SP-8121 Rotating Shaft Seals

https://ntrs.nasa.gov/search.jsp?
R=19780022641 2019-11-15T09:11:16+00:00Z
NASA NASA SP-8121

SPACE VEHICLE
DESIGN CRITERIA
(CHEMICAL PROPULSION)
LIQUID ROCKET ENGINE TURBOPUMP

ROTATING-SHAFT SEALS
FEBRUARY 1978
NAT ION AL AER ONA UTIC S AND SPA CE ADM INIS

TRA TIO N
FOREWORD
NASA exp erience h as indica ted a need for unifonn criteria for the design of sp ace vehicles.
Accordingly , criteri a are being developed in the following areas of t echnology :
Environm ent
Stru ctures
Guid ance and Control
Chemical Propulsion
Individu al componen ts of this work will be issu ed as separate monographs as soon as they
are comple ted. This doc um ent, part o f the series on Chemical Propu lsion , is on e such
monograph. A list o f all monographs issu ed prior to this one can be found in th e final pages
of this docum ent.
These monographs are to be regard ed as guides to de ign and not as ASA requirem ents,
except as m ay be specified in fonn al project sp ecifications. It is exp ec ted , however, th at
th ese docum ents, revi ed as ex perience may indicate to be desirable, eventually will provid e
unifoffi1 design prac tices for NASA space vehicles.
This monograph, " Liquid Rocke t Engin e Turbopump Rotating-Shaft Seals," was prepared
under the direc tion of Howard W. Douglass, Chief, Design Criteria Office, Lewis Research
Center ; proj ect managem ent was by Harold W. Schmidt and M. Murray Bailey. The
mon ograph was writte n by R. E. Burcham of the Rocketdyne Division of Rockwell
in tern atio nal Corpora ti on and was edited by Rus ell B. Keller, Jr. of Lewi . To assure
technical acc uracy of thi d ocum ent, scientist and engin eers throughout the technical
community parti cip ated in interviews, consultation , and critic al review of th e text. In
particul ar, Paul F. Brown , Pratt & Whitney Aircraft Division , United T echnologies
Co rp o rati on ; Paul S. Bu c km ann , Aeroj et Liquid Rocket Comp any ; and Lawrence P. Ludwig
and J ohn Zuk, Le wis Resea rch Ce nter, individuall y and collec tively revie wed th e monograph
in de tail.
Comm ents concernin g th e techni cal content of th e e m o nographs will be welcom ed by the
a tio nal Aeron auti cs and Sp ace Admini tra ti on , Lewi Resea rch Cente r (Design Criteri a
Offi ce), Clevelan d , Ohi o 441 3 5.
F bru ary j 978

Page intentionally left blank
GUIDE TO THE USE OF THIS MONOGRAPH
Th e pu rpo se of thi s mo nog ra ph is to orga ni ze and present , for effective use in design , the
igni fica nt expe ri e n ce a nd kn o wl edge acc umulat ed in development and operational
progra ms to da te. It rev iews and assesses c urrent design practices, and from them establishes
firm guid ance fo r ac hi evin g grea te r co nsiste ncy in design , increased reliability in the end
produ c t , a nd grea te r e fficie ncy in th e d e ign effort. The monograph is organized into two
maj o r sec ti o ns th a t a re preced ed by a brief introduction and complemented by a set of
re fe re nces.
Th e St ate of th e Art , sec tion 2 , revi ews and discusses the total design problem, and
identifies whi ch d es ign ele me nt s are involved in successful design. It describes succinctly the
current technology p ert ainin g to th ese elem ents. When d e tailed information is required , the
bes t ava il a bl e re fe re nces are cited . This sec tion se rves as a survey of the subject that provides
ba ckground ma t e ri al a nd prepares a prop e r technological base for the D esign Criteria and
Recomm e nd ed Prac ti ces .
Th e Design Criteria. shown in it ali cs in sec tion 3 , state clearly and briefly what rule, guide ,
limita ti o n , o r st a nd ard mu st be imposed on ea ch essential design element to assure
successful des ign . The Design Criteria can serve effe ctively as a checklist of rules for the
proj ec t ma nage r to use in guiding a design or in assessing its adequacy.
Th e Reco mm e nd ed Prac ti ces, also in section 3 , state how to satisfy each of the criteria .
Wh e neve r poss ibl e, th e best pro cedure is described ; when this cannot be done concisely,
appro pri a te re fe re nces are provid ed . The Recommended Practices , in conjunction with the
Desigll Criteria. provid e positive guidance to the practicing designer on how to achieve
successful design .
Both sec tion h ave bee n organized into decimally numbered subsections so that the subjects
within simil arl y numb ered sub sections correspond from section to section. The fo rm at for
th e Co nte nt s di spl ay thi s continuity of subject in such a way that a particular aspect of
design ca n be fo llowed through both sections as a discrete subject .
Th e d esign crite ri a monograph is not intended to be a design handbook , a set of

specifi cation s, Or a d e ign manua l. It is a summary and a systematic ordering of the large and
loosel y organized body of existing successful design techniques and practices. Its valu e and
its me rit should b e judged on how e ffectively it makes that material available to and useful
to th e d es ign e r.
iii
L_
Page intentionally left blank
CONTENTS
Page
I. INTRODUCT IO I
2. STATE OF THE ART 3
3. DES IGN CR ITER IA and Reco mm en ded Practices 103
APPE DIX A - Conversion of .S. Customary nits to SI Units 133
APPEND IX B - Glossary 135
REF ERE CES . . . . 145
NASA Space Vehicle Design Criteria Monogra ph s Is ued to Date 151
SUBJECT STAT E OF TH E ART DESIGN CRIT ERI A
SEAL SYSTEM 2. 1 7 3. 1 103
Press ure Environmcn t 2. 1.1 12 3. 1.1 103

Thermal Envi ronme n t 2. 1.2 12 3. 1.2 103
Vac uum Environmen t 2. 1.3 12 3.1 .3 103
Rubbin g Speed 2.1.4 13 3. 1.4 103
Cooling and Lubri cat ion 2. 1.5 13 3.1.5 104
Leakage Drains 2. 1. 6 14 3.1.6 104
Fluid Separati on 2. 1. 7 14 3. 1. 7 104
Fail-Safe Provi ion. 2.1. 8 15 3.1. 8 105
Purge Requ iremcnt 2.1.9 15 3.1.9 105
SEAL ASSEMBLY 2.2 16 3.2 105
Prcssure Capabili ty :2.-. 1 16 3.2.1 105

Temperature Capab ilit y ] .2.2 32 3.2.2 106
Spced Limi la ti ons ].2.3 32 3.2.3 108
Wcar Life 2.2.4 36 3.2.4 II I
Leakagc ? ? .J-
_._ 38 3.2.5 III
Mi sa li gnmc nt Tolera nc es 2.2.6 48 3.2.6 III
Axia l Operating Lengt h 2.2.6.1 49 3.2.6.1 III
Rad ial Locatio n 2.2.6.2 49 3.2.6.2 11 2
Mating Surface ormali ty 2.2.6.3 49 3.2.6.3 11 2
Rota ti onal Eccentricit y 2.2.6.4 50 3.2.6.4 112
Vibratio n Co ntrol 2.2. 7 50 3.2.7 112
v
SUBJECT STATE OF THE ART DESI GN CRITE RIA
Contamination All owa nces 2.2.8 51 3.2.8 11 3
Seal Mounting Requirements 2.2.9 52 3.2.9 113

Seal Pilot 2.2.9.1 52 3.2.9.1 113
Fl ange Seal 2.2.9.2 53 3.2.9.2 11 3
Fl ange Loading 2.2.9.3 53 3.2.9.3 11 4
Shims 2.2.9.4 54 3.2.9.4 114
Provision for Seal Removal 2.2.9.5 54 3.2.9.5 114
Locking Devices 2.2.9.6 54 3.2.9.6 114
Mating Rin g Mounting 2.2.9.7 54 3.2.9.7 114·
SEAL COMPO ENTS 2.3 55 3.3 11 5
Sea l Materia ls 2.3.1 55 3.3.1 115

Co mpatibility 2.3.1.1 55 3.3.1.1 115
Corrosion 2.3. 1.2 59 3.3. 1.2 11 6
Hydrogen Embrittlement 2.3. 1.3 59 3.3.1.3 116
Ma terial Properties and
Temperature Effect 2.3. 1.4 60 3.3.1.4 117
Wear and Friction 2.3. 1. 5 60 3.3.1.5 117
Cleaning 2.3.1.6 64 3.3.1.6 117
Passiva tion 2.3. 1. 7 64 3.3.1.7 11 8
Face-Contact-Seal Rubbing
Elemen ts 2.3.2 64 3.3.2 11 8
Seal Face Wi dth 2.3.2.1 66 3.3.2. 1 11 8
Seal ose Height 2.3.2.2 67 3.3.2.2 119
Seal Rin g Antirotation Device 2.3.2.3 68 3.3.2.3 119
Seal Rin g Pil ot 2.3.2.4 69 3.3.2.4 120
Sea l Insert Reten ti on 2.3.2.5 69 3.3.2.5 120
Insert /Carrier Separat ion 2.3.2.6 70 3.3.2.6 120
Insert Disto rtion 2.3.2.7 71 3.3.2. 7 12 1
Lapped-J oint Sea l-Ring
Carrier Assembly 2.3.2.8 72 3.3.2.8 12 1
Spray -Coated Seal Ri ngs 2.3.2.9 73 3.3.2.9 12 1
Circumferential-Seal Rubbin g Eleme nts 2.3.3 74 3.3.3 122

Segmented Shaft Sea ls 2.3.3.1 74 3.3.3. 1 122
Hydrostatic/ Hydrodyna mic Face

Seal Elemen ts 2.3.4 75 3.3.4 123
Self-Energized Hydrostatic Seals 2.3.4. 1 76 3.3.4.1 123
VI
SUBJECT STATE OF THE ART DESIGN CRITERIA
Externally Pressurized Hydrostatic Seals 2.3.4.2 77 3.3.4.2 124

Hydrody nami c Sea ls 2.3.4.3 77 3.3.4.3 124
Hybrid Seals 2.3.4.4 78 3.3.4.4 124
Ci rcum feren ti al-Sea l Clearance

Elemen ts 2.3.5 78 3.3.5 125
Labyrin th Seals 2.3.5. 1 79 3.3.5. 1 125
Fl oat in g-Ring Seals 2.3.5.2 87 3.3.5. 2 126
Arch-B ound Segment ed Sea ls 2.3.5.3 88 3.3.5.3 128
Face-Contact-Seal Secondary Elements 2.3.6 89 3.3.6 128

Meta l Bell ows 2.3.6.1 90 3.3.6.1 128
Pl astic Lip Sea ls 2.3.6.2 94 3.3.6.2 129
Elastomers 2.3.6.3 96 3.3.6.3 130
Piston Rings 2.3.6. 4 97 3.3.6.4 131
Spring Load 2.3.7 98 3.3.7 131
Pressure Balance 2.3.8 100 3.3.8 132
VII
LIST OF fiGURES
Figure Titl e Page
One-sea l system with back-up labyrinth seal for cO Illaining liquid propell ants 8
2 Two-seal sys tem with back-up labyrinth sea l for conta ini ng li quid hydrogen 8
3 Two-seal syste m for co nta inin g liquid hydroge n 8
4 Three-seal system fo r co ntainin g liquid hydroge n and (H2 + H2 0) h ot gas 9
5 Three -seal system fo r separatin g liquid oxygen and oil 9
6 Four-seal sys tem fo r se parating liquid oxyge n and RP-I 9
7 Four- ea l sys tem fo r se parating liquid oxyge n an d (H2 + H20) hot ga 10
8 Four-seal sys tem for se parating li quid oxyge n and

(H2 + H2 0 ) h ot gas (radial stack) . . . . . . 10
9 Four-seal sys tem for se parating liquid Ou orin e and

hydroge n hot gas (m inim um space) . . . . . . 10
10 Two six -sea l sys tems fo r epa rating high-pressure liquid ox idi Le r and fuel \I
II Face-contac t metal-b ell ows seal
12 Face-contact pl as tic-lip seal 22
13 Face-co ntac t piston-rin g seal 23
14 Face -co ntact elastomer seal 23
15 Th ree configurat ions for a se lf-energized hydro ta tic face sea l 24
16 Two co nfigura ti ons fo r an ex terna ll y pressurized hyd rostatic face sea l 25
17 Four configura ti ons for a hydrodynamic face sea l 26
18 Two co nfi gurati ons for a hybrid face ea l . . . 27
19 Face co nt ac t/clea rance ea l 28
vii i
Figure Titl e Page
20 Circumferential shaft-ridin g segmented seal 28
21 Circumferential shaft-riding arch-bound segmented seal 29
Circumferen tial hydrodynamic segmen ted seal . 29
23 Circumferential floating-ring cont roUed-gap sea l 30
24 Circumferential clearance labyrinth seal 30
25 Circumferential wear-in labyrinth seal 30
26 Seal fluid pressures in representative turbopumps 3]
27 Values for seal rubbing peeds in representative turbopumps 34
28 Values for FV factor for face-contact sea ls in representative turbopumps 37
29 Values for PV factor for face-contact seals in representative turbopumps 37
30 Viscosity of oxygen as a function of temperature and pressure 41
31 Viscosity of para-hydrogen as a function of temperature and pressure 42
32 Flow regime in seal leakage 43
33 Tensile yield trength of typical alloys as a function of temperature 61
34 Elongation of typ ical alloys as a function of temperature . . . _ 62
35 Th ermal expansion of typical seal materials a a function of temperature 63
36 Four types of rubbing elements for face-contact seals 65
37 Relation of face-con tact-seal face width and diameter 66
3"8 Method for reducing pressure separating force between sea l insert and carrier
for seals with higher pressure at the in side diameter . . . . _ . . 71
39 Method for reducing seal insert distortion due to thermal contraction

and cen troid misalignmen t . .. ... . . . . . . . . . . 71
40 Pressure forces on a lapped-joint seal ring 73
41 Pressure forces on a segmented shaft-seal ring (balanced de ign) 74
lX
Figure Title Page
42 Si x configurations for a labyrinth seal 79
43 Varia tion of seal fl ow coefficien t with Reyn old s number, various

labyrin th seal con figura tions . . . . . . . . . . . . . . 80
44 Leakage function for labyrinth seal s as a function of labyrinth pressure ratio 81
45 Optimum pitch for tee th of a straight lab yrinth as a function

of di ametral clearance . . . . . . . . . . . . . . . 82
46 Honeycomb co n truction for sta ti onary part of a wea r·in labyrinth seal 84
47 Configuration and pres ure profil e for a res trained pressure-b alanced
plastic wear·in labyrinth seal . . . . . . . . ..... . 85
48 Configuration of a sp rin g-l oaded, segmented , wear-in labyrinth seal 86
49 Pressu re forces on a fl oating-rin g seal 88
50 Pre sure-i nduced radial forces on an arch·bound segmented seal 89
51 Vari ous configurati ons for metal ·bell ows secon dary elements for face seals 91
52 Vari a ti on of bel lows effec ti ve span and mea n effective diam eter as a res ult
of press ure-indu ced defl ec ti on . . . . . . . . . . . . . . 93
53 Li p-seal elemen ts 95
54 Press ure forces on a lip seal 95
55 Current practices fo r spring loadin g face·contact sea ls 99
56 Press ure-indu ced fo rces on face sea ls 100
57 Current practice and reco mm end ed limi ts for PrV factor for
face -co ntact sea ls (3 -hr li fe) . . . . . . . . . . . . 107
58 Current practices and reco mm end ed limits fo r FV factor for

face ·con tact sea ls (3 ·hr life) . . . . . . . . . . . . 109
59 Current practices and recommended limits for PV factor for

face -co ntact sea ls (3-hr li fe) ....... . 110
~--~--- ---
LIST OF TABLES
Table Titl e Page
Summary of Chief Design Features of Representative

Turbopump Rotat ing-Sh aft Seals . . . . 4
II Advantages and Oi advantage of Various Types of Turb o pump

Rot3t in g-Shaft Seals in Current se . ... 18
III Summary of Materia ls Currently Used for Turb op uillp Rotat in g-Sh aft Sea ls 56
IV Reco illm ended Pressure-Bal ance Ratios for Turb op llillp Rotatillg-Shaft Sea ls 132
XI
L
LIQUID ROCKET ENGINE TURBOPUMP
ROTATING-SHAFT SEALS
1. INTRODUCTION
Seal for the rotating shaft of a liquid-propellant rocket engine turbopump are device that
prevent or minimize the leakage of propellants or fluids between the rotating shaft and the
stationary pump or turbine housing. The shaft seal must provide effective sealing during
high-speed rotation at all extremes of operating conditions for the specified life of the
engine. Because of the serious consequences of leakage of highly reactive oxidizers and fuels ,
the shaft seal is one of the most critical components on a rocket engine . A seal failure can
result in mixing of incompatible fluids, and the ensuing reaction may result in total
destruction of the engine and vehicle. Excessive propellant leakage can lower the engine
efficiency and may result in depletion of the propellant supply prior to completion of the
flight mission. This monograph ha been prepared to delineate the techniques that lead to
seal design in which the probability of eal failure is reduced to a practical minimum.
The design and application of rotating-shaft seals for turbopumps in rocket engine ystems
requires consideration of many factors that nonnall y are not critical in more conventional
applications. The extreme operating conditions combined with the high reliability and
fail-safe requirements dictate a thorough analy is of each detail design factor with
con ideration of all possible failure modes and operating variations. Most of the problems
with turbopump seals have been related to material compatibility with the rocket engine
propellants and to operation at the extremely low temperatures (- 297° to - 423 °F)* of
the cryogenic fluids. The low-temperature problem i compounded on some turbopump
seals by the extreme temperature gradient that occur when the cryogenic fluid must be
sealed along a shaft that is adjacent to a high-temperature (I 200°F) turbine. Additional
thennal problems are created by the heat generated at the seal face by rubbing friction and
viscous shear at the high rotational speeds (SOO ft/sec). Rubbing-surface temperatures in
excess of 1000°F have been measured on liquid-oxygen seals when the fluid environment
was at -297°F.
The shaft seals currently in use on rocket engine turbopumps are primarily the face-contact
and circumferential** types. In general, the face-contact seals (face seals in rubbing contact
with the mating surface) are used for sealing liquid propellant, and the circumferential seals
•Factors for converting U.S. customary units to the International System of Units (SI units) are given in Appendix A .
•• Seal terminology used in the text ba ically is that presented in ASLE SP-l (ref. I). Symbols, materials, and abbreviations
are defined or identified in Appendix B.
are used for sealing turbin e hot gases. Circumferential seals may employ either
rubbing-contact seal elements or clearance elem ents. Circumferential clearance seals,
particularly labyrinth types, are used extensively where reliability is the primmy
consideration and the increased leakage over that of fac e and circumferential clearance seals
is acceptable.
The monograph deals primarily with the experience and knowledge accumulated in
development and operational programs to date ; therefore, most of the discussion is
concerned with the face-contact, circumferential, and labyrinth seals. Because future
requirements indicate the need for more advanced configurations such as the hydrostatic
and hydrodynamic seals, these designs are also treated.
The monograph is organized around the sequence of tasks normally involved in seal design:
(1) Seal System. - The arrangement of seal assemblies, drains, and purges is selected
to minimize the severity of seal operating conditions, provide allowances for all
extremes of operation, and allow for a single seal failure without destructive
failure of the turbopumps. The seal designer must be involved at the preliminary
turbopump design layout to ensure that the required operating conditions are
consistent with reliable seal performance.
(2) Seal Assembly. - The basic type of seal assembly is selected to satisfy the
operating and performance requirements. The design problems and performance
limitations that are related to the seal assem bly and turbopump application are
considered. The seal assembly and seal system designs are iterated until an
acceptable compromise between operating conditions and performance is
established.
(3) Seal Components. - The detail design analysis necessary to establish the
requirements and configuration of the detail components of the seal assembly is
performed and integrated with the seal system and seal assem bly designs. The seal
component design analysis treats the following design factors:
• Leakage • Resonant frequencies

• Seal materials • Dynamic response
• Seal interface configuration • Power loss
• Secondary element • Heat transfer and thermal analysis
• Pressure balance • Stress and deflection analysis
• Spring load • Dimensional tolerances
• Secondary friction
These tasks are considered in the monograph in the order and manner in which the designer
must handle them. Within the task areas, the critical aspects of the requirements that the
seal design must satisfy are presented.
- - - - - - - - --- ---
2 . STATE OF THE ART
The technology for turbopump rotating-shaft seals has evolved through approximately 20
years of rocket engine development. Because of the severe operating conditions for
turbopump seals, most turbopump seal designs have required considerable development and
have extended the state of the art, particularly in the areas of material compatibility,
low-temperature capability, and high-speed service.
Compatibility of seal materials with rocket engine propellants has been a major problem.
The highly reactive propellants are not compatible with many of the conventional seal
materials. The strong oxidizers such as liquid oxygen (L0 2 ) or liquid fluorine (LF 2 ) are
capable of combustible reactions that can result in total destruction of the engine and
vehicle. Many of the early seal experiments resulted in explosive reactions. Much effort was
expended over the years to compile data on material compatibility and to make effective
use of the information in successful seal design.
The extremely low temperatures (-297°F to - 423°F) of cryogenic propellants precluded

the use of conventional elastomer seals and resulted in the development of the plastic-lip
and metal-bellows designs. Welded metal-bellows have been extensively developed during the
past 10 years and are currently the most common seal configuration in use with cryogenic
fluids. Because cryogenic fluids do not provide significant lubrication, material
combinations that are self-lubricating have been developed.
The low temperature of the cryogenic propellants in combination with the high temperature
of the turbine hot gases results in severe temperature gradients through the seal system and
across the seal assembly. The hardware is chilled to cryogenic temperature before start and
must withstand the thermal stresses caused by heating from the hot gas during operation.
Additional thermal gradients are caused by the heat generation at the seal interface by
rubbing fliction and viscous shear. Therefore, considerable development effort has been
directed at providing for transient and operational thermal contraction/expansion, thermal
stresses, and thelmal distortion.
A summary of the chief design features of rotating-shaft seals for representative rocket
engine turbopumps is presented in table I. The seal types are grouped according to the fluid
being sealed. The materials for the seal nosepiece, mating ring, secondary seal or bellows
element, and housing are listed for each seal. Seal diameter, pressure, and speed indicate the
operating conditions. The seal face load and pressure-balance ratio show the design
parameters. Face load is calculated from the sum of the spring load and the hydraulic load
caused by fluid pressure differential. Hydraulic load is based on the pressure-induced closing
force minus an assumed interface-pressure-induced separating force equivalent to one-half
the pressure differential (linear profile). The integration of these data with the successful
test experience fonTIS the basis of the discussion under the separate design su bjects.
L _
- = - - - - - - - - - - - -~- - - - - - - - - - - -------~~~=~-
Table I. - Summary of Chief Design Features of Representative Turbopump Rotating-Shaft Seals
5ucceu(ul test experience

Total PV
Haterials Fluid Unit
Sealed Seal Shaft Rubbinj Sprtng load, factor , Balance Total Total Wear Dynamic Engine
Seal type pressure, load ,
fluid
dim, speed. speed, load, lb. psi x Cps ratio number test, life, leakaee
Secondary psie psi h,
Nose·
piece
Mattng
material 0'
bellows
lIous 1n8
in. 'pm fps 1bf
CD CD
(xl03)
@ o tested hr
® ® progTalll
Llquld Shah RIding Carbon No plating 304 3.5 600 22000 367 NA NA NA NA NA NA NA NEltVA
Hydrogen Segmented Carbon CDJ83 A286
Liquld Fact' Contac.t Ca rbon Chrome on 1 AHS 4530 I AH5 5735 2.09 85 30800 280 16 31 50 14 .75 NA NA NA NA RL-IO
Hydrogen Metal Piston Ring P5AG Inconel or 4650
5667
Liquid Face Contact Carbon Chrome on I AHS 4530 I AH5 5735 1. 70 400 30800 214 38 76 16 .55 NA NA NA NA RL-IO
Hydras .. n Meta 1 Pisto n Ring
'5'" Inconel
5667
o r 4650
Liquid Face Contact Carbon Chrome on 1 AHS 4530 1 Inconel 1.59 500 30800 235 17 52 12 .6 NA NA NA NA Rt-l0
Hydrogf'n Hetal Platon Ring '5M; Inconel or 4650 5665
5665
Gaseous Face Contact Carbon Chrome on I 347 lnvar 36 I 2.531 20 34000 375 11 4.2 .7 17 48 1 SerM Pl'ioebus
Hydrogen Welded Bello.... a P5N 310
- 400o F
Gaaeoua Face Contact Carbon Chrome on I 347 AMS 5646 1.23 20 12300 66 NA 17 1.1 .6 NA NA NA NA Rt-IO
Hydrogen We Ided Be 110wa P5AG JJ1S 5735 or 5512
.j:::.. -3600F
Gaseous Face Contact Carbon Chrome on 1 AMS 4530 1 Incone 1 2 . 02 450 30800 269 12 55 82 22 .6 NA NA NA NA RL-JO
Hydrogen
-1100r
Metal Piston Ring .5'" AHS 5665 o r 4650 5665
Gaseous Shaft Riding Carbon Tungsten NA 3.0 700 24000 312 NA NA NA NA NA NA 150+ 4+ I NA NERVA
Hydrogen Segmented Carbon CDJ83 carbide on
A286
~;~bon ~~~~:10~ 1 ~~~onel

Caseous Face Contact 321 2.950 50 28400 434 25 29 12.6 .7 100+ 100+ 3 SCFM )·2
Hydrogen We lded Be 110wa
- 400o F
Ho t Gas Shaft Riding Carbon Chrome on Inconel 3.768 100 28400 460 15 62 28.6 1.5 100+ 100+ 20 SCFH )·2
Inconel X 600
~~0~~20
SegmC!nted Carbon G84SC
Hot Cas Face Contact Carbon Chrome on 1 Inconel 321 3.339 75 9000 130 19 '.1 40 5.2 .8 100+ 100+ 3 SCFM )·2
HZ + H20 We Ided Be llows P5N Inconel X 750
IOOOOF
Hot Cas Face Contact Carbon t W- 5 on 1 Incone 1 321 4.53 130 6000 120 11 51 38.6 4.6 .95 .38 NA NA M-1
5000r CN Welded Bellows Erl05 lnconel X 718
2
Hot Cas 1 Face Contact Carbon No p l ating I JJ1-350 3'7 2.668 95 25000 271 NA <A 40 11 . 65 100+ 10+ .9 NA Titan III
A- SO + N20 4 Welded Bello..... CDJ8) 440C
Liquid Face Contact AI203 on A1203 on Inconel Inconel .620 50 75000 205 80 16.4 .6 .03 NA 1 SCFM R&D
Fluorine WC!\ded BC!110..... Inconel InconeL 718 718
600 718
Liquid
Fluorine
Face Contact
We 1ded Be 110ws
K162B
~~6~~·ting 1 ~~~onel Incone)
750
1.500 50 28000 170 15 25 60 10 .6 .03 NA 2 SCFM R&D
Liquid Face Contact AI20) No plating 1 1ncooel Incont'l 1.82 600 11000 88 14 26 44 3.9 .6 21 6.3 .2 NA Modified
Fluorine Machined Bellows K162B 718 718 Rt- IO
(continued)
~-- --~I
Table l. - Summary of Chief Design Features of Representative Turbopump Rotating-8 haft Seals (continued)
Succe •• [ul test experience
Sealed SuI type

Haterial . Seal
diam,
Fluid
pressure,
Shaft
speed.
I
Rubbini'
speed,
Spr i ng
lo.d,
Toul
load,
lben
Iunit
load,
psi
PV
factor,
pS~x~o!~s
Balance
ratio
Toul
nWllber
I I
Total
test,
Wear
life,
Dynamic
leakage
Engine
fluid Nose-
piece
Hat i ng
I118tedal
Secondary
0'
bellows
Hou sing
In. psii' rpm fps lb£
CD CD o o tested hr
o
hr @ program
Liquid Fate Contact Carbon Chrome on I 347 304 2.630 225 6750 77 31 95 158 12.2 0 .97 30 35.8 1 I to SCfl1 I Thor
Oxygen Welded Bellows P692
I
4130
I I
LiquLd
Oxygen
Liquid
Face Contact
Welded Be llows
Face Cont.ct
Carbon
P692
Carbon I
Chrome on
loDO
ChrOIDe on I
lnconel
718
lncon e l
304
321
2. 630
2 . 974
200
200
6800
S650
78
112
82
41
106
150
176
210
13.7
23.5
.7
.85
1_
85 29
100+ I
I 10 SCFH
I 15 SCfH
I H-l
I J-2 I
Oxygen Welded Bellovs P5N Incond X 750
Liquid
Oxygen
Fa ~e Contact
Welded Sello"'!!
Carbon
P692
I Chrome on
lnconel X
I lnconel
718
321 2.935 200 9000 115 41 150 210 24 .85 25 15 I 15 sent I J - 2S
I
Liquid
Oxygen
Face Contact
Welded Bello"'s
Carbon
P5N
I LW-5 01'\
(nconel X
I (nconel
718
321 6.902 450 4000 120 20 471 196 23.5 .92 21 2.4 I 1. 7 I 2.3 SCFM I H-I
I
Liquid F!!ce Contact Carbon I Chrome on I ln conel AMS 5646 1. 70 400 12300 92 17 67 41 3.8 .65 NA NA I NA I M I RL-IO
Oxygen
li quid
Welded Bello"'!!
Fa~e Contact
P5JC
Carbon
AMS 5646
Chrocne on I
750
Hylar 303 6.463 140 6000 170 58 127 51 8.7 .7 10<>t 100+ I I H SCFH I F-I
I
VI Oxygen
Liquid
Oxygen
'Lip Secondary
Face Contact
Up Secondary
P692
Carbon
P692
440C
Chrome on
4130
I Kel-f 302 2.650 225 6800 78 J2 82 130 10 . 85 100+ 100+ ' I 10 SCn1 I Thor
I
liquid
Oxygen
Fate Contact
Lip Secondary
Car bon
P692
Chrome on
4130
I Ke1-F 302 2.630 250 10000 115 J2 86 137 15.8 .85 10<>t 100+ I I 10 SC FH I AtLa!! I
Fa~e ~ I I I (n~onel
Liquid
Oxygen
liquid
Hybrid
Welded Bellow!!
Shaft Riding
Carbon
1>692
Cubon
Chrome on
Inconel X
LW-S on
718
lnc o nel
718
321
2.65
4.33
50
385
25000
6000
290
113
34
NA
NA
NA
NA
NA
NA
NA
.6
NA
14
.66
I
I
10
NA
I IS SCFM
13. SGPH
I ADP
I H-I
I
Oxygen
liquid
seglllented Carbon
Fate Contact
P5N
Carbon I
Inconel X
Chrome on I lncone 1 321 2.950 200 28000 360 15 64 89 32 .7 10<>t 100+ I 1 .01 Ibm/sec I J-2
I
Hydrogen Welded Bellowl P5N Inconel X 750
Liquid
Hydrogen
Face Contact
Welded Bellows
Carbon
P5N
I Ch r ome on
lnconel X
I (ncone 1
7 18
(nconel 3 . 515 350 28000 430 20 95 101 43.5 .7 25 15 1 .02 Iblll/sec I J-2S I
600
liquid
Hydrogen
Fa~e Contact
Welded Bellows
Carbon
P5N
I Chrome on I 347 (nvar 36 2.531 150 34000 375 25 31 11.6 .7 17 48 1 .006 1bm/sec I Phoebu s I
Uquid
Hydrogen
Fate Contact
Welded Bellows
Carbon
POJN
I LW- S on
Inconel X
I Inconel
750
(ncone l 4 . S3 300 18000 356 11 133 100 35 . 6 .8 .13 I .15 I NA I H- I
Liquid Shaft Cleaunce Carbon Lw-5 on 321 S.24 210 15500 385 NA NA NA NA HA NA . 56 I NA I 50 CPM I M-I
Hydrogen Arch Bound P5N Inconel X
Segmen t ed Carbon
Liquid Shaft Riding Carbon Tunglten 304 3.0 150 24000 312 NA NA NA NA NA NA 200+ I 10+ I NA I N~VA
Hydroge n Segme nt ed Car bon CDJ83 carbide on
A286
(continued)
Table 1. - Summary of Chief Design Features of Representative Turbopump Rotating-S haft Seals (co ncluded)
PV Successful test experience

Haterials Total Unit factor,
Sealed Seal type
Seal Fluid Sha ft Rubbine Spdng load. load, Balance Total Wear Dynamic Ens ine
psi x fps Toed
diam, pressure, .peed , speed, l oad, lb. psi ratio test~ life, leakage
fluid Secondary (xl03) number
Nose- Mat ing in. psiS <P' fps lbf hr hr progrul
piece IUterial
0< Housing
eD ® (1) ® tested @)
bellows
®
FLOX ® Face Contac t
Lip Seconda r y
K1628 No platinRi Kel-F
K162B
)02 2.626 225 10000 11 4 )0 61 97 11 .7 ).1 10 SCFM R&D
H 0, Face Contact Carbon LW-5 on I AM- 350 )00 2.5 50 8360 91 NA NA 31 2.8 .7 40+ 3+ 1+ I NA TLt.an III
2 Welded Bellow. EY -I D5 321
A- 50 @ race Contact Carbon LW- 5 on I AH- 35D )47 1.887 50 23500 195 10 NA )) 6.4 .7 40+ 3+ 1+ I NA TLtan III
Welded BellO'oiI E"i-105 321
RP-l Face Contact Car bon Chrome on I Buna-N )00 2.36 4 250 10000 107 10 64 10) 10.6 .85 100+ 100+ 4+ I 5 CC/HR Atl as
V-Pac king Secondary C39 4130
RP-l Face Contact Carbon Chrome. on I Buna-N )00 3.177 200 6500 90 22 75 99 8.9 .85 100+ 100+ 4+ I 5 CC/HR Thor
V-Pa cking Secondary C39 4130
RP-I Face Con tact Carbon Chrome. on I AM-350 )02 3. 174 200 7000 97 18 91 117 11.4 .97 )6 46 4+ I 2 CC/HR H-l
Welded Bellow. CCA-72 4130
RP-l Face Contact

O-Ring Secondary
Carbon
C39
Chrome. on
440C
I Viton A I Al 2024 6.497 )00 6500 181 47 304 "6 22.9 .8 100+ 100+ 100 CC/HR F- l
0\
RP-l / Face Cont act Carbon Chrome on I AH ~350 Kaatell oy I 10.142 120 6 500 280 )00 414 81 22.8 . 65 12 4.8 NA I NA F-l
Hot Gas Welded Bellow . P2003 Hut e lloy
SOCOr
Hot a.. Shaft JUding Ca r bon Chrome. on 4 16 9.500 50 6500 270 NA NA NA NA NA 100+ 100+ 4+ I NA F-l
LOX + RP-l Segmented Carbon GM 4340
1000etr
Hot C.. Shaft Rlding Carbon Tungaten 416 2.000 125 40000 )50 NA NA NA NA NA 100+ 100+ 4+ r NA Atlas
LOX + RP-l Segmented Carbon G84 c arblde on
4)40
Hot C.. Shaft Riding Carbon Tungaten 416 ) . 000 160 29000 )80 100+ 100+ 4+ I NA H- l
LOX + RP - I Floating Ring P33 carbide on
4340
CD Sprint Load + Hydraulic Load (Baaed on Linear Profile) NA - data not available
** - dou not apply
<y To tal Hoae Load ... Non Contact Area
CD Nose Unit Load x Rubbing Speed . psi )( fp s
o Effect ive Clodng Area ... Sealing Dam Area
~ Es timated Minimum Effe c tive Sealing Life
@ Me asured or Estimated Ma xilTlJm
G Annular Grooved Faca
® M1xtu ~ 30% liqui d Fluorine + 70'( Liquid Oxygen
® Aerozine 50 (50\ hydruine • SO\ UO,...)

