Datasets:

ai-aerospace

/

ams_data_full_2000-2020

like 0

Electrical Contact Ring Assembly (ECRA) The elevation axis harness, which consists of 14 power circuits and 26 signal circuits, is routed through the ECRA. The outer structure of the ECRA, shown in Figure 5, is stationary and is hard-mounted to the same bracketry as the azimuth actuator stator. The inner structure of the ECRA is fully supported in the stationary section by a duplex set of angular contact bearings as well as a trailer bearing. A tooling ball mounted in the base of the rotating portion of the ECRA mates with a slot in the azimuth output shaft allowing the gimbal to transmit the rotary motion while permitting slight angular misalignment between the ECRA and the output shaft. Drive Tooling Ball Figure 5. ECRA with Thermal Hardware Attached (prior to lead tape over wrap) The structure consists primarily of aluminum and titanium components. The ECRA uses gold/silver/nickel alloy mono-filament brushes in gold-plated brass grooves. Two outer brush blocks that are part of the stator support brushes that span the gap between the stationary and rotary portions of the assembly. Each groove accommodates two brushes, one leading and the other trailing. The power circuits utilize a three-groove design while the lower-powered signal brushes utilize a two-groove design. The final result is a current-carrying margin in the power and signal circuits of three and four times respectively. This is in addition to the electrical redundancy in the gimbal itself. Component-Level Test Issue During component thermal testing, the ECRA exhibited higher than expected noise in the lines. While the ECRA exhibited acceptable noise performance at temperatures greater than 0oC, the noise levels on several circuits increased considerably at sub-zero temperatures. The primary cause was determined to be water contamination of the brush-groove lubricant. It was determined that the nitrogen environment in the thermal chamber was insufficient to purge the assembly of all water vapor. The primary solution was to perform the thermal testing in vacuum. Additional steps that were taken included a vacuum bakeout without the external housing installed before final assembly, a run-in before testing, and efforts to improve 420

Document

AMS_2008.pdf

the cleaning process and prevent contamination sources after cleaning. These changes to the assembly and test program resulted in power and sign al circuit noise level well below the required values. RF System SDO’s data downlink is carried out by a Ka-band RF transmission of 26.5 GHz from a ¾-meter (30 in) High-Gain Antenna to one of two 18-meter (60-ft) ground station antennas. WR-34 waveguides are used to transfer the signal from the tran smitter to the HGA. They are aluminum, plated with silver. An antitarnish coating was applied over the silver plate. Rotary joints, shown in Figure 6, pass the RF signal with budgeted 0.2-dB insertion loss and allow rotations about the two axes. The stationary and rotating sections of each rotary joint are aligned by a duplex set of ball bearings. The rotating portion of the joint is driven similarly to the ECRA. A ball drives the movement through a slot arrangement, using a tooling ball and a clevis with a close-tolerance gap. Drive Tooling Ball Figure 6 RF Rotary Joint with Drive Tooling Ball (Elevation Axis Shown) Waveguide Failure Originally, there were two 7- to 11-cm sections of corrugated flexible beryllium copper waveguide on the gimbal—one in the azimuth section and one in the elevation section. The design intent was that the accordion-style flexibility would compensate for tolerance stack-up and for slight variations in temperature or CTE mismatches. During initial Qualification Unit vibration testing, both of the flexible sections broke completely due to low-cycle fatigue. One cause for this failure was insufficient waveguide support. Some rigid waveguide spans were 25 cm or more, while the manufacture r recommended 15 cm or less. During vibration, it was shown by analysis that the flex waveguide saw deflections well above the yield stress point. The main cause for failure was improper heat treatment and fabrication steps of the delicate 100-μm thick corrugated sections. Metallographic analysis of the failed waveguide revealed a larger grain size than that associated with the certified heat treatment. There are multiple con ditions that can result in excessive grain size. Regardless of its cause, this condition was a primary contributor to the low cycle fatigue failure 421

Document

AMS_2008.pdf

that occurred. There was also concern with the corrugation process, which le d to variable thickness. The thickness before cold forming is 125 μm. After forming, it was to be no less than 100μm. The uncertainty added to the difficulty in analyzing the part. Other problems with the waveguide related to the braze joints between the 1mm thick rigid sections and the flex sections, a 10:1 thickness ratio. Heat from the braze operation could also have increased brittleness in the proximal region where all failures occurred. The added thickness at the joint compounded the stress concentration on the thin section. Figure 7. Waveguide Resolution - Elevation Waveguide with “P-Trap” section Failure Resolution For the elevation waveguide, analysis confirmed that the 90° bend in flex could be replaced with rigid waveguide in a slightly longer, convoluted path. The addition of two more 90° bends created a shape similar to a plumbing expansion section or a P-Trap, shown in Figure 7. The extra path mitigated misalignment and thermal effects. Tolerance stack-up in the axial direction away from the P-Trap was accommodated through the use of aluminum shims. Because of volume constraints, the azimuth section could not have bends, so two approaches were explored. The first was to procure new flexible waveguides made with properly heat-treated material, and to redesign the structural supports. New waveguides were ordered and tested with sufficient supports, and this arrangement was deemed acceptable. The second solution, shown in Figure 2, was to replace the flexible section of waveguide with a slip-joint section. This was also developed and tested, and it was found to work well. Ultimately this slip-joint approach was determined to be more robust with no discernible failure mode, and it was selected for use in the azimuth Waveguide. RF Path Performance The Ka band transmission performance was tracked throughout the development process. Individual waveguide sections were scanned at various points during manufacture, test and integration. Upon delivery from the vendor the RF performance, as measured by Insertion loss and VSWR, was part of the 422

Document

AMS_2008.pdf

End-Item Data Package. Prior to integration into the gimbal, the waveguide was assembled on the bench and throughput loss measured. The budgeted and actual losses are shown in Table 1. Table 1. Gimbal Waveguide Total Throughput Loss Loss (dB) Budgeted to each Gimbal -1.45 Highest measured in Flight Unit 1 -0.86Highest measured in Flight Unit 2 -0.92 Pointing Capability The characteristics of the RF system, including the transmitting and receiving antennae and the power available for transmission, drove the need for an overall allowable random pointing error of + 0.30 degrees. This value includes spacecraft position knowledge, attitude knowledge and control. The gimbal portion of this pointing budget was 0.14 degree. This random error is measured on the ground to be 0.042 for the first gimbal and would have been 0.062 (not including boom-to-gimbal co-alignment) for the second. The total budgeted error for the gimbal, including biases that can be calibrated out, is 0.87 degree. Based on ground measurements, this error is 0.175 degree for the first gimbal and would have been 0.356 for the second gimbal. The alignment budget and measurements are summarized below. Table 2. Pointing Budget (Degrees) Budget Ground Measure Known on groundGround-toOrbitRandom Budget TotalsFlight 1 Flight 2 Hardware Alignment Errors Gimbal to boom axis co-alignment error .13 .043 Gimbal to HGA base I/f alignment error .13 .162 .350 Gimbal Interaxial Orthogonality .14 . .007 .015 Gimbal actuator interface launch shift .55** .025 .018 Dynamic Pointing Errors Gimbal/boom dynamic interaction .04 .011* .011* Gimbal tracking error .08 .041 .061 Total on-orbit error (RSS) .23 .55** .09 .87 .175 .356 Total on-orbit error after compensation and onorbit calibration (RSS)*** -- .05 .09 .14 .042 .062 * from Qualification Unit Jitter testing ** Worst case assumption *** Ignores 0.02 degree of thermal effects allowed in budget The launch shift was budgeted base d on worst case interface assumption s, the measured value were variations measured be fore and after vibration testing. 423

Document

AMS_2008.pdf

For gimbal/boom dynamic interaction, the gimbal was in strumented with force gages, and the forcing function was used to derive the jitter error measurements. The item tested was an engineering test unit, flight-like in all structural respects. For the gimbal tracking error, the budgeted amount of 0.08 degree was calculated by summing estimates for the following for each axis: harmonic drive wind -up, gear error, step late ncy, and wobble of the actuator output. The results were added vectorially (Root-Sum Squared). For the measured value of 0.041 degree, the rotational error and the wobble were measured for each axis and added together vectorially. Due to 1-g effects, this error is greater than the value will be on orbit. Harnessing and Multi Layer Insulation Some more general lessons learned deal with leaving more space for harnessing and multi-layer insulation (MLI) and addressing these details earlier in the design effort. Since the gimbal is deployed away from the spacecraft body, it is exposed to the worst radiation environment and temperature extremes on the entire spacecraft, except perhaps for the instrument complement. Protecting against this onslaught required elaborate measures that were frequently at odds with the smooth operation of a high-precision pointing mechanism. The volumetric demands fo r harnessing were especially great. Even though the Tefzel ®-insulated wiring is resistant to radiation, the SDO system designers implemented a policy of over-wrapping exposed actuator wires with Kapton, Lead and Aluminum tape. Becau se of the reliability requirements for a 5-year mission at geo-synchronous orbit, the tape layers plus the 36 wires from a single actuator formed a bundle that was approximately 15.9 mm (5/8 inch) in diameter. After wrapping, the metallic layers were each electrically bonded to grou nd with silver-filled epoxy. MLI over the entire spacecraft has an electrically conductive germanium black Kapton ® (GBK) outer layer. MLI is usually a challenge to bend and position in small pieces and tight quarters. The extra layers brought additional concerns as the mechanism and thermal goals conflicted. The gimbal required ten separate MLI pieces in order to protect its various convoluted surface features, as well as allow for access to the various parts. Some MLI pieces were only 15 cm (6 in) on a side. The bends and seams, such as between moving parts, are potential heat leaks that could expose the actuators to dangerous extremes of temperature. In addition, GBK is sensitive to even light abrasion such as normal hand pressure from an accidental brush against its surface. All these factors contributed to a tough challenge of constructing accurate, intricate pieces of MLI, tightly positioned, and allowing free relative motion between close-tolerance parts. Conclusion The first flight gimbal, shown in Figure 8, has been tested and delivered to the spacecraft. The second gimbal will be replaced by a spare after being subjected to damaging temperatures during the postthermal vacuum bake-out. Integration and testing of the re-built gimbal is scheduled to be complete in March 2008. Launch is scheduled for late 2008. Acknowledgements Richard Barclay, Carlos Lugo – gimbal electronics Steven Wood – mechanical assembly Richard Marriott – materials Michael Dube – ECRA noise investigation Javier Lecha, Joe Schepis – HGAS Jason Hair – HGAS deployment Ken Hersey – RF SpaceDev/Starsys – actuators and ECRAs Kevlin Corporation – RF rotary joints 424

Document

AMS_2008.pdf

Figure 8. SDO Gimbal Mated to Deploy Boom and High Gain Antenna, with MLI installed Second Unit Shown Vertical in Background 425

Document

AMS_2008.pdf

xii REPORT DOCUMENTATION PAGEForm Approved OMB No. 0704-0188 Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operation and Reports, 1215 Jefferso n Davis Highway, Suite 1204, Arlington, VA 22202-4302, and to the Office of Management and Budget, Paperwork Reduction Project (0704-0188), Washington, DC 20503 1. AGENCY USE ONLY (Leave Blank) 2. REPORT DATE 3. REPORT TYPE AND DATES COVERED 4. TITLE AND SUBTITLE 5. FUNDING NUMBERS 6. AUTHORS 7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) 8. PERFORMING ORGANIZATION REPORT NUMBER 9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSORING/MONITORING AGENCY REPORT NUMBER 11. SUPPLEMENTARY NOTES 12a. DISTRIBUTION/AVAILABILITY STATEMENT 12b. DISTRIBUTION CODE 13. ABSTRACT (Maximum 200 words) 14. SUBJECT TERMS 15. NUMBER OF PAGES 16. PRICE CODE 17. SECURITY CLASSIFICATION OF REPORT18. SECURITY CLASSIFICATION OF THIS PAGE19. SECURITY CLASSIFICATION OF ABSTRACT20. LIMITATION OF ABSTRACT NSN 7540-01-280-5500 Standard Form 298 (Rev. 2-89) Prescribed by ANSI Std. 239-18 298-102Unclassified Unclassified Unclassified UnlimitedE.A. Boesiger,* Compiler George C. Marshall Space Flight Center Marshall Space Flight Center, AL 35812 National Aeronautics and Space Administration Washington, DC 20546–0001 *Lockheed Martin Space Systems Company, Sunnyvale, CA NASA MSFC Point of Contact: Don McQueenAn electronic version can be found at http://ntrs.nasa.gov Unclassified-Unlimited Subject Category 37Availability: NASA CASI 301–621–0390 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, NASA Marshall Space Flight Center (MSFC) and Lockheed Martin Space Systems Company (LMSSC) share the responsibility for hosting the AMS. Now in its 39th symposium, the AMS continues to be well attended, attracting partici-pants from both the United States and abroad. The 39th AMS was held in Huntsville, Alabama, May 7–9, 2008. During these 3 days, 34 papers were presented. Topics included gimbals and positioning mechanisms, tribology, actuators, deployment mechanisms, release mechanisms, and sensors. Hardware displays during the supplier exhibit gave attendees an opportunity to meet with developers of current and future mechanism components. 436M–1225Conference Publication May 2008 NASA/CP—2008–215252 actuators, bearings, deployment, design, gimbals, mechanisms, release, test, tribology39th Aerospace Mechanisms Symposium

Document

AMS_2008.pdf

National Aeronautics and Space AdministrationIS20George C. Marshall Space Flight CenterMarshall Space Flight Center, Alabama35812

Document

AMS_2008.pdf

i NASA/CP-20205009766 45th Aerospace Mechanisms Symposium C ompiled/Edited by: Edward A. Boesiger P roceedings of a symposium Hosted by the NASA Johnson Space Center and Lockheed Martin Space Sponsored and Organized by the Mechanisms Education Association 2020

Document

AMS_2020.pdf

iii 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. Sponsored and organized by the Mechanisms Education Association, responsibility for hosting the AMS is shared by the National Aeronautics and Space Administr ation and Lockheed Martin Space. The 45th AMS was scheduled to be held in Houst on, Texas on May 13, 14 and 15, 2020. Unfortunately, the worldwide COVID -19 pandemic did not allow this. These proceedings are published however in order to provide these lessons learned and mechanism design information to the mechanism community. Topics i ncluded instrument mechanisms, release devices , Mars 2020 mechanisms , tribology , actuators and compliant mechanisms . 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, Lockheed Martin Space, and the European Space Agency. Appreciation is also extended to the session chairs, the authors, and particularly the personnel at JS C responsible for the symposium arrangements and subsequent cancellation 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. 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.

Document

AMS_2020.pdf

v CONTENTS Symposium Organi zing and Advisory Committees .......................................................................... ix Nanometric Positioning with IASI -NG‘s Beam Splitter Mechanism Actuator ..........................................1 Francois Barillot, Jocelyn Rebufa, Gladys Jaussaud & Adrien Guignabert Spectrometer Scan Mechanism for Encountering Jovian Orbit Trojan Asteroids .................................. 15 Kenneth Blumenstock, Alexander Cramer, Joseph Church, Jason Niemeyer, Fil Parong, Sam Zhao, Nerses Armani & Kenneth Lee Point Ahead Mechanism for Deep Space Optical Communication – Development of a New Piezo -Based Fine Steering Mirror ……………………………………… ….. .................................... 29 Adrian Guignabert, Thomas Maillard, Francois Barillot, Olivier Sos nicki & Frank Claeyssen Design, Development and Verification of the METimage Mechanisms ............................................... 43 Sebastian Rieger & Armin Jago Challenges of the Development of a Compliant Focus Mechanism Submitted to the Harsh Martian Environment for the ExoM ars Rover Mission ................................................................................. 59 Antoine Verhaeghe, Gerald Perruchoud, Philippe Schwab, Mathias Gumy, Julien Rouvinet & Lionel Kiener Lessons Learned from Q ualification of HDRM for U ltralight LP -PW I Boom for ESA JUICE Mission ........ 73 Maciej Borys, Ewelina Ryszawa, Ł ukasz Wiśniewski, Maciej Ossowski & Jerzy Grygorczuk Development of a F amily of R esettable Hold- Down and Release Actuator s based on SMA Technology and Q ualification of D ifferent Application Systems......................................................... 87 Marcelo Collado, Cayetano Rivera, Javier Inés, José San Juan, Charlie Yeates , Michael Anderson, Francisco Javier Rivas, Mónica Iriarte, Jens Steppan, Calem Whiting & Karine Murray Development and Post -Testing Anomalies of the Parker Solar Probe Clamshell s Development .......... 103 Mark Bryant Mars 2020 Rover Adaptive Caching Assembly: So Many Challenges .............................................. 117 Milo Silverman & Justin Lin Sealing Station Mechanisms for the Mars 2020 Rover Sample Caching Subsystem .......................... 137 Jesse Grimes -York & Sean O’Brien Design and D evelopment of a R obust Chuck Mechanism for the Mars 2020 Coring Drill .................... 151 Anthony Barletta Percussion Mechanism for the Mars 2020 Coring Drill .................................................................. 165 Kyle Chrystal STIG: A Two -Speed Transmission Aboard the Mars 2020 Coring Drill ............................................. 179 Timothy Szwarc, Jonathan Parker & Johannes Kreuser Development of the Next Generation Battery Cell Isolation Switch .................................................. 193 Ruben Betancourt & Michel Knight

Document

AMS_2020.pdf

vi A Fast-Acting Self-Energized, Low -Cost Valve for Air Cannons ...................................................... 199 Lee Brown Development of a Low -Shock Separation Nut Out of the Pyrotechnic Class ..................................... 205 Sebastien Perez, Frederic Miralles & François Degryse Lubricant Degradation in High- Load, High- Cycle Actuator Test Using Heritage Harmonic Drives for the Multi-Angle Imager for Aerosols Instrument ................................................................................. 211 Michelle Easter Multi-Mission Deployable Boom: Spring Mechanism Design, Failure Investigation , and Resolution ...... 227 Christine Gebara & Paul Lytal NSI Performance Improvement T hrough the use of Automation ..................................................... 235 Jason Kozmic , Bill Gratzl & Hobin Lee Microvibrations Modelling and M easurement on Sentinel 4 UVN Calibration Assembly using a Piezoelectric 6 C omponent Force Dynamometer ......................................................................... 241 Benoit Marquet Lubrication Concept Evaluated for Geared Actuators under Starved Conditions ................................ 255 Erik Nyberg, Ichiro Minami & Jonny Hansen Development of a Universal and Scaleable Mechanism Control Electronics Configured to Application Solely by Parameter and Software Configuration ......................................................................... 261 Udo Rapp & Juan Camilo Garcia Hernandez Eddy Current Effects in Spacecraft Mechanisms .......................................................................... 269 Emilia Wegrzyn, Claudia Allegranza, Thomas Adam , Florian Liebold & René Seiler Recovery and Operational Best Practices for Reaction Wheel Bearings .......................................... 277 Michael Dube, Jeff Fisher, Stuart Loewenthal & Peter Ward Laboratory Studies of Spacecraft Fluid Lubricant Mobility and Film Thickness .................................. 287 Peter Frantz , James Helt & Steve Didziulis Efficacy of Lead Naphthenate for Wear Protection in High Vacuum Space Mechanisms .................... 301 Jason Galary Bearing Anomaly for the Sentinel 6 Supplemental Calibration System ............................................. 315 Gale Paulsen, Dylan Van Dyne, Fredrik Rehnmark , Phil Chu & Ted Iskenderian Parker Solar Probe MAG Boom Design, Analysis and Verification .................................................. 329 Weilun Cheng, Calvin Kee & John Wirzburger Development of a Low -Shock Payload Fairing Jettison System ...................................................... 343 Boris Halter, Josef Zemann, S imon Wieser, B eatrice Burkhart, M athias Burkhalter , Alberto Sánchez & Oliver Kunz Deployment Mechanism for an Earth Re- Entry Deployable Decelerator ........................................... 355 Carl Kruger

Document

AMS_2020.pdf

vii Lessons Learned During the Development, Qualification, and Production of the MM Solar A rray......... 369 Thomas Pace Solid -State Hinge Mechanism for Simple Panel Deployment System .............................................. 383 Thomas Rose, William Hensley & William Francis Mars 2020 Motor Bearing Failure , Investigation and Response ...................................................... 397 Dave Suffern, Jeff Mobley & Stephen Smith Mars 2020 m axon Commercial Motor Development from Commercial -Off-the-Shelf to Flight -Qualified Motors, Gearbox es, and Detent Brakes : Overcoming Issues and Lessons Learned ........................... 411 Michael LoSchiavo, Rebecca Mikhaylov , Robin Phillips & Lynn Braunschweig Mars 2020 Center D ifferential Pivot Restraint : Flexurized Spring System Providing Compliance for Rover Mobility Deployment Prior to Landing ........................................................................... 427 Matt Cameron & Kevin Liu Astrobee Free- Flyer Nozzle Mechanism ..................................................................................... 441 Earl Daley Major Design Choices and Challenges that Enabled the Success of the Ejectable Data Recorder System .................................................................................................................... 455 Jeff Hagen, Michael Burlone & Kristina Rojdev Design and Test of the Orion Crew Module Side Hatch ................................................................. 469 Lance Lininger & Kyle Gotthelf Design, Development, Testing, and Flight of the Crew Dragon Docking System ............................... 483 Jaret Matthews, Caitlin Driscoll, Edward Fouad, Andrew Welter, Marc Jamulowicz & Jessica Ipnar Highlights of the Next Generation AIAA Moving Mechanical Assemblies S tandard ............................ 495 Brian Gore & Leon Gurevich Micro-Vibration Attenuation Using Novel Flexible Pivot Design ....................................................... 503 Luc Blecha, Yoël Puyol , Simon Hayoz, Fabrice Rottmeier & Martin Humphries Compliant Mechanisms Made by Additive Manufacturing .............................................................. 517 Lionel Kiener , Hervé Saudan, Florent Cosandier, Gérald Perruchoud, Vaclav Pejchal, Sébastien Lani & Antoine Verhaeghe Flexible Waveguides for RF Transmission across PSP HGA Rotary Actuator ................................... 529 Deva Ponnusamy , Weilun Cheng, Ted Hartka, Devin Hahne, Calvin Kee, Mike Marley & David Napolillo Thermal Vacuum Testing Lessons Learned for Small Stepper Motors and a CubeSat Translation Mechanism ............................................................................................................................ 543 Alex Few, Lynn Albritton & Don McQueen Design and Development of the GPM Solar Array Drive Assembly, Orbital Performance and Lessons Learned .............................................................................................................. 555 Alejandro Rivera, Glenn Bock , Alphonso Stewart, Jon Lawrence, Daniel Powers, Gary Brown & Rodger Farley

Document

AMS_2020.pdf

viii Two-Axis Antenna Pointing Mechanism Qualification for Juice Mission Dual -Band Medium Gain Antenna ......................................................................................................................... 573 Jorge Vázquez, Mikel Prieto, Jon Laguna & Antonio Gonzalez

