Aerospace Chatbots
Collection
A collection of chatbots for aerospace using retrieval augmented generation (RAG) or fine tuning.
•
20 items
•
Updated
page_content
stringlengths 25
9.66k
| type
stringclasses 1
value | source
stringclasses 12
values | page
int64 1
596
|
|---|---|---|---|
NASNCP-2006-2 14290 3 sth Aerospace Mechanisms Symposium Compiled by Edward A. Boesiger Lockheed Martin Space Systems Company, Sunnyvale, California Proceedings of a symposium hosted by the NASA Langley Research Center and Lockheed Martin Space Systems Company and organized by the Mechanisms Education Association held at the Williamsburg Maniott Hotel Williamsburg, Virginia May 17- 19,2006 May 2006 | Document | AMS_2006.pdf | 1 |
The NASA STI Program Office . . . in Profile Since its founding, NASA has been dedicated to the advancement of aeronautics and space science. The NASA Scientific and Technical Information (STI) Program Office plays a key part in helping NASA maintain this important role. The NASA STI Program Office is operated by Langley Research Center, the lead center for NASA’s scientific and technical information. The NASA STI Program Office provides access to the NASA STI Database, the largest collection of aeronautical and space science STI in the world. The Program Office is also NASA’s institutional mechanism for disseminating the results of its research and development activities. These results are published by NASA in tlie NASA STI Report Series, which includes the following report types: 0 TECHNICAL PUBLICATION. Reports of completed research or a major significant phase of research that present the results of NASA programs and include extensive data or theoretical analysis. Includes compilations of Significant scientific and technical data and information deemed to be of continuing reference value. NASA counterpart of peer- reviewed formal professional papers, but having less stringent limitations on manuscript length and extent of graphic presentations. 0 TECHNICAL MEMORANDUM. Scientific and technical findings that are preliminary or of specialized interest, e.g., quick release reports, working papers, and bibliographies that contain minimal annotation. Does not contain extensive analysis. 0 CONTRACTOR REPORT. Scientific and technical findings by NASA-sponsored contractors and grantees. CONFERENCE PUBLICATION. Collected papers from scientific and technical conferences, symposia, seminars, or other meetings sponsored or co-sponsored by NASA. SPECIAL PUBLICATION. Scientific, technical, or historical information from NASA programs, projects, and missions, often concerned with subjects having substantial public interest. TECHNICAL TRANSLATION. English- language translations of foreign scientific and technical material pertinent to NASA’s mission. Specialized services that complement the STI Program Office’s diverse offerings iiiclude creating custom thesauri, building customized databases, organizing and publishing research results .. . even providing videos. For more information about the NASA STI Program Office, see the following: Access the NASA STI Program Home Page at http://www.sti. nasa.gov E-mail your question via tlie Internet to help@sti.nasa.gov Fax your question to the NASA STI Help Desk at (301) 621-0134 Phone the NASA STI Help Desk at (301) 621-0390 Write to: .1 NASA STI Help Desk NASA Center for Aerospace Information 7 12 1 Standard Drive Hanover, MD 2 1076- 1320 | Document | AMS_2006.pdf | 2 |
NASA/CP-2006-2 14290 3 8th Aerospace Mechanisms Symposium Compiled by Edward A. Boesiger Lockheed Martin Space Systems Company, Sunnyvale, California Proceedings of a symposium hosted by the NASA Langley Research Center and Lockheed Martin Space Systems Company and organized by the Mechanisms Education Association held at the Williamsburg Marriott Hotel Williamsburg, Virginia May 17- 19,2006 National Aeronautics and Space Administration Langley Research Center Hampton, Virginia 2368 1-2 199 | Document | AMS_2006.pdf | 3 |
The use of trademarks or names of manufacturers in the report is for accurate reporting and does not constitute an official endorsement, either expressed or implied, of such products or manufacturers by the National Aeronautics and Space Administration. Available from: NASA Center for Aerospace Information (CASI) 7 12 1 Standard Drive Hanover, MD 2 1076- 1320 (301) 621-0390 National Technical Information Service (NTIS) 5285 Port Royal Road Springfield, VA 22 16 1-2 17 1 (703) 605-6000 | Document | AMS_2006.pdf | 4 |
PREFACE The Aerospace Mechanisms Symposium (AMS) provides a unique forum for those active in the design, production and use of aerospace mechanisms. A major focus is the reporting of problems and solutions associated with the development and flight certification of new mechanisms. Organized by the Mechanisms Education Association, the National Aeronautics and Space Administration and Lockheed Martin Space Systems Company (LMSSC) share the responsibility for hosting the AMs. Now in its 38fh symposium, the AMS continues to be well attended, attracting participants from both the US. and abroad. The 38fh AMs, hosted by the Langley Research Center (LaRC) in Williamsburg, Virginia, was held May 17, 18 and 19, 2006. During these three days, 34 papers were presented. Topics included gimbals, tribology, actuators, aircraft mechanisms, deployment mechanisms, release mechanisms, and test equipment. Hardware displays during the supplier exhibit gave attendees an opportunity to meet with developers of current and future mechanism components. The high quality of this symposium is a result of the work of many people, and their efforts are gratefully acknowledged. This extends to the voluntary members of the symposium organizing committee representing the eight NASA field centers, LMSSC, and the European Space Agency. Appreciation is also extended to the session chairs, the authors, and particularly the personnel at LaRC responsible for the symposium arrangements and the publication of these proceedings. A sincere thank you also goes to the symposium executive committee who is responsible for the year-to-year management of the AMs, including paper processing and preparation of the program. The use of trade names of manufacturers in this publication does not constitute an official endorsement of such products or manufacturers, either expressed or implied, by the National Aeronautics and Space Administration. ... 111 | Document | AMS_2006.pdf | 5 |
CONTENTS ... Symposium Schedule ............................................................................................................................... vi11 Symposium Organizing and Advisory Committees ................................................................................... xii Precision Linear Actuators for the Spherical Primary Optical Telescope Demonstration Mirror ................ 1 Jason Budinoff & David Pfenning The CRISM MotodEncoder Assembly and Diaphragm Bearing Assembly Design ................................. 11 Jeffrey Lees & Ed Schaefer Gear Teeth Particles and Bearing Failures .............................................................................................. 25 William Greenwood & Jeffrey Dabling Failure of Harmonic Gears During Verification of a Two-Axis Gimbal for the Mars Reconnaissance Orbiter Spacecraft ....................................................................................................... .37 Michael Johnson, Russ Gehling & Ray Head Stacer Driven Deployment: The Stereo Impact Boom ............................................................................. 51 Robert Ullrich, Jeremy McCauley, Paul Turin, Ken McKee & Bill Donokowski Heritage Adoption Lessons Learned: Cover Deployment and Latch Mechanism .................................... 65 James Wincentsen Problems and Product Improvements in a Qualified, Flight Heritage Product ......................................... 75 Chuck Lazansky & Scott Christiansen SoftRide Vibration and Shock Isolation Systems that Protect Spacecraft from Launch Dynamic Environments ................................................................................................................ 89 Conor Johnson, Paul Wilke & Scott Pendleton Summary of the New AlAA Moving Mechanical Assemblies Standard .................................................. 103 Brian Gore Lessons Learned From the Development, Operation, and Review of Mechanical Systems on the Space Shuttle, International Space Station, and Payloads .......................................... 11 3 Alison Dinsel, Wayne Jermstad & Brandan Robertson Reliability and Fault Tolerance in ISS Thermofoil Spaceflight Heaters .................................................. 127 Victor Bolton Development, Pre-qualification and Application of an Active Bearing Preload System ........................ .133 Simon Lewis & Martin Humphries Development of a Dual Mode D-Strut@ Vibration Isolator for a Laser Com'munication Terminal ........... 141 Dale Ruebsamen, James Boyd, Joe Vecera & Roger Nagel Design and Testing of a Low Shock Discrete Point Spacecraft Separation System .............................. 149 Pete Woll & Daryn Oxe V | Document | AMS_2006.pdf | 7 |
Faying Surface Lubrication Effects on Nut Factors ................................................................................ 157 Deneen Taylor & Raymond Morrison Torque Loss and Stress Relaxation in Constant Torque Springs .......................................................... 163 Robert Postma Mechanical Design of a Multi-Axis Martian Seismometer ...................................................................... 169 Franck Pecal, Nicolas Paulin, Doug Mimoun & Gabriel Pont Commercial off-the-shelf Components in Reaction Wheels .................................................................. 187 Andrew Haslehurst Design of the ATMS Scan Drive Mechanism ......................................................................................... 197 Curtis Allmon & Dave Putnam Lessons Learned From the Windsat BAPTA Design and On-Orbit Anomalies ..................................... 209 Steve Koss & Scott Woolaway JWST NlRSpec Cryogenic Light Shield Mechanism .............................................................................. 223 Kathleen Hale & Rajeev Sharma Development Tests of a Cryogenic Filter Wheel Assembly for the NlRCam Instrument ....................... 229 Sean McCully, Charles Clark, Michael Schermerhorn, Filip Trojanek, Mark O’Hara, Jeff Williams & John Thatcher Cryogenic Nano-Actuator for JWST ....................................................................................................... 239 Robert Warden Space Shuttle Body Flap Actuator Bearing Testing for NASA Return to Flight ...................................... 253 Tim Jett, Roamer Predmore, Michael Dube & William Jones, Jr Bearing Development for a Rocket Engine Gimbal ................................................................................ 269 Christian Neugebauer, Manfred Falkner, Ludwig Supper & Gerhard Traxler Effect of Test Environment on Lifetime of Two Vacuum Lubricants Determined by Spiral Orbit Tribometry ........................................................................................................................... 283 Stephen Pepper Influence of Oil Lubrication on Spacecraft Bearing Thermal Conductance ........................................... 291 Yoshimi Takeuchi, Matthew Eby, Benjamin Blake, Steven Demsky & James Dickey Mars Exploration Rover Potentiometer Problems, Failures and Corrective Actions .............................. 303 Mark Balzer Mechanism Development, Testing, and Lessons Learned for the Advanced Resistive Exercise Device ..................................................................................................... 31 7 . Christopher Lamoreaux & Mark Landeck Radarsat Range Adjustment Mechanism Design .................................................................................. 331 Xilin Zhang & Sylvain Riendeau vi | Document | AMS_2006.pdf | 8 |
Come-Along Tool Development for Telerobotic In-Space Servicing of the Hubble Space Telescope ....................................................................................................................... 345 Jonathan Penn Planetary Airplane Extraction System Development and Subscale Testing .......................................... 359 John Teter Jr "Digital" SMA-Based Trailing Edge Tab Actuators for Aerospace Applications ..................................... 373 Robert McKillip Jr Development of a Forced Oscillation System for Measuring Dynamic Derivatives of Fluidic Vehicles .................................................................................................................................. 387 Bo Trieu, T. Tyler, 6. Stewart, J. Charnock, D. Fisher, E. H. Heim, J. Brandon & S. Grafton vii | Document | AMS_2006.pdf | 9 |
SYMPOSIUM SCHEDULE WEDNESDAY, 17 MAY 2006 8:OO 8:OO 9:oo 9:30 1 1 :45 1 :oo Wednesday Presenters' Breakfast - Terrace Room CHECK-IN AND REFRESHMENTS - Auditorium INTRODUCTORY REMARKS - Auditorium James Wells, Host Chairman NASA Langley Research Center, Hampton, VA Stuart Loewenthal, General Chairman Lockheed Martin Space Systems, Sunnyvale, CA CENTER WELCOME Stephen G. Jurczyk, Deputy Director NASA Langley Research Center, Hampton, VA SESSION I - ACTUATORS Stephen Sandford, Session Chair NASA Langley Research Center, Hampton, VA * Precision Linear Actuators for the Spherical Primary Optical Telescope Demonstration Mirror Jason Budinoff & David Pfenning, NASA Goddard Space Flight Center, Greenbelt, MD The CRISM Motor/Encoder Assembly and Diaphragm Bearing Assembly Design Jeffrey Lees & Ed Schaefer, Johns Hopkins University Applied Physics Laboratory, Laurel, MD - Gear Teeth Particles and Bearing Failures William Greenwood and Jeffrey Dabling, Sandia National Laboratories, Albuquerque, NM - Failure of Harmonic Gears During Verification of a Two-Axis Gimbal for the Mars Reconnaissance Orbiter Spacecraft Michael Johnson, Jet Propulsion Laboratory, Pasadena, CA; Russ Gehling & Ray Head, Lockheed Martin Space Systems, Denver, CO LUNCH BREAK - Box lunch is provided SESSION II - MECHANISMS Ted Hartka, Session Chair Johns Hopkins University Applied Physics Laboratory, Laurel, MD Robert Ullrich, Jeremy McCauley, Paul Turin, Ken McKee & Bill Donokowski, Space Sciences Lab, University of California, Berkeley, CA e Heritage Adoption Lessons Learned: Cover Deployment and Latch Mechanism James Wincentsen, Jet Propulsion Laboratory, Pasadena, CA - Problems and Product Improvements in a Qualified, Flight Heritage Product Chuck Lazansky & Scott Christiansen, Starsys Research Corp., Boulder, CO Stacer Driven Deployment: The Stereo Impact Boom 2:30 BREAK ... Vlll | Document | AMS_2006.pdf | 10 |
2~45 SESSION 111 - “BIG PICTURE” John McManamen, Session Chair NASA Johnson Space Center, Houston, TX SoftRide Vibration and Shock Isolation Systems that Protect Spacecraft from Launch Dynamic Environments Conor Johnson, Paul Wilke & Scott Pendleton, CSA Engineering, Inc., Mountain View, CA Summary of the New AIAA Moving Mechanical Assemblies Standard Brian Gore, The Aerospace Corporation, El Segundo, CA * Lessons Learned From the Development, Operation, and Review of Mechanical Systems on the Space Shuttle, International Space Station, and Payloads Alison Dinsel, Wayne Jermstad & Brandan Robertson, NASA Johnson Space Center, Houston, TX 411 5 SESSION IV - POSTER PREVIEW Michael Johnson, Session Chair Jet Propulsion Laboratory, Pasadena, CA * Reliability and Fault Tolerance in ISS Thermofoil Spaceflight Heaters Victor Bolton, The Boeing Company, Houston, TX * Development, Pre-qualification and Application of an Active Bearing Preload System . Simon Lewis, European Space Tribology Laboratory, Warrington, Cheshire, U.K.; Martin Humphries, Sula Systems Ltd., Wotton-under-Edge, Gloucestershire, U.K. * Development of a Dual Mode D-Strut@ Vibration Isolator for a Laser Communication’Terminal Dale Ruebsamen, James Boyd, Joe Vecera & Roger Nagel, Honeywell Defense and Space, Glendale, AZ Design and Testing of a Low Shock Discrete Point Spacecraft Separation System Pete Woll, NEA Electronics, Chatsworth, CA; Daryn Oxe, Lockheed Martin Space Systems Company, Sunnyvale, CA * Faying Surface Lubrication Effects on Nut Factors Deneen Taylor, NASA Johnson Space Center, Houston, TX; Raymond Morrison, The Boeing Company, Huntington Beach, CA * Torque Loss and Stress Relaxation in Constant Torque Springs Robert Postma, The Aerospace Corporation, El Segundo, CA 6:30-9130 RECEPTION & DISPLAYS - ADAMS BALLROOM/PROMENADE OF THE MARRIOTT Invited component and software suppliers display current products and provide tutorials. THURSDAY, 18 MAY 2006 7:15 Thursday Presenters‘ Breakfast - Terrace Room 7:45 Light Refreshments - Auditorium 8:15 SESSION V - GIMBALS - Auditorium William Jones Jr., Session Chair Sest, Inc., Middleburg Heights, OH * Mechanical Design of a Multi-Axis Martian Seismometer Franck Pecal & Nicolas Paulin, EADS SODERN, Limeil Brevannes, France; Doug Mimoun, IPGP, Saint Maur, France; Gabriel Pont, CNES, Toulouse, France * Commercial off-the-shelf Components in Reaction Wheels Andrew Haslehurst, Surrey Satellite Technology Ltd, Guiford, Surrey, U.K. Design of the ATMS Scan Drive Mechanism Curtis AIlmon & Dave Putnam, Lockheed Martin Space Systems Company, Sunnyvale, CA ix | Document | AMS_2006.pdf | 11 |
10:15 10:30 12:oo 1 :oo 3:30 3:45 0 Lessons Learned From the Windsat BAPTA Design and On-Orbit Anomalies Steve Koss, Naval Research Laboratory, Washington, D.C.; Scott Woolaway, Ball Aerospace & Technologies Corp., Boulder, CO BREAK SESSION VI - JWST Casey DeKramer, Session Chair Swales Aerospace, Beltsville, MD JWST NlRSpec Cryogenic Light Shield Mechanism Kathleen Hale & Rajeev Sharma, NASA Goddard Space Flight Center, Greenbelt, MD 0 Development Tests of a Cryogenic Filter Wheel Assembly for the NlRCam Instrument Sean McCully, Charles Clark, Michael Schermerhorn, Filip Trojanek, Mark O'Hara, Jeff Williams & John Thatcher, Lockheed Martin Space Systems Company, Palo Alto, CA 0 Cryogenic Nano-Actuator for JWST Robert Warden, Ball Aerospace & Technologies Corp., Boulder, CO LUNCH BREAK - Box lunch is provided SESSION VI1 - BEARINGS & POTS Din0 Christopoulos, Session Chair Raytheon Space & Airborne Systems, El Segundo, CA Space Shuttle Body Flap Actuator Bearing Testing for NASA Return to Flight Tim Jett, NASA Marshall Space Flight Center, Huntsville, AL; Roamer Predmore, Swales Aerospace, Beltsville, MD; Michael Dube, NASA' Godciard Space Flight Center, Greenbelt, MD; William Jones, Jr, Sest, Inc., Middleburg Heights, OH Bearing Development for a Rocket Engine Gimbal Christian Neugebauer, Manfred Falkner, Ludwig Supper & Gerhard Traxler, Austrian Aerospace GmbH, Vienna, Austria Effect of Test Environment on Lifetime of Two Vacuum Lubricants Determined by Spiral Orbit Tribometry Stephen Pepper, NASA Glenn Research Center, Cleveland, OH Influence of Oil Lubrication on Spacecraft Bearing Thermal Conductance Yoshimi Takeuchi, Matthew Eby, Benjamin Blake, Steven Demsky & James Dickey, The Aerospace Corporation, El Segundo, CA * Mars Exploration Rover Potentiometer Problems, Failures and Corrective Actions Mark Baker, Jet Propulsion Laboratory, Pasadena, CA BREAK SPECIAL PRESENTATION - NASA Engineering & Safety Center, Mechanical Systems Super Problem Resolution Team John McManamen NASA Johnson Space Center, Houston, TX The NESC is an organization established in the wake of Space Shuttle Columbia accident to provide an independent look at high risk, complex technically issues. Presentation will include problems the SPRT has been engaged in and focuses in more detail on an ongoing investigation regarding the Space Shuttle Solid Rocket Booster Holddown Post stud hang up problem and the completed assessment of the Orbiter Rudder Speedbrake gear micro-pitting problem. 530-1 0130 SYMPOSIUM BANQUET AT JAMESTOWN SETTLEMENT 5:30 1O:OO Bus leaves Marriott to Jamestown Bus leaves Jamestown for hotel X | Document | AMS_2006.pdf | 12 |
FRIDAY, 19 MAY 2006 7:15 7:45 Light Refreshments - Auditorium Friday Presenters’ Breakfast - Terrace Room 8:15 SESSION Vlll - TOOLS & AIRCRAFT - Auditorium Gerard Migliorero, Session Chair ESNESTeC, Noordwijk, The Netherlands Mechanism Development, Testing, and Lessons Learned for the Advanced Resistive Exercise Device Christopher Lamoreaux & Mark Landeck, NASA Johnson Space Center, Houston, TX Radarsat Range Adjustment Mechanism Design Xilin Zhang &. Sylvain Riendeau, MDA Space, Inc., Ste-Anne-De-Bellevue, Canada * Come-Along Tool Development for Telerobotic In-Space Servicing of the Hubble Space Telescope Jonathan Penn, Swales Aerospace, Beltsville, MD Planetary Airplane Extraction System Development and Subscale Testing John Teter Jr., NASA Langley Research Center, Hampton, VA “Digital” SMA-Based Trailing Edge Tab Actuators for Aerospace Applications Robert McKillip Jr., Continuum Dynamics, Inc., Ewing, NJ * Development of a Forced Oscillation System for Measuring Dynamic Derivatives of Fluidic Vehicles Bo Trieu, T. Tyler , B. Stewart, J. Charnock, D. Fisher, E. H. Heim & J. Brandon, NASA Langley Research Center, Hampton, VA; S. Grafton, Vigyan, Inc., Hampton, VA 11 :15 PRESENTATION: An Overview of LaRC 11 :45 TECHNICAL SESSIONS CONCLUSION PRESENTATION OF THE HERZL AWARD CLOSING REMARKS Edward Boesiger, Operations Chairman Lockheed Martin Space Systems Company, Sunnyvale, CA BUSES DEPART HOTEL FOR LaRC TOUR 1130 2:00-4:00 NASA LANGLEY RESEARCH CENTER FACILITY TOUR xi | Document | AMS_2006.pdf | 13 |
SYMPOSIUM ORGANIZING COMMITTEE James E. Wells, Host Chair, NASA LaRC Robin Tutterow, Host Co-Chair, NASA LaRC Stuart H. Loewenthal, General Chairman, Lockheed Martin Edward A. Boesiger, Operations Chairman, Lockheed Martin Carlton L. Foster, NASA MSFC Claef F. Hakun, NASA GSFC Christopher P. Hansen, NASA JSC Wayne Jermstad, NASA JSC Patrice Kerhousse, ESNESTeC Alan C. Littlefield, NASA KSC Edward C. Litty, JPL Fred G. Martwick, NASA ARC Donald H. McQueen, Jr., NASA MSFC Wilfred0 Morales, NASA GRC Robert P. Mueller, NASA KSC Fred B. Oswald, NASA GRC Minh Phan, NASA GSFC Donald R. Sevilla, JPL Mark F. Turner, NASA ARC SYMPOSIUM ADVISORY COMMITTEE Obie H. Bradley, Jr., NASA LaRC (ret) Robert L. Fusaro, NASA GRC (ret) Ronald E. Mancini, NASA ARC (ret) Stewart C. Meyers, NASA GSFC (ret) William C. Schneider, NASA JSC (ret) xii | Document | AMS_2006.pdf | 14 |