Anticipated operating conditions for turbopump seals in the future will require the current
state of the art to be extended in the area~ of long life, high speed , high pressure, and low
leakage. For exam ple, the operating conditions for the seals in the turbopumps on the
SSME * includ e liquid-oxygen pressures up to 650 psia, hot (900°F) gas pressure up to
4000 psia, and rotational speeds up to 400 ft/sec, all combined with a wear-life req uir ment
of 10 hl'. Current designs generally are capable of satisfactory operation at one of these
conditions; howeve r, new concepts will be required to satisfy all the co nditions
concurrently. The current designs for rubbing-contact seals provide low leakage at pressures
up to approximately 500 psi or speeds up to approximately 500 ft/sec; however, the wear
lif i limited to approximately 4 hours. The current designs for circumferential floating-ring
controll ed-gap or labyrinth seals are capab le of meeting the higher pressure-speed-life
req uirements, but the leakage rate is excessive for many applications. The seal concepts that
appear to have the best potential for satisfying all of the future requirements are the
hydrostatic or hydrodynamic designs. These seals maintain a small (~ 0.0001 to 0 .0004 in.)
clearance gap tha t essen tia lly eliminates ru b bing con tact and minimizes leakage. The test
experience in cryogeni c flu ids is limited; however, the feasibility of the design concepts has
been demonstrated in seve ral programs.
2.1 SEAL SYSTEM

The eal system consi ts of the arrangement of seal assemb lie , drains , and purges required
to satisfy the overall de ign objectives. The first and one of the most important steps in the
design of turbopump seals is the tradeoff analysi for selection of th e total seal system.
Reliable operation of the turbopump requires a seal system designed to minimize the
severity of seal operating conditions, provide allowances for all possible extremes of
operation, and allow for seal failure without destructive failure of the turbopump. Many
methods to accomplish these goals exis t; however, it is usually necessary that the design
features be incorporated as part or th e preliminary turbopump layout, since many of the
seal system design requirements are interrelated with other turbopump design features and
may be difficult to incorporate la ter.
The seal system design generally is iterate d with consideration of the individual seal
assembly capabilities until an acceptable compromise between operating conditions and
perf0l111an Ce is reached. The relative severity of the operating conditions is evaluated by
comparing the fluid-pressure/ rubbing-speed relation for specific types of seals and fluids
with the current practices and recomm ended limits. This relation provides a guide to the
feasibility of alternate configurations for the system design before the detail seal-assembly
designs are established.
The types of turbopump seal systems currently used to contain and separate rocket engine
propellants are illustrated and described in figures I through 10 .
•Space Shuttle Main E ngine .
7
HIGH PRESS LR E
One-sea l system; used when propella nts can be

8ACK·UP safely drained to atmosp here or w h en externa l
LABYRI><TH --\--f--~
leakage is not harmful . Labyrin t h seal provi d es
backup in event of face seal failure.
propella nts.
F igure 1. - O ne-seal syste m with back- u p labyrint h for contain ing liquid
DRAIN
Two·sea l system with a drain; used w h en pro-

pellants (e .g., liquid h ydrogen ) may be hazar-
dous when mixed with atmosph eric air or
when externa l leakage would be harmful .
Labyrin th sea l provide s backup in event of
face-sea I fa ilure.
BACK·U P
LABYRI><TH
MK9 liQUID HY~OCEN TI..RBOPUMP
h ydrogen .
F igure 2 . - T wo-sea l syste m wit h back·up labyrint h for contain ing liquid
STATIC lI fT-lJ'F SEAl (PRESSURE ACTUATEDI
T wo-seal syste m consisti ng o f o ne face-con t act

ACTUATION TUR81NE HOT GAS
PRES SURE-- - - - 1HZ ~ HzOl sea l and one static liftoff seal; used when pro-
pe ll ants can be safe ly drained into the turb ine
HYDROGEN SEAl
hot-gas area and w h en minimu m static leakage
UAKAGE DRAINS
INTO TLR 81M: is des irab le . Th e stat ic li ftoff seal provide s
static sea li ng d u ri ng nonrota tion periods and
lIOUID
HYDROGEN MATI NG RI NG is press u re act u ated to lift off t h e mating ring
d u r in g rotation .
LIQUID HYDROGEN SEAl
J2·S LIQUID HYDROG£N TUR BOPlMP
F igure 3 . - T wo-sea l syste m for conta in ing liqu id h ydrogen .
J
Compact three -seal system consisting of two
face-contact seals and one circumferential sea l
TLOI BIr-L SEAL with a drain; used when propellants and tur-
bine hot gases can b e safely mixed in a com -
mon drain cavity . The primary seal drain
HOT C"S
IH2 .. HzOl utili zes a relief valve to maintain a d esired pres-
sure level to reduce the pressure differe ntial
LIQUID
~YDROGEN across th e primary seal and ens ure coo ling of
the secondary sea l. Th e secondary sea l pro-
vid es additional sea ling capacity in t he event
of a primary sea l fai lure a nd preven t s moist-
J2 liQUID HYDROGEN TLRBOPlNP ure from fr eez ing on the primary seal.
F igure 4. - Three-sea l system for containing liquid hydroge n and (H 2 + H20) hot gas.
Three-sea l sys tem cons isting of two face -

contact sea ls and one purged labyrinth inter-
GAsm . 016" -+~Hg~~,~ ~ medi ate sea l; used to separate two incompat-
ibl e fluids such as liquid oxygen and lubricat-
LIQUID
OXYGEN ing oi l. Th e leaka ge from the face-contact
....TlNG RING -I~~~
sea ls is drained out through separate drains .
SHI M Th e labyrinth inter mediate sea l, which is gas
purged , provides se paration of the drain cavi-
ties during normal operation, but may not be
effective in the event of a seal failure.
Figure 5. - Th ree-sea l system for se parating liquid oxygen and oil.
lOX CAVITY DRAIN 0 PlACES)

Four-seal system consisting of two fac e-
contact sea ls and a purged, doubl e circumfer-
LIQUID
OXYGEN ential intermediate sea l; used to separate two
lOX SEAL incompatible fluids such as liquid oxygen and
RP-l. Th e leakage from the face-contact seals
~~~~-RP -l SEAl
INTERMEDIATE SEAL --H-.'='" is drained through separate drains. The double
IP~G£DI
circumferential sea l, which is inert-gas purged,
MATING RI
provides effect ive se paration of the drain cavi-
ties in th e eve nt of a seal failure . The slinger
prevents direct impingem e nt of high-velocity
SHAfT
fluid and assists in routing the leakage through
F· t TUR 80P I)IflP
the drain .
F igure 6. - Four-seal system for separating liquid oxyge n an d RP-1.
9
l
INlERM(OIAT[ SUol p~G{
fUR l lNE SEAL DIiIAIN MAN IFOCO

Compact four-seal system consisting of two
face-contact seals and a purged double circum-
PRIMARY SEAl MAIN MANIFOlD MATING RING ferential intermediate sea l; used to separate
two incompatible fluids such as liquid oxygen
SE,,-
and hydrogen-rich hot gas. Th e principle of
LIQUID OXYGEN HOT GAS operation is the same as that in the system
IH1 ' Hil'
sh own in figure 6. Overheating and distor-
tion of the mating ring may be a problem
with two seals rubbing on the same ring.
PRIMARY SEAl --~
J2 LIQUID OXYGEN T~ BOPUMP
Figure 7 . - Four-sea l system for separating liquid oxygen and (H 2 + H 20) hot gas.
TlIt.lN[ IHTIRFAC[
Minimum-axial-space four-seal system consist-
ing of four face-contact seals stacked radially;
used to separate two incompatible fluids. Prin-
ciple of operation is the same as that in the
system shown in figure 6. Th e face-contact
seals generally provide a more effective seal
than the circumferentia l shaft seals; however,
the increased radial space required, addition-
al complexity, and potential mating ring
overheating and distortion may prevent their
M-) LIQUID OXYGEN T~80PUMP
use as dry-running intermediate seals (ref. 2).
Figure 8. - Four -seal system for separating liquid oxygen and (H 2 + H2 0) hot gas (radial stack).
Compact m inimum-radia l- and axial-space four-

seal system consisting of two face-contact seals
TLIlBIN' SUI. and a purged, double circumferential intermed-
TLR81N( Ml:El iate seal; used to separate two in compatible
fluids. Principl e of operati on is the same as
that in the system shown in figure 6. Mating-
HYOI!OGEN
BEAAING---r=P--.J-.hR"SI'" HOT GAS ring distortion may be a problem due to use
of structura lly stressed parts for the mating
ring. Separate, nonloaded mating rings are pre-
R &- 0 LIQU ID FUJORIN£ TUR80PUMP ferred.
Figu re 9. - Four-seal system for separat ing liquid fluorine and hydrogen hot gas (minimum space).
10
.__ J
HEll UM RETURN TO LOW
DRAIN PURGE DRAIN PRESSURE AREA
HIGH
PRESSURE FUEL
HIGH PRESSURE
LIQUID OXIDIZER
IRCUMFERENTIAL CI RC U....HRENTI AL
CLEARANCE SEAL ClEARANCE SEAL
ITO REDUCE PRESSURE ) ITO REDUCE PRESSURE I
PR IMAR Y OXIDIZER
SEAL
Six -seal system consisting of four face-contact sea ls a nd two c ircu mferent ia l c learance sea ls; used to
separate two high-pressure incompatible f luids wh en ax ia l space is not c riti cal. Th e circumferential
clearance seals (e.g . , floating ring and labyrinth) d ec rea se th e high pressure to a va lu e acceptable for
face-contact sea ls by allowing recirculation bac k to a low-press u re area such as t h e pump inlet . The
leakage from the primary seals is drain ed through se parate drains. Th e inert-gas purge between the two
face-contact in t er mediate seals provides an effective separation of th e drain cavities in the event of
seal failure.
(a) Si x-seal system u si ng intermediate face sea ls
OVERBOAA o..
V ENTS
CARB ON
FUEL
SE A L
MATING R I N G
MATING RING
L A B Y R I N TH
CA R BON INT E R M EDIATE

S E AL RING
MATING RING
Si x-seal system consisting of two face seals, two labyr int h seals, and two ring seals with both radial and
c ircu mferential contact faces; used to separate two high -pressu re incom patibl e fluids when axial space
is not critical. Th e face seals decrease th e high pressures to a mod est value suc h t h at the ring seal can
ensure low overboard leakage. Should a face sea l fail, a laby rint h sea l limits the overboard leakage to an
acceptab le level. For ground test operat ions, where back pressu re is high, the ring seals can be center
pressurized with helium; this ensures safe separation of the two propellants by providing a hel ium da m
between b ot h vent ca viti es (refs . 3 , 4, an d 5 ).
(b) Si x-seal system using intermed iate ring sea ls
F igure 10. - T wo six-seal sy ste ms for separating high-pressure liquid oxidizer and fuel.
11
J
2.1.1 Pressure Environment
Hi gh-pre ure e nvironm e nt fo r ea ls ge ne rall y re ult in lo we r turb o pump reli a bilit y a nd
large r prop ell a nt leak age. High pres ures a lso in c rease th e di ffic ul ty of separa ting
inco mp atibl fluids. The pre lire leve l is redu ce d to be co ns is te nt with th e c urre n t st ate of
the art for sp ecifi c typ es o f ea ls a nd fluid by utili z ing u p trea m la b y rinth o r
circumferenti al clea ran ce eal in co nju nc ti on wit h lo w-pre sure return bl eed th a t a ll ow
recirculation to a low-pressure area (fig . 10) .
High differential pressures a re redu ce d by m a in ta inin g an e tab li sh ed pre su re I ve l

down tream of th e prim a ry seal. Th e prim ar y sea l m ay be dra in ed direct ly in to th e turbin e
hot-gas area ( fig. 3) if th e fluid s a re co mp a tib le. o r a re li c!' va lve m ay be utili ze d in th e dra in
in conjunc tion with a n additi o na l sea l in e ri es w ith th e prim ary sea l (fig . 4 ) to red uce the
diffe re ntial pre sure . It is a lso poss ible to m a inta in a pre sure leve l dow ns trea m of the
prim a ry seal b y press uri z ing the drain cavity with a n in e rt purge.
Seal lo cated nea r pump impe lle rs gen e rall y a re ubj ected to exce ive prc urc osc ill atio ns
ari in g from th e high-frequ en cy pu lses o f disch arge . F a tigu e fa ilure of th e sea l co mp o ne nts
such a th e seco nd ary-e lem e nt m e tal bell o ws o r pl as ti c lip m ay res ult if th e sca l is not
pro t ec te d. The amplitud e o f pressure oscill a ti o ns at th e ea l is minimi zed by u t ili zing
lab y rinths up tream o f th e seal as d ampin g d ev ic s (fig. 4 ).
2.1.2 Thermal Environment

The low t emp erature of th e c ryoge nic p ro pell a nt and the high te mp e rature of the turb in e
ho t ga e result in seve re temp e rature gradi e nts in th e sea l ys te m a nd ac ross th e sea l
a se mbli es. The th erm al stres es crea ted by th e di ffe re n t ia l th e rm a l ex p a n io n/ co ntract io n
are minimi zed b y d e ig nin g th seal ho u in gs a nd m a ting rings to prov id e m inimum res traint
t o the rm a l d e fl ec ti o ns a nd by u in g m a te ri als w ith simil a r th e rm a l ex p a n io n/ co ntrac ti o n
ra t es. High te mp e ra tures a t the ea l loca ti o n a rc minimi ze d by d e igning th e sys te m to
i o la te the ea l in a cooled a rea . Ba rrie rs o r labyr in th dev ices a re prov id ed to p reve n t the
dir c t fl o w of h o t gase o nto th ea l a sem bly.
2.1.3 Vacuum Environment

Many fubbing material t e nd to wea r excessive ly o r fa il co mpl e te ly und e r vacuum
condition beca use th e o x ygen and wate r vap o r t ha t a re required t o fo rm lubri ca t ing sur face
oxid s are a b en t. It i g n e ral prac ti ce t o design th e tu rb o p u m p sea l ys tem such th a t th e
rubbing seal ar no t ex posed t o a h ard vac uum. Th e seal drain s a re ro ute d to a safe di p o al
area inte rn al to the turbopump or e ngin e (e.g .• th e turbin e ex h au t o r pump inl et), o r th e
drain are e limin a ted and the eal lea k age is a ll o wed to vent direc tl y int o th e turb in e (fig. 3).
Pres ure-re lie f valves m ay be Ll sed in th e seal dra in sys te m o r a dra in cav ity purge m ay be
use d to e nsure a posit ive pre ure in th dra in .
12
L
Atmospheric drains norm all y are long e nough 0 that on ly partial diffusion of the
drain-cavity a tmo phere will OCC lIr. AI 0, th e sea l lea kage normally is sufficie nt to maintain
a positive pressure in th e drain system ; th e re fore, it may not b necessary to provide p ec ial
allowances for vacuum ope ration.
2.1.4 Rubbing Speed

High sea l rubbing speeds res ult in excessiv e h ea t ge neration and require additional coo lin g
ca pa c it y to preve n t th el111al fai lures. Ru bbing speeds are min im ized by u ti l izing th e mall es t
possibl e sea l diam e ter. Co mpromises in the pump d e ign to reduce the shaft rotational speed
are co ns ide red whe n th e rubbing speed i excessiv e. Large r-di am ete r impell er m ay be u ed
or th e pUIllP ope ra ting h ea d and flow re lat ions m ay be adju ted. For reli a bl e operatio n, the
rubbing-speed / sea led-fluid rela tions for pecific type of ea ls a nd fluids mu st be w ithin the
curre nt tate of the art.
2.1.5 Cooling and Lubrication

Rotating- h aft sea ls req uire ad eq u ate coolin g to di ipate the heat gene rated a t th e seal face
by rubbing fri c tion and viscous shear of th e sea led fluid. Also, sea l m a te ri als generally
require some form of lubrica tion for satisfactory p e rform ance. Th e sea led fluid usually
provides ad equ ate cooling and lubri ca tion if prov i io ns are made to all ow the fluid to flow
through the seal cavi t y . Bleed hol es through the imp ell er fro m th e sea l cav it y to the imp eller
inlet, o r through the shaft back to th e pump inl e t, a re used where po ibl e. Pumping vanes
on th e impeller rear shro ud also are utilized to e nsure rec irc ul at ion of th e ealed fluid .
Insta ll ation of the seal in a d ea d en d cavity that may all ow acc umul a ti o n of a vapor pocket
and resu lt in poor cooling a nd lubri ca tion is avo id ed .
Dry-running face-contact intermediate seals (fig. 10(a)) may ove rh eat a a result of lack of
coolin g and lubri ca tion. Circumfe rential sh aft seal (fig. 6) generally are used for
dry-running interm edi ate sea ls. Cooling an d lubri ca ti on of dry-running seals are provided
when required by purging with a n inert fluid or by inj ectin g a coo ling flu id into the seal
cav ity or onto the sea l mat ing ring. Separa te cool ing and lubri ca tion sys t ems, wh ich are
isolated from the ealed fluid , a re LLsed when th e coo ling fluid is not compatible with th e
ealed fl uid .
Mating rings usuall y are loc ated toward the sealed fluid with pressure on th e outside (fig.
10(a)) in order to ensure fl ow of coolant across th e back sur face of th e ring fo r m axim um
h eat transfer to the sealed fluid . Mating rin g. also are drilled to all ow coolant flow through
the ring secti on. Pumpin g vanes or holes in the m atin g ring a re used to ensure coolant flow
aro und th e seal face. However, modi fication of th e mating-ring sectio n by in corporating
van es or holes may res ult in distortions of th e m a tin g su rface.
13
Overheating and distortion of the mating ring may be a problem with two seals rubbing on
the sa me ring (figs. 4 and 8). If space is not limited , series seals on separate mating rings are
prefe rred (fig. 10(a)).
2.1.6 Leakage Drains

Adeq ua te d rains are provided in the seal system to ensure safe disposal of the seal leakage.
Seal drain s are sized to accommo date the maximum anticipated leakage without building up
significant back pressure. The drain effective flow area is calculated with conven ti onal flow
equations by es timating the maximum anticip ated seal leakage and establishing the
al lowable drain cavity pressure. The drain pressure differential and leakage rate then
esta blish th e required drain size.
Seal leakage rate for a cryogenic fluid is calculated on the assumption of liquid conditons
up stream of th e seal. For calculation of the required drain size, the calcula ted liquid leakage
i then conv erted to an equivalent gas volum e at an assumed temperature. This m e thod
provid e the most conservative design , sin ce the cryogenic fluid may be partia ll y liquid and
partially vapor.
Sea l d rains fo r li quid hydrogen are routed to a safe disposal area because of the hazard of
mixing hydrogen with atmospheli c ai r. Lea king hydrogen is not all owed to accumu late
inside of th e vehic le stru ct ure. The general practice on liquid-hy drogen turbopump s is to
eliminate ex te rnal drain s and all ow the seal leakage to vent into the turbin e area (fig. 3) .
2.1.7 Fluid Separation

Keep ing in co mpatibl e fluids separated generally is a critica l requ irem ent of oxid ize r ea l
systems. The potentia l ex plosion haza rd of mixing highly reactive oxid izers and fue ls
dic tates the use of a positive system to maintain safe separation . Separation of in compatible
flu ids on the ame sha ft no 1111 all y is main tain ed by uti lizing ( I ) two face-con ta ct seal to
minimi ze th e leakage. (2) separate drains fo r each p ropell ant to vent the lea kage to a afe
di posa l area, and (3) a purged double circLlmfer ntial seal (fig. 6) , two intenn ediate
face-co ntact seal (fig. 10(a)), o r two in te rm ed iate ring seals (fig. I O(b)) with a purge
betwee n them to separate th e drain cavities. The purge pressure must be high enough to
prov id e a pres ure barrier between the drains that will prevent prop ell ant m ixing ca used by
leakage through the intermediate seal. Seal ystems that require a purge for fail -safe
operation must utili ze a fa il-proof purge system .
Separation is improved by utili z ing a rotating slin ge r (fig. 6) or shou ld ers on the
inte rmediate mating ring (fig. 7) to preve nt direct im pingement of hi gh-ve locity fluid and to
ass ist in routing th e leakage out through the drain. Sling rs are more effective with viscous
14
L_ _
fluid s. Fluids th at are subjec t to capillary action and th erefore tend t o leak into the purge
cavity are prevented from contacting th e interm ediate seal by use of an ad equ ate reservoir
with a gravity drain .
2.1 .8 Fail-Safe Provisions

Fail-safe o perati o n is provid ed fo r by designing th e seal sys tem to allow failure o f a single
seal without ca using des tru ctive failure o f th e turb opump o r aborting a flight mission . The
effect of a seal failure is minimi zed by utili zing lab y rinth seals upstream of th e primary seal
to redu ce the prop ellant lea kage and by providing ad equ at e drainage to safely disp ose of the
leakage . Systems fo r separa ting incomp atible fluid s mu st utilize a positive intelm edi ate sea l
to main tain separati o n of the drain cavities (fig. 6) . T o ensure a press ure barrier be tw een the
drain cavities, th e interm edia te seal cavit y norm all y is pressuIi zed to a level high er th a n th e
maximum val ue anticipa ted in th e seal drain. R edund ant seals in series (fig. 4 ) are also
utilized t o prevent excess ive lea kage in th e event o f a seal failure.
A theo retical fa ilure analysis i perfOlm ed to es tim a te all of th e op erating p aram e te rs and
the p ossible results for each di ffe rent failure conditi on. All diffe rent mod es o f operati o n and
variations of perform an ce a re co n id ered.
2.1.9 Purge Requ irements

To prevent co ndensatio n and freezing of moisture, seal cavities in c ryoge ni c sys tems a rc
pu rged with gaseous nitroge n o r gaseou s h elium to rem ove trapp ed air pri o r to chilld owl1 .
Hydrogen systems are purged with gaseous helium to rem ove air th a t co uld ig ni te the
hy drogen. G aseo us nitrogen is no t used in liquid-h ydrogen sys tem s bec3 u e th e syste m
temperature is below the freezing p oint of nitrogen .
Liquid-fluorine system s are th oroughly purged and dried to remove all traces of mo isturc
and thus preclude any combustible reaction of water and fluorin e.
Seal cavities th at are ex p osed to hot gas genera ted by th e co mbu sti o n of oxygc n a IlL I
hydrogen, which co ntain s significa nt free moisture, are purged to rem ove mo isture that m ay
accumula te in the seal cavities and then freeze durin g subsequ ent chilld own s.
Purged cavities are provid ed with an inlet and an outlet p ort to all ow th e purge gas to fl ow
through the cavity. The purge flowrate and length o f tim e are es tablish d to en un:
compl e te removal o f air and moisture. After tes tin g, th e purge is kept on o r air is preve nted
from re-entering the cavity until th e h ardware re turn to ambi ent temp era ture.
J
15
- - - - - - - --
2.2 SEAL ASSEMBLY

The seal assembly consists of a group of detail part or a unitized as embly - including
sealing surfaces, provlslons for initial loading, and a secondary sealing
mechanism - necessary for accomplishing a complete single sealing function. The d ign
problems and performance considerations related to the selection of the seal as emb ly and
its application to the turbopump are presented in the following discussion.
Face-contact seals are used to minimize leakage at pressures up to approximately 500 p i

and speeds up to approximately 500 ftjsec; circumferential clearance seal are used for
higher pressures and speeds. The low temperatures of the cryogenic propellants require seal
types constructed of materials that maintain adequate ductility as the temp rature
decreases. High surface speeds or long-life application require low face-contact loads or
clearance-type seals. Circumferential clearance seals, particularly labYlinth seals, are used for
maximum reliability when the increased leakage is acceptable. The hydrostatic and
hydrodynamic seals are considered for effective sealing at high-speed , high-pressu re, and
long-life conditions.
A comparison of the advantages and disadvantages of current types of turbopump

rotating-shaft seals is given in table II , which appears on pages 18 through 21; basic featu re
of the seal types are shown in figures 11 through 25 that follow table ll.
2.2.1 Pressure Capability

Pressure levels for primary shaft seals nom1ally are controlled to values less than 500 p ig to
allow the use of minimum-leakage face-contact seals. Pressure levels greater than 500 psig
are reduced by the seal system design methods described in section 2.1.1. Pre ures up to
approximately 1000 psig are feasible with turbopump-type face-contact seals when the
ntbbing speed is low, life requirement is short, size is small, and cooling and lubrication are
adequate.
The fluid pressures in various seal applications in representative turbopumps are hown in
figure 26~~' The maximum pressure for current face-contact seals is 500 psig ; in most
applications, the pressure is approxin1ately 200 psig.
The pressure limit for ntbbing-contact-type seals nom1ally is controlled by the capa bility for
balancing the differential pressure forces on the seal face to maintain a reasonable contact
load . The seal usually can be design ed to withstand the structural loads caused by high
pressure; however, it is not feasible to eliminate the increase in face-contact load a th fluid
pressure increases. Therefore, the face-contact load increases in proportion to the fluid
pressure increase, and the pressure limit is detelmined by the load-speed-life relationship
for specific materials and fluids. Because the face load is proportional to the fluid pressure,
*Figure 26 appears on p. 31.
16
th e re lat io n of fluid p re sure a nd rubbin g speed ro r spec iri c ty p es of sea l an d flu id s prov id es
a guicl e fo r eva lu a tin g th e relati ve seve rit y o f th e a ppli ca ti o n befo re the d e tail cal d esig ns
are es tab li sh ed .
T h e effec t o r hi gh pre sure o n face-co nt ac t be ll ows-t y pe sea ls is m o re p ro no un ce d th a n th e

effect o n elas to me r o r pis to n-rin g ty pes beca u e of th e ch a nge o f be ll ows effect ive di ameter
with pressure . Th e bell o ws ea l can be pre lire b ala nce d fo r a p ec ific p ress ure ra nge;
h o weve r, if th e pre sure va ri a ti o n is grea t e no ugh to ca use th e bell o ws efrec ti ve di am e te r to
ch a nge igni fica ntl y, t h e sea l press ure bala nce will be adv e rsely affec ted . Hi gh e r p ress ure w ill
inc rease th e balan ce ra ti o a nd ca u e ove rl oadin g o f th e ea l face . Lowc r pressure w ill red uce
th e bala nce ra tio a nd th e face-contac t load and m ay re ul t in excessive leakage o r face
e pa ra ti o n. The re fo re, bell o ws sea ls ge ne rally are no t used in appli ca ti o ns w ith la rge pres ure
va ri a ti o ns unless face-load co ntro l i n o t criti cal. Turb opump b ell ows sea ls h ave ope ra t ed
su cces full y with press ure levels o f 500 p ig and press ure variati o n of a p p rox im ately 100
psi. Bell o ws-ty pc sea ls th a t a re d e igned fo r high press ure ge ne rall y in co rpo ra te h eavy pl a tes
with a narrow sp a n to w ith sta nd th e press ure-indu ced fo rces; the refo re , th e be ll o ws spring
ra te is high er, a nd th e ax ia l trave l may be limited unle ss additi on al sp ace is ava il a bl e to all o w
an in crease in t he numb e r of co nvo lut ions.
Circ um fe re n tial sh aft-riding segm e nted ca rb o n sea l (fig. 20) no rm all y a re not u ed as
dry-gas eals in turbo pumps ab ove approx im a te ly 100 psig beca use o r th di ffic ulty of
bala ncin g th e radi al pres ure loa d. High e r pres ure (> 100 pig) m ay res ult in high co ntac t
loads and exces ive wea r o r heat ge nera ti o n. The c irc umfere nti a l segm e nted sea l ca n be used
at high e r press ures with liquid s o r lubricant beca use of th e hydrod y nami c lift e ffec t from
the flui d we dge th at i d eve lo p ed a t the co ntac t surface (similar to th e effec t in a j o u rna l
bea rin g): ho weve r, th e lea kage ra te is signi fica ntl y hi gh e r with liquid s beca use th e segm nts
te nd to wind up an d li ft away fro m th e sh aft surface. Recent ex pe rim e nts indi cate th at it
m ay be feas ibl e to utili ze a h y dro d y nami c co ncep t ( Ray leigh s tep o r spiral g roove (fig. 22))
a t th e co ntact surface to control th e fluid-rilm thi ck ne .
Circ um fe re ntia l sh aft-ridin g flo atin g-rin g co ntrolled-gap seal (fig. 23) a rc ca p ab le of
op era tin g a t very high pre sures (> 1000 psig), beca use the co ntac t-l oa d inc rease w ith
pressure is n egligibl e. The sea l supp o rt the diffe re nti al pressure radi al load as co mpress ive
ho op stress in the fl oa ting rin g ' h oweve r, th e ax ial co ntac t lo ad in c rease w ith pressure, a nd
the radial fo rce req uired to rep osi ti on the n oa ti ng rin g, if th e sh aft ro ta tes ecce n tri call y .
in creases sligh t ly . Th e pressure n 0 n11 all y i limited o nl y by th e cap a bilit y of th e tl oa tin g rin g
to supp o rt th e radial load stru c turall y. La rge pres ure va ri a ti o ns m ay ca use exee sive ch a nges
in cl earan ce ga p and ca n res ult in failure beca use o f ring seizure on th e sh aft.
Th e h y drost a ti c and h ydrod y nami c eals (figs. 15 thro ugh 18) th eo re ti ca ll y a rc ca p able o f
o p e ra ting at pressures high e r th a n 500 ps ig, sin ce th e diffe re nti al press ure load is supp o rt ed
b y a fluid film : h o weve r, these sea ls are ve ry se nsi tive to disto r t io ns 0 f th e in te r race .
17
Tab le II. - Advan tages and Disadvan tages of Various Types of Rota tin g-Shaft Seals in Current Use
Seal type Figurc Advantages Disad van tages
!-ace-contact metal bellows II Low leakage across face ; positive seco ndary sea l ; wide High cos t ; welding is cri tical; stress levels and
tempera ture range (423° to IOOO° F) ; materials fatigue life difficult to predict ; low fat igue life
compatib le with toxic and reactive propell ants; sim- in ex treme vibration and oscillating pressure
plified cons tru ction ; bellows provides spring load environments; variation of effect ive pressure-
and secondary seal ; little space required; materials balance diameter and resu l tan t face load.
not age-limited ; high reliabilit y.
Face-co nwcl plastic lip 12 Low-tempera ture (-320°F) capability ; lip provides Lip stress level and fatigue life difficult to pre-
vibrati on damping; one-piece nose fo r minimum face diet ; lip subj ect to damage and wear by con-
distortion ; co nstant pressure-balance diameter ; re- taminatio n particles; reti ability lower th an bel-
sistant to ext reme vibration enviro nm ents. lows in most applicati ons; lip drag at high pres-
sure.
Face-contact pisto n ring 13 Wid e tempera ture range (423° to 1000°F); material High leakage through seco ndary seal; wear of
00
co mpatib le with toxic and reac tive propellants; ma - piston ring and m ating surface can cause seal
terials not age-timi ted ; cons tan t pressure-balance hangup , excessive drag, and in creased leakage;
diameter for close face-load co ntrol; resistant to ex- piston ring subject to fr etting damage.
treme oscillating pressure environm ents; high-
pressure ca pability.
Face-co ntact elas tom er 14 Low leakage through secondary; low cost; resistant Limited temperature range (-65 ° to 500° F) ; ma-
to extreme vibration and oscillating pressure en- terials age-limited; requires lubrication for con-
vironment s; elastomer pr ovid es vibration damping; sistent performan ce; lub rica nt may deteri orate or
design highly developed and standardized . be washed off; subject to hangup or inadequ ate
response due to elasto mer extrusi on at high pres-
sure; materials and lubricant not compatible with
some propellants.
~------ --~.------
~ ---- ~ -->--
S~ I I-cIH: r[!iLcd hydrostatic l5 Seal face supp orted by fluid film from sealed High leakage ; hlgh cost; rubbing contact dUIing
rac~ pressure; no rubbing contact during steady-state transient operation; sensitive to face or mating-
operation ; no external pressurization; no dilu- ring distortion ; sensitive to rruxed liq uid and vapor
tion of propellant with purge gas; long wear life ; fluids ; spring load and pressure balance critical;
high-speed and high-p ressure capability; lift in- marginal stability; lift dependent on pressure dif-
dependent of speed . ferential.
1· :-. tern ally pressurized hy- l6 Seal face supported by fluid film from externally Requires external inert pressurizing system or
drosta ti c face pressurized so urce; no rubbing con tact during hlgher pressUIe propellant source; high cost; dilu-
transient or stead y-s tate operation ; long wear life; tion of propeUan t with pressurizing fluid; spring
high -speed and high-pressure capability; low propel- load and pressure balance Critical; marginal sta-
lant leakage; lift ind ependent of speed and sealed bility in cryogenic fluids ; relatively large space
pressure. required.
Hydrod ynami c face l7 Seal face supported by fluid ftlm developed from Rubbing contact during startup and shutdown or
hydrodynamic lift forces due to rotation ; no low-speed transients; face geometry easily damaged
\0 rubbing con tact during steady-state operation; max- by rubbing contact; sensitive to face or mating-ring
imum lift force developed at minimum face clear- distortion; sensitive to mixed liquid and vapor
ance; long wear life; high-speed and high-pressure fluids; marginal lift force with cryogenic propeUants;
capability ; lift independent of sealed pressure ; no sensitive to abrasive environments.
dilution of propeUant with purge gas.
Hybrid face l8(a) Hydrostatic action provides lift at low rotational Sensitive to face or mating-ring distortion; sensitive
speeds ; hydIOdynamic action pIOvides additional to mixed liquid and vapor fluids ; spring load and
lift and stability at high speed independent of pressure balance critical; high static leakage; toler-
sealed fluid pressure; no rubbing contact during ances on face geometry critical; face geometry easily
tran sient or steady-state operation; long wear life ; dam aged b y rubbing contact; relatively large space
high-speed and high-pressUIe capability . required.
Hybrid face l8(b) Design allows for rubbing con tact during transient Rubbing contact may cause thermal distortions that
operation ; face contact load reduced at high speed destIOY hydrodynamic lift; face geometry designed to
by hydrodynamic lift force ; lift independent of allow for reasonable wear may not provide significant
sealed pressUIe; longer wear life and higher speed hydrodynamic lift ; face material must be compatible
and pressure capability than rubbing contact seal. with the face unit load and velocity rubbing limits;
amount of hydrodynamic lift force is difficult to pre-
dict.
(continued)
Tab le II. - Advan tages and Disadvan l agc~ or Various Types of Rota lin g-Shu ft Sea ls in Curren I Use (com: 1uded)
Sea l ty pe Figu rr Adva nta ges Disadva ntages
Face contact/clea rance 19 Seal face travel limited to provide minimum face Lea kage d ~ p c ndent o n rc'pl'atabilit y o f th.: mating
clearance afte r initial wear-in at a (h ed position ; rin g position; rubbing surfaces mu) be damaged
no additional face w.:ar after initia l wear-in; high- dllring wear-in: wear Dr over heatin g m ay occ ur if
speed a nd high-pressure ca pabilit y a ft er wear-in; the mating ring p ositi on b not rl'pl'atab le; c\-
simplifi ed co nstru cti o n : high re liab ilit y for stcady- ccss ive leakage or facl' overload ma)' occ ur during
state re peatable o perati o n. star tup and shutd o\\ 11 transil:nt s.
Circum feren tial shaft-riding 20 Lower leakage than other ty pes o f cir c umkrential Highe r lea kage tha n facc-contact-type sca ls: leakagl'
seg mented shaft seals; segments adjust to d iameter a nd radial o f liquid s ma y be c~cessive bccause o f h ydrod y-
locatio n variations to ma int a in rubbing co ntac t ; nami c lift of thc scg m c l1h : lim ited to low press ure
unlimited axia l trave l ; less scnsitive to temperature beca usc of unbalan cl'd prcs,urc' loa d : fri c ti o na l torq uc
differe n tia ls; wid e te mp e raturc ra ngc (-423° t o and hllrsepo,,"l'r Ie,,\\.', ,liT hi)! h co mpared \~ith baJ -
tv 1000° F) ; materia ls no t age limited ; little axia l a nc ed fac·c' seal: (OIl1Pie'\ alld l"f lli ca l co nstru c ti o n ;
o space; d esign h igh ly deve loped. rl' la !t \~ I ~ hi)!h cO\t
Circumferential sh aft-rid ing 21 Lower leakage than flo a ting-ring seal ; segml'nh arc IJ lg her kakagc' th :tlll"llllvc'n!tona l segmen ted sea l:
arch -bound segmented arch-bound at o pera ting diameter to support pres- diffi c ult to prcdic-t c'\act c)pcra ting di a me ter: oprr-
sure load in compressive hoop stress; 10\\ (lln tact ating diam c'tl' r Illust be rl' peatablc : ove rh ea tin)! llr
load for long life, high-speed and high-p rc'ssurl' capa- da nla)!l' nla~ llCl"Ilf dUfln!! \\ l'ar-in ; ma chining. to ler-
bility ; w ide temperature ra nge (-423° to I OOllo l J. ane.:, murc cr it ic-a !. c·llmpk\ a nd c riti ca l co nstru ct ion.
rda li ve ly high co,!.
Cir cumferential h yd rody- 22 Sea l segments supported by nuid film develo ped Jl ig hcr Icak<I)!l' than shaft-fldin g sC)!ITI c nt ed seal ; rubbing
nam ic segmented from hy d rodynami c lift forces du l' to ro ta ti o n ; co nta c t du ring tranSient upcra ti u n : co nform a tion of
h igh er speed a nd pressure capab illt ) a nd long.cr life seg m cn ts tllmating rin!! ,urfner cr iti ca l; to lerances o n
than sha ft-riding segmented sea l: lowe r lcaka)1c than sep. l1l l'n t facc eril lcal: face geo m e tr y easily damaged b y
noating-ring seal: segml' nt ' adjust to variations In rubbing con tact: hi/!h CllS!.
dia m ete r a nd radiali llcallu n .
~.-~--- ~ .- - - - - - -- ----~-----__v_-.- ---.-----
Circumferential floating- 23 Unlimited speed capability; high-pressure capa- Higher leakage than segmented shaft seal; clear-
ring controlled~ap bility; negligible horsepower losses; negligible wear; ance gap sensitive to temperature differential
unlimited axial travel; wide temperature range between ring and shaft; clearance gap may be de-
(-423° to lOOO°F); low clearance gap; simple con- creased by radial deflection of the ring caused by
struction; less axial space; high reliability ; effective differential pressure load ; subject to fretting dam-
for liquids , gases, and mixtures. age.
IV Circumferential clearance 24 Consistent predictable leakage ; maximum reliability; High leakage; clearance gap must be large enough
labyrinth unlimited speed, pressure, life, and axial-travel to allow for mislocation tolerances and shaft
capability; minimum horsepower losses; maximum radial movements.
temperature range; no rubbing; materials not critical;
simple construction; low cost; wide temperature
range (-423° to 1800°F).
Circumferential wear-in 25 Lower leakage than clearance labyrinth; high relia- High leakage; torque may be high during wear-in;
labyrinth bility ; unlimited speed, pressure, and life ; materials transient temperature differentials may cause high
allow wear-in for minimum clearance gap. drag torque after wear-in ; damage may occur during
wear-in ; wear-in materials expensive.
SEALING RING (CARBONI
EPOXY -BONDED TO CARR I ER
SEALING RING (CARBONI
EPOXY-BONDED TO CARRIER
RETAINER WASHER CARRIER
VIBRATION
DAMPER
WAVE SPRING
VIBRATION
DAMPER
BELLOWS
RETAINER RING
RETAINER RING
CARRIER ASSEMBLY
(al LlQUIO-OXYGEN SEAL (bl RP-' SEAL
Fi gure 11_ - Face-contact m eta l-b el lows seal.
SPRING RETAINER
COMPOSITE
LIP RETAINER
UPSTREAM CUP (HOUSINGI

PRESSURE SIDE
DOWNSTREAM
DRAIN SIDE
SEAL RING PILOT

SEAL NOSE
SEAL RING SHELL (COMPOSITEI
Figure 12. - Face-contact plastic-lip seal.
22
LOW PRESSURE
HIGH
PRESSURE
PISTON RING SECONDARY SEAL
CARBON INSERT SEAL NOSE
Figu re 13. - Face-con tact piston-ring seal.
SEAL RING (CARBON) SEAL RING (CARBON)
ANTI ROTATION
GROOVES (2) ANTI ROTATION
GROOVES (2)
SPRING
SPRING
CUP
CUP
V-PACKING ELASTOMER
O-RING (SYNTHETIC
RUBBER)
EXPANDER RING
SHELL
(a) V-PACKING ELASTOMER (b) O-RING ELASTOMER
Figure 14. - Face-contact e lastomer sea l.
23
EQUILI BRI UM
FACE VIEW 10) RECESSED PADS W)TH ORIFICE COMPENSATION
RECESSED STEP
EQUILI BRI UM
UPPER LIMIT
Ib) RECESSED STEP
TAPERED RECESS LOWER LIMIT
HIGH PRESSURE
UPPER LI MIT
EQUILIBRIUM
Ie) CONVERGENT TAPERED FACE
Figure 15. - Three configurations for a self-energized hydrostatic face seal.
24
EQUILIBRIUM
LIMIT r SPR I NG
0/000000000
HIGH PRESSURE
RECESSED PAD HOUSING

LOWER LIM IT
EXTE RNAL PRESSURIZING SOURCE
, . ) PISTON -RING SECON DARY ElEMENT
EXTERNAL PRES SUR I Z ING SOURCE
LI MIT
LOW
LOWER LIM I T
EQU ILI BR I UM
HOUS I NG I. D.
Ibl BE LLOWS SECONDAR Y ElEMENT
Figure 16. - Two configurations for an externally pressurized hydrostatic face sea l.
25
A - MATING RING ROTATION
SEAL DAM
SECfION A-A SHROUDED RAYLEIGH STEP LIFT PAD
MATING RING ROT,l,TION

LIFT PAD
SHROUDED RA YL EIGH
STEP LIFT PAD
HI GH PRESSURE VLh~----r---"r
PRESSURE
RECESSED AREA SECTION B-B SEA L FACE

(a) SHROUDED RAYLEI GH STEP
,----------;-RECESSED SPIRAL GROOVES HIGH PRESSURE
HIGH
PRESSURE
RECESSED
SPIRAL
SEAL DAM GROOVES
SEAL DAM
SEAL FACE SECT I ON
(b) SPIRAL GROOVES
, - - - - . , . - - R ECESSED CIRCULAR GROOVES

HIGH HI GH PRESSURE
PRESSURE
SEAL DAM
RECESSED CIRCULAR GROOVE
SEAL DAM
SEAL FACE SECTION
(e) CI RC ULAR GROOVES
~-t------;>~REC ESSED ANNULAR GROOVES HIGH PRESSURE
' - - - - f - SEAL DAM LIFT PAD

ANNULAR GROOVE
SEAL DAM
SEAL FACE SECTION
(dl ANNULAR GROOVES
Figure 17 . - Four configurations for a hydrodynamic face sea l.
26
SPI RAL
HIGH PRESSURE
REC ESSED SPI RAL

GROOVES
SEAL DAM
CO NVERGENT
CURVED FACE
MATING RING CONTOUR
ROTATION
SEAL DAM
SEAL FACE SECTION
l al COMBINA TION SELF- ENERGIZED

HYDROSTATI C ICO NVERGENT CURVED
FACE) AND HYDRODYN AMI C IS PI RAL - GROOVEI
RECESSED AR EA
SU PPORT PAD S
- M A T I N G RI NG ROT ATI ON
A
SECTI ON A- A TAPERED LIFT PAD
TAPERED LIFT PADS
SUPPORT PAD
TAPERED LIFT PAD SEAL DAM
HIGH
PRESSURE
SEC TION B- B SE AL FACE
I b) COM BINATION RUBBING CON TA CT AND

HYDRODYNAMIC ( TAPERED LIFT PAD)
Figure 18. - T wo configurations for a h y brid face seal.
27
l
SEAL FACE STOP ( AFTER WEAR - IN )
PISTON RING
SECONDARY SEAL
CARBON INSERT SEAL NOSE
Figure 19. - Face contact/ clearance seal.
CARBON RING
CARBON RING ASS~BLY
DOWEL PIN
HOUS ING
SNAP RING
o 0 0
SNAP RING (WAVE SPRING AND WASHER
REMOVED TO SHOW CARBON RING)
CARBON RING ASSEMBLY
F igure 20. - Circu mferen ti a l shaft -riding segm e nted seal.
28
L_~~_
_ _
GARTER SPRING
WASHER.
PIN
ARCH-BOUND
SEGMENTS
Figure 21 . - Circumferen tial shaft-riding arch-bound
segmented seal.
SHROUDED
RAYLIEGH STEP
PAD ( TYP ) - - - ,
GARTER SPRING
PIN
Fi gure 22 . - Circumferential hydr odynami c segm ented sea l.
29
SEALING RING
(CARBON)
RETAINER
RING TAB
RETAINER RING
ClrCLJmf e . OUS ING
F igure 23 . _ ' -n. ng controlledH-gap seal.
rentl al float Ing
C il'c umfere ntial c learance lab yn. nth seal.
ABRADABLE
MATERiAl
ROTOR
Fi gure 25. - CircLJ m f erential wear -In lab yrlnth sea l
30
-- ---
i
FACE CONTACT
FACE HYDRO D YNAMIC HYBRID (ANNULAR GROOVED FACE)
800 800
500 500
C1 r-
en
Cl.
400 400
I
uj
a: r-
::J
en
en
w lOO 300
a:
I I
Cl.
0 r-- r-
:; - - r-
...J
LL 100 200
I I
W
100
- r- r-
- ~~
~
- 100
o
cc
~~NKl~~~~'73
0:: ~
I 0
IT
~ ~ ~
0
'"
0(
n ~
z
0(
.... N lC ~-"7
000
:3II: 3cc 3a:
rT
GO
~ ~ ~
00
--
n ~~
~
~~ I~II
<I)
~
~
~
'7 ~
z
0(
o
~~iJ.J.
N
t-X~2.,,,.t--<a::~1I: '" '" II: ::I i= .... .!. ::I :i ::I ::I
II: II: .... ~ ~ i N N
........ ::I ....
LOXIRP-l LOXlLH 2
LIQUID OXYGEN
LIQUID
FLUORINE
0@ LIQUID HYDROGEN GH 2 RP-l
HOT GAS
(j) FLOX 130% LF2 + 70% LOX)

TURBOPUMP
@ N:z04 FLUID
Figure 26. - Seal fluid pressures in representative turbopumps.

2.2.2 Temperature Capability
Th ' use o f c ryoge ni c tluid s limits the ch o ice o f sea l mate ri als beca use th e lo w temp era tures
indu ce brittle ness o r los of du ctility: frequ e ntl y temp era ture i th e d e te rminin g fac to r in
th e elec ti o n o f th e seal ty pe. The m o t ucce sful ty pe of seal fo r c ryoge ni c h as bee n the
face-co ntact we ld ed bell o w (fig. 11 ) . F ace-co nt ac t seals with a pi a tic (Kel-F o r Myla r) lip
eco nd ary seal (fig. 12) also h ave bee n used succe full y; h oweve r, th e ir reli abilit y ge ne rally
is lowe r. Me tal-pis to n-rin g seco nd ary sea ls ( fi g. 13) are ca pable o f opera ti o n at lo w
temp e ra ture: howeve r, th e grea te r lea kage a nd drag of th e pisto n rin g m ay n o t be
acce ptable. Sh aft-ridin g segme nted ca rb o n seal (fig. 20) al 0 h ave been u ed to seal
cry ogenic liquids, bu t th e lea kage is signi fica ntly high r th an th at of face-co ntac t seals, and
th e segm e nted ea l a re limited to lo we r press ure beca use o f th e load indu ced by unb Rlanced
pre sure.
High temp erature of th e st'aled fluid n 01111 all y i no t a limiting fac to r in the selec tion of
turb o pump seal , excep t a it i rela ted to the cooling cap acity of th e seal rubbing faces .
Turb opump sea ls ge ne rall y depend on th e sealed fluid to dissip a te th e hea t gen era t d at th e
eal face; th ere fo re, a the te mp e rature of the fluid in creases the coo lin g ca pacity decreases.
High temp e ra tures also lim it th e cho ice of eaJ m ate rial beca u e of tru ctural effec ts,
the rm a l degrada tion o f the ma te ri a l pro pertie , and oxid ation. Th e m ax imum temp erature
of th e turbin e ho t ga in the seal area usu all y i Ie s than 1000° F, a temp erature th a t allow
th e u e of co nventio nal h o t-gas sea ls (e .g. , egm ented-carbon , fl oa ting-gap , lab y rinth ,
we ld ed-b ell o ws, a nd pi to n-rin g typ es) if the h ea t generated a t th e eal interface can be
di sip ated sa ti fac to ril y . Me tal-b ell o ws face seal with ca rb on inse rts requi re u ffi cien t
inte rfe rence at th e in se rt 00 to e nsure re tenti o n at high te mp era ture (sec. 2. 3 .2. 5) .
Severe tem pe ra tu re gra dien ts through th e seal sys tem and ac ros th e seal ass em bly occ ur
whe n the c ryogenic fluid s and hot gases are sealed along th e sa m e h aft in adj acent a rea.
Addition al temp erature gradients are cau ed b y th e h ea t genera ted a t the sea l inte rface.
Therm al tr s es cau ed by diffe rential ex pansi on o r contrac tion a re m inimi zed by LI ing
ma terials with similar ex panion/contrac ti o n ra te and by all owing th e de fl ec ti on s to occ ur
with a minimum am o unt of restraint. The therm al de fl ec tions are all o wed fo r by m aking
initial clearances su ffic i nt to preve nt excessive th e1111 al interfe ren ces and by adju stin g th e
al dim e nsio n to com pe nsat fo r th e th erm al de fl ec tio ns.
2.2.3 Speed Limitations

The surface peed o f a rubbin g-co ntac t ty pe o f sea l norm all y is lim ited by th e h ea t
ge nera tion at th e sea l in te rface a nd th e coo l ing ca paci t y of th e inter face m a teri al a nd
surro unding enviro nm e nt. The plim a ry so urce o f hea t ge nera tion is th e fri cti o n to rqu e at
the seal interface ca used by the contac t lo ad and fric ti o n of th e inte rface rubbing m ate li als.
Addition al h ea t is ge ne ra ted by visco u h ea r of th e scaled fluid : h o weve r. thi s h ea t nOll11 all y
32 I
L J
is not significant on turbopump rubbing-contact seal, which usually operate with boundary
lubrication because of the low viscosity of cryogenic fluids. Rubbing-contact seal running
with more viscou fluids (RP-l fuel, lub e oil) generally operate with hydrodynamic
lubri cation; the heat generation due to viscous shear therefore becomes significant and is
considered.
The noncontact hydrodynamic and hydro tatic seals norm all y are not limited by surface
speed, because th eal face a nd rotating mating surface are separated by a fluid fi lm ;
however, the heat generated by vi cous shear must be considered , since it may limit the
peed becau e of the ensitiv it y of the seal interface to thelmal distortion. The higher
leakage rat of this type of eal n0ll11ally provides adequate cooling capacity , particularly
with cryogen ic fluids.
The rate of heat dis ipation from the eal interface detell11ines the re ultant interface
temperature and i equa l in importance t o the h eat generation rate in establishing the peed
limi t. The analytical method used to calculate the temperature profile and thennal
di stortion in the sea l face an d mating ring a re given in references 6 an d 7 .
The maximum allowab le surface speed n 0ll11all y is esta bli sh ed by the temperature limi t at
the eal interface required to prevent thelmal fa ilure of the interface materials or structura l
fai lure of the sea l components and mating ring. In some cases, the limit m ay be e tab lished
by the temperature required to prevent vaporization or thermal d ecomposition of the sea led
fluid.
The structural load on high - peed seals cau ed by centrifugal force also are con id ered,
becal! e dynamic deflections of the mating-ring surface can adversely affect th e interface
pres ure profil e. The centrifu gal tre e ma y be ignifi cant on high-speed mating rings as a
result of the decrea e in y ie ld st rength of ome m ateria ls at high temperature. Th e the lmal
load due to th e rubbin g-surface temperature gradi ent m ay also cause di stortions of the seal
interface. Failure will occur if the temperature gra di ent ca us s interface di tortions and
re ults in hi gher contact loads that compound the h ea t generation problem.
The value for rubbing speeds in various seal appli cations in rep re e ntative turbopump are
hown in figu re '27. The highe t speeds (up to 450 ft / sec) are required on liquid and gaseous
hydrogen and hot-gas app li catio ns. With liquid oxygen and RP-l , rubbing speeds usual ly are
lower than :200 ft /sec.
The heat-generation rate on rubbin g-contac t sea ls is a direct function of the re ultant
contact load (sprin g load plus pres ure clos ing force minu pre sure open ing force),
coellic ient of friction of the rubbing material, and the velocity of the rubbing surfaces, as
hown ill the following equ at ion:
33
1 ------ .- .---.----.
I
FACE CONTACT
FACE HYDRODYNAMIC HYBRID (AN NULAR G ROOVED FACE)
SHAFT · RIDING SEGMENTED
SHAFT · RIDING FLOATING RING
500 !iOD
r- ~ '""" ~ ~
- I
04;00 .4QC)
~ ~~I ~il
u
w
(/)
;::
LL'
I -
300 300
~ -~ ~ ~rll~
6
w
w
c..
(/)
W
.j:>.
l?
Z
en
en
:::J
200
- ~ ~~
- ~~~Ir-~I
~~ ~
200
I
a: ,-_ r- r ,- _
100 100
r~ - ~ - ~ ~ II~III
o o
a: a:!1J 0 0 ID~ 000 00 ~a: ~ Z
o 0':'0"'"
V') 00'":' 'L- V"I(J) 0)--.)0 .) <i(
I"7~t:"~:r""'o.lI"7~ ooIJralI..... <i(!:: N~~~r; .j~~ l.l..J~ ("of I-X"'7"7"7 "7~"7NN'7t:
I-x...., ...... u..I-<I(o::::Ea: 0::0:: a: ::EI- ~ ..... ~~::e a:a:a: ::ea::a::: -', <i(I-Xu..u.. u....c;x-',~::II-
LlOUID OXYGEN FC0~~II~E CD ~ LIOUID HYDROGEN GH 2 RP-1 LOXlRP·1 LDXlLH2 I