Document

AMS_2020.pdf

ix SYMPOSIUM ORGANIZING COMMITTEE Host Chair – Brandan Robertson, NASA JS C General Chairman - Edward A. Boesiger, Lockheed Martin Space Deputy Chairman - Stuart H. Loewenthal, Lockheed Martin Space (retired) William Caldwell, NASA ARC Damon C. Delap, NASA GRC Jared A. Dervan, NASA MSFC Adam G. Dokos, NASA KSC Michael J. Dube, NASA NESC Carlton L. Foster, NASA MSFC (retired) Lionel Gaillard, ESA/ESTeC Claef F. Hakun, NASA GSFC Christopher P. Hansen, NASA JSC Louise Jandura, JPL Alan C. Littlefield, NASA KSC (retired) Ronald E. Mancini, NASA ARC (retired) Fred G. Martwick, NASA ARC Donald H. McQueen, Jr., NASA MSFC Robert P. Mueller, NASA KSC Benjamin J. Nickless, NASA LaRC Joseph W. Pellicciotti, NASA HQ Minh Phan, NASA GSFC Joseph P. Schepis, NASA GSFC Donald R. Sevilla, JPL James E. Wells, NASA LaRC Jonathan P. Wood, Lockheed Martin Space

Document

AMS_2020.pdf

1 Nanometric positioning with IASI -NG ‘s Beam Splitter Mechanism Actuator Francois Barillot *, Jocelyn Rebufa* , Gladys Jaussaud* and Adrien Guignabert * Abstract This paper presents a piezoelectric motor which provides linear motion and very high resolution (40 nm steps). First, the space application ( IASI-NG instrument onboard METOP -SG satellite) and associated performance requirements are presented. The internal architecture of the motor and its main components are then explained. A first focus is done on the experimental verification of the threaded interface lifetime which is a key element of the mechanism. A second focus is on the nanometric position test bench. Achieved results are provided for resolu tion, motion quality and position stability. Finally, results from the vibration test campaign are presented. Introduction Many space projects have shown need for stable sub -micrometer positioning linear actuators. They are typically needed to adjust the mirror position of sensitive optical instruments after launch during initialization or throughout flight life to accommodate aging and other long- term variations. In addition, as long periods can be expected between position changes, it is mandatory that the actuator remains passive (i.e. , not powered) once the adequate position is achieved. This need is met in the IASI -NG space instrument , where a linear actuator offering a 30- nm step resolution and an unpowered position stability of 0.30 µm over 6 months was requested. Combined with the requirement of surviving launch, t hese specifications are beyond the capacity of existing linear piezo motors. For example, in [1] the piezomotor survival against vibration loads was not proven. Such external forces will apply directly on both the motor friction surface and the piezo ceramic and may damage the motor. This can be circumvented using a launch lock mechanism at a price of added mass and complexity . In order to meet such a need, Cedrat Technologies (CTE C) has built a hybrid actuator, starting from its patented Fine Stepping Piezoelectric Actuator (FSPA) [2], but using a combination of its magnetic and piezoelectric technologies to reduce electrical requirements. This new linear stepping actuator first generates a rotating movement and then turns it into a translation movement. It offers nanometric positioning resolution combined with the ability to hold its position without power and during launch without the need for any launch lock mechanism. IASI-NG & Beam Splitter Mechanism The Infrared Atmospheric Sounding Interferometer New Generation (IASI -NG) is a key payload element of the second generation of European meteorological polar -orbit satellites (METOP -SG) dedicated to operational meteorology, oceanography, atmospheric chemistry, and climate monitoring. It will provide operational meteorology data such as temperature and humidity atmospheric profiles and also monitor other gases like ozone, methane or carbon monoxide on a global scale. The instrument is developed under the lead of CNES, who is responsible for the development and procurement of the IASI -NG System (Instrument, Ground Processing software, Technical Expertise Center). Airbus Defence and Space was selected for development of the “Space Segment” mainly consisting of the Instrument itself. * CEDRAT Technologies, Meylan, France ; francois.barillot@cedrat -tec.com Proceedings of the 45th Aerospace Mechanisms Symposium, NASA Johnson Space Center, 20 20

Document

AMS_2020.pdf

2 The heart of the instrument is a modified Michelson type interferometer based on the Mertz concept . This interferometer embeds a beam splitter blade. The blade tip- tilt positioning is critical with respect to IASI NG performance. Therefore, specific actuators are implemented to readjust in orbit the Beam Splitter Tip/Tilt alignment: the two BSMA. They provide a linear motion to lever arms acting either in push- push or pushpull to provide both axis tilt. Act uation forces shall be sufficient to counteract the stiffness of the mechanism and to maintain its position after setting; it means a sufficient passive unpowered holding force (actuator non-powered). Moreover, an additional “vibration mode” (small oscillations at high frequency) is needed to assess the dynamic sensitivity. Figure 1. BSMA A ctuators “ Push-Pull” Configuration The major requirements for the BSMA are the following: • Movement: 45-nm resolution step motion over a ± 40-µm range with up to 20N force • Launch: withstand launch including up to 50N force on output shaft • Unpowered position stability: o 0.15 µm over 24h/1K stability o 0.30 µm over 6 months o 1.4 µm long term • Vibration mode: 0.2 to 0.8 µm oscillations with 10 Hz to 70 Hz frequency • Driver capability: o Piezo: Max voltage 120V, Max current 0.1 A o Mag: Max voltage 50V, Max current 0.3 A • Cold redundancy BSMA Architecture Figure 2. BSMA EM With Connector Saver

Document

AMS_2020.pdf

3 Overall architecture The core BSMA is based on the combination of the following elements: • Fine Stepping Piezoelectric Actuator (FSPA): these piezoelectric motors provide a 320- µm displacement range and a resolution smaller than 50 nm, • Parallel Pre-stressed Actuators (PPA): these actuators are used to generate the sine oscillations for the vibration mode. • Eddy Current Sensor: these sensors are used to monitor the effective position of the BSMA output shaft. In order to achieve cold redundancy, each of these elements ha s to be doubled inside the BSMA. The two FSPAs motors are connected by a lever arm which sums the position of the two motor s. This architecture divides by a factor of two the displacement generated by the FSPA. This improves motor r esolution by a factor of two but at the cost of a doubled stroke for the motor. Figure 3. BSMA I nternal Architecture Electrical interconnection of the components is achieved through a multilayer PCB that also includes the eddy current sensors. A single SUB -D connector is then attached on the side of the BSMA to connect the harness. FSPA Actuator The FSPA is a new brand of patented piezoelectric motors from Cedrat Technologies. It is a combination of a Rotating Stepping Piezoelectric Actuator (RSPA) and a differential screw. This FSPA piezomotor is mainly proposed as a product for industrial and laboratory applications , among CTEC range of SPA piezo motors [3]. However, its compact design and vacuum compatibility is an opportunity for cost -effective space applications such as micro satellites or constellations nano satellites . FSPA piezo motors main advantages are: • 5-mm displacement range with up to 120N driving force, • compact casing (Ø50 mm x 45 mm) and low mass (150 g), • typical stepping size adjustable from 50 nm to 250 nm, • holding force while unpowered, against external forces , that can exceed 1 kN, • end-stops that passively prevent the motor from exceeding its operational stroke including in case of a faulty command from the user. The high holding force at rest allows to avoid the use of an additional launch- lock mechanism. This FSPA motor can be used for nano positioning of an optical payload. It could even be used as an HDRM (Hold Down and Release Mechanism) for another mechanism.

Document

AMS_2020.pdf

4 Figure 4. FSPA35XS, Drive Electronic s & Load C urve The FSPA stick/slip operating principle is based on inertial forces as other SPA [4] . The consequence is that some short current spikes are needed to generate the pulses. For IASI -NG a requirement is to minimize the need for current spikes in order to facilitate electronic design. A magnetic clutch is then added to the or iginal RSPA motor. This clutch allows to open the contact between the RSPA module and the rotor when it reverts to its original position. The maximum current for piezoelectric components can then be reduced below 100 mA (even using larger ceramics than usual). Figure 5. Kinematic Chain of FSPA Inside BSMA Figure 5 shows the kinematic chain for the FSPA inside the BSMA. The step sequence is the following: 1. Start powering the magnetic clutch to reinforce the torque transmission between the RSPA module and the rotor 2. Power the RSPA module to generate a small rotation. Rotation is transmitted to the rotor through the clutch. 3. Reverse power in the magnetic clutch to cut the contact with the rotor 4. Cut power of the RSPA t o have it return to its initial position. The rotor stays in place. 5. Cut power of the magnetic clutch. Rotor is then locked in place. Each step results with a fraction of a turn rotation ( α) for the rotor. As the rotor is connected to the structure through screw 1, the rotation causes a small translation ( α.p1) of the rotor relative to the structure. Further, the output shaft (which cannot rotate) moves relative to the rotor in the opposite direction( -α.p2) because of screw 2. -50510152025 0 20 40 60 80 100 120Speed (µm/s) Load(N)

Document

AMS_2020.pdf

5 The resulting displacement of the rotor versus the structure is then α( p1-p2). This configuration allows the use of larger threads, which are needed to withstand the loads , while allowing very small steps per turn otherwise not possible. Magnetic clutch A key aspect of the BSMA is then its electro magnetic clutch. This clutch was designed based on the following constraints: • transmit torque during stepping sequence, • complete loss of contact during release, • closed when unpowered, • perform more than 1 million operation cycles without detrimental wear • coil temperature within acceptable range, • mass allowing RSPA module to support clutch without assistance of any launch- lock mechanism. The clutch design is based on an electromagnet principle. Magnetic technology was selected as it allows larger displacement compared to piezoelectric ceramics. This is needed to generate a gap which is large enough to accommodate for manufacturing and assembly tolerances. Figure 6. Partial C ut View Showing Rotor, Clutch and RSPA M odule In practice, the clutch consists of 6 pallets driven by the central coil and magnet. Each pallet can rotate around a flexible blade when the magnetic coil is powered. At the end of the pallet is a friction tooth that will interface with the rotor and transmit torque from the RSPA module. A permanent magnet generates some flux which closes the pallets (and therefore the clutch) when the electromagnet is not powered. Performance of the clutch was optimized using magnetic simulation ( Figure 7) in order to minimize its size and heating while providing enough force to ensure torque transmission. As the rotor turns, the distance between the tooth and rotor contact surface can vary significantly depending on the rotor runout and other manufacturing tolerances. A major point of concern during the design was to ensure that magnetic forces would not vary excessively, remaining high enough to ensure proper torqu e transmission.

Document

AMS_2020.pdf

6 Figure 7. Clutch Magnetic Simulation Figure 8. Flexible Pivot Stress Simulation Specific attention was also needed for the design of the flexible blades in order to ensure a high transmission stiffness while preserving a low flexure stress and flexure stiffness. High strength steel was used to ensure a >99% reliability of the clamp including the fatigue effect over lifetime. Prestressed Piezoelectric Actuator ( PPA) A double PPA is placed between the lever arm and the output shaft. This PPA is composed off 2 stacked piezoelectric components to provide redundancy. The supplied piezo components have been validated for space application with a LAT (Lot Acceptance Test). This LAT sequence, composed of several test group samples, was established thanks to previous work with agencies (ESA and CNES). Differential Screw Wear Test Bench Guided lever arm A lever arm connect s the two F SPAs and the output shaft. The resulting position of the shaft is the mean position of the two FSPAs. Redundancy is then achieved as each FSPA can move the output shaft independently from the other one. Figure 9. Schematic of BSMA for Max Position, Lever Tilt Angle is Exaggerated However, a major consequence of this architecture is that the lever arm tip displacement s are not straight. The rotation of the arm induces a side displacement which must be supported by the threaded interfaces. This design was selected due to its compactness and considering that : • Static load is rather small (<100N) compared to allowable screw tension in static conditions , • Speed is very low (allowing contact heat to dissipate) , • Limited lifetime requirement (100 operations) .

Document

AMS_2020.pdf

7 Screws excessive wear As expected, the alignment of the clutch with the rotor was a major difficulty during assembly. Several assembly methods had to be tested on the first engineering models to identify the best options. During one of the tests and due to a mishandling, one of the screw s was damaged and had to be replaced. After disassembly, it was found that the other screws showed excessive deformation and wear (even considering the specific history of the model). Figure 10. EM1 Screws, Left Screw Shows Excessive Wear, Right Screw Was Damaged Due t o Mishandling Following this result, it was decided to upgrade the design of the BSMA. The major design improvements were to: • replace initial soft screw material with a high strength stainless steel with space heritage, • use both liquid and solid space qualified lubricants to minimize the risk of dry contact. Test bench verification In order to validate the updated design, it was decided to build a dedicated test bench. Two major elements were to be considered for the test bench design: • Kinematic Accuracy: movement shall be as close as possible to the effective movement in the BSMA, • Load Accuracy: preload shall be the same as in the BSMA, • Accelerated Test: rotation speed must be accelerated to get an acceptable test duration. A geared electric motor was used to drive the differential screw rotor. Transmission to the rotor was done using gear wheels. Gears would not be acceptable for normal use as their resolution is much t oo large for the application, but this was acceptable for the accelerated test, where each cycle was a complete turn of the rotor. Figure 11. Differential Screw Test Bench

Document

AMS_2020.pdf

8 Figure 12. Position and T orque Measurement (400 C ycles Achieved) The results using the test bench validate the applied changes : Although acquisition noise reduced measurement quality , no significant torque trend was observed during the test (see Figure 12). Comparison of initial and final friction coefficient showed no sign ificant trend either . Finally, an inspection of the tested parts was performed by CNES which concluded that thread wear was acceptable. Nanometric Step Size and Stability Test Bench Test Bench Architecture A test bench was designed to analyz e the output position of the motor over its full stroke, in vacuum conditions and under static load. A second key feature of the test bench is the ability to measure the position stability against time of the motor at the nanometric scale over a period of more than 100 hours. This test bench exploits previous successful experience on long- term nanometric stability measured on PPA piezo electric actuators [ 5]. For this purpose, a 3- channel interferometric displacement sensor ( Figure 13a & b) was used in vacuum to achieve a sub- nanometer precision with high repeatability. Each channel measures the length of the laser beam between a fiber -based sensor head (or optical collimators) and a mirror. The position target mirror s show n in Figure 13 b made identically an d bonded at the same time with the same cure process . A load spring wa s included as well as a force sensor to verify BSMA’s behavior while loaded with a constant axial force. The test bench wa s built on a baseplate with closed loop thermal control. The go al was to stabilize the test bench temperature within a 0.2K peak -peak range. Stroke Verification The first test aims to verify the extreme stroke of the 2 FSPA motors. Extreme stroke means reaching the mechanical end- stops of each motor while the other motor is at maximum operational position. This test allows to validate that, in case of a motor failure, the other motor will be able to compensate and preserve BSMA capability to reach any position inside the operational range. Figure 14 shows the output shaft position. Positions for the nominal and redundant motors are estimated from the commands sent to the BSMA.

Document

AMS_2020.pdf

9 Figure 13 - BSMA Stability Test Bench Figure 14 - Extreme Position Verification Test in A ir (a) Overview of the loading assembly (b) The 3-channel interferometric displacement sensor without loading assembly (c) Inside overview of the test bench (d) Outside overview of the test bench in ISO7 clean room Glued target mirror (a) Full stroke (b) Direction change End-stop verification of margins

Document

AMS_2020.pdf

10 Step Sizes A major characteristic of the BMSA is its ability to generate very small steps. Figure 15 shows analyses of step size for nominal and redundant motor when cycling over the operational range and against its lifetime. Performing a direction change does not impact step size nor shows any sign of backlash ( Figure 15b). Average displacements (red points) for 3 consecutive steps are very stable for both motors and over the complete range ( Figure 15a). Step size appears noisier for forward steps compared to backward. This result remains to be investigated to separate effective step variations from acquisition and post -processing noise. Figure 15 – Step Size Analyses The step size distribution appears to follow a normal law with a standard deviation of 4.3 nm for the nominal motor and 4.9 nm for the redundant motor ( Figure 15 b). In other words, the variation for more than 99% of steps is lower than ± 10 nm. Averaging on 3 steps decreases significantly the step size scattering (less than 12% variation for 3 steps averaging). Lifetime Test The lifetime qualification was still underway at the time of this document. Air lifetime is complete and shows a slow decrease of the average step size. Less than 4 nm of step size decrease has been verified after 1 million steps (320 times the cycle life ). The step size distribution shows a comparable standard deviation before and after the lifetime test in air. Figure 16 - Step Size Against Lifetime (a) Step size against the stroke (Flight Model FM1) (b) Step size for nominal & redundant motor averaged on 5 steps (Flight Model FM1) (a) Approximative step size against lifetime (Qualific ation model – cycles in air ) (a) Step size distribution over the operational stroke before and after 1.2 million steps

Document

AMS_2020.pdf

11 Test Bench Verification for Nanoscale Stability A major difficulty with the stability test is that a temperature variation of 0.1K is sufficient to create a position deviation higher than 70 nm on the test bench references. First, the test bench intrinsic stability , without BSMA , was verified . For this purpose, an aluminum board was used instead of the BSMA to measure the drift ( Figure 17). Figure 17 - Test Bench Stability Verification without BSMA Figure 18 - Test Bench Stability without BSMA The center position show s a very stable beh avior while the side references were equally drifting ( Fig. 18a). The conditions of the test show a stability of around ± 100 mK over more than 200 hours ( Figure 18b). The phenomena explaining this drift is still under investigation. The difference between the central mirror and the reference significantly reduce short term thermal noise. A second verification was to ensure that thermal stability was below 100 mK in short -term (30 min) and 300 mK in long- term (60h). The Figure 19 shows the temperature prof ile during the stability test presented on Figure 20 (FM1). Reference measurements are then used to compensate test bench thermal expansion. This method was found necessary to verify stability requirements. However, it remains difficult to compensate very fast or very slow variations due to complex heat propagation in the test bench. For this reason, it remains key to reduce temperature variations to a minimum. (a) Position stability (b) Thermal stability of the dummy plate

Document

AMS_2020.pdf

12 Figure 20 - Thermal Measurement During Stability of FM1 Stability Test One crucial objective of this test bench is to verify the stability of the BSMA position over time. The first step of the test is to place the BSMA in stable environmental conditions (vacuum, regulated temperature) and then wait more than 60 hours for test bench and BSMA internal components to stabilize. A single step is then performed, and position is monitored for another 30 hours. The stability test shows that the ± 150-nm stability requirement over 24 hours is achieved. Figure 20 - FM1 Post-Processed Position during S tability Test (Difference between output shaft position and left reference) Figure 20 shows the results of the stability test for the flight model FM1. A difference between the output shaft position and references is performed in order to remove the main effects from temperature variation. Moreover, a correction slope is removed from the test bench verification test without BSMA described in the previous paragraph. The post -processed stability curved shows an exponential decrease, then the step and a rather stable behavior. A possible explanation for the exponential effect at the beginning would be the PPA ceramics stabilization following air -vacuum transition. It is interesting to note that the step is not followed by a loss of stability. Mechanical Environment Testing Random vibration testing was performed on the qualification model and flight models. Three axes were tested. Flight models were protected against co ntamination during the test by a plastic film.

Document

AMS_2020.pdf

13 Figure 21 - BSMA Random Vibration Testing on CTEC E lectrodynamic Shaker (Flight Models) The power spectral densities presented on the Figure 22 were applied on all 3 axes. The RMS levels were gradually scaled to different amplitudes up to the maximum RMS values shown in the Table 1. Figure 22 - Random PSD Profiles for M echanical Environmental Testing The eddy current sensors embedded in the BSMA were monitored during the tests to detect any change in position of the rotor. Above a given vibration level the BSMA rotor moved slightly during the Y and Z axis vibration tests (Table 1). It is interesting to note that no damaged occurred during the tests and neither the step size or stroke were influenced when comparing the functional tests before and after vibration. Investigations are still ongoing for a better understanding of the rotor movement. Table 1 – BSMA reliability against vibration levels BSMA Excited axis Random level (g RMS) Actuator damaged Step size or performance change Position shift during test OX 5.7 No No No OY 4.5 No No No 6 No No Yes OZ 8.25 No No No 11 No No Yes

Document

AMS_2020.pdf

14 Conclusion A new patented piezoelectric motor is presented with associated experimental results achieved on the qualification and flight models. This motor is undergoing qualification to be used inside IASI-NG Instrument onboard METOP -SG satellite. The major requirements for the BSMA we re fulfilled: • Movement: 40- nm resolution step motion over a ± 40-µm range with up to 20- N force • Position stability experimentally verified: less than 100nm variation over 48h, • Launch capability without launch- locking mechanism : No damage at all tested levels, some movements occurred for highest levels. Acknowledgment The authors thank Fr édéric DI GESUS and Francois FAURE from AIRBUS DS as well as Laurent CADIERGUE from CNES, for both their technical and financial support. References 1. S. Henein, P . Spanoudakis, P .Schwab, I . Kjelberg, L. Giriens, Y . Welte, L .Dassa , R. Greger, U .Langer, Design and Development o f The Point -Ahead Angle Mechanism For The Laser Interferometer Space Antenna (LISA ), Proc. Conference ESMATS 2009 2. F. Barillot, K. Benoit, C. Belly, A. Guignabert, O. Freychet, Fine Stepping Piezoelectric Actuator (FSPA) for IASI- NG, Conference ACTUATOR, Bremen (G), 25 -27 june 2018 – Proceedings B5.5 p26 3. Stepping Piezo Actuators (SPA) , Cedrat T echnologies , 2019, from https://www.cedrattechnologies.com/en/products/piezo- motors/stepping- piezo- actuators.html 4. C. Belly, T. Porchez , M. Bagot, F. Claeyssen, CEDRAT TECHNOLOGIES, Improvement of Linear and Rotative Stepping Piezo Actuators using design and control, B2.3 Proc ACTUATOR 2012, Pub Messe Bremen (G), June 18- 20, 2012 5. T. Porchez, F. Barillot, C. Belly, Nanometric positioning with piezo actuator and high stability strain gages , Proc. Conference ESMATS 2015

