a function of altitude
important to the lubrication process and thus to its survivability.
(Jastrow, 1960). The low-pressure environment contributes to
 Strategic _ .
Defense / :_pacec_an

Ini_"" Complexity

sta,c_ Tdbology
/ problems
Voyage,,/ / I"
Shuttle "_ /

_r
o

I.-

/ /"" lubrication (life) .... - Trlbology

/ _t*_LubricalJon ..-." ....... "" solutions
/ / dls_bu¢lon .... ---" ""
. . / /C oe .....-.--""

--- I I I I I
1950
1960 1970 1980 1990 2000 2010
Year

Figure l.--Growth of tribology requirements with advances in space. Kannel and Dufrane (1986).

1975). The application of liquid lubricants in the space

10, _ environment presents unique challenges. This report reviews
the state of the art of liquid lubrication in space as well as
the problems and their solutions.
1°°

Lubricant Type and Selection

r Corridan (1959), Freundlich and Hannan (1961), and Clauss
(1961) conducted early ball bearing tests in vacuum. They used
AISI 440C stainless steel instrument bearings with bore
- I I sizes from 3.18 to 9.53 mm (1/8 to 3/8 in.) and with oil-
0 4o0 800 1200 1600 20O0 2400 28OO impregnated, machined phenolic composition retainers. They
Al_tude, krn
obtained good performance with a silicone oil (General Electric
I I I I I Versilube F-50) (Bisson 1964). The operating times of these
0 400 80O 1200 1600 bearings were 1000 to 4600 hr, considered adequate for the
Altitude, mile,=
early 1960's. Today they would not be acceptable. Young and
Figure 2.--Pressure as a function of altitude. Jastrow (1960).
Clauss (1966) tested both AIS152100 and AIS1440C bearings
with phenolic cages lubricated with chlorophenyl methyl
Perhaps the only advantage of operation in space is the
polysiloxane oil at 10 -7 to 10 -9 torr and 422 K (300 *F).
influence of microgravity. Eliminating the weight of a system The bearings failed in less than 15 000 hr. Silicone-base fluids
from its bearing supports minimizes component wear and
are not considered to give good lubrication for long-term
reduces the probability of rolling-element fatigue. However,
operation because of their poor boundary lubricating properties
the bearings are preloaded to give precise location and to avoid
and their tendency to creep (Roberts et al., 1990).
ball skidding (Robbins, 1975). The resultant bearing load is Young and Clauss (1966) conducted additional tests with
usually 1 percent or less of its dynamic load capacity. (The dibasic-acid-ester-base lubricants: MIL-L-6085 (diester-base
dynamic load capacity is defined as a theoretical load that,
oil with corrosion inhibitor) and a dioctyl sebacate. Both
when applied to the bearing, will result in a life of 1 million
lubricants resulted in failures in less than 3600 hr. Failure in
inner-race revolutions.)
these bearings and in the silicone-lubricated bearings were
Spacecraft and satellites are lubricated with dry films, associated with either thickening or loss of lubricant. These
liquids, metallic coatings, greases, or combinations of these
tests suggest that the ester-base lubricants are less satisfactory
lubricants. Liquid lubrication systems are generally used for than the silicone-base fluids in vacuum even though the esters
mission life requirements of over a year, (Ahlborn et al.,
are superior lubricants for rolling-element bearings in air.
 YoungandClauss(1966)alsoperformedtestswith 2.0x103

naphthenic andparaffinicmineraloils.Thenaphthenic-oil-
lubricated bearingsfailedin743hrbecause ofacomplete loss
ofoil.Theparaffinic-oil-lubricated bearings lastedmorethan Z
1.5

12000hr withoutfailure.Thesetestssuggest thatin the

mid-1960's theparaffinic-base oil hadthebestpotential for 1.0

long-lifeapplications in space.
Inthe1970's twolubricants weredeveloped andsuccessfully tn .5
ground testedintheUnitedKingdom: asuperrefined mineral
oil(BP110) andatriester fluid(BP135)(Roberts etal.,1990). t
Theproblemsassociated with high vaporpressure and 0 30 60 90 120 150 180X106

highoutgassing wereavoidedby usingmolecular seals, Kinematic visCOSity at 38 "C, m2/sec

antimigration barriers,andoil reservoirs (Parker,1987).

Hence, thekeytolong-lifereliable applicationistomaintain Figure
various
4--Measured
lubricants at
bearing
25 °C
torque
for an
as a function
R-6 bearing
of lubricant
at 480 rpm.
viscosity
Kannc]
for
and
a lubricantpresence in thebearing cavity. Dufrane (1986).
Superrefined mineraloilsbecame thelubricants of choice
forsuchdevices asmomentum wheels, reaction wheels, and elastohydrodynamic (EHD) film thickness between the balls and
despinmechanisms. Themostpopular superrefined mineral the races.
oil lubricants areKG 80(Kendal) andApiezon C (Shell).
Interestin fluidsof variousviscosities alsoprompted the Perfluorinated Polyalkylethers
development ofsuperrefined gyroscope (SRG) lubricants.
These Roberts et al. (1990) report that in more exposed applica-
SR6 fluidsrepresentahomologous groupthatallowstheuser tions, or where operation at temperatures as low as 200 K
toselecta fluidhavingspecific viscosity characteristicsfor
(-100 °F) is required, or where a high viscosity index is
a givenapplication (KannelandDufrane,1986).Typical deemed important, the perfluorinated polyalkylethers (PVPE)
viscosity-temperature characteristics of theselubricants are
are widely used. Because minimal attenuation of infrared
giveninfigure3.Kannel andDufrane (1986) conducted torque radiation occurs with these fluids, they have been used as
testsonanR-6instrument bearing withthese SRGlubricants. lubricants within infrared scanners on spacecraft.
Theresultsof thesetests(fig. 4) clearlyshowthetorque
Table I (from Kannel and Dufrane, 1986) presents measured
penalty associated
withincreasing theviscosity ofthebearing evaporation rate data for superrefined mineral oil and PFPE
lubricant.However, therearetimeswhenthepenalty ofhigher
lubricants. PFPE lubricants have the lower evaporation rates.
torque mustbeoffsetbythebeating's abilitytodevelop agood
In general, evaporation rates tbr a given type of fluid decrease
as the viscosity is increased. For longer life applications the
10 -3 lower loss rates associated with higher viscosities are desirable.
However, the high-viscosity lubricants have the disadvantage
of higher operating torques.
10"-4
The general conclusion has been that the PFPE lubricants are
the most desirable for spacecraft applications. Three com-
number
mercial PFPE lubricants have been widely used: Krytox PR
_SRG 160
143AC (DuPont), Fomblin Z25 (Montecatini Edison), and
100
_EE 10 -5 Brayco 815Z (Bray). These fluids generally have the lowcst
KG 8O

6O
TABLE I.--EFFECT OF VISCOSITY ON EVAPORATION RATE OF TWO
LUBRICANTS IN A 0.133-MPa (10-6-torr) VACUUM

[Kannel and Dufrane (1986).]

4O

Fluid Published Evaporation

3O
viscosity rate at 40 °C,

20 at 40 *C, mg/cm2-hr
m2/sec

14 × 106 18
10 Superrefined paraffinic mineral oil (SRG 30)
27 13

I I 1 Superrefined
Perfluoro ether
paraffinic
a
mineral oil (SRG 40)
8 2.2
10 -6
1E-O 28 ,19
0 50 100 Perfluoro ether _
357 .0002
Temperature, °C Perfluoro ether _

Figure 3.--Viscosity as a function of temperature for a homologous series aT_pe and manufacturer not specified tn reference

of superrefined mineral oils. Kannel and Dufrane (1986).

 vapor pressure and highest viscosity index of lubricants
TABLE II.--PROPERTIES OF PFPE FLUIDS
considered for space application. Fomblin, which contains an
[Mori and Morales (1989a).]
acetal linkage, has a lower vapor pressure and a higher
viscosity index than Krytox. However, as more long-term
experience has been gained with these lubricants, degradation
Property
t Fluid

v
omblin
l
I Krvtox
of the lubricant has been observed. This degradation has $200
resulted in high bearing torque noise and excessive wear Demnum z25 ] z&56
(Stevens, 1983). Average molecular weight 840C
95{)01 11 OR)O
Kinetic viscosity at 20 °C, cS 500 + 25
The work of Baxter and Hall (1985) with Fomblin Z25 255] 2717
Viscosity index 210
suggests that the degradation is caused by the presence of Pour point, °C 355 i .......
-53
chemically active surfaces, wear particles combined with Density at 20 °C, g/ml l. 894 1.85t661 1.92:
- 15
exposed radicals in the fluid, or both. Zehe and Faut (1989) Surface tension at 20 °C, dyne/cm 19
and Carre (1987) suggest that metal oxide surfaces in contact Vapor pressure, torr: 251 19 i
At 20 °C 5×10 tt
with acetal-containing PFPE lubricants (i.e., Fomblin Z25) will At 100 °C 1 × 10 -7 [xlO-g t:[
2.9x10 IxlO-_ t._
3x10
inevitably result in acidic breakdown of the ethers. They
conclude that little can be done to the PFPE chains to block
the acidic attack without compromising the viscosity- formed. There were no gaseous products from the Krytox or
temperature qualities of the acetal groups. Demnum fluids.
Mori and Morales (1989a,b) studied the effects of Krytox Solutions to the problem of thermo-oxidative breakdown
16256 and Demnum $200 (Daikin) in addition to Fomblin Z25. should focus on the surfaces in contact with the liquid. Some
Fomblin Z25, a copolymer of perfluoromethylene and success in blocking thermo-oxidative degradation of linear
perfluoroethylene oxides, is the only one of three fluids PEPE'S below 570 K (566 °F) has been achieved with the use
that contains an acetal linkage. Krytox 16256 is a of additives containing nitrogen or phosphorous atoms. The
poly(perfluoropropylene oxide) with pendent -CF 3 groups. action of these additives may depend upon their ability to block
Demnum $200 is a poly(perfluoropropylene oxide). The acidic sites on the surface (Jones et al., 1983, 1985). Other
properties of PFPE fluids are sununarized in table II. approaches may involve surface modification to limit
Mori and Morales (1989a) found that PFPE decomposition decomposition, since it is the catalytic action of the resulting
and the resulting reaction products are dependent on the surface that accelerates the breakdown (Zehe and Faut, 1989).
particular molecular structure of the PFPE fluid when irradiated Table III (from Roberts et al., 1990) lists factors that both
by x-rays under ultra-high-vacuum conditions. They also promote and retard PFPE degradation.
studied the reaction of the PEPE fluids with AISI 440C stainless
steel during sliding under 10 -m torr at room temperature Performance Characteristics
(Mori and Morales, 1989b). All three fluids reacted with the
Thomas (1980), in order to provide data for ball bearing
AISI 440C material during sliding. Fomblin Z25 decomposed
torque calculations, performed sliding tests in air using AISI
during sliding and gaseous products, mainly COFe, were
52100 balls with the more common spacecraft lubricants

TABLE III.--FACTORS INFLUENCING PFPE DEGRADATION

[Roberts et al. (1990).]