Precision Linear Actuators for the Spherical Primary Optical Telescope Demonstration Mirror Jason Budinoff* and David Pfenning* Abstract The Spherical Primary Optical Telescope (SPOT) is an ongoing research effort at Goddard Space Flight Center developing wavefront sensing and control architectures for future space telescopes. The 03.5-m SPOT telescope primary mirror is comprise9 of six 0.86-m hexagonal mirror segments arranged in a single ring, with the central segment missing . The mirror segments are designed for laboratory use and are not lightweighted to reduce cost. Each primary mirror segment is actuated and has tip, tilt, and piston rigid-body motions. Additionally, the radius of curvature of each mirror segment may be varied mechanically. To provide these degrees of freedom, the SPOT mirror segment assembly requires linear actuators capable of <lO-nm resolution over a total stroke of 5 mm. These actuators must withstand high static loads as they must support the mirror segment, which has a mass of -100 kg. A stepper motor driving a differential satellite roller screw was designed to meet these demanding requirements. Initial testing showed that the actuator is capable of sub-micron repeatability over the entire 6-mm range, and was limited by 100-200 nm measurement noise levels present in the facility. Further testing must be accomplished in an isolated facility with a measurement noise floor of <5 nm. Such a facility should be ready for use at GSFC in the early summer of 2006, and will be used to better characterize this actuator. Introduction Future large (>6 m) space telescopes such as the James Webb Space Telescope, SAFIR, and beyond require segmented primary mirrors to package into launch vehicle payload fairings of diameters less than their apertures. Requisite architectures to align or “phase” the individual segments into a single optical surface after launch and deployment are required. Current techniqyes used on large ground-based telescop:s such as Keck include precision segment edge sensors and various types of wavefront sensors . However, phasing a large number of segments requires a significant amount of computing resources which can reduce observing efficiency. Maintaining a “phased” array of mirror segments in a challenging environment such as low-earth or L2 orbits remains to be seen. The SPOT research testbed will explore a new method of phasing segmented mirrors. The SPOT telescope architecture has possible application to a robotically assembled telescope for ISS, as shown in Figure 1. 2t- F Figure 1. A possible application of the SPOT telescope architecture: a telescope mounted on the “top” (zenith) end of the 21 truss Wavefront Sensina and Control A relatively recent technique utilizing image-based wavefront sensing has been pursued by the GSFC optics branch, Code 551. By placing a point source and camera at the center of curvature of a spherical * NASA Goddard Space Flight Center, Greenbelt, MD Proceedings of the 3gh Aerospace Mechanisms Symposium, langley Research Center, May 7 7-79,2006 1 | Document | AMS_2006.pdf | 15 |
mirror, a direct measurement of its surface wavefront error is possible. Taking various defocused images at the center of curvature and using an iterative transform solver, the phase error of the reflected wavefront can be determined3. From the phase error tip, tilt, defocus, and other Zernike terms (currently truncated to the first 15 terms) can be recovered. This information is used to position the mirror segments. To further develop this approach to phasing mirror segments, the SPOT internal research & development project was started in 2004. u, B - P or, - 0" E$ c) - 0 (1) Fuewue Caineia hlteiface 8 / Souice at ROC (3) WFS Alpritluils (DSP) _---_---___ DSP Processor Figure 2. SPOT Testbed Schematic Nanometer-level positioning of -1 00-kg mirror segments was required, as well as a high-load nanometer displacement actuator to mechanically bend the mirror segments to adjust their radius of curvature. As the program had limited funding, low-cost actuators were designed to meet these requirements. SPOT Background The Spherical Primary Optical Telescope (SPOT) is a GSFC internal research & development program initiated in 2003. The goal of the SPOT effort is to develop a robust architecture which will reduce the cost of large-aperture, segmented primary mirror space telescopes. The SPOT telescope architecture is based upon two key technology developments: 1 ) a high-rate, center of curvature, iterative transform phase- diversity phasing algorithm, and 2) a low-cost mirror segment. The SPOT demonstration telescope is a 03.5-m segmented spherical primary. The primary consists of 6 identical hexagonal segments measuring 876 mm point-to-point, in a 1-ring configuration, without a central segment. However, only 2 segments are being fabricated for this effort. Two segments are the minimum amount required to successfully demonstrate the phasing architecture. Each segment has rigid-body position control in tip, tilt, and piston. Each segment also has mechanical radius-of-curvature control. Some of the relevant requirements for the SPOT mirror segments are given in Table 1. 2 | Document | AMS_2006.pdf | 16 |
Table 1. Pyrexm Mirror Segment Requirements Tipmilt Range Tipmilt Resolution Tipmilt update rate Position Hold Actuator Thermal Stability Static Load Mirror Requirement I Value I Units I Note Size I 0876 I mm (inch) 1 Point-to-point hex f 2.0 degree 0.05 arcsecond As allowed by focus resolution 1 HZ 0 Amp Power-off hold 35 kg Mirror mass -1 00 kg, assume 3 Micronddeg As small as practicable I Focus update rate I I I I I I actuators I To provide rigid-body positioning in tip, tilt, and defocus of the segments, which will weigh -50 kg each, a tripod mechanism with custom actuators was designed at GSFC. Design of the Segment Assembly Tripod The mirror segment must have 3 rigid-body degrees of freedom: tip and tilt rotations and piston, a vertical translation. A segment assembly is shown in Figure 3. Figure 3. The SPOT Mirror Segment Assembly Tripod | Document | AMS_2006.pdf | 17 |
Kinematics of the Seqment Assemblv & Grubler’s Mobilitv Criterion The mirror requires only tip tilt and piston adjustment, 3 degrees of freedom. A hexapod would provide 6 degrees of freedom; but we don’t need 6. Therefore, a tripod was selected to provide the rigid body motions required. The end joints of each strut must constrain a number of degrees of freedom. For example, a ball-in-socket joint constrains 3 translations but is free to rotate, allowing 3 rotations. From kinematics, Grubler’s mobility criterion states that F, the number of degrees of freedom in a system, can be defined by: i= 1 Where h = 6, the degrees of freedom in the space the mechanism will be operating in n = the number of links in the system j =the number of joints in the system f, =the degrees of freedom allowed (unconstrained) at the th joint For the SPOT tripod, each leg consists of 2 links and 3 joints. Ground is considered a rigid link, and the mirror is the “end effector” or output link. For this system, the following values are used: h = 6, we shall consider the system exists in 6 degrees of freedom n = 8, ground and the mirror are each one link, and each leg has 2 links j = 9, each leg has 3 joints (base, linear, and upper) x 3 legs f, = 2, base joint of leg 1, XY flexure allowing 2 rotations f2 = 2, base joint of leg 2, XY flexure allowing 2 rotations f3 = 2, base joint of leg 3, XY flexure allowing 2 rotations f4 = 1, linear joint of leg 1, allowing 1 translation f5 = 1, linear joint of leg 2, allowing 1 translation f6 = 1, linear joint of leg 3, allowing 1 translation f7 = 2, upper joint of leg 1, XY flexure allowing 2 rotations fa = 2, upper joint of leg 2, XY flexure allowing 2 rotations f9 = 2, upper joint of leg 3, XY flexure allowing 2 rotations Using the above values, the system degrees of freedom are calculated as: i=l The number of system degrees of freedom is 3, corresponding to tip, tilt and piston. The use of XY flexures, allowing f, (for I = 1 ..3, 7..9) = 2, is justified. Ball joints could be used, but the additional passive degree of freedom at each leg (roll) would have to be subtracted out of the Grubler criterion equation to keep F = 3. The XY flexures will allow hysteresis-free angular motion at the cost of increased force proportional to displacement. Actuator Design The actuator is shown in Figure 4. 4 | Document | AMS_2006.pdf | 18 |
Figure 4. The SPOT linear actuator, shown with a six-inch ruler for scale. Figure 5. An exploded view of the SPOT linear actuator Stetmer Motor/Harmonic Drive Gearhead The HD14 1OO:l harmonic Drive gearhead Phytron 500 step per revolution or 0.72 degree step size 3-phase stepper motor, 2.5 Amp winding Agilent HEDL 5540 500 line (A quad B = 2000 counts) incremental encoder on motor output ZSS 52.500.2.5.Kl -HEDL-HDl4/100 5 | Document | AMS_2006.pdf | 19 |
Figure 6. The Phytron ZSS 52 / HD14 actuator Differential Satellite Roller Screw Several options exist for rotary-to-linear motion: lead screw, ball screw, or roller screw. Generally the most precise of these is the satellite roller screw. A differential roller screw was selected. After several months of vendor interaction, a differential roller screw was selected and sized. The smallest, readily available precision roller screw has a pitch of 0.5 mm. A lead of this size can produce 10 nm steps using the 0.72" stepper motor and 1OO:l harmonic drive. Using a differential roller screw, the effective lead can be reduced by 2 orders of magnitude, but at the cost of a stroke limitation to -6 mm. The theoretical attainable step size drops to 0.4 nm (see below). A RollvisTM differential satellite roller screw utilizes equal thread pitch on the nut and the shaft, but varies the nut/shaft thread pitch diameters and the number of starts on the nut and shaft. The effective lead of such a differential satellite roller screw can be calculated by: Where D, = Nut thread pitch diameter, mm D, = Shaft thread pitch diameter, mm P =thread pitch in threaddmm N, = number of thread starts on nut N, = number of thread starts on shaft, negative for opposite handedness to nut starts For the SPOT actuator, the values were varied within reasonable limits until a minimum value for effective lead was found. Using the following values: D, = 29 mm Ds=19mm P = 0.5 threadslmrn Nn = 4 Ns = -6 | Document | AMS_2006.pdf | 20 |
The minimum effective lead, Lff, was found to be 0.02 mm per revolution, or -21 microns per revolution. A bind condition determines the total stroke, which for this differential roller screw is -6 mm. Rollvis Swiss S.A., a Swiss manufacturer of precision roller screws, fabricated the roller screw as Model RV160/19,02.R1.604350, custom designed for maximum resolution. The 5 rollers roll around the shaft and are held in a rotating retainer ring at each end of the nut. A sun gear at each end of the nut and mating roller gears at the ends of each roller keep them in proper clocking as they rotate around the shaft. The shaft, nut and rollers are 410 stainless steel. The lubricant for the roller screw is lsoflex Topas NCA 52, manufactured by Kluber Lubrication. It is a synthetic oil with a calcium thickener. The differential satellite roller screw is shown in Figure 7 and 8. Figure 7. Custom Differential Satellite Roller Screw Figure 8. Nut end details showing the timing gear teeth Helical Couding A standard flexible shaft coupling from Helical Products Company, Inc. was used to couple the motor output shaft to the satellite roller screw shaft. Such couplers allow torque to transmitted despite small axial misalignments between the shafts. A model HRM-125-12mm-12mm coupling was used. This coupler uses 2 pairs of cup-point set screws to secure the motor and screw shafts. The coupling is 17-4PH H900 stainless steel. The coupling is shown in Figure 9. 7 | Document | AMS_2006.pdf | 21 |
Figure 9. Helical Coupling Bearinas Barden 1 O6HCDUL back to back duDlex oair. ABEC-9. SAE52100 steel 30-mm bore diameter, 15-degree contaci angle, 14 7.1 4-mm (9/32") diameter 440C balls, 36-kg (80-lb) heavy preload Static load capacity 1005 kg (221 6 Ib) Machined p ca Winsorlube oil Figure 10. Barden 106HC Duplex bearing pair Lea End Flexures XY flexures Crossed flexure Torsional rate Fat iq u e/c ycl e I if e 8 | Document | AMS_2006.pdf | 22 |
Where K = torsional spring rate r = radius of notch cut b =thickness of flexure section t = width of notch E = Modulus For the SPOT tripod, each leg consists of 2 links and 3 joints. Ground is considered a rigid link, and the mirror is the “end effector” or output link. Conclusion Further testing to fully characterize nanometric step size, repeatibilty, and linearity must be accomplished in a quiet facility. The actuator performance will also be measured with an actuator built into a mirror segment. Actuator positioning performance will be indirectly measured by mirror radius of curvature change per commanded step. References 1. 2. 3. 4. 5. 6. 7. 8. 9. Budinoff, Jason G. “SPOT Mirror Segment Assembly Requirements - Revision C, March 2004 Howard, Joseph ”Optical Design Study for NASA’s Spherical Primary Optical Telescope” SPIE Budinoff, Jason G, Michels, Gregory J. “Design & Optimization of the Spherical Primary Optical Telescope (SPOT) Primary Mirror Segment” SPIE 5877-42 Dean, B, Smith, S, Budinoff, J. “Image-Based Wavefront Sensing for the Control of Space Optics” USAF AMOS Maoi Technical Conference, September, 2000 Chanan, G., Troy, M., Ohara, C. “Phasing the Primary Mirror Segments of the Keck Telescopes: A Comparison of Different Techniques” Proc. SPIE, 4003, 188-201, 2000 Martinez, L.M., Yaitskova, N., Dierickx, P., Dohlen, K. “Mach Zender Wavefront Sensor for Phasing Segmented Telescopes” Tsai, Lung-Wen, ‘Robot Analysis: The Mechanics of Serial and Parallel Manipulators” 01 999 John Wiley & Sons, Inc. Paros, J.M., Weisbord, I. “How to Design Flexure Hinges” Machine Design vol 37, ppl51-156, 1965 Lobontiu, Nicolae “Compliant Mechanisms: Design of Flexure Hinges” 02002 CRC Press LLC 5524-1 9 | Document | AMS_2006.pdf | 23 |
The CRlSM Motor/Encoder Assembly and Diaphragm Bearing Assembly Design Jeffrey Lees" and Ed Schaefe; Abstract This paper will describe the thin section angular contact bearings and WS2 dry film lubrication used on the compact Reconnaissance Imaging Spectrometer for Mars (CRISM) motor/encoder and diaphragm bearing assemblies. Introduction CRISM will use targeted observations to search for evidence of aqueous activity and to characterize the geology and composition of surface features' on Mars (Figure 1). Global measurements acquired repeatedly throughout the Martian year will provide information on atmospheric water vapor, CO, and aerosols complementary to that from other MRO instruments. The Optical Sensor Unit (OSU) consists of an optical system, a cryogenic system, and focal plane electronics gimbaled about a single axis to allow scanning over k60" from nadir. Its mechanical design builds on proven technology from previously successful APL instrument designs. The base housings are fabricated from titanium that provides high stiff ness and thermal isolation within the same component. The gimbal bearings are a precision assembly designed to operate in a -60°C environment. The gimbal is driven directly by a brushless DC motor paired with a BE1 20-bit incremental position encoder. The encoder disk is co-mounted directly to the bearing shaft beside the motor rotor and the read heads are mounted to the bearing housing alongside the motor stator. The electrical signals and purge are passed through a twist capsule in the center of the motor/encoder bearing assembly. A second bearing pair is mounted in a parallel diaphragm bearing housing that provides high stiffness in the lateral directions to the gimbal axis and flexibility along the gimbal axis to compensate for differential expansion of the instrument and spacecraft. The spectrometer housing is passively cooled to -90°C using a flexible link to the anti-sunward radiator. The anti-sunward radiator passes through the center of the diaphragm bearings in a thermally isolated mount, and thus rotates with the OSU; its FOV is independent of gimbal position. 4 I Figure 1. CRISM instrument Johns Hopkins University Applied Physics Laboratory, Laurel, MD Proceedings of the 38 Aerospace Mechanisms Symposium, Langley Research Center, May 17- 19,2006 11 | Document | AMS_2006.pdf | 25 |
CRISM Thermal Design The CRlSM thermal design provides both active cooling of the IR focal plane to cryogenic temperatures to reduce dark current and passive cooling of the spectrometer housing to -80°C for low background. Simultaneously it keeps the electronics section near -40°C (Fig. 2). Cryogenic cooling is provided by three Ricor K508 integral Stirling cryocoolers. The multi-cooler configuration requires “thermal switching” between coolers. A cryogenic diode heat pipe assembly consisting of heat pipes and a thermally isolating mounting assembly connects the active cooler with the focal plane while isolating if from the two dormant coolers2. Each of the three diode heat pipes is connected to the focal plane on one end and to a cooler on the other. The focal plane electronics are mounted in the bottom of the gimbal housing and are maintained at -40°C. The housing along with tantalum plates provide EM1 shielding and minimize cable length to the focal planes. Table 1 lists the CRISM expected flight operating temperatures. Table 1. CRISM Flight Operating Temperatures -40”c 7, I Component ~ I I Min. I Max. I mousina I “C I -1 IR FPA 120 Crvo-Coolers I “C I -25 I 20 c T / 113K -20°C 4- -7 7 -1 20°C I Figure 2. CRISM Thermal Zones 12 | Document | AMS_2006.pdf | 26 |
CRISM Bearing System The CRlSM instrument rotates the OSU do" from nadir. The OSU weighs 20.9 kg (45.9 Ibf) and is supported by the motor/encoder (Figure 3) and the diaphragm (Figure 4) bearing assemblies. The motor- encoder side used a thin-section angular-contact duplex pair mounted back to back and was designed to take the non-axis moments, the entire thrust load, and its share of the radial loads associated with launch event (Figure 5). These bearing were designed for a 266.9-N (60-lbf) axial preload. The diaphragm side used a thin-section angular-contact duplex set mounted face to face and was designed to take only the radial loads associated with the launch event. The diaphragm bearings were designed for a 66.7-N (15- Ibf) axial preload. 7 Spacer/Seal Bearing Spacers Switch Titanium Spindle 2x Limit Titanium Housing Motor Stator Motor Rotor - Encoder Disk Purge Fitting - Twist Capsule i Figure 3. CRISM MotorEncoder Assembly Preloading Angular contact bearings should be preloaded as lightly as necessary to achieve the desired results. A duplex pair is a pair of bearings that have a pre-determined amount of preload built into them. This was accomplished by grinding the inner or outer ring a sufficient amount to eliminate all internal clearance within the bearing commonly referred to as the preload offset. There are, however, several disadvantages to preloading bearings: Increased running torque 0 0 Sensitivity to misalignment Sensitivity to differential thermal expansion 13 | Document | AMS_2006.pdf | 27 |
Parallel Corrugated Diaphragms m I s 7 1 I k- Bearings Flex link L Interface r /- Titanium Housing Figure 4. CRISM Diaphragm Bearing Assembly Anti-Sunward Radiator The CRlSM duplex bearing pairs were separated by titanium spacers so that the preload offset would remain constant over temperature. However, the difference between the 440C inner and outer rings and the titanium shaft and housing resulted in a reduction of clearance as temperature decreased (Figure 6). A reduction of clearance results in a decrease of the contact angle. However, the bearings are only going to experience substantial axial loads during the launch. The bearings were tested to -196°C and continued to rotate freely. The few disadvantages of preloading are more than offset by the following advantages: Reduces axial and radial runout of the rotating shaft. Required for the encoder disk to read head alignment Reduces the shaft deflection under load and improves its assembled stiffness Removes free play in the bearing set, keeping the bearing set loaded in-order to avoid skidding of the balls Minimizes the peak stresses that occur during the maximum loading events by ensuring the load on the bearings is shared by more balls in each bearing 0 Decreases bearing noise In addition to the axial preload, the CRlSM bearings employed a light interference fit, 12.7 pm (0.0005 in) in the bearingkhaft fit and 15.2 pm (0.0006 in) in the bearing/housing fit. 14 | Document | AMS_2006.pdf | 28 |
Encoder Disk- Interface Bearing - Bearing 7 Spacer Bearin - 0.80 0.70 h C ~ 0.60 0 9 .- v $ 0.50 C (d 5 0.40 !!! - (d 2 0.30 E C - 0.20 0.10 Bearing - Retainer Figure 5. CRlSM motodencoder spindle - - L - Clearance at LMC I I l~l~l~l~l~l~l~l~l -200 -180 -160 -140 -120 -100 -80 -60 -40 -20 0 20 Temperature (C) 12.5 12.0 h w Q) U v e, w 11.5 2 - c 0 (d C c 6 11.0 10.5 Figure 6. CRlSM motor/encoder bearing clearance 15 | Document | AMS_2006.pdf | 29 |
CRlSM Bearing Lubrication The thermal analysis of the CRlSM motor/encoder and diaphragm bearing assemblies indicated that they would operate at -40°C and -90°C respectively. The analysis indicated that the diaphragm bearing assembly would be too cold for oil or grease lubrication. Our tests showed that -40°C was beyond the acceptable temperature range for Pennzane synthesized hydrocarbon based oils and greases and -90°C was beyond the accepted range for perfluorinated polyether based Brayco oils and greases. It was desired to use identical lubrication in both bearing assemblies for the following reasons: 0 a To simplify bearinghubrication testing 0 To have nearly identical bearings and identical lubrication in both bearing assemblies Once the bearings and lubrication were selected, an single qualification test could be applied to both assemblies The three choices of dry film lubrication that were considered were: 1. Ion plated lead 2. Sputtered MoS2 (molybdenum disulfide) 3. WS2 (tungsten disulfide) Ion-plated lead bearings were successfully used in the Compact Remote Imaging Spectrometer (CRISP) tracking mirror assembly on the Comet Nucleus TOUR (CONTOUR) spacecraft. The CRISP tracking mirror assembly bearings were a Barden precision angular contact duplex pair with a bronze cage. They were purchased through BE1 as part of their motor/encoder assembly. This was done since they were the same bearings used in the BE1 motor/encoder and they had used an identical set previously in the SABER3 instrument on the TIMED spacecraft. However, we were unable to procure ion-plated lead thin- section bearings at the time. If it were possible to procure them in the future, we would definitely recommend trying them. There is a great deal of literature touting the benefits of sputtered MoS,. Dry film lubrication tests for MoS2 and WS2 were conducted for the CRISP cover and release mechansim3. Sputtered MoS2 is supposed to work great in vacuum, however, moisture absorption can cause severe performance degradation. Most of the CRISP cover and release mechanism qualification testing was conducted in ambient conditions including a test on NASA’s Low Gravity Experiment aircraft (the “Vomit Comet”). Tests comparing sputtered MoS2 and WS,, showed WS2 to be clearly superior to sputtered MoS, for this application. Sputtered MoS2 seemed to exhibit tremendous stiction in tests conducted in ambient conditions where as stiction was virtually undetectable with WS2. In fact, we believe that the cover and release mechanism would not have worked at all using sputtered MoS, in ambient conditions. Sputtered MoS2 was also used in the bearings for the MDlS instrument on the MESSENGER spacecraft launched in 200X. In addition to the well known issues of sputtered MoS2, this instrument suffered two additional issues: 1. Smoothness 2. The sputtered MoS2 coating is a hard and brittle coating The MDlS instrument was able to meet its performance requirements with the sputtered MoS2 coated bearings. However, the initial flight assembly had to be replaced and a new assembly built up because the sputtered coating had become cracked and resulted in large unacceptable torque spikes. Based on our experiences, we would NOT recommend sputtered MoS2 coatings. WS,, Dicronite, was chosen for the CRlSM motor/encoder and diaphragm thin section bearings. However, we were unable to locate any specific examples or find any heritage on WS2-lubricated bearings used in space applications in the available literature. However, based on our previous success with WS2 on the CRISP instrument, we were optimistic that WS2 could also work in bearings. We also knew that the bearing companies offered it as an option and that the WS2 coaters coated bearings. We decided to conduct our own WS2 coated thin-section Teflon-toroid bearing test. 16 | Document | AMS_2006.pdf | 30 |