Document

AMS_2020.pdf

15 Spectrometer Scan Mechanism for Encountering Jovian Orbit Trojan Asteroids Kenneth A. Blumenstock*, Alex ander K. Cramer *, Joseph C. Church* , Jason A. Niemeyer*, Fil A. Parong*, Sam Zhao*, Nerses V. Armani **, and Kenneth Y. Lee* * Abstract This paper describes the design, testing, and lessons learned during the development of the Lucy Ralph (L’Ralph) Scan Mirror System (SMS), composed of the Scan M irror Mechanism (SMM), Differential Position Sensor System (DPSS) and Mechanism Control E lectronics (MCE) . The L’Ralph SMS evolved from the Advanced Topographic Laser Altimeter System (ATLAS) Beam Steering Mechanism (BSM) , so design comparisons will be made. Lucy is scheduled to launch in October 2021, embarking upon a 12- year mission to make close range encounters in 2025 and 2033 with seven Trojan asteroids and one main belt asteroid that are within the Jovian orbit. The L’Ralph instrument is based upon the New Horizons Ralph instrument, which is a panchromatic and color visible imager and infrared spectroscopic mapper that slewed the spacecraft for imaging. T he L’Ralph SMM is to provide scanning for imaging to eliminate the need to slew the spacecraft. One purpose of this paper is to gain understanding of the reasoning behind some of the design features as compared with the ATLAS BSM. We will identify similarities and differences between the ATLAS BSM and the L’Ralph SMM that resulted from the latter’s unique requirements. Another purpose of this paper is to focus upon “Lessons Learned” that came about during the development of the L’Ralph SMM and its MCE , both mechanism engineering issues and solutions as well as Ground Support Equipment issues and solutions that came about during the validation of requirements process . At the time of this writing, the L’Ralph S MM has been flight qualified and delivered to the project . Evolution of the L’Ralph Scan Mirror Mechanism Let us consider the ATLAS Beam Steering Mechanism ( BSM ) depicted in Figure 1 as our basis of comparison [ 1]. The ATLAS BSM ’s purpose is to point rather than to scan. It ha s two degrees of freedom provided by a custom flexure design, locating orthogonal axes of rotation behind a relatively heavy glass mirror with a dielectric coating. With the flexure axes location well behind the mirror , the actuators provide some counterbalancing, but the tungsten counterweight attached by its titanium shaft is the predominant means for balancing. Four custom voice coil actuators with redundant windings provide two-axis actuation and damping. A non-redundant inductive sensing system with two axes incorporate s two pairs of differential sensors which view aluminum target areas at four locations of the moving plate behind the mirror. The angular range of each axis is ± 5 milliradians ( mrad) . On-orbit operating temperature is 10 °C to 35 °C. The L’Ralph SMM depicted in Figure 2 requires only a single axis of rotation with an angular range of ±36 mrad , more than seven times the BSM angular range. The mirror is lightweighted aluminum rather than glass with the sensor targets incorporated into the backside of the mi rror. The i nductive displacement sensor technology has a full range linear displacement resolution compatible with both BSM and SMM, though in terms of angular resolution, BSM and SMM are quite different since their angular ranges are quite different . Due to SMM required redundancy, t wo sensor pairs were moved as close together as t he target diameters would allow. This achieved the needed angular range and resolution, such that t he inductive sensor pairs operated over their full linear displacement range of ±0.25 mm. With a single axis system, it was convenient and beneficial to locate conventional flex pivots on a rotation axis that passes through the mirror. This allowed locating the axis such that mass was balanced without the need for a shaft and counterweight, saving significant mass and eliminating additional structural dynamics. * NASA Goddard Space Flight Center, Greenbelt, MD ** ATA Aerospace LLC, Greenbelt, MD Proceedings of the 45th Aerospace Mechanisms Symposium, NASA Johnson Space Center, 2020

Document

AMS_2020.pdf

16 The flex pivots selected were typical brazed type rather than the higher strength electron- beam welded type. Elect ron-beam welded type flex pivot fatigue life is recommended to be derated according to a technical note by Donegan, Richard J., “Weld versus Braze,” (n.d.) at the Riverhaw k Co., Inc. website [2]. It has become a common practice of the GSFC Electro- Mechanical Systems branch to incorporate significant damping into actuators of precision pointing and scanning mechanisms as a means to reduce the effect of both external and internal disturbances, reducing controller burden, thus improving success in meeting performance requirements . In linear actuators, t his is accomplished by a pair of connected copper sleeves that surround both the OD and ID of the bobbin. Often, it would be desired to have even higher damping than we incorporate, but it can become impractical due to actuator size limitations . Scan rates must be low enough to incorporate high damping such that actuator power to overcome damping is not significant. For the SMM, the actuator is a stretched version of the BSM actuator in order to meet SMM stroke needs , but damping is roughly doubled due to a much lower on- orbit operational temperature range of -120°C to -89°C, which reduces resistance thus increasing damping. At the low temperature extreme, the system is nearly critically damped. The SMM Mechanism Control Electronics (MCE) incorporates a Field- Programmable Gate Array ( FPGA) . The Jovian orbit puts the spacecraft rather far from the sun, reducing solar array effectiveness . This resulted in the imposition of a power requirement for the MCE of less than 4 watts , significantly lower than the BSM MCE which required 13 watts . Fortunately, the sensor system is a relatively low power device requiring 0.4 watt. The BSM MCE utilized high resolution ADC s with low -noise op amps for the feedback signal. Identical components were unnecessarily used in greater quantities for telemetry, along with multiplexers, all of which are rather power -hungry components. For the SMM MCE , significantly reducing the number of telemetry signals , implementing standard noise op amps, and lower resolution ADCs with built in multiplexing, reduced power considerably . For the controller feedback signals, the heritage components were kept to maintain optimal closed- loop performance . Power savings also resulted from the reduction in components since only single -axis control was needed. Substituting a low quiescent power amplifier for driving the actuator pair saved an additional 0.5 watt over the previously used power op- amp. Scan Mirror Optical Challenges The primary challenge of the Scan Mirror design depicted in Figure 3 was to satisfy the flatness requirement of the optical surface, 45 nanometers RMS, whi le balancing size and inertia. In order to minimize thermal distortion effects, the primary material for the structure of the Telescope Assembly, including all mirrors, was selected to be aluminum 6061- T651. All components which contribute to the system alignment were thermal cycled during the fabrication process to provide dimensional stability in the oper ating temperature environment. The optical surfaces were diamond turned and then silver coated for optimal reflectance in the specified wavelength range. The Scan Mirror geometry was selected to minimize flatness impacts from gravity release, diamond turning “fling ,” assembly, and on- orbit thermal gradient effects. Pocketing and tapering of the mirror were used to reduce mass while maintaining stiffness. All of these sensitivities were predicted and the design guided by finite element and other analysis tools . The Scan Mirror also had to possess features for mechanical interface, alignment references and fiducials, as well as conductive targets within the back of the mirror for the inductive displacement sensors. The Scan Mirror is supported by three a luminum blade -style flexures to provide a secure and reliable load path that is tolerant to mounting process imperf ections and thermal gradients. Alloy 7075- T651 was selected to provide the strength necessary to survive launch loads without the coefficient of linear thermal expansion ( CTE) change that would be necessary with a more traditional t itanium or other alternative material flexure. The flexures feature clearances at pin locations to absorb manufacturing tolerances. The flexures are bolted to the mirror and moving housing with accompanying plates that possess cavities which, during assembly, are injected with epoxy to register the flexure position to the pins. Figure 4 depicts a flexure as well as an accompanying plate configured for this process, known as “liquid pinning.”

Document

AMS_2020.pdf

17 Figure 1. Previous Development ATLAS B eam Steering Mechanism Figure 2. New Development L’Ralph Scan Mirror Mechanism

Document

AMS_2020.pdf

18 Figure 3. Lightweighted Aluminum Scan Mirror Figure 4. One of Three Mirror Mount Flexures with Associated Plate for Liquid Pinning Prototype testing early in the design process uncovered a high sensitivity of mirror flatness to the flexure bolt preload torques. This finding resulted in t wo significant design changes. The first change was to remove mirror material under the flexure installation locations creating an undercut (Figure 3 ) such that strains could not propagate to the opti cal surface directly. This change was effective but did not alone minimiz e errors to acceptable levels. A second change was implemented to result in lower bolt preload in the cold operating environment while also maintaining preload at ambient conditions t o prevent interface gapping under launch vibration. This was achieved using a low -CTE titanium bolt and high- CTE aluminum standoff (Figure 4). Figure 5 provides thermal gradient sensitivity analysis results of the mirror surface. Vibration and low temperat ure interferometric tests confirmed the soundness of this strategy. Figure 6 provides one of the interferometric test results.

Document

AMS_2020.pdf

19 Figure 5. Thermal Gradient Sensitivity Analysis Figure 6. Prototype Flexure Torque Interferometric Assessment Strength testing of the aforementioned liquid pin joints was performed to assess shear l oad capability of the bonds. This testing was critical to the convergence of a design capable of withstanding launch loads. Injection and bond geometry, surface preparation, and adhesive selection were all adjust ed as a result of these tests. The first attempt revealed that the injection inlet and outlet size and cavity depth would only allow very low viscosity polymers to flow , limiting selection to materials whose strength was insufficient. Increasing the size of these features also allowed the use of larger injection needles and more manageable injection pressure. Sanded, grit -blasted, and etched/primed surfac e preparations w ere evaluated. The grit blasting process proved too aggressive for use with these very small samples, as the cavity containing walls were eroded, allowing adhesive to migrate out of the cavity. There was no significant difference between the sanded and etc hed/primed samples, however the latter approach was selected for f light because the primed surfaces were expected to maintain good surface preparation longer than sanded surfaces. An in -process low temperature interferometric evaluation wa s performed after the surfaces we re prepared but before the cavities were injected, s ince the stability of the bond surface over time is critical. Flight design and process selection involved coupon testing, depicted in Figure 7 .

Document

AMS_2020.pdf

20 Figure 7. Liquid Pin Joint Shear Load Measurement Test Setup and Coupons after Joint Failures Scan Mirror System Performance Verification The SMM Engineering Test Unit ( ETU) vibration tests , which included 22 g sine burst, swept sine, and random, were deemed very successful with no evidence of damage to the structure, no degradation of mirror figure, and no damage to the flex pivots . Post-test inspections did however suggest that the neutral angle of the mirror had changed slightly with respect to the fixed housing as a result of flex pivot shift despite being clamped. To correct this issue in the flight design, alignment and preload- angle clocking features were milled into the flex pivots to engage with set screws in the cl amps. Non-destructive X -ray cross- section evaluations of the pivots were performed before and after these modifications to ensure that the brazed joints were not disturbed. Figure 8 is an X -ray cross- section of the flex pivot with the clocking features . The major challenge of performance verification for this system was to accurately measure commanded mirror position across the on -orbit operational temperature range of - 120°C to - 89°C. A primary and redundant DPSS is used in L’Ralph to provide mirror position feedback for closed- loop control. Each DPSS was tested by the vendor at various temperatures within the operating temperature range to validate performance, but that testing was performed with a flat double- sided aluminum target on the moving portion of a linear stage placed between an opposing differential sensor pair on the fixed portion of that stage. The sensor arrangement in L’Ralph is different, with each sensor of a pair arranged side-by-side and its relatively large angular rotation of ±36 mrad might add some non- linearity since the target becomes less orthogonal to each sensor as the mirror moves away from mid- range . Furthermore, each sensor target at the back of the mirror is at the bottom of a counterbore, s o there is possibility of a non-linearity contribution by the cylindrical conductive surface surrounding each sensor. Any non- axial motion of the mirror resulting from flex pivot behavior could be yet another contributor to sensor non- linearity. Therefore, validation of commanded mirror position was a necessity at the mechanism level. The ATLAS BSM also required mirror position validation while in a thermal vacuum ( TVAC ) test chamber. The solution was t o use a n Inter-target Differential Electronic Autocollimator (IDEA) developed by Leviton Metrology Solutions, Inc . This is a very compact optical instrument as compared with a typical autocollimator . One of its features is the ability to measure the mirror angle of interest while calibrating against a reference mirror, which we located on the fixed portion of the mechanism under test . With IDEA looking through a window on the chamber, motion of the mec hanism mount within the chamber is calibrated out by the fixed reference measurements . The IDEA system that was custom developed by the vendor for BSM was limited to measuring an angle somewhat beyond that of the BSM range of motion of ±5 mrad. A new versi on of IDEA was developed by the vendor with the capability to measure with some margin beyond that of the L’Ralph angular range of ± 36 mrad .

Document

AMS_2020.pdf

21 The flight SMM was installed into the TVAC chamber shown in Figure 9 . At specific temperature plateaus, whether at qualification or operational, the SMM dwelled for approximately f our hours in order to attain sufficient thermal settling of the mechanism. The flex pivots provided a constricted thermal path requiring somewhat long dwells to achieve a reasonable temperature gradient that would be good enough for validating mirror angle. Thermal cycle plots at each side of a flex pivot and mirror are shown in Figure 1 0. Controller Performance Chan ge Over Temperature Controller parameters were initially optimized at ambient temperature resulting in a 30- Hz bandwidth. Frequency response measurements were taken at various TVAC temperatures as shown in Figure 11 . It was found that controller performanc e became less optimal as temperature decreased. As a result of reduction in resistivity of the copper damping sleeve, damping increased with decreasing temperature, ultimately by about a factor of 2.6 going from 25° C to -130°C. Though high damping is very beneficial in terms of disturbance rejection, it was necessary to optimize the controller within the SMM operational temperature range. Thus, controller parameters were chosen to provide optimal performance at operational temperatures. A process of system identification was performed while undergoing TVAC to determine damping as a function of temperature. Fortunately, controller performance, though not optimal at 25° C, was adequate to perform ambient testing without the need to change controller parameters. Figure 8. Flex Pivot X -Ray Cross -Section Figure 9. TVAC Chamber before Closure

Document

AMS_2020.pdf

22 Figure 10. Fixed Pivot Clamp, Moving Pivot Clamp, and Mirror Temperature Transitions Figure 11 . Closed -loop Bode Plots of the S can Mirror System at Various Temperatures

Document

AMS_2020.pdf

23 Mirror Position Sensor System Non-Linearity and Thermal Error Correction Non- Linearity Correction The original MCE FPGA design had only gain and offset terms for converting DPSS voltage to mirror angle. It was decided to expand this to a third- order polynomial correction to compensate for higher -than- expected non-linearity in the DPSS output. The DPSS tuning process balances non- linearity, thermal stability, and resolution; the latter two were deemed more important to correct in hardware, though non-linearity could be corrected in firmware. At the mechanism level, calibration was performed with our autocollimator as a reference. At the instrument level, calibrations will be performed by scanning across a known star field prior to each encounter. This correction is a polynomial which is computed on each 5 kHz position sample before it is passed to the feedback input of the mechanism’s closed- loop controller. Performance testing at each environment started with sensor recalibration, where static DPSS voltage measurements were compared to autocollimator measurements at over 200 points across the range of motion. The remaining scripted test s were divided into three parts; stability, repeatability, and scan tests. For the stability tests, the SMS was commanded to dwell for several minutes each at various fixed positions across the optical range of the mirror wh ile data such as pointing stability, pointing error, and standby power were captured. R epeatability tests commanded the SMS to a series of test positions in a cyclic manner to ascertain how repeatably the mechanism was able to position the mirror. Finally, the SMS was commanded to scan across several optical ranges at several scan rates to obtain data such as optical smear, torque, and scan power. These tests were automated, and test results were trended across the entire mechanism test campaign to verify the performance of the mechanism against system requirements. Mechanism performance testing showed all requirements were achieved except for one: the p ointing resolution requirement. DPSS gain and non- linearity vary with temperature, but because of recalibration prior to each test, there was no noticeable performance degradation across the range of qualification temperatures. The pointing resolution issue appears to be a result of thermal gradients present during the sensor calibration step, which ul timately led to large residuals in the calibration data and thus inaccurate sensor parameters for the controller. To obtain valid calibration parameters, it was necessary wait until mechanism temperature was stable to within ±1.5 °C of the desired temperature plateau. As a result, there were consistent pointing errors, especially at the end- of-travel positions where non- linearity is higher, and least -squares polynomial fitting has the highest residuals. Rate and smear are the driving requirements for this mechanism, so relaxation of the pointing resolution specification was deemed acceptable by the project. Thermal Error Correction Initial tests of the ETU SMS reveal ed a drift in mirror position with a time constant on the order of 20 seconds. In stability tests where the SMS was commanded to step and hold at a mirror position, the DPSS would report the SMM mirror position as stable at the commanded location, while autocollimator measurements would show the SMM mirror had overshot and was slowly settling into position. The inverse was true in tests where a constant current was used to drive the SMM into one of its hard -stop locations: the autocollimator would show the mirror angle to be static as expected, while DPS S measurements would show the SMM mirror had undershot and was slowly drifting to the expected location. Further testing showed that the initial amplitude of this drift scaled linearly with the size of the step: a larger step meant a larger initial error. Stepping from one end of travel to the other, a 70- mrad motion, led to approximately 200 µrad of initial error. This drift turned out to be a ther mal phenomenon correlated with self -heating of the sensor heads themselves. An experiment was run by the vendor where a single sensor head was heated approximately 1°C, which resulted in a voltage error equivalent to less than 1 mrad, decaying with a simil ar time constant to what was seen in the SMS . The vendor found by measurement under normal operation, power dissipation of a single sensor head changed by a few milliwatts . Based upon the sensors head ’s approximate thermal resistance, moving the SMM from one end of travel to the other would lead to a fractional °C change in temperature for a sensor pair. Based on the results of the sensor heating experiment, that temperature delta would cause a voltage error consistent with what was obs erved in SM S

Document

AMS_2020.pdf

24 testing. Further experimentation by the vendor included measur ing the temperature change of the sensor heads with thermocouples for full -range moves of their test fixture. Results from this testing show both that full-range steps cause the temperature deltas predicted by the sensor head heating test, and that the time constants of both the temperature change and position error are roughly correlated. To minimize impact on the existing controller design, correction for this drift was handled wit hin the FPGA firmware but outside the feedback loop. An algorithm that predicted the sensor drift was summed with the position command input . The results in Figure 12 show a comparison of full -range step response utilizing the correction algorithm to correct for the sensor drift (“Cmd Shaping” in orange) and without it (“No Shaping” in blue). The top plot shows voltage measurements from the DPSS converted to angle, while the bottom shows angle measurements from the autocollimator. When tu ned properly, correcting the sensor drift in this way was able to reduce the overshoot error by a factor of 10, from 150 microradians ( µrad) down to 15 µ rad. TVAC testing has shown that the parameters of this correction vary both with pressure and temperat ure. Analysis of that data is ongoing. Figure 12 . Command Shaping Performance Improvement Mechanism Control Electronics Development Strategy Aggressive schedule demands required a compressed MCE development schedule. Three distinct builds were designed: An Engineering Model (EM), an ETU, and flight . This design flow allowed early design validation, while providing the necessary hardware for interface and controller development and testing. The first EM MCE build utilized commercial equivalent parts instead of flight grade parts. Dual footprints were incorporated into the layout to accommodate package differences between flight grade and commercial parts. To prevent schedule delay due to a roughly 1- year electronic part lead time, an alternate (but functionally identical) part was used for the power op- amp. A reprogrammable FPGA module was used for development in place of the one- time-programmable FPGA that wo uld be used on the flight boards. A build plan change added t wo EM MCE b oards to facilitate controller testing and development, and interface testing. This also meant the schematic and Printed Circuit Board ( PCB) layout were verified before ETU/ flight and allowed for a dry run of the assembly process. Board l evel electrical checkout and functional testing procedures were also developed at this stage. A minor net -swap issue was found in the schematic design and fixed with a white wire on the EM, and the schematic and layout were updated for flight PCBs. Since the SMM was not available early enough for initial EM board testing, a mechanism simulator was built. The structure was 3- D printed, and commercial flex pivots and a prototype aluminum mirror were used

Document

AMS_2020.pdf

25 to closely mat ch the expected plant dynamics. Commerci al voice coil actuators and a commercial DPSS allowed open- loop and closed- loop testing. This rapid prototype allowed controller and FPGA development to move forward months before the ETU SMM would be available. It was also useful as a stand in for the ETU SMM and invaluable for testing the optical verification setup. Views of the mechanism si mulator are shown in Figure 13 . The ETU builds utilized parts appropriate for environmental testing. The second ETU build was assembled with the updated flight PCB. By working out assembly and test procedures on the EM and ETU builds, it was possible for flight assembly and test to be completed very rapidly. Because the flight FPGAs are not reprogrammable, a flight FPGA was burned and installed on an EM board first to verify the design was successful , buying down risk. The flight MCE is shown in Figure 14 . Figure 13 . Mechanism Simulator Figure 14 . L’Ralph Flight Mechanism Control Electronics Flexibility of the Controller w as Crucial to Project Success As previously discussed, the digital FPGA- based controller allowed for correction of position sensor system non-linearity and thermal error. It also facilitated system identification at low temperature to determine damping by allowing adjustment of controller parame ters. However, at the start of the project, a digital controller was not the baseline. During the proposal phase, an analog controller with a simplistic digital section was chosen as the baseline. One reason for that selection was that a digital controller utilizing an FPGA was deemed too power hungry

Document

AMS_2020.pdf

26 to meet the power constraints for this mission. Power had been originally determined using values from an FPGA- based digital contr oller used on another project. A second reason is that a grassroots cost estimate resulted in a significantly lower cost for the an alog design over the FPGA design. Once the project was awarded, the product design lead, the primary author of this paper, had tremendous concern about lack of flexibility and resulting excessive risk of the baselined MCE approach. While there were perceiv ed benefits in terms of power and cost for the baseline choice, subsequent analysis found that power could be met with an FPGA -based design, and further review found that manpower beyond what was baselined would be required to develop the seemingly simple discrete electronics. The project ultimately agreed to change the plan and develop an FPGA -based MCE . As a lesson learned, it became apparent near the end of th is project how important this decision was. If the MCE was not FPGA -based, it would have lacked the flexibility to correct position sensor issues as well as assist with system identification when in TVAC . Without the FPGA -based MCE , some crucial specifications would not have been met thus negatively impact ing the science. As a rule, i t is of critical importance that the design architecture be correct from the beginning. With today’s schedule and budget constraints, it is very difficult to change course when problems arise. Thus, diligence must be exercised to get it right the first time because there is typically not time for an additional iteration. Sound arguments and p erseverance are necessary ingredient s to gain project approval if it becomes necessary to change the path forward to that with th e lowest practical risk. Conclusion The L’Ralph SMS team had the benefit of leveraging the previous ATLAS BSM development. The SMS at first was thought to be a straightforward task since compared to the BSM, it appeared to be less challenging. The SMS would be a single axis system rather than a dual axis system, utilize off-the-shelf flex pivots rather than require development of a custom flexure, would have a similar DPSS and a similar MCE, and was specified to have about seven times coarser resolution than the BSM. Yet, the requirements were quite different resulting in a n SMM that looks very different from the BSM, and a new set of development challenges came about. The team was highly motivated, enjoyed the challenges, and gained a new set of lessons learned that we are sharing with the aerospace mechanisms community via this paper. T he team can be proud of the SMM final product, which support s an incredible mission that will advance our knowledge of planet formation. Figure 15 and Figure 16 are views of the flight SMM. Acknowledgements Michael G. Edick ***, Theodore J. Hadjimichael*, William M. Hansell **, Douglas B. Leviton ****, David W. McClaeb* , Joseph C. McMann*****, Armando Morell**, Matthew A. Owens ***, Gary L. Sheridan** , and Patrick L. Thompson* . ____________________________ * NASA Goddard Space Flight Center, Greenbelt, MD ** ATA Aerospace , LLC , Greenbelt, MD *** Florez Engineering, LLC , Laurel, MD **** Leviton Metrology Solutions, Inc., Boulder, CO ***** Northrop Grumman Corp. , Greenbelt, MD

Document

AMS_2020.pdf

27 Figure 15. L’Ralph Flight Scan Mirror Mechanism Front View Figure 16. L’Ralph Flight Scan Mirror Mechanism Rear View References 1. Blumenstock, Kenneth A. , et al. " ATLAS Beam Steering Mechanism Lessons Learned." Proceedings of the 43rd Aerospace Mechanisms Symposium, (May 2016), pp. 1-14. 2. Donegan, Richard J. “ Weld versus Braze. ” [Engineering considerations regarding electron beam welded versus brazed flex pivots ]. Retrieved from http://riverhawk.com/weld -versus -braze/ , (n.d.)