OXIDIZERS FUELS HOT GAS I
G) FLaX 130% LF2+ 70% LOXI

TURBOPUMP
CV N 20 4
FLUID
Figure 27. - V alues for sea l rubbing speeds in repres entativ e turbopumps.
. _(F Y) f
q- - - (I)
J
wherL'
(1 = hea t generation rate, Btu /sec
F = res ultant total co ntact load, Ibf
Y = su rfa ce speed. rtf e c
r = coeffi cient of fri ction

J = 777 .6 ft-lbf/ Btu
The eq uati o n ca n also be ex pressed in terms of hea t-gene ra ti on rate p r unit area a nd
co ntact unit loa d :
• I _ (PY) f
q - -- (2)
J
where
el' = heat ge ncrat io n rate per unit conta ct area, Btu / sec-in 2
P = co nta ct unit load, psi
Th L' co nta ct load is a function of th e sea l sprin g loa d , press ure closin g fo rce, inte rfa ce
pressure profile. and hydrodyn ami c li ft. Th e spring loa d and pressure clos ing fo rce are
known va lu es. Th e e fl"c ctiv e inte rface pres ure profile ca n va ry from app rox im ate ly 0. 2 to
0.8 times tile pressu re differential. The hydrodynami c lift can vary from ze ro to a va lu e
sulTicient to maintain Iluid-film race se para ti o n. Th e coeffi cient of fr ic tion i related to the
lu bri ca ti o n and hyd rody nami c li ft: it ca n va ry from appro ximately 0.05 to 0.4 for th e
cOlTI mo nl y used sea l materi als.
Because th e heat ge ne ration i proportional to th e produ ct of face load an d velocity , and

beca u e th e speed limit n0 l111 all y is cs tab lished by th e inte rface temp eratu re lim it , it is
co nv eni ent to cs tabli h a face load-v eloc it y relat ion to provid e a design guid e fo r es tablishing
th e face load and urface peed limi t . Thi simpl e relation d oes not account for all of th e
va ri ab les ; th erefo re, a sepa rate rela ti o n mu t be e tablish ed to all ow a basis of comparison
fo r eac h sea l type, face materi al co mbination , and sealed fluid. If th e coeffi cient of friction
and the hea t-tran fer ca pacity of th e seal materials and surrounding e nvironm ent are imi lar,
th e load-velocity relat io n ca n be utili zed with reasona ble acc uracy.
35
Load-velocity relations have been established for turbopump seals on the basis of unit face
load (PY factor) and face load per unit circumferential length (FY factor). Th FY factor is
proportional to the total rate of heat genera tion and is independent of face area. The PY
factor is proportional to heat generation per unit area and i a be tter indication of
heat-transfer capacity. The FY relation is useful for main taining practical limits and al 0 is
convenient for relating the face load requirements to other de ign parameters (spring load ,
secondary seal friction, in ertia forces, pressure forces) .
The values for FY (lbf/ in. X ft/sec) and PY (lbf/in.2 X ft/sec) factors for face-contact seals
in various applications in representative turbopumps are shown in figures 28 and 29.
2.2.4 Wear Life i

The turbopump-seal life requirements in operational engines are low (~ 2 hr) in com pari on
I
with life requirements in the SSME 00 hr) and in most other applications (500 to 10000
hr). Therefore, relatively high seal wear rates can be accepted and may be the only feasible
compromise for some of the more severe applications. The average wear rate of the carbon
seal face on the liquid-oxygen and liquid-hydrogen eals in the J-2 turbopumps ranges from
0.005 to 0.010 in. /hr. The variation in measured wear rates for the arne type of seals may
be as much as ± 100 percent.
Allowance is made for the high wear rate by designing the seal contact face with sufficient
height to wear away while continuing to maintain an effective seal for the required life. The
carbon nose height on mo t turbopump seals is approximately 0.050 in ., which would
provide a minimum wear life of 2.5 hours at a maximum wear rate of 0.020 in ./hr. Usually ,
it is not practical to obtain longer wear life by using greater nose heigh ts, because of the
structural weakness and distortion of the relatively thin cylindrical sections at high
pressures. Also, it may not be feasible to design the seal with sufficient axial travel to allow
for the additional movement required to compensate for greater seal face wear.
Seal wear life of more than 10 hours in liquid oxygen has been demonstrated on a
combination hydrodynamic and rubbing-contact face seal (table 1). The seal face was
grooved to provide additional hydrodynamic lift at high speed to reduce the face contact
load. The hydrodynamic/hydrostatic noncontact seals potentially are capable of much
greater wear life; however, the reliability of hydrodynamic/hydrostatic seals has not been
demonstrated with rocket engine propellants. oncontact types of seals are considered for
turbopump applications that require a wear life greater than approximately 4 hours, except
when the PY factor is low and good lubrication is available.
The wear life of rotating seal faces is highly dep endent on the lubrication available at the
seal interface. Fluids such as RP-l and lube oil, which are capable of developing
hydrodynamic lubrication at the seal interface, make possible relatively long wear life.
36
~XI~ __________________________________________________________________~
u
w
~
t ~ r-----------------T-~r;--------------------------------------~~
x -
z - -
;::::
~ ~ r---~_=r~-------1_+_+_r1_--------_+~--------------------------_+~~
...J .-
a: r-
o -
ti
<t
u..
>
u..
e...J ej
n
a: a:
LO X FLOX AP-l
TURBOPUMP
FLUID
Figure 28. - Values for FV factor for face-contact sea ls in representati ve turbop umps.
-
40
U
W
en
;::
u..
-
X ...-
N ~
Z
;::::
u..
CD
...J r-r-f-.
r-
ri · ~
.--....-
0
I-
U r-
<!
u..
r-
,... r-
>
_r-
l- I- - .- .--
0- 10
- l-
o
II:
0
:I:
i:~Z~
I-
0
::; -
13 j
:I: I-
~ ~ ~i~~~
CIt
000
..........
-
'"" -.:.
0
N
n
~
00
l0
: :: ::;
'"
..
h
..
z
l-
;::
'"::5 13
I- r
± :i: ....
l- II: "- ~ « ~ a: ~ II: ~ .. I-
LOX LH2 GH 2 H2 + H~ LF2 GXY AP-l
TURBOPUMP
FLUID
Figure 29 . - Values for PV factor for face-contact seals in representative turbopumps .
37
I
_J
- - - --I
Cryogenic fluids such as liquid oxygen and liquid hydrogen normally do not develop
hydrodyn amic lubrication or fluid-film interface separation, because of the low viscosity
and change of fluid tate (liquid to vapor) ac ross the seal interface. Therefore, rubbing
contact generally exists, with boundary lubrication or self-lubrication from a film deposited
on th e mating su rface. The wear rate m ay be a function of th e oxide film that forms on
most metals ex posed to oxygen and moisture. Considerably better wear rates are obtained
by electing material co mbin ations that utilize the chemical prop erties of the sealed fluids to
assis t in the f0D11ation of a self-lubricating film at th e seal interface.
It is generally accep ted that seal face wear rate increases with face contact load. The wear
phenomenon generally is explained as caused by adhesion and shearing of the
co ntact-surface microasp eriti es; therefore, higher wear at greater contact load would be
theo retically correct beca use of the larger shearing forces of the additional asperities that
would be in contact as a result of additional surface deform ation. This relation would not
necessarily app ly when hydrodynamic lubricatio n exists ; therefore, m any ex perim ental
results indicate no corr lation of wear and face load within th e ranges evaluated. The general
practice is to comprom ise o ther req uirements and design fo r minimum face load when
longer life is required.
The effect of rubbing speed on wear rate, when conv erted to wear per di tance traveled,
does not ap pea r to be signi ficant, except fo r the effects of the resultant surface-temperature
in crea e. The primary effect of high er speed is to increase the heat-generation rate. The
design o bjec tive fo r long-life eals is to minimize th e PY fac tor.
Most wear and fr iction measurements on carbon m ateri als in cryogenic fluids or dry
co nditions indicate a significan tly higher wear rate and coefficient of friction during the
initi al run-in period. This phenomenon has been observed in turbopump seal t ests and
generall y i ex plained as cau ed by the self-lubrica ting qualities of the carbon film that is
deposited on th e mating urface during th e run-in perio d. The wear rate is high during
f0D11ation of the ca rb o n film and gradually tapers off as the self-lubricating film is formed.
Materials in fubbin g co ntac t that do not form a elf-lubricating film continue to wear at th e
high initial rate.
2.2.5 Leakage
Turbopump seals are selected and designed for the minimum possible lea kage co nsistent
with the required ope rating conditions. Generally, the operating conditions are the fixed
parameters th at control the design, and leakage is th e resulting variable. The most important
co nsideration is reliable operation, and in severe applications, seals designed fo r minimum
lea kage m ay fail prematurely because of overheating. High-sp eed seals may require high
lea kage rates to provide cooling of the seal interface. Provision for th e higher leakages
u uall y can be m ade with proper design of the seal ystem .
38
L
Rubbing-contact seals with low-speed, low-pressure, and short-life requirements may utilize
high spring loads and large pressure-balance ratios to ensure a high face-contact load for
minimum leakage ; however, as the speed 311d life requirements increase, the face-contact
load must decrease until the point is reached where face separation occurs or until a
controlled clearance is maintained at the seal interface. The seal leakage increases as the
face-con tact load decreases or as the face separation increases, as a result of the larger
effective leakage path. The effects of thermal, pressure, and cen trifugal distortions of the
seal interface may have more influence on leakage than the contact load. In some cases,
lower leakage may result from a lower face-contact load , because less thermal distortion is
caused by heat generation at the seal interface.
Seals designed for minimum cryogenic fluid leakage normally are face-contact
welded-bellows types with the maximum face-contact load allowed by the load-velocity
relations for specific materials and fluids. For less severe temperature applications (- 65 0 to
500° F) , elastomeric secondary seals may be selected instead of the welded-bellows
component. A thorough heat-transfer and stress analysis of the seal face and m ating ring
usually is performed to minimize seal interface distortions that can cause excessive leakage.
Effective sealing at load-velocity-life relations greater than those allowed by the current
state of th e art for rubbing-contact seals may require hydrostatic or hydrodynamic seals.
The other noncontact seal types (e.g. , floating ring and labyrinth) normally will not provide
adequate control of leakage for a primary liquid application but may be satisfactory in
hot-gas applica tions or in situations where a backup seal to control leakage can be provided.
Various th eo retical methods to predict seal leakage have been developed , ideal conditions
being assumed. Most leakage theory is based on steady-state laminar flow through very small
uniform flow channels with full fluid / film interface separation. Experimental stu dies h ave
illdicated reasonable co rrel ation with the th eo ry when the variables are known and
adequately co ntrolled. However, the variables are much more difficult to control on rocket
engine turbopump seals beca use of the ex treme themlal gradients and two-phase fluid
conditions; therefore, th e theoreti cal relations mu t be supplemented with empirical data
for practical solu tion.
Theoretical prediction of seal lea kage requires knowledge of th e leakage-path geometry and
fluid condition. The geometry of rubbing-co nta c t seals usually is altered by the wear
process ; therefore, any predic tions based on ini tial su rface m eas urem en ts are not valid after
wear-in. Al 0 , the effect of thennal distortion on cryogenic seals generally is much greater
than the effect cau ed by the nonnal variation of surface finish es. Exp erimental
measurements have indicated that the effects of th ennal di tortion du e to chiJJdown to
- 320°F can in crease th e static leakage rate by as much a 500 perce nt. Additional interface
thermal distortion cau ed by heat generation would be ex pec ted during high-speed rotation.
0
As noted , rubbing-surface temperatures in exccs of 1000 F have bee n measured on
liquid-oxygen seals where the nuid environm ent was at - 297° F .
39
The condition of cryogenic fluid s at th e seal interface ra rely is known with any acc uracy ,
and because of heat transfer from the rubbin g surface, th e fluid usuall y ch anges from p artial
liquid to vapor as it fl o ws ac ros the inte rface . Beca us the lea kage is proportio nal t o the
fluid viscosity and de nsity , the di ffere nce between liquid and va por is significant. The
viscosity variation with press ure and temp era tu re i give n fo r oxyge n in figure 30 and for
para-h ydrogen in fi gure 3 1. The vapori za ti o n proce s may crea te high er pre sure regio ns
across the seal interface a nd th ere by affec t th e pres ure profile. The fl ow process can ch ange
from lamin ar in compre sible-liquid fl ow to compressible-gas ch oked o r turbul ent fl ow.
The theoretical methods for predicting seal leakage are discussed in refe rence 8. Flow
regim es rangin g from molecul ar flow to laminar flow are discussed for liquid s and gases. An
empirical relation betw een initial surface geom etry measurements and static-leakage-path
effec tiv e gap is establish ed.
A more sophistica ted theore tical approach for predic ting seal leakage, whi ch co n iders the
effec ts of misaligned seal face , in terface wavin ess, fluid inertial forces, a nd inte r face fluid
film cavitation , is prese nted in re ference 9 . The an alysis assum es a fu ll tl uid fil m wi thin the
interface clea rance space and a known leakage-p ath effective gap. Th eo ret ica l methods are
presented for pr dic ting leakage in the turbulent-flow regim e and in the ex tend ed regions o f
th e lamin ar-flow regim e. The e ffects of rotationally indu ced turbul e nce are di sc ussed. This
theory applies to most high-pressure eals for cryoge nic propell ants beca use of the high
lea kage rates, high speeds , and low fluid viscosity .
The flow regimes fo r leakage are defin ed by either the Rey nolds numbe r o r the m o lecul ar
mea n free path , as sh own in fi gure 32. The flow regim e mu t be es tablished to de temlin e th e
applica ble theore tical lea kage relationship. Beca use th e leakage fl ow is required fo r the
Rey nold s numb er ca lc ula tio n, it is necessary to as um e a flow regim e and iterate th e leakage
and Rey nolds numb er calcul ations.
Th molec ular regim e exists when the mol ec ular mea n free path i equ al to o r grea te r th an
th e e ffec tive lea kage gap . As th e leakage gap is increased rela tive to th e m ea n free path , th e
flow en ters th e transition regim e. This regim e is th e com bination of mol ec ular and laminar
(re f. 10). The lamin ar regim e is entered when the leakage gap is increased to the p oint wh ere
molec ular collisions with th e wall are no longer significant. The tran ition from laminar to
turbul ent fl ow may be ca used by either in creased lea kage or high er rotational speeds.
Rotation ally indu ced turbul e nce ma y ca use the otherwise lamin ar lea kage to becom e
turbulent (ref. 9).
The flow regim e criteria are summarized below ( re fs. 8 through 1 1):
A.
Molec ular regim e : ;;;. 1. 0
h
40
- -- . _ - --
.,.
.. H
~ .
-.
_ ~ TT I - nTT· rTl IT
" 1 " - - IJ
f+
·ar
tt- ~ _ t-- - H- [-I
-t- .
-l
I
' 01
' 0.
"
!";; "-J
I"
P
P-
-
t"-
H~
±: -
I- f-:1 _
1-1 .~
=F
H
-
I -H
I-
I-
' 04'4 ~ E>- 1-1- -
- c-
~
- ~ f:s - t -
~ ~~ - ~
f"- . -
P-k - ~ '- j
: i~
-
·0
k - I-
P>-.ts: - -
IJ F-~~ ~
l' - - -j
-
I-
~ I:<: "- r l'
H f' 1- -
-- I- H
f-
' 0I I
I-H
-1 -
!'; H~ - ~ ~ r-: ~ - 1"1"- .L J,lt
1-1-
t=-
ex:
Piiili-' J • - -
~ t--=I>-- -
I
t-!- -/"i f-'" po.. k:- N -N r--, - - =t
lL ~ -
<::"1
I'..;
t--l
j-.. .
- . • .l-
u:
CD .~
~
IS
"-
- J.
.! I- H r-I-
I-
..J 0
" r- 1
.p.
>-' ' 00
01
I'<.;
- 1=
1- <00 t-
~ 00
::::
-I I-
U
Ul
, I •• .' It C::P': -
I- ~:;;;;; -I-
> i=l=H ..-0::
' -l-
f-I- f=I=
-+- !±~ f- f- f - C--
fF
>-- t=
00
\= l± --
' v-
t
r- 1
00 o. M
-+ -1 H I- r-I-
I-
I- r-t-
-r
I I TTl I I I + il~ T11J T u I II 1 1-1-11
.1.
-
i
' 00
,. 40 44 .1 51 54 eo.. TO 10 10 100 .. I ... ,,. tOO III UO ' 00 no 400
11f...l
4 $0 ~ 100 700 eoo 100
J.
IOQD
TE MPERATUR E, OR
F igure 30 . - V iscosity of oxygen as a function of temperature and pressure.
-- - -------
.. 00
c.'"
0."
0 .10 ,.
0." I . ",,-0. ..............
~
I-
o . ~I. ~s:s
I--
0 . <0I.
0 . '"
,f-- - i-
lI Q.ItO s,o l\AA 1100
:--.,
2'-. "-
a: 0.'"
,< 0
I
e:-
-
Ll..
Ll..
III
...J
-
-
>-- o 10I I,
.j:>.
f-
C/l
0
0'",.
0 00I
... - 5Q:a)I"5, .
-
tv U 0 01I . 1'<0
C/l
> 0 . 06 ~
oas,. r-....
,.., ",. -
004I.
,.. ""
-
""
0 . 03,- - - ,- r- r- ~f= "" -
YII"UI $oOl\Nl 'l)o
,., -::::
,. -
002!. -
-
~
0 .01
-"" - liD -..., - 210 - 250 - 2Jl) -2H.. -, . -,so - '00 -'" so 100
''''
200 .<0 JllO CIO SO) SCI
TEMPERATURE , OF
Fi gure 3 1. - Vi scos ity o f pa ra -hydr ogen as a fun c ti on of te mpe rat ure a nd pressu re.
- -- .~-- --~-~--
A = mean free pat h
h = effect ive leakage gap
TURBULENT
FLOW
Re = 1000 (leakag e)
LAMINAR FLOW
A
h= 0 .01
~
o....J
u..
TRANSITION FLOW
(MOLECULAR AND
LAMINAR)
SEAL AVERAGE CLEARANCE ...
Figure 32 . - Fl ow regimes in sea l leakage.
43
A
Transition regim e: 11 = 1.0 to 0 .01
A
Laminar regime: 11 ~ 0 .01
Leakage Re ~ 1000
Rotational R e ~ 2000
Turbulent regim e:
Leakage R e ~ 1000
Rotation al R e ~ 2000
The Reynolds number is defined by the following equ ations (refs. 8 through 11):
Uh _Q __ _ w
Leakage R e == v == 7rdv 7rdll (3)
· al R e =[wh
- =- NT-
27r - h
R o t a t IOn (4)
v v
where
A = molecular mean free path at average pressure, in.
h = effect.iv leakage gap, in.
U = leakage flow ve locity , in ./sec
v == kinematic viscosity , in 2 /sec
Il == visco ity, lbm / in.-sec
Q = vo lum e flowrate , in. 3 /sec
w= weight fl o wrate , Ibm / sec

d = seal face average diam eter, in.
44
I
L
r = seal face ave rage radius , in.
w = angul ar ve locity, rad / ec
= rotational speed, rev/ ec
Mol ec ul ar flow. - The mol ecular regim e in gene ral does not app ly to turbopump seals
because th e effec tiv e leakage gap is re latively large. The mol ec ul ar m ea n free path of liqu ids
genera ll y is suffi ciently sh ort to ensure flow beyo nd the mol ec ular regim e fo r mos t practical
sea l : th ere fore, th e th o ry presented is for compressible-fluid flow.
The leakage flow for co mpress ible gas is given by the fo ll ow in g equ ation ( ref. 8) :
0.532 A P ( P2 - Pj ) bh 2
W = - - -=-=:::----::-- --- (5)
RT J1 a L
whe re
P = average pressu re, (P 1 + P 2 )/ 2, psia
P2 = upstrea m press ure, psia
PI = downstrea m pressure, psia
b = seal face ave rage circ umfere ntial length, in .
R = specific gas co nstant , ( in .-lbf)/ (lbm-" R)

T = absolute temperature , oR
J.1 a = absolute viscosity, lbf- ec /in. 2
L = seal face radi al lengt h , in .
Tran iti on flow. - The transition regime generally appli es to turb opump seal only for
leakage of low-pressure compre sible gas und er sta ti c conditi o n. The leakage in this regime
is equa l to the sum of th e laminar flow and the m o lecul ar flow . The mol ecul ar correction
) 45
I
L
usually is negligibl e and m ay be neglected in the upp er region of the transiti o n regim e. Th e
leakage flow for compressible gas is given by the following equ ation ( ref. 8) :
w=w lamin. ar +w m olec ul ar

(6a)
With simplified arrangem ent of tell11S :
bh3 (P2 _ P2 )
W = _-::-:-=-=2~--=-_1_ (6b)
24RT J.l a L
Laminar compressible flow. - The lamin ar compressibl e-fl o w rcla ti o n appli es to

low-pre sure gas and c ryoge ni c-fluid lea kage. Lo w-pressure and lo w-tl o wra te c ryogc ni c tluid s
ge nerally will vap o ri ze be fo re entering the seal inte rface; therefo re, the fl o w process is b es t
described by compressible theory . The lea kage fl ow can becom e ch o ked a t th e ex it fo r
pressure rati os grea ter th a n abo ut 4 : I (6 r/ h > 100). An approx im a te qu as i-o ne-dim ensio nal
an alytical m o del th a t consid er th e in erti a term s has bee n co nst ru c ted fo r a na lys is or
co mpressible ch oked fl o w ae ro s a sealing dam ( re f. 12). The lea kage fo r ubso nic tl ow is
giv en by th e fo ll o win g equ ati o n ( re f. 8) :
bh3 (P2 _ P2)

2 1
W = -'7-:-=:------::-- (7 )
24RT J.l a L
Lamin ar liquid fl ow with out in ert ia. - The theo ry of lamin ar liquid flow w ith o ut inert ia
generall y is satisfac to ry fo r predic ting m ost lo w-speed o r stati c liquid le akage. Cryogcn ic
fluids at m edium pressure (~ 200 psi) ge nerall y fa ll in to this reg im e. Th e rad ial leakage tlow
is given by th e foll o win g equ ation (re f. 8) :
(8 )
46
L_
where
p = density, Ibm /in. 3
I'
2
= eal face outer radius , in.
I'
I
= seal face inner radius , in .
Laminar liquid now with in ertia. - The centri fugal effects caused by rotation are negligible
with low-density c ry oge nic tluid s (e.g. , liquid hydroge n) , but may b e significa nt with high e r
density cryogen ics (e.g., liquid oxygen) at low pressure and high speed. The inertia effects
with more viscous fluids may be signifi cant at low peed if the pres ure is also low. The
laminar radial leakage now with con id eration of th e inertia effects is given by the following
equation (adptd. from ref. 9):
3
w. - prrh [ P P + -3 P (..)2 (r2 - r 21 )] (9)
- 6 J1 I n( r I' r I ) rI - r2 20 m 2
a 2
where
Pr I = pres ure at sea l face inner radius, psia
Pr2 = pressure at sea l face outer radius, psia
pm = mass density = p ig, lbm-sec 2 /in. 4
g = acceleration du e to gravity, in. /sec2
Turbul ent liquid flow without inertia. - Most seals for high-pressure cryogen ic pumps
op rate in th e turbulent regime because of the high leakage now, high rotational speeds, and
low fluid vi cosity. The turbulence may be induced by either leakage or rotation. The
turbulent radial leakage flow without inertia is given by th e following equation (adptd. from
ref. 9):
VI = 26.8 g (~
12)1/7 [ Pm_ (P r2 -
P rl )J4 17 (lO)
J1. r 3/4 _ r- 3 / 4
1 2
I 47
I
L
Turbulent liquid flow with inertia. - The rotational Reynolds number (fwh/v) indicates
that rotation ally induced turbulence will exist on most cryogenic turbopump seals because
of the low fluid viscosity . The radial leakage flow with rotational induced turbulence is
given by the following equation (adptd . from ref. 9):
(11 )
Empirical data. - Reasonable estimates of dynamic leakage of face-contact seals for

cryogenics may be obtained by utilizing either the laminar-flow-with-inertia or the
turbulent-flow-without-in ertia theory (depending on the Reynolds number) , with the
assumptions of liquid conditions and a leakage path effective gap of approximately 200
mi croinches (h :::::: 200 ,uin.) This assumption allows for the normal thermal distortions that
generally occur on conventional turbopump cryogenic seals; it results in calculated leakage
rates that are in the approximate range of most measured values. A smaller effective gap (h
:::::: 100 ,uin.) is possible on designs that compensate for the major thermal effects. A larger
effective gap (up to h :::::: 400 ,uin.) may result if the thermal gradients are extreme or if the
effects of thermal distortion are not adequately considered in the design. When two-phase
fluid conditions exist, a leakage range can be estimated by calculating the leakage based on
both liquid and gas conditions.
An empirical relation for estimating static leakage of liquid hydrogen at low pressure (60
psig) is presented in reference 13. A leakage parameter for carbon face-contact seals based
on face geometry wa obtained from a correlation of transition flow theory and
experim ental test results. The results indicate that a leakage path effective gap of
approxim ately 25 to 50 ,uin . may be assumed for predicting static cryogenic leakage of solid
carbon ring seals. An effective gap of 50 to 100 ,uin. should be assumed for seal-ring-insert
designs becau e of the additional thermal distortions.
Additional emprical relations for correlating gaseous helium static leakage on new and used
carbon face seals with static leakage of liquid hydrogen and liquid oxygen are presented in
references 14 and 15 . The relations were established to allow prediction of static leakage of
propellant on th e J-2 engine.
2.2.6 Misalignment T olerances

The alignment of the rotating mating surface and the stationary seal nosepiece is more
cri tical on high-speed seals because of the larger forces required to compensate for the
inertia effects and ensure adequate dynamic response or tracking of the nosepiece.
L
48 I
~J
Turbopump hardware generally is machined with precision tolerances to maintain initial
install ation alignment.
The dy nami c loading on the rotating shaft and the pump casings usually is very high in
comparison with th at in most other application . Therefore, the initial alignm ents usuall y
are altered by the dynamic deflections. The effec t of the de fl ections can be minimized by
loca ting the seal as closely as possible to th e sh aft bearing support.
The seal design must compensate for the misalignments if the performan ce requirem ent are
to be sa tisfied. The effec ts of tolerance variations, dynamic de fl ections, th erm al ex p ansio ns
and contractions, shaft transient mov ements, and seal face wear on ax ial op erating length ,
radial lo cation, ma ting surface nOlmality , parallelism , eccentri city , wobbl e, and run out mu st
be considered .
2.2.6.1 AXIAL OPERATING LENGTH
Sha ft-ridin g seals (e.g., segm ented carbon and flo a ting ring) are used wh en th e vari ati o n in
ax ial o pera ting length is greater th an th e travel capability of face-contac t seals. The travel
lim i t of a face seal i dep endent on the axial sp ace avail able and generall y is e tablish ed by
th e bell o w o r spring load variation or stress level. Allowable travel ca n be in creased by u sing
longe r, lo we r-rate springs . Turbopump fac e seal usu aJJ y are limited by avail able space t o an
ax ial trave l of app ro xim a tely ±O.O15 in. in sma ll sizes ( I -in. di am eter) , ±0.050 in . in m edium
sizes (3- to 6-in . di ame ter) , and ±O. l 00 in. in large sizes ( 1O-in . di am e te r) .
2.2.6.2 RADIAL LOCATION
The radi al loca ti o n of th e sea l face rela tiv e to th e ce nte r of ro ta ti o n is h eld within
approx im ak ly 0. 00 2 in. / in . o f di am e ter on high-speed face seals in o rd er to preve nt
exce sive wipin g ac ti o n o n th ea l face. The e ffec t of th e wiping ac ti o n is co ntroversial. It
may ca use lea kag o r additi o na l face wear by shifting the wea r trac k. It m ay improve th e
hea t d is ip a ti o n ca pac ity by ex po in g a portion of th e rubbin g sur face to th e ealed fluid a t
eac h revo lut io n. The radi al loca ti o n o f sh aft-riding seals does not affec t pe rfo rm a nce , bu t is
ge ne rall y hdd within approx im a tely 0.005 in. / in. of diam e ter beca use of inte rn al c lea rances.
2.2.6.3 MATING SURFACE NORMALITY
In hi gh-s peed (> 10000 rpm) cals, the nonnality of th e sea l matin g sur face to th e ax is o f
ro tati o n ge nerall y i maintain ed within 0.0001 in. /i n. of di am e te r ( total indica to r
reading - T.I.R .) to preve nt exces ive mating-rin g wobbl e. Wobbl e o f th e m atin g ring at high
. peed results in excessive face lo ading beca use o f th e in erti a o f th e sta ti o nary nose piece as it
49
attempts to tra ck the mating surface. Th e ca pability of th e sea l face to track a wobbling
mating ring is limited by th e in e rtia of th e no cp iece a nd th e resu ltant c lo sing force
(re u ltant press ure force p lus spring force minus friction drag du e to seco ndary sea l or
vibration damp e r) . Excessive leakage, edge chipping of the seal nose, or exee s ive w ea r ma y
occur if th e limit is exceed ed. Bellows-type ea ls nonn a lly have better dynamic re ponse
than other secondary ea l. because friction drag is lower.
2.2.6.4 ROTATIONA L ECCENTRICITY
Th e mating surfa ce diameter for high-sp eed (> 10000 rpm) shaft-riding sea ls u ually is h eld
co nce ntric to th e ce nte r of rotation w ithin 0.0005 in ./ in. of di a mete r (TI.R.) to preve nt
excess iv e rotati o nal ecce ntri city. Th e effects of inertia forces and dyn amic res ponse req uire d
to track the eccen tric motion mu st be considered. Segmented carbon sea ls tend to leak
exces iv e ly when th e run o ut exceeds th e s ta ted limit , because the tracking ca p ab ility of th e
egme nts is limited. Floating-ring shaft ea ls are les sensitive to eccentricity , but tend to
wear at la rge r value because of the high er in ertial forces.
2.2.7 Vibration Control

Turbopump vibration. - Ro cke t e ngin e turbopump sea ls are ex posed to cxt re m c vibration
e nvironm e nts that ca n ca u e premature failure unl ess prop e r co nsid e ration is given to
vibration in ea l de ign. The sea ls on th e J-2 e ngin e are exposed to vibration in each of th e
two m ajo r axes at leve ls up to 15-g peak at a frequency range from 30 to 500 Hz an d up to
30-g peak at a frequency ra nge from 500 to 2000 Hz throughout the operational life of the
sea l. Seve ral c riti ca l frequency range, which co rre pond to e ngin e or turbopump resonant
points, ex is t.
Sea l vibration. ~ Face-contact seals running dry or in poor lubrica nts such a c ryoge ni c
tluid s may exc ite a e lf-gen e rate d natural- frequency vibration as a result of the stic k-slip
co ndition s at the sea l rubbing face. Th difference b e twee n th e tati c and dynamic
coeffic ie nt of friction ca n initiate a circumferential vibration mode at the natural frequency
of th e ea l face, which exc ites the axial vibration mod es. There are two basi c mod es of axial
vibration. One is vibration of the seal face composite as a free body supported by th e
bellows o r sp ring with th e motion res tri cted by th e mating ring surface; this mod e causes
face sepa ration, leakage, and fatigue failure. Th e other mode is vibration or surging of the
bellows or springs between th e seal face and th e housing ; this mode causes fatigue failure.
Mating-ring runout and shaft axial motion hav e bee n assumed to be th e primary exciting
force for initi a tin g sea l reso nant-frequency vibration ; howev e r, ex p e rim e ntal tes t results (ref.
16) and man y other turbopump tes ts indi ca te that the stich:-slip co ndition is the major cause
of und a mp ed vibration of fa ce-contac t seals. Lubri ca tion at th e seal rubbing surface
50
eliminates th e sti ck- lip condition. The re fore, the rubbing-material comb inati on and typ e of
sea kd or coo la nt fluid hav:e a great effect on self-excited seal vibration.
Sea l vibration ca n be controlled by des igning the eal to avo id th e major resonant
frequ encies of the turbopump and by providing seal vibration dampe rs. The general practice
is to design the seal for a natural frequency th at is high er th an any ex p ec ted on the
turbopump ; this practice prevents resonant eaJ vibration exc ited by the turbopump o r
engin e operation. Mechanical-friction vibration d ampe rs are prov id ed at th e seal face outsid e
diam ete r and a t the bellows component outsid e di amete r (fig. II ) to limit th e amplitude of
th e self-excited vibration modes. Some expe rim ent h ave indi ated that effective d amping at
eith er location will prevent both vibration m odes. The fr ict ion d rug force req uired fo r
effectiv e vibration damping of turbopump-typ e weld ed-bellows eals is ap prox imately 5 to
10 percen t of the seal spring load.
The second ary-seal friction drag on conven tiona! elasto m er, piston-ring, an d lip ea l
generally will provide effec tive vibration damping. Surging of th e seal load ing prings has n o t
been a significant problem ; therefore, spring dampers usu all y are not provid ed .
2.2.8 Contamination Allowances