CRISM Bearing Tests Identical sets of sputtered MoS2-coated and WS,-coated thin-section Teflon-toroid angular contact duplex-pair bearings were purchased simultaneously (Figure 7). Prior to running the tests, an initial set of pictures using a scanning electron microscope were taken of both types of bearings (Figure 8 through Figure 10). Figures 8 through 10 show that MoS2 appears to have a much rougher surface than WS2. We believe this is why the MDlS bearings never “felt” smooth. Figure 7. WS, and Mo S2 life test bearings Figure Sb. MoS2 Coating 17 | Document | AMS_2006.pdf | 31 |
rigure iua. w& boaring Figure IUD. ~os~~oaring Following the test of the WS2-coated bearings in which 57952 cycles were completed, we were concerned that the bearings seemed much rougher than they were prior to the test. After a visual inspection of the bearings, a significant amount of debris in the bearings was found. Post-test pictures were taken with a standard microscope, Figure 11 through Figure 13. There were several types of contamination found: 0 White particles 0 0 Brown particles on the toroids Brown film on the raceways We were concerned that the contamination found in the bearings following the test would present a real problem, both in terms of lifespan and smoothness. However, upon review of the data, the rougher feel was not adversely affecting the motor control system. The bearings were disassembled and further inspected. We could find no evidence of wear in the raceways or the WS2 coating. The life test was also significantly longer than the expected lifetime of the instrument at Mars. Bearinq Test Conclusion The WS2 coating is more than adequate for a slow moving oscillating gimbal requiring dry film lubrication. We were satisfied enough with the WS2 performance that we did not test the sputtered MoS2 bearings, we had found the solution we were looking for. The Teflon toroids would likely continue to break down and contaminate the bearings with additional particles that the instrument may no longer be able to rotate as precisely as is required. We believe the Teflon toroids will ultimately be the life limiting factor of these 18 | Document | AMS_2006.pdf | 32 |
bearings. Although this may be an advantage for continuously rotating higher speed mechanisms acting as an additional lubricant Figure 11. Post life test bearing Figure 12. Post life test bearing raceway Figure 13. Post life test b6amlllY mabcway CRISM Bearing Anomalies Bearina Ball Spacing The APL Space Department has used different sizes of thin-section Teflon-toroid-spaced bearings on several programs and tests. One feature that we have always noticed with these types of bearings is the potential for a “large” gap to occur between a ball and toroid (Figure 14). We attribute both the high running torque and the poor feel to the non-uniform ball-toroid spacing. We believe that non-uniform ball-toroid spacing can result in some various amounts of pressure between all the toroids and balls. This can result in unpredictable friction between the ball and toroid causing an increase in the running torque. It may also result in a stick-slip situation between the ball and toroid resulting in torque spikes or non-smooth rotation. Additionally, it could lead to rapid wear of the toroid. The toroid wear particles could wind up in the raceways as contamination, also causing anomalous 19 | Document | AMS_2006.pdf | 33 |
torque spikes. Grease and oil lubricated bearings may exhibit the same problems as the CRISM dry film lubricated bearings to a far less noticeable degree due to the grease or oil between the ball and toroid. The evidence for this conclusion with the CRlSM bearings is two fold: 1. The significant amount of Teflon contamination found in the bearings following the run-in procedure described in Appendix 1 2. The drastic change in torque level and smoothness following the high-pressure air blow through cleaning procedure described in Appendix 1 A Figure 14. Thin section bearings, same size, different manufacturer ReDroducible Torque SDikes Another type of torque spike was also noticed with the CRlSM bearings that could be easily reproduced based on the operation of the spindle. During the run-in procedure in Appendix 1, unidirectional rotation resulted in extremely smooth running torque. However, CRISM was intended to oscillate do”. A reproducible torque spike occurred following a change in direction of rotation. Rotating the spindle backwards and forwards, sometimes referred to as “safe cracking”, torque spikes, equal to or greater than the nominal running torque, would result within several degrees of rotation following the reversal. A ball can never remain rolling between surfaces that form an angle to each other5. All angular-contact ball bearings create an angle between the two raceways. Therefore, as the bearings rotate, the balls produce a gyroscopic motion in addition to rolling. Pressure or drag friction between the toroid and the adjacent balls also seems to deflect the toroids based on the direction of rotation. Thus, the gyroscopic motion of the balls and pressure between toroids and adjacent balls appears to cause to cause the toroids to align themselves based on the direction of rotation. This behavior appears to be attributable to the vast majority of toroids aligning themselves (Figure 15). When the spindle reverses, as do the balls, the toroids flip and align themselves in the opposite direction (Figure 16). As the toroids flip and align themselves in the opposite, direction, a torque spike resulted. This effect was significantly reduced following the cleaning procedure. Once the balls and toroids became more evenly spaced, the drag friction between the toroid and adjacent balls was reduced, thus reducing the ability of the toroids to align themselves. This was also noticed following the Christmas holiday. As the bearings sat over the holiday, pressure between the balls and toroids either slightly re-spaced the balls or caused the Teflon to cold flow resulting in bearings the felt much better after having sat for an extended period. However, following another run-in, the bearings quickly resorted to their pre-holiday behavior. 20 | Document | AMS_2006.pdf | 34 |
L Figure 15. Toroids in “down” position Unfortunately, we never photodocumented this effect, thus, it was extremely difficult to find pictures of the flight hardware demonstrating this effect clearly. Figure 15 clearly shows all the toroids uniformly aligned in the “down” position. Figure 16 shows most of the toroids in aligned in the “up” position, including most of the ones at the top of the bearing. This effect can not be reproduced with a single unmounted bearing. CRISM Bearing Assembly Procedure The CRlSM bearings were removed from the manufacturer’s packaging and visually inspected to verify that they conformed to the documentation and were marked properly (Figure 17). Once we were sure that the bearings were correctly marked and free of contamination, they were assembled as follows: 1. The bearings and spacers were stacked and aligned per drawing and documentation (Figure 18) 2. The bearings and spacers were placed in an assembly fixture (Figure 19) specifically designed to keep them aligned during assembly 3. The bearing retainer was placed on the assembly (Figure 20) 4. The bearing shaft (Figure 21) was cooled in liquid nitrogen 5. The cooled bearing spindle was quickly assembled in the bearing assembly fixture. Weights were placed on the assembly to ensure that it remained seated against the top bearing. 6. The bearing assembly was quickly placed in a N2 purged vessel and allowed to equilibrate for 24 hours (Figure 22) 7. The spindle bearing assembly was then visually inspected and checked for “feel”. It was noted that there was a significant amount of pressure on the spacers between the two bearings on both the Engineering Test Unit (ETU) and the flight unit. The ETU and the flight assembly were the first bearing assemblies with spacers that we had assembled using the liquid nitrogen technique. Previously, the bearings were duplex pairs without spacers. We were concerned that with the shaft being significantly cooler than the bearings and spacers, that as it warmed and expanded, there would be very little pressure between the bearings and spacers or worse, a gap. However, that never materialized 8. The entire spindle bearing assembly was then cooled in liquid nitrogen 9. The cooled spindle bearing assembly was quickly assembled in the titanium motor/encoder/bearing housing 10. The assembly was then purged and allowed to equilibrate for 24 hours 11. The assembly was visually inspected and checked for “feel” 12. We were not happy with the “feel” of the flight assembly and took the steps outlined in Appendix 1 There were two main issues with the flight bearing assembly, 1) high running torque; 2) torque spikes that made the assembly not feel smooth as it rotated. We were confident that contamination was not the problem from pre- and post-assembly inspections causing either of these problems. 21 | Document | AMS_2006.pdf | 35 |
A Figure 17. CRISM riignr uearings I I.: .. I .. ,C , . ' , -,. c I. ; riyure IO. Deuririya stuFncu uriu uiiy~icu s- a 22 | Document | AMS_2006.pdf | 36 |
Appendix 1 Date Operation 12/04/2003 Initial torque measurements Table 2 lists the operations and measurements made to the flight motor/encoder bearing assembly. Most measurements were an average running torque. Some measurements were a peak torque when there was a significant torque spike to the running torque. On several tests where many measurements were made, an average value is reported and denoted by and Avg. in the Torque column. Torque mN-m (oz-in) 111.6 (15.8) Table 2. Bearing Assembly Operations 12/16/2003 11 1.6 (1 5.8) nominal 190.7 (27.0) peak Run-in Q60 RPM, 19 minutes (1 140 Revolutions) 143.0 (20.25) Avg. Run-in Q60 RPM, 25 minutes (1500 Revolutions) 381.3 (54.0) Cool off 254.2 (36.0) 12/24/2003- 1 /4/2004 Loosen retaining nut Run-in Q60 RPM, 20 minutes (1200 Revolutions) Tighten retaining nut, 17.0 N-m (1 2.5 ft-lb) Stored in N2 purged container in cleanroom 158.9 (22.5) 190.7 (27.0) 1 /5/2004 Changed torque measuring method for all future measurements. Peak torque measurements ~ Loosen retaining nut and re-measure torques Set-1 238.7 (33.8) Avg. Set-2 317.8 (45.0) Avg 1 /6/2004 Run-in Q60 RPM, 20 minutes (1200 Revolutions) Peak torque measurements The best these bearings have ever “felt” Removed retaining nut, rotor, and spacer 148.3 (21 .O) Avg. | Document | AMS_2006.pdf | 37 |
Acknowledgements This work was performed at Johns Hopkins University Applied Physics Laboratory under a contract with the National Aeronautics and Space Administration. The authors would like to thank all those who helped design, fabricate, integrate and test the CRISM instrument. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not constitute or imply its endorsement by the United States Government or the Johns Hopkins University Applied Physics Laboratory, Laurel, MD. References 1. S. Murchie et. al., “CRISM (Compact Reconnaissance Imaging Spectrometer for Mars) on MRO (Mars Reconnaissance Orbiter)”, Instruments, Science, and Methods for Geospace and Planetary Remote Sensing, SPlE Vol. 5660, pp. 66-77. 2. Bugby, D., J. Garzon, 8. Marland, 6. Stouffer, D. Mehoke. M. Fasold, “Cryogenic Diode Heat Pipe System for Cryocooler Redundancy,” SPlE Optics and Photonics Conference, Cryogenic Optical Systems and Instruments XI, San Diego, CA, 31 July -4 August, 2005. 3. J. Lees, E. Schaefer, “Design and Testing of the CRISP Tracking Mirror Cover and Release Mechanism”, Proceeding of the 36‘h Aerospace Mechanisms Symposium, Glenn Research Center, April, 2002, NASNCP-2002-211506, pp.63-76. 4. Esplin, Roy. A satellite-based multichannel infrared radiometer to sound the atmosphere (SABER). In Optical Remote Sensing of the Atmosphere held in Salt Lake City, Utah, 5-9 February 1995. pp. 130-1 32. 5. Harris, Rollina Bearina Analvsis, 3rd Edition, John Wiley & Sons, Inc., pp.451-452. 24 | Document | AMS_2006.pdf | 38 |
Gear Teeth Particles and Bearing Failures William H. Greenwood* and Jeffrey G. Dabling* Abstract Torque is transmitted from rotary solenoids to rotate drive arms that advance a ratchet wheel as part of a safety mechanism in missile warheads. The small volume constraint led to single gear teeth to transmit the torque from the rotary solenoid. High contact forces and compliant gear teeth caused many fine particles to be generated at the rubbing surfaces of the gear teeth. The particles were pulled into the ball bearings of the adjacent solenoids causing early failure while having no ill effect on the ball bearings of the drive arms. A temporary solution of custom plastic shields allowed the prototype units to proceed to environmental and flight tests. A subsequent build replaced the gear teeth with ball bearing followers. Introduction Missile warheads and bombs usually have safety mechanisms [l ] to prevent unintended explosions in the event of accidents. These safety mechanisms are small and made of stainless steels for high temperature integrity in the event of accidents involving jet fuel or propellant fires. The safety mechanisms often have ratchet-type wheels, and are usually driven by rotary solenoids (Figure 1). Figure - Mounting plate, ratchet wheel, and rotary solenoid * Sandia National Laboratories, Albuquerque, New Mexico Proceedings of the 3dh Aerospace Mechanisms Symposium, Langley Research Center, May 17- 19,2006 25 | Document | AMS_2006.pdf | 39 |
A new generation of safety mechanisms is nearing production. The safety mechanism of Figure 1 is among the new generation and transmits torque from the rotary solenoid to the drive arms by way of gears. Due to the limited stroke of the solenoid and the volume constraint, the gears are actually single gear teeth cut as part of the solenoid rotor and drive arm (Figure 2). Figure 2. Rotary solenoid and single gear teeth Earlier generations of these safety mechanisms often used full gears to transmit torque from the ratchet type wheels to other shafts for thousands of ratchet wheel cycles. One life test of the first generation of safety mechanisms involved 10,000 ratchet wheel cycles; the main gear had teeth that were no longer involutes but appeared triangular. Much wear debris was adjacent to the gear but did not migrate to ball bearings or cause operational failure. The first units of the new intent safety mechanism were operated in January 2003 and failed to operate after only a dozen to a few hundred ratchet wheel cycles, as shown in Table 1. A failure to operate at such low number of cycles was unexpected. The original units had molybdenum disulfide (MoS2) applied to the gear teeth, but had no lubrication on the ball bearings of the solenoid or the drive arms. Molybdenum disulfide is used since the mechanisms must sit for decades unused and then be called upon to function. The safety mechanisms are usually hermetic units and have a nitrogen and helium atmosphere, although the ratchet cycle tests in this report were all performed in the laboratory atmosphere as part of early qualification. The ball bearings are generally lubricated but a new process for the bearings was not yet ready, so the early ratchet wheel cycle tests were performed with dry (non-lubricated) ball bearings. 26 | Document | AMS_2006.pdf | 40 |
Table 1. Ratchet wheel rotations to failure, MoS2 on gears, no MoS2 on bearings Powder at gear teeth Initial investigation The energized stroke of the solenoid is on the order of five milliseconds and the spring (de-energized) return of the drive arms is also on the order of five milliseconds. The initial investigation by high-speed video of the mechanism in operation showed erratic operation times for the inboard rotor during energized strokes when the unit was close to failure. Failed units were disassembled and visually examined. Drive arm bearings and drive springs showed no problems, and no rubs were detected between the arms and adjacent surfaces. The wear zone on the gear teeth showed a powdered layer rather than a burnished appearance as shown in Figure 3. The powdered layer was 0.1 to 0.2-mm thick at the edges and 0.05 mm or thinner in the center. The powdered layer was composed of many micron size particles and was readily scraped or wiped away from the gears. In addition, many small dark particles were observed on the mounting plate under the gear teeth and at stop pins. The quantity of particles was much greater than observed in previous safety mechanisms. An additional troubling test was a higher than expected minimum operate voltage. The minimum operate voltage occurs when the solenoid torque is nearly equal to the drive arms spring torque. The solenoids were designed to operate the mechanism as low as 16 volts and up to the maximum supply voltage of 22 volts. The ratchet wheel cycle tests are operated at 22 volts to give more solenoid torque and greater forces and loads on the bearings and pins. From the measured solenoid torque and the measured drive spring forces, the ratchet wheel units were expected to operate at 14 volts but when initially assembled were measured to operate at about 15 volts. This higher minimum operate voltage indicated a small loss, in the solenoid bearings, gear teeth or drive arm bearings, and the gear teeth contact between the solenoid and the drive arms was suspected as the probable cause. The powdered layer that was observed at the gear teeth contact in the first wheel cycle tests gave more evidence than the gears had substantial sliding contact and not the rolling contact desired in well-formed gear trains. The solenoid torque at 22 volts is approximately double the solenoid torque at 15 volts. The failure to operate after dozens or a hundred ratchet wheel cycles at 22 volts means an apparent frictional loss increased from near zero at the start to approximately 50% of the solenoid torque. If the large frictional loss was solely at the gear teeth, then slower and erratic spring return times of the drive arms and solenoid rotors were also expected but not seen in the high-speed videos. The powdered layer at the gear contacts was obviously troubling, but it did not appear that the gear teeth interaction caused such high loss of torque. From the failed units, a solenoid rotor was manually stroked and no high friction was measured. Several failed units were cleaned of the powdered layer at the gear teeth contact, and the minimum operate voltage improved somewhat to 18 to 20 volts, but was still significantly higher than the initial 15-volt measurement. The powdered layer at the gear teeth did not appear to be the high fiction loss that caused the units to stop at 22 volts. 27 | Document | AMS_2006.pdf | 41 |
al Figure 3. Gear tooth wear zone has a powdered layer. The solenoids from failed units were rotated manually and some roughness in rotation was noted. However, a quasi-static torque test of the solenoid at 16 and 22 volts showed nearly the same torque as measured before the installation into the ratchet wheel assembly, but the measured friction at zero volts was somewhat higher. The solenoid bearings in failed units were visually examined at 1 Ox magnification, and the balls in the bearings in the inboard rotor had a roughened or a worn surface (Figure 4). The solenoid ball bearings were replaced, and the solenoids were re-assembled into their failed ratchet wheel units. The rebuilt units were then measured for minimum operate voltage and found to be 15 volts, the same as new units. The parts that degraded as the ratchet wheel units were operated were the solenoid bearings. The rebuilt units (with new solenoid bearings) were then cycle tested again and failed at similar cycles as those of Table 1. The solenoid bearings could be degrading due to high dynamic loads or from particles from the gear teeth entering the bearings during operations. The axial load from the magnetic forces when the coil is energized was calculated to be 71 N (1 6 Ibf) compared to the dynamic load rating of 11 6 N (26 Ibf). The small oscillatory solenoid stroke of 0.2 rad (12 degrees) is a more severe service than continuous rotations since only a small portion of the balls see the load repeatedly. Therefore, the high axial magnetic force could be degrading the unlubricated bearings in such a few cycles. However, the innermost bearings of the rotors support all of the axial forces; whereas the failed units showed bearing wear to be most severe in the inboard rotor and approximately equal from the innermost and outermost bearings of the inboard rotor. The rapid bearing degradation did not appear to be caused by the relatively high dynamic axial load. 28 | Document | AMS_2006.pdf | 42 |
,A ri F- A PA Figure 4. Solenoid ball bearings before (left) and after (right) life cycle test. When ball bearings fail from high loads, the initial signs are many small pits as bearing material spalls, then the ejecta from the pits are rolled into the bearing surfaces that cause more localized high loads and accelerated degradation and eventually a non-rotating (seized) bearing. As the bearings were visually scrutinized, many patches or scabs were observed and some indentations but not pits. The scabs appeared to be a particle or perhaps several particles that are rolled into the balls as the bearing is operated. The scabs were foreign particles and not bearing ejecta. The parts that degraded and caused the early wheel cycle failure were the solenoid bearings; the cause of the bearing degradation was the particles from the gear teeth. The problem resolution was to sharply reduce the particles generated at the gear teeth. Attempts to Reduce the Gear Teeth Particles The gear teeth were lubricated with a proprietary molybdenum disulfide process. A tribological investigation showed some alumina particles imbedded in the surface where the solid lubricant was applied. The lubricant applicator did indeed prepare the surfaces with alumina beads. A possible cause was identified; some of the alumina beads from the surface preparation were imbedded into the gear surface and caused grinding at the mating gear surfaces and generated the great amount of metal particles. Gear teeth surfaces were cleaned of the MoS2 and alumina beads and were then life cycle tested, but the same failure to operate after a few hundred ratchet wheel cycles was observed (Table 2). The post- mortem inspection again showed many metal particles at the gear teeth wear zones and scattered on the nearby mounting plate and stop pins. The inboard solenoid rotor had much rougher manual rotation than the outboard rotor. The wear zone on the inboard rotor gear tooth was broader at the base as shown in Figure 3, rather than a more parallel wear zone observed in the outboard rotor gear tooth. 29 | Document | AMS_2006.pdf | 43 |