Document

AMS_2020.pdf

29 Point Ahead Mechanism for Deep Space Optical Communication Development of a N ew Piezo-Based Fine Steering Mirror Adrien Guignabert *, Thomas Maillard* , Francois B arillot *, Olivier Sosnicki * and Frank Claeyssen* Abstract The purpose of this paper is to present the development of a novel tip- tilt mechanism, with integrated optics, designed for the JPL Deep Space Optical Communication (DSOC) module of the up coming Psyche mission (2022 launch). This paper presents the design, assembly and tests of the produced models. Regarding the design phase, an emphasis was put on the mirror calculations to ensure that the required flatness would be maintained after integration, and that the part would withstand the thermal/mechanical envi ronment. The actual optical measurement s performed after assembly are also presented. The qualification results for a new alpha- case removal process for titanium parts are presented. Tests results are especially interesting regarding the temperature behavi or of the mechanism, impact on the strok e, and strain gage sensor feedback. Introduction In the up coming NASA Psyche mission (2022 launch) , JPL is planning the assessment of a first Deep Space Optical Communication (DSOC) module. In this module, a Point Ahead Mechanism (PAM) aims at steer ing the optical downlink signal towards anticipated earth position during DSOC communication phases. As a background, f or 20 years, Cedrat Technologies ( CTEC ) has provided various piezoelectrically - actuated Beam Steering Mirrors as well as Fast Steering Mirrors for space missions (PHARAO for CNES , ATLID for A irbus DS) as well as for optronic equipment in defense [1-5]. More recently, CTEC has also been active in Free Space Optical Communication with a new large- stroke Fast Steering Mirror [4]. In this context , CTEC was subcontracted by L3Harris, to design, manufacture and test the performance of the PAM engineering and flight models for JPL PSYCHE DSOC . The developed PAM is a new tip- tilt mechanism based on low -voltage Amplified P iezoelectric Actuators (APA®), exploiting its space heritage. This paper presents the design, assembly and tests of the produced PAM models, covering the involved technologies and failure modes: piezo materials for actuation, strain gages for indirect angular position sensing, mechanical parts treatments, tip- tilt mechanical structure, mirror flatness , etc. Mechanism Design Specifications and T imeline The main specifications for this mechanism were to ensure an angular stroke of ± 2,8 mrad throughout the full operational temperature range of the mission ( -25/+50°C full perf, - 40/+65°C reduced perf ) and a mirror surface flatness under 63 nm while remaining inside a very limited volume and surviving launch vibrations . The project really started in summer 2018 with a preliminary feasibility study aimed at validating these specifications , which ended positively in Fall 2018. The initial design was fi rst based on ATLID mechanism but specific requirements made it progressively , noticeably different in the end. The required schedule for the final mechanism development, i.e, . delivering Flight models less than 1 year after the actual project start, was ve ry unusual and challenging. * Cedrat Technologies, Meylan, France; Adrien.guignabert@cedrat -tec.com Proceedings of the 45th Aerospace Mechanisms Symposium, NASA Johnson Space Center, 20 20

Document

AMS_2020.pdf

30 Mechanism Overview The piezo actuators are wired in 2 push- pull configurations (1 per axis) to allow direct mirror rotation control. The PAM itself is composed of the following parts: - A bracket baseplate (in aluminum): The APA® are fixed on it with screws. - 4 APA® (in titanium): They provide the required displacement and are fixed to the baseplate and to the mirror support via flexural pivots. The APA® are equipped with Strain Gauge ( SG) sensors that are bonded in place - 4 circular pivots ( in titanium ). - A mirror support (in INVAR) which holds the mirror. - A guiding blade ( in titanium ) soldered onto the central cylinder that stiffens the assembly. - A Silicon Carbide (SiC) substrate- based mirror from Mersen OptoSiC® Figure 1. PAM Overview Including Piezo and O ptical Technology from Cedrat Technologies Strain Gauge Sensing In order to be able to monitor the mirror angle, an indirect solution using strain gages placed on each piezo actuator is selected, based on space heritage from other projects, especially ATLID on this matter, which enabled an important development on SG assembly process. The initial SG redundancy requirement was lifted, because it required an important wiring complexity (32 instead of 16 wires), amongst other additional constraints. The project used constantan, 350- ohm SG. There are 2 S G per piezo stack , mounted in one full Wheatstone bridge per rotation axis to maximize the sensitivity while minimizing thermal drift. All SG wires and printed circuit board ( PCB) traces are the same length to limit offset drift. New Piezo Actuator Design The existing CTEC actuators were either slightly too short in stroke or not stiff enough to ensure the mechanism survival during launch. The mechanism consists of 4 APA®, deriv ed from CTEC standard APA120S but specifically designed for the application needs. Based on CTEC space heritage, the APA® shell was made from Ti6Al4V titanium, allowing a theoretical infinite fatigue lifetime in the specified operational conditions and an interesting stiffness/mass ratio. Another benefit of the use of titanium rather than steel as used in standard products , is the reduction of the thermal stroke effect due to a better CTE match (9 ppm/K) to the piezo stack (typically - 3 to + 1ppm/K) compared to a high performance stainless steel for example (10- 11 ppm/K) . A total of 18 APA® were assembled and tested, the results are indicated in Table 1.

Document

AMS_2020.pdf

31 Table 1. PAM Actuators Measured Results Full stroke (170 V pp) 1st coupled resonant frequency µm Hz Average measured 149.2 6963 .6 Standard deviation measured 2.0 60.9 Design value (worst case) 130.4 6151 Difference measureme nt/design value 14% 13% The design values are based on expected worst case parameters , i.e., low piezo gain and stiffest shell when calculating the stroke. A conservative approach was used to ensure that the required mechanism stroke would always be reached, which resulted in average margins of 14% for the stoke and 13% for 1st resonant frequency . Integrated Optics – Mirror Development The PAM is designed around one of its core components: the mirror. For this new project, CTEC used its experience on piezo- optical aerospace projects and collaborated with partners in order to fully integrate the optical hardware development, from the SiC mirror design to its integration and optical verifications. Mirror design One of the main design constraints of an embedded optics mechanism is to keep the mirror surface deformation to a minimum in order to limit the induced optical wave front error below the requirements. I n this case, a maximum of 63- nm mirror surface flatness is the requirement . In order to ensure the specification would be reached, CTEC developed tools and performed specific simulations in the early design phase, specifically includin g verification of induced surface figure error caused by mechanical biases and thermal deformation , as well as optimi zation of mirror shape and dimensions. Figure 2. Mirror Surface Deformation Simulation Meshing and Boundary Conditions (Screw Tension and Torque Case) (Left ) - Mirror Surface Deformation Evaluation Results ( Right ) The mirror design process is based on a finite element simulation of each mirror deformation contributor: mirror clamping to its support, actuator height variation, screw tightening and thermal operational. Each case used representative boundary conditions. The resulting displacement on each of the mirror surface points is then exported. A MATLAB program is then run to post process the results from the simu lation. Extracting the RMS reference plane from this data set, it calculates the distance of each point to this plane, the peaks and valley of the deformed mirror and the resulting RMS deformation. The specific contribution of each evaluated case on the mi rror deformation is then summed up to give an estimation of the total expected mirror surface deformation.

Document

AMS_2020.pdf

32 PAM development also included a regular mirror optical verification at different stages of the assembly , with the intention of having the ability to stop the process should a mirror appear to be out of the acceptable range or show significant surface figure changes from one step to another. With the recent CTEC experience on this matter, such regular controls also allowed us to learn a lot regarding the impact of each assembly step as well as providing a safer project assembly process . Figure 3. Mirror interface (left) and surface verification with interferometer (right) Mirror Verification Results A total of 4 PAM were produced and the mirror Reflected Wavefront Error, Peak -to-Valley (RWE PV, basically 2 times mirror surface flatness error SFE ) was measured in different phases, i.e., mirror in the initial free condition, once clamped on its mount , and eventually after final integration in mechanism. Table 2. PAM Mirror Surface Flatness Measurement Results PAM model 1.Free coated mirror 2.Mirror clamped on mount 3.Mirror integrated in final mechanism Calculation: Mechanism contribution to mirror deformation RWE PV (nm) RWE RMS (nm) RWE PV (nm) RWE RMS (nm) RWE PV (nm) RWE RMS (nm) RWE PV (nm) RWE RMS (nm) EM 38.1 10.1 34.9 8.9 41.3 10.6 3.2 0.5 EQM 22.2 6.2 27.2 6.6 38.7 10.3 16.5 4.1 FM1 23.2 5.5 19.4 4.6 26.6 6.0 3.4 0.5 FM2 28.9 8.1 33.8 8.2 34.1 9.5 5.2 1.4 Average 28.1 7.5 28.8 7.1 35.2 9.1 7.1 1.6 With an average 35. 2 nm and a maximum of 41. 3 RWE PV, the <127- nm RWE specification was reached with significant margin. The last column evaluates the proper mechanism contribution to mirror deformation based on the difference between final and the initial RWE PV measurements. The mechanism average contribution to the mirror deformation appears to be limited to 7. 1 nm PV and 1.6 nm RMS .

Document

AMS_2020.pdf

33 Eventually, we can compare these results to the expected worst -case deformations evaluated during the design phase. The t hermomechanical contribution could not be evaluated so only mechanism induced deformation are considered . Table 3. Comparison of mirror measurements and expected design worst cases Average measured impact of mechanism on mirror RWE (nm) Average measured impact of mechanism on mirror SFE (nm) Evaluated worst case for mirror SFE (mechanism integration only) (nm) PV 7.1 3.5 5.7 RMS 1.6 0.8 3.1 All measured values are within the expected range of mirror deformation, indicating that the conservative simulation approach was correct in that case. Mechanism Production and Assembly Titanium Parts – Alpha Case Removal Process Qualification The use of titanium for some parts, justified by its interesting mechanical properties and heritage on previous CTEC mechanism s (for similar parts) is not without drawbacks . The main issue is due to the use of wire electro discharge machining (WEDM) for the “flexible” parts manufacturing, i.e. , the actuator amplification shell and the guiding blade at the center of the mechanism. This manufacturing technique is required due to the parts geometry. Figure 4. WEDM titanium parts, guiding blade (lef t) and actuator shell (right), different scale The WEDM locally heats the material which can induce the formation of an alpha case on the surface, with the adverse effect of lifetime reduction (up to 30% according to [6]) , and possibly unexpected failures . This alpha case, usually a few µm thick, can be removed through proper chemical etching. For this mechanism, the already qualified manufacturer was not compatible with the required schedule (overload), hence a backup option had to be identified and qualified. The new supplier proposed an acid etching process . The qualification process was the following: • Machine a set of guiding blade and actuator shell, identical to flight design (supplier, process, material batch, dimensions) • Perform an alpha case an alysis before acid etching, parts in initial state, as well as an interstitial hydrogen contamination measurement. • Pass the parts through the acid etching process , adjusted to remove the required thickness • Perform an analysis after the chemical etching: alpha case and interstitial hydrogen contamination • The process is qualified if analys es show no trace of alpha case after etching and no more than 150 ppm interstitial hydrogen contamination.

Document

AMS_2020.pdf

34 Figure 4. Alpha Case Analysis, Before ( Left) and A fter the A cid Etching ( Right)-Optical Microscope (x1000) No alpha case was found on that actuator shells ; on the guiding blade however , the average alpha case thickness was 8 µm (6 to 11 µm). It was decided that a 15- µm thicknes s removal through acid etching would be enough to remove the alpha case. The final analysis after the acid etching indicates that the process efficiently removed the alpha case (see Figure 4), no trace was left on the part. The hydrogen contamination, that can be induced during the acid etching (penetration of hydrogen inside the material compound and local embrittlement) remained within the boundaries (86 ppm for guiding blade and 26 ppm for actuator shells). The process is then qualified and was used successfully to treat flight parts. Piezoelectric Stacks – Lot Acceptance Tests (LAT) As one of the critical items in the mechanism (brittle, sensitive part), the piezo stacks are handled with special care. One single piezo batch is procured for the project, with high quantity margin (at least x2). An LAT is then performed on 4 piezo stacks taken from this batch, prepared (SG gluing & cabling) exactly as flight piezos. This LAT includes thermal cycling and lifetime tests representative to the final envir onment (this includes mounting them in their actuator shell) . Regular basic electrical verifications (capacitance, insulation), stroke measurements are performed before and after the tests to detect any deviation. A final destructive physical analysis is performed to inspect inner features of the piezo stacks mainly to detect potential voids in the ceramic and electrode delamination. Piezo stacks are cut in two and sections are inspected. Figure 5. Solder Inspection, Example of a Crack The destructive physical analysis reveal ed that the material is dense enough, no voids we re detected. Electrodes were perfectly in place and no trace of delamination was found. In compliance with CTEC

Document

AMS_2020.pdf

35 previous experience, small 45° cracks we re detected in the vicinity of the electrodes. The se cracks are believed to be caused by the thermal expansion and contraction of the electrode during soldering. Since they are covered in epoxy potting, th e cracks’ progression is contained. However , one unusually placed and lengthy crack was detected. Unusual by its location, starting from piezo edge instead of from the electrode, and by its length, 2.2 mm compared to the usual few hundred µm. Figure 6. Piezo N°3, Black Wire Side Unusual Crack ( Length 2 .2 mm) Raising questions , investigations were performed but no clear specific root cause was identified. It was found that this crack was slightly visible from the exterior and most importantly not going through the entire stack (staying near the electrode, under the epoxy ). Figure 7. Exterior View of the C rack One possible root cause for this crack is stress concentration due to a slight mispositioning inside the actuator shell and to the selected prestress level . Simulations representing this mispositioning allowed CTEC to map the induced stress concentration and the result is visually similar to the observed crack pattern. Figure 1. Photo of Piezo Mispositioning and S imulation of C orresponding Principal Deformation Map (Cut View) Matching Crack Pattern

Document

AMS_2020.pdf

36 To avoid this problem, the final flight actuator shell will be larger (9 mm instead of 5 mm) than the one used for the LAT, making it impossible for the piezo face to be in contact with an edge like this. Also, as a precaution, all piezo stacks were inspected before assembly in flight hardware and no exterior sign of similar cracking was found . No other unusual findings were identified , and the piezo batch wa s accepted for integration. Mechanism Integration One of the main constraints of this mechanism integration is that the support bracket (black anodized aluminum part) has its interface surface (fixed to the L3 optical array) perpendicular to the mechanism base plane. This geometry severely limits access to the mechanism for torque wrench es, operator hands, pliers etc. Hence the assembly process had to integrate this specific constraint with the unusual approach to preassemble the actuation mechanism outside its final base plate, then transferring it onto the final support bracket. The final mechanism once assembled is shown in Figure 9. Figure 9. PAM EQM P icture After Assembly The most sensitive part of the mechanism is its SiC coated mirror and many precautions were taken to protect it during the assembly. A specific cover (POM- C) was designed and was used for most assembly phases, especially to protect the mirror surface from wrenches or pliers when needed. In order to ensure additional protection during transportations phases (models are transported several times during assembly and test phases), a transparent Plexiglas box was designed. Figure 10. Mirror Protection Cover Once Installed ( Black Part), PAM T ransportation Box

Document

AMS_2020.pdf

37 Cables and PCB I ntegration A total of 16 SG sensor wires and 8 piezo power wires have to be routed from the mechanism to the interface cables. The use of a connector on the PAM side was not possible due to size constraints. The selected design option was an interconnection PCB , similar to what was used in previous CTEC space mechanisms like the ATLID Beam Steering Mirrors(1). The multilayer PCB provides many benefits : • Allows pre- routing of all SG bridges easily • Easier wire handling and more reliable interconnection (from AWG36 SG wires to AWG26 pigtail cable wire, large number of wires) • More control of SG trace length (and impedance) to ensure a reduction of SG offset thermal drift • SG and piezo signals are routed through different layers, insulated and shielded from each other (ground planes in the middle of the PCB ) • Interconnection PCB was used in previous similar spa ce projects with positive feedback Figure 11. PAM PCB routing overview (merged layers for easier visualization) Wires are routed from their starting point on the piezo actuators (SG and power), through the support bracket via holes , up to their dedicated PCB pad. Wire s are regularly secured along their paths with epoxy dots. One of the constraint s of SG wire is for them to have the same length within each W heatstone bridge (one for each axis) to ensure they maintain the same total resistance within the target range. Figure 12. PAM Cabling to Interconnection PCB

Document

AMS_2020.pdf

38 Qualification and T est Campaign Static Performance The 4 mechanism performance parameters are verified at different steps of their acceptance tests. An initial good health verification wa s also performed to ensure that the piezos and SG we re correctly cabled. At the time of this paper publication, only 3 mechanisms were tested, results are presented in Tab le 4. Figure 13. EQM Y A xis Full Stroke Measurement ( Mirror Angle vs P ush-Pull Voltage) and Admittance Sweep for EQM Y axis – Coupled Resonance Frequency Identification Table 4. Static Measurement Results for PAM EM/EQM and FM2 Parameter Required value EM EQM FM2 PAM total stroke at ambient ( -10/+150V) X axis >6 mrad 8.0 mrad 8.7 mrad 8.5 mrad Y axis >6 mrad 8.6 mrad 8.6 mrad 8.5 mrad SG response at ambient - Functional tests FTM -01-B (cf, III,8) Offset X axis / -0.71 mrad -0.27 mrad -0.46 mrad Y axis / -0.76 mrad -0.35 mrad 0.11 mrad Gain X axis / 1.57 mrad/V 1.49 mrad/V 1.76 mrad/V Y axis / 1.38 mrad/V 1.54 mrad/V 1.67 mrad/V 1st coupled resonance frequency X axis >800Hz 1495 Hz 1301 Hz 1257 Hz Y axis >800Hz 1242 Hz 1272 Hz 1272 Hz PAM mass Total mass / 497.2 g 497.9 g 495.1 g The mechanism total stroke is compliant with the specification with significant margin, since design anticipated some stroke loss at cold temperature. The total stroke is similar for all models/ax es, the slight variation observed is expected and linked to the piezo actuators ’ stroke/stiffness variation. SG parameters are noted for information, only the variation of these parameters ( over temperature) matters. Note that SG measurements are taken after bridge output conditioning (5V excitation, 337 and 352 V/V gain for X and Y).

Document

AMS_2020.pdf

39 Thermal Vacuum Testing The Qualification campaign for the EQM included both non-operational (NOP) and operational (OP) thermal vacuum cycling. Target vacuum is <10- 5 mbar. Tests A,B,C are functional tests (stroke, SG and admittance sweep). Figure 14. NOP and OP T hermal Vacuum Cycling Test Schematic The main objective of the OP Thermal Vacuum Cycle is to characterize the variation of PAM stroke (expected loss in cold) and SG performance throughout the OP temperature range. Only the results of this test are detailed in this paper. The OP Thermal Vacuum Cycle consist s of testing the mechanism (full stroke, SG parameters) at different incremental temperatur es (every 10°C). Given that the PAM applied voltage was not constant during the full test (4 days with electronic and room temperature variations ), the stroke result s are normalized with respect to the voltage range applied, called stroke/voltage gain. The stroke results are shown in Figure 15. Figure 15. X and Y Stroke/ Voltage Gain Through OP Temperature Range – EQM Results The stroke/voltage gain plot for both axes show the expected bell shape, with a maximum at ambient temperature and loss at cold and hot temperatures . A 7% loss can be seen at -30°C and around 2% loss at +60°C. A conservative 20% and 15% stroke loss was ass umed in the design phase (based on past experience on worst cases), explaining the high final stroke margin. The other noticeable feature is the thermal hysteresis , which should not be there: it is expected for the mechanism to have the same stroke for same temperature (the stroke vs voltage hysteresis observed is however nominal and expected) . The reason for this is probably a consequence of the fact PAM temperature was not fully stabilized, tests were performed slightly too quickly after reaching target temperatures and resulted in this thermal hysteresis. The actual temperature data is measured on the PAM support bracket (interface) located under the mechanism, piezo temperatures were not monitored to avoid any potential damage/contamination. Hence it is not possible to directly measure or control piezo temperature.

Document

AMS_2020.pdf

40 In order to confirm this thermal delay hypothesis, a simplified thermal -equivalent first order model of the PAM was created, using measured temperature time constants (between PAM interface and the piezos) and a bell -shaped stroke vs temperature look -up table for the piezos. Figure 2. Simplified PAM Thermal Equivalent Model Results (X: T emperature in °C, Y: Stroke in mrad), Left: 10h Cycle Duration ( Close to W hat was Tested), Right: 100h Cycle Duration where Hysteresis is Greatly Reduced. The model can recreate this hysteretic behavior, which appear s to be greatly reduced after greatly increasing thermal stabilization time (x10) . However, g iven that the measurements cannot be automated, an operator has to be there at each step (hence excluding nights and weekends ). The EQM test already lasted 3 days and 30 days of testing is not practical . Options to improve the OP Thermal Vacuum Cycle were discussed with the customer. SG parameters (offset and gain) were also measured for each temperature step, the results for EQM X - axis are shown in Figure 17. Figure 17. EQM X A xis SG Gain vs T emperature ( Left), EQM X Axis SG O ffset vs T emperature ( Right) The results are also affected by the piezo temperature delay . As anticipated, the SG gain is quite stable with less than 1% variation through the temperature range. The offset is directly impacted by the mechanism thermomechanical excursion, which was meas ured at 0 .4 mrad total (0. 2 to 0. 4 mrad on X and Y SG offset). Vibration Test All models except EM are planned to go through vibration tests. The tests consist of a random vibration verification as well as a shock test, for each axis. A low-level frequency sweep is performed before and after each test (to assess potential modal landscape changes ).

Document

AMS_2020.pdf

41 Regarding the random vibration test , the preliminary levels for whi ch the mechanism was analyzed to were 10 grms (20- 2000 Hz span with a 50 to 800 Hz 0.08 g²/Hz plateau). Later in the project, when the updated and refined DSOC full system vibration simulations were performed, the specified levels at PAM interface location had to be increased. The final random vibration levels are much higher and follow the power spectral density (PSD) shown in Table 5. Table 5. Final PSD for PAM Random Vibration Test X -Axis Test Spec Y-Axis Test Spec Z-Axis Test Spec FREQ(Hz) ASD(g²/Hz) FREQ(Hz) ASD(g²/Hz) FREQ(Hz) ASD(g²/Hz) 20 0.100 20 0.1 20 0.1 50 3.500 50 3.5 50 3.5 150 3.500 395 3.5 340 3.5 2000 0.010 2000 0.01 2000 0.01 Grms = 28.03 Grms = 42.0 Grms = 39.5 Updated mechanism simulation showed that th ese new levels remain acceptable, regarding mechanical stress in the parts. The cables and wires, however, we re more affected, raising concerns with the updated vibration levels. Wire and cable epoxy staking upgrade s are being implemented to mitigate the risk of wire damage, mainly by reducing the free length between tie- down locations . Figure 18. Overview of SG Wires, Initial Configuration ( Left) and R einforced Cabling ( Right) At the time of paper publication , full level vibration tests are planned on a representative model. Mechanism Delivery Status At the time of the paper publication , the engineering model (EM) has been delivered. The EQM vibration test is still to be performed for closure of design verification. FM2 acceptance test series has started with thermal vacuum cycling. FM1 was unfortunately irreparably damaged during a subcontracted bake- out failure (decompressive explosion, not implosion, something not expected in a vacuum test), and a new model will be assembled using spare parts and tested. Conclusion In this paper, the development, procurement, integration and test ing of a novel double tilt PAM mechanism is presented. The mirror integration verification method is explained and the comparison with actual measurements indicate that the results were quite reliable for this case. The test results now available are presented and indicate that the PAM mechanism is working as intended. Some measurement artifacts on the OP Thermal Vacuum Cycle (thermal hysteresis) were investigated and some improvements will be implemented for incoming FM2 tests.