Rock et engin e fluid systems usually are cleaned and filt ered , filters in the 10 to J OOjJ. range
being used ; also, the turbopump shaft seals are protected fro m atmosph ric dirt. Therefore,
foreign-particl e contamination is not a serious problem , and special protec tion systems (e.g.,
neutral fl uid inj ection , buffer zones, centrifugal separators, slingers) ord in aril y ar not
utiliz ed. However, m any turbopump shaft seals are located in sump areas th at tend to
collect foreign-particle settlings (metal filings , machin e chips, weld slag, braze flux , wear
debris from other components). The sensitivity of the seal to foreign -particle dam age must
be co nsidered in the effort to achieve high reliability. A seal that is not ca pab le of operating
with a small amount of foreign-p article contamination would not be practical fo r m ost
turbopump applications.
Rubbing-contact carbon face seals h ave demonstrated satisfactory reli ability when ex posed
to the foreign p articles normally found in turbopump systems. The particles generally are
too large to enter between the rubbing surfaces, and little d amage from scoring or gouging
of the seal face occurs. However, the particles tend to congregate in the second ary seal area
an d cause m alfun ctions as a resu lt of hangup and leakage. Plastic-lip secondary seals tend to
wear an d become gouged, the result being excessive lea kage. Also , foreign particles tend
to becom e lodged in th e plastic, and the abrasive ac tion cores th e m ating su rface. The sam e
problem exists to a lesser ex tent on elastomer secondary seals. Piston-ring seco nd ary seals
also are subject to abrasive damage from foreign particles. The most reli able secondary
element for abrasive service is the b ellows seal.
51
-I
Hydrodynamic and se lf-ene rgized hydrost ati c seals are very sensitiv e to fo reign-p articl e
contamination. The parti cles tend to en te r th e face clea ran ce gap and go uge o r wea r th e eal
face. The face dam age ca n result in loss of th e lift pote ntial and sub sequ ent failu re. The
extern ally pressuri ze d h ydros tati c seal (fig. 16) i resistant to d am age by sealed fl uid
contamin atio n , because th e face is ep arated by the purge fluid . The cleanlin ess of the purge
fluid is cri ti cal.
The possibility that the sealed fluid wilJ fO lln abrasive c rys tals as a res ult of th enn al
decompositio n caused b y h eat transfe r at th e seal rubbing face o r ch emi cal reac ti o n mu s t be
considered . The load-tim es- p eed limit m ay be es tablished by the critical t emp erature. A
neutral fluid bu ffe r m ay be required to prevent the sealed fluid from con tac ting th e seal
face .
The h ard er seal face m aterials ( tungsten carbid e, titanium ca rbid e , aluminum o xid e) are
most resistant t o abrasive dam age .
2.2.9 Seal Mounting Requirements
2.2.9.1 SEAL PILOT
Mounting pilo ts fo r turb o pump seals mu st co mp ensa te for ex trem e thermal gradients and
differential therm al co ntrac tion or ex p ansion of dissimilar m aterials. Excessive dis to rtio ns
and high stresses betwee n th e turb opump h ousin g and th e seal h o using pilo t m ay result fro m
in crea ed interfe ren ce if provisio ns for the th ermal diffe renti al are n o t alJ owed. Th e ge ne ral
prac tice is to use m ateri als with similar th enn al contrac tion rates fo r th e eal h o using. Wh en
dissimil ar m a te ri als are used , th e seal pil ot fit a t ambient temp erature mu st be adju ted to
allow fo r th e th e lm al ch anges.
Appli ca ti o ns th at utili ze simil a r m ateri als m ay require spec ial m o unting me th o ds il' the
temp erature gradi e nt be twee n th e turb opump h ousing and th e seal h ousing is excess ive.
Di am e tral pilo ts may no t m aintain adequ ate seal rad ial alignm e nt beca use o f th e la rge pilo t
clearance required to all o w for th e th eml al ch anges. A m o untin g device co nsisting of three
equall y spaced radi al pin s th at e ngaged radi al slo ts in th e seal h o using was used o n th e seal
on the J- 2 liquid-hydroge n turbin e to elimim ate ca rb o n breakage ca used by exces ive
inte rfere nce at th e eal pil o t. The seal was mounted o n t he pump h o usin g, which was chill ed
do wn to approx im ately - 400° F , and th e fl ange was ex pos d to th e turbin e h ot gas a t
app rox im ately 1000° F. Therefo re, th e sea l fl ange temp e rature in creased m o re th an th a t of
th e pump h o usin g pilot, and thu s exces iv e inte rference develo ped. The three-pin m o unt
all owed th e ml al gro wth whil e m aintainin g radi al a lignm ent. Fl ex ible pil o ts th at all o w radial
de fl ec ti o n with o ut tress in g th e sea l hou sin g m ay also be utili zed .
52
2.2.9.2 FLANGE SEALS
Many f1 ange gas ke t materi als tend to shrink away from the mating surfaces at low
temp erature. T he redu ced sea ting stress m ay result in excessive leakage. The gaske t seating
stress ca n be maintained by spring loading th e gasket or the fl ange to compensate for the
dimensional change. The conventional elastomer O-rings (Teflon, Buna, Viton) are not
sa tisfac to ry below a bo ut - 100°F beca use of excessive therm al-contrac tion differential and
insu ffic ie nt resil iency to comp ensa te fo r th e dim ensional ch ange .
Generall y it is mo re sa ti fac to ry to loca te th e stati c seal on an axial surface rather than in

th e rad ia l direc ti o n beca use o f th e g rea tl y in c rea e d diffic ulty o f sealing a t lo w t emp er a ture .
Axial sea l ca n be c lamp ed ac ross a thin sec ti o n. to minimize the dimensional change caused
by th erm al co ntrac ti o n. Lo w-temp erature gas kets are mad e as thin as possible. The most
effective static seal fo r ex treme temp eratures are the m etallic spring-loaded
pressure-ac tu ated types. Elas ti c de f1 ec ti o n of the seal provid es the necessary resiliency to
comp ensa te fo r dim e nsio nal va ri a ti o n at the sealing joint. The seal seating stress is increased
at high er press ures fo r mo re effec tive sealing by the inherent pressure lo ading feature. The
all-m e tal co nstru c ti o n all o ws the seal to be used with reactive and corrosive propellants at
temp e rat ures from - 4 23 ° F to 1500°F. The seals usually are plated with a softer material
(copper. sil ve r, go ld , Kel- F, Te fl o n) to illlpro ve the complian ce to irregular surfaces and
redu ce th e requi red seatin g stres .
2.2.9.3 FLANGE LOADING
The required f1 ange load is de te mlin ed by th e sea ting stress for gaskets and the load to
de fl ec t th e elas tic memb ers fo r spring-loaded seal s. The e ffect of thermal changes and
press ure loadin g o n th e f1 ange is ta ken into account to en ure ad equate load control at the
ex treme of opera ti o n. The thenn al chan ges a nd press ure-se parating force on the flange
mu st no t red uce th e fl ange preload belo w th e minim um sea ting stress req uired for effective
sealin g. When p ossible, the f1 ange is designed to cause th e pressure force to increase the
fl a nge sea tin g stres . The e ffec t o f th erm al co ntrac tion can b e used to increase the gasket
sea ting st res by utili zing m a te ri al combin a tio ns th at result in more shrinkage in the loading
membe rs th an in th e eal eleme nt. Bolts th a t shrink more than the flange generally will
comp ensa te fo r th e hrinkage o f thin (0.005 to 0.0 15 in.) gaske ts.
The seal f1 ange can be effec tiv ely spling load ed by th e use of several small-diameter
lo ng-l ength bolts fo r clamping. The elas ti c elonga tion ca used by bolt preload generally is
sufficie nt to co m pe n ate fo r gas ket shrinkage. Low-temp erature gaskets are loaded by elastic
members to prevent loss o f preload du e to thermal contrac tion . The use of large ring nuts to
clamp ga ket fl anges is av o id ed beca use o f th e low elastic elongation and the nonuniform
lo ading ca used by th e nut surface no t being ex ac tly p arall el to the seal surface. Several small
bolts aro und the fl ange will pro vid e mo re uni fonn loading without precision tolerances.
53
2.2.9.4 SHIMS
In most turbopumps, the seal axial operating length must be adjusted by using shims to
position the seal or mating ring. The tolerance stackup between the sta tion ary mounting
surface and the rotating mating surface generally is too large to allow installation of the eal
without shims. Most turbopump face seal are installed to a tolerance of ±0.005 to 0.010 in.
Usually the shim is combined with the fl ange gasket to provide the static seal; therefore, the
shim must satisfy the gasket requireme nts previously disc ussed . Gasket shims often are
ground metal washers of various thicknes es with narrow con tact lan ds around the inner and
outer edges to increase th e seating stress. The shims generall y are coated with a softer
material (copper, silver, gold, Kel-F, T eflon ) to provide more effec tive static sealing. Shims
also are constructed with recessed grooves to allow the use of spring-loaded,
pressure-actuated static seals (fig. 7). Bonded lamin ated shims usually are not atisfactory as
gaskets because of the irregular seating su rface and leakage between the lamin ations.
Laminated shims have been used between asbestos rubber gaskets in liquid oxygen on the
H-l turbopumps (fig. 5) ; however, leakage occurred, an d the later de ign s were changed to
either ground m etal wash ers or spring-load ed eal.
2.2.9.5 PROVISION FOR SEAL REMOVAL
Seals frequently become stuck or wedged in th e pump housings as a result of di tortion or

yielding of the pilots. Con iderable difficulty is avo ided by u e of pecial pulling tools to
ass ist in seal removal.
2.2.9 .6 LOCKING DEVICES
The bo lts or nuts used to mount th e ea l are ec urely locked to preve nt loosen ing and
bac king out. A loose bolt or nut in a liquid oxyge n sy tem ca n r suit in a pump explosion
ca used by ignition of th e rubbing me tal. Wh e n po sibl e, bolts an d nuts are trapped in
po iti o n by the adj ace nt hard ware.
2.2.9.7 MATING-RING MOUNTING
Excc ive seal lea kage du e to ma tin g- rin g di sto rtion ca n res ult from improper m ating-ring
mounting. Free-floating no nlo aded mating rings are preferred for minimizing di stortions
from th e clamping forces; however, this method generally is not feasib le a t c ryogenic
temp e ratures because of the difficulty o f sealing around a free-floating ring. Th erefo re, most
c ryoge nic mating rings are clam ped ax ially along the turbopump haft . The distortions du e
to th e clamping load s are minimi zed by e nsurin g that th e ma ting surfaces are flat and
54
L
normal to th e direc tion of applied load. When possib le, th e surfaces are lapped flat within 3
helium light bands.
Th e mating rings must be d esign ed to minimize th e bending mom e nts from th e c lamping
lo ads, th ermal loads, and pressure forces . One method i to e para te th e clamping stre es
from th e seal mating surface by using a th in web to support th e mating ring. Th e other
approach is to make the mating ring rigid enough to withsta n I th e c lamping fo rces without
being distorted.
The joints between the mating nng and shaft or shaft pacers, which are ex posed to
press ure, are ealed to prevent lea kage.
The mating rings are prevented from rota ting relative to the shaft an.d thu are preve nted
from damaging the static seal and pilot. The ax ia l c lamping force genera lly i adequate to
prevent rotation , except when the differe ntial the rm al co ntrac tion or th e Poisson e ffec t of
the ce ntrifugal defl ec tions causes the axial ta c kup to loosen . T hi probl em was so lved on
the H- l turbopump by uti lizing radia l sp li nes be twee n th e mat ing ring and shaft.
2.3 SEAL COMPONENTS

The design considerations relat ed to th e d e ta il requ irements a nd co mpon ent of th e seal
assembly are presented in the fo ll owing disc ussion . Th e det ail co mpon ent d e ign ana lysis is
an esse nti al part of th e seal design and mu st be integ rated with the de ign processes for the
seal system and sea l a embly.
2.3.1 Seal Materials

The selection of mat e ri als for turb op ump sea ls i based primarily o n co mpatibi li ty with the
fluid m edium , with con id crat ion o/" temperat u re Jim itat ion , eorrosion-re ista nce, th ell11al
contraction and ex pansion , hea t co ndu c tivit y, th e rma l stabil ity , st re ngth, ductilit y,
hardness, modulus of elas ti c ity , res ili e ncy. fatigue. c ree p, wea r a nd friction, sel f- lubricity ,
fabricability , availab il it y , and cos t.
A summary of materials currently LI cd for ro tat ing-shan sea ls on ro c ket e ngin e turbopump s
is presented in tab le III. Additional mate rial co mpat ibi lity co nsiderat ion are di eu sed in
references 17 through 20.
2.3.1.1 COMPATIBILITY
The selection of materials for tu r bopump eals is lim ited by the requireme nt that the
material must be compatible with the rocket e ngine propell ants. S tro ng oxid izers sLich a
55
_J
I
Table III. - Summary of Materials Currently Used for Turbopump Rotating-Shaft Seals
Face comb i nat ions Seco ndary Housing and

Mat i ng- Insert Metal
structural
Fluid (nos e - mat e ) CD r i ng ba se seal Bell ows
element s
adhesive plat in g
Li quid Carbon P69 2- Chrome 4130 Ke l -F 34 7 302 Epon 901 / B3 Chromium

Oxygen Car bon PSN- Chrome Inco nel 7S0 My l ar I nconel 7S0 303 Tefl on Fus io n Cadmium
Carbon PS N- LW5 I ncone l 718 Incone1 718 304 Sil ver
Ca rbon SAG- Chrome 440C 32 1 Coppe r
Inconel 750 Nic kel
Inconel 718 Go l d
Inconel 600
Invar 36
Ca rpente r 42
Li qui d
Fl uorine
K1 62B- Kl62B
A1 0 - Kl 62B
2 3
Kl62B
Inconel 750
CD Inconel 750
Inconel 718
Inconel 750
In conel 718
CD Chrom ium
Copper
A1 0 -A1 0 Inconel 718 Nicke l
2 3 2 3 Go l d
FLOX : Kl62B- K162B K162B Kel-F In co nel 7S0 0) Teflo n Fusion CD
30 % Fluorine Inconel 718
VI 70% Oxyg en
0\
Nitrog en
Tetrox i de
Carbon EYIOS-LWS 32 1
CD AJ1 350 32 1
347
CD
Liqui d
Hydrog en
Carbon P5N-Chr ome
Carbon SAG-Chrome
Inconel 7S0
I nco nel 718
CD I ncone l 750
Incone l 718
0) Epon 422
HT 424
CD
Car bon P03N- LWS 310
50% Hydrazine
50% UDMH
Carbon EY10S-LWS 321 CD AM 3S0 347
CD
RP- l Carbon G39-Chrome
Ca r bon CCA72-Chrome
4130
4340
Viton A
Buna N
AM 3S0 300
Alumi num
Epo n 901/B3
Epo n 422
CD
440C
Hot Gas Carbon PSN-Chrome Inconel 7S0 CD I ncone1 7S0 In co ne l 750 Epon 422 0)
LOX + Ui Carbon G84-Chr ome I ncone1 71 8 Inconel 718 Inco ne l 718 HT 424
2 AM 35 0 In cone1 600
1000°F Max.
Hot Gas Car bon P2003-Chrome 4130
CD Inconel 750 Has tell oy Epon 422 0)
LO X + RP - l Car bon G84 - Chr ome 4340 Inconel 71 8 Inconel 750 HT 42 4
1000°F Max. Car bon G84- LW I 440C AM 350 Inconel 718
- -
8 Lis t ed in ord er of pref ere n ce
CD Dat a not avail ab l e
CD Same as fo r liquid oxyg en

li q uid oxyge n and liquid fluorin e a re cap able o r cO lll bu ti o n with all avail able seal m aterials
except full y ox idi zed oxides, fluorid es, and ca rbid es. Hy pergo li c reac tio n (i.e. , ign iti o n o n
co ntac t without an external energy source) ca n occur with so m e p ro pe ll ant and m at erial
combin ation s. Combu stion generally will co ntinu e until eith er th e ox idi ze r o r th e
co mbu stible m ateri al i consum ed ; th erefo re. the res ult of co m bus ti o n may be ca ta trophi c
fa ilure of the engine and p o tenti all y th e to t al ve hi cle.
Me tals. - Man y metals (Monel, nickel-base alloys , and stainless steel) are sufficiently
res i t a nt to ox ida ti o n lea din g to cO lll bu ti o l1 th at th ey m ay be u ed as st ru c tural
co m ponen ts o r liquid-ox idize r eal : howeve r, rub b in g co n tac t of meta l aga in st me tal in
stro ng ox idi zers is avo ided beca use or th e po te nti al igniti on haza rd . Th e protec ti ve ox id e
coa ting may be d es troyed by th e rubbing co ntac t. and th e fri ct io na l heat ge ne ra ted at th e
urface a peri ties may be su f fi c ie ntl y high to initi a te co mbu sti o n. I n tances of
Ill e tal-t o-m e tal ru bbin g in liquid ox idize rs w ith ou t ig ni t io n have occ urred : h o wever, thi s
res ult is un p redic table, and th e h aza rd i ge ne rall y too grea t to ri k. Liquid-ox idi ze r pumps
have exp lo d ed as a resul t o f rubbin g co ntac t o f m e tal ur face.
Me tal-t o-m e tal co ntac t at seal n osepiece pil o ts, antiro ta ti o n tangs, and fri c tion-typ e
vibrati o n d amp ers in liquid o xidiz ers usu all y is acce pta ble, provid ed th at care is t aken to
minimi ze the co ntact energy and ox idation-resistant m aterials are u tili zed . M.o st m etals are
resist a nt to d eto n ati o n from impac t load s in liquid o xidizers. 0 instances of m et al reac tion
in liquid oxyge n have been re ported during the standard 70-ft-lb f imp ac t tes ts, and no
metal-oxygen reac ti o n a t seal antirotation t angs h ave b een noted .
The m Ll rtL'llsiti c s teel all oys (e.g., 17-7 PH, AM-350, 4130, 4340) are no t comp a tibl e with
cryogL' nic p ro pell ant beca use o f the loss of du ctility at low t emp e ra ture. So me o f the e
ma te ri a ls (e.g., 4130 and 4340) h ave bee n used at cryoge ni c t mp e rature fo r t ru c tura l
co m pone nts when du ctility o r elo nga ti o n is not critical ; h o weve r, th e e m at e ri al are no t
used a tl e xing elem e nts.
Meta l pla tings (silver, gold , cadmium , chro mium , nic kel, copper) ge nerall y a rc ve ry res ista nt
to igni t io n in liquid ox idize rs. So m e excepti o n are silv er and cadmiuIn in li q uid o r gaseo u
tlu o ri ne. Static seal and fas te ne r u ed in flu o rin e usu all y a re pla ted with co pp I' o r go ld.
Th e m et al platin gs used for most prop ell ant a re co pper o r silve r for st a tic sea ls, cadmium o r
silve r fo r fas tern ers, and chromium o r cadmi u m fo r co rrosio n pro tec tion. Go ld is the mos t
resis t ant to oxid a tio n a nd used ro r criti ca l :1 ppli ca ti o ns. Silve r is resistant t o ox id a ti o n 111
liquid oxyge n and th erefo re m ay be used in tran:-. iL' nt o r intennitte nt rubbing co ntac t.
Carb o n . - Ca rb o naceo us m a te rials h ave d emo n trated ad equ ate res i tan ce to CO Ill busti o n
and ignitio n in rubbin g co ntac t again t hard -c h ro m e-plated tee l a nd tungs te n/ eh ro mium
carbid e ( LW5 )-coa ted steel m atin g rings in liquid ox ygen. Carb o naceo us m at e ri als with
reac tive im p regn ants are p o te nti ally capa ble o f igniti o n in liquid oxyge n: h o weve r, no
combus tible reactio ns are kn o wn t o h ave occurred with th e m a t ri als Ii ted in table Ill. Th e
57
failure mod e for carbons exposed to high rubbing-surface temp eratures (~ 500 to 1000° F)
in liquid ox ygen is gen erally cracking and breakage ca used by th ermal stresses and chemical
eros io n. Imp ac t tests of carbon (P5 , P69 2, P03N) in liquid o xygen at energy levels of 70
ft-Ib f rc ulted in th e carbon being pulverized without any reaction s. Carbon aceous materials
are no t L1 sed in liquid fluorin e because o f lac k of comp atibility and inst ances of explosive
reac ti o ns. Hyg roscopic carbons (i. e., those th a t tend t o absorb w ater) are not used in
cryoge ni c se rvice because o f possible freezing.
Pl as ti cs. - Plasti c material s (Kel-F , T e flon , Mylar) have dem o nstrated ad equ ate resistance to
ox idati o n and igniti o n both duri ng rubbing co ntact at fluid pressures up to 1000 psi and
impact (7 0 ft-I b f) tes ts in liquid oxyge n. The comp atibility at very high fluid pressures
(5000 psi) m ay be ma rgin al. Th e p lasti c ma teri als are n o t comp atible wi th stro nger oxidizers
such as liquid flu o rin e. Th e co mp a ti b ilit y o f Mylar is m argin al in liquid oxygen at imp act
energy levels in excess of 70 ft-Ib f. De to nati ons h ave occurred during imp ac t tests at en ergy
levels of 80 ft-Ibf; however, n o known reac ti o ns have occurred durin g liquid oxygen t esting
o n lip seals or gaske ts . Pl as tic m ate rials ge nerall y are not used belo w - 32 0° or above 600° F.
El astom ers. - E la ·to meri c ma teri als (e.g., Vito n A and Bun a- ) are n ot useful with
cry ogenic prope ll ants because they lack elo nga ti o n and resilie ncy at low temp erature «
6 5° F). These ma teri als are no t sa tisfac tory above approxim a tely 500° F becau se of thermal
degrad ation . Chemi ca l co m pa tibility is a pro bl em with som e prop ellants. The effec ts of
shrinkage and swe llin g o f e las to me ri c ma te ri als in some prop ell ants must be c onsidered.
Som e elastom eri c ma teri als are ubj ec t to surface de te J1 0 rati o n ca used by high ozone
content in atm os ph e ri c air: Vit o n A is resistant to dete riorati on by ozo ne. Most elastomeric
materials, exce pt Vito n A , are age-limited and require replacement after sp ecific time
pe ri ods from th l' cure date. Viton A ge nerall y is used fo r turbopump RP-l and lubricating
o il seals.
Adhesive . - Th e co mp atibilit y of ep oxy adh es ives with liquid oxyge n is m argin al. Imp act
tes ts have indica ted th at reac ti on will occ ur at imp ac t energy levels in excess of 2 kg-m
(14.46 Ib f-ft ). Epo n 90l /B3 is th e most nea rl y comp atible epoxy adh e iv e available and h as
bee n used successfull y to b o nd ca rb o n in se rts to m etal carri ers on th e J-2 liquid oxygen
sea ls. Th e o nl y o the r bo ndin g me th od sa tisfac tory in liquid o xygen is T eflon fu sion
ho ndin g. whi ch i co mpatibl e wi t h liquid oxygen but is not as stro ng o r as con istent as
l' P OXY adh es ives. Epo xy adh esives are not u sed in liquid oxyge n unl es th e bond ed surfaces
arc pro tecte d fro m direc t co ntact with th e circulating fluid and th ere is no possibility of
hi gh-ene rgy impac t loads or rubbing co nt ac t. 0 adhesiv e bonding m ethods compatible with
liquid tlu o rin e are kn o wn .
Ce rami cs and ce rm e ts. - The cerami c and cerm e t materials (aluminum o xide, titanium
carbid e. tun g ·ten ca rbid e, tungsten/ chromium carbid e) are among the most inert and
ox id ati o n-re istant ma terj als availa ble; th ere fo re, th ey are generally used wh en ch emi cal
co mp atibility . reac tiv e o x idizers , or high-temp e ra ture ox idizing environm ents are a problem.
58
The ceramics are used at temperatures up to approximately 2000°F in oxidizing
environmen ts.
2.3.1.2 CORROSION
Corrosion protection usually is provided by utilizing corrosion-resistant materials (stainless

steels and nickel-base alloys) for seal construction. Chromium and cadmium plating are used
to protect ferrous metals (e.g., 4130 and 4340) in mildly corrosive environments such as
liquid oxygen , RP-J , lube o il , and (LOX + RP-l) hot gas; however, plating will not provid e
adequa te protection when moistu re is present in hydroge n or fluorin e e nvironm e nts ,
pa rti c ul arly at high temperature. The base meta l m ust be corrosion resistant to prevent
corros ion und e rn ea th the plating du e to p lating porosity. The hot gas gen rated by the
combu tion of liquid oxygen and liquid hydrogen (hydrogen-rich steam: H2 + H 2 0) will
corrode sta inless steel through hard-chrom e pla te.
The e lectrolytic ac tion caused by th e combin a tion of free hydrogen, moisture, and h eat
re liltS in excessiv e corrosiv e pitting of the 300- and 400-series stainless steels if
foreign-particle co ntamin ants are present on the me tal surface. This problem was solved on
the ea ls o n th e J-2 engine by utili zing Inco nel allo ys (600, X-750 , and 7 18).
The combustion of h ydroge n a nd fluolin e or th e presen ce of moist ure in fluorin e will

produ ce h yd rofl uori c ac id ( HF) , whi ch is very co rrosive to aluminum all oys, ferrous m etals,
and 400-serie stainle s stee ls. The 300-selie stain less steels are satisfactory for dry HF but
not for wet HF. Th e nickel-base alloy, co ppe r, chro mium , and carbon P5 appear to be
sat i factory ror both wet a nd dry HF .
2.3 .1.3 HYDROGEN EMBR ITTLEMENT
Stru ctu ral elemen ts in hy d rogen env ironm e n ts may ex peri ence bli ttJ e-fra cture fail ures und er
load as a result of hydrogen embrittl eme nt of the meta l. The emblittlem ent may be caused
by reac tio n of h ydroge n with the me tal co nstitu e nts or by absorption of hydrogen by the
m etal. The nickel-base and Mon el alloys app ea r to be affected by hydrogen emblittlem e nt in
hydrogen environme nts at temperatures above approx imate ly - 200° F. The ni ckel-base
alloys (Inconel X-750, ln co nel 718) are satisfactory for use in hydrogen enviro nm ents at
tem p eratures belo w approxim a te ly -:WO° F, provid ed the proper h ea t treatment is used .
Hydrogen embrittl eme nt has not been a problem with th e Incon el X-750 weld ed-m etal
bellows sea l used on the J-2 liquid hydroge n turbopumps. A luminum and copper alloys are
not affected. Th e stabl e stain less s teels (those that rem ain austen itic and do not transform
to marte nsite during defo rmat ion: 310, 3 16, 347, A286) are resistant to hydrogen
embrittlement a nd gene rall y a re Llsed in hydrogen environm ents above - 200°F. Additiona l
d iscussion of hydroge n embrittl emcnt may be found in refere nces 21 and 22.
59
2.3.1.4 MATERIAL PROP ERTIES A ND T EMPERATURE EFFECTS
Th e basi c materi al proper t ies o r ·t re ngth . du ctilit y, elo nga ti on, modulu s of el a ti cit )',
re ili ency , creep , h ardn ess, th erm al co ntrac ti o n and ex pansion, h ea t co ndu ctivit y, and
fati gue life are co n id ered fo r th e materia ls used in turb opump seals. Th e effects o f
op eratin g temp eratures ( 4 23° to 1000° F ) on th ese prop erti es mu st be t aken into acco unt.
In most ca es, th e variati o n o f materi al properti es with temp erature is no t a lin ea r
relationship , and ex trapol ati on o f d ata measured throu gh narrow temp erature ranges ca n
re ult in gross erro rs. F o r some t y pi ca l ea l materi als, th e vari ati on with temperature is given
fo r tensile y ield strength in figure 3 3, fo r elonga ti o n in fi gure 34 , and fo r th erm al ex pansion
in fi gure 35. dditi o nal materi al properti e at c ryoge nic temperature are given in reference
23. Th e properti es at elevated temperatures are given in re ference 24 .
2.3.1 .5 WEAR AND FRICTION
Mo. t current turb opump eal s are rubbin g-co ntac t t y pes th at op erate in th e bound ary
lubri ca ti o n regim e o r with sel f-lubri ca ti o n f rom a film d eposited o n th e m atin g urface. Th e
wea r and fri c ti o n properti es o f th e sea l-face and m atin g- urface m at ri al combin ati ons in th e
sca led fluid are highl y d epend ent on th e lubri ca ti o n avail abl e at th e seal interface. Mo t
fri c ti on data have been o bt ain ed with butt on-ty pe spec im ens; how ever, thi s t es t procedure
doc no r accou nt ro r th e hy drod y nami c lift and hea t buildup th at no rm all y occ ur at sea l
inter faces. Th e h y d ro dy nami c li fe te nd to redu ce th e re ul ta nt face-co ntac t loa d and
th erefo re th e coe ffi c ient of f ri cti o n. Th e e lTeer o f high er rubbing p eed and co ntac t load i
in crease d inter face temp erature, whi ch ge nerall y in c rea e th e coe f ficient o f f ri cti on. Th e
chemical reac ti on betw ee n th e sea l in te rrace m ateri al and th e seal ed fluid has a igniri ca nt
erfec t o n th e fOnll ati o n o r a . elf-Iubri catin g film at th e se al rubbing face.
Th e ce rami c and ru se d-cc nne t ma terials depcnd o n the fo rm ati o n o f an ox id e o r flu orid e
film fo r lubrica ti o n in liq u id oxygc n o r li q uid Ilu orin c. Tit aniuJ1l ca rbid e ( K entaniuI11
K 16 2B) and aluminum ox id e ha vc dClll o nstrated sati sfac to ry wear and fri c ti on in liquid
fluorin e beca use a elf-lubri catin g i'!u o rid c rilm rorm s wh en th ese m ateri al arc ex posed to
flu orin e. Th e appli ca ti o n or ruse d Ilu () rid e coa tin g to th e m atin g surface of ce ramic and
ce rm et materi al. is benefi c ial in redu cing wea r an d fri cti o n in liquid oxyge n, liquid fluorin e,
liquid sodium , hy droge n, vac uum , and air ( refs. 25 and 26 ).
Compo ite materials (AmC erM et 701-6 5) con sistin g o f po rou s sintered nickel-chrom e alloy
matrix infiltrated with in organic fluorid es (62% B aF 2 , 38% CaF 2 ) have dem o nstrated
satisfactory wear and fri c tion wh en t es ted in dry ga eou s helium and ex posed to fluorin e.
Carbonaceous materials used for turbopump seals (t abl e III) are impreg nated with inorganic
resins, metal fluorides , silver, and variou s oth er additiv es to improve th e w ea r and fricti on
characteristi cs and as ist in the formation of a self-lubricating film on th e m ating surface.
The mechanism of transfer- film form ation with ca rbon aceous materi als and the effec t on
60
L
260~----~------r------~------1
110
Ul
~
J:
~
~
UJ
140
,
a:
~
Ul
Cl
...J
UJ
>=
UJ
...J
Ul
Z
UJ
~
40
TEMPERATURE ,
Figure 33. - .
Tensl·1 e yield streng th of typical alloys as a function 0 f temperature.
61
L....- __
I-
Z
W
U
c::
w
"-
z'
o
f=
<l:
tel
Z
o
...J
W
o 400 500
TEMPERATURE , o f
Figure 34. - Elongation of tYPical

. alloys as a function of temperature.
62
0. 008
0.007
0 . 006 -
; .
0 . 005 1----;-- - -,
z
w 0. 004
([
::J
I-
<{ O. 003
([
W
Cl.
~ O. OOZ
w
I-
o 0. 001
I-
u.
o
o
.....
w
19
Z
<{
I
U
I_ / 1 - ---=1- . ~ - L. i : .oj j-- I. - - - l- '-r-~I-]:

- O. 008
• O. 009
.
-II: -.
~=--- C:==;=;~~~~ ~:~'~---1~-:--':
- +·f .
:- -~ - -~.!-.:. . ·t-'T H -:- --'-
.' . . -·;-~ft:l
--f---- --- '1 t-:'" i-·J
- 0.010 ~ - ;-. -- , .. - ---;--1- r r~'-: - I r· - ! .~ ++
-500 -cIOO ~ -roJ -100 100 200 300 400 500 600 700 800 900 1000 1100 1200
TEMPERATURE, OF
Figure 35. - Thermal expansion of typical seal materials as a function of temperature.
I 63
I
L
wear and friction are discussed in reference 27. The factors that affect wear and friction of
carbon materials are discussed in reference 28. The wear and friction of carbon materials in
liq uid nitrogen and hydrogen are discussed in reference 29.
2.3.1.6 CLEANING
Parts exposed to strong oxidizers such as liquid oxygen and liquid fluorine must be
thoroughly cleaned to remove all traces of hydrocarbon contamination prior to exposure.
Seals to be used in oxidizers are specifically designed for ease of cleaning. The individual
components are cleaned and are maintained in a clean condition prior to assembly. The
cleaning process generally consists of rough cleaning by bmshing in trichloroethylene, then
hot vapor degreasing or ultrasonic cleaning or both. The cleaned parts are sealed in a plastic
bag for protection until ready for assembly. The final cleaning and assembly usually are
accomplished in a controlled-atmosphere white room (ref. 19).
To ensure compatibility with liquid oxidizers, seal faces that are lapped with diamond or
silicon compounds suspended in mineral-oil bases must be thoroughly cleaned by bmshing
and ultrasonic cleaning in trichloroethylene. However, even with this technique, it may not
be possible to completely remove the contamin ation from soft or porous materials. For a
Bearium B-lO seal in tended for use in liquid flu orine, the lapping medium was changed to
aluminum oxide particles suspended in trichloroethylene to prevent surface reactions.
Carbonaceous materials for oxidizer service are generally lapped dry on a clean lapping stone
and then wiped clean with a lint-free cloth moistened with trichloroethylene. Carbonaceous
materials generally are not cleaned by flushing in solvent because of the possibility of
reaction with the bonding agents used in some grades of graphite carbon. The carbon
wear-life may be decreased by exposure to cleaning solvents. Carbon used in oxidizer service
is machined either dry or in a compatible coolan t.
2.3.1.7 PASSIVATION
Materials to be exposed to liquid fluorine are passivated with gaseous fluorine prior to liquid
exposure to allow buildup of a protective fluoride film and to react any surface
contamination that may be present. The passivated surfaces must be protected from contact
with moisture to prevent its reaction with the fluoride film (ref. 19).
2.3.2 Face-Contact-Seal Rubbing Elements

Four configurations for the rubbing element of a face-contact seal are shown in figure 36: a
solid seal ring (fig. 36(a)), a seal ring insert (fig. 36(b)), a lapped-joint seal ring (fig. 36(c)),
or spray-coated seal ring (fig. 36(d)). The seal-ring-insert configuration generally is used for
64
I
INTERFERENCE
FIT
REll EF
FACE
WIDTH CARRIER
RADIUS
SEAL
INSERT - - -"
FACE
00
"--+--+--- RA 0 IUS '----ADHESiVE
~--NOSE HEI GHT < - - - - - - ' - - CHAM FER
FACE
10
'----LAPPED FLAT
(~) SOLI 0 SEAL RING

(b) SEAL RING INSERT
SEAL RING PLATE SPRAY COATING
l--- LAPPED-JOINT
STATIC SEAL
BEARING
SUPPORT
PADS
(e) LAPPED-JOINT SEAL RING (d) SPRAY-COATED SEAL RING
Figure 36. - Four types of rubbing elements for face-contact seals .
65
turbopump seals beca use of th e materi al requirements for the second ary elem ent and
antirotation ta ngs . Seal ma terials sa tisfac to ry for rubbing contac t cannot be weld ed to metal
bell o ws and may not have sufficient wea r re istan ce for th e secondary seal or imp act
strength fo r the antiro ta tion tangs. The lapped-join t seal ring is consid ered when distortion
of sea l-in se rt retain er is a problem. The spra y-coated seal fa ce or solid se al nng
configurati ons are pre f rred for th e ceramic and ce rm et m a terials.
2.3 .2.1 SEAL FACE WIDTH
The eal face wid th ge nerally i esta blished by the structural and face-con tac t unit lo ad
requirements; additi o nal factors are face di stortio n, hea t-transfer cap acity , pres ure balance
and vari ation o f inter face pre ure force, hy drod ynamic lift potential, wear, and leakage.
The current prac tice fo r re la ting face-co ntac t-seal fa ce width to face di am eter is shown in
fi gure 37. The facv width on curre nt designs vari e from 0 .040 in. on a 0.615-in . diam eter
al to 0.1 60 in . o n J 10. 142-in .-diam eter seal. The diam eter ratio (ID /OD) varies from 0.87
fo r the sm all seal to 0 .97 for the large seal. Mos t comm ercial eals th at op erate with
0 .18
0.111
0 .14
z 0 .12
-£
~
0 0 .1 0
~
w
<.>
<
u.
0 .08
...J
<
w
VI
0 .08
0 .04
0.02
0 2 3 6 8 9 10 11 12
" SEAL FACE DIAMETER . IN .
Figure 37. - Relation of face-contact-seal fac e widt h and d iam et er.
66
\
.-----
hydrodynamic lubrication and negligible face distortion have wider seal face (!D / OD::::::
0.75 to 0.85) to reduce the contact unit load and to in c rease the h yd rodynamic li ft so that
face wear is reduced.
The seal face width on high-speed cryogen ic sea l usually i a compromise among mll1lmUm
contact unit lo ad, h eat-transfer capacity, pressure - force balance, an d the effects of face
distortion. Wider eal faces decrease the unit load but result in face di tortion and greater
variation in interface pressure load. The variation of the average interfac p ressure profile
du e to face distortions and fluid vaporization causes a larger force variation because of the
larger face area. Therefore, the closing fo rce mu st be in creased to prevent face separation at
the maximum interface pressure; this increase may result in face overload at minimum
interface pressures. The effects of face disto rtion are amplified on wider faces by the large r
contact surface and greater temperature gradient acros the face. The optimum face width
for pressure-force balance and minimum distortion effects would be a knife edge.
Increasing the face width decreases the heat generation per unit contact area and all ows
more heat di sipat ion by conduction into the seal nose ring and rotating mating ring;
however, the heat dissipated by convection to the ci rculating sealed fluid decreases as the
face width is increa ed (refs. 6 and 7). Therefore, the optimum face width for heat transfer
is dependent on the relative cooling capacity of the heat flow paths. Since seals on cryogenic
turbopumps depend on the cooling capacity of the circu lating fluids, the face widths
generall y are narrower than those in most other app lica tions.
The face-contact unit load can be decreased without in creasing the eal face width by adding
pressure-vented support pads (fig. 18(b)) around the seal face to assist in load support. The
support pads reduce the seal no e structural req uirements and allow a narrow (0. 030 in.)
seal face to be utilized for more precise balancing of the pressure forces. This concept is not
widely used for insert-type c ryogenic seals b ca use of the possibili ty of face separatio n and
excessive leakage ca u ed by thenTla l di tortions. The di tortions m ay cause the eal face to
twist and be supported entirely by the pads. The concept ha been used succes fully in
gas-turbine engines and h as a potentia l capab ility for tu rbin e seals, p rovi led tha t
compensation is made fo r the thell11a l distortions (e.g., centroid alignment, lapped joint, or
spray-coat d or so lid seal rings).
2.3.2.2 SEAL NOSE HEIGHT
The eal nose height (fig. 36(a)) is established by the stru ctural requireme nt and the
allowance necessary for face wear. The stress level at the junction of the eal no e and the
base of the seal ring can become excessive as nose height is in crea ed. Thermal stre ses
develop as a result of the temperature gradient fro m the seal face to the base of the seal ring.
Bending and shear stresses result from the pressure deflection of the seat nose cylindrical
section and the deflections caused by in terference between the seal in ert and carrier.
67
Seal-nose cross sections with a height-to-width ratio of 0 .3 to 0.8 usually are employed on
turbopump seals to minimize fa ce distortions and nose stresses.
It is essential that th e junctions of the nose OD and ID to the seal ring base (fig. 36(a)) be
provided with minimum fillet radiu of 0.020 to 0.030 in. Without the fillet, nose breakage
may occur beca use of the stress-concentration factor at the sharp corn er. I t was necessary to
add the nose fill et on the H- l and J-2 liquid-oxygen seals to elimin ate fai lures. Th e edges of
the seal face are radiused 0.005 to 0.0 lOin . to elimin ate edge chipping caused by corn er
loading. All sharp corners on th e seal ring are either cham fered or radiused.
The seal nose height used on current turbopump seals to provide for face wea r ranges from
0.030 to 0.060 in. Turbopump seals are considered to be worn out when the nose is worn
within 0.0 lOin . of th e seal ring base.
The seal ling base on seal inserts must be a minimum of 0.005 in. above th e in ert carrier
(fig. 36(b)) to prevent the carrier from rubbing the m ating ring in th e event of nose wearout
or breakage. This design feature is particularly important on liquid-oxidizer sea ls because of
the com bustion hazard created by rubbing metals.
2.3.2.3 SEAL RING ANTI ROTATION DEVICE
Because of balance requirements and centrifugal forces , high-speed (> 10 000 rpm) seals
generally incorporate stationary seal rings and rotating mating rings. The stationary eal ring
is prevented from rotating by antirotation tangs or drive lugs betw een the seal ring and
housing. Metal-bellows-type seals do not require antirotation devices, because of the
torsional rigidity of the bellows element. Antirotation tangs have been used as vibration
damp ers on some bellows seals; however, the tang-to-slot clearances are very criti cal for the
tangs to be effective as a vibration damper without hanging up. Radi al splines h ave also been
used for a combination antirotation device, seal pilot, and vibration damp er.
Various arrangements of tangs, blocks, and pins that engage slots have been used on
0
turbopump seals. Usually two tangs or blocks are located 180 apart with a sufficiently large
contact surface area to prevent indentations in the slots from the impact loads. The
tang-to-slot clearance is minimized to reduce impact loads. Pins usually will wear an
indentation into the slots and thereby may restrict the axial movement. Also, high-impact
loads tend to loosen press-fit pins and may cause failure if the pins are not restrained. When
possible, the tang is made of material softer than the slot m aterial to minimize axial hangup
from slot indentations. To prevent t ang breakage on brittle m aterials such as carbon or
aluminum oxide, the an tirotation slots are located in the seal ring. Rad iused fillets are
provided at all comers. Materials resistant to fretting and galling are utilized for the
antirotation device.
68
. _ _ _ _ 1
2.3.2.4 SEAL RING PILOT
Except for metal-bellows seals, which are located by the bellows element, seal rings are
piloted by the stationary housing to maintain radial alignment. The pilots usually are
located at the secondary seal to minimize the relative radial motion required for the seal.
The pilot length-to-diameter ratio must be small (:::::: 0.05 to 0.1) to allow angular
misalignment of the seal face without interference at the pilot. Pilot diametral clearances of
approximately 0.003 in. / in. of diameter at operating conditions generally are adequate for
radial alignmen t.
The pilot clearance at ambient conditions is adjusted to compensate for the dim ensional
changes caused by thenllal contraction and pressure deflection. Close-fitting radial splines
that allow differen tial thel111al contraction while main taining radial alignm ent also have been
used to compensate for the dimensional changes at the seal ring pilot. The diametral
clearance of a 6-in. carbon seal ring in a stainless-steel housing decreases by 0.0 15 in . at
- 320° F as a result of differential thelmal contraction. The pilot clearance may be decreased
furth er if the seal housir~g is installed in an aluminum pump casing that is strong enough to
deflect the seal housing at the higher thelmal contraction rate of aluminum . The same
considerations apply to other radial clearances.
Turbopump seal pilots are subjected to high impact and vibration loads that can cause
fretting damage. Hard-chrome plating h as been effectively utili zed to elimin ate this kind of
damage.
2.3.2.5 SEAL INSERT RETENTION
Seal ring inserts generally are retained in the carrier with an in terference fit and adhesive
bond. The amount of interFerence depends on the required operating temperature, the
relative coefficients of contraction and expansion, modulu of elasticity , an d allow able
stress level. Seal insert materials usua ll y are stro ng in comp ression and weak in tension;
therefore, the inserts are maintained in compres ive hoop stress with the interference at the
OD of the insert. Carbonaceo us materials generall y have a low modulus of elasticity (1.5 to
3.0 X 10 6 psi) and therefore may be installed with relatively large interference fits (0. 003 to
0.006 in. / in. of seal diameter). The ceramic and cermet materials are limited to lower
interferences because of the high modulu of elasticity (50 to 60 X 10 6 psi).
The seal insert materials contract and expand less than most steel alloys; therefore, the
effect of the maximum operating temperature range (including the temperature increase
generated by friction) on the interference fit and stress level is considered. Special steels
with l ow contraction and expa nsion rates (Invar 36 , Carpenter 42 , molybdenum steel) m ay
be LI tilized to minimize the ch ange of insert in terference with tempera ture. Hea t- transfer
calculations or estimates of operating temperatures are made to ensure that ad quate
69
interference for insert retention remains at operating conditions . A minimum interference of
approximately 0.001 in ./in. of diameter usually is adequate for carbon inserts.
Seal inserts are bonded to the insert carrier to obtain additional retention and to provide a
positive seal between the insert and carrier. Leakage through the insert interference fit ha
been a significant problem on cryogenic seals that are not bonded or have defective bonds.
A liquid-nitrogen leak test to verify the bond sealing quality was incorporated into
acceptance pec ifica tions for the liquid-oxygen seals on the H-I and F-I engines.
The adhesive bond i applied to the bottom of the insert and the carrier counterbore (fig.
36(b» in accordance with the manufacturer' proce specification. The insert is installed by
pressing it into the ca rrier at room temperature. Heating the carrier or chilling the insert to
decrease the interference for ease of installation is not satisfactory because of the adverse
effects on the bond qua lity. The edge of the carrier counterbore is radiused an d the co rn er
of the insert i chamfered to all ow installation without damage.
The bond trength of epoxy adhe iv e drops sharp ly at high temperature (e.g., Epon 422
decreases from 2800 psi at 70°F to 500 psi at 800°F). Adhesive bonding is not used as the
on ly retention method on seals where the ins rt temperature may exceed approximately
300° F. Many adhesives are brittle at low temperature and therefore are not satisfactory in
cryogenic fluids. The adhesive generally used are listed in table Ill.
2.3.2.6 INSERT/ CARRIER SEPARATION
Insert-type e:l.\s that are exposed to pressure on the inside diameter may fail as a result of
the pressure sepa rating force between the insert and carrier. Either insert breakage caused by
co mplete separation or sea l face distortion caused by partial separation may occur. The
pressu re eparating force is minimized by making the insert OD the same a th seal no e OD
(fig. 38). The interference fit mu t be sufficient to ensure that the retention force is greater
than the separati ng forc at ope rati ng cond itions.
Epoxy ad h e ives are used to rninimize the possibility of in ert/carri er eparation by