Table 2. Ratchet wheel rotations to failure, no MoS2 on gears, no MoS2 on bearings CP008 CP009 12 Feb 2003 180 Rough manual rotation 14 Feb 2003 260 Rough manual rotation Several units with the wear resistant outboard gear were tested with the same disappointing result: the units failed to operate after a few ratchet wheel cycles, and there was a thick powdered layer at the wear zone on the gear teeth (Table 3 and Figure 5). The wear zone on the new outboard gear appeared to be even heavier than the wear zone on the trials from Tables 1 and 2. The rotors spun roughly after the life test. The bearings were observed to have a roughened surface on the inboard rotor, but only slightly roughened on the outboard or stator bearings. The long stringers observed in Figure 5 on the edge of the gear are composed of many fine particles that are magnetized. These magnetized stringers appear at the edges of other parts as well as the gear teeth. A schematic of the solenoid is shown in Figure 6. The solenoid bearings from left to right will be numbered 1 through 4 in many of the descriptions to follow. The outboard rotor bearings are 1 and 2 while the inboard rotor bearings are 3 and 4. Figure 5. Wear particles generated due to a Nitronic 60 gear tooth interface 30 | Document | AMS_2006.pdf | 44 |
Table 3. Ratchet wheel rotations to failure, Nitronic 60 gear, no MoS2 on gear or bearings Date Wheel cycles p CP011 20 Feb 2003 72 Bearings 3 and 4 bad, bearings 1 and 2 fair CP037 20 Feb 2003 280 Bearings 3 and 4 bad, bearings 1 and 2 fair CP034 21 Feb2003 220 Bearing 3 very bad CP045 CP054 CP021 Outboard rotor Outboard coil -- 27 Feb 2003 110 I 27 Feb 2003 90 Inboard rotor rotates very roughly 03 Mar 2003 51 8 Inboard rotor bearing seized Stator 24.9 mm L Inboard coil Inboard rotor I_- 19.2mm -4 Inboarc \ Ball bearings i gear Figure 6. Double rotary solenoid; outboard and inboard rotors move independently. A new trial was performed with MoS2 to the applied gear teeth, but with the surfaces prepared by acid etching rather than alumina bead blasting as in the original condition of Table 1. The desire was a drastic reduction of wear particles at the gear teeth. Again, low ratchet wheel cycles to failure were measured and a heavy powdered layer at the gear teeth was again observed (Table 4). The third replication that failed at wheel cycle 518 was encouraging, but the two replications that failed at low cycles indicate that the condition of Table 4 is similar to the first three conditions. The unit that failed at 51 8 cycles was examined carefully in the post-mortem, but resulted in no clues regarding its much longer operating life. Table 4. Ratchet wheel rotations to failure, MoS2 on gears without grit blasting 31 | Document | AMS_2006.pdf | 45 |
A Temporary Solution - Bearing Shields As our early efforts to decrease particle generation at the gear teeth were unsuccessful (Tables 2-4), we looked for a quick solution; so that environmental and flight tests could continue. A shielded bearing was suggested as a way to prevent the gear teeth particles from entering the solenoid bearings. However, there is not room for standard shielded bearings. The inboard rotor is 3.8-mm thick; the original bearings are 1.6 mm thick, which leaves a 0.6-mm gap between the ball bearings. A shielded bearing of the same inside and outside diameters has a thickness of 2.4 mm that for a pair of bearings is wider than the rotor. The temporary solution was to make thin washers, 50 microns (.002) thick, out of acetate plastic that had the same inside and outside diameters as the bearings and were placed next to the bearings of the inboard rotor. Plastic was used since they could be made quickly. One shield was placed on the inboard rotor bearing next to the outboard gear (bearing 4) as shown in Figure 7, and the other shield was placed between the inboard rotor bearing (bearing 3) and the stator bearing (bearing 2). These shields were placed in test units with no MoS2 on the gears or bearings and had encouraging results as shown in Table 5. Finally, most units run at least 500 wheel cycles. The initial solenoid torque tests showed no increase in friction due to addition of the plastic shields. After the life tests, the shields showed a darkened annulus where the bearings were sliding, but the plastic was not grooved. A workable solution appeared possible. The shields appeared to be protecting bearings 3 and 4 since tested units now had some bad bearings at position 1 with some good bearings at position 4. -- I Figure 7. Plastic shields were used to keep particles out of the solenoid bearings 32 | Document | AMS_2006.pdf | 46 |
Table 5. Wheel rotations to failure, plastic shields over bearings, no MoS2 on gears or bearings -- -- unit # Date Wheel cycles -- CP046 03 Mar 2003 524 Failed, bearing 1,3,4 - bad, bearing 2 - good CPO14 03 Mar 2003 520+ No failure, bearing 1 - bad, 2 - fair, 3, 4 - good CP020 04 Mar 2003 820+ No failure, bearing 1,3 - bad, , bearing 2,4 - good CPO16 13 Mar 2003 404 Failed, bearing 1, 4 - bad, bearing 2 ,3 - fair Ball bearings with MoS2 lubricant were received and put into solenoids with the plastic shields next to the inboard ball bearings and were cycle tested with very favorable results (Table 6). A fix or a solution to our early wheel cycle failure was found. The solenoid used in the first test of CP028 showed an unrelated assembly issue, resulting in a frictional rub between the inboard rotor and the stator. After being rebuilt with an additional 13-micron shim added under the inboard rotor but with the original bearings, the second cycle test of CP028 performed very well. Table 6. Wheel rotations to failure, plastic shields over bearings with MoS2, no Mo~ on gears Unit#m U ate wh eel cycles CP046 21 Mar2003 101 4+ No failure, bearing 1, 2, 3, 4 - good CP048 25 Mar 2003 1 ooo+ No failure, bearing 1,2 -fair, bearing 3,4 - good CP028 25 Mar 2003 52 Failed, bearings - good, inboard rotor rub CP028 26 Mar 2003 1 ooo+ No failure, bearing 1, 3, 4 - good, bearing 2 - fair As previously discussed, the dynamic axial load was 71 N versus the manufacturer’s rated load of 116 N. Since our load condition is a small reversing stroke rather than continuous rotation, the duty is more severe than the rated condition. The dry film lubrication used in the bearings of the units of Table 6 appears beneficial when compared to the dry bearings of Table 5 and confirms the bearing manufacturer’s recommendation that lubrication becomes more important as the bearings are more severely loaded. The drive arm bearings are very close to the gear teeth but did not show any wear when compared to the solenoid bearings. Many of the gear particles stayed on the gear teeth as shown in Figures 3 and 5, but some of the particles will migrate to other parts as shown in Figures 7 and 8. The electric coils in the solenoids produce a magnetic flux field when the coils are energized. The flux field is in the shape of a toroid that makes a complete loop about the coil from the inside coil diameter to the outside coil diameter. Most of the flux passes through the rotor and stator due to the high permeability of electrical iron, but some flux passes through the air between the races of the bearings. Loose magnetic particles will align on these flux lines and then move along the flux lines to high magnetomotive force. Some of the particles are pulled into the bearings where they become scabs as the balls roll aginst the races. The large axial loads on the solenoid rotors are present only when the coils are energized. The wear particles are pulled into the bearings and rolled flat when the coils are energized; the rotors are returned to their home positions by helical extension springs. Hence, the slower and erratic motion of the rotor was observed during the energized stroke only and not during the return stroke. Likewise, a manual rotation and a zero-volt solenoid torque test would not show the large friction that is present when the coil is energized. The inboard rotor motion was erratic and caused the operational failure since the inboard rotor bearings are very close to the gear teeth and the resulting wear particles. When the plastic shields are in place as shown in Tables 5 and 6, the inboard bearings (3 and 4) are not consistently worse than the outboard bearings (1 and 2). 33 | Document | AMS_2006.pdf | 47 |
Why so many gear teeth particles? The tribological investigation continued to evaluate material pairs or lubrication that would decrease the gear particles. Some standard pin on disk tests at a high contact stress showed long life of the MoS2 lubricant and little particle generation. This caused us to calculate the Hertzian contact stress [2] at the gear teeth. The contact stress was 753 MPa (109 ksi). This exceeds the yield strength of the Hiperco 50A, 365 MPa (53 ksi), used for the inboard rotor, the wear resistant stainless steel, Nitronic 60, 414 MPa (60 ksi), used for the outboard gear in one of the trials, and the drive arms PH 13-8 Mol condition H1150 , 620 MPa (90 ksi). The contact stress was undoubtedly higher on the single cantilevered gear tooth of the inboard rotor since it deflects during loading and the wear zone on the tooth was more triangular than rectangular. Figure 3 shows just such a wear zone on the drive arm gear tooth that mates with the inboard rotor gear tooth. The single gear tooth was analyzed for bending stresses at the base during the initial design but the Hertzian contact stresses were not calculated for the contact load at the gear teeth. The high contact stresses mean plastic deformation at the load zones, non-rolling contact and resultant particle generation. Long term solution - eliminate the gear teeth The plastic shield was a fix that permitted further testing of the first build. However, the many particles were undesirable and created a potential for other problems. A long-term solution was desired where few particles were generated. Increasing the gear pitch or the length of the gear teeth to decrease the contact stress by a factor of two or three was not possible due to the volume constraints. In addition, the single gear tooth on the inner rotor and the drive arms made these expensive parts and difficult for part 34 | Document | AMS_2006.pdf | 48 |
acceptance. Changing the solenoid design from a single shaft to two solenoids side by side was considered, but rejected as too drastic a change. The change selected was to use ball bearings as followers between the solenoid and drive arms. The space was limited but the small 1.0-mm by 3.2-mm ball bearing could be made to fit. See Figure 9 and compare to Figure 2. The contact load was high but within the manufacturer's limits of 4,000 MPa (580 ksi). This change was implemented on the subsequent group build, and has been successful with long ratchet wheel cycles before mechanism failure. There are a few dark particles after a ratchet wheel cycle test, but they are located at stop pins and not at the bearing followers that actuate the drive arms. '* I --3 L 4 1 Figure 9. Ball bearings now transfer the torque from the solenoid to the drive arms Conclusion Due to volume constraints and the desire for a simple assembly, single gear teeth were made as integral part of arms and rotors in a small safety mechanism. The gear teeth were discovered in the prototype to generate many very small particles that were pulled into the bearings of the adjacent solenoid and caused very early failures in life tests. The cause of the particle generation was contact stresses exceeding the yield strength of the gear teeth. The lesson learned was to prioritize our efforts in analysis and failure mode prediction on new or unfamiliar design concepts that are incorporated into a mechanism. An ancillary lesson is to consider and evaluate additional concepts when moderate or high risks are identified in the base design. 35 | Document | AMS_2006.pdf | 49 |
References 1. D.W. Plummer and W.H. Greenwood (1993). A Primer on Unique Signal Stronglinks, SAND93-0951, Sandia National Laboratories, Albuquerque, NM. 2. J.E. Shigley and C.R. Mischke (1 989). Mechanical Engineering Design, 5'h edition, McGraw-Hill, New York City, NY. 36 | Document | AMS_2006.pdf | 50 |
Failure of Harmonic Gears During Verification of a Two-Axis Gimbal for the Mars Reconnaissance Orbiter Spacecraft Michael R. Johnson', Russ Gehling**and Ray Head** Abstract The Mars Reconnaissance Orbiter (MRO) spacecraft has three two-axis gimbal assemblies that support and move the High Gain Antenna and two solar array wings. The gimbal assemblies are required to move almost continuously throughout the mission's seven-year lifetime, requiring a large number of output revolutions for each actuator in the gimbal assemblies. The actuator for each of the six axes consists of a two-phase brushless dc motor with a direct drive to the wave generator of a size-32 cup-type harmonic gear. During life testing of an actuator assembly, the harmonic gear teeth failed completely, leaving the size-32 harmonic gear with a maximum output torque capability less than 10% of its design capability. The investigation that followed the failure revealed limitations of the heritage material choices that were made for the harmonic gear components that had passed similar life requirements on several previous programs. Additionally, the methods used to increase the stiffness of a standard harmonic gear component set, while accepted practice for harmonic gears, is limited in its range. The stiffness of harmonic gear assemblies can be increased up to a maximum stiffness point that, if exceeded, compromises the reliability of the gear components for long life applications. Introduction The Mars Reconnaissance Orbiter Mission During its two-year primary science mission, the Mars Reconnaissance Orbiter will conduct eight different science investigations at Mars. The investigations are functionally divided into three purposes: global mapping, regional surveying, and high-resolution targeting of specific spots on the surface. This detailed mapping of the surface of Mars will provide future landed missions with the high resolution data required to land safely in a desired area. The instruments on board the MRO spacecraft consist of five types: cameras, a spectrometer, a radiometer, a radar, and engineering. Refer to Figure 1 for an overall view of the MRO spacecraft's science deck. Cameras HiRlSE (High Resolution Imaging Science Experiment) This visible-camera can reveal small-scale objects in the debris blankets of mysterious gullies and details of geologic structure of canyons, craters, and layered deposits. CTX (Context Camera) This camera will provide wide area views to help provide a context for high-resolution analysis of key spots on Mars provided by HiRlSE and CRISM. MARC1 (Mars Color Imager) This weather camera will monitor clouds and dust storms. Spectrometer CRISM (Compact Reconnaissance Imaging Spectrometer for Mars) This instrument splits visible and near-infrared light of its images into hundreds of "colors" that identify minerals, especially those likely formed in the presence of water, in surface areas on Mars not much bigger than a football field. ' Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA ** Lockheed Martin Space Systems, Denver, CO Proceedings of the 38 Aerospace Mechanisms Symposium, Langley Research Center, May 17- 19,2006 37 | Document | AMS_2006.pdf | 51 |
Radiometer MCS (Mars Climate Sounder) This atmospheric profiler will detect vertical variations of temperature, dust, and water vapor concentrations in the Martian atmosphere. Radar SHARAD (Shallow Radar) This sounding radar will probe beneath the Martian surface to see if water ice is present at depths greater than one meter. Enaineerinq The engineering instruments facilitate spacecraft navigation and communications. High Gain Antenna *, II J VADTR Tnstniment Deck Once the science phase is completed (two years after the mapping orbit is established), the MRO mission enters a second phase, communications relay. In this phase, the communication equipment on-board MRO will be used as a communications relay between the Earth and landed crafts on Mars that may not have sufficient radio power to communicate directly with Earth on their own. This capability allows landed crafts to use smaller antennas with reduced mass, improving the lander’s science complement potential. Due to the mapping nature of the mission, the instrument deck of the spacecraft must always be facing the surface of Mars. Additional pointing requirements include maintaining sun pointing of the solar panels and keeping the High Gain Antenna Earth pointed for communication purposes. The solution to this extreme panel and antenna pointing choreography was to put a two-axis gimbal at each of the appendages: two solar array wings and one High Gain Antenna. The path from Earth to Mars orbit and mapping of the surface consists of launch, cruise, orbit insertion, aerobraking, and mapping phases. The spacecraft configuration of the solar array and the High Gain Antenna are different for each of these phases. During launch, the solar array wings are folded in half and the High Gain Antenna is positioned directly over the spacecraft bus to fit into the launch vehicle fairing (Figure 2). Once MRO is launched and in the cruise phase of the mission, the solar array and High Gain Antenna are pointing in roughly the same direction to capture sunlight and communicate with Earth (Figure 3). For Mars orbit Insertion and aerobraking, the appendages are moved slightly from the cruise configuration to produce an aerodynamically stable configuration. The High Gain Antenna and the solar array wings are the predominant source of atmospheric drag on the spacecraft and must be positioned to keep the spacecraft stable throughout the aerobraking maneuvers, which last for about 6 months (see Figure 4). For the mapping phase of the mission, the solar array wings and High Gain Antenna are almost 38 | Document | AMS_2006.pdf | 52 |
continuously articulated so the wings remain sun pointed and the High Gain Antenna maintains a lock on Earth. This continuous motion must be performed while the spacecraft maintains precise pointing for high resolution imaging and high-speed data transmission to Earth (see Figure 1 for a mapping configuration). Because of the stringent pointing stability requirements, the gimbals were required to be exceedingly smooth and quiet. The different phases for the MRO appendages lead to a large range of requirements for the two-axis gimbal assemblies. The gimbals must be capable of carrying launch loads through their output bearings for the launch phase. The cruise phase is very benign with minimal load on the gimbals. The orbit insertion and aerobraking phases put a significant load on the output bearings and gears, since the loads have a significant component in the backdriving direction for the gimbal actuators. Once the spacecraft is in the mapping phase, the high resolution capability of the instruments on board require that the gimbal assemblies do not produce any significant disturbance to the spacecraft platform while they are continuously scanning to maintain the required pointing of the attached appendage. The gimbals must withstand all of these load combinations and still maintain extreme pointing accuracy and smooth operation once at Mars. In addition to the smooth motion, the lowest structural frequency in the mapping configuration is determined by the natural frequency of the deployed appendages. A major contributor to the frequency of the deployed appendage is the gimbal actuator output stiffness for each axis. Figure 2. MRO Launch Configuration Figure 3. MRO Cruise Configuration Figure 4. MRO Orbit Insertion & Aerobraking Configuration Gimbal Actuator Confiauration Each two-axis gimbal consists of two identical gimbal actuators, structurally connected with application specific components. The core of the actuator is a 130-mm diameter two-phase brushless dc motor with a large number of poles in order to maintain smooth rotor velocity. The brushless motor is commutated using a resolver with the same number of poles as the motor to simplify the commutation logic. The motor directly drives the wave generator of the output harmonic gear component set through a bellows coupling. The bellows coupling was used to minimize speed ripple that would cause disturbances while operating. The harmonic gear is a size-32 HDC, standard-cup-type unit. The flexspline is mounted to the actuator housing and provides the torque reaction mount. The circular spline is mounted in a pair of preloaded angular-contact ball bearings. A multi-speed output resolver is installed between the angular-contact bearings for a compact assembly that measures the output position to the accuracy required for the MRO mission pointing. This arrangement of drive components provides a zero-backlash actuator with minimal mechanically generated disturbance sources and applies all of the externally generated loads directly to the harmonic gear teeth. A photo of a completed flight gimbal actuator assembly is shown in Figure 5. 39 | Document | AMS_2006.pdf | 53 |
The need for the harmonic gear to take the external loads led to the selection of the size of the harmonic gear in concert with the required output stiffness. The available volume and mass allocation for the actuators was minimal on the MRO spacecraft, requiring use of the smallest gears possible to achieve the required load and stiffness capability. The loads from the Mars orbit insertion and aerobraking phases needed to be taken into account along with all other applied loads during the mission. The mass of the solar array wings and the antenna assembly loading the output gear teeth in the acceleration environment of orbit insertion was one source of applied load. The force from aerodynamic loading on the large area array and antenna during aerobraking was another source of loading. These conditions together defined the magnitude of the loads that would be applied to the harmonic gear output teeth in flight. It was determined that the applied loads could be carried with appropriate margins by a size-32, cup-type harmonic gear component set. The output torsional stiffness of the actuator axes affects the spacecraft dynamics. The spacecraft sensitivity to jitter disturbances and the attitude control system authority dictated a minimum natural frequency for the deployed solar array wings and High Gain Antenna. Since the gimbal actuator was a significant contributor to the overall appendage stiffness and the harmonic gear teeth were the load reacting devices, an output stiffness was required of the chosen harmonic gear that exceeded its standard specification significantly. The magnitude of this stiff ness increase for the selected size of harmonic gear was within the range of experience for this type of application on other programs. Since there was no significant difference in the stiffness requirement when compared to other heritage programs, this was not considered a significant risk to the program. The selection of the harmonic gear materials from the available set involved comparing the MRO requirements with those of previous flight programs in order to maintain as much heritage as possible. A fundamental tenet for this and many programs was to use only corrosion resistant materials in all space mechanisms as is commonly done in the medical, semiconductor, and food processing industries. The materials of the various pieces of the MRO harmonic gear component set are listed in Table 1. This combination of materials had been used successfully in several programs with stiff ness and total lifetime revolutions requirements that were similar to MRO. Table 2 lists the other heritage applications of the same material combination with similar functional requirements. To maximize the life cycle capability of all 40 | Document | AMS_2006.pdf | 54 |