Factors promoting degradation Faclors retarding degradation

Starved conditions
Fully llooded conditions
(e.g.. grease plating)
High specific film thickness {_x > 4)
Low specific fihn thickness (k < 1)
Fomblin Y type of perfluorinated oil
Fomblin Z type of perfluorinated oil base (branched structure); Krytox
base (linear structure) rarely used in Europe

Aluminum/titanium substrates at any' Hydrocarbon contamination

ambient temperature
AISI 440C bearing steel; ceramic
AISI 52100 bearing steel
coatings give further improvement
High ambient temperature (> 360 °C) (e.g., TiC -coated balls)

!Sliding surlaces Low ambient [emperature

Vacuum environment Surfaces where rolling motion takes

place

Normal atmospheric conditions

 TABLE IV.--COMMONLY USED LIQUID LUBRICANTS FOR
SPACE TRIBOLOG1CAL APPLICATIONS
[Thomas (1980). ]

Vapor Description and comments

Lubricant Manufacturer Viscosity
at 20°C, pressure
cS at 20°C,
torr

4.0× 10 u Vlineral oil with no additives: used as

Shell 250
ApiezonC
a reference oil

British 7O 7.9×10 '# Synthetic: triester base; boundary

BP 135 a
lubricant and antioxidant additives
Petroleum
520 3.7xl0 _ Mineral oil base; high viscosity;
BP 1 l(I a 3ritish
refined to give low vapor pressure;
Petroleum
boundary lubricant additives .

520 <10 _ Petroleum base with boundary lubricant

KG 80 Kendall
additive tricresylphosphate: also
Refining
contains an antioxidant additive
Company
240 <5.0x 10 ,2 Synthetic fluorinated oil: high density,
Fomblin Z25 Montecatini
low surface tension; high temperature
Edison
and high viscosity index; no boundar'
lubricant additives; high thermal
resistance

ainu bnger illarkelCd b,*ralor_ tcslcd hut nc_cr I]ov, n ua spaccclall

TABLE V.--COEFFICIENT OF FRICTION AND WEAR Lubricant Viscosity,

cS
RATE OF STEEL ON STEEL 1N LIQUID-
LUBRICATED SLIDING CONTACT O BP 135 70
lThomas ( 1980).] i'-1 Z 25 240
<> APC 250
Specific A BP110 520
l_ubricant Coulomb _OxlO '4
_7 KG 80 52O
friction wear rate after
1
coefficient 1000 revolutions,
in'/N-n1 25J

19.1_+0.9× 1015 20
Unlubricated, "failed" a0.46

after 100 re,,olutions _5

,20+_ .03 .93 ± .05 Z
Apiezon C
BP 135 .12 +- .015 1.04 +- .05 I I I I I I I J
.13+.015 .41 +_ .02
BP 110 .9.0
.13+.02 .69 +_ .03
KG 80
.49 + .03 30x10 -4
Fomblin Z25 .12+-.02

aFriclmn io_c rapidl) I_ Ihi_ _aluc

described in table IV. The results of these tests are presented

in table V. At slow speeds, where there is essentially no
_s[ Ibl I 1 t I I 1 I I I
elastohydrodynamic lubrication, and with light loads the 0 20 40 60 80 1O0 1 20 140 160 180
lubricating properties of the five oils were similar. Apiezon C, Time, hr
which had no additives, exhibited the highest mean friction
(a) Air.
coefficient of 0.20. The other four lubricants had lower friction (b) Vacuum (10 _' torr).
coefficients, ranging from 0.12 to 0.13. With the exception of
Figure 5,--Ball bearing torque as a function of lubricant viscosity. Bearing
the Apiezon C, which had no boundary lubricant additives, bore, 20 ram; contact angle, 15°; cage, one-piece phenolic (outer-race
all of the lubricants performed identically in the sliding tests riding); number of balls, 10: speed, 10 rpm; thrust load, 40 N. S|evens
in air. (1983).
Stevens (1983), using the same lubricants as Thomas (1980),
Figure 5 shows the mean torque, or the average of the torque
performed lubrication tests to determine the wear and torque variations, as a function of time lbr each lubricant. Although
of satellite, 20-ram-bore, angular-contact ball bearings
the torque differences between the five lubricants are not
operating at 10 rpm. That speed corresponds to the boundary considered significant, there appears to be a trend of increasing
lubrication regime. The tests were conducted both in vacuum
torque with increasing lubricant viscosity. This trend correlates
and in air. The results of these tests are shown in figure 5.
 withthetrendshown infigure4 byKannel andDufrane (1986). Ftom and Haltner (1968) list early satellite applications using
Theseresultsstronglysuggest thatthemagnitude anddif- chlorophenyl methyl silicones. The most common of these
ferencesin bearingtorquearedueto viscous effects.The fluids was General Electric Versilube F-50. However, the use
variationintorqueaboutthemean valuewasreported tobe of F-50 for space applications appears to have met disfaw)r
similarforthefivelubricants.
Thehighest torque wasobtained in the 1970's. The fluid polymerizes both in vacuum and in
withtheKG80lubricant, whichhasa viscosity of520cSat air because of microasperity interaction. Experience with
20 °C. Thelowesttorquewasobtained withtheBP135 chlorophenyl methyl silicones has also shown that, where there
lubricant, whichhasa viscosity of 70cSat 20°C.Onthe are high sliding velocities, large amounts of wear can occur
basisof workby ToddandStevens (1978)witha cageless in angular-contact ball bearings and sliding surfaces. Meeks
three-ballbearing,Stevens(1983)predicted thatthetorque et at. (1971) conclude that the estimated life belbre significant
duetospin,hysteresis, andmicroslipwouldbe14x 10.4 torque fluctuations (greater than two times average) for
N-re.IntestsbyStevens (1983)withKG80andApiezon C F-50-1ubricated, small bearings run at 55 rpm below 10 8
atspeeds of0.2and1rpmandathrustloadof40N,thetorque torr is less than l year. This result is due to thermally induced
ofthese bearings measuredat10rpmranged from15× 10 4 chemical changes in the oil. Meeks el al. (1971) predict that the
to 17x 10 4 N-re.At 10rpmandahigherloadof 400N ultimate failure life of these bearings witl be less than 2 years.
thetorque wastesswithKG80(120× 10-4N-m)thanwith Fleischauer and Hilton (1990) discuss the use of
Apiezon C (200 × 10 4 N-m). These results suggest that at
polyalphaolefin (PAO) and polyolester (f,f_) oils us possible
the higher load the torque is dependent on the sliding friction
substitutes for linear perfluorinated polyalkylethers. They
coefficients reported in table V by Thomas (1980). Hence, suggest that these oils can be synthesized and blended to
it may be reasonably concluded that at fractional speeds bearing
produce in a controlled way viscosities, vapor pressures, pour
torque is related to ball spin, hysteresis, and microslip. At light points, and other properties that can meet various spacecraft
loads and extremely low speeds the torque is related to requirements. These oils can also be blended with standard
lubricant viscosity, primarily because of cage friction. At low additive packages similar to those in conventional mineral oils.
speeds and high loads the torque is related to the Coulomb
However, there does not appear to be a definitive data base
friction coefficient (table V). In satellite applications the only that would allow using these oils in a satellite or spacecraft
load the bearing experiences in the space environment is the
system at this time with a reasonable degree of confidence.
applied bearing preload, which is generally low. If the speed
is high enough for a sufficient elastohydrodynamic film to
form, the bearing torque will be a function of the lubricant's
viscosity at the operating temperature of the bearing. Elastohydrodynamic and Boundary
In the 1960's Young et al. (1963) and Young and Clauss Lubrication Effects
(1966) performed extensive lubricant research in vacuums
ranging from 2 x 10 7 to 4 x 10-9 torr. The following types Lubricant Function
of liquid lubricant were endurance tested with both AISI 52100
and AISI 440 ball bearings: The primary function of a liquid lubricant in space appli-
cations is to separate surfaces in relative motion so that the
(1) Paraffinic and naphthenic mineral oils
surfaces do not sustain major damage and to keep the
(2) Chlorophenyl methyl polysiloxanc oils
coefficient of friction between the surfaces relatively low. A
(3) Mixed isomeric five-ring polyphenyl ethers
number of lubrication regimes, depending on the type of
(4) Diphenyl bis-n-dodecyl silane
intervening film and its thickness, can be identified. These
(5) Dibasic-acid-ester-base oils
lubrication regimes can be depicted by the Stribeck-Hersey
These lubricants are listed here in the order of the endurance
curve shown in figure 6 (from Jones, 1982). This curve takes
lives obtained with the bearings. The best results and the the [brm of the friction coefficient as a lunction of the
longest lives were obtained with bearings having a paper-based
parameter ZN/P, where Z is the viscosity o1"the liquid, N is
phenolic cage lubricated with a paraffinic-base mineral oil the velocity, and P is the load.
containing additives used for gyroscope bearings. The phenolic
At high values of ZN/P, which occur at high speeds, low
cage is used to retain and supply the oil. Linen-based phenolic
loads, and high viscosities, the surfaces are completely
cages with the same lubricant did not provide as good results. scparated by a thick (0.25 #m; 10 --s in.) lubricant film. This
The linen phenolic cages tended to dry out and crack. Tests
is the hydrodynamic lubrication regime, where friction is
were run with the chlorophenyl methyl polysiloxane "as determined by the rheology of the lubricant. For nonconformal
received" and after the 50 percent most volatile fraction was
concentrated contacts, where loads are high enough to cause
removed by distillation. These tests were run to determine
elastic deformation of the surfaces and pressure-viscosity
whether removing the most volatile material would lengthen
effects on the lubricant, another fluid film regime, elasto-
the operating time in vacuum. The as-received material gave
hydrodynamic lubrication, can be identified. In this regime
at least twice the life (Young et al., 1963).
film thickness may range from 0.025 to 2.5 _m (10 _' to
 I
10 4in.).Asthefihnbecomes progressivelythinner, surface I • _ Seizure
interactionsbegin.Thisregime of increasing friction,which I 0

I
combines asperityinteractions andfluidfihneffects, isreferred I
toasthe"mixed-lubrication regime."Finally,atlowvalues _o
I
I
oftheZN/Pparameter istheboundary lubricationregime.This E
I
regimeis highlycomplex,involvingmetallurgy, surface t .--Transition

topography, physicalandchemical adsorption, corrosion, f4 Z'

catalysis,
andreaction kinetics. Its mostimportant aspectis 8 / 1
thelbrmation of aprotective surface fihntominimize wear •"r t._ I"
I
andsurfacedamage.Theformationof surfacefilms is I
governed bythechemistry ofthefilm-lbrming agentaswell
I
asbythesurface ofthesolidandotherenvim.nm, entalfactors.
Theefl'ectiveness of surfacefilmsin mmmltzmg wearis A X B
y Z

determined bytheirphysical properties,whichinclude shear Rela_ve load

strength,thickness, surface adhesion, fihncohesion, melting Figure 7--Wear rate as a function of relative load, depicting various lubrication
pointor decomposition temperature, andsolubility. regimes. Beerbower (1972).