Document

AMS_2020.pdf

42 The upcoming tests are the random vibration and shock with increased levels. The cable wiring is being reinforced in order to ensure the mechanism survival. Acknowledgment The authors want to thank all partners involved, among others: Marshall Bernklow and Richard Aigbaeken from L3- Harris -SSG , as well as Dan McDonald and Joseph Kovalik from JPL Optical Communication Laboratory for their support during the project. References 1. R. Le Letty, F. Barillot, H. Fabbro, F. Claeyssen, Ph. Guay, L. Cadiergues, Miniature Piezo Mechanisms for Optical and Space applications Proc ACTUATOR Conf, Pub. Messe Bremen (G), June 2004, pp 177-180 2. E. Prevost, A. Weickman, S. Belmana, F. Bourgain, O. Sosnicki, F. Claeyssen, Beam Steering Mechanism For Earthcare Atmospheric Lidar Instrument Atlid – An Ultra -Stable Piezoelectric Tip Tilt Mechanism, Proc. ICSO, Biarritz, Oct. 2016 3. F. Claeyssen, T. Maillard, O. Sosnicki, F. Barillot, A.Pages, C.Belly, A.Bataille, M.Logeais, G.Aigouy, T.Porchez, F. Bourgain, Beam Steering Mirrors from space applications to optronic solutions , Proc. OPTRO Conf, Paris , Feb. 2018 4. F. Claeyssen, K . Benoit, G . Aigouy, T .Maillard, M. Fournier, Olivier Sosnicki, Large- Stroke Fast Steering Mirror For Space Free -Space Optical Communication, Proc. OPTRO Conf, Paris, Feb. 2020 5. F. Bourgain, O. Sosnicki , C. Belly, F. Barillot, F. Claeyssen, An Improved Accurate Beam Steering Piezoelectric Mechanism for ATLID Instrument , Proc. Actuator 2014 Vol. pp. 293-296 6. T. Mower ., Degradation of titanium 6Al -4V fatigue strength due to electrical discharge machining , International Journal of Fatigue, vol.64, pp 84- 96

Document

AMS_2020.pdf

43 Design, Development and Verification of the METimage Scanner and Derotator Mechanisms Sebastian Rieger * and Armin Jago* Abstract This paper presents the major design choices and lessons learned from the development and verification of the METimage Scanner and Derotator mechanisms. The modification of a standard drive unit design to project -specific requirements is described . This includes a presentati on of design modifications that were implemented to meet long-term storage requirements. Tests with notable lessons learned are explained and the derived lessons learned are discussed. Details on the design and development of the optical components , the METimage Solar Calibration Device mechanism, and the control electronics are not presented in the paper . Introduction METimage is a cross -purpose, medium resolution, multi -spectral optical imaging radiometer for meteorological applications onboard the MetOp- SG satellites. It is capable of measuring thermal radiance emitted by the Earth and solar backscattered radiation in 20 spectral bands from 443 to 13, 345 nm [1]. The instrument is developed by Airbus Defence and Space on behalf of the German Space Administration . The METimage instrument is based on th ree key optical assemblies which include mechanisms. These are the Scanner Assembly , the Derotator Assembly , and the Solar Calibration Device (see Figure 1). All three mechanisms are developed by Airbus and the mechanisms team is embedded in the instrument team. The optical elements and the mechanism control electronics are developed by external partner s. This paper focusses on the Scanner mechanism and the Derotator mechanism. Figure 1. Scanner, Derotator, Solar Calibration Device (left to right , not to scale) As shown in Figure 2, the Scanner mechanism is located at the Nadir -side entrance of the optical head of the instrument . The Scanner mirror is tilted by 45 degrees and reflects the optical beam into the telescope at an angle of 90 degrees. The Derotator is located inside the instrument, between the telescope and the detectors. It rotates at exactly half the speed of the Scanner in order to achieve a regular imaging geome try in the focal plane . During sun calibration phases, the Solar Calibration Device mechanism rotates one of its diffusors such that it is exposed to sunlight; the sunlight is reflected into the field of view of the scanner . * Airbus Defence and Space GmbH, Friedrichshafen, Germany sebastian.rieger@airbus.com / armin.jago@airbus.com Proceedings of the 45th Aerospace Mechanisms Symposium, NASA Johnson Space Center, 2020

Document

AMS_2020.pdf

44 Figure 2. METimage Scanner and Derotator embedded in the optical head of the instrument Design Overview The Scanner mechanism and the Derotator mechanism consist of a drive unit (blue/cyan in Figure 1) and the scanner mirror and derotator mirror assembly , respectively. The Scanner drive unit design comprises two sets of angular contact ball bearings (hyperstatic layout), whereas the Derotat or drive unit comes with a single bearing pair (isostatic layout). The motor and encoder concept is identical to the standard drive unit of Airbus Defence and Space for both mechanisms (Figure 3). This design has been flown on several missions, such as MHS (NOAA, MetOp) and the FY3 satellites; it also forms the design baseline for the MWI, ICI and MWS scanning mechanisms for MetOp- SG. The standard design was adapted to me et the needs of the respective project. Figure 3. Standard drive uni t concept (left) and c ross-sectional views of Scanner ( center ) and Derotator (right) drive unit s Telescope assemblyOptical beam Scanner assembly Warm optical assembly Cryogenic subsystem Derotator assemblyInstrument support structureAperture stopNadir baffle Angular -Contact Ball Bearing Pair Brushless DC Motor Optical EncoderOptics Interface Optics Interface Angular -Contact Ball Bearing PairBrushless DC Motor Optical Encoder Stationary interfaceAngular-Contact Ball Bearing PairRotating interface Rotating interfaceStationary interfaceBrushless DC Motor Angular-Contact Ball Bearing Pair Scanner Derotator

Document

AMS_2020.pdf

45 Adaptation of Standard Design to METimage Requirements In order to reduce programmatic and technical risk, existing (flight -proven) concepts, designs and components were selected wherever feasible. This came with the following benefits: • Use of high TRLs was made and, hence, qualification effort was reduced • Common procurement with other projects could be performed • A well- advanced design was available early in the project • Testing of flight -representative mechanisms/components was possible at an early stage • Flight -grade components of early models could be re- used for flight models The METimage Scanner and Derotator designs are based on the existing standard drive unit concept but were modified in order to meet the specific METimage needs. Performance requirements The performance drift error (PDE) of the Scanner and, to a lesser extent, of the Derotator is one of the key performance parameters of the METimage instrument. The PDE has a direct impact on the image quality as it affects the co- registration of the instru ment . The Scanner PDE during earth view shall not exceed 25 µrad over a period of 10 ms; the Derotator PDE shall not exceed 100 µrad. The PDE is the maximum pointing error in a window of 10 ms, relative to the pointing error at the beginning of this window . Significant effort was spent on both mechanisms and control electronics in order to obtain designs compliant to this requirement . The major mechanism contributors to the PDE are • Bearing friction variations • Motor disturbance torques • Encoder measurement accuracy Bearing friction and friction variation were minimized by reducing the diameter and the preload of the bearings. Consequently, different bearing dimensions were selected for Scanner and Derotator, respectively; the large diameter of the Derotator s haft and the corresponding Derotator bearing size would have resulted in too high friction for the Scanner mechanism. The disturbance torques of the motor occur mainly at the following frequencies (per motor revolution): • Rotor pole number: caused by magnets passing an imperfection on the stator • Stator slot number: caused by imperfections on the rotor passing the stator slots • Product of rotor pole number and stator slot number, divided by their greatest common divisor : motor cogging, caused by the fact that the magnetic field is dependent on the distance between magnets and stator slots The brushless DC motor design of the standard drive unit comes with a keyway at the stator outer diameter (Figure 4). This keyway was introduced as a positioning feature for projects where drive electronics without modifiable commutation angle offset are used. For these cases, the motor needs to be integrated into the mechanism in a p re-defined orientation. This positioning feature is neither needed for the METimage mechanisms, nor can the corresponding motor disturbances be accepted. Therefore, the keyway was removed for the METimage motor design. Figure 4. Motor stator positioning keyway that was removed in the METimage design

Document

AMS_2020.pdf

46 Manufacturing tolerances and imperfections on magnets or stator sheets might cause un acceptable disturbances. In order to mitigate this risk while maintaining cost -efficient motor des ign and manufacturing without unnecessarily tight tolerances , feed- forward harmonics suppression w as implemented into the controller of the mechanism drive electronics. The feed forward allows compensation of known, repeatable disturbances by means of pre- programmed, position- dependent motor current variation. Motor cogging was reduced by skewing of the stator sheet metal stack (refer to Figure 5). Skewing also reduces t he torque constant of the motor slightly . This was acceptable for both the Scanner and the Derotator. Motor disturbances could have been minimized by using an iron-less motor . However, the torque constant of an iron- less motor was expected to be approx imately 40-50% lower than the one of the baseline (iron) motor. This would not have been acceptable from a power and torque budget point of view. Figure 5. Motor stator skewing The Performance Drift Error is calculated from Absolute Performance Errors (APE). The APE is the difference between actual position and target (commanded) position for any given point in time. During flight, the actual position is determined with the absol ute optical encoder that is part of the mechanism. Therefore, measurement uncertainties of the encoder have a direct impact on the APE and PDE: • Low-frequency encoder errors (below the controller bandwidth) are followed by the mechanism. This results in a deviation to the target scan profile. The actual (physical) performance of the mechanism is affected, but this cannot be determined from the position data. • High-frequency encoder errors (above the controller bandwidth) are not followed by the mechanism but corrupt the recorded position data. The actual (physical) performance of the mechanism and, hence, the co- registration of the instrument is not affected . In order to minimize the impact of measurement uncertainties on the mechanism performance, stringent requirements were imposed on the encoder. The performance of all encoders was checked with an external reference encoder during mechanism assembly. The second major performance requirement is off -axis motion of the scan mirror (“wobble”). The ti meindependent (also referred to as asynchronous or random) wobble of the Scanner mechanism shall be <10 µrad peak -peak ; the one of the Derotator shall not exceed 20 µrad. In contrast to the synchronous (repeatable) wobble, the asynchronous wobble cannot be corrected during ground processing. The time- independent wobble is caused by any non- deterministic behavior of the bearings. Nondeterministic behavior of the bearings can be attributed to the balls, which rotate and spin in a un predictable manner. The variation of the ball diameter , as defined in ISO 3290, was identified as the major contributor to the time- independent wobble. Surface roughness is approx imately one order of magnitude lower, and the deviation of the spherical form of the ball has a minor influence as the ball is the softest element in the bearing. Therefore, balls of grade 3 were selected for the bearings . The distance between the rings of the bearing pairs was maximized as much as possible.

Document

AMS_2020.pdf

47 Power consumption requirement s The scan profile of the Scanner and Derotator was optimized for integration time per pixel during earth view. The scan period of 1.728 seconds is defined by the satellite orbit and the spati al resolution of 500 m at Nadir. Consequently, a dynamic scan profile with a low velocity during earth view and sun calibration and a high- acceleration/deceleration phase outside this range was defined (refer to Figure 6). The need to maximize earth view duration was traded against power and exported torques requirements. Figure 6. Scan profile of Scanner and Derotator The combined peak power consumption of the Scanner and Derotator mechanisms had to be less than 40 W. The average power consumption over one revolution had to be less than 8.0 W and 6.0 W, respectively, for the Scanner and Derotator mechanism. A multitude of parameters affect the power consumption of a mechanism. The major ones are: • Bearing friction • Motor losses ( mainly iron and copper losses) • Inertial moments during acceleration phases • Encoder power consumption In order to re- use existing designs as much as possible, t he encoder was not modified. Motor power could have been reduced by approx imately 30% by using NdFeB magnets; however, motor magnets made from SmCo were preferred for programmatic reasons (refer to section Long- term storage). Bearing friction was minimized by sizing the bearings to the actual needed dimension, and by reducing the preload to the minimum possible value that was required to ensure that the mechanisms withstand the launch environment (Scanner: 88- mm pitch diameter, Derotator: 124 mm) . In addition, the Derotator bearing concept was modified for power as well as lifetime reasons : instead of two bearings in hyperstatic layout, one bearing in an isostatic layout was chosen. This was found feasible in a trade of power against performance requirements. Inertial moments that need to be overcome by the motor occur during acceleration and deceleration phases. They are dependent on the rotating inertia and the acceleration. The resulting inertial moment is overcome by motor phase current incr ease and, hence, causes higher power consumption. Together with the instrument team, a scan profile was found that satisfied the instrument needs and for which the Scanner

Document

AMS_2020.pdf

48 and the Derotator were compliant to the power requirements with SmCo magnets . This permitted avoiding NdFeB magnets. Cleanliness and contamination control METimage is an optical instrument. The mechanisms carry optical components and, thus, had to meet stringent cleanliness and contamination requirements. In general, the CVCM (Collected Volatile Condensable Materials ) and RML (Recovered Mass Loss ) limits of Table 1 applied to the mechanisms . Table 1. Particulate and molecular contamination requirements Mass of material concerned CVCM [%] RML [%] > 100 g < 0.01 < 0.1 10 – 100 g <0.05 <1 < 10 g <0.1 <1 The CVCM and RML limits were considered for selection of surface treatments (paints) and lubricants. It was decided not to use polyurethane- based paints as these paints show high outgassing levels over long time. Black Keronite a nd Acktar Frac talBlackTM were implemented for all thermo- optical coatings. A surface modification process developed at Airbus which creates black surfaces with high solar absorptivity ( titanium: 0.98, aluminum: 0.84) and infrared emissivity (titanium: 0.94, aluminum: 0.95) while at the same time providing an excellent bonding pre- treatment ( e.g. for bonding heaters , thermistors ) [4] was not selected by the time the decision for the coatings was made. This was due to the – by that t ime – unclear handling constraints of the surfaces treated with the laser. However, laser process es developed by Airbus were applied for increasing friction coefficient s (refer to section Temperature ranges ) and for performing REACh - compliant bonding pre- treatment for motor bonding [11]. Figure 7. Blackening by surface modification with laser (left: laser -treated titanium prototype , right: microscopic view of laser -treated surface) [4] The heritage bearing lubricant for most scanning mechanisms at Airbus Friedrichshafen is grease Maplub SH051 -a, in combination with phenolic resin cages that are impregnated with Nye 2001a. These MAC - based lubricants were chosen due to their excellent lifetime (in terms of revolutions) which exceeds the lifetime of PFPE -based lubricants by far [2]. However, according to current literature, PFPE- based greases and oils like Braycote 601EF and Fomblin Z25 have a better outgassing behavior; for instance, the vapor pressure of Fomblin Z25 is reported to be one order of magnitude better than the one of Nye 2001a [3].

Document

AMS_2020.pdf

49 Table 2. Lubricant outgassing data Lubricant Vapor pressure at 20°C [mbar] Vapor pressure at 100°C [mbar] TML [%] CVCM [%] Reference Requirement N/A N/A 1 (RML) 0.1 Table 1 Fomblin Z25 1.60E -13 2.80E -09 0.04 0.01 [3] 9.35E -10 (25°C) 3.53E -09 0.17 N/A [10] Nye 2001a 5.33E -12 1.33E -08 0.40 N/A [3] N/A 1.80E -09 0.08 0.12(*) Test [6] (*) Correcting the CVCMs for bulk evaporation losses was not possible due to QCM saturation. Therefore, noncompliance (0.12% vs. specified <0.10 %) was accepted. In order to assess the suitability of the heritage MAC -based lubricants for METimage, an outgassi ng test of the batch of Nye 2001a foreseen for the METimage Scanner and Derotator was performed. This test yielded that the actual outgassing behavior of Nye 2001a was far better than expected and, except for the CVCM, was in the same order of magnitude as the one of Fomblin Z25 [6]. It is not clear whether this observation is valid only for the tested batch, or whether literature data on Nye 2001a are in general too conservative. Temperature ranges The specified non-operating temperature range of the Scan ner and Derotator mechanism is -35°C to +55°C (design temperatures). In order to ensure zero slippage, the following modifications of the standard drive unit design were implemented: • Iso-static mounts added at the interface between mechanism and instrument panel (Figure 8) • Laser treatment of slippage -critical surfaces performed (Figure 9) Iso-static mounts are widely used design features for optical instruments. They compensate thermal expansion mismatches in radial direction. This prevents slippage of interface bolts due to thermal loads. Figure 8. Iso-static mo unts of Scanner As mentioned in section Cleanliness and Contamination Control , a surface functionalization process that increases friction between adjacent parts has been developed and qualified at Airbus. This process is based on a laser treatment of the surfaces between two friction partners. Application of this process increases the friction coefficient significantly , which helps to prevent bolt slippage. For instance, the friction coefficient for laser -treated surfaces of titanium on titanium (Ti6Al4V) is >0.5, the one for titanium on aluminum 7075 is >0.8.

Document

AMS_2020.pdf

50 Figure 9. Laser -modified surfaces for increased friction coefficient Lifetime The lifetime requirement of the Scanner mechanism is 201 million revolutions; the requirement for the Derotator mechanism is 101 million revolutions . Both values include on-ground operation and margins as per ECSS -E-ST-33-01C. The design operating range of both mechanisms was specified as 10°C to 34°C. Due to local dissipation, bearing temperatures were calculated to reach more than 50°C during operation. Lifetime analysis per ISO 281 of the standard drive unit concept (hyperstatic bearing layout) with the bearing size required for the Derotator (pitch diameter 124 mm) yielded that this design was not feasible for the given temperature environment and lifetime requirement. Due to its design principle, the Derotator has to accommodate the conic optical path ( 105 mm to 89 mm in diameter) and the bearing size could not be reduced. Therefore, the Derotator bearing layout was changed to an isostatic one (one bearing pair in back - to-back configuration), the contact angle was increased to >30 deg, and the preload was decreased by approx imately 25%. This yielded a drastic improvement in calculated life: 60 billion revolutions for the isostatic design vs. 62 million revolutions for the original hyperstatic design. The Derotator performance requirements, which are less stringent than the Scanner ones, were still met. Long- term storage The specification required that the mechanisms be designed, manufactured and qualified to sustain at least 15 years of storage on satellite level plus 5 additional years on- ground lifetime. This requirement imposed careful selection of materials and processes with respect to corrosion, stress corrosion cracking, long- term stability, creep, etc. The following major design choices were made: • Nickel plating of motor rotor yoke and stator sheet metal stack • Selection of SmCo magnets (instead of NdFeB ) Nickel plating of steels is a widely used low-risk process. However, the plating of the stator sheet metal stacks was a major concern. Protrusions of t he bonding varnish between the metal sheets cannot be avoided during stator baking. The protrusions can be removed on the outer diameter of the stator by turning. On the inner diameter and in the stator slots, the bonding varnish cannot be removed. As nickel does not adhere to bonding varnish, a continuous nickel layer cannot be achieved on the inner diameter of the stator and in the stator slots (refer to Figure 10, left) . This was confirmed in a damp- heat test (7 days, 50°C, 95% relative humidity ): the stator sample showed significant levels of corrosion at these locations, whereas the outer diameter and the entire rotor yoke did not show corrosion at all. Corrosion at the inner diameter of the motor stator, which is only 1 mm away from the rotor magnets, could not be accepted. In order to increase the corrosion resistance of the stator while maintaining the robust nickel layer, it was decided to add a secondary moisture barrier on top of the nickel layer. Parylene, a polymer primarily used as moisture barrier on printed circuit boards , was selected. Paryl ene is less robust with respect to handling than nickel, but adheres well to many substrates, including the bonding varnish of the stators. Hence, the chosen approach came with the following advantages: • At least one continuous moisture barrier available at all locations on the stator, even at spots where the nickel did not adhere well Lasered titanium surface Blank titanium

Document

AMS_2020.pdf

51 • All handling surfaces (outer diameter) have two continuous layers, i.e. any potential damage to the outer Parylene layer (e.g. during AIT) could be accepted The entire stator was coated with Parylene after nickel plating and before wiring and potting. Figure 10. Bonding varnish protrusions visible after Nickel plating (left), c orrosion on nickel -plated stator sample after humidity test ( top right ) and sample with P arylene layer after humidity test (bottom right ) Two types of magnet materials were considered during the design of the motor: NdFeB and SmCo. By the time the MetOp- SG projects were started, an alert had been raised on a previous NdFeB type used on MetOp (first generation) . The nickel -plating of the motors had delaminated and the magnets had corroded. SmCo is resistive to corrosion but is known to be brittle and its remanence is approx imately 30% lower than the one of NdFeB , which causes higher power consumption of the mechanism. More corrosion- resistant NdFeB types had become available since the development of the MetOp satellites . However, our magnet supplier recommended not using nickel plating due to several issues observed with such platings on magnets in the past. Hence, the new type of NdFeB and a coating or plating would have had to be qualified for space and long- term storage. From an accommodation point of view, motor designs with SmCo magnets as well as with NdFeB magnets were feasible. The higher power consumption of a SmCo magnet motor was traded against technical and programmatic risk associated with a qualification program for the MWI, ICI, MWS and METimage projects , with two agencies, four instrument primes, and the satellite prime involved. After a scan profile had been agreed that permitted the use of SmCo magnets from a technical point of view , the decision was made to use SmCo magnets . It should be noted that this resulted in some handling issues during AIT; even our most experienced AIT personnel struggled with the brittleness of the magnets and – despite being extremely careful – damaged a couple of magnets.