excluding the high-pressure fluid from the joint; h owever , the porosity of the adhe ive may
allow seepage of the high-pressure fluid into the joint after prolonged exposure periods.
Mechanical lock are usecl for insert retention when it is not feasible to provide ufficient
retention force with an in terference fit. The mechan ical lo cks are sp ring lo aded against the
insert with a lo ad gr ate r than the sepa ratin g force in o rd er to prev nt slight movement that
can cause distortion of the eal face.
70
NOSE 00
2 NOSEOD~
INSERT INSERT
J 00 00
5 PRESSURE
PRESSURE~
SEPARATING PRESSURE
FORCE FORCE . FORCE
6
I
7
P=~!~~~~NG
~
8 FORCE
9
10
Lv Irm CARRIER
00
I
PRESSURE CARRIER [ [ [ D ' PRESSURE
L,~FORCE
11 00 FORCE
12
BELLOWS MED OR L.ELLOWS MED OR
SECONDARY SEAL DIA. SECONDARY SEAL DIA .
13
(a) Orig inal method (b) Improved method
4
F igure 38. - Method for reducing pressure separatin g force between seal insert and carrier
for sea ls with high er press ure at the inside diameter.
2.3.2.7 INSERT DISTORTION
Seal inserts with a thennal-contraction rate lower than that of the carrier material gene rally
are distorted when chilled to cryogenic temperature because of the bending moment created
by the increased interference and misalignment of the insert and carrier centroid s (fig. 39).
CARRIER
CENTROID CARRIER
INSERT CENTROID
CENTROID
INSERT
~ BENDING MOMENT
FACE CENTROID
DISTORTION
(a) Original method (b) Improved method
Figure 39 . - Method for reducing seal insert distortion due to thermal

contraction and centroid mi sa lignm ent.
71
The probl em is minimi zed by utili zing special st ee ls (lnvar 36, Carpenter 4 2, mol y bd enum
steel) that h ave co ntrac ti on rates similar to th at of th e eal insert. The ins rt and carri er
cen troid s are aligned as closely as possible to reduce the bending mom ent. The bending
force is minimi zed by redu cing the insert c ro s- ectio nal area to lower th e load required to
defl ect the insert. Th e ca rri er piece co ntracts without igni[i ca nt distortion wh en th e
bendin g mom ents are elimin ated .
The effec ts of seal- face therma l distortion ca n be minimized with prefere nti al lapping (cold I
lapping or taper lappin g) to co mpe nsate fo r th e di torti on. Cold lap ping is accomp li h ed by
chilling the seal in liquid nitroge n and lapping the face fl at whil e cold . The cold fl a tnes is
repeata ble when chill ed down in o peration. T aper lapping req uires the face to be lapped
with a reve rse tap er co rresponding to th e chilldown di sto rtion 0 th at the distorti on causes
the face to return to fl atn ess. Th e lapped taper ge nerall y is grea te r th an th e anticipated
distortion in order to pr vent face sep arati o n an d to all ow the face to wear in at th e actual
co ndi tions.
2.3 .2.8 LAPPED-JOINT SEAL-RING CARRIER ASSEMBLY
Seal-ring distortion can be redu ced significa ntly by utilizing a lapped-joint eal rin g (fig.
36(c)) that elimin ates th e th erm all y indu ced loads ca used by carri er contrac ti o n. The lapped
joint allows relative mov em ent betw een th e eal ring and carrier plate to comp ensa te for th e
differential therm al co ntraction. The joint is lo ad ed ax iall y by th e sealed pres ure to effec t a
static eal. The seal rin g is res train ed from rotation by a device that locks it to the h ousing
or th e ca rri er plate. On bellows seals, antirota ti on t angs to the hou ing a re preferred ,
beca u e they provid e vibration dampin g.
The press ure closing fo rce to lo ad the carrier plate aga inst the seal ring is provided by
relievin g either the seal ring or the ca rrier pl ate lapped urface to all ow th e pressure drop to
occur on a m ailer effec tive area than th e second ary seal or bellows effec tiv e area (fig. 40).
The reli ef diam eter is es tablish ed to provid e sufficient clo ing fo rce to effect a satisfac tory
stati c seal at th e lapped joint and to prevent sep ara tion und er adver e pre sure co nditi ons.
The relieved surface is provided with bearing supp ort pads to prevent [ace distorti ons from
th e pressure-force bendin g mom ents.
The lapped-jo int eal designs m ay be limited to lower pressure beca use of lack of stru c tural
upp o rt and relatively sm all cross section of the seal rin g. The effec t of pre ure de fl ectio n
m ay be significant on carbon seal rin gs as a res ult of the low modulu s of elasticity. Insid e
pres ure m ay also be a problem beca use of the low tensil e strength of m os t seal ring
materi als. These potenti al probl ems a re elimin ated by re in forcing the carb on rin g with a
me tal re tention band th a t maintain s the ca rbon in compressive hoop stress. The m e tal band
will ca use thermal di st o rtion of th e ca rbon unless the bending mom nt a re elimin ated by
centroid alignmen t.
72
l
SEPARATING CLOSING
PRESSURE PRESSURE
RESULTANT
PRESSURE
FORCE
SECONDARY SEAL OR
BELLOWS EFFECTIVE DIA
PAD
RELIEF PLATE
DIAMETER
Figure 40. - Pressure f orces on a lapped-joint sea l ring.
2.3.2.9 SPRA Y-COATED SEAL R INGS
Spray-coa ted seal rings (fig. 36(d)), instead of the insert designs, generally are used for the
ce ramic materials beca use of the retention, distortion , and Jea kage problems with cerami c
inserts. The ceramic materials cannot be attached direc tly to a metal-bellows second ary
elem e nt, an d sa ti sfacto ry econd ary-seal materials that allow the use of a solid seal ring may
not be availab le. Seal-face them1al distortion generally is not a problem with the
spray-coated seal rings because of the low resistance to defl ectio n offered by the relatively
thin cross section of m ate rial. The th em1a1 distortion is minimized by applying the coating
to both sid es of the seal ring to equ ali ze the them1al loads.
The ceramic coating is ap pli ed by either th e Dame-plating process or the plasm a-fl am e-spray
process. The plasma-spray process u ually is employed for aluminum-oxide coating beca use
it h as better resistance to thermal shock . Sin ce the ceramic materials h ave contra ction rates
lower than those of most metals, them1a1 tresses are developed at th e coating bond. The
stresses may be u fficien tJ y high to ca use bond failure or breakage of the coating. The
thermal st re s is minimi zed by using a thin coating (0.0 I 0 to 0.020 in.) that redu ces the
forces and the temperature gradient. Aluminum oxid e generally is sprayed on a transition
layer of ichrome to redu ce the thermal stres es and improve the bond trength . Pure
ichrom e i sp rayed on th e base m etal, fo llowed by a mixture of 50-p ercent ichrom e an d
50-percent aluminum oxide, and finally pure aluminum oxide. More gradual gradations m ay
be required for severe applica tions.
The seal rin g base mu st provid e a fo und ation to support the spray coating and allow
sufficient surface for bonding. A rai ed tapered tructure imilar to th at shown in figure
36(d) is preferred. All sharp co rn ers and edges must be radiused or ch amfered.
73
2.3.3 Circumferential-Seal Rubbing Elements
2.3.3.1 SEGMENTED SHAFT SEALS
Segmented shaft seals (fig. 20) provide an effective seal for low-viscosity fluids (gas, LOX,
LH 2 ) by adjusting to the shaft operating diameter and radial location to maiJ1tain rubbing
contact. Viscou fluids (RP-I and oil) tend to develop a hydrodynamic wedge that lifts the
segments and thu causes high leakage. The sealing-ring segment gaps are sealed off with
overlapping backup egmen ts. The segments are loaded radially against the shaft by a
circumferential garter spring and the unbalanced differential-pressure-induced force. The
segments also are loaded axially against the stationary housing by a wave spring and the
un balanced differen tial-pressure-induced force.
The forces induced by differential pressure may be partially balanced by relieving the
contact surfaces (fig. 41). The relieved surfaces are provided with bearing pads for load
support. It is not practical to balance the pressure forces completely because of the seal
construction; therefore, dry-nll1ning segmented seals are limited to lower pres ures (~ 100
psid) than are balanced face-contact seals.
GARTER
SPRING
LOAD
UNBALANCED
RADIAL
~
PRESSURE
LOAD HIGH PRESSURE
r------:::.t - WA VE SPR I NG LOAD (F )

w
LOW PRESSURE
FRICTION FORCE F
f
~g
-- UNBALANCED AXIAL
PRESSURE LOAD
F
f
= F
w
f ~
f = COEFFICIENT
OF FRICTION
Figure 41. - Pressure forces on a segmented shaft-seal ring (balanced design).
74
The garter-spring load is made higher than the friction force caused by wave pring in order
to relocate the segments and maintain shaft contact at low pressures. At pressure exceeding
50 psid, the unbalanced differential-pre ure load i sufficient to maintain the segment
loading. The unbalanced differential-press ure load is approximately 0.2 to 0.3 lbf/ in. of
circumference for each 10-psid pressure increment (assuming a eal nose width of 0.05 in.
and lin ear press ure profile). The resultant haft contact load must be consi tent with the
load-speed-life relations for specific material and fluids.
The segments are prevented from rotating relative to the h ousing by antirotation pins
located at th e egment gap. Severe conditions may require the use of pin block to
distribute the antirotation load over a larger contact area to prevent breakage of the
segments.
Th e design of segmented eals for extreme temperatures (cryogenic fluid or hot ga)
cons id ers the effec t of th e differential thermal contractio n or expansion of the segment
material on th e operating diam ter and inte rnal clearances. The design diamete r of the
segments i made equal to the shaft diam e ter at operating co nditions in orde r to minimize
the required wear-in and prevent breakage.
Additional discussion of the theory and practice of circumferential segmented shaft seals is
given in refere nce 30.
2.3.4 Hydrostatic/Hydrodynamic Face Seal Elements

Th e hydrostati c and hydrodynami c mechanisms for cont rolled fluid-film upport of the seal
face pote ntially arc capable of extending the pres ure-speed-life limitations of
rubbing-contact sea l and therefore are conside red for high-pres ure, high- peed , long-life
applications. The feasibi li ty or the hydrostati c/hydrodynamic concepts has been
de monstrated in several programs on development of je t engine com pres or and mainshaft
cal (refs. 31 through 36) and in other appli cations (ref. 37, 38 , and 39); however,
successful operation in a rock et engine has not been demonstrated.
Because the hyd rosta ric/ hyd rody nam ic seals are sens i tive to the adverse cond itions tha t
exist in turbopumps, anticipated developm e nt problems are related to rubbing material
co mpatibility in liquid oxidizer, wear of critica l face geometry by tran ie nt rubbing
contact, variation of th e ea l interface pressur profile cau cd by face distortion and
vaporization of rhe cryogenic tluid , dynamic in stabi lit y, low h ydrodynamic-Lift potential ,
and high lea "age ra tcs. The disastrous fa il ure mode res ul ti ng from face-geom e try dam age or
10 0 f Ii rt poten tial must be consiclered for re Iia bil i ty evalu a tions a nd fa ii-sa fe req u irem en ts.
• The theoretical a naly i of the hyclro tatic/ hydrod y namic principle is based on fundamenta l
lu brication and tlui I-!low theory that has been modified to ati fy the co nclition of a face
75
seal. The theory assumes steady-state ideal fluid conditions, which seldom ex ist on actual
turbopump seals. It is expected that empirical relations based on expe rim ental test resu lts
and fundamental th eOlY, similar to those developed for hydrostatic and h ydrod ynamic
bearings, will be required for successful turbopump seal design. These relations are not
cu rrently available for rocket engine propell ants. Theoretical methods and design
considerations currently available for hydrostatic/hydrodynamic seal lements are given in
references 31 through 53.
2.3.4.1 SELF-ENERGIZED HYDROSTATIC SEA LS
The self-energized hydrostatic seals (fig. 15) utilize the sealed pressure differential to
maintain controlled face separation. The lift force induced by the h yd ro tatic pressure
depends on a minimum pressure differential that ma y not exist until after rotation starts;
therefo re, transient rubbing contact usually occurs. The seal face material mu t be selected
for rubbing compatibility in liquid oxidizers and sufficient wear resistance to prevent
damage to the critical face geometry.
The self-compensating hydrostatic-pressu re-indu ced lift force requires leakage flow across
the seal interface to create a pressure profile proportional to the clearance gap. The
hydrostatic face seals generally operate with a larger effective interface clearance th an a
rubbing-contact seal; therefore, the leakage rate i significantly higher.
The recessed pads with olifice compensation (fig. 15(a)) develop elf-comp ensating
press ure-induced lift forces that are dependent on the relative flow between the ou ter sea)
face and the pad orifices. The pressure-indu ced lift force increases fo r small face clearances
and decreases for larger clearances so that force balance is maintained at the design
clearance. A minimum of three separate pads around the seal face is utilized to provide face
alignment stability . The volume of the recessed pads is minimi zed to prevent dynamic
instabilities caused by slow resp onse to pressu re changes (Jow fluid-film tiffness). The
dynamic stability is improved by eliminating the recessed pad s; how ever, the
press ure-induced lift forces become more difficult to predict as a result of indefin ite
pressure boundaries and flow effects . The orifice size must be large enough to prevent
clogging by contaminants in the sealed fluid.
The recessed-step (fig. 15(b)) and convergent-tapered-face (fig. 15(c)) designs develop
maximum pressure-induced lift force as the face clearance decrease , because the pressure
drop across the seal face is higher than the drop across the recessed step or convergent
surface. At larger face clearances, the effec t of the recessed step or co nvergent surface
becomes negligible, and the pressure drop occurs across the entire surface; this condition
reduces the pressure-induced lift force and main tains force balance at the design clearance.
•
The depth of the recessed step or taper must be very small (~ 0 .0001 t o 0 .0005 in. ), the
dim ension depending on the fluid -film design thickness required to obtain adequ ate
76
L__
fluid-film tiffness for dynamic tability. The fluid-film tiffness depends primarily o n sea l
face area, pressure differential, and fluid-film thickne s. There i an optimum reces depth
for a given design fluid-film thic kn e s and press ure differential.
2.3 .4.2 EXTERNALLY PRESSURIZED HYDROSTATIC SEALS
The externall y pressurized h ydro tatic seals (fig. 16) maintain contro lled face eparation by
fluid-film support from an external pressure so urce . The pressure-indu ced lift force increases
for small face clearances and dec reases for larger clearances to m aintain force balan ce at th e
design clearance. The ex terna! press ure SOurce a110w the sea l face to be Ii ft ed prior to start
of rotation to elimin ate transi ent rubbing contact. Sin ce the press ure-induced lift force i
indepe nd ent of th e sealed pressure, sealed-fluid viscosity , and rotational speed, thi concept
may be utilized when th e other types of hydrostatic/hydrodyn amic seals are not feasib le.
Howeve r, the ad diti onal comp lexity of the pre suri zing system, avail ability of the
pre suri zing fluid , and dilu tion of the propell ant with th e pressurizing fluid m ay pose
problems. Dyn amic in sta bility of the seal ring also has been a significan t problem . The
stability is improved by in creasing the fluid -film stiffnes by th e m ethods d esc rib ed for th e
self-energized hydrosta tic seal.
2.3.4.3 HYDRODYNAMIC SEALS
The h ydrodyn amic seal (fig. 17) maintains controll ed face separation by m eans of li ft forces
induced by the hydrodynamic pressure develop ed at the seal interface by the rotational
speed. The hydrodynamic lift is independent of the ealed pressure and is proportional to
the rotational speed. Therefo re, the hydrodynamic seal may be utilized at pressures below
the minimum required for hydrostatic action ; however, a minimum speed is required to
develop sufficient pressure-induced lift forces for face separation, and rubbing contact
gene rally occurs during the start and stop transi ents. The potential problems of
rubbing-material compatibility in liquid oxidizers and wear of the critical face geometry
exist. The available hydrodynamic lift i marginal with cryogenic fluids beca use of th e low
viscosity and vaporization of the fluid . Two-phase (liquid and vapor) fluids may disrupt th e
force balance by their effects on the interface press ure profile. The hydrodynamic lift fo rce
and fluid -film stiffness are maximum at minimum fac e clea rances, and the rate of in crea e is
high ; therefore, hydrodynamic seals generally are more stable than the hydrostatic typ e at
low clearances and low press ures.
The shrouded-Rayleigh-step concept (fig. l7(a) ) is press ure balanced like a conve ntional
rubbing-contact face seal ; therefore, the failure mod e caused by loss of lift pote ntial is not
as disastrous as that of the other concepts. T es t programs (ref. 34) have indica ted that thi s
concept is ca pable of d eveloping hydrodynamic lift with low-viscosity fluid s (l OaO°F gas) if
compensation for seal face distortions can be provided. Because sea ls for cryogenic
77
____ _ ~J
turbopumps also operate with thermal differentials and low-viscosity fluids, it is reasonable
to assume that the Rayleigh-step concept has potential capability for turbopump seals (ref.
39). Other hydrodynamic mechanisms (surface waves, surface micro asperities,
nonsymmetric rotation) have been investigated; however, the curren t state of the art is not
sufficiently developed for practical application on seals for rotating shafts in rocket engine
turbopump .
2.3.4.4 HYBRID SEALS
The hyblid- eal concepts (fig. 18) utilize the com bined hydrostatic and hydrodynamic or
combined rubbing-contact and hydrostatic/ hydrodynamic pIinciples. The hydrostatic action
provides the lift force at low rotational speeds, and the hydrodynamic action provides
additional lift force and dynamic stability at high speed independent of the sealed pressure.
Use of the concept shown in figure 18(a) has resulted in improved operation with mix tures
of oil and gas at low pressures (ref. 45). The concept has not been evaluated with cryogenic
propeliants.
The combined rubbing-contact and hydrodynamic concepts (fig. 18(b)) offer the advantage
of a seal designed for rubbing contact to allow for transient operation and for reduced
face-contact load at higher rotational speeds as a result of the additional hydrodynamic lift
force. The reduced face load allows higher speeds and longer wear life than the conventional
nlbbing-contact seals. The potential problems with this concept are thermal distortions of
the eal face by frictional heat and wear of the clitical face geometry. The depth of the
tapered lift pads must be approximately 0.001 in. or les with low-viscosity fluids to develop
significant pressure-induced lift forces; therefore, the allowance for face wear is very low.
Other methods (e.g., use of face grooves) provide more wear allowance but less lift force
induced by hydrodynamic pressure.
2.3 .5 Circumferential-Seal Clearance Elements

Clearance-type seals are used when the pressure-speed-life limits of rubbing-contact eals are
exceeded and the increased leakage is acceptable. Circumferential-seal clearance elements
consist of labyrinth devices, floating rings, and arch-bound* segmented rings. Labyrinth
seals have the best reliability, but the leakage is approximately ten times greater than that of
the arch-bound seals and approximately five times greater than that of the floating-ring
seals. Floating-ring seals generally provide the best compromise between sealing effectiveness
and reliability for high-pressure, high-speed, long-life applications. The arch-bound
segmented-ring seal provides the most effective sealing of clearance-type seals; however, the
potential reliability is lower becaus of the more complex design and required wear-in.
'Ends of the segments bottom Ollt against each other, so that there are no gaps in the ring (fig. 21).
78
l
2.3.5.1 LABYRINTH SEALS
Labyrinth seal elements (fig. 42) are clearance devices that restrict fluid leakage by
dissipating the kinetic energy of fluid flow through a series of flow constrictions and cavities
that accelerate and decelerate the fluid or change the direction of flow abruptly to create
the maximum flow friction and turbulence. The ideal labyrinth transforms all of the kinetic
energy at each throttling into internal energy in each cavity . Practical labyrinths, h owever,
generally transfer significant kinetic energy from one throttling to the nex t. Therefore, the
ideal thermodynamic and fluid-mechanics relations are modified with empirical factors for
practical solu tion.
~ ~~ ~
V777T/T//T/77l 17 / / / / / / / /1
ANGLEO TEETH
STRAIGIfT LABYR INTH STRAIGIfT LABYRINTH
ABRAOABLE
STAGGERED LABYRINTH INTERLOCKING LABYRINTH WEAR - IN LABYRINTH
Fi gure 42. - Six configurations for a labyrinth sea l.
Empirical relations for compressible fluids are available for m any different types of
labyrinths ; however, data on incompressible fluids are more limited, and usually it is
necessary to either conduct experimental tests to establish the flow factor or estimate a
factor from extrapolated data.
The theoretical and empirical method used to op timize the laby rinth geometry and predict
leakage are given in references 54 through 65. Em pirical flow coefficients for incompressible
fluid based on experim ental tests in water are given in fig ure 43 fo r some typical turbopump
labyrinth seals.
79
r-
~~
1.0
0 .9
Q) h = 0 .020
O. B
Lh
~=tn -
0.7
® h 0 .033
0 .6
<D h ~
0 .020
~=th
-a @ h = 0 . 033
0 .5
t-
Z
w
@ h ~
0.019
U 0 .4
I 1-1. G) h = 0 .019
11. I I ~L h
11.
W .~-- .. - -... @ h = 0 .031
~*h
0
U II :.' ® h - 0 _019
l ·:'· .
0.3 - @ =
i:
~IJ1llJt- . :
h 0 .032
- .-'
0
~~h -
...J
00 h
0
\L
(i) 0 . 033
...J
~
-+~I-t " 1 . l.~-: f-:f': - :L:-________

W
l
(J)
0.2
LABYRINTH DIAMETER 9.0 IN .
'-" \ ' \" I ,

:-1 :ji !. TEST FLUID
TEST SPEED
WATER
3600 RPM \150 FPS)
- '--. - .~- ~-I - -r :r-,

1 ':
I
.
I,
-rI :·- -: :~:~::' _
1 - - - -__
_ 1
1
j',:.::_.1
I __ _! .
! \.
j ! ' __ ,:_
4J - ';2g
Q
~H Q -
3
FLOWRATE . IN . /SEC
iLLl4'.1-+4:
-.jI -. - - . 2
A - CLEARANCE AREA. I N.
0.1
2 3 4
i -'4
5
' 6 - -+'""H;--,
-t 7 B 9 10 x 10
4
Qd
R. - All
~H
d -
- DIFFERENTIAL HEAD. IN .
DIAMETRAL CLEARANCE . IN.
II - KINEMATIC VISCOSITY . IN?ISEC
REYNOLDS NUMBER R.
Figu re 43. - V ar iation of sea l flow coefficient with Reyno ld s number,

var ious labyr inth sea l configurat ions _
2.3 .5.1. 1 Labyrinth Geometry
Optimization of the labyrinth geometry has a significant effect 011 ealing e ffec tive ness.
Lea kage flow through a step or staggered labyrinth is approximately 50 perce nt of th e
leakage of a straight labyrinth for similar cond itions (fig. 43). Step labyrinths require more
radia l space, are more difficult to manufacture, and may produce an und e irable thru st load
because of th e unbalanced pressure force. Stagger d labyrinths are more compl ex to
machine and require special assembl y t chniques.
The relative sharpnes of the tooth (ratio of tooth tip thickn ess to clearance) can vary the
leakage as much as 20 percent (ref. 60). The tee th tips usually are mad e sharp (0.005 to
0.0 15 in. rad.) to minimize the flow coefficient and ru bbing ar a for improved wea r-in
capability. Test data indicate that a tooth angle of attack of 40° provid es th e optimum
resistance to flow (ref. 62).
Leakage is reduced significantly by increasing the number of tee th or throttlings ; how ever,
the effect diminishes with large nu mbers of teeth . The vari ation of the leakage fun ct ion l/J
with throttling is shown in figure 44.
'.0.--- - - - - - - - - - - - - - - --,
t----_ Num ber o f thro ttl ingl "t
..---. N
Q.
-N - O.S
Q. ~
..... ~
!:. .f
I +
c~
0 .6
~
0:)
Z
o
i=
u 0.4
Z
::J
u.
w
(!)
«
>l
«
~
0.2
r--------------E~ __
o 0 .2 0.4 0 .6 O.S 1.0
LAB Y RINTH PRESSURE RA T IO P, / P

2
Figure 44. - L eakage fu nction for labyri nt h sea ls as a function of

labyrinth pressure rat io (adptd. f ro m ref . 56).
81
I
I
I
Tooth pitch affec ts th e inte rstage cav it y size and i opt imi zed re lat ive to tile number of
tee th for a give n sea l le ngth. The e ffec t of tooth pitch is m o re sig ni ficant on straight
labyrinths th a n on step lab y rinth. The op tilllulll pitch for straight laby rinth s i displayed on
fi gure 45 as a function of di am e tra l c lea rance. The pitc h o n step lab y rinth s is minimi zed to
obtain th e m ax imum numb e r of con tri c ti o ns. Th e o ptimum cav it y depth is app rox im ate ly
equ al to th e tooth pitc h (ref. 6 \ ).
0.30
~
z
:£ 0.25 - -
u
I-
/
0:
~
:J 0. 20
~
I-
0..
0
/
J:
I- 0.15 / -- -
1/
Z
a:
>-
'"...J
« 0. 10 r-
I-
J:
l?
«
a:
0.05 /
I-
(/l
/ 0. 010 0. 020 0.030 0.040

DIAMETRAL CLEARANCE , IN .
Figure 45 . - Optimum pitch for teeth of a strai ght labyrinth as a function

of diam etra l cl earance (adptd. from ref . 56).
Changing th e lo ca tion of th e lab y rinth t eeth from the rotating to th e stati o nary p art or vice
versa doe not appear to h ave a significant ffect on sealing effec tiv e ne . Th e primary
co nsid e rations are th mate rial co mbin a tion s for pro p e r wear-in , m a nufac turing diffi c ult y,
e rosion or cav ita tion re ista nce, a nd pote nti a l h andling damage to the sharp tee th. The tee th
generally are locat ed on the rotatlng part for we ar-in-ty pe lab y rinths th at utili ze a soft
a brad a ble or thin hon eycomb m a te ri a l for the t at ionary p art. L ocat in g the tee th o n th e
stationary part reduces th e prob a bility of handling damage to the sh a rp tip a nd a ll ows ea y
re placement of a less ex pensive p a rt in th e ev e nt of d am age. L oca tin g th e teeth o n th e
outside stationary me mb e r te nd to minimi ze the d am age from rubbin g co ntac t , prov id ed
that the rotating surface is wea r resi tant a nd the supporting tru c ture for th e sta ti o na ry
member is fl ex ibl e . Dim e nsional stability and fatigue may be problems with fl ex ibl e sea l
memb e r.
l
Th e the rm a l ex pa n io n ca used by th e hea t ge ne ra ted thro ugh rubbing co ntact is ta ken into
acco un t in o rd e r to p reve n t d am age ca use d by the co mp o undin g tend e ncy of d ecreased
clea ra nce a nd additi o nal rubbin g. Th e t empe ra ture o f the tee th gen e rally in crease rap idly
beca use of th e mall sec ti o n in rubbing co n tac t, except whe n th e sealed fl uid h as su fficie nt
coo lin g ca pac it y to m a ke use o f th e la rge ex posed urface a rea a nd tra nsfe r th e hea t from
th e too th a rea qui c kl y. A h ea t-t rans fe r a nalysis i required t o eva lu a te th e th e rm al e ffect
ca used by ru bb ing co ntac t.
Th e o pe rat in g clea ra nce a nd sea l di ame te r a re minimi zed fo r m aximum sea ling effective nes .
Th e leakage is direc tl y prop ortion al to th e o pe ra tin g clea ra nce (as uming a consta nt fl ow
coefficie n t) and ap p rox im a tely pro p o rti o nal t o th e qu are of th diam e te r. Sm all seal
di a me ters redu ce th e fl o w area a nd gen erall y all o w closer cleara nces as a result o f be t te r
dim c nsio n al co ntro l. Th e clea ra nce ca n be m ad e to in c rease o r dec rease a t o p era ting
te mp erat ure by select in g m a te ri al with diffe re nt th e ll11 a l ex p a nsio n ra tes. In om e cases it
m ay be adv a ntageo u to provid e a la rge c lea ra nce for as embl y and utili ze the
th e ll11 al-ex pa nsio n diffe re nti al to redu ce th e cl ea ra nce at op e ra tin g co nditi o ns. The effect
o f press ure d e tl ec ti o n a nd ce ntrifu gal growth o f the roto r a re also co nsid e red .
2.3.5.1.2 Wear-in Labyrinths
Wear-in la b y rinth s a re used to mJl1In1J ze the op e ratin g clea ran ce. Co nve ntion al lab y rinths
req LI ire ad d i ti o nal c lea ra nce to all o w fo r accLlm u la tion 0 f dim e nsio nal to lerances, d y n amic
d e fl cc ti o ns, and th erm al diffe re nti als. Th e wea r-in la by rinths can be installed with
prac ti call y ze ro c lea ra nce o r eve n a sli ght inte rfe re nce. Closer op era ting clea ran ces are
o btain ed if the radi al loca tion of th e sea l h ou in g is adju stabl e a t in st all ati o n t o comp ensa te
fo r co nce ntric ity tol e ran ces. Th e m a te ri als a re selec ted to d e fo rm easil y or wear away during
initi a l co ntact, 0 th at minimum o pe ratin g c lea ra nce a nd negligibl e d am age t o th e rubbing
pa rts are e n ured. Th e gro oves fOlm ed by th e wea r-in o n a traight laby rinth redu ce the flow
coeffi cie nt to a va lu e be twee n th a t o f a traight la b y rinth and th a t o f a st aggered la bYlinth .
Th e fl o w coe ffic ie nt w ill in c rease if th e sh all) edges of th e la b y rinth tee th are ro und ed off
durin g wea r-in .
Turb o pump wea r-in lab Ylinths fo r n on-ox idi zin g fluid s gene rall y h ave a strip of m e tal
ho neycomb ma te ri al ( ref. 66 ) for th e sta tion a ry part. The hone y co mb (fig. 4 6) usu ally
co nsis ts of ln co nel 600 , Ha t ell oy C, or st ainl ess- tee l foil 0.00 2 t o 0 .005 in . thi c k with a
1/ 16 -in. ce ll width ; ce ll d epth is o ne to tw o tim es th e width . The cell s are o ri e nted n0 n11 al
to th e direc ti o n o f ro ta tion . Foil thi c kness and cell size are a trad eo ff be tw ee n ease of
de fo rm a ti o n and e rosio n res ist a n ce . Th e h o neyco mb strip usu all y is brazed to th e supp o rt
rin g. Th e adv a ntage of th e h o ne yco mb mateJi a l i th a t the ed ges of th e thin fo il cells are
eas il y d e fo n11 ed o r be nt by th e initi al contac t o f the sh arp la b y rinth tee th o r sh a ft 0 th a t
o nl y th e required o pe ra tin g clea ra nce res ults. The fac t th a t th e fo il i d efo nll ed ra th e r th an
wo rn a way preve nt co ntamin a ti o n o f th e fluid sys tem with wea r d e bris.
3
DETAIL OF
HONEYCOMB
HONE Y COMB
BRAZE
SUPPORT RING
CASING
OPTIONAL ATTAC HI NG
METHODS
Figu re 46. - H oneycomb construct ion for stationary part of a

wear-in labyrinth seal (ref. 66).
Abradable materials const ru cted from sintered metal fibers h ave been used uccessfully for
wear-in labyrinths; h owever, the wear debris m ay be a problem in some app li cations.
Carbonaceous ma terials also have been used to provide wear capabi lit y; however, carbon is
more wear resistant than the honeycomb or ab radab le mat lial , and damage to the eal may
occur during wear-in if th interference i too large. Carbon h a a lower expansion rate than
teel, is weak in tension, a nd is re latively brittle . Therefore, the carbon u ua ll y is pressed
into a stee l ring that provides st rength and thennal-expansion contro l.
2.3.5.1.3 Plastic Labyrinths
Laby rinths fo r liquid oxidizer require e ith e r (I) ve ry large c learances to e limin ate the
possibility oC rubbing co ntact o r (2) compatibl e material that can rub w ithout the h azard
or combu tion. In c ryog nics, the low temperature ( - 297° F) a l 0 limits the number of
usab le m aterial because of thermal-contraction differentials and loss of elongat io n. K el-F
m ate ri a ls a re compatibl e with liquid oxygen; h owever, their thelmal contract ion rates are
approx im ately four times g reater than that of steel, and the materials a re not structura ll y
table. PIa tic rin gs ca nn o t be reta in ed with a n interference Cit, because of co ld now at
ambient or elevated temperatur . Th e interrcrence fit i decrea ed after temperature
cyc lin g, a nd the g reater co ntract io n rate cau es the plastic to loose n at low temperature .
Th e Ke l-F tructural problem was so lved on t he J- 2 a ncl F-I programs by locking relatively
thin ect ions of Ke l-F into a metal h ous in g with the rubbin g portion expo ecl (fig. 47). The
thin plastic cction is rest ra in ed from its natural thermal contraction rate with retention
lock on the metal hou sing. Dim e nsio nal stab ility a nd thermal cont raction rate thus a re
determined by the tronger housing. The housing is designed to provide a tight fit to the
plastic. A stat ic sea l nange and pressure vents are provided on high-press ure sea ls to
minimize th e st resses on the Kc l-F cau ed by the diITcrential-pressure-induced load. The
span length between the retention lock is made hort (~ 0.5 to 1.0 in.) to minimize the
84
HIGH PRESSURE
METAL HOU S ING LOW

PRESSURE
HIGH
PRESSURE
S IDE
RETENTION
L O CK (TYP)
S URFACE (KE L . F)
LOW PRESSURE
SIDE
HIGH PRESSURE
PRESSURE PROFILE
Figure 47. - Configuration and pressure profil e for a I'estrained pressure-balanced

plastic wear-i n labyri nth seal.
stres o n and llelkction or th e Kel- ca u cd by th e th e rmall y indu ced load. The wear-in load
is minimized by locating the lab y rinth teeth in the o ft e r pi a tic mate rial. Improved
resi. tance to e ros io n an d cav ita ti on damage is ob tain ed by locating the teeth in th e meta l
1'0 tor.
2.3.5.1.4 Segmented Labyrinths
The o ute r ring of wear-in 1 ~lby rinlh s call be segill ellted and spring loaded to provi Ie
addit io nal all owance 1'01' r,l Liial Illi sa li gnillent. Th e labyrinth segme nts arc pu shed radiall y
inward by the sprin g alld pressure load to a d ia mete r small e r than the roto r. The addit iona l
we,Jr- in depth provides Illinimum e lTect ive clearan ce 1'0 1' large ecce ntriciti es. Th e segme nts
arc rree to be pu hed radiall y o utward by th e ro to r to p reve nt excessive co nta ct loads during
wear-in o r tran sie nt opera ti o n. Hi gh-pressure ea ls may require pre ure balan cing to preve nt
excess ive co ntact loa ds durin g wea r-in . The segme nt di ame te r afte r wea r-in is co ntro ll ed by
limitill g the radia ll y inward trave l with a T-slo t arrangem e nt (fig. 4 8) o r by all ow ing the
egments to beco me a rch bound . A lapped jo int 1'01' a stati c sea l is provid ed be tw ee n th e
egille ll t. ami ho using. Antirota ti o n dev ices a re provid ed to preve nt segment ro tat ion.
85
STATIONARY
HOUS ING
W A VE SPRING
RADI AL ST O P
(T Y P) SEGMENTED RING
A BR ADAB L E
MATERIAL
L A PPED JOIN T HIGH PR ESSURE

ROT O R SIDE
Figure 48. - Configuration of a spring- loaded, segmented, wear-in labyrinth seal.
2.3.5.1.5 Labyrinth Erosion Resistance
Labyrinth devices, in general, provide the greatest life and reliability of all seal types.
However, erosion or cavitation damage and fatigue cracking may limit the operating life.
The high fluid velocities caused by large pressure differentials may result in erosion of the
labyrinth surfaces. Fluid turbulence in the labyrinth can create localized low-pressure
regions where the pressure is below the fluid vapor pressure. The energy released by the
collapse of the vapor bubbles may erode the exposed surfaces. Surface erosion apparently
attributable to surface fatigue failure is also caused by impingement of high-energy fluid .
The theory of erosion is controversial, and very little data for labyrinth devices are available.
A general study of erosion (ref. 67) indicates that the erosion resistance of different
materials increases approximately with the 2.5-power of hardness and linearly with material
strength. The harder high-strength materials (tool steel, Stellite, maraging steel) are more
resistant to erosion than th e soft low-strength materials (aluminum, Monel, brass, bronze).
The erosion rate generally starts slow, increases to a maximum value, and then decreases
with increasing exposure time . Some hard and brittle materials (tungsten carbide, titanium
carbide, ceramics) appear to have good initial erosion resistance, but erosion rates increase
with exposure time. Additional discussion and references on erosion are given in reference
67.
Fatigue cracking of the labyrinth structure may occur if the flow frequencies correspond to
the natural frequency of the structure. Flow turbulence may excite a resonant frequency
vibration in the labyrinth components. Stiffening the structural members generally prevents
vibration failures of labyrinths.
86
'-------
2.3.5.2 FLOATING-RING SEALS
Floating-ring elements (fig. 23) consist of an inner carbon ring for wear resistance and an
outer steel ring for strength and thermal expansion/contraction control. The outer ring
material usually is selected to provide the same thermal expansion and contraction rate as
the shaft material, so that a constant clearance gap is maintained as the temperature
changes. The outer ring is sufficiently strong, relative to the inner ring, to control the
diameter of the composite ring.
Extreme temperature environments may result in a temperature differential between the

floating ring and the shaft that will cause a variation of the clearance gap. If the ring
temperature is lower than the shaft temperature, the clearance will decrease, and ring seizure
may occur. This potential failure mode is minimized by allowing sufficient initial clearance
to compensate for the maximum thermal gradient.
The inner ring is maintained in compressive hoop stress with an interference fit. The amount
of interference is established to maintain the minimum contact unit load higher than the
maximum fluid pressure. The fluid pressure acts on the interference joint and causes radial
deflection of the inner ring if the interference unit load is lower than the fluid pressure.
With a carbon inner ring and steel outer ring, the interference load will decrease at higher
temperature because of the difference in thermal expansion rates. An e'x treme range of
operating temperatures requires selection of an inner ring material that has a thermal
expansion rate similar to that of the outer ring.
The load induced by unbalanced radial pressure (fig. 49) is supported by the composite ring
in compressive hoop stress. The radial deflection caused by the compressive stress is
proportional to ring rigidity. The radial section and modulus of elasticity are selected to
minimize the deflection. The initial clearance is adjusted to allow for the deflection and
provide the desired operating clearance.
The axial force induced by differential pressure (fig. 49) loads the floating ring against the
stationary housing to provide a static seal. Low-pressure seals require a wave spring to
provide sufficient contact load to maintain a static seal, High-pressure seals are pressure
balanced by relieving the axial contact surface and minimizing the housing-to-shaft clearance
to reduce the unbalanced load induced by axial pressure. Because of the increased friction
force , high axial loads increase the radial load required to reposition the floating ring. The
seal-ring wear rate is higher when the friction load is excessive because of the larger wiping
force exerted by the shaft required to center the ring with the center of rotation.
The floating-ring element usually is restrained from rotation with two or more antirotation
tangs or pins that engage slots. Unrestrained rings have been used; however, if the ring
rotates with the shaft as a result of partial seizure or sticking, failure generally occurs. The
centrifugal force on high-speed seals may cause excessive deflection and failure . In the
87
L _J
CARBON
RI.N G
UN BALANCED RA D IAL~
PR ESSU R E LOAD
HIGH PRESSURE
!
t:s:,,,,"-.. "-..
LOW PRE SS UR E
L UNBA L A NCED AXI A L
PRESSU R E LOAD
j
FRICTI O N
FORCE
Figure 49 . - Pressure forces on a floating-ring seal.
turbine on the H-I engine, thermal decomposition of the lubrication additive led to shaft
sticking that caused centrifugal failure of the floating ring; this problem was solved by
adding anti-rotation tangs to the ring. Machined tangs generally are more reliable than
press-fit pins because the pins tend to loosen with repeated impact loading.
Additional discussion on controlled-gap seals is given in reference 68 _
2.3.5.3 ARCH -BOUND SEGMENTED SEALS
Arch-bound segmented shaft seals (fig. 21) are designed such that the sealing segments are
butted together to form a solid ring at the operating diameter. The differential
radial-pressure load (fig. 50) is supported by compressive hoop stress in the ring instead of
bearing contact on the shaft surface. The reduced rubbing-contact load increases the
pressure-speed-life limitations over those of conventional segmented shaft seals.
To allow the segments to wear in, the design diameter of the arch-bound segments at
operating conditions is slightly smaller than the shaft operating diameter. To minimize the
required wear-in, the segment design diameter is adjusted to compensate for the deflection
caused by the load induced by radial pressure and differential thermal contraction or
expansion . The bearing contact load decreases asymptotically during wear-in until most of
88
UNBALAN CE D
R ADI AL
PRESSURE
LOAD
1IIIill HIGH
P RESSUR E
LO W
PRE SS URE
Figure 50. - Pressure-induced radial forces on an arch-bound segmented seal.
the load induced by radial pressure is supported by the butted segments in compressive hoop
stress. The segment wear becomes negligible as the contact load approaches zero.
The seal construction is similar to that of the conventional circumferential segmented seals
except for the butted joints. Since the sealing ring joints are butted together, additional
backup rings to seal the joints are not required, and the design may be simplified by using a
single ring similar to that in figure 21. A garter spring around the outside diameter is used to
hold the segments together prior to pressure application _ The segments are prevented from
rotating relative to the housing with antirotation devices. The axial force induced by
differential pressure is minimized by pressure balancing as in the floating-ring seal design. A
steel washer and wave spring ordinarily are used to hold the segments against the axial seal
surface prior to pressure application.
The contact load on high-pressure, high-speed seals may be excessive during wear-in, and the
contact surfaces may be damaged before the wear-in is completed. Damage is prevented by
conducting the wear-in at lower speed and pressure or by adjusting the design diameter until
a line-to-line fit with the shaft diameter is obtained at operating conditions. The segment
design diameter may be made large enough to provide a slight clearance at operating
conditions to eliminate wear-in_ Some trial-and-error adjustments generally are required
because it is difficult to predict the exact diameter at operating conditions.
2.3 .6 Face-Contact-Seal Secondary Elements
Face-seal secondary elements are devices that seal between the seal face ring and the
stationary housing. The secondary element provides for the motion required for the seal
89
-----,
face to track the rotating mating-ring surface and the axial travel necessary to compensate
for face wear and relative movements between the shaft and housing. Secondary sealing
must be maintained during reciprocating br wobbling motions without creating excessive
friction drag or hysteresis. High secondary-seal friction requires increased spring force to
provide the dynamic response necessary to maintain face contact. The larger spring force
results in a higher face-contact load, which lowers the pressure-speed-life limits. Therefore,
on high-speed, high-pressure, or long-life seals the secondary seal friction is minimized. The
secondary elements generally used for turbopump seals consist of metal bellows, plastic lip
seals, elastom eric O-rings, and metal piston rings.
2.3.6.1 METAL BELLOWS
Bellows elements are designed to provide the secondary-seal function, the spring force
required to maintain face contact, and structural support for the seal face ring. The bellows
element is extendible in the axial direction to compensate for relative axial motions, and
sufficiently rigid in the torsional and lateral directions to act as the antirotation and piloting
device for the seal face ring. The bellows element thus allows significant simplification of
the seal design by eliminating the secondary seal, antirotation device, seal ring pilot, and
loading springs. The all-metal construction eliminates periodic replacement of aged
elastomer components; it also extends the minimum operating temperature from - 65° to
- 423°F and the maximum from 500° to 1500°F: Many turbopump applications require
all-metal construction because of the extreme temperatures and reactive environments.
Three different basic types of metal bellows are used (fig. 51): welded, formed, and
machined . Welded bellows (fig. 51(a)) generally are used because they require significantly
smaller space for a given spring rate, compression range, and pressure capacity. The
welded-bellows element usually can be designed to require less space than conventional
elastomeric seals. Since the available space 'on turbopumps usually is limited, the smaller
space requirement and lower spring rate of the welded bellows are significant considerations
and may be determining factors.
The configuration of welded bellows is varied to satisfy specific requirements. Five designs
are shown in figure 51(a). The nested ripple design is used to minimize the spring rate and
compressed length; the pressure capacity of the nested ripple design is significantly increased
relative to the spring rate by the use of double-ply bellows. The single-sweep design provides
higher pressure capability and less change of bellows mean effective diameter (MED) with
pressure. The flat-plate design provides a constant MED at low pressure differentials. The
toroidal design provides constant effective diameter and high-pressure capacity; however,
because of the larger space and the higher spring rate required, toroidal bellows generally are
not suitable for use on turbopump seals. The nested ripple design often is the best
compromise for turbopump seals and is the most commonly used configuration.
90
L_
NESTED RIPPLE
Minimum spring rate and co mpressed length s. Large .

co mpress ion range. Requires MED (Mean Effective Di ameter)
ca libration. Fat igu e life difficult to predict.
SINGLE SWEEP
High er pressure capab ility, less va riati on of MED , and higher

spring rate than nested ripple.
DOUBLE PLY
High er pressure capability and lower spring rate than above.