BesidestheStribeck-Hersey curve(fig.6)already described,

anidealizedplotofwearrateasa function of relativeload
canalsodelineate thevariouslubrication regimes andsome Region X-Y is the boundary lubrication regime. The degree
weartransitions.Region O-A offigure7 (fromBeerbower, of metal-to-metal contact and the wear rate increase as the load
1972)encompasses thehydrodynamic andelastohydrodynamicincreases. Wear is mild and tends to be corrosive to the left
lubrication
regimes, thelatteraspointA isapproached. Since of B and adhesive to the right of B. The location of B is quite
nosurfaceinteractions occurin thisregionexceptduring variable and depends on the corrosivity of the lubricant
startupandshutdown, littleor nowearoccurs,except that formulation. For a noncorrosive lubricant adhesive wear can
causedby rolling-element fatigue,whichcanoccurwithout occur at X. On the other hand, a corrosive additive can extend
surfaceinteractions. RegionA-X is themixed-lubrication the boundary regime to Z' before boundary film failure
regime,wheresurface interactions beginto occuratA and occurs. Region Y-Z is the severe wear regime, where severe
become moreprevalent aspointX isapproached. Wearislow adhesion and scoring occur. Mechanisms cannot operate
becausefluidfihneffectsstillexist. successfully in this regime, and therefore the location of this
transition point is quite important. At Z total surface failure
occurs, followed by seizure.
h ~ 0.025 to 2.5 pm(lO '6 to 10-4in.) In the boundary lubrication regime many properties of the
liquid lubricant become important: shear strength, film
thickness, melting point, and chemical reactivity with the
surface. Operating variables that affect lubricant film perform-
O. 00_5 p.rn(10 -7in.) I
h~ I _. ance include load, speed, temperature, and atmosphere, as
already discussed. Additives present in the lubricant to serve
specific functions, such as antiwear, antifoam, and antioxidant
additives and viscosity improvers, also affect behavior
(Jones, 1982).
.150 --_

Elastohydrodynamic Theory

Elastohydrodynamic (END) lubrication theory is well

documented by Dowson and Higginson (1966) and Hamrock
and Dowson (1981). The Grubin (1949) EnD fihn thickness
O
formula is as follows:
o .3 \ I/
H = 1.95 G°73U °73W o(_1
.ool I I I
I I I
I I t
where
ZN = IViscosit_)(Vel°ca_)-
p (Load)

H film thickness, h/R,

Figure 6 --Coefficient of friction as function of viscosity-velocity-load G materials parameter, o_E'
parameter (Stribcck-Hersey curve). Jones (1982).
 speed parameter, uqo/E'R _
The measure of the effectiveness of the lubricant film is the
W load parameter, F/E'R_
X ratio (i.e., the central film thickness divided by thc composite
o_
pressure-viscosity exponent, GPa I (psi -I) surface roughness of the rolling-element surfaces h,./o).
E' Usually the root mean square (rms) surface finishes of the
reduced modulus of elasticity.
contacting bodies o] and 02 arc used to determine the
composite surface roughness as follows:
1 1 (l-v_ l-v!'_
E'-2,, E, /
o= aT+o
It
average surface speed, (uj + u2)/2, m/sec (in./sec)
The )x ratio can be used as an indicator of rolling-clement
atmospheric viscosity, N-sec/m 2 (lb-sec/in.2)
performance and life. For h < 1, surface smearing or
R, equivalent radius in rolling direction, m (in.), deformation, accompanied by wear, will occur on the rolling
surface. For 1 < X < 1.5, surface distress may be
1 1 1
- + accompanied by superficial surface pitting. This type of pitting
R, R,. = R,.2 was observed by Stevens (1983) in oil-lubricated satellite ball
bearings. For 1.5 < X < 3, some surface glazing can occur
F normal applied load, N (lb) with eventual roller failure caused by classical subsurface-
El,E2 Young's modulus for body 1 and body 2, GPa (psi) originated rolling-element fatigue. At X _> 3, minimal wear
h film thickness, m (in.) can be expected with extremely long life; failure will eventually
be by classical subsurface-originated rolling-element fatigue.
pt,p: Poisson's ratio for body I and body 2
The most expedient, although not the least expensive, way of
The variations in EHD theory and the resulting formulas are attaining a high X ratio is to select a high-quality surface finish.
beyond the scope of this report. Suffice it so say that the The effect of film thickness is to reduce the magnitude and
variations between the EHD film thicknesses calculated by the instability of the torque, as shown in figure 9.
different formulas are less than the variations between the A good example of an effective lubricant film is provided
various sets of experimental data used to verify the theories by the NASA Solar Maximum Mission satellite, known as Solar
(Coy and Zaretsky, 1981). This is illustrated in figure 8, which Max. The 3-ton spacecraft came crashing into the atmosphere
shows the film thicknesses calculated by the theories of on December 2, 1989, after 9 years in orbit. In 1984, Solar
Hamrock and Dowson (1977), Grubin (1949), Cheng (1972), Max became the first satellite to be captured and repaired in
and Chiu (1974). A theoretical model for oil starvation in a space. It had been in orbit for 4 years when NASAspace shuttle
rolling-element bearing was formulated by Chiu (1974). astronauts replaced its inertial reference unit and returned the
Figure 8 also compares the results from Chiu's (1974) analysis original unit to Earth. The unit had three gyroscopes, each
with his experimental data and experimental data of Coy and with two ball bearings. One of these gyroscopes was
Zaretsky (1981). Oil starvation in the Hertzian contacts even disassembled and its bearings examined. The AISI 440C
under flooded conditions appears to be the primary cause of bearings had a 7-mm bore. They had run continuously at
the deviation of the experimental data from classical EHD 6000 rpm. The cages were made from a phenolic material
theory (Coy and Zaretsky, 1981). impregnated with KG 80 lubricant. The two bearings were
G = 5000 found to be well lubricated, with oil still in the cages. Iron
10-3 _" Hamrock and Oowson (1977} -.--, and chromium debris particles were found in both bearings.
-- G=_oo
The surfaces of the bearing raceways and balls had superficial
"_ --" Cheng (1972) --_ /.,,_ pits and scratch-like deformations. The wear debris, which
Orubin (1949)_ /_"'_ ,,.-,.,.__ O 00 oOQ0 was considered small, was attributed to the pits. This
micropitting appeared similar to and confirmed that reported
--" _1_-_. oO0 0 e_._--Chiu(1974}
by Stevens (1983). It was concluded that these bearings would
have lasted for their predicted life of 5 years (Uber, 1985),
=
__ 10 -5 and this conclusion was borne out by the additional 5-year
.5 __._oO-- operation of the Solar Max satellite.
,n • Chiu (1974)
O Coy and Zaretsky (198t)
E
i5 Design Criteria
lO-6 I _lJlJltJ I IlJlJhJ I IlflJll[ I illllhJ
Ahlborn, et al. (1975) report that they designed their
10 -8 10-7 10-6 10-5 10-4
bearings to operate at X < 1 and maximum Hertz stresses less
Dimensionless Contact lubricant flow number, GU = _Uho/R z
than 1.03 x 10 9 N/m 2 (150 000 psi). However, their test
Figure 8. Theoretical effect of kinematic starvation and inlet shear heating data indicate that failure to meet these criteria does not mean
on fihn thickness and comparison with experiment, Coy and Zarctsky ( 1981 ), that long life is precluded. They conducted two separate tests
 Torque T
measurement, 0.17
N-m

measurement. -4-_--5 sec

percent
(no film) _ (a) I [ I '

To. ue
measurement,
N-m
0.17

measurement,
percent 100 Yo
(very thin film) mE
Time

(a) Dry bearing.

(b) Two drops of oil in bearing.

Figure 9.--Angular-contact ball bearing torque and film thickness measurement with no lubricant and marginal lubrication. Bearing bore, 90 mm; contact
angle, 15"; thrust load, 267 N (60 lb); number of balls, 21; speed, 100 rpm; lubricant, Apiezon C. Torque spikes imply instability. Kannel and Dufrane (1986).

environment (10 -5 torr) caused the mineral oil to degas.

in vacuum with R-6 ball bearings (six bearings in each test,
Gases entrained in conventional oil lubricants include oxygen,
with different lubricants) at a maximum Hertz stress of
water vapor, and nitrogen. Oxygen and water vapor help
8.27 x 10 s N/m 2 (120 000 psi) and X of 0.3 and 0.4. There
protective surface films to form on the steel surfaces. These
was no detectable wear on the balls or races after 20 and
films minimize wear for two solid surfaces in contact. When
14 million revolutions, respectively. In another test they
the fluids are degassed, the beneficial effects of the oxygen
operated larger bearings for more than 2 years at 90 rpm at
and the water vapor are lost, and the wear of the surfaces in
a X of 0.5 with no change in performance. Stevens (1983)
contact increases. The lubricant becomes less effective in a
concludes that the predominant form of damage that occurs
vacuum environment than in air (Buckley, 1971).
on the races and balls is micropitting, some of which is caused
Reichenbach et al. (1964) repeated these tests with SAE 90
by repeated indentation of the steel surface by small, hard
EP gear oil (fig. 10(b)). SAE 90 EP is a high-viscosity oil
features. (Micropitting is also characteristic of microasperity
containing a full commercial additive package. Here again,
interaction under boundary lubrication conditions.) In tests
the wear was higher in vacuum than in air. Buckley (1971)
conducted by Stevens the scale of the damage was small and
concludes that the wear data indicate that the additives improve
had no significant effect on the torque characteristic of the
the load-carrying capacity of the fluid. However, it can be
bearing, even down to zero speed.
argued that if the additives are effective in vacuum, there
should be no significant difference between the amount of wear
Boundary Lubrication obtained in air and in vacuum.
Stevens (1983) concludes that the amount and type of
No definitive research has been published comparing the
damage change very little with the type of oil used in rolling-
effect of lubricant additives with various base stocks on rolling-
element bearings in space applications. As a result, the choice
element torque characteristics and wear both in and out of
of lubricant for variable-speed bearings in spacecraft appli-
vacuum. Reichenbach et al. (1964) conducted experiments with
cations can be made on the basis of vapor pressure and viscosity
fluid lubricants in air and in vacuum to determine the influence
without regard to boundary lubricating properties. Most space
of a modest vacuum of 10 -5 torr on the wear behavior of
mechanisms operate at relatively low speeds. From conven-
AISI 52100 steel. The experiment consisted of crossed
tional industrial experience, even though the bearings are
cylinders in contact, with one of the cylinders rotating. After
operating in the boundary lubrication regime, no significant
the cylinders had rotated in contact under fixed load for a set
damage or wear will occur at low speeds. This is not the case
time, the wear scars on the cylinders were measured. The wear
at moderate to high speeds. In a space application, without
scar diameters on the cylinders lubricated with a superrefined
effective coatings or boundary lubrication films or both, it can
paraffinic mineral oil are shown in figure 10(a) as a function reasonably be expected that adhesive wear and plastic
of contact load. The lubricant contained no additives. The wear
deformation or smearing of the surfaces will occur at these
scar increased with increasing contact load. The wear scar was
speeds. Figure 1 l(a) shows a ball-to-race contact region for
also greater in vacuum than in air because the vacuum
 2.2
440C steel ball