Document

AMS_2020.pdf

52 Figure 11. SmCo magnets damaged during AIT Breadboard Testing The breadboard model of the BepiColombo Antenna De- Spin Mechanism was used to demonstrate the feasibility of the chosen control loop design and to obtain an understanding of the required motor and encoder performance at an early stage of the project. As its design comprised a single bearing pair in back - to-back layout, it was also used to assess the suitability of a back -to-back bearing layout for the Derotator. The Antenna De- Spin Mechanism breadboard was hardware- in-the-loop during controller testing with an xPC and Simulink models of the controller. The outcome of the test s was: • Chosen cascaded control loop (motor current, velocity, position) suitable • Motor stator to be skewed in order to reduce cogging torque disturbances • Encoder spikes at high speeds during acceleration phase to be filtered • High-frequency error of encoder has significant impact on measured PDE of mechanism Consequently, a skewed motor design was selected (also refer to section Performance Requirements ), a median filter was implemented into t he controller design, and the high- frequency error specification of the encoder was narrowed. Lubricant Testing and Re-lifing An outgassing test of Nye 2001a oil was performed in order to assess the suitability of this oil and the Maplub SH051- a grease for the project. As pointed out in section Cleanliness and Contamination, the oil outgassing behavior was significantly better than expected from literature. The grease Maplub SH051- a has been discontinued. This type of grease was used for most heritage scanning mechanisms at Airbus Friedrichshafen. To date, no successor showing similar performance has been identified. Maplub SH type b lifetimes are expected to be approx imately one order of magn itude lower than that of the type a greases [2]. In addition, torque peaks at low speeds and de- mixing at temperatures higher than 40°C were reported for the Maplub type b greases in an information note distributed by supplier MAP. The torque peaks were later confirmed by analysis and bench testing [7] . Based on these findings, the project decided to keep using the discontinued Maplub SH051- a grease. The batch of grease available for the MWI, ICI, MWS and METimage projects was manufactured in 1999 and its shelf life has formally expired. It was tested in a spiral orbit tribometer at the European Space Tribology Laboratory (ESTL) and compared against test data of a freshly manufactured sample of the same grease. The tribological performance (friction and lifetime) of the 20-year-old Maplub SH051- a grease is presented in Table 3; it was demonstrated to be no worse than that of a fresh ly manufactured grease batch. The variance of spiral orbit tribometer test data was higher for the re- lifed grease. Therefore, mixing of the re- lifed grease was recommended as a precaution , although i mprovements gained by physical re- mixing of the gr ease were anticipated to be marginal at best at bearing level. This is due to the fact that the amount of grease in a bearing is several orders of magnitude higher than the amount applied to the ball in a spiral orbit tribometer test. [5]

Document

AMS_2020.pdf

53 Table 3. Mean tribological data for re- lifed vs. fresh Maplub SH051- a grease [5] Grease Lifetime (orbits/µg) Initial friction coefficient Mid-life friction coefficient Re-lifed grease 11,050 0.07 0.093 Fresh grease 9,569 0.07 0.100 Life Tests Life testing of Scanner and Derotator was performed in order to demonstrate that the mechanisms met the specified lifetime requirements of 201 million revolutions for the Scanner and 101 million revolutions for the Derotator, respectively. Life test approach The bearings were considered the only life- limited (in terms of revolutions) items of the mechanisms . The following approach was chosen: • Build up full mechanism mo del (with optics dummy and industrial encoder) • Perform cumulated environmental testing • Perform accelerated life test in thermal -vacuum conditions Full models of Scanner and Derotator were built up. This was necessary as different bearing layouts and types had been selected for the two mechanisms (refer to sections Design Overview and Lifetime). Life - testing of standalone bearings was ruled out; such an approach would not have been flight -representative: pre-conditioning of the bearings by representative environmental testing would not have been possible and mounting conditions of the bearing in the mechanisms could not have been considered. Environmenta l testing of the life test models was performed in order to stress the bearings in a similar manner as the flight bearings. The worst -case environmental loads of a flight mechanism would be three times proto- flight testing (mechanism level, instrument level, satellite level) plus launch loads and in- orbit loads. Therefore, t he following environmental pre- conditioning was performed : • Vibration testing along three axes with qualification loads and 4x proto- flight duration (= two times qualification duration) • Shock testing along three axes (3 full -level shocks per axis) • Thermal -vacuum testing (1 cycle non- operating temperatures, 7 cycles operating temperatures) The predicted contact stresses during vibration were much higher than the ones during non-operational temperature cycling (2,000 MPa vs. 1, 500 MPa for the Scanner bearings ). Therefore, no additional benefit would have been obtained from doubling the thermal vac ( TV) test duration (and cost) . The life test was performed under consideration of the following constraints: • The life test has to be accelerated in order to finalize it in reasonable time. The duration of the life tests at nominal speed would have been 11 years. • The baseline temperature of the life test should not be overly conservati ve; the average temperature over one orbit , including hot case margins , was considered a reasonable choice. • Analytical predictions yield ed that the bearings operate in the mixed lubrication regime; a ny acceleration of the life test (= speed increase) requi red an increased temperature during the test in order to keep the oil film thickness constant • It was decided to limit the temperature to 60°C at the bearing; there is little experience with fluidlubricated bearings permanently operated beyond this temperature for such a long duration. • The Hertzian pressures in the life test environment had to be similar to the ones on orbit • The number of load changes (rotational acceleration/deceleration, stops) in the accelerated test had to be the same as in a non- accelerated test • The rotational acceleration should be the same for the non- accelerated test and the accelerated test in order to avoid overstressing the bearing cages

Document

AMS_2020.pdf

54 Accelerated testing Accelerated testing of bearings lubricated with fluid lubricants (oils and/or greases) is not trivial. It is widely accepted that the lubrication regime should be mainta ined when accelerating a test in order to keep the test representative. For lubrication containing grease instead of pure oil, no tribological validation can be performed beyond any tribological criticism. In addition, there is no practical means to measure the oil film thickness in an assembled mechanism [8, 9]. T he test data evaluation performed to select the METimage life test temperatures was based on the assumption that , when varying speed and temperature, same torque is an indication for same oil film thickness . Despite the mentioned uncertainties, keeping the oil film thickness constant was considered the most suitable and widely accepted approach when accelerating a test with fluid lubricants. In addition, this approach is considered conservative due to accelerated tribochemical degradation of lubricants at higher temperatures. The maximum possible acceleration factor was initially determined from bearing film thickness calculations. The final value was confirmed by friction measurements of the bearings of the life test models . The friction measurement was performed with the bearings integrated into the mechanism; the motor was not mounted as its resistive torque would have corrupted the bearing friction measurements . The friction measurement s were performed at different temperatures and operating speeds in a dry nitrogen atmosphere. The results of the Derotator bearings are shown in Figure 12. Subsequently, the f riction torque ((3) in Figure 12) at the nominal average speed of the mechanism (1) and at the nominal average temperature (2) over one orbit was determined. Assuming th at same friction torque meant same oil film thickness , the life test temperatures (4) for different speeds could directly be read from the plot. Respecting the 60°C limit this yielded an acceleration factor of 3 for both mechanisms . Figure 12. Bearing friction measurement results and life test acceleration factor for Derotator The MWS and MWI/ICI projects ran their life tests with acceleration factor s of 3.9 (MWS, discontinuous scan profile like METimage) and 5 (MWI/ ICI, constant speed), respectively , and at temperatures between 40°C and 50°C . The life tests of the IASI -NG mechanisms developed by Airbus Toulouse for CNES were performed with an acceleration factor of approx imately 11 at a temperature of 62 .5°C. This was possible due to the lower operating temperature of IASI -NG. Life test profil e The profile for the accelerated life test of the Scanner is shown in Figure 13. The major constraints were: • The average velocity was increased as described in the prev ious paragraphs. 0100200300400 0 10 20 30 40 50 60 70 80Torque [mNm] Speed [rpm]0°C 10°C 20°C 30°C 40°C 50°C 60°C (1)(3)(2) 32.5°C Nominal average speed 17.5 rpmAcceleration x 2x3x4(4) 50°C (4) 60°C (4) 65-70°C

Document

AMS_2020.pdf

55 • The same number of load direction changes as for a life test at nominal velocity had to be performed. This implied that that a dynamic profile with one acceleration phase and one deceleration phase per revolution had to be performed. • The acceleration/deceleration during the life test was kept the same as in the nominal scan profile in order not to over -stress the cages of the bearings • The ratio between constant velocity phases and acceleration/deceleration phases was kept constant (approx imately 69% constant speed, 15.5% acceleration, 15.5% deceleration) • Controller capability and power availability had to be given for the accelerated test case (the peak power consumption of the profile was optimized by inverting the acceleration and dec eleration phases) Figure 13. Scanner life test profile for acceleration factor 3 Life test results Both the Scanner and the Derotator life tests are still running. The environmental campaigns (vibration, shock, thermal -vacuum c ycling) have been finalized successfully. The life tests are continuously monitored automatically by EGSE and the set current is recorded. Once per day, the accelerated profile is interrupted and current measurements are performed at pre- defined constant speeds. Motor current is used as an indicator of bearing friction. The motor current during the first 55 million revolutions of the Scanner life test is shown in Figure 14. The low-speed motor current slightly decreased during the initial phase of the test. This is attributed to a smoothening of the bearing internal surfaces. The motor set current readings became noisier during summer (approx imately day 230 onwards) . This behavior is attributed to the fact that – besides bearing heating – no active temperature control of the TV chamber is performed. Temperature variations of the facility, whic h are more pronounced during summer when the air conditioning is working with high effort to keep the room temperature within the specified limits, directly impact the temperature in the TV chamber. The spikes at the beginning of the test and around day 90 are related to automatic switch- offs of the setup. The industrial encoder that is used for the life tests instead of the flight encoders provided erroneous position feedback to the control loop, which reacted by increasing the current beyond the limit defined for automatic switch- off. The initial root -cause analysis at the beginning of the test came to the conclusion that the analog- to-digital converter of the encoder, which had been placed inside the TV chamber, was not vacuum- compatible. After replacing the original analog- to-digital converter and mounting it outside the chamber, the test was re- started and the position spike issue seemed to be solved. When it returned around day 90, a more thorough assessment had to be performed. After initial assembly of the mechanism, the industrial encoder had been calibrated using its auto- calibration & reference search feature and its 0100200300400500600700 0 0.25 0.5 0.75 1 1.25 1.5 1.75 2Velocity [deg/s] Time [s]Nominal profileLife test profile

Document

AMS_2020.pdf

56 performance had been demonstrated in a test with an external reference encoder. The encoder stator had then been temporarily dismounted for the vibration test. After re- integration, encoder performance was again demonstrated using an external reference encoder. Subsequently, TV testing was started and the mechanism including encoder was operate d successfully under thermal -vacuum conditions. No position spikes were observed in any of the tests. The root -cause analysis finally led to the conclusion that the industrial encoder should have been re- calibrated using its auto- calibration & reference se arch feature after the second integration to the mechanism. AIT personnel had integrated the encoder stator twice in almost identical positions and orientations, i.e. within some microns; this is why the encoder worked at all after the vibration test. However, the missing proper calibration led to a spontaneous loss of reference from time to time, which caused the position spikes and, consequently, triggered the overcurrent protection. During the encoder performance test after vibration and during the TV test, this issue had – for either good or bad luck – simply not occurred. Based on this assessment, the TV chamber was opened and the industrial encoder was re- calibrated. The encoder spikes did not re- appear. Figure 14. Scanner motor current during first 55 million revolutions Summary and Conclusion The METimage Scanner and Derotator mechanism designs were established by modifying a standard drive unit design to project -specific requirement s. Breadboard- and component -level testing was performed in order to reduce the design risk at an early stage. Flight -representative life test models (except for the industrial encoder) were built up. Environmental testing of the life test models has been completed Set current at accelerated velocity and scan profile Measured current at constant velocity (CV) 6 deg/s 156 deg/s 480 deg/s 642 deg/s

Document

AMS_2020.pdf

57 successfully. The accelerated life tests of both models under thermal -vacuum conditions have been started and are ongoing. The following major lessons learned have been derived from the project so far : • Despite efforts to minimize the development and qualification activities, significant modifications of the standard drive unit design were required in order to meet the specific requirements of the Scanner and the Derotator, respectively. • The tribological performance (friction and lifetime) of 20- year-old Maplub SH051 -a grease was demonstrated to be no worse than that of a freshly manufactured grease batch. • The actual outgassing behavior of the used batch of Nye 2001a is far better than expected and, except for the CVCM, was in the same order of magnitude as the one of Fomblin Z25. • Design for long- term storage comes with very specific cha llenges concerning material choice. For programmatic reasons, SmCo magnets were selected instead of the technically more feasible NdFeB ones. A secondary moisture barrier had to be added on top of the traditional Nickel plating on the motor stator in order to guarantee corrosion resistance at all locations. • Representative accelerated life testing of fluid lubricated bearings is (still) a major challenge. The widely accepted approach of maintaining oil film thickness by increasing the test temperature i s considered conservative but comes with uncertainties (oil film thickness calculation, lubrication regime assessment , validity for greased bearings). Dynamic scan profiles require particular attention in order to ensure that the cages are stressed in a re presentative manner during the life test. • A surface functionalization process that increases friction between adjacent parts was implemented to prevent bolt slippage. The qualification of this process yielded excellent results: the friction coefficient for laser -treated surfaces of titanium on titanium (Ti6Al4V) is >0.5, the one for titanium on aluminum 7075 is >0.8. Acknowledgements The work described was performed on behalf of the German Space Administration with funds from the German Federal Ministry of Transport and Digital Infrastructure and co- funded by EUMETSAT under DLR Contract No. 50EW1521.

Document

AMS_2020.pdf

58 References 1. Wallner, O., Ergenzinger, K. and Schmülling, F. “METimage Instrument Development Status .” Proceedings of ICSO 2018 (2018) 2. Buttery, M. “An Evaluation of Liquid, Solid, and Grease Lubricants for Space Mechanisms Using a Spiral Orbit Tribometer .” Proceedings of the 40th Aerospace Mechanisms Symposium (2010) 3. ESR Technology Ltd. “ESA Space Tribology Handbook ”, 5th edition (2013) 4. Süss , M., Strobel , V., Zapata, A. “Black functionalised surfaces by using laser technologies .” European Congress on Advanced Materials and Processes (2017) 5. Buttery, M. “GEN -ESTL -TM-0215 Airbus MetOp- SG Grease Assessment.” Issue 2 (2016) 6. de Heij, P. “TEC-QTE-8772 Dynamic outgassing test and vapour pressure measurement of Nye 2001a vacuum lubricant oil – test report .” Issue 1 (2016) 7. Busquet, M., et al “Space grease tribological behavior for reformulation: numerical and experimental investigations .” Proceedings of ESMATS 2017 (2017) 8. Lewis, S. “ESTL/TM/199 Guidelines for Accelerated Testing of Liquid Lubricated Mechanisms .” (1997) 9. Lewis, S., et al “Accelerated Testing of Tribological Components - Uncertainties and Solutions .“ Proceedings of the 44th Aerospace Mechanisms Symposium (2018) 10. Orlandi, M. "QM 5633 VBQC - Kinetic Outgassing of Fomblin Z25 - Test Report." Issue 1, revision 1 (2009) 11. Süss, M., Strobel , V. “Laser Surface Modification – An E2E success story for cost efficient bonding preparation. ” Deutscher Luft - und Raumfahrtkongress (2018)

Document

AMS_2020.pdf

59 Challenges of the D evelopment of a C ompliant Focus Mechanism Submitted to the Harsh Martian Environment for the ExoMars Rover Mission A. Verhaeghe *, G. Perruchoud , P. Schwab* *, M. Gumy *, J. Rouvinet * and L. Kiener * Abstract The CLUPI ( CLose- UP Imager) instrument is a high -resolution camera mounted onto the Drill of the ExoMars 2020 rover mission carried out by the European Space Agency (ESA) and Roscosmos. The CLUPI development is under the responsibility of Thales Alenia Space Switzerl and whereas the Principal Investigator is Dr. Jean- Luc Josset from the Space Exploration Institute. For the development CLUPI instrument, the CSEM developed and delivered three models of a flexurebased Focus Mechanism. The CLUPI Focus Mechanism (CFM) des ign utilizes flexure guides to allow very accurate frictionless adjustment of the focal distance of the imager. Such design must also comply with very stringent requirement from the ExoMars mission, especially regarding the low Martian temperatures and the launch/landing load environment . This article presents the three main challenges encountered during the development of the mechanism and how these were addressed: resilience, performance and reliability. This article then draw s the lessons learnt from t his development including potential design improvements for a similar design and general rules to applicable to any development involving compliant mechanism. Introduction ExoMars 2020 is an ESA -Roscosmos led mission which will investigate the presence of past and present life on Mars. Equipped with a drill and a chemical analysis laboratory, it will be the first mission to sample and analyses the Martian underground down to two meters in depth. Figure 1: CLUPI Instrument mounted onto the ExoMars flight model * CSEM SA, Neuchâtel, Switzerland ; antoine.verhaeghe@csem.ch Proceedings of the 45th Aerospace Mechanisms Symposium, NASA Johnson Space Center, 2020

Document

AMS_2020.pdf

60 The Close- up Imager (CLUPI) on board the ExoMars 2020 mission is a high- resolution camera with the primary objective of providing close -up images of the collected samples before their chemical analysis. Detailed information on the CLUP I instrument was provided in a previous publication [ 1]. The instrument is developed by Thales Alenia Space Switzerland, under the PI -ship of Pr. Josset of the Space Exploration Institute. It is equipped with a focus mechanism which extends it capabilities and allows to acquire highresolution images of the surrounding geological environment. Figure 2: CLUPI Focus mechanism FM Figure 3: CLUPI Instrument FM The CLUPI Focus Mechanism is developed by CSEM based on flexible structure technology and the use of Off- The-Shelf components. This approach was chosen to meet the tight schedule constrains of the mission without a significant impact on the mechanism reliability. Mechanism Design The CLUPI Focus Mechanism illustrated in Figure 4, needs to accurately position a mobile set of lenses with respect to a fixed one. It maintains the alignment between the two optics while adjusting the distance between them to change the instrument’s focal length. The main requirements driving the design of the mechanism are: • mass under 220 g • operational stroke from - 4.3 mm to +4.3 mm • concentricity and co- alignment at reference position better than 50 µm and 0.1 de gree • stability of the concentricity and co- alignment during operation better than 20 µm and 0.1 degree • compatibility with the ExoMars environment Figure 4. CLUPI Focus mechanism design overview

Document

AMS_2020.pdf

61 Flexible Guiding Structure The guiding function of the mechanism is carried out using a flexible structure system composed of three flexible guide plates. Each plate is a deformable parallelogram, having four flexure blades acting as joints as illustrated by Figure 5. By combining these three plates in an equilateral prism as illustrated in Figure 6, quasi -isostatic linear guiding is obtained. This guiding allows linear movement along the Z axis (optical axis) and blocks all other degrees of freedom. The obtained guiding has a non- linear rigidity along the main axis which increases from 34 N/m at rest position to 95 N/m at operational end- of-stroke The displacement stroke is limited by flexible end- stops made of a stack of two blades. These end- stop contact points are at ±4.6 mm an d they dampen excessive displacement of the mobile stage up to ±5.0 mm. Figure 5. Flexible guiding plate Figure 6: Mechanism guiding structure COTS Actuator and S ensor In order to drive the mechanism, compact and contactless solutions have been implemented. To save costs and development time, off -the-shelf components were chosen: • a Voice -Coil Motor from Moticont • a Linear Variable Differential Transformer (LVDT) sensor f rom Singer Instrument & Control Ltd Figure 7: LVDT sensor & voice- coil actuator

Document

AMS_2020.pdf

62 Launch Locking System To prevent mechanism movement and to avoid excessive stresses in the guiding flexure blades during the launch and Mars landing phases of the mission, a launch locking system has been implemented. This system is illustrated in Figure 8. The launch lock consists of a non- explosive actuator (NEA) from Eaton actuating a grooved ring and locking stages. In the locked position, the locking stages k eep the mobile stage under pressure exerted by the ring. When the NEA is released, it pulls on the grooved ring which rotates , aligning recesses in the ring above each of the three locking stages . Theses recesses allow the locking stage to retract thus freeing the mobile stage. Figure 8: Launch locking system Encountered Challenges During the mechanism development a great variety of challenges were encountered Many issues were resolved with solutions compatible primarily with cost and schedule constraints. These challenges can be classed into three main domains: resilience, performance and reliability. Resilience The biggest challenge of the CLUPI CFM development was to design a mechanism resilient to the ExoMars harsh mechan ical environment. Such an environment can be summarized by the following points: • Random vibration up to 33 Grms • Sinusoidal vibration up to 25 g (between 30 Hz and 100 Hz) • Shock up to 1500 G’s When designing flexible structure, meeting strong vibration requirement s is always a trial due to the intentional removal of friction and thus very low damping. For the CLUPI development , this task was made even more complex due a significant change in the random vibration specification which occurred dur ing the detailed design phase (after PDR). Therefore, to meet the resilience required, multiple iterations of design were needed with associated analysis and tests. The complete process is described in Figure 9. The mechanical load specifications for random vibration were updated multiple times after PDR due to parallel progress in the overall instrument design. The specifications, their rationales and their random vibration levels are shown in the Table 1 .