of the actuator components, Penzane 2001-3PbNp Oil was utilized in the motor bearings and Rheolube 2004 grease was chosen for the lubricant throughout the harmonic gear. Other flight applications with the same material combinations for the harmonic gear components had also used Braycote@ because of a much lower operating temperature requirement than MRO. Since the required operating temperature range did not necessitate the use of bray oil or grease, the Penzane lubricant family was chosen since it tends to demonstrate more life capability over the Bray series when the operating temperatures are moderate. Table 1. Commercially Available Materials for Note: The Melonite process is a Nitrocarburizing case hardening per SAE-AMS-2753B Hardware Verification The motor and gearbox assembly was the same for all of the six axis applications on MRO, allowing one life verification 6 be performed on the worst case loaded design that would encompass all of theother five axis applications on MRO. The life verification program for these actuators consisted of operating a harmonic drive assembly alone in ambient environment with standard lubrication, followed by operation of an assembled flight-like actuator assembly in a vacuum with thermal cycling and a representative wiring harness for loading of the output gear. The lone harmonic drive gear assembly was operated in a standard gear test fixture used at the harmonic drive vendor. This was performed at ambient pressure and temperature in a bath of low viscosity commercial oil at an input speed of 1750 RPM and with an applied load on the output that matched the cable loading from harnesses. The wave generator was driven for 32 million revolutions with no sign of unusual wear or failure of the bearing or the gear teeth. This was done without incident to a rotational life of five times the flight requirement of 6.2 million input revolutions, indicating harmonic gear rotational lifetime was a low risk. Next an Engineering Development Unit (EDU) actuator was operated in a flight-like configuration. The EDU actuator used the flight housings with output hardware that supported the High Gain Antenna cable management system. The cable management system was incorporated to cycle the cabling as well as provide flight like output loading to the harmonic gear. The High Gain Antenna application was chosen because it has the largest number of cables across any of the gimbal axes. The EDU actuator was operated over a total output angle of 340 degrees. The EDU actuator was run in a thermal/vacuum environment, with the temperature slowly cycled from -25°C to +40°C at the rate of one thermal cycle every 18 hours. The actuator was driven at a motor rate of 125 RPM for approximately 3.8 million input revolutions, and then run at 65 RPM for the remainder of the time. The EDU motor was driven with an industrial stepper motor driver, severely limiting visibility into the performance of the actuator. The stepper motor driver was used for this operation for several reasons: flight actuator drivers were not available for the start of the running, the motor is a two-phase brushless dc (not three phase), and the rotary life was believed to be low risk so the limited visibility was not considered to be significant. The EDU life setup is shown in Figure 6. 41 | Document | AMS_2006.pdf | 55 |
Figure 6. Engineering Development Unit Configuration in Vacuum Chamber Flexsp I i n e Table 2. Materials and Surface Finishes of Heritage Hardware from Previous Programs with Similar Stiffness and Life requirements as the MRO Applications Lubricant Program (Harmonic Drive Size) MRO (size 32) 15-5 PH H1150 Melonite 15-5 PH H1150 Melonite 15-5 PH H1075 Melonite 15-5PH H1075 (size 25) Penzane 2001 -3Pb Rheolube 2004 Braycote 602 Bray 81 52 oil Penzane 2001 -3Pb Rheolube 2004 Penzane 2001 -3Pb Rheolube Circular Spline Program #5 (size 32) 15-5PH H1075 2004 15-5 PH H1075 15-5 PH H1150 Rheolube 2000 with Melonite 3% lead Napthenate 15-5 PH H1075 15-5PH H1150 15-5 PH H1150 Melonite (sbe 40) I Melonite I I 2001 -3Pb Rheolube 42 | Document | AMS_2006.pdf | 56 |
The EDU actuator operated up to 6.1 million input revolutions, when the output telemetry indicated the actuator was not following the input signals properly. Later analysis revealed there were indications of improper operation as early as 4 million input revolutions that were not diagnosed due to limitations in the test setup with the stepper motor driver. The EDU actuator was removed from the chamber, disassembled, and inspected. Figures 7, 8, and 9 show the condition of the harmonic gear teeth at this inspection. Note that the tooth profile was completely obliterated and damaged across the entire width of the flexspline teeth and most of the width of the circular spline. I i Intacl Figure 7. EDU Circular Spline Life Damage Figure 8. EDU Flexspline Life Damage Figure 9. Magnification of EDU Flexspline Life Test Damage 43 | Document | AMS_2006.pdf | 57 |
Once the dramatic failure of the harmonic gear assembly was seen, a detailed review of the manufacturer’s documents uncovered that there were some problems with the Melonite coating on the group of flexsplines that included the life unit, labeled Lot B. Two additional Melonite processing groups of flexsplines had been received at the time of the failure, labeled Lot C and Lot D. Operation of these units was started to determine if the Melonite processing on the EDU life test unit was the source of the failure. A harmonic assembly from each of the two remaining Melonite process groups was placed in a harmonic gear fixture (not in the gimbal actuator) and operated in vacuum with thermal cycling over a 12-hour period, an output load, and an input speed of 130 RPM. Additionally, one of the two units was tested using the same Penzane family of lubricants as the EDU and the other was tested using Braycote@ 602. Both assemblies failed at approximately two million input revolutions. Figure 10 shows the failed splines that were tested with the Penzane family of lubricants. Figure 11 shows the failed splines that were tested with the Braycote@ 602 lubricant. A significant result of these two tests was that the lubricant type made no difference at all, with both assemblies failing at nearly the same number of revolutions with the same type and level of damage. I Figure 10. Circular Spline and Flexspline from Harmonic Only Test #1 Tested with Penzane Figure 11. Circular Spline and Flexspline from Harmonic Only Test #I2 Tested with Braycote@ 602 44 | Document | AMS_2006.pdf | 58 |
Since the lubricant type made absolutely no difference in the life capability of the harmonic drive gear teeth, it was clear the problem was related directly to the material combination of the components and their internal stress level. The appearance of the failure surfaces gave the impression of a possible galling condition at work, but it was not clear if galling was the initiator of the failure or a consequence of the damage once the failure had been initiated. Rough mathematical analyses were performed to estimate the contact stress at the harmonic gear teeth from the preload and under the applied load in the operation. The constant external load in the harmonic fixture runs represented the worst-case load from the cable management system and was responsible for a roughly 40% increase in tooth contact stresses over the preload. The estimated contact stresses in the EDU life test were in the realm of 750 MPa. The galling threshold listed for a 15-5 PH stainless steel contact pair is around 14 MPa. This indicated that the contact stresses compared to the galling threshold for the selected materials was a strong candidate for the cause of the failures. To minimize the schedule time to a solution, the next group of tests used Nitronic 60 (another available harmonic gear material) with a listed galling threshold value greater than 345 MPa. Nitronic 60 was identified as a candidate in addition to other standard commercial materials for the harmonic gear assemblies, like nodular iron. Also, the output stiffness of the assembly was reduced so the internal harmonic gear preload (and with it the internal tooth stresses) could be reduced as well. At this point in the project schedule, the flight solar array panels had been fabricated and their stiffness was measured, allowing reduction of the stiffness margin for these panels in the MRO spacecraft stability analysis. This made it possible to reduce the gimbal actuator stiffness requirement significantly. To investigate these issues within the remaining program schedule, three readily available harmonic drive assemblies were procured with different material combinations and tested on the harmonic gear fixture. The material combinations consisted of a unit with a nodular iron circular spline and an E4340 flexspline. The second unit had a Nitronic 60 circular spline and an E4340 flexspline. The third assembly was composed of a Nitronic 60 circular spline with an E4340 flexspline processed with a Melonite surface. Additionally, the internal preload of the harmonic gear (to obtain the new required output stiffness) was reduced in order to lower the gear tooth internal stresses. Table 3 lists all of the life test units and the material and lubrication configuration for each. F Table 3. Selected Harmonic Gear Material and Lubricant Configurations te: LM stands for Lockheed Martin 45 | Document | AMS_2006.pdf | 59 |
The test units for LM Harmonic Assemblies #3, #4, and #5 were tested in the harmonic drive test fixture, in the environment, with the applied load and input speed listed in Table 4. The stiffness of the units was measured at the start of the testing as shown in column 2 of Table 4. The life tests already discussed are listed in Table 4 for completeness. Test Unit Table 4. Life Unit Stiff ness, Environments, Conditions, and Revolutions at Failure Initial Stiffness with 45/140 N.m Applied Approx. Applied Torque Load Input Revolutions (N-mlrad) Environment on Output Speed to failure vendor #I (in actuator) LM Harmonic Assembly #1 (same unit as Harmonic drive vendor Unit #1) LM Harmonic Assembly #2 LM Harmonic Assembly #3 LM Harmonic Assembly #4 Approx. 68,000 > 56,000 Approx. 68,000 69,000/75,000 56,000 55,600/60300 56,500 (28 N.m) 43,800 (28 N.m) 54,600/61,000 43,800 (28 N.m) 37,300 to 46,300 (28 N.m) Ambient Vacuum, -25"C/+4OoC 18 hour cycles Vacuum, -1 5"C/+40°C 12 hour cycles Vacuum, -1 5"C/+40°C 12 hour cycles Vacuum, -1 5"C/+4OoC 12 hour cycles First 3M revs Vacuum, -1 5"C/+40°C 12 & 24 hour cycles Next 3M revs Vacuum, +23"C constant Next 8M revs Vacuum, -1 0°C constant Vacuum, -lO"C/O"C constant Vacuum, -1 O"C/+l 0°C constant 11.3 N.m HGA Cable Harness 11.3 N.m 11.3 Nom 11.3 N.m 11.3 N.m 11 -3 N.m HGA Cable Harness 1750 RPM 125 RPM & 65 RPM 130 RPM 140 RPM 140 RPM 140 RPM 140 RPM 60 RPM No Failure 4M 2M 2M 4M No Failure Q 14M No Failure Q 9.1 M No Failure Q 13M Key results of the above are as follows: 0 The high input speed of the harmonic gear only operation performed at the harmonic drive vendor allows the lubricant to support high contact stresses that would otherwise result in complete failure of the gear teeth. The difference from the EDU Assembly #1 unit and the LM Harmonic Assembly #l was the Melonite lot and coating details. The results of the LM Harmonic Assembly Unit #1 showed that the Melonite coating was not involved in the failure. The applied output load, while small when compared to the maximum torque capability of the harmonic gear, made a difference in the revolution life to failure. Note that the EDU life unit exhibited failure around 4 million revolutions, while assemblies #1 and #2, with a constant applied load, failed at 2 million revolutions. This difference could be due to the EDU assembly #1 using a cable wrap harness, with a variable load depending on output position, compared to a constant load. 0 0 46 | Document | AMS_2006.pdf | 60 |
LM Harmonic Assemblies #1 and #2, with failure at a similar number of revolutions using very different lubricants, indicated the lubricant was not a significant player in the failure mechanism. This eliminated the lubricant as a variable in further failure investigation. The LM Harmonic Assembly #3, with nodular iron and a reduced preload, had a longer life to failure than the stainless steel, supporting the galling hypothesis. Note the regions of damage in Figure 12. The LM Harmonic Assemblies #4 and #5 used Nitronic 60 for the circular spline, the highest galling threshold material that could be obtained in a harmonic gear assembly. The late date of the testing permitted a reduction of the output stiffness to two-thirds of the initial values used for the gimbal actuators. The stiffness of the unit in Harmonic Assembly #4 was reduced as the operation progressed and a method of setting the harmonic gear preload was established. The result of the run was no failure at all with some minor wear of the harmonic gear teeth, as shown in Figure 13. The schedule dictated that the first successful combination be used, so the final material selection was a Nitronic 60 circular spline and a flexspline of E4340 with no additional surface processing. Note that schedule dictated changing more than one variable at a time. The setting of the internal preload of the harmonic gear assembly during assembly at the harmonic drive vendor was critical in achieving the required output stiff ness without compromising the reliability of the harmonic gear assembly. 0 0 Figure 12. Circular Spline and Flexspline fromHarmonic Only Tex#3 .A n.ri+h +ha ,-ir,.nalmr rnlins A =:, ('. Figure 13. Circular Spline and Flexspline trom harmonic vnty I esr lf4 Note the slight wear region on the circular spline showing where the flexspline was engaged with it. 47 | Document | AMS_2006.pdf | 61 |
Harmonic Gear Assemblv Internal Preload Settinq The cup type harmonic gear assembly has zero backlash due to the angle of approach, along the rotation axis, of the flexspline teeth relative to the circular spline teeth (Figure 14). This arrangement leads to a two-sloped stiffness curve of the output of a harmonic gear. As torque is applied to the gear, the cup flexes to allow more of the width of the teeth to engage with the circular spline. As more tooth area is engaged, more load sharing occurs and the stiffness increases. Once the angle between the teeth of the splines is reduced to near zero, the stiffness has reached its maximum value. As the torque is increased from this point, the stiffness is related to deflection of the individual teeth and the change in roundness of the housing and circular spline. Figure 15 shows a stiffness curve for a typical harmonic gear assembly prior to increasing the internal preload for stiff ness improvement. Flexspline tooth angle - Cup flexes to allow more tooth face engagement / /Flexspline Circular Spline tooth angle Figure 14. Harmonic Drive Cross Section Showing Significant Source of Variable Stiffness Note the largest stresses on flexspline teeth occur at the point of initial contact, the open end of the cup. Drawing courtesy of Harmonic Drive, LLC Increasing the output stiffness of a harmonic drive assembly involves increasing the diameter of the wave generator in the area where it forces contact between the teeth of the two splines. This is accomplished by using a different wave generator plug with a larger major diameter of the oval. As the diameter is increased, the flexspline cup is deflected in the direction of engaging more of the face of the teeth. This has the same effect as increasing the torque on a nominal unit in the low stiffness region. As the wave generator plug size is increased, the low stiffness region gets smaller. This trend continues until the stiffness curve is essentially straight. At this point, a further increase in the diameter of the wave generator plug will increase the overall gear assembly stiffness and significantly increase the internal tooth stresses. Figures 16 through 19 show the how the shape of the stiffness curve changes with different wave generator plugs in a harmonic drive assembly. The low stiffness region is very evident in Figure 16. A larger plug, after insertion into the wave generator bearing may nearly eliminate the low stiffness region, as seen in Figure 17. Figures 18 and 19 show the next two larger size wave generator plugs, without a significant change in the shape of the stiffness curve. The desired operating point for the flight plug is smallest wave generator plug that exhibits a fairly straight curve. If there is any uncertainty between units, the smaller one would always be installed to guarantee that the gear teeth were not being jammed together with high, and unknown, internal stresses. The wave generator plug used in this example would be Figure 17’s. 48 | Document | AMS_2006.pdf | 62 |
Measured Initial Harmonic Drive Stiffness Rotation Angle Figure 15. Typical Harmonic Gear Assembly Stiffness Note two regions of stiffness: low near zero torque and larger at high torques. The following figures are from measured data on the fliaht harmonic aear assemblies: (Note: the following figures' axes are rotated relative to Figure 15) .,~ __-......._.I *_. . :, r. . ,. . .. . . . . . . . . . . . .. . , .. I Applied Torque Applied Torque Figure 16. Stiffness Curve Showing Low Stiffness Region Figure 17. Stiffness Curve with Minimal Low Stiffness Region 49 | Document | AMS_2006.pdf | 63 |
ji Applied Torque Applied Torque Figure 18. Next Larger Plug Size Over Figure 17 Figure 19. Next Larger Plug Size Showing Showing a Small Change in Shape Showing Little Change Conclusions and Lessons Learned The most significant result of the failure investigation was determining that the internal stresses due to the preload and the cable harness loading caused the premature failure of the EDU harmonic gear assembly. In order to maintain reliability for long life applications, increasing the wave generator plug size (a service provided by the harmonic drive vendor) is an acceptable technique up to the point where the low stiffness region is eliminated. This is the maximum stiff ness enhancement achievable without compromising the reliability of the gear assembly for long life. Any further increase in the wave generator plug size will increase the stiffness at the cost of reducing the life of the unit. For minimal life applications, increasing the stiffness beyond this point may still be acceptable. Stiffness enhanced harmonic gears are very sensitive to the externally applied load and test environment. Since the failure mode is galling, the presence of any gas (nitrogen, for example) severely compromises the test results. Life capabilities from previous heritage programs had been successful and so the initial gear material for the MRO gimbal actuators was considered acceptable and robust. However, some of the heritage operation had been performed in nitrogen, instead of vacuum. When enhanced stiffness is required in a harmonic gear application and it is not being used in a preload configuration represented by Figure 16, the situation is sensitive to possible galling of the harmonic gear teeth. Performing harmonic gear component operation at loads above the planned level to increase the tooth contact stresses should be considered. This will demonstrate if internal stress margin exists in the hardware. Also note that running a harmonic gear at a high input speed to reduce the operating time is not adequate. The high- speed condition may function with no incidents while low speed operation may catastrophically quit functioning. Finally, operating a unit at nominal contact stress levels to a larger number of revolutions than planned is a necessary, but not complete, margin demonstration program. A catastrophic failure may be lurking just a few megapascals away from the nominal value. References Harmonic Drive, LLC web site, www.harmonic-drive.com/suppor-t/principals.htm, description of harmonic gear assembly operation Acknowledgements This work was performed at Lockheed Martin Space Systems, Denver, Colorado under a spacecraft system contract to the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the National Aeronautics and Space Administration. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not constitute or imply its endorsement by the United States Government, the Jet Propulsion Laboratory, Pasadena, California, or Lockheed Martin Space Systems, Denver, Colorado. 50 | Document | AMS_2006.pdf | 64 |
Stacer Driven Deployment: The Stereo Impact Boom Robert Ullrich’, Jeremy McCauley*, Paul Turin*, Ken McKee* and Bill Donokowski* Abstract The Impact’ Booms carry 3 scientific instruments each on the twin NASA Stereo2 satellites. When stowed for launch the boom is 1.54 m in length, when deployed the boom extends to 5.80 m. The boom consists of 5 concentric graphite epoxy (GlE) tubes extended by the spring force of a Stacer. The Stacer is fabricated from a flat strip of Elgiloy spring material, rolled with a constant diameter and fixed helix angle. It supplies the motive force for deployment, and requires no external power once released. The deployed boom exhibits excellent rigidity, the natural frequency first mode occurring at 1.96 Hz. Discussed is the implementation of a Stacer to deploy the 5 segment telescoping boom and some of the activities performed during its design, qualification and testing. Mission Introduction The NASA Stereo mission consists of twin, three-axis stabilized satellites orbiting and viewing the Sun in the plane of the ecliptic at -1 AU. Spacecraft A (Ahead) will be sent into an Earth preceding path with an Earth-Sun-Spacecraft angle increasing at e rate of 22’ per year, while Spacecraft B (Behind) is sent into an Earth lagging orbit, also at a rate of 22 per year. The imagers on board will yield true ‘stereoscopic’ views of coronal mass ejections, while other instruments perform concurrent in-situ measurements of a large portion of the electro-magnetic spectrum. The telescoping boom was conceived to interface 3 instruments from the Impact suite: the Magnetometer (Mag), the Solar Wind Electron Analyzer (SWEA), and the Supra-Thermal Electron - Downward looking instrument (STE-D) to the Stereo spacecraft. The program requirements demanded a new concept, as existing hardware was deemed too expensive or unsuitable. The boom was initially developed via three ‘proof of concept’ models for the tube locking mechanism, and a final mock up using the Stacer to deploy four concentric, telescoping graphitejepoxy tubes from the center of the fixed 5th tube. An engineering model (EM) was then built to verify end to end design via qualification testing. The challenge for this mechanism was demonstrating that the design met the requisite GEVS SE3 force (torque) margin. Two flight models (FMs) were then produced, with the EM being refurbished as a flight spare. The FMs are currently mounted to the spacecraft and mission I & T is progressing. Launch is scheduled for May 2006 from Cape Kennedy on a Delta II. Figure 1 Magnetometer, SWEA and STE- D mounted on the end of the Istowed) Figure 2. Magnetometer ’ Space Sciences Lab, University of California, Berkeley, CA ’ In-situ Measurements of Particles And Coronal mass ejection Transients * Solar - TErrestrial RElations Observatory ’ See References Section Proceedings of the 3dh Aerospace Mechanisms Symposium, Langley Research Center, May 1 7- 19,2006 51 | Document | AMS_2006.pdf | 65 |