Requires MED cali brat ion.
FLAT PLATE
Constant MED for low pressures . Limited compression range

and pressure ca pab ility.
g
TOROIDAL
Constant MED and high pressure ca pability. High spring

rate and limited compression range. Large space required.
(a) Welded
U-SHAPED
Most rel iabl e and least expens ive. High spring rate ; large
space required. Requires MED calibration.
(b) Formed
RECTANGULAR
Con stant MED and h igh pressure capability. Very high

spring rate and limited compression range . Th ickness variations
due to mach ining tolerances .
(c) Machined
Figure 51. - Vari ous configurations for m eta l-bellows secondary eleme nts for face seals .
91
If space is not limited, formed bellows (fig. 51(b)) generally are used. The fatigue life of
welded bellows is difficult to predict because of the variation of the stress-concentration
factor at the weld joint. The fonned bellows is more reliable and less expensive than the
welded type; however, the stress level is al 0 difficult to predict because of thickness
variations and residual stresses caused by the fom1ing process. Machined bellows (fig. 51(c))
generally are not used on turbopump seals becau e of the high spring rate and thickness
variations caused by machining tolerances. Many valiations of the basic types of bellows are
available for special applications.
The analytical techniques for the design of metal-bellows diaphragms are discussed in
reference 69. The theoretical analysis of bellows tresses, derivation of the mathematical
formulas, and design cliteria for some selected configurations are given in reference 70.
Simplified fom1Ulas and curves for bellows analysis are presen ted in reference 7l. The
theoretical relations are based on specific conditions and generally require empirical
coefficients for each bellows configuration. Most bellows suppliers have developed empirical
relations for the specific configurations they manufacture. The supplier's data usually are
more accurate than the values provided by the simplified theory.
The theoretical relation of the design variables for the bellows axial spring rate is given by
the following equation (adptd. from ref. 70):
0.431 REf n
K = P (12)
a
where
Ka = axial spring rate, Ibf/in.
R = bellow mean radius, in.
E = Young's modulus, psi
t = bellows plate thickness, in.
np = number of plie
n c = number of convolutions
.
s = half of bellows pan (1.e.,
radial OD - radial
2
ID) ' .
111.
C f = constant for bellows configuration
92
L _ __
I
I
I The bellows stresses due to pressure and deflection are estimated by utilizing the referenced
I theory and the manufacturer's empirical data for the specific bellows configuration. The
pressure-induced stresses increase as the bellows span, pitch, and diameter increase.
I Increasing the plate thickness and number of plies reduces the pressure-induced stress. The
deflection stresses increase as the deflection and the plate thickness increase . Increasing the
span and number of convolutions reduces the deflection stresses. The total stress in a
bellows is the sum of the pressure and deflection stresses. Since the variables are
interrelated, it is usually necessary to iterate the design to minimize the stresses.
The relative stress level or pressure capacity is evaluated by pressurizing the bellows and
measuring the pennanent deformation or change of free length. Welded bellows usually
yield locally at the initial pressurization; this yield ing causes the free length to increase or
I
Q
decrease, the change depending on the bellows design and direction of pressure. To prevent
I additional defOlmation during operation, the bellows usually are stabilized with a proof
pressurization at the minimum operating compression, followed by stress-relief heat
treatment. The effect of bellows free-length variation is minimized by utilizing a mechanical
stop to establish the free length. Mechanical stops also are used to prevent excessive
deflection stresses caused by overcompression of the bellows.
The total force applied to the seal face by the bellows is the sum of the pressure and spring
loads. The pressure-induced load is equal to the product of the differential pressure and
bellows effective area. Both spring load and effective area vary with pressure and pitch
becau~e of the variation of the effective span caused by deflection of the bellows plates. A
portion of the bellows span is inactivated by the plates being pressed together by the force
induced by differential pressure (fig. 52). A larger portion is inactivated as the pitch is
HI GH PR E SS U RE SIDE
f INACTIVE PORTION
SPA N
t
E F FECTIVE
SP AN --y
I MEAN
DIAMETER
MEAN EFFECTIVE
DIAMETER (MED)
Figure 52. - Variation of bellows effective span and mean effective diameter
as a result of pressure-induced deflection .
93
decreased because of the closer spacing of the plates. The spring load is minimized by using
thin plates and the maximum number of convolutions; however, in some cases a lower total
load (spring plus pressure loads) is obtained by using thicker plates and a larger pitch to
minimize the change of effective span and diameter caused by pressure-induced deflection.
Accurate determination of the bellows load requires calibration testing to measure the total
load (spring plus pressure) at the limits of operating pressure and compression. The
calibration usually consists of installing the bellows at the nominal operating length in a test
fixture and increasing the pressure through the required operating range to measure the load
increase. The calibration is repeated for the maximum and minimum operating compression.
The total load is converted to an equivalent mean effective diameter by calculating the
bellows effective pressure area from the load increase due to pressure. The load increase is
the result of the combination of the effective pressure area and the increased spring rate
caused by the reduced effective span. Therefore, it is necessary to measure the load at the
limits of operating compression to determine the total variation . The calculated effective
diameter is used to establish the seal face dimensions for the desired pressure balance. On
different bellows of the same configuration the variation of effective diameter usually
remains within acceptable tolerances. The effective diameter may be predicted with
sufficient accuracy for most applications by utilizing the manufacturer's test data for similar
configurations.
2.3.6.2 PLASTIC LIP SEALS
Plastic lip seals are used as secondary seals for cryogenic fluids when the temperature is
below the minimum ( - 65°F) for elastomeric O-rings. The plastic materials (Kel-F and
Mylar) maintain adequate elasticity down to approximately - 320°F , however, the resiliency
is insufficient to compensate for the thermal contraction. The lip-seal design provides for
pressure and spring loading to compensate for the thermal contraction and maintain sealing
effectiveness. The higher thermal-contraction rate of the plastic material is used to increase
the lip diametral interference at low temperature for improved sealing.
The lip seal is supported in the seal housing with a radiused lip (fig. 53) to provide the
desired contour for conformation to the seal-ring secondary diameter (fig. 12).
The lip seal is made thin (0.005 to 0.010 in .) to provide flexibility for conformance. The edge
of the radiused support is made thin (0.010 in.) to provide the maximum support without
interfering with the lip contour and is radiused (0.005 in. R) to prevent its cutting into the
plastic. The lip length (generally 0.050 to 0.060 in.) is established to provide sufficient
sealing surface at the worst condition of misalignment ; the lip length also determines the
radial pressure load on the lip. A short lip may not provide sufficient conformance for
effective sealing; a long lip results in increased radial load and a greater friction drag force .
94
- LAPPED SURFACE
LIP
T"'CKNesS ~I
SEAL
LIP ,, ~ LABYRINTHS
RADIUS _
y LIP SUPPORT
RADIUS
LIP
SUPPORT
CORNER
, --.1L
RELIEF
RADIUS LIP LENGTH
LIP LIP
DIAMETER SUPPORT EDGE
(RADIUSED)
LIP RETAINER LIP SEAL SEAL HOUSING
Figure 53. - Lip -sea l e lements.
The static seal betwe n the lip an d h ousing i maintained by spring loading the lip with a flat
retainer ring against a labyrinth surface on the hou sing. The labyrinths are m achined sh arp
(0.005 in .) and then lapped flat with a narrow (0.005- to O.OlO-in.) land to provide a true
surface. The retainer ring is relieved along th e lip radius to allow pressure loading of the lip
seal. The differential pressure fo rce (fig. 54) increases the lip load again st the hou sing for
improved sealing.
HIGH PRESSURE
LOW PRESSURE
Figure 54 . - Pressure forc es on a lip seal.
95
The dynamic seal between the lip and seal ling is maintained by providing for a diametral
interference at operating conditions. The effective interference is increased by the
differential pressure load , which deflects the lip radially inward against the seal ring.
Plastic lips have failed as a result of fatigue cracking caused by cyclic deflection of the lip
between the lip support and seal ring. The unsupported portion of the lip i deflected by
axial movement of the seal ring or by pulsating pressure. This failure mode is minimized by
reducing the unsupported length of the lip seal. The housing-to-seal-ring clearance is
minimized, and the lip support edge is made thin to provide the maximum lip support.
Larger clearances also may allow the lip seal to fail by folding backwards in the clearance
space.
Lip seals are worn and scored by foreign-particle contamination. The particles tend to
become lodged in the plastic and cause damage to both the lip and the seal ring. Metal seal
rings generally are hard-chrome plated to minimize the damage.
Plastic lip seals are formed from flat washers that are heated in a contoured fixture to a
temperature that allows plastic flow. Kel-F generally is formed in hot water at
approximately 180°F; Mylar is formed at 325°F. The forming fixture is designed to prevent
residual stresses in the lip, which may cause distortions and wrinkles after the material has
relaxed. Antirotation devices were incorporated in the forming fixture for the lip seals for
the H-l engine after seal wrinkling was traced to the fix ture being rotated during the
forming operations.
2.3.6.3 ELASTOMERS
Elastomeric O-rings and V-packings (fig. 14) are the most commonly used secondary-seal
elements for conventional applications in the temperature range from -6 5 to 500°F. The
economy and reliability of the elastomer generally dictate its use for conventional
applications. However, the reliability of elastomers may be marginal in the more severe
turbopump applications as a consequence of the large temperature ranges, cyclic pressures,
inadequate lubrication, reactive or toxic fluids, high-frequency motions, vibration, and
shock loads. The excessive friction hy.steresis of poorly lubricated elastomers also may be a
problem when good dynamic response of the seal is required .
The V-packing elastomer has the advantage of spring loading for maintaining contact at low
pressure and in the lower temperature ranges where material resiliency is marginal . However,
experience on turbopump seals indicates that the O-ring elastomers are more reliable.
Considerable difficulty was experienced with the V-packing on the fuel seal in the H-l
engine because the packing twisted around and "hung up" the seal ring.
The design of turbopump secondary elastomer seals generally follows the standard
procedures developed for dynamic reciprocating O-rings. The military design standards for
96
elastomers are given in reference 72. The design considerations for O-ring seals and the
design standards for military and industrial elastomers are given in reference 73.
Design considerations for turbopump secondary elastomer seals may vary slightly from
military or industrial standards to satisfy specific conditions. The high reliability
requirements generally justify using precision tolerances to maintain better control of the
O-ring squeeze, concentricity, and clearance gap. The squeeze is minimized to reduce O-ring
friction hysteresis. The groove dimensions are adjusted to compensate for thermal expansion
or contraction and elastomer shrinkage or swelling caused by incompatibility with the fluid.
Floating backup rings are used in the O-ring groove to minimize the clearance gap and
prevent O-ring extrusion .
The elastomer and all metal parts in contact with it are lubricated with a compatible
lubricant to minimize friction and prevent spiral failures. Spiral failures occur when cyclic
motion causes a portion of the O-ring to roll and twist (ref. 73). Rolling and twisting of the
O-ring or V-packing also can cause the seal to hang open and leak. To minimize friction and
increase wear life, the metal sealing surface generally is hard-chrome plated and polished to a
finish of 5 to 10 ,uin. rms or coated with Teflon. Hardened steel also is used for the sliding
surface to improve wear life . The soft metals (e.g. , aluminum) generally are not satisfactory
as a sliding surface for dynamic elastomers because of excessive friction and wear.
The most commonly used elastomers for turbopump secondary seals are Viton A and
Buna N (table III). Viton A is generally preferred since it is not age limited and is resistant
to deterioration by ozone. Various shapes and compositions of TFE in combination with
metal and rubber spring-loading devices to compensate for lack of resiliency have been used
to reduce friction and wear ; however, the experience on turbopump seals is limited. The
TFE compositions generally allow the operating temperature range of the elastomer to be
extended to - 100°F minimum and 600°F maximum.
2.3.6.4 PISTON RINGS
Piston-ring secondary seals (fig. 13) were commonly used in extreme temperatures or
reactive fluids prior to the development of reliable metal bellows as secondary elements.
Most current turbopump applications utilize metal bellows as secondary elements when
all-metal construction is required. However, the piston-ring secondary seal has the advantage
of a more nearly constant pressure-balance diameter and may be required for close control
of face load in some applications. The piston ring is also more resistant to extreme
oscillating pressure environments that may cause fatigue failure of metal bellows. The
combination of higher pressure capacity and closer face-load control generally dictates the
use of piston rings at pressures exceeding 500 psi.
97
--1
The piston-ring secondary seal inherently has significantly higher leakage than the other
types of secondary seals. The sealing effectiveness depends on very precise control of the
mating surfaces to maintain sealing contact. In contrast to the elastomer and plastic lip, the
metal ring will not conform to irregular surfaces and therefore is sensitive to machining
tolerances, thermal distortions, and bending deflections. The piston ring is satisfactory for
applications where a controlled leakage rate is acceptable.
The metal-to-metal sliding contact between the piston ring and the mating seal-ring surface
also is the source of several problems. High-frequency motions cause wear and deterioration
of the sealing surfaces, and vibration causes surface deterioration by fretting erosion.
Excessive friction drag may result in poor dynamic response and seal hangup. The material
combinations are selected to provide ll1bbing contact with minimum friction and wear. The
effects of thermal contraction and expansion also are considered in the material selection.
The sealing surface that slides against the piston ring generally is hardened steel or
hard-chrome-plated steel for wear resistance. The piston rings usually are constructed from
cast-iron alloys. Carbonaceous materials and TFE compositions also have been used.
The design and material considerations for piston rings are discussed in reference 74.
2.3.7 Spring Load

The stationary face-contact seal ring is loaded against the rotating mating surface with a
spring device to ensure sealing contact. The spring force generally is a compromise between
the minim um load for effective sealing and the maximum load for the load-speed-life
relationships.
The total seal-face load consists of the spring force plus the resultant pressure-induced
forces; therefore, the spring force is determined in conjunction with the pressure balance to
establish the desired total load. The spring force is significant on low-pressure seals because
of the negligible pressure-induced force; conversely, the spring force generally is negligible
on high-pressure seals because of the large pressure-induced forces.
The minimum spring load for sealing effectiveness is established by the dynamic response,
secondary-seal or vibration-damper friction, and variations in interface pressure profile. The
spring force for dynamic response is established by calculating the force required to
accelerate the seal ring at the rate required for the seal ring to track the runout of the
rotating mating surface. In some cases, the acceleration rate may be established by sudden
movements of the turbopump shaft during transient operation. The resultant closing force
(spring plus pressure) must be greater than the sum of the seal-ring inertia, secondary
friction, and interface average pressure profile forces for effective sealing. Allowance for
variation in interface pressure profile is provided by either additional spring load or a larger
pressure-balance ratio.
98
L J
- - -- - - - . _ -- - -- -
The maximum spring load is limited by the load-speed-life relationships for specific fluids
and materials. The load and speed limitations based on current practices are reasonably well
defined for a wear life of approximately 3 hr. The life relationships, however, are not as well
defined, because of the limited long-duration testing in rocket propellants. Therefore, the
spring load is minimized on seals that have wear life requirements greater than
approxima tely 4 hr.
The maximum spring load also may be establish ed by the ru bbing-friction power loss at the
seal interface. In small turbomachinery, it is possible for the seal power loss to become a
significant portion of the total turbopump power. The heat input caused by seal power loss
may be significant on systems that utilize recirculating coolant fluid. Heat-balance
calculations generally are perform ed to establish the allowable heat input.
I
I
The current practices for spring loading face seals in various applications in representative
turbopumps are sh own in figure 55. Most turbopump seals are designed for a spring load of
Z 10
u.
II)
...J
w-
U
«
u.
...J
w
«
'"
u.
0
:I:
f-
l.?
Z
W
...J
f-
Z
;:)
a:
...w
0
«
0
...J
l.?
Z
a:
...
'"
~ ±
N
~ ::E
s
~
C< ~
~ ~'"
iE ..: ~
v>
N l::
:IE ::E
S
~
C<
S
.it ~
~s
><
:lEai
' ~
C<
N N .. ..
o
C<
0
C<
S!I
~
C<
lOX lH2 CH 2 H2 • H~ lF2 FlOX RP- l
TUABOPUMP
FLUIO
Fig ure 55 . - Current pract ices fo r spring loading fa ce-contact sea ls.
99
approximately 2 lbf/in. of seal face circumference. The minimum practical spring load is
approximately 0.3 lbf/in. A maximum spring load of approximately 10 lbf/in. was used
successfully to minimize leakage on the liquid-oxygen seals on the H-l engine and the RP-l
seals on the F- I engine; however, spring loads higher than approximately 4 lbf/in. are not
used unless the cooling capacity of the sealed fluid is sufficient to dissipate the heat
generated by the rubbing friction.
The spling loading devices consist principally of coil springs, wave splings, and metal
bellows. Sevenl small coil springs around the seal ring provide the most uniform face
loading. The spling rate usually is minimized to reduce the load variation with operating
compression. Low-rate springs and bellows are preloaded and restrained with a mechanical
extension stop when higher operating loads are desired.
2.3.8 Pressu re Balance

Face seals are pressure balanced as closely as possible to minimize the variation in face load
caused by the forces induced by differential pressure (fig. 56). Pressure balance is
particularly important when fluid pressures are greater than approximately 100 psi, because
of the large pressure-induced forces on the seal face. It is not feasible to balance the
HIGH PR ESSURE SIDE
MIN IMUM PROFILE

I DIVERGENT SURFACES I
NOM INAL PROF ILE

(PARALLEL SURFACES I
MAXIMUM PROFILE
(CONVERGENT SURFACESI
INTERFACE
PRESSURE
PRESSURE
CLOS ING
PROF ILE FORCE
SEAL
FACE
AREA
SECON DARY SEAL OR BELLOWS

EFFECTI VE DIAMETER
SEAL FACE
I SEALI NG DAM I
- LOW PR ESS URE 5 I DE
F igure 56. - Pressure-induced forces o n face seals.
100
l
pressure-induced forces completely because of the variation of the seal interface pressure
profile. Practical designs require a safety margin that allows for the maximum variation in
interface pressure profile so that face separation can be prevented.
The average interface separating pressure is equal to the pressure profil e factor ex. times the
differential pressure:
average in terface separating pressure = ex. (P I - P2) (13)
The theoretical pressure profile factor for steady-state laminar or turbulent flow of an
incompressible fluid between parallel sealing surfaces is approximately 0.50 (refs. 9, 75 , and
76). This factor can increase to approximately 0.80 for convergent surfaces and can decrease
to approximately 0.20 for div ergent surfaces.
The theoretical pressure profile factor ex. for subsonic laminar flow of a co mpressible fluid
between parallel sealing surfaces varies between approximately 0.50 and 0.67; the lower
pressure ratios (PI /P2 ) give factor tending toward 0 .50, and the higher press ure ratios
produce factors tending toward 0.67, as shown by the following equation:
(14)
The theoretical pressure profile factor for flow of a compressible fluid between parallel
sealing surfaces with isentropic entrance condition and choking (sonic flow) at the exit
varies between approximately 0.67 and 1.00; flow just barely choked gives factors tending
toward 0.67 , and very highly choked flow gives factors that approach 1.00. Leakage f10w
can become choked at the exit for pressure ratios greater than about 4: I (I:::.r/h > 100). The
experimental data indicate closer agreement to the theoretical anlaysis when an
entrance-loss coefficient of 0.6 is used (ref. 12). Since entrance losses tend to reduce the
average interface separating pressure, the pressure profile factor for choked flow is redu ced
when the entrance losses are considered. The theoretical pressure profile factor for choked
flow with an entrance-loss coefficient of 0.6 varies between approximately 0.5 and 0.65 for
pressure ratios from 1 to 10 .
The pressure profile factor is affected by turbulence (both flow induced and rotation
induced) , change of fluid state, viscosity, seal clearance, and in terface geometry. The effect
of rotation on the pressure profile factor usual ly is not significant in sealing cryogenic or
compressible fluids.
101
Pressure balance is achieved when the separating force on the seal induced by interface
pressure is equal to the pressure-induced closing force . The pressure balance ratio is defined
as the ratio of the effective closing or balance area (fig. 56) to the seal face or dam area:
. closin o area
pressure balance ratIO = f I:> (15)
ace area
A seal is pressure balanced when the pressure balance ratio is eq ual to the in terface-pressure
profile factor. Therefore, a balance ratio of 0.5 corresponds to the theoretical pressure
balance for laminar incompressible flow across parallel sealing surfaces.
The balance ratio used on practical designs varies from 0.55 to 1.0, the value depending on
the operating conditions and the desired safety margin. High-pressure rubbing-contact seals
are relatively close-balanced to minimize the face contact load ; therefore, the margin for
variation in the interface pressure profile is reduced , and the seal is more sensitive to face
separation caused by face distortion or fluid vaporization across the interface. The high
pre sure may cause additional deflections of the eal face and mating ring surfaces; these
deflections increase the interface pressure profile and result in greater separating forces. The
maximum sealing effectiveness is achieved with large balance ratios. A balance ratio of 0.8
or greater generally provides for the maximum variation of the interface pressure profile
that results from converging surfaces, choked flow, and fluid vaporization. Low-pressure
cryogenic seals usually are designed with a balance ratio of 0.85. Seals for high-pressure
incompressible fluids generally have a balance ratio of 0.55 to 0.6; seals for high-pressure
compressible fluids have a ratio of 0 .7 . Most conventional seal designs use a balance ratio of
0.65.
The pressure-balance ratio usually is selected in conjunction with the spring load to provide
a total (spring plus pressure) face-contact load that is consisten t with the load-speed-life
relations and the minimum load requirements for effective sealing.
The variation in the face and balance diameters because of thermal contraction or
expansion, pressure deflection, and din1ensional tolerances must be considered for accurate
detem1ination of the balance ratio. Precision tolerances ordinarily are used on turbopump
seals to minimize the balance-ratio valiation. The face diameters on carbon inserts usually
are machined after installation to reduce the tolerance accumulation. Face diameters on
bellows seals usually are machined after the bellows effective diameter is measured.
102
L_
3. DESIGN CRITERIA and
Recommended Practices
3. 1 SEAL SYSTEM
3.1.1 Pressure Environment

The system design shall minimize the severity of the seal pressure environment.
Reduce high pressure levels (> 500 psig) by utilizing upstream labyrinths or circumferential
clearance seals in conjunction with low-pressure return bleeds that allow recirculation back
to a low-pressure area (fig. 10). Reduce the differential pressure by maintaining an
established pressure level downstream of the primary seaL Minimize the amplitude of
pressure oscillation by utilizing labyrinths upstream of the seal as a damping device (fig. 4).
3.1.2 T hermal Env ironment

The system design shall minimize exposure of the seals to high-temperature
environments and severe thermal gradients.
Minimize high-temperature environments at the seal location by designing the system to

isolate the seal in a cooled area. Provide barriers or labyrinth devices to prevent the direct
flow of high-temperature fluids onto the seal assem bly or its components. The thermal
stresses created by the differential thermal expansion/contraction should be held to a
minimum by designing the seal housings and mating rings to provide minimum restraint to
thermal deflections and by using materials with similar thermal expansion/contraction rates.
3.1.3 Vacuum Environment

The system deSign shall preclude exposure of rubbing seals to hard vacuum.
Route the seal drains to a safe disposal area internal to the turbopump or engine (e.g., the
turbine exhaust or pump inlet); when the fluids are compatible, eliminate the drains and
drain the seal leakage directly into the turbine (fig. 3).
3.1.4 Rubbing Speed

Seal rubbing speeds shall be at a minimum consistent with the turbopump design
requirements.
103
Use the smallest possible seal diameter to minimize the rubbing speed.
3.1.5 Cooling and Lubrication

Th e system design shall provide for adequate cooling and lubrication of th e seals.
Cooling and lubrication of the primary shaft seal and mating ring should be provided by
allowing the sealed fluid to flow through the seal cavity.
Where possible, utilize bleed holes through the impeller from the seal cavity to the impeller
inlet or through the shaft back to the pump inlet. Pumping vanes on the impeller rear
shroud also may be utilized to ensure recirculation of the sealed fluid. The seal should not
be installed in a dead end cavity that may allow accumulation of stagnant vapor.
Dry-running intermediate seals should be circumferential shaft seals with an inert-fluid purge
for cooling.
Additional cooling and lubrication should be provid ed when required by purging with an
in ert fluid or by injecti ng a cooling fluid into the eal cavity or onto the eal mating ring.
Separate cooling and lubrica tion systems isolated from the sealed fluid should be used when
th e cooling fluid is not co mpatible with the sealed fluid.
For maximum hea t dissipation , locate mating rings with the pressure on the outer diam eter
of the sealing interface. The use of one mating ring with two rubbing contact seals should be
avoided because of overhea ting and di tortion problems. Separate mating rings for each seal
should be used .
3.1 .6 Leakage Drai ns

The system design shall provide drains for safe disposal of expected seal leakage.
Size the seal drains to accommodate the maximum anticipated leakage without building up
a significant back pressure. Calculate the drain effective flow area with conventional flow
equation by es timating the maximum anticipated seal leakage and establishing the
allowable drain cavity pre ure. The drain pressure differential and leakage rate will then
e tabli h th e required drain size.
3.1.7 Fluid Separation

Th e system design shall provide safe separation of incompatible fluids.
104
Incompatible fluids on the same shaft should be separated by utilizing two face-contact seals
to minimize the leakage; provide separate drains for each propellant to vent the leakage to a
safe disposal area and either a purged double-circumferential seal (fig. 6) or two
interm ediate face-contact seals with a purge between to separate the drain cavitie (fig.
10(a)) . The purge pr ssure hould be high enough to provide a pressure barrier between the
drain to prevent mixing of the propellants by leakage through the intenl1ediate seal.
3.1.8 Fail-Safe Provisions

The system d esign shaLL provide that a single seal failure w ill not cause failure of
the turbopump.
Conduct a theoretical failure analysis to estimate the operating parameters and the possible
re ult for each different failure condition; all different modes of operation and variations of
perfolmance should be considered. Size the seal drains to provide for the maximum leakage
that would result if a single seal fails. The intennediate seal purge pressure should be
e tablished higher than the maximum drain back pressure. Redu ndant seals in series should
be used when failure of a single seal would cause turbopump failure .
3.1.9 Purge Requirements

Th e system design shall provide seal cavity purges adequate to remove trapped
flUids and to dry accumulated moisture.
Provide purged cavities with an inlet and outlet port to allow the purge gas to flow through
the cavity. The purge flowrate and length of time should be established to ensure complete
removal of air and moisture prior to system chilldown. After the test, the purge should
remain on or air should be prevented from re-entering the cavity until the hardware returns
to am bient temperature.
3.2 SEAL ASSEMBLY
3.2.1 Pressure Capability

The seal shall perform satisfactorily at the maximum expected operating pressure.
Select the seal type (table II) that will satisfy the pressure-speed-life relation with the
minimum leakage and maximum reliability. The recommended limits of the fluid
pressure-speed relation for turbopump face-contact seals (3-hr life) in liquid oxygen, liquid
105
L
hydrogen, liquid fluorine , gaseous hydrogen , RP-l and hot gas are shown by the solid
curves in figure 57 . The curren t practices are represen ted by the circular da ta poin ts.
The recommended limits are estimated by assuming that the face-contact load increases in
proportion to the fluid pressure and that the limit is established by the seal interface heat
buildup, which is a function of a constant fluid-pressure/speed relation for each fluid. The
limiting pressure-speed relation was established by the relative success of the current
applications. The limits shown are to be used as approximate guides before the seal detail
design is established. The limits are based on commonly used seal materials (table III) and a
wear life of approximately 3 hr. Long-life seals will require a more conservative
pressure-speed factor or noncontact seal interfaces. The following recommendations are
based on current turbopump practices:
• Face-contact metal-bellows seals (fig. 11) should be used for cryogenic or reactive
fluids at pressures up to approximately 500 psig.
• Face-contact piston-ring seals (fig. 13) should be considered for high-pressure (>
500 psig) cryogenic or reactive propellants when face-load control is critical.
• Face-contact elastomeric seals (fig. 14) should be used at pressures up to

approximately 1000 psig for conventional lubricated applications in the
temperature range of - 65°to 500°F.
• Circumferential shaft-riding segmented carbon seals (fig. 20) should be used for
low-pressure « 100 psig) purged intermediate or hot-gas applications .
• Circumferential floating-ring controlled-gap shaft seals (fig. 23) should be used for
high-pressure ( > 100 psig) hot-gas, purge-gas, or long-life (> 4 hr) applications
when the increased leakage is acceptable.
• Circumferential labyrinth seals (fig. 24) should be used for high-pressure and
long-life applications when reliability and economy are the primary considerations
and the increased leakage is acceptable.
• Face-type hydrostatic/hydrodynamic seals (figs. 15 through 18) should be

considered for combined high-pressure, high-speed , and long-life applications
when minimum leakage is desired.
• Variations or combinations of the basic seal types in table II should be considered

for special problems.
3.2.2 Temperature Capability

The seal shall perform satisfactorily at the operating temperature of the
application.
106
LIQUID OXYGEN LIQUID HYDROGEN
PfV LIMIT = 60000 PfV LIMIT = 200 000
500 500
0
en 400 en 400
Cl.
U.
Cl.
u.
00
300 300
0' 0
w
w
w w
Cl. 200 Cl. 200
en 0 en
100 00 100
CD
0 100 200 300 400 500 0 100 200 300 400 500
PRESSURE , PSIG PRESSURE, PSIG
GASEOUS HYDROGEN LIQUID FLUORINE

PfV LI M IT = 50 ()()() PfV LIMIT = 50000
500 500
0
en 400 400
Cl.
u. ~
u.
300 300
0' 0'
w w
w w
Cl.
CIl
200 Cl.
en
200
8
100 100 0
0 100 200 300 400 500 0 100 200 300 400 500
PRESSURE, PSIG PRESSURE, PSIG
RP-l FUEL H2 + H20 HOT GAS

PfV LIMIT = 80000 PfV LIMIT = 20000
500 500
en 400 en 400
Cl. Cl.
U. u.
300 300
0' 0 0'
w w
w w
200
Cl. 200 Cl.
en CIl
100 00 100
0 100 200 300 400 500 0 100 200 300 ..ao 500
PRESSURE, PSIG PRESSURE, PSIG
0 Current practices
//// Recommended limits
Figure 57. - Cu rrent practices and recommended limits for PfV factor for face-contact seals (3-hr life) .
1 107
,
I
___ J
~I
The seal assembly configuration should allow utilization of materials that have satisfactory
properties over the maximum range of anticipated operating temperatures. The following
temperature limits for various seal types should be observed:
Seal type Temperature limits, of

Metal bellows -423 to 1500
Plastic lip -320 to 200
Piston ring -423 to 1500
Elastomeric -65 to 500
Segmented carbon - 423 to 1000
Floating ring -423 to 1500
Labyrinth - 423 to 1800
3.2.3 Speed Limitations

The seal design shall be consistent with the speed and load-velocity limitations.
The surface speed limit should be established by estimating the resultant seal interface
temperature on the basis of the heat generation caused by rubbing contact or viscous shear
and the heat dissipation to the surrounding environment. Consideration should be given to
the temperature limits of seal-face and mating-ring materials and thermal decompositon of
the sealed fluid . The stresses and deflections caused by thermal and centrifugal forces also
should be considered. The heat-transfer methods given in references 6 and 7 are
recommended for calculating the seal interface temperature profile and thermally induced
deflections.
Estimates of the face load-velocity limits should be made by utilizing the load-velocity
relationships for specific materials and fluids. The recommended limits are shown by the
curves in figures 58 and 59 as a function of velocity and face load. The current practices are
shown by the circled data points on the same figures for comparison. The recommended
limits are estimated by assuming that the limit is established by the seal interface heat
buildup, which is proportional to a constant face load-velocity relationship for each fluid .
The limiting load-velocity relationship was established by the relative success of the current
applications. The limits, which are to be used as approximate guides, are based on
commonly used seal materials (table III) and a wear life of approximately 3 hI. For longer
life, a seal will require a more conservative load-velocity factor or noncontact seal interfaces.
The face load on high-speed seals should be minimized by utilizing designs with minimum
spring-load requirements and precise pressure-balance control. The circumferential clearance
seals (e.g., floating-ring and labyrinth) or the fluid-film seals (hydrostatic/hydrodynamic)
should be considered when the surface speed is higher than approximately 500 ft/sec .
108
L1aUID OXYGEN L1aUID HYDROGEN
FV LIMIT : 2000 FV LIMIT 4000 E
500 500
400 400
In 0 0
Q.
U.
If
u.
300 300
ci ci 0
w w 0 0
w
Q.
200 w 200
Q..
In In
100 100
0 4 8 12 16 20 0 4 8 12 16 20
LOAD/UNIT LENGTH, LBFIIN. LOADNNIT LENGTH, LBF/IN.
GASEOUS HYDROGEN LIQUID FLUORINE

FV LIMIT = I 500 FV LIMIT = 1 000
500 500
400 400
In In
Q.. Q.
U.
300 u.
300
ci Q
w w
w
Q.
200 w 200
Q..
In In
100 100
0 4 8 12 16 20 0 4 8 12 16 20
LOAD/UNIT LENGTH, LBF/IN. LOAD/UNIT LENGTH, LBF/IN.
RP-1 FUEL H2 + ¥ HOT GAS

FV LIMIT = 2500 FV = 800
500 500
400 400
In
Q.. In
U. Q..
300 u. 300
ci 0'
w
w 200
w 200
Q.. W
In Cl-
In
100 cQ) 100
0 4 8 12 16 20 0 4 8 12 16 20
LOAD/UNIT LENGTH, LBF/IN. LOAD/UNIT LENGTH, LBF/IN.
o Current practices
/ / / / Recommended limits
Figure 58 . - Current practices and recommended limits for FV factor for face -contact seals (3-hr life).
109
~
LIQUID OXYGEN LIQUID HYDROGEN

PV LIMIT = 25000 PV LIMIT = 50 000
500 500
400 400
Ul Ul
0
~ ~
u. 300 u. 300
0
0" 0"
w 200
w
200
00
w w
~ ~
Ul Ul
100 100
0
40 80 120 160 200 0 40 00 120 160 200
UNIT LOAD. PSI UNrr LOAD. PSI
GASEOUS HYDROGEN LIQUID FLUOR INE

PV LIMIT = 20000 PV LIMIT = 20 000
500 500
0
400 400
Ul 0 Ul
~ ~
u. 300 u. 300
0 0
w w
w 200 w 200
CL CL
Ul Ul
100 100
0 40 80 120 160 200 0 40 00 120 160 200
UNIT LOAD, PSI UNIT LOAD, PSI
RP-1 FUEL H2 + H2 0 HOT GAS