Asperity [-R a (cfa) <:0.011am

I
1.8

1.4

1.0

E
E. .6
440C steel race Martensite -.-' ,,
M23C6 --/ L Ra(cla ) &O.03 urn
E L ,
ell
(a) 10 .urn
"1o
.2 I I I I

1.8 TiC-coated
440C steel batr
O Vacuum (10-5torr) ,,--R=(cla)< 0.007
/% Air (760 tort)

1,4 /,.,'-TIC coating

1.0

.6

.2, I I I I I I 440C steel race Martensite_ / /

Z. Ra(cla ) :.-0.03
0 20 40 60 80 100 120 M23 C 6 --J'

Contact load, kg (b) 10 l.Lm

I I I I I I I
0 200 400 600 800 1000 1200 (a) Uncoated ball.

Contac! load, N (b) TiC-coated ball.

(a) Superrefined mineral oil. Figure l l.--Representation of lubricated ball-race contact with and without
(b) SAE 90 EP gear oil. titanium carbide coating on ball. Boring et aI. (1987).

Figure 10--Cross-cylinder load-carrying tests. Material, AISI 52100; diameler,

(TIC), using a chemical vapor deposition process. Boving et
0.64 cm: cylinder speed, 95 rpm; experiment duration, 1 hr: ambient
temperature, 20 °C (293 K). Reichenbach el al. (1964). al. (1981) tested TiC-coated AISI 440C stainless steel, oil-
and-grease-lubricated, 6-mm-bore ABEC 5 ball bearings under
a radial load of 1 N and an axial load of 10 N at a speed of
a conventional liquid-lubricated bearing. The surface
24 000 rpm. Comparisons were made with uncoated bearings
roughness is an arbitrary scale. The steel microstructure
operated under the same conditions. The average pretest
corresponds to the indicated magnification and is to scale. The
vibration level of the coated bearings was between 0.20 and
race surface (R,,(cla)_< 0.01 #m) is approximately three 0.25 g, which corresponded to the pretest vibration level of
times rougher than the ball surface (R,,(cla)= 0.03 #m).
the uncoated bearings. At 24 000 rpm and after several
While the bearing operates in the mixed or boundary
thousand revolutions of operation, the TiC-coated bearings
lubrication regime, or even when the lubricant film mostly
operated more smoothly than the uncoated bearings. The
prevents the balls and the races from coming in contact,
deceleration time of the TiC-coated bearings at different
collisions of the surface peaks cannot be avoided. Such
intervals during the experiment, from 100 to 25 000 hr,
collisions lead to microwelds, which instantly rupture and
remained between 100 and 120 sec. After 3000 br of operation
roughen the surface by material transfer. The resulting heat wear tracks could be observed on the races and balls of the
leads to lubricant breakdown. The magnitude of this uncoated bearings. The lubricant blackened because of wear
breakdown depends on the component surface roughness, the particles caused by adhesive wear and their reaction to the oil
applied load, and the lubricant properties (Boving et al., 1987). and the grease. However, when the TiC-coated races were
In order to overcome this potential problem, Boving et al. examined after 25 000 hr, no traces of wear were observed.
(1981, 1987) coated both balls and races with titanium carbide The lubricant remained clear.

I0
 Boving et al. (1987) tested both TiC-coated and uncoated Molecular

AFBMA grade 3 A1SI 440C balls in uncoated races (fig. 1 l(b)). seal _ : [

The race surface was approximately five times rougher than Satellite /-- Locking
the ball surface. Gyro spin axis ball bearings with the uncoated device [or
_nterface _. , , launch
steel balls running uninterrupted at 30 000 rpm for 20 000 hr
underwent a steady lubricant breakdown with increased noise
Nylon oil
and vibration levels. McKee (1987) tested TiC-coated balls reservoir
under identical conditions; there was practically no lubricant on shaft

breakdown and the noise and vibration levels remained low lip

during the entire operating period.

Lammer et al. (1980) tested TiC-coated balls lubricated with
Fomblin Z25 oil (perfluorinated polyether), again with no
lubricant deterioration or surface reaction. These tests were
run at low rotationless angle movements under high loads and
motionless modes for extended times. Therefore, we can o,, !!i
reservoir
conclude that TiC coatings may be able to extend the useful
life of liquid-lubricated bearings in a space environment. "'-. oil-impregnated
Pretoad
in housing -, , _1_ _ _-- phenolic
bAanl? _-.cage .
_3'ear_COglwa_t

Bearing Torque Instability

diaphragm-__, _.. _ ,
Cage Instability and Noise
Power slip

One type of failure observed in space despin mechanical

assemblies (PMA) is ball bearing cage (separator) instability.
This instability has caused abnormal torque variations and
serious pointing errors in the antenna. An associated proNem
is the maintenance of low electrical noise in the interface
Figure 12.--Schematic of medium-s_'ed (despin) mec 'hanism. Hostenkamp (1980).
betwecn the stationary and rotating elements, such as the slip
ring assemblies shown in figure 12. Typical DMA mission friction in a bearing, which is observed as erratic torque. The
requirements have been 1 to 5 years. For missions of these
torque fluctuation measured in figure 9 is an example of a
durations it has been possible to conduct real-time life tests
bearing operating in an unstable mode (Kannel and
to generate the desired level of confidence in achieving the
Dufrane, 1986).
mission goal and to define the operating characteristics of the The key to cage stability lies in the tangential stiffness of
system. A number of laboratory life tcsts of these systems have the ball-race interfaces and the rate at which energy can be
been conducted (Parker, 1987; Ward, 1984; Feuerstein and
absorbed at these interfaces. When the cage strikes a ball, the
Forster 1975). In typical r)MA systems the angular-contact ball
ball will slip and generate a reactive force against the cage.
bearings are AREC7 specification and have bore sizes ranging
If the slip occurs easily, such as when the balls and races are
from 90 to 150 ram. They generally operate at speeds from
completely separated by a film of low-viscosity fluid, the
30 to 60 rpm and are oil lubricated. The current trend is for reaction force is small. Under such conditions the cage energy
longer missions of 7 to 10 years. Tests have been run for 7 is absorbed by shear losses in the lubricant film. Conversely,
to 10 years with little or no bearing wear and no torque or if the balls and races are separated by a highly viscous fluid,
torque noisc problems. Generally, power system failures occur the ball slippage will be small, the shear losses will be much
before any lubrication system problems become apparent less, and the cage motion will be undamped. Depending on
(Fleischauer and Hilton, 1990).
the parameters of the particular bearing, cage instability can
The ability to control gross bearing torque and bearing be a serious problem (Kannel and Dufrane, 1986).
torque variations (or noise) is essential to achieving long life
and reliability. Gross bearing torque can be controlled by
Cage Stability Criterion
properly selecting the lubricant and the lubricant quantity in
the bearing. Bearing torque variations are much more difficult Kannel et al. (1976) developed a cage stability criterion
to control. One major source of torque variations is erratic given by the following equation:
motions of the cage separating the balls. Under idea[ conditions
the cage simply rotates with the balls. However, under some
conditions the cage incurs rapid secondary motion, which is
known as cage instability. Cage instability generates erratic
 where was ball guided. The bearings were completely stable with
the lower viscosity SAE 10 but were continuously unstable
DP damping
parameter,
dimensionless with the other conventional, higher viscosity spacecraft
C_, ball-race traction parameter, N-sec/m (lb-sec/in.) lubricants. Changing the bearing nominal contact angle from
M cage mass, N sec2/m (lb-sec2/in.) 23 ° to 28 ° did not affect cage stability. Other tests run by
Kannet el el. (1976) with race-guided cages did not result
C,1 ball-cage spring rate (linearized), N/m (lb/in.)
in any instability. Varying the cage-race clearances up to
0.018 mm (0.007 in.) did not produce bearing cage instability.
However, cage instability can occur in bearings with race-
guided cages. Changing the bearing contact angle as befi)re
where
did not affect cage stability. Kannel et al. (1976) also varied
the quantity of lubricant in the bearing. They showed that cage
p,,, average contact zone viscosity (- 106 cP [or a
stability can be achieved by increasing the lubricant supply,
mineral oil). cP
but overall bearing torque is increased. These tests suggest
A contact area, Ill-" (in. 2) that bearing torque instability may be controlled by using race-
h film thickness, m (in.) guided cages, by decreasing lubricant viscosity, or by
increasing the lubricant supply to the hearing.
As h gets small or #,, gets large, C, gets large.