Document

AMS_2020.pdf

63 Figure 9: Improving resilience throughout the complete CLUPI f ocus mechanism development Table 1: CFM random vibration specifications Specification Rationale Random Levels Early specification Very early specification – Only quasi -static - PDR From the last issue of instrument specification which followed a mission configuration change (Pasteur Payload IRD v8) X: 19.05 Grms Y: 16.90 Grms Z: 18.00 Grms 1st CFM specification Derived from PDR: the level at CFM interface are amplified by the instrument structure All: 36 .04 Grms 2nd CFM specification Following random test failures, a formal request was issued to reduce the instrument and CFM random levels. This led to the issue of the 2nd CFM specification. X: 17.56 Grms Y: 17.86 Grms Z: 27.49 Grms 3rd CFM specification Testing of the rover drill box on which the CLUPI is mounted showed higher random level than expected. Along with instrument design update this led to the 3rd CFM specification. X: 19.27 Grms Y: 18.03 Grms Z: 33.44 Grms While the specification evolved, the design was improved and tested many times to e nsure that the CFM would survive the mission mechanical environment . These design iterations were presented in detail in [ 2] and can be summarized in the following steps ill ustrated by Figure 10. 1. Initial design with titanium blades . This design was implemented in the first breadboard (BB). 2. Random vibration test of the breadboard (BB) led to rupture by fatigue. Investigation of the rupture is described in [ 3]. 3. Updates of the d esign which were implemented in the Updated Breadboard. This update includes the following changes: a. Change blade material to stainless steel Marval X12 which is known to have good fatigue resilience as stated in [ 4] b. Implementation of anti -buckling pins whi ch limit the off -plane displacements of the guiding plate, thus limiting the stresses in the blades. These anti -buckling pins are described in [ 3] 4. Second random vibration test led to rupture due to the guiding plate arm vibration mode 5. Updates to the design which were implemented in the Second Updated Breadboard. This update implement ed transverse blades on the pivot flexures of the guiding plates to shift the first guiding

Document

AMS_2020.pdf

64 plate mode from 830 Hz to above 2500 Hz. Random vibration analysis of this solution showed very conservative (low) stresses. 6. Shock test and analysis showed higher acceleration levels on the mobile stage, which may lead to excess ive buckling of the transverse blades. While the test was successful , the design needed to be improved. 7. The l ast update to the design consist ed of a refinement to the transverse blades: shortened and thickened to raise their bucking limit from 9 N to 30.8 N. This update was implemented on Engineering Qualification and Flight Models (EQM/FM). 8. Successful random vibration test on EQM. Figure 10: CFM guiding design update While m aking the CFM resilient to the launch and re- entry loads of the ExoMars mission was an arduous task, CSEM managed to meet the ever -increasing mechanical load requirements. Performance The CLUPI instrument is a close- up imaging system. On the Martian ground , it aims to replace the geologist ’s eye and magnifying glass. To do so it is equipped with an opto- electronical system capable of acquiring high resolution color images of targeted object s from 10 cm to infinity. Furthermore, the instrument is equipped with a processing unit able to perform z -stacking of multiple images of the same object. This algorithm is used to compensate the short depth of focus of the instrument: CLUPI can acquire a set of images with various working distances and reconstruct a completely sharp image (along with a 3D -map) of the imaged object [ 1]. In order to fulfil these instrument capabilities, the CLUPI Focus mechanism has quite stringent requirements with respect to positioning precision and guiding performance. Performance requirement s are lis ted in Table 2 along with their potential impact on the instrument performance:

Document

AMS_2020.pdf

65 Table 2: CFM performance requirement and their impact of the CLUPI performance CFM requirement Impact on CLUPI performances Positioning repeatability of 20 µm Poorer repeatability will impact the sharpness of the images acquired due slightly off -focus lens Positioning resolution of 5 µm Poorer resolution may impede the proper functioning of the z - stacking algorithm and of the au to-focusing algorithm. Mobile stage parallelism at rest position better than 0.1° Lack of parallelism at rest may impact the image quality Mobile stage concentricity at rest position better than 50 µm Lack of concentricity at rest could impact the image quality Mobile stage parallelism stability along stroke better than 0.1° Lack of parallelism stability may impact the image quality and the ability to perform the z -stacking due to erroneous coregist ration of successive images Mobile stage lateral shifts along stroke (concentricity stability) better than 20 µm Lack of concentricity stability could impact the image quality and the capability to perform the z -stacking due to erroneous co-registration of successive images To achieve such performance the following approaches were implemented: 1. Completely frictionless design along with a LVDT sens or and Voice- coil actuation: with such a configuration, the mechanism positioning capabilities are mainly driven by the control loop performance and the electronics used to drive the actuator and sensor. With a frictionless mechanism and considering that CLUPI operates only when the rover is immobile , the control loop can be very simple and achiev e excellent performance, especially after a proper sensor calibration. 2. Fine tolerance chain between the mobile and the fixed opt ics: the parallelism and concentricity at rest are guaranteed by a very stringent tolerance chain throughout the flexurized guiding structure. 3. Axisymmetrical construction: the ax isymmetry of the guiding structure improves the parallelism and concentricit y at rest , since (undesired) lateral motions tend to counteract each other, instead resulting in very small (benign) rotation of the optics about the optical axis . The axisymmetry of the guiding also guaranties parallelism and concentricity stability throu ghout the movement as the errors induced by the movement in a perfectly axisymmetric construction also remain axisymmetric , i.e. normal and centered on the guiding axis. Such approaches were in general well implemented with the following exceptions: • The CFM design is not a perfectly axisymmetric construction. To fulfill this approach, the sensor and actuator should have been either centered onto the optical axis or repeated around the optical axis. Early in the design, the trade- off selected a single set of off -centered sensor and actuator for mass and cost saving. This resulted in an off -centered actuation force which generates a significant error of concentricity and parallelism stability. Such err or was identified early as a potential nonconformity, based first on FEM analyses and then on breadboard testing. • The LVDT sensor was selected for its compactness. Because of its miniature size, the radial gap between the mobile rod and the fixed coil body is very narrow (about 0.04 mm). With such a tiny gap, it is not possible to guaranty the absence of contact between the two parts during the movement especially with the errors introduced by the lack of axisymmetry . Therefore, even with a perfect initi al alignment of the two parts of the LVDT, during the movement, the mobile and fixed part make contact, generating tiny friction forces which can climb up to 50mN. This residual friction force impacted the positioning performance, which thankfully remained within the requirements. In the end the compliance of the EQM and FM to performance requirement s is provided in Table 3.

Document

AMS_2020.pdf

66 Table 3: EQM and FM compliance to CFM performance requirement s Requirement (measurement) EQM FM Positioning repeatability of +/-20µm ~ At the limit  Positioning resolution of 5 µm ~ At the limit  Mobile stage parallelism at rest position better than 0.1°  < 0.034°  < 0.043° Mobile stage concentricity at rest position better than 50 µm  < 61 µm Exceeded by ~ 20%  < 30 µm Mobile stage parallelism stability along stroke better than 0.1°  < 0.068°  < 0.12° Exceeded by ~20% Mobile stage lateral shifts along stroke (concentricity stability) better than 20 µm  < 68 µm Exceeded by x 4  < 114 µm Exceeded by x 6 It is worth noting that after delivery of the instrument to ESA, the optical performance was assessed in the instrument calibration campaign to be acceptable. Reliability Reliability of systems for space is an essential requirement: the system must work reliably throughout its complete lifetime as no repair can occur after launch. In the case of the CLUPI mechanism the reliability of the guiding structure was quite straight forward once on Mars and safely released from its launch lock, since normal operation does not generate any stresses significant for fatigue consideration, the mechanism is safe from any risk of failure. Considerable efforts were made to ensure that the actuator and sensors were compatible to 300 sols in the Martian environment [3]. Additionally, all routing, soldering and contact crimping was made according to the relevant ECSS standard. Thus , the major of risk of failure of the mechanism is related to the launch lock system: if the launch lock fails to release, the CLUPI instrument capability would be considerably reduced as the instrument would only be able to image target at a fixed defined distance (about 26 cm). In order to reduce this risk, the following mitigations wer e implemented: • The locked position of the optics corresponds to a working distance at which some critical operation of the instrument can still be c arried out : while locked, the instrument is capable to image the drilled core sample collected by the rover drill and placed into the sample drawer [ 1]. • The actuation system of the launch lock system is a space- qualified non- explosive actuator (NEA) manufactured by Eaton. • Extensive testing of the launch locking system was carried out at breadboard level in order to ensure that the coating of the friction surfaces and the actuation forces were selected with sufficient margin: not enough margin and the launch lock may not release; too much margin and the NEA released energy propagates to the guiding struct ure and may damage it. This early testing included many force measurements to assess the friction coefficient best and worst -case values. Furthermore, high-speed imagery of the unlocking process allowed for a better understanding of the propagation of the released energy onto the guiding structure [3]. Despite these efforts , the launch locking mechanism remained a very sensitive component with its reliability and efficiency excessively dependent on the resetting procedure. Two important observations were made during the qualification of the EQM, one of which led to a partial failure during the thermal -vacuum qualification test.

Document

AMS_2020.pdf

67 The first observation was made during the EQM qualification test to random vibration loads. During this test, it was observed that the mechanism sine sweep responses before and after a random cycle did not match as they should according to the applicable ECSS standard. Such phenomenon usually indicates a mechanical failure inside the tested item. After careful inspection of the mechanism , no damage was found. Further testing and investigation revealed that , under the random vibration environment, components of the launch lock system settled into a different position, thus impacting the sine sweep response of the mechanism. In the end, during the qualification test, the launch lock resetting procedure had to be revised. In this update the final pressure applied onto the locking ring is along the unlocking direction. Wh en doing so, the ring moves back about 0.05 mm and is then in more stable configuration. While this observation did not impact the reliability of the mechanism, it illustrates how sensitive the mechanism is to the resetting procedure. The second and more critical observation was made during the thermal -vacuum qualification and acceptance test of the EQM and FM. In this test, after the first non- operational cycl e, the units were placed at the low operational temperature in order to perform a functionality check . At the beginning of the EQM functional ity check, a representative current was applied through the NEA leads to actuate it and release the launch lock. When doing so it was observed that the NEA fuse wire burned (electrical discontinuity). However, t he launch lock did not release immediately. Following the failure, it was decided to warm up the thermal vacuum chamber to prepare it for opening and investigation. During the warming up, about 1h after the NEA firing, the EQM launch lock released as illus trated in Figure 11. Figure 11: Temperatures & LVDT reading during firing & release of the EQM launch lock The opening of the thermal vacuum chamber was then interrupted and the chamber placed again at low operational temperature in order to test the FM model similarly. In the case of the FM model the unlocking occurred 74 s after firing, and without any temperature change. Such an observation is even more complex to interpret since the observed delay s are both very different and far from the expected behavior (<2 s). Following these observations, the thermal vacuum qualification test continued without any further issue and the unlocking failure was in vestigated. In order to investigate the issue, many unlocking tests at low temperature were conducted on the EQM. The first tests performed were done with a mechanical release system instead of the fuse spool. Such test s all had unlocking times under 1s. Then representative test s with fuse spools were done in order to reproduce the issue. In those tests, similar behavior was observed with very dissimilar unlocking times. The unlocking times for all the troubleshooting tests are described by Table 4.

Document

AMS_2020.pdf

68 The strong difference of behavior between the test s with a mechanical release system and the tests done with a fuse spool indicate that the issue was related to the spool and not to the locking ring and stage mechanism. This behavior was then discussed wit h Eaton, the NEA supplier, which identified t wo potential root causes: • The actuation current applied is under the required minimum. Actuation is still possible however the fuse wire tends to melt to slowly which has sometimes resulted into spool getting stuck closed . When the right current is applied, the fuse wire burns very fast (flashing behavior) and most of it sublimates . This issue has also been seen on other development s [5]. However, such a root cause does not explain how after some time the spool released . In similar cases , the spool remain ed stuck indefinitely . • The resetting procedure used could wear the surface of the spool . Such a wear could create barbs of plastic which prevent opening. After some time under the pressure of the plunger, the plastic barbs creep and the spool finally releases . Table 4: Unlocking test duration and temperatures Test description Unlocking duration (s) Test temperature Thermal vacuum chamber test EQM 3592 [-59°C; 21°C] Thermal vacuum chamber test FM 74 [-56°C; -49°C] Mechanical unlock 1 0.78 -56°C Mechanical unlock 2 0.49 -56°C Mechanical unlock 3 0.55 -56°C Mechanical unlock 4 0.20 -57°C Mechanical unlock 5 0.28 -71°C Spool unlock 1 3117 -57°C Spool unlock 2 0.42 -56°C Spool unlock 3 25 -57°C Spool unlock 4 4457 -57°C Spool unlock 5 with new reset procedure 0.42 -57°C Spool unlock 6 with new reset procedure 3.48 -57°C Spool unlock 7 with new reset procedure 250 -56°C Spool unlock 8 with new reset procedure 3.20 -57°C Spool unlock 9 with new reset procedure 106 -57°C In order to solve this issue, the approach implemented considering the very tight schedule and the quite advance status of the development was to update the launch lock resetting procedure. In this update, tooling was developed to avoid any wear of the spo ol during its installation into the NEA. With this solution implemented further testing was performed. It can be seen in Table 4 that the durations obtained with the updated procedure are significantly better than the ones before. While the implemented solution did not fully solve the issue, considering the mission schedule and the issue associated risk, the deviation from expected behavior was accepted. The main argument backing the acceptance is that despite the sometimes very long delay, the launch lock did not ever get stuck: i t always released. Lessons Learnt for F uture Development Looking retrospectively at the CLUPI focus mechanism development, it seems that many issues could have been averted and that better solutions could have been implemented to solve encountered issues . This observation is to be put into perspective with the mission major constrains whi ch were: mass, cost and

Document

AMS_2020.pdf

69 schedule. Furthermore, these solutions may appear straightforward, now that the development is complete, but they were certainly not as obvious at the beginning of the project. That is precisely why it is important to look back to t he now completed development and identify important lessons that would help future development s. These lessons learned presented here are two- fold. There are specific lessons about what could be done from the beginning if a very similar mechanism were to be developed. Then there are more generic lessons about aspects which had an important impact on the CLUPI Focus Mechanism but are also applicable to a lot of other mechanism developments . Improvement for a F uture Focus Mechanism Solutions were found to answer the issues and address the challenges of the CLUPI Focus Mechanism . However, these solutions had to be compatible with the very tight mission schedule. For the development of a future focus mechanism such solutions could be implemented at earlier s tage in the development and thus be better integrated in the design. In some case, alternative more elegant solution s could be implemented. All these solutions were presented in detail in [ 6] and are summarized hereafter: 1. Improvement of mechanical interfaces: It would allow a reduc tion of the amplification and transmission of the mechanical load to the focus mechanism. Figure 12: Improvement of bracket design to improve mechanical interface 2. Transverse Blades Integrated to the Guiding Plate Structure: A more symmetrical design of the guiding plate would allow a better optimization of the flexure parameter s to increase resilience without impacting the guiding movement stiffness. Figure 13: Transverse blades integrated to the guiding plate structure

Document

AMS_2020.pdf

70 3. Coalignment of the Actuation: Centering the actuation would enable full symmetry on the design and drastically improve the guiding alignment stability performance. Figure 14: Actuation and s ensing centered around the optical axis 4. Implementation of a M odel-Based Control -Law: With a model -based control law, the residual LVDT friction could be modelled as a known perturbation (measured in calibration or on- line) to improve the positioning accur acy 5. Open Locking Ring to Reduce Friction During Unlocking: Such a solution maintains high friction loads in the locked position to dissipate vibration energy during launch but minimize s the friction during unlocking. Figure 15: Launch Locking System with Open Locking Ring Lessons Learnt for F uture Mechanism Development Apart from potential design improvement, the CLUPI focus mechanism gave important insight about general development of mechanism. These lessons are numbered hereafter along with a brief description about how it became an issue within the CFM development. 1. Resilience of C ompliant Mechanism is Always Critical: Discuss the mechanical environment requirements with the mission and your customer . Mechanical requirements must be backed up by strong rational es and all must be done to minimize the loading applied to the mechanism.

Document

AMS_2020.pdf

71 In the CLUPI focus mechanism development , the mechanical environment specification severity increased drastically at a late stage of the project. This led to additional (and extensive) design iterations at CSEM in an already very tight schedule. Efforts were done to make sure that the mechanism would be compatible with the interfacing structure. Further improvement s could have been implemented on the interfacing structure to reduce its impact on the amplification of random levels. 2. In Compliant Mechanisms , Avoid Poorly Constrained Intermediate Stages: internal modes of the guiding structure are often the root -cause to compl iant mechanism failure. While overall structural modes are dampened by the assembled structure and the mobile stage modes are dampened by a launch lock , intermediate stage modes cannot be easily locked an d are not dampened. In the CFM development , the intermediate stage mode (originally at 830 Hz) was an important source of stresses in the mechanism. It is not damped by any phenomenon but the internal material dampin g. Thus, the intermediate stage arm behaves as a tuning fork and stress es the blades above their limit. The solution was to better constrain the arm with a transverse blade. This solved the problem but could have been implemented at an earlier stage of the project thus saving the cost and time of developing further breadboards. 3. Most Compliant Mechanisms Need Launch Locks : This is especially true when the guiding rigidity is low, which is needed achieve long strokes. In these cases, the development of launch locki ng must be started early in the development of the compliant mechanism. This allows for a better integration of the locking system in the mechanism and may even allow for a locking of the intermediate stages. At the very beginning of the CFM development n o launch lock was designed. Not that it was not foreseen as necessary by the engineering team, but because the cost of the launch lock development was not included in the first feasibility study. Furthermore, the feasibility study report did not mention explicitly the need for a launch lock . Due to this missing information, the customer was misled to believe no launch lock would be required. Thus, during mass budget negotiations , no allowance was made for such a system. It was only shortly before the instrument delta PDR that the need of a launch lock was identified and flagged as critical. T he late-coming launch lock design was then highly constrained by an already very tight schedule and mass budget. 4. Test as You Fly, as Early as P ossible: While mission representative test s are ofte n associated with extensive costs, they are the only way to anticipate issues with multiple root -causes. The failure of the CFM launch lock during qualification is a good example to justify the need for “test as you fly” method: m any tests were performed to make sure that the unlocking would behave as expected: • Launch lock friction and force characterizati on measurements  successful • NEA te sted alone, with representative actuation current and low temperature  successful • Launch lock mechanism tested at ambient with mechanical release representative  successful However, the first time the launch locking system was tested integrally in -flight representative condition was during the thermal vacuum qualification test , during which the system failed to unlock within an acceptable duration.

Document

AMS_2020.pdf

72 Conclusion Despite a very rocky path filled with potholes and obstacles , CSEM managed to lead the development of the CLUPI Focus mechanism up to delivery and integration within the instrument. The mechanism qualification campaign showed that it can survive all the mission environment requirements. Its mechanism performance was measured to be acceptable and this was confirmed during the instrument calibration campaign performed by ESA and the Space exploration Institute. The mechanism reliability was deemed acceptable even if it is greatly impacted by the sensitivity of the launch locking mechanism to the resetting procedure. Overcoming these issues lead to a successful development from which a lot of lessons were learnt. This was only made possible by an efficient collaboration between CSEM and its dedicated partners: Space Exploration Institute, Thal es Alenia Space Switzerland , ESA, EATON, Singer Instrument Ltd, Petitpierre SA, Arcofil SA and Fisba Optics AG. Presently, the mechanism was successfully integrated onto the CLUPI flight model which was mounted onto the ExoMars rover drill. Hopefully, with a launch currently planned for July 2020, the first images of the Martian ground taken by CLUPI and its focus mechanism will be acquired in spring 2021. References 1. Josset, J. -L., Westall, F., Hofmann, et al. (2017). The Close- Up Imager Onboard the ESA ExoMars Rover: Objectives, Description, Operations, and Science Validation Activities. Astrobiology , 17, 595– 611. https://doi.org/10.1089/ast.2016.1546 2. Verhaeghe, A., Perruchoud, G., Schwab, P., Gumy, M. (2019 ). Challenges of the development of a compliant Focus Mechanism submitted to the harsh Martian environment for the ExoMars Rover mission. Presented at Final Presentation Day, ESTEC, Noordwijk. 3. Verhaeghe, A., Perruchoud, G., Schwab, P., Gumy, M. (2017). Development Challenges of a Focus Mechanism Design for ExoMars Mission Submitted to the Harsh Martian Environment and Utilizing Off-the-Shelf Equipment. In 17th E uropean Space Mechanisms and Tribology Symposium. Presented at the ESMATS, Hartfield. http://esmats.eu/esmatspapers/pastpapers/pdfs/2017/verhaeghe.pdf 4. Spanoudakis, P., Schwab, P., Kiener, L., Saudan, H., Perruchoud, G. (2015). Development Challenges of Utilizing a Corner Cube Mechanism Design with Successful IASI Flight Heritage for the Infrared Sounder (IRS) on MTG; Recurrent Mechanical Design Not Correlated to Recurrent Development. In 16th European Space Mechanisms and Tribology Symposium. Presented at the ESMATS, Bilbao. http://esmats.eu/esmatspapers/pastpapers/pdfs/2015/spanoudakis.pdf 5. Yoon, J., Betancourt, R. (2019). Battery Cell Bypass Switches – A Case Study in Test Like You Fly Merits. In 18th European Space Mechanisms and Tribology Symposium. Presented at the ESMATS, Munich. http://esmats.eu/esmatspapers/pastpapers/pdfs/2019/yoon.pdf 6. Verhaeghe, A., Perruchoud, G., Schwab, P., Gumy, M., Rouvinet, J. (2019). Lessons Learnt during the Development of a compliant Focus Mechanism for the EXOMARS Rover Mission. In 18th European Space Mechanisms and Tribology Symposium . Presented at the ESMATS, Munich. http://esmats.eu/esmatspapers/pastpapers/pdfs/2019/verhaeghe.pdf

Document

AMS_2020.pdf

73 Lessons Learned from Q ualification of HDRM for U ltralight LP -PWI Boom for ESA JUICE M ission. Maciej Borys*, Ewelina Ryszawa* , Łukasz Wiśni ewski*, Maciej Ossowski* and Jerzy Grygorczuk* Abstract The paper presents the complications appearing during the qualification campaign of the Langmuir Probe – Plasma Wave Instrument (LP -PWI). The article focuses on one subsystem of LP -PWI: the Hold Down and Release Mechanism (HDRM). After the qualification vibration test , the HDRM was supposed to open to release the LP -PWI boom. However , the mechanism was blocked. The analysis present ed reveals several root causes of this failure. The following root causes were identified: an optimistic functional analysis, in correct integration processes , and neglected finishing of the parts . The second part of the paper shows the improvements implemented in the HDRM and the result of these changes. Finally, the lessons learned from the qualification process of the LP -PWI and HDRM failure are presented. Introduction The LP -PWI is one of the instruments within the Radio & Plasma Wave Investigat ion experiment on board ESA’s JUICE (Jupiter Icy Moon Explorer) mission. The experiment consists of four identical LP -PWIs mounted on the edges of JUICE Spacecraft (S/C) . The main objective of the instrument is to provide crucial information about the bulk plasma surrounding Jupiter’s icy moons. The instrument is built in cooperation with the Swedish Institute of Space Physics in Uppsala under contract with ESA Prodex . The article begins with a general description of the LP -PWI architecture and an introduction to the main environmental and technical requirements . The state- of-the-art section presents an overview of the boom with a short description of the main subsystems. The paragraph on system over view gives a detailed description of HDRM . The next chapters of the article focus on the qualification campaign and problems that appeared after the vibration test. The core of the article is the analysis of the HDRM ’s failure to open. The root causes of this failure and further solutions are described. As a summary , the paper presents all HDRM design and process improvements that helped overcome al l the problems encountered with the opening of the HDRM . Final conclusions and lessons learned can be found in the last chapter . State of the A rt The HDRM is a part of the bigger instrument - the LP-PWI (Figure 1). The LP -PWI is a two -section boom ended with Langmuir Probe (LP) [1]. The LP is a spherical sensor made of titanium and covered with TiA lN (Titanium Aluminum Nitride). The LP on board JUICE mission is 100 mm in diameter , which is two times larger than those on board Rosseta [2] and Cassini missions [ 3]. The two sections of the boom are made of CFRP (Carbon Fibers Reinforced Plastic) tubes ended with titanium interfaces and covered with Single Layer Insulation . The interfaces allowed to link the tubes with two hinges : Base and Central . Both are very similar . They are equipped with clock springs , which drive the boom, and a latching mechanism. The Central Hinge opens to 180 deg, while the Base Hinge opens to 135 deg. In a deployed configuration all LP -PWIs are positioned at an angle of 45 deg to the S/C longer edge. * ASTRONIKA Sp. z o.o., Warsaw, Poland Proceedings of the 45th Aerospace Mechanisms Symposium, NASA Johnson Space Center, 2020

Document

AMS_2020.pdf

74 Figure 1. LP-PWI overview (CAD model deployed configuration – left; CAD model stowed configuration – right; as built model stowed configuration - bottom) The boom has two HDRMs: Base HDRM and HDRM#2, which hold the CFRP tubes in a stowed configuration during the launch. The Base HDRM is placed close to the Langmuir probe while HDRM#2 is placed close to the Central Hinge . The position of HDRM#2 was optimized to maximize the first natural frequency of the boom. Both mechanisms are identical in terms of functionality ; the difference between them lies in the structure: HDRM#2 is standalone while Base HDRM shares the structure with the Base Hinge. In a stowed configuration the boom’s length is under 1.6 m; it deploys to the length of 3 m. The LP -PWI’s angular position must be kept within ±0.5 deg cone with a tip in Base H inge. The length of the boom must be kept within the tolerance of ±5 mm. During the mission the LP -PWI will be exposed to highly demanding environmental conditions . There are four identical LP -PWI booms located on three different corners of the S/C. As consequence, the level of Sun illumination is different for each boom . The S/C trajectory comprises a close Sun approach with high Sun illumination, as well as long time spent in the shadow during Jovian Tour with Sun eclipse. This leads to a wide range of temperatures on the LP, ranging from -220°C up to 200°C . However , the main driver behind LP-PWI’s architecture was a limited mass budget – in order to comply with requirements, each LP -PWI needed to have a mass under 1.3 kg. The low mass of the unit did not allow for using hold and release mechanisms available on the market due to the mass requirement . The HDRM also needed to be part of the boom structure. In the LP-PWI, the HDRMs are also the main support of the boom as well as the interface with the S/C. Each HDRM contains mechanical and electrical interfaces with the S/C. The design concept of the HDRM was based on the Hold Down and Release Mechanism from MUPUS instrument on Rosseta mission [ 4]. A similar design was later also used in the DRAGON -8U Nanosatellite Orbital Deployer [ 5]. However , this specific application ( for LP -PWI) forced a major tailoring of the mechanism.