PV LIMIT = 25000 PV LIMIT x 10000
500 500
400 400
Ul Ul
~ ~
u. 300 u. 300
0"
w
0"
200 w 200
w w
CL ~
Ul Ul
100 100
0 40 80 120 160 200 0 40 80 120 160 200
UNIT LOAD, PSI UNIT LOAD, PSI
0 Current pt"actices
//// Recommended limits
Figure 59. - Current practices and recomm end ed limits for PV factor for face -contact seals (3-hr life).
110
L~~_
3.2.4 Wear Life
The seal wear life shall satisfy the turbopump life requirements.
Provide sufficient height on the seal contact face to allow for the maximum anticipated
wear rate. The wear rate should be minimized by providing some form of lubrication (e.g.,
hydrodynamic, boundary, deposited film) and by using the lowest possible face-contact
load. Except when the PV factor is low and good lubrication is available for contact seals,
the clearance or fluid-film seals should be considered when wear-life requirements exceed 4
hr.
3.2.5 Leakage
The seai ieakage shall be the minimum possible consistent with the required
operating conditions.
Face-contact metal-bellows or elastomeric seals with the maximum face contact load
allowed by the load-velocity relations for specific materials and fluids should be used to
minimize seal leakage. The face-contact load should be decreased as the speed and life
requirements increase. The hydrostatic/hydrodynamic concepts should be considered for
effective sealing at load-velocity-life relations greater than those allowed by the current state
of the art for rubbing-contact seals.
Conduct a thorough heat-transfer and stress analysis of the seal face and mating surface so
that seal interface distortions that can cause excessive leakage (refs. 6 and 7) can be
anticipated and minimized.
To size the drains and predict propellant losses, estimate the seal design leakage by using the
available theoretical methods and empirical relations (refs. 8 through 15). It is
recommended that the dynamic leakage for cryogenic face-contact seals be estimated by
utilizing either the laminar-flow-with-inertia or the turbulent-flow-without-inertia theory
(the choice depending on the value of the Reynolds number), with the assumptions of liquid
conditions and a leakage-path effective gap of approximately 200 pin. The static leakage
should be estimated on the basis of a leakage-path effective gap of approximately' 50 pin .
3.2.6 Misalignment Tolerances
3.2.6.1 AXIAL OPERATING LENGTH
The seal axial operating length shall provide for installation tolerances, thermal
expansion/contraction differentials, dynamic deflections, and face wear.
111
-- - - - - - - - -
Design th e seal bellows or loading springs and the internal clearance to allow for the
maximum toleran-ce variations. Tolerance stackups with con id eration of th ermal effects and
stress analysis with consideration of dynamic deflections hould be performed.
3.2.6.2 RADIAL LOCATION
The radial location of th e seal face relative to the cen ter of rotation shall prevent
excessive wiping action.
It is recommended that the radial loca tion of face-contact seal be held within
approximately 0.003 in. /in. of diameter on high-speed (> 200 ft /sec) seals. Shaft-riding seals
should be held within approximately 0 .005 in. /in. of diameter.
3.2.6.3 MATING SURFACE NORMALITY
Th e nomality 0/ the seal mating surface shall be such as to preven t excessive

mating ring wobble.
It is recommended that the normality of the seal mating surface (as installed) to the center
of rotation be held within approximately 0.000 1 in. /in . of diameter (T.I.R.) on high-speed
eals.
3.2.6.4 ROTATIONAL ECCENTR ICITY
Shaft-riding-seal deSigns shall provide for radial location of the mating sur/ace that
prevents excessive ro tational eccentricity.
The mating surface diameter on high-speed shaft-riding seals should be held concentric to
the center of rotation within 0.0005 in ./in. of diameter (T.I.R.).
3.2.7 Vibration Control

The seal vibration damping shall prevent resonant vibrations that resu lt in
excessive leakage or fatigue failure.
Control seal vibration by designing the seal for a natural frequency that is higher than any
expected on the turbopump or by providing mechanical-friction dampers (fig. II) at the
seal-face outside diameter or at the bellows componen t. The friction drag force for effective
112
_______ J
- -- - - - --- - -- - -- _ l
damping of bellows seals should be approximately 5 to 10 percent of the seal spring load.
The secondary-seal friction on elastomeric, piston-ring, and lip seals usually is adequate for
effective damping.
To eliminate vibration caused by stick-slip conditions, select rubbing material combinations

to provide self-lubrication or provide lubrication at the rubbing seal face.
3.2.8 Contamination Allowances

The seal shall toLerate the foreign-particle contamination allowable in the
application.
Use rubbing-contact face seals with tungsten carbide, titanium carbide, or aluminum oxide
face materials and metal-bellows secondary element for maximum resistance to abrasive
environments in those applications in which the materials are compatible. Special protection
systems (neutral fluid injection , buffer zones, centrifugal separators, slingers) should be used
if the contamination is evere. Where possible , use system filters in the range 10 to 10011 to
clean the sealed fluid.
Avoid installing th e seal in a sump area that would tend to collect foreign-particle settlings.
3.2.9 Seal Mounting Requirements
3.2.9.1 SEAL PI LOT
The seal pilot shall provide precise radial location with allowance for thermal
differen tials.
Use seal housing materials that have thennal contraction/expansion rates similar to that of
the pump housing, or adjust the ambient fit to allow for the thennal differential.
Use three equally spaced radial pins engaging radial slots or radial splines to compensate for
temperature gradients larger than allowable for diametral pilots.
3.2.9 .2 FLANGE SEAL
The f lange static seal shall co mpensate for thermal contraction.
Use metallic, spring-loaded, pressure-actuated static seale, to compensate for the dimensional
variations caused by thennal contraction at the sealing joint.
I
1 113
I
_J
3.2.9.3 FLANGE LOAD ING
The flange loading shall maintain the seating stress required for effective sealing.
To maintain flange loading, use several small bolts around the flange with a sufficient
preload to cause elastic elongation. Avoid using large ring nuts, because of the low elastic
elongation and nonuniform loading.
3.2.9 .4 SH IMS
The shimming provision shall be adequate to maintain the required installation

tolerance.
To adjust the seal axial installed length, use ground metal gasket-shims coated with a softer
material (copper, silver, Teflon) ; use spacers with grooves for spring-loaded static seals.
3.2.9.5 PROV ISION FO R SEAL REMOVAL
Th e seal design shall provide for removal of stuck or wedged seals.
Include threaded holes in the flange or recessed grooves that can be engaged by special
pulling tools.
3.2.9.6 LOCK ING DEVICES
Bolts and nuts used to mount the seal shall be securely retained.
Use positive locking devices to prevent loosening of bolts and nuts. When possible, bolts and
nuts should be trapped in position by the adjacent hardware.
3.2.9.7 MATING-RING MOUNTING
The method for mounting the mating ring shall provide for minimum distortion,
static sealing, and antirotation.
Use free-floating nonloaded mating rings to minimize distortion caused by clamping loads.
The shaft mating surfaces for clamped rings should be flat (3 helium light bands) and normal
to the direction of applied load. The mating ring should be isolated from all structural loads.
114
L I
Use gaskets, dispersion coating, plating, or elastomers to seal the joints between the mating
ring and shaft or spacers exposed to pressure.
Use antirotation pins, radial splines, or clamping friction to prevent mating-ling rotation
relative to the shaft.
3.3 SEAL COMPONENTS
3.3.1 Seal Materials
3.3.1.1 COMPATIBILITY
Seal materials shall not be adversely affected by th e sealed fluid or by the

operating temperature.
The seal materials based on current practices given in table III are recommended. Additional
material recommendations are given in references 17 through 20.
For structural components, use metals that maintain adequate ductility at cryogenic
temperatures (e.g., Monel, nickel-base alloys, and stainless steels). Avoid using the
martensitic steels (e.g., 17-7 PH, AM-350 , 4130, 4340) for flexing elements at low
temperature because they become brittle.
Avoid metal-to-metal rubbing contact in liquid oxidizers because of the explosion hazard.
Metals resistant to impact detonation and -oxidation should be used for seal-ring pilots,
antirotation tangs, and friction-type vibration dampers.
Structural metals subject to surface reaction with the fl uid shou ld be protected with
chromium or cadmium.
Fasteners that are subject to thread galling should be plated with copper, silver, or cadmium
for applications in the temperature range of - 423 to 1000°F.
Static seals and gaskets should be plated with copper or silver for improved sealing at
temperatures from - 423 to 1000°F.
Gold or copper pia ting is recommended for seals used in iiq uid fluorine.
For rubbing-contact seals, use carbon (P5N, P692 , P03 ,P5AG, EY I OS , CDJ83, CCA 72,
G84 , G39) seal faces rubbing against hard-chrome-plated steel and tungsten/chromium
lIS
carbide-coated mating rings. Avoid using carbon aceo us materi als in ox idi z ing atmosp heres at
temperatures higher than approximately I OOO°F beca use of oxidation and ch emica l ero ion.
Special high-temp erature carbons (e.g., P2003 , G84. CDJ83) should be used when
temperatures exceed 800°F.
Plastic materials (e.g., Kel-F , T eflon , Mylar) hould be used for gaskets. tatic sea ls. and
secondary seals in mo t cryogenic or reactiv e propellants (except fluorine). Avoid u ing
plastic materials at temperatures below - 320° or above 6 00°F. Pi a tic materials ho uld not
be used for stressed compon ents above approximate ly :WO°F beca u e plastics a re su bject to
cold flow or c reep .
Avoid using plas ti c materials in liquid oxygen whcn imp act ene rgy leve ls excee d 70 rt-Ibf or
when pressures are extremely high (;;;;. 5000 psi). onmeta lli c materials inte nd ed 1'01' use in
oxygen should be tested in accord ance with test procedure o utlin ed in refe rence 77.
Elastomeric materia ls (Viton A , Bun a ) sh ould not be used at tempera tures below
approximately - 65°F or above approximately 500° F. Use Viton A for resistance to
deterioration by ozone.
Epoxy adhesive Epon 901/B3 i recom m nd ed for bonding carbon in e rt s to be used in

liquid oxygen. Avoid using epoxy adhesive if the materi al will be ex pose d to imp ac t ene rgy
levels in exce s of 2 kg-m (14.46 ft-Ibf). Do n o t use epoxy in liquid flu o rin e.
Use ceramic (aluminum oxide, titanium carbid e, tungsten ca rbid c) a nd cermet material ror
maximum oxidation resistan ce at high tcmperature (up to 2000°F). Th e ceram ic and cermet
materials are recomm ended for u e with liquid flu o rin e and oth er highl y reactive
propellan ts.
3.3.1 .2 COR ROS ION
S eal materials shall not be subject to co rrosioll tltat could adJlerse/y aJIeel
performance.
Use corrosion-resistant stainless steels and nickel-ba e all oys for seal co ns tru ct ion . Protect
ferrous metals (e.g. , 4130 and 4340) with chromium or cad mium plat in g.
Use nickel-base alloys for m etallic elem ents expo ed to hydrogc n-ri ch team (H 2 + H 2 0).
3.3.1.3 Hydrogen Embrittlement
Seal structural components in hydrogen environments shall be resistant to

hy drogen embrittlem en t.
116
L
Th e ta bl e stainles tee ls (e .g., 300 eri es and A-2 86) o r aluminum all oys . ho uld bc used fo r
hydroge n e nviro nm e nt above 20QoF. Th e ni c kel-base all oys sho uld be u "cc.l ,.It tc mp era ture
below 200°F . Mate ri als subjec t to hydroge n e mb rittl eme nt sh ould be protec ted wi th
co ppe r p lat ing ( refs. 2 1 a nd 22 ).
3.3.1.4 MATERIAL PROPERTIES AND TEMPERATURE EFFECTS
Sea l IIwleria l IJml Jerri es alld /lieir "ar ia rioll \I' i rli /eIlI IJera { lI r e sli all be slI i rab le .lC)!·
t li e alJp lica rioll.
Select th e ea l ma teri,i/s o n th e bas is 0 1' th e requirements es tabli shed by a he,lt-tra nsl'e r and
st re s a nalys is <.In c! th e mate ri al pro pe rti es give n in rek re nces 23 and 24. Co nsid e r the
va ri at ion of mate ri al ph y ica l pro pe rti es wit h te mp e ra ture (fig . 33,34. a nd 35 ).
3.3.1.5 WEAR AND FRICTION
Rllbb illg-co/l/aci .\ea l llw{erials /() be IIsed i ll OTOgC'lI ic or d lT C'II I'iro II III ell ts sli all
.lC)J'J1I a se((llIiJri('{l I illg .lil lll \\'11('11 C'.\)J()sed to rli e .ll lI id lII C'd ill lll .
U e ca rbo naceous mate ri als with in o rga nic re in s. me tal nu orid es. sil ve r, anci vari ous o th e r
ad di t ive to assist in th e fo rm at ion of a elr-Iubri catin g film o n th e ma tin g ur face.
Use ma te ria l co n ta inin g tit anium ca rbid e o r aluminum ox id e to fo rm a lubrica tin g tlu o rid e
lill11 in liq uici flu o rin e.
F u ed flu o ri de coa tin gs sho uld be ap pli ed to th e ma tin g surface o f ce rami c a n I ce rmet
ma teri al to red uce th e wea r and fri c ti o n in li q uid oxyge n. liquid tlu o rin e. liquid sodium .
hydroge n, vac uum . a nd air ( refs. 25 a nd 26) .
Additi o nal reco mm e nd a ti on fo r sea l ru bb in g mate ri als ba ed o n curre nt pract ice a re give n
in tab le 1I1.
3.3.1.6 CLEANING
Par ts C'xposC'd to liquid oxidi::.ers shaLl bC' ./i·C'C' of allli ),d roca r boll co ll talll i /lalio/l.
Me talli c pa rts ho ulel be cl ea ned by ( 1) bru hin g in tri chl o roe th y len e, (2) ho t va po r
degreas in g, o r (3) ultraso nic clea nin g (re f. 19). o n-m e talli c m a te ri als (pi a tic, elas tome r.
and ca rb o n ) sh o uld be clea ned by h a nd wiping with a lint-free clo th m o iste ned with liquiel
Freon . vo id flu hin g with clea nin g o lve nts. II parts sh ould be clea ned a indi vidu al
11 7
l
co mp on e nts pri o r to as embl y . Protect th e p art a ft er clea ning by sealin g in a pl astic b ag
until rea dy fo r assembl y.
O xidi zer ea ls sh o uld be d esign ed to allow clea ning of the sep ara te compone nts. Avoid
trapp ed areas whi ch co uld prevent ad equ a te clea ning.
Ca rb o n ma teri als fo r o x idizer service sh ould be m achined and lapp ed e ith er dry o r in a
co mp atibl e fluid . Av oid usin g diamond o r sili co n lapping compounds susp e nd ed in min e ral
o il ba e , beca u c of th e di ffic ult y of c lea nin g soft o r porous ea l m a teri als.
3.3.1.7 PASSIVATION
Ma terials n"jJosed/ o liquid f 7u orill e shall be passive ill co ntac t with f luorin e.
Build Lip a pro tec ti ve Ilu orid e film a nd reac t surfa ce co ntamin a ti o n by p assiva ting the
ma te ri a l in gaseo us flu o rin e p ri o r to li q uid x po ure . A void ex p o ure to m oisture afte r
pass iva ti o n ( ref. 19).
3.3.2 Face-Contact-Seal Rubbing Elements
3.3.2.1 SEAL FACE WIDT H
Th e sea l face width sh all provide max im um sealing effectiveness co nsistent with
wear-life and h eat-trallsfer req uiremellts.
Th e sea l face width h o uld be a minimum o f 0.03 0 in . to provid e sufficient m argin for
mac hinin g t ole ra nces a nd edge chippin g. Th e minimum width for larger eal in creases as a
res ult o f th e stru c tural load a nd unit co ntac t load requirem ents. The 0.030-in. width m ay b e
use d fo r la rge r sea l if pre ure-ve nted supp o rt p ad s ( fig. 18(b )) are used on th e sea l face to
a is t in loa d supp o rt.
U e th e mi nim um prac ti ca l face wid til fo r p rec i e pre ure bala nce and m ax im um h ea t
tra nsfe r to th e ea led Iluid .
Th e max imum face wid th sh o uld be e ta blished b y the load vari a ti o n du e to th e inte rface
press ure pro fi le o r by th e effec t of face di sto rti o n o n lea kage and th e pressure pro fil e . Th e
th eo re ti ca l lea kage is redu ce d in direc t pro p o rti o n to th e face-width in crea e ; h ow ever the
adv e rse e ffec t o r additi o nal face dis tortion usu all y is grea ter on turb o pump ea ls.
\I
J
Use the maximum practical face width to reduce the unit contact load for minimum wear
and to achieve maximum heat transfer by conduction into the seal ring and mating ring.
The recomm ended face width in relation to fac e diameter for turbopump face-contact seals
is given in figure 37.
3.3.2 .2 SEAL NOSE HEIGHT
The seal nose h eight shall aLLolV for tile maximum face wear and shall be
consistent IVith the stru ctural req uirements.
Establish the minimum nose height (fig. 36(a)) on th e basis of predicted maximum wear rate
and Ii fe req LI irem en ts.
Establish th e maximum nose height on the basi of (1) the structural limitations required to
preve nt failure at the jun ction of the nose and sea l-r ing base and (2) the potential face
distortion caLi ed by th ermal and press ure stresses.
The nose he ight 011 turbopump seals should be from 0.030 to 0 .060 in. The h eight-to-width
ratio should be ap pro x im a tely 0.3 to 0.8.
U e a minimum fillet radius of O.O~O Lo 0 .030 in. at th e jun ction of the no e and ca l-ring
base on brittle m ate ri a l (carbon).
Allow a minimum or 0.005 in . ca rbon base above th e m e tal ca rrier (fig. 36(b)) on in ert
des ign s fo r oxidizer se rvi ce .
3.3.2 .3 SEAL-RING ANT IROTAT ION DEVICE
Thc all/irotatioll dcvice shall pr(!)lellt til e seal rillg .limn mtatillg but shall IlOt
rcstrict a.ria/llw»clllellt.
Use two or more tang or block on th e stationary hou sing that e ngage slot on th e ca l ring
/lange or radial splin e. Th e tang contact area should be sufficient to prevent ind e ntations in
th e lots fro m the impact loads. Minimi ze the tang-to-slot clearance. Th e tang material
should be o f tel' than th e slot material.
Avoi I usin g antiro tation pins beca use the y re ult in a higher contact loa I and te nd to wear
indentations in the lot.
119
3.3.2.4 SEAL-R ING PI LOT
Th e sea l-rillg pilo / s/zalllllailllain preci.'c! mdia//oca/ion ol the sea / and allolV th e
required lace angular misa lignm ent.
Loca te thc ea l-ring pil o t (fig. 12) at th e cco ndary ca l and maintain a Icngth-t o-di ame tc r
ratio of approximately 0 .05 to 0.1. The pil ot diam etral c1carancc sho uld gc nerall y be abo ut
0.003 il1. j in. o f diam c ter at o pera ting co nditi o n . Adjust th ' ambi en t cleara nce to provide
for th el111al differential and pres ure dellection. Radial sp lin c may al 0 bc u ed to
maintain radial alignmcnt a nd co mp ensatc for th crmal diffe rential. Wh cn nccessa ry. usc
hard-chromc plating on metalli c sea l ring at th e pil o t diam etc r to elimin atc fre tting damage.
3.3.2.5 SEAL INSERT RETENTION
Til e sea l-ring insen slloll be rewined at o /J C'rating CU llditio ll S.
ea l-ring in erts sho uld be retain ed in th e ca rri e r with an interference fit at th e 00 of thc
in ert and with adh c i c bonding applicd to th e botto m o f th e insert (fig. 36(b)). The
diallletral interferencc ror carbon in se rt s at operating co nditions hould be a minimum of
appro ximately 0 .00 I in. j in. of diamcte r. sc sp ec ial steel s with low xp an ion and
co ntra ction rates (e .g., Invar 36, Carpcntc r 42 , mo lybd enum stcel) to minimi zc th e
interfcrencc changc with tempcra turc.
Th e cpo xy adhesives reco mm e nded for bo ndin g are Ii ted in tabl e III. void using ad hesive
bonding a th e o nl y rete nt io n meth od when th e insert tempe rature may exceed
0
approximately 300 F.
3.3 .2.6 INSERT/ CARRIER SEPARATION
Tlt e int erferencC' ./i/ () II in.\'er/ · witll inside /Jressure shall be su.llicien/ /0 ellsure
llial til e re/en /i()11 j(J rc(' is grC'a ter tll an the separating lorce.
E tabli h th e required interi"eren cc o n th c basis o f the relative valu es o f th e press ure

sepa ratin g force betwcen th e insert and ca rri e r and th e retention force ca u ed by fri ction at
th e in ert 00. Th e fri ct io n rorce i detc nnin cd by th e unit contact load a t th e in er t 00
required to de lkct th e in se rt radially inward by th e amount of th e interference . The in sert
rigid it y and the coe ffi cient of fr iction dete rmin e th c necessa ry interference. Th e effec t of
therma l differentials and pressure de fl ec tion s on th c interfere nce must be co nsid ered.
Consid er mec hani ca l lock for insert retention wh en it is not fea ibl e to provid e sufficient
retention force with an inte rference fit . The mechani cal loc k hould be pring loaded aga inst
1_
120
the insert with a load greater than the separating force to prevent slight movements that can
cause seal-face distortion.
Use epoxy adhesives to exclude the high-pressure fluid from the insert-to-carrier joint.
The recommended method for reducing the pressure separating force for seals with higher
pressure at the inside diameter is shown in figure 38(b).
3.3.2.7 INSERT DISTORTION
Inserts retained in a metal carrier with an interference fit shall not be distorted
excessively when chilled to cryogenic temperature.
To minimize the interference change with temperature, use the special steels (Invar 36,
Carpenter 42, molybdenum steel) with a thermal-contraction rate similar to that of the
insert material.
Align the insert and carrier cross-section centroids (fig. 39(b)) as closely as possible to
minimize the bending moment. The bending force should be minimized by reducing the
cross-sectional area of the insert.
3.3.2.8 LAPPED-JOINT SEAL-RING/CARRIER ASSEMBLY
The lapped joint shall (1) allow relative movement between the seal ring and
carrier plate to minimize distortion and (2) maintain sufficient closing force to
effect a static seal.
The seal ring and carrier plate surfaces should be lapped flat (3 helium light bands) and
relieved to provide a resultant pressure closing force (fig. 40). The relieved surface should be
provided with bearing support pads, excep t where distortion can cause the bearing support
pads to separate the seal-ring static seal.
3.3.2.9 SPRAY-COATED SEAL RINGS
The spray coating shall maintain a bond to the seal-ring base adequate to prevent
chipping and [laking.
The plasma-spray process is recommended for aluminum oxide coatings. The thenl1ally
induced stress should be minimized by using a thin (0.010 to 0.020 in.) coating of
aluminum oxide sprayed on a transition layer of Nichrome. Use pure Nichrome on the base
I
121
J
metal, then a mixture of 50-percent Nichrome/50-percent aluminum oxide, and finally pure
aluminum oxide.
Use a raised tapered structure similar to that shown in figure 36(d) for the coating base. The
coating should be applied to both sides of the seal ring to equalize the thermal load and
minimize the thermal distortion .
3.3.3 Circumferential-Seal Rubbing Elements
3.3.3.1 SEGMENTED SHAFT SEALS
3.3.3.1.1 Segment Loading
Th e radial and axial loading on a segment shall be sufficient for effective sealing
without exceeding the load-speed-life limitations fo r specific seal materials and
sealed fluids.
To maintain shaft contact, the segments should be loaded in th e radial direction with a
garter spring (fig. 20). The garter-spring load should be approximately 0.1 to 0.2 lbf/in. of
circumference.
Axial lo ading should be maintained with a wave spling (fig. 20) providing approximately 0 .5
to 1.0 lbf/in. of circumference. The garter-spring load should be higher th an the friction
force caused by the wave spring.
The differential axial and radial pressure forces should be partially balanced by relieving the
contact surfaces (fig. 41). The resultant shaft contact load should be consistent with the
load-speed-life limitations.
3.3.3.1.2 Segment Joints
Th e segment jOints shall provide fo r static sealing.
The segment joints should be overlapped with backup segmented rings and lapped flat or
machined to th e same radius for conformance (fig. 20).
3.3.3.1.3 Segment Antirotation Device
Th e antirotation device shall prevent the segments f rom rotating but shall not
restrict th e required radial movement.
122
------- -- -
Use antirotation pins at the segment gaps to prevent rotation. Pin blocks (fig. 21) may be
used to distribute the load over a larger contact area. Provide sufficient clearance between
the pins and segments to allow the required radial movement. Consider dimensional
tolerances and thermal differentials.
3.3.3.1.4 Segment Diameter
The segment diameter shall conform to the shaft diameter at operating

conditions.
The segment design diameter should be established with consideration for the thermal
differential so that the segment has the same radius as the shaft under operating conditions.
3.3 .4 Hydrostatic/Hydrodynamic Face Seal Elements
3.3.4.1 SEL F-EN ERG IZED HYDROSTATIC SEALS
The face geometry shall provide a self-energized hydrostatic fo rce that is balanced
with the pressure and spring closing forces at the design clearance.
Use the recessed pads with orifice compensation or the recessed step or the convergent
tapered face (fig. 15), the choice depending on the specific design requirements.
The recessed-pad design should utilize a minimum of three separate pads around the seal
face to provide for face misalignment stability. The volume of the recessed pads should be
minimized to prevent dynamic instabilities caused by slow response to pressure changes. The
orifice size should be large enough to prevent clogging from contaminants in the sealed
fluid.
The depth of the recessed step should be determined by calculating the required pressure
drop across the recessed area to provide the restoring force necessary to maintain the design
face clearance.
The convergent-taper deSign also should be established by the relative pressure drop across
the tapered area. The average pressure force on the seal face should be equal to the pressure
and spring closing forces at the design clearance.
Conduct a thorough stress and thermal analysis so that the seal face and mating ring
distortions can be minimized (refs. 6 and 7) . Conditions that result in a divergent interface
surface must be avoided because the restoring force is inherently unstable.
Reference 35 is recommended for a discussion of the detail design procedures.
123
3.3.4.2 EXTERNALLY PRESSURIZED HYDROSTATIC SEALS
The interface hydrostatic force resu lting from the external pressurization shall be
equal to the pressure and spring clOSing forces at the design clearance.
Utilize a minimum of three separate pads around the seal face to provide for face
misalignment stability. The volume of the recessed pads should be minimized to prevent
dynamic instabilities. The orifice size should be established to provide the necessary pressure
drop at the required purge flow to maintain the design clearance.
The required orifice pressure drop should be determined in conjunction with the flow
analysis of the face/pad region and the supply system.
Detail design procedures are discussed in references 35 and 46.
3.3.4.3 HYDRODYNAMIC SEA LS
The hy drodynamic lift force shall be sUfficient to balance the net pressure and
spring closing forces at th e design clearance.
The face geometry should be established with consideration for the rotational speed and
fluid viscosity. The secondary-seal diameter or the bellows effective diameter and the spring
load should be established to prevent the closing force from exceeding the available
hydrodynamic lift force. The hydrodynamic design should provide that the seal be pressure
balanced to obtain the desired operating film thickness at the seal interface (ref. 48).
Use materials that are dimensionally stable and suitable for intermittent rubbing contact in
the fluid medium to be encountered.
Conduct a thorough stress and thermal anlaysis so that distortions of the seal face and
mating ring can be minimized (refs. 6 and 7).
Detail design procedures for various applications are discussed in references 31 through 53.
3.3.4.4 HYBRID SEALS
Th e hydrostatic action shall provide the lift force required to maintain the design.
clearance at low rotational speeds, and th e hy drody namic action shall provide
additional lift and dy namic stability at high speeds.
124
l_
The design procedures for the hydrostatic and h ydrodynamic concepts should be integrated
to optimize the hybrid-design configuration. Design methods are discussed in reference 45.
3.3.5 Circumferential-Seal Clearance Elements
3.3.5.1 LABYRINTH SEALS
3.3.5.1.1 Geometry
The labyrinth geometry shall provide maximum sealing effectiveness.
The labyrinth clearance should be a minimum consistent with the installed and operating
tolerances . Step or staggered labyrinths should be used where possible. The teeth tips should
be sharp (0.005 to 0.015 in. R). The number of teeth should be maximized (fig. 44)
consistent with the optimum tooth pitch of approximately 0.1 in. per 0.01 in. diametral
clearance (fig. 45) for straight labyrinths. The cavity depth should be approximately equal
to the tooth pitch.
Detail design procedures for various applications are discussed in references 54 through 65.
3.3 .5.1 .2 Wear-In Labyrinths
Wear-in labyrinth materials shall be qompatible with the fluid environment and
shall be easily deformable or worn away at initial contact.
Metal honeycomb or sintered-metal-fiber abradable materials should be used in nonoxidizing

fluids. The honeycomb foil thickness and cell size should be based on a tradeoff of ease of
deformation and erosion resistance.
The thickness generally should be 0.002 to 0.005 in., the cell width approximately 1/16 in.,
and the cell depth one to two times the width. Cells are oriented normal to the direction of
rotation. The honeycomb strip should be brazed to a support ring for attachment.
Abradable materials sh ould not be used in applications where the wear debris will
contaminate the system. Carbonaceous materials may be used if the interference is not too
large.
Kel-F is the recommended material for wear-in labyrinths in liquid oxygen.
I
I 125
L_
I
3.3.5.1 .3 Plastic Labyrinths
The housing for plastic labyrinths shall provide structural support adequate to
maintain dimensional stability.
The plastic material should be restrained from its natural con traction rate by locking
relatively thin sections into a metal housing (fig. 47). The h ousing sho uld be strong enough
to control the thelmal contraction and provide dimensional stability. A static seal flange and
pressure vents should be provided on high-pressure seals. The length of the plastic span
between retention locks should be approximately 0.5 to 1.0 in.
3.3 .5.1.4 Segmented Labyrinths
Segmented labyrinths shall provide for radial movement with positive stops to
control the wear-in diameter.
Segmented labyrinths should be spring loaded to allow wear-in to a diameter smaller than
the rotor. The segment diameter after wear-in should be controlled by limiting the radially
inward travel with a T-slot arrangement (fig. 48) or by allowing the segments to become
arch-bound. A lapped joint should be provided between the segment and housing for a
static seal. Antirotation devices should be provided to prevent segment rotation.
High-pressure seals should be partially pressure balanced to prevent excessive contact loads
during wear-in.
3.3.5.1.5 Erosion Resistance of Labyrinth Materials
Labyrinth materials shall be resistant to surface erosion caused by cavitation or

impingement of high-energy fluid.
The hard high-strength materials (tool steel, Stellite, maraging steels) are recommended for
labyrinths that may be exposed to severe cavitation or to high-energy-fluid impingement.
The soft low-strength materials (aluminum, Monel, brass, bronze, plastic) should not be used
when surface erosion may be a problem.
3.3.5.2 FLOATING-RING SEALS
3.3 .5.2.1 Clearance
The floating-ring element shall maintain a controlled clearance to the shaft under
all extremes of operating conditions.
126
1--
Use an ou ter-ring material that has the same thermal expansion and contraction rate as the
shaft material, so that a cons tan t clearance gap is maintained as the temperature changes.
The ring radial section should be sized to provide sufficient rigidity to minimize the radial
deflection due to the unbalanced pressure while minimizing the seal ring inertia to achieve
maximum response to eccentric shaft rotation. The seal ring-to-shaft operating clearance
should provide adequate leakage control and sufficient margin to allow for the thennal and
pressure deflections; an operating diametral clearance of 0.0005 to 0.001 in./in. of diameter
is recommended.
3.3.5 .2.2 Wear
Th e inner ring and shaft surface shall withstand the wear resulting from the
intermittent contact required to center the ring.
The inner ring materials should be selected for minimum friction and wear. The material
should be compatible with the sealed fluid and self-lubricating if the fluid is not a lubricant.
Carbonaceous materials are recommended for most applications . The shaft surface should be
plated with a hard wear-resistant material (h ard -chrome plate, ch romium carbide, or
tungsten carbide). The recommended seal materials based on current practices are given in
table III.
3.3.5.2.3 Axial Loading
The /loating-ring axial load shall be adequate for static sealing but shall not
restrict radial movement.
The differential axial pressure force (fig. 49) on high-pressure seals should be partially
balanced by relieving the axial contact surface and minim izing th e housing-to-shaft
clearance. Use a wave sp11ng on low-pressure seals to maintain axial contact.
3.3.5.2.4 Antirotation Device
The antirotation device shall restrain the floating ring from rotating without
restricting its free radial or axial movement.
The outer steel ring should have two or more mach ined tangs that engage slots in the
housing to prevent rotation.
127
3.3.5.3 ARCH -BOUND SEGMENTED SEALS
The segments shall become arch-bound after wear-in at the operating conditions
to form a solid floating ring with minimum shaft contact load.
Adjust the segment design diameter to compensate for the deflection caused by the radial
pressure load and the thermal contraction or expansion differentiaL The diameter should be
approximately 0.00 1 in. smaller than the shaft at operating conditions to allow wear-in . The
recommended materials based on current practices are given in table III.
3.3.6 Face-Contact-Seal Secondary Elements
3.3 .6.1 METAL BELLOWS
3.3.6.1.1 Function
The bellows element shall act as the secondary seal and provide structural support
for the seal ring.
The bellows element should be leak proof and allow for motion in the axial direction. It
should be sufficiently rigid in the torsional and lateral directions to act as the antirotation
and piloting device. Mechanical stops should be used to establish the free length and prevent
overcom pression.
3.3.6.1.2 Spring Force
The bellows elements shall provide the spring force necessary to maintain face
contact.
The bellows type, diameter, plate thickness, span, convolutions, and number of plies should
be consistent with the space available and with the stress requirements, so that the desired
spring force and compression range are provided. The bellows stress and spring rate should
be estimated with the use of available analytical techniques (refs. 69 through 71) and the
supplier's empirical data for specific configurations.
Minimize the spring rate by using the maximum number of convolutions that can be fitted
into the available space with the convolution pitch necessary to provide the required
compression range and the plate thickness necessary to satisfy the pressure and deflection
stress requirements. Double-ply bellows may be used to increase the pressure capacity
without significantly increasing the spring rate.
128
3.3.6. 1.3 Pressure Capacity
The bellows element shall withstand the maximum differential pressure without
significant permanent defo rmation.
The bellows pressure capacity should be estimated with the available analytical techniques
(refs. 69 through 71) and the supplier's empirical data for specific configurations. The
bellows should be stabilized with a proof pressurization at the minimum operating
compression followed by a stress-relief heat treatment. The change in free length as a result
of pressurization should not exceed ±O .005 in . after the bellows have been stabilized.
3.3.6.1 .4 Effective Diameter
The bellows effective diameter shall be sufficiently constant to allow adequate

pressure balance.
When accurate pressure balance is required , the bellows should have minimum variation of
effective diameter with pressure and pitch. Thicker plates, smaller span, and a larger pitch
should be used to minimize the effective diameter change.
The bellows should be calibrated to measure the effective diameter or total load (spring plus
pressure) at the limits of operating pressure and compression before the seal face diameters
are established for force balance.
3.3.6.2 PLASTIC LIP SEALS
3.3.6.2 .1 Load ing
Th e lip-seal elements shall provide for pressure and spring loading of the lip .
Establish the lip-seal length and housing support geometlY (fig. 53) to provide a differential
pressure force that will increase the seal seating stress at higher pressure. The seal should be
spring loaded against the housing with a flat retainer ling to compensate for the thermal
contraction.
3.3.6.2.2 Geometry
Th e lip contour and thickness shall provide conformance to the seal ring and
flexibility for misalignment.
129
-------
Dimension the lip-seal radius and housing support radius to provide a smooth transition with
the maximum lip support between the housing and seal ring. The lip-seal thickness should be
approximately 0.005 to 0.0 lOin. to maintain flexibility for conformance and to minimize
the thermal loads.
3.3.6.2.3 Interference
The lip seal shall maintain a diametral interference with the seal ring at extreme
operating conditions.
Dimension the lip-seal diameter to provide a diametral interference of approximately 0.005

in./in. of diameter at room temperature. The material should be selected to allow the
interference to increase at operating temperature . Size the seal ring-to-housing clearance and
the lip length to maintain an interference under the wor t conditions of misalignment.
3.3.6.3 ELASTOMERS
3.3.6.3.1 Squeeze
The groove dimensions shall provide squeeze adequate to maintain sealing contact
and prevent excessive friction at extreme conditions.
Control the groove dimensions with precision tolerances. A tolerance stackup to establish
the dimensional limits should be performed. The effects of thermal contraction or
expansion and shrinkage or swelling should be considered to establish the squeeze. An
operating squeeze of 5 to 10 percent is recommended for secondary elastomer seals.
Detail design requirements are given in references 72 and 73.
3.3.6.3.2 Sealing Surface
The elastomer sealing surface shall minimize friction and wear.
The sealing surface should be either hardened steel or hard-chrome plating, finished to 5 to
10 }.Lin. rms. The metal surface may also be coated with 0 .001 - to O.003 -in .-thick Teflon to
minimize friction and wear.
I
130
I
J
3.3.6.3.3 Lubrication
Tlie lubrica tio ll shall b e suffic ient to prevent rolling, tlVis /ing, or sticking oj' th e
elL/SWiller.
Th e e la tomer and all m etal p arts in contact with th e e lastomer hould be lubri ca ted with a
compatible lub ricant ( re f. 73) . Silico ne oil or grease is recommended ror Buna and Vitoll:
do not , however, use Bun a or silicon e oil or grease in liquid o r ga eo u oxygen.
3.3.6.4 PISTON RINGS
The ring alld cylinder sealillg sUlfaces shall provide lill C'-to-lin e con /a ct lvi/Ii
minimum Fiction and wear.
The piston ring sliding surface should be h ard en ed tee l or h ard c hrollle plate, finished to 8
to 16M in . rms , an d round within 0.0003 in ./in. of diameter. Th e cy lindri ca l taper of the
ealin g diameter should not exceed 0.0003 in ./in . of le ngth , and th e squareness of th>
diam e ter should not exceed 0.0005 in. / in. of groove depth (ref. 74) at all operating
conditions. The effects of axial thermal gradients on the angular displ acem ent of the groove
seal surface should be co nsid ered.
3.3.7 Spring Load

Th e spring load sha ll provide effective sealing alld shall be consistent lVith the
load-speed-lile rela tiolls.
Esta blish the minimum load for effec tiv e sealin g by ca lcu lat ing the force required to
accelerate the sea l ring at the rate required to track th e rotating m ati ng surface run o u t or to
follow sudd en shaft moveme nts. The resultant closing fo rce shou ld be grea ter than th e sum
of the seal-rin g inertia, eco ndary friction, and interface press ure profile forces.
Establish the maximum load by the load-speed-life relatio nship s for specific fluid s a nd
materi als or by th e seal powe r-l oss I imi ta tions. Calcula te th e all owa bl e hea tin pu t by
heat-balance an alysis.
The recommend ed sprin g load per unit c ircumfe re nti al le ngth o f sea l face are as follow s:
Sealed fluid Reco mm e nd ed pring load , Ibf/ in.
Liquid oxygen 2 to 4
Liquid h ydroge n I to 2
Gaseous h yd rogen 1 to:2 (continued)
,
/31
r
l_
Hot ga (H2 + H 2 0) I to :2
RP-I I to 4
Liquid fluorin e 2 to 3
3.3.8 Pressure Balance

The pressure balallce shall prevellt ex cess iJl e JCI(:e load at high pressure alld prolJid e
adeqlfa t e lJIargill j f)r I'Clriatioll ill the illtel jcl c(! IJressli r e profile.
Sea l with difrerenti al prc urc in excess of 100 psi should be press ure balanced. The
pres ure-bal ance ratio should be selec ted with co n id eration for th e maximum variation of
interface pre sure pro ril e and th e total (spring pIll press ure) face-load requireme nts. The
foll ow in g pre sure-bal a nce ra ti os are recom mend ed (ta ble I V):
Table IV. - Recommended Pressure-Balance Ratios for Turbopump Rotating-Shaft Seals
Sealed fluid condition Seal requirement Recommended balance ratio
High-pressure Minimum face load 0.55 to 0 .6

incompressible
Low-pressure Minimum load and adequate 0.6 to 0 .7

incompressible margin for interface
(conven tional designs) pressure variation
High-pressure Minimum face load 0.65 to 0.7

cryogenic·
Low-pressure Maximum sealing effectiveness 0.8 to 0.9

cryogenic·
High-pressure, sonic Minimum face load 0.67 to 0 .75

(choked) flow, . as determined by analysis (ref. 12)
compressible
Low-pressure , subsonic Minimum load and adequate 0.6 to 0.7

laminar flow, margin as determined by the following equa-
compressible tion (eq. (14)) plus required design
margin :
t
a""'3I I + I +Pd 2
P
(PdP 2 )
J
*Listed as a separate category because flujd can change from liquid to vapor as it crosses seal interface.
132
_ _ _ _ _ _ _ _ _ ---.J
APPENDIX A
Conversion of U.S. Customary Units to SI Units
Physical quantity U.S. customary unit SI unit Conve rsion factor

.
Angle deg rad 1.745xlO - 2
Angular velocity rad/sec rad/sec 1.
. 2
Area ffi. cm 2 6 .452
Density Ibm/in .3 kg/m 3 2.768x104
Energy Btu J I.054x10 3
Force lbf N 4.448
FV factor Ibf/in. X ft/sec N/cm X m/sec 5.338xlO - 1
Gas constant (specific) in.-Ibf/(lb m-OR) J/(kg-K) 4 .483xIO - 1
Heat generation rate Btu/sec J/sec 1.054xlO 3
Heat generation rate Btu/(sec-in.) J/(sec-<:m) 4.l 50x[Q2

per unit length
Heat generation rate Btu/(sec-in. 2) J/(sec-<:m2) 1.634xlO 2

per unit area
Impact energy level ft-Ibf J 1.356
Leakage rate gpm (gal/min) cm 3 /min 3.785x10 3
sefm (std eu ft/min) em 3 /min 2.832x104
Length ft m 3.048xlO- 1
in. em 2.54
j.lin. j.lm 2.54xlO- 2
Load/unit length lbf/in. N/em 1.751

(continued)
133
Conversion of U.S. Customary Uni ts to SI Units (concluded)
Physical quantity U.S. customary unit SI unit Conve rsion facto r*
Mass Ibm kg 4.536x l O- 1
Mass density lbm-se c2 /in .4 kg-sec 2 /cm 4 l.090x lO- 2
Modulus of elasticity psi (Ib f/ in 2 ) N/cm 2 6 .895x lO - 1
Pressure psi N/ cm 2 6 .895x l O- 1
PV fac to r Ib f/in ? X ft/sec N/cm 2 X m /sec 2. 102x l O- 1
Ro tati onal speed rev/sec rad/sec 6.283
rpm (rev/ min) rad/sec 1.047x l O- 1
Sp ring rate Ib f/ in . N/cm 1.751
Surface speed ft/sec (fps) m/ sec 3.048x lO - 1