The larger the value of DP, the more likely that the cage Blocking
will be unstable. An approximate criterion tor cage stability
is given in figure 13. To check stability, values of the ball- Loewenthal (1988) describes another tbrm of torque
cage friction coefficient)"and a cage restitution factor e, must variation experienced in oil-lubricated, oscillating gimbal
be known, where bearings such as those used in the Hubble Space Telescope
high-gain antenna (HGA) drive. The bearings were a 66.7-N
(15-1b) preloaded pair of A541 (27-ram bore) size, duplex
angular-contact ball bearings mounted face to Pace. Each
bearing contained 24 balls with a diameter of 3.18 mm
An accurate assessment of cage stability requires analyzing (0.125 in.), an inner and outer ball-race conformity of
the cage motions with comprehensive computer models. 51.8 percent, and a one-piece, inner-land-guided, phenolic-
Cage instabilities reported lot DMA bearings have typically laminated cage. The bearing was lubricated by KG 80 oil.
occurred in bearings lubricated with a Vackote lubricant During testing under repeated cycling the gimbal drag torque
(basically Apiezon C). At room temperature this lubricant increased from a nominal 14 N-mm (2 in.-oz) to as high as
pr_xtuces a damping parameter DP greater than unity, implying 127 N-ram (18 in.-oz). This drag approached the stall torque
an instability. The damping parameter can be greatly reduced of the drive motor. The torque trace for these bearings is shown
by using a lower viscosity lubricant, which hypothetically in figure 14. The bearing oscillated _+96 ° at a cycling rate
should produce a stable cage (Kennel et el., 1976). in order of 0.5 deg/sec. The highest torques only occurred at the end-
to ewduate this hypothesis, Kannel et al. (1976) conducted of-travel reverse point shown at the -96 ° location in
bearing tests with an SAE 10 mineral oil (50 cP), Vackote figure 14(b).
(192 cP), and Apiezon C (176 cP). The cage for these tests Loewenthal (1988) reports that this torque anomaly was
observed with six UGA gimbals. The gimbal design was
essentially identical to an earlier gimbal that exhibited no such
1.0 torque anomaly. Todd (1981) had reported the same
phenomenon (termed "blocking") in hard-preloaded pairs of
.8 -- ball bearings oscillated over a 90 ° arc. Loewenthal (1988)
,6 stat}le discusses the various causes of blocking. Among these are a
.4 _e url
ball speed variation caused by bearing misalignment: as the
Cage stable balls advance or retard from the average speed they squeeze
_ 2 the cage's ball pockets. In oscillatory bearings the distance
0
errors between the balls and the resulting cage loading or
"windup" increase with rotation, reaching a maximum at the
0 .2 .4 .6 ,8 10
end of travel and then decreasing as rotation is reversed. A
Restitution factor, ec
common approach to reducing this problem is to use alternating
Figure 13. Cage rc_,titution [a,.'tor Ior determining cage stability. Kennel and toroid ball separators so that the balls can nlore freely adjust
Dufrane (1986).
their spacing. The toroidal material can be Teflon or polymfide

12
 increasing the percentage of conformity reduces both spin
20x10 -3 torque and drag torque. At the 54-percent confomaity the
7 Clockwise
i predicted spin torque was 45 percent less and the predicted
drag torque was 39 percent less than that with the baseline
bearing having a 51.8-percent conformity.
oI Results of tests run with the baseline bearing and the

MO
54-percent-conformity bearing with and without toroidal
_ " separators are shown in figure 15. The measured torque
Z -20
reduction was greater than had been predicted. From these
30x 1 -3 data the blocking phenomenon appears to be the primary result
o
of spin-generated, transverse ball creep, which increases the
balls" tendency to climb the bearing raceway shoulder, rather
-u 20 '_- Counterclockwise
than of ball speed variation (Loewenthal, 1988). But according
to M.J. Todd of the European Space Tribology Laboratory,
0 Risley, England (in a letter to this author dated Jan. 15, 1990),
if all balls crept equally, there would be no blocking. This
-1(?
would then suggest that both misaligmnent and transverse
-20 --/" Clockwise creep must be present for blocking to occur.
-30 / I I 1 (b>i
-1 O0 -50 0 50 1 O0 Raceway Deposits
Rotation, deg
Phinney et al. (1988) also studied the blocking phenomenon
(a) Torque characteristics at startup.
in oscillating duplex gimbal bearing applications. In addition,
(b) Torque characlcrb4ics at failure.
they investigated torque spikes, another common occurrence.
Figure 14. Startup and failure torque charaderistics for preloaded pair of
They concluded that torque spikes were caused by debris piling
27 ran>bore, duplex angular-conlact ball bearings under oscillatory motiolL
up just before the ends of the ball paths, under repeated
C)cling rate, 0.5 deg/sec; bearing nominal preload, 66.'7 N (t5 lb); lubricant,
KG 80. Loe_'enthal (1988).
cycling, at fixed angles too short to overlap the ball tracks.
Debris piles up behind the balls by compaction before a low
wave builds up and carries any particles off to the side and
impregnated with oil. Loewenthal (1988) reports that although
the toroidal separators reduced the +'runaway" torque at the back into the running track behind the ball. The debris could
end of travel, drag torque continued to increase over time by only build up when the ball tracks did not overlap. Once the
ball tracks overlapped, the debris was redistributed such that
60 percent from startup values.
it could no longer build up when travel was shortened. Phinney
Todd (1981) suggests that not only misalignment (hence,
et al. (1988) attribute the source of the debris to particles
ball speed variation) but also transverse creep of the spinning
ball are necessary ingredients of blocking. "Tight" (51.8 generated during assembly by press fits of the hardware and
the tools used for the installation. They report that bearings
percent) ball-race conformity readily produced blocking
are run in a 360 ° rotation in the gimbal to distribute any
because of the sensitivity of ball speed to changes in contact
assembly debris. This is in addition to the run-in they receive
angle. However, under "loose" (57 percent) ball-race
conlbrmity, where the bah speed was less sensitive to contact before assembly (Phinney et al.. 1988).
There can also be other causes of torque spike. An example
angle, blocking was never observed. A disadvantage of using
is the formation of solid friction polymer from the lubricant,
the loose conformity is that at satellite launching the resultant
Hertz (contact) stress on the bearing raceway can exceed which becomes deposited in the running track and the cage
acceptable limits. However, Loewcnthal (1988) determined pockets (Hunter et al., 1987). Meeks et al. (1971) discuss the
that the confi)rmity for the Hubble Space Telescope riGA oil polymerization of F-50-1ubricated bearings. They suggest
an inverse correlation between the theoretical oil-fihn thickness
bearings could be increased from a baseline of 51.8 to 54
and the amount of polymer observed. Since the F-50 oil
percent without exceeding allowable Hertz stress limits.
Lowenthal (1988) performed a transverse creep analysis, polymerizes (or crosslinks) at approximately 589 K (601 ° F),
considering the elastohydrodynamic film in the contact to they speculate that this temperature is reached at the
behave elastically at low strain rates until the shear stress microasperity contact points with very low-viscosity oil
reached some limiting value. At this limiting value the film (fig. l l(a)). A polymer deposit will act like solid debris.
was considered to shear like a plastic material. By integrating Additional debris is generated from the ball pockets and the
the local traction forces across the contact, he computed the land-guided faces of the cage. The accumulation of this debris
net traction forces in the direction of rolling as well as in the will cause torque fluctuation and can eventually jam the bearing
transverse direction. Loewenthal (1988) concluded that (Harris, 1969).

13
 30X10 -3

ConFormity,
percent

O St
25 [] ._
"_ Baseline
Open symbols denote conformity
with phenolic cage
Solid symbols denote conformity
with PTFE toroid separators
2O

E
.o_
o= lS
I

Gimbal Inverted -/

10

I I I I I I I I I I
0 5430 1000 1500 2000 2500 3000 3500 4000 4500 5000

Number of cycles

Figure" 15, Bearing torque for prcloaded pair t)l" 27-ram-bore, duplex angular-comacl hearings under oscillator_, m,fion v, ilh cmnbinati_ms of ball-race

confi)rmity and separator type. Speed, (1.5 deg/sec: oscillatory motion, 960: bearing nominal preload, 66.7 N ( 15 lb): lubricanl, KG 80, I,oewenthal (1988).

Lubrication Methodology and Supply

'_ --0.0005 /
"_1 7 0.0740 [ +0.0000_ in. diam
Controlled Leakage Chopper I_,=- [ +0.0005 '_
min'or --, "_l II 0.0745 _---0.0000 ,t in. diam

Ahlborn et al. (1975) discuss the design criteria for space I '3 ....... -,_ ' I
liquid lubrication systems. They state that, whenever possible,
sealed systems should be avoided. Seals, whether static or
rotating, offer lower reliability than an open system design
because the seals may fail. However, sealed systems are used
when considerations such as preventing the contamination of _ reservoirs, vacuum
_i "-_ Sintered-nylon
optics or planetary surfaces are important.
If a lubricant film that provides reasonably constant friction
- impregnated with
force is maintained on moving contacting surfaces, a space RotaTing shaft: .-'" diester-base oil
2750 rpm _"
mechanism can last for an indefinite time. Therefore, the
Figure 16.--Radi_meter spindle assembly li)r TIROS [l Inete_)rological
lubricant evaporated from the contacting surfaces into the space
satellite. Weinreb (1961L
environment must be replenished. Ahlborn et al. (1975) recom-
mend that the lubrication system be designed to lose only 10 on the fact that, on a molecular scale, even smooth surfaces
percent of the internal lubricant supply to space during the life appear rough, and according to Knudsen (1950) the direction
of the mission. This goal can be achieved by controlling leakage. in which a molecule rebounds after a collision with a wall
The premier work related to controlled leakage was reported is statistically independent of the angle of incidence. For this
by Weinreb (1961) lor the bearings in a mechanism for the reason the molecular flow resistance of small orifices can be
TIROS, meteorological satellite. The mechanism, shown in made relatively high. The vapor pressure inside the chamber
figure 16, was a five-channel infrared radiometer that consisted can be maintained and vaporization of the lubricant can
of five optical mirrors mounted on five gears and eight ball be minimized.
bearings driven by a low-power motor whose output torque The mechanism shown in figure 16 was designed with
was 2.12 × 10 -4 N-m (0.03 in.-oz). The design was based lubricant reservoirs of oil-impregnated sintered nylon. The

14
lubricant was a MIL-L-6085A diester oil with a vapor
pressure of approximately 10 .4 torr. When the outside t
i 5_
a
pressure was below 10-2 torr, a molecular flow occurred
around the shaft through the small clearance. The clearance I ' T
was a nominal 0.0127 mm (0.0005 in.). By using the :quation (a)
derived by Knudsen (1950), it was possible to calculate the
rate at which oil escaped from the bearing assembly. The
amount of oil was determined in the reservoirs from the
required life of the satellite and the escape rate of the lubricant.
A problem with this type of lubricating system is the potential -J-- I
of condensing oil vapor on the optical mirrors. a

Silversher (1970) reports a simplified equation, based on

a modification of kinematic gas theory, that accounts for the
(b)
escaping oil vapor loss through an aperture:
M I/2 "a"

where
w = 0.0583 7rP(_)(];)A

g
T a
I
-f-i
w weight loss through aperture, g/see (c)
P vapor pressure of gas, torr (a) Straight escape padl.
(b) Elbow escape path.
M molecular weight of oil vapor
(c) T-shaped escape path.
T temperature of oil vapor, K
Figure 17.--Measurahlc escape paths. Silversher {1970).

a gap width of aperture, cm

f channel length; or escape path, cm 10-8
0_1 on
beating Oil in nylon
A area of aperture, cm 2 reservoirs

This equation is valid if the ratio f/a is 16 or greater.