The Science Flowdown Requirements The requirements for the Impact boom were based on the scientific needs of the three instruments mounted to it: the Mag (built at Goddard Space Flight Center); the SWEA, supplied by CESR, Toulouse, FR; and the STE-D, supplied by UCB - SSL. The magnetometer for this mission is very sensitive: the magnetic (B) field at 1 Au heliocentric orbit is - 3 to 4 orders of magnitude smaller than near Earth. This slight field strength was a driver for the EMVEMC design for the spacecraft and the devices near it. To avoid 'sensing' the spacecraft, the Mag needed to be 3 meters away from it. This requirement set the minimum boom length. To ensure a low magnetic signature from the boom assembly, no ferritic alloys were allowed for its construction. Titanium screws were used to mount the magnetometer to its tray on the 4'h tube element and the tray itself is made of carbon impregnated PEEK, a high-strength, conductive engineering plastic. To lower Mag exposure to any eddy currents present in the harness or structure of the boom, the mounting tray offset the Mag 200 mm from the nearest metal on the tube. Additionally there were not to be any other instruments closer than 1 m to the Mag. To allow accurate inter-experiment correlation of data, the angular alignment accuracy and repeatability requirement has an allowable deviation of <0.88' (52.5 arcmin) between the Impact boom mounting feet and the magnetometer housing from the stowed condition to the deployed state, for the two axes that form the mounting plane of the Mag. The second experiment mounted on the Impact boom is the Solar Wind Electron Analyzer (SWEA, supplied by CESR, Toulouse, FR). The SWEA has two variably charged hemispheric surfaces that attract electrons into an anode assembly which counts them as they impinge on to it. The SWEA would have a limited field of view when mounted directly to the spacecraft deck, hampering its ability to characterize the electron regime in the volume around it. Proximity to the spacecraft also causes deflections of the electrons due to the almost unavoidable static fields that develop near the spacecraft surface. Since this effect is difficult to model, a better solution was found. The implementation of the SWEA on the Stereo mission is extremely good: it is mounted on the extreme end of the boom, allowing a full 2n radians x 135O field of view (FOV). The demands of the SWEA for mounting to the boom are not complex: power lines, command lines supplied to it, data return lines from it. Figure 3- SWEA 1 STE The final instrument is the STE, a Supra-Thermal Electron detector, mounted on the side of the SWEA pedestal. It needs a clear 80' x 80° field of view looking along the plane of the ecliptic at a 45' angle (aligned with the Parker spiral), and to stay at - -4O'C. The fixed base of the Impact boom is pointed towards the Sun for all science activities, and the only off- points scheduled are for momentum dumping. The boom deploys away from the Sun, so that there is minimal solar input to the boom suite, giving very low operational temperatures for the instruments. Thermal control was a large concern for the instrumenters. 52 | Document | AMS_2006.pdf | 66 |
e Figure 4. Solid Model Section View of Stowed Impact Boom Impact Boom Mechanism The Impact Boom consists of five concentric Gr/E Tubes, ranging from 50 mm to 210 mm in diameter, with a pair of aluminum rings bonded to each end. Each ring pair contains three lock pins, pointing outward at the Sun-ward end, and inward pointing at the release mechanism end, and three sockets, in their corresponding orientations. The pins are spring loaded, and have rollers mounted in their tips. When released, the Stacer spring extends the tubes until the end of travel where the pins drop into sockets, locking the assembly rigid. The mounting feet are integral to the outermost rings on the 0210-mm tube. There is a spool for the electrical harness while stowed, a flyweight brake to govern the deployment speed, a shape memory alloy release mechanism with pretensioning adjustment, deployment assist rods and kick springs to initiate the deployment. Combs at each end hold the tubes in alignment prior to deployment, and during vibration/launch. A provision for individually adjusting the combs to remove any play in the stowed tubes due to fabrication tolerances is also provided. The design was performed in SolidworksTM, utilizing its 3-dimensional solid modeling and multiple configuration capabilities. Tube Details q- .I I The use of telescoping concentric tubes is not a new idea. Each telescope application brings its own set of challenges however. For the Impact Boom, the tubes needed loose tolerances on their cylindricity callout to allow for simple tube manufacture. The deployment / locking scheme required compliance regarding the inter-tube fit since there is a relatively low force available from the Stacer. The boom needed to be very rigid when deployed, so locking pins were utilized at end of travel, rather than relying on spring force to hold them in place. The tubes have three longitudinal concave grooves equispaced about their circumference, running their length, with a precise profile that doesn’t jam the rollers. These grooves kept the pins on track to be aligned with the sockets at the far end of the tube. The tubes are a five-layer Gr/E composite ~ designed to be quasi- isotropic: three 0 -90 layers interleaved with two 45’ layers of 0.1 2-mm woven epoxy pre-impregnated material (Fiberite Hy-E 1034C prepreg). The tubes were fabricated by Vision Composites of Signal Hill, California on internal mandrels, with a slight taper to enhance ease of extraction after cure. The cure regimen was specified to be ‘dry’: the ratio of epoxy to carbon filament was held to a minimum to ensure low surface resistance. This was achieved by using a higher autoclave pressure with a slower ‘warm up to cure’ temperature ramp and a thick layer of absorbent over the bleeder sheet. The process determination was somewhat lengthy, however, the final result met requirements. ’ Figure 5. Tube End Detail Figure6 Partially Deployed Tubes 53 | Document | AMS_2006.pdf | 67 |
At both ends of each tube there are inner and outer interlocking aluminum rings that ‘sandwich’, and are bonded to, the Gr/E. After an extensive search, Loctite Hysol 9309NA was used to form the bond. The thermal environment for the boom is rather severe: it will be in the shadow of the spacecraft for all of the science activity; thermal analysis estimates put the operating low temperature at 30K. There is very little data for epoxies at this temperature, so we performed an FEA for the bond between the aluminum and the Gr/E to establish what parameters minimized the stresses in the glue, the weakest part. Several cases for the glue design were examined: thickness of the bond, edge conditions of the bond and effects of the aluminum ring thickness (internal and external) on the joint stresses. Optimized, the glue line was 1 vo. mue. 1 vo. YUU Figure 7. FEA of Aluminum Ring - Glue Line - Gr/E - Glue Line - Aluminum Ring chosen to be -0.4-mm thick, with a fillet onto a tapered aluminum ring edge. This best case predicted a stress value of 11OvMPA (16 k~i)~, which exceeded the glue published maximum stress value of 38 MPa (5.5 ksi)’. There was concern that the joint would not be sound after exposure to the thermal gradient, so an actual test was needed. A test GR/E tube / aluminum ring assembly was fabricated, with a large cantilever mass attached at the extremity providing -2X expected loads for this test. This test ‘tube’ assembly was installed in a cryogenic liquid helium chamber, which was then installed into a cryogenic liquid nitrogen chamber in turn was placed inside of a refrigerated chest. We performed a multiple cycle thermal test (in a dry air environment), utilizing four candidate gl!es. The cold temperature was set to 25K (-248 C), and nominally, with no crazing or cracking and was warm was 150K (-1 23’C). The 9309 performed accepted for use. As a side note, only one of the tested epoxies exhibited any signs of thermal distress. As a hedge against exceeding our thermal predicts, small solar absorbers were attached to the two joints in the mid boom. These little flags raise the expected temperature by -5 degrees, buying some margin for the assembly (Figure 6). I Figure 8’ Epoxy Thermal Test Chamber Besuner Consulting, Madera, CA 93638 Loctite Hysol Applications Note, April 101, Loctite Aerospace, Bay Point, CA 94565 4 54 | Document | AMS_2006.pdf | 68 |
Lock Din details Once the rings are bonded to the tubes, the locking pins / rollers are installed into precision radial bores in these inner and outer rings. When the segment locks, the pins are pushed into the sockets by custom wound torsion springs, two for each pin. The ‘arms’ of the springs also provide alignment for the pins, keeping the roller in the groove. The pin / roller combination allows any tube dimensional change to be insignificant during the deployment, as the spring compliance takes up any bumps or dips. When the tube section reaches end of travel, the locking pins are pushed into tapered holes, causing the tubes to become rigid with respect to each other. Guide ramps are provided at the end of travel to ensure that the pins are aligned with the sockets. Each locking pin has a slight taper (i.e., <20° included angle) that fits into the corresponding tapered socket. This gives a ‘self-locking’ feature to the pin, offering increased rigidity and prevents the pins from backing out under slight vibrations. With six lock pins engaged per joint, three inward acting, three outward, the boom exhibits great rigidity, and offers redundancy in the event that one pin (or more, up to three maximum, as long as they are not in same ring) does not lock. Figure 9. 50-mm Tube End Lock Pins (EM) Rollers are fitted to the tips of the locking pins to minimize deployment drag when rolling in the tube’s grooves. Repeated deployments have shown no signs of wear to the tube or the rollers. To provide conductivity, the tapered portion of the pins and sockets were Alodined, while the sliding cylindrical parts were Type Ill black anodized to give good wear and low friction sliding properties. The Gr/E exhibits a low surface resistance too, enabling the boom to easily meet the surface resistance requirement of <lo8 ohms per square, throughout its stroke. The drag was measured to be 3.1 N on average for the assemblies. The main function of the rollers is to keep the tubes aligned during deployment, so that the pin engagement is virtually guaranteed at the end of stroke. Shape Memorv Allov Release (SMAR) details The SMAR uses the interesting phase change properties of a 50% titanium - 50% nickel alloy (trade named Nitinol initially) to provide the actuation of the Impact Boom. This device, pioneered by TiNi Aerospace‘ in cooperation with UCB-SSL, takes advantage of the -4% dimensional change in the drawn alloy wire when heated above its transformation temperature to let a ball detent assembly release a large spring loaded retracting pin. Since there was a large design load (50 Gs), >2.5 kN retraction force was needed. A force amplifier was added to the TiNi standard P50 (-200-N [50- Ib] pin puller). The force amplifier contains a stack of Belleville washers, preloaded and held by the P50 pin in another ball detent assembly, providing a final pull force exceeding 3 kN. When an electric current is passed though the Ti-Ni wire, it changes phase, elongates, releasing the primary pin, which then retracts and releases the main pin, which retracts with great force, allowing the Stacer to deploy. The main benefit of using an SMAR, aside from increased safety as no explosives are used, is that the flight unit can be tested over and over again (hundreds of cycles), and is simply resetable with a hand tool, with no temperature or time dependant constraints. I 1 a Flvweiaht Brake Figure 10. Shape After the SMAR has been triggered, kick springs push the tubes out of the combs and the deployment assist device pushes the Stacer out of the canister with a force of -90 N, giving the assembly a good initial velocity. The Stacer continues to provide force throughout the travel, so the deployment velocity would continue to increase until a balance between drag and push is achieved. This balance is never reached by the boom, so the deployment velocity reaches a ‘run-away’ condition rapidly, with the possible issues of lock pin shearing, ring-tube separation or other damage as consequences. As with every Stacer, a means to limit deployment velocity is incorporated. For the Impact Boom, a flyweight brake mechanism is attached to Memory Alloy Release Assembly TiNi Aerospace, San Leandro, CA 94577 6 55 | Document | AMS_2006.pdf | 69 |
P % the Stacer via a woven Dacron (parachute shock) cord. Similar to the device found on (old) dial telephones that prevented the dial from being rotated faster than an old telephone exchange could count, the flyweight brake supplies braking force proportional to the rotational speed of its weight assembly. If the force (speed) increases, the brake shoes are centripetally accelerated against the brake drum, increasing the braking force and slowing the rotational velocity. Over a wide range of forces, the brake typically can control the speed to *lo%. For the Boom, a deployment velocity of -0.5 m/s was chosen. This allows a certain momentum to build, but is slow enough to avoid shearing damage to the lock pins at the end of travel. Harness and SDOOI The power and electrical signals between the data processing unit and the instruments are carried by a cable routed down the center of the tubes, and is stowed on a spool for launch. This harness is a custom- fabricated conductor assembly consisting of seven coaxial cables and five twisted shielded pairs. Built of silver-plated copper with Gortex@’ dielectric, the harness is wound onto a bobbin when stowed and is pulled off when the boom deploys. This ‘straight through’ design provides greater signal strength, higher reliability and allows longer harness length as there are no slip rings or other connections between the data processing unit and the instruments. Care is taken to prewind the harness to avoid kinking or ‘birdnesting’ when stowing. I Figure ,,. Flyweight Brake 81 Harness Spool The Stacer The Stacer is a rolled, constant helical pitch, fixed diameter flat spring. The strip width, thickness, roll diameter, and helical pitch are selectable for each application, allowing each Stacer to be tailored for optimum properties. Stacers range in size from <1 m to >10 m in length, from 4 mm to 55 mm in diameter at the tip, and can provide extensive force from almost nothing to 200 N. Trade studies can balance mass versus length, force, etc. In the last 30 years, more than 650 units have been utilized in aerospace applications, from sounding rocket sensors to gravity gradient booms with large masses on the end. What makes the Impact Boom unique is the use of the Stacer as a spring ‘motor’ without using it as the structure or sensor surface. Most applications have the Stacer with the sensor(s) mounted directly on it, or the Stacer as the sensor, for example as an antenna (a total of six 6-m-long beryllium copper Stacers are used on the Stereo satellites for the Swaves experiment in this manner). I .- -I Figure 12. Picture of a Stacer To accommodate the Mag requirement of low magnetic signature for the boom, Elgiloy@ was selected as the spring material over the more traditional beryllium copper (Be-Cu). This alloy was chosen to minimize any eddy currents that could be developed between the SWEA / STE and the spacecraft. Originally invented in the late 1940’s in Elgin, IL for use in watches, it has been used for exacting Stacer applications several times. Its internal resistance is higher than copper, and cuts down the eddy currents accordingly. It has a higher modulus (E), and can provide greater force in the same physical volum7e as the Be-Cu. Elgiloy is a cobalt ‘super-alloy’, having an E -1 90 GPa and a yield strength of -1 600 MPa . 7 Matweb, http://www.matweb.com/search/SpecificMaterial.asp? = Elgiloy 56 | Document | AMS_2006.pdf | 70 |
At deployment, the formation of the Stacer starts with the initial coil winding out of the storage canister onto a cylindrical tip piece, which is slightly larger than the free coil diameter of the Stacer. Thus the Stacer grabs the tip piece tightly, and the subsequent coils ‘stack’ up on the prior, producing the characteristic spiral appearance. The typical helix angle provides for significant overlap, such that a section taken at any point along the Stacer would yield at least two thicknesses of strip material. Since the ‘outer’ layer of strip is rolled to the same diameter as the inner layer, the outer grips the inner with a force normal to the surface. So between layers, significant inter-coil friction exists and prevents inter-coil slipping for small disturbances. This gives the Stacer one of its more useful properties: it behaves as a thin walled tube for small displacements, with similar bending strength and stiffness. If a larger displacement occurs, the coils slip, dissipating the strain energy, serving as a friction damper. The damping ratio value is typically 5 - 15% for the non-slipping regime, and can reach 30 - 40% with the slipping. Of course, the displacement limit is buckling, as any tube would experience when taken beyond its yielding strength. As described, the motive force for the deployment is a Stacer. When compressed (stowed) it is a very compact package: it fits in a cylinder 050 X 130-mm long. When the Stacer is stowed, the strip is flexed into the canister, laying each coil inside its predecessor, and wound tightly to the outside of the can. When released, this stored strain energy is reclaimed, giving the motive effort needed to move the tubes along their path. The Stacer generates a higher force at the beginning of stroke, -46 N for this application, and the force curve dropped to 1 N at the end of stoke (this was an isolated minimum value obtained from one force test). The force that the Stacer provides is shown in the polynomial fit curve in Figure 13. Impact Boom Stacer Push Force 50 25 20 3 0 1 2 3 4 5 Deployed Length (m) Figure 13. Stacer Force 57 | Document | AMS_2006.pdf | 71 |
For the purposes of torque (force) margin analysis, the initial push is 45 N, final thrust 18 N (the lowest value obtained). This lower value was used to bound the design force available for deployment. The Torque (force) Ratio (tR) requirement from GEVS SE (Sec. 2.4.5.3) is: t~ = tavaill trequird 2 3-0 and the Torque (force) Margin (tM) requirement from GEVS is: (Equation 1) tM = tavail trequired - 1 2.0 (Equation 2) for systems dominated by resistive torques due to friction. This assumes worst case for the boom, taking the lowest force for the Stacer and applying it to the entire stroke. There is additional margin as there is a significant mass at the end of the boom SWEA / STE-D which contributes momentum towards full deployment stroke. Using the given values it can be shown that the force available, the minimum Stacer force of 18 N, divided by the force required, the tube drag of 3.1 N yields a torque ratio of 5.8, and a torque margin of 4.8. The Stacer satisfies the force requirements by analysis. Still, the device must show functionality to prove that manufacturing has been in accordance with design. The graph shows the need for a deployment initiator. The stowed Stacer is in a ‘meta-stable’ condition. If left by itself, it would partially deploy in either direction, therefore a back plate on the canister is required. To ensure that it deploys a deployment assist device (DAD) is incorporated. The final upturn in the force curve is an artifact of how the force was measured. The Stacer in this case is 5-m long, and when it is fully deployed, the coils have tightened onto themselves. The force value was taken at the moment the Stacer began to slip back into the canister. For this case, the coil needs to be expanded significantly, and requires greater effort. There is an additional use for the Stacer after deployment as the secondary EMVEMC shield. While each of the conductors in the harness is shielded, the mission’s low noise requirement demanded a second, ‘over-shield’ for all conductors. Since the harness runs down the center of the Stacer, the Stacer was tied to ground, and serves this purpose. Deployment Sequence Deployment is initiated when a TiNi Aerospace shape memory alloy release device (SMAR, Model P50-810-1 RS) is triggered causing the restraint pin to pull out of the tail of the Stacer tip piece. To give the stacer and tube deployment an initial ‘kick’, a deployment assist device (DAD) is incorporated between the SMAR mount plate and the 50-mm tube base. The DAD consists of three long coil springs compressed when stowed, and when released provide -90 N of push at the very beginning of the stroke. After the first 100 mm of travel, the initial coils of the Stacer are fully formed around the tip piece, and the flyweight brake has been spun up to speed. At this time the DAD has completed its stroke. The Stacer is attached to the base of the 50- mm tube via a swivel, allowing the Stacer to wind down while extending, recapturing the strain energy stored when the Stacer was wound ‘out’ against the canister. At the end of the 50-mm tube travel, the six lock pins pop into their sockets, and transfer the Stacer push force, as well as momentum, to the 90-mm tube, pulling it along until it latches; the process continues with the 130-mm tube and the 170-mm tube, and finally the entire four tube rigid assembly locks onto the 210-mm tube, which is fixed to the spacecraft. While the actual sequence follows this description fairly closely, occasionally the tube drag would cause one or another tube to partially deploy. There is no provision or requirement for any tube to deploy in any set sequence. To control the velocity of the tubes during deployment, the flyweight brake is attached to the 50-mm tube via a lanyard, limiting the speed of deployment to -0.5 m/s, giving a total deployment time of -10 sec. There are position alignment blocks for stowed (launch) condition holding the tubes aligned 1. i -. L-he- Figure 14. Impact Boom after a thermal vacuum deployment 58 | Document | AMS_2006.pdf | 72 |