(rubbing speed)
Tempera ture of K 5
K = - (" F + 4 59 .67)
9
oR K K= ~ (" R)
9
Un it load psi N/cm 2 6.895x l O- 1
Vi scosity, absolute Ib f-sec/ in 2 N-sec/cm 2 6.895x lO - 1
Viscosi ty, dynamic Ibm/(in.-sec) N-sec/cm 2 1.786x lO- 3
Ibm/(ft-hr) N-sec/cm 2 4. 134x lO- 8
Viscosity. kin emat ic in.1 /sec cm 2 /sec 6.452
Wear rat e in./hr mm/h r 2.54x 10 1
Yie ld strength psi N/cm 2 6.895x l O- 1
ksi ( 1000 psi) /cm 2 6.895x I 0 2
Young 's modu lus psi N/cm 2 6 .895xlO- 1
*I:. :\.ccpt for temperature convers io ns, whi ch arc to be made as sho wn, m ul tiply val ue given in U .S . customary unit by
conversion fac lor 10 obtain eq uivalent va lue in 5 1 unil. For a co mplete lisling of conversion factors for basic physical
quantities, sec Mecht ly, E. A.: The Internati onal System of Units . Ph ysical Constants and Conversion Factors. Sec-
ond Revisio n, NASA SP-70J2 , 1973 .
134
APPENDIX B
GLOSSARY
Term* or Symbol Definition
b sea l face ave rage circ um fe renti al width , in.
sea l face effec tive closin g area

balall ce ra ti o
sca lin g face (d am) area
co nstant fo r bell ows configurati on
d sea l face ave rage di ame ter, in .
Young's modulus, psi
F resultant tota l contact load per unit circumferenti al length , Ibf/in.
friction force, lbf
Fw wave spring loa d , Ibf
FV factor total contact load per unit circum fe rential length tim es rubbin g speed ,
Ib f/in . X ft /sec
r coe ffi cient of fri cti on
fps fee t per seco nd
g acce le rati on du e to grav ity, ft /sec 2
effec ti ve (a verage) leakage ga p, in.
ID in sid e di ame ter. in.
mechani cal equi va lent of hea t, ft-Ibf/ Btu
ax ial spring rate, Ib f/ in.
L sea l face radial length , in.
MED (or M.E.D .) mea n effect ive diame ter, in .
N rota tional speed , rev /sec

*As no ted , sea l te rmin ology used in the preceding text basically is that presented in ASLE SP-l (ref.l), which should be
reviewed for co mple te cove rage of terms .
135
Term or Symbol Definition
number of convolutions
number of plies
number of throttlings in a labyrinth seal
OD outside diameter, in.
P face contact unit load, psi
P average pressure, (P 2 + PI )/2, psia
PI downstream pressure, psia
P2 upstream pressure, psia
Prl pressure at inner radius r I, psia
Pr2 pressure at outer radius r2,psia
PV factor total contact unit load times rubbing speed, Ibf/in. 2 X ft/sec
PrV factor flu id pressure time~ rubbing speed, Ibf/in.2 X ft/sec
Q volume flow, in. 3 /sec
q heat generation rate per unit circumferential length, Btu/(sec-in.)
q' heat generation rate per unit contact area, Btu/(sec-in. 2 )
R specific gas constant, in. -lbf/(lbm-OR)
R bellows mean radius, in.
Re Reynolds number
R&D research and development
seal face average radius, in.
rms root mean square
136
L_
seal face inner radius, in.
seal face outer radius, in.
SSME space shuttle main engine
radial OD - radiallD
s half of bellows space (i .e.; ), in.
. 2
scfm standard cubic feet per minute
T absolute temperature, oR
T.I.R. total indicated runout
bellows plate thickness, in .
total load spring load + hydraulic load (based on linear profile), lbf
u leakage flow velocity, in./sec
unit load total contact load -;. nose contact area, Ibf/in .2
v surface speed, ft/sec
w weight flowrate, Ibm/sec
pressure profile factor
molecular mean free path at average pressure, in.
incremental change in a quantity, or difference between two quantities
viscositY,lbm/in .-sec
absolute viscosity, Ibf-sec/in.2
v kinematic viscosity, in. 2 /sec
p density, Ibm/in. 3
mass density pig, Ibm -sec 2 /in .4
flow coefficient
137
w angular velocity , rad/sec
leakage function, 1/1 = V I - (P1 /P2 )2

T4 + In PI
( /P)
2
Material* Identification
Aerozine 50 or A-50 50/50 blend of N2l4 and UDMH, propellant grade per MIL-P-27402
A-286 austenitic heat-resistant iron-base alloy; same as AMS 5735
AM-350, AM-355 semi -austenitic hea t-resis tan t precipi ta tion -hardening stainless steels
AmCerMet 701-65 sintered Ni-Cr alloy filled with fluoride eutectic ; manufactured by
Astro Met Associates, Inc.
AMS 4530, AMS 4650 Cu-Be alloys
AMS 5646 wrought Ni-Cr corrosion- and heat-resistant steel
AMS 5665 nickel-base Cr-Fe alloy; same as Inconel 600
AMS 5735 austenitic iron-base alloy containing 15 Cr, 26 Ni, and 1.3 Mo;same as
A-286
Bearium B- IO trade designation of Bearium Metals Corp. (Rochester, NY) for leaded
bronze
Buna N trade designation for copolymer of butadiene and acrylonitrile
Carpenter 42 trade designation for low-expansion steel alloy (42% Ni, balance Fe)
manufactured by Carpenter Steel Div., Carpenter Technology Corp.
(Reading, PA)
CCA72, CCA-72 trade designation for carbon (graphite) manufactured by Union Carbide
Corp.
CDJ83 trade designation for carbon (graphite) manufactured by Union Carbide

Corp.
cermet material composed of a ceramic with a metal binder; e.g., tungsten

carbide with a cobalt binder
• Additional information on metallic materials herein can be found in the 1972 SAE Handbook, SAE, Two Pennsylvania
Plaza, New York, NY ; in MIL-HDBK-SB, Metallic Materials and Elements for Aerospace Vehicle Structures , Dept. of
Defense, Washington , DC, Sept. 1971; and in Metals Handbook (8th ed.), Vol. 1 : Properties and Selection of Metals, Am.
Society for Metals (Metals Park , OH), 1961.
138
Material Identification
chrome chromium
designation for a fused alumina of grea t hardness
CRES corrosion-resis tan t steel
elastomer polymeric material that at room temperature can be stretched to

approximately twice its length and on release return quickly to its
original length
Epon 422 , Epon 901/B3 trade designations for epoxy adhesives manufactured by Shell Chemical
Co.
epoxy thermosetting resin widely utilized as a binder in the fabrication of

glass-filament/resin composites and as an adhesive
EY-JOS trade designation for carbon (graphite) manufactured by Morganite ,

Inc.
FLOX mixture of 30% LF2 and 70% LOX
gaseous hydrogen
G 2 gaseous nitrogen per MIL-P-2740 I
G39, G84, G84SC trade designations for carbon (graphite) manufactured by U.S. Graphite
Has tell oy trade designation for a series of higll-temperature nickel-base alloys

manufactured by Stellite Div., Cabot Corp.
helium pressurant helium (He) per MIL-P-27407
HF hydrofluoric acid
hydrogen-rich steam (combustion product of L0 2 + LH 2 )
hot gas combustion products or gaseous discharge from a heat exchanger;

hot-gas temperature can reach 1300° F , depending on the process , in
state-of-the-art propulsion systems
HT424 trade deSignation for epoxy adhesive manufactured by American

Cyanamid Co.
hydrazine N2 H4 , propellant grade per MIL-P-26S36
139
Inconel 600, 71 8, trade designations of International Nickel Co. for Ni-Cr-Fe alloys
X ( now X-750)
Inconel 5665 same as AMS 5665 and Inconel 600
Inconel 5667 nickel-base Cr-Fe alloy; same as Inconel X-750
trade designation of Carpenter Steel Division, Carpenter Technology

Corporation, for a nickel-base alloy (36% Ni, balance Fe) with a low
coefficient of expansion
_I ~ trade designation of 3M Corp . for a high-molecular-weight polymer of

chlorotrifluoroethylene
trade designation of Kennametal, Inc. (Latrobe, PA) for titanium

carbide bonded with a Ni-Mo alloy
trade designation of International Nickel Co. for a wrought

age-hardenable alloy containing Ni, Cu, and AI
liqUid fluorine, propellant grade per MIL-P-27405
liquid hydrogen (H 2 ), propellant grade per MIL-P-27201
LOX, L0 2 iiquid oxygen (0 2 ), propellant grade per MIL-P-25508
LW I trade designation for a flame -plating form of tungsten carbide

manufactured by Linde Co.
. , LW-5
LW5 trade design ation for a flame-plating form of chromium carbide
manufactured by Linde Co.
maraging steei ma rtensite- and age-hardening Ni-Fe alloy
molybdenum steel a type of tool steel whose hardness is enhanced by its molybdenum
content
Monel trade designation of International Nickel Co. for an alloy containing

mainly Ni and Cu
Mylar lldde designation of E.!. duPont de Nemours and Co. for a

polyethylene terephthalate film
.1) 0wzine propellant grade per MIL-P-26536
140
Material ldentifica tion
nitroge n tetroxide (oxidizer), propell ant grad e per MIL-P-26539
Nichrome trade designation of Dri ver- Harri s Co . (Harrison , IJ) for a

high-tem perature corrosion-res istan t Ni-Cr al loy
para-hydrogen type of molecular hydroge n having an antipara ll el nuclear sp in :

preferred over ortho-hydrogen for rocket fu els
pl astic high-molecul ar-weight material th at while u uall y firm and hard in its
finished state is at some stage in its ll1 anu raclUre so rt enough to be
fo rm ed into a des ired shap e by appli ca ti on or hea t or press ure or both
P03 , P2003 trade designations for carbon (graphit e) manufact ured by Pure Carbon
Co.
P5AG trade designation of Pu re arbon Co . for carbo n (graphite) im pregnated

wi th sil ver (Ag)
P5 trade designation of Pure Carbon Co . for carbo n (graphite) impregnated

with metal-flu orid e alt
P69 2 trade designation of Pure Carbon Co . fo r ca rb on (graphile) imp reg nate d

with resin
RP-l kerosene-base high -energy hydrocarbon fuel, propell an t grade per

MIL-P-255 76
Stellite designati on for a serie of hard Co-W-Cr-C alloys manufactured by

Stellite Div. , Cabot Corp .
sy nthetic rubb er any of a gro up of man-mad e elas tomers whose properti es approximate
one or more of the properti es of natural rubber
T6, T73 , T86, T351 designati ons for heat-treating and tempering processes for aluminum
alloys
Tefl on (TFE) trade designation of E.l. duPont de emours and Co. for polymer of
tet rafl uoroe thylene
tool steel hard steel intended for use in tools; co ntains a high percentage of C or
othe r hardness-imparting elements (e.g. , Mo)
UDMH unsymmetrical dimethyihydrazine, propell ant grade per M IL-P-25604
141
Vit on A trad e de ignati on of E.!. duPont de Nemours and Co . fo r copolymer of

vin ylid ene nu orid e and hexa nu o ropropylene
17-7PH semi-austenitic precipit ati o n-hard ening stai nl ess stee l
300 se ri e (e.g.. se ri es of austeniti c stainl ess steels

304.3 10.347 )
400 sc ries (c .g .. se ries o f martensiti c and fe rritic stainl ess stee ls

4 16.440C )
201 4 . 2024 wrought alumi num all oys with Cu a th e prin cip al all oyin g elem ent
41 30 , 4340 high-strength. martensite-h ard enin g, low-all oy stee ls
606 1,6 15 1 wrought aluminum all oys with Mg and Si as th c prin cip al all oy ing
elcmen ts
7075, 70 79 wro ught aluminulll all oys with Zn as th e prin cip al all oy in g elcment
ORGANIZA Tl ON
Abbreviation Iden tifica tion
AEC At omi c Ene rgy Commission
AFML Air Force Mate ri als Laboratory
AFRPL Ai r Fo rce Rocke t Propul sion Labora to ry
AIS I Am cri ca n Iro n & Stee l In stitute
AMS Aero pace Materials Spec ifica ti on
ASLE American Society of Lubri ca ti on En gin ee rs
ASM Am eri ca n Soc iety fo r Metals
ASME Am eri ca n Soc iety of Mechanica l En gin ee rs
ASTM Am eri can Society fo r Testin g and Materials
DMIC Defense Metals In fo rmati o n Center
142
Abbrevia tion Identification
liT illinois ln sti tu te of Technology
SAE Society of Automotive Engineers
WADD Wright Air Development Division
WPAFB Wright Patterson Air Force Base
143
144
~-
J
REFERENCES
1. Brown, P. F. , Jr., ed .: A Glossary of Seal Terms. ASLE SP-l , ASLE (Park Ridge, IL), 1969.
2. Roesch, E.; and Pasternak, I.: Development of Large Size Bell ows Face Type Seals for Liquid Oxygen
and Oxygen / Hydrogen Hot Gas Service at Moderate to High Pressures. ASA CR-65818, 1966.
3. Colbert, J. E. ; Mosier , S. A. ; and Bailey , T. E.: Flox/Methane Pump-Fed Engine Study. ASA
CR-72485, PWA FR-3040 , Pratt & Whitney Aircraft, Florida Research and Devel opment Center
(West Palin Beach, FL), May 1969.
4. Bail ey, T. E.: Research on a Hydrogen -Fluorine Propulsi on System , Final Report. NASA CR-72074,
PWA FR-1585 , Pratt & Whitney Aircraft, Florida Research and Development Center (West Palm
Beach, FL), Oc tober 1966.
5. Munk, W. R.; and Bailey, T. E.: Hydrogen-Fluorine Engine Evolution (U) . ASA CR-72481 , PWA
FR-3030 , Pratt & Whitney Aircraft, Florida Research and Development Center (West Palm Beach ,
FL), October 1969. (CONFIDENTIAL)
6. Russell , T. E.; et a1.: Gas Turbine Face Seal Thermal Deformation and Computer Program for
Calculation of Axisymmetric Temperature Field. ASA T D-5605 , 1969.
7. Robinson, R. P. ; and Burton, R. : Temperature Gradients in Seal Rings: A Method of Analysis. Paper
12, ASLE 4th Int'l Conf. on Fluid Sealing (Philadelphia , PA) , May 5-8:1969.
8. Bauer, P.; Glickman , M.; and Iwatsuki, F.: Analytical Techniques for the Design of Seals for Use in
Rocket Propulsion Systems. Vol. II - Dynamic Seals . AFRPL-TR-65-61 (AD 464959) , IIT Research
Ins titute, May 1965.
9. Sneck, H. J.: Study of Dynamic and Static Seals for Liquid Rocket Engines, Final Report. Vol.
2 - The Effects of Geometry and Inertia on Face Seal Performance. NASA CR-89737 , S-67 -1150 ,
General Electric Co., Mar. 3 1, 1967 .
10 . Hirschfelder, J. 0. ; Curtiss , C. F.; and Bird , R. B. : Molecu lar Theory of Gases and Liquids.
Wiley-lnterscience, 1964.
II. Hinze, J. 0.: Turbulence: An Introduction to its Mechanism and Theory . McGraw-Hill Book Co. ,
1959 , p. 67.
12. Zuk , 1. ; Ludwig, L. ; and Johnson, R. : Qu asi-One-Dimensional Compressible Flow Acro s Face Seals
and arrow Slots. I-Analysis. ASA TN D-6668 , 1972 .
13. Golf, L.: Final Rep ort - Investigation of Positive Type Shaft Seals - Leakage Correlation. R-7986,
Rocketdyne Div. , North American Rockwell , In c., ov . 14, 1969.
145
L
l
*14. Anon.: LOX Pump Seal Leakage Control Progress Summ ary . LER 71524005 , Rocketdyne Div.,
North Am. Aviation, Inc., Sept. 21,1967.
* 15 . Anon .: Boattail Leakage Analysis . LER 71524006, Rocketdyne Div. , North Am. Aviation , Inc. , Dec.
31 , 1967.
16. Hudelson , J. C.: Dynamic Instability of Undamped Bellows Face Seals in Cryogenic Liquid . ASA
TN 0-3198 , 1966.
17. Boyd , W. K. ; Bery, W. E.; an d White , E. L.: Compatibility of Materials with Rocket Propellants and
Oxidizers. DMIC Memo. 201 , Battelle Memorial Institute , Jan. 29 , 1965.
18. Key , C. F.; and Riehl, W. A.: Compatibility of Materials with Liquid Oxygen. NASATMX-54611 ,
December 1963.
19. Schmidt, H. W.: Fluorine and Fluorine-Oxygen Mixtures in Rocket Systems. ASA SP-3037 , 1967.
20. Butner, M. F.: Final Report, Propellant Lubrication Properties Investigati on. WADD-TR-61-77 , Part I
(AD 259 143), June 1961.
21. Walter, R. J. ; and Chandler, W. T.: Effects of High-Pressure Hydrogen on Metals. Paper presented at
the ASTM-ASME-ASM Symposium on Effects of Gaseous Hydrogen on Me tals, Materials Engineering
Congress (Detroit , MI) , October 1968.
22. Fletcher, E.; and Elsen, A.: Hydrogen-Induced, Delayed, Brittle Failures of High Strength Steels.
DMIC Memo. 196 , Battelle Memorial Institute , 1964.
23. Schwartzberg, F. R.; et al.: Cryogenic Materials Data Handbook . 2 vols. AFML-TDR-64-280 (AD
713619 and AD 713620) , Air Force Materials Laboratory (WPAFB, OH), July 1970.
24. Anon .: Aerospace Structural Metals Handbook . AFML-TR-68-115, Vols. 1, 2, and 3 , Air Force
Materials Laboratory (WPAFB, OH) , 1969.
25. Sliney, H. E.; Strom, T. N.; and Allen G. P.: Fused Fluoride Coatings as Solid Lubricants in Liquid
Sodium, Hydrogen , Vacuum, and Air. NASA TN 0-2348 , 1964.
26. Hady, W. F. ; Allen , G. P. ; Sliney, H. E.; and Johnson , R. L.: Friction Wear and Dynamic Seal Studies
in LiqUid Fluorine and Liquid Oxygen. ASA TN D-2453 , 1964.
27. Hady, W. F. ; et al.: Wear and Friction of Mechanical Carbons in LiqUid Oxygen as Influenced by
Transfer Films. Trans . ASLE, vol. 6 , no. 3, July 1963 , pp . 201-208.
28. Paxton, R. R. ; and Shobert, W. R.: Testing High Speed Seal Carbons. Trans. ASLE , vol. 5, no. 2 ,
November 1962 , pp . 308-314.
"Dossier for design criteria monograph " Liquid Rocket Engine Turbopump Rotating-Shaft Seals." Unpublished. Coll ected
source material available for inspection at Lewis ReseaIch Center, Cleveland, Ohio.
146
L_ __J
29. Wisander, D. W.; and Johnson, R. L.: Wear and Friction of Impregnated Carbon Seal Materials in
Liquid itrogen and Hydrogen. Advances in Cryogenic Engineering, vol. 6, Plenum Press. Inc . ( ew
York), 1961.
30. Anon.: Dynamic Sealing: Theory and Practice . Parts r through VI. Koppers Company, Inc .,
Baltimore, MD, September 1976.
31. Hawkins, R. M.: Developmen t of Compressor End Seals, Stator Interstage Seals, and Stator Pivot
Seals in Advanced Air Breathing Propulsion Systems. NASA CR-72819 , 1970.
32. Povinelli, V. P. ; and McKibbin, R. H.: Development of Mainshaft Seals for Advanced Breathing
Propulsion Systems - Phase II. NASA CR-7273 7, 1970.
33. Povinelli, V. P.; and McKibbin, R. H.: Development of Mainshaft Seals for Advanced Air Breathing
Propulsion Systems - Phase Ill. NASA CR-72987, 1971.
34. Zuk, J. ; et al.: Design Study of Shaft Face Seal with Self-Acting Lift Augmentation. 1. Self-Acting
Pad Geometry . NASA TN 0-5744 , 1970.
35. Parks, A. J .; et al.: Development of Mainshaft Seals for Advanced Air Breathing Propulsion Systems.
NASA CR-72338 , 1967.
36. Johnson, R. L.; and Ludwig, L. P.: Shaft Face Seal with Self-Acting Lift Augmentation for Advanced
Gas Turbine Engines. NASA TN 0-5170, 1969.
37. Lynwander, Peter: Self-Acting Seals for Helicopter Engines , Final Report. ASA CR-134940 , LYC
75-78, Avco Lycoming Division, October 1975.
38. O'Brien, Michael: Development of Spiral-Groove Self-Acting Face Seals , Final Report. ASA
CR-135303 , LYC 7741, Avco Lycoming Division , June 1977.
39. Burcham, R. E.; and Boynton, L. L.: Small High-Speed Self-Acting Shaft Seals for Liquid Rocket
Engines, Final Report. NASA CR-135167, RI/ RD 77-195, Rocketdyne Div., Rockwell International
Corp. , Ju!); 1977.
40. Ludwig, L. P. ; Strom, T. N.; Allen, G. P.; and Johnson, R. L.: Improving Performance of Face
Contact Seal in Liquid Sodium (400 to 1000°F) by Incorporation of Spiral-Groove Geometry. NASA
T 0-3942,1967.
41. Ludwig, L. P.; and Johnson, R. L.: Design Study of Shaft Face Seal with Self-Acting Lift
Augmentation. III - Mechanical Components. ASA TN 0-6164 , 1971.
42. Zuk, J.; Ludwig, L. P.; and Johnson, R. L.: Design Study of Shaft Face Seal with Self-Acting Lift
Augmentation. II - Sealing Dam. NASA TN 0-7006, 1970.
43. Zuk, J. ; and Ludwig, L. P.: Investigation of Isothennal, Compressible Flow Across a Rotating Sealing
Danl. I - Analysis . NASA TN 0-5344, 1969.
147
44. Johnson , R. L. ; Loomis, W. R. ; and Ludwig, L. P.: Performance and Analysis of Seals for Inerted
Lubrication Systems of Turbine Engines. ASA T D-4 761, 1968.
45. Gardner , J. F.: Combined Hydrostatic and Hydrodynamic Principles Applied to Non-Contacting Face
Seals. ASLE 4th Int'I Conf. Fluid Sealing (Philadelphia, PA) , May 5-8,1969 , ASLE SP-2 , 1969 , pp .
88-98.
46. Wolf, J. E.: Hydrostatic Seals for Turbopumps. Paper 66 -FE-18, ASME Fluids Engrg. Conf. (Denver,
CO), April 25-28, 1966.
47. Ludwig, 1. P. ; Zuk, J.; and Johnson, R. 1.: Use of the Computer in Design of Gas Turbine Mainshaft
Seals for Operation to 500 ft/sec (152 m /sec) . ASA TMX-52886 , 1970.
48. Ludwig, 1. P. ; Zuk, J.; and Johnson, R. 1.: Design Study of Shaft Face Seal with Self-Acting Lift
Augmentation. IV - Seal Force Balance. ASA T D-6568,1972.
49. Ritter, J. A.: Design and Experimental Evaluation of a Turbopump Lift-Off Seal. NASA CR-72796 ,
1971.
50. Maillare , F. R.; Nelson, 1. H. ; and Buckmann, P. S.: Evaluation of Wear Ring Seals for High-Speed ,
High-Pressure Turbopumps. J. Lubrication Tech. , Trans. ASME, Series F, vol. 91 , 1969 , pp. 438-450.
51. Anon.: Hydrostatic Bearing Feasibility Program - Final Report. AFRPL-TR-65 -120, AFRPL,
December 1965.
52. Chiang, T. ; and Cheng, H. S.: An Analysis of Flexible-Seal-Ring Vibrations. Trans. ASLE, vol. 11 , no.
3, July 1968 , pp. 204-215.
53. Winn , 1. W. ; Thorkildsen, R. 1. ; and Wilcock, D. F. : Design of One-Piece Jet-Engine Compressor End
Seals. Paper 68-Lubs-14, ASME Lubrication Symposium (Las Vegas, NY), June 17-20, 1968.
54 . Zabriskie, W. ; and Sternlicht, B. : Labyrinth-Seal Leakage Analysis. J. Basic Eng., Trans. ASME, Series
D, vol. 81,1959, pp. 332-336.
55. Vermes, G.: A Fluid Mechanics Approach to the Labyrinth Seal Leakage Problem . J. Eng. Power ,
Trans . ASME, Series A, vol. 83, 1961, pp. 161-169.
56 . Egli, A. : The Leakage of Steam Through Labyrinth Seals. Trans. ASME , vol. 57, 1935 , pp. 115-122 .
57. Martin , H. M.: Labyrinth Packings. Engineering, vol. 85, January 1908, pp. 35 -36.
58. Robinson , C. S.: Flow of a Compressible Fluid Through a Series of Identical Orifices. J. Appl. Mech.,
vol. 15 , no. 4 , December 1948 , pp. 308-310.
59. Kearton , W. J. ; and Keh , T. H.: Leakage of Air Through Labyrinth Glands of Staggered Type_ Proc.
Inst. Mech. Engrs., Series A, vol. 166, no. 2, 1952, pp. 180-188; discussion, pp . 189-195.
148 I
60. Kearton, W. J.: The Flow of Air Through Radial Labyrinth Glands . Proc. Inst. Mech. Engrs., vo!' 169,
no. 30, 1955 , pp . 539-550; discussion, pp. 551-552.
61. Jerie, J.: Flow Through Straight-Through Labyrinth Seals. Proc. Seventh Int'I Congo App!. Mech., vol.
2, Part I, 1948, pp. 70-82.
62. Trutnovsky , K.: Beruhrungsfreie Dichtungen. VOl-Verlag Gmb H (Dusseldorf), 1973.
63. Stepan off, A. J.: Centrifugal and Axial Flow Pumps. Reyjsed , John Wiley & Sons (New York), 1960.
64. Eckert , B. ; and Schnell, E.: Axial- und Radial-Kompressoren. Second ed., Springer-Verlag (Berlin),
1961.
65. Pfleiderer, C.: Die Kreiselpumpen fUr Flussigkeitem und Gase. Fourth ed., Springer-Verlag (Berlin),
1955.
66. Watkins, S. , Jr. : Honeycomb Rotor Seals. Prod. Eng., vol. 36 , no. 1, January 1965, pp. 61 -65.
67. Heymann , F. J. : Toward Quantitative Prediction of Liquid Impact Erosion. ASTM STP 474,1970.
68. Blair, R. W.: Controlled Gap Seal. Paper 687A, SAE Mtg. , Jan. 9-13,1956.
69. Trainer, T. M.: Final Report on the Development of Analytical Techniques for Bellows and
Diaphragm Design. AFRPL-TR-68-22 , AFRPL, March 1968.
70. Anderson , W. F.: Analysis of Stresses in Bellows: Part I - Design Criteria and Test Results. AEC
NAA-SR4527 , Atomics International Div. , Rockwell International , October 1964.
71. Winborne, R. A.: Simplified Formulas and Curves for Bellows Analysis. AEC Res. Rep .,
NAA-SR-9848, Atomics International Div., Rockwell International, August 1964.
72. Anon.: Packings ; Installation and Gland Design , Hydraulic , General Requirements For. Military
Specification MIL-P-5514, Dept. of Defense , 1963. I
73. Anon.: Parker O-Ring Handbook . Catalog OR 5700, Parker Seal Co. , January 1970.
74. Anon.: Engineer's Handbook of Piston Rings, Sealing Rings, and Mechanical Shaft Seals. Ninth Ed .,
Koppers Co., Inc. , 1975.
75. Denny, D. F.: Some Measurements of Fluid Pressures Between Plane Parallel Thrust Surfaces with
Special Reference to Radial Face Seals. Wear , vol. 4,1961, pp. 64-83.
76. Cheng, H. S.: On the Elastohydrodynamic Film and Pressure ProfIle of High Pressure Face Seals.
MTI-65 TR3 (N 161-25938), Mechanical Technology, Inc . (Latham, NY), February 1965.
77. Anon.: Flammability, Odor, and Offgassing Requirements and Test Procedures for Materials in
Environments That Support Combustion. NASA NHB 8060.1 A, February 1974.
149
_J
150
N ASA SPACE VEHICLE DESIGN (RITERIA
MONOGRAPHS ISSUED TO DATE
ENVIRONMENT
SP-800S Solar Electromagnetic Radiation , Revised May 1971
SP-801O Models of Mars Atmosphere (1974) , Revised December 1974
SP-8011 Models of Venus Atmosphere (1972) , Revised September 1972
SP-8013 Meteoroid Environment Model - I969 (Near Earth to Lunar Surface),

March 1969
SP-8017 Magnetic Fields- Earth and Extraterrestrial, March 1969
SP-8020 Surface Models of Mars (1975), Revised September 1975
SP-8021 Models of Earth's Atmosphere (90 to 2500 km), Revised March 1973
SP-S023 Lunar Surface Models, May 1969
SP-S037 Assessment and Control of Spacecraft Magnetic Fields , September 1970
SP-S03S Meteoroid Environment Model-1970 (Interplanetary and Planetary),

October 1970
SP-S049 The Earth's Ionosphere, March 1971
SP-S067 Earth Albedo and Emitted Radiation , July 1971
SP-S069 The Planet Jupiter (1970) , December 1971
SP-SOS4 Surface Atmospheric Extremes (Launch and Transportation Areas) ,

Revised June 1974
SP-S085 The Planet Mercury (1971) , March 1972
SP-S091 The Planet Saturn (1970), June 1972
SP-8092 Assessment and Control of Spacecraft Electromagnetic Interference,

June 1972
151
SP-8103 The Planets Uranus, eptune, and Pluto (1971) , November 1972
SP-8105 Spacecraft Thermal Control , May 1973
SP-81ll Assessment and Control of Electrostatic Charges , May 1974
SP-8116 The Earth's Trapped Radiation Belts, March 1975
SP-8Il7 Gravity Fields of the Solar System , April 1975
SP-8118 In terplanetary Charged Particle Models (1974) , March 1975
SP-8122 The Environment of Titan (1975) , Jul y 1976
STRUCTURES
SP-8001 Bu ffeting During Atmospheric Ascent , Revised ovember 1970
SP-8002 Flight-Loads Measurements During Launch and Exit , December 1964
SP-8003 Flutter, Buzz, and Divergence , Jul y 1964
SP-8004 Panel Flutter, Revised Ju ne 1972
SP-8006 Local Steady Aerodynamic Loads Duri ng Launch and Exit, May 1965
SP-8007 Buckling of Thin-Walled Circular Cylinders , Revised August 1968
SP-8008 Prelaunch Ground Wind Loads , ovember 1965
SP-8009 Propellant Slosh Loads , August 1968
SP-80l2 atural Vibration Modal Analysis , September 1968
SP-80l4 Entry Thermal Protect ion, August 1968
SP-80 19 Buckl ing of Thin-Walled Truncated Cones , September 1968
SP-8022 Staging Loads , February 1969
SP-8029 Aerodynamic and Rocket-Exhaust Heating During Launch and Ascent ,

May 1969
SP-8030 Transient Loads From Thrust Excitation, February 1969
SP-803l Slosh Suppression, May 1969
,
152
j
SP-8032 Buckling of Thin -Walled Doubly Curved Shells , August 1969
SP-8035 Wind Loads During Ascent , June 1970
SP-8040 Fracture Control of Metallic Pressure Vessels , May 1970
SP-8042 Meteoroid Damage Assessment , May 1970
SP-8043 Design-Development Testing , May 1970
SP-8044 Qualification Testing , May 1970
SP-8045 Acceptance Testing , April 1970
SP-8046 Landing Impact Attenuation for on-SurFace-Planing Landers , April

1970
SP-8050 Structural Vibration Prediction , June 1970
SP-8053 uclear and Space Radiation Effects on Materials , June 1970
SP-8054 Space Radiation Protection , June 1970
SP-8055 Prevention of Coupled Structure-Propulsion Instability (Pogo), October

1970
SP-8056 Flight Separation Mechanisms, October 1970
SP-8057 Structural Design Criteria Applicable to a Space Shuttle , Revised March

1972
SP-8060 Compartment Venting, ovember 1970
SP-8061 Interaction with Umbilicals and Launch Stand, August 1970
SP-8062 Entry Gasdynamic Heating , January 1971
SP-8063 Lubrication, Friction, and Wear , June 1971
SP-8066 Deployable Aerodynamic Deceleration Systems, June 1971
SP-8068 Buckling Strength of Structural Plates, June 1971
SP-8072 Acoustic Loads Generated by the Propulsion System, June 1971
SP-80n Transportation and Handling Loads , September 1971
153
SP-8079 Structural Interaction with Control Systems , November 1971
SP-80S2 Stress-Corrosion Cracking in Metals, August 1971
SP-8083 Discontinuity Stresses in Metallic Pressure Vessels , November 1971
SP-809S Preliminary Criteria for the Fracture Control of Space Shuttle

Structures, June 1971
SP-8099 Combining Ascent Loads, May 1972
SP-8104 Structural Interaction With Transportation and Handling Systems,

January 1973
SP-8108 Advanced Composite Structures, December 1974
GUIDANCE AND CONTROL
SP-801S Guidance and Navigation for Entry Vehicles , November 1968
SP-8016 Effects of Structural Flexibility on Spacecraft Control Systems, April

1969
SP-8018 Spacecraft Magnetic Torques , March 1969
SP-8024 Spacecraft Gravitational Torques , May 1969
SP-S026 Spacecraft Star Trackers, July 1970
SP-80n Spacecraft Radiation Torques , October 1969
SP-S028 Entry Vehicle Control, November 1969
SP-8033 Spacecraft Earth Horizon Sensors, December 1969
SP-8034 Spacecraft Mass Expulsion Torques , December 1969
SP-8036 Effects of Structural Flexibility on Launch Vehicle Control Systems,

February 1970
SP-8047 Spacecraft Sun Sensors, June 1970
SP-SOS8 Spacecraft Aerodynamic Torques , January 1971
SP-80S9 Spacecraft Attitude Control During Thrusting Maneuvers, February

1971
154
SP-8065 Tubular Spacecraft Booms (Extendibl e, Reel Stored) , February 1971
SP-8070 Spaceborne Digital Computer Systems , March 1971
SP-8071 Passive Gravity·Gradient Libration Dampers, February 1971
SP-8074 Spacecraft Solar Celi Arrays , May 1971
SP-8078 Spaceborne Electronic Imaging Systems, June 1971
SP·8086 Space Vehicle Displays Design Criteri a, March 1972
SP·8096 Space Vehicle Gyroscope Sensor Appli cations, October 1972
SP·8098 Effects of Stru ct ural Flexibility on Ent ry Vehicle Control Systems ,

June 1972
SP·8102 Space Vehi cle Acceleromete r Appli cations , December 1972
CHEM ICAL PROPULSIO
SP·8089 Liquid Rocket Engine Injectors, March 1976
SP·8087 Liquid Rocket Engine Fluid·Cooled Combustion Chambers , April 1972
SP·8124 Liquid Roch,t Engine Self-Coo led Combustion Chambers, September

1977
SP·8 113 Liquid Rocket Engin e Combustion Stabilizati on Devices, Novemb er

1974
SP·8120 Liquid Rocket Engine ozzles, July 1976
SP·8 107 Turb opump Systems for Liquid Rocket Engines , August 1974
Sp·8 109 Liquid Rocket Engine Centri fugal Flow Turb opumps, December 1973
SP·8125 Liquid Rocket Engine Axial Flow Turb opumps, April 1978
SP·8052 Liquid Rocket Engine Turbopump Indu ce rs, May 1971
SP·8 110 Liquid Rocket Engine Turbines , Janu ary 1974
SP·8081 Liquid Propellant Gas Generators , March 1972
SP·8048 Liquid Rocket Engine Turbopump Bearings , March 1971
155
l
SP-S I0l Liquid Rocket Engine Turb opump Shafts and Couplings_ Septemb er
1972
SP-S I00 Liquid Rocket Engin e Tu rbopump Gears , March 1974
SP-SOSS Liqu id Rocket Meta l Tanks and Tank Components , May 1974
SP-S094 Liquid Rocket Valve Components , August 197 3
SP-S097 Liqui d Rocket Valve Asse mblie _ ovember 1973
SP-S090 Liquid Rocket Actuators and Opera to rs. May 1973
SP-SI19 Liquid Rocke t Disconnec ts,Coup ling . Fittings. Fi xed Joints. and Seal,
September 1976
SP-S I23 Liquid Rocket Lin es. Bell ows. Fl ex ibl e Hoses. and Filt ers. April 1977
SP-S I1 2 Press uri za ti on Syste ms for Liquid Rockets. October 1975
SP-SOSO Liquid Rocket Pressu re Regu lato rs, Relief Valves. Check Val ves. Bu r
Disks, and Explosive Valves, March 1973
SP-S064 Solid Propellan t Selec tioil and Characteri za ti on , Jun e 197 I
SP-S07 5 Solid Propellan t Processing Fac tors in Rocke t Motor Design, October
1971
SP-S07 6 Solid Prope llant Grain Design an d Interna l Balli st ics, March 1972
SP-S073 Solid Propellant Grain St ructural In tegrit y Anal ysis. Jun e 1973
SP-S039 Solid Rocket Moto r Perfo rm ance An alysis and Pred icti on, May 1971
SP-80 5 1 Solid Rocket Motor Ign it ers , March 197 1
SP-8025 Solid Rocket Motor Metal Cases , April 1970
SP-8093 Solid Rocket Motor Internal Insulat ion, Decemb er 1976
SP-811 5 Solid Rocket Moto r ozzles , Jun e J975
SP-8 114 Solid Rocket Thru st Vecto r Control , December 1974
SP-8041 Captive-Fired Tes tin g of Solid Rocket Motors , March 1971
156 NASA-Langley, 1978

SP-8121 Rotating Shaft Seals

Uploaded by

Copyright:

Available Formats

SP-8121 Rotating Shaft Seals

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

SP-8121 Rotating Shaft Seals

Uploaded by

Copyright:

Available Formats

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

NASA NASA SP-8121

LIQUID ROCKET ENGINE TURBOPUMP

NAT ION AL AER ONA UTIC S AND SPA CE ADM INIS

F bru ary j 978

Th e d esign crite ri a monograph is not intended to be a design handbook , a set of

2. STATE OF THE ART 3

3. DES IGN CR ITER IA and Reco mm en ded Practices 103

APPE DIX A - Conversion of .S. Customary nits to SI Units 133

APPEND IX B - Glossary 135

REF ERE CES . . . . 145

NASA Space Vehicle Design Criteria Monogra ph s Is ued to Date 151

SUBJECT STAT E OF TH E ART DESIGN CRIT ERI A

SEAL SYSTEM 2. 1 7 3. 1 103

Press ure Environmcn t 2. 1.1 12 3. 1.1 103

SEAL ASSEMBLY 2.2 16 3.2 105

Prcssure Capabili ty :2.-. 1 16 3.2.1 105

Vibratio n Co ntrol 2.2. 7 50 3.2.7 112

Seal Mounting Requirements 2.2.9 52 3.2.9 113

SEAL COMPO ENTS 2.3 55 3.3 11 5

Sea l Materia ls 2.3.1 55 3.3.1 115

Circumferential-Seal Rubbin g Eleme nts 2.3.3 74 3.3.3 122

Hydrostatic/ Hydrodyna mic Face

Externally Pressurized Hydrostatic Seals 2.3.4.2 77 3.3.4.2 124

Ci rcum feren ti al-Sea l Clearance

Face-Contact-Seal Secondary Elements 2.3.6 89 3.3.6 128

Spring Load 2.3.7 98 3.3.7 131

Pressure Balance 2.3.8 100 3.3.8 132

Figure Titl e Page

3 Two-seal syste m for co nta inin g liquid hydroge n 8

4 Three-seal system fo r co ntainin g liquid hydroge n and (H2 + H2 0) h ot gas 9

5 Three -seal system fo r separatin g liquid oxygen and oil 9

6 Four-seal sys tem fo r se parating liquid oxyge n and RP-I 9

7 Four- ea l sys tem fo r se parating liquid oxyge n an d (H2 + H20) hot ga 10

8 Four-seal sys tem for se parating li quid oxyge n and

9 Four-seal sys tem for se parating liquid Ou orin e and

II Face-contac t metal-b ell ows seal

12 Face-contact pl as tic-lip seal 22

13 Face-co ntac t piston-rin g seal 23

14 Face -co ntact elastomer seal 23

15 Th ree configurat ions for a se lf-energized hydro ta tic face sea l 24

16 Two co nfigura ti ons fo r an ex terna ll y pressurized hyd rostatic face sea l 25

17 Four configura ti ons for a hydrodynamic face sea l 26

18 Two co nfi gurati ons for a hybrid face ea l . . . 27

19 Face co nt ac t/clea rance ea l 28

20 Circumferential shaft-ridin g segmented seal 28

21 Circumferential shaft-riding arch-bound segmented seal 29

Circumferen tial hydrodynamic segmen ted seal . 29

23 Circumferential floating-ring cont roUed-gap sea l 30

24 Circumferential clearance labyrinth seal 30

25 Circumferential wear-in labyrinth seal 30

26 Seal fluid pressures in representative turbopumps 3]

27 Values for seal rubbing peeds in representative turbopumps 34

28 Values for FV factor for face-contact sea ls in representative turbopumps 37

29 Values for PV factor for face-contact seals in representative turbopumps 37

30 Viscosity of oxygen as a function of temperature and pressure 41

31 Viscosity of para-hydrogen as a function of temperature and pressure 42

32 Flow regime in seal leakage 43

33 Tensile yield trength of typical alloys as a function of temperature 61

34 Elongation of typ ical alloys as a function of temperature . . . _ 62