Several examples of escape paths are shown in figure 17.
10 -@
Controlled-leakage sealing was augmented by using liquid Oil in
3henolic cage
lubricants with low vapor pressure. Silversher (1970) reports
that it is sometimes difficult to correlate a high-average-
I I 1
molecular-weight product with a uniform low vapor pressure. I0 -I( I
70
100 90 80
Many liquid lubricants contain a wide range of molecular
Oil remaining, wt %
weights. The low-molecular-weight fractions generally
volatilize first, so that the rate of fluid loss will not necessarily Figure 18,--Oil evalx_ration rate variations in vacuum at 40 °C. Ahlborn et al.

be constant. This is illustrated in the data of figure 18 (from (1975)

Ahlborn et al., 1975). The rate of evaporation is greater from

the porous nylon reservoirs than from the bearing surfaces or
approximately 3.3 g (an effective porosity of approximately
the phenolic cage. As a result Ahlborn et al. (1975) assume
25 percent). The reservoirs (two for each bearing) were placed
that only 30 percent of the total oil in the reservoirs is available
close to each bearing, one on the shaft and one on the housing,
to lubricate the bearings or other contacting surfaces.
to ensure that at least one reservoir was always at a higher
The most extensively documented use of evaporation as a
temperature than the bearing. Narrow clearance relative to the
method for space lubrication was with a medium-speed despin
moving counterpart was provided for the aluminum-backed
mechanism manufactured by Dornier System GmbH in 1973
reservoirs to limit oil loss from the immediate bearing vicinity.
(Hostenkamp, 1980). A schematic of the bearing and
Barrier films (Tillan M-2, derived from FX706) were applied
lubrication system is shown in figure 12. The launch loads
to the clearances to aid in this purpose.
bypassed the bearings through an off-load device. The A thermal vacuum real-life test of 7 years was successfully
lubrication system for the bearings was a combination of BP
conducted by the European Space Tribology Laboratory
21 I0 grease with BP I l0 base oil (table IV) thinly applied to
(Parker, 1987). During the 7-year test period the thermal
both balls and races and the BP 1 l0 base oil impregnated in
conditions were changed at 4-week intervals to simulate flight
porous phenolic cages and nylon reserwfirs. There was conditions; the mechanism speed was 60 rpm. Earth eclipses
approximately 0.25 g of oil per cage (an effective porosity
were simulated at 6-month intervals. The testing was success-
of approximately 2.3 percent). Each reservoir contained

15
 fully continued
fi_ran eighth year but at normal ambient
temperature. Initial sarnpte

The vacutnn pressure took several years to stabilize because

g
.14 _ weight,
of continuous outgassing from the mechanism. This outgassing .12

resulted in the pressure substantially increasing during the I [] 4.20380

_.10 i-- 0 4.19905
wanner test cycles. Residual gas analysis by, mass spectrometry
II
II r Vacuummoregna_on 0 4.20595
suggested that a gradual increase m the relative pressure of
.08 /1--/ _ w,tn KG 80 o,I
hydrocarbons was due to small amounts of oil being released
into the test environment from the mechanism (Parker. 1987).
During testing the current demanded by the motor to enable _- 04

the speed control system to maintain a constant speed gradually

.02 _'" _ extrac[lon in Freon TF
increased. This was not, however, attributed to an increase
in friction torque within the mechanism but to aging of the
j j f i J
0 40 80 120 160 200 400
motor rotor's permanent magnets. The thermal distribution
Soaking lime, days
within the mechanism remained substantially constant except
during eclipse simulations. Toward the end of these simulations {a) Vacut, m bake al - 100 °C tbr 24 hr.

(b) Soxhlet exlracti{m in Frcon TF.

a small increase in motor power dissipation was detected.
{c) VacuUlll imprcgnati{m with KG NO oil.
Changes of thermal gradient across the bearings in the various
Figure 19.- Absorption of oil by phenolic rclainers as a function of lilnc.
test modes did not result in any detectable change in mechanism
Fleischauer and Hiltnn (1990L
performance. A one-per-revolution ticking noise was present
during the early part of the life test. This was attributed to
surfaces. Hence, it may be concluded that long-term effective
relaxation of the flexible diaphragm used to preload the lubrication using phenolic cages as lubricant reservoirs can
bearings (Parker, 1987).
best be achieved if the cage is ball guided rather than race
Post-test examination of the dismantled mechanism revealed
guided. However, as previously discussed, ball-guided cages
its components to be in excellent condition. The oil-lubricated
are more likely to be unstable (Kanncl et al., 1976).
bearings were virtually like new without any measurable wear.
From late 1984 into early 1985 the scanners of the U.S.
The oil and grease lubrication was highly satislactory although Earth Radiation Budget Experiment (ErBE) and NOAA-9
some oil was lost from the bearings because of the inadequate
satellites, after several months in orbit, began experiencing
application of barrier films. The oil had marginally contam- problems (Watson et al., 1985). A motor torque increase was
inated the power brush and slip ring components without any
detected on several of the units. This suggested an impediment
adverse effect (Duvall, 1985). In a letter to this author dated
to the tree rotation of the instruments, which were mounted
Nov. 1, 1989, H.M. Briscoe, Consultant in Space Mechanisms
on a ball bearing assembly. The motor drove the instruments
and Tribology, Tykes Barn, Great Britain, suggested that the back and forth at I rps. The scan angle was 176 °. On one
oil film on the slip rings was beneficial because the brush wear
occasion the rotation of an instrument stopped completely. The
was less than expected. He speculated that oil leakage from
torque increase was attributed to the ball bearing assembly.
the bearing chamber may be a reasonable means of lubricating
The operating conditions of the ball bearing assembly were
slip rings.
duplicated in a laboratory bearing endurance test at I0 ; torr.
The bearing was an AISI 440C, 40-ram-bore, angular-contact
Lubricant-Impregnated Cages ball bearing with 34 balls and a nominal contact angle of 20 °
to 25 °. The preload was 89 N (20 lb). The bearings porous
Fleischauer and Hilton (1990) state that the phenolic-based
phenolic cage, which was inner-land guided, was saturated
cage materials commonly used in satellite bearings have
with Brayco 815Z, a perfluoroalkylether. The phenolic cage
irrepmducible, time-dependent oil absorption properties
was the only source of liquid lubricant tbr the bearing. Both
(fig. 19). These data were obtained from samples that had been
the races and the balls were passivated and chemically treated
vacuum impregnated with oil and then allowed to soak in the
with tricresylphosphate (TCP). The test duration was 8.5 million
oil for the times indicated. From the data it appears that the
cycles. The results of these tests revealed ineffective transfer
cage material continuously absorbs oil. This would suggest that
of lubricant from the phenolic cage to the contacting surfaces.
the phenolic material acts more like an oil sink than a source.
Also, the use of Tee to chemically treat the ball and race
Tyler et al. (1976) suggest that it is possible for lubricant
surfaces resulted in the inability of the PFPE to wet the bearing
feed to occur from the phenolic cage to the balls but not from surfaces (Morales et al., 1986).
the phenolic cage to the bearing race lands. Layers of cotton
material in the phenolic are parallel to the bearing race lands.
Lubricant teed, they conclude, cannot occur in a direction Barrier Films
perpendicular to these layers. As a result, the lubrication of
In order to minimize lubricant loss or migration, both from
the interface must depend on the initial oil coating on the two
the bearing and the satellite, creep barrier coating films are

t6
 in figure 20. In this system, a positive commandable
being used on noncontacting bearing surfaces and on the
lubricator, the reservoir is sealed to prevent lubricant contami-
surfaces of molecular seals. However. normal procedure is
nation. The degassed lubricant supply is stored in a flexible
to avoid the application of barrier films within the bearing and
metal bellows. Pressure is maintained by an external spring
apply them only to escape gaps. The U.S. Naval Research
pack. Opening the release valve permits oil to inflate an
Laboratory developed a barrier film for ihstrument bearings.
The film, a fluorinated methacrylate, acts as a low-energy adjustable-stroke metering bellows. Subsequent closing of the
release valve and opening of the metering valve starts the l]ow
barrier across which most lubricants cannot creep. It is stable
in air to 423 K (302 °F) and in vacuum to 373 K (212 °F). to the applicator. Metering pressure is sufficient to overcome
the characteristic backpressure of the applicator and to provide
It outgasses at 373 K (212 °F) at a rate less than 2 × l0 _
the desired flow rate. A metering orifice provides flow rate
g-cm2/sec. Ahlborn et al. (1975) observe that because there
are differences between the various commercial barrier films. adjustment capability. The applicator is supported rigidly, in
the space between the outer race and the inner-race-guided
they should be evaluated tbr a specific application before being
retainer. In a 110-ram-bore bearing with 12.7-mm (0.5-in.)
used. They observe that some lubricant tbrmations appear to
diameter balls the space is approximately 3.81 ,rim (0.15 in.)
"poison" the surlilce of the coatings and render them
wide. A standoff distance, typically five times the expected
ineffective. They state that the mechanical and chemical
launch-induced axial ball movement, separates the applicator
stability of these fihns is improved by high-temperature
tip from the ball path. A combination of toroida[ tip shape and
vacuum baking.
Teflon coating enables the applicator to support an oil droplet
that spans the standoff distance. In the operating bearing the
Positive Feed Systems passing balls wipe off a portion of the distended hemispherical
droplet. In this manner oil is slowly and unilbrmly transferred
Large spacecraft, starting with Skylab (fig. 1), have brought
from the balls to the retainer ball pockets and to both races.
a new set of tribological problems. Smaller spacecraft could
The droplet is continuously replenished by' flow from the
be stabilized by spinning the spacecrali or by using control
metering system during the 2- to 4-rain relubrication cycle.
jets. Large, extended-mission spacecraft require large control
The positive commandable lubricator is designed for high-
moment gyroscopes (CMG's) that are capable of handling the
surface-energy oils such as Apiezon C. The high surface
large slew loads resulting from astronauts moving around the
energy of hydrocarbon oils provides the stabilizing force li)r
craft and from changing the orientation. The large CMG'S can
the lubricant droplet on the applicator tip. The system is best
be heavily loaded and operate at high speeds. Small bearings
suited for large bearings, where there is sufficient space
using lubricant-impregnated cages may no longer suffice fi)r
between the retainer dynamic envelope and the race to accept
these units, and positive lubrication of large bearings may be
the applicator (James, 1977).
required. The lubrication system has to be highly reliable but
NASa research (Loewentbal et al., 1985) with a terrestrial
extremely compact to meet weight and space requirements
experimental 46-cm (19-in.) diameter, 58-kg (128-1b) flywheel
(Kannel and Dufrane, 1986).
showed the feasibility of using a wick lubrication system in
A major tribological challenge is ensuring that the lubricant
a vacuum environment to lubricate moderate-speed bearings.
gets into the bearing to lubricate the ball-race interfaces.
Devices such as centrifugal oilers attached to the rotating shaft The flywheel with its lubrication system is shown in figure 21.
In this system a lightly spring-loaded wick saturated with oil
have been successfully used in some applications. Another
contacts a conical sleeve adjacent to the bearing inner race.
method (Glassow, 1976) uses a single-stroke pump immersed
Frictional contact against the sleeve causes a small amount
in a vented oil reservoir to pump the lubricant to the bearing.
of oil to be deposited. This oil migrates along the sleeve to
James (1977) propose; the entirely different approach shown

Metering bellows'-m
,-- Rotating

Adjustable stop --., I ,/ inner race

m Metering valve

, j inner-race-
Release vane-.. "., _ ,"/- guided

Applicator --,, retainer

Stationary

I
L outer race

_,,,.-_....... H.._'A Delivery tube --/

Standoff "*_''_
'.- Lubricant
reservoir
L_ Spring pack

Figure 20.--Positive commandable lubricator for satellite bearing application. James (1977).