relative to each other, and carrying the vibration loads. These also incorporate ‘kick’ springs to aid in their deployment, and to alleviate any possible “stiction” from the alignment blocks. The boom is not retractable once deployed. Re-stowing is achieved by removal from the spacecraft, and hand retraction of each set of pins followed by each tube segment being (de) telescoped; after which the Stacer is compressed into its canister; and the harness and flyweight brake are rewound. Finally, the SMAR is reset reinstalled, and preload is set. Verification The Impact Boom’s qualification activities were based on GEVS SE, as modified by JHU-APL for mission specific needs. The test regime selected for the Stereo mission was Protoflight, meaning new (unflown or non-heritage) hardware is tested with a combination of prototype (EM) levels (i.e., temperature or vibration) with flight (FM) durations. This method is typically used to shorten development times by eliminating the engineering / qualification model fabrication and test period. However, the Impact Team did build up an engineering model, and tested all 3 assemblies to the protoflight levels. The main changes and additions pertinent to this paper: Level 300 cleanliness, UV + Visible light inspection, no silicones used for fabrication, and testing for silicone residuals. Vibration levels were taken from the Delta II user’s manual* modified by APL analysis for the ‘stacked’ configuration. Stringent EMVEMC levels were levied, due to the extremely sensitive radio receiver and magnetometer on board. r Testing procedures were standard NASA mission fare. The main tasks to be performed for this application were: demonstration of sufficient force (torque) margin for extension of telescoping sections throughout the Boom’s stroke; thermal design validation at 25K (discussed previously); and thermal vacuum cycling and deployment verification at hot and cold operational temperatures. As the team worked on the testing it became clear that Stacer thrust force is not easy to measure accurately or repeatably. The deployment of a Stacer is a ‘stick - slip’ affair: and once stopped, it sticks, then when released slips, giving a wide range of force values due to the hysteresis built into the inter coil friction. For consistency, the force value used at any point was the force needed to start the Stacer being pushed back into the canister, after overcoming the ‘stiction’. This does not accurately convey the sliding force, but is as close as can be statically measured. Attempts to measure Stacer force dynamically were fruitless. Another difficulty lay in measuring the drag from rollers and harness. Each tube has a 1.1 -m stroke, and pulling steadily for that distance vertically while monitoring force is a challenge. The weight of the tube assembly was subtracted from the pull out force, giving the drag value. Finally it was seen that proving force margin analytically was not conclusive as the uncertainties in each measurement, when combined, exceeded the margin requirements. A different path was chosen: show that the boom deploys while using 113 of the available Stacer force. By definition, there is sufficient margin. This is how the boom was verified. Figure 15. Thermal Vacuum Chamber Delta II Payload Planner’s Guide, The Boeing Company, Huntington Beach, CA 92647 8 59 | Document | AMS_2006.pdf | 73 |
When testing deployables, the desire is to prove beyond question that the mechanism will deploy in space, however, it must be tested here on the ground first. How many times? GEVS provides a minimum, and each program defines how many additional operations. This brings up wear margin: the design must show that it is sufficiently robust to survive testing and flight without degradation. The EM served this purpose, getting many deployments more than the FMs did. After identifying these values, a test plan was developed, reviewed and implemented. Testing large deployables in a simulated space environment is difficult, and ensuring that the test actually verifies functionality is critical. Deploying the boom horizontally was initially considered since it is easier to develop a 6-m- long test rig that rests on the floor. After a few small efforts in this orientation, it was realized that the only way to ensure that roller drag on the tubes was representative of actual orbital deployment would be to deploy the boom vertically. While several interim off load pulley systems were used, all the verification deployments were performed on the thermal vacuum gantry. To this end, a tall vacuum chamber was designed and built to allow the tests to be performed (Figures 14 & 15). Inside the chamber ‘chimney’ a gantry that allowed a counterbalance pulley system to provide G negation was installed (Figure 16). The distance from the top of the boom to the pulley was maximized to provide the least possible restorative (centering) force to the I sections of the boom during deployment. Figure 16. Gantry Detail Counterweiaht DescriDtion To demonstrate the force (torque) margin, the masses to be used for the counterbalance force had to be chosen to show that the Stacer would be energetic enough to deploy the boom. The mass of each of the tubes (in flight configuration) was added to give neutral balance, plus the Stacer neutralization mass (determined by bare Stacer vertical deployments to be 164 9). This mass (5214 g) was decreased by 213 of the Stacer minimum force (31 N * 0.67 = 20.7 N, converted to kg: 20.7 N / 9.98 kgm/s2 = 2.07 kg) and subtracted from the counterweights. All 10 verification deployments were ‘force margin’ deployments and were successful. After deployment, each boom was inspected for wear, with no signs of degradation of rollers or Gr/E. The EM has been deployed -20 times, and is still in good condition. Initially, the counterweights far exceed the G negation requirements for deployment as only one tube is being deployed, while it is being pulled by the counter weight for 4 tubes. This is not invalid for our testing needs, as the flyweight brake dissipates the extra force, keeping the velocity in correct range. The area of interest is the very end of travel, where the Stacer force is lowest, and the full mass is being acted on. It was this point that the gantry design was built on. Maanetometer Alicmment Verification After deployment, the alignment of the Mag needed to be measured to determine compliance with the specification. A very accurate digital level (resolution: 0.01 degree a.02) was used for this activity. The measurements were taken relative to the origin and coordinate system established in the Interface Control Document (ICD)’ for the boom. The relative angles for the mounting feet for the X - 2 and X - Y planes were recorded at the mounting foot. The angles for the same planes (translated out to the Mag tray) were then measured. The difference was taken, yielding the deviation of the Mag tray from the mounting plane. Figure 17 shows the measured differences of the deviations for the FMs for each plane. See Appendix A for ICD Details 9 60 | Document | AMS_2006.pdf | 74 |
I Impact Magnetometer Alignment 0.090 0.085 0.040 0.045 0.050 0.055 0.060 0.065 0.070 0.075 -1 Average Deviation in the X-Y Plane (degrees) Circle Diameter = Std. Deviation Figure 17. Magnetometer Alignment Error For the final determination of pointing error, the root of the sum of the squares was calculated to give the magnitude of the deviation (the directiqn is not of interest as long as the specification is met), Flight Model 1 measured deviation "used up" 0.10 of its allowable 0.88O margin, while FM2 used 0.11 . The Impact Boom deploys accurately, repeatably, and exceeds requirements by a large margin. Lessons Learned Desian for test. The final validation of any design is complete when the device performs as expected in its orbital environment. GEVS SE gives guidelines, developed over many programs, as to what tests must be run, and how much extra (or over-) testing is needed. For the boom to be tested, a vacuum chamber was required to be built in a vertical orientation. The initial plan had been to use a horizontal track, using an existing facility. This plan did not give sufficient demonstration of the booms ability to deploy in a straight line as the track would have given alignment to the sections through out their travel. Additionally, the drag induced by the lock pin rollers when the tubes deploy horizontally far exceeded the Stacer deploy force. At the time the decision was made to go vertical, the Stacer should have been sized to allow a non- counterweighted deployment. There would have been a mass hit, but being able to leave off the gantry would have been a great savings, as the chamber would not have had to be as tall. Safetv One person was iniured durina initial installation of the stacer into its canister. As the final Dortion was stowed, the safety ietaining pinwould not fit into the hole provided for it. When additional force was tried, the operator lost control of the stacer and it deployed in an uncontrolled manner to -1 m, when it was grabbed, cutting their finger deeply through two sets of latex gloves. Some points arose from the review of the accident: a) Check fit all safety related parts, sub-assemblies and fixtures. The size + tolerance of the hole for the safety pin were too small after plating to allow the pin to be inserted. This would have been an easy test, prior to assembly. b) Have back up hardware and personnel: the operator was working virtually alone, and had no back up person there to hold the stacer while the pin was being inserted. After the finger was cut, a colleague from another part of the lab had to run to help control the stacer and finger damage. Another 61 | Document | AMS_2006.pdf | 75 |
point for spring loaded deployables: don’t rely on a single safety path. A lock down plate in front of the stacer assembly would have caught the stacer, preventing damage. These recommendations have been implemented in the procedures, and stowing fixture used now: the stowing procedure has a minimum requirement of two persons to proceed. Materials SDecifications Often materials can be useful beyond supplier’s data, one only needs to verify what limits to the previous testing exist. The search for a cryogenic temperature suitable glue was fruitless. In the end a test was performed to establish suitability, after much effort. The question to add after “At what temperatures does the product perform satisfactorily?” is “Has it been tested beyond that?” A fair number of days could have been saved by realizing the manufacturers don’t have all the information. Marain While most programs have a margin requirement, it is good to carry some margin for additional mass demands while designing. This is almost rhetorical. The proposal mass for the instruments at the end of the boom was 1.2 kg. After deliberation, it was determined that the data would be significantly better if pre-processing were done closer to the detector, so additional circuit boards were added out at the end, raising the deployed end mass to 2.2 kg. This drove the size of the release mechanism from being ‘off the shelf‘ to a new, custom version, requiring additional testing, with the usual learning curve associated with new mechanisms. The entire structure needed ‘beefing up’ to accommodate the added loads. Summary The Impact Boom has completed qualification and acceptance verification for use in flight for the NASA Stereo mission. This application has shown the use of a Stacer spring can be implemented for major deployables as a motive force, not only as a sensor or sensor support. This represents a major cost savings from traditional motor driven deployables, with their associated high cost electronics. Currently, the launch is planned for 26 May 2006, with deployment of the boom occurring within a 3 - 30 day window after launch. Acknowledgements Thanks go to Dr. Janet Luhmann, Principle Investigator for the Impact Suite, and the entire Impact Boom Team at the University of California Berkeley, the Space Sciences Lab. Also to NASA Goddard Space Flight Center for working with us to get this new hardware off the ground, and to all the reviewers whose inputs and criticisms helped make the boom robust enough to survive qualification. Thanks also go to the Stereo crew at APL for their continued support for the program. 62 | Document | AMS_2006.pdf | 76 |
I .- Figure 18. Stereo “B” Spacecraft: Impact Boom location References Space System Reliability and Safety Office, Code 302; “General Environmental Verification Specification for STS and ELV, Rev A; (June 1996), National Aeronautics and Space Administration, Goddard Space Flight Center, Greenbelt, MD 20771; June, 1996; 32.4.5.3 - 2.4.6.2 Appendix A 1543.1 [60.757 Overall inc Blanket at each end Figure 19. Impact Boom ICD Detail: “B” Stowed 63 | Document | AMS_2006.pdf | 77 |
4260 [I 67.72-1 (Deployment Travel) r SWEA Deployed Position m ‘i 2 r ” r SWEA Stowed Position +X ___) - 1267.4 [49.907 (from Boom 4299.5 Origin [I to 69.272 Magnetometer) ------I - 5803.1 1228.477 (Overoll incl Blankets at each end) -1 DETAIL B Boom Origin (Bottom of Insulator Pad) Figure 20. Impact Boom ICD Detail: “B” Deployed 64 | Document | AMS_2006.pdf | 78 |
Heritage Adoption Lessons Learned: Cover Deployment and Latch Mechanism James W incentsen. Abstract Within JPL, there is a technology thrust need to develop a larger Cover Deployment and Latch Mechanism (CDLM) for future missions. The approach taken was to adopt and scale the CDLM design as used on the Galaxy Evolution Explorer (GALEX) project. The three separate mechanisms that comprise the CDLM will be discussed in this paper in addition to a focus on heritage adoption lessons learned and specific examples. These lessons learned will be valuable to any project considering the use of heritage designs. Introduction Because of future JPL mission needs for meter class space telescopes, there was an internal technology demonstration to develop a complete mechanism set for single-time deployable cover to protect the optics. Because the task was tracked for a fast technology development, the decision was made to fully adopt the GALEX cover deployment design early in the project cycle to save cost and schedule. For reference, the baseline design aperture opening diameter was 0.83 m (32.7 in) and the outer diameter of the instrument was 1.1 m (43.3 in) in diameter, roughly twice the size of GALEX. CDLM Overview The three mechanisms that comprise the CDLM are the Latch, Hinge, and two Energy Absorbers. When the Latch releases the cover, two push off springs and the Hinge mechanism rotate the cover approximately 4.66 rad (267 deg) and impact the crushable honeycomb filled Energy Absorbers. Deployment time is approximately 3.4 seconds. The Hinge mechanism is un-dampened. After deployment, the cover remains against the canister. An overview of the instrument and placement of the mechanisms are presented in Figure 1. Latch Mechanism The Latch mechanism, attached to the cover, employs a Starsys paraffin thermal actuator as the prime mover. An interfacing slotted Latch Arm is affixed to the Cover Ring. The Latch Arm is spring loaded to rotate away from the aperture opening upon release. A detail view of the latch area is shown in Figure 2, the Latch Arm is shown in Figure 3, and a cross section of the Latch mechanism is shown in Figure 4. The Locking Piston passes through and retains the Latch Arm. The Push Piston has a small-diameter tip and passes through the Latch Arm slot. To deploy the cover, the actuator heater is energized, which translates both the Push Piston and Locking Piston, forcing the Locking Piston clear of the Latch Arm. Once the Locking Piston is clear of the Latch Arm, the Push Piston slips through the Latch Arm as it rotates. The mechanism locks open by means of torsion spring loaded arms that snap into a grove on the Locking Piston. Microswitches sense the motion of the arms and provide telemetry of the mechanism state. Power to the actuator is discontinued once one of three criteria are meet: Hall effect sensors mounted on the Energy Absorbers register a deployed cover (discussed later), PRTs mounted on the actuator reach a maximum temperature, or a time limit circuit is exceeded. Both the temperature and time are based on a look up table derived from thermal vacuum Latch test data. After power is terminated and the actuator * Jet Propulsion Laboratory, Pasadena, CA Proceedings of the 38th Aerospace Mechanisms Symposium, Langley Research Center, May 17-19,2006. 65 | Document | AMS_2006.pdf | 79 |
cools, a Push Piston return spring resets the actuator for further ground testing. Resetting of the cover and Latch Arm is manually performed. Latch Mechanism Cover 7 Hinge Mechanism 7 Energy Absorber Mechanisms Energy Absorber Structure Figure 1. Cover Deployment L and Latch Cover Ring Magnets Instrument Canister Microswitches - LatchArm - 1 Figure 2. Latch mechanism and Cover/Cover Ring 66 | Document | AMS_2006.pdf | 80 |
Figure 3. Latch Arm assembly, rotated Push Piston and actuator return spring ot in Locking Piston Torsion Arms Figure 4. Latch mechanism cross section 6- Hinge Mechanism Figure 5. Hinge Mechanism 67 | Document | AMS_2006.pdf | 81 |
Hinge Mechanism The Hinge mechanism works by means of two sets of nested compression springs acting against a lead screw / carrier nut combination. A clever and complex design, the Hinge is fully single fault tolerant. For some failure modes, such as the loss of a spring, the Hinge is two-fault tolerant. A graphic of the Hinge mechanism is presented in Figure 5 and a cross section is shown in Figure 6. In the cover-closed position, the compressed springs react against the Hinge mid section and the spring support end, which are restrained from rotation by shoulder screws and rollers running in slots in the housing. Attached to the spring support end is the lead screw, which is also restrained from rotation. The lead screw passes though a carrier nut, which is constrained from translation, but allowed to rotate. As the lead screw translates, the carrier nut rotates. The cover is attached to the clevis, which in turn is fitted on roller bearings on the carrier nut. Pins fixed in the carrier nut and clevis engage during rotation, driving the cover open. Each side of the hinge mechanism works independently of the other. If one side jams, the Clevis rotates freely on the bearings. If a Clevis bearing freezes, the carrier nut and Clevis can rotate on the inner bearing. rn Should screw / Rollers Spring Support End Housing L Clevis Translating Rotating Fixed I Lead Screw Figure 6. Hinge mechanism cross section Energy Absorber Mechanism The GALEX Energy Absorber used a compression spring in combination with a ratcheting plunger. Once the plunger was pushed in, the ratchet held the plunger and compressed spring fixed. Hall effect sensors imbedded in the striker and magnets mounted on the cover provided telemetry of a deployed cover. The cover magnets also served to latch the cover open. One disadvantage of the GALEX design was that after each cover deployment test, the Energy Absorber had to be disassembled to be reset. It was requested by the project to simplify the GALEX Energy Absorber design with replaceable, crushable honeycomb. We desired to keep the new energy absorber function as similar as possible to the GALEX mechanism due to the support structure design. The housing and Hall effect sensor striker were left relatively unchanged, but the compression spring and ratcheting device were replaced with crushable honeycomb core. The honeycomb core is bonded to a simple disposable aluminum plate which is attached to the plunger. The push rod and Hall effect sensor striker are then attached to the plunger, creating the plunger subassembly. As the magnets mounted on the Cover impact the striker during a deployment, the core is crushed against the Energy Absorber End Cap. See Figure 7 for an exploded view of the crushable honeycomb Energy Absorber mechanism and Figure 8 for a graphic of the Plunger subassembly. The Energy Absorber honeycomb core is replaced after each cover deployment. The Push Rod is segmented in two pieces with a left-hand thread so that the Hall effect sensor striker and front segment 68 | Document | AMS_2006.pdf | 82 |
Push Rod can be removed from the plunger subassembly without stressing or twisting the sensor leads. The mechanism is reset by first removing the striker and Front Cap and then the Plunger subassembly is removed from the housing and the core replaced. During prototype cover deployment tests, several types of honeycomb were experimented with, for the goal of reducing cover bounce-back and obtaining the cleanest cover capture. Aluminum core, 3/8-5052- 1 .O, foil thickness 0.01 8 mm (0.0007 in) and crush strength of 172 KPa (25 psi), trimmed to three cells was found to work well, however the 50.8-mm (2-in) long samples available came close to bottoming out. New core, 88.9-mm (3.5-in) long, was selected to allow for enough travel with margin (see following paragraph for more detail). Deployment with this core proved to be the best and the cover bounce back was limited to approximately 0.04 rad (2.5 deg) (the cover does not rebound off the striker; instead the Plunger is pulled back to its limit stop). Because of the core’s long length compared to its cross-sectional area, the core appears to initially buckle uniformly, then folds between the mid-section to base. Crush tests were conducted with the core and the force required to continue crushing gradually dropped from the initial buckling. This result corresponded well for this application as the impact force falls off considerably after some energy is absorbed. See Figure 9 and Figure 10 for before and after cover deployment images of 50.8-mm (2-in) and 88.9-mm (3.5-in) core respectively. The kinetic energy of the deployed cover and energy to be absorbed is equal to that of the Hinge mechanism compression spring’s potential energy at the cover closed position. The average crush load multiplied by the crushed length gives the energy absorbed, and thus, with some extra length for margin, defined the length of required core.’ Figure 11 is Hexcel’s honeycomb crush strength curve, which illustrates the peak load, average crush load, and the energy absorbed. Core used in the Energy Absorbers were pre-crushed slightly to remove the peak force spike. Crush test data from the three-cell, 88.9-mm (3.5-in) long core is presented in Figure 12. Housing Endcap 7 Figure 7. Energy Absorber exploded view 69 | Document | AMS_2006.pdf | 83 |
Figure 9. Three-cell core, 50.8-mm (2411) long I Figure 10. Three-cell core, 88.9-mm (3 70 | Document | AMS_2006.pdf | 84 |
Peak Lead crush load bsd eMnat.d by precnuhing Bottomed * I I I I I I I I 1 out 8 mn. Area under curve I energy absorbed Deflection Figure 11. Hexcel Honeycomb crush strength curve’ 318-5052-,0007, 3 Cell, 3.5in Samples Displacement vs Force Sample -1 -5 Displacement [in] Figure 12. Force / Displacement graph for 3-cell, 88.9-mm (3.5-in) core 71 | Document | AMS_2006.pdf | 85 |