Document

AMS_2020.pdf

75 HDRM System overview The core of the HDRM is a Preload Jaw which provides the preload to the CFRP tubes and keeps them in the stowed configuration (Figure 2). The Prel oad Jaw consist s of a stainless -steel V-shape spring with two Tension Clamp s attached to each arm of the spring. The Tension Clamps press the Tube Interface , which is attached to the upper 2nd section of the boom. The upper section lay s down on the separators , which are a part of the lower 1st section of the boom. The pressing on the T ube Interface is transfer red through separators to the lower section and it holds both of the CFRP tubes . The Tube I nterface is equipped with overlay made of Vespel SP1. The T ension Clamps are made of nitride d titanium , which is essential for further investigation in this article. The materials are a good and stable friction pair with friction coefficient 0.21-0.24 in vacuum, in room and low temperature ( -80°C) [6]. Vespel SP1 also has high strength, which does not degrade significantly in the higher temperatures (tensile strength: 86.2 MPa at 25° C; 41.4 MPa at 260°C). These features make Vespel SP1 the best choice for this application. Figure 2. HDRM design In a closed position of the HDRM , the Preload Jaw is bent and supported by two levers. The levers are connected to each other above the CFRP tubes with a Vectran string. The string is wrapped around the resis tors placed in the lever. The resistors play a role of thermal knives that cut the Vect ran string . In a single HDRM, there are two resistors (one on each lever): primary and redundant. The geometry of the lever combined with the inclination angle of the T ension Clamp (α) lowers the loads seen by V ectran string with a ratio 28.5. The opening of the HDRM ( Figure 3) is initiated by a weakening of the Vectran string by heat from deployment resistors . The levers are motorized by t heir own torsion springs and they are released after cutting of the Vectran string . The constrain resulting from the closed levers disappear s, and the V-shape spring comes back to its open position. The T ension Clamp s stop pressing the T ube Interface and the boom’s sections are ready to be opened.

Document

AMS_2020.pdf

76 Figure 3. Closed HDRM #2 - left; Open HDRM #2 - right HDRM preloading The project entailed the delivery of 4 identical LP -PWIs, hence all of the eight HDRMs shall be similarly preloaded. The preload cannot be determined only by its hardware. The manufacturing tolerances did not allow for reaching the required repeatability in the preloading. Naturally, it cannot rely only on the hardware dimensioning and needs to be adjusted during integration process ( Figure 4). The nominal preload for one HDRM was selected to a range of 160-170 N. A higher loading of HDRM could lead to the cumulation of too high a stress in the structure. The preload is applied to the Preload Jaw by the T ension Tube. During integration, the Tension Tube is loaded and tightens the V -shape spring when HDRM stays in a closed position. When all parts are preloaded, the tube is blocked by 6 blocking screws tightened to the HDRM walls. On the one hand, this solution releases the mechanism from manufacturing inaccuracy , but on the other the final preload depends on the correctness of the integration process . Figure 4. Application of the preload in HDRM Qualification Campaign – Failure to Open after the Vibration Test Qualification c ampaign overview The full qualification model (QM- 1) of the LP -PWI was subjected to several functional and environmental tests during the qualification test campaign. The functional tests were focused on the deployment reliability. The boom was designed to be deployed in micro -gravity. On ground , the LP-PWI can only be opened in a horizontal pos ition. For the deployment test , the boom requires a 3x4 m flat table and ball support s attached to the CFRP tubes . During the deployment , the ball supports move on the table and offload the boom from gravity force. The dynamic opening of the boom excluded any other offloading methods. Due to the necessity of opening on a large ground support , the deployment following the environmental test w as replaced with a limited release. The limited release was restricted to the opening of the HDRMs in vertical position without opening the hinges. Blocking screws

Document

AMS_2020.pdf

77 The environmental test included the thermal -vacuum (TVAC) cycling , shock and vibration tests. The TVAC test was performed on a reduced qualification model (QM-2), which was limited to one hinge, one HDRM, and shorter CFRP tubes with a dummy mass . The unit pass ed the thermal cycling in the range of - 180°C and 100°C , and was successfully deployed in cold conditions (-50°C). After the test, the QM- 2 parts were used in a full model QM-1. The QM-1 was subjected to vibration and shock test s. The v ibration test included resonance search, sine and random vibrations , and was performed on qualification level. The boom was subjected to vibr ations in all 3 axes (Figure 5). The model passed the vibrations , but the limited release after the test was not fully successful . Figure 5. Configuration of LP -PWI on the shaker It is worth mentioning that the qualification test campaign was preceded by a test campaign of the engineering model in Phase B . The design of the HDRM had not been modified significantly as compared to the previous phase. The Breadboard (Engineering) Model (BBM) pass ed over 50 deployments and random vibrations at acceptance level (-3 dB from qualification level) . The HDRM was considered well tested on BBM level and the failure on QM came very unexpected. HDRM opening failure After the vibration test on QM- 1 (in all 3 axes) followed by limited deployment test , the Base HDRM did not open. The Clamp got stuck on the Tube Interface after the opening of the levers ( Figure 6) - the Clamp s were block ed. In order to release them , several finger taps onto the V-shape s pring were needed . At the same time HDRM#2 opened without any problems. Figure 6. Failure to open of HDRM Even before the issue with the actual deployment, t he first sign indicating that something changed in the Base HDRM was the modal response recorded by an accelerometer placed on the levers (Figure 7). The resonance searches in Z axis before and after random vibration show large differences between these stages . These response changes could not be linked with the behavior of the remaining subsystems of the boom. The f irst mode was (at 500 Hz) increased by over 200 Hz. The mode around 1700 Hz moved close to 2000 Hz. Main mode recorded at 1200 Hz totally changed its shape. Generally, f or the higher frequencies (above 800 Hz) the modal responses start to highly deviate from each other.

Document

AMS_2020.pdf

78 The increase and the major change of modal response s turned out to be the first symptom that something was blocked in the HDRM . After the unsuccessful HDRM opening it became clear that t he blocked Clamp transfers the loads from the CFRP tubes to the levers differently . This caused the change in the modal response. Figure 7. Resonance response of the lever in Base HDRM in Z -axis before and after vibration of QM- 1 Root Cause of the F ailure After the failure of the Base HDRM the most critical task was to find the root cause of the problem. At the beginning , blocking of the Tension Clamp seemed very unlikely . The design of the HDRM was well known and had been used in two others flight instruments . However, a deeper analysis revealed 4 potential root causes of this failure. All of them affect each other , therefore it is impossible to clearly identify which problem was the most critical for the system. Preload change during vibration As mentioned above, the QM-1 was not the first model that was teste d. The qualification phase was proceeded by the breadboard (engineering model) testing phase. The HDRMs in BBM and in QM- 1 were practically identical in terms of the Preload Jaw design. The B BM passed similar tests during the breadboard test campaign. However, the BBM was subjected to random vibrations with a level reduced to -3 dB (from qualification level ) due to the limitations of the available shaker. Hence, the BBM had never seen full qualification loads before the QM vibe test. The preload in LP -PWI was between 160 N and 170 N. This value was established by scaling the preload form the MUPUS instrument. The vibration load s in the Preload Jaw were estimated at 1120 N. It was clear that the preload was not high enough to avoid gapping during v ibrations. Unfortunately, a higher preload could not be applied due to the strength limitation of the HDRM structure. Strengthening the structure would affect the mass budget and therefore could not be applied . There was also a strong conviction that if the BBM survived the vibrations , then the QM-1 should also pass the test. However, t he qualification level of random vibrations appeared to be critical. The gapping between the Tension Clamp and the Tube Interface caused micro movement of the Clamps (Figure 8), which led to a higher tension of the V -shape spring and leveraged the preload.

Document

AMS_2020.pdf

79 Figure 8. Position change of Tension Clamp on Vespel overlay Stacking in the softer material The design of the HDRM assumed the protruding of the Tube I nterface from the Tension Clamp . The edges of the harder Clamp touch the relatively soft interface surface. In case of gapp ing during the vibration, the Tension Clamp moved and start ed to press one edge into the Vespel SP1 surface (Figure 8). This led to an uncontrolled increase of contact pressure between the parts . As a consequence , the Clamp was blocked on the Vespel SP1 overlay . Mechanical analysis – low margin. The mechanical analysis of the HDRM’s performance is relatively complex . The behavior of the Preload Jaw is very hard to model. The V -shape spring is bent and preloaded at the same time. At first , the motorization margin was not calculated for the Prel oad Jaw , due to the Clamp rotation theory described above: the preload force pulls the Tension Clamp at a certain distance from the contact area with the Tube Interface. This creates the torque that rotates the Clamp. The rotated Clamp loses contact with the T ube Interface, hence the friction between them can be neglected. The mechanism was considered safe from being blocked. The previous experience from BBM (over 50 successful deployment s) seemed to confirm this theory . The movement of the Clamp s during the opening is very dynamic and it was very difficult to observe the behavior of the mechanism in real time. Finally, the QM -1 vibration tests showed that the Preload Jaw can be blocked on the T ube Interface. This situation was later repeated manually in a lab. The failure of the HDRM opening sparked a reconsideration of the mathematical model and the behavior of the V-shape springs. The approac h was changed - if a more accurate model cannot be built then the most conservative one should be used. The new mathematical model excludes the rotation of the Clamp during the opening. Now the Clamp slides from the interface and maintains the contact with the T ube Interface. The Preload Jaw is fully symmetrical, so the model is presented only for one Clamp. Figure 9 presents the forces acting on the Clamp after the release of the levers. All forces ’ values are presented for one Clamp .

Document

AMS_2020.pdf

80 Figure 9. Force relation in Tension Clamp The Clamp slides down from the interface driven by an opening force. The opening force is a summary of the force from V -shape spring’s bending (FB) and the sliding force (S). The sliding force is a result of the initial preload (P) and its change introduced by residual force after vibrations (Fv). The contact force (Q) is also a result of the residual vibration force and the preload force. The contact force is additionally reduced by a part of the bending force (F B’). The friction force (F F) simply result s from the contact force lowered by a part of the bending force (F B’’) and multipli ed by the friction coefficient (μ) between the Clamp and Tube Interface. The success criteria for the HDRM opening was presented in Eq.1. The opening force must be higher than the friction force. The Eq. 2 presents the dependences between the forces and the angle of the Clamp (α). (𝑄𝑄−𝐹𝐹𝐵𝐵′′)𝜇𝜇< 𝐹𝐹𝐵𝐵′+𝑆𝑆 Equation 1 ((𝑃𝑃+𝐹𝐹𝑉𝑉)𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐−𝐹𝐹𝐵𝐵𝑐𝑐𝑠𝑠𝑠𝑠𝑐𝑐)𝜇𝜇< 𝐹𝐹𝐵𝐵𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 +(𝑃𝑃+Fv)𝑐𝑐𝑠𝑠𝑠𝑠𝑐𝑐 Equation 2 The QM design was confronted with the new mathematical model . The calculation showed that the QM had a barely positive motorization. The nominal working point of the HDRM was located below but still very close to the motorization line (Figure 10). The change of the preload combined with a slight increase of the friction coefficient could block the HDRM, presented by the movement of the working point into the red area on the plot . The conclusion was that t he margin of safety for the HDRM working point was too low to be accept ed. The friction coefficient can increase when the contact surface is worn out or when the parts were not properly finished. There is no possibility to check the friction coefficient on the unit itself. The contact force can be locally increased by a wedging of the Clamp into the interface. Both effects lead to the increase of the friction force and blocking of the mechanism.

Document

AMS_2020.pdf

81 Figure 10. Motorization of Preload Jaw – QM-1 (α=10 deg; FB=6N; μ=0.22 ) Preload application Another cause of the preload change following the vibration test could be a wrong method of its application. The l ower preload cause s a higher gaping than it was expected. Generally, the preload is adjusted during the integration process by the loading and movement of the tension tube (Figure 4). In BBM the preload was applied in a vertically positioned boom. The Tension Tube was loaded with an adjusted mass. In QM- 1 the method was modified. The preload was applied through the force meter with a stiff hook attached to the tension tube. The HDRMs walls were constrained and could not move due to the horizontal positioning of the boom. In this configuration, the wedging of the tube between the walls was highly possible (Figure 11). The force meter indicated the correct value of the preload, but this force was not necessary fully transferred to the Preload Jaw . It could have been consumed by the block ing of Tension Tube between the HDRM walls. This way of the preload adjusting did not allow for the full control of the preload in HDR M. As a consequence, the final pr eload could have been randomly lower, which led to a higher gapping during the vibration test. The final position of the tube as well as its tilting were not measured. Additionally, t he tilted tube tighten ed the Preload Jaw asymmetrically . The asymmetrical position of the V -shape spring causes higher pressure of one side of the Clamps , which increases the tendency of pressing the edge into the soft Vespel SP1 overlay . The modification between the BBM and the QM-1, which was intended to be an improvement , introduced an additional uncertainty to t he mechanism.

Document

AMS_2020.pdf

82 Figure 11. Tilted and blocked Tension Tube. Design and process improvements The blocked Preload Jaw in the HDRM was a critical and unexpected failure. The problem was on the critical path of the project . The design needed to be improv ed and the test campaign had to be continued. The design changes could not be too deep in order not to affect the design of the rest of the boom: the LP-PWI was in the midst of a qualification test campaign, and any major changes could challenge the validity of the previous QM tests. Additionally, most of the flight parts had already been manufactured, hence the quantity of the modified parts had to remain as few as possible. The i mprovements focused on: - Control ling the friction coefficient between the Tension Clamp and the T ube Interface. - Lower ing the risk of remaining the residual vibration loads in the Preload Jaw. - Improv ing the relation of motorization margin to the preload force. - Better control of the preload adjustment . The LP -PWI qualification model QM- 1 with the following modifications is called QM -bis in the next chapters. Design changes As mentioned above, the modification had to be conservative and affect as few parts as possible. First , the Tension Clamps were redesigned ( Figure 12). They were widened to protrude from the Tube Interface and the edges were manually rounded. More attention was put on the surface finish. The inclination angle was increased to 15 deg. Applyi ng a higher angle was not possible as it resulted in too high loads on the levers and could have led to an inadvert ent opening during vibration. The V -shape spring was also widened. The previous spring’s model was optimized to withstand the preload and keep a low mass. The modified spring increased the bending force at the expense of the mass budget .

Document

AMS_2020.pdf

83 Figure 12. Tension Clamp design changes The change of the Tension Clamp led to the modification of the T ube Interface. The titanium part of the interface was already attached to the CFRP tube and replacing them was not possible. Only the Vespel SP1 overlay could have been replaced. Because the Clamp angle had been changed , the overlay needed to be adjust ed as well. The simplest solution was to round the contact surface of the overlay and create a linear contact with the Tension Clamp (Figure 13). This solution also had other benefits. In the previous design, the contact area between the Clamp and the T ube Interface was random. The manufacturing tolerances of the parts did not allow for perfect surface- to-surface contact. As a consequence, the Clamp pressed the overlay in random point s under not necessarily a correct angle. Rounding of the overlay improve d the control over the contact area and inclination angle . The disadvantage of this solution included higher Hertzian stresses of the linear contact , but it remained within accepted range. As a plastic material , Vespel SP1 can be easily deformed under pressure without risk of breaking. In this case, after exceeding the Hertzian stresses , the overlay simply adjust s its shape to the Tension Clamp and creates the sufficient surface to surface contact area. Figure 13. Vespel SP1 overlays modification Change of the preload application process The process of preload adjust ment in the HDRM was also improved. The load was applied with a mass attached to the tension tube, similarly as in the BBM . During the process the boom was put in a vertical position, hence the HDRM’s walls did not have to be screwed to the integration plate. T hanks to this , the HDRM walls were not rigidly fixed and a clearance remained between them and the Tension Tube. The tube was able to move freely without the risk of tilting and blocking. The position of the T ension Tube was observed from the beginning of preloading to the blocking by the screws. A lot of effort was put to ensure

Document

AMS_2020.pdf

84 that the final position of the tension tube is parallel. After constraining the tube in HD RM, its exact position was measured with the accuracy of 0.01 mm. The measurements were taken on both sides of the T ension Spring. All deviations of readings higher than 0.05 mm meant it was neces sary to re apply the preload. This inspection assured that the tube was not tilted by more than 0.15 deg. Test Campaign Continuation All design changes were focused on improv ing the motori zation margin of the P reload Jaw. Figure 14 presents the comparison between the motori zation of the Preload Jaw in QM- 1 and in QM-bis. The increase of the Tension Clamp ’s inclination angle and the modification of the V-shape spring made the HDRM much more independent from the preload and friction coefficient changes. The wi dening of the Tension Clamp and rounding of its edges excluded the possibility of the titanium edge sticking into the Vespel SP1. A wellfinished surface of the overlays and the C lamps assured as low friction coefficient as possible. The margin of safety for the HDRM working point was significantly increased . Figure 14. Motorization of Preload Jaw – QM-bis (α=15 deg; FB=8 N; μ=0.22 ) The modified qualification model (QM- bis) was vibrated once again in accordance with the same test specification. During the vibration s, it was apparent that the applied changes improved the mechanism. The comparison of modal responses (performed before and after the nominal random vibration test) from the accelerometer placed on the lever was more accurate than it was in the QM-1 (Figure 15). Previously , a major increase of the modes ’ frequencies and a great difference of the modes at frequencies above 800 Hz were observed . In the QM-bis the first mode decreases its frequency after the test. It can be explained by the initial setting of the mechanism. For higher frequencies the resonance characteristics remained similar before and after random vibration. Any major shift was not observed, and the main mode is almost identical.

Document

AMS_2020.pdf

85 Figure 15. Resonance response of the lever in Base HDRM before vibration and after in Z -axis for QM- bis The QM-bis passed the vibration test – the modal responses were very similar. For most of the parts it was the second vibration with the qualification level which additionally proved the endurance of the boom. The limited release just after the vibrations was successful as well - both HDRMs were opened. The qualification campaign c ould resume. All other tests , including full deployments and shocks, were completed successfully. The LP-PWI was fully deployed over 50 times during the qualification test campaign (QM-1: 8 deployments; QM- bis: 49 deployments ). In the end, the LP -PWI reached TRL7. Conclusions and Lessons Learned The experience from the qualification campaign of LP -PWI is essential and can be used for other project s. Several lessons learned were extracted: Great heritage and successful breadboarding does not guarantee reliability of the mechanism. Several issues were neglected at the beginning of the project. The first motorization model of the Preload Jaw was too optimistic . The preload value was simply scaled from another instrument without confronting it with the real loads acting on the unit. These steps were taken due to the lack of a good analysis in the early phase of the project, but also due to strong faith in the reliability of the mechanism. The good heritage of HDRM and successful breadboard campai gn discouraged implementation of any improvements. In simulations and mathematical models, the most conservative approach should be considered as the baseline . The analysis of the behavior of the Preload Jaw was neglected in the project . The i nability to build a detailed mathematical model was replaced with an optimistic approach. The mechanism was considered unlikely to block. If a sufficiently detailed mathematical model is too complicated to achieve , the most conservative approach should always be used. In addition, in the early design process , the mechanism shall demonstrate clearly defined constraints and parameters that are verifiable and controllable during the manufacturing, assembly and testing process. When the project constraints appear, the design compromises are embedded and sufficient level of risk awareness needs to be tracked and not forgot ten. The harder material should always protrude from a softer one. One of the basic mistakes in the design was the decision to make the Tube holders wider than the Tension Clamp . In consequence, the edges of the Clamp could wedge into the tube holders’ overlays. The overlays made of Vespel SP1 are much softer than the nitride titanium Clamp . The wedged edge of the Clamp generates much higher Hertzian stresses and deforms the overlay. In the connections where one material is much softer than the other one, a contact of the hardest edges shall be avoided.

Document

AMS_2020.pdf

86 Parts inspection The inspection and rejection of the par ts with questionable quality at an early stage is critical. The project was under high time pressure. The risk of using several parts with surface or shape defects was underestimated. Not enough attention was put on the surface finishing and rounding of the parts’ edges. The accepted parts were not in a bad condition and they passed the standard acceptance criteria, however, the standard inspection did not consider their specific application. In some cases, the parts were not properly finished. The elements that are part of the interface or a component of the friction pair should always have excellent surface finish and the inspection shall assume their future function. In addition, if the quality of the critical parts r aises any question, the risk of accept ing them may lead to costs much higher than those resulting from the delays or production of new ones. Acknowledgments The work described in this article was performed under a c ontract with the European Space Agency’s Prodex No. 4000119065/16/NL/JK. The Primary Investigator of the project is IRF Uppsala led byJan- Erik Wahlund (PI), Jan Bergman (Instrument manager), Victoria Cripps and Sara Gomis (Instrument Product Assurance Manager s). The project is supervised by ESA’s expert Ronan Le Letty. Astronika’s engineers who also participated in the research and development activity described herein are: Tomasz Kuciński (project managers at initial stage of the conntract ), Kamil Bochra, Mateus z Duda, Michał Bogoński , Mateusz Grzyb, Karol Jarocki, Łukasz Bereś, Filip Szymański, Paweł Miara, Henryk Gut . The authors would like to thank the above mentioned for their support. References 1. M. Borys, L. Wiśniewski, J. Grygorczuk, et al. ‘’LP-PWI deployable boom for JUICE mission – innovative features and breadboard model development ’’ 17th European Space Mechanisms and Tribology Symposium (ESMATS), Hatfield, UK, September 20- 22, 2017 2. A.Sjögren, ‘’Modelling of Rosetta Langmuir Probe Measurements’’ UPTEC F09 063, Uppsala, Sweden, Octo ber, 2009 3. K.S. Jacobsen a,, J. -E. Wahlund b , A. Pedersen, ‘’Cassini Langmuir probe measurements in the inner magnetosphere of Saturn’’, Planetary and Space Science 57 (2009) 48– 52 4. Grygorczuk, J., Banaszkiewicz, M., Seweryn, K., Spohn, “MUPUS Insertion dev ice for the Rosetta mission” Journal of Telecommunications and Information Technology (1/2007), pp50- 53. 5. M Dobrowolski, J Grygorczuk, B Kędziora, M Tokarz, M Borys ‘’DRAGON- 8U Nanosatellite Orbital Deployer ’’; Proceedings of the 42nd Aerospace Mechanisms S ymposium; NASA/CP -2014- 217519; 487-496 6. J. Grygorczuk, M. Dobrowolski, L. Wisniewski, “Advanced Mechanisms and Tribological Tests of the Hammering Sampling Device CHOMIK ” 14th European Space Mechanisms and Tribology Symposium (ESMATS), Constance, Germany, September 28- 31, 2011

Document

AMS_2020.pdf