17
 ,- Drive
/
Midspan // spindle
sland //
bearing -_ ,- Release
/
/ coupling
//
//-- Vacuum line / /
/

i oo matic \

/
/
/

I
iI
Carbon lace seat / I
1 piston and I
/ /
.J rotor i L Upper
/- Spring I rotor
!t release // preload I
/ washers -- / I bearing

J 1
I
I

Elaslomer
bearing

/ ")-- Oil
Oil reservoir ,( /
" / wick
\\ 1 feed
bearing

Lower rotor bearing _'

_L Rotor CD 90 46861

Figure 21.--Schematic of NASA flywheel rotor module. Loewcnthal et al. (1985)

its large end and into the bearing under the centrifugal force
field. The wick absorbs oil from the reservoir; the oil travels
Absolute
through the wick by capillary action to replenish the oil housing

entering the bearing. In a terrestrial system, because of gravity, pressure,

millitorr
the oil droplets leaving the bearing and the oil vapor that
40-50
coalesces against the cooler housing walls eventually return 6O
-- I ..C],,. I 85-1 O5
oo
to the reservoir, thus closing the cycle. Whether condensation
/"
and reuse of the oil can be accomplished in a space system =
I_O-2oo ,6.-'--'-_
is an open issue. 50 --

The rotor system was operated at speeds from 10 000 to

20 000 rpm at absolute housing pressures from 195 down to
45 millitorr. The peak rated power of the flywheel, 120 kW
at 20 000 rpm, corresponds to a maximum discharge torque
of 57 N-m (504 in.-lb). The bearing outer-race temperature
profiles are shown in figure 22. These data show that the ® £3- _" O" "_
bearings will operate at lower temperatures when the housing " 30 t31 I I I I
0 4 8 12 16 20x103
pressure is low: the aerodynamic heating of the rotor is less
Rotor speed, rpm
at lower pressure. The data confirm that the wick feed system
Figure 22.--Flywheel lower support bearing temperature characteristics with
provides adequate cooling, particularly at the lower vacuum
speed and housing pressure. Bearing type. 20-ram-bore, deep-groove ball
levels (Loewenthal et al., 1985).
bearing; thrust load, 712 N (160 Ib). Locwenthal et al. 11985).

18
 Boundary lubrication with antiwear and extreme-pressure
Concluding Remarks
additives in the lubricant in a satellite (vacuum) environment
A search of the literature revealed that spacecraft lubrication is not well defined. There is an issue as to whether these
additives are at all effective in the absence of oxygen. The
systems have, for the most part, performed adequately over
role of additives needs to be defined for these applications.
the past three decades. That is, spacecraft and satellites have
reached their required lifetimes without a lubrication-related The long-term effects of atomic oxygen and other space
failure. However, this success was achieved because the life radiation on lubricant behavior and stability remain unknown.
The effects can be important for space missions longer than
of satellite systems is generally limited by their batteries,
electronics, and thermal and optical systems. However, as 10 years.
these technologies are improved, spacecraft systems will Alternative cage materials with lower wear rates need to
be considered as replacements for phenolic-base materials,
probably become limited by the lubrication system.
Problems related to spin-stabilized satellites involve the especially in oscillating bearings, in order to inhibit the buildup
of wear debris and the potential jamming of the bearing.
despin mechanical assemblies and the slip ring assemblies,
Alternative lubrication systems providing for positive lubri-
which are required to operate with low torque ripple and low
electrical noise. Tests have verified current designs, and in cant feed to the contacting surfaces may need to be considered
for future space missions. In addition, alternative lubricants
many satellites power system failures have occurred before
lubrication failures have surfaced. However, the ability to need to be seriously examined as replacements for the

control gross bearing torque and noise is essential to achieving perfluorinated polyalkylethers (PFP_:) for long-term space
missions.
high reliability and long life.
In the fourth decade of the space age tribologists and
Spin bearings for control moment gyros, operating at speeds
aerospace design engineers face continuing challenges to
up to 12 000 rpm, have been lubricated successfully for current
achieve longer lifetimes with liquid lubrication in space.
design lives with oil-impregnated phenolic cages. Future
systems will require much longer lives as well as higher load
capabilities.
Gimbal bearings, which operate in an oscillatory mode and
rarely make a full revolution, are a major concern because
Lewis Research Center
of the blocking phenomenon and debris buildup discussed
National Aeronautics and Space Administration
herein. Careful consideration must be given to the design and
Cleveland, Ohio 44135, February 16, 1990
use of these bearings in order to achieve long lifetimes.

19
 References

Ahlborn, G.H.: Hinricks. J.T.; and Perrin, B.J. (1975) Long Life Lubrication Flora, D.G., and Haltner, A.J. (1968) Lubricants for the Space Environment,
S)stems Paper presented at European Space Tribology Symposium, AIChE Paper 18D. Paper presented at AIChE Symposium on Materials lor
Frascati. Ital,,. Apr. 9-11. IAvail. NTIS. 75N22762.) Reentry and Spacecraft Systems--Spacecraft Materials. Part II--Materials
Conference. Philadelphia, PA.
Baxter, B.H.: and Hall. B.P. (1985) The Use of Perlluoroether Lubricants

in Unprotected Space Environments. Nineteenth Aerospace Mechanisms Freundlich, M.M.; and Hannan, C.H. (1%1) Problems of Lubrication in Space,
Symposium, NASA CP-2371, pp. 179 207. Lubrication Engineering, vol. 17, no. 2, pp. 72-77.

Beerbov, er, A. (19721 Boundary Lubrication. Scientific and Technical

Glasgow, F.A. (1976) Assurance of Lubricant Supply in Wet-Lubricated Space
Application b}_recasts, Dept. of Army. DAHC-19 69-C. Bearing. Proc. lOth Aeros!_we Mechani._tL_ Symposium, NASA CR-148515,
pp. 90 103.
Bisson, E.E (1964) Friction and Bearing Problems in the Vacuum and
Radiation Environrnent of Space. Advanced Bearing Technology, NASA Grubin, A.N. (1949) Fundamentals of the Aertxlynamic Theory, of Lubrication
SP-38, E.E. Bisson and W.J. Anderson. eds., pp. 259-287. of Heavily Loaded Cylindrical Surfaces, Chapter 2, Investigation of the Contact
of Machine Counts, Translation of Russian Book No. 30, Central Scientific
Boring. H.J.; Hintermann, HE.; and Steh[e, G., (1981) TiC-Coated Ball
Institute for Technology and Mechanical Engineering, Mo._ow, ed., Ketova,
Bearings lot Spin-Axis Gyro Application. Lubrication Engineering. vol. 37,
Kh.F. (Available from Department of Scientific and Industrial Research, Great
no. 9, pp. 534-537.
Britain, Trans., CTS-235 and Special Libraries Asstx:iation, Trans. R 3554.)
Boring, H.; Hintermann, H,E.: and Hanni, W., (1987) Ball Bearings with
Hamrock, B.J.; and Dowson, D. (1977) Isothermal Elastohydrodynamic
C VD-TiC Coated Components. Proceedings _(the 3rd European Space
Lubrication of Point Contacts. Part Ill--Fully Flooded Results, Journal of
Mcclla_u._m,_ and Triholog 3 Symposium, ESA-SP 279. J. Hunt, ed..
Lubrication Technology, Vol. 99, No. 2, pp. 264 276.
European Space Agency, Paris, France, pp. 155-159.

Hamrock, B.J.; and Dowson, D. (1981) Ball Bearing Lz_bricathm: The

Buckle',, I) H [1971)Friction. Wear. and Lubrication in Vacuum. NASA
Elastohydrodynami(:_ of Elliptical Cantacts, Wiley, NY.
SP 277. pp. t54 158. Buckle v, D,H. (1985)

Harris, C.L. (1%9) Vacuum Evaluation of Lubricants and Techniques for Space-
Carre. D.J ( 1987} The Perlbrmance of Perfluoropo[yalkylether Oils Under
Ext'_)sed Components, ASME Paper 69-LUB 30.
Bt_undary Lubrication Conditions, STLE Paper 87--TC ID-I.

Hostenkamp, R.G. (1980) Design Features of the Dornier Medium Speed

Cheng_ H.S. (1972) Isothermal Elaslohydrodynamic Theory for the Full Range
Mechanism. Proceedings of the Secorul Spate Tribology Workshop, ESA
of Pressure Viscosity Coefficients. Journal of Lubrication Technology,
w_l. 94. no. I, pp. 35 43. SP-158, G. Levy, ed., European Space Agency, Paris, France, pp. 29-33.