Heritage Adoption Lessons Learned Sufficient review of a heritage design is necessary before adoption as heritage designs may impose unnecessary limitations, constraints, or failure modes on interacting mechanisms or systems. Additionally, a flight readiness review of a heritage design is necessary, as it can not be assumed that all necessary information regarding the design, such as as-built drawings, assembly instructions, test procedures and test data, are available. Of particular interest are the GALEX heritage adoption lessons learned during the CDLM technology development. GALEX cover deploy mechanism was obviously built to a redlined set of drawings; however, these drawings were not available during this task. As GALEX was an aggressively “faster, better, cheaper” mission, it is thought that resources were not available to complete the drawing package. Much of any schedule savings in using the heritage design was exhausted going through more than 70 drawings to look for, and correct, interference, material, and lubrication issues. Some issues were found only after fabrication and assembly, necessitating the rework or re-fabrication of built parts. Accurate build and assembly histories are required to adopt heritage designs. Very limited test data, such as cover deployment time and cover impact force, was available from GALEX. The lack of test data necessitated the building of a schedule intensive deployment test fixture and mockup cover and duplicating cover deployment tests. Fortunately, a spare Latch mechanism and an engineering model Hinge mechanism were available for testing. An image of the deployment fixture is shown in Figure 13. Mockup cover Early adoption of the Hinge mechanism restricted the design of the crushable honeycomb Energy Absorbers. Not until the deployment test fixture and prototype energy absorbers were built and tests run did it become apparent that the cover impact force was much lower than expected. If the honeycomb crush strength was too high, the cover would bounce off the Energy Absorbers as the impact force dropped off during impact. Conversely, the honeycomb had to be strong enough to resist the Plunger subassembly inertial loads during vibration. It was found during testing that the impact force necessary to sustain honeycomb crushing, and to effectively keep the cover from rebounding, required the Hinge torque output to be increased (by the use of larger springs). By increasing the un-dampened Hinge torque output, the amount of kinetic energy to absorb was also increased. This was further complicated by adopting the same GALEX composite cover thickness, even though the cover grew substantially in size. An alternative design, with a stiffer cover, could have placed the crushable honeycomb on brackets close to the hinge axis where the impact forces would be higher, allowing for a more stable and compact 72 | Document | AMS_2006.pdf | 86 |
honeycomb design and reducing the torque output and energy absorption. The overall design would have been much simpler, reducing drawing, fabrication, and installation costs and schedule with the removal of the Energy Absorber structure and simplification of the Hinge mechanism. Additionally, the crushable honeycomb would have been replaceable without disassembly of the Energy Absorber mechanism. GALEX placed the Latch mechanism on top of the cover due to mounting surface constraints and contamination control. However, the proposed configuration could have allowed the Latch mechanism to be mounted to the Cover Ring instead of on the cover. This would have removed the necessity of routing Latch cabling over the Hinge, eliminating cable parasitic torque drag. Additionally, the simplified cabling could have received cable stiffening micrometeorite shielding. With the Latch mechanism mass removed from the cover, the crushable honeycomb could have been placed close to the hinge axis instead of along the canister. Cover deployment depended on Latch Arm rotation. The GALEX latch arm was positioned close to the instrument aperture opening and was rotated away from the aperture to reduce stray light issues. The rotation of the latch arm rotate introduced some potential single point failures. After the GALEX latch arm was adopted, it was discovered than an earlier mission that originally designed the Latch mechanism instead utilized a fixed latch arm with a slot in the path of cover deployment. Because the proposed design did not face the same stray light issues as GALEX, a potential failure mode was unnecessarily adopted and additional work was necessary to reduce risk. The Latch mechanism is zero-fault tolerant in some cases as it depends on a single actuator (with redundant heaters) and a single set of pistons to translate prior to cover deployment. An alternate latch mechanism design using two Starsys pin pullers in a toggle type configuration could have been more desirable and would have been fully single fault tolerant. Additionally, the mechanism would have been less complex and would possibly have fewer parts and less expensive to fabricate. Both the Latch and Hinge mechanisms were complex, with many tightly toleranced parts. These mechanisms were expensive and schedule intensive to fabricate. Simpler alternate designs described above would have potentially saved more schedule than building the heritage designs. The Hinge mechanism is only capable of cover deployment with the Hinge axis aligned with gravity (vertically). It would have been preferable to incorporate a hinge mechanism that was capable of deploying the cover in any orientation as the instrument will not be positioned vertically during I&T and ATLO, making an end-to-end test impossible. Instead, cover deployment tests will be conducted before CDLM delivery to I&T. Once in I&T and ATLO, first motion tests will verify Latch mechanism functionality and the cover will be sweep to ensure there are no obstructions. The Energy Absorber honeycomb was finalized before flight-like cable was installed over the Hinge axis during prototype cover deployment tests. It was expected that the cable could be wrapped in a way to provide a positive torque to aid the cover rotation, but was initially found not to be possible. Further prototype deployment tests with the cable showed that the honeycomb needed to be changed. The crushable honeycomb Energy Absorber prototype effort cost significant schedule. In keeping the Energy Absorber housing similar to GALEX (and thus limiting the redesign of the support structure), the Energy Absorber still must be disassembled to be reset after a cover deployment test. This disassembly is only moderately less time consuming than a GALEX Energy Absorber reset. While the residual end-of-travel Hinge torque output and magnetic latch provide enough force to keep the cover captured during spacecraft maneuvers, a positive latch mechanism would have been more preferable. 73 | Document | AMS_2006.pdf | 87 |
Conclusion Heritage designs offer projects an attractive means of reducing cost and schedule. However, without a through review and investigation of the state of readiness, a heritage design may impose unnecessary limitations and restrictions, in addition to cost and schedule impacts. It is recommended that trade studies be completed of designs similar in function to that of the proposed heritage design prior to adoption. Only then, potential shortcomings of the heritage design may become apparent. Before adopting a heritage design: Perform trade studies of designs similar in function to that of the heritage hardware. Look for best solution. Thoroughly review heritage drawing package for completeness Verify the heritage design will meet project requirements Review heritage test data and test plans and verify they meet current projects requirements. If they do not, study impact and feasibility of revised testing. Review heritage design for failure modes. It can not be assumed that all modes were found, or that new modes will not be introduced. Perform all prototype testing with as flight-like hardware configurations as possible Acknowledgements The research described in this paper was performed by the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration. The author gratefully acknowledges the contributions of JPL team members: Mark Baker, Kevin Burke, Keith English, Ted Iskenderian, Mike Johnson, Ellyn McCoy, Doug Packard, Don Sevilla, and Brad Swenson. References 1. “HexWeb Honeycomb Energy Absorption Systems, Design Data”, Hexcel Corporation, March 2005. 74 | Document | AMS_2006.pdf | 88 |
Problems and Product Improvements in a Qualified, Flight Heritage Product Chuck Lazansky’ and Scott Christiansen* Abstract This paper will discuss improvements to an existing, qualified, flight heritage launch restraint and release mechanism. The changes made to the design are the result of customer feedback, test failures, and optimization of performance margins. Specific test failures and their resolutions will be discussed. Impacts to field units, process controls, product heritage, and qualification status will be summarized. Conclusions and lessons learned will include aspects of what “qualified product” means and insights around what is required to improve a product based on lessons learned through production and customer use. Introduction A Launch restrainthelease device must meet a demanding set of requirements to be reliable and robust. Most importantly, the device must never release prior to command (during ground handling, transportation, launch, etc.) and must always release when properly commanded. A combination of redundancy and robust design features are typically used in pursuit of meeting these demanding requirements. The QWKNUT has been designed with these goals in mind, and has been shown through qualification and flight use to meet these requirements. Figure 1. Gen 1 QWKNUT Mechanism Product Description The QWKNUT is a device which utilizes a segmented nut to maintain and release an axial preload. The device accepts a standard, hardened, %-28 bolt to carry the nominal 13345-N (3000-lbf) load. Preload is released when the four nut segments are opened by activating the mechanism. Release of the device is initiated by a redundant pair of shape-memory alloy (SMA) wires within the QWKNUT, which are linked directly to the latch. The QWKNUT requires an electrical pulse (3-5 Amps, -10-75 msec) similar to that used for pyrotechnic release devices. The pulse causes resistive heating of the SMA wire above its transition temperature, resulting in a strain of the wire and release of the latch. Functionally, the mechanism can be separated into two parts: the preload-bearing part, and the latch- release part. The preload-bearing part consists of four nut segments (Figure 3). Axial load from the bolt exerts a radial load on these nut segments, which are retained radially by a set of rollers and a bearing outer race. This outer race has slots which correspond to each roller, such that rotation of the race allows the rollers to drop into these slots, allowing the nut segments to open and release the bolt. The outer race is preloaded with redundant coil springs acting to rotate the race and release the device. Starsys Research Corporation, Boulder, CO Proceedings of the 3d’ Aerospace Mechanisms Symposium, Langley Research Center, May 17- 19,2006 75 | Document | AMS_2006.pdf | 89 |
The outer race is prevented from turning by two latch pawls (called “toggles”) that engage the OD of the race. The latch-release paH of the mechanism consists of a lever-arm which reacts the radial load from the toggles, and retains them in the outer race (Figure 2). Nominally, the lever arm is held against a stop feature, in the latched position, by a retention spring. The lever arm is acted on directly by the SMA wire, which upon activation rotates the lever arm and frees the toggles, causing release of the mechanism. REDUNDANT COMPRESSIMJ SPRINGS =ATE TCRQUE TO ROTATE OUTER RAE ROTATES OUTER RACE LATCHED RELEASED Figure 2. Latch-Release Section Detail OUTER RACE ROLLERS DROP INTO NOTCHES NUT SEGMENTS ARE SEGMENTED NUT FREE TI7 OPEN LATCHED RELEASED Figure 3. Preload-Bearing Section Detail A key feature of the QWKNUT is that it is fully resettable. The reset procedure entails using a small tool to rotate the outer race back towards its latched position. The reset process is the exact reverse of the release sequence. Once the outer race is rotated, the rollers move radially inward closing the nut segments. The toggles then re-engage the outer race, and the lever arm, which is spring loaded, moves over both toggles to latch the mechanism. The straightforward reset process and multiple use capability is an attractive feature in a release device, allowing a “test as you fly” approach to ground verification. For additional ease of ground use, the original version of the QWKNUT contains cutoff switches, which discontinue power to the SMA circuits once release occurs. The switches are not rated for Flight use, and additional lead wires are included which bypass the cutoff switches for Flight. 76 | Document | AMS_2006.pdf | 90 |
History of Product The product was developed and fully qualified in 1999 with two test units. The initial production run of 4 units was delivered, and successfully flown, with the FalconSatl program. Following the initial run, roughly six production runs were completed for various programs, totaling roughly 50 units. During this time, we became aware of improvement opportunities as we worked to support customers with specific flight applications. These were primarily usability issues related to damage during handling and operation, and included: 0 Cutoff switch arm breakage 0 Reset difficulty 0 0 0 Strain relief of lead-wires SMA wires burned by excess pulse energy Mechanism damage due to bolt over-engagement Some of these issues were addressed with engineering changes during these first 6 production runs. In 2004, engineering proposed more extensive design changes and an internal research and development effort was initiated. Production of the original design (Gen 1) was to continue until the new design(s) (Gen 2 and 3) and the delta-qualification program was completed. There were two versions of the new design to be qualified to suit different programmatic needs. The Gen 2 design did not include an in-line current shut-off feature. The Gen 3 design included a redundant, in-line current shut-off rated for Flight use as well as ground operation. While the above designs were being developed, the final planned production lot of the Gen 1 design was in process (circa May 2004). During this time, we encountered a failure in vibration testing of several Gen 1 Engineering units at levels well below qualification. Test Failures Four significant test failures have occurred during the product life to date. These failures occurred sequentially and were linked together as the process of investigation and resolution were carried out for each. As each of the failures was investigated, our understanding of the mechanism increased. In some cases, the failures required an exhaustive process of properly addressing any and all affected hardware in the field. As painful as it was, the investigation process created invaluable opportunities to roll improvements into the Gen 2 and 3 designs, requiting in an extremely robust final product. Starsys’ FRB (Failure Review Board) process involves a systematic approach to failure resolution, utilizing tools such as failure trees, fishbone diagrams, and detailed tracking of closure actions. Regular reporting of progress and senior technical oversight continue until root cause and corrective action plans are determined, and the investigation is closed. From June of 2004 until November of 2005, the QWKNUT was under nearly continuous FRB activity for each of the failures described. These failures are an example of how a qualified design can fail, and evolve, over the product life-cycle. Engineering Unit Vibration Failures In June of 2004, an Engineering Unit QWKNUT (Gen 1) undergoing development vibration testing released the nominal 13345-N (3000-lbf) preload under the relatively low ievel of 12 Grms compared to the 35 Grms qualification level. In reviewing the test setup, there were a number of potential causes including cross-axis noise on the vibration table, and a non-standard, massive bolt interface. However, two more EDU units at Starsys Research were vibration tested and also showed an intermittent release of preload at levels below the 35 Grms qualification. We initiated an FRB action to investigate the anomaly. With roughly 38 Gen 1 flight units delivered to customers and staged for flight use, there was concern regarding the scope of the anomaly and status of units in the field. The engineering team generated a complete failure tree. As potential causes were eliminated, the failure tree pointed to the latch release portion of the mechanism. High-speed video of the failure event showed the lever arm moving slightly during vibration with respect to the toggles, and then moving suddenly, and fully, in the direction of release. A cause for this behavior could not be found initially. Physical and dimensional inspection of the parts did not reveal any clear discrepancies. The engineering analysis 77 | Document | AMS_2006.pdf | 91 |
showed relatively low inertial forces on the lever, which could not overcome the lever retention spring and drive the lever in the direction of release. Tests were developed to assess latch performance. The lever retention spring force was measured, and also the net force required to release the lever arm from the toggles (force-to-release, or FTR). The results of the testing were that some units showed low lever spring force (up to 50% below nominal), and this was strongly correlated to vibration failure at 35 Grms. All units which failed in vibration showed spring force below nominal values. The low spring force was traced to yielding of the spring caused by installation damage, as well as overstress during manual releases. Though low spring retention force was strongly correlated to vibration failure, it did not explain the physical cause of the lever motion under vibration. The true root cause was elusive. Driven in part by the need to resume production, the vibration performance was initially addressed by increasing latch retention spring force. This was accomplished with the addition of a second retention spring. The dual-spring design provided strong margin for load- holding under vibration, and was not vulnerable to handling damage. Validation testing of the proposed change supported the assertion that release margins and vibration margins were better balanced with the second spring, and that cycle life was not impacted. A formal Delta-Qualification test program of the new configuration was planned in parallel with production. The latch performance tests were also added to the production process as additional screens. However, when production resumed, the in-process testing showed inconsistent results for FTR screening. There was variability within a single mechanism that could not be explained. Production was stopped, and the investigation resumed with the focus on the dry-film lubrication layer at the latch interface. The lubrication was to be removed from the lever of a production unit to determine the source of the inconsistent FTR measurements. Drive springs (2X) provide bias torque for Flywheel ! ! ! ! ! ! ! ! ! ! i Section through latch area of QWKNUT “%._. ,,...- During launch Lever :’ Spring preloads Lever On orbit SMA wire pulls Lever Arm off Lever Arm creates &/;&n force n Toggle of is Lever perpe dicular to the Lever Toggle Pivot reacts torque to the Body t Figure 4. Latch-Toggle Section Detail Failure Mechanism The physical cause of the failure was discovered when lubrication was removed from the critical latch surface on the lever. The shape of the surface did not meet the intent of the design. The curved surfaces contained “flat” areas approximately 0.76 mm (0.030 in) in width, at each end (Figure 5). These flats were not easily seen with the dry-film lubrication present. The presence of the flats immediately explained the 78 | Document | AMS_2006.pdf | 92 |
intermittent vibration behavior and the variability in release force measurements. If the toggle is on the “flat” area of the lever, a tangential driving force results and leads to release of the device. If a toggle is located on the curved surface of the lever, no driving force is present and no release occurs (Figure 4). The width of the flats corresponds closely to the nominal location of toggle contact when the device is latched. In this case, small variations in toggle position on the critical surface caused the variability in latch performance. The physical cause was confirmed when the discrepant lever arm in the failed unit was replaced with a conforming part from another lot, and the unit passed 3 repetitions of vibration exposure (35 Grms, 3 minutes). Root Cause The root cause of the failure was found to be inadequate form control of the critical latch surface. Though the latch surface was depicted on the drawing with a continuous curvature, the dimensional controls were not adequate to ensure this. A total of 5 production lots of lever arms were received. The first 4 production lots of parts met the design intent. The 51h lot was procured from a new supplier who used a different machining process, which did not produce the intended surface at the critical latch point. The inspection process did not detect the different geometry since the drawing did not specifically control the feature shape. The inadequate drawing controls, combined with variation in manufacturing process (change of vendor, machine, or even machinist) created the opportunity for the discrepancy. Fortunately, the discrepant lot of levers was confined to the current production run, and no field units were affected. Continuous curvature between 45” surfaces \ Burnish mal indicating Tc contact poin Lever from previous GOOD lot - MoS, coating intact; continuous curvature of radius is evident “Flats Burnisl indicat contac Lever from discrepant lot - MoS, stripped down to anodize Figure 5. Lever Surface Discrepancy Corrective Actions The corrective actions for the vibration anomaly included adding proper engineering controls to the critical lever surface to ensure design intent is met. The curved surface is now tightly controlled on the drawing using runout and true position callouts. 100% inspection of this surface is required per an inspection process defined on the drawing, followed by engineering review before parts are certified and released. Process controls were also added at the assembly level to verify proper latch performance. These include measurement of retention spring force and force required to release the latch and defined pass/fail criteria (Figure 6). The latch release point is also screened against pass/fail criteria. The dual-spring design change was implemented (and delta-qualification tested) to improve margins for vibration performance, and prevent handling damage. Finally, acceptance level vibration testing under nominal preload is performed on all units. Once the lever surface was restored to the intended geometry, the device has proven to be extremely robust with respect to vibration exposure. Delta qualification tests were performed at greater than 42 Grms with the lever arms intentionally shimmed 0.508 mm (0.020 in.) in the direction of release (to approximately 70% of the release point) and all units passed the test. 79 | Document | AMS_2006.pdf | 93 |
End of preview. Expand
in Dataset Viewer.
Aerospace Mechanism Symposia PDF documents parsed by page. All symposia documents from the year 2000-2022 are included. No splitting was used. Original documents here: https://github.com/dan-s-mueller/aerospace_chatbot/tree/main/data/AMS
- Downloads last month
- 0
Size of downloaded dataset files:
13 MB
Size of the auto-converted Parquet files:
7.01 MB
Number of rows:
5,387