Hunter, S.D.; Kingsbury, E.; and Prahl, J.M. (1987) Oil Film Thickness
Chiu, Y.P. (1974) An Analysis and Prediction of Lubricant Film Starvation
Measurement and Analysis for an Angular Contact Ball Bearing Operating in
in Rolling Contact Systems, ASLE Transactions, vol. 17, no. 1. pp. 22-35.
Parched Elastohydredynamic Lubrication. Svng_._4um on Thin FluM Fibrt_, M.J.
Chmss, F.J. (1961) l,ubricants and Self-Lubricating Materials lot Spacecraft Brown, S. Carmi, and RL. Mullen, eds., ASME, New York, pp. 43-46.
Mechanisms. Lockheed Aircraft Corp., LMSD-894812. (Avail. NTIS,
AI)-266954.) James. G.E. (1977) Positive Commandable Oiler for Satellite Bearing
Lubrication. NASA TM-79356, pp. 87-95.
Corridan, RE. (1959) Summary Re_rt: Bearing and Lubricant Study.
Jastrow, R. (1960) Lunar and Terrestrial Atmospheres. Advances in the
Advanced Technical Laboratory Division, American Standard Corp.
Astronautical Sciences, Vol. 5, Plenum Press, New York, pp. 338-345.
Coy. J.J.: and Zaretsky, E V. (1981) Some Limitations in Applying Classical
Jones, W.R.. Jr. (t982) Boundary Lubrication--Revisited, NASA TM-82858.
EHD Film Thickness Formulas to a High-Speed Bearing, Journal of
Lubrication Technology, vol. 103, no. 2, pp. 295-304. Jones, W.R., Jr., et al. (1983) Thermal Oxidative Degradation Reactions of
Linear Perfluoroalkyl Ethers, Ind. Eng. Chem. PrtM. Res. Dev., vol. 22, no.
Dolan, F.J. ; and McMurtrey, E.L. (1985) Long Term Evaluation of Grease
2, pp. 166-170.
Type Lubricants in Selected Environments, Man's Permanent Presence in
Space, AIAA, New York. Jones, W.R, Jr., et al. (1985) The Effects of Metals and lnhibitors on Thermal
Oxidative Degradation Reactions of Unbranched Perfluoroaklyl Ethers, Ind.
Dowson. D.; and Higginson, GR. (1966) Elasto-H_'dro_/_'namic Lubrication,
Pergamon Press. London. Eng. Chem. Prtxt. Res. Dev., Vol. 24, No. 3, pp. 417-420.

Kannel, J.W.; Bupara, S.S.; and Pentlicki, C.J. (1976) Investigation of Cage
Duvalt, J.A. (1985) Report on the Dismantling and Examinatic.n of Dormier
',rod Bearing Instability in Despun Antenna _Bearings Due to Changes in Lubricant
System GmbH Medium Speed Mechanism on Completion of a Seven Year
Real Life Test. ESA-ESTL-65. Properties, AFML TR-75 38-PT-2. (Avail. NTIS, AD-A031096.)

Kannel, J.W.; and Dufrane, K.F. (1986) Rolling Element Bearings in Space.
Feuerstein. S.; and Forster, A.S (1975) A Concept of Despin Mechanical
Assembly Life Testing Involving Total CharactefiTation. SAMSO-TR-75-203. Twentieth Aeroa_we Mechanisms S3_posium, NASA CP 2423, pp. 121-132.

Knudsen, M. (1950) Kinetic Theor)' r_[" Gases; Some Modern A.wects ,

Fleischauer, P.D.: and Hilton, M.R. (1990) Assessment of the Tribological
3rd ed. Wiley, NY.
Requirements of Advanced Spacecraft Mechanisms. Assessnwnt of the State
_¢ the Art atut Triboh_gy of Advanced Spacecraft Mectumisms, WRDC- Lammer, J.; Jochimisen, G.: and Hintermann, H.E. (1980) TiC-Coatings
TM-qO 138, 1990.
for Bearings in Spacelab-lnstrument-Pointing System, Paper presented at
ASLE-ASME International Lubrication Conference, San Francisco,
Aug. 18-21, 1980.

20
 Silversher, H.I. (1970) Controlled-Leakage Sealing of Bearings for Fluid
Loewenthal,
SH.(1988) TwoGimbal BearingCase Studies;SomeLessons Lubrication in a Space Vacuum Environment, Aerospace Mechani._ms,
Learned.
Twenty-Second Aero.wace Mechanisms Symposium, NASA
Part B. Bearings and Suspensions, vol. 1, G.G. Herzl, ed., J.W. Stacey
CP-2506, pp. 253-269.
Inc., San Francisco, CA, pp. 27-33.
Loewenthal, S.H.:, Scibbe, H.W.; Parker, J.J.; and Zaretsky, E.V. (1985)
Stevens, K.T. (1983) The Performance of Fomblin Z25 Oil in Ball Bearings
Operating Characteristics of a 0.87 kW-hr Flywheel Energy Storage Module.
Operating in Vacuum, ESA-ESTL-56. (Avail. NTIS, 85N17150.]
Proc. 20th bztersociety Energy Conversion Engineering Conference

(IECEC), pp. 164-174. Thomas, A. (1980) Boundary Lubrication of Steel by Various Oils of Low
Vapor Pressure. Proceedings of the Second Space Tribology Workshop,
McKee, F.B. (1987) Gyro Spin Axis Bearing Performance Using Titanium
ESA SP-158. G. Levy, ed., European Space Agency, Paris, France, pp. 6366.
Carbide Coated Balls, Paper presented at 1983 Ball Bearing Symposium,
Orlando, FL, DOD/lnstrument Bearing Working Group, Mar. 9-12, 1987. Todd, M.J. (1981) Investigation of Torque Anomaly in Oscillating PDM
Bearings, ESA-ESTL-49. (Avail. NTIS, 83N26091.)
Mceks, C.R., Christy, R.I., and Cunningham, A.C. (1971) Accelerated
Vacuum Testing of Long Life Ball Bearings and Sliprings. Fifth Aerospace Todd, M .J. ; and Stevens, K.T. (1978) Frictional Torque of Angular Contact
Mechanisms Symposium, NASA SP-282, pp. 127-135. Ball Bearings With Different Conformities. ESA-CR(P)- 122 I, European
Space Agency, Paris, France.
Morales, W.: Jones, W.R., Jr.; and Buckley, D.H. (1986) Analysis of a
Spacecraft Instrument Ball Bearing Assembly Lubricated by a Perfluoro- Tyler, J.C.; Brown, R.D.: Carper, H.J.; and Ku, P.M. (1976) Analysis of
alkylether. NASA TM-87260. Film Thickness Effect in Slow-Speed Lightly-l-x_aded Elastohydrodynamic
Contacts. Part ll--Measurement of Film Thicknesses in Vacuum, AFML
Mori, S.; and Morales, W. (1989a) Degradation and Crosslinking of
TR-74-189-PT-2. (Avail. NTIS, AD-A021512.)
Perfiuoroalkyl Polyethers Under X-Ray Irradiation in Ultrahigh Vacuum.
NASA TP-2910. Uber, J.M. (1985) Analysis of DRIRU Bearings and Lubricant From Solar
Max Repair Mission, SMRM Degradation Study Workshop, NASA
Mori, S.; and Morales, W. (1989b) Tribological Reactions of Perfluoroalkyl
TM-89274, pp. 147-169.
Polyether Oils With Stainless Steel Under Ultrahigh Vacuum Conditions
at Room Temperature, Wear, vol. 132, 1989, pp. 111-121. Ward, W., (1984) Evaluation of a 90-mm-Bore Bearing Operating in a
Simulated Space Environment, Lubrication Engr., vol. 40, no. 5,
Parker, K. (19871 Seven Year Thermal Vacuum Real Life Test of a Medium
pp. 290-298.
Speed (Despin) Mechanism Manufactured by Dornier System, ESA-
ESTL-64. (Avail. NTIS, 88N17038.) Watson, ND., et al. (1985) Earth Radiation Budget Experiment (ERBE)
Scanner Instrument Anomaly Investigation, NASA TM-87636.
Phinney, D.D.; Pollard, C.L.; and Henricks, J.T. (1988) Experience With
Duplex Bearings in Narrow Angle Oscillating Applications. Twenty-Second Weinreb, M.D., (1961) Results of T1ROS II Ball Bearing Operation in Space,
Aero.space Mechanisms Symposium, NASA CP-2506, pp. 211-226. NASA TM X-51354.

Reichenbach, G.S. : Shaw, R., Jr.; and Foster, R.G. (1964) Lubricant Behavior Young, W.C.; Clauss, F.J.; and Drake, S.P. (1963) Lubrication of Ball
in High Vacuum, ASLE Transactions, vol. 7, no. I, pp. 82-90. Bearings for Space Applications, ASLE Transactions, vol. 6, no. 3,
pp. 178-191.
Robbins, E.J. (1975) Tribology Tests for Satellite Application: Simulation
of the Space Environment, Paper presented at European Space Tribology Young, W.C.; and Clauss, F.J. (1966) Lubrication for Spacecraft Application,
Symposium, Frascati, Apr. 9-1 I. (Avail. NTIS, 75N22764.) Lubrication Engineering, vol. 22, no. 6, pp. 219-227.

Roberts, E.W.; Rowntree, R.A.: and Anderson, J.C. (1990) A Review of Zehe, M.J.; and Faut, 0.D. (1989) Acidic Attack of Perfluorinated Alkyl
European Trends in Space Tribology and Spacecraft Mechanism Tech- Ether Lubricant Molecules by Metal Oxide Surfaces. NASA TM-101962.
nology. Assessment of the State of the Art and Tribology of Advanced
Spacecraft Mechanisms, WRDC-TM-90-138, 1990.

21
 Nahonal Aeronautics and
Space Adm_n_strahon
Report Documentation Page
I. Report No.
2. Government Accession No.
3. Recipient's Catalog No.
NASA RP- 1240

4. Title and Subtitle

5. Report Date
Liquid Lubrication in Space
July 1990

6. Performing Organization Code

7. Author(s)
8. Performing Organization Report No
Erwin V. Zaretsky
E-5094

10. Work Unit No.

505-63-1B
9. Performing Organization Name and Address

1. Contract or Grant No.

National Aeronautics and Space Administration
Lewis Research Center
Cleveland, Ohio 44135-3191
13. Type of Report and Period Covered
12. Sponsoring Agency Name and Address
Reference Publication

National Aeronautics and Space Administration

Washington, D.C. 20546-0001 14, Sponsoring Agency Code

15. Supplementary Notes

16. Abstract

The requirement for long-term, reliable operation of aerospace mechanisms has, with a few exceptions, pushed
the state of the art in tribology. Space mission life requirements in the early 1960's were generally 6 months to a
year. The proposed U.S. space station scheduled to be launched in the 1990's must be continuously usable for 10
to 20 years. Liquid lubrication systems are generally used for mission life requirements longer than a year.
Although most spacecraft or satellites have reached their required lifetimes without a lubrication-related failure,
the application of liquid lubricants in the space environment presents unique challenges. This report reviews the
state of the art of liquid lubrication in space as well as the problems and their solutions.

17. Key Words (Suggested by Author(s)) 18, Distribution Statement

Tribology; Lubricants; Space lubrication; Bearings; Unclassified - Unlimited
Liquid lubrication
Subject Category 37

19 Security Classif. (of this report)

20. Security Classif. (of this page) 21. No. of pages 22. Price*
Unclassified Unclassified 23 A03

NASAFORM1626OCT86 *For sale by the National Technical Information Service, Springfield, Virginia 22161

NASA-_y, 1990
i